This page intentionally left blank

Pearson
Boston Columbus Indianapolis New York San Francisco Upper Saddle River Amsterdam
Cape Town Dubai London Madrid Milan Munich Paris Montreal Toronto Delhi
Mexico City Sao Paulo Sydney Hong Kong Seoul Singapore Taipei Tokyo
Library of Congress Control Number: 2011923252
Publisher: Julie Levin Alexander
Publisher’s Assistant: Regina Bruno
Executive Editor: Marlene McHugh Pratt
Senior Acquisitions Editor: Stephen Smith
Development Editor: Pamela A. Powell
Associate Editor: Monica Moosang
Editorial Assistant: Samantha Sheehan
Director of Marketing: Dave Gesell
Marketing Manager: Brian Hoehl
Marketing Specialist: Michael Sirinides
Marketing Assistant: Crystal Gonzalez
Managing Production Editor: Patrick Walsh
Production Liaison: Julie Boddorf
Production Editor: Lisa S. Garboski, bookworks
Manufacturing Manager: Ilene Sanford
Manufacturing Buyer: Alan Fischer
Art Director: Christopher Weigand
Cover Designer: Karen Salzbach
Cover Photo: Jamie Novak, St. Paul Fire Department
Composition: Aptara®, Inc.
Printing and Binding: R.R. Donnelley/Willard
Cover Printer: Lehigh-Phoenix Color/Hagerstown

Credits and acknowledgments borrowed from other sources and reproduced, with permission, in this textbook
appear on appropriate pages within the text. Unless otherwise stated, all photos have been provided by the authors.

Note: A portion of this book was prepared by David J. Icove, a former employee of the Tennessee Valley
Authority (TVA).The views expressed in this book are the views of Dr. Icove and his coauthor. They are not
necessarily the views of TVA or the United States Government. Reference in this book to any product,
process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute
its endorsement or recommendation by TVA or the United States Government.

Copyright © 2012, 2007 by Pearson Education, Inc., publishing as Pearson, 1 Lake Street, Upper Saddle River,
New Jersey 07458. All rights reserved. Manufactured in the United States of America. This publication is protected
by Copyright, and permission should be obtained from the publisher prior to any prohibited reproduction, storage
in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. To obtain permission(s) to use material from this work, please submit a written request to
Pearson Education, Inc., Permissions Department, 1 Lake Street, Upper Saddle River, New Jersey 07458.

Many of the designations by manufacturers and sellers to distinguish their products are claimed as trademarks.
Where those designations appear in this book and the publisher was aware of a trademark claim, the designations
have been printed in initial caps or all caps.

Library of Congress Cataloging-in-Publication Data

DeHaan, John D. (John David)
Kirk’s fire investigation / John D. DeHaan, David J. Icove.––7th ed.
p. cm.
Includes bibliographical references and index.
ISBN-10: 0-13-508263-3
ISBN-13: 978-0-13-508263-8
1. Fire investigation. I. Kirk, Paul Leland, 1902–1970. II. Icove, David J., 1949-III. Title. IV. Title:
Fire investigation.

TH9180.K5 2012
363.37'65––dc22 2011005408
®
Pearson is a registered trademark of Pearson PLC.

10 9 8 7 6 5 4 3 2 1

ISBN 10: 0-13-508263-3

ISBN 13: 978-0-13-508263-8
DEDICATION

For Mick Gardiner, in recognition of his unlimited dedication to improving fire investigation
across the UK and the rest of Europe through better training.

To the memory of Monty McGill—friend, colleague, teacher, and investigator of great

integrity, skill, and courage.
This page intentionally left blank
CONTENTS

Preface xxiii
Acknowledgments xxvii
About the Authors xxxi
NFPA 1033 Correlation Matrix xxxii
Fire and Emergency Services Higher Education (FESHE) Grid xxxiv

Chapter 1 Introduction 1
Fire Investigation 2
The Fire Problem 4
Fire Statistics in the United States 4
Fire Statistics in the United Kingdom 5
Role of the Fire Investigator in Accurately Reporting the Causes of Fires 5
The Detection of Incendiary Fires 6
Reporting Arson as a Crime 6
Problems Associated with Estimating Incendiary Fires 8
Scientifically Based Fire Investigation 10
Comprehensive Methodologies for Fire Investigation 10
The Scientific Approach to Fire Investigation 11
Applying the Scientific Method 11
Steps in the Scientific Method 12
Levels of Confidence 15
Legal Opinions Regarding Science in Investigation 16
Chapter Review 17
Review Questions 17
References 17

Chapter 2 The Elementary Chemistry of Combustion 19

Elements, Atoms, and Compounds 20
The Oxidation Reaction 21
Carbon Compounds 22
Other Elements 22
Organic Compounds 23
Hydrocarbons 23
Petroleum Products 25
Carbohydrates 26
Pyrolysis of Organics 28
Conclusions about Organic Compounds 28
State of the Fuel 28
Significance of State of Fuel 29
vii
 Difficulty in Classifying Some Hydrocarbons 29
Solids 30
Liquids 30
Chapter Review 31
Summary 31
Review Questions 31
References 31

Chapter 3 Fundamentals of Fire Behavior and Building

Construction 32
Basic Combustion 33
Flaming Fire 34
Structure of Flames 36
Smoldering Fire 38
Explosive Combustion 41
Heat 42
Heat and the Rate of Reaction 42
Heat and Temperature 43
Heat Release Rate 43
Heat Transfer and Heat Flux 45
Direct Flame Impingement 50
Flame Plume 51
Sequence of a Room Fire 54
Beginning or Incipient Stage 54
Growth, a Free-Burning Stage 56
Fire Growth to Flashover 56
Post-Flashover Stage 61
Decay Stage 62
Flow of Hot Gases 63
Effects of Environmental Conditions 65
Temperature 65
Humidity 66
Wetness of Fuel (Fuel Moisture Content) 67
Wind 68
Oxygen Content 68
Elements of Building Construction 69
Structural Shell 69
Fire Resistance Ratings 72
Internal Structure 73
Chapter Review 82
Summary 82
Review Questions 82
References 82
viii Contents
Chapter 4 Combustion Properties of Liquid
and Gaseous Fuels 85

Types of Fuel 86
Gases 86
Liquids and Their Vapors 86
Solids 86
Physical Properties of Fuels 87
Vapor Pressure 87
Flammability (Explosive) Limits 88
Flash Point 91
Flame Point/Fire Point 94
Ignition Temperature 94
Ignition Energy 97
Boiling Points 98
Vapor Density 99
Heat of Combustion 104
Hydrocarbon Fuels 104
Natural Gas 104
Liquefied Petroleum Gas 105
Petroleum 106
Gasoline 106
Kerosene and Other Distillates 107
Diesel Fuel 107
Lubricating Oils 107
Specialty Petroleum Products 108
Nonhydrocarbon Liquid Fuels 108
Alcohols, Solvents, and Similar Nonhydrocarbons 108
Alternative Fuels or Biofuels 108
Combustion of Liquid Fuels 109
Pyrolysis and Decomposition of Liquids 112
Fuel Gas Sources 112
Gas Lines 112
Natural Gas 113
LP Gas 113
Chapter Review 121
Summary 121
Review Questions 121
References 121

Chapter 5 Combustion Properties of

Solid Fuels 123

Pyrolysis 125
Crown Fires and Fireballs 126
Nonpyrolyzing Fuels 127
Contents ix
 Combustion Properties of Wood 127
Components of Wood 127
Ignition and Combustion of Wood 128
“Low Temperature” Ignition of Wood 131
Charcoal and Coke 137
Wood Products 138
Paper 140
Plastics 143
General Characteristics 143
Behavior of Plastics 145
Special Considerations for Fire Investigators 149
Paint 156
Metals 158
Magnesium 159
Aluminum 159
Coal 160
Dust Explosions 160
Combustion Products of Solid Fuels 161
Flame Color 161
Smoke Production 162
Chapter Review 164
Summary 164
Review Questions 164
References 164

Chapter 6 Sources of Ignition 167

Introduction to Ignition Sources 168
Primary Ignition Sources 168
Matches 169
Lighters 171
Torches 171
Candles 173
Secondary Ignition Sources 173
Sparks/Arcs 174
Hot Objects/Hot Surfaces 174
Friction 176
Radiant Heat 177
Chemical Reaction 178
The Role of Services and Appliances as Ignition Sources 179
Gas Appliances as Ignition Sources 179
Portable Electric Appliances 186
Kerosene Heaters 187
Stoves and Heaters 187
x Contents
 Oil Storage 187
Electricity 188
The Role of Hot and Burning Fragments in Igniting Fires 188
Windblown Sparks 189
Fireplaces and Chimneys 190
Long-Term Heating (“Low-Temperature” Ignition) 194
Trash Burners, Incinerators, and Bonfires 197
Hot Metals 197
Mechanical Sparks 198
Firearms Residues 200
Smoking as a Fire Origin 201
Cigarettes 201
Bedding and Furnishings 203
Cigarettes and Flammable Liquids and Gases 205
Pipes and Cigars 206
Plantings 206
Spontaneous Combustion (Self-Heating) 207
Characteristics of Self-Heating 207
Self-Heating Oils 209
Self-Heating of Vegetation 213
Other Materials Subject to Self-Heating 215
Implications for the Fire Investigator 215
Other Sources of Ignition 215
Lightning 216
Implications for the Fire Investigator 216
Ignition by Electric Lighting 219
Ignition from Discarded Batteries 223
Animal Interaction with Sources of Ignition 223
Assessing Ignition Sources at the Fire Scene: The Ignition Matrix 225
Chapter Review 227
Summary 227
Review Questions 227
References 227

Chapter 7 Structure Fires and Their Investigation 231

Beginning the Investigation 233
During the Fire 233
Immediately After the Fire Is Extinguished 234
During the Clearing of the Scene 234
After Cleanup 235
Investigative Information during Suppression 236
Responsibility of the Firefighters 236
Minimizing Post-Fire Damage 237
Overhaul 237
Salvage and Security Concerns 238
Contents xi
 Documenting the Fire Scene 238
Photography and Photographic Equipment 238
Digital Images 239
Enlargements and Film 240
Photography for the Fire Investigator 240
Sketching 243
Measurement and Scanning Systems 245
Notes 245
Reconstructing the Pre-Fire Conditions 249
General Principles of Fire Behavior 249
Fire Patterns 249
Tracing the Course of the Fire 250
Implications for the Fire Investigator 252
Examination of a Structure Fire Scene 252
General Considerations 252
Interviews with Firefighters 253
Interviews with Witnesses 254
Search Patterns and Practices 255
Fire Behavior Indicators 257
Burn Patterns 258
Heat Level (Heat Horizon) 265
Smoke Level (Smoke Horizon) 268
Low Burns and Penetration 268
Floor Burns and Penetrations 276
Char Depth 279
Appearance of Char Surface 281
Surface Effects 282
Displacement of Walls and Floors 282
Spalling 283
Ghost Marks 286
Calcination of Gypsum Board 288
Annealed Furniture Springs 289
Glass 290
Melting Points of Materials 295
Clean Burn 296
Myths and Misconceptions about Indicators 296
Evidence (Documentary or Witnesses) of an Abnormally Fast Rate of Fire
Spread or Collapse 296
Evidence of Abnormally High Temperatures (Melted Metals, etc.) 296
Spalling of Concrete 296
Crazing of Glass 297
Irregular Damage to Floors and Floor Coverings 297
Black, Heavy, Oily Soot on Windows/Black Dense Smoke 297
Annealing of Steel Springs and Steel Structural Materials 298
Floor-to-Ceiling Heat Damage 298
Deep Char 298
Progression to Flashover 298
Alligatoring (Shiny or White) 298
Arson Evidence 298
Trailers 298
Containers 298
xii Contents
 Contents Inventory 299
Ignitable Liquids 300
Detectors—Electronic and Canine 302
General Considerations 304
Protected Areas 304
Utilities 310
Elimination of Electrical Ignition Sources 310
Arc Mapping 311
Appliance Condition 312
Trash 312
Detection Systems Mapping 313
Interior Fires from Exterior Sources 314
Roof and Attic Fires 314
Timelines 315
Collection and Preservation of Evidence 316
Debris Suspected of Containing Volatiles 316
Other Solid Evidence 319
Liquids 319
Testing of Hands 319
Testing of Clothing 320
Chain of Evidence 320
Analysis and Hypothesis Testing 321
Chapter Review 323
Summary 323
Review Questions 323
References 324

Chapter 8 Wildland Fires and Their

Investigation 327

Fire Spread 329

Fuels 330
Fire Spread 331
Moisture Content 331
Intensity of Wildland Fire 331
Fire Behavior 332
Effect of Wind 333
Effect of Tall Fuels 333
Other Effects 333
Determination of Origin 333
Investigation Methodology 334
First Evaluation 334
Other Sources of Information 335
The Scene Search 335
Burn Indicators 338
Documentation 344
Contents xiii
 Sources of Ignition 345
Power Lines 346
Lightning 346
Burning or Hot Fragments 349
Campfires 349
Cigarettes 349
Incendiary Fires 349
Modeling 353
Collection and Preservation of Physical Evidence 354
Cigarettes, Matchbooks, and Other Fragile Evidence 354
Shoe and Tire Impressions 355
Charred Matches 355
Debris Suspected of Containing Volatiles 355
Containers 355
Weather Data 355
Chapter Review 357
Summary 357
Review Questions 357
References 357

Chapter 9 Automobile, Motor Vehicle,

and Marine Fires 359

Automobiles and Motor Vehicles 360

Fuel Tanks 360
Fuel Tank Connections 361
Fuel Pumps, Fuel Lines, and Carburetors 361
Fuel Injection Systems 361
Vehicle Fuels 362
Other Combustible Liquids 363
Engine Fuel System Fires 364
Electrical Systems 367
Miscellaneous Causes 369
Considerations for Fire Investigation 371
Combustible Materials 372
Miscellaneous Ignition Mechanisms 372
Vehicle Arson 373
Considerations for Fire Investigation 373
Protocol for Vehicle Examination 373
Safety 373
Photography and Sketches 376
Importance of Scene Preservation 377
Exterior Examination 378
Evidence of Stripping 378
Considerations for Fire Investigation 379
Motorhomes and Other Recreational Vehicles 387
Characteristics of Motorhomes 387
Fire Risk 387
xiv Contents
 Propane Tanks 388
Considerations for Fire Investigation 388
Mobile Homes (Manufactured Housing) 390
Construction and Materials 390
Considerations for Fire Investigation 391
Heavy Equipment 392
Boats and Ships 393
Ships 394
Tankers 395
Ship Construction and Firefighting Techniques 396
Motives for Vehicle and Marine Arson 397
Chapter Review 399
Summary 399
Review Questions 399
References 399

Chapter 10 Electrical Causes of Fires 401

Basic Electricity 403
Static Electricity 403
Current Electricity 404
Direct and Alternating Current (DC and AC) 407
Electrical Units 407
Electrical Calculations 408
Series and Parallel Circuits 410
Electrical Systems 412
Conductors and Insulators 412
Current-Carrying Capability (Ampacity) 413
Protection—Overcurrent and Short Circuit 416
Fuses 416
Circuit Breakers 418
Thermal Protectors 419
Surge Protection Devices 419
Overcurrent Devices and Fire Investigation 420
Ground Fault Interrupters 420
Arc-Fault Circuit Interrupters 421
Open Neutral 422
Electrical Service Distribution 422
Service Entrance 423
Receptacles 424
Ignition by Electrical Means 425
Conduction Heating 426
Overheating by Excessive Current 426
Overheating by Poor Connection 427
Insulation Breakdown (Degradation)—Carbon Tracking 429
Arcs and Sparks 435
Contents xv
 Aluminum Wiring 437
Electric Transformers and Motors 438
Fixed Heaters 440
Appliances 440
Electric Lighting 443
Electric Blankets 448
Extension Cords 448
Heat Tapes and Heat Cable 448
Batteries 449
Investigation of Electricity-Related Fires 449
Post-Fire Indicators 450
Mapping of Arc Faults 454
Arcing Through Char 457
Laboratory Examination 458
Chapter Review 464
Summary 464
Review Questions 464
References 465

Chapter 11 Clothing and Fabric Fires 467

Types of Fabric 468
Natural Fibers 469
Petroleum-Based Synthetic Fibers 469
Non-Petroleum-Based Synthetic Fibers 471
Fire Hazards 471
Influence of Weave and Fiber 471
Clothing Ignition 472
Regulation of Flammable Fabrics 472
Regulation of Flammable Fabrics 472
Furniture Testing 476

Flammability Testing 478

Flammability Tests for Federal Regulations 478
Flammability Tests for California Regulations 481
General Observations 482
Considerations for Fire Investigators 483
Chapter Review 485
Summary 485
Review Questions 485
References 485

Chapter 12 Explosions and Explosive Combustion 487

Chemical Explosions 488
Key Terms and Concepts 489
Diffuse-Phase Explosions 490
xvi Contents
 Gases 490
Vapors and Vapor Density 498
Deflagrations 501
Ignition 503
Condensed-Phase Explosions 505
Chemical and Physical Properties 506
Types and Characteristics of Explosives 507
Propellants or Low Explosives 507
High Explosives 509
High Explosive Categories 511
Components 512
High-Order/Low-Order Explosions 513
Mechanical Explosions 514
Acid, Gas, or Bottle Bombs 515
BLEVEs 517
Electrical Explosions 518
Investigation of Explosions 519
The Scene Search 520
Speed and Force of Reaction 523
Scene Evaluation and Hypothesis Formation 530
Evidence Recovery 530
Laboratory Analysis 532
Incident Analysis 533
Chapter Review 536
Summary 536
Review Questions 536
References 536

Chapter 13 Chemical Fires and Hazardous Materials 539

Gases 540
Hydrocarbons 541
Other Gases 542
Liquids 543
Solvents 543
Miscellaneous Liquids 546
Solids 546
Incendiary Mixtures 547
Oxidizing Salts 548
Reactive Metals 549
Clandestine Laboratories 549
Clandestine Drug Laboratories 550
Marijuana Cultivation 554
Clandestine Explosives Laboratories 555
Contents xvii
 Warnings 555
NFPA 704 System 555
Federal Hazardous Materials Transportation System 556
Chapter Review 558
Summary 558
Review Questions 558
References 558

Chapter 14 Laboratory Services 559

Availability of Laboratory Services 560
Forensic Laboratories 560
Fire Testing Laboratories 561
Expert Qualifications 561
Identification of Volatile Accelerants 562
Gas Chromatography 562
Gas Chromatography/Mass Spectrometry (GC/MS) 566
Sample Handling and Isolation of Volatile Residues 569
Identification of Volatile Residues 573
Interpretation of GC Results 582
Chemical Incendiaries 585
Improvised Mixtures 585
Laboratory Methods 586
General Fire Evidence 587
Identification of Charred or Burned Materials 587
Burned Documents 588
Failure Analysis by Forensic Engineers 589
Evaluation of Appliances and Wiring 590
Miscellaneous Laboratory Tests 591
Spoliation 595
Non-Fire-Related Physical Evidence 596
Fingerprints 596
Blood 599
Impression Evidence 600
Physical Matches 603
Trace Evidence 603
Chapter Review 606
Summary 606
Review Questions 606
References 606

Chapter 15 Fire-Related Deaths and Injuries 611

The Team Effort 613
Species of Remains 614
Identity of the Victim 614
xviii Contents
 Cause of Death 615
Manner of Death 616
Victim Status at Time of Death 616
Death Due to Fire versus Death Associated with Fire 616
Pathological and Toxicological Examination 616
General Considerations 616
Destruction of the Body 619
Effects of Fire 625
Other Pathological Findings 632
Carbon Monoxide Asphyxiation 635
The Carbon Monoxide Hazard 637
Effect of Rate of Absorption 639
Sources of Carbon Monoxide 640
Investigation of Carbon Monoxide Asphyxiations 642
Other Toxic Gases 643
Hydrogen Cyanide and Other Toxic Gases 643
Toxic Gases from Sulfur-Containing Polymers 644
Other Mechanisms 644
Burn Injuries 645
Manner of Death 649
Chapter Review 650
Summary 650
Review Questions 651
References 651

Chapter 16 Arson as a Crime 655

The Crime of Arson 658
Arson Law 659
Elements of Proof 659
Direct and Circumstantial Evidence 660
Motive 661
Profit 662
Vandalism 664
Juvenile Fire Setting 665
Excitement and Thrill Seeking 665
Revenge, Retaliation, or Spite 667
Concealment of another Crime 668
Extremism (Social Protest and Terrorism) 669
Mixed Motives 670
Irrational Fire Setting 670
The Arson Set 671
Arranging the Fire—Location 672
Fuels 673
Method of Initiation 681
Contents xix
 Deductions from the Interpretation of Evidence 687
Criminal Investigative Analysis or Profiling 688
Analytical Reasoning 689
Elimination of Accidental or Natural Causes 690
Chapter Review 692
Summary 692
Review Questions 692
Court Citations 692
References 693

Chapter 17 Other Investigative Topics 695

Safety and Health 696
Fire Modeling 698
Mathematical Fire Modeling 698
Zone Models 699
Field Models 699
Models for Specialized Applications 700
ASTM and Critical Modeling Issues 701
What Should We Ask about Any Model We Use? 701
Fire Assessment 702
Documentation 702
Model Evaluation 703
Testing Complex Computer Models 706
Critical Analysis of Cases 707
Search and Seizure 708
Search and Seizure Court Decisions 710
Sources of Information 713
Spoliation 714
Public-Sector Investigators and Spoliation 715
Private-Sector Investigators and Spoliation 715
Consequences of Spoliation 715
Chain of Evidence 718
Report Writing 718
Report Summary 719
The Scene 719
The Investigation 720
Report Conclusions 720
Report Writing Basics 720
Courtroom Testimony 721
The Expert Witness 721
Pretrial Preparation 726
Testimony 727
Scientific Method 727
Chapter Review 732
xx Contents
Summary 732
Review Questions 734
Court Citations 734
References 735

Suggested Reading 737

Glossary 745
Index 751

Contents xxi
This page intentionally left blank
PREFACE

The last 25 years have seen a dramatic change in the standards of performance expected of
fire and explosion investigators. Such changes have been brought about by Supreme Court
decisions, a development of professional standards including several certification pro-
grams, and acceptance of published texts and guides such as NFPA 921. These years have
also seen a dramatic improvement in the frequency and quality of interactions among fire
investigators, fire scientists, and engineers involved in fire safety and fire protection.
The intuitive extrapolation or interpolation of data to explain fire development or fire
indicators had been standard practice among fire investigators, and it has been faulty far too
often. The scientific method has finally been recognized as the core analytical process that
leads to accurate and defensible conclusions in fire investigations. That method, however,
requires reliable data and information, which often have been lacking in fire topics. The
integration of the wealth of information, knowledge, and experience of fire engineers and
of those scientists involved in the chemistry and physics of fire development into fire inves-
tigation has proceeded along many paths—personal, educational, and professional—and
on an international basis. The ATF Fire Research Laboratory at Beltsville, Maryland; the
Center for Fire Research at NIST; the Fire Research Station in the United Kingdom; and
numerous private researchers have all contributed significantly to fire investigations for
many years. Recently, Vyto Babrauskas published a virtual encyclopedia of information
about the ignition and combustion of materials. The Ignition Handbook is a comprehen-
sive summary of decades of fire research in an accessible, fully referenced source. This
seventh edition of Kirk’s Fire Investigation includes new material reflecting “new” knowl-
edge from that remarkable book. Fire engineers are now involved directly in investigations
and also teach investigators how to apply fire engineering principles.
The National Fire Protection Association’s NFPA 921: Guide for Fire and Explosion
Investigations has focused the attention of investigators and the legal profession on the
scientific principles behind investigation. Coauthors David Icove and John DeHaan were
active technical consultant members in the development of NFPA 921 (DeHaan from 1991
to 1999, and Icove from 1990 to date), and this seventh edition of Kirk’s Fire Investigation
reflects a closer parallel between practices and information in both sources that can only
enhance the accuracy and reliability of all investigations. Correlations are also offered to
information in NFPA 1033: Standard for Professional Qualifications for Fire Investigator.
Eight years ago, a companion book for this text was released. Forensic Fire Scene
Reconstruction was created to explore more of the engineering principles behind fire
behavior and the mechanisms of production of many of the post-fire indicators that Kirk’s
describes. Taken together, these books provide a sound basis for the fire expert to use when
evaluating fire scenes, preparing reports, or offering testimony. Today’s investigators are
being held to a higher standard of professional practice than ever before. It is no longer
adequate to claim expertise based on years of experience alone. A professional must demon-
strate that what he or she is doing follows the practices and the knowledge base of the
relevant professional community. Such knowledge and practices are based on texts such as
this one, Fire Scene Reconstruction, and NFPA 921, which are continuously peer reviewed
and revised to reflect the most current knowledge. The revisions in this edition follow the
same path as in previous editions but include many new photographs, published experi-
mental data, and case examples. There is revised material on ignition and fire dynamics,
supported by new references and color photos. This text reflects an international database
as offered by fire and explosion investigators, scientists, and engineers from all over the
world. It is offered in the hope it will augment the knowledge and improve the skills of
investigators everywhere and help them find the right answers for the right reasons.

xxiii
 It is hard to believe that it has been 30 years since John DeHaan took over responsi-
bility for the text that what was then Fire Investigation by Dr. Paul Kirk. So many years
have passed that a whole new generation of fire investigators is now practicing, many of
whom have asked, Why is it called Kirk’s Fire Investigation? It is clear they are not aware
of Professor Kirk’s contribution to the discipline. Paul L. Kirk was a professor of bio-
chemistry and criminalistics at the University of California at Berkeley, but it was his spe-
cialty of microchemistry that focused his attention on physical evidence and its analysis.
Professor Kirk was part of the Manhattan Project (where separation and identification of
trace quantities of particular chemicals was a critical step in the development of the
atomic bomb). After the war, he focused on analytical chemistry as an adjunct to criminal
investigation. He was in charge of the criminalistics program at Berkeley until his death in
1970 and launched the careers of many criminalists who now practice around the world.
He wrote the landmark text Crime Investigation in 1953 and maintained a private crimi-
nalistics consulting practice. It was this practice that led to his involvement in fire and
arson investigation, where he was consulted in a wide variety of fire and explosion cases.
He published Fire Investigation in 1969 as the first textbook on fire investigation written
by a scientist rather than a field investigator. It became a standard reference and was still
in print some 11 years after his death. His concern with using science to solve the puzzles
of fire and explosion presaged the current emphasis on using the scientific method to
investigate fires by more than 30 years. It is clear that good, knowledgeable investigators
have been using that approach for years, even if they were not aware of it.
In honor of Dr. Kirk’s pioneering work in bringing science to fire investigation, his
name is included in the title, and the spirit, of this text.

J. D. DeHaan, Ph.D., F-ABC, IAAI-CFI, FSSDip, CFEI-NAFI

D. J. Icove, Ph.D., P.E., CFEI

What’s New
The Seventh Edition is one of the most adventurous editions over the last decade. John
DeHaan has been joined by David Icove to produce the keystone textbook in the fire
investigation field. The following highlights are changes to this edition that set it apart
from previous ones.
■ Completely updated chapters with learning objectives
■ Reference tracking of the National Fire Academy–developed Fire and Emergency
Services Higher Education (FESHE) curriculum
■ New case examples and results of recent fire tests
■ Substantial new artwork and photographs, many in color
■ Updated bibliographic references and appendices, which can be found on the
MyFireKit for this text

For additional review materials, appendices, and suggested

readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyBradyKit link to Buy Access
from there.

xxiv Preface
 As an added bonus, Kirk’s Fire Investigation, Seventh Edition, features a MyFireKit,
which provides a one-stop shop for online review materials, appendices, suggested read-
ings, chapter support materials, and other resources.
You can prepare for class and exams with multiple-choice and matching questions,
Web links, study aids, and more! To register for MyFireKit for this text, please visit www.
bradybooks.com and follow the MyBradyKit link.

Scope of This Book

Kirk’s Fire Investigation is divided into the following chapters.
Chapter 1, “Introduction,” describes the field of fire investigation, which is the for-
mal process of determining the origin, cause, and development of a fire or explosion.
Chapter 2, “The Elementary Chemistry of Combustion,” describes how a fire is a
chemical reaction that produces physical effects. This chapter also introduces some of the
concepts and terms that the fire investigator might encounter when reviewing laboratory
reports or meeting with experts.
Chapter 3, “Fundamentals of Fire Behavior and Building Construction,” describes
fire as an exothermic oxidation reaction that proceeds at such a rate that it generates
detectable heat and light. This chapter also describes the basic elements of building con-
struction and materials to give the fire investigator an understanding of how they can
affect fire growth and patterns.
Chapter 4, “Combustion Properties of Liquid and Gaseous Fuel,” discusses some of
the concepts that are basic to understanding how gaseous and liquid fuels burn, the con-
ditions and limitations that apply to the combustion of such fuels, and the conventional
methods of expressing combustion properties in terms of laboratory tests.
Chapter 5, “Combustion Properties of Solid Fuels,” describes the more complex igni-
tion and combustion of solid fuels, which usually depend on pyrolysis to create com-
bustible gases and vapors.
Chapter 6, “Sources of Ignition,” describes a wide variety of heat sources and their
thermal properties. This chapter also discusses the fundamental processes of heat transfer,
heat release rate, fire propagation, and assessment that must be applied to each possible
situation.
Chapter 7, “Structure Fires and Their Investigation,” describes the principles of fuels,
ignition, and fire behavior with which investigators should be reasonably familiar before
undertaking the probe of a fire. This chapter also discusses the necessity of having a clear
understanding of the purposes and goals of the investigation and a rational, orderly plan
for carrying it out to meet those purposes, as well as the value and limitations of post-fire
indicators and the basic physical processes that create them.
Chapter 8, “Grass and Wildland Fires and Their Investigation,” describes how the
investigator who understands fuels, fire behavior, and the effects of environmental condi-
tions is in a better position to interpret the subtle and sometimes delicate signs of fire pat-
terns in wildland fires and therefore is better able to identify the origin and cause, no matter
what type of fire is involved.
Chapter 9, “Automobile, Other Motor Vehicle, and Ship Fires,” describes the fuels,
ignition sources, and dynamics as encountered in vehicles with which vehicle fire investi-
gators must be familiar to carry out correct investigations.
Chapter 10, “Electrical Causes of Fires,” describes the basics of electricity, its con-
trol, uses, and measurement of which the investigator must have a working knowledge to
accurately assess its effects on fuels in the fire environment. The chapter also discusses the
many ways in which electrical power can cause fires and diagnostic indicators of use to
investigators.
Chapter 11, “Clothing and Fabric Fires,” describes how despite governmental regula-
tions, fires in which fabrics are the first materials to be ignited are still a very common
Preface xxv
 occurrence. The chapter discusses the nature of common fabrics and upholstery materials
and their contributions to both ignition hazard and fuel load in current studies.
Chapter 12, “Explosions and Explosive Combustion,” describes the range of explo-
sion violence—from the diffused type of rolling, progressive flame resulting from the com-
bustion of a rich mixture of flammable gases or vapors in air, to the violent, almost
instantaneous detonation of condensed-phase explosives. The chapter also discusses the
chemical structure and mechanisms of low and high explosives, as well as mechanical and
electrical explosions.
Chapter 13, “Chemical Fires and Hazardous Materials,” describes how to identify
the common causes of fires involving chemicals and hazardous materials.
Chapter 14, “Laboratory Services,” describes the role of laboratory services in fire
and explosion investigation and the types of examinations that can be requested. These
include not only fire debris analysis but the application of a wide range of physical, chem-
ical, optical, and instrumental tools on a variety of substances.
Chapter 15, “Fire-Related Deaths and Injuries,” identifies the common causes of
(and factors contributing to) fire-related deaths and injuries and describes indicators and
postmortem tests. Postmortem fire effects on the body are a primary focus.
Chapter 16, “Arson as a Crime,” explains the basic concepts of arson as a crime along
with its common motives, and describes common incendiary devices and mechanisms.
Chapter 17, “Other Investigative Topics,” identifies specialized tools that are relevant
and useful in fire investigation, including fire modeling. The chapter also discusses on-scene
safety considerations for the fire investigator and the implications of the scientific method
in analyzing fire events.

xxvi Preface
ACKNOWLEDGMENTS

Peer review is important for ensuring that a textbook is well balanced, useful, authoritative,
and accurate. The following agencies, institutions, companies, and individuals provided
invaluable support during the peer-review process of this edition.
Vyto Babrauskas, Ph.D.
Fire Science and Technology Inc.
Issaquah, WA
Nick Carey
Fire Investigation Group, London Fire Brigade,
London, UK
Detective Mike Dalton
Knox County Sheriff’s Office/Fire Investigation Unit
Knoxville, TN
Gary Edwards
Fire Science Program Director/Instructor
Montana State University–Billings College of Technology
Billings, MT
Tom Goodrow
ATF National Academy (Retired)
West Chatham, MA
Gary S. Hodson, IAAI-CFI
Sgt., Provo UT Police Department (retired)
Adjunct Instructor, Utah Valley University
Special Investigator, Unified Investigations and Sciences
Provo, UT
Jeffrey Lee Huber
Professor of Fire Science
Lansing Community College
Lansing, MI
Judith Kuleta
Bellevue College
Bellevue, WA
John E. Malooly
President of Malooly & Associates, Inc.
Senior Special Agent (Retired), U.S. Department of Justice, Bureau of ATF
Chicago, IL
Matthew Marcarelli
Lieutenant, City of New Haven Fire Department
Adjunct Instructor, Connecticut Fire Academy
J. Ronald McCardle
Major, Florida Bureau of Fire & Arson Investigations (Retired)
Currently Instructing, Consulting, & Researching in Fire and Explosion Causation
Florida

xxvii
 C. W. Munson
Chemeketa Fire Technology
Salem, OR
J. Graham Rankin, PhD
Forensic Science Program
Marshall University
Huntington, WV
Steve Riggs
Public Agency Training Council
Indianapolis, IN
James P. Ryan
Fire Investigator, Arson Bureau
New York State
Department of Homeland Security and Emergency Services
Academy of Fire Science
Montour Falls, NY
Nathan Sivils
Director, Fire Science
Blinn College
Bryan, TX
Aaron S. Woolverton
Adjunct Professor
Austin Community College
Austin, TX
Rather than extend an already substantial list from previous editions, we focus on
those persons who contributed the most to this Seventh Edition. As with any other evolu-
tionary process, the result is the product of many generations. You know who you are from
those earlier lists. Rest assured that your contributions are still greatly appreciated. We
want to acknowledge the following individuals who reviewed and generously offered coun-
sel and new material to improve this edition: Doug Wood; Morris Polich & Purdy, San
Francisco, CA; Daniel Madrzykowski, P.E., National Institute of Standards and Technol-
ogy, Gaithersburg, MD; Major Ron McCardle, Florida State Fire Marshal’s Office (retired);
Dr. Niamh Nic Daéid, University of Strathclyde; Kim R. Mniszewski, P.E., FX Engineering;
Jeff Morrill, MorrFire Investigations; Gordon Damant, Sacramento, CA; Steve Riggs, Pub-
lic Agency Training Council; Senior Special Agent Paul Steensland (retired), United States
Forest Service, Susanville, CA; Luis Velazco, Luis Velazco Investigations, Ltd., St. Simons
Is., GA; and Special Agent Dino Balos and Steven J. Avato, Bureau of Alcohol, Tobacco,
Firearms and Explosives, Falls Church, VA. Your efforts were substantial, and your ideas
were greatly welcomed. Our special appreciation is offered to Doug Wood and his staff for
all the new material on spoliation and other legal issues.
Vyto Babrauskas, Mick Gardiner, Gary White, Steve Mackaig, Ron Parsons, Chris
Korinek, Rick Korinek, Bob Svare, and Mark Svare all reviewed critical portions of text
and contributed significantly to the accuracy of this book. In addition, Vyto Babrauskas
and Dan Madrzykowski always generously shared ideas, insight, and information when-
ever we asked, for which we are deeply grateful.
A number of people generously shared their case histories, test results, and photo-
graphs for this edition, including Jim Albers, Santa Ana Fire Dept. (retired); David Barber,
Goleta, CA; Steve Bauer; Dr. Roger Berrett; Lou Bilancia, Synnovation Engineering and
HTRI Forensics; Calvin Bonenberger, Fire Marshal, Lafayette Hill Fire Department; Dr.
Bernard R. Cuzzillo; Chris Bloom, CJB Consultants, Grants Pass, OR; Joe Bloom, Bloom
xxviii Acknowledgments
Fire Investigation, Grants Pass, OR; Helmut and Peter Brosz, Brosz & Assoc., Markham, On-
tario, Canada; Paul Carolan, Toronto Fire Department, Canada; Randy Crim, Fire Marshal,
Lake Jackson, TX; Donna Deaton, U.S. Forest Service (retired); Andrew Derrick; Denise De-
Mars, Streich DeMars, Inc.; Jack Deans; Det. Richard Edwards, Los Angeles County Sheriff’s
Department (retired); Ryan B. Fields, Orca Fire Investigation, Medford, OR; Capt./Inv. Bruce
Fusselman, Phoenix Fire Department, Phoenix, AZ; Nick Carey, John Galvin, and Paul
Spencer, London Fire Brigade, London, UK; Edward Garrison, Fire/Explosion Investigator,
Raleigh, NC; Thomas Goodrow, Fire/Explosives Technical Specialist, ATFE (retired); Tony
Grissim, Leica GeoSystems; Don Perkins, Curt Hawk, and Diane Spinner, Fire Cause Analysis;
Gerald Haynes, Forensic Fire Analysis, LLC, Fredericksburg, VA; Dr. Robin Holleyhead;
Science &Justice; Chief Kurt Hubele, Richland Fire Department, WA; John Jerome; Capt.
Thomas Kinkaid, Knoxville, Fire Department; Chris W. Korinek, P.E., and Richard E. Ko-
rinek, P.E., Synergy Technologies LLC; Ken Legat, Christchurch, New Zealand; SA/CFI
Michael A. Marquardt, ATFE, Grand Rapids, MI; Vic Massenkoff, Contra Costa County Fire
Department; Marion Matthews, U.S. Forest Service; Lamont “Monty” McGill, retired Fire In-
vestigator and Bomb Technician (deceased); Wayne Moorehead, Forensic Consultant; Jamie
Novak and Cameron Novak, Novak Fire Investigations and St. Paul Fire Department; Dr. Said
Nurbakhsh, California Bureau of Home Furnishings, North Highlands, CA; Chief Mike
Oakes and Tony Hudson, Clallam County Fire Investigation Team, Port Angeles, WA; Keith
Parker, Marin County Fire Department, Woodacre, CA; David W. Powell, SYTEK Consul-
tants, East Syracuse, NY; Steve Riggs and Tim Yandell, Public Agency Training Council; Susan
Sherwin, Scottsdale, AZ; Stuart Sklar, Fabian, Sklar & King, P.C., Farmington Hills, MI;
Robert Toth, Iris Investigations; Inv. Jeff Weber, San Jose Fire Department, CA; and Capt. San-
dra Wesson, Little Rock Fire Department, AR. Our special thanks to Det. Michael Dalton and
Inv. Greg Lampkin, Knox County Sheriff’s Office, Knoxville, TN, for all their great photos.
We also want to offer our special acknowledgment to Jamie Novak, who seems to be
able to burn (and blow up) more buildings than anyone else. Jamie generously opened his
awesome collection of photos, and many were selected for this book. We all know that
fire and explosion investigation is largely a visual endeavor, and such photos are vital to
the usefulness of an investigation text. The Bureau of Home Furnishings and the Fire
Research Station (Garston, UK) have repeatedly provided us with opportunities to do
special tests and share the results.
Firsthand observations of fire behavior and patterns are a critical element in qualifi-
cations for every fire expert. Organizations such as Gardiner Associates (UK), the Euro-
pean Working Group on Fires & Explosions, ATF (Glynco, GA), Florida State Fire
Marshal, and various chapters of the IAAI are to be commended for the special efforts
they have made in providing “live burn” training and research. The efforts of Dan
Madrzykowski and his team from NIST at CCAI conferences were greatly appreciated for
the careful science they demonstrated. The Training Committee of the CCAI has done an
exemplary job of providing extraordinary opportunities for firsthand fire observations.
Its motto is “Build it and they will come—to burn it and learn.” The Forensic Fire Death
Investigation Courses provided recently by the San Luis Obispo Fire Investigation Strike
Team provided unique opportunities to observe the effects of fire on human cadavers.
These results will benefit investigators through the illustrations presented here.
Our very patient editors, Monica Moosang, Pam Powell, and Barbara Liguori, saw
this text through all manner of crises. John DeHaan’s office manager, Shirley Runyan,
performed exemplary work in myriad roles—text, proofreading, graphics, and especially
in unraveling the mysteries of electronic communication. Our deepest thanks to everyone.
Finally, we want to acknowledge the personal inspiration we gain from working with
fire investigators such as Monty McGill, Jamie Novak, Jack Malooly, Ross Brogan, Jeff
Campbell, Nick Carey, Randy Crim, Bob Toth, Wayne Miller, Jim Allen, Mike Dalton,
Jim Munday, and Mike Marquardt. Their dedication to finding the right answer through
scientific analysis is an example for all of us.
Acknowledgments xxix
This page intentionally left blank
ABOUT THE AUTHORS

This textbook is coauthored by two of the most experienced fire scientists in the United
States. Their combined talents total more than 80 years of experience in the fields of
investigation, fire behavior, fire protection engineering, criminalistics, fire science, and
crime scene reconstruction.

An internationally recognized forensic scientist, Dr. DeHaan is the senior author of Kirk’s
Fire Investigation and coauthor of Forensic Fire Scene Reconstruction, the two leading
textbooks in the field of fire and arson investigation. He is also a former principal mem-
ber of the NFPA 921 Technical Committee on Fire Investigations.
Dr. DeHaan has been a criminalist for more than 40 years and has gained consider-
able expertise in the analysis of fire and explosion evidence as well as shoe prints, and in
instrumental analysis and crime scene reconstruction. He has been employed as a crimi-
nalist by the Alameda County Sheriff’s Office, the U.S. Treasury Department, and the
California Department of Justice.
His research into forensic fire scene reconstruction is based on firsthand fire experi-
ments on fire behavior involving more than 500 observed full-scale structure, and 100
vehicle, fires under controlled conditions, as well as laboratory-scale studies. Dr. DeHaan
has testified as an expert witness in civil and criminal trials across the United States and John D. DeHaan, PhD,
overseas. He is currently the president of Fire-Ex Forensics Inc. and consults on civil and FABC, CFI–IAAI, CFEI-
criminal fire and explosion cases across the United States and Canada and overseas. NAFI, FFSS, FSSDip
Dr. DeHaan graduated from the University of Illinois–Chicago Circle in 1969 with a
BS degree in Physics and a minor in Criminalistics. He was awarded a PhD in Pure and
Applied Chemistry (Forensic Science) by Strathclyde University in Glasgow, Scotland, in
1995. Dr. DeHaan is a Fellow of the American Board of Criminalistics (Fire Debris), a
Fellow of the Forensic Science Society (UK), and holds Diplomas in Fire Investigation
from the Forensic Science Society and the Institution of Fire Engineers, and a Certified
Fire Investigator certification from the International Association of Arson Investigators.
He is also a Certified Fire and Explosion Investigator in the National Association of Fire
Investigators.

An internationally recognized forensic fire engineering expert with more than 40 years of
experience, Dr. Icove is coauthor of Kirk’s Fire Investigation and Forensic Fire Scene
Reconstruction, the two leading textbooks in the field of fire and arson investigation. He
is also coauthor of Combating Arson-for-Profit, the leading textbook on the crime of
economic arson. Since 1992 he has served as a principal member of the NFPA 921 Tech-
nical Committee on Fire Investigations. As a retired career federal law enforcement agent,
Dr. Icove served as a criminal investigator on the federal, state, and local levels. He is a
Certified Fire and Explosion Investigator (CFEI).
He retired in 2005 as an Inspector in the Criminal Investigations Division of the U.S. David J. Icove,
Tennessee Valley Authority (TVA) Police, Knoxville, Tennessee, where he was assigned for PhD, PE, CFEI
the last 2 years to the Federal Bureau of Investigation (FBI) Joint Terrorism Task Force
(JTTF). In addition to conducting major case investigations, Dr. Icove oversaw the devel-
opment of advanced fire investigation training and technology programs in cooperation
with various agencies, including the U.S. Fire Administration of the Federal Emergency
Management Agency (FEMA).
Before transferring to the U.S. TVA Police in 1993, he served 9 years as a program
manager in the elite Behavioral Science and Criminal Profiling Units at the FBI, Quantico,
Virginia. At the FBI, he implemented and became the first supervisor of the Arson and

xxxi
 Bombing Investigative Support (ABIS) Program, staffed by FBI and ATF criminal profil-
ers. Prior to his work at the FBI, Dr. Icove served as a criminal investigator at arson
bureaus of the Knoxville Police Department, the Ohio State Fire Marshal’s Office, and the
Tennessee State Fire Marshal’s Office.
His expertise in forensic fire scene reconstruction is based on a blend of on-scene
experience, conduction of fire tests and experiments, and participation in prison inter-
views of convicted arsonists and bombers. He has testified as an expert witness in civil
and criminal trials, as well as before U.S. congressional committees seeking guidance on
key arson investigation and legislative initiatives.
Dr. Icove holds BS and MS degrees in Electrical Engineering and a PhD in Engineering
Science and Mechanics from the University of Tennessee. He also holds a BS degree in Fire
Protection Engineering from the University of Maryland–College Park. He is presently a
Research Professor in the Department of Electrical Engineering and Computer Science at
the University of Tennessee, Knoxville; serves on the faculty of the University of Mary-
land’s Professional Master of Engineering in Fire Protection; and is a Registered Profes-
sional Engineer in Alabama, Arkansas, Louisiana, Tennessee, Texas, Virginia, Maryland,
and Pennsylvania.

xxxii About the Authors

NFPA 1033 Matrix
NFPA 1033 identifies the professional level of job performance requirements for fire
investigators. This standard specifies the minimum job performance requirements for
service as a fire investigator in both the private and public sectors. Job performance
requirements (JPRs) for each duty are the tasks an investigator must be able to perform to
successfully carry out that duty and are summarized in the following matrix:

Professional Levels of Job Performance for Fire Investigators

as Cited in NFPA 1033, 2009 Edition

General 4.1.2 Employ all elements of the scientific method as the operating analytical
Requirements process
for a Fire 4.1.3 Complete site safety assessments on all scenes
Investigator 4.1.4 Maintain necessary liaison with other interested professionals and entities
4.1.5 Adhere to all applicable legal and regulatory requirements
4.1.6 Understand the organization and operation of the investigative team
and incident management system
Scene 4.2.1 Secure the fire ground
Examination 4.2.3 Conduct an interior survey
4.2.4 Interpret fire patterns
4.2.5 Interpret and analyze fire patterns
4.2.6 Examine and remove fire debris
4.2.7 Reconstruct the area of origin
4.2.8 Inspect the performance of building systems
4.2.9 Discriminate the effects of explosions from other types of damage
Documenting 4.3.1 Diagram the scene
the Scene 4.3.2 Photographically document the scene
4.3.3 Construct investigative notes
Evidence 4.4.1 Utilize proper procedures for managing victims and fatalities
Collection/ 4.4.2 Locate, collect, and package evidence
Preservation 4.4.3 Select evidence for analysis
4.4.4 Maintain a chain of custody
4.4.5 Dispose of evidence
Interview 4.5.1 Develop an interview plan
4.5.2 Conduct interviews
4.5.3 Evaluate interview information
Post-Incident 4.6.1 Gather reports and records
Investigation 4.6.2 Evaluate the investigative file
4.6.3 Coordinate expert resources \

4.6.4 Establish evidence as to motive and/or opportunity

4.6.5 Formulate an opinion concerning origin, cause, or responsibility for
the fire
Presentations 4.7.1 Prepare a written report
4.7.2 Express investigative findings verbally
4.7.3 Testify during legal proceedings
4.7.4 Conduct public informational presentations

Sources: NFPA 1033, 2009 ed.; S. Sklar presentation, 2008.

xxxiii
 Fire and Emergency Services Higher Education
(FESHE) Grid
The National Fire Academy-developed Fire and Emergency Services Higher Education (FESHE)
curriculum serves as a national training program guideline which is a requirement for many fire
service organizations and training programs. The following grid outlines the needs for Fire Investi-
gator I and II levels and where specific content can be located in this text:

Course 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 App App

Requirements E H
Identify and explain the responsibilities of the fire X X
department from a firefighter’s perspective when
responding to the scene of a fire, including the
possibility of incendiary devices often encountered. (FI-I)
Define criminal law and explain the constitutional X X X
amendments (4th, 5th, 6th, 8th, and 14th) as they
apply to fire investigations. (FI-I)
Analyze the precedents set by constitutional law case X X
studies that have affected fire investigations. (FI-I)
Define and explain the common terms used in fire X X X X X X X X X X X X X X X X X
investigations
Describe the basic elements of fire dynamics and how X X X X
they affect cause determination including fire behavior,
characteristics of fuels and methods of heat transfer.
(FI-I)
Analyze the relationship of building construction X
on fire investigations including types of constructions,
construction and finish materials. (FI-I)
Evaluate fire protection systems and building X X
services and discuss how their installation affects
the ignition of fires in buildings. (FI-I)
Discuss the basic principles of electricity. (FI-I) X
Explain the role of the fire investigator in recognizing X X
health and safety concerns including potential
hazardous materials awareness. (FI-I)
Describe fire scene investigations and the X X
process of conducting fire investigations using the
scientific method. (FI-I)
Explain how an investigator determines the point of X X X X X
origin in a room. (FI-I)
Identify the types of fire causes and differentiate X X
between accidental and incendiary causes. (FI-I)
Describe and explain the basic procedures used for X
investigating vehicle fires. (FI-I)
Identify the characteristics of arson and common X
motives for the firesetter. (FI-I)
Identify and analyze the causes involved in line of X X X
duty firefighter deaths related to structural and
wildland firefighting, training and research and the
reduction of emergency risks and accidents. (FI-I)
Explain the rule of law as it pertains to arrest, search X X X
and seizure procedures and their application to fire
investigations. (FI-II)
Recognize and interpret fire scenes common to X X X X X X X X
various types of fires. (FI-II)
Describe the chemistry of combustion and the X X X X X
relationship of atoms, elements, compounds, and
organic compounds on fire. (FI-II)
Explain the nature and behavior of fire including the X X X X X
effects of heat. (FI-II)

xxxiv
Course 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 App App
Requirements E H
Explain and identify the combustion properties of X X X X X
liquids, gases, and solid fuels. FI-II)
Identify and explain electrical causes of fires. (FI-II) X X
List and explain the procedures for lifting fingerprints, X X X
evidence collection and preservation. (FI-II)
List and identify the make-up and use of incendiary X X
devices, explosives, and bombs. (FI-II)
List the procedures for documenting fire scenes, X X X X
including sketching, photography, and report
writing. (FI-II)
Analyze fire-related deaths and injuries and describe X X
methods of documentation. (FI-II)
Identify the techniques for interviewing and X X X
questioning suspects and subjects. (FI-II)
Explain the role of the fire investigator in courtroom X
proceedings including courtroom demeanor and
testifying. (FI-II)
Identify and list the sources and technology available X X X X X X X X X X X X X X X X X
for fire investigations. (FI-II)

xxxv
This page intentionally left blank
 CHAPTER
1
Introduction

Courtesy of Jamie Novak, Novak Fire

Investigations and St. Paul Fire Dept.

KEY TERMS

accidental fire, p. 2 explosion, p. 2 incendiary fire, p. 3

arson, p. 4 fire, p. 2 suspicious, p. 7
combustion, p. 10 fire behavior, p. 10

OBJECTIVES

After reading this chapter, you should be able to:

■ Describe the role of fire investigators in their determination of the origin, cause,
and development of a fire or explosion.
■ List the four major peer-reviewed publications in the field of fire investigation.
■ Identify the common job performance requirements for fire investigators.
■ Explain the problem of fire loss statistics in the United States and in the United
Kingdom, particularly in the estimation of incendiary fires.
■ Demonstrate an understanding of the role of the scientific method as it applies to
fire and explosion investigations.
■ Discuss the levels of confidence as they apply to expert opinions in fire and explosion
investigations.

1
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

Fire Investigation
This textbook focuses on the field of fire investigation, which is the formal process of
determining the origin, cause, and development of a fire or explosion.1 Often called origin
fire ■ Rapid oxidation and cause (O&C) investigation, the probe is launched after a fire or explosion is extin-
with the evolution of guished and investigators then strive to determine what circumstances caused or con-
heat and light; uncon- tributed to the incident. Due to the complex nature of the event, where fire often deforms
trolled combustion.
or distorts the evidence, fire investigation is among the most difficult forensic sciences to
explosion ■ The sudden practice. Not knowing whether the fire was intentionally set, investigators must undertake
conversion of potential to prove or disprove any allegations or suspicions and to piece together the pre- and post-
energy (chemical or fire events. If the investigation indicates a fire was deliberately set, reasonable accidental
mechanical) into kinetic
energy with the
fire causes must be evaluated and eliminated.
production of heat, Fire investigators must have broad and up-to-date knowledge of topics such as fire
gases, and mechanical science, fire chemistry, thermodynamics, thermometry, fire dynamics, explosion dynam-
pressure. ics, computer fire modeling, forensic sciences, fire analysis, scientific methodology, pho-
accidental fire ■ A fire
tography, hazardous materials, building construction, gas and electricity utility systems,
occurring without human behavior (related to evacuation, victims, firesetting, witnesses, and information),
design or intent. forensic lab analysis, failure analysis, criminal and civil law, and analytical tools.
The latest edition of Kirk’s Fire Investigation is considered the leading worldwide
authoritative textbook in the field. The textbook is used both for training and as the
basis for certification in many jurisdictions. A majority of the overall methodology of
fire investigation relies upon Kirk’s as well as on the following three peer-reviewed
sources:
■ The second edition of Forensic Fire Scene Reconstruction, published in 2008, which
is also considered a leading forensic textbook, serves as a “bridge textbook” that
ties the concepts of fire engineering analysis to fire investigations.
■ The National Fire Protection Association’s NFPA 921: Guide for Fire and Explosion
Investigations, approved by the NFPA Standards Council on December 14,
2010, and issued on January 3, 2011 (NFPA, 2011). The 2011 edition of NFPA
921 supersedes all previous issues and is considered by practitioners
and judicial authorities the standard of care for conducting fire and explosion
investigations.
■ The National Fire Protection Association’s NFPA 1033: Standard for Professional
Qualifications for Fire Investigator, approved by the NFPA Standards Council on
May 30, 2008, with an effective date of July 18, 2008. The 2009 edition of NFPA
1033 supersedes all previous issues and was approved as an American National
Standard on July 18, 2008.
NFPA 1033 identifies the professional level of job performance requirements for fire
investigators.2 This standard specifies the minimum job performance requirements for
2 Chapter 1 Introduction
 TABLE 1-1 Professional Levels of Job Performance for Fire Investigators as Cited
in NFPA 1033, 2009 Edition

General Requirements for a Fire 4.1.1 Employ all elements of the scientific method as the operating analytical process
Investigator 4.1.2 Complete site safety assessments on all scenes
4.1.3 Maintain necessary liaison with other interested professionals and entities
4.1.4 Adhere to all applicable legal and regulatory requirements
4.1.5 Understand the organization and operation of the investigative team and incident
management system
Scene Examination 4.2.1 Secure the fire ground
4.2.2 Conduct an interior survey
4.2.3 Interpret fire patterns
4.2.4 Interpret and analyze fire patterns
4.2.5 Examine and remove fire debris
4.2.6 Reconstruct the area of origin
4.2.7 Inspect the performance of building systems
4.2.8 Discriminate the effects of explosions from other types of damage
Documenting the Scene 4.3.1 Diagram the scene
4.3.2 Photographically document the scene
4.3.3 Construct investigative notes
Evidence Collection/Preservation 4.4.1 Utilize proper procedures for managing victims and fatalities
4.4.2 Locate, collect, and package evidence
4.4.3 Select evidence for analysis
4.4.4 Maintain a chain of custody
4.4.5 Dispose of evidence
Interview 4.5.1 Develop an interview plan
4.5.2 Conduct interviews
4.5.3 Evaluate interview information
Post-Incident Investigation 4.6.1 Gather reports and records
4.6.2 Evaluate the investigative file
4.6.3 Coordinate expert resources
4.6.4 Establish evidence as to motive and/or opportunity
4.6.5 Formulate an opinion concerning origin, cause, or responsibility for the fire
Presentations 4.7.1 Prepare a written report
4.7.2 Express investigative findings verbally
4.7.3 Testify during legal proceedings
4.7.4 Conduct public informational presentations

Sources: NFPA 1033, 2009 Edition; S. Sklar, personal communications, 2008.

service as a fire investigator in both the private and public sectors. Job performance
requirements for each duty are the tasks an investigator must be able to perform to suc-
cessfully carry out that duty. These are summarized in Table 1-1.
Not only must fire investigators refer to and rely on these guides and standards, as
well as on expert treatises, they must conduct their investigations using a systematic and
scientifically guided approach. In cases of incendiary fires (i.e., those set intentionally to incendiary fire ■ A
destroy), in addition to determining the origin and cause of the fire, the investigator has deliberately set fire.
the added responsibility of identifying the person who set the fire and providing proof to
a trier of fact. Incendiary fires represent a significant percentage of all fires and, because
they are set specifically with the intent of destroying property, an inordinately high per-
centage of dollar losses. Finally, a majority of the fire investigators who testify as expert
witnesses usually have certification in the field by one or more internationally recognized
professional organizations.
Chapter 1 Introduction 3
 The Fire Problem
Internationally, fire continues to be the most costly of all public safety problems today, as
it has been for the past several decades. The losses in human lives and injuries due to fires
and explosions continue to occur. Fire-caused property losses are far in excess of those
caused by all classes of crime and rival those produced by hurricanes and earthquakes.
The major issue in identifying and quantifying the fire problem is the failure to obtain
accurate incident data. This problem is not limited to any single country.

FIRE STATISTICS IN THE UNITED STATES

For example, in the United States, fire incident data are collected by the U.S. Fire Administra-
tion and the National Fire Information Council using the National Fire Incident Reporting
System (NFIRS) on a voluntary basis. Shrinking state budgets have often resulted in the
downgrading of the priorities of fire incident data collection programs. Uniform Crime
Reporting (UCR) statistics are amassed through the Federal Bureau of Investigation’s (FBI’s)
arson ■ The intentional National Incident Based Reporting System (NIBRS), but arson incident data are collected
setting of a fire with only when a fire or police investigator actually submits a report into NIBRS. Subsequently,
intent to damage or arson cases go unreported when fire departments fail to pass along and fill out a separate
defraud.
police report on intentionally set fires.
Although the National Fire Protection Association (NFPA) collects data on fire inci-
dents, its estimates are based on data the association receives from fire departments that
respond to its National Fire Experience Survey.3 NFPA estimated that public fire depart-
ments in the United States responded to a total of 1,451,500 fires in 2008. Of these, an
estimated 515,000 were structure fires, which represented a decrease of 2.9 percent from
the number in 2007. The total monetary property loss for structure fires was $12.4 billion,
an increase of 16.2 percent over the previous year’s total. The average overall loss per
structure fire was $24,002, an increase of 19.7 percent from the 2007 average. Residential
fires accounted for about 78.2 percent of all structure fires (estimated at 403,000), with
291,000 occurring in one- and two-family dwellings and accounting for 56.5 percent of all
structure fires. In addition, in 2008 there were an estimated 236,000 vehicle fires and
594,000 “outside” fires.
No accounting can accurately reflect the loss in lives and injuries that results from
fires. In 1987, the NFPA estimated that 5,810 civilians died as a result of fires, and 28,295
suffered injuries.4 By 1999, these losses were reduced to 3,570 deaths in structure fires
(and 470 in vehicle fires) and 21,895 injured.5 In 2003, estimated losses were 3,925 deaths
(an increase of 16 percent over 2002 losses) in structure fires and 475 in vehicle fires, with
18,125 injuries.6
NFPA estimates on civilian fire deaths totaled 3,320 and 16,705 civilian fire injuries
in 2008, a decrease from 2007 totals of 3.2 and 5.5 percent, respectively. An estimated
2,780 civilians perished and 13,560 were injured in residential fires. Of these, 390 died
and 3,975 were injured in apartment fires, and 2,365 died and 9,185 were injured in one-
and two-family dwelling fires.7 Thus, even though the overall estimated number of fires
has remained steady since 1999, injury and death rates are dropping.
During the period 1994–2007, an average of 100 firefighters died on duty each year
(not including the 343 firefighters who died at the World Trade Center in 2001).8 There
were an estimated 36,880– 44,210 fireground injuries reported to U.S. firefighters per
year between 2003 and 2007, reflecting a fluctuating rate of 22–27 fireground injuries
per 1,000 fires per year since 1987.9 The National Safety Council reports that accidental
fire-related burn injuries are the sixth leading cause of all accidental deaths.10
In addition to these direct losses are the incalculable indirect fiscal costs. The actual
dollar property losses previously listed probably represent only 10 percent of the total
costs. There are inestimable losses due to lost business income, unemployment, reduced
property values, and reduced tax bases as a result of nearly every major fire. In a 2005 Fire
4 Chapter 1 Introduction
Protection Engineering article, Frazier estimated that the total cost of fire in the United
States is on the order of $130–$250 billion per year.11 Personnel and equipment costs are
incurred in the suppression of every fire, historic structures are lost forever, and wildland
fires destroy watersheds, timber, and wildlife, which cannot be replaced at any cost.
It should be noted that virtually all U.S. fire “statistics” are only educated guesses, in
the authors’ opinion. The Consumer Product Safety Commission (CPSC) once estimated
that fewer than 5 percent of structure fires are even reported to authorities. Most of them
were presumably too minor and resulted in no major injuries or property losses. Many
state fire marshals do not forward their own statistics to NFIRS (sometimes because of
staffing shortages). California, which represents more than 10 percent of the U.S. popu-
lation, has not been represented in NFIRS analyses for several years. Further, without ade-
quate training for the person checking off the boxes in NFIRS reporting forms, those
responses are often no more than guesses based on first impressions and on-the-fly assess-
ments made by incident commanders.

FIRE STATISTICS IN THE UNITED KINGDOM

The United States is not alone in the problem of fire loss. The Office of Fire and Rescue
Services, Communities and Local Government (CLG) in the United Kingdom (UK) col-
lects, maintains, and analyzes UK fire statistics. In May 2006, the CLG assumed respon-
sibility for the fire and rescue service. Previously, responsibility had been held by different
UK government departments: the Home Office; the Department for Transport, Local
Government (DTLG); and the Office of the Deputy Prime Minister (ODPM).
In the United Kingdom, a total of 385,000 fires were reported in 2007, of which
77,000 (23.5 percent) were in buildings. In 2008, the number of building fires fell by 8
percent to 77,000; and the number of dwelling fires fell by 6 percent, to 51,000. The total
number of fires in buildings other than dwellings (commercial premises, schools, etc.) also
fell 12 percent, to 231,000. Outdoor fires totaled 186,000, and there were 44,000 road
vehicle fires.12
In 2003 the ODPM estimated that the average costs (property damage, health care,
and lost employment) were £24,280 (approximately $47,000) per domestic fire, £58,100
(approximately $107,000) per commercial fire, and £4,670 ($8,600) per vehicle fire. The
ODPM estimated that £1.1 billion ($2.03 billion) in losses were incurred from material
damage and business interruption claims in the United Kingdom in 2003, but another
£6.6 billion ($12.2 billion) in costs fell on society in the form of suppression costs,
cleanup, environmental damage, and disruption to households.13
In 2007, there were 443 fire-related deaths in the United Kingdom, a 2 percent increase
from the 2006 figure. The UK study noted that a majority of fire-related deaths (greater
than three quarters) occur in dwelling fires, with more fire-related deaths per 1,000 fires
than in any other location.14
Findings from the 2004/05 Survey of English Housing on all outbreaks of fire expe-
rienced by households in England suggested that the fire and rescue service attended
approximately one fifth of all domestic fires. The explanation for this was that many of
the fires recorded in that survey were minor and could be put out by someone in the home
or bystanders; therefore, the fire and rescue service was not called.15

ROLE OF THE FIRE INVESTIGATOR IN ACCURATELY

REPORTING THE CAUSES OF FIRES
What role does the fire investigator play in this tragedy? By accurately and efficiently
identifying the causes of fires, whether accidental or incendiary, investigators can make a
real and substantial contribution to reducing these terrible losses. Think of the number of
people who would have died had the flammability of clothing and bedding not been iden-
tified in the 1960s or had the problems with instant-on televisions not been tracked down
Chapter 1 Introduction 5
 in the 1970s, or with coffeemakers and portable heaters in the 1980s. What of the scores
of potential victims in several large cities in the 1990s had the serial arsonists there not
been identified and incarcerated? Saving lives and preventing injuries are the real reasons
fire investigators spend long hours searching a fire scene or doing laboratory analysis and
fire research.
The role thorough fire investigation plays in the regulatory and code-making process
is also significant. Much of the progress in reducing the frequency of fires and fire deaths
is due to the advances made in improving life safety and fire safety codes. Although these
codes are written on the basis of the best predictive models, the only way to be sure they
have the desired effect and do not create other, unanticipated risks is to document the
effects of fire in code-compliant buildings and transport. In the early 1900s builders were
proud of the fire-resistant buildings they had built of stone, concrete, steel, and glass. It
was not until major tragedies like the Triangle Shirtwaist fire in 1911 that people realized
that building contents and not just the structures were major factors in fires, and having
adequate means of escape was critical to survival.16 Manual alarm and firefighting systems
were also found to be woefully inadequate.

The Detection of Incendiary Fires

Few other investigations are as daunting as a fire scene. Because a fire scene has to be
investigated before it can be established whether the crime of arson has been committed,
every fire scene must be considered a possible crime scene (from the standpoint of scene
security, processing, and evidence collection) until clear proof is found that the fire was
accidental. As a crime scene, the typical fire scene is the antithesis of the ideal. There has
been wholesale destruction of the physical material and substance of the scene, first by
the fire itself and second by fire suppression and overhaul by the firefighters. Naturally,
this destruction affects evidence of both accidental and incendiary causes. There has been
contamination of the scene by vehicles, personnel, equipment, hoses, and large quantities
of water. Scene security and control of unauthorized persons are difficult because of the
attraction fire offers to the public and the press and may be further complicated by the
large area involved in many fire scenes. Often, the scene investigation is carried out days
or weeks after the incident, after weathering and vandals may have further compromised
whatever evidence remains. Considering these negative factors, it is not surprising that
police and fire personnel have often let fire investigation fall through the cracks of respon-
sibility. As a result, it is surprising that fire causes are ever accurately determined and civil
and criminal cases successfully brought to trial.17

REPORTING ARSON AS A CRIME

The Uniform Crime Reports (UCR), published by the FBI of the U.S. Department of
Justice, provides enlightening statistics on arson in the United States. Statistics for 2007,
published in 2008, are used in this chapter. Annually, the FBI publishes Crime in the
United States, which catalogs the entire crime picture of the United States. Note that a
2006 study of statistics for previous years indicates that there is a fairly constant pattern
in arson statistics for past years.18 UCR reported in 2008 that the number of arson
offenses decreased 3.6 percent compared with those reported in 2007.19
The FBI’s UCR Program defines arson as
any willful or malicious burning or attempting to burn, with or without intent to defraud,
a dwelling house, public building, motor vehicle or aircraft, personal property of another, etc.

The FBI collects data only on the fires that through investigation were determined to
have been willfully set. Fires are not included in the data collection whose cause is labeled
as “suspicious” or are “unknown or undetermined.”
6 Chapter 1 Introduction
 The FBI points out that arson rates are calculated based on data received from all law
enforcement agencies that provide the UCR Program with data for 12 complete months.
Unlike other national reporting systems, UCR Program data collection does not include
any estimates for arson, because the degree of reporting arson offenses varies from agency
to agency.
According to the 2008 UCR, 14,011 law enforcement agencies (providing 1–12 months
of arson data) reported 62,807 arsons. Of those agencies, 13,980 provided expanded
offense data on about 56,972 arsons. These data show that arsons involving structures (res-
idential, storage, public, etc.) accounted for 24,750 fires, or 43.4 percent of the total num-
ber of arson offenses. Mobile property was involved in 28.9 percent of arsons, and other
types of property (such as crops, timber, and fences) accounted for 27.7 percent of reported
arsons.20 See Table 1-2 for a detailed breakdown of the 2008 statistics.
The 2008 FBI statistics showed that the average dollar loss per incident due to arson
was $16,015. Arsons of industrial/manufacturing structures resulted in the highest aver-
age dollar losses (an average of $212,388 per arson incident). Regarding arson rates, the
2008 UCR data showed that nationwide, the rate of arson was 24.1 offenses for every suspicious ■ Fire cause
100,000 inhabitants in the United States. has not been deter-
In contrast, the NFPA statistics for 2008 estimated that arson caused 30,500 struc- mined, but there are
ture fires and 17,500 vehicle fires, a decrease of 6.2 percent from 2007 totals. NFPA indications that the
fire was deliberately
noted in its 2008 report that in Version 5.0 of the NFIRS the classification of arson was set and all accidental
changed to intentionally set from the previous incendiary category. NFIRS no longer car- fire causes have been
ries a suspicious fire category in its reporting system.21 eliminated.

TABLE 1-2 Arson in the United States by Type of Property, 2008

NUMBER PERCENT PERCENT PERCENT OF

PROPERTY OF ARSON PERCENT NOT IN AVERAGE TOTAL OF ARSONS CLEARANCES
CLASSIFICATION OFFENSES DISTRIBUTIONa USE DAMAGE CLEARANCES CLEAREDb UNDER 18
Total 56,972 100.0 $16,015 10,277 18.0 37.4
Total structure: 24,750 43.4 19.0 29,701 5,520 22.3 35.2
Single occupancy 11,435 20.1 21.4 28,788 2,389 20.9 25.7
residential
Other residential 4,062 7.1 13.6 29,343 983 24.2 25.0
Storage 1,629 2.9 20.7 18,021 298 18.3 48.7
Industrial/ 251 0.4 22.3 212,388 48 19.1 41.7
manufacturing
Other commercial 2,331 4.1 15.7 55,531 450 19.3 28.0
Community/public 2,667 4.7 16.2 15,799 857 32.1 65.3
Other structure 2,375 4.2 21.4 13,668 495 20.8 47.1
Total mobile: 16,454 28.9 8,766 1,461 8.9 20.3
Motor vehicles 15,572 27.3 8,186 1,314 8.4 18.7
Other mobile 882 1.5 18,991 147 16.7 34.7
Other 15,768 27.7 2,099 3,296 20.9 48.7

Source: Crime in the United States: Uniform Crime Reports (Washington, DC: U.S. Dept of Justice, Federal Bureau of Investigation, 2008).
Note: 13,980 agencies; 2008 estimated population 250,243,947.
a
Because of rounding, the percentages may not add to 100.0.
b
Includes arsons cleared by arrest or exceptional means.

Chapter 1 Introduction 7
 The NFPA statistics for 2008 estimated that intentionally set fires in structure fires
resulted in an estimated 315 civilian deaths and $866 million in property losses, an
increase of 18.2 percent in property losses over those in 2007. Deliberate vehicle fires
caused $139 million in property damage, a decrease of 4.1 percent from 2007 data.22
These estimates are almost certainly very low, since many U.S. jurisdictions do not report
arson incidents (structural or vehicular) (in the authors’ experience). Many investigators
think that the total of intentionally set structure fires is closer to 40 percent.23 These sta-
tistics may become further compromised as departments devote larger portions of their
resources to Emergency Medical Services (EMS) operations and less to fire investigation.
This is a startling contrast to the United Kingdom experience. The UK authorities
(CLG) estimate that some 31 percent of structure fires in 2003 were arson caused and 26
percent in 2004/05 after an intense national campaign,24 and that deliberately set fires in
structures and vehicles took 117 lives, caused 3,200 casualties, and cost the economy
there some £2.8 billion ($5.2 billion) (including costs in anticipation of arson fires).25
In the United Kingdom, the CLG report notes that “deliberately set” fires include
those where deliberate ignition was identified or was “merely suspected” and recorded by
the fire and rescue service as “doubtful.” “Accidental” fires included those for which the
listed cause was “not known” or “unspecified.” The number of deliberate primary fires
(buildings, vehicles, outdoor structures, and any involving casualties, as defined by CLG)
fell for the third consecutive year, from 79,700 in 2005 to 72,600 (9 percent) in 2007. Of
the 21,400 deliberate fires in buildings recorded in 2007, more than half (59 percent)
occurred in buildings other than dwellings. Of these 12,600 deliberate fires in other build-
ings, around a third occurred in private garages or sheds. The number of “accidental
fires” decreased by 6 percent, to 82,000, in 2007 from 2005 (in all major categories).
Arson-caused road vehicle fires accounted for 56 percent of all deliberate fires and 68 per-
cent of all vehicle fires in the United Kingdom in 2007.26

PROBLEMS ASSOCIATED WITH ESTIMATING INCENDIARY FIRES

The U.S. estimates are certainly on the conservative side, since many fires are never prop-
erly investigated because of lack of time, or are misidentified inadvertently as accidental
fires due to lack of experience, or intentionally to avoid the complications that arise from
identifying a fire as a criminal act. These complications can range from administrative,
such as requiring more investigative time or involving another public agency, to political,
such as generating negative publicity about the community. The sharp contrast between the
UK estimates and the NFPA estimates that only 6.1 percent of the reported structure fires
were incendiary demonstrates a systematic problem with NFPA’s data gathering and statis-
tical processing. Recent estimates (2003–2007) represent a decrease of 15.7 percent from
2002’s already low percentages (traditionally around 15 percent of all fires).
One significant reason for this decrease is that the NFIRS changed its categories in
2003, eliminating the outmoded “suspicious” causation category. Only fires that were
specifically determined to be intentionally set are now reported in the “incendiary cate-
gory.” In the past, estimates of half of “suspicious” and “undetermined” fires were added
to the “incendiary” total. Fire investigators across the United States suspected that even
the “old” percentage (15 percent) was far below the reality of actual investigations. In
response to a nonscientific random survey by DeHaan, many experienced public-sector
investigators estimate that 35–40 percent of all the structure fires they see (and an even
higher percentage of vehicle fires) are deliberately set. Lack of evidence, lack of time and
resources, and limited cooperation of other public agencies make the selection of “unde-
termined” or even “accidental” causation the expedient solution. These problems have
recently been exacerbated by budgeting problems at all levels of state and local govern-
ments, which have forced severely curtailing or even ceasing fire investigation services in
both fire and police agencies. Consequently, the decision-making process has reverted to
8 Chapter 1 Introduction
fire company officers who may or may not be trained or prepared to identify evidence
supporting an incendiary cause.
In 2000, the National Institute of Justice published Fire and Arson Scene Evidence: A
Guide for Public Safety Personnel for first responders in police, fire, and medical
services.27 It was hoped that by familiarizing these professionals with the basics of fire-
related investigation and physical evidence a higher percentage of them could contribute
toward successful prosecutions. With the cutbacks in public services experienced across
the United States, it is not known whether this guide has been successfully employed.
From the small percentages of fires actually reported as intentional via NFIRS, it would
appear not.
The situation is made even more critical by the governmental agencies that find them-
selves unable to cope with the onslaught of criminal cases of homicide, rape, assault, and
drug-related activity and decide to ignore arson as a crime. This practice is conveniently
expedited by categorizing arson as a property crime so that it can be allotted reduced pri-
ority and less tactical support than the more urgent crimes against persons. It must be rec-
ognized that arson is a crime against people. It is aimed against people where they live,
where they work, and where they do business or go to church. Fire is just as much a
weapon as a gun, a knife, or an axe. When aimed at persons, fires may be set directly on
them or, more often, set deliberately to trap the victims in a structure or vehicle. Every set
fire is a crime against the personnel of fire and police agencies who are obligated to
respond to and enter fires in lifesaving efforts.
Fire is a most horrendous weapon. Successful fire investigation, culminating in an
accurate determination of cause, is the only way to minimize these terrible losses.
Accidental fires can be prevented by the identification and elimination of hazardous prod-
ucts and processes or careless practices, as well as the development and enforcement of
better building codes and fire safety practices. As the recent British experience with vehi-
cle arsons has shown, incendiary fires can be reduced by recognizing them, identifying
those responsible, and prosecuting them.28 Sadly, many cities, large and small, in the
United States have been forced to eliminate or severely reduce their investigation units,
leaving line officers to do what they can “on the fly.”
With such massive losses involved, it is not surprising that fire losses are one of the
most common subjects of civil litigation. Yet, for all its economic and legal importance,
fire remains one of the most difficult areas of investigation. It is challenging because the
destruction of the fire and, sometimes, its suppression create difficulties in reaching firm
and unassailable conclusions as to the fire’s cause. It is the investigator’s responsibility not
only to identify the origin and cause of the fire but to be prepared to defend those con-
clusions in a rational, logical manner supported by scientifically valid data. In the absence
of a rational reconstruction or scientific data, conclusions cannot be substantiated, and
successful prosecution for arson is virtually impossible. Civil cases, even with their
reduced requisite of proof, are made more difficult to decide fairly without such scientific
analysis.
From the standpoint of public safety and economic considerations, it is important
that investigative methods for accurately determining the causes of fires be understood
and applied to the fullest extent possible. There should be no presumption that a fire is
either accidental or incendiary. It is vital that the investigator not prejudge any fire as to
cause without evaluating all the collected evidence. If an investigator decides that a fire is
arson before collecting any data, then only data supporting that premise are likely to be
recognized and collected. As Sherlock Holmes pointed out, “It is a capital mistake to the-
orize before one has data. One begins to twist facts to suit theories instead of theories to
suit facts.”29 The same caution is necessary for those who decide that all fires are acci-
dental, no matter what their true cause may be. Although fire has been known since pre-
historic times, it is still poorly understood by many people, including some firefighters,
fire investigators, and others whose services depend on such understanding.
Chapter 1 Introduction 9
 Scientifically Based Fire Investigation
There is a natural tendency for everyone to be impressed by the magnitude and violence
of the fire itself and the degree of destruction that results from a large fire. In the case of
insurance investigators, this tendency is aggravated by the fact that they are expected to
assess this damage and possibly pay a claim for it. For the accurate investigation of fires,
this preoccupation with the magnitude and results of the fire is unfortunate. Cause, not
extent, should be the investigator’s concern, and too much attention to extent will obscure
and complicate the search for the origin and cause.

COMPREHENSIVE METHODOLOGIES FOR FIRE INVESTIGATION

Because of the difficulty and complexity of a complete and accurate fire investigation, and
preconceptions, there is a special need for every investigator to develop a comprehensive
analytical approach to the task. This analytical approach recognizes that to be a successful
fire investigator, a person must understand and master numerous facets of fires, fuels, people,
and investigative procedures. The investigator must really understand how a fire burns,
what factors control its behavior, and that all fires do not necessarily behave in a precisely
fire behavior ■ The predictable fashion. Such deviations from typical or expected fire behavior must be corre-
manner in which a fuel lated with the fire conditions and their causes established. These causes usually involve the
ignites, flame develops, nature of the fuel involved or the physical circumstances and environment of the fire.
and fire spreads. Unusual
fire behavior may reveal
One of the most important, yet often neglected, areas of fire investigation is an under-
the presence of added standing of the fundamental properties of the fuels involved. These properties, which
fuel or accelerants. include the chemistry, density, thermal conductivity, and heat capacity of the fuel, deter-
mine the ignitability and flame spread characteristics of the fuel. These, in turn, control
the nature of ignition as well as events that follow. These properties are determined pri-
marily by the chemistry of the fuel, and the investigator should have a thorough knowl-
edge of the relevant chemistry and what happens in the reaction. Different ignition
sources, whether glowing, flaming, or electric arc, will affect fuels differently, and so the
investigator must rely on scientific data about the interaction of such sources with the first
fuel ignited to help establish whether a particular source is even competent to accomplish
ignition and what time factors may be involved. The focus tends to be simply on identi-
fying fuels in fire scenes, but ventilation factors are equally important (as without air,
there won’t be much of a fire to investigate). Investigators need to document and evalu-
ate the sizes of rooms, windows, and doors, and changes in their condition during the fire.
Ventilation factors must be considered as carefully as fuel factors. The fundamental
requirement for a fire is the presence of fuel (in suitable form), oxygen, and heat (in suit-
able quantities), all linked by a self-sustaining chemical reaction. These factors must be
taken into account by investigators when evaluating every stage of a fire’s development
from ignition to extinguishment. The physical principles of heat transfer and effects on
surfaces must also be understood, since they are fundamental to fire spread and to the
production of fire patterns on which investigators rely after the fire.
Any attorney who is involved in fire cases, whether criminal or civil, should be gen-
erally familiar with the same areas of chemistry. An attorney who has no understanding
of what happens in chemical reactions and the results of those reactions will be seriously
handicapped when questioning both lay and technical witnesses. Highly significant facts
of the fire, which may be crucial in causing an erroneous decision in court, may not be
recognized. Such knowledge can also influence the decision whether to try a case or settle
it on the best available terms. Thus, it is advantageous for both the investigator and the
attorney involved to have a basic understanding of the simple chemistry of fuels, fire, and
combustion ■ Oxidation combustion and the physical laws that control the production of post-fire indicators.
that generates detec- Because most fires investigated involve either residential or industrial structures, there
table heat and light.
is a tendency to think of fires only in those terms. This error becomes apparent when a forest
or wildland fire is under investigation. Although wildland fires involve the same basic
10 Chapter 1 Introduction
combustion of fuel in air, the dynamics of fire spread and the analysis and interpretation
of diagnostic signs of fire behavior are quite different from those for structural fires.
Factors such as weather, topography, and fuel density must be assessed in such fires. Today
we commonly face the problems of investigating fires in the urban/wildland interface,
where houses and trees coexist. Wherever the fire is of a highly local but very specialized
nature, such as a fire in an industrial facility, the investigator must carefully analyze what
is found and seek technical assistance if the issues are even marginally in doubt. Such fires
are increasingly common when dealing with new products or complex processes of modern
technology with which even experienced investigators have no familiarity.
Explosions, closely akin to fires and sometimes accompanying them, also must be
understood by investigator and attorney alike. Explosions, like fires, can be accidental or
intentional in cause. The difference between a fire and an explosion may be so small as
to allow confusion and may almost be a matter of definition. The interrelation and dis-
tinction between the two types of chemical events must be clearly understood so that the
combined event is properly interpreted when it occurs. The explosion that accompanies a
fire is usually quite different from that produced by a bomb, commercial explosive, or a
premixed fuel–air mixture, although all involve the rapid exothermic oxidation of a fuel
or starting material. That distinction is sometimes lost both in the investigative and legal
phases of subsequent actions. The investigation of a bombing may be very different in
character from that of a fire-related explosion, but both rely on systematic collection and
analysis of data and careful formulation of accurate conclusions. These distinctions and
comparisons are developed in greater detail in Chapter 12.

THE SCIENTIFIC APPROACH TO FIRE INVESTIGATION

The scientific approach to fire investigation to be discussed here extends not only to the
review of the entire case but also to the examination of the scene, where patient, thorough,
and systematic evaluation of the scene and its contents will usually pay off in information
critical to the case. It must be appreciated that the destruction of some scenes is so com-
plete due to prolonged burning that virtually nothing can be ascertained reliably from the
ashes regarding the origin, cause, and development of the fire. There can be virtually no
reliable interpretation of burn patterns to establish direction or duration of fire exposure.
Even such scenes can yield information about the contents of the structure, which can
reveal fraud. There is a curious tendency for some investigators to show an inherent bias
as to cause. These investigators appear to believe that all fires have a single cause, for
example, electrical appliances or smoking materials; they then expend a vigorous effort to
prove that the fire had this cause, rather than spend the time seeking the true cause.
Prejudgment of fire causation is as dangerous as the prejudgment of any other unexplained
event or crime. It is unfortunate that this appears to be an occupational issue within many
fields of investigation, but perhaps it is more noticeable in fire investigation than in other
areas. Adhering to a systematic approach (which is not to say it is dogmatic) such as the
one described here will minimize the unfortunate effects of such prejudicial tendencies.
There is also an inclination to extrapolate observations about fuels and combustion
in small-scale fires to full-blown incidents without knowing whether they are, in fact,
appropriate or accurate. It is here that the data from fire engineering and fire dynamics
studies can best be applied, but only if the user is fully aware of the conditions and limi-
tations of the information.

Applying the Scientific Method

In recent years, the term scientific method of fire investigation has been more widely intro-
duced, sometimes to the confusion of fire investigators.30 The concept does not mean that
a fire investigator has to be a scientist, or to analyze things in a laboratory, or to replicate
Chapter 1 Introduction 11
 a fire scene in a field test or computer model. It means to apply the scientific method of
logical inquiry to the investigation. It is an approach most good fire investigators already
apply; it simply means the investigation follows a series of logical steps.
The universally accepted methodology cited by NFPA 921 in conducting investiga-
tions into the origin and causes of fires centers on the use of the scientific method, which
embraces sound scientific and fire protection engineering principles. NFPA 921 defines
the scientific method as
the systematic pursuit of knowledge involving the recognition and formulation of a prob-
lem, the collection of data through observation and experiment, and the formulation and
testing of a hypothesis.31

The basic concepts of the scientific method are simply to observe, hypothesize, test,
and conclude. The scientific method is considered the best approach for conducting fire
scene origin and cause determinations, analysis, and reconstruction. The process of the
scientific method is that it continuously refines and explores various working hypotheses
until arriving at a final expert conclusion or opinion, as illustrated in Figure 1-1.

STEPS IN THE SCIENTIFIC METHOD

As applied to fire investigation and reconstruction, the scientific method is a logical and
iterative seven-step systematic process. Properly applied, this systematic process addresses
how to correctly formulate, test, and validate a hypothesis.
As illustrated in Figure 1-1, the following is a summary of this seven-step process:
1. Recognize the Need. A paramount responsibility of first responders is to protect a
scene until a full investigation can be started. After initial notification, the investiga-
tor should proceed to the scene as soon as possible and determine the resources
needed to thoroughly conduct the investigation. It is necessary not only to determine
the origin and cause of the event but to determine whether future fires, explosions,
or loss of life can be prevented through new designs, codes, or enforcement strate-
gies. In public-safety sectors, determination of the origin and cause of every fire is
often a statutory requirement. In some cases, the fire investigator may be called on
to establish responsibility for the fire.
2. Define the Problem. Devise a tentative investigative plan to preserve and protect
the scene, determine the cause and nature of the loss, conduct a needs assessment,
formulate and implement a strategic plan, and prepare a report. This step is usually
focused on the determination of the cause and nature of the loss but may also include
determining who has the primary responsibility and authority to interview witnesses,
protect evidence, and review preliminary findings and documents describing the loss.
3. Collect Data. Facts and information about the incident are collected through direct
observations, measurements, photography, evidence collection, testing, experimenta-
tion, case histories, and witness interviews. All collected data should be subject to
verification of how they were legally obtained, their chain of custody, and notation
as to their reliability and authoritative nature. Data collection should include the
documentation of the property; construction and occupancy; fuel loading; the
processing and layering of the debris and evidence as found; examination of the
fire (heat and smoke) patterns; char depths and calcination; and the application of
electrical arc mapping surveys where relevant.
4. Analyze the Data. Using inductive reasoning (which involves formulating a con-
clusion based on a number of observations), analyze all the data collected. The
investigator relies on his or her knowledge, training, and experience in evaluating
the totality of the data. This subjective approach to the analysis may include knowl-
edge of similar loss histories (observed or obtained from references), training in and
understanding of fire dynamics, fire testing experience, and experimental data. Data
12 Chapter 1 Introduction
 FIGURE 1-1 A flowchart
Step 1 - Recognize the Need Step 7 - Select the outlining the scientific
• Respond to loss, protect scene, assess resources Final Hypothesis method as it applies to
• Investigate to identify conditions or persons fire scene investigation
responsible, prevent future similar losses “Expert Conclusion and reconstruction.
or Opinion” Derived from NFPA 921
• Make authoritative conclusion [NFPA 2008 and 2011,
as to area and point of fire pt. 4.3]. Source: Icove, D. J.,
origin(s), and cause and DeHaan, J. D. Forensic Fire
Step 2 - Define the Problem
• Identify competent ignition Scene Reconstruction, 2nd ed.,
• Devise and implement a tentative investigative and sources Pearson Brady Publications,
strategic plan • Establish level of confidence Upper Saddle River, NJ, 2009.
• Determine primary responsibility and authority of final hypothesis
• Determine cause and nature of fire • Prepare expert report

Step 3 - Collect Data

• Describe, observe, and document incident
• Photograph, sketch, and collect evidence
• Conduct tests and experiments
• Interview witnesses
• Assemble historical loss histories of similar incidents

Step 4 - Analyze the Data

(Inductive Reasoning)
• Evaluate, where applicable, fire pattern damage,
heat and flame vectoring, arc mapping, fire
engineering analysis
• Test predictions with established scientific knowledge
of fire dynamics, loss histories, fire tests, and
experimental data
• Recommend changes in working hypothesis

Step 5 - Develop a Working Hypothesis

• Develop a tentative working hypothesis from data
analysis
• Address causal or mathematical relationships
• Formulate and test alternative hypotheses with
established scientific and fire engineering knowledge

Step 6 - Test the Working Hypothesis

(Deductive Reasoning)
• Compare hypotheses to all known facts, incidence
of prior loss histories, authoritative fire test data,
sound published treatises, and experiments
• Recommend collection and analysis of additional
data
• Solicit peer review to identify alternative hypotheses
• Modify working hypothesis
• Eliminate all other fire causes, if possible

Chapter 1 Introduction 13
FIGURE 1-2 The working Fire
hypothesis relies on a Dynamics Professional Fire
wide spectrum of factors. and Guidelines and Pattern
Source: Icove, D. J., and Modeling Standards Analysis
DeHaan, J. D. Forensic Fire
Scene Reconstruction, 2nd ed.,
Pearson Brady Publications,
Upper Saddle River, NJ, 2009.
Fire Human
Testing Behavior

Working
Hypothesis

Witness
Fire Scene Statements
Assessment

Fire and
Explosion Environmental Forensic
Loss Interactions Evidence
Histories

evaluated can include the pattern of fire damage, heat and flame vectoring, arc
mapping, and fire engineering and modeling analysis.
5. Develop a Working Hypothesis. A hypothesis is defined as “a supposition or con-
jecture put forward to account for certain facts, and used as a basis for further
investigation by which it may be proved or disproved.”32 Based on the data analy-
sis, the investigator develops a tentative or working hypothesis to explain the fire’s
origin, cause, and development that is consistent with on-scene observations, physi-
cal evidence, and testimony from witnesses. The hypothesis may address a causal
mechanism or a mathematical relation (e.g., plume flame height, impact of differing
fuel loads, locations of competent ignition sources, room dimensions, impact of
open or closed doors and windows). Figure 1-2 shows how working hypotheses
take into account a wide spectrum of factors.
6. Test the Working Hypothesis. Deductive reasoning involves a conclusion based on
previously known facts. Using deductive reasoning, compare the working hypothe-
sis with all other known facts, the incidence of prior loss histories, relevant fire test
data, authoritative published treatises, and experiments. Use the hypothesis testing
to eliminate all other reasonable origins and causes for the fire or explosion, recom-
mend the collection and analysis of additional data, seek new information from
witnesses, and develop or modify the working hypothesis. Interactively repeat steps
4, 5, and 6 until there are no discrepancies with the working hypothesis. A critical
feature of hypothesis testing is to create alternative hypotheses that also can be
tested. If the alternatives are in opposition to the working hypothesis, their evalua-
tion may reveal issues that need to be addressed. By testing all hypotheses rigor-
ously against the data, those that cannot be conclusively eliminated should still be
considered viable.
The scientific process as used in fire investigation is fundamentally different
from that used in other fields, because in most fire cases, the solution or final
hypothesis will not be tested by burning another house, vehicle, or wildland to see
if the outcome is similar. The investigator must gather adequate reliable data and
apply robust, reliable analyses (usually of basic indicators of fire movement, inten-
sity, or duration) to develop a working (primary) hypothesis. Because these findings
will often be “tested” in an adversarial legal proceeding, the fire investigator is
14 Chapter 1 Introduction
 expected to have devised and tested alternate explanations (hypotheses). The evalu-
ation of all these hypotheses is sometimes called abductive reasoning. The final out-
come of this process is the selection of the hypothesis that fits all available data and
the rejection of all other reasonable hypotheses. When applied in medical evaluations,
this process is called differential diagnosis. This process best describes most modern
fire investigations. Even though it does not follow every sequential step of the pure
“scientific method,” it is a well-recognized analytical process.
It should be noted that this application of the scientific method is not the same
as that used in experimental science. There, observation of an event causes the
scientist to formulate a hypothesis about a possible explanation and then design
experiments to gather data (information) with the specific intent of testing or reject-
ing that primary hypothesis. Only if the primary hypothesis has been rejected (or
shown to be incorrect) will the experimenter return to the process and create another
testable scientific hypothesis. The crucial test for scientific inquiries is to use the
“proven” hypothesis to predict the outcome of further tests conducted by others
(demonstrating its reproducibility and reliability). In fire/explosion investigations,
it is very rarely possible to re-create an event to fully test the final hypothesis.
7. Select Final Hypothesis (Conclusion or Opinion). An opinion is defined as “a belief
or judgment based upon facts and logic, but without absolute proof of its truth.”33
When the working hypothesis is thoroughly consistent with evidence and research,
it becomes a final hypothesis and can be authoritatively presented as a conclusion
or opinion of the investigation.

LEVELS OF CONFIDENCE
Upon selection of a final hypothesis, the confidence of an expert opinion is often deter-
mined. Levels of confidence may also vary with the profession of the person preparing the
report. For example, fire protection engineers apply science and engineering principles to
protect people, homes, workplaces, the economy, and the environment from the devastat-
ing effects of fires. Fire protection engineers also analyze how buildings are used, how fires
start and grow, and how fires affect people and property. They commonly state in their writ-
ten reports that their expert opinion is “to a reasonable degree of engineering certainty.”
Forensic opinions should be offered only at a very high level of certainty; that is, no other
logical solutions offer the same level of agreement with the available data. In forensic origin
and cause determination, the investigator applies scientific knowledge to re-create the path
of spread of the fire, trace it back to its origin, and, there, establish the cause of the fire. Such
opinions are often expressed “to a reasonable degree of scientific certainty.”
When various hypotheses about the cause of the fire or explosion are tested, if only
one hypothesis fits the available data and others are conclusively eliminated, the opinion
can be expressed to a reasonable degree of scientific certainty (within the limitations of
the available data). All scientific conclusions are subject to continual retesting or reeval-
uation if new reliable data are presented. For example, early observational “data” led the
ancients to believe Earth was at the center of the universe. Later data showed that belief
wrong and suggested a heliocentric universe. Modern data, of course, show our entire
solar system to be circulating in an enormous galaxy moving across the universe.
If two (or more) hypotheses about the origin or cause of a fire or explosion exist and
neither can be demonstrated to be false, the degree of certainty or confidence is reduced
to “possible” or “suspected,” and the conclusion should be “undetermined.” The posing
of alternative hypotheses is in the U.S. adversarial judicial system a fair way of testing the
strength or certainty of expert conclusions.
An expert opinion should be expected to withstand the challenges of a reasonable
examination by peer review or through cross-examination in the courtroom, according to
NFPA 921.34 Curiously, NFPA 921 currently discusses only two levels of confidence with
Chapter 1 Introduction 15
 respect to opinions—probable and possible. There is no mention of a “conclusive” opin-
ion! A “conclusive” opinion can be offered when all the available data fit the final
hypothesis and all the reasonable alternatives can be tested and have been conclusively
accepted or rejected. Obviously, if there is not a conclusive opinion, but the weight of
available data suggests one alternative is vastly more likely than the others, a “probable”
opinion can be offered. If two or more hypotheses are equally likely, then a “possible”
opinion is justified.
The collection and interpretation of information, and the decision-making process,
will be addressed in a later chapter of this book, but the steps presented here are what
every professional fire investigator follows even when he or she does not stop and enu-
merate them.

LEGAL OPINIONS REGARDING SCIENCE IN INVESTIGATION

Some of the most significant changes in fire investigation in the last 15 years have been
the result of legal interpretations of the Daubert, Kumho Tire, Joiner, and Benfield deci-
sions (which will be discussed in some detail in Chapters 16 and 17). Today’s fire inves-
tigator must demonstrate to the court that he or she has the training and qualifications to
offer an opinion about the fire, has data (information) to support that conclusion, has fol-
lowed an appropriate protocol in the investigation and decision-making process, and has
followed, in general, the scientific method of inquiry. A basic protocol is outlined in the
NIJ Guide described earlier. More detailed protocols are discussed in Forensic Fire Scene
Reconstruction35 and NFPA 921: Guide for Fire and Explosion Investigations.36 Protocols
will be discussed in Chapters 7 and 8. Note that a protocol is not a cookbook with all the
steps laid out in mandatory order but an outline of a systematic, comprehensive approach
to scene investigation.
The complexities encountered in fire investigation are sometimes overwhelming to
the investigator, but patience and adherence to fundamental scientific principles of com-
bustion, fire behavior, and heat transfer will usually result in a reasonable and defensible
diagnosis of the fire if there are sufficient reliable data. These elementary principles of
combustion, fuels, and heat effects, as well as the equally important principles of scientific
investigation, are the concerns of this book.

16 Chapter 1 Introduction
 CHAPTER REVIEW

Review Questions
1. Give three examples of direct costs associated 4. What are some of the negative factors to overcome
with fire losses. Give three examples of indirect when investigating a fire scene?
costs associated with fire losses. 5. List the steps an investigator might follow when
2. What is the fundamental organizational problem using the scientific method of fire investigation.
that interferes with the complete investigation of 6. Name some reasons why the reported number of
fires in most communities? incendiary fires is considerably less than the
3. What points of logic support the idea that arson actual number that occur.
should be considered a crime against people
rather than a property crime?

References
1. NFPA 921: Guidelines for Fire and Explosion 13. Office of the Deputy Prime Minister, “Economic Cost
Investigation (Quincy, MA: National Fire Protection of Fires in England and Wales—2003” (London:
Association, 2008), pt. 3.5.59; (2011), pt. 3.3.62. ODPM, 2005).
2. NFPA 1033: Standard for Professional Qualifications 14. McEvoy and Gamble, “Summary Fire Statistics.”
for Fire Investigator (Quincy, MA: National Fire 15. Office of the Deputy Prime Minister, Statistical
Protection Association, 2009). Bulletin, “Fires in the Home: Findings from the
3. M. J. Karter Jr., “Fire Loss in the United States, 2004/2005 Survey of English Housing” (London:
2008” (Quincy, MA: National Fire Protection ODPM, 2007).
Association Report, Fire Analysis and Research 16. D. Von Drehle, Triangle: The Fire That Changed
Division, August 2009). (See also NFPA Journal, America (New York: Atlantic Monthly Press, 2003).
September–October 2009). 17. J. D. DeHaan, “Challenge of Fire Investigation” in
4. M. J. Karter Jr., “U.S. Fire Loss in 1987,” Fire Journal Proceedings Interflam 2001 (London: Interscience
(September 1988). Communications, 2001).
5. M. J. Karter Jr., “1999 U.S. Fire Loss,” NFPA Journal 18. Crime in the United States: Uniform Crime Reports,
(September–October 2000). 2006 (Washington, DC: U.S. Department of Justice,
6. M. J. Karter Jr., “Fire Loss in the United States, 2003” Federal Bureau of Investigation).
(Quincy, MA: National Fire Protection Association 19. Crime in the United States: Uniform Crime Reports,
Report, Fire Analysis and Research Division). 2008 (Washington, DC: U.S. Department of Justice,
7. Karter Jr., “Fire Loss in the United States, 2008,” 9. Federal Bureau of Investigation).
8. R. F. Fahy, P. R. LeBlanc, and J. L. Molis, “Firefighter 20. Ibid.
Fatalities in the United States in 2008,” NFPA Journal 21. Karter, “Fire Loss in the United States, 2008,” 14.
(July–August 2009): 60–67. 22. Idem.
9. M. J. Karter Jr. and J. L. Molis, “Firefighter Injuries 23. J. D. DeHaan and D. J. Icove, Forensic Fire Scene
for 2007,” NFPA Journal (November–December 2008): Reconstruction, 2nd ed. (Upper Saddle River, NJ:
46–54. Pearson/Brady, 2009), 234.
10. National Safety Council, “Report on Injuries in 24. “Communities and Local Government Fire Statistics,
America” (NSC: Itasca, IL, 2008). United Kingdom, 2007,” (Garston, Watford,
11. P. Frazier, “Total Cost of Fire in the United States,” Hertfordshire: CLG, Fire Statistics and Social Research
Fire Protection Engineering 26 (April 2005): Branch, Building Research Establishment, March
32–36. 2008).
12. P. McEvoy and J. Gamble, “Summary Fire Statistics, 25. ODPM, 2005.
United Kingdom, 2007” (Garston, Watford, 26. McEvoy and Gamble, “Summary Fire Statistics.”
Hertfordshire: Communities and Local Government, 27. National Institute of Justice, Fire and Arson Scene
Fire Statistics and Social Research Branch, Building Evidence: A Guide for Public Safety Personnel
Research Establishment (August 2009). (Washington, DC: U.S. Department of Justice), 2000.
Chapter 1 Introduction 17
28. ODPM, 2005. American Society for Testing and Materials Committee
29. A. C. Doyle, “A Scandal in Bohemia” in The Complete E30.40 on Technical Aspects of Products Liability
Sherlock Holmes (Garden City, NY: Doubleday), 1970. Litigation (withdrawn 1995), 1989).
30. S. J. Avato, “Scientific Methodology,” Fire and Arson 33. Ibid.
Investigator (July 2000): 35–36. 34. NFPA 921, 2011, pt. 18.6
31. NFPA 921, pt. 3.3.139; 2011, pt. 3.3.62. 35. Icove and DeHaan, Forensic Fire Scene
32. ASTM E1138-89: Terminology of Technical Aspects of Reconstruction.
Product Liability Litigation (West Conshohocken, PA: 36. NFPA 921, 2011.

18 Chapter 1 Introduction
 CHAPTER
2
The Elementary Chemistry
of Combustion

Courtesy of Jamie Novak, Novak Fire

Investigations and St. Paul Fire Dept.

KEY TERMS

absolute temperature, p. 29 flame, p. 21 normal hydrocarbons, p. 24

aliphatic, p. 24 flaming combustion, p. 28 organic, p. 23
aromatic, p. 24 flammable (same as inflammable), p. 22 oxidation, p. 21
atom, p. 20 heat of combustion, p. 27 paraffinic, p. 24
Btu, p. 27 hydrocarbon, p. 23 pyrolysis, p. 28
combustible, p. 28 inorganic, p. 23 saturated, p. 24
diatomic, p. 20 joule, p. 27 vapor, p. 21
exothermic, p. 21 naphthenics, p. 26 volatile, p. 26

OBJECTIVES

After reading this chapter, you should be able to:

■ Understand what is a chemical reaction and how it is involved in fire.
■ Describe the basis of how elements can combine with other atoms to form
molecules by chemical interaction.
■ Explain simple oxidation reactions involving carbon, carbon monoxide, or
sulfur with oxygen.
■ List several carbon-containing organic compounds.
■ Describe four items formed during the thermal decomposition of wood (pyrolysis).
■ List and describe three physical states of fuel.

19
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

F
ire is a chemical reaction that produces physical effects. As a result, the fire investiga-
tor should have some familiarity with the simpler chemical and physical properties and
processes involved. Because fire consists of a number of chemical reactions occurring
simultaneously, it is important to understand first what a chemical reaction is and how it is
involved in a fire. This chapter also introduces some of the concepts and terms that the fire
investigator might encounter when reviewing laboratory reports or meeting with experts.

Elements, Atoms, and Compounds

Under normal conditions, all matter is composed of elements or combinations of elements
called compounds. An element is a substance that cannot be decomposed into simpler
substances by chemical or physical treatment. Examples of elements are iron, carbon,
oxygen, chlorine, sulfur, hydrogen, and chromium. There are 92 naturally occurring ele-
ments, along with a number of unstable, synthetic elements that are never encountered
outside the laboratory. In descriptions of chemical reactions, each element is represented
by a symbol, usually the first letters of an element’s English or Latin name. For example,
hydrogen is represented as H, oxygen as O, carbon as C, lead (plumbum) as Pb, and iron
(ferrum) as Fe. See Appendix A for a list of the elements and their physical properties.
atom ■ Smallest unit All elements are composed of very small units called atoms. All the atoms of a single
of an element that still element are the same as each other in terms of their size, weight,* and chemical properties,
retains its properties. but are different from those of every other element.
Although the atom is not the indivisible block of matter that it was once thought to
be, in a chemical reaction like fire, it does not break up into its component particles.
Therefore, the atom is the smallest unit of an element that takes part in a chemical reaction.
The atoms of nearly every element can combine with other atoms to form molecules
by chemical interaction. A molecule may be defined as the smallest unit of an element or
compound that retains the chemical characteristics of the original substance. For example,
the atoms of hydrogen, the simplest and lightest element, like those of most other elements,
do not exist for very long as single or free atoms. They prefer to be in combination with
other atoms. If the hydrogen is pure, two atoms of hydrogen will combine with each other
diatomic ■ Molecules to form a relatively stable diatomic (having two atoms) molecule of hydrogen gas repre-
consisting of two atoms sented by the chemical formula H2. The letter H represents the hydrogen atoms, with the
of an element.
subscript 2 meaning that there are two hydrogen atoms in that molecule.
To understand these chemical combinations, it should be understood that every atom
consists of a central core of subatomic particles called protons (positively charged) and
neutrons (neutral or uncharged) in various numbers, surrounded by a cloud of negatively
charged electrons circulating about the nucleus like planets around our sun. It is these

*Isotopes of an element have different atomic weights and radioactive properties, however.

20 Chapter 2 The Elementary Chemistry of Combustion

electrons and the ease with which they are shared with other atoms that determine the
chemical reactivity of the element. This chemical reactivity tells us how easily the atom
will break the bonds or ties that hold it in its molecule and how easily it will form bonds
with other elements. The electrons surrounding the nucleus of each atom are what deter-
mine how that atomic element will prefer to combine with other atoms in particular
ratios. For example, at normal temperatures oxygen is a diatomic gas represented by the
formula O2, which constitutes about 21 percent of the air. Occasionally, however, it will
combine as a triatomic molecule, O3. This compound is called ozone and is produced
most often by a high-voltage electrical arc in an oxygen environment. The ozone mole-
cule is not very stable and will readily give up its “extra” atom to form a stable oxygen
molecule, O2. The free atomic oxygen produced then very quickly reacts with other atoms
in the vicinity to form oxides or with another free oxygen atom to form additional O2
molecules. Oxygen is the most critical component of all ordinary fires because it is essen-
tial for common combustion. Eliminating it will extinguish nearly all fires.

THE OXIDATION REACTION

Fire is not only a number of simultaneous chemical reactions, it is, overall, a series of
oxidative reactions. This means that atoms in the fuel are being oxidized, that is, combin-
ing with the oxygen of the air. There are many kinds of chemical reactions, but the ones
of most interest in a flame are oxidations. flame ■ A luminous
It is interesting to see what happens in a very simple oxidation, common to most fires. cloud of burning gases.
When hydrogen, an excellent fuel, is oxidized, two diatomic molecules of hydrogen com- oxidation ■
bine with one diatomic molecule of oxygen to form two molecules of water. Simply, the Combination of an
chemical formula for that reaction can be written as element with oxygen;
chemical conversion
2H2 ⫹ O2 S 2H2O involving a loss of
electrons.
Because water is a more stable compound than the gases that form it, the reaction
occurs with great vigor and the output of much heat and is called an exothermic (heat- exothermic ■
producing) reaction. If the gases are mixed before being ignited, a violent explosion will Generating or giving
off heat during a
result upon ignition. When they are combined in a flame, a very hot flame is produced.
chemical reaction.
Because hydrogen is found in almost all fuels, even in the complex molecules that
make up wood, plastic, or oil, the burning of virtually any common fuel results in the pro-
duction of water vapor in large quantities. This water vapor is sometimes seen condens- vapor ■ The gaseous
ing on the cold glass windows of a burning structure. The conversion of chemically bound phase of a liquid or
hydrogen in the fuel into water vapor during the fire also invariably leads to the produc- solid that can be
returned to that phase
tion of much heat, although less than what is produced when pure hydrogen is burned. by the application of
The net energy content of complex fuels is less than that of pure hydrogen or pure carbon pressure.
because some of the energy is consumed in breaking the chemical bonds that hold the oxi-
dizable atoms in their molecular structure.
Compounds may be represented by their numerical or empirical formulas, such as H2O,
CO2, and O2, or by structural or molecular formulas in which lines are used to represent the
sharing of electrons by the atoms of a molecule. In this way water, H2O, is represented as
O
H H
and carbon dioxide, CO2, is
O C O
The lines represent the forces or bonds between the atoms of elements making up the
molecule. A single line indicates that one pair of electrons is shared, two lines indicate two
pairs shared, three lines mean three pairs, and so on. In this way, the chemist shows not
only how many atoms are in a molecule but in what order they are arranged and the nature
of the chemical bonds between them, giving a great deal of information in a small space.
Chapter 2 The Elementary Chemistry of Combustion 21
 CARBON COMPOUNDS
Another element, carbon, represented chemically as C, is of very special interest in connec-
tion with fires because it, like hydrogen, is almost invariably a major component of fuels. In
flammable (same as fact, it is the element around which most of the flammable compounds are built.
inflammable) ■ A com- Some materials, such as diamonds and graphite, are pure carbon, but in crystalline
bustible material that
forms that are very difficult to burn. When diamond does burn, it produces the same prod-
ignites easily, burns
intensely, or has a rapid ucts (CO2 and CO) and energy as any other form of carbon. When graphite burns, it is
rate of flame spread. consumed so slowly that high-temperature crucibles are often made from it. Charcoal and
coke are both industrial products consisting primarily of carbon, but quite impure carbon.
These materials do not ignite easily. However, when afire, they produce considerable heat
and are consumed at a slow rate. The general chemical equation ordinarily applied to the
oxidation of carbon is

C(solid) ⫹ O2 (gas) S CO2 (gas)

Carbon dioxide (CO2) is always produced in fires involving carbonaceous fuels. It is

an end product of nearly all combustions, including those that occur in the animal body.
Exhaled breath is markedly enriched by carbon dioxide formed by oxidation of foodstuffs
in the tissues. This is an oxidation like that which occurs in a fire, although the foodstuffs
are not carbon alone but compounds of carbon with other elements. As a practical matter,
in all fires there is also another reaction, which may be quite secondary or may become
primary depending on the supply of oxygen. It is

2C(solid) ⫹ O2 S 2CO

Carbon monoxide (CO), the product of this reaction, is the commonly known gas
that produces an asphyxiating effect (which will be dealt with in some detail in Chapter
15). Although CO invariably accompanies combustion reactions involving any carbona-
ceous fuel, it will not reach dangerous proportions in a properly adjusted gas appliance.
Its concentration is very dependent on fire conditions: very low in freely burning flames
but very high in oxygen-deficient fires or smoldering fires.
The three reactions given (those producing H2O, CO2, and CO) merely indicate the
basis of the final combustion products and do not define the complex mechanisms by
which the products are actually formed; however, they constitute the three most basic
reactions of the fire. Water and carbon dioxide are the major products of nearly all fires,
with carbon monoxide in somewhat lower concentrations in the effluent gases. Some
fuels, notably those of coal or petroleum origin, are composed almost entirely of carbon
and hydrogen and have only small amounts of other elements. A possible exception is
sulfur, which is an impurity in most raw fuels. It also is oxidized, to form sulfur dioxide,
according to the simple reaction

S(solid) ⫹ O2 S SO2

Sulfur dioxide (SO2) is the very sharp smelling gas often noted around metal smelters
and other industrial installations.

OTHER ELEMENTS
Numerous other elements are found in the types of fuels most commonly encountered by
fire investigators. Nitrogen does not burn in the sense of generating an exothermic reac-
tion. Its role in a fire is complex, and yet it is inconsequential as a fuel. The nitrogen in
air (~70%) absorbs heat and expands proportionally. Nitrogen may be encountered when
it is part of a molecule such as a nitrate, which delivers extra oxygen to a reaction. The
role of such oxidizers will be discussed in later chapters dealing with explosives and haz-
ardous materials.
22 Chapter 2 The Elementary Chemistry of Combustion
 Other elements present in wood include sodium, silicon, aluminum, calcium, and
magnesium. These are the constituents (in the form of their respective oxides) that form
the white or gray ash that remains after wood is burned. They do not contribute signif-
icantly to the combustion itself, with certain exceptions that will be mentioned later in
this text.

Organic Compounds
Most of the compounds we have discussed up to this point fall into the general class of
molecules called the inorganics. They are the combinations of sulfur, lead, chlorine, iron, inorganic ■ Containing
and, in fact, all elements except carbon. The compounds based on carbon are so numer- elements other than
carbon, oxygen, nitro-
ous and so vital to life processes that they are considered a branch of chemistry them-
gen, and hydrogen.
selves. The chemistry of carbon-containing compounds is called organic chemistry, and
most of the important fuels involved in structure and wildland fires are organic com- organic ■ Compounds
pounds. For the fire investigator, hydrocarbons and carbohydrates are the most signifi- based on carbon.
cant organic compounds and are discussed next.

HYDROCARBONS
The compounds that form good fuels are legion, but they fall into relatively few classes.
Preeminent as a class are the hydrocarbons, compounds composed solely of carbon and hydrocarbon ■
hydrogen. The simplest of these is methane (CH4), which is the chief component of natural Chemical compound
gas. Written structurally, containing only
hydrogen and carbon.
H
H C H
H
the formula indicates the important fact that carbon has four valences or combining
points on each atom. Hydrogen has only one, which means that four hydrogen atoms
can combine with a single atom of carbon. The complete combustion of methane yields
CO2 and H2O as final combustion products, but even this seemingly simple reaction
goes through numerous intermediate reactions. The combustion of methane in oxygen
is thought to involve about 100 such elementary reactions.1 These reactions involve the
loss of hydrogen atoms and then various recombinations that produce ethene and acety-
lene within the flame as well as unstable molecular species such as —OH, —CH2O,
and —CHO.
These unstable species are called free radicals. The more complex the molecule,
the more pathways the intermediate reactions can follow and the greater the variety of
free radicals that are produced. In hydrocarbon combustion, the —CHO and —OH
free radicals (as well as CO) combine with O and H to form H2O and CO2. Free rad-
icals can exist only at relatively high temperatures, but as they cool they can condense
and form the pyrolysis products such as the brown “varnish” or oily residues that are
seen adhering to surfaces after a fire. (These processes are explored in later sections of
this text.)
In addition, carbon is somewhat unique in that its atoms have a very strong capacity
to combine with one another to form chains, rings, and other complex structures. This
capacity to form chains might be illustrated by the compound butane, which is packaged
as a liquid petroleum gas, whose formula is C4H10, and is written structurally as
H H H H
H C C C C H
H H H H
Chapter 2 The Elementary Chemistry of Combustion 23
normal hydrocarbons These straight-chain compounds can be extended almost indefinitely. They are called
■ (n-alkanes)
normal hydrocarbons or n-alkanes. Such chains can be branched to produce more complex
Hydrocarbons having
straight-chain struc-
structures without including elements other than carbon and hydrogen. Because there are
tures with no side many possibilities for the location of such branching, a large variety of compounds may
branching; aliphatics. exist that have the same empirical formula but different structures. For example, the com-
pound isobutane has the same empirical formula as butane but the structure

H H H
H C C C H
H H
H C H
H

Here the prefix iso- indicates a branched structure. Both chain length and point of
aliphatic ■ branching can be varied almost indefinitely, so that there are an enormous number of pos-
Hydrocarbons with sible compounds. All these compounds are called aliphatic or paraffinic (meaning they
straight-chain structures have the maximum number of hydrogen atoms attached to carbon atoms with no double
of the carbon atoms;
normal hydrocarbon.
bonds) and are named according to the number of carbon atoms in the longest carbon
chain in the molecule. See Appendix B for some examples of the nomenclature for these
paraffinic ■ compounds.
Hydrocarbon compounds
Another important class of hydrocarbons has a six-membered ring of carbon atoms
involving no double or
triple C—C bonds; alka- as part of its structure. The simplest such compound is benzene (C6H6):
nes; saturated aliphatic
hydrocarbons. H
C
H C C H
H C C H
C
H

Note that there is an alternating sequence of double and single bonds between the
carbons in the benzene ring. Because these ring structures impart a characteristic smell to
aromatic ■ many of their compounds, they are called aromatics. One or more of the hydrogen atoms
Hydrocarbon compound of a ring compound can be replaced with any number of chemical structures, some of
whose structure is based which can be very large and complex. Again, the simplest of the substituted benzene-ring
on a benzene ring.
or aromatic hydrocarbons is toluene (C7H8), with the structural formula

H
H
C
H C C C H
H C C H H
C
H

saturated ■ Just because the aliphatic compounds listed thus far are all saturated, or lacking double
Hydrocarbons that bonds, does not mean that all aliphatic compounds are in this category. For example, the
have no double or
gas ethene (not ethane, which is saturated) is C2H4, written structurally as
triple C—C bonds.
H H
C C
H H
24 Chapter 2 The Elementary Chemistry of Combustion
while propene (C3H6) is
H H
H
H C C C
H
H
Another category of unsaturated hydrocarbons involves triple bonds between adja-
cent carbons. Such compounds are usually designated by the suffix -yne as in pentyne (or
1-pentyne to designate the carbon atom to which the triple bond attaches):
H H H
H C C C C C H
H H H
The most common triple-bond hydrocarbon encountered in fire investigation is ethyne
(C2H2), or, as it is more commonly known, acetylene:
H C C H
Another class of hydrocarbons commonly encountered in fuels has rings consisting of
five, six, or seven carbons but without the combination of single and double bonds that
characterize the aromatics. These cycloparaffins, as they are called, appear as in cyclo-
hexane (C6H12):
H H
H H
C
H C C H
H C C H
C
H H
H H
See Appendix B for additional hydrocarbon nomenclature and structure.

PETROLEUM PRODUCTS
All hydrocarbons are good fuels, but very few are ever used commercially as pure com-
pounds. In fact, it is very expensive to isolate any single compound from petroleum or
coal tar mixtures in a pure state. Once isolated, the pure compounds might not have the
desired physical or chemical properties and would have to be blended with another com-
pound in any event. Virtually all commercial fuels associated with fires are mixtures of
relatively large numbers of individual compounds but with sufficient similarity in chemi-
cal structure that their combustion behavior may be quite similar.

Petroleum Distillates
Petroleum products are first separated from one another by heating raw petroleum and col-
lecting the vapors as they distill off at various temperatures. Such products are labeled
petroleum distillates. Each of these fractions or cuts will contain a mixture of all the com-
pounds that boil off between two set temperatures. For example, petroleum ether is a cut of
distillates that boil off at temperatures between 35°C and 60°C (95°F–140°F), while
kerosene contains those that boil between 150°C and 300°C (300°F–572°F).2 Most of the
compounds found in these two fractions are aliphatic straight-chain or branched alkanes,
and most are saturated—that is, with no double bonds. Some aromatics will also be present.
Cycloparaffins and aromatics have better characteristics as motor fuels, and their
proportions are enhanced through chemical processing of the crude oil by such processes
Chapter 2 The Elementary Chemistry of Combustion 25
 as cracking and reforming. They are then blended with various distillate cuts into fuels
such as gasoline with the desired combustion properties. Gasoline is not a true petroleum
distillate but rather a blended product that typically covers the boiling point range from
40°C to 190°C (100°F–375°F). Because all these common petroleum products are com-
plex mixtures of hydrocarbons, not all their components evaporate or burn at the same
volatile ■ A liquid rate. Their evaporation starts with the lightest and most volatile compounds and gener-
having a low boiling ally progresses to heavier compounds as the temperature rises (petroleum products that
point; one that is have undergone such progressive partial evaporation are sometimes called “weathered.”
readily evaporated
into the vapor state. For a discussion on the forensic impact of weathered petroleum distillates commonly found
in fire scene debris, see Chapter 14 and the work by Stauffer, Dolan, and Newman3).
During combustion, the process has the same progression but occurs at a much faster
rate, as radiation from the flames causes the temperature of the liquid pool of fuel to rise.
As a result, the residues of partially burned petroleum products found in fire debris have
different chemical and physical properties (vapor pressure, flash point, specific gravity,
viscosity, etc.) from those of the original fuel. At the same time, the flammability proper-
ties of the fresh or unburned fuel are determined in large part by the lighter, more easily
vaporized components of the mixture. Winter-blend gasolines, for instance, may contain
6 to 10 percent by weight methyl butane and 5 to 6 percent n-pentane, while summer-
blend gasolines may contain 4 to 8 percent methyl butane and 4 to 5 percent n-pentane
to maintain the same bulk properties of vapor pressure and flash point at very different
temperatures.4 The physical and chemical properties of many of these liquid fuels will be
discussed in more detail in Chapter 4.
Hydrocarbons with carbon backbone (chain) lengths greater than C24 are waxes at
room temperatures. In general, the longer the chain length, the higher the melting and
boiling points of the compound will be. Mixtures of various high boiling point hydrocar-
bons will be encountered as petrolatum (liquid or petroleum jelly) or paraffin wax. The
heaviest (longest chain) hydrocarbons are found in asphalt.

Nondistillates
A number of new petroleum products are not true petroleum distillates but are used in a
variety of consumer and industrial products that will be encountered in fire debris. One
class of compounds consists of blends of isoparaffinic (branched hydrocarbon chain)
compounds, which have good solvent properties but little of the odor commonly associ-
ated with petroleum solvents. They are used in insecticide solvents, hand cleaners, lamp
naphthenics ■ Class of oils, and some lighter fuels. Another class of petroleum products, called naphthenics, has
hydrocarbons based on had the aromatic and paraffinic components stripped out (by catalytic reactions or
cycloalkanes. absorption/elution methods). These products consist largely of cyclohexane and cyclohep-
tane with various substitutions. They have found very wide use in solvents and torch
fuels. Others are blends of aromatics in various volatility ranges used in insecticides and
adhesives for their specific solvent properties. There are also products that consist entirely
of n-alkanes (sometimes just one or two, such as n-dodecane and n-tridecane) intended
for use in liquid fuel candles. All these product types constitute possible volatile fuels but
are synthesized products and not true petroleum distillates. Their identification and char-
acterization will be discussed in more detail in Chapter 14.

CARBOHYDRATES
Perhaps the most important type of organic compound concerned in the study of combus-
tion processes is in a different group, known as carbohydrates. These make up the bulk
of wood, which is the most common fuel of structural fires. They differ from the hydro-
carbons in very significant ways. Here the molecules are very large and complex. More
significantly, they contain a relatively high content of oxygen; that is, they are already par-
tially oxidized. The process of burning wood is simply a completion of the oxidation that
started in the natural synthesis of the fuel itself.
26 Chapter 2 The Elementary Chemistry of Combustion
 Carbohydrates are so named because their chemical formulas include the elements
carbon, hydrogen, and oxygen in multiples of the simple formula

— CH2O —

which is a combination of a carbon with water (hydrate). Actually, the simplest carbohy-
drate has the formula
C6H12O6

which is the empirical formula for several sugars, of which the most common are glucose
(or dextrose), the sugar found in the blood; and fructose, which is obtained from grapes and
other fruit. Cellulose, the major constituent of wood, is made up of many units of glucose
attached to one another in chains. [This process takes place when each glucose molecule loses
a water molecule (the reaction is represented by the formula C6H12O6 S C6H10O5 ⫹
H2O), and the molecules link up together at the open “holes” left in the structure.] Because
cellulose is simply an infinite replication of glucose units (a natural polymer), the main reac-
tion that occurs on burning cellulose is the same as that for burning glucose:

Glucose: C6H12O6 ⫹ 6O2 S 6CO2 ⫹ 6H2O

Cellulose: C6H10O5 ⫹ 6O2 S 6CO2 ⫹ 5H2O

Ordinarily, as with other fuels, not all the carbon is oxidized to carbon dioxide. In
most fires, less oxygen is available and some CO (carbon monoxide) is produced instead
of CO2. Note that the hydrogen originally present is not as effective a fuel as it is in
hydrocarbons, since it has already been partially oxidized in forming the original mole-
cule. This partial oxidation accounts in part for the fact that wood fires do not readily
achieve the high heat output of many other fuels. heat of combustion ■
Fats are a form of carbohydrates and can be significant fuels in some fires. Whether The quantity of heat
released from a fuel
they are vegetable oils (such as cotton, corn, or linseed) or animal fats, they are carbohy-
during combustion,
drates (called fatty acids) with the characteristic molecular structure that they are straight- measured in kilojoules
chain or branched hydrocarbons at one end and —COOH groups at the other, and are, per gram or Btus per
in turn, substituted onto a glycerol backbone, as shown here: pound.

H
H C OH H H H H H H H H H H H H H H
H H H
H C OH H C C C C C C C C C C C C C C COOH
C C C
H C OH H H H H H H H H H H
H H H
H
Glycerol (Glycerin) Linoleic acid (C18H32O2)

Because of their largely carbon–hydrogen makeup, fats can burn readily and yield sig- joule ■ The SI unit of
nificant amounts of heat. measurement of heat
The amount of heat that any fuel can produce when burned is called its heat of com- energy (1 J ⴝ 0.238
bustion, and it is a measure of the number of joules or Btus (British thermal units) of heat cal ⴝ 0.00095 Btu).
released per kilogram (or pound) of fuel. The heat of combustion is a means of estimat- Btu ■ British thermal
ing how much total heat a given amount of fuel can produce; it is not a measure of the unit. A standardized
rate at which the heat can be released, or of the flame temperature a fuel will produce. measure of heat, it is
(See Table 4-5 for some typical heats of combustion.) Heat of combustion is measured by the heat energy
required to raise the
calorimetry techniques. Heat of combustion is represented by the symbol ΔHc, where Δ temperature of 1 pound
(delta) represents change, H represents heat (enthalpy), and the subscript c denotes combus- of water 1 degree
tion. Other constituents of wood, such as resins, not only have a higher heat of combustion Fahrenheit.

Chapter 2 The Elementary Chemistry of Combustion 27

 than cellulose, but their chemistry produces a higher flame temperature. Therefore, resinous
woods (such as pine or fir) are more likely to produce hotter fires than are nonresinous woods
such as birch or balsa.

PYROLYSIS OF ORGANICS
Except for the simplest hydrocarbon fuels that merely need to evaporate to produce mole-
cules simple enough to combine directly with oxygen in a flame, all fuels have molecules
that have to be broken into small enough “pieces” to undergo combustion. The primary
effect of heat on wood or other solid fuel is to decompose or pyrolyze it. The words
pyrolyze or pyrolysis stem from the Greek words pyro (meaning fire) and lysis (meaning
pyrolysis ■ The chemi- decompose or decay). Therefore, pyrolysis can be defined as the decomposition of a mate-
cal decomposition of rial into simpler compounds brought about by heat. Pyrolysis of wood, for example, yields
substances through the
action of heat, in the ■ burnable gases such as methane;
absence of oxygen. ■ volatile liquids such as methanol (methyl alcohol) in the form of vapors;
combustible ■ A mate- ■ combustible oils and resins, which may be as vapors of their original forms or
rial that will ignite and pyrolyzed from more complex structures; and
burn when sufficient ■ a great deal of water vapor, leaving behind a charred residue, which is primarily
heat is applied and
when an appropriate
carbon or charcoal.
oxidizer is present. The gases and vapors generated diffuse into the surrounding air to burn via flaming com-
flaming combustion ■ bustion. The charcoal then burns with a considerably greater heat output on a weight-
Rapid oxidation of for-weight basis than does the original wood, since the noncontributing oxygen has been
gases and vapors that driven off, largely as water. The carbon combines directly with the oxygen in the air in
generates detectable
heat and light.
contact with it to support glowing combustion. (From Table 4-5: the ΔHc of wood is
around 16 kJ/g, while that of charcoal is over 34 kJ/g.)

CONCLUSIONS ABOUT ORGANIC COMPOUNDS

It is apparent from the preceding considerations that fuels differ not only in their chemi-
cal nature but also in their mode of oxidation, the heat output of the reaction, and in
other respects. They also have different characteristics with respect to the temperature at
which they ignite, the quantities of oxygen used per unit of fuel, the rate at which they
will oxidize, and other properties related to their physical state. Some of these character-
istics will be discussed in Chapters 3, 4, and 5 of this text as they relate to the combus-
tion of gaseous, liquid, and solid fuels.

State of the Fuel

Solids, liquids, and gases are the three physical states of matter of interest to the fire inves-
tigator. The state in which a fuel exists must be related to its other combustion proper-
ties, as we shall see in Chapters 3 and 4.
A solid is a substance whose molecules are held in a fixed, three-dimensional relation-
ship to one another by molecular forces. Therefore, a solid has fixed volume and shape.
Some solid fuels (like candle wax) melt to a liquid (which then evaporates to form a
vapor). A few solids (naphthalene) evaporate directly to a vapor state (a process called
sublimation). Many solids do not melt but decompose by pyrolysis.
A liquid has “looser” bonds that allow the molecules to move about but not escape.
Thus, a liquid has no fixed three-dimensional shape but still retains a fixed volume.
Gasoline, like every other liquid fuel, itself does not burn when in the liquid state. Such
liquids are, however, easy to vaporize, and the vapor that is formed burns like any gas—
by mixing with oxygen and combusting as a flame. Note that because the molecular
bonds in liquids are typically not as strong as those in solids, less heat is required to break
these bonds in liquids compared with those in solids.
28 Chapter 2 The Elementary Chemistry of Combustion
 The molecules in a gas have only very weak forces between them because the mole-
cules are generally very far apart (depending on temperature and pressure). Thus, a gas
will expand to fill any available volume and diffuse (more or less) freely in air. There are
relatively few gaseous fuels at room temperature. However, many materials become
gaseous at the temperatures that occur in a fire, and virtually all materials can be con-
verted to gases by applying sufficiently high temperatures. Another hydrocarbon, paraf-
fin wax, requires much higher temperatures first to melt, then to vaporize it to a form
whose vapors will burn in the familiar form of a candle flame.

SIGNIFICANCE OF STATE OF FUEL

The classification of liquids, solids, and gases is far from absolute. They are to a large
extent interchangeable, and the conversion from one state to another is a function of the
temperatures and pressures that exist during a particular fire. The fire investigator must
keep in mind that a change in physical state (such as from a bulk liquid to a finely divided
aerosol when hydraulic fluid leaks under high pressure) can dramatically change the
ignitability of the fuel and the size of the resulting fire. In general, the physical state of the
fuel controls, in many ways, the ignitability of the fuel, its burning rate once ignited, and
the quantity and nature of combustion products that result.

DIFFICULTY IN CLASSIFYING SOME HYDROCARBONS

Among the hydrocarbons are a number of materials that are difficult to classify as gases
or liquids. The liquid petroleum gases, for instance, of which propane and butane are the
prime examples, can exist briefly in both states at ordinary temperature and pressure.
Because they are so very volatile, when exposed they readily vaporize, leaving no liquid
residue. Propane, with a boiling point of -42°C, vaporizes much more rapidly than
butane, which is heavier (longer carbon chain) and has a higher boiling point (-0.6°C).5

GASES AND THE IDEAL GAS LAW

Hydrogen, methane, ethane, propane, and acetylene are the most common gaseous fuels. Like other
gases, these gases also obey (usually) the ideal gas law, which relates the amount of gas to its pressure,
volume, and temperature.
The ideal gas law relationship is

pV ⫽ nRT

Where:

p ⫽ pressure
V ⫽ volume
n ⫽ number of moles of gas
T ⫽ temperature (absolute temperature, measured in Kelvin)* absolute temperature ■

R ⫽ universal gas constant (to make all the units come out right!) Temperature above
absolute zero (in K
Increasing the temperature increases the volume if pressure is held constant or increases pressure or °R).
if the volume is held constant. It is the heat-driven expansion of gases (or liquids) that gives rise to buoy-
ancy. The same number of molecules occupying a larger volume produce a lower density.
If the amount of gas present is measured in moles (one mole of any material is equal in weight to its
molecular weight and always contains the same number of molecules (Avogadro’s number, or 6.023 ⫻ 1023).
One mole of oxygen (O2) then weighs 32 g, and one mole of butane (C4H10) weighs 58 g. At standard tem-
perature and pressure [0°C, 760 mmHg (1 atm)], one mole of gas will always occupy 22.4 L (liters).

*Temperature above absolute zero (0°C ⫽ 273 K). See Chapter 3, Table 3-2.

Chapter 2 The Elementary Chemistry of Combustion 29

 When vapors such as these are placed under pressure, they condense to form liquids. (By
definition, a vapor can be condensed to a liquid by the application of pressure, where a
gas is above its critical temperature and pressure and cannot be condensed by pressure
alone.) Commercial-grade butane and propane [and the mixture often called liquefied
petroleum gas (LPG)] are stored and transported in tanks, largely in liquid form. On
release of the the pressure, as when the overlying gas in the tank is fed to a burner, more
of the liquid vaporizes and maintains the same pressure (critical pressure) in the tank until
all the fuel has evaporated.

SOLIDS
Solid fuels, discussed in greater detail in Chapter 5, as a rule do not burn as solids (there
are a few exceptions). When a solid is burning, a portion of its surface may be smoldering,
even glowing, yet a nonglowing solid fuel may be surrounded by flames, and we say that
it is burning. This is not strictly true, for the flames are the result of gas–gas combustion,
so it is the gaseous pyrolysis products that are mixing with the air to produce the flames
and not the solid fuel. The smoldering phase generally follows when the thermal decom-
position of the solid has slowed to the extent where some oxygen can reach the hot solid
surface. (Very porous solid fuels may allow diffusion through the fuel as well.) If the sur-
face is hot enough, a gas–solid interaction can take place that we observe as a glowing
combustion. In a wood fire, the turbulence of the flames or the presence of a strong draft
often makes it possible for glowing and flaming combustion to occur at the same time on
the same piece of fuel.
Very finely divided solid fuels (such as wood dust or grain dust) can burn very
quickly, even to the point of creating explosive forces, because the small volume of the
individual particles permits rapid heating and decomposition into combustible products,
and the large surface area permits rapid mixing with surrounding air.

LIQUIDS
A variety of liquid fuels exist. Examples include gasoline, pools of liquefied plastics, paint,
organic heat-transfer fluids, and oils used in cooking and manufacturing processes. Of
these liquid fuels, gasoline and pooled liquid plastics are of particular interest to the fire
investigator.
Gasoline, being much more complex and having heavier molecules, has a higher
boiling point (actually a range, as described earlier) and is mostly a liquid at ordinary
temperatures and pressures. Thus, it is not necessary to seal it in a tank to retain it as
a liquid. A small opening or vent to the exterior of the tank allows only a minor escape
of the most volatile portions of the gasoline as vapors. When gasoline combusts in the
cylinders of an automobile engine, most of it is actually vaporized in the carburetor (or
by the action of injecting it at high pressure), and the remainder is vaporized in the hot
cylinder, which means it undergoes a totally gaseous reaction of combustion. With a
cold motor, in which some of the gasoline may not vaporize, the portion that remains
liquid is not combusted and tends to escape past the pistons into the crankcase. If a
gasoline engine is exposed to conditions below the flash point of the fuel, say -45°C (-
50°F), it will not start until it is externally heated to the point where the fuel can pro-
duce an explosive mixture of vapor and air in the cylinders (or where mechanical
dispersion by an injector system can create the necessary vapor pressure). Such inter-
pretations may appear to be elementary, but if overlooked in the interpretation of a fire,
they can lead to incorrect conclusions.
Plastics that melt and form pools of liquefied fuel as they burn will be discussed in
more detail in Chapter 5.

30 Chapter 2 The Elementary Chemistry of Combustion

 CHAPTER REVIEW

Summary
Although fire is a complex phenomenon, it relies on processes can be reduced to such simple reactions will
only a handful of basic chemical reactions. The oxida- be better prepared to examine the nature of the combus-
tion of carbon, hydrogen, and sulfur accounts for the tion reaction in more depth. Combustion—or, more
largest volume of combustion products of common precisely, the processes of oxidation, heat transfer, and
fuels. Investigators who understand that virtually all fire fuel chemistry—will be treated in succeeding chapters.

Review Questions
1. The smallest unit of an element that takes part in 6. What is the heat of combustion of a fuel?
a chemical reaction is a/an ________. 7. Name four classes of petroleum products.
2. What natural element is necessary for normal 8. What chemical structure is characterized by the
combustion? designation alkane?
3. What three compounds are produced by the 9. What chemical structure is characterized by the
combustion of hydrocarbon fuels? designation aromatic?
4. What is the difference between hydrocarbons and 10. Define an organic compound.
carbohydrates? 11. According to the ideal gas law, if the temperature of
5. What is the difference between glucose and a fixed amount of gas rises from 300 K (27°C) to
cellulose? 600 K (327°C), what will be its change in volume?

References
1. W. C. Gardiner, “The Chemistry of Flames,” Scientific dissertation, Strathclyde University, Glasgow, Scotland,
American 246, no. 2 (1982): 110–25. 1995) (Ann Arbor, MI: University Microfilms, 1996),
2. ASTM D3699: Specification for Kerosene (West 11, 288–89.
Conshohocken, PA: American Society for Testing and 5. R. H. Perry and D. Green, Perry’s Chemical Engineers’
Materials, 2008). Handbook, 7th ed. (New York: McGraw-Hill, 1997),
3. E. Stauffer, J. A. Dolan, and R. Newman, Fire Debris table 3-1.
Analysis (Burlington MA: Academic Press, 2008).
4. J. D. DeHaan, ”The Reconstruction of Fires Involv-
ing Highly Flammable Hydrocarbon Liquids” (PhD

Chapter 2 The Elementary Chemistry of Combustion 31

 CHAPTER
3
Fundamentals of Fire
Behavior and Building
Construction

Courtesy of Jamie Novak, Novak Fire Investigations

and St. Paul Fire Dept.

KEY TERMS

ambient, p. 35 flame spread, p. 40 overhaul, p. 40

backdraft, p. 41 flammable liquid, p. 36 plume, p. 35
buoyancy, p. 35 flashback, p. 41 radiation, p. 47
conduction, p. 45 flashover, p. 49 self-heating, p. 40
convection, p. 46 fuel load, p. 54 smoldering combustion, p. 33
entrainment, p. 35 glowing combustion, p. 38 stoichiometry, p. 39
fire-resistive, p. 73 heat release rate, p. 43 vented, p. 64
firestorm, p. 47 heat transfer, p. 45 ventilation, p. 38
flameover, p. 56 incipient, p. 54 watt, p. 43

OBJECTIVES

After reading this chapter, you should be able to:

■ Describe the four conditions that must exist for a fire to occur.
■ Recognize the details involved in the typical flaming combustion of organic fuels.
■ Describe several states of combustion.
■ Explain the three basic methods of heat transfer.
■ Calculate the visible flame height given the heat release rate and effective diameter
of a fire.
■ Discuss the four phases of fire development.
■ Explain the effects of environmental conditions on fire.
■ Distinguish between Type III (ordinary) construction and Type V (wood frame)
construction as seen in single-family residences in the United States.

32
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

F
ire is an exothermic oxidation reaction that proceeds at such a rate that it generates
detectable heat and light. As indicated in the previous chapter, two types of combustion
(or fire) are (1) flaming and (2) smoldering. Flaming combustion, as defined in Chapter 2, is
a gaseous combustion in which both the fuel and oxidizer are gases. Smoldering combustion smoldering combustion
■ The direct combina-
involves the surface of a solid fuel with a gaseous oxidizer (usually the oxygen in air). Nearly
tion of a solid fuel with
all destructive fires are flaming combustion. The smoldering fire is not uncommon, either atmospheric oxygen to
alone or in combination with flames. For example, the nonflaming fire in a mattress or a pile generate heat in the
absence of gaseous
of sawdust is a good illustration of a smoldering fire (which can do extensive damage), as is the flames; see glowing
charcoal fire used for barbecues. Many destructive flaming fires began as small smoldering combustion.
fires.
The differences between flaming and smoldering combustion are a result of the nature
and condition of the fuel, as well as the availability of oxygen, all of which influence the rate
at which heat is being produced. If the smoldering mattress or sawdust is stirred up, it may
develop into a flaming fire because more fuel is brought into contact with oxygen. On the
other hand, the glowing charcoal has little or no flame because the compounds that could
be volatilized into gases (and thereby contribute to an open flame) were lost at the time the
original fuel was charred or coked. The differences between a wood fire and a fire of charcoal
made from wood are discussed later, in the section dealing with the properties of fuels.

Basic Combustion
For a fire to occur, the following conditions must exist:
■ Combustible fuel must be present.
■ An oxidizer (such as oxygen in air) must be available in sufficient quantity.
■ Energy as some means of ignition (e.g., heat) must be applied.
■ The fuel and oxidizer must interact in a self-sustaining chain reaction.
The first three elements listed have long been described as the fire triangle, but the
fourth must also be present if the fire created is to be continuous (self-sustaining), thus
creating what is called the fire tetrahedron1 (see Figure 3-1). Although these requirements
appear obvious, it is true, nevertheless, that statements have been made in court describ-
ing fires burning on bare concrete floors devoid of fuel. Absolute absence of any source
of heat is sometimes claimed. In general, disregard of the simple requirements just listed
is not at all uncommon. It must be remembered that removal of any one of the four ele-
ments results in extinguishment, and a careful analysis of all the factors is required to
establish what happened, since mechanisms of ignition, combustion, and even sources of
fuel may not be obvious at first glance. Some are transient and require careful analysis of
possible mechanisms before their presence and effects can be reasoned out.
Chapter 3 Fundamentals of Fire Behavior and Building Construction 33
 FIGURE 3-1 The fire
tetrahedron: fuel, oxidizer,
and heat interacting in a
self-sustaining chemical
reaction.

HEAT
OX
YG EL
EN FU

CHAIN
REACTION

Experience with the common gaseous, liquid, and solid fuels is so widespread that
misjudgments of their contributions to a fire would not be expected. However, there are
many less common materials (e.g., plastics and metals) with which experience is less wide-
spread. In addition, the mere presence of a suitable fuel in conjunction with the other con-
ditions does not guarantee that a fire will result. The fuel must be present in a suitable
physical state to be ignited. For example, common furnace oil is an excellent fuel but
when spread on a concrete slab, it generally resists every effort to ignite it, even with a
blowtorch. If that same oil is soaked into a cloth wick, it is easily ignited. Another exam-
ple of the importance of the fuel’s state is newspaper, which is so helpful in starting fires
but can be used to extinguish some small fires. Proper interpretation of a fire situation
requires knowledge of the properties of the fuel, its physical state, and the character and
dynamics of the fire itself. Thus, the knowledge of availability and suitability of fuels in
specific instances may require more than everyday experience.

Flaming Fire
A fire characterized by flame is the most common type. Here the flame actually is the
fire—the production of gaseous reaction products with the evolution of heat and light.
The color of the light emitted is determined in part by the elements in the reacting mix-
ture. The color emitted by a hydrocarbon gas burning well mixed with air (e.g., a cor-
rectly adjusted oxyacetylene torch) is clear blue. The gaseous flame is made more visible
when carbon and other solid or liquid by-products resulting from incomplete combustion
are raised to incandescent temperatures and glow—red, orange, yellow, or white, depending
on their temperature. The reactions of a flaming fire are, under ordinary circumstances,
oxidative, with oxygen from the air as the oxidizer. An important fact to remember is that
there cannot be a flaming fire unless a gas or vapor is burning. This holds true whether
the fuel is a gas to begin with or a vapor evaporated from a liquid or distilled or driven
off from a solid. The flame is a totally gaseous reaction. Liquid fuels can produce only
flaming fires, since liquids themselves do not burn and do not pyrolyze to leave a solid
char (except in unusual circumstances) to support smoldering.
Solid fuels often produce predominantly flaming fires, although they are frequently
accompanied by glowing fire. Wood decomposes readily under the influence of heat to
generate combustible gases; therefore, it tends to burn in flaming combustion, especially
in the early stages of the fire. Later, the charcoal formed will continue to combust as a smol-
der fire. Coal, in contrast, has less volatile components and tends to give off proportionately
34 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 C FIGURE 3-2 Typical
H2O Buoyant smoke plume flaming combustion of
organic fuel showing
HCN CO2 decomposition region
HCN where volatized fuel
decomposes to simpler
H 2O species before combustion,
CO CO2 and intermediate products
are formed.
CO2
H 2O Entrainment
(mixing, diffusing, and cooling)
C CO2

CO2
CO H 2O
H2
H2
C
CH3OH CH
CH HCN CO
CH3 C
CHO
Radiant heat CN CH O Radiant heat
C 2
CH4
H2 Flame plume
C
CH
CO
CH CH3
C CH C C
2 CH
CH2
CHO H
H CH2
CH3 CH CH2
C CH3
O2
C H CH2O O2
CH Decomposition region H
Air N2 N2 Air
Vaporized fuel
Entrainment
Pyrolysis zone (supplying O2)

Fuel

smaller quantities of combustible gases than wood and a lesser amount of open flame.
Figure 3-2 illustrates the dynamics of flaming combustion of a solid or liquid. Note that
heat generated by the flame radiates onto the surface of the fuel, raising its temperature,
evaporating it if it is a liquid, and pyrolyzing it if it is a solid. [The equilibrium surface tem-
perature of a solid fuel is controlled by its chemistry. For most fuels it is on the order of
350°C to 500°C (660°F to 930°F)].2 The molecules of vapor boiled out of the liquid or
pyrolyzed out of the solid are dissociated or torn apart by the chemical conditions within
the flame (sometimes called reducing conditions). They can then readily combine with oxy- plume ■ The convective
gen coming from the outside. The plume of hot gases (and solid [soot] or liquid [aerosols] column of hot gases
products from incomplete combustion) rises by buoyancy (being hotter and therefore less generated by a flame.
dense than the ambient air in the room). This movement draws surrounding air and its buoyancy ■ Tendency
oxygen into the flame, entraining and mixing it with the fuel gases, thereby maintaining or ability to rise or float
the reaction. Entrainment is the drawing of air (or other gases) into the buoyant plume. in air or liquid as a result
If a flaming fire is confined to a closed room where the oxygen being used in the flames of a difference in density.
is not replaced, the oxygen concentration in the surrounding air will fall from its normal ambient ■ Surrounding
level of about 20.9 percent. The rate at which the concentration of oxygen drops depends conditions.
on the size of the fire, the size of the room, and the amount of fresh air that can leak into
entrainment ■ The
the room from any small openings. Generally speaking, when the oxygen concentration
mixing of two or
falls below about 15 percent in the vicinity of the flame, the combustion rate of ordinary more fluids as a result
combustibles begins to decrease. Eventually, the concentration reaches a limiting oxygen of laminar flow or
concentration that will no longer support flaming combustion, and the flames will die out. movement.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 35

 The oxygen limit that will actually extinguish the flames is dependent on the nature of the
fuels involved and on the temperature of the combustion gases—the higher the tempera-
tures produced, the lower the oxygen concentration can drop and still support flames. For
instance, oxygen concentrations between 5 and 8 percent have been observed in room tests
flammable liquid ■ A with flammable liquids where the ceiling temperatures were on the order of 900°C to
liquid having a flash 1,000°C; and in post-flashover room fires, oxygen concentrations between 0 and 5 percent
point below 38°C
have been observed when temperatures in the hot gas layer were over 1,000°C.3
(100°F).

Structure of Flames
Most of the flames we encounter in fire investigation are diffusion flames. This means
that the gases or vapors that are supporting the flame diffuse outward (or upward) from
the surface of the fuel and the oxygen for combustion diffuses toward the fuel from the
surrounding air. If the concentration of fuel and the concentration of oxygen are plotted
as a function of distance from the fuel surface, a plot like Figure 3-3c results. At some distance,

Vertical
fuel
(solid)
Pool fire

Oxygen
Flames Radiant in air
heat
Radiant Vapors
heat from fuel

Fuel
Oxygen
in air
(b)

Flames
Fuel Vapors generated
surface by pyrolysis of
fuel surface
(a)

21%
Saturation
Concentration of

Concentration of
oxygen (%)
fuel vapors

Combustion zone

0
0 Distance from fuel surface

(c)

FIGURE 3-3 (a) Diffusion of fuel vapor away from vertical fuel surface as oxygen diffuses toward fuel surface.
(b) Diffusion of fuel vapor upward from horizontal fuel surface. (c) Plot of concentration of O2 and fuel vapor
as a function of the distance from the fuel surface.

36 Chapter 3 Fundamentals of Fire Behavior and Building Construction

the concentrations of fuel vapors and oxygen are correct, the mixture can be ignited, and
a flame can be maintained in that zone as long as the rate of fuel generation is maintained.4
(The heat from the flame radiating onto the fuel surface helps maintain that supply from
liquid or solid fuels.) The rate at which fuel is delivered to the flame is dependent on the
total heat flux reaching the fuel surface and the latent heat of vaporization (for liquids)
and latent heat of gasification (for solids). In combustion explosions, the fuel vapors and
air are mixed together and then ignited in what is called a premixed flame. In gas appli-
ances, this mixing is created by the gas jet. If properly adjusted, the gas is fed into the
combustion zone at a rate and at a concentration at which nearly complete (and con-
trolled!) combustion occurs in the desired location.
A small flame such as that of a candle is called a laminar flame, because the zones
of high fuel concentration, dissociation (reducing), mixing, and combustion are well
defined in layers surrounding the candle wick. The airflow around the flame is such
that an envelope of very high temperatures is maintained around the source of fuel. In
a candle, this high-temperature zone allows the combustion of the soot created by the
pyrolysis (described in Chapter 2) of the wax to take place at very high temperatures
(with the soot particles radiating visible light [incandescence] in the yellow-white colors
of very high temperature surfaces). By the time the soot escapes the high-temperature
zone it has been (almost) completely consumed and turned into CO2 (as in Figure 3-4).
If the flame becomes too large, the smooth flow of the air cannot be maintained, or if
the envelope is disturbed by a sudden draft, this, in effect, tears open the high-temperature
zone, allowing some of the soot and other intermediate products to escape unburned.
The presence of the ragged tips of flames like these are a sign that the combustion has

Soot-formation region
(luminous)

Inner edge of
diffusional combustion
1200°C
1400°C
1200°C
800°C
200°C
1000°C
Hydrocarbon
cracking region
(dark)
800°C
10 mm

Outer edge of
diffusional combustion
500°C

Pseudo-premixed
combustion
Molten wax
(blue luminosity)
(C20H42)

20 mm

(a) (b)

FIGURE 3-4 (a) Typical laminar flame of candle showing blue luminosity at base. Courtesy of John D. DeHaan.
(b) Temperature distribution in laminar candle flame showing a zone of very high temperature. Source: From Fire,
H. Rossotti, St. Anne’s College, Oxford University, United Kingdom.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 37

FIGURE 3-5 Gasoline
pool fire on plywood
produces a typical
turbulent flame plume.
Courtesy of Jamie Novak,
Novak Fire Investigations and
St. Paul Fire Dept.

become turbulent (as in Figure 3-5). Turbulent combustion dominates nearly all fires,
glowing combustion ■ causes the flicker that we associate with most fires, and in some cases can be extremely tur-
The rapid oxidation of bulent.
a solid fuel directly with
atmospheric oxygen
creating light and heat Smoldering Fire
in the absence of flames.
A smoldering fire is characterized by the absence of flame but the presence of very hot
ventilation ■ A tech- materials on the surface of which combustion is proceeding. If the temperature of that sur-
nique for opening a
burning building to
face is high enough [500°C (900°F) or higher], a visible glow or incandescence can be seen.
allow the escape of As we shall see later, the color of the incandescent glow on the surface is related to its
heated gases and smoke temperature, as is also true of the airborne particles in the flame. Some visible colors and
to prevent explosive their related temperatures are illustrated in Table 3-1. (Note that these are not the temper
concentrations (smoke colors visible on polished metal surfaces after heating and cooling.) “Glowing” and “smol-
explosions or backdrafts)
and to allow the
dering” are often used interchangeably [although Babrauskas describes glowing combustion
advancement of hose as a non-self-sustaining condition (usually as a result of forced ventilation), while smolder-
lines into the structure. ing is self-sustaining solid gas combustion].5

TABLE 3-1 Visual Color Temperatures of Incandescent Hot Objects

APPROXIMATE APPROXIMATE
COLOR TEMPERATURE (°C) TEMPERATURE (°F)
Dark red (first visible glow) 500–600 930–1,100
Dull red 600–800 1,110–1,470
Bright cherry red 800–1,000 1,470–1,830
Orange 1,000–1,200 1,830–2,200
Bright yellow 1,200–1,400 2,200–2,550
White 1,400–1,600 2,550–2,910

Source: Data taken from C. F. Turner and J. W. McCreery, The Chemistry of Fire and Hazardous Materials (Boston:
Allyn and Bacon, 1981), 90. See also D. D. Drysdale, An Introduction to Fire Dynamics, 2nd ed. (Chichester, UK:
Wiley, 1999), 53.

38 Chapter 3 Fundamentals of Fire Behavior and Building Construction

 A charcoal or coal fire is a good example of smoldering combustion. In the past it was
common to see a blacksmith heating iron objects in a fire made with coal through which
a forced-air draft passed. These fires showed very little flame, but the glowing fire induced
by the air draft was intensely hot, more so than most flaming fires. Similar situations are
sometimes seen in the deep piles of hot coals left on collapse of large wood-frame struc-
tures. When there is sufficient draft through the coals and a layer of insulative ash to retain
the heat, the temperatures reached can melt copper, steel, silver, and even cast iron, pro-
ducing unusual artifacts. The duration of such high temperatures in glowing coals can be
considerably longer than the high temperatures reached during the open-flame stages of the
fire, making it possible to melt even massive solid objects.
Because the smoldering fire involves two phases, a solid and a gas, the ratios of oxidizer
and fuel are not of direct concern, although the chemistry of the reaction still determines the
total consumption of each. The proportions of fuel, oxygen, and final products are called
the stoichiometry of the reaction. Perfectly balanced chemical reactions, which result in the stoichiometry ■ Balance
complete reaction of all starting materials with no waste, are called stoichiometric or ideal of chemical reactants
reactions. Although glowing/smoldering combustion is often considered the end (decay) and products.
stage of a fire, it can occur at any time during a fire depending on the ventilation conditions.
Figure 3-6 shows glowing combustion on the surface of a burning wood timber where the
incoming draft is so strong it is pushing the vapors generated from the wood aside and forc-
ing very rapid glowing combustion of the surfaces. The vapors burn as flames to the sides
and rear of the timber, where they can mix with air.
Smoldering combustion can continue even in an oxygen-deficient atmosphere (below
5 percent), in part due to the very high temperatures at the oxidizing surface. In the com-
bustion of ordinary solid fuels, a transition from flaming to smoldering may be seen as
the oxygen level drops in an enclosure, which can revert back to open flames if a fresh
supply of oxygen is introduced. In this instance, unlike in a flaming fire, the fuel and oxidizer

FIGURE 3-6 Glowing

combustion on timber
with a very strong draft
pushing combustion gases
away from the burning
surface. Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 39

 are not premixed or even mixed at the moment of combustion but react as solid surfaces
over which the oxidizing gas passes. Inasmuch as fuels are generally organic materials that
are often reduced to charcoal (or coke) as the volatile decomposition products are distilled
out and burned, most fires reach a smoldering state in their later phases. It must be
remembered that the smoldering material here is not the original fuel but a derived fuel
resulting from the earlier portions of the fire.
An important feature of the glowing or smoldering fire—especially when there is a
forced draft, as in the blacksmith’s forge—is that the limited flame that accompanies it
results chiefly from the burning of carbon monoxide (CO) to carbon dioxide (CO2). The
carbon monoxide is generated by incomplete combustion of carbon due to insufficient
oxygen and is further oxidized with a hot, but rather small, generally bluish, flame.
In the investigation of most fires, the smoldering phase generally represents either the
incipient stage or the last part of the fire process. As an incipient fire, smoldering may be
very prolonged and may do a great deal of damage before flames occur. The investigator
and indeed the firefighter may be much more concerned with flaming fire and its effects
overhaul ■ The fire- in the earlier stages of the fire. Much of the post-fire cleanup or overhaul by firefighters
fighting operation of at fire scenes, however, is directed at seeking out smoldering spots and ensuring that they
eliminating hidden do not rekindle into a destructive fire later. The investigator is primarily concerned with
flames, glowing embers,
or sparks that may
the initial small fire, whatever its nature, that grew into the large one.
rekindle the fire, usually Some exceptions to these observations must be considered. A fair proportion of flam-
accompanied by the ing fires follow a period of smoldering. When the fire is ignited in a suitable organic fuel
removal of structural or in an environment in which ventilation is very limited, it does not generate the heat
contents. release rate necessary to burst into flames. This slow combustion may proceed for an
appreciable time before the flaming fire erupts and yet do a great deal of damage. The
rate of spread of a smoldering fire may be on the order of 0.0001 m/s (0.2 to 0.25 in./min)
flame spread ■ The as compared with the 0.001 m/s to 0.01 m/s (0.04 to 0.4 in./s) typical of flame spread
rate at which flames across solid fuels.6 The conditions for transition from smolder to flame are not well
extend across the understood at this time and have not been successfully modeled.
surface of a material
(usually under specific
Smoldering fires may be produced by the ignition of latex rubber or natural fiber
conditions). (cotton) mattresses or upholstery padding in residences; hay or grain piles in barns; or
coal dust, sawdust, or other finely divided cellulosics in industrial premises. Fires that
appear to start at times when there is no obvious source or reason for ignition are often
attributed to such prolonged ignition mechanisms. Proof of such mechanisms requires
careful consideration of all the following factors:
■ initial source of ignition (there must be some source of heat to trigger the process);
■ evidence of prolonged smoldering (adequate fuel mass);
■ nature of the fuel (its ability to form a rigid porous char that will allow air (oxygen)
to percolate into the mass of fuel); and
■ cause (or sequence of events) that caused the fire to reach the open-flame state, that
is, “burst into flame.”
Stay Safe

Even if the fire never reaches flaming combustion, considerable destruction may
The extensive produc- result from the slow, long-term burning of closely packed fuels such as paper, grain, cloth,
tion of carbon monoxide
typical of smoldering
or upholstery stuffing. The extensive production of carbon monoxide typical of smolder-
fires makes them serious ing fires makes them serious life safety risks. The smoke and heat generated by even mod-
life safety risks. est smoldering fires can seriously damage other contents of a room or building.
Even in the absence of open flame, a smoldering fire is still a fire because it can
accomplish the destruction of the fuel by the thermal processes of pyrolysis and combus-
self-heating ■ An tion. All smoldering fires produce high concentrations of CO and visible smoke, since
exothermic chemical or such combustion is never complete. The mechanisms of self-heating, by which a mass of
biological process that
can generate enough
fuel can reach its ignition conditions in the absence of any outside ignition source will be
heat to become an considered in Chapter 6. A destructive smoldering fire need not be accidental. Some
ignition source. incendiary fires are started in a suitable fuel but in a restricted space where the oxygen
40 Chapter 3 Fundamentals of Fire Behavior and Building Construction
supply is deficient and flames are suppressed. Enough air may be available to sustain a
smoldering fire until the fuel is exhausted or more oxygen becomes available.
Another consequence of smoldering results from the discarding of previously flaming,
but supposedly extinguished, materials that are still combusting. Discarded matches that
are not specially treated will continue to smolder for some time after the flame is extin-
guished and can initiate a fire in a suitable fuel. The smoldering cigarette certainly falls
into this category (see later section). Hot coals, spilled from fireplaces or removed from
barbecues or other similar sources, will start larger fires if they contact suitable fuel.
These will be discussed in Chapter 6 on sources of ignition.

Explosive Combustion
Explosive combustion may not be recognized as a fire. It requires consideration here
because this type of combustion does accompany fire, often as an initiating factor and some-
times during a conventional fire when favorable circumstances develop. Because this type of
combustion is considered in a later chapter, it will be discussed only briefly at this point.
Explosive combustion can occur when vapors, dusts, or gases, premixed with an
appropriate amount of air, are ignited. Under these circumstances, the combustion that
results is not different from that which happens in the burning of these materials, except
that the premixing allows the entire combustion to occur in a very short period of time.
Thus, all the heat generation, creation of combustion products, and the expansion of
those products—which normally would require an appreciable time—become an almost
instantaneous event that is recognized as an explosion by its mechanical effects. The event
may be very forceful and produce great damage, including the blowing apart of an entire
building, or it may be so small as to produce only an audible pop.
Sometimes, the first phase of a fire is an explosion, as flammable gases or vapors have
accumulated from some source, mixed with appropriate amounts of air, and become
ignited. Such explosions are sometimes followed by flaming fire, especially if there is
residual unburned fuel after the initial combustion. As we shall see later, the brief dura-
tion of such events precludes ignition of all but the thinnest, lightest fuels to create a fol-
lowing fire. In such instances, it is the cause of the initial explosion that is the concern of
the investigator, since it is known that flammable gases were generated by some source
and that they were later ignited. It remains for the investigator to identify the source of
the fuel vapor, a suitable ignition source, and the conditions that brought them together.
When a vapor explosion occurs during a fire, it is usually an indication that a new
source of fuel was made available locally and in some quantity. The most common source
of this material is a sealed can, bottle, or drum of an ignitable liquid that breaks or bursts
because the heat of the fire has expanded the contents beyond the strength of the con-
tainer. Being heated above its normal boiling point, this fuel vaporizes immediately upon
release and forms an explosive mixture, which is then ignited when an ignition source backdraft ■ A defla-
grative explosion of
such as a flame is present. From the investigative standpoint, this event has two values.
gases and smoke from
First, it proves that the container was sealed initially and was therefore not involved in an established fire that
starting the fire. Second, it proves that an extensive fire and much heat surrounded the has depleted the oxygen
container before it burst, and this situation was the result of fire progression from else- content of a structure,
where, an unusual accident, or the work of an arsonist. most often initiated by
introducing oxygen
These facts can be of importance in the interpretation of many fires. Another cause
through ventilation or
of an explosion during a fire is a smoke explosion, also called a backdraft or flashback. structural failure.
In energetic fires in confined spaces, a great deal of fuel may be pyrolyzed into com-
bustible gases and vapors that cannot burn because there is insufficient oxygen. If fresh flashback ■ The ignition
air is suddenly admitted to the compartment (by the failure of a window or the opening of a gas or vapor from
an ignition source back
of a door) and mixed with this hot, fuel-rich mixture, a destructive combustion explosion to a fuel source (often
can occur. The dynamics of explosive combustion of gases and vapors are discussed in seen with flammable
Chapters 4 and 12. Explosions involving detonating (high) explosives are not combustion liquids).

Chapter 3 Fundamentals of Fire Behavior and Building Construction 41

 based, as they do not involve the combination of a fuel with atmospheric or chemical oxy-
gen. They occur when the explosive material itself dissociates (comes apart) and then
recombines with an extremely high rate of reaction and heat release rate. These chemical
reactions will be discussed in more detail in Chapter 12.

Heat
Aside from the significance of the chemical reactions that produce a fire, the most funda-
mental and important property of the fire is heat, which is a measure of energy. Heat ini-
tially starts a fire, and the fire produces heat. As will be discussed in the chapter on
ignition, every method by which a fire may be ignited involves the application or genera-
tion of heat, and nothing but heat is ever necessary to start the fire when the fuel and its
environment are suitable.
However, most of the destructiveness of a fire is the direct result of the heat gener-
ated. Heat produces the damage to the structure, intensifies the fire, is the means by which
the fire spreads and enlarges, and provides the greatest barrier to the extinguishment of
the fire. When the heat can be properly studied and understood in connection with a fire,
the sequence and cause of that fire will generally be very clear. There are several special
considerations of heat, as applied to a fire investigation, that require understanding:
1. Heat as it applies to igniting the fire (This topic is developed further in Chapter 6
on ignition.)
2. Heat as it applies to the increase of the rate of chemical reactions (including fires)
as they develop
3. The transfer of heat as the factor controlling the spread of fire when the additional
conditions of available fuel and oxygen are met
4. The effects of heat on materials that survive the fire to bear indicators of the fire’s
intensity, duration, or direction of spread

HEAT AND THE RATE OF REACTION

The rate of all chemical reactions is dependent on temperature. With rare exceptions, all
reactions have a higher rate at greater temperatures. The Q10 value is the increase in rate
of the reaction that results from raising the temperature 10°C (18°F). For most oxidation
reactions, this value is two or more. Therefore, it can be generalized that the rate of a
combustion reaction usually doubles with every increase of 10°C. The great importance
of temperature stems, at least in part, from the fact that the fire generates much heat and
raises the temperature of the reacting components, thus increasing the rate of reaction.
This in turn generates more heat, thus again increasing the rate of reaction. This is the key
to the fourth element of the fire tetrahedron—the chain reaction. If it were not for the
diminishing availability of fuel and oxygen, combined with loss of heat to the larger sur-
roundings, that brings this chain reaction under control, every fire would become a vio-
lent and rapid holocaust. Because these factors do act to control fire, the most immediate
significance of the concept of rate of reaction is in connection with those processes that
lead to spontaneous combustion. This subject is treated separately in a later section; how-
ever, it can be briefly considered in the following light.
Assume that a system of fuel and oxygen is properly contained, insulated, and pro-
portioned, and a very limited exothermic reaction is proceeding. This might be a slow oxi-
dation, bacterial activity, or some other type of chemical transformation in which heat is
generated. If the heat cannot escape as fast as it is generated, the temperature will rise. As
it rises it speeds up the very reaction that is generating the heat, thus leading to an increas-
ing rate, which results in greater and greater quantities of heat. This heat will finally be
sufficient to raise the temperature of the system to the ignition temperature of at least part
of the fuel. Naturally, when this temperature is reached, the fuel ignites to a free-burning
42 Chapter 3 Fundamentals of Fire Behavior and Building Construction
fire, and we say that it resulted from spontaneous combustion. If the heat can escape (due
to convective, radiative, or conductive losses), the reaction rate does not increase as rap-
idly, and the reaction may continue at a reduced rate until the temperature stabilizes at
some point and will not increase further.

HEAT AND TEMPERATURE

Heat is energy in a kinetic form, or energy of molecular motion. Except at absolute zero
(⫺273°C, ⫺460°F, 0 K) all matter contains heat because its molecules are in motion.
Temperature is merely an expression of the relative amount of this energy that a body pos-
sesses. Heat or temperature is described using four units of measure: Celsius (°C), Kelvin
(K), Fahrenheit (°F), or Rankine (°R). The conversions are as follows:
TC ⫽ (TF ⫺ 32) 5>9
TK ⫽ (TF ⫺ 32) 5>9 ⫹ 273
TF ⫽ TC (9>5) ⫹ 32
TR ⫽ TF ⫹ 460
Table 3-2 presents selected temperatures in those four units of measure, as well as conversions.
The total amount of heat contained in a body is determined by the mass of that body
and by its thermal capacity. The thermal or heat capacity of material is a measure of how
much heat must be added to a given mass of material to increase its temperature. The value
is given in joules/kilogram kelvin (J/kg · K) (since it is dependent on the temperature of the
mass as well). Thermal capacity is usually represented by the symbol cp. The thermal
capacities of some materials are shown in Table 3-3 (see also Appendix C). The fundamen-
tal properties of a material that influence its ignitability are thermal capacity, thermal con-
ductivity, and density. These properties will be explored in later sections of this text. For
the purposes of fire investigation, it must be remembered that heat is a measure of energy, watt ■ Unit of heat
release; 1 watt ⫽
while temperature is merely a means of comparing the heat content of two objects. 1 joule/second; unit of
power or work (in elec-
HEAT RELEASE RATE trical circuits, equivalent
to voltage multiplied
The amount of heat produced by a fire or any heat source is measured in calories or, more by amperes).
commonly, in British thermal units (Btu) or joules (or kilojoules, kJ) (1 Btu ⫽ 252 cal ⫽ 1,055
J ⫽ 1.055 kJ). The rate at which that heat is being produced by the heat source is measured heat release rate ■ The
most conveniently in joules per second, Btu per second, or calories per second. Because 1 J/s rate at which heat is
generated by a source,
⫽ 1 watt, the rate is measured in watts (W) or Btu/s (0.95 Btu/s ⫽ 1000 # W ⫽ 1 kilowatt, or usually measured in
1 kW). This# value, the heat release rate, is usually represented by Q (the dot meaning per watts, joules per second,
unit time). Q is the quantity that characterizes the size of a fire. Another way of looking at or Btu per second.

TABLE 3-2 Temperatures Measured in Different Units

K °C °F R
773 500 932 1392
373 100 212 672
298 25 77 537
273 0 32 492
255 ⫺18 0 460
233 ⫺40 ⫺40 420
0 ⫺273 ⫺460 0
Kelvin Celsius Fahrenheit Rankine

Chapter 3 Fundamentals of Fire Behavior and Building Construction 43

 TABLE 3-3 Heat Capacities and Thermal Conductivities of Some
Common Materials

HEAT CAPACITY (cP) THERMAL CONDUCTIVITY THERMAL INERTIA

MATERIAL (@20°C) (J/kg · K) (k) (W/m · K) (k␳ c)a (W2/m4K2)/s
Copper 380 387 1.3 ⫻ 109
Aluminum 900* 273* —
Steel (mild) 460 45.8 1.6 ⫻ 108
Glass (window) 840 0.76 1.7 ⫻ 107
Brick (common) 840 0.69 9.3 ⫻ 105
Gypsum 840 0.48 5.8 ⫻ 105
Wood (oak) 2,850 0.17 3.2 ⫻ 105
Polyurethane foam 1,400 0.034 9.5 ⫻ 102
Cotton 1,300* 0.06* —
Air 1,040 0.026 —
†
Water 4,180 0.60 2.5 ⫻ 106

Sources: D. D. Drysdale, An Introduction to Fire Dynamics, 2nd ed. (Chichester, UK: Wiley,1999), 33.
*
R. H. Perry and D. Green, Perry’s Chemical Engineers’ Handbook (New York: McGraw-Hill, 1984); SFPE Handbook
of Fire Prevention Engineering, 2nd ed. (Quincy, MA: NFPA, 1995), table B-7.
†
SFPE Handbook of Fire Protection Engineering, table B-4, p. A-30.
a
ρ ⫽ density of the material in kg/m3

the heat release rate of a fire is to consider the amount of fuel being consumed in it per second
and multiplying that mass burning rate by the heat of combustion. In a real fire, there is a
certain inefficiency due to incomplete combustion, so the relationship is not perfect. The heat
release rate is a most useful way of comparing fires or predicting their behavior and their influ-
ence on other fuels nearby. The heat release rates of some typical fires are shown in Table 3-4.

TABLE 3-4 Heat Release Rates of Some Typical Fires

Smoldering cigarette 5W
Wooden kitchen match or cigarette lighter 50 W
Candle 50–80 W
Office wastebasket with paper 50–150 kW
Small chair (some padding) 150–250 kW
Armchair (modern) 350–750 kW (typical)—up to 1.2 MW
Recliner (synthetic padding and covering) 500–1,000 kW (1 MW)
Sofa (synthetic padding and covering) 1–3 MW
Pool of gasoline (2 qt, on concrete) 1 MW
Christmas tree (dry, 6–7 ft) 1–2 MW (typical)—up to 5 MW
Living or bedroom (fully involved) 3–10 MWa

Sources: J. Quintiere, Principles of Fire Behavior (Albany, NY: Delmar, 1998), 114–22; J. Krasny, W. Parker, and
V. Babrauskas, Fire Behavior of Upholstered Furniture and Mattresses (Park Ridge, NJ: Noyes, 2001).
a
Depending on ventilation.

44 Chapter 3 Fundamentals of Fire Behavior and Building Construction

HEAT TRANSFER AND HEAT FLUX
An integral part of every fire is the transfer of heat, both to the fuel (which is critical to
continuity of the fire) and away from regions of combustion, whatever their type. The
principles of heat transfer are relatively simple and are vital to the understanding of the heat transfer ■ Spread
fire itself. The rate at which heat is falling on a surface (or passing through an area) is of thermal energy by
called the heat flux and is measured in watts per square centimeter (W/cm2) or kilowatts convection, conduction,
2 2 2 or radiation.
per square # meter (kW/m ) (1 W/cm ⫽ 10 kW/m ). The heat flux is represented by the
symbol q – (the double prime being notation for “per unit area”).
Heat is transferred in three fundamental ways: (1) conduction, (2) convection, and (3)
radiation. All these processes play a part in fires; however, the relative importance of each
will vary with the intensity and size of the fire, as well as with the shape and content of
the environmental system that is burning. Each will be considered separately.
Conduction
Conduction is the transfer of heat energy through a material by contact between its mov- conduction ■ Process
ing molecules. Because actual contact of the vibrating molecules is necessary, conduction is of transferring heat
limited, for the most part, to localized action. Its effects are most noticeable in solid mate- through a material or
between materials by
rials where molecular contact is at its highest and where convection (large-scale, physical direct physical contact.
movement of the molecules) does not occur. Heat always travels from hot areas of a solid
to cold ones by conduction. This can be illustrated readily by heating one end of a metal
rod and observing the temperature at the other. Heat traveling through the rod will cause
a rise in temperature at the other end, but with considerable lag. The amount of heat flow-
ing through the rod is proportional to the time, the cross-sectional area, and the difference
in temperature between the ends, and is inversely proportional to the length.
The rate at which heat is transmitted through a material by conduction is measured
as the thermal conductivity (k), in units of W>m ⴢ K (when the heat impinging on a surface
is measured in W/m2, distances or thicknesses in meters, and temperatures in °C or K).
Scientists and engineers find it convenient to express such relationships in the form of an
equation, because it shows the way in which factors interrelate and allows for prediction
of results. The relationship for conductivity is given by
# #
q ⫽ kA(T2 ⫺ T1 )>l or by rearranging, k ⫽ q l>A(T2 ⫺ T1 )
Where:
#
q ⫽ the rate at which heat is being conducted (in kJ/s or kW)
k ⫽ thermal conductivity of the material
A ⫽ the area through which the heat is being conducted
T2 ⫺ T1 ⫽ the difference in temperature between the “hot” and “cold” zones
l ⫽ the length (or distance) through which the heat is being conducted7
#
Plugging different values into this equation shows that heat is conducted faster (q gets
larger) if T2 is much larger than T1, if the area is larger, or if the distance through which
the conduction is taking place is smaller. The thermal conductivity of a material (k), is
usually specified at 0°C or 20°C, since it varies with temperature. It also varies dramati-
cally among materials. The thermal conductivities of some common materials were
shown in Table 3-3 (see also Appendix C).
An illustration of the importance of thermal conduction might involve the relative
behavior of a piece of metal—for example, a length of wire or a galvanized metal sheet—
as compared with the behavior of a piece of wood in the same fire. The thermal conduc-
tivity of metals is high, so that if the metal piece is heated, the heat rapidly spreads to
unheated areas. If the temperature in these portions rises above the ignition temperature
of any fuel material in contact with the metal, this fuel may be ignited at a distance from the
initial source of heat. A piece of wood, in contrast, being a poor conductor of heat, may
be intensely heated and even burn at one location for a long interval. The heat, however,
Chapter 3 Fundamentals of Fire Behavior and Building Construction 45
 does not spread through it well, and the wood will not be expected to ignite at a distance
from the point at which the heat is applied. One side of a simple board will show deep
charring in a fire, and the opposite side, unless exposed to flame from another source,
remains uncharred and apparently normal wood.
The importance of the thermal conductivity of the material through which the heat is
being conducted in the development and consequence of fire is considerable. Not only is
conductivity involved in the ignition of materials, in which the heat must flow from the
heat source to the fuel and to some degree into it before a fire will start, but it also has a
marked effect on the degree of fire damage, as we shall see in later sections. The thermal
conductivity of easily ignitable fuels such as wood, plastic foam, or paper is very low,
which means that heat applied to a surface tends to accumulate there, raising the local
temperature, possibly to the fuel’s ignition temperature. The conductivity of metals is very
high, meaning that heat applied is quickly dissipated into the bulk of the metal, and the
local temperature does not approach ignition temperatures as readily. Copper conducts
heat almost 2,000 times more efficiently than does wood. After a fire, thermal damage to
the insulation on a copper wire may be seen to extend for some distance from the actual
point of heat application or the location of a heat-producing electrical fault. Wood may
be damaged very heavily on only one side, and even then only in the area of direct appli-
cation of heat. Other areas not directly exposed to fire may be undamaged. Poor thermal
conductors such as felt, plastic foam, and the like are used as insulation. They also tend
to be poor conductors of electricity as well as heat. The high thermal conductivity of some
metals means that it is more difficult to melt them by application of localized heat. For
example, the melting point of copper (1,082°C; 1,981°F) is often exceeded in the com-
bustion zone of a wax candle, but only the finest gauge copper wire is likely to be melted
in a candle flame because larger cross sections permit the heat to be conducted away so
quickly that the temperature of the bulk metal never exceeds the melting temperature. The
high thermal conductivity of metals contributes to their rapid loss of heat when taken out
of a flame. A hot metal particle may contain a great deal of heat due to its high thermal
capacity but loses its heat to its surroundings much more quickly than will wood of the
same mass.
Convection
convection ■ Process Convection, as a means of transferring heat, is extremely important in fires. It can be
of transferring heat by defined as the distribution of heat by means of a circulating medium or the transfer of
movement of a fluid. heat to or from a moving medium. As such, it will occur in gases and liquids but obvi-
In convective flow, the
warm fluid becomes
ously not within solids. Convection is, however, responsible for transferring heat from
less dense than the solids to liquids or gases (and vice versa) by convective heat transfer. In most fires it is
surrounding fluid and driven by differences in density caused by temperature variations (buoyant flows). Such
rises, inducing a buoyant flows are usually referred to as convective or convection flows.
circulation. When water is heated, its temperature can rapidly be raised despite its very low heat
conductivity. The temperature rises rapidly because the heated water on the bottom of a
container expands, becomes less dense, and is displaced by the heavier cold water, which
is heated in turn. Another very familiar example is the heating of buildings by furnaces or
heaters. Here the fire is small and controlled, but the heated air coming from it is often
circulated exclusively by convection. Forced-draft furnaces have a fan or blower to sup-
plement this convection.
In fires the moving masses of hot materials are the gaseous products of combustion,
along with surrounding air, which is also heated. These expand and become lighter and
move upward at a rapid rate. This buoyancy is the form of heat transfer that accounts for
most heat movement in a normal fire. It also largely determines the fundamental proper-
ties of fire with respect to its movement, spread, and ultimately its pattern.
Convective heat transfer is represented by the equation
#
q ⫽ h(T2 ⫺ T1 )A (W)
46 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 #
where, once again, q represents the rate at which the heat moves, A is the area through
which it is being transferred, T2 and T1 are the temperatures of the two materials, and h
is a value that represents the efficiency of the heat exchange between those two materials.
Because convective heat transfer varies with different combinations of materials, surface
textures, speeds of movement, and so on, it is a value that one looks up in a reference.
For a buoyant flame plume in air, h is between 5 and 10 W/m2 · °C. Application of the
preceding equation, with a typical flame temperature of 800°C, would give a convective
heat transfer to a nearby surface on the order of 4 to 8 kW/m2.8 This is considerably less
than the radiative heat transfer of 40 or 50 kW/m2 one would expect from the same flame
at near contact.
In very large fires, especially outdoors, the upward movement of hot gases may be so
great as to contribute to the formation of what is called a firestorm. In such cases, drafts firestorm ■ Overwhelm-
are created that can suck lighter fuels at ground level into the fire, and the buoyant cur- ing progression of fire
rent lifts masses of burning gases and solid debris hundreds of feet into the air, debris that through structures or
wildlands caused by a
falls as burning brands among other fuels downwind. The burning gases can form a sep-
combination of convec-
arate mass of flames called a fireball or firewhirl, which can ignite fuels that come into tive and radiative
contact with it. The intensity of the fire produces radiative heat of such magnitude that processes.
fuels some distance away are heated to their ignition temperatures and burst into flames.
The firestorm is the mechanism responsible for the terrible damage in infamous fires such
as those in Dresden, Germany, and Tokyo, Japan, (both as a result of air raids during
World War II), and in Chicago, Illinois, and Peshtigo, Wisconsin (accidental fires, one
urban, the other wildland in origin, both in October 1871).
Radiation
Transfer of heat by radiation is less commonly appreciated than transfer by conduction radiation ■ Transfer of
or convection, and yet it plays the most critical role in fire growth and spread, particu- heat by electromagnetic
larly in larger fires. Radiation aids fire spread across a surface, promotes ignition of other waves.
fuels, and may produce burn patterns that survive the fire. All objects having a tempera-
ture above absolute zero (0 K, ⫺273°C, ⫺460°F) radiate energy in the form of electromag-
netic energy. Radiated heat energy can be transferred to another body without any
contact or circulating medium. As we shall see later, radiation plays a critical role in the
spread of a fire later in its development. At low temperatures, this radiation is solely in
the form of infrared emission. At temperatures above 500°C (900°F), some of the radia-
tion is in the visible portion of the spectrum, and the color perceived depends on the tem-
perature of the body, as described in Table 3-1. Infrared radiation behaves exactly like
visible light except that it is not detectable by the unaided eye. The infrared radiation
from the sun is the primary source of heat on Earth. As soon as the sun is visible, the heat
it radiates can be felt, just as an observer feels the radiated warmth of an open fire. Such
warmth is felt because it is absorbed by the body, and the absorption of heat is greater
than the radiation of heat by the body under these circumstances. In the same way, a per-
son in a cold environment is chilled because his or her body radiates heat faster than it
receives heat from the environment. In surroundings where everything is at the same tem-
perature, all objects are both radiating and absorbing radiant heat at the same rate. The
result is no change in temperature. If the temperature is measured in K (°C ⫹ 273), the
intensity of the total emitted radiation (E) is related to the temperature (in K) raised to
the fourth power (T 4), as in the Stefan-Boltzmann relationship:
E ⫽ ⑀␴T 4 (W/m2)

where ⑀ is the emissivity of the source and ␴ is the Stefan-Boltzmann constant.9 This
means that a modest increase in the temperature of a surface, say, doubling it from a room
temperature of 300 K (27°C; 81°F) to 600 K (327°C; 621°F), increases the intensity of radi-
ant heat emitted by that surface by a factor of 16 (2 to the 4th power ⫽ 16). The effects of
radiant heat are determined by its magnitude and duration. Table 3-5 gives some examples
Chapter 3 Fundamentals of Fire Behavior and Building Construction 47
 TABLE 3-5 Effects of Thermal Radiation on Thick Solids (Wood, Plastic, Human Tissue) in Still
Air @ 20°C

RADIANT
# HEAT EQUILIBRIUM SURFACE
SOURCE FLUX (q – ) (kW/m2) TEMPERATURE* OBSERVED EFFECT
Direct summer sun 1 40°C (100°F) None
Distance from fireplace 2–4 45°C (120°F) Pain after 30 s
Proximity to fireplace 4–6 54°C (130°F) Pain after 8–10 s
2nd degree burns to skin in 20–30 s
Near proximity to fireplace 10 100°C–150°C Scorching of some materials
(200°F–300°F) Melting of some thermoplastics
Face of fireplace 20 150°C–250°C Some cellulosics and synthetics ignite in
(300°F–500°F) ⬍60 s
Inside of fireplace 30 300°C–400°C Autoignition of many fuels in 0–30 s
(600 °F–800°F) (wood in ⬎ 60 s)
Adjacent to flames 50 400°C (800°F) Autoignition of nearly all materials in ⬍5 s
Post-flashover 120–150 ⬎500°C (⬎800°F) Rapid combustion

*Estimated from various sources including ASTM E1321-97: Standard Test Method for Determining Material Ignition and Flame Spread
Properties.

of the effects of thermal radiation. These effects are useful in estimating the intensity of
heat exposure on surfaces after a fire as an aid to reconstructing the positions of heat
sources and “target” (receiving) surfaces.
The heat flux is not the only factor that determines the transfer of radiant heat. There
is also a dependency on the view factor; that is, does the receiving surface sit at right
angles to the path of the radiation, or is it exposed uniformly to radiation from the
source? The intensity of the radiation being received from a point source of heat falls off
with the square of distance.10 In other words, if an object is 1 m (3 ft) away from a small
wastebasket fire, directly facing it, and is receiving 6 kW/m2 radiant heat, and the object
is moved twice as far away (to 2 m; 6 ft), the radiant heat intensity (heat flux) drops to
1/4 of 6 kW/m2, or 1.5 kW/m2. The heat flux on a surface some distance r away from a
small, localized fire is represented by
# #
q – ⫽ Xrad Q >4␲r2
#
where Q is the total heat release rate of the wastebasket fire, and Xrad is the fraction of
11
heat being released# as radiant heat (for a trash fire, that might be 0.4). One can see that
as r gets larger, q – falls off very rapidly (the inverse-square relationship).
If an object is very close to a large fire or hot surface, the fire can no longer be rep-
resented as a single point source, and the summed contributions from various parts of the
large source overwhelm the inverse-square law. The relationship becomes
#
q –1 ⫽ ⑀␴T 24F12
where F12 represents the view factor between
# the source (at temperature T2) and the receiv-
ing or target surface (surface 1), and q –1 is the heat flux falling on that target surface.12
Calculation of the view factor F12 requires an estimation of dimensions of the radiant sur-
face and distance of separation, as in Figure 3-7. The value is calculated as the ratio a/c,
and the appropriate curve is selected in Figure 3-8. The y-value is calculated from ratio b/c.
The intersection of curve x with the vertical line corresponding to y yields the factor F12
on the vertical axis.13
48 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 Hot surface FIGURE 3-7 A small
target facing a large heat
b
source (a by b in size) a
distance c away is subject
to radiant heating from a
wide angle of view
Field of view Target
a

.3 FIGURE 3-8 The config-

x=∞
uration factor is taken
2
.2 from this nomograph,
1 where x ⫽ a/c and y ⫽
.7 b/c. The closest curve for
.5 x is selected. Where that
.1 .4 curve intersects with the y
.3 value determines F12.
.07 From Hamilton and Morgan.
Configuration factor, F12

.2 NACA Tech. Note 2836,

.05
December 1952.
.15

.03 0.1

.02

.01

.007

.005
NACA

.003
.05 .1 .2 .5 1.0 2 3 5 7 10 20

The calculation of the view factor for such exposures is much more complex, but
notice that distance alone is no longer a factor. For instance, if the hot gas layer in a room
ignites, the entire ceiling layer represents a radiant heat source. If we measure the heat
flux on an object 1.2 m (4 ft) below the ceiling layer and the heat flux on an object 2.4 m
(8 ft) below the ceiling layer, the heat fluxes will be very similar. That is one reason why
radiant heat from a large heat source seems to bring so much of a room’s contents to igni-
tion at about the same time. These factors will control how much radiant heat is being
received on the target surface. The “edge” or “corner” effect is one reason why the tar-
gets in the center of a room ignite long before the carpet at the edges as a room
approaches flashover. flashover ■ The final
The effect of the heat flux is also subject to the nature of the exposed surfaces and stage of the process of
fire growth; when all
especially to the degree to which they reflect heat or infrared rays. Heat (or visible light)
exposed surfaces of
falling on a surface may be absorbed or reflected. Objects that radiate heat readily also combustible fuels
absorb it readily. In general, dark-colored objects both absorb and radiate heat better within a compartment
than light-colored ones, but some light-colored materials absorb infrared (heat) energy are ignited, the room is
very effectively. However, polished surfaces, such as chrome plate, resist both absorption said to have undergone
flashover.
and emission. The same mirror that reflects visible light also reflects the infrared or heat
rays in the same manner, but the glass from which it is made absorbs the infrared, and its
Chapter 3 Fundamentals of Fire Behavior and Building Construction 49
FIGURE 3-9A Massive fire in condominium building igniting build- FIGURE 3-9B Structure fire igniting vehicle (and ground cover) by
ing across the street by radiant heat (and fire balls). Courtesy of Calvin radiant heat. Courtesy of Greg Lampkin, Knox County, TN, Fire Investigation
Bonenberger, Fire Marshall, Lafayette Hill, PA, FD. Unit.

temperature rises. Many of the post-fire indicators on which investigators rely to indicate
fire travel and intensity are the result of radiant heating, so an understanding of the fac-
tors that contribute to it is very important.
Radiant heat is a major factor that makes firefighting difficult, because close prox-
imity to a large fire heats up equipment and firefighters, making operations difficult if
not impossible to carry out safely. When a fire is burning in a portion of a structure, all
surfaces that face the fire will be heated by radiant heat, and when this surface temper-
ature reaches the autoignition point of the material itself, it will burst into flame. In
very large fires, it is not uncommon for adjacent buildings and nearby vehicles to be
ignited solely by the radiant heat (as in Figures 3-9a and b). This radiative ignition is a
major factor in fire growth in a compartment, as will be discussed shortly. Charred
areas inconsistent with fire spread from item to item are commonly due to radiant heat
from the room fire and should be so recognized by the fire investigator. Generally, they
are not as highly localized as char damage caused by direct flame impingement but can
exhibit “shadow” effects where there are intervening solid objects to block some of the
radiation. When there are no intervening or overlying solids, or prolonged post-
flashover burning, radiation char patterns tend to be uniform in degree of charring.
They can complicate the search for a fire origin because radiant heat can produce char-
ring at low levels in a room even when the fire generating the heat is not as low in the
room (see Figure 3-10). However, the ignition of floor coverings by radiant heating and
their burning can result in char damage to the floor that is not as uniform as direct radi-
ant heat char (and can be very irregular). These effects will be discussed in more detail
in Chapter 7.

DIRECT FLAME IMPINGEMENT

In a real fire, flames are spread not only by the three basic mechanisms of heat transfer
described previously but also by direct flame impingement, and its role cannot be ignored.
On a microscopic scale, direct flame contact or impingement is a combination of both
convective and radiative mechanisms. If the plume of hot gases from the established flame
rises by buoyant flow into contact with more fuel, heat is transferred from the plume by
convective heat transfer and radiation until the new fuel pyrolyzes and generates gaseous
fuels that are ignited by the flames of the plume. The flames then spread upward and,
50 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 FIGURE 3-10A As fire
Ceiling
extends across the ceiling,
heat radiating downward
Hot gases
from the ceiling layer
and flames
chars and ignites exposed
surfaces; areas under
Radiant
chair and table are pro-
heat
tected until later in the
fire. When flashover is
completed, all fuels in the
room will be on fire, and
the undersides of tables,
chairs, and shelves will be
exposed to flames from
floor-level burning.
Floor

Lightly charred Heavily

in protected scorched and
areas charred areas

FIGURE 3-10B A struc-

ture fire demonstrating
ignition by radiation.
Intense localized fire on
the floor and railing of
the porch was ignited by
radiation coming from
fire plume exiting the top
portion of the door.
Courtesy of John D. DeHaan.

more slowly, outward. Because fresh fuel is continually being added to the reaction with
little loss of heat, the fire can spread quickly through the fuel presented. As more fire is
created, the faster fuel is involved and the temperature of the room or enclosure rises,
thereby increasing the rate of oxidation of other fuels, and the fire spreads at a faster and
faster rate as long as the supply of oxygen is maintained.

Flame Plume
A fire burning in the middle of a room or out in the open creates a rising column of hot
gases (plume) by the buoyant flow just described. The temperature in a simple fire is at a
maximum in the well-mixed combustion zone just above the fuel package. As the gases
(and solids and liquid aerosols mixed therein) rise they lose their heat, some by mixing with
room air (entrainment) and convective transfer to the surrounding air, and some by radiant
Chapter 3 Fundamentals of Fire Behavior and Building Construction 51
 FIGURE 3-11 Temperature distribution in
Temperature distribution buoyant flame and smoke plume
in buoyant flame and
smoke plume along verti- Height
cal centerline. above fuel

Smoke plume

Tip of flame

Flame plume
Mixing zone

0 25 800
Ambient
Temperature (°C)

loss (particularly from soot particles and aerosol droplets of pyrolysis products from
incomplete combustion). As a result, if we measured the temperature of the flames, we
would see a predictable pattern to the temperature as a function of height above the fuel
package. Temperatures decrease from 700°C to 900°C (1,500°F to 1,800°F) just above the
fuel to about 500°C (930°F) at the tip of the flame, where it becomes very turbulent and
intermittent.14 As the gases continue to rise they continue to lose heat, and the tempera-
ture continues to decrease in a predictable pattern until the gases have lost so much heat
that they are no longer buoyant. (Remember that most of the combustion products—CO2,
SO2—are heavier than air at the same temperature; CO is nearly the same density; only
water vapor (H2O) is lighter.) A temperature plot of flame temperature versus height
would look something like Figure 3-11. The relationship between the flame height and the
heat release rate of the fire is quite predictable. One form of the relationship for a symmet-
rical pool fire burning with no nearby walls or ceiling is the Heskestad equation:
#
Zf ⫽ 0.23Q 2>5 ⫺ 1.02D
#
where Zf is the visible flame height in meters, Q is the heat release rate in kilowatts, and
D is the diameter of the pool (in meters).15
Entrainment draws fresh air into the base of the fire to keep the reaction going and
diffuses and cools the rising gases in the plume, as in Figures 3-12a and b. If the fire is
against a noncombustible wall, the entrainment of cooling air is reduced by 50 percent,
as in Figure 3-12c. If fire gases have more time to rise before cooling to the point they are
no longer incandescent, the visible flame plume is taller. If the fire is in a corner, the
entrained air is reduced even more (as in Figure 3-12d),# and the flame plume is even taller.
Radiant heat from nearby walls further increases Q. This situation is dramatically illus-
trated in Figure 3-13. These fires can be represented by the following equation developed
by Alpert and Ward:
#
Zf ⫽ 0.17(kQ) 2>5
where k is called the wall factor. The wall factor is 1 for no wall, 2 for a single wall, and
4 for a corner.16 These relationships allow us to predict how big a fire will look to an
observer (or what kind of effect its gases will have on materials that come into contact
with it at various heights) given the heat release rate of the fire. Due to the turbulence of
flames, the height of the flame tip will oscillate up and down, and the tip will sometimes
separate from the main body of the flame. The human eye tends to average these varia-
tions, and the height is normally fairly close to the predicted Zf.
52 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 Fire Entrained air

FIGURE 3-12B Top view of

FIGURE 3-12A Side view of entrainment around periphery of
entrainment—unrestricted fire. unrestricted fire.

Fire Fire

FIGURE 3-12C Top view of FIGURE 3-12D Top view of corner

reduced entrainment of wall fire. fire—restricted entrainment.

FIGURES 3-13A & B Fires in similar vertical cushions—free-standing (away from walls). Maximum flame
height: about 0.7 m (30 in.) above cushion. Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 53

FIGURE 3-13C Maximum flame height against noncombustible FIGURE 3-13D Maximum flame height in corner (about 3 m).
wall (about 1.5 m). Courtesy of Jamie Novak, Novak Investigations and Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.
St. Paul Fire Dept.

Sequence of a Room Fire

fuel load ■ All com- A generalized room fire in a room with a normal fuel load goes through fairly predictable
bustibles in a fire area, stages of development in its lifetime. Each stage has characteristic features, and the effects
whether part of the
on fuels vary in each. The fire investigator must appreciate what these effects are to prop-
structure, finish, or
furnishings. erly interpret the post-fire scene. The length of time each stage takes will vary (dramati-
cally in some cases) with the circumstances of ignition, fuel, and ventilation, but the
incipient ■ Beginning sequence from incipient or beginning stage to growth to decay stages will remain essen-
stages of a fire. tially the same (see Figure 3-14).

BEGINNING OR INCIPIENT STAGE

The beginning or incipient is the first stage of fire development after the moment of igni-
tion. The flames are localized in the first fuel ignited, and a free-burning open flame is
typical. The room has a normal oxygen content (about 21 percent) throughout, and the
overall temperature has not yet begun to rise. There is a plume of hot gases rising from
the flame. The contents of this plume will depend on the fuels being burned but usually
include soot (carbon and other solids), water vapor, carbon dioxide, carbon monoxide,
sulfur dioxide, and traces of other toxic gases (as well as heated nitrogen entrained from
54 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 Flashover
Post-flashover
steady-state
Incipient burning
Growth period Decay period
1400 Carpets,
1200 wall coverings
Heat release rate (kW)

ignited
All fuels Major fuels
1000
Sofa involved consumed
800 ignited Smoldering
Second item
600 ignited
(chair)
400 Extinguished

200

30 60 90 120 150 180 210 240 270 300 330 360 390
Time (s)
Ignition of
first fuel
(chair)

FIGURE 3-14 Total heat releases rate versus time for a fire in a typical small furnished room. As each item
ignites it contributes to the growth of the fire. In this case, the critical heat release rate (and hot gas tempera-
ture) is reached and the room goes into flashover, followed by steady-state burning, and then a decay period
as the fuel is exhausted. Finally, the fire dies down to smoldering embers before being extinguished.

the surrounding air). Convection or buoyant flow carries these products and heat to the
upper part of the room and draws oxygen in at the bottom of the flames to sustain com-
bustion. If there is a solid fuel above the flame, convection and direct flame contact spread
flames upward and outward, producing the typical vertical pattern of char (with the
flammability/flame spread of the fuel determining the nature of the pattern). If there is no
fuel above the flames, the gases in the plume rise, losing heat to the room air by radiation
and convective (often turbulent) mixing. An incipient fire has such a low heat release rate
(HRR) that overall temperatures may not be affected measurably (and sometimes not
even the temperature of the upper “smoke” layer) (see Figures 3-14 and 3-15).

4000
3500
Rate of heat release (kW)

3000
2500
2000
1500
1000
500
0
0 200 400 600 800 1000 1200 1400
Time (s)

FIGURE 3-15 The heat-release curve of a lighter-ignited fire in a bedroom. The incipient fire was less than 100 kW
for more than 720 seconds (12 minutes). The growth phase involved the bed mattress and box spring (mattress
platform). At ~1100 seconds (18 minutes) the fire spread to the face of a nearby wooden dresser. Rapid fire growth
on the dresser caused ignition of the smoke layer to trigger flashover. Fire was quickly extinguished by hose spray.
Peak HRR was 3.9 MW. Courtesy of Bureau of Electronics, Appliance Repair, and Home Furnishings, North Highlands, California.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 55

 GROWTH, A FREE-BURNING STAGE
The fire grows in intensity as more fuel becomes involved. Convection and radiation
spread the flames upward and outward from the initial fuel package until nearby fuels
reach their ignition temperatures and begin to burn. Heat radiating from the flames may
begin to bring other fuels to their ignition temperatures, spreading the fire laterally. The
speed of this lateral spread is highly dependent on how close the other fuel packages are
and their physical arrangement. The closer they are to one another, the faster the fire will
spread and grow in heat release rate. Until the initial fire reaches a moderate size (around
200 kW), radiant heat transfer to target surfaces even a small distance (0.5 m) away is
minimal, and most fire spread is driven by convective heat transfer above the fire and
direct flame contact. [Radiant heat transfer to a vertical surface facing the fire is balanced
against convective transfer from that surface to the room air, thus reducing its net gain.
A surface above the flame (or in contact with it) will not have the same convective heat
losses so will get hotter faster.] Hot gases generated by the fire form a layer at the ceiling.
The temperature of this layer rises (at a rate that depends on the heat release rate of the
fire, the size of the room, and how well insulated the ceiling and walls are) as the fire con-
tinues to burn and spread.
This upper layer of the room with its low oxygen content; high temperature; and
soot, smoke, and partially burned pyrolysis products (including toxic gases such as car-
bon monoxide, hydrogen cyanide, hydrogen chloride, acrolein, and the incompletely
reacted chemical species called free radicals) is now very different from that near the floor.
The floor layer is still relatively cool and contains fresh air (with an oxygen concentration
close to normal) that is being mixed into the flames below the layer of reaction products.
Survival is still possible in the room if one stays in the cooler lower layer and does not
breathe the gases higher up. If ventilation to the room is limited, the flames will produce
more incompletely burned materials including pyrolysis products, carbon monoxide, and
solid fuels, which can be seen as soot and smoke. The level of this fuel-rich layer grows
lower and lower in the room (deepens), and its temperature increases as the fire contin-
ues and the flames grow. If the layer deepens to reach the level of the fire in the room, its
reduced oxygen content will reduce the combustion rate, possibly to the point of extin-
guishing any flaming fire. The mass of combustible gases, aerosols, soot, and smoke accu-
mulates, often until one or more of the fuels present in the layer reaches its ignition
temperature whether from accumulated heat from the flaming fire or direct contact with
open flames. The layer of airborne “fuel” then catches fire, and the flames may extend
flameover ■ The flam- throughout the room at ceiling level. This is the process often called flameover or
ing ignition of the hot rollover. The flame front in the roiling cloud of fuel and air can reach speeds of 3 to 5
gas layer in a develop-
m/s (10 to 15 ft/s). With flaming combustion proceeding in that layer, the temperature of
ing compartment fire.
the ceiling layer increases even faster. The rapid increase in temperature of the hot gas
layer causes expansion of the layer, filling more of the room. In observed tests, flameover
in a residential living room caused flames to fill the compartment nearly to floor level in
7.5 seconds after ignition, prompting firefighters to exit for their safety.17

FIRE GROWTH TO FLASHOVER

Even in the absence of flaming ignition of the hot gas layer (flameover), the radiant heat
from the hot gas layer at the ceiling is striking all other materials in view in the room with
an intensity (heat flux) that is dependent on the temperature of the layer. When this layer
reaches a critical temperature of approximately 600°C (1,100°F), it is generating approx-
imately 20 kW/m2 (2 W/cm2) radiant heat at floor level in a normally proportioned room.
This is sufficient to bring most other fuels, such as furniture and floor coverings in the
room, to their ignition temperatures. At this time, all fuels in the room become involved
in flaming combustion, and the room is said to have undergone flashover.18 Once this
occurs, all exposed fuels in the room are involved, temperatures throughout the room
56 Chapter 3 Fundamentals of Fire Behavior and Building Construction
reach their maxima, and survival for more than a few seconds is impossible. Once the
floor or floor covering ignites, the flames cause high temperatures near floor level, which
meet and often exceed the high temperatures that existed only in the ceiling layer prior to
flashover. Noncombustible floors, of course, will not contribute flames or heat, so the
floor temperatures will remain lower than in the rest of the room. The enormous turbu-
lence caused by the generalized flames fills the room with hot gases as the two-layer pat-
tern breaks down. The rapid extension of open-flame fire across the hot gas layer of a
room is often the event firefighters refer to as flashover because it means that complete
ignition is imminent. Flashover to the fire scientist refers to the process of extension, radi-
ation, and ignition of all fuels in a room or compartment. See Figures 3-16a–i for a graphic
representation of the process. It will be discussed in greater detail in a later section.
As each fuel package in a compartment ignites, it contributes a quantity of heat to
the room at a rate influenced by what fuel is present, how it is arranged, and even how
it is ignited. If we plot the heat release rate (or average temperature of the hot gas

FIGURE 3-16A Room fire at 5 minutes is still limited to paper in FIGURE 3-16B Fire at 6 minutes has spread to papers and the
wastebasket; HHR ⬍ 100 kW. Courtesy of BRE (Incorporating the Fire fabric on the sofa by direct flame impingement; HRR ⬍ 200 kW.
Research Station), Garston, United Kingdom, by permission. Courtesy of BRE (Incorporating the Fire Research Station), Garston, United
Kingdom, by permission.

FIGURE 3-16C Fire at 7 minutes has spread to draperies by direct FIGURE 3-16D Fire at 7.5 minutes. The draperies are fully
flame contact and radiant heating from sofa. HRR increasing very involved. Flames in the smoke layer are not sustained; HRR ⫽
rapidly as vertical flame spread dominates; HRR ⫽ 1,000 kW (1MW). 2,000 kW (2 MW). Courtesy of BRE (Incorporating the Fire Research Station),
Courtesy of BRE (Incorporating the Fire Research Station), Garston, United Garston, United Kingdom, by permission.
Kingdom, by permission.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 57

FIGURE 3-16E Fire at 9 minutes. Radiant heat ignition of tops of FIGURE 3-16F Fire at flashover transition (10.75 minutes). All
furniture, and the flames in the smoke layer become continuous contents are on fire, and there is a large plume exterior to room;
(rollover); HRR ⫽ 2.5 MW. Courtesy of BRE (Incorporating the Fire HRR ⫽ 5.2 MW. Courtesy of BRE (Incorporating the Fire Research Station),
Research Station), Garston, United Kingdom, by permission. Garston, United Kingdom, by permission.

FIGURE 3-16G Post-flashover fire at 11.75 minutes. The larger FIGURE 3-16H Decaying fire at 13 minutes. Massive furniture
furniture is fueling the majority of fire; HRR ⫽ 2.5 MW. Courtesy of (sofa on right and cabinet with folded draperies on left) is still on
BRE (Incorporating the Fire Research Station), Garston, United Kingdom, by fire; HRR ⫽ 1.4 MW. Courtesy of BRE (Incorporating the Fire Research
permission. Station), Garston, United Kingdom, by permission.

FIGURE 3-16I Decayed

fire at 15 minutes; HRR ⫽
900 kW. Courtesy of BRE
(Incorporating the Fire
Research Station), Garston,
United Kingdom, by permission.

58 Chapter 3 Fundamentals of Fire Behavior and Building Construction

 6000 FIGURE 3-17 Heat-
Flashover release curve for living
room fire in Figure 3-16.
5000 Courtesy of BRE (Incorporating
the Fire Research Station),

Heat release rate (kW)

Garston, United Kingdom by
4000
permission.
Flames in smoke layer
3000
Fire dies down
Curtains alight Sideboard fully alight
2000
Fire hoses applied

1000

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Time from ignition (min)

layer) for a typical room fire as a function of time, we might see something like Figure
3-17. As each item (chair, sofa, carpet, etc.) ignites, it contributes to the hot gas layer,
which is continuously being depleted by conductive losses through the ceiling and
walls, radiative and convective losses, and mass flows out doors, windows, and vents.
When enough heat is released quickly enough into a room, the room can go to
flashover. It is obvious that the higher the heat release rate (HRR) of significant fuel
packages, the better the chances that their additive nature can make the development
of flashover possible. If we examine various fuel packages using a furniture calorime-
ter (see Chapter 11), we can estimate their respective contributions. Organizations
such as the Building and Fire Research Laboratory at the National Institute for
Standards and Technology (NIST) and the California Bureau of Home Furnishings
have conducted hundreds of tests to examine the HRR of various items. Depending on
their covering and padding, upholstered chairs have a maximum HRR of 150 to 700 kW,
sofas: 250 to 3,000 kW, and mattresses: 150 to 2,800 kW.19 Other fuel packages can
play a big part: Christmas trees: 800 to 1,600 kW,20 wood shelves with 100 videocas-
settes: 800 kW,21 a pool of 1.9 liters (L) of camp fuel: 900 to 1,000 kW.22 If one large
(700 kW) or several small (150 to 250 kW each) packages reach their maxima at or
near the same time, the average-dimensioned room can go to flashover. A complica-
tion is that external radiant heat can increase the HRR of a typical package, so an
encompassing fire involving multiple fuel packages can grow at a faster rate than a
simple additive process would indicate.
The development of a well-ventilated, nonaccelerated fire in a furnished room is
shown in Figure 3-16. The room was set up with a modern (flame-retardant) sofa (meet-
ing Home Office Statutory Instrument 1324-1988 or BS 5852) with one wall open and
venting beneath a large calorimeter. The fire was set in papers in a wastebasket just in
front of the sofa. Figure 3-16a shows the incipient fire, still confined to the wastebasket
after 5.0 minutes of burning. One minute later, the front of the sofa has ignited and the
fire is growing (Figure 3-16b). At 7.0 minutes (Figure 3-16c), the draperies have ignited,
and within 30 seconds, the draperies have spread the fire across the rear of the room
(Figure 3-16d). By 9.0 minutes (Figure 3-16e), the smoke layer is beginning to ignite
(rollover), and most of the fuel packages in the room have ignited by radiant heat. By
10.75 minutes, the fire has reached flashover, in which all the fuels in the room are on
fire, and a large plume of waste gases is venting out of the room and igniting outside the
room (Figure 3-16f). At this point (post-flashover), the maximum heat release rate of 5.2 MW
was recorded by the calorimeter. By 11.75 minutes, the majority of the fuels have been
Chapter 3 Fundamentals of Fire Behavior and Building Construction 59
 burned, and the fire is going into a decay phase (Figure 3.16g). At 13.0 minutes, many of
the fuel packages have burned out except the remains of the sofa in the right rear corner
and the wood credenza on the left (which contained a number of folded draperies) (Figure
3-16h). The fire continued to burn for some 25 minutes before extinguishment. The heat
release curve of this test, Figure 3-17, shows the various stages described. Due to the pro-
longed growth phase (~10 min), the post-flashover steady-state burning was very brief, as
much of the fuel had already been consumed, but the decay phase is well defined. The
increase at 17 to 20 minutes is due to the delayed combustion of a wood cabinet filled
with combustible draperies.
It should be noted that the flashover transition triggers several significant effects in
the room, which may or may not be visible to an observer. Starting with a hot gas layer
temperature that exceeds 600°C (1,100°F), the temperatures throughout the room spike
quickly, especially at floor level (as floor coverings ignite). The overall oxygen content in
the room drops dramatically as the ventilation limit is reached and exceeded. This causes
a dramatic increase in the concentration of CO (and also of HCN somewhat later). The
radiant heat flux increases on target surfaces in the lower portions of the room. This trig-
gers ignition (or at least off-gassing) of exposed fuels. Their subsequent ignition to flames
depends on their individual thermal inertias.23 When monitoring compartment fires, it is
not uncommon to observe these events at slightly different times. The “time” of flashover,
then is not a single exact point but is largely dependent on what indicator one is relying
on. Window glass tends to shatter as a room fire approaches flashover, and the rustle and
popping of a limited room fire becomes a violent rumble due to the turbulence of the post-
flashover combustion. These events are audible to bystanders.
The approximate HRR needed to trigger flashover can be calculated if the dimensions
of walls, ceiling, windows and doors, and insulative properties are known. The simple
calculations in FPETool or in the NRC Fire Dynamics Tools spreadsheets can be used.24
The Fire Dynamics Tools spreadsheets include the following three calculations:
Babrauskas:
#
Q FO ⫽ 750AO (hO ) 0.5
Thomas correlation:
#
Q FO ⫽ 7.8 AT ⫹ 378 AO (hO ) 0.5
McCaffrey/Quintiere/Harkleroad (MQH):
#
Q FO ⫽ 610 冤hK AT AO (hO ) 0.5冥0.5
Where:
#
Q ⫽ heat release rate needed to include flashover
FO
AT ⫽ total surface areas of walls, floor, and ceiling (m2) (not including areas
of windows or doors)
AO ⫽ area of opening (m2)
hO ⫽ height of opening (m)
hK ⫽ convective heat transfer coefficient (for specific wall covering)
For a room 3 m wide ⫻ 4 m long ⫻ 2.4 m high (10 ft ⫻ 13 ft ⫻ 8 ft) with a single
door 0.8 m (32 in.) wide ⫻ 2 m (78 in.) high, these estimates are
#
Babrauskas Q FO ⫽ 1,697 kW
#
Thomas Q FO ⫽ 1,292 kW
#
MQH Q FO ⫽ 2,437 kW
There is a spread (1,300–2,400 kW) of estimated HRR needed for flashover, since each for-
mula is based on different experimental data and assumptions about proportions (room
60 Chapter 3 Fundamentals of Fire Behavior and Building Construction
size versus vent size) and thermal properties, so if the HRR calculated or estimated from
the fuel packages present in a room falls anywhere in the 1,300–2,400 kW range, the possi-
bility that that room will be driven to full involvement has to be considered.
Any fire is a chemical balancing act among fuel, air, heat, and the rate at which the
fuel is being consumed. If the ventilation is limited, the fire will progress at a slower rate,
the temperature rise will be more gradual, the smoke production will be enhanced, and
the ignition of the airborne fuel (smoke layer) may be delayed until sufficient air is avail-
able for its open-flame ignition to occur. This ignition may not happen within the room
at all; the fuel-rich smoke will then be seen igniting as it leaves the door and window
openings to mix with the oxygen-rich air outside. This combustion can be sustained if a
large enough fire is burning within to maintain a flow of smoke and gases. If the fuel
does not burn quickly enough or does not produce enough heat when it does burn, heat
losses, through convection and conduction, will prevent the fire from going to full
involvement. One interesting result of observing room fires ignited with a significant
(one liter or more) amount of a flammable liquid is that if the liquid is distributed over
a sufficient number of ignitable substrates, the radiant heat or direct flame contact from
the liquid-fueled fire can cause ignition of all the exposed ordinary fuels in the room
within seconds of ignition without the formation of the hot gas layer. Under these con-
ditions, full room involvement has been achieved without the hot gas layer/radiant heat
flashover process.25
A well-insulated room is more likely to retain enough heat to develop a fully developed
fire than is one with highly conductive walls and ceilings. Not all room fires reach
flashover. If the ventilation openings are too small, of course, there will not be enough ven-
tilation to support a large enough fire to bring about flashover. If vent openings are too
large, the fire has to be very energetic to overcome the losses and form the hot gas layer
that is critical to flashover. If the hot gas layer never reaches its critical temperature, there
will not be enough radiant heat to involve the remainder of the room. Flashover is not a
specific moment or event but a transition from a fire that is growing by involving one fuel
package after another to a fire in which all the exposed fuel in the compartment is on fire.

POST-FLASHOVER STAGE
Once the fire has reached post-flashover burning, all involved fuels will continue to burn
as quickly as oxygen is made available to them until all available fuels are nearly con-
sumed. It is not a threshold that can be crossed again and again. A post-flashover fire in
a typical room is usually a ventilation-controlled fire. The heat release rate of the fire is
no longer controlled by how much of the exposed fuel is involved, because it is all on fire.
The size of the fire from that point on may be limited by the amount of air that can enter
the compartment. This is usually determined by the buoyant flows of hot gases and air
into and out of the room. The area, height, and arrangement of ventilation openings into
the room determine that flow, as noted in the relationship
# #
Q max ⫽ ¢Hair ⴢ m max ⫽ 3,000(0.5Ao (ho ) 0.5 ) ⫽ 1,500Ao (ho ) 0.5
Where:
¢Hair ⫽ 3,000 kJ> kg
˛ ˛

#
m max ⫽ maximum buoyant flow
AO ⫽ area of opening m2
hO ⫽ height (m)
A single entry door (0.8 m wide and 2 m tall) can allow enough air in (and enough
waste gases out) to support a nearly 3 MW fire in a room.26 Windows positioned high up
in a wall will be largely filled with buoyant waste gases (including products of incomplete
Chapter 3 Fundamentals of Fire Behavior and Building Construction 61
 combustion) escaping from a room and will not be as effective in allowing fresh air in as
windows placed lower down in the walls would be. In a post-flashover fire, the locations
of the most energetic combustion (and accompanying thermal effects) are no longer con-
trolled by where the fuel packages are but by where the best ventilation is located. Recent
tests by Carman have demonstrated that post-flashover rooms can exhibit marked venti-
lation effects dependent on where fresh air rich in oxygen can enter a room or where the
fire is burning in a continuously oxygen-deficient area. Entrained flows of air through a
doorway can produce “clean burn” fire patterns on walls opposite the door, mimicking
patterns from localized fuel packages.27 Such effects can also produce burn patterns across
carpets from doorways or windows. Areas where oxygen can enter the room and mix
with the combustion gases will support the most intense combustion. The mixing is
extremely turbulent to the point at which it is not a steady, torchlike flame but moves about
in a chaotic fashion.28 This flame behavior can produce irregular heat effects on nearby
surfaces. If a post-flashover room fire is suddenly deprived of oxygen (such as by struc-
tural collapse), flaming combustion will rapidly decrease to return later when more air is
available, but it cannot return through flashover. Some of these effects will be illustrated
in Chapter 7.

DECAY STAGE
Eventually, the available fuel is exhausted, and open-flame burning gradually becomes
less and less prevalent. (During the fire, if the oxygen concentration drops much below
16 percent, open-flame combustion will decrease even in the presence of unburned fuel
and may even cease completely at oxygen concentrations below 5 percent.) As an ade-
quately ventilated fire dies down the amount of fuel vapors that remain to be pyrolyzed
out of the solid fuels decreases, and the amount of flaming combustion dies back. As a
result, smoldering combustion becomes predominant, which can continue as long as there
is suitable fuel. This decay stage is often called the “smoldering” phase of a room fire.
High temperatures may continue as long as there is suitable fuel, depending on the venti-
lation and insulation present. Some products continue to pyrolyze to form carbon monox-
ide and other gaseous fuels as well as toxic gases, solid soot, liquid aerosols, and other
fuels in the form of smoke.
If the room is not adequately ventilated, these combustion products can build up to
form an ignitable vapor mixture. When a supply of fresh air is introduced in the presence
of an ignition source, such vapor accumulations can deflagrate into a secondary fire,
sometimes called a backdraft, flashback, or smoke explosion. This ignition can proceed
with explosive speed, and although the pressures produced are low compared with fuel
vapor–air deflagrations, they can produce pressures on the order of 0.5 to 1.5 psi (3 to
10 kPa, 30 to 100 mbar) or higher.29 This pressure is enough to do structural damage and
endanger lives. Although such smoke explosions are rare, they can occur, especially when
smoky fuels such as latex or urethane foam rubber are present. An excellent study by
Fleischmann, which explored the dynamics of these events, showed that even under con-
trolled conditions, ignition of the layer could not be assured, since it involved having the
right fuel-air mixture in contact with a suitable flaming ignition source.30 One well-
documented case involved a relatively small but smoky fire that filled an apartment with
a hot, fuel-rich smoke. When fire crews entered the apartment, the smoke leaving the
apartment ignited in the stairwell, trapping three firefighters.31 The use of positive-
pressure ventilation (PPV) in firefighting may introduce enough fresh air to a hot, fuel-
rich environment to trigger such events. Testing by NIST has shown that PPV used prior
to suppression causes higher temperatures and heat fluxes in the fire room and in areas
“downstream” of the vent from that room.32
This generalized sequence does not occur in every fire. The entire fire may be ignited
by a smoldering process, which then has to grow to the point at which it has a high
62 Chapter 3 Fundamentals of Fire Behavior and Building Construction
enough heat release rate to initiate and sustain flaming combustion. Thus, smoldering
should not be considered only to be the final end product of a fire. Glowing combustion
can also take place at the same time as flaming combustion. On wood, for instance, if the
pyrolysis and volatilization occur at such a rate that oxygen can reach the hot solid char
surface, glowing combustion can take place. This situation is noted most often in strong
draft conditions where the impressed airflow can push aside the outflowing gases, or late
in a fire in which the rate is dropping due to exhaustion of the fuel. An example of such
a fire was shown in Figure 3-6.

Flow of Hot Gases

Buoyant flow is the dominant mechanism by which fire gases move. Normal air (and
most smoke) has a density of 1.2 kg/m3 at 20°C (293 K). Its density (␳) then varies with
its absolute temperature (from the ideal gas law, introduced in Chapter 2):
␳T1>␳T2 ⫽ T2 (K)>T1 (K)
Thus, the density of gas at flame temperature (800°C; 1073 K) is
␳T2 ⫽ 1.2 (293>1073)

␳T2 ⫽ 0.33 kg>m3

Flame gases then will rise rapidly due to their much-decreased density. As the gases cool,
of course, they lose some of that buoyancy in proportion to their cooling. The horizon-
tal flow of a gas (or other fluid) due to gravity-driven forces is called advective flow. As
a hot gas flows along a ceiling it loses its heat to the underside of the ceiling by convec-
tive transfer. Thus, its temperature (and its convective heat transfer rate) is at a maxi-
mum where it first contacts the cold ceiling. The horizontal, buoyancy-driven advective
flow along the underside of a ceiling is called the ceiling jet. When that jet hits a verti-
cal surface (like a wall), it piles up against it like water waves pile up against a breakwa-
ter (sometimes called momentum flow or “mushrooming”). If it flows under a door
header or window soffit, it leaves a weak turbulent flow on the downstream side of the
soffit (or other barrier), as shown in Figure 3-18. If the flow takes it outside, it forms a
door or window plume, which then continues its vertical rise driven by its buoyancy. As
the plume rises vertically, entrainment can push it back against the vertical wall above
the opening, especially if there is no balcony or other horizontal projection to push it

Eddy flow
Ceiling jet

Momentum flow

Momentum flow
FIGURE 3-18 Buoyant
(mushrooming)
flow against a vertical
barrier causes downward
momentum flow (mush-
rooming). Eddy flow on
other side produces a
protected area.
Doorway

Chapter 3 Fundamentals of Fire Behavior and Building Construction 63

FIGURE 3-19
Entrainment of window
plume against vertical
High heat
wall is reduced by a flux
balcony or “eyebrow” Entrainment
above the opening.

Flames
Flames

Window plume Window plume

Airflow in Airflow in

away from the wall (see Figure 3-19). Such plumes, if they are flaming, can quickly cause
failure of the spandrel panels or windows directly above the opening through which they
vented ■ The fire has vented, as shown in Figure 3-20. This is the mechanism by which many high-rise fires
extended outside the inflict damage on multiple stories. Fires such as the Caracas Parque Central (2004) (16
structure or compart-
stories involved) or Madrid’s Windsor building (2005) (12 stories involved) are exam-
ment by destroying the
windows or burning an ples.33 A similar fire venting from the Cook County Building in Chicago (October 2003)
opening in the roof or (Figure 3-21) did not spread upward from its 12th floor of origin because the windows
walls. of each floor were set back from the face of the building by one foot or so. That inset
protected the windows on floors above. The total heat flux against a wall or window
above such an entrained window plume from a post-flashover fire can be as high as 80

FIGURE 3-20
Entrainment of window
plume on sheer wall
causes rapid failure of
window immediately
above. Courtesy of John D.
DeHaan

64 Chapter 3 Fundamentals of Fire Behavior and Building Construction

 FIGURE 3-21 Deeply
inset windows on this
Chicago high-rise limited
damage to floors above
origin to smoke. Courtesy
of NIST.

to 120 kW/m2, well above the 20 kW/m2 needed to crack ordinary window glass or the
35 kW/m2 needed to cause the window to fall out or the 20 to 40 kW/m2 needed to
ignite most combustible materials.34

Effects of Environmental Conditions

The investigator of fires will usually be concerned with the role played by such environ-
mental conditions as the outdoor temperature or humidity that preceded the fire. These
factors need to be taken into account at times and should be documented. It is well known
that when an area has become excessively hot and dry, there is greatly increased hazard
of forest and grass fires, and structural fires may be more common or of greater severity.
Wetness, an environmental condition subject to rapid alteration in a local region (e.g.,
water from fire hoses), is also a real factor in the rate of burning, the danger of ignition,
and the intensity of the resulting fire. Although these factors may be secondary to the
investigation, they merit careful consideration by the investigator, since they sometimes
can affect the ignition, rate of fire spread, and even the post-fire burn indicators.

TEMPERATURE
All other things being equal, fires are more likely to become severe on a hot rather than
on a cold day. However, some of the most destructive structure fires have occurred in
weather when the water from fire hoses froze on the firefighters’ ladders. The reason for
the greater fire hazard on a warm day is easily explainable.
1. Heat dehydrates; therefore, fuel in a heated environment is dryer (has a lower mois-
ture content) than in an unheated one. Dry materials ignite more readily than damp
ones, although the heat from the fire itself once established is usually sufficient to
dry out most materials ahead of the flames when there is no renewal of the wet
condition.
2. As developed elsewhere in this chapter, the rate of a combustion reaction (such as
fire) is approximately doubled with every rise of 10°C. In a fully developed fire, this
factor in the initial environment is negligible because the heat from the fire itself is
Chapter 3 Fundamentals of Fire Behavior and Building Construction 65
 so much greater than an ambient temperature. However, the warmer the environ-
ment, the faster the reaction in the extreme initial (incipient) portion of the fire,
when the fire’s own heat is still not so overwhelming. In other words, it is easier to
ignite a fire with hot than with cold fuel.
The variation of temperature in normal interior living environments has little influ-
ence on most fires simply because the variations are not great enough, nor the influence
direct enough, to make any appreciable difference. The pre-fire temperature may have
great influence on self-heating processes, the vapor pressure of flammable liquids, and
even the distribution (mixing) of vapors. For these reasons, the pre-fire temperatures
(interior and exterior) should be documented by a thorough investigator.

HUMIDITY
Firefighters tend to be greatly concerned with humidity because they note a greater inci-
dence of fires on dry days than on wet days. Again, this observation refers far more to
exterior fires than to interior ones. When it is hot and dry for protracted periods, it is well
known that wildland fires become an extreme menace; because the fuel moisture increases
markedly on cool, humid days, there is far less hazard. The tendency to apply the same
qualitative reasoning to interior structural fires is often a misguided one, although not
always. The question as to the actual importance of humidity is not a simple one to
develop, because it involves several secondary factors.
Humidity refers to the amount of water vapor in the air. At saturation, this value is
determined by the air temperature. The greater the temperature, the greater the amount
of water that can be held by the air without condensing into fog, dew, or rain. Thus, it is
not so much the absolute humidity that is relevant as the relative humidity, that is, the
percentage of the saturation value that is actually present. A relative humidity of 50 per-
cent means only that the air contains half of the saturation value for that temperature.
However, if the temperature rises without any change in moisture content, the relative
humidity drops to a lower percentage, without changing the total moisture. For practical
purposes, it is always the relative humidity that is quoted and of importance because this
is the best measure of the moistness of the air at any given temperature.
The effects of humidity on fire stem from several basic factors. When porous objects
are exposed for sufficient time to low relative humidity, they dry out (lose internal mois-
ture) and are therefore more combustible than when less dry (see next section on fuel
moisture). However, the difference in fuel moisture content is not so great as is generally
believed; the moisture content of many materials, such as paper, varies within surprisingly
small limits with ordinary changes in humidity. The relationship between humidity and
fuel moisture content is shown in Figure 3-22 for fine cellulosic materials. Fuels outside
are also exposed to wind, which causes lower fuel moisture content than the temperature

25

20
moisture content (%)
Equilibrium

15
FIGURE 3-22
Approximate relationship 10
between moisture con-
tent of fine cellulosic fuels 5
and relative humidity at
room temperature (25°C).
Courtesy of Vytenis Babrauskas, 0
0 20 40 60 80 100
Fire Science and Technology.
Relative humidity (%)

66 Chapter 3 Fundamentals of Fire Behavior and Building Construction

and exterior air humidity alone would dictate. Perhaps the greatest effect from this stand-
point is on living vegetation, a major fuel of forest fires. Plants not only dry out as would
a sheet of paper, but, much more important, they transpire moisture from pores in the leaf
structure. Despite the closing of these stomata in dry weather, a leaf may lose more mois-
ture proportionately than a sheet of paper when exposed to drying conditions. The
process is enhanced by wind, as the moving air forces faster moisture loss. The difference
is clear when it is noted that dry periods correlate with increased wildland fires more than
with increases in primary structural fires. For ignition to occur, the small amount of
energy of some ignition sources has to be transferred to the fuel to trigger a self-sustaining
reaction. If some of this energy goes into evaporating moisture from the first fuel,
there may not be enough left to promote ignition. Storage and watering conditions of
cut Christmas trees have been shown to be critical to their relative risk of ignition.35 One pub-
lished set of tests showed how difficult it was to ignite Christmas trees using modest igni-
tion sources (such as matches) when the tree had been thoroughly hydrated.36
Humidity can also play a role in self-heating processes, for example, as in the fuel
moisture content of agricultural materials, where too little will preclude the initial growth
(and heat generation) of microorganisms. An excess may play a part in promoting self-
heating in charcoal and other materials. These mechanisms will be discussed in more
detail in Chapter 6.
Another effect that is rarely given its proper significance is the relation of humidity to
small electrical arcs resulting from static discharge and from small sparks produced by
striking materials such as a steel object with a rock. Static electricity tends to accumulate
excessively in dry periods. This is not due to the insulating value of dry air (since moist air
actually has a higher dielectric constant than dry air). Rather, the charge dissipates by
leakage via surface moisture when the relative humidity is high, so it rarely reaches the
potential necessary to produce a visible arc. Mechanical sparks tend to be more vigorous
in dry atmospheres, since dry air conducts heat more poorly than moist; therefore, the
particle retains its heat longer. The spark can also more readily ignite dry fuel materials
than moist ones that it may contact.
Most important to understand is that once the fire is actually burning, the humidity
alone has little or no effect on it. The effects are virtually confined to the primary igni-
tion, not to the magnitude of the ensuing blaze. The reason is rather obvious, namely, that
the heat generated by the fire is so great that all humidity considerations of the fire envi-
ronment are totally overshadowed. It is also true that when fuel in a humid environment
is exposed for a considerable period of time to localized heat, it is dried by that heat,
regardless of the ambient humidity.

WETNESS OF FUEL (FUEL MOISTURE CONTENT)

Fuel moisture content is not the same as humidity, which relates to wetness of air. The
former is of far greater importance than humidity when the degree of wetness exceeds
that which is in equilibrium with the air (controlled partially by humidity). An object that
is exposed to air reaches equilibrium (fuel moisture) with it and is stated to be “dry” as
long as excess moisture in the form of rain, dew, or water from an artificial source is not
applied to it. Although all “dry” fuel materials burn readily, the excess moisture of wet
fuel must first be evaporated before the fuel can be raised to its ignition temperature.
Because 4.4 kJ is required to evaporate one gram of free water, it is apparent why so much
extra heat must be applied to wet fuel before it can ignite.
Wetness of fuel must not be confused with humidity, because they are totally differ-
ent even though related in some aspects. Dew or even heavy fog tends to moisten and
accumulate on surfaces and makes them hard to ignite. Fuel that is wet internally (high
fuel moisture) from absorption of moisture is even more difficult to burn than fuel mois-
tened on the surface only. The principle is identical, however, in that the water boils out
Chapter 3 Fundamentals of Fire Behavior and Building Construction 67
 at a temperature far below the ignition point and keeps the fuel temperature low only
until there is no more free water to boil. Because of the excess water content of a water-
logged piece of wood, there is little comparison between it and a piece of wood that has
stood in humid air until it is in equilibrium with it, that is, is “dry.” Preliminary studies
indicate that measuring the moisture content of wood or char after a fire may give an indi-
cation of whether prolonged heating occurred prior to its ignition (such as in the forma-
tion of reactive charcoal) or whether fuel moisture conditions were suitable for ignition
by low-energy ignition sources such as cigarettes or sparks from “hot work.”37 The pre-
fire humidity in the area of ignition of grass and other vegetation should be documented,
since it may greatly affect ignitability.
Tests by the California Bureau of Home Furnishings on dry Christmas trees produced
maximum heat releases on the order of 800 kW, but a similar tree thoroughly wetted
before ignition had a peak heat release rate of only about 200 kW.38 Tests published by
NIST on tests of Scotch pine Christmas trees verified these results.39 This effect of fuel
moisture is true whether the fuel is a solid fuel with water present or complex organic
materials such as meat or fat, which incorporate water as part of their structures. If suf-
ficient heat can be applied to drive off this water, meat or fat can certainly burn (as any
hapless barbecuer can attest). The effect of excess water is the basis for fighting fires with
water. If water can be added faster than the fire can boil it away, the fire eventually will
be extinguished because the temperature is lowered. The steam generated by putting
water on a large fire in a room may be sufficient to displace enough air in the vicinity of
the flames to further reduce the combustion rate. Some exceptions exist. Some fuels, such
as magnesium metal fires, cannot usually be fought with water because water on hot
masses of magnesium generates the excellent fuel hydrogen gas, and its combustion
increases the size of the fire spectacularly.

WIND
Wind is usually considered to be a factor only in fire spread in wildland fires, but a strong
wind can affect fire development in a building, particularly after windows begin to fail.
Open doors can admit enough wind to promote more horizontal spread than would be
expected from normal dynamics, and additional movement can be induced as windows
break by thermal shock. In some cases, the induced wind through the structure can over-
whelm normal spread. In one documented fire test in a large wood-frame building, the
area of origin was actually less damaged than were other areas where the fire was driven
and then enhanced by a strong wind blowing through window and door openings.40
Recent tests by NIST have demonstrated that even moderate external winds entering a
high-rise building through an open window cause much higher temperatures and heat
fluxes in the fire room and quickly spread hot gases and smoke into adjacent hallways.41
As mentioned previously, wind causes drying of living grasses and forest products
beyond what would result from high temperatures alone, so wind conditions in the hours
and days prior to a wildland fire may be critical factors in ignition. The careful investiga-
tor should establish the wind, temperature, and other weather conditions that existed at
the time of the fire.

OXYGEN CONTENT
Although most fires occur in normal oxygen concentrations (20.9 percent in air at sea
level), oxygen concentrations may play a role in both ignition and fire growth. Oxygen
concentrations below 15 percent in air will retard ignition and fire growth.
Concentrations below 10 percent will make flaming fire growth impossible (except in
post-flashover fires), but smoldering fires will propagate even at oxygen concentrations
below 3 percent. Elevated oxygen concentrations will dramatically enhance ignitability of
fuels (even by weak ignition sources) and cause very rapid flame spread. Fuels that are
68 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 FIGURE 3-23 Plastic
O2 cannula burning with
bright flame. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

normally difficult to ignite, such as rubber insulation or petroleum lubricants, are very
susceptible to ignition in oxygen environments. Fires in surgical or dental facilities may
be very severe in areas (such as wall cavities) where plastic oxygen lines fail due to fire
exposure. Ignition of oxygen-saturated surgical drapery, dressings, clothing, bedding, and
even hair during surgeries has been reportedly caused by small electrical arcs in instru-
ments such as cauterizers. The transition from smoldering to flaming combustion is dra-
matically enhanced by excess oxygen. Recently, smokers using supplemental (medical)
oxygen have discovered that a dropped cigarette can readily ignite clothing or bedding,
sometimes with tragic consequences (Figure 3-23).

Elements of Building Construction

It is very important for an investigator to be familiar with buildings—how they are con-
structed, what materials were used in their construction, and how they react to fire. The
manner and materials of construction can have an enormous effect on manner and rate
of fire spread as well as on the nature of post-fire patterns of damage.

STRUCTURAL SHELL
The outermost portion of a building is often called the structural shell. The structural
shell is the outside or supporting portion of a structure, whether it is residential, business,
or industrial in use. It includes especially the outer walls and supports and occasionally
the columns and steel skeletons at the center of the building, which are sometimes used
in modern construction as supporting structures. The materials, arrangement, and inter-
connections of these structural elements can have major effects on the spread of fire and
the resulting collapse of burning buildings. The most comprehensive source of useful
information in these disciplines is Francis Brannigan’s Building Construction for the Fire
Service, and the reader is referred there for a detailed review.42 Structural shells can be
considered to have one of the basic constructions seen in Figure 3-24 and discussed next.
Masonry
In this category are concrete (cast and block), tile, brick, and similar materials, even
adobe. Shells constructed of masonry are clearly not flammable and will not directly con-
tribute to a fire. To assume that they are not subject to fire loss is a mistake, because
Chapter 3 Fundamentals of Fire Behavior and Building Construction 69
 Ceiling joist

Finished floor
Subfloor
Ordinary
Pocket
construction
Floor girder

Support
Concrete Brick
column
block outer
Footing
Footing Masonry
load-bearing Rafter
wall

Joist
Header Top plate
Sheathing
Wall studs
Corner
post Wall stud
Wood frame Exterior Sole plate/toe plate
covering Rough floor
Header Joist
Sill
Sill
Foundation Foundation

Footing
Footing
End view
Side view Web members
or ties (2 × 4)
Core (hollow or filled) Lightweight
prefabricated
trusses
Heavy timber
Shingles
1/2-in. Gang nails Bottom
Finished floor plywood chord
Lightweight
Planking wood frame
Heavy timber Wire and Wall stud
stucco
Floor girder coating Toe plate
Support columns or piers Mud sill
Concrete slab
Footing Masonry Roof (foam, tar,
load-bearing wall rubber membrane)
Steel or wood Roof deck (wood,
Parallel chord
trusses metal, concrete)
truss
Masonry
Tilt-up Precast concrete Steel wall
panels column

Concrete slab

FIGURE 3-24 Typical ordinary, wood frame, and heavy-timber construction, and commonly used terms.

70 Chapter 3 Fundamentals of Fire Behavior and Building Construction

within such shells there are ordinarily walls, floors, and other structures of wood, com-
bustible furnishings, and other combustible materials.
Masonry walls are found both in heavy-timber (industrial) structures and in ordinary
construction of many residential and light commercial structures. They can form the
structural shell, supporting the ends of beams and girders, as well as the load-bearing
walls within the structure. Masonry walls can be two interconnected wythes or thick-
nesses, with a hollow space in between, or the space may be filled with insulation, mor-
tar, cement, or rubble, or even be a single thickness. Hollow masonry walls offer an
opportunity for flames and hot gases to travel upward within a building.
A fire that develops within such a masonry shell may actually be more intense than if
the walls were directly involved because of the insulation effect and chimney action that may
be present. Some of the worst conflagrations have involved this type of structure. The large
thermal mass of masonry walls, floor, and ceiling holds in the heat produced by fire within.
The higher temperatures can cause more complete combustion of the contents and inhibit
fire suppression activities. Furthermore, higher and sustained temperatures in nonreinforced
masonry can also lead to the premature unexpected failure of exterior supporting walls. The
remaining rubble stays hot for considerable periods due to this “oven effect.”

Wood
In some portions of the world, wooden shells for residences are the rule, and almost
everywhere at least some residential buildings are basically of wood. Business and
industrial buildings are less commonly constructed in this manner at present, but many
older buildings still in existence were so constructed. These may be referred to as “fire
traps” and often they are, but the reasons relate more to the interior construction, fin-
ish, and content than to the shell, which is rarely the point of fire initiation, regardless
of what material is used for its construction. Wood stud walls are particularly danger-
ous from a fire spread standpoint when they are used in balloon frame form. In these
multistory structures, the wall studs form a continuous cavity from floor to attic.
Balloon frame construction has not been used in the United States since the 1940s.
Platform construction minimizes fire spread, as each stud frame is no more than one
story, and there is separation between floor, wall, and ceiling cavities, as was seen in
Figure 3-24.
A special case of wood structural shells is that in which the exterior coating is some
form of plaster such as stucco. The presence of such material on the outside of the exte-
rior walls has little influence on the fire hazard, except that such a building is less likely
to ignite from an external source such as a brush fire or a fire in an adjoining building.
However, interior fires, which make up the vast majority of all structural fires, will be
quite independent of the existence of a stucco coating. The stucco and wire mesh cover-
ings on wood-framed structures may remain standing after the structural wood has
burned away. Manufactured or modular homes often have wood-framed exterior shells.
They will be discussed in more detail in a later chapter.
Metal framing and studs are being used as a substitute for wood framing. Top plates,
sole plates, sills, and studs are made of formed galvanized sheet steel. They interlock to
form direct replacements for wood framing. Sheathing (inside and out) is then affixed
using adhesives or screws. The performance of these wall panels in fire exposures appears
to meet or exceed those of traditional wood-framed walls.

Glass and Metal

Some modern buildings have shells that are virtually all steel with panels of aluminum
and glass hung between the vertical members to form the walls. Although the steel and
aluminum panels themselves have no intrinsic fire hazard, they are sometimes covered
with synthetic coatings of plastic sheet or insulating materials that can contribute greatly
to the fuel load of the building.
Chapter 3 Fundamentals of Fire Behavior and Building Construction 71
 The large glass panels are subject to rapid failure on direct exposure to flames from
fires in adjacent areas or lower floors. With their failure, the fire can ignite draperies or
furnishings immediately within, and the fire leapfrogs upward through the building. Such
failures are thought to be responsible for fire spread in tragic fires in high-rise buildings
like those in Madrid (2004), in the First Interstate Bank Building (Los Angeles, 1988), and
in the One Meridian Plaza (Philadelphia, 1990).43 The flexible plastic or rubber seals that
retain the metal and glass panels in such buildings can also fail on fire exposure, allow-
ing flames to penetrate the wall even if the panel itself remains in place. Similar failure of
sealants between floor decks and exterior walls can allow fire to move upward between
floors even when the windows remain intact and in place.
Sandwich Panels
A relatively recent innovation, especially in industrial and commercial structures, is the
use of sandwich panels of sheet steel or aluminum glued to both sides of a rigid insulating
material such as rigid polyurethane or polyisocyanurate foam or expanded polystyrene,
or onto a built-up frame with mineral wool. The panels can be fitted to a metal external
framework or interlocked together to form walls, corners, and ceilings. They offer easily
maintained surfaces, good thermal insulative capacity, and ease of construction and porta-
bility, all at relatively low cost. They are often encountered in food processing or refriger-
ated storage facilities (see Chapter 5). Panels may also be faced with fiberglass or rigid
plastic, gypsum wallboard, plywood, or chipboard. Depending on the type of foam used,
they can be subject to very rapid collapse if exposed to fire.44 (Figure 3-25.)

FIRE RESISTANCE RATINGS

The NFPA currently uses the following classes of construction per NFPA 220, Types of
Building Construction.45
Type I (Formerly “Fire Resistive”)
Only noncombustible materials are permitted for the structural elements of the building
(materials must meet relevant NFPA fire resistance specifications. Structural materials
include walls (interior and exterior), partitions, columns, floors, and roofs. The use of
some combustible materials is permitted for nonstructural elements such as roof coverings,
some insulation materials, and limited amounts of wood for interior finish and flooring.

FIGURE 3-25 Sandwich

Panel Building. Walls are
glued laminates of gypsum
wallboard, expanded
polystyrene foam, and
chipboard. Floor and
ceiling are reinforced with
stamped steel channels.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

72 Chapter 3 Fundamentals of Fire Behavior and Building Construction

Type II: (Formerly “Non-combustible”
Structural elements must be noncombustible or limited combustible (as permitted by the
local code) that are protected to have some degree of fire resistance (1 hour or 2 hours)
for exterior walls. Other components may be completely unprotected per the local code.
Type III (Formerly “Ordinary Construction”
or “Exterior-Protected Combustible”)
All or part of interior structural elements may be of combustible materials (or other mate-
rials as permitted by local code). Exterior walls must be of noncombustible materials. All
concealed spaces must be appropriately fire-stopped. They can be further classified as
Type III (211)-Protected, where floors and structural elements have a 1-hour fire resist-
ance rating, or Type III (200)-Unprotected, where there are no specified fire resistance rat-
ings for floors or structural elements.
Type IV (Formerly “Heavy Timber”)
All structural members (columns, beams, arches, floors, and roofs) are of unprotected
wood (solid or laminated) with large cross-sectional areas. No concealed spaces are per-
mitted. Noncombustible or limited combustible materials (minimum fire rating of 2
hours) can be used in both interior and exterior load-bearing walls. Brick and stone are
the most common wall materials. “Mill” construction often denotes floor supports
designed to fail during an extreme fire, leaving the supporting walls or posts intact.
Type V (Formerly “Wood Frame”)
Structural members are entirely of wood or other permitted materials. Exterior walls may
or may not be required to be fire resistive based on the horizontal separation distance to fire-resistive ■ A struc-
the adjoining structure. Type V construction is divided into two subclasses: Type V (111), ture or assembly of
materials built to pro-
where there is a 1-hour fire resistance throughout including the exterior walls, and Type
vide a predetermined
V(000), which has no fire resistance ratings except exterior walls when the horizontal sep- degree of fire resistance
aration distance is less than 10 ft. as defined in building or
fire prevention codes
(calling for 1-, 2-, or
INTERNAL STRUCTURE 4-hour fire resistance).
As used in this chapter, internal structure refers to floors, carpets and pads, ceilings, walls
and interior finish, and doors.
Floors
In multistory buildings especially, the material of which the floors are constructed is often
critical to fire development. Wood floors—however they may be finished—are likely to be
breached from beneath in any major fire, spreading the fire upward from floor to floor
and limiting the chances for isolating it before it grows too large. Concrete floors, fre-
quently used between stories of industrial and commercial buildings, tend to prevent the
spread of flames from story to story, thereby making containment of the fire much easier.
However, under sustained direct fire exposure, concrete floors can buckle and crack,
allowing fire to penetrate. Such decks can be lightweight concrete poured onto corrugated
steel panels or precast panels with internal steel reinforcement to provide strength and fire
resistance. Combustible forms of wood or plastic foam may also be left in place around
the concrete, accelerating its failure. The lowest floor of a structure will usually be
breached only when the fire begins on the floor itself (or in some open space beneath it),
when burning debris falls onto it from above, or when post-flashover radiant heat is
maintained for long enough.
Some residences built on steep slopes may have an elevated lower floor, allowing expo-
sure to exterior grass or trash fires. A thorough external examination of a burned building
will reveal such exposure problems. Even on flat land, the ground floor is sometimes raised
sufficiently to allow a fire to develop beneath it. The ventilation provided by the footings
or posts under such buildings (usually old frame dwellings) produce perfect conditions for
complete involvement of all wooden structural members. Such buildings can be destroyed
Chapter 3 Fundamentals of Fire Behavior and Building Construction 73
 virtually completely, leaving only charred stubs of the posts and the remains of a noncom-
bustible roof covering. Gas may escape under the floor, and trash or grass fires that are too
close to the building may also get under it, providing accidental hazards.
The material and construction of every floor can have a great influence on the rate of
fire spread and can contribute to the overall fuel content of the structure and therefore
must be considered whenever the overall fire pattern is being assessed. Floor decks com-
prised of plywood or chipboard over manufactured wood trusses have been shown to fail
in as little as 10 minutes of direct flame contact.46 Manufactured floor joists that use ply-
wood or oriented strand board (OSB) can fail in less time than will joists of solid wood.
These considerations are important when considering the time of development and spread
in a building. Fire penetration rates of wood floors are discussed in Chapter 5.
Babrauskas has compiled extensive references on fire-related failures.47 Floor coverings
vary greatly in their fire resistance. Tile, terrazzo, and heavy linoleum are highly resistant,
while carpet, bare wood floors, and some vinyl floor coverings can burn more easily from
above or beneath.
Carpets and Pads
Although not part of the structure itself, carpet and its padding are often overlooked as
potential contributors to the fuel load. As illustrated in Chapter 5, carpet is made of a
fiber pile woven or glued to a backing. The pile may be of a natural fiber like wool or
cotton or, more commonly, a synthetic like nylon, polyester, or acrylic fiber (or a blend
of two or more different types). The backing was traditionally made of a natural fiber
like jute, but today it is much more common to find a synthetic fiber, like polypropy-
lene, which is characterized by its white, flat, ribbonlike appearance. These carpets may
have a rubber or urethane-foam backing affixed or may be installed over a separate
fiber, butyl latex rubber, or urethane foam pad. A plastic-film moisture barrier may be
integral with the pad or the carpet backing. These changes affect the spread of fire and
the contributions to the fuel load and heat release rate of the furnishings. Wool and cot-
ton carpets require high heat fluxes to sustain their ignition and tend to smolder rather
than burn with open flame. Synthetic carpet piles tend to ignite more easily than natu-
ral fibers (and some, not intended for interior use, will ignite very readily) and tend to
support open-flaming combustion. As a room fire approaches flashover conditions car-
pets will ignite and add to the heat release rate of the contents. When carpets are made
with a jute backing, the jute tends to char but remains largely in place, offering some
protection to the padding underneath. When a polypropylene woven backing is used,
being a thermoplastic [melting point 165°–186°C (330°F–365°F)], it melts when heated.
This allows the entire carpet to melt and shrink, exposing the underlying pad to the radi-
ant heat and promoting its rapid involvement in the fire. As both synthetic carpet and
pad melt and burn in a post-flashover fire they tend to form molten puddles of burning
plastic, creating localized and irregular patterns of floor damage surrounding areas of
undamaged floor. Such patterns can be mistaken for the pool/halo patterns formed by
burning ignitable liquids, as shown in Figure 3-26.48 The localized flames from such
floor coverings can be sustained long enough to melt and oxidize aluminum floor mold-
ings. It is strongly suggested that a comparison sample be taken of the unburned carpet
and pad in any room suspected of being an origin. This comparison sample can be tested
to establish the type of floor covering used (content, depth, construction), which may
influence its contribution to the fire as well as reveal the volatile residues it can con-
tribute to the debris.
Ceilings
In multistory buildings, the same considerations for floors also apply to ceilings, because
the ceiling of one floor is often the lower side of the floor of the story above. One very
important difference exists, however. Because hot gases rise and fires burn upward pref-
erentially, the ceiling is always exposed to heat when any substantial fire exists in a room.
74 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 FIGURE 3-26 Brief post-
flashover fire caused the
synthetic carpet to melt
and burn in a highly irreg-
ular pattern. Courtesy of
Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

If the ceiling is noncombustible, as with gypsum board, the fire is inhibited from burning
through to the story above. As a typical accidental fire develops in a room the buoyant
hot gas layer induces more damage to the upper one third of the room prior to flashover.
A combustible ceiling is exposed continuously to the hottest gases from the fire and is
quickly breached. In room fire tests conducted by the authors, the time to penetrate ceilings
has been observed as follows: 1/2-in. plywood—5 to 6 minutes, mineral fiber suspended
(T-bar) ceiling—5 to 10 minutes, wood lath and plaster—10 to 15 minutes, 1/2-in. drywall—
15 to 20 minutes of direct flame contact.
The situation with floors is different in that the buoyant flow takes the hot gases
upward and the radiant heat from the flame plume heats more fuel above the original
fire, so the fire burns upward, away from the floor, with the hot gases gathering in the
upper portions of the room. Because the floor is exposed to less heat, it is much less
likely to be breached during the growth of a fire involving ordinary combustible fuels,
except where falldown of burning debris occurs. A fire accelerated with an ignitable liq-
uid, however, may generate pockets of burning fuel that may char and even penetrate the
floor (around joints and seams). In post-flashover room fires where the floor covering is
ignited by the radiant heat, the heat fluxes and temperatures at floor level meet (and
often exceed) the high temperatures that previously existed only near the ceiling. At this
point damage to the floor becomes much more severe. In cases where the post-flashover
burning has not occurred or has not been prolonged, then the fire damage to the floor
is normally much less than seen in the upper portions of the room. Falldown of com-
bustible ceiling coverings can cause both falldown fire damage and protected areas, as
shown in Figure 3-27. In an accidental fire, until flashover occurs, involvement of fuels
at floor level tends to occur late in the fire (except sometimes in the vicinity of the ori-
gin), where an accelerated incendiary fire may produce floor-to-ceiling damage (full-
room involvement) early in the fire. Prolonged post-flashover burning may obliterate
this distinction. When the ceiling in a room is made of combustible material like wood,
the critical heat flux needed to ignite it is often reached early in the fire development.
Early critical heat flux results in quick failure, and the fire can spread upward very rapidly.
In a test conducted by one of the authors, the flame plume from a wood-and-cardboard
bonfire reached the lath-and-plaster ceiling of a room, causing rapid heating of the wood
lath. The ceiling failed within 10 minutes of ignition of the fire, allowing venting of the
Chapter 3 Fundamentals of Fire Behavior and Building Construction 75
 FIGURES 3-27A & B Two views of post-flashover room fire. Note rectangular areas of floor damage and
adjacent protected areas from falldown of low-density cellulose ceiling tiles. Furring strips and nails on ceiling
(and remaining tiles) help identify ceiling lining that aided fire spread. Courtesy of Jamie Novak, Novak Investigations
and St. Paul Fire Dept.

hot gas layer into the attic above. When this caused failure of the roof, it prevented the
development of full flashover conditions in the room, since without the hot gas layer
there was inadequate radiant heat to support post-flashover burning. At the other
extreme, if the ceiling is wood with a noncombustible layer above it (such as the origi-
nal plaster ceiling), the fire development in the room is extremely rapid, and time to
flashover is shortened significantly, as the hot gases cannot escape the room. The pres-
ence of combustible ceilings or walls makes development to flashover much more likely
and can dramatically reduce (often by 50 percent) the time needed for a room fire to
progress to flashover.49 Figure 3-28 shows a very combustible cellulose ceiling tile, once
very common.
As we saw earlier in this chapter, a fire of 650 kW heat output (a level attainable by
full flaming combustion of most large upholstered furniture items) has the potential for
generating a flame plume that will reach the ceiling of a normal residential room from the
floor (2.4 m; 8 ft), even if located away from the walls. Such flame contact will produce a
scorching of the ceiling material and may char the covering away completely if the flames
last long enough. Smaller fires (lower HRR) near walls or in corners may produce a ceiling-
high plume due to the corner effect.
76 Chapter 3 Fundamentals of Fire Behavior and Building Construction
 FIGURE 3-28 Low-
density cellulose ceiling
tiles may be found in
older structures. They can
support fire at ceiling
level and prolonged
smoldering fire as falldown
at floor level. Courtesy of
John D. DeHaan.

Walls and Interior Finish

The surfaces of walls are especially subject to attack by fire, both because of their verti-
cal orientation and because of their large surface. Next to ceilings, walls constitute the
most vulnerable portions of the main structure of any building. With the exception of
masonry walls (and those walls with steel substructure), it is normal for the main struc-
tural components, like wall studs, to be of wood and therefore to be combustible. These
structural components are usually covered and are relatively immune to attack for a lim-
ited time if the covering is noncombustible. Interior wall surfaces were at one time made
of plaster, which is itself noncombustible. The support may be of metal wire or wooden
lath. Even though the plaster will not burn, it will conduct heat and therefore will not pro-
tect underlying combustible lath and studding of the wall for as long as gypsum board, in
more common use today, even when intact. Even a small opening, however, will allow
flames to enter the wall cavity and play directly on the lath, quickly igniting it. Such walls
are often papered, which again adds to the problem of fire resistance. A single layer of
wallpaper is not especially hazardous; it quickly chars away without adding a lot of fuel
to the fire. When a wall has been repeatedly papered without removing the old paper, the
multilayer surface that results creates an increased hazard. As long as the paper is glued
tightly to the plaster, heat applied to its face is conducted into the plaster. When the paper
lifts away from the plaster, it ignites easily. Both the paper and the adhesive with which it
is attached are combustible, and thickening the layer increases the amount of fuel avail-
able to the fire. Paint used similarly has generally less effect, although very thick layers of
paint have been shown to contribute briefly to the intensity of a fire involving them and
ignite “falldown” spot fires.50 (See Figure 5-24.)
The use of various types of wallboard as a substitute for plaster has been a more
recent development of wide application. Early wallboards were of cellulose materials and
were generally quite combustible. The introduction of gypsum board or sheetrock is pos-
sibly the most important development in improving fire resistance in housing and com-
mercial buildings. The terms gypsum board, drywall, sheetrock, gyp rock, and plasterboard
are often used interchangeably to denote a rigid panel of calcium sulfate (plaster) coated
on both sides with a heavy paper bonded to the plaster. Wallboards come in various thick-
nesses, and some may come pre-coated with a vinyl coating rather than raw paper. Some
earlier forms may not be paper covered on both sides and may come in nonstandard
Chapter 3 Fundamentals of Fire Behavior and Building Construction 77
 thicknesses. Fire-rated gypsum board is thicker than standard board and contains glass or
mineral fibers to offer additional mechanical strength to the board even after fire expo-
sure. Even fire-rated gypsum board has a finite heat capacity and can conduct heat to
underlying combustible members, eventually allowing them to char. As plaster (calcium
sulfate dihydrate) is heated, it gives up its water of hydration and loses its mechanical
strength by the time it reaches ~200°C (400°F). Water-resistant wallboards are made with
a cement instead of plaster and are used in bathrooms and enclosures where
moisture/water exposure is likely. One older variant is called button board, in which the
gypsum board is pre-perforated, and a coating of cement-based plaster is applied over it.
Such products significantly reduce the speed of fire spread and can help contain even
a large fire in a compartment due to the dehydration process and thermal mass of the cal-
cium sulfate. The paper coating is bound tightly to the gypsum, and heat applied is con-
ducted away into the bulk material, allowing it to resist rapid combustion. The fire
properties of gypsum wallboard have been extensively examined by NRC-Canada and
others.51 As the room approaches flashover the paper across the entire exposed wall sur-
face may ignite, adding briefly but substantially to the heat release rate of the fire. The
vinyl coatings used on wood and gypsum walls can also add to the fuel load in the room.
The flame speed ratings of such wall coverings are determined by tests such as ASTM E84
and E1321 and can be used to evaluate the relative risks of fire spread. There is also a
new generation of tests that more closely simulate real-world performance by fitting the
material to be tested to both the walls and ceiling of a corner-configuration life-size test
cell (ISO 9705 and UBC Standard 26-3).52
Wood surfacing, which was originally standard in nearly all primitive construction,
was replaced by plaster and other materials but has again become popular because of the
aesthetic appeal of the many manufactured plywood and veneer boards that are currently
available. These materials are relatively inexpensive, highly decorative, readily applied,
and durable. They can also raise the fire hazard by a very significant amount in even more
ways than solid wood walls, as described earlier. If ignited, they tend to burn rapidly
because of their thinness. They allow the fire to penetrate into the interior of the wall,
where it can spread on both sides of the board. And some plywoods have been known to
delaminate, with separation of the very thin layers from which they are constructed.
When delamination occurs, the thin layers exposed to hot gases from both sides ignite
very rapidly, causing very intense fires. No matter how plywoods ignite, they always add
to the fuel load of the room, with commensurate effects on fire behavior and intensity. It
is not likely that materials as useful as these will fall into disuse, but it can be recom-
mended that they be secured over a layer of gypsum board, which at least prevents burn-
ing from both sides and keeps the fire restrained, both in total intensity and from entering
the interior of walls.53 Figures 3-29a–g illustrate the influence of wall-covering material
on vertical fire spread, where the same initial fire is ignited against gypsum board, marine
plywood, and thin-veneer paneling. Laminated wall materials can delaminate during a
fire, and thin combustible layers can curl to the floor, producing fire damage to floors and
walls, as in Figure 3-30.
It should also be noted that nonstandard materials can be encountered as wall or even
ceiling coverings that can have very fast flame-spread characteristics. These nonstandard
materials include carpeting, unprotected urethane or polystyrene foam, or the kraft-paper
covering of fiberglass insulation. These materials may burn away so completely, even in a
short-duration fire, that they are not readily detectable by post-fire inspection alone. If
there is any doubt, the investigator should establish what the wall and ceiling linings were
from either interviews or pre-fire photographs. Careful examination of the walls behind
baseboards or metal receptacle plates may reveal the presence of combustible coverings.
The structural and covering materials of walls, floors, and ceilings should be documented
for each room involved in the fire (particularly the room of origin) on forms such as that
in Appendix J.
78 Chapter 3 Fundamentals of Fire Behavior and Building Construction
FIGURE 3-29A Small (150 kW) trash fire FIGURE 3-29B A small (150 kW) trash fire FIGURE 3-29C As it self-extinguishes fire
ignited at the base of a gypsum board wall at the base of a 3⁄4-in. marine plywood wall leaves behind an inverted V pattern typical of
does not spread. Post-fire damage is limited penetrates into the wall and sustains some limited-fuel fires. Courtesy of Susan Sherwin,
to surface scorching of paper coating vertical spread before self-extinguishing. Scottsdale, AZ.
(unpainted). Courtesy of Susan Sherwin, Courtesy of Susan Sherwin, Scottsdale, AZ.
Scottsdale, AZ.

FIGURE 3-29D A similar fire ignited against FIGURE 3-29E Within 2 minutes, the fire FIGURE 3-29F Flames reach the ceiling
wood-veneer paneling. Courtesy of Susan in Figure 3-29d travels rapidly upward. with nearly vertical spread from ignition.
Sherwin, Scottsdale, AZ. Courtesy of Susan Sherwin, Scottsdale, AZ. Notice the V pattern at the ceiling. Courtesy
of Susan Sherwin, Scottsdale, AZ.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 79

 FIGURE 3-29G The
burn pattern is nearly
vertical, spreading into a
V only where the plume
encountered the ceiling.
Courtesy of Susan Sherwin,
Scottsdale, AZ.

FIGURE 3-30 Wood paneling on walls contributed to fire spread from ignition in left rear corner. Note paneling
curling from ceiling downward (especially along seams). Falldown of burning, curled remnants contributed to
ignition of chair and sofa on right. Had this fire burned longer, the exposed paneling would have burned away
completely, leaving only irregular heat/smoke patterns on walls and remains behind baseboards or switchplates.
The investigator must establish whether such paneling was present. Courtesy of Investigator Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

80 Chapter 3 Fundamentals of Fire Behavior and Building Construction

FIGURE 3-31A Kitchen fire test—under- FIGURE 3-31B One door has failed— FIGURE 3-31C Both doors have failed. Very
ventilated. Fire in cabinets. Courtesy of Jamie intense fire near door—near flashover. intense fire at doors—post-flashover. Courtesy
Novak, Novak Investigations and St. Paul Fire Dept. Courtesy of Jamie Novak, Novak Investigations of Jamie Novak, Novak Investigations and St. Paul
and St. Paul Fire Dept. Fire Dept.

Doors
Doors provide critical separation of interior spaces, so when they fail, fire behavior can
change. See Chapter 5 for times of door failure. The nature (construction and materials)
and starting position of all doors in the fire’s path should be documented by physical evi-
dence. Changes in their position (from recovery or firefighting) should be documented via
interviews. Figures 3-31a–c show the effect of door position on fire behavior and patterns.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 81

 CHAPTER REVIEW

Summary
It should now be apparent that fire is a complex on fuel load, ventilation, and the like). These fairly
process subject to many conditions that influence its predictable patterns may already be apparent to the
growth, spread, and development. These conditions astute reader, but they will be brought into clearer
include the physical state of the fuel, its mixture with focus in later chapters, which deal with the interpreta-
oxygen, and the transmission of heat throughout the tion of the diagnostic signs left after the fire.
fire environment. The physical state of the fuel (solid, Similarly, building construction offers a complex
liquid, or gas) has such an important effect on the array of configurations and materials. Understanding
combustion reaction that the different states will be the role of the structural shell, fire resistance ratings,
examined separately in upcoming chapters. The and internal structure is particularly important for the
behavior of a fire follows certain patterns (depending fire investigator to understand fire spread.

Review Questions
1. What conditions must exist for a fire to occur? 6. Describe a diffusion flame.
2. Define the two basic types of combustion, and 7. What is the difference between heat and
give an example of each. temperature?
3. What needs to be burning to produce a flaming 8. Define heat release rate.
fire? 9. Describe the three basic means of heat transfer—
4. Give a brief description of the four stages of a conduction, convection, and radiation.
typical fire development—incipient, growth, 10. What is a ventilation limit?
free-burning, decay. 11. What roles can concealed spaces play in allowing
5. Define flashover as the term applies to fire fire spread in some buildings?
development in a room.

References
1. Society of Fire Protection Engineers, Handbook of Fire 7. Quintiere, Principles of Fire Behavior, 49.
Protection Engineering, 4th ed. (Quincy, MA: SFPE, 8. Ibid., 53–54.
2008), 4–158, fig. 4–6.4. 9. Drysdale, An Introduction to Fire Dynamics, 60.
2. D. D. Drysdale, An Introduction to Fire Dynamics, 10. Quintiere, Principles of Fire Behavior, 58.
2nd ed. (Chichester, UK: Wiley, 1998), 173. 11. Idid.
3. J. D. DeHaan, “The Dynamics of Flash Fires Involving 12. Idid., 56–57.
Flammable Hydrocarbon Liquids,” American Journal 13. Idid.
of Forensic Medicine and Pathology 17, no. 1 (1996), 14. Drysdale, An Introduction to Fire Dynamics, 126–33.
24–31. 15. Ibid., 132.
4. J. G. Quintiere, Principles of Fire Behavior (Albany, 16. R. J. Alpert and E. J. Ward, “Evaluation of
NY: Delmar, 1998), 33–36. Unsprinklered Fire Hazards,” Fire Safety Journal 2,
5. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire no. 2 (1984): 127–43.
Science Publishers, 2003). 17. J. Albers, “Flashover in Santa Ana,” California Fire
6. J. G. Quintiere, “Surface Flame Spread” in SFPE Service (November 1995), 18–21.
Handbook of Fire Protection Engineering, 4th ed. 18. Drysdale, An Introduction to Fire Dynamics,
(Quincy, MA: NFPA, 2008), 2–235. 292–96.

82 Chapter 3 Fundamentals of Fire Behavior and Building Construction

19. V. Babrauskas and J. Krasny, “Fire Behavior of 37. J. Bloom, “The Use of a Wood Moisture Detector in
Upholstered Furniture,” NBS Monograph 173 Structural Fire Investigation,” Bloom Fire
(Gaithersburg, MD: U.S. Department of Commerce, Investigations, 1998.
National Bureau of Standards, 1985); G. H. Damant 38. Damant and Nurbakhsh, “Christmas Trees.”
and S. Nurbakhsh, “Heat Release Testing of Furniture 39. J. Stroup, Scotch Pine Christmas Tree Tests, FR4010
and Bedding Products,” Fifth Fire Safety Conference and (Gaithersburg, MD: NIST, December 1999); V.
Exhibition, Orlando, FL, February 1993. Babrauskas, G. A. Chastagner, and E. Stauss,
20. G. H. Damant and S. Nurbakhsh, “Christmas Trees— “Flammability of Cut Christmas Trees” in
What Happens When They Ignite?” Fire and Materials International Association of Arson Investigators
18 (1994). Conference, Atlantic City, NJ, 2001.
21. Kent Fire Brigade (United Kingdom), “Video Tapes 40. J. Novak, St. Paul Fire Department, personal
Fed Fatal Fire,” Fire ( June 1994). communication, August 1995.
22. DeHaan, “The Dynamics of Flash Fires,” 24–31. 41. S. Kerber and D. Madrzykowski, “Evaluating
23. D.W. Belles, “Illustration of Flashover in a Furnished Positive-Pressure Ventilation in High-Rise Structures,
Room,” NFPA Journal 79, no 2 (1985). NISTIR 7468 (Gaithersburg, MD: BFRL, November
24. J. R. Lawson and J. G. Quintiere, “Slide Rule 2007).
Estimates of Fire Growth,” NBSIR 85-3196 42. F. L. Brannigan, Building Construction for the Fire
(Gaithersburg, MD: National Bureau of Standards, Service, 3rd ed. (Quincy, MA: NFPA, 1992).
1985); see also H. E. Nelson, “FPETool: Fire 43. Fire Destroys Madrid High Rise,” 5; G. P. Gary,
Protection Engineering Tools for Hazard Estimation,” “Report on the Investigation of the First Interstate
NISTIR 4380 (Gaithersburg, MD: Building and Fire Bank Building, Los Angeles, CA.” Paper presented at
Research Laboratory, NIST, October 1990); N. Iqbal the Kansas City Arson Seminar, Kansas City, MO,
and M. H. Salley, Fire Dynamics Tools (Washington, May 1988; D. L. Ziegler, “The One Meridian Plaza
DC: U.S. Nuclear Regulatory Commission, November Fire: A Team Response,” Fire and Arson Investigator
2004). (September 1993).
25. J. D. DeHaan, “Compartment Test Fires: 1996–98,” 44. S. Colwell and D. Smith, “Comparative Fire Tests on
CAC News (Spring 2001). Sandwich Panels—Summary Report,” Fire Engineers
26. Quintiere, Principles of Fire Behavior, 188. Journal 60, no. 205 (March 2000); M. Harris, “Testing
27. S. Carman, “Progressive Burn Pattern Development in Your Metal,” Fire Engineering Journal/Fire Prevention
Post-Flashover Fires” in Proceedings Fire and (June 2005): 27–28.
Materials 2009 (London: Interscience 45. NFPA 220, Types of Building Construction (Quincy,
Communications, 2009). MA: NFPA, 2005).
28. DeHaan, “Compartment Test Fires.” 46. Illinois Fire Service Institute Floor Panel Tests, 1998;
29. R. J. Harris, Investigation and Control of Gas A. R. Earls, “It’s Not Lightweight Construction; It’s
Explosions (New York: E & FN Spon, 1983), 126. What Happens When Lightweight Construction
30. C. M. Fleischmann, P. J. Pagni, and R. B. Williamson, Meets Fire,” NFPA Journal (July–August 2009):
“Exploratory Backdraft Experiments,” Fire Technology 38–45.
(Fourth Quarter, 1993). 47. V. Babrauskas, Ignition Handbook, 942–65; V.
31. R. W. Bukowski, “Modeling a Backdraft: The Fire Babrauskas, “Charring Rate of Wood as a Tool for
at 62 Watts Street,” NFPA Journal (November– Fire Investigators” in Proceedings Interflam 2004
December 1995). (London: Interscience Communications, 2004), vol. 2,
32. S. Kerber and W. D. Walton, Effect of Positive 1155–70.
Pressure Ventilation on a Room Fire, NISTIR 7213 48. R. J. Powell, “Testimony Tested by Fire,” Fire and
(Gaithersburg, MD: U.S. Fire Administration, Arson Investigator: (September 1992), 42–51; J. J
National Institute of Standards and Technology, Lentini, “The Lime Street Fire: Another Perspective,”
March 2005). Fire and Arson Investigator (September 1992):
33. J. A. Moncada, “Fire Unchecked,” NFPA Journal 52–54; J. D DeHaan, “Fire: Fatal Intensity,” Fire
(March–April 2005), 47–49; “Fire Destroys Madrid and Arson Investigator (September 1992): 55–59;
High Rise,” Fire Prevention and Fire Engineers Journal J. Albers, “Pour Pattern or Product of Combustion?”
(March 2005), 5. California Fire/Arson Investigator (July 2002):
34. Babrauskas, Ignition Handbook, 527–33. 5–10.
35. C. B. Wood, “Pining for Water,” NFPA Journal 93, 49. B. Karlsson and J. G. Quintiere, Enclosure Fire
no. 6 (November–December 1999); Damant and Dynamics (Boca Raton, FL: CRC Press, 2000).
Nurbakhsh, “Christmas Trees.” 50. J. M. McGraw and F. W. Mowrer, “Flammability of
36. M. Bishop and D. DeMars, “Accidental Ignition of Painted Gypsum Wallboard Subjected to Fire Heat
a Christmas Tree? Very Unlikely!” Fire and Arson Fluxes” in Proceedings Interflam 1999 (London:
Investigator 50, no.1 (October 1999). Interscience Communications, 1999), vol. 2, 1325–30.

Chapter 3 Fundamentals of Fire Behavior and Building Construction 83

51. S. M. Cramer et al., “Mechanical Properties of Materials 2003 (London: Interscience Communications,
Gypsum Board at Elevated Temperatures” in Proceedings 2003), 1–8.
Fire and Materials 2003 (London: Interscience 53. N. Benichou, M. A. Sutton, and V. R. Kodur, “Fire
Communications, 2003), 33–42. Resistance Performance of Lightweight Framed Wall
52. R. W. Bukowski, “Prediction of the Structural Fire Assemblies” in Proceedings Fire and Materials 2003
Performance of Buildings” in Proceedings Fire and (London: Interscience Communications, 2003), 9–20.

84 Chapter 3 Fundamentals of Fire Behavior and Building Construction

 CHAPTER 4
Combustion Properties
of Liquid and Gaseous Fuels

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

KEY TERMS

accelerant, p. 106 flame point, p. 94 ignition temperature, p. 94

adiabatic, p. 104 flammable liquid, p. 94 odorant, p. 113
autoignition temperature, p. 94 flammability limits, p. 88 olefinic, p. 107
BLEVE, p. 120 flash point, p. 87 vapor density, p. 99
boiling point, p. 98 ignitable liquids, p. 112
combustible liquid, p. 94 ignition energy, p. 97

OBJECTIVES
After reading this chapter, you should be able to:
■ List and describe the three types of fuels.
■ Describe the physical properties of fuels.
■ Explain how the flammability limits of a vapor/air mixture are affected as a function
of initial temperature.
■ Explain the relationships of the terms flash point, flame point/fire point, ignition
temperature, and ignition energy.
■ Define the terms boiling point and vapor density.
■ Describe how the heat of combustion (defined in Chapter 2) of a fuel is of importance
to the fire investigator.
■ Explain the difference between hydrocarbon and nonhydrocarbon fuels.
■ Describe the behavior of liquid pools on porous and nonporous surfaces.
■ Describe the distribution and effects of radiant heat from a burning pool of an
ignitable liquid.
■ Discuss several failure modes of fuel gas lines.

85
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

N
ow that we have examined the fundamental chemical and physical properties of fire
and fuels, we are ready to investigate the combustion mechanisms of particular fuels
in greater detail. Unfortunately, there is no subject about which even experienced fire
investigators are so likely to err. Most fire investigators are not chemists, but some chemical
knowledge is essential to the proper interpretation of fires, as we have seen fires to be strictly
chemical reactions. To compound the investigator’s confusion about fuels and the tests used
to evaluate them, there is a whole vocabulary of terms that apply to tests devised in the labo-
ratory to define the properties of fuels of investigative significance. It is the purpose of this
chapter to discuss some of the concepts that are basic to understanding how gaseous and liq-
uid fuels burn, the conditions and limitations that apply to the combustion of such fuels, and
the conventional methods of expressing combustion properties in terms of laboratory tests.

Types of Fuel
Combustion properties have different significance, or at times perhaps no significance,
when applied to various types of fuel. Thus, the concept of flash point, discussed later, is
essential information when studying liquids, is rarely ever used for solids, and would have
no meaning if applied to gases. The relationship between the type of fuel and the type of
property is clarified later.

GASES
Flammable materials found in the gaseous form will burn whenever mixed with the
proper amount of air and properly ignited. Flash point has no significance with such
materials, nor has boiling point. However, the flammable range of mixture with air, the
vapor density, and ignition energy and temperature are important properties.

LIQUIDS AND THEIR VAPORS

Vapors from liquids are the materials that directly support the flame over a liquid fuel.
For the liquid/vapor combination, vapor pressure, flash point, and, to a lesser degree,
boiling point are all important considerations. Ignition temperature, range of flammable
concentrations of vapor/air mixtures, and vapor densities are critical properties of the
vapors themselves. Liquids themselves do not undergo combustion, except in exotic cases;
instead, it is the vapors emanating from liquids that can ignite and burn as flames.

SOLIDS
Solids burn by direct combination of oxygen with their surface (smoldering combustion),
as volatilized materials that have melted and vaporized, or as complex fuels that pyrolyze
86 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
to form combustible gases and vapors and leave a noncombustible solid residue. Reactive
metals such as magnesium, sodium, potassium, or phosphorus, and carbonaceous fuels
such as charcoal, burn only at their surface, as a glowing fire. If the vapors being
pyrolyzed from common solid fuels are disregarded, most of the properties just mentioned
do not apply to solids. Nonpyrolyzable solids have no flash point in the strict sense, no
vapor density, and no single ignition temperature. They have a flammable mixture range
only when they are finely divided and stirred into air as a combustible dust suspension.
The physical properties that control their flammability are such factors as density, con-
ductivity, and thermal capacity. Their porosity and melting point will also influence their
performance as fuels under some conditions. The chemical properties of a solid fuel will
determine the nature of volatile pyrolysis products and the rate at which they are gener-
ated. The physical and chemical properties of solid fuels and their ignitability will be
examined in greater detail in Chapter 5.

Physical Properties of Fuels

Physical properties of fuels of particular interest to the fire investigator include vapor
pressure, flammability or explosive limits, flash point, flame point/fire point, ignition tem-
perature, ignition energy, boiling points, vapor density, and heat of combustion.

VAPOR PRESSURE Pressure

Any liquid exposed to air will evaporate as molecules escape its surface. If this process
occurs in a sealed system, an equilibrium state will be reached in which the vapor reaches
saturation, and there is no further evaporation. The partial atmospheric pressure exerted
by the vapor at this stage is called the saturated vapor pressure, which is most often meas- Saturated
vapor
ured in millimeters of mercury (mmHg) (atmospheric pressure is 760 mmHg) or kilopas-
cals (kPa), where 1 atmosphere (atm) is 101.32 kPa. The vapor pressure is a direct
measure of the volatility of a liquid and is determined by its molecular weight, its chem-
Liquid
ical structure, and the temperature (see Figure 4-1). The higher the temperature, the more
liquid will evaporate and the higher the vapor pressure. Table 4-1 displays several pressure
FIGURE 4 -1 Vapor
equivalents.
pressure is the pressure
As seen in Figure 4-2, the relationship between temperature and vapor pressure is not created by a liquid (or solid)
linear but is a logarithmic function with controlling constants that are different for each evaporating to saturation
liquid. For a given class of compound (alkane, alcohol, branched alkane), the lower the in a sealed vessel.
molecular weight, the higher the vapor pressure. The vapor pressure of an aromatic or
that of a normal alkane is lower than that of a branched alkane of the same molecular
weight at the same temperature, so a branched (iso) alkane will have a lower flash point flash point ■
than an alkane of the same molecular weight. For example, benzene (molecular weight Temperature at which
78): flash point, 11°C; hexane (MW  82): flash point, 22°C; isohexane (MW  82): an ignitable vapor is
first produced by a
flash point, 29°C. The vapor pressure is a fundamental physical property that has a liquid fuel.
great influence on the flammability of liquid fuel. When the vapor pressure of a liquid
reaches 760 mmHg, the liquid is at its boiling point, so the temperature at which that

TABLE 4-1 Pressure Equivalents

kPa mmHg psi atm bar in. w.c.

689.2 5170 100 6.80 7 —
101.32 760 14.7 1 1.013 (1013 mbar) 406.7
6.89 51.7 1 0.068 0.07 (70 mbar) 27.67
0.69 5.2 0.1 0.007 0.007 (7 mbar) 2.77

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 87

FIGURE 4-2 Calculated 900
saturation vapor pressure
(in mmHg) as a function of 800
temperature (°C) for three
common hydrocarbons: 700
n-pentane, n-hexane, and n-pentane

Vapor pressure (mmHg)

n-octane. Upper dashed 600
line is atmospheric pressure
(760 mmHg). Middle 500
(long-dashed) line corre-
sponds to the upper explo- 400
sive limit (UEL) (8% in air,
61 mmHg) for n-alkanes. 300
The lowest dashed line
corresponds to the lower 200
n-hexane
explosive limit. Note that
at all temperatures above 100
5°C, pentane and hexane n-octane
are above their UEL at 0
10 0 10 20 30 40
saturation.
Temperature (C)

occurs is recorded as its boiling temperature. When the temperature of a liquid fuel is such
that its vapor pressure reaches the percentage of 760 mmHg that is equivalent to the
lower flammability limit of the fuel (see next section), the fuel is said to be at its flash
point, as discussed later in this section. The vapor pressure is also dependent on the con-
tour of the surface of the liquid, being higher for a convex surface than for a flat surface
at the same temperature. This partly explains why an aerosol of small droplets or a liquid
distributed on the fibers of a porous wick is often easier to ignite than a pool of the same
liquid. Also playing a critical role in such ignitions is that the larger total surface area and
small size of the aerosol droplets allows them to absorb heat more quickly from their sur-
roundings. A thin film of liquid on a fiber of a wick can absorb heat very quickly from
applied heat flux as well as from the surrounding air and supporting fiber. The increased
temperature causes it to evaporate more quickly. Being a thin film, it has no bulk liquid
to “cool” it by convective heat transfer. The capillary action of the porous wick quickly
replaces the evaporated fuel to sustain the evaporation or combustion.

FLAMMABILITY (EXPLOSIVE) LIMITS

Mixtures of flammable gases or vapors with air will combust only when they are within
particular ranges of concentration. When a gas is present at a concentration below its
flammability limits ■ lower explosive limit (LEL, or lower flammability limit—LFL), it is considered too lean a
The lower and upper mixture to burn and cannot be ignited. At concentrations above its upper explosive limit
concentrations of an (UEL), the fuel/air mixture is too rich to burn and will not ignite. The explosive range
air/gas or air/vapor
mixture in which com-
between these two limits is illustrated in Figure 4-3 and Tables 4-2 and 4-3. Some mate-
bustion or deflagration rials such as hydrogen, carbon monoxide, and carbon disulfide have very wide ranges
will be supported. within which flame will propagate to create an explosion. However, most fuels, such as
petroleum distillate hydrocarbons, have quite narrow explosive ranges.
Fuel/air mixtures are often considered to be in either closed or open systems, as dis-
cussed next.
Closed Systems
Within a closed system, if the mixture will not explode, it will not ignite. For this reason,
the explosive range and flammability range of a gas or vapor may be thought of as one
and the same. The explosive ranges (and vapor pressures) of possible fuels at a fire scene
must be carefully considered, for they can play a significant part in determining acciden-
tal or incendiary origin and in ruling out some hypothetical situations. As can be seen
88 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 100% by volume 0% by volume FIGURE 4-3 Schematic
of flammable gas/air mix-
tures above and below
explosion limits. Based on
No combustion possible Brannigan, F. L., R. G. Bright,
due to too little oxygen and N. H. Jason, eds., Fire
(mixture too strong) Investigation Handbook, NBS
Handbook 134. Washington,
DC: U.S. Government Printing
Increasing fuel content Office, 1980, p. 139.

Increasing air content

Limited combustion possible
around ignition sources
Upper explosion limit (UEL)

Explosion occurs
as soon as
ignition is applied

Lower explosion limit (LEL)

No explosion
due to too little gas
(mixture too weak)

0% by volume 100% by volume

Gas content Air content

from Figure 4-4, flammability limits are temperature dependent: the higher the initial tem-
perature, the wider the range. Technically, the explosive limits and flammability limits are
determined under different conditions—constant pressure versus constant volume—and
are therefore very nearly the same.
Open Systems
In an open system, other variables control the ignitability of the fuel/air mixture. If a mix-
ture of vapor and air is initially too lean for an explosion, efforts to ignite it may raise the
temperature and increase the amount of material volatilized sufficiently to cause ignition.
More important, probably, is the too-rich mixture that will not ignite. The application of
a brief arc will not be expected to cause a fire; however, prolonged application of a flame
may add so much additional heat that the flammability limit is affected or may produce

Saturated A FIGURE 4-4 The flam-

Upper mability limits of a vapor/
vapor-air
limit air mixture as a function
mixtures
of initial temperature at a
Volume % fuel in air

constant pressure, show-

Autoignition ing the higher the tem-
Flammable
perature, the wider the
Mist mixtures
explosion range. From
Zabetakis, M. G., “Flammability
Characteristics of Combustible
C D Lower Gases and Vapors.” U.S. Bureau
limit of Mines, U.S. Government
B printing Office, 1965, p. 3.

TL TU AIT
Temperature

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 89

 TABLE 4-2 Flammability (Explosive) Limits and Ignition Temperatures
of Common Gases

EXPLOSIVE IGNITION MINIMUM

LIMITS TEMPERATURE IGNITION
(IN AIR) (MINIMUM) ENERGY
Fuel Lower Upper °C °F mJ
Natural gas 4.5 15 482–632 900–1,170 0.25
Propane (commercial) 2.15 9.6 493–604 920–1,120 0.25
Butane (commercial) 1.9 8.5 482–538 900–1,000 0.25
a
Acetylene 2.5 81 305 581 0.02
Hydrogen 4 75 500 932 0.01
Ammonia (NH3) 16 25 651 1,204 —
Carbon monoxide 12.5 74 609 1,128 —
Ethylene 2.7 36 490 914 0.07
Ethylene oxide 3 100 429 804 0.06

Sources: Fire Protection Handbook, 19th ed. (Quincy, MA: National Fire Protection Association, 2003), table 8-9.2;
and SFPE Handbook of Fire Protection Engineering, 4th ed. (Quincy, MA: Society of Fire Protection Engineers and
the National Fire Protection Association, 2008), table 3-18.1.
a
Higher concentrations (up to 100) may detonate.

so much turbulence as to mix in additional air and thus start a fire. Because of the initial
overly rich mixture, this resultant fire is expected to develop into what is known as a
rolling fire of some intensity. It must be emphasized that the reason fire is ignited in both
cases is that the system was altered by introduction of heat from the ignition source. Such
an alteration is not possible in a closed system and will not be effective in producing a fire
any more than in producing an explosion. Thus, on a warm day, when the gasoline/air
mixture inside a tank of gasoline is richer in gasoline vapor than the upper explosive limit,
an arc could safely be produced inside the tank with no ignition. At very low tempera-
tures, when fewer vapors are produced and therefore may form a mixture with the air
that is within the explosive range, an explosion may result (see Table 4-2).
Vapor pressure considerations play a critical role in evaluating incidents where the
flame external to a partially full container of flammable liquid is thought to have caused
the vapor inside the container to explode by combustion. As we saw in Figure 4-1, the
vapor pressure of pentane (the major, most volatile liquid component in automotive gaso-
line) is, at saturation, well above its UEL (8 percent) for any reasonable temperature
(above -20°C; -4°F).1 This means that if there is a significant quantity of liquid in the
container and a reasonable time has elapsed, the concentration of vapors inside the con-
tainer will be far too rich to support combustion. A flame can be sustained at the mouth
of the vessel if there is enough vapor emanating to mix with the surrounding air, but no
explosion will occur. (Heating of a closed vessel containing liquid can cause a mechanical
explosion if the heating is sufficient to create a vapor pressure that exceeds the mechani-
cal strength of the vessel. See Chapter 12 for a discussion of such mechanical explosions.)
To test the theoretical (calculated) concept, empirical testing of plastic fuel containers
filled with gasoline was conducted by Novak and by DeHaan in cooperation with the New
Zealand Police Service.2 In each case, gasoline vapors emanating from the filler spout were
ignited readily. There was no explosion, but there was a small clear flame about 12 cm (5 in.)
high in each test, as shown in Figure 4-5a. As the plastic container melted, the opening
90 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
increased in area (Figures 4-5b and c), and heat was absorbed by the gasoline (sometimes
to boiling). The flame plume got larger, but there was no propagation into the can to cause
an explosion. Eventually, the entire top of the can melted or burned away, exposing the
entire horizontal surface of the gasoline, as seen in Figures 4-5d and e.
The maximum size of the fire is controlled by the area of the pool exposed multiplied
by an experimentally determined kilowatts per area factor. For gasoline, that factor is
around 1,800 to 2,000 kW/m2. For a typical 20-L (5-gal) container 38 cm  38 cm
(15 in.  15 in.) in size, this means a maximum fire of about 400 kW. Estimates of the
HRR of these tests based on plume height revealed a maximum fire of ~150 kW (seen in
Figure 4-5c), the reduction being due to the “lip” of plastic around the pool, which
reduced entrainment efficiency. Even when the test was conducted with a flat-style 10-L
container half full, lying on its side with no cap (Figure 4-6a), there was no explosion. In
that test the upper side of the reclining can melted away over a period of minutes (Figures
4-6b–d). Eventually, the plastic containers failed as the gasoline overflowed from expan-
sion, and a very large pool fire resulted, but there was no explosion (Figure 4-6e). The
melted remains of such containers (Figures 4-5e–g and 4-6f) often contain pockets of
unburned fuel and bear critical manufacturing data.
For liquid fuels with lower vapor pressures, propagation or ignition can occur within
a fuel container. This is now recognized as a problem with ethanol-based motor fuels,
where ignition inside fuel storage tanks is possible at normal temperatures.3 Guidry
reported an incident in which denatured (ethyl) alcohol (whose vapor pressure curve is
much lower than that of pentane and has a much broader explosive range) caused an
explosion within a container.4 Under the ambient temperature and use conditions, an
ignitable vapor concentration existed inside the container. When a flame ignited the
stream of liquid pouring from the container, the flame propagated into the container. This
caused an explosion inside the container that split the container and spewed burning alco-
hol over the person holding it, inducing serious burns.
In cases involving the home storage of gasoline, burn injuries can occur when the can-
ister is tipped or knocked over, spilling its contents and dispersing vapors within the area
of the container. In a case study by Kennedy and Knapp of 25 burned children less than
6 years old, the research showed that in each case vapors from overturned or spilled gaso-
line were ignited by a pilot light from either a water heater or dryer. The study noted that
children at this age are totally unaware of the hazards gasoline introduces into their envi-
ronment. Other problems cited in the study are that older children and adults who may
misuse gasoline as an inhalant, solvent, or fire starter thus also may place themselves in
harm’s way by unwittingly exposing the vapors to a competent ignition source.5
The tragic explosion of flight TWA 800 off the coast of New York in July 1996 was
caused by a very unusual set of conditions when a nearly empty jet fuel tank was heated (by
ambient summertime conditions and the operation of onboard air conditioners). The fuel
was then exposed to a reduced atmospheric pressure as the aircraft gained altitude (increas-
ing the fuel’s partial vapor pressure). At 4,000 m (13,000 ft) an accidental electrical ignition
source within the tank ignited the vapors as they reached their LEL. The resulting internal
fuel tank explosion caused massive damage to the fuselage and brought the plane down.6
Automotive fuel tank explosions in movies are created by using high explosives. In real life,
automotive fuel tanks very rarely explode, but when a tank does, it is usually as a result of
a collision that ruptures it and allows all the liquid to drain out; the concentration of vapor
in the tank can then fall below the UFL, and the vapor may explode if ignited.

FLASH POINT
Characteristics of Flash Point
The term flash point is basic to the description of many common liquid fuels, since it can
be used to assess the fire hazards of ignitable liquids under various conditions (ignitable
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 91
FIGURE 4-5A Vapors of one gallon of FIGURE 4-5B Burning vapors at 10 minutes. FIGURE 4-5C Gasoline boiling as vapors
gasoline in a plastic container ignited at Courtesy of Jamie Novak, Novak Investigations and burn at 20 minutes. Courtesy of Jamie Novak,
open spout. Courtesy of Jamie Novak, Novak St. Paul Fire Dept. Novak Investigations and St. Paul Fire Dept.
Investigations and St. Paul Fire Dept.

FIGURE 4-5D Melted remains of plastic FIGURE 4-5E Melted gasoline container FIGURE 4-5F Underside of melted container
can with burning gasoline. Courtesy of Jamie after fire (top). Courtesy of Jamie Novak, Novak bears molded information. Courtesy of Jamie
Novak, Novak Investigations and St. Paul Fire Dept. Investigations and St. Paul Fire Dept. Novak, Novak Investigations and St. Paul Fire Dept.

FIGURE 4-5G Container

cut open showing pockets
that often trap residual
liquid fuel. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul Fire
Dept.

FIGURE 4-5 Ignition test of 3.8-L (1-gal) plastic container filled with gasoline, with open filler spout.

92 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

FIGURE 4-6A Ignition—no explosion FIGURE 4-6B One minute—quiet flame at
mouth of filler.

FIGURE 4-6C Five minutes—filler neck is melt- FIGURE 4-6D Ten minutes—side of can melting,
ing, flames growing longer. exposing more gasoline to air.

FIGURE 4-6E After 14 minutes, can melts in a FIGURE 4-6F Melted container after fire.
pool fire.

FIGURE 4-6 Plastic 10-L gas can lying on its side, half full of gasoline, with no cap. All Courtesy of New Zealand
Police Service, Canterbury CIB.

liquids being those that fall under the common classifications of flammable or com-
bustible liquids). The flash point of a material is the lowest temperature at which it pro-
duces a flammable vapor. At its flash point temperature, the vapor pressure of the fuel (as
a percentage of 760 mmHg) is equal to that fuel’s lower limit of flammability. For exam-
ple, the flash point of n-decane is 46°C (115°F). Its vapor pressure at that temperature is
6 mmHg, which is 0.78 percent of 760 mmHg. This value is very close to the measured
lower explosive limit of 0.75 percent. A liquid fuel must be able to generate a vapor in
sufficient quantity to reach that lower limit in air before it can burn. This does not mean
that the vapor will ignite spontaneously at this temperature but only that it can be ignited
by a flame, a small arc, or other local source of heat. For example, the flash point of low-
octane automotive gasoline is about 43°C (45°F). This flash point value means only
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 93
 that gasoline produces a vapor that can be ignited at any temperature above about – 43°C,
not that gasoline will spontaneously ignite at such an extremely low temperature.

Determining Flash Point

The flash point is determined by placing a small sample of the fuel in the cup of a testing
apparatus and heating or cooling it to the lowest temperature at which an arc or small
pilot flame will cause a little flash to occur over the surface of the liquid. No fire results
from such a test other than the little flash of flame, which immediately extinguishes itself.
Several recognized flash point testers are used today, including the Pensky-Martens,
Cleveland, and Tagliabue (Tag) testers. Each device has a particular temperature range
and fuel type for which it is most accurate. The Tag closed-cup tester, for example, is the
accepted method (described in ASTM D56) for testing nonviscous liquids with a flash
point below 93°C (200°F).7 The Pensky-Martens tester (closed-cup) is best suited for liq-
uids with a flash point above 93°C (200°F) and is described in ASTM D93.8 The open-
cup tests are used for evaluating hazards of ignitable liquids where they are encountered
in open-air situations—during transportation or spills, for instance. The Tag open-cup
method is described in ASTM D1310,9 and the Cleveland open-cup method is described
in ASTM D92.10 Most tables of flash point data for fire safety purposes list the closed-
cup flash point of most materials. As a general rule, the open-cup flash points for the same
materials would be a few degrees (10 percent) higher than the closed-cup values. The
flash point tests used in the industry are described in detail in the NFPA Fire Protection
Handbook and will not be repeated here. The same source contains extensive tables of
flash point data for all types of flammable liquids. Some representative flash points are
listed in Table 4-3 for general reference.
NFPA11 classifies ignitable liquids according to their reported flash points:
flammable liquid ■
A liquid having a flash ■ Class I: Flashpoint below 38°C (100°F) (flammable liquids)
point below 38°C ■ Class II: Flashpoint between 38°C and 60°C (100°F–140°F) (combustible liquids)
(100°F). ■ Class III: Flashpoint above 60°C (140°F) (combustible liquids)
combustible liquid ■
A liquid having a flash FLAME POINT/FIRE POINT
point at or above 38°C
(100°F). The flame point or fire point is the lowest temperature at which a liquid produces a vapor
that can sustain a continuous flame (rather than the instantaneous flash of the flash
flame point ■ point). It is usually just a few degrees above the rated flash point. Although it is not often
Temperature at which
a flame is sustained by
reported, the flame point is probably a more realistic assessment of a fuel’s contribution
evaporation or pyrolysis to a fire environment than the flash point. The fire points of a few common fuels are
of a fuel. shown in Table 4-3. The exact values of flash point and flame point temperatures of tested
fuels depend on the following factors:12
ignition temperature ■
(same as autoignition) ■ the size of the ignition source,
The minimum tempera- ■ how long it is held over the source each time,
ture to which a sub-
■ the rate of heating of the liquid, and
stance must be heated
in air to ignite inde- ■ the degree of air movement over the fuel.
pendently of the heat-
ing source (i.e., in the
Note that Glassman and Dryer (see Table 4-3) reported minimum fire points for alcohols
absence of any other to be equal to their minimum flash points and observed a significant difference between
ignition source; non- values obtained using flame igniters and those using electrical arc ignition, apparently as a
piloted ignition). result of the infrared absorption characteristics of alcohol fuels.
autoignition tempera-
ture ■ The temperature IGNITION TEMPERATURE
at which a material will The ignition temperature is important to all considerations of a fire. Sometimes called
ignite in the absence of
any external pilot source
the autoignition temperature (AIT) or spontaneous ignition temperature (SIT), it is the
of heat; spontaneous temperature at which a fuel will ignite on its own without any additional source of igni-
ignition temperature. tion. For purposes of comparing liquid and gaseous fuels, ignition temperature can be
94 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 TABLE 4-3 Flash Points and Flame (Fire) Points of Some Ignitable
Liquids

FLASH POINT FLASH POINT

(CLOSED CUP)* (OPEN CUP) FIRE POINT
Fuel °C °F °C °F °C °F
Gasoline (auto, low octane) 43 45 — — — —
Gasoline (100 octane) 38 38
Petroleum naphthaa 29 20
JP-4 ( jet aviation fuel) 23 to 1 10 to 30
Acetone 20 4
Petroleum ether 18 0
Benzene 11 12
Toluene 4 40
Methanol 11 52 1(13.5)† 34(56) 1(13.5) 34(56)
Ethanol 12 54
n-Octane 13 56
Turpentine (gum) 35 95
b
Fuel oil (kerosene) 38 100
Mineral spirits 40 104
n-Decane 46 115 52† 125 61.5† 143
Fuel oil #2 (diesel) 52 (min) 126
Fuel oil (unspecified) — — 133‡ 271 164‡ 327
Jet A 43–66 110–150
p-Xylene 25§ 77§ 29§ 84§ — —
JP-5 (Jet aviation fuel) 66 151

Sources: *Fire Protection Guide to Hazardous Materials (Quincy, MA: National Fire Protection Association, 2001),
except as noted.
†
Calculated minimum limit for open-cup tests with flame igniter. Higher values in parentheses are for spark source
ignition. Source: I. Glassman and F. L. Dryer, “Flame Spreading across Liquid Fuels,” Fire Safety Journal
3 (1980–1981): 123–38.
‡
SFPE Handbook of Fire Protection Engineering, 4th ed. (Quincy, MA: NFPA, 2008), table 2-8.5.
§
ScienceLab.com, MSDS p-xylene, 2008.
a
Generic name for miscellaneous light petroleum fractions used in such consumer products as cigarette lighter fuels
and fuels for camping stoves and lanterns.
b
The flash point of kerosene is set by law in many jurisdictions and may be higher in compliance with local laws.

measured by injecting a small quantity of fuel into a hot-air environment (usually a closed
container) at various temperatures. In the standard test, ASTM E659, the test chamber is
an electrically heated 1-L glass flask (sometimes called a Setchkin apparatus).13 When the
container is at or above the fuel’s ignition temperature, the fuel will burst into flame upon
injection. For example, low-octane gasoline, with a flash point of -43°C (-45°F), requires
a minimum temperature of about 280°C (536°F) to catch fire (100 octane gasoline has an
AIT of 456°C). This is the minimum temperature that must be reached by the match,
spark, lighter, or other igniting instrument in order for that material to burn. With one or
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 95
 two exceptions, such as catalytic action, the temperature of at least a portion of the fuel
must be raised to its ignition temperature before fire can result from any fuel. Ignition
temperatures of common materials are generally so high as to rule out spontaneous com-
bustion, except for a very small category of materials and only under very special condi-
tions. One exception, for instance, is carbon disulfide (CS2), which is encountered as an
industrial cleaning solvent. Carbon disulfide has various reported AITs of 90°C to 120°C
(194° to 240°F). In general, for a given series of hydrocarbons, the longer the carbon
backbone, the lower the AIT of the material. Note the sequence of AITs for n-alkanes in
Table 4-4, where n-pentane has an AIT of 260°C, while n-octane and higher alkanes have
an AIT of around 208°C.
Catalytic oxidation of some fuels can occur at temperatures much lower than the AIT,
but these reactions are very rare and occur under special conditions. For example, plat-
inum in a finely divided state may serve to allow (catalyze) combustion of certain flam-
mable gases in air at low initial temperatures. In a fuel cell or a hand warmer this may be
important, but it is very rare in general fires. For solids, in which pyrolytic decomposition
must precede ignition, there is another value called the piloted ignition temperature,
which will be discussed in Chapter 5.

TABLE 4-4 Minimum Autoignition Temperatures, Flammable Ranges, and Specific Gravities
of Some Common Ignitable Liquids

IGNITION TEMPERATURE MINIMUM IGNITION FLAMMABLE RANGE

FUEL (°C) (°F) ENERGY (mJ) (% IN AIR @ 20°C) SPECIFIC GRAVITY
Acetone 465 869 1.15 2.6–12.8 0.8
Benzene 498 928 0.2 1.4–7.1 0.9
Diethyl ether 160 320 –– 1.9–36.0 0.7
Ethanol (100 percent) 363 685 –– –– 0.8
Ethylene glycol 398 748 –– –– 1.1
Fuel oil #1 (kerosene) 210 410 –– ––
Fuel oil #2 257 495 –– ––
Gasoline (low octane) 280 536 –– 1.4–7.6 0.8
Gasoline (100 octane) 456 853 –– 1.5–7.6 0.8
Jet fuel (JP-6) 230 446 –– ––
Linseed oil (boiled) 206 403 –– ––
Methanol 464 867 0.14 6.7–36.0 0.8
n-Pentane 260 500 0.22 1.5–7.8 0.6
n-Hexane 225 437 0.24 1.2–7.5 0.7
n-Heptane 204 399 0.24 –– 0.7
n-Octane 206 403 –– 1.0–7.0 0.7
n-Decane 210 410 –– –– 0.7
Petroleum ether 288 550 –– 1.1–5.9 0.6
Pinene (alpha) 255 491 –– ––
Turpentine (spirits) 253 488 –– –– 1

Sources: Data taken from Fire Protection Guide to Hazardous Materials (Quincy, MA: National Fire Protection Association, 2001); C. F. Turner
and J. W. McCreery, The Chemistry of Fire and Hazardous Materials (Boston: Allyn and Bacon, 1981).

96 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

 Virtually all fires originate where there is a localized high temperature in a region in
which an appropriate fuel/air mixture occurs. The local region may be very small, as in
the case of an electrically induced arc, a mechanical spark, or a minute flame. The impor-
tant concept is that at this very small point in space, a temperature in excess of the
ignition temperature occurs in the presence of appropriate fuel and air (or oxygen), and
this energy is transferred to an adequate volume of ignitable fuel. Such circumstances are
not unusual, but from the investigative standpoint, it must always be accepted that these
circumstances constitute a minimum requirement for any fire whatever to result. That
tiny initial flame then propagates through adjoining mixtures in the flammable range.
Many times, no source of such relatively high temperatures appears to be present, and the
origin of a fire seems very mysterious. In such instances it should be remembered that it
is actually not difficult to produce a source of energy in a very small area. Fires can be
kindled by rubbing two sticks together, by the flint-and-steel method (which strikes a very
small but hot spark), and by other similar means. In the start of a fire, it is only the heat
content of the source (as reflected in part by its temperature) that counts—not the area
over which the temperature holds. A nail in a shoe heel striking a rock may ignite a fire,
not because the shoe is hot but only because an almost infinitesimally small region (the
spark) reaches a very high temperature. Except in industrial fires, where pressurized ves-
sels and delivery lines make for special cases, it is rare indeed for any large amount of fuel
to be heated past its ignition temperature. If even a small portion of it is so heated, and
the temperature is above the ignition point, a fire can result. This is why so many fires
occur under conditions that do not seem to be hazardous. As long as the fuel gas or vapor
concentration is within its flammability range in the vicinity of this small but intensely
energetic source, there can be ignition. The fundamental consideration is not the temper-
ature of an ignition source compared with the ignition temperature of the fuel but the
transfer of heat energy from the source to the fuel. An ignition source may be present that
is at a temperature that is technically higher than the AIT listed for that fuel, but there
still may not be ignition if enough energy is not transferred to the fuel.
It should be noted that the AIT is a minimum environmental temperature measured by
a particular laboratory technique (ASTM E659 being the most common) and is dependent
on the size, shape, and even surface material and texture of the confinement vessel used in
the test apparatus. The real-world ignition sources that a fuel might encounter will usually
have different size, shape, and confinement and so will almost always have to be at a tem-
perature considerably higher than the listed AIT. AIT data are more for reference and
comparison than for a specific determination. Some minimum ignition temperatures of
common fuels are listed in Tables 4-2 and 4-3. Because the minimum ignition temperature
is dependent on the method used for its determination, that value should be used prima-
rily for comparisons between fuels. Testing has revealed that “hot surface” ignition of
gasoline in open air, for instance, requires that the temperature of the surface be at least
200°C (360°F) above the listed autoignition temperature [i.e., 500°C700°C (930°F
1,300°F)]. It is thought that the hot surface produces rapid evaporation, which mini-
mizes contact time.14 The material and texture of the hot surface will also affect minimum
ignition temperatures. The “grade” of gasoline is also critical: the higher the grade or
octane rating, the higher the AIT. (See the discussion in Chapter 9.)

IGNITION ENERGY
Every fuel gas or vapor has a minimum ignition energy. This is the amount of energy that ignition energy ■ The
must be transferred to the fuel to trigger the first oxidation. For most fuels, this is a very quantity of energy that
small amount of energy; for hydrocarbon fuels, the minimum ignition energy is on the must be transferred
into a fuel/oxidizer
order of 0.25 millijoule (mJ) (see Table 4-2). This amount is dependent on the concentra- combination to trigger
tion of the vapor and is often at its minimum when the fuel vapor is at its stoichiometric a self-sustaining
or ideal concentration (as shown in Figure 4-7). For a vapor/air mixture to ignite, there combustion.

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 97

 FIGURE 4-7 Ignitability
curve and flammability
Limits of
limits for methane/air
flammability
mixtures at atmospheric

Spark energy (mJ)

pressure and 20°C.
From Zabetakis, M. G., Ignitability
“Flammability Characteristics limits
of Combustible Gases and
1
Vapors,” Washington, DC:
U.S. Bureau of Mines, U.S. 0.5
Government Printing Office,
1965, p. 2.

0.2
2 4 6 8 10 12 14 16 18
Volume % methane in air

are four required conditions. First, an ignition source must have adequate energy (above the
minimum ignition energy for that fuel); second, the fuel must be present at a concentra-
tion within its flammability range; third, there must be contact between the source and
the fuel while it is in that range; and finally, the contact must be of sufficient duration for
enough energy to be transferred from the source to the fuel. Unless all four of these con-
ditions are met, there will not be an ignition.
Several of the flammability properties are interrelated. For instance, the minimum
ignition energy of a fuel is dependent on fuel concentration, as can be seen in Figure 4-7.
This means that a very energetic source can ignite mixtures closer to the flammability lim-
its for that fuel, while a very weak source may be competent only when the mixture is
near its stoichiometric mixture. In addition, the flammability limits themselves are
dependent on starting, or initial, temperature, as was shown in Figure 4-4: the warmer
the starting (pre-fire) mixture, the wider the limits.
Although gaseous-phase explosions are treated in depth in Chapter 12, there are sev-
eral observations about fire behavior that are related to concentration to be discussed
here. The speed with which a flame front progresses through a fuel/air mixture is depend-
ent on concentration. The flame speed reaches a maximum when the mixture is near its
stoichiometric or ideal mixture. As a result, overpressures (caused by rapidly expanding
gases) are also at or near their maxima. A premixed hydrocarbon fuel/air mixture near its
stoichiometric mixture theoretically can produce pressures up to 120 psi (8 bar) of pres-
sure under ideal conditions.15 A lean mixture propagates at a slower speed and produces
less heat and, therefore, lower pressures. Because all the fuel is consumed in the reaction,
there is no subsequent fire or flame. In contrast, a rich mixture results in a less vigorous
explosion with less mechanical damage due to lower pressures and slower speeds. The
excess fuel produces smoke or soot and a following flaming fire. Thus, the heat effects
and blackening are often greater and the blast damage less when the mixture is rich than
when it is lean. The explosive limits of some of the more hazardous liquids and gases are
discussed in Chapter 13.

BOILING POINTS
Boiling points are sometimes important in fire investigations, but they are of secondary
significance compared with other properties discussed in this chapter. Most liquids may
boiling point ■ The be expected to be heated to their boiling point in a fire. However, the very important
(pressure-dependent) distinction between a liquid fuel that is reasonably pure and will therefore have a rea-
temperature at which
sonably definite boiling point and one that is a mixture of many components of variable
a liquid changes to its
gas phase. boiling points must be grasped, since boiling point ranges are used to characterize many
fuels encountered in fires (See a later section of this chapter for examples of boiling
point ranges.)
98 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 A pure, single-component liquid fuel will have a definite boiling point at which an
entire sample may be distilled without a change in temperature of the vapors. The gen-
eral public has available to it relatively few pure liquid fuels compared with the much
larger number of mixtures that have a boiling range rather than a boiling point. Such mix-
tures range from gasoline, with more than 200 individual compounds, to a large number
of manufactured mixtures for special purposes such as cleaning or industrial solvent
action. When heated, such mixtures allow the distillation of their most volatile con-
stituents first, followed in turn by constituents of decreasing volatility (a process often
called weathering, which will be discussed more fully in Chapter 14 ). In general, it is the
most volatile constituent of a mixture that is of most significance as it relates to the fire
hazard. When a liquid fuel that has a boiling point range (e.g., a mixture such as gaso-
line) burns, the more volatile species evaporate first and fastest from the surface layer. The
bulk liquid beneath the surface will retain the lighter species for some time, since they will
have to migrate by diffusion (aided by convective flows) through the bulk liquid to reach
the surface before they can evaporate.16 A consequence of such fractional distillation is
that the residue of a petroleum distillate may have a lower autoignition temperature than
the bulk liquid, possibly posing a risk of delayed ignition.17
In initiating a fire, materials of high boiling point will rarely be above their flash point
at ambient temperatures, and only the more volatile (low boiling point) will be expected
to show special hazard. As a rule, the boiling point and the flash point tend to parallel
each other, so the low-boiling material will usually have a low flash point as well. The
presence of a mixture does not greatly diminish the fire hazard of the most volatile com-
ponent in that mixture. A mixture of gasoline and diesel fuel, for instance, exhibits a flash
point indistinguishable from that of gasoline alone until the gasoline represents less than
about 5 percent of the mixture.

VAPOR DENSITY
The fundamental property of a vapor that predicts its behavior when released in air is its
vapor density. The vapor density, that is, the density of a vapor relative to that of air, may vapor density ■ The
be simply calculated by dividing the molecular weight of the vapor by the mean molecular ratio of the weight of
a given volume of gas
weight of air, which is approximately 29. This relationship is represented by the equation
or vapor to that of an
Vapor density (gas or vapor)  Molecular wt of gas>Molecular wt of air equal volume of air.

For example, the lightest gas, hydrogen, has a molecular weight of only 2; therefore,
its vapor density (V.D.) is
V.D.  2>29  0.07
Methane, the simplest hydrocarbon fuel gas found in swamp gas and natural gas, has
a molecular weight of 16. Its vapor density, then, is roughly 16> 29, or 0.55. Ethane, the
other major component of natural gas, has a molecular weight of 30, making it almost
exactly the same density as air, and giving it a vapor density of 1.03. The vapor densities
of some common fuels are shown in Table 4-5.
Gases or vapors with a vapor density greater than 1.0 are heavier than air and tend
to settle through the air into which they are released until they encounter an obstruction,
such as a floor, after which they tend to spread outward at this level, much in the same
manner as if they were liquids. In contrast, vapors lighter than air tend to rise through
the air until an obstruction, such as a ceiling, is encountered, after which they spread at
the high level. The tendency to spread is less absolute than with liquids because the dif-
ference in density as compared with air is much less than compared with any liquid, and
air currents produce mixing. Some mixing with the air is inevitable, and the nearer the
vapor density is to that of air, the greater the mixing of fuel and air. This effect of greater
mixing is due to the process known as diffusion. All gases, regardless of their molecular
weight or vapor density, tend to diffuse into each other when the opportunity exists.
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 99
 TABLE 4-5 Vapor Densities of Common Fuel Gases and Vapors

FUEL VAPOR DENSITY (AT STANDARD TEMPERATURE AND PRESSURE)

Hydrogen 0.07
Methane 0.55
Acetylene 0.90
Carbon monoxide 0.97
Ethane 1.03
Propane 1.51
Butane 1.93
Acetone 2.00
a
Pentane 2.50
b
Hexane 3.00
a
Lightest liquid component of gasoline.
b
Lightest major component of camping fuels.

Differences in diffusion rates of gases depend on their vapor density: when a fuel
vapor is much heavier than air (V.D. 1.0), its diffusion rate will be much less than
when its density is much closer to that of the air into which it is diffusing, that is, when
its vapor density is close to 1.0. The diffusion of a vapor is a very slow process. In a closed
container, diffusion of a vapor will eventually produce a uniform distribution after some
hours. In an open system such as a room, diffusion of vapors from a volatile liquid does
not produce a uniform distribution but rather a gradient of concentration from very high
(saturation vapor pressure) to very low, as represented in Figure 4-8.
Methane, as noted, is lighter than air and thus tends to rise in a room (as long as its
temperature is the same as that of the room air). Its rate of diffusion is greater than that
for heavier molecules, and so it tends to mix with air and form a gradient of concentra-
tion, as illustrated in Figure 4-8a. The same is true for natural gas, which is a mixture of
methane and ethane and other light gases, so its vapor density will be between 0.55 and
1.0, depending on the relative proportions of the components. Ethane is rarely encountered
by itself, but because its vapor density is close to 1, it will diffuse readily in air and neither
rise nor settle in air, there being no gravitational separation. (It should be noted that diffu-
sion acts to keep vapors diffused and mixed in air and once mixed in air, vapors will not
“settle out” to re-form layers of high concentration.) The same is true for acetylene (C2H2)
with a molecular weight of 26 and therefore a vapor density of 0.9, as well as ethene
(C2H4) and carbon monoxide (CO), each with a molecular weight of 28 and therefore a
vapor density of 28/29, or 0.97. All other commonly encountered fuel gases and vapors
have a vapor density significantly greater than 1 and will tend to sink when released in air.
For example, propane, the next heavier hydrocarbon, (V.D.  1.51) will mix readily with
air, but it will tend to sink as it does so. [It should be noted that a flammable propane/air
mixture (2  9.6 percent) will have a vapor density very close to that of air.] Butane (V.D. 
1.93) sinks more readily and tends not to diffuse as quickly. All higher hydrocarbons, char-
acteristic of most petroleum products, will of course have slower diffusion rates (diffusivity)
along with their greater vapor density. Gasoline vapors, for instance, contain a mixture of all
the most volatile components of the liquid blend. This mixture, being considerably heavier
than air, will sink rapidly to the floor or ground surface. Gasoline vapors will therefore
pour into low regions or down drains and will carry their intrinsic hazard to the lowest
possible region (see Figures 4-8b and c for a graphic representation).
100 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 FIGURE 4-8 Typical
Rich
mixtures of natural gas or
flammable fluid vapors in
Explode closed rooms. (a) Natural
gas (methane) in a room.
(b) Gasoline in a room.
(c) Gasoline (or butane)
in a room with various
Lean elevations.

(a)

Lean

Explode
Liquid

Rich

(b)

Lean
Source

Explode

Rich

(c)

It should be remembered that from a fire hazard standpoint, the vapor density and
diffusion rate of the lightest (most volatile) component of a mixture is what will deter-
mine the spread and ignitability of vapors, not the properties of the bulk liquid. For
instance, gasoline has a range of components, some of which, like n-decane, have a very
low vapor pressure at room temperature and a correspondingly high flash point. It is the
isobutane and n-pentane content of automotive gasoline that generates the bulk of the
vapors at ordinary room temperatures and produces the significant ignition risk.
Gasoline’s heavier components such as the xylenes do not contribute significantly to the
vapors until the gasoline has been evaporating for many minutes, by which time the
lighter components will have already been ignited.18 In general, the vapors of fresh auto-
motive gasoline will behave exactly like those of n-pentane in their movement.
It is apparent that various concentration layers will exist (with a gradient of interme-
diate concentrations between the heaviest and lightest layers). Some of these vapor layers
will be within their explosive (flammable) limits, while the mixture in other layers will be
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 101
 either too rich or too lean to be ignited. These distributions represent release into a sealed
room with still air in which there is no physical activity and in which the vapors are at
the same temperatures. In assessment of real-world distribution of vapors, the effects of
temperature and air currents must not be overlooked. Mechanical activity such as a fan,
furnace, operating machinery, or even someone walking will stir vapors into mixtures
throughout a room. The draft created by a water heater burner, furnace burner, or open
window may be enough to redistribute vapors and may draw them into contact with an
open-flame ignition source at the same time. Even butane if warmed sufficiently, will rise
to the top of a room much like methane because the vapor density drops rapidly with
increased temperatures. Vapors from a flammable liquid cooled by evaporative cooling
will be even denser, and their convective flow will enhance their tendency to settle.19
When a gas mixture is formed from gases at different temperatures, the temperature alone
may control its movement initially until the temperature differences are equalized. The
alert investigator must be aware of all these influences (including timer- or thermostat-
controlled equipment that may cycle on with no one in attendance) that can distribute
and circulate vapors and possibly provide an ignition source as well. In the United States,
gas-fired water heaters have been required since 2005 to be designed so as to make it very
difficult for their burners to ignite gasoline vapors. One design places the entire heater in
an 18-in.-deep “bucket” to keep vapors out of the burner chamber. The other uses a flash
suppressor on the combustion chamber to prevent propagation outward. These designs
will be described in more detail in Chapter 6.
The concentration gradient established by a heavier-than-air gas (such as LP gas)
leaking into still air has been shown to be very steep.20 This means that a person stand-
ing in the room may not be able to detect the leak by its odorant while a layer below knee
height in the same room is within its explosive range. Recent experiments have shown
that thermal currents, even in a small space such as a camper or RV, result in almost com-
plete mixing of propane into air, with no detectable gradient.21
One interesting property of dense vapors (V.D. 2.5) is their propensity to flow hori-
zontally, like a viscous liquid. Pentane vapor released from a pool at room temperature
will diffuse upward only a few centimeters, and then the mass of vapor will slump side-
ways and spread outward along the surface at a rate of about 0.05 m/s in still air. This
so-called advective flow can spread vapors along a surface much more quickly than diffu-
sion, but this flow is easily overcome by even a modest draft.22
Careful consideration of the distribution of fuel gases and vapors must be made when
assessing the likelihood of ignition by possible sources in the compartment. If an ignition
source like a candle flame is high up in a room, LP gases or gasoline vapors are going to
require a long time to diffuse sufficiently to reach their flammable range in the vicinity of
the candle flame unless there is some external means of circulation like a fan or strong
draft, or the vapors are released under pressure (jet). Similarly, a kerosene heater burning
normally at floor level in a room is unlikely to ignite natural gas leaking into the room
until the gas can fill the room from the top down in the absence of mechanical circula-
tion. The mere presence of an ignition source and a fuel source in the same room is not a
guarantee that there will be ignition. The assessment of possible ignition sources for such
vapor/air mixtures should take into account all these contributing factors, including the
heights of, and protective enclosures around, appliance burners.
It was once thought that the distribution of blast effects from a vapor/air deflagration
was directly linked to the preignition distribution of vapors. However, experiments by one
of the authors have demonstrated that even when a highly localized (floor) layer of
n-hexane vapors is ignited, the pressures are equalized throughout a moderate-sized com-
partment within 5 milliseconds (ms).23 This means that the walls of the average compart-
ment will be exposed to the same pressure at the same time from within and will therefore
fail where it is weakest structurally. It is frequently noted in explosions in wood-frame struc-
tures that it is the bottom of the wall that moves rather than the top, despite whether the
102 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 FIGURE 4-9 Pocketed
burning of natural gas
accumulated under this
wooden floor produced
significant charring of
joists and floor. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

deflagration was fueled by natural gas or LP gas. This result can mean that the bottoms of
the walls were not as well anchored to the foundation or slab as the tops of the walls were
to the structure above. The bottoms of wood-frame walls are also much more likely to be
weakened by termites or dry rot. When a fire follows an explosion, however, if the turbu-
lence of the explosion has not completely dissipated the layer of fuel vapor, the fire will tend
to burn at the interface between the richest layer of fuel and the air (as discussed previ-
ously). Thus, a fire following a hexane (or gasoline) vapor explosion will tend to burn atop
or within a remaining layer of dense vapor (where the fuel-rich layers will burn longest).
The distribution of fire damage low down in the room after a vapor ignition then indicates
a heavier-than-air vapor has been involved. The reverse will occasionally occur on the
underside of a natural gas layer trapped or pocketed on the underside of a floor or roof
structure, as in Figure 4-9. In either case, the location and distribution of postexplosion fire
damage may give important clues as to the vapor density of the fuel responsible.
The volume of vapor that can be generated by the evaporation of a known volume of
a volatile liquid may be important in evaluating the ignition potential of accidental or
even intentional spills. If total evaporation can be assumed, the volume of vapor gener-
ated (at standard temperature and pressure) by 1 U.S. gallon (3.8 L) of liquid can be cal-
culated from the empirical relationship
Vapor (in cubic feet, ft3)  111  Specific gravity of liquid*  Vapor density of liquid
For example, 1 U.S. gallon of diethyl ether, with a specific gravity of 0.7 and a vapor den-
sity of 2.55, will produce
111  0.7>2.55  30.4 ft3 of diethyl ether vapor
Because diethyl ether has a flammability or explosive range of 2 to 36 percent in air,
30.4 ft3 of vapor, if uniformly distributed, will produce an explosive mixture in a room
of 84 to 1,520 ft3 in volume. In metric units the equivalent relationship for 1 L of liquid
(0.001 m3) is
Vapor (in cubic meters, m3)  0.85  Specific gravity  Vapor density

*See Table 4-4 for specific gravity values.

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 103

 For commercial propane, the liquid/vapor expansion rate is on the order of 272; that
is, 1 L of liquid propane will produce about 272 L of vapor at standard temperature and
pressure. Commercial butane has an expansion factor of approximately 234.24 The
expansion rate for LP gas itself will vary with the composition of the LP gas, since the
rates for each component are different.

HEAT OF COMBUSTION
Another property of fuels that is often of great importance to the investigator is the heat
of combustion of the fuels involved, as introduced in Chapter 2. If the approximate
amounts of fuels in a room can be estimated by weight, the intensity of burning observed
can be compared with the possibilities of various combinations of fuels. As we saw in
Chapter 3, the rate at which heat is released is more critical to reconstructing the events of
a fire than the total amount of heat, but it is sometimes useful to be able to recognize when
the fire damage to a room simply does not tally with the amount of heat that should have
been produced by the fuels known to be present. The causes for such differences can then
be explored. The scientific method question is: Was there so much damage that it cannot
be accounted for by assessment of fuel loads normally present? Was there reduced fuel
because stock had been removed prior to the fire? Was there reduced ventilation available,
or was there enhanced combustion due to oxygen enrichment? Was the lack of damage due
to the incomplete or inefficient combustion of the available fuels? Some of the most com-
mon fuels, such as wood, have such variable properties as to give only approximations of
the actual output in a specific instance. Table 4-6 summarizes the heat of combustion of a
few of the most important fuels from the standpoint of the fire investigation.
adiabatic ■ Conditions The heat of combustion is used to calculate a theoretical adiabatic flame temperature
of equilibrium of tem- for each fuel. These calculated values are sometimes erroneously used to estimate the
perature and pressure.
potential effect the flames will have on surrounding materials. Work by Henderson and
Lightsey indicates that the actual measured flame temperatures produced by such fuels
when burned in air are much lower than previously thought.25 The measured tempera-
tures they reported are included in Table 4-6. The roles of the flame temperatures and heat
of combustion in fire investigation will be explored in later chapters. It should be noted
that the real-world maximum open-flame temperatures of nearly all “normal” fuels burn-
ing in turbulent diffusion flames is about 900°C (1,700°F), and this figure is used in most
calculations. Some fuels (such as styrene, crude oil, and urethane foam) that burn with
very smoky flames, so that the heat produced cannot be readily lost (low emissivity), pro-
duce very high temperatures in the flames. Fuels that burn very cleanly with no soot or
pyrolysis products (alcohols) also have a low emissivity and high flame temperatures.

Hydrocarbon Fuels
By far the most important fuels from the standpoint of fire investigation are the hydro-
carbons, which range from the light gas methane to much heavier compounds, including
oils and asphalts. (For a discussion of the chemistry of hydrocarbons, see Chapter 2.)
Because of the importance of hydrocarbons to fire investigation, some of their specific
properties will be discussed in order of their molecular weight or chemical classification.

NATURAL GAS
Natural gas, consisting chiefly of methane, is of obvious significance. Natural gas from
different geological formations shows considerable variation in composition and may
contain some noncombustible gases. It is typically 70 to 90 percent methane, 10 to 20
percent ethane, and 1 to 3 percent nitrogen, with traces of other gases.26 For most pur-
poses, the investigator may consider its properties to be the same as those of methane,
without introducing serious error.
104 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 TABLE 4-6 Heats of Combustion and Flame Temperatures of Some Fuels

HEAT OF COMBUSTION FLAME TEMPERATURE (°C)

FUEL (MJ/kg) NET ADIABATIC ACTUAL
Acetylene 49.9a (55.8MJ/m3) 48.2b 2,325
b 3 c
Butane 49.5 (112.4 MJ/m ) 45.7 1,895
3 c
Carbon monoxide (11.7 MJ/m ) 10.1
Charcoal 33.7 34.3 2,200d 1,390d
Coleman fuel 770e
Coal (anthracite) 30.9–34.8
Cotton 16
Ethane 47.5 (60.5 MJ/m3)c 47.4 1,895
Ethyl alcohol 29.7 26.8 840e
Fuel oil #1 (kerosene) 46.1
Hexane 48.3 (164.4 MJ/m3)c 43.8 1,948/2,200e
Hydrogen 130.8 (12.1 MJ/m3)c
Gasoline 43.7 810e
3 c e
Methane 55.5 (34 MJ/m ) 50 1,875
3 c
Propane 50.4 (86.4 MJ/m ) 46.4 1,925e 970e
Polyethylene 43.2
Polypropylene 43.2
Polystyrene 40
Polyurethane 23
Wood 20 16–20 1,590f ~600–1,000f
a
Fire Protection Handbook, 19th ed. Quincy( MA: National Fire Protection Association, 2003), table A.1.
b
SFPE Handbook of Fire Protection Engineering, 4th ed. (Quincy, MA: Society of Fire Protection Engineers and the
National Fire Protection Association, 2008), tables 1-5.3 and 1-5.6.
c
R. J. Harris, The Investigation and Control of Gas Explosions (New York: E and FN Spon, 1983), 6–7.
d
R. W. Henderson and G. R Lightsey, “Theoretical Combustion Temperature of Wood Charcoal,” National Fire and
Arson Report 3 (1985): 7.
e
R. W. Henderson and G. R Lightsey, “The Effective Flame Temperatures of Flammable Liquids,” Fire and Arson
Investigator 35 (December 1984): 8.
f
R. W. Henderson and G. R Lightsey, “Theoretical vs. Observed Flame Temperatures during Combustion of Wood
Products,” Fire and Arson Investigator 36 (December 1985).

LIQUEFIED PETROLEUM GAS

Liquefied petroleum gas (LPG) is commonly used in rural areas where natural gas is not
available. It is a mixture of propane and n-butane, with small quantities of ethane, ethyl-
ene, propylene, isobutane, and butylene.27 Depending on its source and its intended use,
it can be quite variable in its composition, ranging from almost pure propane to a heav-
ier mixture largely composed of butane. For LPG sold in the United States, the investiga-
tor can consider LPG the same as propane in its physical and combustion properties
without serious error (in some other countries, LPG can comprise primarily butane). If
butane is known to have been involved (from lighters, torches, or aerosol containers), it
is best to use its reference values for calculations.
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 105
FIGURE 4-10 Boiling Jet fuel A Fuel oil #2
point ranges of straight Kerosene
petroleum distillates. Diesel
MPD Waxes
Straight-run Oils Asphalts
LPG gasoline

ⴚ50 0 100 200 300 400 500 600 650

Boiling point ranges (ºC) of straight petroleum distillates

PETROLEUM
Petroleum in its crude state is a thick oil varying in color from light brown to black. It con-
tains a very large number of compounds and different types of compounds, which are sep-
arated to a considerable extent in the manufacture of petroleum fuels. Products obtained
from the distillation of petroleum include gases, petroleum ether, straight-run gasoline,
kerosene, medium distillates including paint thinner–type mixtures and solvents, diesel
fuels, heavy lubricating oils, petroleum jelly, paraffin, waxes, asphalt, and coke. The distil-
lation curve of petroleum products is represented in Figure 4-10.

GASOLINE
Gasoline, widely used in vehicles and appliances and perhaps the most important fuel of
petroleum origin, is a mixture of volatile, low-boiling, and midrange hydrocarbons. It con-
tains hydrocarbon compounds with boiling points between approximately 32°C (90°F)
and 205°C (400°F). The average molecular weight of gasoline is sometimes taken to be
close to that of n-octane, about 114. The petroleum distillate described as “straight-run
gasoline” contains the raw hydrocarbons that distill off between 40°C and 200°C (100°F
to 395°F) and is used as camping fuel and sometimes called “white gas” or naphtha. [The
range of compounds included is often measured by the “carbon number” or molecular
length of n-alkanes included. In this case, this distillate “cut” includes C4 (butane ) to C12
(dodecane).] Modern automotive gasoline contains more than 200 hydrocarbons in a com-
plex mixture (which is not just the straight cut or fraction but a blend of many compounds,
especially aromatics) whose relative component concentrations vary little from refinery to
refinery. This means that it is not generally possible to identify one oil company’s product
from another based on a comparison of the basic hydrocarbon “profiles” of various gaso-
lines, Recently, forensic chemists have noticed that component ratios are changing and sub-
accelerant ■ A fuel stitutions are being made, so laboratory identification of gasoline as an accelerant is no
(usually a flammable longer as straightforward as it once was. Before gasoline is sold as a motor fuel, however,
liquid) that is used to
a variety of additives are added by the refinery. These additives often include compounds
initiate or increase the
intensity of speed of to improve the burning characteristics of the fuel. Dyes were once added to distinguish
spread of a fire. leaded from unleaded motor fuels. Today, nearly all automotive gasolines are unleaded,
and dyes are rarely seen except to identify lower-taxed fuels for agricultural use.
Oxygenates, especially ethanol and methanol, are very common in automotive gaso-
lines today. In many areas they are completely replacing MTBE (methyl t-butyl ether), once
commonly used as an oxygenate additive. The additive packages used by one manufacturer
are usually different from those used by others and offer some means of distinguishing
fresh, unburned, and uncontaminated gasolines. Unfortunately, due to the industry prac-
tice of exchanging raw petroleum feedstocks, additive packages, and even finished gasoline
depending on market and supply conditions, it is not possible in most cases to identify a
particular retail brand of gasoline from a sample of the fuel as it is delivered to the retail
customer. However, recent advances in gas chromatography and data interpretation are
making it possible to compare unburned fuels against suspected sources with considerable
specificity. See Chapter 14 for a further discussion of identification of such products.
106 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
KEROSENE AND OTHER DISTILLATES
Kerosene is defined by ASTM D3699 as having a boiling point range from 175°C to
300°C (350°F to 572°F) with a minimum flash point of 38°C (100°F).28 Kerosene and the
distillate fuels of higher boiling ranges were considered useful only for illumination until
the diesel motor became a common device. Although diesel fuel is similar to kerosene, it
spans a wider range of less-volatile components. The modern jet engine and other turbine-
driven engines further increased the use and economic significance of kerosene-like fuels.
The classification scheme used for laboratory characterization of ignitable liquids (those
liquids considered either flammable or combustible) includes an intermediate range called
medium petroleum distillates (having a boiling point range between about 125°C and
215°C (250°F to 400°F). This range includes carbon numbers from C9 to C17. Many
household and commercial products fall into this class—paint thinners, mineral spirits,
some charcoal starters, and even some insect sprays. This classification scheme will be dis-
cussed in more detail in Chapter 14.
Kerosene and similar petroleum distillates have long been of significance in the set-
ting of deliberate fires. Their lower volatility presents a lesser hazard to the user than does
gasoline. The liquid evaporates more slowly, so that less haste is required in its ignition,
and there is much less danger of explosion. Kerosene-class compounds have very low
autoignition temperatures, on the order of 210°C (410°F) (see Table 4-4). Interestingly,
despite their lower AITs, their high flash points make them less likely to aid the spread of
fire (see Table 4-3). Their compositions involve several series of higher-molecular-weight
hydrocarbons. Kerosene (fuel oil #1, Jet Fuel A) generally contains paraffinic and olefinic olefinic ■ Hydrocarbons
hydrocarbons in the C10 to C16 range. (This range denotes hydrocarbons having 10 to 16 containing double
carbon atoms linked together; see Appendix B.) These have boiling points between 175°C carbon ¬ carbon bonds
(denoted C “ C); nonsat-
and 260°C (350°F to 500°F) and therefore fall in the fraction that follows (with some urated hydrocarbons;
overlap) gasoline in the simple distillation of crude oil. alkenes.

DIESEL FUEL
Diesel fuel contains predominantly paraffinic hydrocarbons with low sulfur content and
additives to improve combustion. It covers the boiling point range from 190°C to 340°C
(375°F to 650°F), which includes carbon numbers from C10 to C23. Fuel oil, or domestic
heating oil, is the next-heavier product, representing a distillation range of approximately
200°C to 350°C (400°F to 675°F). (Diesel fuel and domestic heating oil may both some-
times be described as fuel oil #2, depending on local usage.) The product for home use is
carefully controlled to minimize sulfur and water content, which could corrode heating
units having minimal service. Those fuels intended for industrial service have higher sul-
fur and ash content and cover a range of products and applications. Diesel fuels have a
higher autoignition temperature (250°C or higher) and a low vapor pressure. In a fire,
unless present on a porous wick or as an aerosol, diesel fuels only reluctantly support fire.
Dyes may be added to denote diesel fuel intended for farm use, which is often taxed at a
lower rate than that for road use.

LUBRICATING OILS
Lubricating oils include a tremendous variety of hydrocarbons with a wide range of vis-
cosities and thermal and frictional characteristics to suit every lubrication need.
Although lubricating oils are not readily combustible at ordinary temperatures, at ele-
vated temperatures they can add to the fuel load of an established fire. Release of an oil
under pressure creates an aerosol of fine droplets that are more easily ignited than the
bulk liquid. It should be kept in mind that used motor oil (which is often stored in a
casual fashion) often contains a significant percentage of gasoline residues. These con-
taminated oils can have a much lower flash point than the pure oil and can be the source
of accidental fire.
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 107
 SPECIALTY PETROLEUM PRODUCTS
Specialty petroleum products are not true petroleum distillates but are used in such a
wide variety of consumer products that they will show up in fire debris even if they are
not involved directly in the fire’s ignition. These products are manufactured or blended
for specific chemical and physical properties. Isoparaffinic blends (see Chapter 2) are
made in various boiling point ranges and may be found in copier toners, cigarette
lighter fuels, charcoal starters, lamp oils, and even in some waterless hand cleaners.29
Aromatic blends, sometimes called LPAs for low-paraffinic aromatics, are blends of aro-
matics and naphthenic hydrocarbons intended for use in adhesives and some insecti-
cides. They can be found as background volatiles in household furnishings and are
chemically similar to the aromatic fractions found in automotive gasolines.30 There are
also single-component products such as limonene (used as a cleaner) and n-tridecane
(fuel for liquid candles). They are all combustible liquids and are rapidly replacing
many traditional petroleum distillates in consumer products. Some are being encoun-
tered as arson accelerants.

Nonhydrocarbon Liquid Fuels

Many materials other than hydrocarbons have considerable significance to the fire inves-
tigator because they can be used as accelerants and because they can play a part in the
ignition and spread of accidental fires. These nonhydrocarbon liquid fuels include alco-
hols, solvents, and similar materials, as well as alternative or biofuels, as discussed next.
The chemistry of nonhydrocarbons was discussed in Chapter 2.

ALCOHOLS, SOLVENTS, AND SIMILAR NONHYDROCARBONS

Methyl alcohol was formerly widely used in spirit (mimeograph) duplicating machines,
as an alternative motor fuel, and as a thinner for shellacs and finishes. Ethyl alcohol, espe-
cially in reagent (190 proof) or high-proof beverage form (85 or higher proof), can also
be used to start fires even though its heat output is not very high. Isopropyl (rubbing)
alcohol is flammable at the concentrations packaged for retail use. All the simpler alco-
hols can provide fuel for the incendiary fire and accidental fire as well and are difficult to
detect in post-fire debris because of their extreme volatility and water solubility. Ketones
such as acetone and methyl ethyl ketone are widely used as solvents for paints and
finishes, as are ethyl acetate and related compounds. Although more and more paints are
water based, thus requiring no volatile solvents, many wood finishes employ turpentine
as thinner or linseed oil (or related organic oil) as part of the finish. A tremendous vari-
ety of proprietary mixtures used as mop cleaners, dust attractants, insecticide solvents,
copier toners, glue solvents, and plastic resin solvents are encountered at fire scenes. They
require mention because of a predisposition to think of liquid fuels only in terms of petro-
leum distillates. Such thinking can be totally erroneous in many fires. Industrial and com-
mercial fires are often fed and intensified to extremes by special-purpose cleaning agents
and solvents. The general properties of many of these compounds have already been dis-
cussed. Because of the very large number of specific compounds and the variety of mixtures
in which they may be found, no attempt will be made here to examine each of them. Special
considerations of some of the more hazardous of these fuels will be made in Chapter 13.

ALTERNATIVE FUELS OR BIOFUELS

Currently, there are motor fuels in wide use that are not petroleum derivatives. Alternative
fuels or biofuels are based on vegetable materials–either processed food oil or ethanol pro-
duced by fermentation of corn or other vegetable materials. These fuels can be pure biofuels
or, more commonly, a blend of biofuel and petroleum product.31
108 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
Biodiesel
Biodiesel (sometimes denoted as Bl00) is produced by chemically treating vegetable oils
(corn, peanut, safflower, canola, etc.) or, more rarely, animal fats to yield a mixture of four
fatty acid methyl esters. (The ratios of the few components depend on what oil or fat was
the starting point.) There are presently more than 770 commercial outlets for B100
biodiesel in the United States and countless informal producers. Biodiesel can be used as a
direct substitute for diesel fuel in vehicles with minor mechanical modifications. A 20/80
blend of biodiesel and petroleum diesel (denoted as B20) is also in widespread use, since it
requires no mechanical modifications for use in most motor vehicles. Its performance as a
fuel in a motor or in an open fire will be dominated by its diesel fuel constituents. Biodiesel
fuels may contain residues of methanol and water from the production process. Because
the molecular weights of biodiesel components are similar to C18 to C23 alkanes, the flash
points are going to be similar (100°C). The presence of significant quantities of residual
methanol will obviously greatly reduce the flashpoint.

Ethanol
Pure ethanol is not suitable as a motor fuel, so the most common variety is E85 (85 per-
cent ethanol and 15 percent gasoline). This product is already being sold across the
United States for use in “Flex Fuel” vehicles. Motor gasolines in the United States have
contained up to 10 percent ethanol as an oxygenated additive to reduce carbon monox-
ide content. As a motor fuel, this product must meet the same flash point, vapor pressure,
and ignition temperature standards as gasoline fuel. The effect on hazard, then, is
expected to be minimal. However, its detection after a fire with water extinguishment
presents a problem to be discussed in Chapter 14.

Combustion of Liquid Fuels

It must be remembered that when an ignitable liquid is poured or spilled on a surface,
only its vapors (rather than the liquid itself) actually combust. The behavior of the liquid
and its vapors is generally predictable. On a nonporous surface (linoleum, vinyl, painted
concrete), the viscosity and surface tension of the liquid will determine the depth of the
freestanding pool that results. The depth, of course, determines how large an area the
pool will cover for a given quantity of fuel. Viscous liquids like kerosene or diesel fuel
form a deeper pool (~1 mm) that will not cover as large an area as a nonviscous fuel like
methanol or gasoline (whose freestanding pool may be only 0.1 to 0.5 mm deep).32
On other surfaces, the depth of the pool will be determined by the nature of the sur-
face, as shown in Figure 4-11. A semiporous surface like raw wood or concrete will allow
some penetration, 2 or 3 mm or so, with a proportional reduction in the area of the pool.
Carpets or other porous surfaces allow very deep penetration, so the resulting pools can
be very small. The prefire evaporation rate from such porous surfaces is considerably
faster than the rate from a freestanding pool (of the same diameter) of the same liquid at
the same temperature due to the capillary drive of the substrate (the wick effect).33 As
described previously, prior to ignition the vapors generated by a pool will rise by diffusion

Upward diffusion of vapor FIGURE 4-11 Liquid

Horizontal advective Liquid residue Penetration into pools on nonporous and
Liquid flow of vapor on surface substrate porous surfaces with
vertical and horizontal
Vapor Vapor movement of vapors.

Nonporous surface Porous or semiporous surface

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 109

 FIGURE 4-12 Distribution
and effects of radiant
heat from burning pool
of ignitable liquid.
Central radiant
heat into pool Radiant heat at edges
absorbed by substrate

Protected by overlying pool

Substrate Localized scorching

or other heat effects

and spread sideways by advective flow, as in Figure 4-11. Upon ignition, the flames are
established everywhere the vapors have diffused into the surrounding air in a flammable
mixture. The radiant heat from flames near the center of the pool is in part absorbed by
the liquid remaining in the pool, and the radiant heat from flames at the edges is absorbed
in part by the adjoining floor surface that is not protected by the liquid. This radiant heat
may be enough to damage or even ignite the floor exterior to the pool, as in Figure 4-12.
Meanwhile, the heat being absorbed by the liquid is being distributed through the liquid
by convective circulation, and the temperature of the entire pool begins to rise. The liquid
does not really cool the surface beneath it, but it can protect it because the temperature, obvi-
ously, cannot exceed the boiling point of the liquid. For a low-boiling-point fuel like methyl
alcohol, the temperature of the floor under the pool will not exceed about 65°C (150°F)
(the boiling point of methyl alcohol), and so the floor under the pool escapes all damage
until the alcohol evaporates. With gasoline, there is progressive distillation and consump-
tion of the components from light to heavy until only the heaviest remain (controlled by
the diffusion mechanism through the bulk liquid described earlier). The bulk liquid
temperature is limited only by the boiling point of the heaviest constituents, such as n-
dodecane with a boiling point of 216°C (421°F). As the lighter components boil off, the

TABLE 4-7 Properties of Petroleum Products

AUTOIGNITION
PETROLEUM DISTILLATE BOILING POINT/RANGE FLASH POINT TEMPERATURE
Ethanol 78.5°C (173°F)* 13°C (55°F) 363°C (685°F)
*
Toluene 110.6°C (231°F) 4.4°C (40°F)
Gasoline (low octane) 32°C190°C (90°F375°F) 43°C (45°F) 257°C (495°F)
*
Methanol 64.7°C (148.5°F)
Medium petroleum distillate 125°C215°C (250°F400°F) 13°C (55°F) 220°C (428°F)
Mineral spirits Varies 40°C (104°F) 245°C (473°F)
VM & P Naphtha (regular) -2°C (28°F) 232°C (450°F)
Kerosene (C10–C16) 175°C300°C (350°F500°F) 38°C (100°F) 210°C (410°F)
n-Dodecane 216°C (421°F)* 74°C (165°F) 203°C (397°F)
Fuel oil #1 175°C260°C (350°F500°F) 43°C72°C 210°C (410°F)
(110°F162°F)
Diesel fuel (fuel oil #2) 200°C350°C (400°F675°F) 52°C (125°F) 257°C (494°F)
n-Pentane 36°C (97°F) 40°C (40°F) 260°C (500°F)

Sources: Fire Protection Guide to Hazardous Materials (Quincy, MA: National Fire Protection Association, 2001).
*
Merck & Co., The Merck Index, 11th ed. (Rahway, NJ: Merck, 1989).

110 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

liquid can reach temperatures such that some floor materials may be scorched. When
heavy petroleum products like diesel fuel are burning, the temperatures of the liquid can
be as high as 340°C (650°F), which is above the ignition temperatures of solid fuels like
some woods. (See boiling points in Table 4-7.) (As discussed earlier, partial evaporation
of petroleum distillates changes the ratio of components, since lighter ones evaporate faster
than heavier ones.) This mechanism explains why some fuels produce halo-type burns,
and others scorch and char the floor more extensively. Because they are so shallow, free-
standing pools of most ignitable liquids burn away quite rapidly (the rate being dependent
on the nature of the surface and the size of the pool, with very small pools (0.1 m; 4 in.
in diameter) burning away faster than larger pools. When there are cracks or crevices,
localized burning can take place for longer periods of time, inducing localized charring of
the floor at the joints and seams, as illustrated in Figure 4-13a. If extra ventilation can then

(a) FIGURE 4-13A Pocketed

burning from liquid in
cracks or seams in a floor
may be enhanced by any
upward draft and radiant
heat feedback within
cracks. Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

(b) FIGURE 4-13B Radiant

heat striking a wood
plank floor can initiate
charring of the exposed
Radiant heat edges, causing a beveled
appearance. Combustion
between the planks can
be enhanced by draft
from below. This can be
very difficult, if not
Draft impossible, to differenti-
ate from (a).

FIGURE 4-13C
Gasoline/diesel mix
burning in seams of
wood floor. Fuel on flat
surfaces has already
burned off.Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 111

 be drawn through these cracks, severe but localized floor damage can result, as shown in
Figure 4-13c. Downward radiant heating can produce “beveled” edges on floorboards, as
pictured in Figures 4-13b and 7-21. It can be very difficult, if not impossible, to differen-
tiate between these effects.

Pyrolysis and Decomposition of Liquids

Although some unstable liquids decompose at fairly low temperatures, most fuel liquids
reach their boiling point and boil away before being subject to pyrolytic decomposition.
Even when liquids are exposed to an intense general fire, convection within a deep pool
(as opposed to a thin film) helps mix the cooler and warmer liquids. The liquid at the surface
boils off into the flame, with the heat absorbed from the flames being lost by evaporative
cooling. This process maintains the temperature of the entire pool at or just below the
boiling point of the liquid; the liquid pool protects the surface on which the pool is sitting.
When dealing with viscous liquids such as asphalt, plastic resins, or some foods, it must
be remembered that with impaired circulation, portions of the liquid in contact with an
intense heat source can reach and exceed their pyrolysis temperatures and begin to char
and decompose like a solid. The pyrolysis products then enter the flame in the same way
as those driven off a solid.
In tank or pool fires involving heavy crude oils, the residue formed on the burning
surface (by distillation of the lighter components) may become denser than the surround-
ing liquid. The hot residue can then settle into the tank and contact water in the bottom,
producing a steam explosion that propels the burning fuel above outward with disastrous
results. This situation can also occur with nonpetroleum liquids like cooking oils that can
pyrolyze and form a true char.
Liquid fuels are categorized by their flash point (see Table 4-3.) Flammable liquids
are all those that have a flash point below 100°F (37.8°C) (NFPA Classification: Class I),
with Class IA and IB liquids having flash points below room temperature, 73°F (22.8°C).
Combustible liquids are those with a flash point above 100°F (37.8°C) [NFPA Class II,
flash point between 100°F and 140°F (37.8°C and 60°C), and Class III, flash point
above 140°F (60°C)]. Class III combustible liquids are further divided into Class IIIA
(flash point between 140°F and 200°F) and Class IIIB (flash point above 200°F
(93°C)].34 In the United Kingdom, highly flammable liquids are those that have a flash
point below 32°C (90°F); flammable liquids are those with a flash point between 32°C
and 60°C; and combustible liquids are those with a flash point above 60°C (140°F).35
ignitable liquid ■ Ignitable liquids are a category of liquid fuels that includes both flammable and com-
Classification for liquid bustible liquids.
fuels including both
flammable and com-
bustible classes.
Fuel Gas Sources
Among the many sources of fuel gases, of particular interest to the fire investigator are gas
lines, natural gas, liquefied petroleum gas, containers, and appliances, as discussed next.

GAS LINES
Although gas lines do not normally represent a source of ignition, they do represent a
source of readily ignitable fuel. Because of the properties of gas, outlined previously and
in Chapter 12, it is clear that gas enclosed in a pipe does not normally pose any hazard.
Unless mixed with appropriate quantities of air, the gas is not flammable. Only when the
gas escapes (or air can enter) is there any danger associated with it, because then the gas
can mix with air to form a combustible or explosive mixture. Such an escape is possible

112 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

 FIGURE 4-14 Typical
residential natural gas
pressure regulator.
Courtesy of John D. DeHaan.

because of (1) leakage due to inadequately sealed joints or corroded pipes, (2) mechani-
cal fracture of lines from external causes, and (3) failure of lines because of excessive heat,
which may melt essential sealing components of the line, valves, regulator, or gas meter.
Each of these requires separate consideration.

NATURAL GAS
Fuel gas in the form of natural gas or LP gas is delivered to appliances via a system of
tanks, regulators, pressurized delivery pipes, and flexible connectors. Natural gas in
the United States is piped underground via transmission pipelines that may be pressur-
ized to as high as 1,200 pounds per square inch (psi). Distribution pipelines (mains)
operate at typical pressures of about 60 psi (but may be up to 150 psi). A line regula-
tor (as shown in Figure 4-14) at each service location then drops the pressure to 0.14
to 0.36 psi, or 4 to 10 inches water column (w.c.) for delivery to residential appliances
(1 psi ⫽ 27.67 inches w.c.).36 Pressures in piping inside commercial buildings (for
industrial furnaces) may be up to 5 psi (but may be higher under special circum-
stances).37 The delivery piping may be steel, wrought iron (copper tubing if the gas is
low sulfur), brass, coated aluminum alloy tubing, or stainless steel. Cast iron pipe is
not to be used. Plastic pipe or tubing can be used only on outside underground instal-
lations. Natural gas has very little odor of its own, so an odorant mixture (usually, t-butyl odorant ■ Organic
mercaptan, thiophane, or another mercaptan) is added during pumping into the delivery chemical (often a mer-
captan) added to natu-
pipeline.
ral gas or LP gas prior
to its distribution to
LP GAS make it detectable by a
person with normal
Characteristics and Uses sense of smell at a con-
Although natural (or manufactured) gas is common in urban areas, many rural and centration less than
isolated regions make use of LP (liquefied petroleum) gases. These are normally either 20% of its LEL.

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 113

 propane, butane, or a mixture of the two with small quantities of other hydrocarbons.
Propane has a boiling point of -42°C (-44°F) at atmospheric pressure and is kept in liq-
uid form by pressure (its vapor pressure at the ambient temperature). Commercial
propane contains propylene and traces of other gases and ethyl mercaptan as an odor-
ant. Butane (pure n-butane) has a boiling point of -0.5°C (33°F) at atmospheric pres-
sure, so it can briefly exist as a free liquid at very low temperatures. It, too, contains
butylenes and other trace gases and an odorant such as ethyl mercaptan.38 LP gas is sup-
plied in fixed or portable containers of various sizes that are normally refilled by trans-
ferring liquid fuel under pressure from a delivery truck or facility. Most installations are
designed so that only gas is vented to the appliance rather than the liquefied fuel (the
most common exception being LPG tanks for forklifts and similar equipment, in which
the tank is designed to deliver liquid fuel to the motor). Only gas is vented by having the
delivery tube above the maximum fill level of the tank or cylinder, as shown in Figure
4-15. When the cylinder is filled to its normal maximum (80 percent of capacity), liquid
propane spills out of the spill tube via the bleed valve, and the operator knows the tank
is full. If the tank is improperly positioned or overfilled and then connected to an appliance,
the liquid can overwhelm the delivery system and spill out via the main delivery valve.
The pressure within the vapor space of the tank is controlled by the temperature of
the tank and its contents. This can range from 28 psi for propane at 0°F (-18°C) to 127
psi at 70°F (21°C) to 286 psi at 130°F (54°C). A pressure relief valve on the container is
usually set to bleed off gas when pressures exceed 250 psi.39 A one- or two-stage regula-
tor reduces the tank pressure to a working pressure of 11 to 14 inches w.c. (0.4–0.5 psi)
for introduction into a typical nonindustrial appliance.40 Operating pressures for undi-
luted LP gas systems in buildings shall not exceed 20 psi according to NFPA 54, National
Fuel Gas Code. Flexible rubber connectors (hoses) may be found on connections to
portable cylinders such as for campers and barbecue grills. The same comments with
regard to leaks and failures listed previously apply to LP gas delivery systems, with some
additional conditions. The portability of tanks exposes them to weather and mechanical
damage from handling (especially as they are refilled). The tanks are labeled to reflect
their capacities and the dates of their manufacture and pressure (hydrostatic) testing. The
necessity to disconnect and reconnect them repeatedly for filling makes them susceptible
to damaged connectors that can cause leaks. Flexible connections can crack with age or
misuse.
Because both propane and especially butane are heavier than air, their escape into the
air is somewhat different from that of methane, which is a major constituent of natural
gas. Being heavy, propane and butane vapors behave somewhat more like the fumes of

FIGURE 4-15 Cross

Vapor withdrawal
sections of typical tube
horizontal and vertical Pressure-relief To appliance
propane containers. valve Bleed valve Bleed valve
Courtesy of Fire & Arson
Investigator, April 1999.
80% fill line
80% fill
lines

Pressure-relief
valve

Cutaway diagram of typical horizontal propane

Cutaway diagram of typical vertical container for cooking or heating. (Tanks for vehicles
propane container are positioned to deliver liquid, not vapor.)

114 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

gasoline, which accumulate at low levels in enclosures, where they are less likely to be
smelled but more likely to encounter ignition sources. When propane (with a vapor den-
sity of 1.5) is involved, its density is sufficiently close to that of air to mix more readily
with air than will the heavier fuel gases. In its flammable range, a propane/air mixture is
almost the same density as air itself, so it may not “settle” at all. Modest mechanical stir-
ring in a room, or even the thermal stirring caused by sunlight heating one side of an enclo-
sure, has been shown to mix propane quite uniformly in an enclosed space.41 Release of
gas under pressure causes turbulent mixing.
It is of interest that LP gases apparently give rise to fires (and asphyxiations) somewhat
more frequently than does natural gas. The reasons are related to the nature of the fuels
themselves and also to the method of installation and uses of the related appliances and
related storage and distribution equipment. While natural gas is normally delivered from
heavy iron pipe, permanently installed and under the specifications of local building ordi-
nances, LP gases are more often delivered through semiflexible copper tubing installed in
an uncontrolled manner by amateur installers, and from removable tanks. In addition, the
supply of gas has to be replenished from tank trucks or with portable tanks rather than
from a permanent installation. Thus, there is considerably increased chance of poor con-
nections, broken tubing, leaking valves, and similar deficiencies. LP gas water heaters have
also been installed to replace costly electric heaters in small closets where ventilation is
inadequate when the door is closed. LP gas is delivered at considerably higher pressures
and through differently sized orifices than is natural gas. Because of its higher heat of com-
bustion and higher density, it also delivers 2.5 to 3 times as much energy per unit volume.
Equipment fitted for one type of gas can seriously malfunction if used with the other. If
malfunction of a gas appliance is suspected, the investigator must ensure that the correct
type of regulator, burner orifice, and control valve(s) are all in place to match the type of
gas used. See Chapter 15 for additional information on gas appliance failures.
Pressurized Containers
Pressurized containers of LP gas, propane, or butane are kept on hand in many types
of occupancies to fuel everything from tools like handheld propane torches to camping
stoves and barbecues (see Figure 4-15). These containers come in various sizes from
1-lb to 28-lb capacity (rated by weight of water equivalent) and are usually not hard to
recognize even when fragmented by a BLEVE, which is defined later in the chapter.
However, there is another source of LP gases that is not as readily identified. A variety
of aerosol products, from lubricants, paints, and insecticides to air fresheners, hair
sprays, and deodorants, are pressurized with dimethyl ether or mixtures of butane,
isobutane, and propane instead of the inert fluorocarbons used for so many years.42
These aerosol containers may hold as much as 120 ml of pressurized liquid fuel, which
vaporizes immediately upon release. Such small quantities do not present an ignition
risk when uniformly diluted throughout an average-size room. However, when confined
to a cabinet or an appliance, they can form an explosive mixture on release within that
space. Failures of single aerosol cans inside a vehicle will produce explosive concentra-
tions that can do considerable damage. Such products are sometimes used to excess
(in defiance of the manufacturer’s advice), and entire apartments and houses have
been damaged by the release and subsequent ignition of multiple cans, as shown in
Figure 4-16. The solvents used in such products (kerosene for insecticides, alcohols for
cosmetics) pose an additional risk of fire if released in the vicinity of an open flame.
Natural gas is also shipped and stored as a cryogenic liquid (CNG) with a boiling
point of –160°C (–260°F). When released, CNG causes the formation of a dense vapor
cloud denser than air, which poses a significant ignition risk as it spreads.43
Leaks
A gas line may show leakage from inadequate sealing of the joints during installation,
from deterioration of joints, or from corrosion severe enough to penetrate the walls of the
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 115
FIGURE 4-16 Six cans
of aerosol roach killer set
off in the cabinets of this
kitchen provided an
ignitable concentration
for the pilot light of the
stove (left foreground) to
ignite. The blast damaged
the ceiling of the kitchen
and adjacent living room
and lifted the roof. Note
the scorch mark on the
inside of the cabinet door
from the flame plume
from the can. Courtesy
of Captain Investigator Bruce
Fusselman, Phoenix Fire
Department, Phoenix, AZ.

line. Rigid metal lines can fracture if subjected to severe vibration, mechanical movement,
or loading. Leaks may also occur from worn or defective valves, regulators, and fittings.
In any of these events, it is expected that the odorant intentionally added to all natural
gas will serve as a warning of escaping gas, because natural and some manufactured gases
are inherently odorless. Leakage has also followed lightning strikes whose induced cur-
rent caused perforations of stainless steel and copper gas lines. Goodson recently reported
that corrugated stainless steel tubing (CSST) with a plastic covering intended for residen-
tial gas delivery systems has been shown to be perforated by the passage of induced
current from lightning strikes, causing perforations 0.04 to 0.15 in. across. These perfo-
rations allowed gas leaks into attics and other enclosed spaces where ignitable concentra-
tions could accumulate.44 Sanderson has described multiple penetrations of polyethylene
gas piping caused by discharge of static electricity built up on the interior of such piping
by the passage of particulates (dust and rust) through the pipe. These very small diame-
ter perforations would release gas at low rates that could accumulate in concealed spaces
(such as the steel pipe riser near the meter).45 Chemical corrosion can also cause multiple
leaks in any metallic tubing, as shown in Figure 4-17.
Assuming that such a leak exists and that it is not noted, the sequence of events is as
follows. Initially, the gas is too low in concentration to be ignited from any source. When
the gas reaches its lower combustible limit it will have formed an explosive mixture
throughout the area in which the gas concentration has reached this limit. This will

FIGURE 4-17A Corrosion of propane gas FIGURE 4-17B Internal close-up of chemi- FIGURE 4-17C Electric arc–induced perfo-
flex line caused sufficient leak to cause cal corrosion. Courtesy of Peter Brosz, Brosz & ration of CSST and ignition of gas. Courtesy
explosion. Courtesy of Peter Brosz, Brosz & Associates Forensic Services Inc.; Professor Helmut of Jamie Novak, Novak Investigations and St. Paul
Associates Forensic Services Inc.; Professor Helmut Brosz, Institute of Forensic Electro-Pathology. Fire Dept.
Brosz, Institute of Forensic Electro-Pathology.

116 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

normally occur only in a confined or partially confined space, but that space will be under
great danger of an explosion—not a fire, initially at least. If there is a source of ignition,
such as a pilot flame, or if a match is ignited by a smoker at this time, an explosion will
occur. It can be very forceful, which is characteristic of a stoichiometric mixture, and with-
out continuing flame. If the mixture is not ignited until it accumulates to form a rich mix-
ture, it will generate less mechanical force but will sustain a flame after the explosion. If
ignited right at the leak, where there may be some gas/air mixture within the combustible
limits locally, a continuing fire is expected. This fire can play, in a blowtorch effect, on
other fuels in a direct line and ignite them. The distance the flame will project depends on
the pressure of the gas being fed to the line. Typical residential delivery pressures of 0.5 psi
or less produce a horizontal jet of flame only 1 or 2 ft (0.3–0.6 m) in length. Industrial or
service pressures are much higher, and the flame-jet length will be proportionately longer.
It should be remembered that due to the odorants used in utility gas, most gas leaks are
detected by residents long before combustible mixtures are produced.

Mechanical Fracture
Release of gas from a mechanically fractured line is expected rarely. However, it is proba-
bly far more common than is generally believed and occurs even in the absence of natural
catastrophes such as earthquakes, landslides, and similar major causes of destruction. For
example, a gas line was laid under a street over which there was heavy trucking. Failure to
tamp the soil properly around the pipe allowed it to buckle under the recurrent loads until
the line parted, filling the soil with gas, which eventually percolated through porous soil to
accumulate inside a structure, which led to an explosion and fire. In another instance, a
small gas stove was attached to the gas line with a copper tube and with a soldered con-
nection. The stove was moved around somewhat, with recurrent flexure on the connection.
Finally, the joint parted and resulted in both a fire and multiple asphyxiations by the com-
bustion products. Misuse can also lead to leaks, as depicted in Figure 4-18. The corrugated
flexible connector pipes used on many gas appliances are subject to corrosion by ammo-
nia or sulfur-containing contaminants in the gases and can become dangerously leaky even
when not moved. Many such lines today are coated to minimize such corrosion.
Gas (natural or LP) released from a buried pipe has been seen to migrate consider-
able distances along the loose earth or void space surrounding some gas lines (or through
porous soils). This migration can result in gas accumulations inside buildings some

FIGURE 4-18 Improvised

use of portable LP tanks
for household furnace
(note empty cylinders and
wrenches). Courtesy of
Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 117

FIGURE 4-19 An illegal
bypass connection around
the gas meter (in the cup-
board to the right) leaked
and allowed an explosion
and fire that destroyed
this apartment. Courtesy
of Fire Investigation Unit,
Metropolitan Police, London,
England.

distance from the actual leak. This process is enhanced if the overlying soil is sealed by pave-
ment, ice, or even frozen soil, which minimizes upward losses. Such percolation through soil
can also scrub the gas of its odorant, which means that a leak can exist without witnesses
reporting the usual gas odors. Both natural gas and LP gas can also lose their odorants by
chemical adsorption onto the walls of new pipe (steel or plastic) or new containers, or by
oxidation by the red iron (oxide) rust inside old, disused or newly installed steel pipe.46,47
Improperly plumbed gas appliances that are connected using rubber tubing or even
garden hose have been found responsible for leaking fuel that resulted in a fire or explo-
sion, as illustrated in Figure 4-19. Such connections leak most often at joints or connec-
tions to metal fittings. The low delivery pressures in most residential structures rarely lead
to failure by bursting, but such rubber products can degrade from age, sunlight, or expo-
sure to chemicals and can crack or split as they become brittle.
Another type of mechanical failure is failure of pressure regulators in gas delivery sys-
tems. A failure due to corrosion, mechanical damage, or freezing in any of these regula-
tors can result in massive overpressures being delivered to the appliances. These
overpressures, in turn, result in flames exiting the appliances as the burners and vents are
unable to accommodate the extra flames and exhaust volume.
As with leakage, release of gas under those circumstances can lead to fire when the gas
is ignited directly as it first emerges from the fractured line. If the leaking gas mixes with
enough air before it is ignited, there will be an explosion in the accumulated mixture. This
explosion is almost always followed by a continuous flame at the source of the fuel gas.
Failure from Heat
In general, failure of a gas line from excessive heat will be limited to a general fire with which
the gas is not causally related. Lines are frequently joined by low-melting alloys such as sol-
der; if such a joint is heated by exterior flames or even hot gases, the joint may fail. In very hot
structure fires, it is not uncommon to have brass or bronze fittings partially melt (see Figure
4-20). Because these fittings are regularly used in connection with gas lines, such failures will
be found after the fire. Severe fire heating can cause differential expansion of brass connectors
and steel/iron pipe fittings, sometimes resulting in “loose” connections after cooling.
Because there must already be a fire to melt the attachment in these situations, it is
clear that the escaping gas will merely feed additional fuel to the fire locally and will produce
118 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels
 FIGURE 4-20A New LP
tank fittings. Courtesy of
Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

FIGURE 4-20B Post-fire.

Partially melted valve and
fittings from fire exposure
may require X-ray exami-
nation to reveal whether
they are open or closed.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

a characteristic “blowtorch” type of burning in front of the opening. A melted connec-

tion is not to be considered as a cause of the fire but only as a result of and a contribut-
ing factor to the intensity of the fire. The gas cannot spread under these conditions to
augment the flames except ahead of the line opening. Under no conceivable circumstances
will gas escaping into a burning fire avoid burning locally to be consumed at a distance
from the opening. The large blowtorch type of flame will have the effect of causing a
burned-out area immediately in front of the opening or charring to adjacent surfaces, as
pictured in Figure 4-21. Its effect on the remainder of the fire will be negligible in virtu-
ally every circumstance. At worst, the blowtorch type of flame may alter the fire pattern
because of local intensity, which tends to deflect air currents to the hottest part of the fire.
Despite these rather obvious facts, it is quite common for claims to be made that
escaping gas severely augmented the total fire throughout a building. This would necessi-
tate the highly unlikely situation that the unburned gas passed through, around, or over
burning areas to reach distant portions of the fire. Very unusual release conditions might
provide mixtures that are too rich to ignite or gases that are moving too quickly to be
ignited by a fire in the vicinity of the leak (until they are sufficiently diluted or slowed),
but such movement would normally be expected to mix enough air and flames into the
gas stream by turbulence that ignition would readily occur immediately.
Another dangerous situation occurs when ignitable liquids are packaged in sealed
containers, such as solvents in metal cans or glass bottles. When these containers are
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 119
FIGURE 4-21 Plume
damage (clean burn) on
the back of the stove
from the same gas
plume. (The stove had
been moved away from
the wall prior to fire.)
Courtesy of Det. Richard
Edwards (retired), Los Angeles
County Sheriff’s Department,
Whittier, CA.

exposed to a fire, pressure built up by the expanding liquid prevents the liquid from evap-
orating. Because evaporation allows a liquid to maintain its temperature at or below its
boiling point, the temperatures inside the containers exceed the boiling point of the liquid
and may trigger pyrolysis of the liquid. The containers are then filled with a mixture of
BLEVE ■ Boiling-liquid, pressurized liquid, vapor, and (sometimes) gaseous pyrolysis products, which can add
expanding-vapor explo- greatly to the fuel load of the fire when the containers erupt. These eruptions can occur
sion. A mechanical with explosive force and suddenly and dramatically affect the course and intensity of the
explosion caused by the general fire. Failure of a fluid-filled container due to overpressure caused by extreme heat-
heating of a liquid in a
sealed vessel to a tem-
ing is often called a BLEVE (boiling- liquid, expanding-vapor explosion), even if the fluids
perature far above its and gases involved are not in themselves combustible. See Chapter 12 for further discus-
boiling point. sion of such mechanical explosions.

120 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

 CHAPTER REVIEW

Summary

Flammable gases and liquids are common fuels in acci- to ignition as well as to the fully developed fire. As we
dental structure fires and constitute a convenient source have seen, the fire environment is a complex one, and
of fuel for a very high percentage of incendiary fires. the intensity and duration of the heat produced may
Although such fuels can be generally classified and even have dramatic effects on the physical and chemical
compared on the basis of their physical and chemical properties of the fuels present. This is an example of the
properties, the investigator should be careful in using value of real-fire experience in observing flammable liq-
these properties blindly in assaying a fuel’s contribution uids and gases as they behave in actual fires.

Review Questions

1. What is saturation vapor pressure and how is it 7. Name five common petroleum-based fuels and
measured? describe each.
2. What is meant by the lower explosive limit? 8. What are three differences between natural gas
3. What happens at the flash point of a liquid fuel? and LP gas?
4. What is the difference between flash point and 9. At what pressures is natural gas delivered in
autoignition temperature? distribution lines? Regulator to residential line?
5. What is the equation used to determine the vapor To appliance burner?
density of a gas or vapor? 10. What is the difference between flammable,
6. Which compounds have the widest flammability combustible, and ignitable liquids?
ranges?

References

1. D. Drysdale, An Introduction to Fire Dynamics, 2nd ed. 8. ASTM D93-08: Test Methods for Flash Point by Pensky-
(Chichester, UK: Wiley, 1999). Martens Closed Cup Tester (West Conshohocken, PA:
2. J. D. DeHaan, “Exploding Gas Cans and Other Fire ASTM, 2008).
Myths,” CAC News (Third Quarter 2004): 20–26. 9. ASTM D1310-01: Test Method for Flash Point and
3. D. Gardiner, M. Barden, and G. Pucher, “An Experimen- Fire Points of Liquids by Tag Open Cup Apparatus
tal and Modeling Study of the Flammability of Fuel (West Conshohocken, PA: ASTM, 2007)
Tank Vapors from High Ethanol Fuels,” (Washington, 10. ASTM D92-05a: Standard Test Method for Flash and
DC: National Renewable Energy Laboratory, Fire Points by Cleveland Open Cup (West
Subcontract Report NREL/SR-540-44040, Golden, Conshohocken, PA: ASTM, 2005).
Colorado, U.S. Department of Energy, October 2008). 11. R. Friedman, Principles of Fire Protection Chemistry
4. D. Guidry, private communication, May 2005. and Physics, 3rd ed. (Quincy, MA: National Fire
5. C. S. Kennedy and J. F. Knapp, “Childhood Burn Protection Association, 1998), 44.
Injuries Related to Gasoline Can Home Storage,” 12. Friedman, Principles of Fire Protection Chemistry and
Pediatrics 99 (March 1997): e3. Physics, 114.
6. M. A. Dorsheim, “Fuel Ignites Differently in Aircraft, 13. ASTM E659-78: Standard Test Method for
Lab Environments,” Aviation Week and Space Autoignition Temperatures of Liquid Chemicals
Technology, July 7, 1997, 61–63. (Quincy, MA: ASTM, 2000).
7. ASTM D56-0: Standard Test Method for Flash Point 14. American Petroleum Institute, “Ignition Risk of
by Tag Closed Tester (West Conshohocken, PA: ASTM, Hydrocarbon Vapors by Hot Surfaces in the Open
2005). Air,” API Publication 2216 (January 1991).
Chapter 4 Combustion Properties of Liquid and Gaseous Fuels 121
15. R. J. Harris, The Investigation and Control of Gas in Fire Debris Analysis: Biodiesel Basics,” J. Forensic
Explosions (New York: E & FN Spon, 1983), 39. Sciences 52, no. 2 (March 2007), 371–79.
16. J. D. DeHaan, “The Reconstruction of Fires Involving 32. DeHaan, “The Reconstruction of Fires.”
Highly Flammable Hydrocarbon Liquids” (PhD disser- 33. J. D. DeHaan, “The Influence of Temperature, Pool
tation, Strathclyde University, Glasgow, Scotland, Size, and Substrate on the Evaporation Rates of
1995; Ann Arbor, MI: University Microfilms, 1996). Flammable Liquids,” paper presented at Interflam,
17. P. Kennedy and A. Armstrong, “Fraction Vaporization Edinburgh, June 1999.
of Ignitable Liquids: Flash Point and Ignitability Issues,” 34. NFPA, Fire Protection Handbook, 3– 44.
paper presented at Fire and Materials 2007, 10th 35. Drysdale, An Introduction to Fire Dynamics, 204.
International Conference and Exhibition, Interscience 36. NFPA 921: Guide for Fire and Explosion Investigations
Communications, San Francisco, January 29–31, 2007. (Quincy, MA: NFPA, 2008 and 2011), pt. 9.3.3,
18. DeHaan, The Reconstruction of Fires. p. 92.
19. Ibid. 37. NFPA 54: National Fuel Gas Code (Quincy MA:
20. R. L.Valentine and R. D. Moore, The Transient Mixing NFPA, 1999), 16.
of Propane in a Column of Stable Air to Produce a 38. R. Lapina, “Propane Fire & Explosion Investigation,”
Flammable but Undetected Mixture (New York: Fire and Arson Investigator (April 2005): 46– 49.
American Society of Mechanical Engineers, 1974). 39. B. V. Ettling, “Venting of Propane from Overfilled
21. L. Thatcher, RVFOG Results, personal communica- Portable Containers,” Fire and Arson Investigator
tion, March 1999. (April 1999): 54–56.
22. DeHaan, “The Reconstruction of Fires.” 40. NFPA 921, (2008 and 2011) pt. 9.4.3.2.
23. J. D. DeHaan et al., “Deflagrations Involving Heavier- 41. Thatcher, RVFOG Results.
than-Air Vapor/Air Mixtures,” Fire Safety Journal 36 42. J. D. DeHaan and W. Howard, “Combustion
(2001): 693–710. Explosions Involving Household Aerosol Products” in
24. NFPA 58: Standard for Storage and Handling of LP Proceedings Fourth International Symposium of the
Gases (Quincy, MA: NFPA, 2008). Detection and Identification of Explosives, Chantilly,
25. R. W. Henderson and G. R. Lightsey, “The Effective VA, December 1995. See also Fire and Arson
Flame Temperatures of Flammable Liquids,” Fire and Investigator (September 1997): 25–27.
Arson Investigator 35 (December 1984). 43. P. K. Raj, “Where in a LNG Vapor Cloud Is the Flam-
26. NFPA, Fire Protection Handbook, 19th ed. (Quincy, mable Concentration Relative to the Visible Cloud
MA: NFPA, 2008), 8–120. Boundary?” NFPA Journal (March–June 2006): 68–70.
27. Ibid. 44. M. Goodson and M. Hergenrether, “Investigating the
28. ASTM D3699-08: Specification for Kerosene (West Causal Link between Lightning Strikes, CSST and
Conshohocken, PA: ASTM, 2008). Fire,” Fire and Arson Investigator (October 2005):
29. Exxon Corp. Isopars. Houston, TX, 1990. 28–31.
30. D. M. Gialamas, “Is It Gasoline or Insecticide?” CAC 45. J. Sanderson, “Electrostatic Pinholing,” Fire Findings
News (Spring 1994). 13, no. 4 (Fall 2005): 1–3.
31. R. J. Kuk and M. V. Spagnola, “Extraction of 46. Gas Odorization Manual, American Gas Association,
Alternative Fuels from Fire Debris Samples,” (1996).
J. Forensic Sciences 53, no. 5 (September 2008): 47. NFPA 921: Guide for Fire and Explosion Investigations
1123–29; E. Stauffer and D. Byron, “Alternative Fuels (Quincy, MA: NFPA, 2011) Pt. 9.9.92, p. 104.

122 Chapter 4 Combustion Properties of Liquid and Gaseous Fuels

 CHAPTER
5
Combustion Properties
of Solid Fuels

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

KEY TERMS

cellulosic, p. 125 point of origin, p. 138 spontaneous ignition, p. 132

char, p. 128 pyrolysis, p. 125 synthetic, p. 143
flame resistant, p. 156 pyrophoric, p. 132 thermoplastic, p. 145
piloted ignition, p. 128 soot, p. 143 thermosetting (resin), p. 125

OBJECTIVES
After reading this chapter, you should be able to:
■ Describe different modes by which fuel vapors are generated from a solid fuel.
■ Define pyrolysis.
■ Demonstrate a clear understanding of the combustion properties of wood, paper,
plastics, paint, metals, and coal.
■ Differentiate the ignition and combustion processes of thermoplastics from those
of wood.
■ Describe the reasons why wood does not have a fixed ignition temperature.
■ Explain the importance of dust particle size and concentration in dust explosions.
■ Explain the role of flame color and smoke production during fire development.

123
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

I
n addition to the liquids and gases examined in Chapter 4, there is a large group of very
common fuels for which accurate and precise combustion data cannot generally be
tabulated—solid fuels. Many of the typical combustion properties may have little or no
meaning for this group. For this reason, it is not often possible to list such familiar properties
as flash point or explosive range for the most common fuels of all: wood, plastic, paper, and
the like. Despite this limitation, the properties of these fuels are of the greatest importance,
and to some extent they can be described, either by numerical data or in general terms.
The ignition and combustion of solid fuels are more complex than that of liquid or
gaseous fuels because solid ignition usually depends on pyrolyzing enough of the material
into a combustible fuel and having those products mix with the correct amount of air to be
ignited. The ignition of any solid fuel is a surface phenomenon, which means that the temper-
ature of the surface is the critical factor, not the bulk temperature. The ignition of the typical
solid depends on providing a high enough heat flux to bring about the necessary pyrolysis (as
introduced in Chapter 2). The criterion of “high enough” depends, as we have seen, on the
properties of the potential fuel, that is, its thermal conductivity, thermal capacity, and density,
as well as its rate of absorption of heat. Ignition has been defined as the initiation of a self-
sustaining combustion.1 Because combustion can be either flaming or smoldering, when we
examine ignition processes we have to be aware of the difference between those processes.
As Babrauskas has pointed out, some ignition temperature data are the surface temperatures
of fuels as they ignite into flames, and others are the temperature of an oven into which the
object is placed that results (after some time elapses) in smoldering combustion.2
Reasonably pure cellulose is rare in nature, occurring in substances like cotton or linen
fiber. Thus, an undyed cotton dress or linen shirt is fairly pure cellulose. A wooden beam is
composed largely of cellulose, but in a different structural arrangement with many addi-
tional complex organic substances.
The chief reason that any numerical values at all can be attributed to the solid fuels is
that when heated, most, if not all of them, undergo heat decomposition or pyrolysis with
the production of simpler molecular species that do have definite and known properties.
Even so, precise values are not often available, because in some instances the pyrolysis of
solid materials has been inadequately studied and because the pyrolysis of a single solid
material gives rise to a great many simpler products. The resulting complex mixture varies
with the thermal and environmental conditions and has physical and chemical properties
unlike those of any pure compound. The properties of solid fuels and pyrolysis, and the
mechanism of their combustion, will be examined in this chapter.
124 Chapter 5 Combustion Properties of Solid Fuels
Pyrolysis
As discussed earlier, the most significant part of all flaming fires is the flame itself, in
which the combustion reaction takes place solely between gases. This remains true even
though the fuel feeding the flame is a solid: wood, cloth, plastic, paper, or even coal. How
then does a solid fuel maintain a gaseous reaction in the flames that burn around it? There
are several routes by which this can occur, as shown in Figure 5-1. Some solid fuels can
evaporate directly to vapors, by a process called sublimation. Naphthalene, a flammable
solid formerly used as mothballs, and methenamine, used for ignition tests, sublime at
room temperatures. Other solid fuels, such as candle waxes, melt and then evaporate.
Thermoplastics melt, decompose into smaller molecular species, and then evaporate.
Another class, such as polyurethane, decomposes when heated to form volatile liquid
products, which evaporate. The final category is the very large class of products that
when heated, decompose to yield both volatile products and a charred matrix. This cate-
gory includes wood, paper, other cellulosic products, and most thermosetting resins. The cellulosic ■ Plant-
key to most of these mechanisms is the phenomenon of pyrolysis, which occurs within the based materials based
solid fuel as a result of being strongly heated. on a natural polymer of
sugars.
In chemistry, pyrolysis is defined as the decomposition of a material brought on by
heat in the absence of reactions with oxygen. In fires, pyrolysis of fuel also includes ther- thermosetting (resin) ■
mal decomposition that takes place in the presence of oxygen (since oxygen is nearly Polymer that decom-
always present in the combustion zone). Oxidation by-products are therefore common in poses or degrades as it
is heated rather than
real-world combustion products. The phenomenon has been known for a very long time, melts.
but its mechanism is only now beginning to be understood, and then to only a limited
degree. To understand what occurs during the pyrolysis of fuel, it is necessary to remem- pyrolysis ■ The chemi-
ber that all practical fuels are organic in nature; that is, they are complex compounds cal decomposition of
substances through the
based on carbon. action of heat, in the
Nearly all fuels of importance to the fire investigator are vegetable in origin or are absence of oxygen.
derived by decomposition, bacterial action, or geologic processes from “living” material,
both animal and vegetable. This is true of oil and coal as well as of natural gas. Wood, the
most common fuel in ordinary fires, is the direct result of life processes in which very com-
plex organic structures are synthesized by natural processes within living cells. To date, no
one can write a complete structural formula for wood, although it is well known that its
major constituent is cellulose, a very large molecule synthesized from many glucose (sugar)
molecules in long chains of undetermined length (see page 27). In addition to cellulose,
there are many other compounds in wood: hemicellulose and lignin are the most prevalent
(up to about one quarter each), and there are various resins, pitches, oils, and other sub-
stances in various quantities. Certain softwoods, such as pine, have large quantities of
volatile oils called terpenes and oleoresins (the source of commercial rosin), while most
hardwoods contain little or no resin and only low concentrations of some of the terpenes.
Organic compounds (including the constituents of wood) when heated are subject to
complex degradations to simpler compounds that are more volatile and therefore more

Sublimation

Melting Evaporation
FIGURE 5-1 Different
modes by which fuel
LIQUID

VAPOR
SOLID

Melting Decomposition + vapors are generated

from a solid fuel. Source:
evaporation
Introduction to Fire Dynamics,
Decomposition + Decomposition + 2nd ed., D. Drysdale. 1999.
melting evaporation New York: John Wiley & Sons
Limited. Reproduced with
Decomposition + evaporation
permission.

Chapter 5 Combustion Properties of Solid Fuels 125

 flammable than the original ones. It is these compounds that oxidize in the flame. If a
sample is pyrolyzed by heating it electrically in an inert atmosphere, the decomposition
products may be separated by gas chromatography, identified by mass spectrometry, and
studied by a variety of tests. Such studies have been important in recent years in elucidat-
ing the mechanism of heat decomposition. Pyrolysis products from wood include water
vapor, methane, methanol, and acrolein, for instance. For the purposes of fire investiga-
tion, it is not important to understand the exact mechanism of pyrolysis or all the inter-
mediate products that result. However, it is important to understand the basic facts of
heat decomposition because they are at the heart of understanding the basic nature of fire
fed by solid fuel.
In the sense that they do not evaporate or sublime directly into fuel vapors, wood,
paint, plastics, and coal do not burn. When they are heated, they decompose into smaller
molecules with greater volatility and flammability, and, of course, into carbon, which is
combustible directly. This is the process of pyrolysis, and it is fundamental to nearly all
fires. It explains how fuel vapors generated by the solid are ignited to support the flames
visible above the solid fuel surface. There is a zone just below the surface where pyroly-
sis occurs, and beneath (or behind) there is a region where no pyrolysis has yet occurred.
The gaseous materials formed by the pyrolysis of the large, nonvolatile molecules of the
solid fuel are the materials that burn in the flame. Without them there could be no flame.
For instance, recent work by Stauffer has demonstrated that polyethylene pyrolyzes into
a homologous series of alkanes, alkenes, and dienes; polystyrene forms aromatic products
while burning; and polyvinyl chloride first releases noncombustible hydrogen chloride
(HCl) gas and then subsequently yields burnable organic fragments.3 Recent work by
DeHaan has demonstrated that animal fats pyrolyze into homologous series of alkenes,
aldehydes, and alkanes in ratios different from those produced by polyethylene.4

CROWN FIRES AND FIREBALLS

The pyrolysis process explains why fires may become “crown” fires in the burning of a
forest. When foliage is heated to the point that it exudes a large quantity of volatile gases
and vapors that are flammable, the vapor cloud can burst explosively into flames, and the
entire bush or tree becomes almost instantaneously enveloped in fire. This is the phenom-
enon that gives rise to the comment that a tree or bush “exploded” into fire. The mate-
rial composing it simply was heated to the point that it decomposed by pyrolysis to form
a large quantity of volatile, flammable gases that when mixed with air, underwent rapid
combustion, or deflagrated. Pyrolysis also explains the mechanism by which fires are set
at a distance by radiant heat from a large fire nearby. The radiation is absorbed until the
temperature produced starts pyrolytic action, and temperatures rise until the autoignition
temperature of the pyrolysis products is reached. At this time, fire erupts all over the fuel
exposed to the radiated heat.
An additional curious fire phenomenon that is explainable only on the basis of
pyrolysis products is the fireball, or “firewhirl.” It may be generated in a very intense
fire and can travel through the air for appreciable distances while burning. The “ball”
is formed by the emission of quantities of pyrolysis products far in excess of the air
locally available to burn them. This means they accumulate in a limited region, burn-
ing primarily on the periphery. The strong updraft created by the fire carries the pyrol-
ysis products upward into the air, where they continue to burn independently of the
original fuel from which they were formed. In wildland fires, these products can drift
out of the direct updraft and carry a mass of flaming gases into areas not yet on fire. It
is possible for such fireballs to entrap personnel fighting the fire. Fatalities from fire-
balls are known. Anyone who has viewed a very intense fire will have observed this
phenomenon, often in miniature, as masses of flame that detach from the fire and rise
into the air, as shown in Figure 5-2.
126 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-2 Fire plumes
above very energetic fires
can become detached
and move away from
their source as firewhirls
or fireballs. Courtesy of
Calvin Bonenberger, Lafayette
Hill, PA.

NONPYROLYZING FUELS
Not all solid fuels have to undergo pyrolysis before they can combust. Reactive metals
like sodium, potassium, phosphorus, and magnesium combust in air when oxygen com-
bines directly on the exposed surface. The resulting heat vaporizes the fuel and produces
hot gases and incandescent oxides (ash), but the fuel does not first pyrolyze into simpler
compounds. A few “solid” fuels like asphalt and wax melt and vaporize when heat is
applied, and the vapors support the flames. Carbon (as charcoal) can undergo combus-
tion without flames (smoldering combustion) as a solid–gas interaction on its myriad sur-
faces as oxygen diffuses into it.

Combustion Properties of Wood

Far more wood is burned as fuel in structural and outdoor fires than any other solid
material. Thus, its properties as a fuel, in regard to its behavior during combustion, are
generally of greater importance to the fire investigator than those of any other solid com-
bustible material.

COMPONENTS OF WOOD
The term wood is generic and covers a wide variety of materials, natural and man-made,
whose chief component is vegetable in origin. The major constituent of wood is cellulose
(苲50 percent), while numerous other constituents are present: hemicellulose (苲25 percent)
and lignin (苲25 percent), with resins, salts, and water making up variable percentages.5
The chemical makeup of cellulose was discussed in Chapter 2. Wood products, which
are materials composed chiefly of cellulose, include all kinds of manufactured boards and
panels, formed wood items such as furniture, and a host of paper and cardboard prod-
ucts manufactured, for the most part, from wood pulp.
Wood is obtained from many varieties of trees, some resinous, some not; some dense,
others light. Woods vary greatly in their water content, volatile components, and other
Chapter 5 Combustion Properties of Solid Fuels 127
 chemical properties. In addition to having diverse origins, wooden materials are also
greatly altered by manufacturing processes that result in a variety of prepared sheet,
board, and formed materials with all types of treatment and finishes.

IGNITION AND COMBUSTION OF WOOD

Wood, being a chemically and physically complex cellulosic fuel, introduces a number of
variables. Each of the major constituents has different temperatures at which it undergoes
pyrolysis: hemicellulose at 200°C to 260°C (400°F to 500°F), cellulose at 240°C to 350°C
(460°F to 660°F), and lignin at 280°C to 500°C (536°F to 930°F).6 The thermal conduc-
tivity of wood varies with orientation, as does its permeability to air, both being signifi-
cantly higher in the direction of the grain than across it. This will affect the ignitability of
a wood mass or surface, for some portions will be more easily ignited than others. Heat
must penetrate the wood to trigger pyrolysis and charring, as illustrated in Figure 5-3.
The production of volatile oils and resins is also faster along the grain than at right angles
to it, adding further variation.
Wood discolors and chars relatively quickly at temperatures above 200°C to 250°C
(400°F to 480°F), as shown in Figure 5-4, but prolonged heating at temperatures above
107°C (225°F) will have the same effect.7 (There are limited data on long-term [days
or weeks] exposure to lower temperatures, but it appears that destructive pyrolysis can
occur very slowly at temperatures down to 85°C [175°F].) As wood chars, its absorp-
tivity of incident heat flux becomes higher due to the darker surface and the lower ther-
char ■ Carbonaceous mal inertia (k␳c) of the char, so its temperature begins to increase faster once charring
remains of burned begins. All this means that wood does not have a fixed ignition temperature; it varies
organic materials. with the rate and manner in which heat is applied to it. It is clear that on heating,
a wood surface has one temperature at which it generates enough volatiles (which
are evidenced by the visible smoke) that can be ignited by the application of a pilot
piloted ignition ■ (external) flame (piloted ignition), and another much higher temperature at which
Ignition aided by the the vapors themselves will ignite without any external pilot flame (non-piloted or spon-
presence of a separate
external ignition source
taneous ignition). Along with all the other variables, the nature of the heating (whether
such as a flame or elec- radiant or convective) will influence both piloted and spontaneous ignition tempera-
tric arc. tures. Other variables include moisture content and size thickness of the sample,

Char base
Pyrolysis
Char layer zone Pyrolysis
zone base

Normal wood

FIGURE 5-3 Normal

combustion of wood
resulting in progressive
formation of char and
pyrolysis zones. Courtesy
of the USDA Forest Service,
Forest Products Laboratory, Char
Madison, WI. depth

128 Chapter 5 Combustion Properties of Solid Fuels

 FIGURE 5-4 When
Douglas fir and similar
softwoods are exposed
to a fire temperature of
1⬙ 700°C to 900°C (1500°F
350 ºF to 1900°F), the pyrolysis
and accompanying char-
220 ºF ring occur at tempera-
½⬙ Inner char tures of approximately
zone 550 ºF 177°C to 288°C (350°F
ASTM E119 to 550°F), as shown. The
depth of the pyrolysis is
Fire temp. dependent on the mois-
ture content, density of
Char rate app. 1500 ºF to 1900 ºF the wood, and intensity
¼0⬙ in./min. of the approaching fire.
Note that the char rate is
only approximated even
under these idealized lab-
oratory test conditions.
Courtesy of the USDA Forest
Service, Forest Products
Laboratory, Madison, WI.

orientation (vertical or horizontal), oxygen concentration, duration of heat exposure,

and especially whether piloted or nonpiloted (autoignition) is being observed.
Due to its complexity, the ignition temperature of “wood” is not a single, clearly
defined value for all woods and thermal conditions, as evidenced by the scattering of tem-
peratures reported here. It is generally agreed that even with fresh (undecomposed)
woods, the piloted ignition temperature will vary with the species of wood, the size and
form of the sample itself, the ventilation, and the intensity, manner, and period of heat-
ing. The autoignition temperature is similarly variable. It might be thought to be depend-
ent on the ignition temperatures of the pyrolysis products formed in the wood, but the
major species—methane, methanol, and carbon monoxide—all have autoignition temper-
atures on the order of 450°C to 600°C (850°F to 1,100°F), which are much higher than
some observed autoignition temperatures for wood. Clearly, other factors not yet identi-
fied are involved, perhaps the glowing combustion of surface char. The auto- or self-ignition
temperature is closely related to the temperature at which the oxidation of the wood
becomes sufficiently exothermic to raise the temperature of the surrounding fuel and
thereby become self-sustaining. Various authors have reported autoignition temperatures
to be on the order of 230°C to 260°C (440°F to 500°F). Browning reported that ignition
temperatures could be as low as 228°C (442°F) for some woods, and temperatures rang-
ing from 192°C (378°F) to 393°C (740°F) have been reported elsewhere in the literature,
but it is not clear whether they were observing flaming ignition or self-sustaining smol-
dering combustion.8 The reported variabilities reflect the effects of different test methods,
availability of oxygen, and even different masses and geometries of the sample.
Browning’s tests were of small wood chips embedded with thermocouples in a heated
chamber, and his criterion for ignition was the oven temperature at which the wood chip
temperature exceeded that of the environment. Many of these earlier studies did not dis-
criminate between smoldering and flaming combustion, so it is not surprising that
reported results vary so much. Because ignition of wood is a surface phenomenon, the
measurement of surface temperature at the location of ignition is the desired result.
Clearly, the complexity of the ignition of wood makes measurement of relevant tempera-
tures very difficult and subject to measurement error. Babrauskas has assembled the
results of numerous studies and concluded that values around 250°C to 260°C (480°F)
Chapter 5 Combustion Properties of Solid Fuels 129
 represent a defensible surface temperature for both piloted or nonpiloted ignition of fresh
whole wood in a reasonably short period of time9 [ignition near a minimum flux (immer-
sion in oven) being smoldering combustion that may transition to flaming]. Piloted
autoignition (surface) temperatures measured in tests where loose specimens were in open
air being heated by a radiant heat source were typically 300°C to 400°C (570°F to
750°F). For smoldering ignition, surface temperatures under 300°C are typical. Higher
temperatures favor gasification and thereby flaming combustion.10
Ignition Variables
Dry wood is more readily ignited than wood with a higher moisture content (for the rea-
sons explored in Chapter 3). Continuous or even recurrent exposure to raised tempera-
tures will dry wood, with the result that the ignition temperatures may be somewhat
lowered. However, the work of McNaughton does not indicate a significantly increased
hazard as the wood dries until the temperatures reach 275°C to 280°C (527°F to 536°F),
by which point both paint and wood are thoroughly charred.11
One recent study of piloted ignition of Masson pine by radiant flux (in horizontal
specimens 10 cm ⫻ 10 cm) revealed surface ignition temperatures of 301°C to 405°C
(574°F to 761°F).12 The higher the moisture content of the sample (0–30 percent MC),
the higher the ignition temperature (301°C at 0 percent MC to 368°C at 30 percent MC,
for example). Janssens reported that the ignition temperature is increased by 2°C for
every 1 percent moisture increase.13 Hemicellulose has the lowest ignition temperature
and lignin the highest. Softwoods, with less hemicellulose and more lignin, have slightly
higher autoignition temperatures (349°C–364°C; 660°F–687°F) than do hardwoods
(300°C–311°C; 572°F–592°F) under the same test protocol.14
Some woods, like pine, will ignite into flames more readily and burn vigorously,
because they contain resinous materials that easily produce volatile flammable vapors
when heated. These add greatly to the limited ability of cellulose and lignin alone to sup-
port the fire. Ease of ignition of wood is correlated with the content of pitch and other
components that readily decompose to generate combustible vapors. Some woods, like
pitch pine and slash pine, contain so many combustible components that they may easily
be ignited with a match flame and will burn furiously, as if soaked in kerosene.
Other than volatile resins found in many species of wood, the main fuel component is
cellulose. Because the reader is probably familiar with the low heat output from such pure
cellulose fuels as cotton or linen fabrics, it is readily appreciated that the cellulose (with its
¢Hc ⫽ 16 kJ/g) itself is not an efficient part of wood combustion. As we have seen, cellu-
lose, as a derivative of glucose, is already partly oxidized, and therefore it has proportion-
ately less fuel (per unit weight) to provide for oxidation in fire. Ease of ignition makes
wood (¢Hc 苲 16 kJ/g) or paper suitable for kindling a fire, but the total heat output is less
on a weight-for-weight basis than that of fuels with a higher percentage of hydrogen such
as coal or petroleum (¢Hc ⫽ 46 kJ/g). Thus, fires that have only wood for fuel may be
somewhat less intense than those that are fed by hydrocarbon liquids, which are often used
as accelerants (because they can release more energy per second per unit of surface area,
sometimes called energy density). Interestingly, charcoal has a ¢Hc of 34 kJ/g.
Decomposed Wood
The ignition temperatures of fresh, undecomposed wood as measured in the laboratory
may be of less importance to the fire investigator, since fresh, new wood is not always
involved in a structure fire. Of more interest is the effect of organic decomposition (decay)
and thermal decomposition (pyrolysis) on the ignitability of wood. Angell reported that
the ignition temperature of southern pine, measured at 205°C (400°F), when sound,
dropped to 150°C (300°F) when the same wood decayed.15 (These values seem far too
low for autoignition and may represent values for piloted ignition or sustained smolder-
ing, possibly as a result of self-heating.) The minimum radiant heat flux for ignition of
wood has been measured at 苲12 kW/m2 for piloted ignition and 20 to 40 kW/m2 for
130 Chapter 5 Combustion Properties of Solid Fuels
autoignition (depending on time of exposure and test protocol). It should be noted that
these samples were small (0.5 in. ⫻ 2 in.) and exposed to convective, short-term heating.
Ignition times were recorded as 427 seconds for the sound wood and 105 seconds for the
rotted wood (which indicate that the samples were exposed to very high heat fluxes).16
The relationship between applied heat flux and piloted ignition has been studied by
Janssens. As we saw in a previous chapter, the higher the heat flux and the lower the ther-
mal inertia, the faster ignition will take place. Because different woods have different den-
sities and thermal conductivities, the ignition time will depend on the variety of wood
involved. Moisture content will affect both, but since most interior wood is in the mois-
ture content range of 9 to 12 percent, its effect can be ignored.
Regarding time to ignition, Babrauskas’s data offered the empirical correlation
#
tig ⫽ 130␳0.73>(qex – ⫺ 11.0) 1.82
#
where ␳ ⫽ density of the wood (kg/m3), and qex – is the radiant heat flux to which the
(thick) wood is exposed (over the range 20–40 kW/m2).17 The lower the density or the
higher the radiant heat flux, the shorter the ignition time. There is a relationship between
environmental (immersion) temperature and ignition time for wood. Very old data pub-
lished by NFPA show that longleaf pine, for instance, does not produce piloted ignition
at 157°C (315°F) even after exposure for 40 minutes, yet piloted ignition can occur after
14.3 minutes at 180°C (360°F), 11.8 minutes at 200°C (390°F), 6 minutes at 250°C
(480°F), 2.3 minutes at 300°C (570°F), or 0.5 minute at 400°C (750°F).18 (Very low
reported ignition temperatures of 180°C–200°C, however, should be interpreted as com-
ing from faulty measurements when only short-term heating has been performed.)

“LOW TEMPERATURE” IGNITION OF WOOD

Of more frequent interest to fire investigators are the effects of slow, prolonged heating of
wood, which dehydrates and then decomposes the wood by pyrolytic action. The normal
combustion of wood is a progressive process with zones of char and pyrolysis that progress
inward from surfaces exposed to heat, as illustrated in Figure 5-4. At temperatures between
100° (212°F) and about 280°C (540°F) wood loses weight slowly as moisture and volatile
oils are released gradually, and a large percentage (苲40 percent) of the wood is turned to char-
coal.19 At temperatures above about 180°C (360°F), the pyrolysis of all three major solid
constituents (i.e., cellulose, hemicellulose, and lignin) reaches its maximum rate, leaving a
smaller percentage (10 to 20 percent by weight) as char. If the heat being accumulated by the
char is retained, and there is an adequate supply of oxygen, the temperature of the mass can
rise to the point at which combustion can take place (ignition temperatures of about 300°C;
570°F). The retention of heat, of course, depends on the amount of thermal insulation avail-
able and the amount of heat that is being lost to convective and conductive processes. If there
is too much insulation, the supply of oxygen becomes inadequate to sustain combustion,
although smoldering combustion can be sustained even at very low oxygen levels. It is for
these reasons that the formation of charcoal by slow heating of wood can play a significant
role in the initiation of some accidental smoldering fires20 as depicted in Figure 5-5.
Failure of the insulation around metal chimneys or fireboxes can expose wooden
building components to such temperatures, while the failure of the integrity of metal or
masonry flues may allow flames or sparks to provide piloted ignition. Obviously, direct
flame exposure from an opening in a flue or chimney will lead to relatively rapid ignition
of exposed wood. There is also a relationship between incident heat flux, piloted ignition
temperature, and time to ignition. Yudong and Drysdale demonstrated that the higher the
heat flux (between 15 and 32 kW/m2), the lower the observed ignition temperature and
the shorter the time to ignition.21
Although whole wood itself may not ignite at “low” temperatures, it is well known
that prolonged exposure to heat at temperatures below 120°C (248°F) over an extended
Chapter 5 Combustion Properties of Solid Fuels 131
FIGURE 5-5 Charring
in ceiling from prolonged
contact with electric
radiant-heat ceiling panels.
Courtesy of Robert Toth, IRIS
Investigations.

period of time can cause wood to degrade to charcoal by the distillation and pyrolysis
process described.22 Shaffer calculated that exposure to temperatures as low as 150°C
(300°F) for long periods can decompose finely divided cellulosics to charcoal, which then
may ignite.23 This charcoal has been referred to as pyrophoric carbon or pyrophoric char-
coal, alluding to the properties of activated (laboratory) charcoal to oxidize with room air
pyrophoric ■ Capable even at modestly elevated temperatures. It has been argued that pyrophoric is a misnomer,
of igniting on exposure since the U.S. Department of Transportation defines pyrophoric materials as those that can
to atmospheric oxygen
ignite spontaneously within 5 minutes of exposure to air. (The Chemical Manufacturers
at normal temperatures.
Association defines pyrophoric as a material that will ignite spontaneously in dry or moist
air at a temperature of 54.4°C (130°F) or below.)24 Bowes, in his comprehensive study of
self-heating mechanisms, points out that activated charcoal is made by heating a carbona-
ceous fuel (often, coconut hulls for laboratory use) under reducing conditions (in the
absence of all oxygen). Although this material is capable of self-heating when exposed to
air, especially when warm, exposure to air over a long period of time simply permits oxi-
dation of the carbon.25 If a wooden surface is exposed to a radiant heat source when air
is freely available, the pyrolysis products will distill away, but the activated form of char-
coal will not be formed in bulk. The resulting char may not be susceptible to self-heating,
and flames could not be supported because the required volatiles have already dissipated.
Bowes’s examination of the problem showed, however, that if a heated cylinder such as a
flue or pipe runs through a massive wood member with minimal clearance, the conditions
will be appropriate for the formation of a charcoal that can self-heat to smoldering com-
spontaneous ignition ■ bustion if the temperatures are high enough spontaneous ignition. Bowes offered the
A chemical or biological opinion that the temperatures produced by ordinary steam pipes would not be adequate to
process that generates create this condition but that superheated steam pipes or flues might. He commented, how-
sufficient heat to ignite
the reacting materials.
ever, that if a suitably hot source (i.e., one with adequately high radiant heat output) was
See self-heating. close to a wood surface overlaid with a noncombustible layer of sheet metal, tile, or simi-
lar barrier to oxygen, a reducing atmosphere could be produced, as shown in Figure 5-6.
The failure of the barrier or some other change that would suddenly bring the heated char-
coal into contact with fresh air could then result in flaming combustion. Bowes presented
the thermodynamic argument that to accomplish this, the temperature of the source would
have to be considerably hotter than steam pipes operating at normal pressures.
132 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-6 Fire in a
wood floor under the
hearth of an improperly
installed fireplace. Courtesy
of Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

Heaters and Flues

One study of the problem of ignition of wooden structures by heaters and flues has shed
further light on this problem. Martin and Margot evaluated a series of fires in houses
where a change had been made to high-efficiency wood stoves, sometimes 5 and 10 years
prior to the fire.26 They discovered that the affected wood surfaces had been faced with
metal or tile, often with an intervening layer of solid insulation. With prolonged heating,
even at temperatures below 120°C (250°F), the insulation would allow enough heat to be
conducted into the wood to trigger degradation while at the same time excluding air from
the surface. The lower the temperature was, of course, the longer the time required. If
after a suitable mass of charcoal had been created and maintained at a high temperature
by the surrounding mass of wood oxygen was admitted (by way of structural collapse,
wood shrinkage, or movement of the responsible heater), the mass of charcoal could sus-
tain self-heating combustion to the point of igniting flaming fire in adjoining fuels. Martin
and Margot’s analysis of the thermodynamics showed that temperatures had to be main-
tained for a very long period of time, that air circulation had to be minimal (to minimize
heat losses as well as to exclude fresh air), and that the involved wooden members had to
be fairly massive for this process to take place.
Degradation to Char
Recent work by Cuzzillo and Pagni has explored a number of factors associated with the
heating (or “cooking” at temperatures well below normal ignition temperatures) of wood.27
Their findings show that the degradation of wood to char over a long period of time
decreases the thermal conductivity and dramatically increases the porosity of wood. Cracking
of the char makes it even more accessible to atmospheric oxygen. (Fresh wood is much less
permeable across its grain than it is along its grain, and it has higher thermal conductivity
along its grain than across it as well.) When wood is heated at low temperatures, charcoal is
formed (even in the presence of some air). If enough is produced to be a critical mass for self-
heating, and the char is sufficiently porous to permit diffusion of oxygen through the mass,
the heated material can be ignited as a self-sustaining, smoldering combustion. If conditions
promote a sufficient increase in heat release rate, there can be a transition to flaming com-
bustion in adjoining masses of wood that have not been degraded completely.
Chapter 5 Combustion Properties of Solid Fuels 133
 Cuzzillo’s work showed that self-heating of the char formed by low-temperature “cook-
ing” is a likely mechanism leading to runaway self-heating when the conditions are right. It
has to be realized that runaway self-heating of such char does not necessarily lead to flam-
ing combustion, because so much of the volatile content has been cooked out of the char.
But a change in ventilation (such as shrinkage of wood timbers, failure of a metal or tile cov-
ering, or even a shift in the direction and intensity of ambient wind) could be the factor to
push the combustion process into flames. The processes that prompt transition from smol-
dering to flaming are not well understood. Flaming may be triggered by better ventilation
that causes a higher heat release rate (above a critical threshold), or the creation of hot spots
in the char that cause piloted ignition. The role of the processes involved in long-term, low-
temperature ignition of wood will be discussed in more detail in Chapter 6.
Areas in which oxygen-deprived “cooking” processes may take place include areas
surrounding light fixtures and fireplaces. Flues, vents, and chimneys (see Figure 5-7) nor-
mally have adequate clearances built into them because they are recognized sources of
heat, but untreated cellulose insulation (wood shavings, sawdust, or paper pulp) can inad-
vertently be brought into contact with such heat sources. Pipes carrying steam under pres-
sure can reach temperatures of more than 150°C (300°F) and should be well ventilated
with adequate clearances where they pass through wooden structural members. In a
recent review, it was concluded that any heat source that can create surface temperatures
on wood members in excess of 77°C (170°F) must be properly separated or insulated (as
UL recommended years ago). In fact, that is a requirement of the City of New York
Building Code.28

FIGURE 5-7 Displaced

metal chimney sections
allowed heat to ignite
wood framing of chimney
enclosure. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

134 Chapter 5 Combustion Properties of Solid Fuels

 FIGURE 5-8 Massive
floor timbers over 80 years
old ignited by hot-water
pipes. Residents smelled
smoke for 2 weeks prior
to detecting smoldering
fire. Courtesy of Thomas
Goodrow, Fire/Explosives
Technical Specialist, ATF
(retired).

The review of low-temperature ignition issues by Babrauskas revealed numerous

well-documented cases where cellulosic materials (including sawdust, wood boards, and
timber) were “ignited” by sources between 90°C and 200°C (196°F–390°F).29 Charring
of wood is known to take place at temperatures as low as 105° to 107°C (221°F–225°F),
and that wood charred at such low temperatures exhibits similar exothermic oxidation as
char produced at very high temperatures (⬎375°C; 700°F). It is also thought that mois-
ture and cyclic heating and cooling play roles in making charred wood susceptible to self-
heating. There are well-documented fire cases where massive wood timbers in a residence
were ignited by prolonged contact with steam heating pipes whose surface temperatures
never exceeded 120°C (250°F), as illustrated in Figure 5-8. As with other self-heating
materials (as will be seen in Chapter 6), if there is a sufficient reactive mass, adequate oxy-
gen, insulation, and time, such self-heating can reach runaway growth achieving smolder-
ing that, in many cases, can transition to open flame.
In the 2004 MagneTek court decision, several scientific errors crept in and were “val-
idated” by the published decision.30 These included a claim that pyrolysis is just a theory
that is not suitably supported by scientific data. Nothing could be further from the truth.
Pyrolysis as described in this text has been known for over 200 years to be part of the
combustion process, and thousands of papers support and prove its mechanisms.
Somehow the court became convinced that wood has a fixed ignition temperature of
400°F (195°C) and that any ignition has to be caused by a source with at least that tem-
perature. The court failed to note that wood is a very complex fuel that undergoes numer-
ous chemical and physical changes as it is heated. The various forms created by these
processes have different chemical and physical properties, some of which include an
increased propensity to self-heat. Even though the term pyrophoric is misapplied to wood
and charcoal (and that is presumably the term the court meant to discredit), the mecha-
nisms of self-heating are well known. No, there is no tested and published “formula” that
Chapter 5 Combustion Properties of Solid Fuels 135
 can accurately predict the relationship between heat (or temperature) exposure and time
to ignition for such a complex and multivariate problem as “low-temperature” ignition
of wood; however, that does not mean that scientifically defensible hypotheses cannot be
offered and tested in fire investigation. Some solar water heaters (which are deliberately
designed to capture the energy of the sun’s heat radiation) can produce temperatures
above 150°C (300°F) both in the heat-collecting element and in the pressurized transfer
medium (if any). When the circulating pumps for these heaters fail, charring of underly-
ing wood structures and roof fires have reportedly resulted.
Investigators tend to associate the time of discovery with the time of first ignition.
This assumption may introduce serious errors to the fire analysis. Due to its slow output
of heat and smoke, smoldering may proceed for an extended period without being noticed
by passersby. When the combustion transitions to flame, it is almost certain to be discov-
ered quickly.

Flame Temperatures
The actual flame temperatures produced by burning wood vary greatly, with such vari-
ables as oxygen content, forced-air draft, resin content, and degree of carbonization hav-
ing significant effects. Using a typical net heat of combustion for wood of 18 MJ/kg
(7755 Btu/lb) gives a calculated adiabatic (ideal condition) flame temperature of 1,590°C
(2,900°F). The actual flame temperature of wood has been measured and reported as
being some 500°C less (estimated to be 1,040°C, or 1,900°F).31 It should be noted that
the combustion of wood is so complex that a calculation of adiabatic flame temperature
is of little relevance. It merely serves to illustrate the contrast between such calculated
values and observed flame temperatures. It should also be remembered that measure-
ments of turbulent flame temperatures produce time averages of rapidly fluctuating
temperatures.

Char Rates
Once wood is ignited, the rate at which it will char is dependent on the heat flux to which
the wood is exposed, its material, and its physical form. The standard exposure for wall
and floor materials uses the E119 furnace, whose temperature and heat flux rise from near
ambient to a maximum after 60 to 90 minutes. Under these test conditions the char rate
for plywood ranged from 1.17 mm/min to 2.53 mm/min (0.05 to 0.1 in./min). For 1/2-in.
(12.7-mm) plywood, burn-through times were 10 to 12 minutes. For 3/4-in. plywood
(18–19 mm), burn-through times were 7.5 to 17.6 minutes. For boards (with tongue-and-
groove edges or no gaps), burn-through times were 10.5 minutes for 19.8-mm pine, 14.17
minutes for 20.6-mm oak, and 24.3 minutes for 38-mm boards.32 The presence of gaps
dramatically decreased burn-through times, reducing them by half that of tightly fitted
boards of the same thickness. Testing by the authors has shown that post-flashover burn-
ing of wood surfaces (such as plywood) can achieve 3 to 4 mm/min (0.12 to 0.16 in./min)
char rates, particularly where ventilation is at a maximum (in doorways for example).33
These rates are in accordance with results reported for “jet” fires against softwoods and
for higher-temperature fires than the E119 protocol, in which for exposures to mild fluxes
(苲20 kW/m2) the char rates are on the order of 0.6 to 1.1 mm/min (0.02 to 0.04 in./min).34
Butler tested the charring rate of one type of wood at a wide range of radiant heat
fluxes and found an arithmetic relationship, as shown in Figure 5-9. Note that at a radi-
ant heat flux of less than 10 kW/m2, the char rate was zero, while at 20 kW/m2 it was
approximately 0.6 mm/min (0.02 in./min). At flame contact heat fluxes of 50 kW/m2
the rate was 1.2 mm/min (0.05 in./min), and at 150 kW/m2 the rate was on the order of
4.2 mm/min (0.17 in./min).35 Babrauskas reported that other researchers reported simi-
lar correlation. With a correlation
# for mass loss rate, the best mathematical fit for surface
char rate was ␤ ⫽ 0.028 qtot ⴖ (where ␤, char rate, is in mm/min) for radiant flames below

200 kW/m2.
136 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-9 Variation of
1 charring rate of wood
with radiant heat flux.
Note the log/log scale
and sharp decrease below
20 kW/m2. From Butler,
C. P. “Notes on Charring Rates

Rate of charring (mm/s)

in Wood.” Fire Research Note
10–1 No. 896, 1971. Reproduced in
Drysdale, D. An Introduction
to Fire Dynamics. New York:
John Wiley & Sons, 1985,
p. 182.

1/40"/min
10–2
0.6 mm/min

10–3
10 100 1000
Radiant heat flux (kW/m2)

Doors often provide the only separation between compartments and provide delays
to fire spread, so an appreciation of fire penetration times is important to timeline recon-
structions. The critical feature is usually the tightness of fit, with loose or badly fitting
doors providing only 3 to 5 minutes of resistance to flame spread. If the door is known
to be a reasonably good fit, a solid-core door 44.5 mm (1-3/4 in.) thick can withstand 14
to 30 minutes of E119 exposure.36 Shoub found “panel” doors [panels 9.5 mm (3/8 in.)
thick] burned through in 5 to 6 minutes of E119 exposure (1.6 to 1.9 mm/min). Hollow-
core doors with 4.8-mm (3/16-in.) wood panels burned through in 9 to 10 minutes.37
Many hollow-core doors in low-cost housing or offices are coated fiberboard and have
been seen to fail within 3 to 5 minutes of fire exposure in room fire tests by these authors.
As noted previously, intensities (and exposure times) of real fires in modern rooms can be
more severe than the E119 test protocol.38
Role of Adhesive
Another factor that may contribute to the influence of manufactured wood products is
the combustion properties of the adhesive used. Often, the adhesive is less readily com-
bustible than the wood layers; if it is more combustible, it may contribute what is effec-
tively a fire accelerant. The resins used in oriented strand board (OSB) have the same
ignition and combustion properties as the bulk solid timber, so burn-through times are
reportedly similar.39 In the evaluation of any fire involving these materials, special atten-
tion to the adhesive involved is essential. Some plastic binders can melt and support a
flame like a large candle, leaving little residue after prolonged combustion.

CHARCOAL AND COKE

Although wood contains oxygen in its structure and could be considered partially oxi-
dized, one of its main combustion products, charcoal, contains no significant oxygen and
is therefore an entirely carbonaceous fuel. Charcoal is an excellent fuel, having a heat of
combustion on the order of 34 MJ/kg. Charcoal is, of course, formed by pyrolyzing and
destructively distilling the volatile materials out of the wood, leaving behind the non-
volatile constituents, chiefly carbon, as shown in Figures 5-3 and 5-4. Charcoal is corre-
spondingly difficult to ignite and, lacking volatile materials, gives little flame. The small
Chapter 5 Combustion Properties of Solid Fuels 137
 blue flames associated with its burning result from the formation of carbon monoxide
(CO) gas, which burns rapidly close to the surface of the charcoal from which it was gen-
erated, to form carbon dioxide (CO2). A charcoal fire is largely a smoldering fire, that is,
one without flame, but intensely hot. Here the solid is reacting with gaseous oxygen to
produce combustion on and slightly below the surface of the fuel.
Coke, formed similarly from coal or petroleum, displays burning characteristics much
like those of charcoal and for the same reasons. It is not a cellulosic fuel, because coke
(and coal) is almost entirely carbon and hydrocarbons.

WOOD PRODUCTS
Not all wood present in structures is in the form of structural lumber. There is an infinite
variety of manufactured wood products in every building. The susceptibility of these
products is not the same as for wood alone. These products deserve special mention
because they greatly modify the burning characteristics of the wood from which they are
derived, and they can add significantly to the fuel load and ease of ignition. As a general
rule of heat transfer, the smaller the dimensions, the thinner the fuel, and the more edges
presented to a heat source, the faster the fuel package will be ignited.
Plywood and Veneer Board
Plywood and veneer board are very common. Both are made from very thin sheets of
wood (peeled from suitable logs) laminated to one another with layers of adhesive. The
nature of the adhesive determines the suitability of the product for interior (dry condi-
tions only), exterior (some weather exposure), or marine (extended exposure to water)
use. Standard plywood is largely made from fir, with cheaper, imperfect sheets and pieces
in the interior layers. Veneer board generally is made with a top layer of hardwood lam-
inated to less expensive base layers.
The authors have investigated a number of fires in which plywood and veneer board
were heavily involved. In no instance was the ignition of the fire traced to such materials,
but the rapidity of buildup and spread of the fire was nearly always markedly increased.
(See fire tests on vertical wood surfaces in Chapter 3.) In one instance of a fire of electri-
cal origin, an extremely thin plywood finish at some distance from the origin had pro-
point of origin ■ The duced such severe local burning that the location was for a time suspected as the point of
specific location at origin. This illustrates the necessity of checking on all possible types of causation and the
which a fire was ignited. behavior of involved fuels before adopting any one of them. Failure to do this has led to
numerous errors on the part of some fire investigators. Plywood and veneer board curl as
they delaminate as they burn. (See Figure 3-30.) This behavior can contribute to rapid
downward flame spread and cause floor-level burns, sometimes on flooring some distance
from the wall. (See Chapter 7 for an example.)
Particleboard or Chipboard
Particleboard or chipboard is made from small chips, sawdust, and waste from wood
and paper mills bonded together under great pressure with a suitable adhesive.
Particleboard is very dense but cheaper to make than plywood and is widely used in
floors, cabinets, and furniture where appearance is not of importance. Particleboard
has very limited strength under wet conditions and tends to swell and crumble when
exposed to water for any length of time. Both plywood and particleboard are often fin-
ished with a thin vinyl covering that can be combustible. A product that is very widely
used for exterior sheathing and manufactured structural components is called oriented
strand board (OSB) or chipboard. These are large splinters, shards, and thin fragments
of wood that are oriented so that their grain is roughly parallel and then glued together
under great pressure and heat using formaldehyde resins. The resulting product is
stronger than particleboard and cheaper than plywood because it can be made from
waste from wood processing.
138 Chapter 5 Combustion Properties of Solid Fuels
Burning Behavior of Thin-Cut Woods
With regard to the burning characteristics of wood products, it must be remembered that
the finer or thinner wood is cut, the better it ignites, and the faster it burns, because the
thin fuel is exposed to heat from both sides, and there is no underlying mass into which
the excess heat can be conducted. As a result, the thin fuel reaches its surface ignition tem-
perature more quickly and ignites. It then burns faster because of the large surface area
exposed to heat and air. A laminated or glued-composite board will burn much like a
solid-wood board of the same dimensions except for the influence of the adhesive or coat-
ing, which can be significant. If the adhesive softens under heat (or if it has been exposed
to water or moist conditions), the layers will delaminate and open like the pages of a
book. This will make the thin layers much easier to ignite, expose a greatly increased sur-
face area to the flames, and cause comparatively fiercer burning. Thus, in this regard, the
adhesive may be more important than other considerations. Plywood from the most com-
mon sources is not much different in its combustion properties from other wood of the
same thickness. Some imported veneer boards are made with unsuitable adhesive and may
greatly accelerate a fire involving them. Particleboard is generally more difficult to ignite
because of its dense nature. Once ignited, it behaves in the same way as densely packed
sawdust and is prone to smoldering combustion that is difficult to extinguish. Prolonged
exposure to moisture causes many particleboard-type composite materials to disintegrate,
with corresponding changes in their ignition and combustion properties.
Exterior Deck Materials
Several new products intended for exterior decks are being seen more frequently today.
These consist of cellulose (wood flour, wood fiber, or ground paper) blended with a poly-
ethylene or polypropylene plastic binder. They can vary from 41 to 65 percent wood fiber,
with 31 to 50 percent plastic and ash (possibly added as a filler or opacifying agent).40
They may be solid in cross section, hollow, or channeled (as seen in Figure 5-10). These
products have been tested for ignitability by burning brand and for fire performance when
ignited by an underdeck burner, or in a cone calorimeter. These composite materials have
been found to exhibit increased heat release rate (i.e., a bigger fire) than the equivalent
solid wood material.41 They also exhibit a tendency to drip molten plastic as they burn
(thus increasing the chances of spreading a fire down onto combustibles below) and to fail
structurally rapidly and without warning (increasing risks of injury to firefighters).42
Other Cellulosic Building Materials
Other cellulosic building materials, not necessarily made from wood but alike in their
combustion properties, include Celotex, Masonite, and similar boards made from com-
pressed and bonded fibrous materials of a cellulosic nature. The low-density products
(such as Celotex) were formerly used as acoustical tiles to cover ceilings and today are
used, in large sheets, as insulation, floor underlay, or for other such purposes (see Figure
3-27). Tests show that some of these materials can support rapid flame spread across their
surfaces and are subject to a smoldering fire along their edges or, occasionally, in their
interiors. In a large surrounding fire, they burn well and contribute much fuel. Thus, it is
common to see that in a room fire, all the acoustical tiles have fallen and burned on the
floor. They fall readily because they are generally glued into place, and the heat softens
the adhesive. If ceiling insulation or ceiling tiles are made of low-density cellulose (paper
fibers), they can contribute significantly to the intensity of the fire, but only rarely are they
involved in starting it. Once on fire, they are difficult to extinguish, since they tend to con-
tinue smoldering. The high-density products (Masonite) are considerably more ignition
resistant.
Blown-in ceiling insulation can be shredded or macerated paper with a flame retar-
dant added. A similar product is also made for application by being blown in while wet
for use in walls, where the dried mass is more resistant to mechanical disturbance and fire
penetration. Blown-in insulation can also be mineral fiber, which is noncombustible. Both
Chapter 5 Combustion Properties of Solid Fuels 139
 types of products are typically gray in color, and careful examination (including labora-
tory analysis) is required to identify the product.43 Loose blown-in cellulose insulation is
a triple threat when it comes to accidental ignition for the following reasons:
■ Its good insulative properties minimize heat loss from even small heat sources
buried in it.
■ Its loose texture means it can be moved or can fall into contact with heat sources.
■ Its fire retardance can be compromised by moisture or mechanical separation, mak-
ing it more susceptible to smoldering.
■ Some brands of cellulose insulation have been formulated with fire retardants
which can be highly corrosive to metals, although these are not common.
Fresh, proper flame-retardant treatment will, however, make the bulk material very
resistant to both flaming and smoldering combustion. Samples of any suspected struc-
tural, decorative, or insulation materials should be retained for laboratory analysis.

Paper
Paper is one of the more interesting substances involved in fires, partly because of its fre-
quent use in starting fires and partly because of its unusual combustion properties.
Everyone who has ignited a fire with paper will realize that newspaper works well, while
a picture magazine makes very poor kindling. The reasons are simple but not always
understood.
The basic ingredient of all paper is cellulose, the same material as that in cotton and
a major constituent of wood, both of which are used in the manufacture of paper.
Although cellulose is readily combustible, not all the extraneous constituents of paper are.
“Slick” papers have a very high content of clay, which is not combustible. Many writing
papers also have a high clay content. The differences in the combustion properties of var-
ious papers are readily demonstrated by burning a piece of newspaper, which leaves a
small and very light ash, followed by burning a piece of heavy, smooth, slick magazine
paper, which leaves a very heavy ash. The ash in most instances is the clay, although many
specialty papers also contain such constituents as titania (titanium dioxide), barium sul-
fate, and other noncombustibles.
Because paper is a cellulosic product like wood, its ignition properties are not eas-
ily measured. Graf studied the ignition temperatures of a variety of papers.44 With few
exceptions, he found the ignition temperature to be between 218°C and 246°C (425°F
and 475°F). At 150°C (300°F) most papers appeared to be unchanged during the dura-
tion of his tests. Many papers tanned in the range up to 177°C (350°F), went from tan
to brown by about 204°C (400°F), and if not ignited, went from brown to black at
about 232°C (450°F). Such events are affected by the rate of heating, manner of igni-
tion, nature of heat exposure, and amount of ventilation; therefore, a precise ignition
temperature cannot be cited. Smith’s data (as reported in the Ignition Handbook) indi-
cate that most uncoated papers ignited at surface temperatures of 260°C to 290°C
(500°F to 550°F) as a result of a validated test.45 Other reported tests generally agreed.
Long-term ignition was reported to occur in paper wrapped around a pipe maintained
at 200°C (400°F).
Paper ignites easily because, like wood, it has a low thermal inertia [calculated from
k␳c, the product of the fuel’s thermal conductivity (k), its density (␳), and thermal capac-
ity (c)], so its surface temperature goes up rapidly, and it is thermally thin, allowing rapid
heat saturation of its entire thickness. The minimum ignition heat flux for various paper
products has variously been reported from 20 to 35 kW/m2.46 If paper is exposed as a
single sheet rather than compressed in a stack, its large surface area promotes rapid
burning, and a high, but brief, heat release rate. When paper is stacked, it no longer has
much exposed surface, and a pile of stacked papers is very difficult to burn, although it
140 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-10 Plastic
and cellulose composites
for building applications.
(a) Plank form for decks.
(b) Board form for fences.
Courtesy of John D. DeHaan.

(a)

(b)

will char on the exposed surface. Figure 5-11 illustrates the effect of thermal thickness.
Ventilation is absent on the inside of such a stack, thus making paper a good insulator
against penetration of heat from a surrounding fire. In fact, a small fire may be smoth-
ered by paper quite effectively, merely by covering the burning surface with a few layers
of it. The air supply is thereby shut off, and the fire is quenched. This is true even of
Chapter 5 Combustion Properties of Solid Fuels 141
FIGURE 5-11 Loose
paper exposed to radiant
heat from a developing
room fire demonstrates
the effect of “thermal
thickness”—scorched
where it curled away and
unaffected where in con-
tact with book beneath.
Courtesy of John D. DeHaan.

newspaper, which is among the most combustible types of paper in existence. A stack of
paper, however, can support smoldering fire within its mass. Single sheets of paper, either
flat or crumpled, have proven very difficult, if not impossible, to ignite by contact with
a glowing cigarette. Carbon paper, however, can be ignited with a cigarette and will sup-
port a smoldering fire. Once burning at a steady rate, cardboard cartons (flats) have
been measured to yield a maximum mass flux of 14 g/m2/s compared with that of wood
cribs, 11 g/m2/s.47
In the interpretation of a fire, it is not the presence of paper that is significant but
its distribution in terms of exposed surface. Sometimes, the presence of a stack of papers
is taken as the origin of a general fire. Although it is possible to burn a stack of folded
newspapers to completion over an extended time, it is surprisingly difficult to do because
of the limited surface area exposed to the air. In contrast, a pile of loosely crumpled
sheets of the same paper is not only easy to ignite but quickly raises the temperature of
nearby combustibles to their ignition points and thus initiates a fast-growing, very
destructive fire. In our experiments, rooms with a small general fuel load (bed, dresser,
carpet, and wooden doors) could be quickly brought to a fully involved stage of fire with
a large pile of crumpled newspaper ignited with a match. The papers burned nearly to
completion, leaving only small, unremarkable flakes of ash. Although fiberglass thermal
insulation is not combustible, it is often faced with a kraft paper layer secured with an
asphaltic binder. The thin paper, if exposed on walls or ceilings, can support extremely
fast fire spread.
At the same time, it should be remembered that a flammable liquid poured onto a pile
of papers is a common arson set. If the papers are closely packed, the accelerant burns at
the surface of the papers, with the absorbent paper acting like a wick. The residue of
unburned papers within such a pile is a good place to look for unburned traces of the liq-
uid accelerant. Here the primary purpose of the paper is to retain the accelerant, not to
contribute largely to the fire by its own combustion.
142 Chapter 5 Combustion Properties of Solid Fuels
Plastics
Plastics are nearly universal in modern structures and vehicles and are found burned in
nearly all fires. As a result, fire investigators need a working knowledge of the kinds of
plastics, their applications, and their burning behaviors.

GENERAL CHARACTERISTICS
There are many kinds of plastics, with a variety of physical and chemical properties.
Some, like Teflon (tetrafluoropolyethylene), are not readily combustible and will exhibit
heat damage only at extremely high temperatures. Others, such as nitrocellulose, are
readily ignitable and violently flammable. Most of the commonly used plastics range
between these two extremes. Called synthetics as a class because they are not natural synthetic ■ Material
products but are manufactured in a factory (usually from petroleum feedstocks), most that is human-made,
usually referring to
plastics are readily combustible. Some may burn to completion only in an intense sur-
organic polymers.
rounding fire, and others will readily burn almost to completion. Many laboratory tests
of the ignition and combustion properties of plastics are used to identify them and clas-
sify them as to fire hazard. It is important to remember that nearly all plastics are made
from long chains of hydrocarbons linked together in various ways. Thus, sufficient heat
input will rupture the chemical bonds holding the chains together, and the plastic will
pyrolyze to simpler, more volatile compounds. These pyrolysis products may be very
toxic, as in the case of styrene monomer and cyanogens; extremely flammable, as in the
case of carbon monoxide (CO) and the short-lived radicals CH and CH2; or readily sus-
ceptible to further pyrolysis. In addition, some plastics and rubbers create pyrolysis prod-
ucts that can have serious consequences in the fire environment. Polyvinyl chloride
plastic releases hydrochloric acid (HCl), and Teflon releases hydrogen fluoride (HF) and
free fluorine (F2) when it finally decomposes in a fire. Some rubbers generate hydrocyanic
acid (HCN) during combustion. For a full description of pyrolysis products from plastics
in fires, the reader is referred to other references.48 In addition, thermoplastic polymers
will melt and flow, sometimes dripping onto a source of flame to fuel it and sometimes
collecting in a large pool of burning plastic, which burns fiercely and resists extinguish-
ment. The behaviors of some of the more common plastics when exposed to a small
flame in air are listed in Table 5-1. As with laboratory tests with other solid fuels, it must
be remembered that the fire behavior of synthetics is largely dependent on their degree of
molecular cross-linking, physical form (sheet, rod, fiber, etc.), test conditions, and pres-
ence of inorganic fillers and fire retardants, so it is best not to generalize about the fire
behavior of a particular type of plastic.
Because the chemistry of plastics is similar to that of petroleum products, they tend
to produce similar sooty flames when burning in air, and the open-flame temperatures
produced by burning plastics are on the same order as those of common petroleum dis-
tillates. The flames of thermoplastics can be sustained by the formation of a pool of
molten plastic, producing very intense fires. The flame temperatures produced by some
plastics are quite high. Polyurethane is one example for which flame temperatures up to
1,300°C (2,400°F) have been measured over polyurethane mattresses during fire experi-
ments.49 The reported maximum mass burning fluxes for plastics are shown in Table 5-2
(for horizontal steady-state burning). Multiplied by the heat of combustion for each fuel,
these mass fluxes can be used to approximate the HRR of horizontal fuel surfaces of var-
ious sizes. Note that these heat release rates are nearly the same as for ignitable liquids
[for polypropylene, for example: 24 g/m2 · s 44 kJ/g ⫽ 1 MW/m2].
The combustion of many polymers contributes largely to the formation of the greasy
soot ■ The carbon-
or sticky dense soot found at many fire scenes and is responsible for the dense black
based solid residue
smoke more frequently noted during the early stages of structure fires. The dense smoke created by incomplete
many polymers produce is rich in flammable vapors, and the ignition of dense clouds of combustion of carbon-
these vapors quickly involves an entire room upon flashover. Only polyethylene, polypropylene, based fuels.

Chapter 5 Combustion Properties of Solid Fuels 143

 TABLE 5-1 Flaming Burning Characteristics of Small Samples of Common Polymers

POLYMER FLAME COLOR ODOR OTHER FEATURES

Nylon Blue with yellow tip Resembles burning Tends to self-extinguish;
grass clear melt
Polytetrafluoroethylene Yellow None Chars very slowly; difficult
to burn
Polyvinyl chloride Yellow with green Acrid Tends to self-extinguish;
base acidic fumes
Urea-formaldehyde Pale yellow Fishy formaldehyde Very difficult to ignite
Butyl rubber Yellow, smoky Sweet Burns
Nitrile rubber Yellow, smoky Sickly sweet Burns
Polyurethane Yellow, blue base Acrid Burns readily
Silicone rubber Bright yellow-white None Burns with white fumes;
white residue
Styrene-butadiene rubber Yellow, very smoky Styrene Burns
Alkyd plastic Yellow, smoky Pungent Burns readily
Polyester (contains styrene) Yellow, very smoky Styrene Burns
Polyethylene Yellow, blue base Resembles burning Burns readily; clear when
candle wax molten
Polymethyl methacrylate Yellow Characteristic Burns
Polypropylene Yellow, blue base Resembles burning Burns readily; clear when
candle wax molten
Polystyrene Yellow, blue base, Styrene Burns readily
very smoky

Source: K. J. Saunders, The Identification of Plastics and Rubbers (London: Chapman and Hall, 1966). See also Fire Protection Handbook,
19th ed. (Quincy, MA: NFPA, 2003), fig. 8.10.6.

and polyoxymethylene (Delrin) generally burn with a clear flame and little soot. Polyethylene
and polypropylene also produce a candle-wax smell, since upon heating, they pyrolyze into
much the same components as wax, with the resulting similar behavior.
Although the autoignition temperatures of plastics are generally on the order of
330°C to 600°C (650°F to 1,100°F), their universal applications in packaging, utensils,

TABLE 5-2 Maximum Mass Flux Rates (Steady-State Burning)

PLASTIC g/m2 · s
Polystyrene (granular) 38
Acrylic plastic (PMHA, granular) 28
Polyethylene (granular) 26
Polypropylene (granular) 24
Rigid polyurethane foam 22–25
Flexible polyurethane foam 21–27
Polyvinyl chloride (granular) 16

Source: From J. Quintiere, Principles of Fire Behavior (Albany, NY: Delmar, 1998), 109.

144 Chapter 5 Combustion Properties of Solid Fuels

 TABLE 5-3 Thermal Properties of Polymeric Materials

THERMAL BEHAVIOR MELTING POINT (°C)

NATURAL POLYMERS ¢Hc (DRYSDALE) (MERCK INDEX)a
Wool Chars above 200°C
Cellulose 16.1 Chars above 100°C
Thermoplastic Polymers
Polyethylene (HD) 46.5 Melts 130°C–135°C (85–110)a°C
Polypropylene (isotactic) 46.0 Melts 186°C (165)a°C
Polymethyl methacrylate 26.2 Melts 160°C (250)°C
Polystyrene 41.6 Melts 240°C (180)°C
Polyacrylonitrile Melts 317°C
Nylon 66 31.9 Melts 250°C–260°C
Nylon 6 — (225)°C
Polyvinyl chloride (PVC) 19.9 — (Chars 200–300)°C
Thermosetting Polymers
Polyesters (Dec. 250)°C
Polyurethane foam 24.4 Decomposes 200°C–300°C:
polyols, isocyanates
Phenolic foams 17.9 Chars
Polyisocyanurate foams 24.4 Chars
Silicones 26.8b Decomposes/oxidizes to SiO2b

Source: From D. Drysdale, An Introduction to Fire Dynamics, 2nd ed. (Chichester, UK: Wiley, 1999), 3.
a
The melting point of many polymers is determined by the amount of cross-linking between the chemical structures and can vary considerably.
b
Buch, RR, “Rates of Heat Release and Related Fire Parameters for Silicones”, Fire Safety Journal, 17, 1991, 1–12.

furnishings, wall coverings, windows, walls, and even exterior structure panels provide
many opportunities for them to come into contact with temperatures of this magnitude
(see Table 5-3).50

BEHAVIOR OF PLASTICS
When considering plastics as fuels, it is helpful to classify them by their behavior when
they are heated. Plastics that undergo reversible melting without appreciable chemical
degradation are called thermoplastics. Plastics that do not melt but rather decompose and thermoplastic ■ Polymer
leave behind a solid char are called thermosetting plastics. There is a small group of mate- that can undergo
reversible melting
rials that decompose into volatile materials that are not rigid and are sometimes called
without appreciable
elastomers; these include polyurethane rubber. The chemical and physical behaviors of chemical change.
some of the common polymers are shown in Table 5-3.
Minimum autoignition temperatures of common plastics are shown in Table 5-4.

Polyvinyl Chloride
The behavior of a particular plastic in a general fire is the result of many factors, which all
should be kept in mind when evaluating its potential contributions to the fire. Vinyl
(polyvinyl chloride—PVC) may burn slowly when tested in solid form in a laboratory but has
been observed to burn rapidly and contribute greatly to flame spread when present as a thin
film on wall coverings (probably due to the adhesives and plasticizers used). PVC wire insu-
lation softens and melts when heated by overheated wiring but chars if exposed to a flame.
Chapter 5 Combustion Properties of Solid Fuels 145
 TABLE 5-4 Autoignition Temperatures of Common Plastics

MINIMUM IGNITION TEMPERATURE

PLASTIC °C °F
Polyethylene 365* 488 910
Polyisocyanate 525 977
*
Polymethyl methacrylate 310 467 872
*
Polypropylene 330 498 928
Polystyrene 360* 573 1,063
Polytetrafluoroethylene 660 1,220
Polyurethane foam (flexible) 456–579 850–1,075
Polyurethane foam (rigid) 498–565 925–1,050
Polyvinyl chloride 507 944

Sources: NFPA, Fire Protection Handbook, 19th ed. (Quincy, MA: NFPA, 2003), Table A-6.
*
V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire Science Publishers, 2003), 244.

Nylon
One of the first synthetic fibers, nylon, is widely used in carpets. Most nylon-fueled flames
tend to be self-extinguishing, but under intense radiant heat flux, nylon fiber carpet has
been shown to burn with some enthusiasm.

Polyurethane and Polystyrene Foams

Unless it is properly treated with fire retardants, polyurethane foam is very flammable, as
it is readily ignited with any open flame. It produces a hot, smoky flame that progressively
decomposes more and more of the foam, with the “melt” joining the burning substance
at the bottom of the “crater.” Rigid (and flexible) polyurethane foam and polystyrene
foam are used as thermal insulation in equipment and in structures and have contributed
to the rapid spread of many destructive and often-fatal fires. The 2003 tragic nightclub
fire in West Warwick, Rhode Island, killed 100 of the 400 patrons present when stage
pyrotechnics ignited the unprotected urethane foam sound insulation covering the stage
walls. The fire growth was captured on videotape and shows progression from localized
flashover in the drummer’s alcove into the main room (苲10 m ⫻ 15 m) in less than 3 min-
utes as the fire spread across the walls with extreme rapidity.51 Plastic foams in foam-core
steel structural “sandwich” panels have become more commonly used in building com-
mercial buildings.52 Polystyrene foam–filled panels are especially susceptible to failure at
joints and seams (see Figure 5-12). Others have been seen to delaminate, exposing the
melting plastic foam to external flames.53 Recent advances in this technology have
improved the fire performance of some sandwich panels, with some being resistant to
both ignition and melting.54 Investigators should take a sample of any polymeric filler for
analysis before reaching conclusions about sandwich panel contribution to fire spread.

Latex or Natural Rubber

Latex or natural rubber foam products have largely been replaced by polyurethane foams,
but the former are still encountered in older furnishings and carpet pads. Although latex foam
products sometimes physically resemble high-density urethane foams, their flammability
properties are very different. Although urethane foams can be ignited by an open flame
(Figure 5-13a) or intense radiant heat from an external source, most cannot be ignited by
direct contact with a glowing cigarette (as in Figure 5-13b) unless they are covered with
a susceptible fabric or are in contact with a sustained heat source (as in Figure 6-29).55
146 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-12 Underside
of foam core sandwich
panel floor unit ignited by
overheated floor furnace
on right. EP foam melting
and igniting to ignite
plywood deck. Steel
channel support in center.
Flame is on edge of
plywood deck. Courtesy
of John D. DeHaan.

Unlike urethane foams, latex foams (Figure 5-14a) can be ignited by a small glowing igni-
tion source (such as a cigarette) and once ignited will sustain a smolder and are capable
of rekindling into a flaming fire many hours after suppression of flaming combustion, as
illustrated in Figure 5-14b. Latex foams are also capable of self-heating to the point of
spontaneous ignition if preheated to a modest level, whereas finished urethane foams do
not self-heat. (The chemical reaction that creates the foamed elastomer, however, is
exothermic, and incompletely cured blocks of foam and waste products have been known
to self-heat to ignition.) Like urethane foams, latex foams can be ignited by a flaming
source and will generate a dense, toxic smoke when burning. This smoke has been linked
to “cold” smoke explosions.56

FIGURE 5-13A Polyurethane foam mattress topper burns readily FIGURE 5-13B A cigarette on polyurethane foam alone produces
after match flame ignition. Courtesy of John Jerome and Jim Albers. longitudinal char but no sustained combustion (glowing or flaming).
If covered with heavy cotton fabric, the PU foam may sustain
combustion. Courtesy of John D. DeHaan.

Chapter 5 Combustion Properties of Solid Fuels 147

FIGURE 5-14A Typical latex foam with curing/ventilation holes. FIGURE 5-14B A close-up of the perforated latex foam from an
Courtesy of John D. DeHaan. upholstered sofa fire that rekindled owing to self-heating. Latex foam,
unlike polyurethane, can self-heat if warmed. Courtesy of John D. DeHaan.

Thermoplastics
Thermoplastics are molten at the temperatures of the burning fuel surface, so they will run
and drip as they burn, as pictured in Figure 5-15. If the droplets are on fire as they fall to
floor level, they can ignite new flames along the vertical upholstered surfaces of sofas.
Plastic skylights, lampshades, and light fixture diffusers have been known to be ignited by

FIGURE 5-15 Burning

synthetic upholstery melts
as it burns. Burning
droplets ignite the carpet
below and the lower
portion of the vertical
upholstery face. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

148 Chapter 5 Combustion Properties of Solid Fuels

FIGURE 5-16A Large polypropylene FIGURE 5-16B Large polypropylene waste container ignited with crumpled paper becomes
waste bins, if not flame retardant, are a very large sustained pool fire. Courtesy of Lamont “Monty” McGill, Fire/Explosion Investigator.
easily ignited to produce very large
sustained pool fires, especially when
empty. Courtesy of Lamont “Monty” McGill,
Fire/Explosion Investigator.

the ceiling temperatures in a general room fire. As these articles burn they soften and sag,
finally falling, sometimes onto combustibles below, producing isolated, discrete points of
origin. During post-fire examination, some investigators have been misled by these numer-
ous separate fires, suspecting an incendiary cause. The investigator should be careful to
attempt to correlate all unusual fire patterns with results from the fuel load before ascrib-
ing a criminal origin to a suspicious fire. Plastics are being used increasingly as a substitute
for glass in windows, particularly in schools where vandalism is a problem. These plastics
have been ignited by matches or lighters and gone on to ignite major structural fires. Large
polypropylene waste bins, if not flame retardant, are easily ignited to produce very large,
sustained pool fires, especially when empty, as seen in Figures 5-16a and b.
Plastic Building Materials
Several new building materials have been introduced for use as exterior decks, railings,
and fences. These are basically extruded plastic (reinforced or nonreinforced) often made
entirely from recycled polyethylene, PVC, or ABS,57 as seen in Figure 5-17. These mate-
rials have been shown to be easily ignited with a match or other small open flame to pro-
duce a rapidly growing fire, with molten burning material dripping or running off as it
burns. The result can be a large pool fire of the molten material or rapid ignition of other
combustibles located below, as seen in Figure 5-16. Such rapid combustion also leads to
loss of structural strength.58

SPECIAL CONSIDERATIONS FOR FIRE INVESTIGATORS

The potential of plastics as a source of fuel in both accidental and intentional fires should
not be overlooked. Years ago, televisions, radios, and other small appliances were
enclosed in wooden or metal housings, which would often contain the small flames of
internal fires; today they may be made of synthetics that fuel the initial fire and may help
spread it. Some but not all plastic appliance housings include fire retardants and will resist
Chapter 5 Combustion Properties of Solid Fuels 149
FIGURE 5-17A Extruded plastic deck planks come in a variety of FIGURE 5-17B Plastic decking ignited with a match. Courtesy of
forms. Courtesy of John D. DeHaan. John Jerome and Jim Albers, Barona Fire Dept.

FIGURE 5-17C Dripping

flaming plastic with rapidly
growing flames above.
Courtesy of Jim Albers and
John Jerome, Barona Fire
Dept.

FIGURE 5-17 Fire test of extruded plastic deck planks.

modest internal fire exposure. The prudent investigator should have such housings tested
for fire retardants before ascribing an origin to such an appliance.
When the fire investigator is faced with reconstructing a fire, these considerations of
fuel type are critical. The materials that make up the bulk of the fuels involved in most
structure fires have changed dramatically over the last 30 or 40 years, affecting the way
in which fires ignite and spread, the combustion products they create, the heat release
rates and temperatures they produce, and the residues they leave behind.

Post–World War II Finishes and Furnishings

A typical house was finished in the 1940s or 1950s with plaster walls over wood lath or
wire mesh, painted or covered with heavy paper or cloth wallpaper. Floors were bare
wood or linoleum with wool carpets and rugs. Window coverings were heavy cotton (bro-
cade) or wool fabrics (or steel-and-wood window blinds). Furniture was often bare wood.
When it was upholstered, it was usually with cotton velour or wool mohair, or leather,
over padding of horsehair or cellulosic (vegetable) fibers (cotton, coir, kapok, Spanish
moss, etc.). Latex foam rubber was found in some cushions and pillows, and mattresses
were stuffed with cotton or other vegetable fiber or even feathers. Rooms were usually
poorly insulated, and glass windows were single glazed.
150 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-18
‘Traditional’ cotton
mattress burned over a
period of hours (unob-
served). Heat release rate
was never sufficient to
ignite nearby combustibles
except for tablecloth on
nearest table. Courtesy of
Diane Spinner, Fire Cause
Analysis.

Such furnishings were all combustible, but how readily? Most of the materials just
listed were susceptible to smoldering ignition sources such as a dropped cigarette, but,
with a few exceptions, most would not be readily ignited (to a self-sustaining fire) by a
short-lived flame source such as a common match. Ignition of all but latex foam and
kapok took many seconds or even minutes of exposure, and all produced large quantities
of smoke prior to ignition to flames. Once ignited, they would burn slowly, often reluc-
tantly, producing small flames, preferring to smolder instead. Heat release rates would be
very low (300–400 kW being typical for a cotton sofa), so spread to other fuels would
require direct flame contact. See Figure 5-18 for an example. Although the localized flame
temperatures would be “normal” [i.e., 800°C to 900°C (1,700° to 1,800°F)], because the
heat release rates were so low (and room insulation so poor), overall hot gas layer temper-
atures would usually remain below the 600°C (1,200°F) range. Progression to flashover
would therefore be rare and would take a long time to develop when it did occur.
Combustion products included carbon monoxide (CO), soot, acrolein (from wood), and
hydrogen cyanide (HCN) (from wool). Carpets and floors burned only with reluctance and
only after a lot of radiant heat had been applied. Post-fire residues of liquid accelerants were
relatively easy to identify based on odor and, later, by simple gas chromatography (GC),
because the furnishings offered little in the way of interfering pyrolysis products.

Changing Contemporary Finishes and Furnishings

By the 1960s and 1970s, things were changing. Wall coverings tended to be paint or
vinyl wallpapers over gypsum wallboard (Sheetrock; drywall), with its better durability
under fire exposure. Floor coverings were often wall-to-wall carpet of nylon or polyester
face yarns with jute fiber backings over a pad of butyl rubber or mixed fiber. Furnishings
were often polyurethane (PU) foam or cotton padding covered by cotton or cotton/syn-
thetic upholstery fabric (or vinyl). Window coverings were cotton or synthetic, or some-
times fiberglass fabrics. After 1972, mattresses were more commonly combinations of a
cotton core with a polyurethane foam layer to improve resistance to smoldering sources
(cigarettes). Thermal insulation in rooms became more common, and double glazing
appeared.
Chapter 5 Combustion Properties of Solid Fuels 151
 As a result, fire behavior began to change. Synthetic fabrics and PU foam added smol-
der resistance to furnishings but made them more susceptible to ignition by short-lived
open-flame sources. Fires became hotter, and heat release rates increased as more synthet-
ics became involved. Fire growth was more rapid in this new style of furniture, and rooms
could approach flashover more readily and in a shorter period of time. Numerous
researchers have confirmed these results.59 Carpets, however, and many other furnishings
resisted ignition even by radiant heating. When carpets did ignite, the face yarns tended
to burn off, but the heavy jute backings would burn slowly and remain in place to pro-
tect the rubber pad beneath. Vinyl products could resist ignition and slow the progress of
the fire, but when they finally became involved, could contribute hydrogen chloride (HCl)
vapors to the fire gases. Gas chromatography could discriminate flammable liquid
residues from all the pyrolysis patterns, thanks to improved chromatography, because the
peak patterns produced by these pyrolysis products were easily distinguished from the
patterns from common flammable liquids.60
By the late 1980s, floor coverings still consisted of various synthetic face yarns, but
the backing came to be made of polypropylene due to its lower cost and better resistance
to odors and mildew (see Figures 5-19a and b). Polyurethane foam carpet padding (new
or rebond) replaced latex rubber or fiber padding exclusively. Under severe radiant heat-
ing, face yarns could melt and even ignite, as could the backing, exposing the combustible
PU foam underlay beneath. Such carpets would tend to resist localized flame and heat
sources, but once alight could burn more completely than before and create much more
heat and floor-level damage than ever before. Combustion products now included a wide
range of aldehydes, ketones, and toxic intermediates known as free radicals.
Furnishings became almost exclusively polyurethane foam padding (rarely with flame
retardants) with synthetic covering—very difficult to ignite with a smoldering cigarette but
readily ignited by even a small flame. Once furnishings were ignited, flames could spread
quickly and engulf all of a large chair or sofa in flames in less than 5 minutes.61 Thermoplastic
upholstery materials, with their low melting points, melt as they burn, producing molten,

FIGURE 5-19A Modern

Carpet Pads. Upper left:
Foam rubber “waffle.”
Lower left: High- density
foam. Lower right:
Rebond polyurethane.
Upper right: “Waste” felt
automotive. Courtesy of
John D. DeHaan.

152 Chapter 5 Combustion Properties of Solid Fuels

 FIGURE 5-19B Carpet
Backings. Top left:
Traditional broadloom
carpet with woven jute
backing. Top right: Jute
backing on wall-to-wall
carpet. Lower left:
Synthetic backing coated
with sealant. Lower right:
Polypropylene and other
synthetic fiber backing.
Courtesy of John D. DeHaan.

burning droplets of materials that fall to the base of the furniture and institute rapidly
growing vertical-face fires on the sides of the furniture, as well as drop-down damage to
floors and carpets beneath (see Figure 5-15). Ignition of a modern mattress and comforter
is shown in Figure 5-20. Recent tests by DeHaan have shown that a modern queen-size
mattress and comforter ignited by a single flame will result in flashover of a large 3.35 ⫻
4.27 ⫻ 2.74-m (11 ⫻ 14 ⫻ 9-ft) bedroom in 5.5 minutes. Calorimetry tests have revealed
that modern queen-size mattresses are capable of producing a heat release rate of over
3 MW.62. Since 2007 all mattresses and futons sold in the United States (even for residen-
tial use) are required to pass a stringent ignition test, 16 CFR 1633. This test involves
exposing the side of the mattress to a sustained 18 kW gas burner flame and mandates
the maximum heat release rate and total heat released after removal of the burner.63 This

(a) (b)

FIGURE 5-20 Open-flame ignition by small gas burner causes rapidly growing fire on (a) common polyurethane
mattress set and (b) polyester/cotton comforter. (Ignition with a match flame would add less than one minute
to these development times.) Courtesy of Bureau of Electronics & Appliance Repair, Home Furnishings.

Chapter 5 Combustion Properties of Solid Fuels 153

FIGURE 5-21 Traditional
cotton padding (right)
is easily ignited with a
cigarette and forms a
loose, powdery black
ash. Polyester fiberfill is
thermoplastic, not ignitable
with a cigarette. It is
easily ignited with a flame
and burns readily, forming
a black, hardened knob
of molten plastic (arrow).
Courtesy of John D. DeHaan.

requirement will significantly reduce the ignitability and size of any resulting fire originat-
ing from new mattresses. Sadly, this standard does not apply to foam mattress toppers or
other bedding, and imported mattresses with forged approval labels have already been
seen in the marketplace. This testing is described in more detail in Chapter 11.
Figure 5-21 illustrates the difference between traditional cotton padding and its mod-
ern replacement for pillow and cushion stuffing, polyester fiberfill. Polyester is a thermo-
plastic that ignites readily on contact with a flame and produces a fast-growing flaming fire.
Today, wall-to-wall carpets often include low-melting-point and readily ignitable
polypropylene face yarns with polypropylene backing over polyurethane underlay (par-
ticularly in low- or medium-price-range installations). Except for “high-end” mattresses,
for which latex foams are being offered once again (due to their better breathability and
reduced moisture buildup), polyurethane foams are nearly universal in mattresses and fur-
niture cushions. Such PU foams are very resistant to smolder ignition but are still easily
ignited with a match. The acquisition of comparison samples of upholstery fabrics,
padding (fillers), and floor coverings has become more important to the fire investigator
than previously.
New regulations for mattresses do not imply a similar improvement for upholstered
furniture. CPSC has recently proposed new regulations for upholstered furniture flamma-
bility, but these would do little more than enforce existing voluntary industry efforts to
make furniture cigarette-resistant. Flaming-ignition behavior would not be regulated.
Flaming-ignition behavior has been regulated in California for more than three decades
by the California Bureau of Home Furnishings. However, these regulations have not
resulted in consumer furniture that is difficult to ignite from flames, nor furniture with
improved HRR characteristics. It is apparent, however, that the simple inclusion of a bar-
rier layer of PVC, aluminized fabric, wool, or even fiberglass can dramatically improve all
types of fire resistance at a very modest increase in cost.64
The polypropylene carpets are becoming a widespread problem. They will pass the
methenamine tablet test (originally DOC FF 1-70, now known as 16 CFR 1630) (meaning
154 Chapter 5 Combustion Properties of Solid Fuels
 FIGURE 5-22 Poly-
propylene carpet ignited
with small pools of
gasoline (at center of
each panel) continued
to burn at 0.5–1.0 m2/hr
until self-extinguished.
Courtesy of John D. DeHaan.

that a dropped match or similar flame of 苲50W HRR will not ignite the carpet to a self-
sustaining fire) and are therefore legal for use in residences. Unfortunately, exposure to a
slightly larger fire (such as a burning crumpled sheet of paper) with its higher heat flux
can initiate a self-sustaining fire that spreads at about 1 m2/hr (10 ft2/hr) with very small
flames 5 to 8 cm (2 to 3 in.) in height along the margins (i.e., a very low heat release
rate).65 If the fire starts in an unoccupied building, it can burn for hours, consuming entire
rooms of carpet without running out of oxygen, due to its very low rate of combustion
(see Figure 5-22). In addition, because of its low critical incident heat flux, such carpet is
more readily ignitable by radiant heat even at some distance away from a large fire (such
as a chair fire). As the face yarn burns, the polypropylene backing melts and shrinks (as
shown in Figure 5-23), exposing the combustible PU foam underlay, which can now ignite
and sustain the spread of fire through and across the carpet. The rate of spread is depend-
ent on the interaction between the carpet and the pad and can be affected by such vari-
ables as whether the carpet is stretched tightly and securely fastened or lying loosely atop
the pad with plenty of air between the two. Urethane foam carpet pads (particularly
rebond) and all-synthetic carpets produced complex “stews” of pyrolysis products that
can sometimes be quite difficult to distinguish from some non-distillate-type petroleum
products. Comparison samples of unburned carpet and pad are very important evidence.
Today’s furniture is markedly improved in its resistance to the most common type of
accidental ignition: a dropped cigarette or other forms of glowing sources (electric
heaters, glowing electrical connections, etc.). The significant disadvantage is little or no
Chapter 5 Combustion Properties of Solid Fuels 155
FIGURE 5-23 Synthetic
carpet melted by radiant
heat from chair fire.
Courtesy of David J. Icove,
University of Tennessee.

resistance to flaming sources. Once alight, such furnishings can be completely involved in
3 to 5 minutes and be reduced to a charred frame in 10 minutes while producing very high
temperatures [in excess of 1,000°C (2,000°F) in the flame plume] and enormous heat
release rates (2 to 3 MW is common for a sofa or recliner today).66 Such performance vir-
tually ensures that the average-size furnished room with an open door will be completely
engulfed in a post-flashover fire within 3 to 10 minutes of first established flame. The
intensity of such fires can obliterate traces of ignition sources, obscure fire patterns (if sus-
tained for more than 5 minutes), induce the speed of spread and kind of damage once
thought possible only for accelerated fires, and make the recovery and identification of
possible accelerant traces very difficult. Gas chromatography/mass spectrometry is now
needed to discriminate possible accelerant traces from pyrolysis products. Further
flame resistant ■ changes for the future include flame resistant performance standards for bed covers and
Material or surface household furnishings. As these products appear fire growth will be affected. The reader
that does not maintain
is referred to Chapter 11 for more information.
or propagate a flame
once an outside source The best the investigator can do is to be aware of the contributions such furnishings
of flame has been can make to the ignition and growth of a fire and to take whatever steps necessary to doc-
removed. ument the kind of furnishings present (new or old? cellulosic or synthetic fabrics and
padding?) and, especially, to collect comparison samples of carpet, pad, upholstery, and
padding whenever possible for later identification in the forensic lab (even partially
burned samples are better than no samples at all). This may be the only way to accurately
reconstruct the train of events of a fire.

Paint
Paint is a greatly misunderstood product, in part because of the tremendous diversity of
surfacing coatings, all described as paint. Historically, paint consisted of a drying oil (such
as linseed or tung oil) in which mineral-based pigments were suspended. The mixture was
then thinned using turpentine or a petroleum distillate. The thinner evaporates shortly
156 Chapter 5 Combustion Properties of Solid Fuels
after application, and so it plays an important part in the flammability of the paint only
while it is still wet. The drying oils and other resins present are combustible even after the
paint has dried and constitute the major fuel contribution of many paints. The mineral
powders once used universally as pigments were not combustible and so tended to retard
the combustion of the paint itself. This is no longer the case. The cost and environmental
hazards posed by most inorganic pigments have caused them to be replaced in large part
by cheaper organic dyes and fillers, which will burn. The newer water-base paints are gen-
erally emulsions of latex, polyvinyl acetate, or acrylics in water when applied and form a
plasticlike polymer coating when they dry. Their properties are then similar to the plas-
tics on which they are based. Varnishes, shellacs, and lacquers tend to be more com-
bustible, since they are often based on naturally occurring resins or on combustible plastic
polymer equivalents.
Most paints and coatings will add slightly to the fuel present in a structure and can
aid the spread of flames from one part of a room to another. In some cases the paint layers
have been seen to soften and peel off the wall or ceiling as they ignite and thus aid the
spread of fire downward if there is readily ignitable fuel below it, as pictured in Figure 5-24.
This is especially true if there are many layers of paint. Then, even “noncombustible” sur-
faces of painted steel or gypsum board can contribute to the fire’s spread.
There are also numerous coatings intended to protect the substrate from fire effects.
These coatings, called intumescents, swell and bubble when heated, providing a noncom-
bustible layer of thermal insulation that slows the incursion of heat into the substrate.67 Such
intumescents are widely used to protect steel structural members as well as combustible surfaces.

FIGURE 5-24 Test fire

of gasoline and diesel
caused ignition of paint
on walls, which peeled
off and fell as flaming
sheets. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

Chapter 5 Combustion Properties of Solid Fuels 157

 Whether a paint or other coating will aid or retard the flames of a small, growing fire
cannot be generalized without testing the specific product, but the contribution of layers
of paint to the fuel load of an ordinary structure cannot be overlooked by the careful
investigator. Recent tests by McGraw and Mowrer using a cone calorimeter revealed
that at applied heat fluxes of 50 kW/m2, multiple layers of latex paint on gypsum wall-
board did not support flame spread but doubled the heat release rate (from paper alone)
to 苲111 kW/m2 (of wall area). At heat fluxes of 75 kW/m2, fire spread was modestly
accelerated, and the heat release rate rose to 134 kW/m2 (of wall area). (At 25 kW/m2
there was no ignition of wallboard—painted or unpainted.)68

Metals
Metals, as fuels, are so rarely of importance that they may well be overlooked. Actually,
most metals can be burned, and many metals, when finely divided, are susceptible to
direct combustion in air. Others can burn when they are in relatively massive form, but
generally only in a very hot environment. Some metals are pyrophoric; that is, they burn
spontaneously in air when finely divided. A metal like uranium, for example, is extremely
hazardous when finely divided because of its tendency to oxidize in air, and the heat of
the slow spontaneous oxidation may be sufficient to ignite a larger quantity of stored ura-
nium powder. Iron powder is, to a considerable extent, pyrophoric, although without the
special hazard that characterizes some other metals. Iron in the form of steel wool or fine
powder will burn readily and requires only a modest ignition source (a match or electri-
cal heating). Ignition temperatures are as low as 315°C (600°F) (for powder) and 377°C
(710°F) (for fine steel wool). (See Figure 5-25.) Very fine iron powder is pyrophoric.69
The important qualification of the dangers associated with metal combustion is the state
of subdivision. The finer the metal powder, the more likely it is to ignite, but rarely with-
out supplemental ignition. The combustion properties of some of the more seriously haz-
ardous metals are treated in some depth in Chapter 13.

FIGURE 5-25 Steel-wool

fire—rapid glowing
combustion with emission
sparks. Courtesy of John
D. DeHaan.

158 Chapter 5 Combustion Properties of Solid Fuels

 The danger associated with burning of metals is confined almost exclusively to indus-
trial plants, especially those in which the metals are handled in a finely divided state.
Many disastrous fires of this type have been fed primarily by metal fuel. In general, such
fires are so intensely hot and violent that they reduce the structure far more completely
than is ever seen in a fire that is primarily fed by wood as fuel. Chemical laboratories or
manufacturing facilities may contain enough reactive metals like sodium, potassium, or
phosphorus to cause special fire hazards (see Chapter 13).

MAGNESIUM
Perhaps the most important metal that can serve as a fuel in structure fires is magnesium.
Magnesium is not readily ignited unless it is rather finely divided—in the form of shav-
ings, filings, or dust. In a large fire, massive magnesium castings—wheels, aircraft com-
ponents, and the like—can burn very fiercely, producing an enormously hot fire (see
Figures 5-26a and b). In a fire where large magnesium pieces have burned, it is more likely
that the magnesium was ignited by a strong surrounding fire rather than the reverse. The
reported ignition temperature of magnesium ranges from 520°C (968°F) (powder) to
623°C (1153°F) (solid). Magnesium/aluminum alloys with more than 10 percent magne-
sium have ignition temperatures in the 500°C to 600°C (930°F to 1,112°F) range.70
A matter that is of special concern with magnesium especially, and with some other
metals that are rarely encountered, is that when water is sprayed on hot magnesium, large
quantities of hydrogen gas will evolve and will either explode or increase the fire by mam-
moth proportions. In fighting a large magnesium fire, it is of the greatest importance that
the firefighters use foams, carbon dioxide, or other essentially nonaqueous materials.

ALUMINUM
Aluminum is much more difficult to ignite than is magnesium because of its tendency to
form a fine adherent film of oxide on the surface, which protects the metal that under-
lies the film. Thus, in structural fires, it is very common to find much aluminum that has
melted but not burned. [Under special test conditions, ignition temperatures have been
measured from 1,500°C to 1,750°C (2,822°F to 3,182°F)].71 Residential fires are hot
enough to exceed the melting point of aluminum (660°C; 1,220°F) in a majority of
instances. This temperature can be reached during fire growth in the hot gas layer,
directly in a flame plume, or in a post-flashover fire at floor level. Thin aluminum roofing

FIGURE 5-26A Magnesium alloy engine block from light aircraft FIGURE 5-26B Magnesium extrusions burn with white-hot flames.
produces a large sustained white flame. Courtesy of David Barber, Courtesy of Lamont “Monty” McGill, Fire/Explosion Investigator.
Goleta, CA.

Chapter 5 Combustion Properties of Solid Fuels 159

 or sheathing may oxidize to white aluminum oxide ash, especially in commercial structures
or mobile homes where aluminum sheathing is common. Even this material, however, is
more likely just to melt or to blow away as ash in the updraft from the fire than to burn.

Coal
From the standpoint of its general use, coal ranks high as a fuel widely used in industrial
power production and general heating purposes. It is of secondary consequence as a prob-
lem for the fire investigator. There is progressively less use of coal for heating structures
than in the past, but it is of increasing industrial importance as a fuel because of new low-
pollution combustion techniques. Both the mining and storage of coal are hazardous due
to the combustibility of coal. For example, under proper conditions, both stored and
unmined coal is subject to self-heating to the point of spontaneous ignition.72 Moisture
appears to be essential; large masses of coal and long time frames are typically needed, as
well as both oxygen and moisture, to trigger self-heating.73 The dimensions (mass) of piles
of coal are critical to self-heating, with deep piles more susceptible (with content percent-
ages dependent on type—anthracite versus bituminous).74
Coal is highly carbonaceous, being an organic fuel made up of approximately 80 to
90 percent carbon, 4 to 5 percent hydrogen, and 12 percent oxygen, nitrogen, sulfur, and
other elements in (usually) trace quantities. The presence of these other elements suggests
that coal contains a variety of compounds that can undergo pyrolysis. Coal differs from
most other organic fuels in that the proportion of such pyrolyzable fuels is relatively
small. The more readily pyrolyzable compounds that are present in a fuel, generally the
easier it is to ignite. Coal requires considerable input of heat from an ignition source to
initiate a self-sustaining, flaming combustion (in a bulk solid) because of the small quan-
tities of pyrolyzable compounds. Coke is made by heating coal in the absence of oxygen
to drive off light fractions and impurities to increase its heat content.
Explosions involving coal dust can occur in mines and in coal-handling equipment
(such as conveyors) in mines, power plants, and railroad facilities. Smaller quantities and
dry conditions are involved in such explosions. Reported autoignition temperatures range
from 200°C to 600°C (400°F to 1120°F) depending on grade (lignite being the lowest).75

Dust Explosions
Any finely divided solid fuel is capable of undergoing explosive combustion if adequately
mixed with air and suitably ignited. Such fuels as coal dust, metallic dusts, sawdust, flour,
rice dust, cornstarch, milk powder, sugar, aspirin, plastic dust, dry copier toner, computer
printer toners, and powdered sulfur have all been involved in such dust explosions.76
Explosive limits (flammability limits) are defined in terms of units such as grams of fuel
per cubic meter, which are difficult values to measure or estimate after the incident.
Minimum explosive concentrations range from 30 g/m3 for fuels like powdered alu-
minum, sulfur, or polypropylene; to 60 g/m3 for coal dust; to 200 g/m3 for PVC powder,
sugar, or soy flour. The minimum concentration is also affected by particle size, with finer
dusts having a lower minimum concentration.77 Ignition energies are also particle-size
dependent, with the finer dusts being the most easily ignited. As with vapor or gaseous
fuels, the limits are sensitive to the energy of the ignition source—the stronger the source,
the wider the explosive range. Flame speeds tend to be lower than for vapor/air explo-
sions, and there tends to be more post-blast burning. Because there are usually layers of
dust on adjacent surfaces that are stirred up as the blast wave passes, there is often a series
of explosions, sometimes increasing in severity as the sequence progresses. If there is an
inorganic oxidizer present as a dust suspension along with the fuel, the resulting explo-
sion can be very energetic. One such mixture is aluminum powder and potassium perchlo-
rate flash powder. If this mixture is suspended as a dust and ignited, the resulting deflagration
160 Chapter 5 Combustion Properties of Solid Fuels
can approach detonation velocities. This mixture has an extremely low ignition energy, and
a major explosion can be triggered by even a small arc, friction, or even a heated surface.

Combustion Products of Solid Fuels

FLAME COLOR
The color of the flame, especially in the very early stages of a fire, can be highly signifi-
cant but will rarely be remembered or recorded with accuracy, although it may be cap-
tured in videos or photos of the fire. It is clear that in most fires the flame color is not
likely to have great significance, because most structural, as well as outdoor, fires involve
only cellulosic material such as wood and some other organic construction materials such
as asphalt, paint, and similar supplemental substances. For all these materials, the flame
will be more or less yellow or orange in color, and only the amount and color of smoke
that they give off will be related to the character of the fuel and the adequacy of the oxy-
gen supply. It should also be noted that flame color, when not complicated by special ele-
ments that produce unusual colors, is related to the flame temperature, which is, in turn,
generally related to the oxygen admixture with the fuel. Thus, the color of the flame is most
often indicative of burning conditions as well as of the actual temperature of the flame.
However, unusual colors of flame are sometimes seen in fires. When this is the case,
it would be poor judgment not to give them consideration. For example, the simpler alco-
hols burn with a relatively blue or purple flame. If such a flame was observed immediately
after the fire started, it could well indicate the use of this type of material as an acceler-
ant. It would also rule out hydrocarbon accelerants, all of which burn with a yellow
flame. Natural gas (if not premixed), like liquid hydrocarbons, generally burns with a yel-
lowish flame, tending to be blue in the periphery. If adequate air is mixed with it, as is
done in a heating burner, the flame is colorless to blue. Hydrocarbon gases or vapors
properly mixed with air burn with blue flames. Without the proper air mixture, they
become yellow and smoky due to incomplete combustion.
One gas that is of special interest relative to flame color is carbon monoxide, which
is extremely flammable. Most of it is produced by oxygen-deficient flames in which the
combustion is not complete, and the gas escapes before it is consumed by further oxida-
tion. Carbon monoxide (CO) also burns with a blue flame. In a confined space, if the
amount of carbonaceous fuel involved greatly exceeds the amount of air available to com-
bine with it, there will be strong generation of carbon monoxide. The CO may escape to
a region richer in oxygen, where it may be reignited and ring the primary fire with blue
flames. Such a situation will provide insight into the combustion of excess fuel with lim-
ited air in the initial fire area.
A good illustration of this property was reported by Kirk in an early edition of this
text. In that case, the most intense part of a large hotel fire was in a subbasement and fed
by hydrocarbons.78 Firefighters noted blue flames coming from a sidewalk entrance to the
basement while black smoke was emerging elsewhere. The blue flames were at a signifi-
cant distance from the subbasement, and there was as yet little fire in the intervening
space. The reasons for the existence of the bluish flame were not immediately perceived,
and the only final explanation that fit the facts was a secondary burning of carbon
monoxide produced in the basement fire because of limited ventilation.
Some other special flame colors may occasionally be of significance. Colored flames
have long been generated for pyrotechnical purposes by adding certain metallic salts that
generate the metals’ spectral colors at flame temperature. Traces of sodium, present in
most fuel, give a yellow flame, difficult to distinguish from the yellow produced in flames
by glowing carbon particles. Some of the yellow that is so common in flames is partially
due to sodium traces, but not specifically so. The significance of sodium in this connec-
tion is therefore minimal. Strontium salts give the bright red color used in many flares and
Chapter 5 Combustion Properties of Solid Fuels 161
 decorative flame and explosion effects. Copper halides impart an intense green color to
the flame, potassium salts give violet, and barium salts, a yellowish green. Brilliant white
flames will accompany the burning of magnesium/alloy components (as shown in Figure
5-26). Many other variations of color will be characteristic either of the elements or of
certain of their compounds.79
Clearly, these unusual colors of flame will only occasionally be of significance in the
investigation of fires. Assume, for instance, that an intense fire burned in the interior of a
motor vehicle. This fire might have been due to various types of fuel material and igni-
tion sources, one of which could have been road flares. If a brilliant red color was noted
in the early part of the fire, this information could assist in locating the early fuel as flares.
Yellow flames could indicate more conventional fuels such as synthetic upholstery. The
investigator who is aware of the potential value of flame color in the interpretation of fire
will be alert for any color that is unusual or not explainable in the most obvious hypoth-
esis, as this could lead to more accurate appraisal of the cause and sequence of many fires.
The firefighters who were first on the scene should be interviewed to determine any
unusual flame conditions on their arrival.

SMOKE PRODUCTION
Although the character of the smoke produced by a fire is variable, observation of the
smoke may be helpful in obtaining an indication of the type of the fuel. When complete
combustion occurs, most materials produce little or no visible smoke. Under these ideal
conditions, all carbon is burned to carbon dioxide, a colorless gas. Incomplete combus-
tion of carbon leads to carbon monoxide, also colorless and odorless but having very
important toxic properties. Incomplete combustion of carbonaceous materials also pro-
duces various highly carbonaceous compounds that are opaque, such as soot, which is
largely carbon.
Hydrogen, universally present in all organic materials, burns to water vapor, which is a
major constituent of all combustion gases. It does not contribute to the smoke characteris-
tics, although it may at times be recognized as steam. Other elements often present in organic
materials, such as nitrogen, sulfur, and halogens, contribute gases to the combustion prod-
ucts. However, only under very special conditions do these alter the color of the smoke. They
may contribute strongly to its odor when present in larger-than-ordinary amounts. Such
effects will be encountered in some industrial fires but rarely in residential ones.
The color of smoke is almost entirely determined by the character and type of fuel
and the availability of oxygen for complete combustion. Most materials, including all
hydrocarbons, will never burn completely even in the presence of excess air, without pre-
mixing of the fuel with air. This is even true of natural gas when burning in large volumes.
As the hydrocarbon molecules become larger, such as those in petroleum oils, more
air is required, and it becomes more difficult to completely mix the fuel with air. Thus, in
an ordinary fire, materials such as petroleum distillates tend to produce a large quantity
of dense black smoke. Therefore, the presence of much black smoke may indicate the
burning of a highly carbonaceous material like the petroleum products sometimes used as
accelerants. Dense black smoke, however, is usually more indicative of the amount of ven-
tilation and the stage of development at which the fire is burning. For example, fire burn-
ing in a restricted space without adequate and complete ventilation will always form more
smoke and darker smoke than when the same fire is well ventilated. The early stages of a
developing structure fire may be dominated by the incomplete combustion and partial
pyrolysis of fuels present. These products of incomplete combustion are often characterized
by the dense clouds of dark smoke they form until flashover occurs, windows break, and
the flammable vapors freely combust. At this point, the smoke becomes lighter colored
and less opaque, sometimes almost transparent. If the initial fuel is cellulosic, the smoke
will be light colored from the outset due to the release of water vapor and absence of soot.
162 Chapter 5 Combustion Properties of Solid Fuels
 Oxygenated organic materials (such as cellulosics) burning freely in the presence of air
very often produce little or no colored smoke. Alcohols, for example, are in this category.
If used as accelerants, they will not be detected by observing the smoke. Wood and most
cellulosic building materials also fall into this category. They produce either a white or light
gray smoke, except when the supply of air is greatly restricted. Even under these circum-
stances, they will not produce a heavy black smoke, such as follows the burning of heav-
ier hydrocarbons. Asphaltlike materials, also common in building construction, and such
materials as tar paper, some paints, sponge rubber upholstery, adhesives, sealing com-
pounds, and most floor coverings, will generally produce a black smoke. These are more
likely to burn late in a fire, so that the production of heavy black smoke late in a fire is not
ordinarily an indication of the use of accelerant. Many plastics, especially polystyrenes and
polyurethane, produce thick black smoke when burned in large quantities in air.
The optical density of smoke produced during oxygen-depletion calorimetry is meas-
ured by a standardized light-path method in the exhaust duct. The value obtained can be
used to predict the visibility factors that affect tenability and escape paths for people in
the fire room.80 Once again, if comparison samples of suspected fuels are available, tests
can be conducted to assess their contributions.
Smoke color, especially colors other than the neutral white, gray, or black, may give
indications of special types of fuel being combusted if they are present in large amounts.
A smoke that has an unusual hue, such as red-brown or yellow, should lead to an inquiry
about the presence of materials other than conventional building materials and furnish-
ings. This is another occasion when questioning the fire crew may reveal useful insight or
even evidence.

Chapter 5 Combustion Properties of Solid Fuels 163

 CHAPTER REVIEW

Summary
Solid fuels account by far for most of the combustibles important if the investigator is to understand the igni-
consumed in both structure and wildland fires. tion and combustion of solids. Although the complete
Because of pyrolysis and solid/air (smoldering) com- chemistry of pyrolysis is too complex to be fully exam-
bustion, the behavior of solid fuels in fire is more com- ined here, its role in decomposing a solid fuel to sim-
plex than that of gases or liquids, which evaporate and pler, oxidizable components must be understood.
then mix more freely with air. Those processes are, Only one element is now missing from our examina-
nonetheless, fairly predictable if all appropriate vari- tion of the fire process—a source of ignition. No mat-
ables and vagaries are considered. For example, an ter whether the fuel is solid, liquid, or gas, without
appreciation of the mechanisms of heat transmission ignition there will be no fire to investigate. Sources of
by radiation, convection, and flame impingement is ignition will be examined in the next chapter.

Review Questions
1. Describe the phenomenon of pyrolysis. 7. What behavior affects the combustion of thermo-
2. Describe four of the routes by which a solid can plastics like polypropylene, nylon, or polystyrene?
be converted to a vapor. 8. What metals burn readily in air?
3. Describe what happens when wood burns. 9. What factors control the phenomenon of a dust
4. What conditions usually accompany low- explosion?
temperature ignition of wood? 10. Describe the difference between cellulosic and
5. Describe the difference between thermoplastic synthetic materials. Which is more easily ignited
and thermosetting synthetic materials. by a glowing cigarette?
6. Why is latex foam different from synthetic foams
in its combustion properties?

References
1. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire Wood in Fire, Report No. 2202 (Madison, WI: Forest
Science Publishers, 2003), 16. Products Laboratory, 1960), quoted in Fire Protection
2. Ibid., 9. Handbook, 17th ed. (Quincy, MA: NFPA, 1991),
3. E. Stauffer, “Sources of Interference in Fire Debris 3–26.
Analysis” in Fire Investigation, ed. N. Nic Daéid (Boca 9. Babrauskas, Ignition Handbook, 944.
Raton, FL: CRC Press, 2004). 10. Ibid., 942–52.
4. J. D. DeHaan, D. Brien, and R. Large, “Volatile Organic 11. G. C. McNaughton, Ignition and Charring Temperatures
Compounds from the Combustion of Human and of Wood (U.S. Department of Agriculture, Forest
Animal Tissue,” Science and Justice 11, no. 4 (October– Service, November 1944); see also Forest Products
December 2004): 223–36. Laboratory Report No. 1464, Wood Products 50, no.
5. D. D. Drysdale, An Introduction to Fire Dynamics, 2 (1945): 21–22.
2nd ed. (Chichester, UK: Wiley, 1999), 175. 12. X. Z. Xiang and J. J. Cheng, “A Model for the Prediction
6. Ibid. of the Thermal Degradation and Ignition of Wood
7. Ibid., 179. under Constant Variable Heat Flux” in Proceedings
8. B. L. Browning, ed., The Chemistry of Wood (New Fire and Materials 2003, San Francisco, CA (London:
York: Wiley-Interscience, 1951); C. R. Brown, “The Interscience Communications): 71–79.
Ignition Temperatures of Solid Materials,” NFPA 13. M. L. Janssens, cited in Babrauskas, Ignition Handbook,
Quarterly 28, no. 2 (October 1934), quoted in Fire 946.
Protection Handbook, 17th ed. (Quincy, MA: NFPA, 14. Xiang and Cheng, “A Model for the Prediction of the
1991), 3–26; H. O. Fleischer, The Performance of Thermal Degradation and Ignition of Wood.”
164 Chapter 5 Combustion Properties of Solid Fuels
15. H. W. Angell, Ignition Temperature of Fireproofed 38. J. Albers, “Straight Facts about the Standard Time-
Wood, Untreated Sound Wood, and Untreated Decayed Temperature Curve,” American Fire Journal (May
Wood (Forest Products Research Society, 1949). 1998): 6–7.
16. Babrauskas, Ignition Handbook, 954. 39. Babrauskas, Ignition Handbook, 965.
17. Ibid., 947. 40. Retrieved April 10, 2005, from www.cnr.berkeley.edu/
18. NFPA, Fire Protection Handbook, 3–26, Table 3-3B. forestry/structures/DeckbdMatls.htm; see also S. L.
19. Drysdale, An Introduction to Fire Dynamics. Quarles, L. G. Cool, and F. C. Beall, “Performance of
20. Babrauskas, Ignition Handbook, 943. Deck Board Materials under Simulated Wildfire
21. L. Yudong and D. D. Drysdale, “Measurement of the Exposures,” paper presented at Seventh International
Ignition Temperature of Wood,” First Asian Fire Safety Conference on Woodfiber-Plastic Composites,
Science Symposium, Hefei, Anhui Province, People’s Madison, WI, May 19–20, 2003.
Republic of China, October 1992. 41. N. M. Stark, R. H. White, and C. M. Clemons, “Heat
22. Drysdale, An Introduction to Fire Dynamics. Release Rate of Wood-Plastic Composites,” SAMPE
23. E. L. Shaffer, “Smoldering Initiation in Cellulosics Journal 33, no. 5 (September–October 1997): 26–31.
under Prolonged Low-Level Heating,” Fire Technology 42. Retrieved April 10, 2005, from www.cnr.berkeley.edu;
23 (February 1980). see also Quarls, Cool, and Beall.
24. Chemical Manufacturers Association, cited in 43. “Insulation Analysis,” Fire Findings 8, no. 1 (Winter 2000).
Babrauskas, Ignition Handbook, 18. 44. S. H. Graf, “Ignition Temperatures of Various Papers,
25. P. C. Bowes, Self-Heating: Evaluating and Controlling Woods, and Fabrics,” Oregon State College Bulletin
the Hazards (Amsterdam: Elsevier, 1984). 26 (March 1949).
26. J. C. Martin and P. Margot, “Approche 45. Babrauskas, Ignition Handbook, 896.
Thermodynamique de la Recherche des Causes des 46. F. W. Mowrer, “Ignition Characteristics of Various
Incendies Inflammation du Bois I.” Kriminalistisk und Fire Indicators Subjected to Radiant Heat Fluxes”
forensische Wissenschaften 82 (1994): 33–50. in Proceedings Fire and Materials 2003 (London:
27. B. R. Cuzzillo and P. J. Pagni, “The Myth of Pyrophoric Interscience Communications), 81–92.
Carbon,” Fire Safety Science—Proceedings of the 47. J. G. Quintiere, Principles of Fire Behavior (Albany,
Sixth International Symposium (Bethesda, MD: IAFSS, NY: Delmar, 1998).
2000), 301–12; B. R. Cuzzillo and P. J. Pagni, “Low 48. F. L. Fire, “Plastics and Fire Investigations,” Fire and
Temperature Wood Ignition, Fire Findings 7, no. 2 Arson Investigator 36 (December 1985); C. J. Hilado,
(Spring 1999). Flammability Handbook for Plastics, 5th ed. (Lancaster,
28. V. Babrauskas, B. F. Gray, and M. L. Janssens, “Prudent PA: Technomics, 1998); Stauffer, “Sources of Interference
Practices for the Design and Installation of Heat- in Fire Debris Analysis.”
Producing Devices near Wood Materials,” Fire and 49. G. Damant, California Department of Consumer Affairs,
Materials 31 (2007): 125–35. Bureau of Home Furnishings, personal communication,
29. V. Babrauskas, “Ignition of Wood: A Review of the 1988.
State of the Art” in Proceedings Interflam 2001 50. Babrauskas, Ignition Handbook.
(London: Interscience Communications), 71–88. 51. D. Madrzykowski et al., “Fire Spread through a Room
30. Truck Insurance Exchange v. MagneTek, Inc., 360 with Polyurethane Foam Covered Walls” in Proceedings
F.3d 1206 (10th Cir. 2004). Interflam 2004 (London: Interscience Communications),
31. G. R. Lightsey and R. W. Henderson, “Theoretical 1127–38.
vs. Observed Flame Temperatures during Combustion 52. G. Cooke, “When Are Sandwich Panels Safe in a Fire?
of Wood Products,” Fire and Arson Investigator 36 Part 1,” Fire Engineers Journal 58, no. 195 (July 1998);
(December 1985). G. Cooke, “When Are Sandwich Panels Safe in a Fire?
32. V. Babrauskas, “Charring Rate of Wood as a Tool Part 2.” Fire Engineers Journal 58, no. 196 (September
for Fire Investigations” in Proceedings Interflam 1998).
2004 (London: Interscience Communications), 53. M. Shipp, K. Shaw, and P. Morgan, “Fire Behavior of
1155–76. Sandwich Panels Constructions” in Proceedings Interflam
33. J. D. DeHaan, “Compartment Fires 1996–98,” CAC 1999 ( London: Interscience Communications), 98.
News (Spring 2001). 54. “Panel Perspectives,” Fire Engineering Journal and
34. R. H. White and E. V. Nordheim, “Charring Rate of Fire Prevention (September 2005): 44–46.
Wood for ASTM E 119 Exposure,” Fire Technology 55. R. H. Holleyhead, “Ignition of Solid Materials and
28 (February 1992): 5–30. Furniture by Lighted Cigarettes,” Science and Justice,
35. C. P. Butler, “Notes on Charring Rates in Wood,” 39, no. 2 (1999): 75–102.
Fire Research Note No FR 896 (Fire Research Station, 56. W. D. Woolley and S. A. Ames, “The Explosion
Borehamwood, UK, 1971). Risk of Stored Foam Rubber,” Building Research
36. Babrauskas, “Charring Rate of Wood as a Tool for Establishment Current Paper CP 36/75, 1975. (See
Fire Investigations.” also: “Fatal Mattress Store Fire at Chatham
37. Ibid. Dockyard,” Fire 67 (1975), 388.)
Chapter 5 Combustion Properties of Solid Fuels 165
57. Retrieved April 10, 2005, from www.cnr.berkeley.edu; 67. P. E. Jackson, “Intumescent Materials,” Fire Engineers
see also Quarls, Cool, and Beall. Journal 58 (May 1998): 194.
58. J. Albers, “Cover Story,” California Fire/Arson 68. J. M. McGraw and F. W. Mowrer, “Flammability of
Investigator (July 2008): 25. Painted Gypsum Wallboard Subjected to Fire Heat
59. J. F. Krasny, W. J. Parker, and V. Babrauskas, Fire Fluxes.” Proceedings Interflam 1999 (London:
Behavior of Upholstered Furniture and Mattresses Interscience Communications), 1320–30.
(Norwich, NY: Noyes/William Andrew, 2001). 69. Babrauskas, Ignition Handbook, 870–74.
60. J. D. DeHaan and K. Bonarius, “Pyrolysis Products of 70. Ibid., 873.
Structure Fires,” Journal of Forensic Sciences 28 71. Ibid., 871.
(1988): 299-309. 72. D. J. Hodges, Colliery Guardian 207, no. 678
61. J. Albers, “It’s Time to Reduce Upholstered Furniture- (1963).
Driven Flashover,” American Fire Journal (July 1999): 73. Babrauskas, Ignition Handbook, 719–24.
6–8. 74. Ibid.
62. S. Nurbakhsh, California Bureau of Home Furnishings, 75. Ibid.
personal communication, 2006. 76. A. B. Spencer, “Dust: When a Nuisance Becomes
63. CPSC, 16 CFR 1633: Standard for the Flammability Deadly,” NFPA Journal (November–December 2008):
(Open Flame) of Mattress Sets. Federal Register 70, 57–61.
no. 50 (March 15, 2006). 77. R. G. Zalosh, “Explosion Protection.” SFPE
64. Holleyhead, “Ignition of Solid Materials and Furniture Handbook on Fire Protection Engineering, 2nd ed.
by Lighted Cigarettes.” (Boston: SFPE, 1995), 3–315.
65. J. D. DeHaan and S. Nurbakhsh, “The Combustion 78. P. L Kirk, Fire Investigation (New York: Wiley,
of Animal Carcasses and Its Implications for the 1969).
Consumption of Human Bodies in Fires,” Science 79. J. A. Conkling, The Chemistry of Pyrotechnics
and Justice (January 1999). and Explosives (New York: Marcel Dekker, 1985),
66. J. F Krasny, W. J., Parker, and V. Babrauskas, Fire chap. 7.
Behavior; S. Sardqvist, Initial Fires: RHR, Smoke 80. ASTM E 1354: Standard Test Method for Heat and
Production and CO Generation from Single Items and Visible Smoke Release Rates for Materials and Products
Room Fire Tests (Lund, Sweden: Lund University, Using an Oxygen Consumption Calorimeter (West
April 1993). Conshohocken, PA: ASTM, 1997).

166 Chapter 5 Combustion Properties of Solid Fuels

 CHAPTER
6
Sources of Ignition

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

KEY TERMS

arc, p. 174 rollout, p. 184 short-circuit, p. 179

piloted ignition, p. 185 self-ignition, p. 208 spark, p. 174
resistance, p. 169 service, p. 179

OBJECTIVES

After reading this chapter, you should be able to:

■ Describe common ignition sources and their typical heat release rates.
■ Recognize secondary sources of ignition.
■ Identify the role of appliances in starting fires.
■ Discuss the role of hot and burning fragments in kindling fires.
■ Explain the role of smoking as a fire origin.

For additional review materials, appendices, and suggested readings,

visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

167
 Introduction to Ignition Sources
Ignition can be defined as the initiation of self-sustaining combustion in a fuel. The mech-
anisms of ignition were explored in detail in Chapters 4 and 5. As we have seen, the fun-
damental properties that influence an object to ignitability are its density, thermal
capacity, and thermal conductivity. Taken together, these properties constitute the mater-
ial’s thermal inertia. Ignition requires transfer of enough energy quickly enough to over-
come the thermal inertia and trigger sufficient combustion to become self-sufficient. If
that occurs, the ignition source is said to be competent (for that fuel under those condi-
tions). As we have seen from previous chapters, ignition involves bringing at least a part
of the fuel to a characteristic temperature by means of conducted, convected, or radiant
heat transfer until it can sustain combustion. In gaseous fuels, this process involves rais-
ing the temperature of only a small volume to ignition. In solids, it usually involves the
surface of the fuel (the exception being self-heating, in which the heat is being generated
within the bulk of the fuel). No matter what the nature of the fuel, without some source
of ignition, that is, without some source of energy, there will be no fire. Without excep-
tion, the source is some type of hot object or mass, chemical reaction, flame, or electric
current. Except for the self-heating case, the temperature of the source must exceed the
ignition temperature of the fuel, and it must be able to transfer enough heat into a suit-
able mass of fuel before there can be ignition. In the case of self-heating, ignition will
cause self-sustaining smoldering combustion, in which oxygen from the surrounding air
diffuses into the surface of the charring fuel to create enough heat to advance the reac-
tion, but smoldering combustion can, of course, be started equally well by an external
heat source. Although ignition of smoldering combustion is sometimes of interest to the
fire investigator, for the most part the onset of flaming combustion is the event of most
concern. Flaming combustion of a solid material occurs when sufficient gases or vapors
are generated from the pyrolyzing fuel to support flames.
As a general rule, the lower the energy of the ignition source, the closer the source
and first fuel have to be for ignition to take place. The ignition source may be small and
inconspicuous compared with the destructive fire that follows. There may be intermedi-
ate stages from initial ignition to full involvement of the first fuel package and then spread
as other fuel packages in the room ignite, and this item-to-item growth must obey the
same physical laws. Material that is intentionally heated by electricity, friction, chemical
reaction, or a controlled flame may be dropped or blown into fuel that then ignites. These
intermediate occurrences must always be taken into consideration in dealing with the
source of ignition. The challenge to the fire investigator is to identify the fuel first ignited
and then determine how the ignition source managed to transfer enough heat to that fuel
for it to kindle into flame.
In this chapter we examine a wide variety of heat sources and their thermal proper-
ties. This text cannot discuss each source in combination with every possible fuel. The
fundamental processes of heat transfer, heat release rate, and fire propagation must be
applied to each possible situation. By applying the scientific method to ignition processes,
the investigator must evaluate likely sources.

Primary Ignition Sources

The primary source of ignition of virtually every fire is heat. As we have seen, the heat
can originate from a hot surface, a hot particle, a chemical reaction, or a distant radiant
heat source, but the small open flame of a common match or lighter is among the most
effective primary ignition source for fires. Without question, such sources are responsible
for a large percentage of both accidental as well as deliberate fires. Other sources include
mechanical sparks (incandescently hot or burning particles) or electric arcs, and the lat-
ter can range from very small to lightning bolt size. All these sources generate enough
168 Chapter 6 Sources of Ignition
localized heat to ignite many fuels, although not all sources will ignite all fuels. Heat may
also be generated by the passage of current through wires (particularly in an overload
condition), through a high resistance connection or through an unintended conductive resistance ■ Opposition
path. Heat is also generated by numerous chemical reactions of the exothermic type, some to the passage of an
electrical current.
very slow, some very fast. To clarify the relationship of various ignition sources to various
fuels, matches, lighters, torches, and candles will be considered briefly.

MATCHES
A very common means for initiating combustion, the match is specifically and exclusively
designed for this purpose, and it is the most basic source of flame. Whether it directly ini-
tiates the destructive fire or merely sets a controlled fire that is later responsible for the
larger fire is immaterial to the basic question of ignition but may be very material to the
investigation of the larger fire.
The match is a stick that is treated to be readily combustible combined with a head
that includes both fuel and oxidizer in a form that can be ignited by localized heat pro-
duced by friction. Matches are found in two general categories: the strike-anywhere
match and the safety match. The strike-anywhere or kitchen match contains an oxidant
such as potassium chlorate mixed with an oxidizable material such as sulfur or paraffin,
a binder of glue or rosin, and an inert filler like silica. The tip, which is more easily
ignitable by friction, contains a high percentage of both phosphorus sesquisulfide (P4S3)
and ground glass. The head of the safety match contains an oxidizer and a fuel such as
sulfur. It will ignite only when struck against the strip containing red phosphorus, glue,
and an abrasive like ground glass. The stick of the match is often treated with a chemical
to suppress afterglow. Paper matches are generally impregnated with paraffin to improve
both water resistance and the burning characteristics of the head.
The heat output of a large kitchen match is on the order of 50 to 80 W (see Table 6-1),
and the average flame temperature (once the incendiary tip has burned away) is on the
order of 700°C to 900°C (1,300°F to 1,650°F). Because the match flame is laminar and
involves wood (or cardboard) and wax, temperatures within the thin combustion zone
near the outside of the plume are similar to those of a laminar candle flame and can be
as high as 1,200°C to 1,400°C (2,200°F to 2,500°F).1 Such energy and temperature are
adequate for ignition of many fuels as long as that energy can be transferred into the fuel.
Such a low heat output is not enough to ignite most fuels by radiant heat, so there is usu-
ally direct contact between the flame and fuel before ignition. This is true of ordinary
combustible solids, but combustible gases or vapors can be drawn into the flame by
entrainment flow, so there may appear to be ignition without direct contact.

U.S. and Foreign-Manufactured Matches

Minor variations in match formulation and manufacturing detail exist among U.S. manu-
facturers, although some matches are custom-made with special head or shaft designs for
promotional purposes. Large differences in appearance (and sometimes in action) are
expected only with matches of foreign origin. For instance, some foreign-made matches
have been known to ignite by frictional contact between their heads.

Considerations for Fire Investigators

It is common to attempt to compare matches from a scene with suspected sources like
partial matchbooks on the basis of their length, width, color, and the paper content of the
cardboard stems. On some occasions, the torn bases of the paper matches can be physi-
cally compared, jigsaw fashion, to the stubs in the matchbook. Most U.S. book matches
today, however, are pre-perforated, which makes them tear uniformly and thus reduces
the chances for success in such comparisons. It is sometimes possible to trace the origin
of partially consumed matches at a fire scene by conducting an elemental analysis on the
unburned match heads.2
Chapter 6 Sources of Ignition 169
 TABLE 6-1 Heat Release Rates and Burn Times of Common Ignition Sources

MAXIMUM
TYPICAL FLAME HEIGHT MAXIMUM
TYPICAL BURN FLAME HEAT FLUX
OUTPUT (kW) TIME (s)a (mm) (kW/m2)
Cigarette 1.1 g (not puffed, laid on solid 0.005 1,200 — 42
surface) bone dry
Cigarette 1.1 g (not puffed, laid on solid 0.005 1,200 — 35
surface) conditioned to 50% R.H.
Methenamine pill, 0.15 g 0.045 90 4
*
Candle (21 mm, wax) 0.075 — 42 70b
Wood cribs, BS 5852 Part 2
No. 4 crib, 8.5 g 1 190 15c
No. 5 crib, 17 g 1.9 200 17c
No. 6 crib, 60 g 2.6 190 20c
No. 7 crib, 126 g 6.4 350 25c
(Crumpled) brown lunch bag, 6 g 1.2 80
(Crumpled) wax paper, 4.5 g (tight) 1.8 25
Newspaper—folded double sheet, 4 100
22 g (bottom ignition)
(Crumpled) wax paper, 4.5 g (loose) 5.3 20
Newspaper —crumpled double sheet, 7.4 40
22 g (top ignition)
Newspaper double sheet newspaper, 17 20
22 g ( bottom ignition)
Polyethylene wastebasket, 285 g, filled 50 200d 550 35e
with 12 milk cartons (390 g)
Plastic trash bags, filled with cellulosic 120–350 200d
trash (1.2–14 kg)f
Small upholstered chair 150–250 — — —
Upholstered (modern foam) easy chair 350–750 — — —
Recliner (PU foam, synthetic upholstery) 500–1000 — — —
Gasoline pool on concrete (2 L) 1000 30–60 — —
(1/m2 area)
Sofa 1000–3000 — — —

Sources: V. Babrauskas and J. Krasny, Fire Behavior of Upholstered Furniture, NBS Monograph 173 (Gaithersburg, MD: U.S. Department
of Commerce, National Bureau of Standards, 1985).
*From S. E. Dillon and A. Hamins, “Ignition Propensity and Heat Flux Profiles of Candle Flames for Fire Investigation” in Proceedings: Fire and
Materials 2003 (London: Interscience Communications), 363–76.
a
Time duration of significant flaming.
b
Centerline—immediately above flame; ⬍4 kW/m2 outside.
c
Measured from 25 mm away.
d
Total burn time in excess of 1800 s.
e
As measured on simulation burner.
f
Results vary greatly with packing density.

170 Chapter 6 Sources of Ignition

LIGHTERS
Basically, there are two general types of lighters: the electric-element type found in vehi-
cles, and liquid-fuel lighters. For most fire purposes, the electric lighter may be disre-
garded as a source of ignition because it depends on the battery of a vehicle to heat the
wire heating element. It is not operative except with its electrical connections to a battery,
which greatly restricts its use.
The basic liquid-fuel lighter ignites the flammable vapors from a fuel-soaked wick or
liquid reservoir by means of a spark produced when a steel strikes a flint. Cigarette lighter
“flints” are made of “misch metal,” which includes a rare earth such as cesium, lan-
thanum, or neodymium whose sparking potential is very high.3 Thus, a spark or hot par-
ticle ignites a fuel vapor under controlled conditions, leading to a small, predictable flame.
Such lighters are a convenient substitute for a match and are, no doubt, responsible for
many incendiary fires. They are rarely left behind and are therefore unlikely to be recov-
ered as evidence. When lighters are recovered, the possible presence of latent fingerprints
(especially on a removable fuel reservoir inside the case) and trace evidence should be con-
sidered. Some lighters ignite the vapors by catalytic oxidation or piezoelectric effect on a
crystal, but since the flame produced is the same as from a “regular” flint lighter, their use
as an ignition source remains functionally the same. The heat release from a wick-type
lighter is on the same order as that of a large kitchen match (50–80 W), and maximum
temperatures are similar to those reported for pool fires of hydrocarbon fuels, ≈1,000°C
(1,830°F).4
Traditional lighters that the user refueled with a light liquid petroleum product
have largely been supplanted by disposable lighters filled with butane. In the latter, the
butane fuel compartment is under pressure, and the fuel is delivered through a meter-
ing valve to be ignited (in the manner described earlier) as a small, laminar jet flame.
The adiabatic temperatures of such flames have been reported as 1,895°C (3,440°F), so
real-world maximum temperatures might be expected on the order of 1,400°C
(2,500°F). Research at the University of Maryland on cigarette lighters with port-type
and Bunsen-type burner designs (adjusted to produce 75-W flames) reported tempera-
tures reaching 2,022 K (1,749°C; 3,180°F) and peak heat fluxes as high as 169 kW/m2
(25 mm from tip axially).5
Tests of a typical disposable butane lighter indicated a fuel release rate of 0.001 to
0.002 g/s and a flame less than 20 mm (0.8 in.) in height. Based on a ΔHc of 46 kJ/g, this
fuel consumption would produce an estimated 40 to 90 W.6 Such lighters can explode if
exposed to high temperatures and are subject to leakage when dropped or exposed to
reduced atmospheric pressure such as in airplane cabins. The metering valve can some-
times be adjusted to release a fairly voluminous stream of flammable gas, and the result-
ing flame can be several inches high (with corresponding increase in heat release rate).
Because these lighters are nonrefillable and intended to be disposable, they are more likely
to be left at the point of ignition of a fire. Innumerable specialty lighters such as barbe-
cue or fireplace lighters with an extended delivery tube are also available, as well as nov-
elty lighters disguised as guns, pens, and toys.

TORCHES
Torches of a wide variety of types and fuels find uses in many industrial and residential
applications. There are “traditional” blowtorches using manually pressurized “white gas”
delivery. There are “drip torches” using kerosene or diesel fuel for setting controlled “back
fires” in wildlands. Handheld or long-wand roofing torches use propane, butane or MAPP
[methyl acetylene (H3C ¬ C ‚ CH), and propadiene (H2C “ C “ CH2)] as fuels.
Acetylene/oxygen and acetylene/air handheld torches are very well known. Certain paint-
stripping devices (heat guns) that are either propane-fueled or electrically powered can pro-
duce very high gas temperatures and high heat fluxes. Examples are shown in Figure 6-1.
Chapter 6 Sources of Ignition 171
FIGURE 6-1 Common
hand torches. From left:
propane, MAPP (with
igniter), blow torch (white
gas), butane microtorch.
Foreground: acetylene/air
hobby torch and oxy-
acetylene welding torch.
Courtesy of John D. DeHaan.

Fire Findings reported that flame temperatures in a propane hand torch ranged from
332°C (629°F ) at the tip to 1,333°C (2,071°F ) at the flame center to 1,200°C–1,350°C
(2,200°F⫺2,468°F ) at the tip of the “inner flame” cone.7 Any of these propane torches
can cause accidental (and intentional) fires if they are applied to suitable fuels. Cellulosic
materials are most susceptible to ignition by torch application especially if they are finely
divided, since brief application of such a flame to a solid wood surface will cause only
surface scorching and (possibly) brief glowing combustion that ceases as soon as the flame
is withdrawn. Finely divided materials (dried leaves, needles, sawdusts) are very suscepti-
ble to ignition, usually in two stages: smoldering followed by flame. More massive mate-
rials are susceptible when the torch flame is played into narrow crevices or spaces, where
radiant losses from surface ignition are minimized. In these cases, ignition can be caused
within the cavity that may not be readily detectable or extinguishable from the outside.
Such crevices can also retain sawdust and detritus from termite attack, making ignition
even more likely. In hot torch work on roofs, torches are used to cause tar to adhere to
roof structures or liquefy the asphaltic adhesive on the flexible membrane as it is laid.
Under either condition, smoldering ignition triggered in crevices or in debris is then con-
cealed under the hot tar or membrane. Such roof fires following several hours of con-
cealed smoldering have been identified.
A special variety of small devices is a new generation of microtorches advertised as
being intended for welding and brazing operations but that are primarily sold by smoke
shops. These use butane as a fuel and premix it in a special torch tip that can produce
extremely hot flames. (See Figure 6-1.) These microtorches are small enough to be concealed
in the hand and can be ignited and left burning. Due to their ostensible application as
welding devices, such microtorches normally lack any child-resistant features. The maxi-
mum temperatures measured in the combustion zones of such a microtorch should be the
same as for a full-size one (1,200°C–1,350°C; 2,200°F–2,470°F). Such torches are often
associated with inhalation of drugs of abuse such as crack cocaine. See Chapter 14 for
further discussion of their role in arson.
172 Chapter 6 Sources of Ignition
CANDLES
Once only a consideration of fire causation during power failures, candles have become
extremely popular in many different forms and functions. The National Candle Association
estimates that candles are used in 7 of 10 American homes and that candle sales increase by
15 percent each year. The USFA estimates that candles are responsible for more than 9,400
residential fires each year. The CPSC estimated an even higher number—12,800 fires causing
170 deaths and 1,200 injuries.8 The NFPA estimated that in 2002 some 18,000 reported
home fires were started by candles, causing 130 deaths, 1,350 injuries, and property losses
of $333 million in the United States alone.9 A traditional paraffin or beeswax candle pro-
duces about 50 to 80 W of heat with a laminar flame with an average flame temperature of
800°C to 900°C (1,470F–1,650°F), but like a match flame, a candle flame has an outer com-
bustion zone that can have temperatures as high as 1,200°C to 1,400°C (2,200°F–2,550°F),
as we saw in Chapter 2. Because of the candle’s steady, prolonged flame and high tempera-
ture zone, it is possible to melt thin copper or even iron wire in a candle flame if it is held
steadily in the very hot combustion zone, but the thermal conductivity of such metals con-
ducts the heat away too quickly to permit the melting of larger-diameter wire.
The structure and dynamics of candle flames were discussed in detail in Chapter 3.
Dillon and Hamins reported that a 21-mm (0.8-in.)-diameter candle had a mass loss rate
of 0.09 to 0.11 g/min. With a net ΔHc of 43 kJ/g, this produced a flame 40 mm (1.6 in.)
high, with an HRR of ~65 to 75 kW. They measured the total heat flux along the vertical
centerline and showed a maximum of about 70 kW/m2 at the tip of the flame, dropping to
~40 kW/m2 at 50 mm (2 in.) above the tip and ~20 kW/m2 at 170 mm (6.7 in.) above the
tip.10 Just 13 mm (0.6 in.) from the centerline, the total heat flux at the level of the flame
tip was about 27 kW/m2, showing the significant amount of heat in the buoyant plume.
These results confirm direct observations of the propensity for a candle flame to ignite fuels
at some distance immediately above the flame but not target fuels at the side of the flame.
A well-designed candle melts at just the right rate to maintain a steady supply of wax
for the wick to deliver to the flame: too little wax and the candle extinguishes for lack of
fuel (as the wick burns up); too much wax and the wick gutters and drowns.11 Ideally the wax
burns completely to carbon dioxide (CO2) and water. The combustion of carbon soot occur-
ring in the outer high temperature zone is essentially complete; virtually no soot escapes the
laminar flame. If the wick is too long and the flame becomes turbulent, soot can escape. If
the wax is contaminated with oils or other organic materials, these may not burn com-
pletely, and a large quantity of soot can be released. If the wick is not symmetrically placed
in the candle, the candle can fail and slump, exposing large lengths of wick. The heat release
rate (and therefore the flame length) of a candle is determined by the amount of exposed
wick. If charred matchsticks are dropped into a candle, they can act as supplemental wicks,
dramatically increasing the size of the flame. The larger flame can impinge on adjacent fuels
or heat the container in which the candle is placed (such as the metal cup of a small chaf-
ing dish or tea light candle), creating increased risk of accidental fires.12 Candles cause fires
when combustibles come into contact with the flame: loose material can move into the
flame, the candle can tip over while lighted (especially with taper or column-style candles as
opposed to pillar or votive styles), or the candle fails as the wick comes loose or otherwise
becomes more exposed. This last situation causes a larger flame. Ignition can also occur
when combustibles are included in the candle as decorations. Candles with aromatic oils are
being used in large arrays for aromatherapy and then left to burn while the occupant goes
to sleep. Large decorative candles with multiple wicks and burning times of several days
mean prolonged risks of ignition of draperies, furniture, and clothing.13

Secondary Ignition Sources

For purposes of this chapter, secondary ignition sources include sparks and arcs, hot
objects and surfaces, friction, radiant heat, and chemical reaction.
Chapter 6 Sources of Ignition 173
 SPARKS/ARCS
spark ■ Superheated, The definition of spark is ambiguous because the term can refer to one of two situations:
incandescent particle.
■ an electric arc of brief duration where electric current is discharging through air or
arc ■ Flow of current another insulator
across an air gap (or ■ a tiny fragment of burning or glowing solid material moving through the air
another nonconductor)
between two The electric spark is not readily distinguished from the electric arc, except for dura-
conductors. tion. Babrauskas has suggested that an arc represents a stable electrical flow, whereas an
electric spark represents an initial flow of electric current across a gap between conductors.
The arc persists as a discharge for some considerable time interval, while an electric spark
is thought of as being virtually instantaneous. It is therefore simpler to regard all such elec-
trical phenomena as arcs of various durations and leave the term spark to represent a solid
particle or molten droplet heated by some process to incandescence. The longer an arc per-
sists, the more time it has to heat its immediate surroundings and to transfer heat to sur-
rounding fuel. The energy that can be released from an electric arc can range from
millijoules to millions of joules. Because the arc can persist anywhere from microseconds
to hundreds of seconds, the total heat released can be within a very broad range—from a
tiny, brief arc of static electricity on a household doorknob to a massive lightning strike.
The capacity of arcs to ignite fuels, then, depends greatly on their duration, current flow,
and the susceptibility (physical and chemical properties) of nearby fuels. A full discussion
of electrical sources of ignition including arcs is found in Chapter 10.
The latter conception of a spark, that is, a tiny fragment of burning or glowing mate-
rial moving through the air, has become the standard usage and will be used here. Sparks,
then, can be produced by a sharp blow between two dissimilar materials such as steel and
flint, by extreme frictional contact between two moving materials, by extreme heating
and melting of conductors in electrical failures, or by the airborne debris of a solid-fuel
fire. Such incandescently hot particles are capable of igniting fires in vapors and some
solids and will be treated later in this chapter.
Mechanical sparks include burning brands or very hot embers, fragments of debris
from an existing fire, droplets of metal from extremely hot combustion or electric arcing
events, incandescent or burning metal from impact or mechanical friction, or soot/partic-
ulates from engines of many types. These will be considered as a special case of “hot
objects” in the following section.

HOT OBJECTS/HOT SURFACES

Most hot objects are heated either by being in or close to a flame, by frictional heating,
or by the flow of electric current through them. In electric appliances, ceramic or wire
heating elements may be present that are designed to carry current and generate heat as
a result. If combustible materials come into contact with these heating elements, they can
cause fires by igniting and then spreading the fire to other fuels nearby. Such appliances,
then, are not primary sources of ignition but lead to loss of control of a designed heat
source. Malfunctions that allow a flow of electricity through conductive paths for which
they were not intended or at rates that exceed the design of intended conductors can also
lead to the production of hot surfaces. Hot objects play an important role in transmitting
the heat of various chemical and physical processes to fuels, thereby spreading a fire. As
we saw in Table 3-1, the visible light emitted by the surface of a hot object can give an
indication of its temperature. The capacity of this object to ignite a fire depends not only
on its initial temperature but also on its mass and thermal capacity, which together deter-
mine how much heat this mass contains.
The evaluation of a possible ignition of a fuel by hot object or surface is not a simple
matter of noting the temperature of the surface or mass or comparing it with the listed
AIT of the fuel. Ignition of any fuel will not occur unless enough heat is transferred into
174 Chapter 6 Sources of Ignition
a sufficient mass of fuel to establish a persistent flame. The transfer of that heat from a
surface depends on the nature and contour of the surface (even its roughness and cleanli-
ness), the nature of contact, and whether the fuel can maintain the contact long enough.
The AIT of a liquid or gaseous fuel, for example, varies with the test method used, the
volume and shape of the test cell, and the material used for the test cell.14 For example,
clean, polished metal hot surfaces produce ignitions at lower temperatures than do simi-
lar metals that have an oxide layer. Very short contact or residence times do not allow
enough heat to be transferred. A volatile liquid fuel dripped onto a flat, hot metal surface
is likely to cool the immediate contact area by evaporation; the resulting vapors rise by
convection away from the hot surface, thus reducing the residence time. The temperature
of the surface, then, would have to be hundreds of degrees hotter than the AIT for that
fuel for there to be ignition, and even then it would not be a guaranteed result.
Interestingly, in general for liquid fuels, the lower the flash point of the liquid, the
higher the temperature of the hot surface necessary to achieve ignition.15 Hot-surface
ignition of liquid or gaseous fuels is achieved at lower temperatures if they are confined
(in a tube or other enclosure). Other variables include surface type, droplet size, and air-
flows near the surface. When hot particles can penetrate or burrow into a solid fuel, res-
idence time is no longer a factor, but the total amount of heat that the mass can carry is
critical. If the thermal energy of the particle is too low (insufficient mass or density, inad-
equate thermal capacity), there may not be ignition even if the object is buried in the fuel
and transfers all its heat. Hot objects that land on a rigid surface are even less likely to
ignite a solid fuel due to the extra convective and radiative losses to the surrounding air.
It is tempting to measure the temperatures of hot surfaces as they occur in open air
and then compare those temperatures against ignition temperatures of potential fuels (as
has been reported in various recent publications, such as Fire Findings), but such temper-
atures are only a guide to the processes taking place (see later section in this chapter.) If
the temperature on the surface of a glass lightbulb in the open is measured at, say, 200°C
(400°F), that value reflects only the steady-state temperature of that surface achieved by
the balance between the heat flux (input) striking the inside of that bulb from the hot fil-
ament and the heat losses from convective, radiative, and even conductive processes on
the outside of the bulb. If the rate at which the heat is lost is changed by placing the bulb
in contact with a surface or by burying it in an insulating material, the temperature of that
surface will rise, often dramatically. The new temperature then poses a very different
“source” for transfer of heat to a potential fuel. In short, burying even a low-wattage heat
source in a deep layer of cellulose insulation poses an ignition risk even when the open-
air temperature of that heat source would make it seem “safe.”
As a reference point, however, knowledge of open-air temperatures of some hot sur-
faces is useful. The surface temperatures of halogen work-light lamps (bulbs) has been
reported to be 593°C to 685°C (1,100°F to 1,266°F) (sufficient to ignite flammable
vapor/air mixtures as well as solids), and the glass face of a 300-W halogen work light to
range from 180°C to 215°C (357°F to 418°F).16 The open-air surface temperatures of
heating elements of electric range tops ranged from 370°C (700°F) for small burners to
590°C (1,100°F) for large burners in older appliances, and 500°C to 730°C (925°F to
1,350°F) for equivalent burners in new ranges.17 Most modern electric ranges have a
great deal of power in each cooking element, as well as high temperature. Goodson and
Hardin reported wattage ratings for various electric range heating elements ranges from
1,200 to 2,500 W.18
As we have seen, with complex fuels such as cellulosics, the increase in temperature
may bring about physical and chemical changes in the fuel that make it more susceptible
to ignition than the original fuel. The heat transfer rate to the fuel surface is probably
more useful than temperature as an indicator of ignitability. A table of minimum critical
radiant heat fluxes is found in Appendix K. The chemical and physical nature of the fuel
has to be considered, and the effects of radiant heat have to be tested as part of the
Chapter 6 Sources of Ignition 175
 hypothesis testing of the scientific method. The higher the radiant flux applied to a fuel,
the lower the observed ignition temperature will usually be, dramatically decreasing the
time to ignition. If the fuel is a thermoplastic, it may melt, sag away from the ignition
source, or form a nonporous mass and become less likely to be ignited. Materials that
char upon initial heating may become more susceptible with initial heating if the charring
produces a porous residue that will support a smoldering combustion. Charring may also
produce a pyrolyzed fuel that requires more energy to ignite.

FRICTION
As a source of ignition, friction is a special case of a “hot object.” Friction between two
moving surfaces generates heat (as in the disc brakes of an automobile, which can become
extremely hot). As a source of fire, rubbing two sticks together is little used on the mod-
ern scene (except traditionally by Boy Scouts). Even they are more likely to supply a bow
to spin a wooden point in a wooden depression because of the much greater speed and
frictional heat that are generated. Wood is used, not primarily because it is a fuel, but
because it is a very poor conductor of heat. It allows heat to be generated by friction faster
than it is carried away and dissipated. With enough effort, the local heat may rise to the
point of kindling a glow in fine tinder, which can then be aroused into a flame by fanning
or blowing. Although this method of igniting fires has historical interest, it is of little or
no practical consequence today.
However, friction in other contexts may be of great importance in igniting fires, espe-
cially in machinery. An overheated wheel bearing or “hot box” on older railroad cars—
frictionally heated because of inadequate lubrication—may cause discharge of hot metal
fragments or cause adjacent fuel material (e.g., cotton-waste oil wick) to catch on fire.
Any bearing that does not have adequate lubrication can become hot through friction,
and contact of the hot object with a readily ignited fuel can lead to a fire. This lack of
lubrication is without a doubt one of the relatively common sources of fire where machin-
ery is in use. Conveyor or power transmission belts can jam or be forced to run against
frozen rollers and cause extreme frictional heating. Once ignited, the moving belt can
spread fire very rapidly.
If frictional heat becomes extreme, one or both of the surfaces may disintegrate, pro-
ducing a shower of incandescent particles. Such disintegration may be the end result of
massive overheating or very localized surface heating where the temperature of the mass
itself may not rise appreciably. The temperatures of hot particles generated by this process
are limited by the melting temperatures of the material involved. For instance, a mechan-
ical spark generated from a copper-nickel alloy surface may have a maximum tempera-
ture of 300°C (570°F) (and not be incandescent at all), while a fragment of tool steel may
be at 1,400°C (2,550°F) and be bright sparkling white.19 This situation can occur with
any moving machinery. Bearings in conveyor systems, for instance, can clog with abrasive
dust or seize from lack of lubrication and create large numbers of incandescent sparks for
long periods of time. Bearings in automotive turbochargers can fail in seconds if lubrica-
tion is interrupted, due to their very high rotational speeds (up to 25,000 rpm).
Not only bearings give rise to dangerous temperatures from excessive friction. High-
speed rotors, such as those found in airplane motors, that come into contact with the hous-
ing have been known to produce enough heat locally to melt the housing and ignite related
hardware. Fragments of overheated brake shoes in malfunctioning train or truck brakes
can be thrown from the moving vehicle for miles before being detected. Mufflers, tailpipes,
tailgates, chains, and other equipment dragging on the road surface can generate substan-
tial showers of incandescent metal particles with enough heat to ignite roadside fuel (dried
grass, leaves, or litter). Machining operations (milling machines, lathes) can generate show-
ers of hot particles that can ignite waste in or under the tool. Mechanical sparks (hot frag-
ments) are discussed in more detail in a later section of this chapter.
176 Chapter 6 Sources of Ignition
RADIANT HEAT
Although the heat radiated from a fire may at times ignite other fires at a distance, such
ignition is not primary but secondary to an already large, established fire. Radiant heat
plays a great role in spreading most fires and was examined in Chapter 3. As a primary
ignition source, radiant heat alone is much less frequently encountered. Radiant heat
from fireplaces, stoves, and heaters has been seen to bring nearby cellulosics to their igni-
tion temperatures and thus start fires. When such devices are permanent fixtures, their
remains will usually still be detectable after the fire. Other sources may not be so easily
identifiable. The investigator must remember that direct physical contact between the
heat source and the first fuel is not always necessary when sufficient radiant heat can
be brought to bear on the fuel surface. Here the reflective and absorptive qualities of the
fuel are critical, as are its density and thermal conductivity. All that is required is for the
target fuel to absorb more heat than it can dissipate and for that heat to raise the local
(surface) temperature above the fuel’s spontaneous ignition temperature. Rays from
direct sunlight (at a typical heat flux of 1 kW/m2) are not intense enough to ignite com-
mon fuels, but if they are concentrated or focused by a transparent object that is round
in cross section (i.e., spherical or cylindrical) or by a concave reflective surface (such as
a shaving mirror or the polished metal bottom of some aerosol cans), they can reach 10
to 20 kW/m2 at the focal point of the light path. If a cellulosic fuel is located at or near
that focal point, the fuel can be heated to its ignition temperature and catch fire. Two
examples of ignition by sunlight focused by shaving mirrors are shown in Figures 6-2
and 6-3. This mechanism is often blamed for starting wildland fires, but experiments by
DeHaan revealed that the conditions have to be just right: the object has to produce a
sharply focused area of sunlight, and the dry fuel has to be located precisely at the focal
point of that lens or mirror before ignition will take place. Empty glass bottles, broken
glass, or flat reflectors do not have a measurable focal length and will not cause ignition
by this means.
The maximum heat fluxes of some common small flaming ignition sources (at a dis-
tance of 25 mm; 1 in.) were shown in Table 6-1. Some tools and appliances, even when
operating as designed, can act as nonflaming ignition sources. Radiant heat from floodlights

FIGURE 6-2 Sunlight

focused on cloth target
by shaving mirror can
reach heat flux sufficient
to ignite fabrics. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

Chapter 6 Sources of Ignition 177

FIGURE 6-3A Shaving mirror (foreground) focused sunlight onto FIGURE 6-3B Shaving mirror (on left) focused sunlight from window
corner of box above it in parked car. Courtesy of Scott R. Schuett, onto towels, which ignited on rack. Courtesy of Curt Hawk, Fire Cause
Lebanon, CT. Analysis.

(particularly quartz halogen), projection lamps, heat lamps, and flameless torches can be
enough to ignite cellulosic fuels at close range.20 A noncombustible covering (sheet metal,
tile, asbestos, or plasterboard, for example) can allow heat to be transferred to a fuel sur-
face beneath while impeding the loss of heat by convection, thus making it susceptible to
ignition even when concealed. Hot surfaces associated with vehicles such as brake rotors,
exhaust manifolds, and catalytic converters have the potential to ignite fuels by both radi-
ant and conducted heat and will be discussed in Chapter 9.
Radiant heat is also linked to the ignition of wood even when the heat flux is not
enough to create ignition temperatures directly. In installations of heaters, furnaces, boil-
ers, and even chimneys and vent flues, radiant heat at low levels may induce degradation
of the wood to charcoal. Charring of wood can occur when temperatures on the order of
105°C (230°F) or greater are maintained for very long periods of time. This mechanism
was discussed in Chapter 5 and will be discussed in greater detail in a later section deal-
ing with furnaces.

CHEMICAL REACTION
A number of chemical mixtures are capable of great heat generation and even formation
of flame. They are only of occasional consequence in ordinary investigation of fires. Some
accidental fires in warehouse-style home-supply stores have recently been linked to leaks
or spills of corrosives (acids or bases) that came into contact with metals, or strong oxi-
dizers (swimming pool chlorines) that came into contact with fuels such as automobile
brake fluids and underwent an exothermic reaction. They play a role in a limited number
of incendiary fires because of the lack of knowledge about them on the part of the arson-
ists. Such reactions are more likely to be encountered in industrial or clandestine drug lab
fires than elsewhere, and they require the attention of the chemical specialist. Salvage and
recycling operations may produce mixtures of chemicals that spontaneously combust on
contact, producing dangerous, large fires.21 Many chemical fires are due to ignition or
decomposition of hazardous chemicals. The properties of the more common hazardous
materials are examined in detail in Chapter 13.
There are, of course, biochemical reactions that generate heat—drying oils, decom-
posing cellulosic and organic materials, and the like—that can contribute to ignition.
They are discussed later in this chapter under the heading “Spontaneous Combustion.”
178 Chapter 6 Sources of Ignition
The Role of Services and Appliances
as Ignition Sources
So many fires are started either by service facilities or appliances, or are attributed to service ■ The conduc-
these origins, that consideration of such possible origins is of great importance. When the tors and equipment for
delivering electricity
origin of a fire is not determined, it is very convenient to decide that it was due to some
from the supply system
malfunction of an appliance, a short-circuit in electrical wiring, or other similar cause. to the equipment of
This decision is especially attractive because, in many fires, there is so much damage to the premises served.
wiring, or melting of appliances, that such an opinion can have some credibility. In addi-
tion, it is indisputable that malfunctions or failures of services and appliances do cause short-circuit ■ Direct
contact between a
many fires. Thus, investigators may find themselves on the horns of a dilemma in the current-carrying
attempt to separate the real from the claimed origins of fires at such times. In modern conductor and another
building construction, electrical wiring is widely distributed as an integral part of nearly conductor.
all parts of the structure, and gas lines and appliances are also very common. In all such
structures, it is the rare fire indeed that does not engulf some such items to some extent
at least, thus possibly focusing attention on them.
The investigator must remember that the mere presence of an electrical wire or a gas
line does not constitute proof of responsibility for igniting a fire. If an electrical wire is
not energized (that is, no voltage applied to it), it cannot provide energy to ignite any fuel.
Even if it is energized, there has to be a mechanism, a reason, for it to create heat in an
unintended manner. It is the investigator’s responsibility to establish this mechanism and
demonstrate how it can create enough heat to ignite the first fuel. Gas lines, even when
charged or pressurized with gas, do not constitute an ignition source (except under the
most exceptional circumstances when they can conduct misdirected electrical energy) but
are merely a source of fuel.

GAS APPLIANCES AS IGNITION SOURCES

Gas appliances used for nonindustrial purposes are relatively limited to a variety of heat-
ing devices: stoves, ranges, room heaters, furnaces, clothes dryers, and water heaters. All
these items have open flame burners that create heat. Thus, regardless of the exact type
of heating device, the investigations, as well as all the possible hazards that attend the use
of the device, are very much the same. The heat output rating of the appliance, whether
the flame was correctly adjusted, how the flame could come into contact with com-
bustibles, and whether the automatic control mechanisms are functioning, are the features
that are of interest to the fire investigator.
One of the few ways in which the gas furnace or dryer in normal operation is likely
to initiate a fire results from the simultaneous failure of the thermostat and the high-level
or “high limit” control. The high-level device is a second thermostat or thermal cutoff
(TCO) located in the circulating-air ducting of a furnace, for example. It is usually cou-
pled in series with the regular thermostat and is adjusted to break the circuit (and close a
valve on the gas delivery line) at a maximum temperature above which the furnace can-
not be safely operated. This temperature varies but is usually above 90°C (200°F). A
simultaneous failure of both thermostats allows the furnace to continue operating indef-
initely with great production of heat and possibly a resulting fire.
The failure of pressure regulators (as mentioned in Chapter 4) can cause an appliance
to become an ignition source. Although such regulators may or may not be part of the
appliance, their failure can cause masses of flame to escape from the burner compartment
of the stove, furnace, or water heater affected. These flames can then ignite nearby com-
bustibles by radiant heat or by direct flame impingement. Even under normal operation,
the burners create considerable heat. Some appliances may not have adequate standoff or
may be installed improperly, thereby exposing combustible walls or floors to radiant or
conducted heat (as illustrated in Figures 6-4a–c). Inadequate air supplies can result in
Chapter 6 Sources of Ignition 179
FIGURE 6-4A Fire in
water heater closet as
found (unit less than
1 year old). Courtesy of
Battalion Chief Kurt Hubele,
Richland, (WA) FD.

FIGURE 6-4B After

cleanup—no floor left in
closet. Courtesy of Battalion
Chief Kurt Hubele, Richland,
(WA) FD.

FIGURE 6-4C Exemplar

heater in next apartment
showed heater installed
on plywood floor with no
stand-off legs. Courtesy of
Battalion Chief Kurt Hubele,
Richland (WA) FD.

180 Chapter 6 Sources of Ignition

excessive soot formation in the combustion chamber. This hot soot can then induce smol-
dering ignition of nearby combustible surfaces.
The “standing” pilot flame that was once standard on gas water heaters (and furnaces,
ranges, and ovens) and offered a small but continuous and potentially competent ignition
source, has largely been replaced by electric igniters. These igniters may be in the form of
a high-voltage arc, an electric spark created by capacitive discharge or piezoelectric effect,
or by a glow bar (electrically heated hot surface). The investigator should examine all gas
appliances carefully before concluding their status as an ignition source.22
Although gas appliances do start fires, they are probably less often the cause of a fire
than is generally assumed. Gas is the chief fuel in many places, and its intrinsic hazards
tend to encourage an abnormal fear of it as a source of fire. Gas can serve as a convenient
“probable cause” when the “poor” investigator finds no other that is provable; this
approach is not recommended. Occasionally, gas is released as a result of a failure within
the electrical system of the appliance and then ignited as it is released.23 The investigation
of any fire involving fuel gases should include examination and documentation of the con-
dition of the fuel supply, the ventilation available to the flame, electrical control system(s),
and the system for removal of products of combustion (vents, flues, and chimneys).
A common residential gas water heater (as diagrammed in Figure 6-5) produces a fire
of 24,000 to 36,000 Btu/hr (6.66 to 10 Btu/s, 7 to 10.5 kW). A steel baffle is inserted in
the central flue to slow the exhaust gases and give them more time to transfer heat to the
water. There is typically a draft hood at the top of the heater to entrain room air into the
exhaust gases and cool them. The vent is usually single-walled metal except where it
passes through walls or ceilings, where it is at least double-walled, as shown in Figure 6-6.
The flue-gas temperature at the top of the tank for steady-state operation of a 30,000
Btu/hr residential water heater is on the order of 300°C to 315°C (569°F to 601°F) with

Vent stack (single-wall)

Water in
Draft hood w/standoffs
Hot water outlet Typical Temperatures during Normal
(Steady-State) Operation
Center of flue (top of tank)
With baffle 300ºC–315ºC (569ºF–600ºF)
Without baffle 565ºC (1,050ºF)
Baffle Exterior of vent connector 73ºC (164ºF)
Top of 4-ft vent stack (gas temp.) 183ºC–205ºC
Fiberglass
(362ºF–400ºF)
insulation
Water tank Without baffle 400ºC (750ºF)
Water 60ºC–80ºC (140ºF–175ºF)
Top exterior of tank 36ºC (96ºF)
Underside of front 63ºC (145ºF)

Exhaust flue

Drain
Burners

Service door Gas controls/thermostat

FIGURE 6-5 Cross section of typical residential gas water heater and some typical temperatures developed.
Some data courtesy Fire Findings. Additional data from tests by author (unpublished).

Chapter 6 Sources of Ignition 181

 FIGURE 6-6 Typical double-wall heater vents. Exterior—galvanized steel. Inner—steel or aluminum. Courtesy of
John D. DeHaan.

the baffle in place. If the baffle is left out, the gas temperature rises to 565°C (1,050°F).
The draft hood allows entrainment of room air, reducing the gas temperature in the flue
(with baffle) to around 200°C to 250°C (400°F to 450°F).24

On-Demand Heaters
Energy conservation issues are prompting the installation of “on-demand” or “tankless”
water heaters instead of traditional 30- to 50-gal reservoir units. These units can be gas
or electrically fueled. They are typically wall-mounted metal boxes approximately 0.9 m
(36 in.) high, 0.5 m (20 in.) wide, and 0.3 m (12 in.) deep. They may be concealed in clos-
ets or service areas. They provide hot water only when they sense water flow. A gas-fueled
burner provides 120,000–200,000 Btu/hr (31–57 kW) of heat to heat 4–8 gal of water
per minute. Due to their very high heat outputs, on-demand units have to be properly
installed and vented. Most gas units have external electrical connections for the ignition
and control devices.25

Fire Cause Considerations

When a gas appliance starts a fire, the fire pattern itself will often, if not usually, reveal
the fact. Associated with the pattern showing the origin to be behind, around, or over the
gas appliance there must invariably be some malfunction, improper installation, or other
detectable fault or defect. Only the combination of factors can be considered as proof of
origin from a gas appliance. The reason for the double requirement is relatively obvious.
The arsonist often will set the fire under or close to a gas appliance to make it appear that
the appliance was responsible for the fire. In such cases, liquid flammables are likely to
be poured around the appliance and the fire started so that the origin seems to be obvi-
ously associated with the appliance. An incorrect assessment can result if care is not exer-
cised in the scene examination. In such an instance, such ignitable liquids are likely to
penetrate below the appliance, and possibly below the floor, and produce the character-
istic charring and damage in regions a gas fire could not reach. Natural gas fires are
unlikely to burn much below their point of release, since the fuel is significantly lighter
than the surrounding air. Fuel oil and the vapors from flammable liquids will settle down-
ward. Propane released under pressure (i.e., from a delivery line or as a jet) mixes with
air sufficiently that it will neither rise nor settle.
Common sense to the contrary, people insist on storing flammable liquids in furnace
or water heater closets. Leaks or spills from stored containers create vapors that are
quickly drawn into the combustion chamber of the appliance (by the convection of either
182 Chapter 6 Sources of Ignition
 FIGURE 6-7 New FVIR
water heater with closed
controls and fuel supply.
Note flash suppressor
mesh in air intake.
Courtesy of Jim Albers, Santa
Ana Fire Dept. (retired).

the pilot flame or the main burner). If there is sufficient vapor mixed with the air, an
explosion (usually followed by sustained fire of the liquid) will result.
Recent changes in water heater design have been aimed at reducing the chances of
ignition of flammable liquid vapors (called FVIR—flammable vapor ignition resistant—
designs) (see Figure 6-7). For many years, building codes have required gas water heaters
to be at least 18 in. (45.7 cm) above floor level if they are placed in a garage (where acci-
dental releases of flammable liquids are likely). Testing has revealed that unless the liquid
is poured near the heater and the room tightly sealed (to prevent losses by advective flow
under doors), this elevation precaution will prevent ignition. If spillage of flammable liq-
uids is a possibility even inside homes, to prevent ignition one design places a standard
water heater into a steel bucket (or “weir”) with sides 18 in. high. Testing has revealed
this design to be highly efficient in preventing ignitions.26 Another new design employs a
flash suppressor screen on the combustion chamber and sealed service, electrical, and
plumbing entry ports. This design will prevent any ignition (of vapors or fugitive gas)
starting in the combustion chamber from propagating into the room.27

Insulation
All approved gas appliances are surrounded by some type of insulation. In circulating
heaters, this insulation consists merely of an outer wall separated from the wall of the fire-
box, through which circulates the cool air that is to be heated. Such a wall will not
become significantly hotter than the ambient air that circulates behind it if air circulation
is maintained both within the heat exchanger and outside the appliance. Presence of com-
bustible trash against such an appliance wall will not normally lead to ignition of the
trash. The same cannot be said of the wall of the combustion chamber, which may well
be hot enough to ignite material against it. However, something would come into direct
contact with it only under very unusual circumstances.
Shifting or even partial collapse of such air-space insulation may create hot spots on
the exterior, as will settling of packed insulation. These hot spots then create a higher ignition
Chapter 6 Sources of Ignition 183
 risk. Improper clearances between vent stacks (flues) and wood framing can result in igni-
tion. The absence of the draft hood or central baffle in a gas water heater will cause
higher-than-design (normal) gas temperatures on the outside of the vent, which can trig-
ger degradation and ignition of nearby wood.

Backfire or Rollout
A common claim that there has been a backfire from a furnace or other heating appli-
ance carries somewhat more credibility than an ignition of combustibles against the
heater. Nevertheless, a backfire is an extremely rare happening that can result only from
gross maladjustment or malfunction within the appliance. To obtain such a “backfire,”
it is necessary for the gas to accumulate in some quantity in the combustion chamber
before it is ignited. Such accumulation can occur as a result of overpressure in the deliv-
ery system, use of the wrong fuel delivery orifice, or blockage of the heat exchanger or
exhaust vent.28 On ignition, a small explosion occurs and flames may blow out of the
open front of the appliance. If there is a suitable combustible fuel in the path of this
small amount of flame, a fire could result. The combination of factors is so unlikely,
however, that any claim of this type must be very carefully scrutinized before being
accepted. With a pilot light, it is difficult to accumulate sufficient gas for any signifi-
cant backfire, although maladjustment of some type may be effective. In any event, so
gross a defect is not difficult to recognize. In the absence of a pilot light, the gas will
not ignite until it diffuses or spreads to some other source of ignition, which in all prob-
ability will require enough gas to produce an explosion. In addition, escaped natural
gas will often pass harmlessly up the vent pipe into a chimney. It is difficult to imagine
any reasonably normal situation in which fires are caused by this type of situation. The
common use of electric ignition systems on gas furnaces, stoves, and water heaters is
replacing a continuous ignition source—the pilot light—with an intermittent source
that does not function until demanded. This system could allow for the buildup of a
volume of gas/air mixture with explosive potential instead of providing a continuous
ignition source to burn off small quantities as they reach their lower flammable limit.
With a fuel oil–fired furnace, the motor-driven pump can produce a sustained fire out-
side the furnace housing if the exhaust or heat exchanger is blocked. The failure of a
fuel gas pressure regulator (either in the appliance or external in the delivery system)
can cause continuous massive “overfueling” that results in the exit of burner flames via
rollout ■ Exiting of appliance service doors or vent openings (rollout).”
burner flames in an
appliance due to mas- Direct Contact
sive overfueling. Another manner in which a gas appliance can start a fire involves direct contact of a
normal gas flame with flammable or combustible material. In general, this will occur
only with open flames, such as gas range tops or broilers, although flammable materi-
als can sometimes be drawn into a protected flame. The electric elements of stoves or
broilers can also ignite solid fuels that come into direct contact, or the contents of skillets
or pans.

Plugged Vent
It has been reported that if the vent of a gas water heater or furnace is plugged, the lack
of a chimney can induce sufficient back pressure that the flames are forced out of the
burner enclosure, possibly into contact with nearby combustibles.29 Examination of such
devices should include careful examination of the exhaust duct for the remains of bird
nests, dead leaves, construction debris, and the like. Blockage of the vent stack alone
should cause the hot gases to vent from the draft hood if it is properly installed. The inves-
tigator should also check to see that the lower flame shield (under the burner) has not
been removed and that the service doors are in their proper locations. (Careless owners
often neglect to replace the service doors after igniting the pilot light and should be inter-
viewed to determine the last time the equipment was serviced or lighted.)
184 Chapter 6 Sources of Ignition
Open Flames
Open flames are most common on kitchen ranges, laboratory and industrial burners, and
special types of heating appliances and are more hazardous as sources of fire. A common
example is the ignition of cooking oils and greases in the kitchen, as well as towels, cloth-
ing, curtains, paper, and the like near open flames. Fires started by stoves and ovens are
the leading cause of residential fires.30 Autoignition temperatures of lard or vegetable
cooking oils are quite high [over 350°C (662°F)], but enough boiling and splattering oils
can splash onto the burner below to cause piloted ignition at lower temperatures. (The piloted ignition ■
piloted ignition temperature is the minimum temperature at which a fuel will sustain a Ignition aided by the
presence of a separate
flame when exposed to a pilot source.)
external ignition source
Skillet fires can be very large (see Figure 10-40). Grease ignitions are especially com- such as a flame or
mon as a cause of fires in restaurants, where the amount of cooking is great, and time spent electric arc.
cleaning stove hoods may be minimal. Such fires may also spread upward in grease-coated
piloted ignition
exhaust ducts, where fans and vents carry the fire up with increasing magnitude. Most
temperature ■ The
building codes require automatic fire protection of the cooking area and lower hood. Such minimum temperature
systems can do little if the flame is established in the exhaust ducting above the hood. at which a fuel will
The ignition temperatures and flash points of cooking oils are listed in Table 6-2. sustain a flame when
Note that the fire points are well above the “normal” cooking oil maximum tempera- exposed to a pilot
source.
ture of 205°C (400°F).31 This suggests that unattended cooking is a major contributing
factor. Tests observed by one of the authors of 500 ml (20 fl oz)of corn oil in a skillet on
a gas stove revealed that heavy white smoke with a marked lachrymatory property was
generated after ~10 minutes, and autoignition occurred about 25 minutes after the start
of the test. The flames were between 0.5 and 0.7 m (19 and 28 in.) high and caused igni-
tion of the cabinets above and to the nearest side of the stove.32 Note in Table 6-2 that
the fire point, AIT, and hot-surface ignition temperatures of all cooking oils decrease
with use.
Small gas appliances that are not part of a permanent installation can be displaced
or upset, with occasional drastic results. Laboratory burners are especially subject to
being tipped over. They can start fires by playing the flame on top of a wooden table,
for example, or against some other combustible material. Similar considerations may
hold for small portable gas heaters sometimes used in homes and similar areas. These
heaters may be permanently installed or temporary and may not meet any standard of
safety or building codes. The investigator is referred to NFPA 54: National Fuel Gas
Code or NFPA 58: Liquefied Petroleum Gas Code, which give detailed instructions for

TABLE 6-2 Properties of Cooking Fats Measured at Fire Research Station

HOT-SURFACE
FAT FLASH POINT (°C) FIRE POINT (°C) AIT (°C) IGNITION TEMP. (°C)
AFTER 8 AFTER 8 AFTER 8 AFTER 8
HEATING HEATING HEATING HEATING
VIRGIN CYCLES VIRGIN CYCLES VIRGIN CYCLES VIRGIN CYCLES
Corn oil 254 227 NA 321 309 283 526 542
Drippings (bacon) 254 241 NA 331 348 276 553 537
Hydrogenated cooking fat 260 210 NA 331 355 273 568 554
Lard 249 218 NA 326 355 282 541 568
Olive oil 234 218 NA 316 340 280 562 543
Peanut oil 260 243 347 335 342 280 552 535

FFrom Ignition Handbook by V. Babrauskas, © 2003, p. 886, Fire Science Publishers. Used by permission.

Chapter 6 Sources of Ignition 185

 the proper installation and operation of all types of gas-burning appliances. Fire Findings
reported the air venting from a 60,000 Btu/hr (17 kW) kerosene-fueled salamander
portable heater had a temperature of 639°C (1,182°F). Due to turbulent mixing, the tem-
perature of the maximum hot-air stream dropped to 238°C (460°F) at 0.3 m (1 ft) and
148°C (298°F) at 0.6 m (2 ft) from the heater.33

Altered Gas Nozzle

An unlikely, but possible, type of fire causation from a gas appliance involves some alter-
ation in the gas nozzle in the venturi, possibly by partial or total unscrewing of the noz-
zle or installation of the incorrect fuel nozzle. If this should happen, much larger than
normal quantities of gas would reach the flame, which, in turn, might become much big-
ger and spill from the top or front of the appliance. The authors have not witnessed such
an accident, but it is conceivable, especially when tampering or ignorant efforts at repair
or adjustment have occurred. Such fires would necessarily occur at the first use of the
appliance after the opening was enlarged, which should again make their detection and
proof of cause relatively simple. The service and repair history of the appliance may reveal
recent problems or repairs. As discussed in Chapter 4, the venturi used must match the
fuel gas and pressure regulator being used. Often, X-rays will reveal internal failures,
blockages, or the open/closed position of fire-damaged valves and regulators.

PORTABLE ELECTRIC APPLIANCES

Heat-producing electric appliances can, of course, cause ignition of susceptible fuels even
when operating normally. (Arcing, failures, and malfunctions will be discussed in Chapter
10.) Electric heaters, toasters, or toaster ovens with exposed Nichrome (resistance) heater
wires offer an ignition source with surface temperatures on the order of 600°C (1150°F)
or greater. Such hot surfaces are not able to ignite most flammable gases or vapors but
can ignite most cellulosics (including foods) if they come into contact or near contact with
them (see Figure 6-8).

FIGURE 6-8 Toaster

pastries ignited by
repeated operation of
toaster cause flames
30 cm (~12 in.) high
sufficient to ignite
nearby combustibles.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

186 Chapter 6 Sources of Ignition

 The heater elements in clothes dryers are especially susceptible to such contact. Dryer
lint is one of the most readily ignitable solids in the average residence. It is usually a mix-
ture of fine cellulosic, synthetic, and natural fibers (wool and hair). Accumulations of lint
on heater grids or on adjacent surfaces of venting (both inside the dryer and out) are one
of the most common “first fuels” in dryer fires.
The radiant heat produced by most portable electric heaters is on the order of
10 kW/m2 or less at the wire guard. This may be enough to ignite cellulosics or char elas-
tomeric and thermosetting materials if contact is prolonged. Cellulosic fuels that are
brought into contact with the heater element itself are likely to be ignited quickly. Today,
most consumer-type portable electric heaters have “tilt” switches to shut power off if the
unit is knocked over, in addition to thermostats and TCOs. Investigators must carefully
examine such heaters to determine what, if any, safety devices are present and if they have
failed or been tampered with.

KEROSENE HEATERS
The fuel in kerosene heaters may be delivered to a constant flame by vertical transport up
a wick that is partially inserted in a fuel reservoir, in the same way as in a kerosene lamp.
A more effective means of fuel delivery is a barometric delivery system in which the fuel
is delivered to a shallow horizontal reservoir from a vertically oriented removable tank.
The flow of fuel is controlled by a barometric valve in the tank. A partial vacuum above
the fuel in the tank prevents the fuel from flowing into the reservoir until the level drops
below that preset by the valve. Modern heaters have a shutoff mechanism that extin-
guishes the wick if the heater is tipped over or subjected to excessive vibration (move-
ment). The main safety problem with kerosene heaters is that the capacity of the tank
exceeds that of the reservoir, so that if there is a loss of the vacuum in the tank, the excess
fuel can flood the reservoir and create a spill external to the heater, which then ignites
with catastrophic result.34 This loss of vacuum can occur as the result of a leak in the tank
or its associated parts, or as the result of fueling the heater with a fuel such as gasoline or
camping fuel with a higher vapor pressure than that of kerosene. In these cases, the vapor
then replaces the partial vacuum that holds back the flow of fuel, resulting in a flood of
fuel that overwhelms the reservoir. Although a newer anti-flare-up device is available, most
older heaters encountered pose a serious fire risk from this type of failure.35 Accidental
spills during refilling can cause significant fires (especially if the burner is not extinguished).
Any residual fuel left in such a heater should be collected and preserved for testing to
establish its actual content.

STOVES AND HEATERS

A variety of liquid fuel portable heaters are also made for camping and boating use. These
range from pressurized fuel delivery systems (Primus stoves) to denatured alcohol fuel
burning on the surface of a noncombustible wick element. These heaters are subject to
spillage of burning fuel if tilted or overturned. Although these heaters are not intended for
use in enclosed spaces, they are often seen in tents and campers. Because such heaters may
consist of a small circular burner pan and a perforated metal shield over it, they may not
be recognized by an investigator as ignition sources (see Figure 6-9).

OIL STORAGE
External heating (fuel) oil storage tanks are common in both residential and commercial
premises. Traditionally, these are welded steel tanks with rigid metal pipe connections.
Those features, combined with the low vapor pressure (high flash point) of typical heat-
ing fuels, mean they do not constitute a high risk of ignition (either accidental or intentional).
However, when such fuels are absorbed on porous materials that can act as a wick, their
low ignition temperatures (~205°C; 400°F) pose a risk. If plumbing connections are
Chapter 6 Sources of Ignition 187
FIGURE 6-9 Alcohol-
fueled units sold in
camping stores can be
used as tent heaters or
cook stoves. Courtesy of
John D. DeHaan.

exposed to an external fire, they can fail quickly. Some jurisdictions allow fuel storage
tanks to be made of high-density polyethylene (owing to their corrosion resistance and
lower costs). Testing has shown that such tanks (in 300- to 400-gal sizes) can be caused
to fail by external fires. If 5 L(1.5 gal) of gasoline is poured over the outside, or a mod-
est cardboard/newspaper fire is ignited underneath the tank, a catastrophic failure and
release of the tank contents can result. Placement of a cloth wick into the mouth of the
tank also causes failure, since it continues to burn within the tank (due to the low vapor
pressure) until the tank melts. The melted plastic of the tank can support a large pool fire
of molten polyethylene, so even an empty tank poses a fire risk.36 Oil-fueled furnaces can
cause fires by creating heavy soot if burners are maladjusted, or more commonly, leaks
from fuel connections can support flames outside the combustion chamber, which then
ignite nearby combustibles.

ELECTRICITY
Electricity represents a source of ignition in the form of arcs, electrically heated wires and
heating elements, and even lightning. For a full discussion of all means of electrical fire
causation, see Chapter 10.

The Role of Hot and Burning

Fragments in Igniting Fires
Hot or burning particles, commonly termed sparks, are in a special category with regard
to kindling of fires. As pointed out earlier in the chapter, the designation spark is an
ambiguous one, since it can refer to incandescently hot metal from a mechanical or elec-
trical source or to hot bits of debris from an established fire. Most fragments from an
established fire are glowing, although they may at times be flaming. Bits of wood, paper,
and other light organic fuels are most susceptible to the formation of glowing fragments,
188 Chapter 6 Sources of Ignition
and larger pieces of paper may travel while still flaming. Bits of paper or tobacco may
become detached during lighting of a cigarette or pipe. These particles naturally have the
propensity to ignite other fuels with which they come in contact if they retain enough
heat. It should be remembered that most such fragments will be carried upward by the
draft from the fire that generates them, and air currents may carry them for significant
distances to ignite other fuel.

WINDBLOWN SPARKS
The question frequently arises: How far can windblown sparks travel and still cause a sec-
ondary fire? Although no precise answer is possible, some factors can be stated. The ambi-
ent wind is highly important, because such sparks tend only to rise with the buoyant flow
over the fire, drift a short distance, and fall nearby, unless propelled by wind. Another fac-
tor is the type of material that is burning. Very small fragments and thin materials such as
paper will normally be totally consumed before they return to some combustible surface.
Fragments of wood, excelsior, or corrugated cardboard will burn considerably longer and
travel farther before igniting something else. For the latter, distances of about 12 m (40 ft)
have been observed. Tests by DeHaan demonstrated that burning wood chips and straw
could be carried up to 15 m (50 ft) from their source in a 32-km/hr (20-mph) wind to ignite
cloth, straw, and wood shavings targets.37 Light items will probably only rarely travel as
much as 6 m (20 ft), except under significant wind conditions. Another factor is the height
that they reach before being blown to one side. The higher they go, the more time they
have to burn up totally or to cool off, and the longer their exposure to any air movement.
In any event, distances in excess of 9 to 12 m (30 to 40 ft) are to be accepted only with the
greatest caution. Documentation of the intensity and direction of wind at the time of the
fire is critical to the evaluation of such mechanisms.
Throughout this discussion of spark travel, the role of wind has repeatedly been referred
to as modifying the fire spread. No investigation involving the distance or effectiveness of
such spark ignition is complete without ascertaining from meteorological records, or at least
from eyewitnesses, the wind conditions at the time in question. It is evident that violent
winds will produce very different effects from mild ones. However, the difference is not
that which might be supposed. A strong wind will blow a spark farther than a weak wind
but will also speed the rate of combustion. With a given size of burning fragment, the wind
effect does not vary greatly with different wind velocities. The most significant effect of the
strong wind, by far, is its ability to blow large fragments of material, which will naturally
burn longer than small ones. A wad of burning newspaper on the ground will not ordinar-
ily travel along the ground for any significant distance if propelled only by a breeze. In a
strong wind, however, it may roll for a considerable distance before it is consumed. The rela-
tion of importance is the time necessary for it to burn as compared with the speed with
which it is moving laterally with the wind. The consideration is far more relevant to mate-
rials at ground level than to those that emerge from chimneys, because very large burning
objects are rarely carried up a chimney to any significant height. In very large fires, how-
ever, the updraft can loft large pieces of burning debris, boards, branches, cardboard, and
rags high enough that they can be carried hundreds of yards by the prevailing wind, ignit-
ing remote fires. This has been observed to be a major means of fire spread in large fires in
both urban and wildland settings, as depicted in Figures 6-10a and b.
In one recent major conflagration, burning wood fire brands (some as large as 2 ⫻ 4s)
were lofted some 60 m (200 ft) above a multistory condo/retail complex under construc-
tion and were blown more than 680 m (0.4 mi) to ignite wooden shake-shingle roofs of
two apartment complexes (see Figure 6-10a). At the time, the winds were 2–27 km/hr
(13–17 mph).38 See Figure 6-10b for an illustrative map.
There are several common origins for hot fragments, and since these are the primary
source of the fire that may result, they will be considered individually.
Chapter 6 Sources of Ignition 189
FIGURE 6-10A Massive fire engulfing six-story residential/commercial FIGURE 6-10B Map showing trajectory of burning brands 0.7 km
complex (wood framed). Courtesy of Don Perkins, Fire Cause Analysis. (0.4 mi) from fire scene to ignite roof of apartment complex 0.8 km
(0.4 mi) away. © 2009 Google––Map data––© Tele Atlas

FIREPLACES AND CHIMNEYS

Fireplaces are blamed for igniting many fires, some of which they undoubtedly do.
Nevertheless, the mere fact that a possible fire origin is close to a fireplace is not to be
considered as positive indication that the fire was initiated by the fireplace. Discussed next
are six ways in which a fireplace can be instrumental in initiating a fire.

Emission of Sparks or Blazing Materials

from the Front or Open Portion
Such sparks or blazing materials may ignite nearby combustible floor coverings or furnishings.
Serious fires from this cause are rare because the top of a wooden floor is very difficult to
ignite to a sustained combustion. Scorch marks are not uncommon and are annoying, but
wood floors do not usually represent the first fuel ignited. Light fuels, such as paper, may
readily be ignited, so prefire actions and conditions may be critical. Some rugs or carpets
are sufficiently flammable to be ignited under these circumstances, and furniture may occa-
sionally be ignited. Most building codes require a noncombustible hearth extending well
to the front and sides of a fireplace opening. Burning fragments of Christmas wrappings
apparently ignited a large pile of crumpled wrappings in front of the fireplace in the fire
shown in Figures 6-11a and b. Thus, there is a hazard from this source, but it is readily
interpreted when encountered (when prefire activities are described).

Intense Radiant Heat from a Fireplace Stove

Filled with Inappropriate Fuels or Overfueled
Sufficient radiant heat can be generated by intense fires in the firebox to create heat fluxes
of 10–20 kW/m2 through the front opening. These can ignite ordinary combustibles in
line of sight of the opening if they are within 1 m (3 ft). Inappropriate fuels include crum-
pled masses of paper, Christmas tree branches, plastics, foams, cardboard boxes, and sim-
ilar commodities. Some fuels are intended for fireplace use (e.g., compressed-sawdust
artificial logs) but may cause overheating if overfueled. The manufacturer of artificial logs
typically includes package instructions requiring that not more than one log be put on the
fire at a time. Overheating of the firebox is also liable to cause ignition of wood materi-
als located behind the fireplace or inside the wall.

Emission of Sparks or Blazing Materials from a Chimney

Chimney fires, which are the greatest danger factor, arise because soot, bird or rodent nests,
creosote, dust, cobwebs, and a variety of combustible materials sometimes are allowed to
190 Chapter 6 Sources of Ignition
FIGURE 6-11A Living room furniture burned when fragments of FIGURE 6-11B Heat and smoke layer enveloped the top of Christmas
burning Christmas wrappings escaped from fireplace to ignite a pile tree but only scorched it (oxygen-deprived smoke layer). Courtesy of
of wrappings on the floor. Courtesy of Derrick Fire Investigations. Derrick Fire Investigations.

accumulate in the chimney, where they may be ignited by a spark or flame from the fire
below and lead to a large blaze within the chimney itself. When this happens, large quanti-
ties of burning fragments will be expelled from the chimney top and greatly increase the
danger of igniting a roof or other adjacent combustible structure. It is clear that this possi-
bility is the most dangerous from the standpoint of starting large fires. It is also clear that a
large proportion of them will be roof fires, although blazing residues (firebrands) or large
glowing sparks (embers) may travel some distance and ignite fires elsewhere than on a roof.
One consideration of the effect of wind on chimney action is also relevant. Wind blowing
across a chimney opening exerts what is known as the Bernoulli effect, that is, a lowering
of the pressure at the chimney mouth. The result is a stronger draft up the chimney, which
can increase the size of fragment that can move up to the top and be ejected into the stronger
wind stream. This is an additional reason for the investigator to be very careful considering
wind velocity when there is a possibility that spark movement started a fire.
Emission of Sparks or Blazing Materials
Resulting from Defects or Holes in a Chimney
Such sparks can ignite timbers or other construction material adjacent to the chimney.
Fires kindled by this method are equally rare but may be very serious when they occur.
Few chimneys become so defective by use or abuse as to pose any real danger of leaking
sparks or burning materials through their walls. However, such defects can result from
anything that causes serious deterioration of the house, such as earth settling or slippage,
earthquakes, or extreme age and decrepitude. Some manufactured chimneys can separate
at their joints due to corrosion or mechanical failure, as shown in Figure 6-12. A chim-
ney fire may damage the chimney lining, so that gaps develop that allow hot gases to
escape the flue and begin to char wooden structures. The next time there is a chimney fire,
such gaps may also permit flames and sparks to contact combustible framing directly.
Because of the characteristic pattern and behavior of a roof fire (to be discussed in
Chapter 7), there will rarely be a serious problem in interpreting the cause of the fire. The
recent history of use should be established by interviews with the occupants.
Overheating of Combustible Materials near
or in Contact with a Fireplace or Chimney
Overheating of combustible materials by radiant heat from or direct contact with the hot
surfaces of fireplaces or chimneys is probably less important in fires involving stone or
Chapter 6 Sources of Ignition 191
FIGURE 6-12 Chimney
cap was dislodged and
allowed hot gases to vent
into wood-framed enclo-
sure eventually igniting it.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

masonry fireplaces than those with metal ones. Fires in the immediate vicinity of chim-
neys are not uncommon, and many people attribute them automatically to overheating of
wood structural elements. The insulative capacity of intact stone or brick is such that
these chimneys rarely transmit enough heat to damage the structure. Metal chimneys and
fireboxes are much more common today. Loose insulation between the double walls of
metal fireplaces can settle, allowing spaces with minimal insulation capacity. The outer
walls can then reach temperatures far in excess of what they should, and these may be
dangerously close to combustibles, as shown in Figure 6-13. Chimney enclosures today
are often of wood construction. They are supposed to be lined with drywall, but this lining

FIGURE 6-13 A close-up

of the fire area around a
metal heater vent shows
extensive charring of
timbers due to inadequate
clearance. Courtesy of the
Fire Investigation Unit,
Metropolitan Police, London.

192 Chapter 6 Sources of Ignition

is not always installed. Manufacturers of metal fireplaces provide detailed installation
instructions, and if these have been violated by the installer, a fire may well occur.
Violations can include disobeying required clearance from combustibles, use of incorrect
fluepipes, blocking of required air-circulation openings, or altering the manufactured
unit. The commonly specified triple-wall flue pipes depend on proper flow of cooling air
through the annular segments, and the design can be defeated by improper installation
that blocks this flow of cooling air. It is occasionally found that a tradesperson “simpli-
fied” the installation of a fireplace by removing protuberant standoffs that are designed
to enforce a certain distance of clearance. By the way, the term zero clearance is an inap-
propriate marketing catchphrase and should not be taken to mean that such a unit can be
installed in any relationship to combustibles that the installer might wish. Fireplaces lack-
ing an Underwriters Laboratories (UL) (or equivalent) listing obviously may be presumed
to be intrinsically defective and unsafe.
Circumstances of recent use may be of great value in establishing the actual cause of
chimney-related fires. Unusually large or hot fires (such as from burning a dry Christmas
tree) or sustained fireplace use due to a very cold weather spell may trigger ignition where
it had previously been avoided by “normal” use.
Similar considerations apply to furnace and water heater flue pipes where they pass
through openings in wooden structures. It must be remembered that in such instances
there is usually a chimneylike insulator structure surrounding the actual chimney or vent.
Any fire that starts from any cause in the vicinity of such a penetration will be drawn up
it, often localizing the fire in the region around the suspected chimney or vent. Such insu-
lators or “thimbles” may be required around exhausts for heaters, stoves, and emergency
generators (see Figure 6-14) For this and other reasons, such a postulated origin of fire
should be scrutinized most carefully before accepting it. Considerations of the properties
of purported first fuels are relevant.
Changes in equipment such as the installation of high-efficiency heaters or fireboxes
(fireplace inserts) can change the airflow of an old chimney or dramatically increase the
surface temperatures of fireboxes and flues, thus creating problems where none existed
before. Such increases may be enough to overwhelm the designed insulation or induce
more heating by conduction through brick or stone, resulting in degradation of wood
structural members in contact. Gaps can also develop between the top front of the fire-
box and the facing brick or stone. Because the chimney and fireplace are not likely to be
destroyed by the fire, evidence of gaps in its structure, added to the fire pattern, should

FIGURE 6-14 Sheet metal thimbles for passing exhaust vents through walls. Courtesy of John D. DeHaan.

Chapter 6 Sources of Ignition 193

 reveal to the investigator its contribution to starting the fire. Some units require that fire-
box side edges be tightly sealed with noncombustible material. A gap there can allow hot
air emanating from the front of the unit to get sucked into the chase behind the fireplace,
causing overheating of wood materials there. This is another possible source of fire.

Embers and “Hot Ashes” Removed from Fireplaces and Barbecues

Embers from fireplaces, campfires, and barbecues have been identified as the cause of
numerous “delayed ignition” fires remote from the fireplace, campfire, or grill. Due to low
heat release rate (HRR) and slow combustion and the insulative properties of ashes and
the charred wood itself, the embers are undetected when removed. Usually, the homeowner
cleans out the ashes and places them in a combustible box, bag, or trash can, assuming
they are “out.” Wood or charcoal embers, insulated by ashes, can continue to smolder for
3 to 4 days under the right conditions and can result in ignition after being removed.39

LONG-TERM HEATING (“LOW-TEMPERATURE” IGNITION)

As we saw in Chapter 5, wood, especially in massive form, requires a considerable amount
of heat to cause its ignition. For flaming combustion, there must be enough heat flux to
pyrolyze volatiles from the wood and ignite those vapors. Temperatures around proper
flues and chimneys in good repair are not expected to reach the 250°C (482°F) minimum
ignition temperatures for fresh wood. However, it has been reported that exposure to much
lower temperatures (below 105°C or 221°F) causes degradation of wood to charcoal.40
This charcoal, also erroneously called “pyrophoric carbon,” is sometimes found in wood
around heat sources like furnaces and chimneys. (Although the term pyrophoric is techni-
cally misapplied here because it refers to materials that ignite spontaneously in air at mod-
erate temperatures [54°C (130°F) or below], it is widely used in fire investigation in
discussing wood that appears to ignite at temperatures far below its usual ignition temper-
atures.) These mechanisms were discussed in Chapter 5 (see Figures 5-4–5-6).
It is apparent to experienced fire investigators that fires have begun in confined spaces
in wooden structural components where exposure to a source of some heat for long peri-
ods of time has caused far more charcoal to form than would be expected from the dura-
tion of the reported (or observed) fire. Our review of fire cases that present the strongest
argument for the formation of self-heating charcoal shows that the circumstances are
often as Bowes predicted: a heat source near wood that is protected by a barrier of metal
or tile or in a very confined area (such as a pipe or flue snugly fitted through a wooden
beam, as pictured in Figure 5-8, for a long period of time, followed by structural collapse
or other change that allows a sudden inrush of fresh air, can precipitate a flaming fire.41
Production of this charcoal is clearly a long-term phenomenon, often requiring many
months, and more typically years, to reach combustible form and critical mass. Martin
and Margot reported that temperatures as low as 105°C (221°F) were capable of degrad-
ing wood but that the lower the temperature, the longer the time required.42
The mechanism for this process is complex and has been the subject of some debate,
especially concerning the applicability of a predictive thermodynamic technique called
Frank-Kamenetskii after its creator.43 There are several important features of such
processes that have to be recognized:
■ The conversion of fresh wood to a charred or “cooked” state can occur at low
temperatures [well below 250°C (482°F) and possibly as low as 77°C (170°F)] over
a long period of time.
■ Fresh wood when heated produces gases and vapors that will support flaming
combustion (but requires significant heat be applied to effect this pyrolysis and then
ignition), whereas charcoal has much fewer volatile components and supports only
very limited flaming combustion but will readily support glowing combustion.
■ Fresh whole wood (sawn lumber) will not self-heat unless very hot, but charcoal will.
194 Chapter 6 Sources of Ignition
■ Fresh wood has a much higher permeability to air along its growth axis than across it.
■ Smoldering combustion can result from runaway self-heating.
■ A mass of fuel undergoing smoldering combustion can transition to open flame if
ventilation causes a high enough heat release rate (either external such as from a
draft of wind or enhanced buoyant flow around or through the mass).

When all these factors are taken into consideration within the context of the previously
described scenarios, we can see that there is a supportable hypothesis for these fires. First,
heat is applied to a mass of wood that has, by its mass or covering, limited exposure to
air (the work by Cuzzillo has shown the permeability of “cooked” wood to air to be a
critical factor).44 The wood degrades to a “cooked” form with the gradual release of the
volatile fuel vapors insufficient to support flames (with the characteristic flat, shrunken
surfaces with sharp-edged segments). The charred, “cooked” wood has a high permeabil-
ity to air enhanced by the multiple cracks into the wood produced by the “cooking,” a
low thermal inertia, and an ability to self-heat, all very different from fresh wood.
Continued application of heat to such a material can induce runaway self-heating, espe-
cially if air is admitted as the result of shrinkage or separation from sealing materials. This
runaway self-heating is the critical step in the process. At this point, there is ignition (self-
sustained smoldering combustion), but it may not be recognized from outside the cavity
or concealed space where it is taking place. If enough air is admitted and the enhanced
ventilation causes a high enough heat release rate, the smoldering combustion can transi-
tion to flaming combustion (with the rapid involvement of adjacent intact wood struc-
tural members). The flaming stage is often detected as the “fire.” Figures 6-15a–c show
some of the critical features of the process. The conditions have to be nearly perfect—
temperatures typically between about 100°C and 200°C (212°F–392°F) with very limited
ventilation, so that heat is not readily dispersed and oxygen generally excluded, and an
adequate mass of wood. There are extensive data on the relationship between fuel mass,
environmental temperature, and time to develop runaway self-heating. These conditions
are sometimes found in accidental fires investigated, but the investigator must be careful
of blaming “pyrophoric carbon” for any fire in the vicinity of a flue or hot water pipe
simply because no other cause can be identified. As Bowes pointed out:
■ The temperatures of heating surfaces have to be at high temperatures [although
nowhere as high (200°C; 400°F) as his calculations indicated];
■ The wood has to have substantial mass (plywood floors are more likely to simply
char through than to ignite);
■ The temperature of the hot surface has to be high enough to maintain heat flux on
the wood to overcome losses due to conduction or convection;
■ The time has to be long enough (weeks, months, or years, depending on the intensity
of the applied heat); and
■ There has to be exclusion (or limited circulation) of air in the vicinity until just prior
to detection of the fire.45

One case involving all five of these elements was the result of a water heater that sat on
a wooden platform “protected” by a steel sheet and an asbestos barrier. Some weeks after
a new water heater was installed, the platform began to burn and the fire was detected.
(A similar case was shown in Figure 6-4.) It should be remembered that smoldering com-
bustion (in charcoal) requires only very low oxygen concentrations and that fresh air with
its richer oxygen supply is required to sustain open flames. Because of the time required for
the production of charcoal, low-temperature ignition of wood can generally be eliminated
as a cause of fires in very new buildings. Interestingly, Babrauskas et al. reported that wood
in contact with a surface with a temperature in excess of 77°C (170°F), even if periodic,
should be considered a potential ignition scenario.46 Both UL and the Building Code of the
City of New York cite 170°F as the upper “safe” temperature for wood surfaces.47
Chapter 6 Sources of Ignition 195
 FIGURE 6-15A An 89-mm cube of
Douglas fir after being heated at
200°C (424°F). Solid-phase ignition
(smoldering) occurred on the end-
grain faces while the side-grain faces
exhibited little damage. Note the
greater shrinkage at the end-grain
faces than at the center. Unlike
isotropically porous materials where
the highest self-heating temperatures
are at the center, the highest temper-
atures in self-heating whole wood are
near the end-grain faces. Courtesy of
Dr. Bernard R. Cuzzillo, Berkeley Engineering
and Research, Berkeley, CA.

FIGURE 6-15B An 89-mm wood cube heated at 200°C (424°F) and split open to show
the hourglass heat pattern within. This indicates that the regions of maximum oxidation
reaction are near the end-grain faces and along the axial centerline. Courtesy of Dr. Bernard R.
Cuzzillo, Berkeley Engineering and Research, Berkeley, CA.

FIGURE 6-15C An 89-mm cube of

Douglas fir that has been heated at
200°C after its end grain has been
sealed with RTV silicone sealant, pre-
venting the diffusion of air along its
centerline. The exposed faces are
sharply faceted into small, flat
shrunken areas. The same pattern of
fissuring extends uniformly into the
center of the cube. This is similar to
exposing the middle of a long piece
of solid timber to a heat source mid-
length. The fissuring allows air pene-
tration into the wood block and could
support runaway self-heating. Courtesy
of Dr. Bernard R. Cuzzillo, Berkeley
Engineering and Research, Berkeley, CA.

196 Chapter 6 Sources of Ignition

 As with most other accidental ignition sources, careful investigation and evaluation
must be carried out before concluding low-temperature ignition or the more general
“overheated chimney” as a cause. In addition to there being much more char than
expected, if the wood timber is cross sectioned, the char should also gradually change
from complete char to partially pyrolyzed to undamaged as one moves away from the
purported heat source. Unfortunately, the processes involved and their interaction
are complex, and there are presently no data showing an exact relationship between tem-
perature of exposure versus time to ignition. Babrauskas has recently published an
informative review on the topic in his Ignition Handbook.48

TRASH BURNERS, INCINERATORS, AND BONFIRES

As sources of hot and burning fragments, trash burners, incinerators, and bonfires rate high.

Trash Burners and Incinerators

The types of material normally destroyed by burning in such devices are those most likely
to form airborne burning fragments, and the necessity of openings in the incinerator to
allow ventilation of the fire allows escape of burning fragments. Accidental fires particu-
larly of dry grass or brush in the area are commonly associated with trash burners and
incinerators. It is fortunate that fires started by trash burners will rarely require much
investigation because of the obvious nature of the origin and the fact that the pattern will
trace directly to the burner in most instances. The strength and direction of the wind and
the ignitability of possible first fuels (including fuel moisture and nature of materials being
burned) are all important considerations.

Bonfires
The bonfire poses a somewhat different problem than the trash burner, despite also being
an exterior fire. To maintain a satisfactory bonfire, it is customary to use relatively mas-
sive wood instead of paper, packing, cardboard, and similar items of trash. The wood
poses relatively much less hazard from rising fragments of burning material than does
trash. However, there is no restraint or protection of the fire from winds, as is true of the
trash burner. In a high wind, a bonfire becomes very dangerous from the standpoint of
starting larger, uncontrolled fires. If it is adjacent to dry grass or leaves, it is always haz-
ardous. Here also, the investigation is likely to be rather simple. An occasional difficulty
is in distinguishing the remains of a small bonfire from those of a general fire that
involved a concentration of fuel at a point that may leave excessive ash and charcoal,
which appear similar to the remains of a bonfire. Careful examination will usually distin-
guish the two. In many instances the remains of bonfires contain, in addition to wood ash,
the remains of food containers or other indication of human activity. The direction of the
prevailing wind and the contour of the land are critical to correctly determining such ori-
gins. The considerations of ignition by such airborne brands include type of fuel being
burned, size of fragments that can escape, and size of the fire (creating the upward buoy-
ancy). The larger the fragments and the more massive they are, the longer they can burn.
Flat brands such as fragments of cardboard or wood shingles have better aerodynamics
than needle or spherical shapes and can remain airborne much longer and travel farther.
See Chapter 8 on wildland fires.

HOT METALS
As a direct source of ignition of fires, metals are only of occasional importance. They can
play a part as sparks of burning metal from grinding or cutting operations or as droplets
of molten metals. Molten metal may carry enough heat to ignite susceptible fuels with
which it comes in contact and is therefore capable of starting fires, as are burning frag-
ments. Because a considerable amount of heat is required to melt the metal, such dangers
are usually associated only with special operations in metal fabrication. Places in which
Chapter 6 Sources of Ignition 197
 such operations are carried out are more than normally susceptible to fires from this type
of ignition. However, the dangers are often exaggerated, since many alloys, including
types of solder, will not generally carry enough heat to ignite fuel material on which they
fall. Babrauskas has published a significant review of such ignitions.49
Molten ferrous alloys, such as steel, are far more dangerous than solders because of
their higher melting temperatures and often higher thermal capacities. Thus, welding and
cutting of steel with an arc or a torch in the presence of susceptible organic fuels is a very
dangerous practice. See NFPA 51: Standard for Fire Prevention during Welding, Cutting,
and Other Hot Work for safety during those operations.
The state of subdivision of a solid fuel is critical in determining its susceptibility. It is,
for example, almost impossible to ignite a large timber accidentally by the short, direct
application of a torch flame or by the molten metal from it. In contrast, finely divided
fuels, such as shredded packing materials, sawdust, shavings, or loose paper may very
well be ignited under the same conditions. Directing a flame into a narrow crack or split
in a piece of wood, however, may cause ignition of both surfaces. The close spacing causes
reduced losses and enhanced feedback, resulting in ignition.
Molten metal can also be produced by the fusion caused by the heat of a short circuit
or failure of a component in electrical equipment. Although ejection of such material from
the equipment certainly presents a hazard, the greatest danger may be from direct ignition
by the short circuit itself, but only if a current protection device does not interrupt the cur-
rent. Thus, the metal may be considered a secondary and less probable source of ignition.
In considering hot or molten metal fragments as sources of ignition, the investigator
must always keep in mind that most metals have a high density, and the hot fragment will
fall rapidly, in contrast to bits of paper or wood, which tend to rise in the convective
plume of the fire itself. Aluminum is of relatively low density. In flat form, such as a roof
section, aluminum sheet may rise in a buoyant fire plume rather than fall, but this will
not be true of solder, molten brass, iron, or copper. Aluminum, because of its low density,
ease of oxidation, and low melting point, is very unlikely to play a role in spreading fire
as airborne debris and in general is not a likely source of ignition, except in the case of
high-voltage power lines. These are most commonly made from aluminum, and wildland
fires can readily be ignited when globs of molten aluminum land on dry vegetation. Such
molten masses are most commonly generated when two power lines unexpectedly make
contact, resulting in “conductor clashing,” because of extreme winds, failed insulators,
failed supports, or excessively stretched conductors, for example.
Despite their higher melting temperature, copper particles are a much less effective
ignition source than are aluminum particles. It should also be remembered that metals can
burn, although most metals are extremely difficult to ignite except when very finely
divided. Magnesium is the chief exception, either alone or alloyed with aluminum, being
more easily ignited.
Occasionally, massive metal pieces heated by torch cutting or welding operations
can ignite ordinary combustibles they contact. This is far more likely if the fuel is a char-
forming material rather than a thermoplastic. This mechanism is extremely unlikely, how-
ever, if more than a few minutes have elapsed between the hot work and observed ignition
(due to the high thermal conductivity of the metal). Investigation and thorough interviews
will be needed to test such hypotheses. Fire watches after hot-work operations are typi-
cally only 30 minutes long, but some smoldering ignitions from hot-work operations have
been known to take hours, instead of minutes, to manifest in a flaming fire.

MECHANICAL SPARKS
Iron, steel, and some other metals will spark when struck by a suitable object, as during
grinding operations or when a portion of a vehicle strikes a stone-surfaced roadway. The
actual mechanism is quite simple. A highly localized frictional contact rips a fragment of
198 Chapter 6 Sources of Ignition
the ferrous alloy out with sufficient energy to heat it beyond its ignition temperature [on
the order of 700°C to 800°C (1,300°F to 1,475°F)]. The particle ignites in the air and
burns (at ⬎1,600°C; 2,900°F), leaving a sometimes incandescent flake of carbon and iron
oxide as an ash.50 These sparks represent small but finite heat sources that can ignite
some flammable vapors. If the “ash” flakes are hot enough and large enough to transfer
sufficient heat to their surroundings, they can ignite readily combustible (finely divided)
solid materials on which they fall. Such fires are seen along railroad tracks where mechan-
ical sparks from overheated brake shoes can be dropped for many hundreds of yards.
Witnesses may report fires associated with passing trains.
Many grass and brush fires along railroads are ignited by particles emitted from diesel
locomotive exhaust or eductor tubes. These particles, consisting of carbon, calcium oxide,
and the oily residues of fuel and lubricant, are emitted along with the exhaust. Under
some conditions of engine operation they are emitted while still burning and can ignite
dry grass along the track or even several feet from it.51 Many states require spark
arrestors on all engines operating in high fire-hazard areas, but they are sometimes over-
looked or even intentionally eliminated during maintenance.
Sparks can also be created by cutting or grinding of iron or steel using a carborun-
dum blade grinder. Incandescent particles have been observed up to 5 m (15 ft ) from a
handheld grinder being used on a steel reinforcing bar. Such sparks reportedly have a
starting temperature of 1,600°C to 2,130°C (2,900°F to 3,866°F) and an energy on the
order of 1 kJ. Impact of a steel tool on concrete (particularly quartz) or aluminum-coated
or contaminated rusty steel can also create mechanical sparks capable of igniting flamma-
ble gases.52
The rail-grinding equipment used on railroads creates large quantities of hot sparks that
can be projected some distance from the track. Because such equipment is not stationary,
its operation is not controlled and may escape consideration, as shown in Figure 6-16.
The exhaust systems of off-road motorcycles and trucks are also liable to emit hot carbon
particles occasionally. Some vehicles may be required to have spark arrestors before being
permitted to operate in forest areas. Catalytic converters on vehicles have been known to
decompose and to spew chunks of their ceramic matrix at very high temperatures into
grass and brush.

FIGURE 6-16 Sparks

from this rail grinder were
thought to be responsible
for igniting a sawdust bin
in a lumberyard. Typical
water spraying would
extinguish fires on the
railroad right-of-way but
would not necessarily
affect hot sparks buried
in sawdust adjacent to
the track. Courtesy of Joe
and Chris Bloom, Bloom Fire
Investigation, Grants Pass, OR.

Chapter 6 Sources of Ignition 199

 FIREARMS RESIDUES
Forest and grass fires, as well as structure fires, are occasionally ignited by the burning
residues of firearms ammunition. Although modern smokeless powder is nearly com-
pletely consumed during the discharge of rifles and shotguns, some flakes escape while
still burning from the muzzle of a short-barreled handgun. Although they are low in mass
and thermal content, these flakes can ignite finely divided cotton and similar combustibles
at very close range (less than 0.3 m; 1 ft).
Black powder, in contrast, produces large quantities of burning particles of powder
and incandescent ash from both short- and long-barreled guns. Although generally lim-
ited to antique handguns, rifles, and cannons, the use of black powder for sporting pur-
poses is undergoing a strong resurgence in popularity.53 Because residues from such
weapons can ignite dry fuels like paper, cloth, wood shavings, leaves, and the like even
at distances of up to 1.3 m (4 ft), the possibility of accidental ignition by this means is
one the investigator must consider in identifying the cause of exterior, and even some
interior, fires.

Military Ammunition
More dangerous sources of ammunition fires are tracer and incendiary bullets, normally
used only in military weapons. Tracer ammunition contains a mixture of strontium
nitrate, magnesium powder, and an active oxidizer like calcium peroxide. Incendiary
ammunition usually contains white phosphorus or a similar material with a low ignition
temperature. Either type is a severe fire hazard because of the high-temperature flames
produced in the fired projectile. Both will initiate destructive fires in dry leaves, brush, or
needles (duff). Such ammunition has, on rare occasions, been used by arsonists to set
remote fires that are inaccessible to fire crews. (See Figures 6-17a and b.)
Virtually all military ammunition is marked by characteristic color bands on each
projectile, but some foreign military surplus ammunition brought into the United States
has been cleaned and repackaged, precluding ready identification.54 Hunters buying this
recycled ammunition have inadvertently discharged these missiles during hunting and tar-
get practice, sometimes with serious consequences. With the characteristic paint removed,
identification of such incendiary projectiles requires X-ray or visual laboratory examina-
tion. Signaling or rescue flares, some of which are available for even small-caliber hand-
guns, present the same sort of risk because they, too, contain pyrotechnic or incendiary

FIGURE 6-17A Unfired tracer bullet with orange nose paint. FIGURE 6-17B Tracer ammunition emitting bright red flare in a
Courtesy of John D. DeHaan. “bench” test. Courtesy of John D. DeHaan.

200 Chapter 6 Sources of Ignition

mixtures that burn for many (up to 40) seconds. Some have a parachute to slow their
descent and thereby prolong their burning time (and make them subject to winds aloft),
while others are simply ballistic “meteors.”55

Exploding Ammunition
Exploding ammunition, the projectile of which contains a percussion cap and a small
powder charge, has at times been sold to the public and represents an occasional fire dan-
ger. Foreign armor-piercing rifle ammunition with a steel insert has been thought to be the
cause of some wildland fires behind target ranges, because the steel is capable of striking
mechanical sparks as it ricochets off bare rock. Exploding paper targets have also been
implicated in accidental wildland fires.

Smoking as a Fire Origin

The widespread prevalence of smoking, as well as the general opinion as to its causation
of fires, calls for some special comments because there are many misapprehensions rela-
tive to this matter. There is no question that many fires are started by smokers, both
indoors and outdoors. Nor is there any question that the carelessness of the smoker is the
primary reason for this high incidence. However, it is common to hear theories about how
smoking caused a fire when the theory is totally inappropriate, and the alleged ignition
would not have been successful under the circumstances.
Smoking involves both burning tobacco and a flame used to ignite it. The roles of
matches and lighters as primary sources of ignition were discussed previously. Whether
the flame is from a match or a lighter, this flame will inevitably start a fire if it comes in
contact with suitable fuel. Thus, the discarding of a match into dry grass, before the
match is extinguished, has been the cause of many exterior fires. Both match flames and
lighters will cause explosions and fires with volatile flammables in the vicinity. Many per-
sons have attempted to commit suicide by turning on the gas, but then they wish a last
cigarette. Sometimes, the explosion that follows is their executioner, but more often it
puts them in the hospital with serious burns. It is important to realize only the flame
causes this type of incident.

CIGARETTES
In the authors’ opinion, cigarettes as a source of ignition have been blamed in many more
instances than warranted. It must be appreciated first that cigarettes vary considerably
from brand to brand in their burning characteristics. Modern tobacco cigarettes contain
a complex chemistry of additives to control the moisture content and burning rates of
both the tobacco and the paper. U.S. cigarettes sometimes used to include a chemical addi-
tive to keep them burning even when not being “puffed.” But in the last few years, ciga-
rettes designed to self-extinguish have become common, instead (see further).
The temperature distribution within the burning tobacco and its ash is complex and
difficult to measure. The most extensive work done in this area was by Baker, who
reported that surface (solid) temperatures reached 850°C to 900°C (1,560°F to 1,650°F)
at the edge of the glowing mass during the puff or draw using X-ray measurement tech-
niques.56 After the draw stopped, the hottest part of the mass shifted to the center, with
temperatures of 775°C (1,427°F), and the temperature of the ash at the edge dropped to
≈300°C (572°F), as illustrated in Figure 6-18. These temperatures are in general agree-
ment with test results reported by Beland and by the NFPA.57 Variations in the types and
cut of the tobacco used will affect the temperatures developed as well as the burning rates.
This is especially true of hand-rolled cigarettes, whether of tobacco or other vegetable
material, whose properties have not been studied in detail. In tests conducted by the
California Bureau of Home Furnishings, the temperature characteristics of many common
tobacco cigarette brands were measured. Maximum temperatures produced ranged from
Chapter 6 Sources of Ignition 201
Puffing
Light gases Sidestream Sidestream
diffusing out smoke gases

Mainstream Condensation B A
smoke and filtration
Direction of gas flow
Natural
Air convection 100 400 500 500 400 100
stream
Smolder
Oxygen 750
775
8 mm 600 °C
300 700 300
B A

Successive coal shapes

over a 15 -s period

FIGURE 6-18A Airflow in a cigarette at rest and while FIGURE 6-18B Temperature distribution in a burning cigarette at
being puffed. Reprinted by permission from Science & Justice. very start of puff. Reprinted by permission from Science & Justice.

400°C (760°F) for one variety to 780°C (1,440°F) for another. Temperatures rose by
as much as 100°C as the burning reached the end of each cigarette. Some brands of
cigarettes consistently self-extinguished when left unsmoked; others always burned to
completion.58 Although the rate at which a cigarette will burn depends on its orientation
and the direction and intensity of any draft, the rate was also found to vary from brand
to brand. Rates from 6 to 8 mm (1/4 to 1/3 in.) per minute and total burn times of 13 to
15 minutes for unaided horizontal burning were measured in a series of tests by the
California Department of Forestry.59
The heat release rate for an average tobacco cigarette has been reported as ~5 W. Such
a low level emphasizes the necessity for a fuel to be in direct contact with the smoldering
tip of the cigarette before ignition can be ensured. A lighted cigarette will only occasion-
ally set fire to dry fuel on which it rests uncovered, such as upholstery, dry grass, and the
like. A linear scorch will result, but it will rarely cause ignition. If the fuel is loose enough
that the burning cigarette can “burrow” into it as it burns, ignition is much more likely.
When the same cigarette is covered by even a single layer of light cloth, the insulation pre-
vents the radiation of the heat, and temperatures up to 100°C (212°F) higher can result.
This temperature difference may be sufficient to allow ignition of the fuel. If the cigarette
“burrows” down into loose cellulosic fuels or is pushed partially between two uphol-
stered cushions, enough heat may build up to cause ignition if the arrangement conserves
the heat without unduly restricting the air supply. On exposed polyurethane foam, which
is often very susceptible to flame, heat from the cigarette melts the rubber away from it
so that as the cigarette burns, it forms a trough in the foam (as in Figure 5-13b). In com-
bination with some fabric coverings, however, polyurethane can support a smoldering
fire.60 Latex rubber foams can be ignited by a cigarette in contact and support glowing
combustion that can transition to flame. Most modern upholstery foam rubbers are made
to resist such ignition even when intentional. On average, a great many lighted cigarettes
will have been improperly discarded for each fire that results from such a discard. (See
Chapter 11 on fabric flammability.) One study showed that cigarettes discarded into
crumpled paper trash transitioned to flames in only about 1 percent of tests.61 One well-
known fire investigator routinely demonstrates ignition of toilet paper rolls by placing a
cigarette inside the roll (as sometimes occurs in public toilets). These often transition to a
flaming fire within 20 minutes. He has been known to do this demonstration as an after-
dinner entertainment (much to the dismay of his wife).62 As discussed in Holleyhead’s
202 Chapter 6 Sources of Ignition
extensive review of cigarette ignition, many cellulosic products including toilet tissue,
cardboard, paper towels, and office carbon paper are all susceptible if the contact and
geometry are suitable.63 This has to be taken as an indication that cigarettes can cause
ignition of such fuels, and it takes only one ignition to start a fire!
It should be remembered that not all cigarettes have the same potential to ignite cel-
lulosic fuels. ASTM E2187 is a method for comparing the ignition potential of tobacco
cigarettes. Recent efforts to reduce accidental fires caused by cigarettes have focused on
redesigning cigarettes to make them self-extinguish if not being actively smoked (as do
cigars) by changing the paper or the tobacco.
The majority of U.S. states have passed regulations mandating that cigarettes sold in
those states be designed to self-extinguish when not being smoked.64 This legislation is
expected to have a significant impact on the frequency of cigarette-induced fires. These
“Reduced Ignition Propensity” (R.I.P.) cigarettes are designed to self-extinguish within 5
minutes of being left unsmoked. This means there is insufficient contact time to ignite
almost any susceptible fuel. In addition, Canada, Australia, and the European Union (EU)
have enacted similar legislation. Early tests have revealed that although such cigarettes
(employing bands of low-porosity paper) do not guarantee that all will self-extinguish and
therefore will not prevent all fires, the frequency of ignition is reduced.

BEDDING AND FURNISHINGS

While changes in the construction materials of modern furnishings have made them less
susceptible to ignition by cigarette, such fires still occur. Cigarettes and cigars, being
sources of smoldering combustion, are much more likely to ignite finely divided cellulosic
fuels like cotton padding than they are thermoplastics and foams that melt rather than
char. Years ago, mattresses and upholstered furniture were made of cushions of cotton
batting, kapok, or sisal (or latex foam rubber, which will also smolder), covered with nat-
ural fiber fabrics of linen or cotton. Thus, a dropped cigarette could burrow into fuel,
which would in turn support a smoldering fire for an extended period of time. Such fur-
niture is still frequently encountered.
Newer furniture is most often made with polyurethane foam cushions covered in ther-
moplastic fabric. Neither of these materials is likely to be ignited by a smoldering source
such as a cigarette. The cushions are sometimes wrapped in cotton batting instead of a
polyester fiber blanket, however, so the risk of cigarette ignition is not entirely absent. The
synthetics, of course, are much more susceptible to ignition by an open flame and, once
ignited, will support a much larger, more energetic fire than will the older furnishings.
Cotton sheets and blankets may similarly be ignited by cigarettes but only when they
are presented as a wadded mass of fuel. Single flat layers of such fabrics will be penetrated
by a smoldering cigarette but not ignited due to the heat lost to convection. Kapok, some-
times used as a pillow stuffing, is very susceptible to glowing ignition sources, but feath-
ers are not (although both will support smoldering once established). Decorative pillows
may be filled with anything and should not be overlooked as an intermediate fuel that is
ignited by a cigarette and then grows into a flaming fire that involves a urethane mattress.
The investigator must be aware of such complex interactions, must search the debris as
well as interview the owner or tenant, and must obtain samples of all the likely first fuels,
if possible.
Despite significant advances in improving the fire resistance of furnishings, smoking
in bed remains a threat to life safety. The reasons are simple. Bedding often contains cot-
ton, which is readily ignited by smoldering sources and is capable of sustaining a smol-
dering fire for an extended period. If the smoker should fall asleep, so that a lighted
cigarette drops into wadded sheets, pillowcases, or onto a cotton-stuffed mattress or pad,
a fire can readily result, and statistics indicate that such fires are common.65 A special
hazard of fires that involves upholstered furniture or bedding occurs when the drowsing
Chapter 6 Sources of Ignition 203
 (a) (b)

(c) (d)

FIGURES 6-19A–D Cigarette ignition of a traditional arm chair taking 42 minutes from placement to flames.
Note location of ignition and failure of other cigarettes on other surfaces. Courtesy of Jamie Novak, Novak Investigations,
and St. Paul Fire Dept.

smoker is awakened by the heat but, being confused, adopts incorrect measures to control
the situation. Such persons have been known to attempt to throw the burning bedding
out the windows or carry it to the exterior and in the process have extended the fire over
large areas of the building.
The time between ignition from a smoldering cigarette to open flaming combustion
of upholstered furniture is highly variable. One test, depicted in Figure 6-19, shows a
delay of 42 minutes between placement and flame ignition. The time depends on the
combination of cellulosic and synthetic materials and the extent and location of contact
with the cigarette. The National Bureau of Standards [now the National Institute of
Standards and Technology (NIST)] reported that of six upholstered chairs in one test,
three went to flaming ignition in 22 to 65 minutes, while the remaining three simply smol-
dered.66 The California Bureau of Home Furnishings reported that of 15 upholstered
chairs, 9 went to flaming combustion in times ranging from 60 to 306 minutes, 5 self-
extinguished, and one was still smoldering when extinguished at 330 minutes.67 Krasny,
204 Chapter 6 Sources of Ignition
Parker, and Babrauskas have published an extensive study of the variability of cigarette
and flaming ignition of furniture, and the reader is referred there for more information.68
In experiments observed by DeHaan, flaming combustion was achieved in older, cellu-
losic-filled upholstered furniture and bedding in times ranging from 22 minutes to over
2 hours from the placement of the burning cigarette. (If the fabric covering is torn, and
the cigarette is placed against the cotton stuffing, the time to flame can be as little as
18 minutes). The correct combination of fabric and padding (stuffing) can reduce the
ignitability of furnishings. Even cotton-padded furniture can be cigarette-ignition resist-
ant if covered with a heavy fabric made of thermoplastic (synthetic) fiber.69 If a piece of
furniture appears to have been ignited by a cigarette, samples of covering and padding
should be recovered for later examination.
An unpleasant accompaniment to bed and furniture fires is the danger of death by
burns or smoke inhalation, even in the absence of flaming combustion. The fact that many
persons who smoke in bed have also consumed excessive amounts of alcohol increases the
total hazard greatly.

CIGARETTES AND FLAMMABLE LIQUIDS AND GASES

Another fallacy about cigarettes relates to the danger that they will readily ignite flamma-
ble liquids or gases. Repeated attempts to cause explosions by inserting a lighted cigarette
into an explosive gasoline vapor–air mixture have resulted in failure, even with the ash
knocked off and the cigarette being puffed strongly. Tests by the authors involving throw-
ing or dropping lighted cigarettes into gasoline vapors above or alongside a spill on pave-
ment have never resulted in an ignition (until a lighted match was dropped into the same
vapor). This is not proof that such an explosion cannot happen, but any claim that it has
occurred should be viewed with the greatest skepticism.
In his extensive review of the dynamics of tobacco cigarettes and flammable vapors,
Holleyhead describes several factors revealed by the tests of Baker and others that make
ignition of gasoline vapors by a glowing cigarette most unlikely.70 Tests have shown the
oxygen levels in the cigarette in the vicinity of the combustion zone to be very low and
carbon dioxide levels to be very high, both factors that significantly reduce the chances of
vapor ignition, as illustrated in Figure 6-20. Furthermore, the residence time of airborne
vapors in the cigarette being puffed is so short that there is not enough time for any but
the most reactive species to ignite. The spacing between the fuel elements in a commercially
made cigarette is such that the quenching distance of all but the most reactive gases is not

Direction of gas flow

Direction of gas flow

10 8 8 4 6 2
18 6 6 8 8 12 16 20

16

10 12 14
18 14 12 10 8 4 2 0

16 6 6 8 8 8

⫺8 ⫺4 0 ⫹4 ⫹8 ⫹12 ⫺4 0 ⫹4 ⫹8
Distance from line of paper burn (mm) Distance from line of paper burn (mm)

FIGURE 6-20A Oxygen concentrations in a cigarette—at the start of FIGURE 6-20B Carbon dioxide levels are very high whether the
a concentration in a cigarette at the start of a “puff.” Note extremely cigarette is being actively smoked or not. This will inhibit ignition
low O2 concentration throughout high-temperature zone preventing of all but the most sensitive gases. Reprinted by permission from Science
ignition of gases/vapors. Reprinted by permission from Science & Justice. & Justice.

Chapter 6 Sources of Ignition 205

 exceeded, suppressing any sustained ignition. It has been suggested by Kirk that the ash
surrounding the burning tobacco may act in the same manner as the screen around a
miner’s safety lamp, acting to suppress any flame by quenching it between closely spaced
elements. Finally, while the surface temperatures of the burning tobacco inside the cigarette
tip are high, the presence of incombustible ash may reduce the efficiency of heat transfer
to the fuel vapor.71 These heat transfer considerations are validated by the published test
data, which showed that only hydrogen, hydrogen sulfide, carbon disulfide, acetylene,
ethylene oxide, phosphine, and diethyl ether are susceptible to cigarette ignition, while
gasoline vapors and methane are not.72 Recent tests by the ATF Fire Research Laboratory
involved contact of burning tobacco cigarettes of different brands with gasoline vapors
from a pool at room temperature. A total of 137 attempts were made using both smolder-
ing (passive) and actively drawn (puffing) cigarettes with no ignitions observed. Further
testing involved multiple applications of 157 burning cigarettes to or near a gasoline-
saturated cotton blouse with no ignitions observed.73 It must be noted that contaminants
or irregularities in the filling or in the paper may occasionally promote a very brief, tiny
flame. Such a flame would be capable of igniting a suitable fuel/air mixture. Although it
is virtually impossible for a lighted cigarette or cigar to ignite most flammable vapors, this
is not the case if either is actually lighted in such an atmosphere. That would ensure the
ignition of a fire, explosion, or both. Recent efforts to reduce accidental fires caused by
cigarettes have focused on redesigning cigarettes to make them self-extinguish if not being
actively smoked (as do cigars) by changing the paper or the tobacco.
The damage to clothing or upholstery when it is scorched or a hole burned through
it by a cigarette is annoying, but at least this damage will very rarely ignite the fire that
burns the house. In any instance where smoking is claimed as the source of a fire, very
careful attention must be given to assure that the claimed fuel is one that could be lighted
by burning tobacco or, alternatively, the flame from a discarded burning match, which is
actually a much more likely source. Consideration must also be given to time factors. Due
to their low HRR, cigarettes require contact time to transfer enough heat to cause igni-
tion of another fuel. This generally means at least 15 minutes have to elapse. (A slight
draft, however, has been observed to shorten this time by 2–3 minutes.)74 The authors
have confirmed those ignition times by their own tests and that only incendiary materials
(like gun cotton, gunpowder, or flash powder) can be ignited immediately on contact with
a glowing cigarette. The use of medical oxygen by a smoker makes all materials saturated
with oxygen much more ignitable by smoking materials. This includes clothing, bedding,
facial hair, and the plastic mask and cannula (as seen in Figure 6-21).

PIPES AND CIGARS

Pipe and cigar smoking is not basically different in its fire potential from cigarettes. Cigars
tend to go out when not puffed, but this does not prevent some fires from being ignited
when cigars fall onto a suitable target. Discarded pipe embers, like lighted cigarette embers,
can burn their way through crumpled papers in a wastebasket until the embers lodge where
there is thicker fuel (and ventilation) and then ignite it. Pipe ashes will also scorch uphol-
stery or clothing, actually burning holes in cloth in many instances. They rarely cause flam-
ing fire except with sawdust, cotton packing, or similar readily ignitable materials. Hot
pipe and cigar ashes are often physically larger than cigarette ashes and have more heat to
transfer to fuels they contact but do not sustain combustion on their own for very long.

PLANTINGS
Finely divided cellulosic materials can be encountered in gardens and decorative plantings
(of both real and artificial plants). Dry potting soils, peat moss, and Spanish moss have been
shown to be readily ignitable by contact with a smoldering cigarette. Smokers often assume
such plantings are in regular (mineral or garden) soil and are therefore noncombustible.75
206 Chapter 6 Sources of Ignition
 FIGURE 6-21 Oxygen
mask readily ignited by
match flame burns with
great intensity. Courtesy
of Jamie Novak, Novak
Investigations, and St. Paul
Fire Dept.

Spontaneous Combustion (Self-Heating)

The claim is frequently made that a fire started from spontaneous combustion. This sug-
gestion at first seems to violate the basic premise that there must be fuel, oxygen, and heat
before there can be combustion and implies that in some mysterious manner fires simply
set themselves on occasion. The term spontaneous chemical causation would be more
accurate, for in such cases the mass of fuel, rather than igniting from one point or even
one surface, ignites throughout its mass at very nearly the same time; and the heat has
been accumulating, not from some exterior source, but from chemical processes occurring
within the fuel mass. When such combustion occurs, it is a wonderful example of the bal-
ance of physical and chemical processes that occur in all fires but on a time scale that
allows their individual roles to be better appreciated. Taken as a whole, however, their
interactions make them hard for the fire investigator to understand. The single driving
force is an exothermic reaction generating heat, and if that heat is not dissipated by
normal mechanisms as fast as it is made, it can build up in a fuel mass and raise the
temperature of the mass. Because reaction rate generally doubles with every increase in
temperature of about 10°C (18°F), the reaction, by generating heat, tends to make itself
go faster and faster with a constantly increasing heat production rate. When there is insuf-
ficient dissipation of the heat, the temperature may rise to the point of ignition of the
entire reacting mass. The process is treated in more detail in the fire investigation litera-
ture.76 Not all chemical reactions are sufficiently exothermic (self-heating) to cause such
ignition, and not all exothermic reactions are sensitive enough to temperature rises
to drive them to ignition. A list of commonly encountered materials, rated according to
their susceptibility to self-heat, is published in the NFPA Fire Protection Handbook, but
a few are discussed here by way of examples.

CHARACTERISTICS OF SELF-HEATING
The general rule that applies to self-heating is that the mass of material needed to accom-
plish ignition is inversely related to the reactivity of the material, and the time needed is
directly related to the mass required (i.e., the lower the reactivity, the larger the necessary
Chapter 6 Sources of Ignition 207
 mass and the greater the time needed). Catalytic reactions (such as the setting of fiber-
glass resin) require only a very small mass because the reaction is highly exothermic and
self-ignition ■ See can take place in a few seconds. Drying oils can trigger self-ignition in a few hours and
autoignition. require as little as 25 to 50 ml (1 to 2 fl oz). Activated charcoal can self-heat in masses
of a few pounds and requires several hours to a few days. Laboratory studies have
demonstrated that charcoal barbecue briquettes are unlikely to self-heat to ignition
unless presented in large masses [more than 50 lb (20 kg)] or at high ambient tempera-
tures (over 100°C).77 However, these data must be understood as applying to pure and
uncontaminated charcoal briquettes. Credible reports exist of fires that originated from
small bags of charcoal briquettes where external ignition sources (e.g., cigarettes) were
successfully excluded. These fires appear to have been due to briquettes contaminated by
chemical substances that promoted combustion. Some charcoal briquette bag ignitions
have taken place, however, after an unwise individual returned “hardly burned” briquettes
to the bag.
Hays and grasses require significant masses (100 kg/220 lb or more) and days to
weeks before ignition, even at moderate temperatures. Coal and bulk materials will not
self-heat to ignition unless they are present in very large masses (many tons) and given
weeks or even months. The higher the starting temperature, the faster the process can
progress. Cotton laundry that has not been allowed to cool down properly from drying
has been known to ignite after just a few hours. The presence of cooking oils that have
not been removed by laundering may contribute to self-heating of laundry, as may
residues of bleaches (oxidizers). It has been suggested that self-heating of residues of cook-
ing oils in clothes dryer lint may lead to some fires in clothes dryers.78 Tests have revealed
that triglyceride oils such as corn, cotton, or linseed oils are not very soluble in water and
will be retained in clothes after laundering (and may be detectable in the residual water
in the washing machine pump if the clothes in the dryer have been fire damaged.79
Ground-up or chipped plastics being stored for recycling have been involved in a number
of apparent self-heating incidents. The example shown in Figure 6-22 was linked to cooking

FIGURE 6-22 An opened

box of chipped plastic
shows deep charring in
the center of the mass.
Self-heating was linked to
cooking oil contamination
and grinding and storage
in bulk in high ambient
temperatures. Courtesy of
Edward Garrison, Fire/Explosion
Investigator, Raleigh, NC.

208 Chapter 6 Sources of Ignition

oil contamination in the plastic bottles that were ground up. The material was stored in
large (4 ⫻ 4 ⫻ 4-ft; 1.2 ⫻ 1.2 ⫻ 1.2-m) boxes with plastic liners for extended periods of
time. Two ignitions were discovered (in late July in very hot weather) a week apart. The
first (shown here) was intercepted, but the second one led to a fire that destroyed the
entire warehouse. The grinding or shredding process creates a great deal of frictional heat
that can promote self-heating; also, metal or rock contaminants can create hot particles
that can initiate ignition deep within the mass.
Some suspected sources of spontaneous heating can be tested by subjecting them to
the Mackey test, which was developed to assess the hazards of oils used to improve the
weaving qualities of cotton and linen fiber, or, preferably, by establishing their critical
temperature for various mass sizes and shapes.80 Some processes can be tested by differ-
ential scanning calorimetry (DSC) to determine the rate at which they yield heat. The like-
lihood that a process can be an ignition source can then be evaluated.81 Because the
environment has so much to do with self-heating (porosity, oxygen supply, thermal con-
ductivity, heat transfer coefficient, etc.), simple chemical or thermal tests of materials may
not reveal all their tendencies to self-heat. Although time-consuming, the most accurate
analysis is to place various masses of the suspect material into ovens and expose them to
slowly increasing environmental temperature. Internal temperature monitoring of the
mass will reveal when runaway self-heating occurs. Detailed laboratory procedures and
data analysis methods are described in the Ignition Handbook.82 Brian Gray warned
about simple reliance on small-scale tests in predicting self-heating in industrial situations
where products have relatively high critical temperatures when tested in small, single
packages but whose critical temperatures as stored “assemblies” are considerably lower.
He cited two cases in which hot storage conditions of bulk product changed the thermo-
dynamics in critical fashion. In one case, drums of calcium hypochlorate (whose genera-
tion of oxygen is an exothermic reaction) were stored in an exterior storage container and
produced a runaway self-heating reaction. In the other case, small packages of instant
noodles were originally packed at a packing temperature of 81°C (177°F), below the esti-
mated critical temperature for the shipping assembly (in cartons on pallets) of 106°C
(222.8°F). A change in manufacturing process caused the product to be packed at 113°C
to 120°C (235°F–248°F), a change that caused a massive fire loss.83
In one well-documented case in New York, a very large warehouse fire was traced to
a storage area where many cases of latex rubber surgical gloves had been stored for sev-
eral years. Because the gloves were packaged in separate boxes in each case (as shown in
Figure 6-23a), investigators could not believe that the latex had self-heated until they
found cases on the periphery of the main fire whose contents displayed the entire range of
self-heating damage from mildly discolored to deeply charred (see Figures 6-23b and c).
The large masses of gloves had been stored in an unventilated area, and the high summer
temperatures finally tipped the balance toward runaway self-heating.84
Powders (dusts) or dried aerosols (from catalytic-setting paints) have been seen to
accumulate on ventilation filters or hot surfaces (such as lamps) and develop into self-
heating masses, causing fires. These processes are often aided by airflow and elevated tem-
perature of the substrate.85

SELF-HEATING OILS
The most commonly encountered consumer product that can self-heat to spontaneously
ignite is a “drying oil” such as linseed (extracted from flax seeds), tung nut, fish, or soy-
bean oil. As the name implies, such oils harden by oxidation of the double bonds of the
fatty acids in the oils, particularly the linolenic acid component.86 As they oxidize they
polymerize to form a tough, natural plastic coating. This coating property is why these
oils have been used in paints and finishes for centuries. The oxidation process relies on
oxygen in the air and generates heat. The reaction rate depends, then, on the amount of
Chapter 6 Sources of Ignition 209
FIGURE 6-23A Self-
heating of latex surgical
gloves stored for years in
a warehouse. Courtesy of
U.S. Department of Justice,
Bureau of Alcohol, Tobacco
and Firearms and Explosives.

FIGURE 6-23B Showing

gradations of damage in
cartons. Courtesy of U.S.
Department of Justice, Bureau
of Alcohol, Tobacco and
Firearms and Explosives.

FIGURE 6-23C Note

limited damage to individ-
ual boxes. Courtesy of U.S.
Department of Justice, Bureau
of Alcohol, Tobacco, Firearms
and Explosives.

210 Chapter 6 Sources of Ignition

drying oil that is presented to the air. A quantity of such oil in a can will not self-heat to
a detectable degree, but if the oil is spread as a thin layer on a porous substrate like cot-
ton cloth, there can be enough surface area to promote rapid oxidation and heating. If
this heat is trapped by insulation in such a manner that oxygen can still diffuse into the
mass, the reaction rate will rise.
Raw linseed oil will demonstrate this effect but often will not spontaneously ignite,
because bulk raw linseed oil contains oleic and linoleic acids, which are not as reactive as
linolenic acid. Raw linseed oil does not dry as fast and hard as boiled linseed oil, which
is made by boiling the product to concentrate the amount of the less saturated, more reac-
tive fatty acids or, more commonly, by adding a drying agent, most often a metallic oxide.
The drying agent triggers rapid setting (and thereby more rapid heat generation). The
effect is that a quantity of cotton rags smeared with boiled linseed oil can self-heat to the
point of autoignition when kept at room temperature for a period of just a few hours (as
pictured in Figures 6-24a and b). No containment or external insulation is necessary, for
the reaction is enhanced by the availability of oxygen, and enough insulation is provided
by the overlying rags. Such a reaction has been observed by DeHaan to occur in a pile of
rags bearing an estimated 100 ml (4 fl oz) of boiled linseed oil, piled in the open against
a concrete wall at an ambient temperature of 25°C (77°F) in just 5 hours.87
Howitt et al. reported that as little as 25 to 50 ml (1 to fl oz) is adequate if other fac-
tors are optimum.88 Reactions monitored in one series of tests revealed temperatures in
excess of 400°C (750°F) within the mass of rags, and these are confirmed by Howitt’s
observations.89 Older literature is consistent with these estimates. Clearly, once the mass
of rags bursts into open flame, there will be enough heat to ignite other ordinary fuels in
the vicinity. This is the mechanism thought to be responsible for the fire in the Meridian
Plaza high-rise office building, and has been identified as the heat source in numerous
fires in furniture shops and home workshops where wood-finishing products containing
boiled linseed oil were used and residues wiped up with cloth rags.90 The key elements
that must be present are presence of a drying oil (most commonly of vegetable or animal
origin), a porous support medium that allows free access of oxygen to diffuse into the
mass and that does not melt when the temperature rises, provision of an adequate oxy-
gen supply, and time for the reaction to occur. The higher the ambient (or starting) tem-
perature, the faster the reaction will proceed. If the ambient temperature is below about
10°C (50°F), linseed oil is unlikely to self-heat to flaming ignition. Many wood stains and
varnishes contain both a petroleum solvent and a linseed oil base. The petroleum solvent
plays no role in the self-heating process except to slow it down by evaporative cooling. If
the product is sprayed and then wiped up, the solvent will largely have evaporated and
will not serve as a delay mechanism.
Self-heating does not occur with hydrocarbon (lubricating) oils or animal lard (except
at unusually high ambient temperatures) and does not usually occur with the cooking oils
commonly found in household kitchens [e.g., peanut, corn, olive, safflower, rapeseed
(canola) at relatively low ambient temperatures]. One exception occurs when such oils are
absorbed into porous foods like flour or onto cotton towels and are then exposed to high
starting temperatures. This mechanism is very likely when the oil has been used to deep
fry fish. Then, the effluent will contain fish oils (that can self-heat), and porous bread or
batter crumbs, all at high temperatures.91 Rags used to wipe down hot cooking griddles
with various oils and then thrown into waste (or laundry) bins while still hot can also self-
heat to ignition. Self-heating fires have occurred when hydrocarbon oils leaked into
porous insulation lagging covering pipes running at an elevated temperature, but such
fires will occur only in an industrial environment.92
Gaw reported that stacks of cotton towels oiled with linseed oil, then laundered and
washed and stored at 93°C (200°F) self-heated to 575°C (1067°F) and then ignited into
flaming combustion. He noted that significant oil remained in the towels after launder-
ing, that bleach had no effect on the amount of oil removed, and that residual oil was
Chapter 6 Sources of Ignition 211
FIGURE 6-24A Sponta-
neous ignition of cotton
rags with linseed oil. Note
burning from center of
mass. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

FIGURE 6-24B Linseed

oil rags smoldering as
they are removed. Notice
discoloration. Courtesy of
John Galvin, London Fire
Brigade.

FIGURE 6-24C Moments

later, rags ignite in flames
as they are exposed to air.
Courtesy of John Galvin, London
Fire Brigade.

212 Chapter 6 Sources of Ignition

recovered from the clothes washer pump even after laundering oiled towels and then a
full load of nonoiled towels.93 A more extensive review of self-heating processes and the
investigation of spontaneous ignition fires has been published by DeHaan.94
The reactivity of oils is measured by their so-called iodine number. While there are
exceptions, generally, the higher the iodine number, the more susceptible the oil is to self-
heating. For example, linseed oil has an iodine value of 173–201, tung oil: 150–176, soy-
bean oil: 137–143, cottonseed oil: 108–110, corn oil: 111–130, and peanut oil: 85–105.95
The NFPA has rated many common products (both natural and synthetic) with respect to
their susceptibility to self-heating, and the reader is referred there for specifics.96 Stauffer
has carried out extensive work in the chemistry and forensic identification of these prod-
ucts and the reader is referred there for more detail.97

SELF-HEATING OF VEGETATION
Aside from occasional serious fires that occur in the storage of chemicals, the most com-
mon fires termed spontaneous are without doubt those that are ignited in hay, grasses,
bagasse, or other vegetation residues. The phenomenon has been known since ancient
times, and considerable experimental work has been performed in attempting to elucidate
its mechanism.98 A case of spontaneous fire in a hay mow was observed and recorded in
detail by Ranke in 1873, who was himself a scientist and who performed further experi-
ments in support of his theory of strong adsorption of oxygen by pyrophoric carbon. This
was apparently the first detailed and well-documented description and served as the start-
ing point for nearly all later work on the subject.99
There are several sources of heat that can contribute to self-heating of hays, grasses,
and the like. First, there is the heat of respiration of microbial populations and living
materials (such as freshly cut grass). This can raise the temperature of a fuel mass to over
70°C (160°F), at which temperatures the cells and microorganisms die. In order for most
vegetable materials to undergo self-heating, there must be a favorable moisture content to
allow a fermentative process to proceed. Unlike the fuels already described, which
undergo chemical changes, self-heating of vegetable materials is often a biological process
dependent on oxygen and living microorganisms, at least at first. If the hay is well cured
(dried), it will not allow destructive fermentation; if it is too wet, not enough oxygen can
diffuse into the mass, too much heat is conducted out, and no fire will occur. Partially
cured (dried) hay is best for self-heating (moisture content between 12 and 21 percent).
The activity of the microorganisms will heat the hay to the thermal death point of the
organisms as a limiting temperature. This, too, is in the neighborhood of 70°C (160°F)
and lower for many organisms. Regardless of insulation that exists in a large mass of the
vegetable material, temperatures in this range fall far short of the necessary ignition point
of the hay, which is on the order of 280°C (540°F) and above.
The biological temperature limit makes clear that microorganisms can raise the temper-
ature only to a point at which some exothermic chemical process(es) can be initiated to raise
the temperature much higher. This phase of the process has been the subject of much atten-
tion and has produced various theories, including the production of “pyrophoric carbon,”
pyrophoric iron, heat from enzyme action, and even auto-oxidation of the oils contained in
seeds. It has been reported that much acid is generated in the early stages of the process,100
and that this is accompanied by marked browning of the hay. This high acidity has been
used as a means of determining whether spontaneous combustion has occurred as con-
trasted with external ignition. Perhaps the most widely accepted theory is that of Browne,
who postulated the formation of unsaturated compounds by the action of
microorganisms.101 These could form peroxides with oxygen of the air, and later form
hydroxy compounds along with heat. Others do not agree that acidic degradation products
are formed.102 Rathbaum’s work explored the complex role of temperature, ventilation, and
moisture content for hay and other grasses.103 There are several possible chemical routes
Chapter 6 Sources of Ignition 213
 that do not rely on biological mechanisms. Gray has demonstrated that bagasse (the waste
material from sugar cane extraction) can self-heat in the hot, dry climate of inland Australia
even in the absence of any microorganisms.104 Apparently, between 70°C and 170°C
(160°F–340°F), chemical decomposition reactions supported by moisture dominate, and
above 170°C (340°F) oxidation reactions dominate as long as oxygen is available.
Whatever the actual mechanism may be, it is certain that it consists of a number of
steps, the most important being chemical in nature. The packing density of the hay or
grass, the size of the stack or bale, and its moisture content when left are critical factors.
Other facts that are of importance to the investigator concern the environment—rainfall
prior to ignition and ambient temperature of storage. Spontaneous fires in stacked hay do
not usually occur in less than 10 to 14 days after stacking and generally require 5 to 10
weeks [under ideal conditions of moisture and very high ambient temperature
(35°C–40°C; 95°F–105°F), ignition could be observed in 6 to 7 days].105 The fire that
occurs is in the center of the stack and burns to the exterior, usually forming some type
of chimney or flue to the exterior. A fire from an outside source (accidental or intentional)
burns toward the inside.106 Unburned hay in a stack in which spontaneous combustion
has occurred will often be very dark in color (with a color gradation ranging from gold-
brown to chocolate colored to blackened to char) and may have a higher acidity than nor-
mal hay. Odors described as caramel-like or tobacco-like have been described. If a
burning stack is spread apart, hay that is not on fire may suddenly burst into flame when
fresh oxygen is introduced if spontaneous combustion is the cause of the fire (which pre-
heats the fuel mass). Depending on the geometry of the stack, some exterior fires may pre-
heat the interior. This is not expected to occur when the hay has been fired exteriorly, but
small interior fires have been found after deliberately ignited test fires.
One of the interesting artifacts of spontaneously ignited hay fires is the formation of
the “hay clinker” (see Figure 6-25). This glassy, irregular mass, gray to green in color, can
be found throughout the pile wherever temperatures were the highest and most prolonged
but are likely to be concentrated near the center. The clinker is composed of the inorganic
residues of silicon, sodium, and calcium from the plant stems as well as impurities baled
with the hay (soil, dust) fused into a glassy mass by the heat developed in the fire. Hicks
has done an in-depth study of hay clinkers that strongly indicates that the inorganic con-
stituents of various grasses may be linked to their propensity to form such glasses.107

FIGURE 6-25A Hay clinker from haystack fire ignited by spontaneous FIGURE 6-25B A much larger hay clinker from a haystack fire
ignition. Clinker is gray-green in color, foamy-glassy in appearance, contains some carbonized stems, as well as the glassy residues of
and is a natural glass formed by the oxides of silicon, calcium, the inorganic oxides. Courtesy of Ryan B. Fields, Orca Fire Investigation,
magnesium, and sodium present as trace elements in grasses and Medford, OR.
feedstocks. Courtesy of John D. DeHaan.

214 Chapter 6 Sources of Ignition

Recently, Tinsley, Whaley, and Icove performed a series of live fire experiments examining
the formation of hay clinkers in externally ignited fires. The result of that study indicated
that the formation of hay clinkers is dependent on a variety of factors, none of which reli-
ably indicate the origin or cause of the fire.108 Hay fires in pole barns with metal roofs may
cause the formation of numerous large masses as opposed to a few smaller clinkers.
Clinkers do not occur in all self-heating hay fires, so their absence cannot be taken as proof
of incendiary cause, and their presence should not be taken as proof of spontaneous
origin. A hay clinker may be mistaken for the residue of an incendiary device, but it is
innocuous. Elemental analysis is essential. It must also be remembered that spontaneous
combustion is an easy decision to reach when the cause of the fire is not known. It seems
certain that many fires that have been attributed to spontaneous causes were actually set.
When other materials of vegetable origin such as fiberboard are involved, some input
of external heat is necessary to trigger enough oxidation. Fiberboard, pressboard, and
plywood are made by subjecting a wood product and a resin to high pressures at high
temperatures. When these products are manufactured, if they are not stacked to allow
excess heat to escape, residual heat remaining in the stacked mass may trigger self-heating.
Sawdust piles and other materials of wood have also been thought to ignite spontaneously,
but they are unlikely to do so unless there is a drying oil or similar external contaminant
present. Sanding waste from freshly varnished wood has been linked to self-heating, triggered
by residual heat from the sanding process. This tendency is aggravated if the wood has
been recently coated with a catalytic polymer finish whose setting involves an exothermic
reaction, not just simple evaporation of a solvent.

OTHER MATERIALS SUBJECT TO SELF-HEATING

In addition to the two best-known examples of systems subject to spontaneous ignition,
oily rags and hay, a number of other fuel systems may ignite spontaneously. One of the
more troublesome and common ones is coal, lying unmined in the ground or at times in
large piles or storage bins above ground, since the wetting of coal can cause generation
of heat.109 As with hay, exposure to water and the accumulated moisture content has
been found to be critical to self-heating.110 Without doubt, the process itself is a slow
exothermic chemical one in which heat is generated faster than it can escape in an envi-
ronment that is well insulated. Other factors of importance are the type of coal (some
grades of which ignite far more readily than others) and the coal size (finer dusts being
more susceptible). Massive quantities are needed, and days or weeks are required before
flaming ignition can be achieved.

IMPLICATIONS FOR THE FIRE INVESTIGATOR

There is a temptation to label many accidental fires as “spontaneous” because there is no
identifiable ignition source or obvious human intervention, but this is not correct. If the
materials and processes cannot be specifically characterized as susceptible to self-heating
under the prevailing conditions, then the cause must be considered to be unknown.
Spontaneous ignition (with very rare exceptions) does not occur instantaneously, and the
time frame for development is linked to the chemistry and mass of the reactant. Flaming
ignition is always preceded by smoke and odors that should be detectable by anyone in
the vicinity for some time prior to flaming ignition.

Other Sources of Ignition

Although accidental fires are invariably ignited by a small flame, electric arc, burning
fragments, or other hot objects, there are some sources of ignition that do not fall neatly
into those classes. In this category are lightning, spontaneous combustion, electric light-
ing, discarded batteries, and animal interaction with sources of ignition.
Chapter 6 Sources of Ignition 215
 LIGHTNING
Long before humans learned to start a fire, it is almost certain that their knowledge of fire
came from fires that were ignited by bolts of lightning. Lightning is a tremendous electri-
cal discharge of natural origin that is not fundamentally different from the electric arc,
except in intensity. It is a massive discharge of the huge static electric charges built up in
clouds of moisture or dust. The clouds and the earth act as the plates of a condenser or
capacitor; the potential across them can reach up to 100 million volts. The discharge,
when it finally occurs, usually consists of several strokes back and forth, each lasting a
few millionths of a second and 40 to 80 thousandths of a second apart. The currents
involved are typically around 20,000 amperes, although much higher values are occasion-
ally reached.111 Lightning strikes are usually accompanied by the physical destruction of
any poor electrical conductor in their path. Windows can be blown out, wood is splin-
tered and shattered, and electric appliances can be exploded (see Figures 6-26a–c). The
air in the path of the main stroke can be heated to a temperature of 30,000°C (54,032°F)
and is expanded at a supersonic speed as a result. The resulting pressure shock wave can
damage nearby structures with explosive force. In a lightning flash, the heat generated by
the flow of current through a struck solid object will often ignite it. This is especially true
when the object is dry and readily ignited. The loss of buildings to lightning-induced fires
is still large, although protective devices now available have markedly reduced this loss
within the last 50 years or so.
Lightning and Trees
Trees are ignited only occasionally by lightning, the result being determined chiefly by
their susceptibility to combustion. An old, dead, dry trunk, for example, is much more
likely to catch fire than a living tree, which is both a better conductor and less flamma-
ble. In general, many bolts of lightning will strike for every fire that is initiated. In
regions where thunderstorms are common, however, the risk is very great. The probabil-
ity of ignition of ground fuels is controlled by their fuel moisture and the depth of the
accumulated ground layer.112 In most instances, such fires are readily traced. When a fire
of undetermined origin occurs in a remote area of mountain or forest, the investigation
as to whether it was caused by lightning may be difficult. Meteorological records will
generally establish whether there was a storm in the questioned area at the right time.
Some forestry services maintain a special lightning watch to monitor and record light-
ning strikes in wildland areas, and there are meteorological satellites that record light-
ning activity. A lightning detection service can be of help in confirming or eliminating
lightning as an ignition source in the United States. Vaisala Inc. maintains a lightning
strike service called StrikeNet, which uses a network of electronic detectors. Its reports
claim accuracy of 500 m as to location and can be requested online for particular time
frames and various distances from a street address or GPS location. A typical StrikeNet
report is shown in Figure 8-19. Search of the area may uncover a tree or other object
that was split or damaged by means other than the fire itself. If this finding corresponds
to the origin of the fire as indicated by study of the pattern, it is very probable that light-
ning was the origin.

IMPLICATIONS FOR THE FIRE INVESTIGATOR

The investigator needs to be familiar with the effects of lightning when it strikes objects.
The most common effects are splitting of wood and mechanical disruption of structures.
Most of these effects arise from sudden volatilization of water inside the tree or structure
by heat from the electric current, so that the effect is that of an internal explosion. The
effects are extremely variable, being determined by both the intensity of the electrical dis-
charge and the localized condition of the object that is struck. (Figure 8-18 shows the
fused earth produced by a lightning strike in soil.) Lightning strikes can induce power
surges in nearby power or phone lines (causing failure and sometimes ignition of appliances
216 Chapter 6 Sources of Ignition
 FIGURE 6-26A Lightning
strike on house roof. Fire
damage to roof where
struck. Courtesy of Greg
Lampkin, Knox County TN
Fire Investigation Unit.

FIGURE 6-26B Limited

fire damage, down pipe
dislodged. Courtesy of Greg
Lampkin, Knox County, TN,
Fire Investigation Unit.

FIGURE 6-26C Interior

wall damage as lightning
followed electrical cables.
Courtesy of Greg Lampkin,
Knox County TN Fire
Investigation Unit.

Chapter 6 Sources of Ignition 217

218

267(513)
202(396) 230(446)
111
101 263(505)
208(407) 91(196) 247(477) (232)
(214) 144(291) 189(282)
137(279) 133(271) 138(280) 145(293)
55(131) 53(127) 52(126)
54(130) 51(124) 55(131) 49(120) 56(133) 51(124)
52(125) 52(126) 54(129)

207(405) 238(460) 254(489)

57(134) 66(151) 64(147)
53(127) 59(138) 62(144)
81 60-W lamp 90 100-W lamp 100 200-W lamp
(178) A-19 bulb (194) A-19 bulb (212) A-23 bulb

87(188) 78(173) 92(198) 83(181) 95(203) 90(194)

182(360) 238(460)
252(486)
102(216) 133(271)
147(297)
89(192) 98(208) 105(221)
94(201) 109(228) 108(226)
75(167) 78(172)
125(257) 149(300) 93(199)
175(347)
77(171) 88(191)
74(165) 72(162) 105(221)
93(200)

FIGURE 6-27 Surface temperatures of bulb and socket for typical incandescent lamps in open air (unlimited circulation) as a function of position. (Numbers in
parentheses are temperatures in degrees Fahrenheit.) Courtesy of GE Lighting.
connected to them). Examination of appliances (including computers and telephones)
throughout a structure or even in adjacent buildings is important in verifying the role of
lightning (or lone power surges) in igniting a fire, as all appliances plugged in at the time
are likely to be damaged.

IGNITION BY ELECTRIC LIGHTING

Incandescent lightbulbs can generate temperatures in open air at the surface of the glass
bulb or envelope ranging from 74°C (165°F) to 267°C (513°F) depending on the wattage
of the lamp and its position, as seen in Figure 6-27.113 Such temperatures can scorch cel-
lulosics or melt synthetics that come into contact but would be unlikely to result in flam-
ing ignition. However, much higher temperatures can be produced if the bulb is enclosed
in a housing or wrapped or buried in insulating material. Bulbs of low wattage normally
produce a surface temperature inadequate to ignite common combustibles. As the wattage
is increased (especially with bulbs in housings of some type that restrict ventilation), tem-
peratures can build up to quite high values. Thus, potentially, a large incandescent light
bulb (⬎50 W) can start a fire if it is in prolonged contact with suitable fuel. As was
pointed out earlier in this chapter, the circumstances of contact between a heat source and
a fuel control whether there is ignition. If a hot surface is placed in contact with a fuel
long enough, the surface may provide insulation and reduce the heat losses from that sur-
face, thereby increasing its temperature and its potential for ignition. (See Figure 6-28).
Gruehn recorded smoldering ignition of a wood floor by prolonged contact of a 75-W
lightbulb resting on it.114 Even low-wattage lamps (⬍15W), if buried or wrapped in an
insulating material, can generate sufficient temperatures to ignite paper, cloth, sawdust,
grain, and the like. Such events are uncommon, not because of the lack of a high enough
temperature, but rather, because something easily ignitable (such as a cellulosic material)
must be in contact with the bulb, and this rarely occurs. The most likely areas for such
ignitions are basements, storage rooms, and shops when packing material is allowed to
be in contact with an operating bulb.
In tests conducted by DeHaan and Albers, a 60-W incandescent bulb buried in crum-
pled newspapers ignited them to flames in just 10 minutes 5 seconds. A 75-W bulb buried

FIGURE 6-28 A 75-W

lamp igniting cellulose
insulation. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

Chapter 6 Sources of Ignition 219

 in sawdust (uncontrolled fuel moisture) caused smoke in 6 minutes 13 seconds, but no
flames were observed prior to suspension of the test at 28 minutes. There was melting and
smoke generation of polyester fiberfill burying a 100-W bulb, but no ignition for 32 minutes.
A 150-W decorative lamp had no effect on polyurethane foam resting atop it for some 30 min-
utes due to the large surface area and limited contact area but a 100-W lamp buried under
polyurethane foam caused ignition after some two hours (see Figures 6-29a–c and 6-30).115
The breakage of any lit incandescent lamp, even low-voltage and low-wattage ones,
introduces two additional sources of possible ignition: the filament and an arc. The oper-
ating temperature of a tungsten filament is on the order of 1,500°C (2,700°F). The fila-
ment begins to burn up as soon as air contacts it. The filament begins to cool as soon as
the circuit is interrupted by failure of the filament, but it may retain more than enough

FIGURE 6-29A A 100-W

lamp buried in polyurethane
foam forces pyrolytic
decomposition and char
formation. Courtesy of John
Galvin, London Fire Brigade.

FIGURE 6-29B Leading

occasionally to flames.
Courtesy of John Galvin,
London Fire Brigade.

220 Chapter 6 Sources of Ignition

 FIGURE 6-29C After
flaming combustion, a
rigid, porous, but very
fragile char is formed
around the heat source.
Courtesy of John Galvin,
London Fire Brigade.

FIGURE 6-30 Test Panel

for Lamp Ignition. The
60-W lamp ignited
crumpled newspaper in
~10 minutes. The 75-W
lamp caused smoldering
of sawdust in 6 minutes,
13 seconds. The 100-W
lamp caused melting and
offgassing of polyester
fiberfill. The 150-W lamp
(center) failed to ignite a
polyurethane foam pad.
Courtesy of John D. DeHaan.

heat to ignite flammable vapors or solids that come into contact with it shortly after fail-
ure. In addition, the brief arc produced when the filament fails produces the same risk
of ignition as any other electric arc if flammable gases or vapors are in the vicinity.
Fluorescent lights operate at much lower temperatures than incandescent lamps, with
the exterior of the tube rarely exceeding 60°C to 80°C (140°F to 180°F).116 Mechanisms
involving failures of connectors for fluorescent lamps and power cords, and compact flu-
orescent lamps, will be discussed in Chapter 10 on electricity. Quartz halogen lamps con-
tain a quantity of iodine or bromine vapor and a tungsten filament operating at very high
temperatures. The surface temperature of the fused quartz envelope can be as high as
600°C to 900°C (1,100°F to 1,650°F), and the temperature of a nearby reflector can be
proportionately high.117 Even brief contact between a fuel and the lamp (or even the
reflector) could result in flaming ignition (see Figure 6-31). Halogen lamps can also ignite
Chapter 6 Sources of Ignition 221
 FIGURE 6-31 Halogen
lamp ignites curtain
valance after less than
one minute of contact.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

fuels by radiant heat and melt diffusers, shades, and fixtures. They have also been known
to shatter while lit, scattering hot fragments of glass and metal onto combustibles.118 An
example of a high-wattage metal halide lamp whose element shattered, with the frag-
ments penetrating both inner and outer enclosures is shown in Figure 6-32. The hot metal
and quartz fragments were believed to have ignited cellulosic materials directly beneath.
Lamps with an extra-shatter-resistant glass shroud are available but are rarely used
because of their significant extra cost.

FIGURE 6-32 Metal halide lamp failure caused ignition. Courtesy of Peter Brosz, Brosz & Associates Forensic Services Inc.; Professor Helmut Brosz, Institute
of Forensic Electro-Pathology.

222 Chapter 6 Sources of Ignition

IGNITION FROM DISCARDED BATTERIES
Most batteries for consumer devices (1.5 to 9 V) can support current discharges of up to
1 A when fresh. If metal can form a conductive path between the terminals, sufficient electric
current heating can occur to ignite cellulosic materials in direct contact. Nine-volt batteries
have both terminals at one end, so a coin or a piece of aluminum foil can act as a heating ele-
ment if contact is made with both terminals. Lithium-ion battery failures can also cause fires.

ANIMAL INTERACTION WITH SOURCES OF IGNITION

The very mysterious origin of some fires may be clarified by considering the intervention
of various animal species carrying out unnoticed activities of their own. It has been sug-
gested that a rat may chew a match head and thereby ignite it. Considering the ingredi-
ents in common matches, it is unlikely that they would present much temptation;
therefore, this explanation must be regarded with skepticism. Some evidence has been
found that a rat or mouse carrying a strike-anywhere match in its mouth could strike it
accidentally against a rough surface such as brick or concrete, lighting it in passing. The
final effect will be the same as with any other match-ignited fire, but the combination of
factors makes it most unlikely to occur.
Less consideration has been given to the possibility that both rodents and birds may
carry lighted cigarettes or other burning objects, which then ignite combustibles, possibly
even rooftops. Many rodents and birds have a strong instinct for acquiring and storing a
variety of materials, either for nest building or other reasons. Paper objects are especially
attractive, for example, to pack rats, whose nests have been found to be lined with facial
tissues. The paper cigarette may be similarly attractive, even if still aglow, although we
know of no proven instance of such an occurrence. (Admittedly, the ignition of a fire on
a clear, dry day at the top of a tall palm tree occupied by birds notorious for their scav-
enging has no other rational explanation other than one has picked up a still-smoldering
cigarette butt for its nest). Some birds are known to carry sticks, nest materials, and other
objects and then drop them down the roof vents of heating appliances. Such a practice is
clearly one that can produce a fire hazard. If the object was a lighted cigarette and it was
dropped onto flammable material, a fire could result without human intent or careless-
ness, other than the discarding of an unextinguished cigarette. The effects of bird nests
and rodent nests in chimneys and flues have been previously noted. Birds have also been
known to build nests on top of or adjacent to high-wattage flood lamps or other inter-
mittent heat sources. When the lamp or appliance is turned on, a delayed ignition can
occur. DeHaan has ignited dried grass and leaf debris in a loose pile in front of a 250-W
flood lamp. (In that case the lamp was mounted at ground level, and windblown leaves
and grass had accumulated around it.)
Mice and rats have been known to chew the insulation off wires, either short-
circuiting them directly or leaving the exposed conductors awaiting some other contact.
A simple short across the exposed conductors will be very unlikely to ignite a fire rather
than simply trip the circuit protection. If foreign materials (from saltwater to portions of
the animal’s anatomy) can sustain current flow, the heat generated can degrade the insu-
lation and lead to ignition of the insulation and other fuels in the immediate vicinity. This
mechanism probably represents the single most significant contribution to fire causation
by animals. Post-fire signs may be minimal once the surrounding insulation has been
destroyed, but unburned insulation has a gnawed appearance, as seen in Figure 6-33. (See
also Figure 6-34.) Snails, slugs, birds, and other animals have been known to short out
electrical equipment, thereby starting fires both inside structures and out. Dogs and cats
have been known to knock over containers of flammable liquids, which are then ignited
by remote ignition sources. They have also been known to knock candles, Christmas
trees, and lamps (such as the heat lamps used to keep pets, poultry, and livestock warm)
into contact with combustibles. Moths have been offered as a mechanism for starting spot
Chapter 6 Sources of Ignition 223
 FIGURE 6-33 Rodent-gnawed wires in this RV refrigerator came close to being a fire cause. Fortunately, only
the bare ground and neutral conductors had been exposed at this point. Courtesy Joe Bloom, Bloom Fire Investigation.

FIGURE 6-34 A grass fire started at the base of this power pole. The remains of the perpetrator can be seen as
the tail and the foot on the left side of the transformer. Note “flash burn” on adjacent pole. Courtesy David J. Icove,
Ph.D., University of Tennessee.

224 Chapter 6 Sources of Ignition

fires by catching fire in a candle flame and then flying some distance to ignite remote
fuels. One author pointed out the absurdity of this explanation: “A moth on fire would
be aerodynamically unstable and unable to fly very far even if the chitin from which the
wings are made supported combustion.”119
There appears to have been little or no study made of the habits of animal species in
relation to fire hazard, since such studies would be very difficult to carry out. Even without
more exact information, it is apparent that fires have at times been started by animals,
and more attention to this possible source of fire hazard is clearly indicated. A good inves-
tigation would include consideration of such mechanisms and a search for evidence to refute
or confirm the hypothesis.

Assessing Ignition Sources at the

Fire Scene: The Ignition Matrix
Nearly every room (or vehicle) offers multiple possible ignition sources—not only from
inherent energy sources such as fixed wiring, fixtures, outlets, heaters (gas or electric),
or other “appliances”—but also from the variety of human activities that may bring
more ignition sources into action. These additional ignition sources include smoking
materials, use or misuse of electric tools or appliances and devices, extension cords and
power strips, candles, and countless more. Today, more than ever before, fire investiga-
tors face the challenge of identifying all potential ignition sources in the area of origin
and then eliminating them, down to one (based on reasonable data or methods). As dis-
cussed in Chapter 1, the scientific method of fire investigation expects that an investiga-
tor will reach a defensible conclusion by demonstrating not only how this source could
have started the fire but also how all other reasonably possible sources have been
excluded. If two or more ignition sources are deemed “possible” at the end of the inves-
tigation, the cause has to be declared to be “undetermined.” Demonstrating these con-
clusions can be difficult, especially in a concise, yet comprehensive manner. Electrical
engineer Lou Bilancia has developed an innovative way of systematically evaluating
numerous ignition sources and also documenting how each was or was not competent
to ignite a particular first fuel. The suggested ignition matrix is merely a grid on paper.
Based on physical examination of the scene, interviews with occupants, or pre-fire pho-
tos or videos, the investigator identifies and lists all potential first fuels in the room in a
vertical column (as shown in Figure 6-35 ). The investigator then lists all the possible
ignition sources in a row across the top of the page. Each square, then, represents the
interaction between an energy source and a first fuel.120 Each combination is then eval-
uated on four primary values:
1. Is this ignition source competent to ignite this fuel? Yes or no.
2. Is this ignition source close enough to this fuel to be capable of igniting it? Yes or no.
3. Is there evidence of ignition? Yes or no.
4. Is there a pathway for a fire ignited in this first fuel to ignite the main fuel? Yes or no.
There are often comments or notes concerning each evaluation, such as the following:
■ This source was not energized (appliance or power cord) or not in use (candle).
■ “Too far away” or “Depends on duration being adequate,” or “Unknown—must
test exemplar.”
■ Physical or visual evidence of burning, ignition was witnessed or recorded in video.
■ Plume developed from first fuel sufficient, or flashover produced ignition.
The matrix can then be color-coded to indicate which combinations are capable, which
are excluded, and which need further data.
A typical bedroom, then, may have 15 or more potential first fuels: bed (mattress, box
spring, frame), bedding (blankets, pillows, sheets, comforters), carpet and pad, dresser,
Chapter 6 Sources of Ignition 225
FIGURE 6-35 Ignition
matrix—identified fuels Source Compact
Cell phone Plug-in room
along vertical axis; Clock radio Cigarette Candle fluorescent
charger refresher
ignition sources along Fuel lamp
horizontal. Courtesy of
1. yes (a) 1. yes (a, c) 1. 1. yes comp 1. yes (a, c) 1. yes (a)
Lou Bilancia, Synnovation
Lace night 2. yes close 2. not close 2. yes close 2. yes 2. yes 2. not close
Engineering and HTRI
table cover 3. no (d) 3. 3. no evid. 3. maybe 3. no (d) 3. no (d)
Forensics. Copyright 2008–10.
4. yes path 4. yes 4. 4. yes 4. yes gravity 4.
All rights reserved. Used with
permission. 1. no (c) 1. yes (a, c) 1. 1. yes comp 1. yes (a, c) 1. yes (a)
Sheets or 2. not close 2. yes close 2. not close 2. not close 2. no 2. yes close
covers 3. no (d) 3. no (d) 3. no evid. 3. 3. no (d) 3.
4. yes gravity 4. yes 4. 4. 4. no 4.
1. no (c) 1. yes (a, c) 1. 1. yes comp 1. yes (a, c) 1. yes (a)
2. not close 2. yes close 2. not close 2. not close 2. no 2. yes close
Curtains
3. no (d) 3. no (d) 3. no evid. 3. 3. no (d) 3. no (d)
4. yes path 4. yes 4. 4. 4. yes shade 4.
1. no (c) 1. yes (a, c) 1. not comp. 1. yes comp 1. no (a, c) 1. no (a, c)
Plastic
2. not close 2. not close 2. 2. yes close 2. no 2. not close
decorative
3. no (d) 3. no (d) 3. no evid. 3. maybe 3. no (d) 3. no (d)
flowers
4. 4. no 4. 4. yes 4. yes shade 4. no
1. yes comp 1. yes (a, c) 1. 1. yes comp 1. no (a, c) 1. no (c)
2. not close 2. not close 2. not close 2. not close 2. yes close 2. not close
Lamp shade
3. no (d) 3. no (d) 3. no evid. 3. indirect 3. no (d) 3. no (d)
4. yes path 4. no 4. 4. yes path 4. yes 4. no

1. Competent ignition source Y/N? Color legend

2. Proximity, ignition close to fuel Y/N? Red – Competent and close
3. Evidence of ignition Y/N? Blue – Not competent
4. Initial fuel path to fuel load Y/N? Yellow – Competent but ruled out

Codes Notes
P – plume or flashover a. if the device failed
W – witnessed b. if fuel was cellulosic with a breeze
F – open flame c. only with open flame
N – not energized d. device was intact

TV cabinet, clothes, posters, chairs, draperies, books, bookshelves, and so forth. Wall and
ceiling coverings must also be included if they are combustible. The bedroom may also
have numerous potential ignition sources (based again on physical examination of the
scene, pre-fire photos or videos, and interviews with recent occupants). These may include
fixed wiring, outlets, switches, and fixtures; table lamps, clock, radio, TV, electric blanket,
extension and power cords, cigarettes, candles, and matches. Although there may be many
competent ignition-fuel combinations in such a room, many will be excluded based on
physical separation alone. A candle on the dresser will not be capable, for instance, of
directly igniting first fuels more than an inch or two away. (Movement of such ignition
sources, however, must always be considered.) If the definable (and defensible) area of
origin is small, then the lists of first fuels and ignition sources will be considerably shorter
than if the whole room is being considered. The investigator must be careful, however, of
excluding fuels or sources merely if they are “outside the area of origin” by a matter of a
few inches.
This approach forces the investigator to consider a whole range of alternative
hypotheses and consider each one on the basis of factors such as heat release rate, heat
flux, separation distances, thermal inertia, and routes of fire spread. A completed matrix
offers a concise demonstration that all potential ignition sources have been considered
and all (but one, presumably) eliminated. Such a matrix is also more easily verified than
the traditional item-by-item list, improving the investigator’s own review process.

226 Chapter 6 Sources of Ignition

 CHAPTER REVIEW

Summary

No matter what the nature of the fuel, its susceptibility that precedes it) may change the temperature of the
to ignition, or the duration of the suspected ignition source, the chemical properties of the fuel, or the phys-
source, it is important to remember that an ignition ical state of the fuel in complex ways. All materials and
source must be of sufficient energy and in contact with reactions must follow certain physical and chemical
(or at least capable of transferring heat to) appropriate laws, and fuels and oxidizers are no exceptions. The
fuel long enough to raise the fuel to its ignition temper- analytical fire investigator must ensure that a postu-
ature to trigger the combustion of that fuel under the lated sequence of events in a fire not only fits observed
ambient conditions present. That combustion may be conditions but also follows these laws. Reliance on
in the form of a self-sustaining smolder or open flames, published data to confirm some ignition hypotheses
and each requires particular conditions. As we have may be enough if the conditions of the published tests
seen, flames, arcs, sparks, self-heating, and heated are a reasonable match to the fire conditions. If they
objects can all initiate fires under the right conditions, are not, properly designed and executed tests may be
and all must be considered whenever a fire scene is necessary to prove or disprove a particular hypothesis.
being examined. It is not enough simply to have an We have now completed an examination of all the
ignition source with a given temperature. The cause of basic elements of fire—chemistry of combustion,
a fire is the source of energy, the first fuel ignited, and behavior of fuels, and the mechanisms of ignition. We
the circumstances under which those are brought are now in a position to better understand the behav-
together that resulted in ignition. The investigator must ior of fire in a structure or wildland and interpret the
realize that the circumstances of ignition (and the contact post-fire diagnostic signs in light of this knowledge.

Review Questions
1. List five different open-flame fires (ignition sources) a potential ignition risk for wood exposed on a
and list their heat release rates. long-term basis?
2. Why are candles a more common accidental 7. Name three sources of frictional mechanical
ignition source today? sparks.
3. What effect does “insulation” have on the tem- 8. Why will cigarettes not ignite gasoline vapors but
perature of an ignition source like a lightbulb? will ignite acetylene or hydrogen?
4. Name three ways in which gas appliances can 9. Name three ways in which animals can cause
start accidental fires. fires.
5. How can kerosene room heaters cause fires? 10. List five conditions for spontaneous combustion
6. What fuel surface temperature is required to ignite to occur.
fresh whole wood? What temperature is considered

References
1. B. Karlsson and J. G. Quintiere, Enclosure Fire 4. D. D. Drysdale, An Introduction to Fire Dynamics,
Dynamics (Boca Raton, FL: CRC Press, 2000), 13. 2nd ed. (Chichester, UK: Wiley, 1999), 165.
2. J. Andrasko, “Identification of Burnt Matches by 5. J. W. Williamson and A. W. Marshall, “Characterizing
Scanning Electron Microscopy,” Journal of Forensic the Ignition Hazard from Cigarette Lighter
Sciences 24 (1978): 627–42. Flames,” Fire Safety Journal 40, no. 5 (July 2005):
3. V. Babrauskas, Ignition Handbook (Issaquah, WA: 491–92.
Fire Science Publishers, 2003), 509. 6. J. D. DeHaan, unpublished data, February 1996.
Chapter 6 Sources of Ignition 227
 7. “Propane Torch Flame Exceeds 2400°F,” Fire 29. N. Dwyer, “Flame Rollout.” Fire Findings 14, no. 1
Findings 13, no. 2 (Spring 2005): 5. (Winter 2006): 12-13.
8. S. E. Dillon and A. Hamins, “Ignition Propensity and 30. J. Nicholson, “Watch What You Heat,” NFPA
Heat Flux Profiles of Candle Flames for Fire Journal (September–October 2006), 67–73.
Investigation” in Proceedings Fire and Materials 2003 31. Babrauskas, Ignition Handbook, 886
(London: Interscience Communications): 363–76; M. 32. J. D. DeHaan, CCAI tests (unpublished data), July 2002.
Ahrens, Home Candle Fires (Quincy, MA: NFPA, 2002). 33. “How Hot Does It Get?” Fire Findings 11, no. 2
9. J. Nicholson, “When You Go Out—Blow Out,” (Spring 2003).
NFPA Journal (September–October 2005): 68–73. 34. J. J. Lentini, “Vapor Pressures, Flash Points, and the
10. Dillon and Hamins, “Ignition Propensity and Heat Case against Kerosene Heaters,” Fire and Arson
Flux Profiles.” Investigator 40, no. 3 (March 1990).
11. M. Faraday, The Chemical History of a Candle 35. R. W. Henderson and G. R. Lightsey, “An Anti-Flareup
(Atlanta: Cherokee, 1993). Device for Barometric Kerosene Heaters,” Fire and
12. D. Townsend, “Fires and Tea Light Candles,” Fire Arson Investigator 45, no. 2 (December 1994): 8.
Engineers Journal (January 2000): 10. 36. D. McAuley, “The Combustion Hazard of Plastic
13. J. L. Sanderson, “Candle Fires,” Fire Findings 5, no. 4, Domestic Heating Oil Tanks and Their Contents,”
(Fall 1997); H. Singh, Hazard Report on Candle-Related Science and Justice 43, no. 3 (July–September 2003):
Incidents (Washington, DC: U.S. CPSC, March 1998). 145–58.
14. R. Holleyhead, “Ignition of Flammable Gases and 37. J. D. DeHaan, unpublished data, July 2003.
Liquids by Cigarettes: A Review,” Science and Justice 38. USFA, Santana Row Development Fire, USFA-TR-153
36, no. 4 (1996): 262–66. (Emmitsburg, MD: FEMA, January 2004).
15. Babrauskas, Ignition Handbook, 213. 39. J. A. Bohn, “Woodstove Ashes Staying Hot for 110
16. S. Corwin, “Halogen Work Light Testing,” Fire Hours,” personal communication, Longmont Fire
Findings 9, no. 2 (Spring 2001): 12–13. Department, February 2003.
17. “How Hot Does It Get?” Fire Findings 8, no. 2 40. B. A. Schwartz, “Pyrophoric Carbon Fires,” Fire and
(Spring 2000): 5. Arson Investigator (December 1996): 7–9; B. R.
18. M. Goodson and G. Hardin, “Electric Cooktop Fires,” Cuzzillo and P. J. Pagni, “Low Temperature Wood
Fire and Arson Investigator (October 2003): 31–34. Ignition,” Fire Findings 7, no. 2 (Spring 1999): 7–10.
19. NFPA 921: Guide for Fire and Explosion Investigations 41. P. C. Bowes, Self-Heating: Evaluating and Controlling
(Quincy, MA: NFPA, 2004), table 5.3.1.1, 19. the Hazards (Amsterdam: Elsevier, 1984).
20. J. L. Sanderson. “Heat-Bulb Testing: Wood, Straw 42. J. C. Martin and P. Margot, “Approche Thermodyna-
and Towels Respond Differently at Close Distances to mique de la Recherche des Causes des Incendies.
Lamps,” Fire Findings 8, no. 4 (Fall 2000); R. E. Inflammation du Bois I.” Kriminalistik und
Lowe and J. A. Lowe, “Halogen Lamps II,” Fire and Forensische Wissenschaften 82 (1994): 33–50.
Arson Investigator (April 2001): 39–42. 43. Drysdale, An Introduction to Fire Dynamics; V.
21. R. J. Hall, J. J. Wanko, and M. E. Nikityn, “Special Babrauskas, “Pyrophoric Carbon: The Jury Is Still Out,”
Hazards Fire Investigation,” NFPA Journal Fire and Arson Investigator (January 2001): 12–14.
(November–December 2008): 68–73. 44. B. R. Cuzzillo and P. J. Pagni, “The Myth of
22. “Hot Surface Ignition,” Fire Findings 13, no. 2 Pyrophoric Carbon” in Proceedings Sixth International
(Spring 2005): 6. Symposium on Fire Safety Science, Poitiers, France,
23. M. Goodson, D. Sneed, and M. Keller, “Electrically 1999, 301–12.
Induced Fuel Gas Fires,” Fire and Arson Investigator, 45. Bowes, Self-Heating.
(July 1999): 10–12. 46. V. Babrauskas, B. F. Gray, and M. L. Janssens,
24. “How Hot Does It Get?” Fire Findings 7, no. 2 “Prudent Practices for the Design and Installation of
(Spring 1999); DeHaan, unpublished data. Heat-Producing Devices near Wood Materials,” Fire
25. “Instant Demand Water Heaters,” Fire Findings 14, and Materials 31 (2007): 125–35.
no. 1 (Winter 2006): 4. 47. Building Code of the City of New York, Sec.
26. J. M. Hoffman et al., “Effectiveness of Gas-Fired [C26-1400.6] 27-792 and Sec. [C26-1409.1]
Water Heater Elevation in the Reduction of Ignition of 27-809. See also UL127: Factory Built Fireplaces
Vapors from Flammable Liquid Spills,” Fire (Northbrook, IL: Underwriters Laboratories).
Technology 39 (2003): 119–32. 48. Babrauskas, Ignition Handbook.
27. J. L. Sanderson, “Manufacturers Vary Methods to 49. Ibid., 507–16.
Incorporate FVIR Technology,” Fire Findings (Fall 50. Ibid., 509.
2004): 1–4. 51. F. D. Maxwell and C. L. Mohler, Exhaust Particle
28. J. Bertoni, “The Essentials of Gas and Oil Fired Ignition Characteristics (Department of Statistics,
Forced Warm Air Furnaces,” Fire and Arson University of California Riverside, February 1973).
Investigator (December 1996): 15–18. 52. Babrauskas, Ignition Handbook, 508–16.

228 Chapter 6 Sources of Ignition

53. V. Babrauskas, “Risk of Ignition of Forest Fires from et al., “Who Knew? Cigarettes and Gasoline Do
Black Powder or Muzzle-Loading Firearms,” Report for Mix,” presentation B84 at American Academy of
U.S. Forest Service (San Dimas T & D Center, 2005). Forensic Sciences 59th Annual Meeting, San Antonio,
54. E. W. Few, “Ammunition Identification Guide,” TX, February 19–24, 2007.
Arson Investigator, California Conference of Arson 74. J. L. Sanderson, “Cigarette Fires in Fabrics,” Fire
Investigators, Spring 1978, 24–29. Findings, 15, no. 3 (Summer 2007): 1–3.
55. “Skyblazer Best in Low-Cost Aerial Pyrotechnics,” 75. D. A. Schuh and J. L. Sanderson, “Cigarette Testing
Powerboat Reports (September 1994). Reveals Dry Potting Soil, Peat Moss Can Be Viable Fuel
56. Holleyhead, “Ignition of Flammable Gases and Sources,” Fire Findings 17, no. 2 (Spring 2009): 1–3.
Liquids by Cigarettes.” 76. J. D. DeHaan, “Spontaneous Combustion: What
57. B. Beland, “On the Measurement of Temperature,” Really Happens,” Fire and Arson Investigator
Fire and Arson Investigator (September 1994); NFPA (January and April 1996).
921, Table 5.3.1.1. 77. F. C. Wolters, P. J. Pagni, T. R. Frost, and D. W.
58. G. Damant, lecture, California Conference of Arson Vanderhoot, Size Constraints on Self-Ignition of
Investigators (Sacramento, CA: California Department Charcoal Briquets (Pleasanton, CA: Clorox Technical
of Consumer Affairs, Bureau of Home Furnishings, Center, 1987).
June 1979). 78. N. D. Reese, G. J. Kloock, and D. J. Brien, “Clothes
59. D. A. Eichman, “Cigarette Burning Rates,” Fire and Dryer Fires,” Fire and Arson Investigator (July 1998):
Arson Investigator (June 1980). 17–19; J. L. Sanderson and D. Schudel, “Clothes Dryer
60. J. F. Krasny, W. J. Parker, and V. Babrauskas, Fire Lint: Spontaneous Heating Doesn’t Occur in Any of 16
Behavior of Upholstered Furniture and Mattresses Tests,” Fire Findings 6, no. 4 (Fall 1998): 1–3.
(Norwich, NY: William Andrew, 2001); G. Damant, 79. D. C. Mann and M. Fitz, “Washing Machine Effluent
“Cigarette-Induced Smoldering of Uncovered Flexible May Provide Clues in Dryer Fire Investigations,” Fire
Polyurethane Foams,” Journal of Consumer Product Findings 7, no. 4 (Fall 1999): 4.
Flammability 2 (June 1975). 80. W. M. Mackey, Journal of the Society of Chemical
61. J. L. Sanderson, “Cigarette Fires in Paper Trash,” Fire Industries 14, no. 940 (1895); D. A. Frank-
Findings 6, no. 1 (Winter 1998).
Kamenetski, Diffusion and Heat Transfer in Chemical
62. R. Berrett, personal communication, November 1998.
Kinetics (New York: Plenum, 1969).
63. R. Holleyhead, “Ignition of Solid Materials and
81. G. Helmiss and W. Schwanebeck, “Aspects of the
Furniture by Lighted Cigarettes: A Review,” Science
Investigation of the Chemical Processes of Self-Heating
and Justice 39, no. 2 (1999): 75–102.
by Means of Quantitative Thermal Analysis,” Journal
64. J. M. Shannon, “Fire-Safe Cigarettes; Keep Fighting,”
of Forensic Sciences 30, no. 535 (April 1985); A.
NFPA Journal (March–April 2009): 6; R. G. Gann
Baxton, E. Stauffer, and O. Delemonte, “Evaluation of
et al., “Relative Ignition Propensity of Test Market
the Self-Heating Tendency of Vegetable Oils by
Cigarettes,” NIST Technical Note 1436 (Gaithersburg,
Differential Scanning Calorimetry,” Journal of Forensic
MD: NIST, January 2001).
65. A. L. Miller, The U.S. Home Product Report, Science 53, no. 6 (November 2008): 1334–46.
1987–91: Forms and Types of Materials First Ignited 82. Babrauskas, Ignition Handbook, 405–36.
in Fire (Quincy, MA: NFPA, February 1994). 83. B. F. Gray, “Interpretation of Small Scale Test Data for
66. E. Braun et al., “Cigarette Ignition of Upholstered Chairs,” Industrial Spontaneous Ignition Hazards” in
Journal of Consumer Product Flammability 9 (1982). Proceedings Interflam 2001 (London: Interscience
67. J. A. McCormack et al., “Flaming Combustion of Communications), 719–29.
Upholstered Furniture Ignited by Smoldering Cigarettes” 84. D. Hevesi, “Warehouse Caught Fire When Gloves
(North Highlands, CA: California Department of Combusted,” New York Times, August 10, 1995.
Consumer Affairs, Bureau of Home Furnishings, 1986). 85. D. Kong, “How to Prevent Self-Heating (Self-Ignition)
68. Krasny et al., Fire Behavior of Upholstered Furniture in Drying Operations,” Fire and Arson Investigator
and Mattresses. (October 2003): 45–48.
69. G. H. Damant, “Cigarette Ignition of Upholstered 86. D. G. Howitt, E. Zhang, and B. R. Sanders, “The
Furniture” (Sacramento, CA: Inter-City Testing and Spontaneous Combustion of Linseed Oil” in
Consulting, 1994); R. Holleyhead, “Ignition of Solid Proceedings International Conference of Fire Safety,
Materials and Furniture by Lighted Cigarettes.” San Francisco, CA, January 1995.
70. R. Holleyhead, “Ignition of Flammable Gases and 87. J. D. DeHaan, unpublished data, 1993.
Liquids by Cigarettes: A Review,” Science & Justice 88. Howitt, Zhang, and Sanders, “The Spontaneous
36, no. 4 (1996): 262–66. Combustion of Linseed Oil.”
71. Ibid. 89. Ibid.
72. Babrauskas, V., Ignition Handbook, 2003, 717. 90. B. Dixon, “Spontaneous Combustion,” Journal of the
73. J. E Malooly and K. Steckler, “Ignition of Gasoline by Canadian Association of Fire Investigators (March
Cigarette,” ATF (May 2005); see also R. E. Tontarski 1992); Ziegler, D. L. “The One Meridian Plaza Fire:

Chapter 6 Sources of Ignition 229

 A Team Response.” Fire and Arson Investigator, IAAI-NSW Annual Conference, Sydney, Australia,
September 1993. September 1990.
91. B. M. Dixon, “The Potential for Self-Heating of 105. L. Lowenweent, “Fire Investigation,” International
Deep-Fried Food Products” in Proceedings FBI Criminal Police Review, no. 344 (January 1981): 2–3;
International Symposium on the Forensic Aspects of see also Proceedings International Symposium: Self-
Arson Investigators, Fairfax, VA, 1995. Heating of Organic Materials.
92. Babrauskas, Ignition Handbook, 403–4; see also 106. Firth and Stuckey, Society of the Chemical Industry.
Bowes, Self-Heating. 107. A. J. Hicks, “Hay Clinkers as Evidence of Spontaneous
93. K. Gaw, “Autoignition Behavior of Oiled and Washed Combustion,” Fire and Arson Investigator, July 1998,
Cotton Towels” in Proceedings Fire and Materials, 10–13.
San Francisco, CA, 2005. 108. A. T. Tinsley, M. Whaley, and D. J. Icove, “Analysis
94. J. D. DeHaan, “Spontaneous Ignition: What Really of Hay Clinker as an Indicator of Fire Cause,” paper
Happens, Part I,” Fire and Arson Investigator 46 presented at ISFI 2010, Adelphi, MD, September
(March 1996); “Part II,” Fire and Arson Investigator 27–29, 2010.
46 (June 1996). 109. D. J. Hodges, “Spontaneous Combustion: The
95. W. Vlachos and C. A. Vlachos, The Fire and Explosion Influence of Moisture in the Spontaneous Ignition of
Hazards of Commercial Oils (Philadelphia PA: Vlachos Coal,” Colliery Guardian 207, no. 678 (1973); N.
& Co., 1921). Berkowitz, “Heats of Wetting and Spontaneous
96. NFPA, Fire Protection Handbook, 20th ed. (Quincy, Ignition of Coal,” Fuel 30, no. 94 (1951).
MA: NFPA, 2008), table 6.17.11, pp. 6–288 to 6–292. 110. A. V. Smirnova and A. K. Shubnikov, “Effect of
97. E. Stauffer, “A Review of the Analysis of Vegetable Oil Moisture on the Oxidative Processes of Coals,”
Residues from Fire Debris Samples,” Journal of Forensic Chemical Abstracts 51, no. 18548 (November 25,
Science 50, no. 5 (September 2005): 1091–1100; E. 1951).
Stauffer, “A Review of the Analysis of Vegetable Oil 111. V. Fuchs, Forces of Nature (London: Thames and
Residues from Fire Debris Samples,” Journal of Forensic Hudson, 1977), 62–63.
Science 51, no. 5 (September 2006): 1016–32. 112. D. J. Latham and J. A. Schlieter, Ignition Probabilities
98. E. J. Hoffman, Journal of Agricultural Research 61, of Wildland Fuels Based on Simulated Lightning
no. 241 (1941); J. B. Firth and R. E. Stuckey, Society Discharges, USD Research Paper INT-411 (Ogden,
of the Chemical Industry 64, no. 13 (1945); and 65, UT: USDA Forest Service, September 1989).
no. 275 (1946); Proceedings International 113. General Electric Co., Incandescent Lamps, General
Symposium: Self-Heating of Organic Materials, Delft, Electric Publication TP-110.
Netherlands, February 1971. 114. R. L. Gruehn, “To Be Prepared Is Everything,” Fire and
99. H. Ranke, Liebig’s Annals of Chemistry 167 (1873): Arson Investigator 39, no. 1 (September 1988): 54.
361–68. 115. J. D. DeHaan, and J. C. Albers, Low-Energy Ignition
100. Firth and Stuckey, Society of the Chemical Industry. Tests, CCAI, November 7, 2007 (unpublished data).
101. C. A. Browne, “The Ignition Temperature of Solid 116. R. A. Cooke and R. H. Ide, Principles of Fire
Materials,” NFPA Quarterly 28, no. 2 (1935). Investigation (Leicester, UK: Institution of Fire
102. Bowes, Self-Heating. Engineers, 1985).
103. H. P. Rathbaum, “Spontaneous Combustion of Hay,” 117. Lowe and Lowe, “Halogen Lamps II.”
Journal of Applied Chemistry 13 (July 1963): 118. Ibid.
291–302. 119. Cooke and Ide, Principles of Fire Investigation.
104. B. F. Gray, “Spontaneous Combustion and Its 120. L. Bilancia, “The Ignition Matrix,” CCAI, November
Relevance to Arson Investigation,” paper presented at 2007.

230 Chapter 6 Sources of Ignition

 CHAPTER
7
Structure Fires and Their
Investigation

Courtesy of David J. Icove, Ph.D.,

University of Tennessee.

KEY TERMS

alligatoring, p. 281 clean burn, p. 281 indicators, p. 257

annealing, p. 289 crazing, p. 293 overhaul, p. 234
area of origin, p. 250 dropdown, p. 264 salvage, p. 236
calcination, p. 288 fire patterns, p. 257 smoke horizon, p. 268
chain of evidence (chain of ghost marks, p. 286 spalling, p. 283
custody), p. 320 heat horizon, p. 268 trailers, p. 271
char depth, p. 279 ignitable liquid, p. 272

OBJECTIVES

After reading this chapter, you should be able to:

■ Explain the comprehensive steps necessary in conducting structure fire investigations.
■ Recognize the necessary investigative information to be identified, collected, and
preserved during fire suppression.
■ Understand the general principles of fire behavior that combine combustion of
common fuels with the dynamics of heat transfer.
■ Explain the methodology for examining a structure fire scene.
■ Recognize the myths and misconceptions traditionally linked to incendiary fires that
have been shown to be unreliable.
■ Explain the systematic documentation of the fire scene through photography,
sketching, evidence collection and preservation, and use of timelines.
■ Understand how to state correctly the conclusions about a fire’s origin and cause.
■ Recognize and interpret damage as indicators of fire travel (i.e., fire patterns).
231
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

B
efore attempting to investigate a fire, the investigator should be reasonably famil-
iar with the principles of fuels, ignition, and fire behavior outlined earlier. He or she
must have a clear understanding of the purposes and goals of the investigation and
a rational, orderly plan for carrying it out to meet those purposes. Having an investigative
protocol and following it are now major expectations of courts following on Daubert,
Kumho Tire, Benfield, and other decisions regarding admissibility of expert witnesses and
their opinions. Critical to evaluating fire patterns accurately is reconstructing the pre-fire
scene as completely as possible. This includes not only physically re-locating fuel packages in
their original locations but also identifying wall, floor and ceiling coverings, determining
window and door openings, and conducting a fire engineering analysis for a reconstruction
of fire spread. No two fire scenes are ever alike, and no routine can be applied in the same
way to each fire. A carefully reasoned plan of attack can be applied as each case requires.
The purposes of the investigation can include the following:

■ To determine the origin of the fire. Where did it start?

■ To determine the cause—the nature of the initial fuel and means of ignition. What
was ignited? What ignited it? How did first fuel and ignition source come together?
■ Was it an accidental fire or an intentionally set fire?
■ If the fire was intentionally set, was it set with willful malice to destroy property, or was it
set to destroy unwanted personal property without intent to harm or defraud another?
■ If the fire was accidental, what factors contributed to its ignition or spread?
■ If life was lost, what contributed to the death? Was it the rapidity of the fire spread
or the toxic gases and smoke? Why or how did others survive?
■ How did fire protection equipment or systems function to minimize losses or fail to
protect the property?

As recommended in Icove and DeHaan, the full forensic fire scene reconstruction involves
the following steps (not necessarily in a fixed order).1

1. Document the fire patterns, scene processing, and fire growth/behavior if possible.
2. Evaluate heat transfer damage.
3. Establish the starting conditions—structure, fuels, heat sources, ventilation, and
environment.
4. Collect and correlate human observations and factors.
5. Conduct a fire engineering analysis.
6. Formulate, evaluate, and test hypotheses and form conclusion.
232 Chapter 7 Structure Fires and Their Investigation
 The preceding steps follow the precepts of the scientific method as described in Chapter 1,
and they also outline the best approach to any fire investigation, as follows:

1. Collect data—observe and document the scene as it exists, then document the
excavation and examination of the scene. Interview firefighters, occupants, reporting
party(ies), and other witnesses.
2. Establish starting conditions—based on the data collected, establish what the structure
and its contents looked like at the start of the fire.
3. Assess likely direction(s) of spread by observing patterns of heat and fire damage, smoke
deposits, structural collapse, and detector activation.
4. Ascertain whether there were any human (witness) observations or video or photographic
recordings. These help establish likely areas of origin (or eliminate possible ones),
starting conditions, and time frames.
5. Analyze the data by applying fire engineering principles to explain and interpret the
information collected to lead to possible areas of origin and ignition sources. Humans
affected by the fire provide information about the fire through tenability analysis.
6. Formulate and test scenarios about the fire’s origin and causation against available
data (including case histories, laboratory tests, and peer-reviewed scientific papers).
Some hypotheses may require additional data that have not been collected and pub-
lished to permit their evaluation. In that case, tests must be conducted to address par-
ticular issues. For example, when fire death cases involved the combustion of human
bodies, there were no reliable data about the heat of combustion, mass loss rates, or
combustion mechanisms of human tissue, so experiments were carried out and pub-
lished in peer-reviewed scientific journals to provide reliable data.

These steps are summarized in Table 7-1 Similar protocols are published in NFPA 921 and
the NIJ Guide previously cited (as well as IFSTA’s Fire Investigator).2

Beginning the Investigation

With respect to the origin of the fire, the investigation may begin at any time. Because every
fire should be investigated to detect the crime of arson, or to seek out dangerous products
or practices to preserve the safety of the community, the investigation may well begin while
the fire is still burning. In other cases, because of the circumstances of the fire or the criminal
or civil litigation surrounding it, the investigation may not begin until many weeks after the
fact. Let’s examine the sequence of events. If feasible, every good fire scene investigation
begins with an initial “noninvasive” survey of the scene and its immediate surroundings.

DURING THE FIRE

When possible, it is often helpful for the investigator to be present during the fire and its
suppression. The region of the structure in which the fire started may readily be apparent
or may be obscured by flames and smoke that prevent close inspection. The investigator
can profit from observing the suppression activities of the firefighters, since these can alter
the course and intensity of a fire as well as cause unusual burn patterns. The investigator
can also photograph or videotape the sequence of the fire and begin the search for wit-
nesses. Preliminary interviews with witnesses to establish what they saw or heard, and
where and when, is an important part of nearly every fire investigation.
The responsibilities of the firefighters with respect to investigations go far beyond
merely extinguishing the fire. They are often the first witnesses on the scene and have
Chapter 7 Structure Fires and Their Investigation 233
 TABLE 7-1 General Protocol for Fire Scene Examination

Survey and Assessment

Safety
Exterior
Find and Interview Witnesses
Fire personnel
Police
Neighbors, occupants, bystanders, etc.
Documentation
Notes, sketches, diagrams, photos
Exterior
Interior
Evidence
Specialized (X-rays, panoramic photos)
Establish Starting Conditions
Examine Fire Patterns
Conduct Fire Engineering Analysis (formulating and testing hypotheses) Using Ignition Matrix
Collect and Evaluate New Data as Needed
Establish Likely Area(s) of Origin
Evaluate Possible First Fuels and Ignition Sources
Test Hypotheses about Ignition Mechanisms
Formulate Conclusions (final hypothesis) about Origin, Cause, and Responsibility (as appropriate)
Prepare Report

enormous control over the fate of all types of evidence at the scene—from burn patterns,
to human remains, to residues of the ignition source. Their involvement in the investiga-
tion is detailed later in this chapter.

IMMEDIATELY AFTER THE FIRE IS EXTINGUISHED

So that access to the scene remains secure, it is important for the investigator to be on the
scene of the fire as soon as possible after extinguishment. Although the condition of the hot
remains is often not suitable for anything approaching a complete investigation, some
exterior surveys may be carried out. The general consequences of the fire may be more
apparent at this time than will be true later when the rubble is cleared out. In some fires,
it is virtually impossible to make an examination until some clearing has occurred because
of the large amount of collapsed rubble usually present. In small fires, however, much can
be accomplished at this stage, because when the fire is localized, so is the damage, and
reasonable access is available. In the largest fires, it may be days before a complete inves-
tigation may even be initiated, much less finished. Potential hazard areas (e.g., structural
collapse, chemical or electrical hazards) can be identified.
overhaul ■ The fire-
fighting operation of
eliminating hidden DURING THE CLEARING OF THE SCENE
flames, glowing It is critical that overhaul (the removal of potential seats of smoldering fire for the pre-
embers, or sparks that vention of rekindles) is kept to an absolute minimum until the investigator arrives and
may rekindle the fire,
usually accompanied by
examines the scene. In scenes where there has been a death, overhaul in the vicinity of the
the removal of struc- body should be strictly limited to dousing of smoldering furnishings with a hand line to
tural contents. prevent reignition.
234 Chapter 7 Structure Fires and Their Investigation
 It is most important for the investigator to be on hand during cleanup and to assume
some supervision of the overhaul operation. It is during this phase that most of the
remaining evidence is destroyed or displaced, thereby reducing its interpretive value.
Significant items of physical evidence are frequently disturbed or lost during cleanup.
Such problems can be avoided if the investigator is present and able to take the necessary
steps to preserve the evidence. Many fire service personnel are instructed in recognizing
evidence so that they can call the investigator’s attention to relevant items as they are
found. This interdependence is most valuable when the investigator cannot supervise all
areas of a large fire scene and must rely on capable and trained assistants. When it is nec-
essary to remove furnishings during cleanup, it is critical that the location of major fuel
packages be documented. It was once thought that the only fire patterns of value were on
the floors, so all “debris could be shoveled or flushed out leaving a clean scene for the
investigator.” Nothing could be further from the truth. It is essential that no debris be
removed by fire crews prior to the arrival of fire investigators (see Wallace and DeHaan).3

AFTER CLEANUP
The investigator now has a chance to examine burn patterns in full, provided they have
not been destroyed by the demolition of unsafe portions of the building during cleanup.
Dismantling and removal of the building (and its contents) should be prohibited or, at
worst, kept to the absolute minimum required for public safety, until the investigation is
completed. The investigator must not be less thorough because much of the building has
been removed. Nor must any remaining rubble or specific items removed from the struc-
ture be overlooked. It is sometimes possible to recover highly significant evidence from
such debris. The documentation of their original location will be more difficult, however.
Investigation after extensive demolition or repair is necessarily handicapped to a con-
siderable degree. Often, photographs or videos are available that will assist in evaluating
the limited remains of the fire scene. These can be very helpful, but they are no substitute
for original investigation at the scene. If photographic, diagrammatic, and interview doc-
umentation is sufficiently comprehensive, most fires can be reconstructed and accurately
investigated long after the fact, but in some cases it is simply not possible to arrive at firm
conclusions as to the cause of the fire if the scene has been extensively cleared. If fire
destruction is nearly complete, very little can be done in the way of fire pattern analysis,
as depicted in Figure 7-1. Only rarely can the investigator hope to prove the use of liquid
flammables from the pattern of the fire, and then only when extensive post-flashover

FIGURE 7-1 This office

complex was destroyed
when an arson fire was
set in a vacant office, and
the fire penetrated into
an unprotected common
attic space that included
10 separate buildings.
This particular office was
the Arson and Bomb
Investigation Unit of the
State Fire Marshal.
Courtesy of John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 235

 burning has not occurred and when the pattern is so distinctive as to leave no doubt as
to whether they were used. Many investigators working under these circumstances are
tempted to limit their activities to the questioning of witnesses, almost to the exclusion of
examining any physical evidence still remaining. It is vital to the success of the inquiry to
remember that the most productive investigation will be done right at the fire scene,
regardless of the time sequence of the investigation as related to the fire itself. However
useful interviews may be for suggesting lines of inquiry or corroborating conclusions, the
irrefutable physical evidence of the fire’s origin and cause will be found only by careful
examination of the fire scene. The systematic examination of a typical structure fire scene
can follow a summary guideline such as that in Table 7-1.

Investigative Information during Suppression

One of the major difficulties of fire investigation is that the fire scene is everything that
the good crime scene investigator would like a scene not to be. The ideal crime scene has
no personnel present besides the investigator and any coworkers. It is invisible to the out-
side world to minimize sightseers and their complications, and it is relatively clean and
orderly with items of evidence readily visible. Fire scenes are, of course, just the opposite.
The fire is usually detected by an outsider who may compromise what one would con-
sider to be ideal “crime scene” security right at the outset. Fire scenes are almost always
peopled by large numbers of firefighters and their supervisors, onlookers, and the press.
A great deal of damage and disorder is caused by the fire and its suppression. Most sig-
nificant of all is that, unlike at other crime scenes where a body, broken window, or a
gunshot indicates that a crime has been committed, suspicions of criminal fire activity
are often not raised until long after the efforts of suppression. Arson is unique in that,
because of the nature of fire, the crime destroys rather than creates evidence as it pro-
gresses. The firefighter’s prime responsibility, of course, is to protect life and property.
The responsibility to protect public safety by the detection of criminal activity is second-
ary to other functions and, in fact, does not fall within the official purview of the fire
service at all in some countries. Although it is recognized that virtually all fire scenes
start with the complications of a foreign influence or presence, the fire investigator with
proper forethought can make the most of the shortcomings.

RESPONSIBILITY OF THE FIREFIGHTERS

In police work, the “first person on the scene” is the most valuable witness as to the
nature of the undisturbed scene and the circumstances surrounding it. If investigators
can familiarize line firefighters with some aspects of fire investigations and physical evi-
dence, these first people on the scene can be a valuable asset to the fire investigator from
the outset. In some progressive jurisdictions, the firefighters are instructed that in the
event of any suspicious nature of a fire, they are to use extra caution in the suppression,
salvage ■ Procedures overhaul, and salvage of that fire scene. All three of these activities are, of course, part
to reduce incidental of the duties of the firefighter, but each of them can vastly complicate the scene, if not
losses from smoke,
make it completely worthless. With the proper precautions, unnecessary damage or dis-
water, and weather
following fires, ruption may be prevented. For instance, the use of excessive quantities of water at high
generally by removal or pressures should be avoided. Large quantities of water can literally flush important evi-
covering of contents. dence away from the scene or can, at best, complicate the scene investigation by causing
collapse and making extensive cleanup efforts necessary. Using excess water in suppres-
sion can cause collapse of plaster and sheetrock forcing investigators to spend many
hours removing hundreds of pounds of sodden plasterboard and waterlogged fiberglass
insulation from the scene. In smaller structures, the use of straight streams of water
should be avoided. “Fog” streams are not only less destructive but usually more effective
as a firefighting technique.
236 Chapter 7 Structure Fires and Their Investigation
 Unnecessary structural damage, such as removal or destruction of windows, walls,
and doors, should be avoided. The use of straight hose streams to break windows or
strip siding off frame buildings may be necessary for suppression, but excessive use
(sometimes out of juvenile destructiveness) is to be strongly discouraged. The use of such
high-pressure streams inside small buildings can strip interior walls and ceilings of their
gypsum board or plaster coverings and destroy invaluable burn patterns on those sur-
faces. High-pressure streams have also been seen to disrupt badly burned bodies and
destroy the fragile remains of incendiary devices. Fog or modified fog streams are more
effective in most interior attacks and are much less destructive. Fire crews have been
known to push the walls of buildings into a burning pile to minimize postfire cleanup.
This practice is a complete antithesis of the preservation of a fire scene. It is understood
that rapid entry to a fire scene may, on many occasions, be gained only by forcible entry
and that subsequent requirements to vent the scene may result in the breakage of many
of the windows. Firefighters should try to note mentally which doors or windows were
broken or open as well as those that were locked on their arrival. The recognition of
signs of forced entry, such as broken door frames or torn window screens, is also very
valuable to the investigation, and such observations should be related to the investiga-
tor. During the early stages of suppression, firefighters are in the best position to detect
possible multiple points of fire origin. The realization that several fires simultaneously
established themselves in separate areas of a building is an important step in the detec-
tion of multiple arson sets, a common technique used by fire setters. The fire investiga-
tor will be interested in not only the location and extent of the fire but also the direction
of suppression efforts, the points at which the fire was attacked, and the manner in
which the structure was vented, since all these can affect the fire patterns within. The
firefighters should note any unusual flames or smoke colors, odors, and weather condi-
tions (wind direction and intensity, temperature, etc.), and condition (operability) of fire
detection and suppression systems in the building. These guidelines are emphasized in
greater detail in the NIJ Guide published in 2000.

MINIMIZING POST-FIRE DAMAGE

The primary sources of post-fire damage to the scene are firefighting, overhaul, and sal-
vage operations.
Overhaul
Overhaul of the fire scene generally entails the removal and destruction of possible pock-
ets of smoldering fire in walls, floors, and upholstered furniture. These overhaul activities
should be limited to that necessary to prevent a recurrence of the fire.4 For example:
■ When walls or floors are torn up, mental note should be made of their condition.
■ Were the rugs in place?
■ Or were they folded up and thrown in a corner?
■ Was the furniture in place?
■ Were the walls and floors intact or had someone removed their covering materials
previously?
Sometimes, an arsonist will intentionally open ceilings, remove attic access panels, or
open holes in fire-resistant plaster or gypsum board walls to expose the internal studding
and improve the rate of flame spread from room to room. Large items of fire-damaged
upholstered furniture, such as sofas and beds, are normally removed to the outside of the
building during overhaul. The dumping of all such materials from an entire house into a
single pile can make the subsequent investigation impossible. When furniture such as this
is removed, someone should remember or note the locations from which various pieces
were removed. When conditions permit, all contents from all the rooms should not be
dumped into a single pile but in separate piles with a location for the contents of each
Chapter 7 Structure Fires and Their Investigation 237
 room established. Only the minimum amount of overhaul necessary for safety should be
allowed until the origin of the fire has been determined.
Gasoline-powered fans, pumps, and saws are often used during overhaul. Use of such
possible sources of contamination should be restricted to exterior placement and then
carefully documented. All fueling of such tools should be carried out in a designated area
remote from the fire. Electric or hydraulic tools are preferable if they are available.
Salvage and Security Concerns
During the salvage portion of the fire scene cleanup, owners, tenants, and employees are
sometimes allowed to enter the fire scene to remove items of value. The fire official
responsible for the scene should not allow any unauthorized persons to enter a fire scene
and especially should not allow materials to be removed from the scene until the investi-
gator has had a chance to evaluate those items. Cases have been reported in which the
“victim” returned to the fire scene to pick up some valuables and then proceeded to
destroy the evidence of one or more arson sets for which he or she was responsible. The
fire crews therefore have the best opportunity to detect suspicious circumstances because,
on entry, they can be aware of unusual conditions, contents, or fire behavior. Smith has
published an excellent in-depth review of firefighters’ responsibilities and their important
role in successful fire investigation.5 It is the responsibility of the fire department to main-
tain custody and control of the scene until the investigation is complete, even if that
entails hiring a private guard. The legal requirements for maintaining control of a scene
will be discussed in Chapter 17.

Documenting the Fire Scene

The general approach to scene documentation is outlined in Table 7-2. Thorough docu-
mentation using photographs, sketches, diagrams, and notes is the foundation of good
investigative process. Such information can be critical in demonstrating to the court what
was relied upon by the expert. It will also be invaluable to the original investigator when
called on years later to recall details of the investigation.
Throughout this chapter it has been repeatedly indicated that the appearance of the
fire scene should be documented with both photographs and sketches. The importance of
these steps cannot be overemphasized. Demonstrating to a court and jury the condition
and exact location of evidence is as important as the evidence itself. This documentation
is the beginning of the chain of custody and can be accomplished only with the use of a
camera and sketch pad.

PHOTOGRAPHY AND PHOTOGRAPHIC EQUIPMENT

Camera equipment need not be terribly complex or expensive, although investing in bet-
ter equipment will usually result in better pictures as a product. A good-quality 35-mm
camera or digital equivalent, preferably one that will accept a variety of special-purpose
(close-up and wide-angle) lenses, is the best all-around choice for fire photography. A
variety of film is available in 35 mm, and processing is less costly per frame than with
other film sizes. Smaller pocket-type cameras with plastic lens elements are to be avoided
except as an absolute last resort. Plastic lenses can distort or diffuse the image captured
by the film or digital array. The problems may not be apparent with snapshots but may
make an enlargement so fuzzy that it is an embarrassment to show it in court. Instant-
print cameras, although handy for some investigative purposes (such as witness inter-
views), are generally more costly to purchase. The cost per photograph is many times that
of conventional photographs, and most investigators discover that their convenience does
not compensate for the high costs and sometimes less-than-ideal picture quality.
Because regular film is so low in cost, and permanent in format once processed, some
investigators prefer that all critical scene photos be taken with traditional film, then scanned
238 Chapter 7 Structure Fires and Their Investigation
 TABLE 7-2 Systematic Documentation

STEP TYPE GUIDANCE PURPOSE

1 Exterior • Photograph exterior and • Establishes venue and location of fire scene in relation
perimeter of property. to surrounding visual landmarks
Sketch exterior. • Documents exposure damage to adjacent properties
Use GPS to obtain location. • Reveals structural conditions, failures, violations, or
deficiencies
• Establishes extent of fire damage to exterior of scene
• Establishes egress and condition of doors and windows
2 Interior • Record and sketch extent • Documents heat transfer damage, heat and smoke
of fire damages, potential stratification levels, and breaches of structural element
ignition sources, and data • Traces fire travel and development from exterior to
needed for fire reconstruction suspected point(s) of fire origin
and modeling. • Documents heat transfer damage, heat and smoke
stratification levels, and breaches of structural element
• Documents condition of power utility and distribution,
furnace, water heater, and heat-producing appliances
• Documents positions and damage to windows, doors,
stairwells, and crawl space access points
• Documents fire protection equipment locations and
operation (sprinklers, heat and smoke detectors,
extinguishers)
• Documents readings on clocks and utility equipment
• Documents alarm and arc-fault information
3 Investigative • Document clearing of debris • Assists in recording fire pattern and plume damage,
and evidence prior to isochar lines
removal; smoke, burn, and • Establishes condition of physical evidence, utilities,
charring patterns; packaged distribution, and protection equipment (breakers,
evidence. relief valves)
• Documents integrity of the chain of custody of evidence
4 Panoramic • Produce multidimensional • Clearer peripheral views of exterior and interiors
photos and sketches. • Establishes photograph viewpoints of witnesses

Source: D. J. Icove and J. D. DeHaan, Forensic Fire Scene Reconstruction, 2nd ed. Upper Saddle River, NJ: Pearson Education/Brady Publishing,
2009.

for electronic manipulation. This option also precludes legal arguments as to the originality
and content of scene photos. Some agencies use two cameras, one film and one digital, for
critical scenes. In addition to a camera and film, a flash unit is essential because most pho-
tos will be taken inside burned buildings with no lights of their own. A high-intensity elec-
tronic flash unit is often needed for photographing charred surfaces in dark interior scenes.
(This is a limitation of built-in flash units on most digital cameras.) The authors have dis-
covered that the heat flux from high-power electronic flash guns can ignite freshly
deposited, susceptible soot and pyrolysis accumulations on walls if discharged close by.
Digital Images
Digital cameras are now in very common use. The quality of the image produced should
be the primary criterion for selection. Some low-cost digital cameras offer a pixel count
that produces adequate small prints but do not have the resolution for any sort of enlarge-
ments. A camera that provides 1,200 ⫻ 2,500 pixels (3 megapixels/image) should be con-
sidered a minimum requirement for ensuring courtroom-quality prints. Cameras that
produce 8-megapixel photos are very common at a moderate price. Storage of images
containing 3 megapixels or more each requires considerable computer capacity.
Digital images provide flexibility in final format and simplify preparation of illustrated
reports. Some digital cameras offer security devices that identify “original” images. Files
Chapter 7 Structure Fires and Their Investigation 239
 copied directly onto a computer or disk must retain their original format without compres-
sion. Many fire departments have adopted formal protocols for preserving digital images.
Agencies or individuals using digital photography, both simple and complex, should
develop a standard operating procedure and departmental policy for preservation of dig-
ital images. Many departments have adopted formal protocols for the preservation of dig-
ital images consistent with the guidelines recommended by the Scientific Working Group
on Imaging Technology (SWGIT).6 These guidelines cover imaging technologies, recom-
mended policies, best practices, and techniques for photographing specific impression evi-
dence. These resource documents are available on the FBI (www.fbi.gov) and International
Association for Identification (www.theiai.org) websites. The SWGIT procedures recom-
mend preserving original images by storing and maintaining them in an unaltered state
and in their native file formats. Duplicates or copies should be used for working images.
Original images should be preserved in one of the following durable formats: silver based
(non-instant) file, write-once compact recordable disks (CDR), or digital versatile record-
able disks (DVD-R).
Enlargements and Film
Enlargements are often necessary for court display. Color prints of sufficient size are very
costly and may be prohibitive in cost if full documentation is to be offered. Many investiga-
tors prefer to take all their scene photos on color transparency (color slide) film such as
Ektachrome and then project these in the court during testimony. This option offers the low-
est cost per photograph and the means of re-creating the scene at life size for the jury. Its
drawbacks are that some courtrooms are not easily darkened for proper projection and that
it is awkward for the jury to review them in the absence of printed labels and projection
equipment. The ideal presentation would probably include transparencies for courtroom
testimony with a duplicate set of small color prints to which the jury could refer. Many
courtrooms today have digital image projection capability. If the format of the storage
medium can be guaranteed to be compatible with the court’s equipment, high-quality digi-
tal images can be shown in open court and later used in print format for record-keeping or
jury review. Color photographs or images are the standard medium. For court displays
involving fatalities, however, the court may insist on black-and-white images only. For the
most part, however, the subtle coloring of burn patterns is lost on black-and-white images,
and the photos contain less information. The investigator must be careful that the color ren-
dition of the final images used correctly and accurately duplicates the original scene.
Photography for the Fire Investigator
If the fire is still in progress when the investigator arrives, he or she should photograph
the building from various angles and at various times. These photos will help document
the extent of damage to the building on arrival. If the building is then totally destroyed
by fire or collapse, information is available to indicate in which direction the fire was pro-
gressing. Other potentially valuable information in such early photos will include condi-
tions of doors or windows, presence of suspicious vehicles, and the appearance of
onlookers. (The photos of spectators from a series of similar fires should be compared to
detect faces that frequently appear at fires.)
When the fire has been extinguished, photographic documentation should follow the
same general pattern as the scene search (exterior, interior, investigative, specialized). The
exteriors of all remaining walls of the structure should be photographed to document
the extent of the damage. The interior of the structure, including ceilings and floors, should
then be photographed to preserve their appearance before the investigation or overhaul. Prior
to exterior photography, the windows and doors on the ground floor (or other openings that
could have provided access) should be numbered sequentially (e.g., clockwise starting at front
door) and numbered markers positioned at each one so that they are visible as each door and
window is documented as to condition and security. The markers should then be moved
inside to help ensure that all windows, doors, and possible entry points are documented.
240 Chapter 7 Structure Fires and Their Investigation
 During the investigation, many points in the burn pattern; partially destroyed appli-
ances; utility materials; the remains of incendiary devices; possible sources of accidental igni-
tion; signs of forced entry, burglary, or other crime; alarm or sprinkler systems; and other
items of possible importance will have been noted. These should be photographed carefully
so as to record their general features, condition as found, and their relationships to other
objects nearby. These photographs can be most important in determining whether there is
evidence that the fire spread from possible ignition sources to other combustible objects or
to adjacent structures. Such a fire spread must cast suspicion on the object that initiated it,
even though it may ultimately be identified as simply an agent of normal fire spread.
The largest single activity related to most investigations after a fire is moving debris
so that critical indicators and items can be uncovered. In many instances this removal
must be done by a work crew, or even machinery, and may require days to accomplish. If
possible, the investigator should be present, with camera, because an immediate record
can be obtained of anything suspicious or informative that may be uncovered before fur-
ther clearing activities may destroy or damage the item in question. Removal of the room
contents and their reassembly outside the structure on a proportioned full-scale map on
a plastic sheet may aid in visualizing the pattern, as illustrated in Figures 7-2 and 7-3.

FIGURE 7-2A The inves-

tigators of this trailer fire
chalked an outline on the
soil and transferred items
to the outline as they were
removed. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

(a)

(b) (c)

FIGURES 7-2B AND C (b) Investigators dismantled the garden shed in this fatal fire, leaving the body and
debris undisturbed. (c) They were removed to a plastic sheet marked with the same grid as the structure.
Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.

Chapter 7 Structure Fires and Their Investigation 241

FIGURE 7-3A This mobile home fire required a large plastic tarp FIGURE 7-3B As debris was removed from each grid within
with equal-scale grids overlaid on a floor plan. Courtesy of Don Perkins, the scene it was placed in the corresponding grid on the tarp.
Fire Cause Analysis. Courtesy of Don Perkins, Fire Cause Analysis.

After a room or area has been cleared of debris, it is often of value to reassemble the fur-
nishings in their original locations (using protected areas to reposition them) and repho-
tograph them in place.
In large rooms, single photos cannot capture all relevant information. With computer
scanning of photos readily accessible, it is possible to scan a series of overlapping photos
and then, with specialized programs, stitch the digital images together to form a panorama
of the room (such as with the Photo Stitch link in Roxio’s PhotoSuite ). In lieu of such pro-
grams, photos can be physically overlapped in a mosaic of images to re-create a view
(because the human eye has a wider field of view than the typical 50-mm focal length of a
camera lens).7
It is expected that ultimately the point of origin will be revealed and, it is hoped, recog-
nized. At this point the investigator should carefully supervise the clearing and record photo-
graphically all the evidence that gave rise to the choice of this as the point of origin. Close-up
photos should be taken of any possible source of ignition or similarly critical evidence in the
vicinity of the suspected origin. This includes possible ignition sources that were considered
and eliminated, because that is part of the investigator’s analysis and opinion.
Whether or not the investigator is later called as a witness in an arson trial or in a
civil action, the photographs will be the best evidence of what was done, what was seen,
and why he or she reached the stated conclusion. The evidentiary value of photos can be
compromised if they are not documented properly with a photographic log. A typical log
is included in Appendix I. Note that the log provides for a description of each photo as it
is taken, including the subject matter and direction in which it was taken. This data can
be recorded on paper as the photos are taken or dictated into a cassette or digital recorder
for later transcription. This log can be in the list format of Appendix I or as a floor plan
with numbered circles and arrows, as depicted in Figure 7-4.
Neither the human memory nor the notebook, important as both are, will ever match
the camera as a means of recording observations. Fire scenes are some of the more diffi-
cult photographic subjects because of the properties of burned wood in relation to light.
In addition, large structures offer special problems because many such scenes have insuf-
ficient light to focus a camera accurately, and flash lamps are ineffective at the distances
that must be used in many instances. Another source of difficulty is that the general reflec-
tivity of a burned scene is so low that it is recommended that the lens diaphragm or f-stop
be opened one or two stops, even compared with a meter reading. Otherwise, photo-
graphs tend to be underexposed to the point of being useless. The “automatic” settings
242 Chapter 7 Structure Fires and Their Investigation
 3 4 FIGURE 7-4 Scene
diagrams are essential to
complete documentation
43'9
and demonstration of the
12'4 3'9 5'10 3'2 18'8
2'6 21'4 3'8"4 reliability of the data. This
9'10 5'11 12'9
31 diagram incorporates a
photo log.
29 32

4'7
28 17
18
30 5

3'5 ⫻ 9'6

5'6 ⫻ 9'5
BATH

BATH
BEDROOM KITCHEN

13'6
13'6

11'10 ⫻ 13' 18'2 ⫻ 13'

8'11
20
26 25 19 16 15
27
25'9

25'9
2 6
24 21 11
CLOSET

12
4'8 ⫻ 6'4

14 33
6'8

BEDROOM BEDROOM LIVING

12'3
10'7 ⫻ 11'9 9'6 ⫻ 8'9 17'4 ⫻ 11'9
CLOSET
5'7

35 34
4'8 ⫻ 5'1 23 22
10 36 13

9
2'2 5'1 4'9 2'9 8'2 6'11
8'9
LIVING AREA
3'2 10'11 9'10 17'10
1'128 sq ft
43'9

8 N 7
1

of digital cameras can be overridden with programmed settings to accommodate such

scenes. The investigator should become familiar with the manual setting of any camera used
to maximize its usefulness.
Burned wood, although black, is highly reflective at certain angles of light incidence.
These glare spots photograph white, appearing lighter in many instances than unburned
wood nearby. The use of a polarizing filter will assist in reducing these glare spots and
giving a truer representation of color intensity. Also, critical areas may be reasonably
reproduced by illuminating them head-on so as to avoid glare angles.
Some investigators use video recorders to preserve an image of a fire scene. Although
such media are convenient to use, they are very limited in their capability for preserving
fine detail, and review of specific images is difficult. Video should not be substituted for
good sketches and comprehensive photography at a scene, but it may be useful for sup-
plementing documentation of a fire scene. Video of the fire in progress can be very useful
in retracing its path. Some digital video recorders have the capability to record still pho-
tos (or video with “still-photo” resolution) and audio “notes,” allowing several means of
information to be captured in one tool.
There are many tricks that photographers learn to allow them to take pictures under
even the most difficult conditions. Space does not permit us to examine them here. The reader
is referred to specialty texts in this area (e.g., Lyons or Miller and McEvoy or cfitrainer.net)
for further guidance.

SKETCHING
Fire scene sketches and diagrams should be considered important supplements to photo-
graphs and written notes, not replacements for them. A sketch is a drawing or layout that
graphically portrays the scene and items of evidence within it. Only sketches record and
Chapter 7 Structure Fires and Their Investigation 243
 reveal spatial relationships between items of evidence found at the scene. They offer the
visual advantage of illustrative evidence while allowing the elimination of distracting or
obscuring detail found in some photographs. A simple floor plan of a structure can cap-
ture more information more clearly than many pages of narrative description. Accurate
and detailed scene sketches allow reconstruction of a scene weeks, months, or years after
the incident, by indicating the placement of various items of evidence. They help refresh
the memory of investigators or witnesses, and they can indicate the perimeter of a crime
scene search with respect to a larger geographical area. A simple plan (overhead) view of
a room or building (as shown in Appendix F) can capture as much information as pages
of narrative and convey it quickly and accurately to a reader. Although careful execution
and neatness certainly improve the visual quality of sketches, the most important require-
ment is that they portray evidence accurately. Because their artistic merit is of secondary
consequence, good scene sketches do not require a high degree of artistic ability.
Composing a detailed room diagram can focus an investigator’s attention on identifying
and locating fuel packages and all potential ignition sources in a room. Such a diagram
will aid in the preparation of an ignition matrix, as described in Chapter 6.
To be useful as overall evidence, fire scene sketches must depict the following:

■ The outline and dimensions of the room, building, or area of interest

■ The locations and relative positions of pertinent evidence found at the scene or critical
features of the scene itself, including significant burn patterns
■ The locations of approaches or possible points of entry such as doors, windows, stair-
ways, skylights, paths, and roadways (including dimensions of windows and doors)

Fire scene sketching requires only paper, pencils, pens, tape measures, a ruler, and a
clipboard for the rough original “on-the-scene” sketch. Paper ruled into a grid of 8 to 12
squares per inch (graph paper) makes scene sketching easier, since it is easier to draw out-
lines of rooms or structures by tracing along the grid, counting off the divisions corre-
sponding to the measurements. Drawing aids such as the Acculine device (see Appendix
F), which helps produce straight lines on a hand-drawn diagram, may be useful. The fin-
ished drawing, to be filed with the report or used in court, requires unlined paper, pens,
rulers, drawing board and, preferably, T-square and triangle or, more commonly today,
the use of a drafting system such as Chief Architect or The Crime Zone, discussed later
in this chapter. True “scale” drawings are done to architectural standards and are done
only in special cases.
The best rule to keep in mind is that the final sketch must be completely intelligible
to the viewer without detailed study. If many types of evidence are to be denoted on a
sketch, a single sketch may not be adequate, and sketches for each individual purpose
would be a good idea. If too much detail is included in a single sketch, the sketch’s major
advantage over a photograph—its simplicity—is lost.
To ensure intelligibility and acceptability as evidence, every sketch or diagram should
always include the following:

■ The investigator’s full name and agency, and the date and time at which the sketch
was prepared
■ The case number of the incident, if available
■ The full names of all persons involved in making the sketch or assisting in
measurements
■ The address or location of the fire scene
■ The scene’s geographic orientation (compass marking) and GPS location, if available
■ A legend that includes the meaning of all symbols used, a numbered list of numbers
or letters used to denote items of interest and their meanings, and a scale of the
drawing (for example, 1 inch equals 1 foot; 1⬙ ⫽ 1⬘)
244 Chapter 7 Structure Fires and Their Investigation
 Because the sketch may be used at a later time to help place various pieces of evidence
at a reconstructed scene, the accuracy of all locations and dimension measurements is crit-
ical to its acceptability as evidence. Use a conventional system of measurement (feet and
inches, meters and centimeters). Use a second person to verify all measurements, if possi-
ble. Do not use paces or “steps” on the final drawings, since they are not standard meas-
urement units. Do not measure a distance in paces and then convert to feet and inches
because this implies accuracy that is not obtainable by a pace measurement. Remember,
the method of sketching and measurement chosen should always give someone viewing
the result the best information possible. If a body is involved, be sure the sketch accurately
depicts the orientation and position of the torso as well as the orientation of all extremi-
ties. For further details on proper scene sketching techniques, see Appendix F. Sketching
can be combined with a photograph to enhance the interpretive value of an exhibit.
Some investigators may find it useful to use different colored pens or markers when
fire patterns to coincide with the “depth” designations described earlier. For instance,
blue could be used to outline (or crosshatch) an area that has only surface deposits, green
to denote areas where there are some thermal surface effects, yellow to denote areas
where there have been some deeper effects, and red to outline areas where the material is
penetrated completely or consumed. This technique can be used with both manual and
computer-aided sketching and offers a supplement to scene photographs or notes that
may not conveniently record depth factors.
Computer-aided drafting and drawing programs specially designed for crime scene
documentation are available now, although one found to be most useful is “Home
Designer Deluxe” offered by Better Homes & Gardens. BH&G is planning to offer a spe-
cial “fire investigation” package that will include a set of graphic symbols and features for
that application.) Microsoft Visio 2000 products for crime scene applications are available
from Microsoft’s website. There also are specialized consumer diagram programs such as
The CrimeZone.9 See Appendix F for an example of Home Architect capability.
Measurement and Scanning Systems
Measurements of rooms, doors, and windows are essential data for the accurate analysis
and reconstruction of most fire scenes. Although they do not have to be precise, they
should be within ⫾5 percent of the actual dimension. A variety of manual and optical
tape measurements (available in Appendix F) are essential.
Investigators who frequently have to diagram large, complex scenes (particularly out-
doors) have used a laser-based surveying method (called Total Station) that measures and
records the distance as well as the angles of position and inclination or declination from
a reference point to each item or point of interest onto a computer program. The program
then plots out all recorded points, designating walls, fences, curbs, locations of evidence,
and the like. Computer-driven laser scanners are now finding more use in rapidly captur-
ing fire scenes with tremendous accuracy. The newest imaging techniques combine wide-
field image capture with measurement data. Laser scanning cameras use an automatic
rotating laser measuring (survey) system and a synchronized digital camera to produce a
fully explorable, 3D virtual reality image that captures dimensional detail with great accu-
racy.10 They can be used to quickly document even large fire and explosion scenes. Systems
by Leica Geosystems and Delta Sphere have been shown to be usable in a variety of
scenes, even in small rooms.11 An example of a laser scan of a burned room is shown in
Figures 7-5a and b. An alternative is a digital scanning camera such as the Panoscan that
captures a 360° image with measurement capability on the virtual reality image, as shown
in Figures 7-6a and b.

NOTES
Notes are the final element in comprehensive fire scene documentation. Information
that is not visual cannot be captured in a photo or diagram. Weather data, witness
Chapter 7 Structure Fires and Their Investigation 245
FIGURE 7-5A Laser scan of fire scene. Courtesy of Precision FIGURE 7-5B Standard photograph of same scene. Courtesy of Don
Simulations. Perkins, Fire Scene Investigation.

Object

FIGURE 7-6A The Panoscan digital imaging system can take FIGURE 7-6B Panoscan image of large fire scene in a commercial
two 360° scans from top and bottom positions on the tripod, kitchen. (When viewed as a computer image, the spherical
allowing a stereo-optical re-creation of distances. Courtesy of distortions are compensated.) Courtesy of Robert Toth, IRIS Fire
Panoscan. Investigations.

246 Chapter 7 Structure Fires and Their Investigation

 ROOM FIRE DATA FIGURE 7-7 Data form
for complete documenta-
Room # tion of room fire.
Room
Length Floor Plan
Width
Height

Note ceiling height changes

Walls: Structure/material Thickness Covering Sample? Y/N

Structure/material Thickness Covering Sample? Y/N
Ceiling: Structure/material Thickness Covering Sample? Y/N
Floor: Structure/material Thickness Covering Sample? Y/N

Openings (door, window, other vents) into room (number on plan above):

Height (bottom to Soffit Depth Open or Closed?

top of opening) Sill Height (above opening) Width Changes During Fire?
1.
2.
3.
4.
5.
6.
7.

HVAC System:
Description
On/off prior to/during fire?

Furnishings (descriptions of major fuel items, including floor and wall coverings, draperies):

Time Line: Alarm time FD arr. time Control time

Detection:
Pre-Fire Events:

interviews, odors, time events, contents, and materials are captured as notes. These may
be written on plain paper or recorded on a protocol form such as the one shown in
Figure 7-7. Note that this form reminds the investigator not only to measure dimen-
sions of rooms, doors, and windows but also to identify and collect samples of wall,
ceiling, and floor coverings.
No matter when the scene investigation is begun, documentation is essential, for it
forms the basis for all that follows—fire pattern analysis, engineering analysis, hypothe-
sis formulation and testing, reporting, and testimony. The days when an investigator
could walk through a scene, visually evaluate fire patterns on the fly, reach a conclusion,
and then testify to that opinion without documentation are gone. The expectations
of both the courts and the profession are much higher today. The various investigation
Chapter 7 Structure Fires and Their Investigation 247
FIGURE 7-8A
Coffeemaker fire involving
other fuels (cereal boxes,
towels). Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

FIGURE 7-8B Fire

spreads into cabinets.
Plume entering air return
grille in suspended ceiling.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

FIGURE 7-8C Extension

above suspended ceiling
very rapid. Light fixture
diffuser softens and falls
out. Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

248 Chapter 7 Structure Fires and Their Investigation

protocols describe the preferred professional standards and practices, such as ASTM
E620,12 ASTM E678,13ASTM E860,14 ASTM E1020,15 and ASTM E1188.16 NFPA 921
contains forms for field data collection. Equivalent forms are included in Appendix I.
The photographic and diagrammatic documentation of a typical fire scene employ the
techniques described in Table 7-2. They follow a logical flow that maximizes evidence
detection and preservation while minimizing the danger of loss or contamination. This
topic will be examined in greater detail in a later section of this chapter.
The purpose of the exterior examination and documentation is to establish the loca-
tion of the fire in relation to other landmarks and record the condition of the structure
prior to any scene disturbance. It also encourages a scene search for physical evidence and
paths of possible entry and exit. Hazardous conditions can be assessed and needed
resources can be requested (e.g., scaffolding for an unstable wall.)
The interior documentation includes recording all rooms of a structure (not just the
fire-damaged areas). This ensures that some information will be available later about pre-
fire conditions such as furnishings; wall, floor, and ceiling finishes; and maintenance con-
ditions. Heat, smoke, and explosion effects are, of course, documented in detail. Interior
documentation also includes the existence and condition of fire protection systems and
possible ignition sources.
Investigative documentation includes anything that will be of assistance in conduct-
ing the analysis and hypothesis testing. This includes ignition sources, fuel loads, witness
observation, measurements, specific fire patterns, and electrical systems (arc fault map-
ping). This documentation will be in the form of physical evidence, photos, diagrams,
sketches, notes, and even video or audio recordings.
Specialized documentation may include X-rays; thermal, infrared (IR) or ultraviolet
(UV) images; or panoramic photos. Some of these specialized techniques will be described
later in the text and in the appendices.

RECONSTRUCTING THE PRE-FIRE CONDITIONS

Before meaningful analysis of fire behavior and fire patterns can be carried out, the inves-
tigator must be familiar with the conditions that existed at the start (or immediately
before) the fire. This means knowing the room dimensions (including ceiling height), inte-
rior finish, fuel packages (size, type, and location), and conditions of windows and doors
(open or closed, sizes, and sill and soffit heights). This is why documentation of unburned
areas of a building is so important—to give some idea of ceiling and wall finishes, the
nature of furnishings, housekeeping standards, and the like. Previous chapters have dis-
cussed the roles that fuel type, physical state, orientation, and ventilation play in a fire.
Even the orientation of room contents and lines of sight can affect what witnesses can see
and when during a fire’s development. If one is going to attempt any meaningful, accu-
rate, defensible fire engineering analysis or computer modeling, reliable data on the start-
ing conditions are essential. Figures 7-8a–c show a test fire in which rapid extension into
the ceiling void occurred because of a ceiling vent above the origin.

General Principles of Fire Behavior

You should now have some appreciation of the basic principles of combustion of com-
mon fuels and the dynamics of heat transfer, fire, and flame. We now apply those princi-
ples to more complex situations involving combinations of fuels and their structural
arrangement.

FIRE PATTERNS
The effects of smoke, hot gases, heat, and flames on materials are what form the fire
patterns that investigators use to trace the course of the fire and, by reversing that
Chapter 7 Structure Fires and Their Investigation 249
area of origin ■ The process, establish the area of origin of the fire. These effects can be generally categorized
general locale in which as follows:
a fire was ignited.
■ Surface deposits (with no irreversible effects on substrate)
■ Surface thermal effects—scorching or melting, discoloration
■ Charring—surface
■ Penetration—charring below surface
■ Consumption—charring completely through a material or actual destruction
These effects are the foundation of the fire patterns investigators use. They reflect the
intensity and duration of thermal exposure.
If no thermal effect is observed on a surface, the investigator can conclude it was not
exposed to enough heat to induce any chemical change. In general, the observed changes
reflect the intensity of the applied heat flux—the higher the heat flux, the higher the sur-
face temperature (dependent, of course, on the thermal inertia of the surface). In general,
the deeper the penetration of the changes, the longer the duration of the exposure.
However, heat penetration is not linear with time (although it becomes nearly so for long
exposures). The higher the heat flux (or the higher the temperature of the exposed sur-
face), the faster heat is conducted into the material. The total observed damage is there-
fore the result of both intensity and duration, so both must be taken into account.

TRACING THE COURSE OF THE FIRE

The fire investigator’s task would be easy indeed if one were present for the duration
of the fire, from ignition to extinguishment. Unfortunately, fire scenes are nearly always
cold and badly burned, and the investigator must reconstruct the sequence of the fire
backward from what is visible afterward to its point of origin. Fortunately, there are
some general guidelines that can be applied to all structural fires to assist in this recon-
struction. With few exceptions, fires, however large, start with a small flame, such as a
match, candle, or lighter, or a spark of some type. The point and source of ignition are
two of the key elements to be determined by the investigator, regardless of the size of
the conflagration that followed this tiny flame (the others being first fuel ignited and
circumstances of ignition). These elements determine the origin of a fire and must be of
primary importance in any fire investigation. With very few exceptions, no defensible
conclusions can be reached about the cause of a fire if the area of origin cannot be
established. The primary exception are fires where the first fuel ignited is a flammable gas
or vapor.
An investigator who spends a large proportion of time surveying what are obviously
the later stages of the fire is wasting much of that time, because such a survey cannot
answer the fundamental question of nearly every investigation—what started the fire?
The insurance adjustor may be interested in the final stages of the fire in which most of
the property damage occurred. The arson investigator may be interested in what contents
were destroyed from the standpoint of helping uncover a motive for a malicious fire.
Otherwise, the predominant focus is on why and how the fire occurred, and the investi-
gation must similarly focus on the tiny initial flame and the first fuel ignited.
Because it is of the greatest importance to trace the behavior of this small flame as it
grows into the large fire, the following rules of fire behavior are offered:
■ Hot gases (including flames) are much lighter than the surrounding air and therefore
rise. In the absence of strong winds or physical barriers (such as ceilings), which
force hot gases to travel elsewhere, smoke and hot gases will always preferentially
move upward.
■ Due to heat transfer processes in the buoyant flame plume, fire will always prefer-
entially move upward much more rapidly than horizontally or downward (with
some downward progression as a result of radiant heat and falldown).
250 Chapter 7 Structure Fires and Their Investigation
■ Combustible materials in the path of the flames will be ignited, thereby increasing
the extent and intensity of the fire. The more intense (higher heat release rate) the
fire, the faster it will rise and spread.
■ If there is not more fuel above or beside the initial flame to be ignited by convected
or radiated heat, or if the initial fire is too small to create the necessary heat flux on
those fuels, the fire will be self-limiting and often will burn itself out. A flame plume
that is large enough to reach the ceiling of a compartment is likely to trigger full
involvement of a room, for it is charging the upper gas layer in the room with gases
that have not cooled by mixing or radiative losses as much as gases in a shorter plume.
As a result, the hot gas layer maintains a higher temperature and is more likely to
reach its critical temperature (~600°C or 1,150°F in a typical room), which then
induces flashover or full room involvement.
■ The fuel load of the room or structure is of paramount importance in the develop-
ment of a fire. This fuel load is not only the structure itself but also the furnishings
and contents, and the wall, floor, and ceiling coverings that feed a fire and offer it
paths and directions with optimum fuel conditions. In evaluating a fire’s progress
through a room or structure, the investigator must establish what fuels were present
and where they were located. The chemical nature of the fuels and their physical forms
will affect their ignitability and their expected heat release rate. In the reconstruction
of a fire, the fuel load is not just the total number of joules or Btu’s of heat that can
be generated but the rate at which that heat is released.
■ Variations on the upward spread of fire or smoke will occur when air currents deflect
the flame, when horizontal surfaces block the vertical travel, or when radiation from
established flames ignites nearby surfaces. If fuel is present in these “new” areas, it
will ignite and spread the flames laterally.
■ Upward, vertical spread is enhanced when the fire finds chimneylike configurations.
Stairways, elevators, utility shafts, air ducts, and interiors of walls and hollow support
columns all offer openings for carrying flames generated elsewhere, and fires may
burn more intensely because of the enhanced draft.
■ Downward spread will occur whenever there is suitable fuel in the area. Fire will
burn downward across a solid fuel but at a rate that is a tiny fraction of its upward
spread rate. Combustible wall coverings, particularly paneling, encourage the travel
of fire downward as well as outward. Burning portions of ceiling and roof coverings,
draperies, and lighting fixtures can fall onto ignitable fuels below and start new
fires that quickly join the main fire overhead (called “falldown” or “drop down”
fires). Radiation from rollover fires or very hot gas layers can ignite floor coverings,
furniture, and walls even at some distance, creating new areas of fire. The dynamics
of such fires were treated in Chapter 5. The resulting fire patterns can be complex
to interpret, and, once again, the investigator must remember to take into account
what fuel packages were present from the standpoint of their potential heat release
rate contributions.
■ Suppression efforts can also greatly influence fire spread. Positive-pressure ventilation
(PPV) or an active hose stream attack on one face of a fire may force it back into
other areas that may or may not have already been involved, and push fire down,
even under obstructions such as doors or cabinets. The investigator must remember
those “unusual” conditions and check with the firefighters present.
■ Fire tends to flow through a room or structure much like a liquid—upward in relatively
straight paths and outward and around barriers.
■ Ventilation from open doors, windows, or vents, or forced by heating, ventilating,
and air conditioning (HVAC) systems or PPV fans can affect fire growth and move-
ment as much as fuel arrangement. All ventilation must be documented.
■ The development of a fire in a compartment (which may be a room or a portion of
a room) can be likened to filling a leaky bucket from a water hose. One can fill
Chapter 7 Structure Fires and Their Investigation 251
 even a leaky bucket if the flow of water is great enough. Similarly, the hot gas layer
in a room can reach its critical intensity (producing a heat flux of ⬎20 kW/m2) if
the fire is large enough to overcome the “leaks” of hot gases due to openings or due
to radiant and conductive losses. As seen in the 1985 Bradford (UK) football stadium
fire, even a compartment that is open on three sides can develop flashover conditions
when the fire gets big enough.
■ Fire intensity in a post-flashover room is often greatest around ventilation sources,
so damage accumulates faster around door sills, windows, or other openings where
fresh air could be drawn into the room. These ventilation effects can even reach
walls opposite the opening.
■ The total fire damage to an object observed after a fire is the result of both the intensity
of heat applied to that object and the duration of that exposure (with the realization
that intensity of heat varies considerably during a fire and that all the exposure may not
have occurred at the same time).

IMPLICATIONS FOR THE FIRE INVESTIGATOR

Fire investigation and the reconstruction of a fire’s growth pattern back to its origin are
based on the fact that fires form patterns of damage that are, to a large extent, predictable.
With an understanding of combustion properties, heat transfer, and general fire behavior,
the investigator can reliably examine a fire scene for particular indicators. Using each indi-
cator as an independent test for direction of travel, intensity or duration of heat applica-
tion, or point of origin, the investigator can approximate the fire’s development. It must be
remembered that every fire forms a pattern that is chiefly determined by the configuration
of the room or structure, the availability of fuel, and ventilation effects. Because of the
upward-reaching tendency of flames, some conical form of fire damage is most common,
with the apex at the bottom being a source of fire (possibly a point of origin) and the fire
rising and spreading out from there. Naturally, this pattern will be altered by the presence
of obstructions or the availability (or lack) of fuel in an area. Thus, interior fires often
appear to have very complex patterns as the result of complex structural arrangement
rather than any unusual property of the fire. These complications have already been noted
in part and will be further discussed as needed in the relevant section on diagnostic signs.
In its simplest terms, the requirement is to identify the origin (the point or space
where the fire began) and then to examine that space for possible ignition sources. The
fuel first ignited at the origin must be identified (even by elimination) for a causal link to
be established between the ignition source and the final destructive fire. If the investiga-
tor cannot identify the fuel first ignited, the ignition source competent to ignite that fuel,
and the circumstances that brought that fuel and that ignition source together, the cause
cannot be conclusively established. This determination is not limited to the identification
of physical residues of both fuel and ignition source but may depend on the hypothesis
testing of observed fire behavior and damage.

Examination of a Structure Fire Scene

GENERAL CONSIDERATIONS
The successful fire investigation begins with the proper attitude on the part of the inves-
tigator. In the authors’ experience, a significant percentage of all structure fires today are
maliciously ignited. Therefore, every fire scene must be approached as a potential arson
fire from the standpoint of preparation, control, search, documentation, and collection of
evidence. The investigator must be careful, however, not to prejudge the cause of any fire
and should consider every one to be undetermined in causation until sufficient proof is
found. Failure to consider a presumption of innocence may lead to a tragic wrongful
prosecution. Arson is a most difficult crime to prove because it is often not apparent until
252 Chapter 7 Structure Fires and Their Investigation
the investigation is nearly complete. As a result, every fire investigation must be treated
with the caution and thoroughness of a crime scene until it can be clearly demonstrated
that the fire was accidental or natural in origin. Even accidental fires may have far-reaching
consequences from a liability standpoint, and the physical evidence may be just as crucial
to investigators from regulatory agencies or the insurance industry. (See ASTM E860,
E1188, and E145917 for guidance on preserving evidence.) There are also concerns of
spoliation of evidence in civil cases where evidence critical to establishing what caused a
fire and what did not has been destroyed or removed by the public-sector investigator. See
Chapter 17 for a discussion of spoliation issues.
This caution extends both to the security of the scene and to the initial dealings with
occupants and all witnesses in that individual accounts of what happened should be care-
fully cross-checked and verified. The scene must remain in the custody of the fire or police
department from the time of the fire to ensure the integrity of the scene and its evidence.
It is the responsibility of the fire department to ensure continuity of possession. If that
custody has not been maintained, a search warrant must be obtained or the owner’s per-
mission secured before any evidence can be rightfully recovered. Some of the legal con-
siderations of fire scenes are discussed in Chapter 17.
It is vital that the investigator be willing to do a complete investigation before he or she
begins. It must be appreciated that this may entail many hours of difficult and unpleasant
effort, sometimes under less-than-ideal conditions. Custody and control of the scene must be
retained until the investigation is complete. Supreme Court decisions have made it clear that
security of the scene must be strictly maintained both before and during the investigation.
The investigator must be willing to examine every fire systematically and thoroughly, paying
close attention to small details. The scene must be sketched and photographed to preserve
its appearance for later court inspection. These sketches and photos must be accompanied
by careful notes and an adequate report to permit an accurate reconstruction at any later
time. Because so many fires are arson caused, the fire investigator must be careful to exam-
ine every fire scene for things that look out of place—objects that do not belong in the area
in which they are found or are in positions that do not appear to be normal or logical.
Because of the tremendous damage that many structure fires do, it is attention to small
details and such vague properties as “out of place” that can make the difference between a
successfully identified fire origin and the frustrating opinion “undetermined cause.”

INTERVIEWS WITH FIREFIGHTERS

Every fire scene investigation follows a certain protocol or logical sequence (as outlined
in Table 7-1) that will vary slightly depending on the circumstances of the individual fire
and its investigation. If a fire investigator is fortunate enough to arrive at the scene while
the fire is still underway, first contact should be made with the fire officials present. The
ranking fire official can provide information as to the time of alarm, time of arrival, and
time required for control, as well as the names of owners, occupants, and witnesses. Copies
of all relevant departmental reports should be requested. The firefighters will be able to
provide valuable information as to the extent of involvement of the structure, behavior of
the fire, unusual odors, and whether any explosions took place. Because the suppression
activities can affect fire patterns inside, the investigator should determine where the structure
was entered and what attacks were made on the fire. Were windows and doors open,
closed, or locked upon arrival? Were entry paths blocked? Were doors or windows
blocked open (to improve draft)? Were doors locked? If so, how was entry made? Did
anyone try the lock before forcing the door? It must be remembered that an arsonist usually
has to gain entry to a building in the same way as a burglar or other criminal. Signs of
forced entry such as damaged locks or doorjambs, or broken windows are sometimes
overlooked by fire investigators who are focused on the fire and its aftermath. Where were
attempts made to ventilate the structure to aid in suppression? Were tactics like positive-
pressure ventilation used? Because such an operation can dramatically affect the course,
Chapter 7 Structure Fires and Their Investigation 253
 intensity, and time factors of a fire, the investigator must be made aware of their use and
of the response of the fire.
The observations of the first-in firefighters are often crucial to the reconstruction of a fire.
What was on fire and what was not? Was the fire localized on one side of the room, just at
floor level, just at ceiling level, or was everything on fire (i.e., post-flashover)? Was the smoke
so thick one could not tell? More departments are using thermal imaging cameras, and that
information may be useful in establishing where the hottest part of the fire was. With the use
of protective clothing and self-contained breathing apparatus (SCBA), observations of tem-
perature and odor will usually be very limited, so visual observations will be most important.
If the time of entry can be established from the radio log or arrival times of other units, a
timeline may be reliably established to at least trace the development of the fire from that
point. Fire crews should also be asked about their use or refueling of any gasoline-powered
tools or equipment, precautions taken, and locations of use or refueling.
The fire crews can also offer important information as to the weather conditions
(wind, temperature, humidity) and the nature of the utilities. Were the gas, electricity,
phone, and cable TV in service? Who disconnected them? Did any of the occupants or
witnesses speak to the firefighters? What did they say regarding the fire, the building, its
contents, or the activities of people around the building recently? The sum total of infor-
mation gathered by the firefighters combined with their own observations often offers the
key to the entire investigation. The skill is to gather the details, compare them with other
witnesses’ accounts, and make accurate assessments and records in order that conclusions
can be properly presented. Many fire departments today request a “debriefing” report
from every firefighter who attended, describing what was seen and done.

INTERVIEWS WITH WITNESSES

While the investigation relies on application of the scientific method, the gathering of
information from witnesses requires the art of communication. Just like fires, no two wit-
nesses are exactly the same. They may be traumatized or excited, helpful or obstructive.
The investigator must be skillful in adopting a demeanor and approach that will encour-
age the free flow of information and enable an accurate picture to be drawn. It is as
important to identify and properly react to the genuine concerns of fire victims as it is to
recognize suspicious behavior.
The investigator should attempt to discreetly canvass the neighborhood for witnesses
to the fire or people with knowledge as to the structure, its owners, occupants, or its past
history. He or she should determine who was in the building last and when and, most
important, what the furnishings and contents of the building were and how they were dis-
tributed. The combustible contents of a structure have the largest single effect on burn
patterns, and the pre-fire fuel load must be identified. Appendix H includes a field inter-
view form developed by the Chicago Fire Department to ensure that critical information
is not overlooked in any interview with firefighters or police officers. Also, an incident
report form used by the same department illustrates comprehensive documentation.
NFPA 921 and Forensic Fire Scene Reconstruction also include witness interview forms
and many other useful forms.
Civilian witnesses to the earliest stages of a fire or explosion can offer information to
the investigator that may be impossible to establish in any other way, especially in cases
where there is extensive destruction before the fire can be reported or where fire service
response is delayed. Their observations could include the following: what was on fire when
the fire was noticed, where in the building they first saw fire and/or smoke, how they first
become aware of the fire (explosion, smell, sound of breaking glass, smoke or fire alarm,
etc.), what unusual smells, sounds, vehicles, or people they noticed just before the fire.
When these interviews are conducted, it may help to walk with the witnesses to where they
said they were to see if lines of vision are unobstructed. Sometimes the process of re-creating
witness position and path triggers additional memories. These interviews will go a long
254 Chapter 7 Structure Fires and Their Investigation
way to helping the investigator establish the starting conditions, that is, what the scene
looked like at the start of the fire, a step essential to accurately interpreting the indicators.

SEARCH PATTERNS AND PRACTICES

As described earlier in this chapter, before beginning the actual search for evidence, the
investigator should examine the exterior of the structure from all sides. This walk-around
survey will give a fair idea of the limits of the fire damage, the areas of heaviest damage,
and the possibilities of an external source of ignition (such as a trash fire) or reveal obvi-
ous forcible entry. View lines can be established at this stage. What was visible to passers-
by or neighbors? Which neighbors have a view of the scene? Once the survey is completed,
it is useful for the investigator to step back from the immediate scene for a moment and
determine what areas surrounding the structure must be searched as well—for example,
along likely paths of entry or exit and the areas under windows or around doors. If a tall
adjoining building is not available, it is often helpful to use an aerial snorkel or ladder
unit to get immediately above the fire scene to facilitate observing, charting, and photo-
graphing the damage. Figure 7-9 illustrates the value of an overhead photo from an aerial
ladder—showing the location of the most damage.
Once an accurate idea of the scene limits is in mind, a perimeter is established, prefer-
ably by means of a physical barrier such as rope, banner, or barricades. All personnel not
essential to the search are then kept outside that perimeter for the duration of the scene
examination. This perimeter method provides both scene security and a good starting place
for the search that follows, since it sets the outer limits of the search pattern. Documentation
also begins before the search, normally at the outer perimeter, in the form of photos, notes,
and sketches. Some investigators find it useful to prepare a rough floor plan at this stage to
help plan the investigation and begin to put the scene features into context.
Most structure fires lend themselves to a spiral search pattern, beginning at the outer
perimeter and working toward the center of the fire. The search pattern will, of course,
depend on the size and shape of the area to be searched and the time and number of people

FIGURE 7-9 Photo from

aerial ladder shows extent
of damage. Courtesy of
Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

Chapter 7 Structure Fires and Their Investigation 255

 available to do it. The exterior areas of the scene must be carefully examined for shoe and
tire impressions that do not appear to be those of fire or police personnel or their vehi-
cles, discarded tools (e.g., from a forced entry), fuel containers, beer or soft drink contain-
ers, matchbooks, or shards of glass from windows. If an explosion preceded or
accompanied the fire, the exterior search will focus on any fragments of the structure or
explosive device. Because the trajectory these fragments followed must be reconstructed
to establish the seat of the explosion, it is vital that the location and distribution of any
such fragments be carefully charted on the crime scene sketch (see Chapter 12). (Special
attention should be paid to the distribution and distance of dispersal of all window glass.)
This process minimizes the chances of destroying evidence that was damaged by the fire
or explosion before its relationship to the scene has been assessed. As the investigator
works through the scene, the process is one of tracing the flow of the fire back to its ori-
gin and cause, which is often (but not always, for reasons to be discussed) in the area most
damaged. Indicators of direction of movement will be discussed in an upcoming section.
Large scenes may require division into sectors that are then searched in sequence, as dic-
tated by access or exigencies (limitations). An example is shown in Figure 7-10.
As a general rule, anything that looks out of place should be checked. Does the item
belong to the scene? Is it clean and fresh, as if recently dropped there, or stained and
weathered, as if exposed for many weeks? Is the grass underneath it still green? Anything
that looks odd or out of place should be charted on the scene sketch, photographed in
place, and then collected. It is always easier to disregard collected items if proven later to
be unrelated to the fire than to try to recover important items once they have been left
behind. All collected objects should be considered potential sources of latent fingerprints
and therefore must not be handled directly by more than one person, and packaged to
minimize damage and contamination by others’ handling them. Vinyl or latex gloves pre-
vent contact with chemical or biological hazards and minimize (but not always prevent)
deposition of fingerprints or DNA.
Proceeding toward the building itself, the investigator looks for signs of attempted
forced entry, such as tool marks, fresh damage to door or window frames, broken locks,
or broken windows. Are there signs of prior efforts at extinguishing the fire—a garden

FIGURE 7-10 String grid

divides this large scene
into sectors for searching.
Courtesy of Don Perkins, Fire
Cause Analysis.

256 Chapter 7 Structure Fires and Their Investigation

hose or an empty fire extinguisher? If so, an attempt should be made to find the people
responsible, since they may have valuable eyewitness information to offer about the early
stages of the fire. Documentation should include photos supported by notes or diagrams
showing the presence or absence of signs of forced entry at all possible access areas, as
well as signs of activity—tire impressions, shoeprints, extinguishers, and the like. All
external areas of the structure must be photographed before disturbing them.
No matter what the nature of the structure or the extent of the fire damage, nearly
every thorough fire investigation starts in the least damaged areas and progresses toward
the most heavily damaged areas—starts with the areas of least confusion and disruption
and proceeds to those most disrupted by fire or explosion. This process in part helps
establish the starting conditions in the fire damaged areas by showing the investigator
what wall and ceiling finishes were like, as well as furnishings. As the investigator works
through the scene, the process is one of tracing the flow of the fire back to its origin and
cause. It can be thought of as a search for what material was first ignited and the source
of heat that ignited it. Logic and knowledge of basic fire dynamics are the most impor-
tant tools to keep in hand. Materials exposed to a violent fire can do some unusual
things and exhibit unexpected properties, but the fundamental precepts of matter,
energy, and heat transfer cannot be violated. There must be fuel, oxidizer, and energy in
appropriate forms and amounts before there will be fire. The combination of fuel and
energy that the investigation indicates as the source of the fire must meet the scientific
and fire engineering requirements of energy release, fuel, and oxygen supply. As fuel
packages and potential ignition sources are discovered they should be plotted on a
detailed scene diagram. For the reasons we have discussed, the area most heavily dam-
aged is not always the area of origin, since the intensity and duration factors will be
influenced by fuel load, manner of ignition, ventilation, and fire suppression. The inves-
tigator must take all these into account when evaluating the damage pattern in the search
for the origin.

FIRE BEHAVIOR INDICATORS

Indicators are the visible, and usually measurable, changes to the surfaces and materials indicators ■ Observable
of a fire scene produced by smoke, heat, or flames. Because they are produced by pre- (and usually measura-
dictable physical processes, they can be used to re-create the path of the fire’s develop- ble) changes in appear-
ance caused by heat,
ment, the location of fuel packages, the nature of ventilation, and sometimes the height, flame or smoke.
intensity, and duration of the flames. There are many indicators that the investigator can
use in reconstructing the path of the fire. None of them are precise. None of them are
foolproof. None should be relied on to the exclusion of all others. Depending on the
nature of the structure and the extent of the fire, it may not be possible to apply some
indicators. The most accurate origin determinations are accomplished by applying a num-
ber of different indicators independently and then combining the results as a series of
“overlays.” The investigator examines the structure methodically, checking and recording
one fire indicator at a time. Each search will reveal some useful information about the
direction of fire travel (sometimes called “arrows”) and the relative intensity in each area.
After several such surveys, stacking the indicators up against each other will reveal points
of similarity and agreement. Further searches can then be concentrated in the general area
of suspected origin indicated by the combination of indicators.
Some indicators of fire travel result from the deposit of soot and other combustion
products on exposed surfaces without any chemical changes. Most other indicators are fire pattern ■
produced by heat-induced changes to surfaces. The intensity and duration of the heat Indicators of damage
applied will dictate the nature and depth of those changes. Because indicators of damage (or relative lack of
damage) created by a
(or relative lack of damage) form patterns, and are created by a combination of smoke, combination of smoke,
heat, and flames, they are sometimes called fire patterns. The diagnostic signs most often flames, and heat
used to indicate fire intensity and direction of travel are outlined in the following sections. effects.

Chapter 7 Structure Fires and Their Investigation 257

 Burn Patterns
Burn patterns are created when applied heat fluxes are above the critical thresholds to
scorch, melt, char or ignite a surface. Where the fire’s destruction did and did not take
place constitutes the burn pattern. Burn patterns are the cornerstone of all fire investiga-
tion because of their universal applicability. Data from a fire scene include the evaluation
of any and all heat transfer patterns. In seriously damaged structures, the question, Where
was the destruction most complete? may accomplish the same end. We know that fire
generally travels upward and outward, leaving, in the absence of barriers or unusual fuel
conditions, a V-shaped or conical pattern in the structure it leaves behind. In the absence
of active suppression, fire will usually have time to burn longer at its point of origin than
at places subsequently involved by spreading flames. As a result, destruction will be more
extensive in that area, all other things being equal. These include fuel loads and ventila-
tion, however, that are rarely equal throughout an entire scene. Figures 7-11a–d illustrate
patterns in a bedroom fire originating from a wastebasket fire.

FIGURE 7-11A Small waste- FIGURE 7-11B “V” patterns on wall and headboard point back
basket fire quickly spreads to toward wastebasket. Note thermal patterns at margins of pattern
drapery and bedding. Courtesy (charring to scorching to no damage). Courtesy of Jamie Novak, Novak
of Jamie Novak, Novak Investigations and St. Paul Fire Dept.
Investigations and St. Paul Fire Dept.

FIGURE 7-11C Damage pattern on bed (including partially collapsed FIGURE 7-11D TV set shows more collapse
springs) becomes more intense toward origin (lower right corner). and glass fracturing on side facing growing
Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. fire as fire grows and spreads around corner
from bed. Courtesy of Jamie Novak, Novak
Investigations and St. Paul Fire Dept.

258 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-12
Temperature distribution of
100 gases in a buoyant plume
(a) unrestricted (b) under
ceiling (temperatures in
300 degrees Fahrenheit).
Smoke
200 Temperature in °F
700 200 1000 900 800 700 600
Ceiling

400 1200
1000
300
Turbulent mixing
1300 400
spreads and cools
hot gas plume
Flames
1500 Margins of thermal
500 500 plume produce
“V” pattern on wall
1700
Fuel

(a) (b)

One of the factors that is not always equal is fire suppression. Fires are attacked at
what appear to be their most critical areas—where other buildings are threatened, where
victims are trapped, where dangerous contents are to be protected, and so forth.
Occasionally, suppression will be focused on the very area of a structure in which the fire
was started, and even though the rest of the building is destroyed, the area of origin may
be the most intact. That is why interviews of the suppression crews are so important to
revealing where the fire was attacked and with what tactic.
Burn patterns that can be used to indicate the direction of fire spread can be observed on
structural elements, wall surfaces, or even furnishings remaining in the building. Because
these elements are of most value when they remain undisturbed, it is most important that
overhaul of the structure be kept to an absolute minimum to prevent their destruction as evi-
dence. Application of heat to a material causes changes first only to its surface and if sus-
tained, changes to its bulk. If we apply the principles of heat transfer and the structure of
flame plumes from Chapter 3, we can understand how these patterns are formed. The tem-
peratures in a flame and smoke plume decrease with height in the plume or distance from the
centerline, as illustrated in Figure 7-12. Because the radiant heat flux intensity is controlled
by the temperature of the source, we can see why the pattern of damage from a moderate-
sized fire to an adjacent noncombustible vertical surface is often in the form of an inverted
“V” (as depicted in Figure 7-13). As room air is entrained into the rising flame and smoke
plume it will be spread out as it rises, but is will also be diluted and cooled. If the adjacent
surface is very susceptible to thermal changes or if combustion products are deposited as the
gases cool, there may be a pattern resembling an hourglass or a “V” in outline.18 The over-
all pattern most often resembles a “V”or “U” with its apex pointing toward the heat source.
Wherever the total heat flux onto the wall is below the critical intensity needed to
affect the surface, there will be a line of demarcation for that heat pattern. Areas outside
that demarcation will not be visibly affected. Note that this process is driven by the buoy-
ant flow of the hot gases, so the “V” (or inverted “V”) is produced on vertical surfaces,
not on horizontal ones. Because the gases cool as they move away from the centerline of
the plume, damage patterns to horizontal surfaces above the fire will tend to be circular,
with a “bull’s-eye” of the most damage directly above the fire.
A typical “V” pattern on a wall as a result of a localized fire is shown in Figure 7-14.
Notice how the pattern spreads out naturally, with the apex of the “V” right at the origin
Chapter 7 Structure Fires and Their Investigation 259
 FIGURE 7-13 Energetic
fire on chair in corner
produced an hourglass
pattern on a noncom-
bustible and on the
ceiling as a result of the
corner configuration. Note
gradations of thermal
damage. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

of the fire. Although it is sometimes claimed that the more vertical the sides of the “V,”
the faster the initial fire (and therefore the more suspicious), one can appreciate that the
nature of any wall covering or vertical target surface, size of the fuel package, height of
the ceiling, and conditions of ventilation determine the shape of the pattern. The angle
and width of the “V” are not dependent on the rapidity of ignition of the initial fuel but
on the fire properties of the target surfaces (see Figure 3-29). The “V” pattern associated

FIGURE 7-14 Fire that

started at end of sofa
produced “V” pattern of
damage on wall behind
sofa and on sofa. Note
clean burn on drywall,
thermal transfer pattern
on side wall, and enhance-
ment of ceiling damage
as a result of corner effect.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

260 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-15 The wide,
shallow “V” pattern on
the wall is due to a very
large plume flattened
by the low ceiling. The
room approached
flashover conditions in
under 6 minutes. Courtesy
of John D. DeHaan.

with most fires is a result of several factors, one being that the natural plume shape of a
large fire is a cone; the interaction of this plume with a wall or a ceiling is the major fac-
tor in producing a pattern. The shape of this cone depends on the size of the fire with
respect to a ceiling that tends to flatten and spread it horizontally, as shown in Figure 7-15.
A vertical wall intersecting this cone may then display a vertical cross section of the cone.
If a fire is located some distance away from the wall, the cone does not intersect the fire
cone at its base, so a shallow “U” or “V” pattern may appear, as in Figure 7-16. The ceiling
directly above a large fuel package may exhibit a circular bull’s-eye pattern of damage
(unless bounded by nearby walls), with the center reflecting the effects of the highest tem-
perature gases, and succeeding rings of less damage. Because hot gases moving across a
surface lose some of their heat by convective transfer, they cool, producing less damage
as they move away. If the fire is located against a wall, the circular pattern is interrupted
by the wall to form a semicircle (as in Figure 7-16).

Ceiling damage
Shallow “V” where
plume intersects wall

Hot gas
plume

FIGURE 7-16 A fuel

package burning in a
room may produce a cir-
cular burn on the ceiling
if the plume is large and
energetic enough. Vertical
Object near Intersecting wall
surfaces intersecting the
fuel has away from fuel
plume may have areas of
lower, more pronounced Fuel package
plume-induced charring.
“V” pattern (ignited)

Chapter 7 Structure Fires and Their Investigation 261

 Direction of
fire travel

Fire travel

Doors

FIGURE 7-17 Even if the wall covering is destroyed, the FIGURE 7-18 Beveling of exposed
“V” pattern is often reflected in the remaining structural combustible surfaces indicates that
members. fire moved from right to left.

A small fire of low energy and limited duration will typically display a triangular,
inverted-“V” or “A” pattern of damage to a nearby adjoining wall. This is especially true
if the wall or its covering is not readily ignitable. If the wall covering is very combustible,
the flame travel can be almost vertical, as shown in Figure 3-29g. If the wall covering has
been destroyed, the “V” pattern is often recorded on the wall studs still remaining, as in
Figure 7-17. Because the sides of an object exposed to an encroaching flame will be more
heavily burned than the protected sides, and the buoyancy of the hot gases causes them to
rise as they flow horizontally, the object will often have a beveled appearance that is more
pronounced on the side and edges facing the oncoming fire, as illustrated in Figure 7-18.
If a fire is traveling very rapidly, the hot gases will flow over recessed areas on the down-
wind side, leaving relatively undamaged areas on the side away from the fire.
Furniture will also exhibit burn patterns of use in determining point of origin and
direction of spread, since an object, whether upholstered or not, will generally be burned
more extensively on the surfaces exposed to an oncoming fire. In large combustible
objects, significant portions of a “V” pattern may be visible, as in Figures 7-19a and b.

Angle of burn pattern

on objects

FIGURE 7-19A The “V” pattern FIGURE 7-19B The chair on the right was ignited first and had time to burn completely
can be reflected in large items of down to the springs and base. The right side of the wine rack was burned away by the
furniture. radiant heat from the chair fire. The broad “A” pattern on the wall behind is consistent
with the location and size of the chair fire. The plastic receptacle plate is also burned more
on its right side (toward the initial fire). Courtesy of John D. DeHaan.

262 Chapter 7 Structure Fires and Their Investigation

If the original location of the furniture can be established by observation or by witnesses,
the patterns can be incorporated into the overall pattern observed in the room.
The before-and-after photos of a nonaccelerated test fire in Figures 7-20a–d demon-
strate a number of these features, including the “V” pattern on the corner walls, as well
as on both sofas. The paint and paper have been burned off the walls and the ceiling in
the corner around the origin. Radiant heat damage to the left side of the TV stand and the
armchair shows more intense heat coming from the left. The fire was approaching flashover
when it was extinguished, so the carpet in the corner where the fire began and the wood
floor beneath it were very badly burned by radiant heat. The carpet and furnishings closer

(a)

FIGURES 7-20A AND B

(a) Before and (b) after
photos of a test fire
ignited by setting fire to
right front corner of the
left sofa. The “V” pattern
extends up and back
from the origin, across
left sofa, and onto the
wall behind. Damage to
right-hand sofa is most
severe in the front center
(directly across from the
origin) and decreases to
right and left. Courtesy of
John D. DeHaan
(b)

Chapter 7 Structure Fires and Their Investigation 263

FIGURE 7-20C Overall
photo of scene shows
floor-to-ceiling plume
damage in the corner,
floor-level damage
between the sofas, and
higher damage farther
away. Courtesy of Keith
Parker, Marin County Fire
Department, Woodacre, CA.

FIGURE 7-20D After

removal of the sofas, the
protected area on the
wall shows the location
of the sofa back. Radiant
heat damage to the left
front corner of the chair
was produced by the sofa
fire. The “V” pattern
from the sofa includes
the left side of the TV
and stand. Note the
damage to the floor
caused by radiant heat
ignition of all-synthetic
carpet and pad. No
ignitable liquids were
used. Courtesy of Keith
Parker, Marin County Fire
Department, Woodacre, CA.

to the photographer were just being affected by the radiant heat from the hot gas layer.
The header formed by a structural beam restrained the hot gas layer from flowing into
the dining room. The protected area on the right (far) wall shows where the back of the
smaller sofa was in contact with it.
In another test fire, depicted in Figure 7-19b, the fire damage shows the beveling effect on
the wine rack resulting from the fire started in the chair in the corner. The chair was reduced
dropdown ■ The to its springs and metal base by the fire, and the floor around it was charred by the radiant
collapse of burning heat and from flaming dropdown of the synthetic upholstery falling on it during the fire. The
material in a room
that induces separate,
thin wooden top of the small table in the foreground is burned through by radiant heat.
low-level ignition; Beveling of exposed edges occurs when heat is applied and the temperatures of edges
falldown. and corners rise faster than those of adjacent flat surfaces (because of their more limited
264 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-21 Beveled
areas (in cross section)
reveal the direction of the
fire development through
the floor. Courtesy of
Lamont “Monty” McGill,
Fire direction Fire/Explosion Investigator.

thermal capacity, they cannot lose heat as quickly to the interior). That is why such edges
are easier to ignite than a flat surface and, once ignited, burn more quickly. This pre-
dictable behavior allows the investigator to use such beveling to re-create the direction of
travel of the fire gases. Directional travel can produce cross-sectional beveling of exposed
wood elements that can indicate the direction of the flow of hot gases around them. Such
beveling, then, can be used to plot pointers or “arrows” of fire spread direction on a plan
of the building.
This beveling effect can be very useful in establishing the direction of burn-through
of floors. Because of the complexity of most structure fires, it is not enough to record the
fact that the floor has been burned through in a particular area. It must be determined
whether the floor has been burned from underneath or from the top. Figure 7-21 illus-
trates the typical appearance, in cross section, of the two cases, that is, burned from
underneath and from the top. It is obvious that joists will protect the flooring immedi-
ately above them when the fire burns from underneath, but all the flooring surrounding
the hole will be destroyed first if the fire is burning downward.
Once these indicators of direction of fire spread (sometimes called vector or arrow
patterns) have been observed throughout the structure, they can be combined with obser-
vations of relative char depth and the heights of heat and smoke horizons in various rooms
to map out the fire’s approximate course of travel.19 The combination of direction and
intensity indicators is sometimes referred to as vector analysis. These vectors can be
mapped out on a floor plan to help establish a likely area of origin. This information, of
course, must be considered in light of what is known about the suppression efforts that
occurred at the scene. It is well known that venting a burning building or actively attack-
ing it on one side can force fire to reverse its course and attack new areas or reburn previously
affected areas. It has been observed that if a fire starts in one room and spreads to another
with a larger fuel load, the subsequent extension of the second fire can obliterate the direc-
tion indicators on adjoining walls, making it look as if the fire started in the second room.
Ventilation through windows, whether wind-driven or from PPV fans during suppression,
is a critical factor in evaluating burn patterns and fire behavior.20
Heat Level (Heat Horizon)
As hot gases are released into a room their buoyancy causes them to rise to the ceiling, and
then their accumulating mass and momentum causes them to start to flow outward. When
hot gases flow across a surface, the flame front is always at a marked acute angle to the
surface, as depicted in Figure 7-22, because of the physical properties of the density-driven
flow. In the same manner as when a dam bursts, and the gravity-driven flow of the “wall”
of water causes it to flow with its face at an angle to the ground (as illustrated in Figure 7-23),
the same occurs with hot gases. If the flame front comes into contact with vertical surfaces,
it produces a characteristic “front” that will tell the investigator the direction of flame
spread at that point at some time during the fire. The effect of such progression across a
thin-paneled, hollow-core door is shown in Figure 7-24.
Chapter 7 Structure Fires and Their Investigation 265
FIGURE 7-22 Hot gases
entering this room pro-
duced a front at an angle
of about 45° from the
ceiling. A layer of gases
moving across a surface
will not assume an angle
of 90° or more because
of the momentum flow
pushing it across the ceil-
ing. Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

Lake Dam (failing)

Advective (gravity-driven) flow

FIGURE 7-23 Advective (gravity-driven) flow of water FIGURE 7-24 Hot gases flowing from left to right (out door)
released from a collapsing dam assumes an acute angle produced characteristic “beveling” damage to a hollow-core
from the ground. door. Courtesy of John D. DeHaan.

266 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-25 A uniform
hot gas layer ignited
paneling near its entry to
the room and scorched
the paneling across the
room. Courtesy of Detective
Mike Dalton, Knox County
Sheriff’s Dept., Knoxville, TN.

As the hot gases continue to accumulate they fill the room from the top down, pro-
ducing a hot gas layer. In large rooms or with very large (high HRR) fires, there may be
times when the gas accumulates (in a corner, for instance) in a deeper layer immediately
above the fire than it does in the rest of the room; eventually the layer will fill the room
to a uniform height throughout (except near ventilation openings, where it rises) (see
Figure 7-25). The larger the fire or the smaller the room, the faster this hot gas layer deepens
until it can begin spilling out windows or under headers into adjacent rooms. As the hot
gas layer spills out into adjacent rooms it begins to fill those rooms from the ceiling down-
ward. This produces “shallower” layers in other rooms, with the “deepest” layer in the
room of origin. This process produces a characteristic stair-step pattern of smoke or heat
damage that will help trace the flow of the fire, as shown in Figure 7-26.

FIGURE 7-26 Fire moving

down this hall from far
end produced a stair-step
or rising heat horizon
pattern. Note damage to
paneling in next room
on left. Courtesy of Greg
Lampkin, Knox County, TN,
Fire Investigation Unit.

Chapter 7 Structure Fires and Their Investigation 267

 The height of significant levels of heat (i.e., the depth of the hot gas layer described
earlier) in a room is revealed by the charring, burning, blistering, or discoloration of paint
heat horizon ■ The or wall coverings. The heat horizon will generally be level, that is, of uniform height from
demarcation (usually the floor when the atmosphere in the room is fairly static. Changes in level indicate points
horizontal) of fire
of ventilation, such as open windows or doors or breaks in the walls or floor. Typical
damage revealed by the
charring, burning, or questions for the fire investigator include the following:
discoloration of paint ■ Do such points of ventilation agree with the known history of the fire, or do they
or wall coverings.
indicate that an attempt was made to encourage fire spread by opening walls, ceilings,
or floors prior to the fire?
■ Are there signs of such intentional venting?
■ If windows or doors were open, were they known to be open before the fire?
■ Were the windows or doors damaged before the fire or during it?
The heat level will often drop markedly in the vicinity of the point of origin or when
there is combustible material ignited lower down or along the base of the wall. Low burns
can also be produced by combustion of floor coverings during full involvement of the
room. As the hot gas layer deepens, the heat level will sink from near the ceiling as the fire
grows. At flashover with involvement of floor coverings, the heat level will be at or near
floor level (bases of walls). Ignition of a premixed quantity of natural gas, LP gas, or flam-
mable liquid vapors in a room tends to produce a scorching of wall surfaces that is uni-
form from floor to ceiling, while post-flashover burning usually produces deeper burning
with its longer duration. In rooms not involved in fire, the heat level may be revealed by
the distribution of melted objects—candles, plastic fixtures, even appliances—in the room.
The temperature of the heat source determines the intensity of heat flux it produces.
Therefore, the temperature profiles in Figure 7-12 will produce commensurate effects. As hot
gases move across a solid surface some of their heat is lost by convective transfer, so they cool
as they move away from the primary contact point. This produces a gradation of effects, as
shown in Figure 7-27, which allows testing of hypotheses about contact and movement.
Smoke Level (Smoke Horizon)
Even when there is no heat effect on a surface, surface deposits can reveal the height at
smoke horizon ■ which smoke has stained the walls and windows of a room (the smoke horizon), which
Surface deposits that reflects the nature and extent of the fire and the amount of ventilation. Smoke staining
reveal the height at can occur even in rooms that have not been damaged by the flames themselves. Such
which smoke and soot
stained the walls and
staining is produced by the condensation of liquid products of incomplete combustion
windows of a room (pyrolysis products) and the deposition of soot. As with heat levels, smoke levels will
without thermal change at points of ventilation and therefore may be useful in estimating which doors and
damage. windows were open during the fire.
The smoke level may also be useful in assessing the cause of fire deaths. For instance,
it may be possible to estimate if a person of a certain height would have been able to see
an exit if standing in the room during the fire or whether he or she would have been over-
come by the cloud of hot toxic gases accompanying the smoke. Occasionally, victims who
did not stand up during a smoky fire escaped, while people who did were quickly over-
come. The heat and smoke levels throughout a building are most conveniently charted on
a separate scene sketch for later comparison with other data.
Buoyant flow of hot smoke can produce a “backsplash” or “mushrooming” effect if the
hot gases have sufficient momentum when they encounter a wall. Just as water waves pile
up against a breakwater, hot smoke can “pile up” against a wall producing a dropdown of
the smoke horizon (sometimes called momentum flow). Flow through a door often leaves a
void or clear area on the “downstream” side of the header, as illustrated in Figure 7-28.
Low Burns and Penetration
We examined the dynamics of the fire environment in previous chapters and can now
appreciate the reasons why low burns on walls or burned floor materials may be significant.
268 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-27
Gradations of thermal
damage on this wall
show the intensity of the
plume as it cooled from
the centerline out and
vertically. Smoke deposits
(nonthermal) surround
the upper portions of the
pattern. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

The lowest point of burning in a structure or room must always be examined carefully,
for it may well bear evidence of the cause of the fire, as well as its origin. Fires ignited in
a room’s normal fuel load most rapidly burn vertically upward from their point of ignition,
leaving the floors and lower reaches of the walls much less damaged than the ceiling and
upper walls. Any localized area with a floor burned or a wall burned right down to the

Header

Clear area on “downstream”

side of header

Momentum flow

Mushrooming effect
visible on wall

FIGURE 7-28
Momentum flow and
doorway flow, leaving
clear area “downstream.”
Fire Deeper layer exists in
room of origin.
Door
Chapter 7 Structure Fires and Their Investigation 269
 floor should be considered deserving of further investigation. Such a burn by itself does
not mean that the fire is incendiary in origin. It means that high temperatures or high heat
fluxes were encountered at floor level. Something normal to the room may have caused
it to burn in this fashion, such as combustible floor coverings, post-flashover burning,
falldown, or collapse. Each possible hypothesis has to be tested and eliminated.
During the free-burning phase of a room fire after flashover, the radiant heat is
extremely intense (120 to 150 kW/m2), and the combustion in the room becomes unsta-
ble and extremely turbulent. The intense radiant heat causes ignition of floor coverings
and the floors themselves, producing a generalized (although not always uniform) burn-
ing. The synthetic carpet and pad materials in common use today melt and ignite read-
ily upon exposure to these post-flashover radiant heat fluxes. In a post-flashover fire,
when the size of the fire is controlled by ventilation, the most intense burning will be in the
vicinity of ventilation openings (see Figures 7-29a and b). The very high flame temperatures

FIGURE 7-29A
Extensive melting of
synthetic carpet caused
by radiant heat (pre-
flashover). Notice the
pattern is very irregular.
“Clean burn” on right wall
is due to ventilation (fresh
air drawn into fire at floor
level). Courtesy of SA/CFI
Michael A. Marquardt, ATF,
Department of the Treasury,
Grand Rapids, MI.

FIGURE 7-29B The

same carpet in a room
fire ignited with gasoline
shows similar irregular
melting. Charred and
scorched area shows
location of gasoline pour.
“Clean burn” beside chair
is a ventilation effect as the
energetic gasoline fire
beside chair entrains fresh
air from front of room.
Courtesy of SA/CFI Michael A.
Marquardt, ATF, Department of
the Treasury, Grand Rapids, MI.

270 Chapter 7 Structure Fires and Their Investigation

of burning carpet, pad, and floor materials can be sustained as long as there is fuel—
much longer than the typical flammable liquid fire—and can induce fire damage down
to floor level. Even before flashover occurs, the room is exposed to high heat fluxes
across a wide area, producing a generalized (although not always uniform) scorching
and melting. Localized low burns surrounded by large areas of undamaged floor are
unusual and deserve close attention. The pre-fire conditions of the floor are important
because even combustible trash or clothing will protect the floor covering from radiant
heat, leaving the exposed areas to be damaged by the full brunt of the post flashover fire.
Figure 7-29 shows two identically furnished rooms, one accelerated with gasoline, the
other ignited by open flame, that were extinguished just as flashover occurred (with ceil-
ing and floor temperatures of 815°C (1,500°F). Note that the synthetic carpets melted in
both rooms in an irregular fashion, probably due to their nonuniform contact with the floor
trailers ■ Long trails of
(from curling). The pour pattern of the gasoline is still visible but would be obscured once fast-burning materials
the carpet and floor were fully ignited by post-flashover radiant heat. Figures 7-30a–d used to spread a fire
illustrate the flames and post-fire patterns of a gasoline trailer in a carpeted room. throughout a structure.

FIGURE 7-30A Ignition of gasoline trailer from door to sofa. Note FIGURE 7-30B Seconds later, flames have traveled along trailer
flash (blue) flame extending through vapor layer from pool. Courtesy against the base of the wall to recliner. Courtesy of Jamie Novak, Novak
of Jamie Novak, Novak Investigations. Investigations and St. Paul Fire Dept.

FIGURE 7-30C Fire extinguished just before flashover. Post-fire (after FIGURE 7-30D Damage to left side of recliner down to floor level
removal of some of suspended ceiling). Trailer pattern on carpet barely should prompt further exam. Drapery collapse may have caused
visible on left. Uniform damage across sofa and to floor level to walls this, but residues of the gasoline trailer in the carpet would betray
to left and right is not consistent with limited damage to other fur- an arson cause. Courtesy of Jamie Novak, Novak Investigations and St. Paul
nishings. Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. Fire Dept.

Chapter 7 Structure Fires and Their Investigation 271

ignitable liquid ■ One possible cause of such low burns is the ignition of a pool or trailer of ignitable
Classification for liquid liquid on the floor. When localized burn damage extends right down a wall to the floor
fuels including both
flammable and com-
surface, it may indicate that an ignitable liquid was used, since other sources of intense
bustible classes. local combustion, such as trash, usually do not burn right to floor level. The flames from
a pile of flammable trash burn upward in the typical convective cone from the fuel sur-
face and contact the wall higher than floor level. Sustained combustion of a pile of cellu-
losic clothing, bedding, or furniture that has collapsed onto a wood floor, however, has
been seen to cause penetrations through floors. Ignitable liquids alone will not burn long
enough to achieve generalized penetrations of wood floors. However, a liquid can soak

FIGURE 7-31A Post-

flashover fire deeply
charred all exposed
surfaces down to floor
level. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

FIGURE 7-31B View looking up the stairwell shows FIGURE 7-31C View looking up the stairwell shows
ignition of top step. Courtesy of Jamie Novak, Novak ignition of lower steps. Courtesy of Jamie Novak, Novak
Investigations and St. Paul Fire Dept. Investigations and St. Paul Fire Dept.

272 Chapter 7 Structure Fires and Their Investigation

into the crack that separates the wall from the floor and take the fire to the very bottom
of the wall. A puddle or trail of flammable liquid is the most commonly detected arson
device. The use of these liquids often produces effects that are most characteristic if the
room has not gone to flashover. If the corners of a room are burned out before other
exposed areas of the room are damaged, that is unusual fire behavior, because corners are
usually dead air spaces that are destroyed last in accidental fires. Post-flashover fires,
however, can generate so much radiant heat and be so turbulent that destruction can
extend into the corners and even down stairs, as shown in Figures 7-31a–c. Older struc-
tures may have uneven floors, and any liquid poured on them in quantity seeks out the
lowest surface. Such pools of poured liquid provide a longer-duration fire with potential
to create more localized damage.
The behavior of ignitable liquids in a fire follows certain physical rules, which inves-
tigators must appreciate if they are to detect such accelerated fires. These physical rules
include the following:
■ Liquids (and their vapors) flow downhill, seeking out the lowest available surface.
They can flow under doors and even under the toeplate of wood-frame partition walls.
■ Liquids will seep into seams or cracks in a floor and, from there, provide a reser-
voir of liquid to sustain more prolonged burning and localized charring.
■ Very volatile liquids (camp fuel, acetone, alcohols, lacquer thinners, and even
gasoline) will flash quickly off a nonporous surface. [In burn tests of 1- to 4-L (1- to 4-qt)
quantities, which produced a freestanding pool less than 0.5 mm deep, a regression rate
of 0.5 to 1 mm/min (0.02–0.04 in./min) and a total burn time of about 1 min was expected
and observed].21 Because of the low-boiling-point liquids involved, such fires produce
minimal surface scorching (see Figures 7-32a and b) compared with the relatively deep
charring induced by burning an overturned cardboard box filled with crumpled news-
paper (Figure 7-32c). Only the reservoirs along cracks and seams provide fuel for deep
charring (as in Figure 4-13e). Scorching and some sooting of the floor surfaces sur-
rounding the pool are likely, particularly on the downwind side of a pool burning in a
draft (as in Figure 7-33). The floor under the pool may be discolored by the solvent action
of the liquid or damaged by the brief exposure to flames as the last of the liquid boils
off. Flooring with very low thermal damage thresholds such as modern vinyl sheet
flooring may be scorched or distorted by a burning gasoline pool.
■ Less-volatile fuels with a higher boiling-point range such as paint thinner or kerosene
will exhibit different effects. Recall that it is the vapors of liquid fuels that burn rather than
the liquids themselves. The more radiant flames from these fuels produce a more pronounced

FIGURE 7-32A Gasoline burn pattern on plywood. Courtesy of FIGURE 7-32B Paint thinner burn pattern on plywood. Courtesy
John D. DeHaan. of John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 273

 FIGURE 7-32C
Cardboard box with
crumpled newspaper
burn pattern on plywood.
Courtesy of John D. DeHaan.

FIGURE 7-33 Sooting

and scorching around the
margins of the discol-
oration pattern on a vinyl
floor reveal the presence
of a flammable liquid.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

charring on the floor surface surrounding the liquid pool, often producing a halo or ring of
damage. The higher boiling point of these liquids allows the temperature of the liquid pool
to rise enough that the floor surface beneath the pool may be scorched (see Figure 7-34).
These liquids can form deeper pools on nonporous surfaces and burn more slowly than
the volatile fuels and have more time to spread into floors and penetrate crevices and porous
edges.The resulting pattern can be a deep ring of char damage around the edges of a lightly
charred area that defines, to some extent, the liquid pool (as in Figure 7-35). However, a
similar pattern can be created by the burning of thermoplastic materials (including uphol-
stery) that melted and burned as a semimolten pool. Residues of unburned polymer should
be found adhered to the floor in such cases.
■ When first ignited, any liquid fuel burning on a carpet will produce a pronounced
ring or halo of damage on the adjacent carpet pile, leaving the carpet pile in the center
of the pour undamaged (the pile acts as a wick) until the liquid is nearly consumed, as
depicted in Figure 7-36.22 When that last liquid is burned, the center of the pool will
scorch, char, and ignite. On some carpets, this sequence will leave a charred area sur-
rounded by a more deeply charred halo. On carpet, the size of the damaged area is not
274 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-34 Test burns
on wood plank floor.
Foreground burn is burn
pattern resulting from
lacquer thinner: The
center pattern is kerosene,
and the rear pattern is a
gasoline and diesel fuel
mixture. Courtesy of John
D. DeHaan.

FIGURE 7-35 Halo

pattern around a pour
of gasoline on a carpet
in an arson fire. Note the
unburned carpet in the
center of the pool.
Courtesy of Jamie Novak,
Novak Investigations and St.
Paul Fire Dept.

FIGURE 7-36 This

polypropylene carpet
continued to burn for
more than 45 minutes
after being ignited by a
small, localized pool of
gasoline (located near
the center of the carpet
panel). Courtesy of John
D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 275

FIGURE 7-37 Linear burn pattern on car- FIGURE 7-38 Floor after a test fire (see Figure 7-20). The carpet was destroyed
pet from burning oxygen cannula tubing. by radiant heat in the corner between the sofas, and the floor was deeply charred
Courtesy of Detective Mike Dalton, Knox County by radiant heat and the falldown of burning upholstery materials (no accelerant).
Sheriff’s Office, Knoxville, TN. Courtesy of John D. DeHaan.

always a reliable indicator of how much fuel was involved or even how large the liquid
pool was prior to ignition because carpets vary greatly in the amount of penetration
(into the pad) and the response of the carpet to radiant heat from the flames.23 Data
and observations from tests of similar carpet and pad can support conclusions about
pool size versus quantity and fire behavior. New all-synthetic carpets can self-sustain
combustion and produce extensive burn patterns that bear no similarity to the original
pool of liquid fuel that ignited them, as in Figure 7-36.24
Oxygen masks and cannulas fed by clear plastic tubing are becoming common in resi-
dences today. If the oxygen-saturated tubing is ignited, a blowtorch effect rapidly burns
the tubing away, leaving behind a narrow trailer of burn damage (see Figure 7-37).
Floor Burns and Penetrations
The investigator must be aware of the effects of radiant heat from burning fuel packages, or
from masses of hot gases in a post-flashover room or moving through ventilation openings
such as doors and windows.25 The mixing of fuel-rich hot gases with fresh oxygen-rich air
as they leave such openings produces very high temperature flames and enhanced radiant
heat output, causing enhanced thermal damage. Floors can be charred and even ignited at
heat fluxes of 20 kW/m2 or greater even if no fuel is on fire in the immediate vicinity. Figure
7-38 shows the charred floor produced by the radiant heat produced by two sofas (Figure
7-20) and the total combustion of the all-synthetic carpet and pad. Areas under the two
sofas were burned as the synthetic upholstery materials ignited and collapsed to the floor.
Due to the effects of radiant heat from burning surfaces and combustion of falldown fuels,
it is not sound practice merely to clear a room of its burned contents and rely on the burn
patterns on the floor as indicators of fire location. Post-flashover burning is very intense but
nonuniform and unstable and can produce floor burns that bear no relationship to fuel loca-
tions, as illustrated in Figure 7-39. Synthetic carpets and pads melt in such exposures, leav-
ing irregular pools of melting and burning residue and equally irregular areas of exposed
floor.26 The ignition of floors and floor coverings can melt aluminum threshold strips that
276 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-39 Prolonged
post-flashover produced
deep and irregular floor
damage in this room.
Courtesy of John D. DeHaan.

tests have shown will not be melted by ignitable liquid fires alone. The sustained flames of
the burning floor coverings add to the radiant heat input from the room fire and can over-
whelm the high thermal conductivity of the aluminum. Figure 7-40 shows the difference in
effect on an aluminum threshold between exposure to gasoline and diesel fuel fire and a
post-flashover fire (⬍10-min duration) with no ignitable liquids.
Often, the investigator is tempted to identify any burning through a wooden floor as
having been caused by the ignition of a flammable liquid. As discussed previously, fire
burns downward into a solid fuel very slowly. While penetrations can be caused by a liq-
uid’s seeping into the seams and joints of a leaky floor and then being ignited along the
seams or underneath, this is usually not the cause. Penetration is more likely to occur due

FIGURE 7-40 Aluminum

carpet strips after fire
exposure. Top: Exposed
to post-flashover fire
(with carpet beneath).
The strip is beginning
to melt. Duration of fire
approximately 10 minutes.
Bottom: Exposed to
gasoline–diesel fuel fire
on plywood. Note the
strip is sooted and
scorched but the metal
is still intact. Duration of
fire about 2 minutes.
Courtesy of John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 277

 to the radiant heat from a prolonged post-flashover room fire; from the sustained burn-
ing of a substantial fuel package like a sofa, bed, or overstuffed chair; or from prolonged
burning of other fuels that fall onto the floor from above while still burning energetically.
Sustained smoldering combustion of such fuels as draperies or upholstered furniture has
been observed to cause isolated, localized penetration down through a wood floor.
Other fuels burning from underneath must also be considered. Unusual floor con-
struction, fuel materials stored on the next floor down, or unusual air draft conditions
can all contribute to floor penetration. Such penetrations are more common in post-
flashover room fires in the vicinity of ventilation openings. The beveling effect of the burn
pattern discussed earlier will be of great help in establishing whether the fire came from
above or below as it penetrated the floor.
An unusual pattern of fire penetration was observed in a test one of the authors con-
ducted on an all-wood structure. The room had a tight-fitting tongue-and-groove floor
over a normal wood subfloor. A substantial fire was ignited without the use of an
accelerant, allowed to burn (post-flashover), and then extinguished before the floor was
significantly penetrated. The burn pattern on the subfloor (diagonal planks) is shown in
Figure 7-41. The diamond-shaped holes correspond to the overlap of gaps between the
subfloor planks and the finished flooring.

FIGURE 7-41 The burn

pattern on the wood
subflooring (on the left)
shows unusual penetration
in spite of the absence of
any flammable liquids. This
room sustained extensive
post-flashover burning.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

278 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-42 Isolated
burns on kitchen floor
resulting from falldown
of cabinet (and its con-
tents) above the sink.
Note “low burn” on front
of cabinet from falldown
fire. Courtesy of Jamie Novak,
Novak Investigations and St.
Paul Fire Dept.

Isolated burns can be produced by burning falldown fuels (as in Figure 7-42), so the
investigator must be careful about ascribing an ignitable liquid cause. Once again, a sketch
of the fire scene is made with the extent and distribution of any low burns or floor pene-
trations recorded on it. This is retained for later comparison with the other indicators.
Observations must also be made of any possible falldown sources—cabinets, draperies,
light fixtures, wall paneling, and ceiling materials.

Char Depth
The depth to which the pyrolysis action of fire has converted an organic material (usu-
ally wood) to its volatile fractions and charcoal is called the char depth. As we saw in char depth ■ The
Chapter 5, heat applied to wood surfaces causes progressive destruction of the cellu- measurement of
lose and other organic constituents of the wood. This charring will progress through pyrolytic or combustion
damage to a wood
a layer of wood at a somewhat predictable rate. Some investigators have been known surface compared with
to place considerable emphasis on the time dependence of char depth, thereby claim- its original surface
ing an ability to precisely time the duration of fire exposure throughout a structure. height.
The char rate of structural wood [which is often cited as being on the order of 2.5 cm
(1 in.) in 45 minutes] is, first of all, not a linear phenomenon. That is, wood chars very
quickly for the first few fractions of an inch and then at a much slower rate, presum-
ably due to an insulation effect of the newly formed char and to reduced air/fuel inter-
action. It is only an average rate that can be projected, so any attempt to interpolate
short times of fire exposure precisely based on fractions of an inch of char depth is
seriously misleading.
Chapter 7 Structure Fires and Their Investigation 279
 Most significantly, it is not the time of burning, but the intensity of the heating (which
varies locally and with time to a great degree) that most affects the char rate, as seen in
Figure 5-9. The higher the intensity of the applied heat, the faster the charring, and the
deeper the resulting char in a given amount of time.27 Anything that affects the intensity,
such as ventilation or localized fuel, will affect the char depth. Different species of wood
will also char at somewhat different rates. The density of the wood is a major factor in
char depth, as it is with regard to the rate of spread of a fire. The same wood will char at
different rates on faces cut across its grain than on faces cut with it (and create different
char patterns) due to the differences in heat conductivity and transport of fluids within
the wood. The adherency of the char is also an extremely important variable with certain
species such as redwood. Variations of exposure created when char flakes off in some
areas and adheres in others further complicate the relationship between char depth and
fire exposure. Edges and corners cannot lose heat to the interior of a fuel mass as well as
flat surfaces and will char more quickly. This difference in heat loss can produce cross-
sectional beveling on exposed wood members that can indicate the direction of flow of
hot fire gases around them. Even the dimensions may have an influence because of their
variable effects on heat capacity and conductivity. To complicate the practical picture fur-
ther, misinterpretation may follow because water may have been applied in one area in
fighting the fire, inhibiting charring there, while water may not have been applied in
another area close by, although the general fire intensity was similar.
Although char depth cannot be used to establish precise times of fire exposure, it can
be used, if applied systematically, to assess the relative fire exposure throughout a struc-
ture. Figure 7-43 illustrates a charred wooden beam in cross section. When testing for
depth of char, remember to measure depth from the original wood surface, not the exist-
ing char layer (which may often be impossible). Use whatever small-diameter, blunt-
tipped tool is most convenient to test for the depth at which intact wood remains. (A
sharp knife point should be avoided, since it tends to easily penetrate into undamaged
wood; some investigators prefer a tire tread-depth gauge or plastic caliper, since it has a
blunt wire indicator.) Use the same tool throughout the scene. Take char depth readings
on similar wood surfaces at the same level(s) throughout—for instance, waist height,
head height, and knee height—logging all observations on an appropriate sketch of the
scene. Some investigators prefer to plot an isochar diagram of the scene by connecting
points of similar char depth. The resulting contours (resembling the isotherms of a
weather map) outline the areas of deepest charring. These areas can then be correlated
to fuel packages and ventilation sources. Remember, with all else being equal, char depth
is most likely to be at its maximum at or near the fire’s point of origin (where it burned
longer) or near another localized source of fuel or ventilation (where it could burn more
intensely). Distribution of char depth must be correlated to the fuel load and ventilation
to be of any value.

Original Depth of char

wood
surface

Measuring tool
FIGURE 7-43 Proper use
of blunt-tipped measuring Wood
tool to estimate char
depth. Courtesy of Lamont Char
“Monty” McGill, Fire/Explosion
Investigator.

280 Chapter 7 Structure Fires and Their Investigation

 Variations in char depth may be of use in positioning fuel packages and establishing
ventilation conditions. Due to the rapid combustion of wood in high heat flux exposures
and the production of such exposures around ventilation points such as doors or win-
dows in post-flashover fires, floor penetrations in or near door openings in such fires have
no value in indicating placement of accelerants. In a post-flashover room fire, the interior
of the room may be burning in an oxygen-deficient condition except around ventilation
sources. There may be deeper charring or clean burn patterns around such vent openings clean burn ■ An area
or in the flow path of incoming fresh air. The analysis of post-flashover rooms by Carman of wall or ceiling where
has revealed that ventilation effects not only enhance combustion around ventilation the charred organic
residues have been
points but can produce clean-burn patterns of fire damage across the room from the ven- burned away by direct
tilation source. This may give the appearance of a “V” pattern where there was no local- flame contact.
ized fuel package.28

Appearance of Char Surface

The appearance of the char surface may be indicative of the fire conditions. Many inves-
tigators rely on a rule that says that large, shiny blisters of char (alligatoring) indicate a alligatoring ■
fast-moving, rapidly burning fire, while small, dull checkering indicates a slow-burning Rectangular patterns of
fire—without realizing that many of the factors mentioned previously affect surface char formed on burned
wood.
appearance as well as char depth. While a deeply seated, flat, or “baked”-appearing char
is most often produced by prolonged heating, as was seen in Figure 6.15c, it can also be
produced by a short-duration fire under some fire conditions (very intense fire with lim-
ited ventilation).
One observation of the authors (which awaits systematic testing) is that charring of
wood surfaces in oxygen-deficient conditions (such as the underside of a wood floor deck)
is more likely to produce flattened, dull, sharp-edged char. This char formation is caused
by heat-driven shrinkage. Insufficient oxygen reduces the combustion rate of the exposed
margins. The undependability of the size of the checkering or alligatoring is illustrated in
Figure 7-44, where 1-in. by 6-in. wood planks in the same wall demonstrate remarkable
variation in char appearance. One study by Ettling showed a relationship between temper-
ature of exposure and size of char blisters (on one species of wood)—the higher the tem-
perature, the smaller the blisters. However, oven exposure at a constant temperature for
various times caused differences.29 In a real fire where both times and temperatures are
unknown, one cannot reliably assess the effects of both. Exposure to low (below 300°C;
600°F) oven temperatures did produce a characteristic sunken appearance of the wood sur-
face with infrequent cracks or fissures that wandered over the surface in Ettling’s study.

FIGURE 7-44 Fir boards

on the wall illustrate the
variability of alligatoring
dimensions. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

Chapter 7 Structure Fires and Their Investigation 281

 FIGURE 7-45 The differ-
ence in appearance in
char and pyrolysis zone
demarcations between
(left) slow- and (right)
fast-developing fires.
Courtesy of Lamont “Monty”
Lengthwise
McGill, Fire/Explosion
cuts
Investigator.
“rip cut”

One indicator that may be more useful is the appearance of the charred wood in cross
section. As we saw in Chapter 5, the thermal changes in wood occur at characteristic tem-
peratures of short duration. As we saw in Chapter 3, an intense heat exposure to a high-
intensity heat produces a sharp, steep gradient between a high-temperature layer and
unaffected material below in a poor conductor like wood. Spearpoint and Quintiere con-
firmed such gradients in wood.30 The transmittal of heat through the wood has less time
to occur in a sudden, fast-developing fire, and the resulting pyrolysis zone is very narrow.
As a result, there is often a sharp line of demarcation between the charred and unburned
layers of the wood. When a charred beam is cut lengthwise, the gradation between the
charred layer and underlying undamaged wood is more gradual with a prolonged, slowly
developing fire. Figure 7-45 illustrates the contrast. No matter how reliable an indicator
such gradients may be regarding the speed of development of a fire, it must be noted that
a fast-developing fire may or may not be accelerated (i.e., arson) depending on what fuels
are found in the room and how it was ignited. An accidental fire involving direct-flame
ignition of a polyurethane mattress can produce flashover in a room just as quickly as
intentionally ignited gasoline.
Surface Effects
Radiant heat fluxes on the order of 10 kW/m2 (or prolonged contact with hot gases over
150°C; 300°F) will cause surface thermal effects such as scorching of wood or thermoset-
ting polymers or melting of thermoplastic coatings. Higher heat fluxes or high tempera-
tures will induce charring even without ignition. Such indicators need to be documented
as to type, location, and depth. Comparison samples of undamaged materials are needed
to allow identification and establishment of critical heat fluxes or temperatures.
Displacement of Walls and Floors
Once again, it is vital that extensive overhaul and dismantling of the structure not have been
conducted, because such activities may make it impossible to determine whether walls,
floors, or ceilings were moved by the effects of an explosion prior to or during the fire. It is
a fairly obvious observation to make if the structure is still standing, as in Figure 7-46. The
mechanisms and effects of explosions will be treated in full in Chapter 12. The amount of
material moved and the degree of fragmentation of the resulting pieces are the critical obser-
vations. Walls may be dislodged at their tops, at their bottoms, or both, depending on the
directional effects of the developing blast or its reflections, or simply because one edge of
the wall was fastened into the rest of the structure more rigidly than another.
The majority of the examination for displacement is easiest to carry out during the
examination of the exterior, but damage is sometimes readily detected upon checking the
interior. Window glass is the most readily broken structural material and may be thrown
large distances by an explosion. Diffuse vapor/air explosions may not break window panes
but may knock them free of their frames and toss them intact some distance away. All
282 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-46 A gasoline
vapor explosion in this
frame structure produced
this displacement of walls
and structural members.
Courtesy of John D. DeHaan.

materials moved by an explosion should be carefully charted and compared against their
original positions so that the direction of force can be assessed. If the structure collapsed
or was razed, careful examination of the floors or foundations where the sole plates of
the walls were attached sometimes may yield indications of sudden disruption. The inves-
tigator must measure and document the displacement of all structural elements.

Spalling
The chipping or crumbling of a concrete, plaster, or masonry surface is called spalling. It spalling ■ Crumbling
can be caused by the effects of heating, by mechanical pressure, or by a combination of or fracturing of a con-
both. Concrete is an aggregate of sand and gravel held together by a crystalline matrix crete or brick surface as
a result of exposure to
(cement) containing a considerable quantity of water of hydration. When exposed to heat, thermal or mechanical
some of this water is released from both the crystalline matrix and the aggregate (gravel), stress.
and the material loses its integrity and crumbles. While concrete is very strong in com-
pression, it is relatively weak in tension. As a result, heating may cause some types of sand
or gravel aggregate to expand more than others, which puts the cement matrix into ten-
sion, resulting in failure. This material property is a variable that many experiments in
inducing spalling have overlooked.31
Moisture trapped in the porous structure of concrete or masonry is also evaporated
into steam by heating. Forces generated by the expanding steam then can disrupt the
cement, causing it to crumble. (The inclusion of polypropylene fibers in concrete can pro-
vide pressure relief as the fibers melt during fire exposure, reducing spalling.) This steam
expansion is especially true of recently poured (green) concrete and masonry. The authors
know of one case where a large fire was built in a just-completed masonry fireplace, and
the resulting steam explosion created enough fragmented stone to fatally injure the
installer. Heat expansion of steel reinforcing components within a concrete floor, beam,
or column can also cause failure of the covering concrete. Steel reinforcing rods have a
higher rate of thermal expansion than concrete, so steel components, if not buried deeply
Chapter 7 Structure Fires and Their Investigation 283
 FIGURE 7-47 Radiant
heat from ignitable liquid
fire on a concrete surface
causes evaporation of
liquid fuel and warms
the surface beneath.
The temperature of the
surface outside the pool
will rise as radiant heat is
absorbed by the concrete. Radiant heat
The temperature of the
surface beneath the pool
is limited by the boiling
point of the liquid in
contact with it.

enough, can heat, expand, and distort to displace overlying concrete, compromising the
mechanical strength of the beam, column, or floor. Other forms of mechanical stress can
cause spalling of concrete or masonry, especially high mechanical pressures or the shock
of rapid cooling of a superheated concrete surface by a stream of cold water. Although
these other causes can contribute to spalling, by far the most common cause of spalling
of structural concrete in structure fires is localized heating as a result of some ordinary
fuel load in the immediate area. The formation of the crumbling surface is quite variable
and depends on the nature of the aggregate used, the age of the concrete, the amount of
mechanical restraint (as in prestressed panels or reinforced slabs), and the rapidity of the
rise in temperature, as well as the temperature itself.32 Because spalling can be triggered
by thermal expansion, it is important to consider the physics of heat transfer from a fire
on a concrete surface into that concrete. As shown in Figure 7-47, the radiant heat from
the flames first raises the temperature of the fuel, causing faster evaporation. Some heat
will pass through the liquid and raise the temperature of the surface beneath the pool.
Excess heat from the surface is quickly transferred to the liquid, keeping the maximum
temperature of the surface at or just above the boiling point of the liquid. As we have
seen, the boiling point of gasoline (a mixture) is from 40°C to 150°C (100°F to 300°F);
therefore, the temperature of the concrete beneath a burning pool of gasoline can never
be much higher than 150°C (300°F), which is not high enough to trigger the thermal
expansion usually needed for spalling. The duration of a flammable liquid fire on con-
crete is also very short (typically 1 to 2 minutes). That is not enough time for a significant
amount of heat to penetrate concrete with its limited thermal conductivity.
As a fire indicator, spalling can indicate the presence of a significant fuel load of ordi-
nary combustibles, as well as sources of intense localized heating such as a chemical
incendiary.33 Experiments by one of the authors and James Novak have shown that con-
crete spalling is very unlikely (although it can take place) when a flammable liquid alone
is ignited on the surface. It must be remembered that a flammable liquid burns only at its
top surface and along its edges while protecting the area immediately beneath the pool
from high temperatures.34 As a result, spalling (if it occurs) is often more pronounced
along the outside edges of the burning pool. Freestanding pools rarely burn long enough
to induce much radiant heat into the surface, with minimal expansion effects. Once the
main pool of liquid is consumed, flames may continue at cracks, holes, seams, and expan-
sion joints where pockets of fuel have collected. As a result, spalling tends to be more pro-
nounced in the vicinity of such irregularities. If a heavier petroleum distillate fuel such as
kerosene is burning, its higher boiling-point range will permit higher temperatures
beneath the pool, and more persistent, more radiant flames will result. Such flames may
have more pronounced effects on the concrete, but such fuels are also unlikely to be used
as accelerants. Edges of spall patterns may be stained or discolored from the tarry residues
284 Chapter 7 Structure Fires and Their Investigation
FIGURE 7-48A Spalling test on interior concrete floor. Pool fire FIGURE 7-48B After self-extinguishment, some staining around
of white gas. Courtesy of Jamie Novak, Novak Investigations and St. Paul margins, but no spalling. Courtesy of Jamie Novak, Novak Investigations
Fire Dept. and St. Paul Fire Dept.

of such combustion, but such discoloration is not proof of the involvement of flammable
liquids, since such staining has been observed in tests where only normal combustibles
were burned. In tests by one of the authors, localized spalling was much more likely if the
floor was covered with carpet or tile. In such instances, the localized burning induced by
a gasoline spill was more likely to involve prolonged burning of the combustible floor
covering. This combustion would involve no protective pool of liquid and was more likely
to induce spalling than a pour of liquid alone. Note that the depth of spalling has no rela-
tionship to the quantity or type of liquid fuel involved.
Spalling is more likely to be produced over large, unprotected areas of concrete floors
by radiant heat developed by normal fire progression, especially where there is extensive
collapse or falldown. See Figures 7-48 and 7-49 for the results of an experiment contrast-
ing the effects of a gasoline fire and a wood bonfire on the same type of concrete (note
the staining around spalled areas on the wood-fire concrete). The symmetry and location
of spalled areas must be carefully considered. Even localized spalling, when accompanied
by staining, may not be the result of a flammable liquid. Figure 7-50 shows several pro-
nounced spalled areas that were produced by burning asphalt falling from a collapsing

FIGURE 7-49A Spalling test on interior concrete floor. Large FIGURE 7-49B Prolonged burning promoted explosive spalling
wood-fueled bonfire. No ignitable liquids. Courtesy of Jamie Novak, during fire. Courtesy of Jamie Novak, Novak Investigations and St. Paul
Novak Investigations and St. Paul Fire Dept. Fire Dept.

Chapter 7 Structure Fires and Their Investigation 285

 FIGURE 7-49C Post-fire
appearance. Note exten-
sive spalling (especially
nearest ventilation
source—door to left).
Note “pool-shaped”
discoloration on concrete.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

roof onto a concrete floor. This is exactly the mechanism most likely to induce spalling—
sudden contact with a very hot, molten, or burning material.
Some investigators have reported that an ultraviolet light (or various wavelengths of
visible light) from a forensic light source played on a concrete surface aids in detecting
ignitable liquid residues as a result of their fluorescence. An early reference addressed this
technique.35 While this suggestion has not been investigated in depth by the authors, the
presence of fluorescent compounds in asphalts, greases, and other petroleum products
indicates that they would be present in the pyrolysis products of many complex synthet-
ics. Such compounds would be expected to fluoresce in the same manner as the petroleum
distillates and thereby interfere with their detection by producing false positives. Preliminary
field tests observed by the authors have revealed few useful patterns on real-world surfaces
that would indicate the presence of ignitable liquids. A new generation of pulsed-laser
fluorescence detectors holds promise for distinguishing pyrolysis products from ignitable
liquid residues, but they await validation.36

Ghost Marks
Localized spalling of concrete surfaces often takes place in conjunction with another diagnos-
ghost marks ■ Stained
tic sign of flammable liquid fires on floors covered with vinyl or asphalt tiles—called ghost
outlines of floor tiles
produced by the disso- marks. When asphalt tile squares are affixed to concrete surfaces, they are glued down with
lution and combustion a more-or-less continuous layer of tarry adhesive. When a petroleum-based flammable liquid
of tile adhesive. is poured on such a floor and set afire, the heat and solvent action tends to curl the tiles up

FIGURE 7-50 Spalling

of concrete floor caused
by the combustion of
extremely hot asphalt
from a collapsing roof.
Note the piles of tar and
gravel near each spalled
area and the staining of
adjacent concrete surfaces.
Courtesy of John D. DeHaan.

286 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-51 Ghost
marks on concrete indi-
cate pour pattern through
a doorway. The damage
of the threshold of the
door frame is in direct
line with the apparent
pour. Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

at the edges. Liquid on the surface then seeps between the tiles and, in some cases, partially
dissolves the adhesive. As the fire progresses toward full room involvement, the combination
of flammable liquid and adhesive provides fuel for the flames, which tend to be more intense
at the seams than in surrounding areas. As a result, more of the adhesive is destroyed, and
the underlying concrete floor may be locally discolored and is more likely to be spalled. The
adhesive in these areas will be burned off the surface leaving subsurface staining in the concrete
along a grid pattern corresponding to the seams between the tiles, as depicted in Figure 7-51.
Depending on the fire conditions and the nature of the floor tile, the authors have observed
in experimental room fires that tiles will shrink even in the absence of flammable liquids,
exposing the floor at the joints to higher heat fluxes and producing similar general effects. In
these cases, the residues of the adhesive tend to remain on the surface of the concrete. Very
faint subsurface staining has been observed on concrete that had a tile floor over it but no
fire damage (as in Figure 7-52). This staining was thought to be a result of the solvent

FIGURE 7-52 Faint stain

lines in concrete removed
from a concrete floor that
was undamaged by fire.
The stains align with the
tile margins, apparently
resulting from wax-
stripping solvents used
on the floor. Courtesy of
John D. DeHaan

Chapter 7 Structure Fires and Their Investigation 287

 action of the floor-wax-stripping solvent used for many years in the room as it soaked
between the tiles. Once again, the observation of such a feature across a wide, general
expanse of floor rather than in small, isolated or discrete areas should indicate by hypothe-
sis testing that it was the result of some general feature of the room and not the result of a
release of massive quantities of a flammable liquid. In the authors’ experience, large quanti-
ties (more than a gallon or two) are used only in a very small percentage of cases, and they
usually result in physical destruction of the building by deflagrating explosion.
Ghost marks, then, should not be considered to be proof of the use of a flammable
liquid, but taken in proper context may be an indicator of such an accelerant. Samples of
flooring and adjacent framing should always be taken for lab analysis.

Calcination of Gypsum Board

As we have seen previously, gypsum wallboard consists of a layer of calcium sulfate (plas-
ter) between layers of heavy paper. The calcium sulfate is normally in the dihydrate form
(two molecules of water per molecule). When exposed to heat, this water of crystallization
calcination ■ Loss of is lost through calcination in two major steps at predictable temperatures (52°C–80°C;
water of crystallization 125°F–175°F and 130°C–190°C; 266°F–375°F). In wallboard each dehydration is
caused by heating. detectable as a loss of mechanical strength (penetration resistance) when a probe is pushed
into the surface. In addition, as the paper and paint burn off the exposed surface, the com-
bustion products appear to condense in the outermost portion of the gypsum. As heating
continues, this layer migrates through the matrix, resulting in a visible color change from
gray to white, as seen in Figure 7-53. As the gypsum is heated and dehydrated it shrinks,
cracks, and loses its mechanical strength until it is completely dehydrated to a fragile
powdery substance. In structure fire experiments conducted by the authors, the detectable
changes (mechanical and visible) were proportional to the intensity and duration of the
exposure to fire.
Mann and Putaansuu recently reported on a series of tests using a cone calorimeter
to reproducibly heat samples of gypsum wallboard, after which the amount of dehydra-
tion was measured using FTIR, and penetration depth with a modified tire tread-depth
gauge.37 These tests indicated that probe penetration with 0.87 kg/mm2 (6 lb) of force fol-
lowed the depth of the interface between the fully hydrated layer and the hemihydrate

FIGURE 7-53
Calcination of gypsum
drywall advancing from
fire exposure (fire on right
side). Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

288 Chapter 7 Structure Fires and Their Investigation

form, as the anhydrous and hemihydrated forms offered less resistance than did the fully
hydrated form. Color changes, however, did not. It was suggested that the deposition and
migration of colored combustion products have numerous variables. Mann and Pataansuu
reported that similar testing of fiberglass-reinforced (X-rated) gypsum board revealed
similar linear results between penetration depth, exposure time, and depth of dehydration.
The presence of the fibers or vermiculite reduces the shrinkage of the gypsum and the for-
mation of cracks that weaken the material. If applied systematically with some care to
ensure reproducible force applications (0.87–1.03 kg/mm2; 6–10 lb), this technique does
yield reliable data on the duration of typical fire exposures for these products. Mann and
Pataansuu reported that the visible “smoke layer” position correlates well to the depth of
the anhydrous/hemihydrate interface due to changes in porosity caused by dehydration
(but not the actual depth of the hydrated layer).
There may be two or three discrete changes in color (white to gray to white), possibly
as a result of the two steps of the dehydration or the chemistry of the condensed pyrolosis
products. The colors of interest are transitions between gray and white. Other colors such
as pink, blue, or green have no causal link to the fire, since they have been linked to impuri-
ties in the gypsum. Data on this color phenomenon are still fairly limited.38 Because gypsum
is usually a uniform manufactured product with a predictable and uniform thermal con-
ductivity, the depth of thermal effects should be directly related to the duration of exposure
with the caveat that extreme radiant heat fluxes will cause faster buildup of surface effects.
Air pockets or voids in gypsum may also give rise to erroneous estimates. A thorough survey
of a room can allow the mapping of areas with the highest “penetrations” corresponding
to the locations of the most severe heating (from fuel packages, plumes, or ventilation
effects). Kennedy and others have examined the processes of heat effects on drywall and
have also demonstrated the reliability of calcination observations.39
Annealed Furniture Springs
As a steel wire spring is heated it will reach a temperature, well below its melting point,
at which it loses its springiness, or its temper. This temperature is called its annealing annealing ■ Loss of
temperature. At temperatures of about 600°C to 650°C (1,100°D to 1,200°F) the spring temper in metal caused
by heating.
will collapse if there is any weight on it and may even collapse of its own weight. This
temperature-related indicator may be of assistance to the investigator in establishing
whether a fire developed within an upholstered sofa or bed or enveloped it from without.
A smoldering fire developing slowly within the upholstery of a sofa or bed often produces
a localized area of intense heat, as pictured in Figure 7-54. The springs in this localized

FIGURE 7-54 Intense

fire inside this mattress
reveals high temperatures
(greater than 1,000°C;
1,800°F) with open
flames just beginning to
show (ignition by ciga-
rette). Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

Chapter 7 Structure Fires and Their Investigation 289

 area will often be more collapsed (or extended if they are in suspension) than those
springs around the margins. When furniture is enveloped by an external fire, heating may
be less intense, and the general degree of spring collapse is often less marked than from a
long-burning internal fire. In a prolonged post-flashover fire, however, collapse will often
be total no matter what the origin of the fire.
One useful indicator of spring collapse is directional, in that the collapse of springs
may be more pronounced in the corner or side facing an oncoming fire, if the resulting
destruction is not complete and ventilation is fairly uniform. Temperatures adequate to
cause collapse of springs by loss of temper can be produced by both flammable liquids
and by ordinary combustibles, so their collapse cannot be reliably used to demonstrate
such an incendiary origin.40 It has also been shown that such springs often taper away,
appearing to have been melted by fire exposure. This tapering is due instead to repeated
exposure to turbulent flames and not to intense heat from an incendiary mixture. Such
appearances are more often noted in structures gutted by fire.41 Objects lying on the
springs or debris falling onto them during the fire will cause more collapse, and any inter-
pretation of indicators from spring collapse must keep such limitations in mind. Like
other fire-direction indicators, spring collapse should be noted and photographed as part
of the general scene search.
Glass
Glass is a structural material found in virtually all structures. It has some properties that
make it a very valuable indicator to the investigator. First, although it looks and generally
acts like a solid, it is in fact a supercooled liquid. As a result, when exposed to stress, glass
is elastic and will actually bend to a substantial degree before breaking. When a large object
strikes a normal (i.e., nonreinforced, nontempered) glass window, the pane bows out until
its elastic limit is reached, whereupon it shatters. This reaction results in two sets of frac-
ture lines, radial and concentric, as shown in Figure 7-55. The edges of mechanically
broken glass will almost always bear a series of curved, conchoidal fracture lines, which
are produced by the release of internal stress within the glass when it breaks. These con-
choidal fracture lines are roughly perpendicular to one side of the glass and curve around
to be roughly parallel to the other, as shown in Figure 7-56. These lines can reveal the side
of the pane from which the breaking force came. Note that it is vital to know whether
one is dealing with a radial fracture or a concentric fracture. For this reason, it is best that
as much of the broken window as is practical be recovered and submitted to the labora-
tory for reconstruction.

FIGURE 7-55 Typical

mechanical impact fracture
allowing the development
of radial and concentric
fracture pattern.

290 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-56
Conchoidal fracture lines
on the edges of glass
fragments can reveal the
direction of the impact.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

Glass can also break as a result of stress produced by nonuniform heating. The result-
ing fractures tend to curve randomly about the pane (as opposed to mechanically caused
fractures, which are almost always straight lines), and the fracture edges bear little or no
conchoidal lines, usually being almost mirror smooth, as in Figure 7-57. Glass expands
when heated, so convected or radiant heat raises the temperature of the exposed portion of
the glass pane, and it expands. The portion of the pane behind (or in) the frame or molding
is protected or shadowed and “sees” no heat, and because of its poor thermal conductivity,
maintains its starting temperature and size. Tensile stress builds up until the tensile strength
of the glass is exceeded, and the pane cracks, usually roughly parallel to the framed edges

FIGURE 7-57 Edges of

thermally fractured glass
reveal virtually no con-
choidal fracture lines.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

Chapter 7 Structure Fires and Their Investigation 291

 FIGURE 7-58 Radiant or
convected heat on the
central portion of the Edges of pane
glass pane causes thermal behind frame are
shielded from heat
expansion of heated por-
tion. The portion behind
the molding or frame is
in shadow and “sees”
no heat and does not
expand. When tensile
stress becomes sufficient,
the glass breaks, typically Convected or
in fractures roughly paral- radiant heat
lel to the frame.

(see Figure 7-58). This process can be predicted by mathematical modeling in terms of tem-
perature differences.42 Studies of this process have shown that it is dependent on the strength
and type of glass; ordinary window glass breaks when there is a temperature difference of
about 70°C (126°F).43 This breakage can occur in any glass object that is subjected to
uneven heating and may reflect the direction from which heat approached. Thick glass can
shatter due to a sufficient temperature difference between the side facing the fire and the one
facing away from it. Double-glazed windows fail in sequence; the pane facing the fire fails
first, then shields the other pane until it falls out of the frame. If the glass simply cracks and
does not fall away, the outer pane may survive even through flashover.44
Not only can one determine whether the glass was broken by mechanical or thermal
stress, from the presence or absence of soot, char, or ash deposits on the broken edges, one
can tell if the breakage occurred before or after the fire (Figure 7-59). The location of the
glass in the fire debris is of utmost importance in interpreting its evidential information.
Glass with relatively little soot found on the carpet of a room, for instance, with all fire
debris on top of it would be a strong indication that the window was broken before or very
early on in the fire. Glass found lying on top of the debris most probably resulted from
breakage during suppression or overhaul. The appearance of deposits on the inside surface
of the window glass is of less evidential value than was previously thought. Because of the
common use of synthetic materials in upholstery and in wall and floor coverings, the pres-
ence of thick, oily soot can no longer be used as an indicator of the use of a petroleum dis-
tillate as an accelerant. Unfortunately, the color, density, and tenacity of smoke deposits are
more closely related to the degree of ventilation of the fire than to the type of fuels involved.

FIGURE 7-59 Broken

edges of glass fragments
reveal deposits of soot
and ash consistent with
having been broken
before the onset of the
fire. Courtesy of the State
of California, Department of
Justice, Bureau of Forensic
Services.

292 Chapter 7 Structure Fires and Their Investigation

FIGURE 7-60A A glass window fractured and softened by heat FIGURE 7-60B Close-up of crazed glass. Courtesy of Jamie Novak,
exposure during an ordinary combustible fire (post-flashover). Novak Investigations and St. Paul Fire Dept.
Crazing was caused by suppression water spray. Courtesy of Jamie
Novak, Novak Investigations and St. Paul Fire Dept.

Glass breakage often occurs just as the room approaches flashover. If enough cracks
develop that the glass can actually fall out, fresh air can be admitted suddenly, possibly
triggering a backdraft coincident with flashover. Sudden heating (as from a flammable
liquid accelerant) may induce more cracking than a slow fire but not so much more that
it will be detectable by post-fire examination of the fragments. If such heating occurs very
quickly, the glass pane usually simply shatters.45 Crazed glass, in which the pattern of partial-
thickness fractures or cracks (Figures 7-60a and b), resembling a complex road map in
the glass, was once thought to be indicative of a very rapid buildup of heat sometime during
the fire. Laboratory tests have shown that crazing, like cratering (pitting of the surface in crazing ■ Stress cracks
small circular defects), is caused by a spray of water hitting a window already heated by in glass as the result of
fire exposure, usually during suppression.46 Narrow slivers or long shards of glass indi- rapid cooling.
cate that an explosion of some type took place, as pictured in Figure 7-61. The presence
of heavy soot or varnish on splinters of glass indicates that the explosion took place late
in the fire, most probably as a result of a smoke or backdraft explosion. (Glass can also
be broken and projected some distance from the window if struck by a straight-stream

FIGURE 7-61 Glass

shards caused by defla-
gration of flammable
liquid vapors. Parallel,
partial-thickness fractures
are typical. Courtesy of John
D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 293

 hose.) Absence of soot, or a very faint coating, indicates an explosion early in the fire,
possibly caused by an explosive device or fuel/vapor deflagration.
Thus, the appearance of the glass can tell the careful investigator a great deal about
the nature of the fire or explosion; it can tell even more if its location in the debris is care-
fully noted and photographed before it is disturbed. Broken glass scattered by an explo-
sion should be charted (with dimensions) and photographed during the exterior examination,
and this documentation can be supplemented with observations of the glass remaining in
and around the window frame. Tempered glass, found in vehicles and many structure
doors, is made by treating the glass with chemicals or heat during manufacture, which
results in glass sheet that is much more resistant to mechanical breakage. When it does
break, it shatters into thousands of small cubic blocks. In general, it is not possible to
reconstruct such a window or to determine whether it was broken by mechanical or
thermal shock. One exception is in vehicles or metal-framed glass doors where there is
often a rigid frame that holds the peripheral fragments in their original configuration even
after the collapse of the majority of the glass. The fracture pattern remaining can some-
times allow discrimination between mechanical and thermal fractures.
Melting glass, whether it is part of a window, furnishings, or electrical lamps, can be
useful in indicating the direction of fire travel. Heat causes glass to expand on the surface
being heated, which often causes a window to bend or “belly-in” toward the source of
heat. When it finally melts at a temperature of approximately 750°C (1,400°F), glass will
soften on the side facing the heat first, and so it will sag or flow in that direction as it loses
structural strength. Note that windows in aluminum frames may react to heat-induced
distortion of the frame rather than the glass. Observed tests show that normal single-
glazed windows more often fall inward toward the heat but can fall outward as well.
An electric lightbulb is first evacuated and then filled with an inert gas, often nitro-
gen (except for very low wattage lamps, which are evacuated). When a lightbulb is heated,
the expanding gas inside creates overpressure, which balloons out the first area of the
bulb to soften in the advancing flame. This “blowout” points back in the direction from
which the fire came, since the side facing the oncoming flame will usually be the first to
soften. The bulb in Figure 7-62 is a good example of this effect. Other glass objects in the

FIGURE 7-62 The

“blowout” on this light-
bulb faces the oncoming
flames and indicates the
left to right direction of
fire travel. Courtesy of
Detective Mike Dalton,
Knox County Sheriff’s Office,
Knoxville, TN.

294 Chapter 7 Structure Fires and Their Investigation

structure—containers, mirrors, and the like—will indicate to the observant investigator
not only what temperatures were developed but also from which direction the flames
came—but only if they remain undisturbed until they are photographed and examined.
Glass bottles, for instance, may be much more intricately fractured on the side facing the
heat source even if they appear to be intact. They can shatter, however, if any attempt is
made to move them.

Melting Points of Materials

Plastics are often the materials first affected in a fire. As we saw in Chapter 5, many plas-
tics begin to soften and melt at low temperatures of 100°C–200°C (212°F–400°F), and
others decompose at similar temperatures. The melting behavior of glass was mentioned in
the previous section as being diagnostic of the conditions created within a structure. Glass
can vary in its melting point over a wide range, depending on its composition. Ordinary
soda-lime glass used in most windows has a melting point of about 700°C to 800°C
(1,300°F to 1,400°F), a temperature that is reached in many wood-fueled structure fires.
As additional indicators of the intensity of a fire, metals are often very useful. When
a fire scene is examined, previously molten aluminum residues are often found. Because
this metal melts at a comparatively low temperature, 660°C (1,220°F), this is not unex-
pected or indicative of a particularly unusual fire when found. Its high thermal conduc-
tivity means that localized flames are unlikely to melt thick metal components. This is
especially true of flammable liquid fires, whose typical durations are so short that even
low-melting-point metals are likely to be affected (see Figure 7-40). Zinc/aluminum die-
cast fittings (sometimes called “pot metal”) have a very low melting point (380°C; 718°F)
and are very common in fire scenes.47 A more useful indicator is melting brass or copper.
Depending on its composition, brass melts over a range of 875°C to 980°C (1,625°F to
1,780°F). Its presence in a burned structure is not unusual, but it may assist in establish-
ing which parts of a room reached higher temperatures than other parts. Copper, com-
monly present as water piping and electrical wiring, will occasionally melt in the hottest
structure fires, its melting point being 1,080°C (1,981°F). In their extensive study of
buildings totally burned during an urban firestorm, Lentini, Smith, and Henderson dis-
covered melted brass and copper fittings in many of the ruins. These included copper
water pipe (although water service to the area was interrupted during the fire, water
should still have been present in these lines). This demonstrated that such melted materi-
als in individual fire-gutted scenes (so-called black holes) cannot be linked to the use of
an accelerant.48 (See Appendix C for melting points of some common materials).
As we saw in Chapter 3, the flame temperatures produced by flammable liquids burn-
ing as a pool in air are not necessarily higher than those produced by ordinary com-
bustible fuels, so the presence of melted metals should not be taken as proof of the
presence of liquid accelerants. It is of critical importance to the fire investigator to note
that molten aluminum can interact with copper and dissolve in it to form a bimetallic
alloy. This alloy can have any percentage of copper and aluminum, but at a ratio of
67 percent aluminum and 33 percent copper, it forms a eutectic metal alloy with a melting
point of only 548°C (1,013°F), below that of either original metal. Molten aluminum can
therefore attack and dissolve away heated copper conductors in electrical systems and
appliances. The reader is referred to the article by Beland, Ray, and Tremblay for a complete
discussion of this phenomenon.49
Zinc falling onto heated steel can cause penetrations by a similar alloying process.
Cast iron, which melts in the range of 1,200°C to 1,450°C (2,200°F to 2,700°F), is some-
times found melted in the ashes and rubble of heavily burned heavy-timber structures.
Because well-ventilated wood and charcoal fires can produce temperatures on the order
of 1,300°C (2,400°F), such melting may occur by strictly thermal processes.50 It has also
been suggested that carbon monoxide combustion and actual reductive smelting of the
cast iron may play a part. It has recently been suggested that black iron can alloy with
Chapter 7 Structure Fires and Their Investigation 295
 calcium salts to form a eutectic with a relatively low melting point. The presence of plas-
ter or gypsum residues in contact with melted iron may be significant.51
Whenever any residues of a molten metal are found at a fire scene, they can help
establish the minimum temperature occurring at that spot in the structure. While such
melted metals cannot and should not be used as proof that the fire was incendiary, the fire
investigator should note their presence, extent, and distribution. Such information can be
of help in establishing differences between temperatures produced by normally fueled and
ventilated accidental fires and those produced by enhanced draft conditions or unusual
fuel loads from chemical accelerants in incendiary fires.
Copper and, to a lesser extent, copper alloys are very susceptible to oxidation when
heated in air. This reaction results in the formation of black copper oxide on the surface
of the metal. If the heating and exposure to air are continued long enough, ultimately the
entire object, for instance, wire, may be changed to oxide. The degree of oxidation of
exposed copper wiring can serve as an indicator of the conditions existing at the point of
exposure. Copper wire carried in conduit is not so subject to oxidation because of the lim-
ited access to air and the general reducing (rather than oxidizing) atmosphere resulting
from the decomposition of insulation materials. See Chapter 14 for a discussion of the
analysis of copper wire residues. The surface of copper wire and other conductors may
turn green as the hydrogen chloride gas (HCl) released by the pyrolysis of the PVC insu-
lation turns it to copper chloride.
Clean Burn
When a noncombustible surface is exposed to soot and pyrolysis products, they condense
on it. If that surface is heated sufficiently, those materials can be burned off, leaving a
“clean burn.” The surface temperatures needed to accomplish this are flame temperatures
(⬎500°C; 1,000°F). Therefore, a clean burn indicates direct flame contact of some duration.
The flames may be from a fuel package playing on the surface or from incursion of fresh
air into an underventilated post-flashover room, as shown in Figure 7-63.

MYTHS AND MISCONCEPTIONS ABOUT INDICATORS

A number of indicators have been traditionally linked to incendiary fires that have been
shown to be unreliable. Most of them have been dealt with in this chapter or previous
ones but are so pervasive that they deserve to be specifically numerated here.
Evidence (Documentary or Witnesses) of an Abnormally Fast Rate
of Fire Spread or Collapse
Fact: Fires are rarely seen from the moment of their ignition, and so their actual rate of
fire development is rarely known. The furnishings in use today will also support more
energetic, much faster growing fires than most people appreciate. The pre-fire fuel loads
and their placement are critical to evaluating fire spread. Lightweight construction tech-
niques often used today in commercial and residential structures can lead to much more
rapid penetration and collapse of floors, ceilings, and roofs than occur with traditional
techniques.52
Evidence of Abnormally High Temperatures (Melted Metals, etc.)
Fact: Flame temperatures of flammable liquid fires are about the same as temperatures of
wood-fueled fires. Higher temperatures (in excess of 1,000°C; 2,000°F) can be produced
by the combustion of many synthetic materials, and post-flashover fires can create tem-
peratures of that magnitude from floor to ceiling in modern furnished rooms in the absence
of any flammable liquids.
Spalling of Concrete
Fact: Radiant heat from sustained fires in ordinary fuels and structural collapse are much
more likely to transfer enough heat into concrete to cause it to spall than is the heat from
short-lived fires of pools of flammable liquids. Temperatures and heat fluxes under pools
296 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-63 The
“clean burn” on the back
wall of this post-flashover
test cubicle resulted from
the entrance of air after
the glass viewport shat-
tered. Courtesy of John D.
DeHaan.

of flammable liquids are too low to induce sufficient heating to cause most concrete to
spall. Flammable liquids on floor coverings can induce sufficiently energetic burning of
those coverings to induce spalling.
Crazing of Glass
Fact: Thermal stress can cause glass to fracture, but rapid buildup of such thermal stress
(thermal shock) causes rapid failure, often producing shattering of the glass, rather than
the intricate partial-thickness patterns associated with crazing. Such patterns are produced
when glass heated by fire exposure is suddenly cooled (usually by hose stream during
suppression).
Irregular Damage to Floors and Floor Coverings
Fact: Heat fluxes on floors in post-flashover room fires are not only very high, they are
not constant, driven by extremely turbulent mixing between the combustion gases and air
coming into the room through doors and windows. As a result, the post-fire damage is
often more intense around these ventilation openings. The reaction of synthetic carpets
and pads to such radiant heat causes irregular melting and combustion. Falldown or col-
lapse of overhead structural materials can cause irregular, intense floor-level damage even
in the absence of flashover conditions.
Black, Heavy, Oily Soot on Windows/Black Dense Smoke
Fact: Black dense smoke is a result of the type of fuel burning in a room and the efficiency
of the combustion. Modern synthetic materials are almost always petroleum-based and
Chapter 7 Structure Fires and Their Investigation 297
 when burned in a typical fire, release large quantities of black soot and dark pyrolysis
products.
Annealing of Steel Springs and Steel Structural Materials
Fact: Steel loses its temper and tensile strength at temperatures above 500°C (1,000°F).
Such temperatures can be produced by smoldering or flaming combustion of ordinary
upholstery materials. Ignitable liquid accelerants are not required.
Floor-to-Ceiling Heat Damage
Fact: As the temperature of the hot gas layer in a room exceeds 600°C (1,150°F), the radi-
ant heat flux exceeds that needed to char and ignite all exposed fuel surfaces. Due to the
large surface area of the hot gas layer, shadowed or protected areas beneath furniture may
not be adequate to prevent ignition of carpets under furniture.
Deep Char
Fact: The rate at which wood chars is dependent on many factors (type of wood, nature
of finish, etc.) but especially on the intensity of the radiant heat (heat flux) falling on it.
Direct flame exposure can cause wood to char at twice the rate at which it chars at igni-
tion, and wood in a post-flashover fire can char at 10 times the rate at which it chars at
ignition.
Progression to Flashover
Fact: Today’s furniture with its synthetic fabrics and padding is capable of producing very
high heat release rates. (A single modern recliner fully involved produces as much heat per
second as a half gallon of gasoline poured on the floor and ignited.) Although modern fur-
nishings are more resistant to smoldering heat sources, they are more susceptible to open-
flame ignition and, once alight, can cause the spread of a fire in a 3-m ⫻ 4-m (10-ft ⫻ 14-ft)
room to flashover in 5 to 8 minutes in the absence of any ignitable liquids.
Alligatoring (Shiny or White)
Fact: The surface appearance of charred wood is dependent on the nature of the wood
and ventilation conditions during its combustion. The presence of white ash on charred
surfaces is the result of prolonged burning in a well-ventilated fire, resulting in complete
combustion of carbonaceous residues, leaving behind the light-colored oxides of inor-
ganic compounds.

ARSON EVIDENCE
There are many other observations that will be made in the course of the scene examina-
tion that may reveal the fire’s origin and cause. Incendiary fire cause may be revealed by
obvious indicators such as separate multiple points of origin, the presence of trailers of
flammable liquid or paper or rags connecting areas of intense burning, or the remains of
timing or ignition devices.
Trailers
What appear to be trailers may be artifacts of normal fuel combustion, as illustrated in
Figures 7-64 and 7-65. Arsonists often use multiple “sets” or ignition sources to ensure
maximum destruction. Very often, unless the devices are carefully synchronized, the first
devices ignite a fire that interferes with the operation of later ones. They may be found
incompletely burned and readily identified. What appear to be multiple points of origin
must be carefully evaluated so as to eliminate their cause by falldown of draperies or
lighting fixtures, by lightning or power surges (causing multiple failures of appliances), or
even by radiant heat or hot fire gases traveling across a room or between rooms.
Containers
Containers are commonly found at fire scenes. While most of them will be normal con-
tents of the house, such as paint cans and household supplies, they may have contained
flammables that could be associated with arson and, if not, may still provide useful
298 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-64 Irregular
char pattern on a wood
floor caused by post-
flashover fire, set by
ignitable liquid on furni-
ture. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

information. If the can contained a liquid, from a small amount up to a full can, and it
was tightly sealed and heated, exposure to heat would cause it to bulge and split around
one or more seams. If the liquid was flammable, it would have boiled out of the can and
contributed substantially to the total fuel. If such a can is not bulged or burst, it is an
indication that it was either totally empty or that the cap was loose. Under some circum-
stances, lightly soldered cans may selectively heat from radiant fire above them, so that
the solder softens and allows escape of the contents without showing a bulge. When
arson is suspected, it is essential that all such containers be collected and submitted for
laboratory examination, with care being taken to preserve possible latent fingerprints.
Contents Inventory
All portions of the structure should be carefully inventoried and documented by the inves-
tigator, whether or not the contents were damaged by the fire. This process serves two

FIGURE 7-65 Intense

post-flashover fire pro-
duced irregular deep char
in this test fire with no
ignitable liquids. Courtesy
of John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 299

 purposes. First, it is for the protection of the investigators, who may be called on to prove
the appearance and contents of the scene at the time of their possession. This is especially
important when the contents are of high monetary value. Second, it ascertains the con-
tents of the structure at the time of the fire. A post-fire inventory should be carefully
checked against an inventory prepared by the owner or tenant. Removal of a building’s
contents before a fire is a strong indication that the fire was planned. The investigator
must be sensitive to the absence of major appliances, tools, equipment, stock, guns, cloth-
ing, jewelry, memorabilia, antiques, sports equipment, and pets. In cases of insurance
fraud, junk or cheap goods may be substituted for valuable goods or stock (which has
been sold or stored elsewhere). Many of the other signs of such manipulation are dis-
cussed in Chapter 16, on arson fires.
We have illustrated in a number of instances how the fuel load in a structure affects
the fire pattern. A thorough inventory of the burned structure provides the investigator
with a clear and complete picture of the fuel load in the building at the time of the fire.
Objects that could contribute significantly to the fuel load of each room—sofas, uphol-
stered chairs, drapes, carpets, and the like—should be marked on a sketch of the scene.
If the contents are removed during overhaul, the investigator may seek the assistance of
the owner or tenant in identifying what furniture was present and where it was located
before the fire.
Ignitable Liquids
Although it is possible to set large, very destructive structure fires without the use of
ignitable liquids, by far the most commonly detected arson means is the pouring or
spilling of an ignitable liquid (usually a petroleum product) onto the floor and contents
of a room followed by direct ignition with a match. Because of the frequent use of such
materials, the investigator who is able to locate identifiable residues of such liquids will
be more effective in detecting incendiary fires. Ignitable liquids are used as accelerants
because they are cheap, easily available, and are volatile and therefore are easily ignited.
Whenever such a liquid is exposed to the high temperatures of a fire, it is going to vapor-
ize quickly even if it does not burn. When the post-fire debris is being examined, only that
liquid that has been protected from the flames and high temperatures will remain. This
protection can be afforded by absorbent materials such as carpet, upholstery, untreated
wood, plaster, or soil or by cracks or crevices in floors or between floors and baseboards,
as shown in Figure 7-66. The water used to extinguish most fires will then tend to pro-
tect these residues from evaporation for some time. Whenever localized burn patterns are
discovered on walls, floors, or floor coverings, nearby debris should be checked for the

FIGURE 7-66 Volatile

liquids are protected from
evaporation by joints in
floors, baseboards, and
the like. A close-up of
burn pattern from
gasoline–diesel mixture
shows the percolation of
unburned fuel oil from
pockets along the edges
and end grain of planks.
This effect was most
pronounced the day after
the fire. Courtesy of John
D. DeHaan.

300 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-67 Electric
hydrocarbon detector uses
a solid-state sensor that
vapors enter by passive
diffusion. Operator must
move the sensor slowly.
Courtesy of Cameron Novak,
Novak Investigations.

odors of common flammable liquids. The appearance of such burn patterns is often
enhanced by cleaning the floor with a broom and then flushing with water. The pattern
can then be photographed and sketched as needed. Seams, crevices, and other protected
areas can then be checked for distinctive smells by nose or portable hydrocarbon detec-
tor (Figure 7-67). Many investigators rely on olfactory detection of volatile fuels, and
some have very sensitive and accurate senses of smell. There are several drawbacks, how-
ever. First, the human nose is very subjective in its perception of odors and can be
“fooled” into false reactions. The human nose fatigues quickly, and inhalation of some
vapors can render the nose incapable of detecting other odors. The fire scene also con-
tains toxic gases and vapors that should not be inhaled in high concentrations.
If a floor is porous or leaky, some of the accelerant may have seeped through and pen-
etrated the soil underneath. Water and debris falling onto the soil then trap the volatiles
for later detection. In recent years, specially trained canines have proved their value in
locating areas of debris where ignitable liquid residues can be identified by laboratory
analysis. While not substitutes for a complete, professional investigation, trained canines
can help expedite the scene examination for likely areas to sample. Their role will be
examined in greater detail later in this chapter.
As we saw in Chapter 4, the flame temperatures produced by accelerants such as gaso-
line burning in air are not significantly higher than those produced by ordinary com-
bustibles (and in some cases are lower). Ignitable liquid accelerants are noteworthy because
of their potential for rapid combustion, localized damage, and unusual distribution. Liquid
accelerants can burn at low levels in a room and often in protected areas, leaving burn pat-
terns that are often distinguishable from those produced by ordinary fuels (barring over-
whelming subsequent fire damage such as post-flashover burning or collapse).
Remember that some ignitable liquids do not have a characteristic odor, and some do
not cause an electronic detector (or canine) to react. It is also possible for a sustained fire
to consume all liquid fuels used as accelerants. Other materials, such as crumpled news-
paper, waxed paper, plastic packing materials, and the like are used very easily to accel-
erate a fire and leave very little residue detectable by chemical or even visual examination.
In the final analysis it is always the experience of the investigator (backed by appropriate
Chapter 7 Structure Fires and Their Investigation 301
 published experimental data) that determines what importance is to be made of such pat-
terns. When indications are that an ignitable liquid has been used, the prudent investiga-
tor will recover samples of the flooring and nearby associated debris for laboratory testing
no matter what odors are present. Water-soluble flammable liquids such as methanol, ace-
tone, or ethanol are diluted and may be washed away by the application of hose streams
during suppression or debris removal. Empty containers of such solvents should warn of
their possible use as accelerants.
For all the reasons described in this text, the investigator should not rely on the pres-
ence of “unusual,” “pour,” or “pool” patterns of floor damage as proof that ignitable liq-
uid accelerants were used in the absence of positive laboratory findings. This is especially
true if there has been structural collapse or post-flashover burning. Time and exposure
are the enemies with which the investigator must always contend when seeking out
volatile residues. When the debris dries out, sun and high temperatures will evaporate
even protected residues in a matter of days. (Some residues have been detected in soil
under houses a few weeks after a fire, but the finding of liquids under the house bespeaks
of very large quantities used.) While cold, wet weather may aid in the retention of
ignitable liquid residues, it may adversely affect burn patterns and other evidence exposed
to the elements. The fire scene investigation must be carried out as soon as is practical to
provide the best chance of detection. It also minimizes the chances of contamination of
the scene (accidental or intentional) from spurious sources of ignitable liquids (such as
from equipment used for salvage).
If the burn pattern is on a carpet or rug, it is often useful to brush it vigorously with
a clean, stiff broom—after carefully visually examining it for the presence of trace evi-
dence (adhering glass fragments or plastic, paper, or fabric from an incendiary device)—
to remove loose debris and enhance the burn pattern for examination. The carpet and any
pad can then be sampled for lab testing. The resulting burn pattern may be enhanced by
sweeping away loose debris and then flushing with clean water. Residues may be found
between the tiles remaining, under baseboards, or soaked into the wood subfloor. In a
multiple-layer floor covering, residues of ignitable liquids (with evidentiary significance)
are not going to be found in the lowest layers if they are not found in the ones above.
No matter from where laboratory samples are to be taken, two things must be kept
in mind: First, volatile liquids are driven off by high temperatures; therefore, they are not
likely to be found in the areas of deepest char. Rather, they are more likely to be found in
places where the liquids would be protected from the heat. Second, many building and
upholstery products involved in fires today are synthetic materials that share the same
petroleum-based origin as most accelerants. To ensure that the lab will be able to distin-
guish minute traces of accelerants from the semivolatile decomposition products of syn-
thetics, comparison samples are very important. A quantity of undamaged material should
be collected from a nearby area that does not appear to have been contaminated with
accelerant. Even in badly damaged structures, comparison samples of floor coverings may
be found under appliances, bookcases, or large furniture. Because these scene samples can-
not be guaranteed to be free from accelerants, they cannot be considered to be control
blanks and should be considered and labeled as comparison samples. True control blanks
would have to come from a noncontaminated source such as a retail store, manufacturer,
or identical unburned vehicle.

Detectors—Electronic and Canine

“Sniffers” or, more properly, portable hydrocarbon detectors are used by many success-
ful investigators. Many varieties are available, with any degree of sophistication desired
(the advantages and disadvantages of all of them will not be treated here). Suffice it to say
that, no matter which hydrocarbon detector is used, it should not be used to the exclu-
sion of other tests examining appearance or odor. Even the best electronic detectors have
a sensitivity on the same order as that of the human nose and therefore offer no complete
302 Chapter 7 Structure Fires and Their Investigation
advantage in that respect. Like the human olfactory system, detectors can be over-
whelmed or fatigued by continued exposure to high levels of volatile fuels. They may not
be sensitive at all to some common accelerants (odorless charcoal lighter seems to be the
most elusive) and thus yield false negatives when there really is an ignitable liquid pres-
ent. Most of them are sensitive to an oxidizable gas or vapor, including the carbon
monoxide (CO), ammonia (NH3), methane (CH4), and methyl alcohol (CH3OH) vapors
produced by the destructive distillation of wood and give false positives. Many modern
sniffers rely on an electronic detection element that is not as sensitive to pyrolysis prod-
ucts or water vapor as were older designs. One successful approach is to use two matched
detectors, one of which is located in a burned “reference” area while the other is used to
examine a burned “target” area. The difference between the two signals is then more
likely to be significant and less likely to indicate simple volatile pyrolysis products.53
Accelerant-detection canines, typically trained to detect one drop of partially evapo-
rated gasoline or paint thinner, have a detection limit on the order of 0.1 part per million
(ppm) [0.1 microliter (μL) in a 1-L can] in samples with a clean background. The pres-
ence of burned debris (particularly carpet and pad) can reduce this sensitivity signifi-
cantly, but at scenes dogs have been able to discriminate traces of ignitable liquid from
fire debris more successfully than the human nose, and they carry out a search of a scene
much more quickly than a person.54 As with drug-detection dogs, their positive alerts can
be used as probable cause to search a person or take a sample from a fire scene, but their
positive indications should not be accepted as proof of the presence of accelerants. Dogs
cannot discriminate between gasoline used as an accelerant and volatile traces of carpet
adhesives or insecticide sprays. Only a proper laboratory analysis can specifically identify
the material, and only then can it be put into context with the scene findings, for an accel-
erant must be identified by its intended use, not by its mere detection.55 Because of the
variety of petroleum products that can be detected at trace levels in normal household
items, the investigator must be careful not to put too much value on detection at very low
concentrations unless proper comparison samples are collected.56
Figure 7-68 illustrates the areas of wood surrounding a char pattern that should be
chiseled or pried up and tested with a sniffer or the nose. Areas yielding any sort of response
can then be cut out and sealed in appropriate containers for laboratory testing. When all
the areas have been checked and sampled, the containers can be left in place, labeled with
numbered cards, and the entire area then photographed to record graphically the disposi-
tion of the samples recovered. Such a process is shown in Figure 7-69.

Liquid pattern

FIGURE 7-68 Drawing

Tongue-and-groove flooring seam showing the areas of
wood surrounding a char
pattern that should be
sampled for laboratory
testing (including the
central area). Courtesy of
Lamont “Monty” McGill
Fire/Explosion Investigator.

Chapter 7 Structure Fires and Their Investigation 303

FIGURE 7-69 After
collecting samples, it is a
good idea to photograph
the area sampled with
cans in place. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

General Considerations
PROTECTED AREAS
Areas that have not been damaged by the flames are sometimes as significant as areas that
are badly charred. In any fire, the greatest damage possible will be at points where active
flame or radiant heat existed for an appreciable time. Because flames burn upward, it fol-
lows that floor areas do not receive as much active flame as other fuels. They will char by
radiant heat in most instances, but it is often noted that areas of the floor are totally
undamaged though surrounded by damaged areas.57 Such areas tend to have a definite
form—circular, rectangular, and so on—and invariably show clearly the location of an
object sitting on the floor at the time of the fire (see Figure 7-70). The object may be com-
bustible and show the effect as well as though it were noncombustible. Cardboard car-
tons often serve to protect the floor under them, because the bottom of the carton does
not burn well due to poor ventilation, even when the remainder of the carton is totally
destroyed.
The importance of such protected areas lies in two factors: (1) locations of objects are
definitely determined, and (2) the general type of the object is defined. For example, a
bucket that originally contained liquid accelerant (or a metal wastebasket where a dis-
carded cigarette ignited the contents) might be placed on the floor in the fire area, and its
location would be proved. Delineation is generally so accurate that the size and shape of
the bottom of the object is readily determined and compared with any suspect item that
is located.
In fires that approached flashover, protected areas on the floor can reveal the loca-
tions of furniture or victims that were moved, or even combustibles like paper or card-
board that were consumed during the later stages of the fire, as in Figures 7-71 and 7-72.
These objects are frequently removed during the firefighting or in clearing the area after-
ward and thus will only occasionally be found still sitting in the original position. When
304 Chapter 7 Structure Fires and Their Investigation
FIGURE 7-70 Protected areas of the carpet show the locations of furniture during the fire. Courtesy of Jamie
Novak, Novak Investigations and St. Paul Fire Dept.

such undamaged areas are found, it is up to the investigator to determine what object pro-
tected the floor and where it is now. Such objects are most often removed during suppres-
sion or overhaul, but more than once have the remains of a container or delay device been
removed after the fire by the perpetrator. This is a primary reason why the scene must be
secured from all persons until the investigation is complete. Such protected areas may help

FIGURE 7-71A Protected areas of carpet show the locations of furniture during the fire. Courtesy of John
D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 305

 FIGURE 7-71B

establish the position of doors, a feature that is critical to estimating ventilation condi-
tions and access possibilities at the time of the fire (see Figure 7-73).
Another situation in which a protected area is important is in the claimed fire in a
metal wastebasket or trash can. If there is not much fuel in the can or the fire is pro-
longed, fire may heat the bottom of the can and scorch the floor beneath. In most cases,
the bottom of the container will protect the area under it as well as retain manufacturer

FIGURE 7-72A Charred “path” on carpet from door (to left) to broken windows (to right) after prelimi-
nary debris removal from a post-flashover fire. Courtesy of Det. Rich Edwards, Los Angeles Sheriff’s Office (retired).

306 Chapter 7 Structure Fires and Their Investigation

 FIGURE 7-72B Piles of
clothing and bedding pro-
tected irregular areas of
carpet. (Photo taken just
after extinguishment
shows piles of clothing
and gray firefighting
foam. Courtesy of Det. Rich
Edwards, Los Angeles Sheriff’s
Office (retired).

data. Its position may then be accurately determined if the floor is discolored or charred
around it. A fire starting inside a plastic wastebasket will tend to cause collapse uni-
formly around the perimeter, as pictured in Figures 7-74a and b. A fire involving such a
container from outside will usually cause asymmetric melting (greater on the side facing the
oncoming fire). By examining the arrangement adjacent to the trash container and the
extent of fire in this region, the investigator may be able to assign the source of the fire

FIGURE 7-73 Protected

area of carpet shows the
position of the door dur-
ing the radiant heat stage
of fire. Courtesy of Lamont
“Monty” McGill, Fire/Explosion
Investigator.

Chapter 7 Structure Fires and Their Investigation 307

FIGURE 7-74A Plastic trash can was melted uniformly around its FIGURE 7-74B Bottom of can still bears identification information.
perimeter when the fire was ignited inside. Courtesy of John D. DeHaan. Courtesy of John D. DeHaan.

to this origin or to rule it out. Without the knowledge of the exact location of the container,
this is likely to be impossible or very difficult. Often, a fan-shaped char pattern will
occur on a wall just over such a point, with its lower boundary at the height of such a
trash container. This circumstance is frequently the very best evidence of the trash con-
tainer as the origin of the blaze. When a plastic gasoline container is ignited in a fire, it
burns or melts from the top down. If there is liquid fuel left in it, the fuel will be trapped
between the ‘folds’ of the collapsing container as in Figure 4-5g. Some liquid may leak
out when the mass is overturned, or it may be detected by a visible or audible “sloshing.”
Protected areas on walls are also frequently observed for similar reasons. The location
of an item of furniture that is no longer present can often be determined accurately from
such an area, as shown in Figure 7-75. The opposite effect is also observed occasionally in

FIGURE 7-75 This fire

occurred on a covered
patio (assuring plenty of
ventilation) and burned
post-flashover. The out-
line of the completely
destroyed sofa (and likely
origin) is still visible on
the stucco wall. Courtesy
of Derrick Fire Investigations.

308 Chapter 7 Structure Fires and Their Investigation

FIGURE 7-76A Protected area on left wall shows there was a mattress in place at the time of the fire.
Damaged area on right wall and damaged area of mattress in foreground shows the mattress was overturned
during firefighting. Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.

instances where the furniture burned and collapsed; instead of protecting the wall behind it,
it enhanced the burn (as in Figures 7-76 a and b). Such protection effects will not necessar-
ily serve to locate the position of a specific piece of furniture, because the marginal dimen-
sions are lost, but they can, at times, contribute significant knowledge of the fire origin, when
considered along with the general fire pattern. Localized floor penetrations are often

FIGURE 7-76B Broken window frame caused localized damage to the wall behind it and allowed smoke to
vent during the fire. Courtesy of John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 309

 FIGURE 7-77 Deep localized charring and pene- FIGURE 7-78 Flashover fire briefly forced flames
tration of the wood floor was the result of a wood- out under door, producing charred door and floor
and-cardboard “bonfire” in the center of this room (repeated several times). Courtesy of John D. DeHaan.
with no ignitable liquids. Courtesy of John D. DeHaan.

caused by combustible furniture burning for prolonged periods, as seen in Figure 7-77. A
combination of the two effects is also sometimes seen. When a piece of furniture is afire, it
may both protect the wall behind it and, by intensifying the local fire, increase the amount
of wall damage on the sides and above the item. This is especially important when the fire
started in the furniture in question. Ventilation effects can also produce low burns, even in
the absence of localized fuel packages. Figure 7-78 shows the flames of a fully involved
room fire being pushed out under the door as the room briefly reached an overpressure con-
dition. This condition could be maintained for only a few seconds, but the flames “pulsed”
from beneath the door several times before the fire found its ventilation limit.

UTILITIES
The condition of the utilities servicing the building should always be inspected and doc-
umented. When utility companies respond to a scene to remove gas or electric meters (to
ensure the safety of scene investigators), they should be asked to evaluate the service for
signs of tampering, bypassing, damage, or malfunction. The investigator should deter-
mine when power, gas, telephone, cable TV, and the like were cut off and by whom.
Sometimes a planned fire will be betrayed by the owner’s cutting off costly services sev-
eral days before the structure was destroyed.

ELIMINATION OF ELECTRICAL IGNITION SOURCES

Remember that in accordance with the scientific method, all logically possible sources of
ignition from natural and accidental causes must be eliminated if there is to be proof of
the crime of arson. The use of an ignition matrix as a means of structuring and documenting
the elimination process was illustrated in Chapter 6. This will, of course, include the
310 Chapter 7 Structure Fires and Their Investigation
utilities, appliances, and electrical wiring (both fixed and power cords). To check out an
electrical system in a burned structure is often the most frustrating portion of the entire
examination. Wires, conduits, and remains of fixtures are often scattered throughout
much of the bulk of the rubble. Wires will generally lack insulation, this having been
burned off in the fire. Because wiring is most commonly of copper, and copper when
heated is readily oxidized to copper oxide, the blackened and brittle condition of the wires
will make them difficult to handle when attempting to trace them. The most significant
item to look for in such an instance is not just a mass of damaged wire but rather fused
copper in localized areas. Such fused masses are most likely to occur in the neighbor-
hood of switches, outlets, and similar connections, and at the fuse box. At any of these
places, it will be common to find wires with rounded, fused ends, demonstrating that
there was arcing, almost certainly by short-circuiting. When there are numerous such
fused ends, it is unlikely that a fire origin has been found but rather that there was fail-
ure of the breakers to shut off current as wires farther away from the breaker or fuse
box shorted together as the insulation failed in the fire. When there is a single heavy
fusion, but other portions of the wiring are generally intact, this single point becomes
highly suspect as a possible source of ignition. Normally, in such an instance, the fuse is
blown or the breaker is open, indicating that this short circuit occurred and triggered the
safety mechanism. Such protective devices respond to a direct short and usually interrupt
power before a fire can be started. High-resistance connections, on the other hand, can
create enough heat for long periods of time without drawing enough current to trip pro-
tective devices. These long-term heat sources are much more likely to ignite ordinary
fuels than momentary short-circuit arcs. Examination of the fuse box or circuit breaker
panel to determine which circuits have been tripped may be helpful in eliminating areas
and the circuits that serve them.
It should be noted also that wire within a metal conduit will rarely short except when
the conduit is heated sufficiently by an external fire to char the wire’s insulation. If a short
circuit is formed within a conduit from any cause, it must be a high-energy one to penetrate
to the exterior and start a fire, although such events are not unknown (involving unpro-
tected circuits, high-current applications [⬎40 A], failed overcurrent protection devices,
or three-phase wiring). To start a fire requires that the conduit itself be heated to the point
of fusion. In a very rare instance, it might be heated above the ignition temperature of
some material in contact with it, which could become ignited. The high heat conductiv-
ity of the metal conduit is actually a very good protection against starting of fires from
wires enclosed in the conduit. See Chapter 10 for more information.
One additional effect that has been noted in more than one instance involves the
zinc die cast used in conduit connectors. Zinc die cast has a relatively low melting point
and will melt in a fire surrounding the conduit. The heat of the fire may also remove
the insulation, so that molten zinc will contact copper. When this occurs, the zinc and
copper will form brass. Either the molten zinc or the brass formed may complete a short
circuit between wires that were not previously in contact. Such an effect is not a cause
of the fire but the result of an external fire, and it should not be interpreted otherwise.
The presence of brass within a conduit or spilled from a broken conduit is actually very
good evidence of an external fire that led to its formation. The diagnostic signs of elec-
trically related fires are discussed in great detail in Chapter 10 of this volume.

ARC MAPPING
The plotting of electric arcing in electrical wiring found at a fire scene has been found to be
of considerable value in testing hypotheses about possible areas of origin. A map of arc faults
can be overlaid on diagrams of other indicators. If the electric circuit involved can be traced
back to its power source, the location of the arc fault (which caused the overcurrent device
to trip or the conductors to part) farthest from the power source is very likely to be the area
Chapter 7 Structure Fires and Their Investigation 311
 closest to the origin. This assumes, of course, that the wiring was equally exposed to the fire.
Wiring concealed in walls will be affected much later than will exposed power cords. Arc-
fault mapping is not used to establish whether an electrical fault caused the fire but as an indi-
cator of area of origin. The technique is described in more detail in Chapter 10 of this text.

APPLIANCE CONDITION
The condition of an appliance that started a fire is not expected to be identical with one
that was in a fire external to it that it did not start, but the differences are sometimes dif-
ficult to spot. The external fire will damage any appliance by burning the enamel, discol-
oring and corroding the steel, and otherwise distorting or warping it. If the fire is external
to the appliance, whatever damage it suffers will show uniformity, or at least agreement,
with the fire patterns on walls, furniture, or other appliances in the vicinity. The appli-
ance is much less likely to be damaged internally if the cabinet or housing is noncom-
bustible. If the appliance started the fire, it will be damaged more internally than
externally and usually have some local damage or effect that is quite distinct from the gen-
eral damage pattern. This will be a point of consequence to the investigator. Because fires
are easily blamed on a faulty oven or a maladjusted furnace, incendiary fires are some-
times ignited in the vicinity of such appliances to throw the investigator off the track.
The key to uncovering this situation is again a study of the fire pattern. In most such
instances, the arsonist will use a flammable liquid that is poured around and under the
appliance and then ignited. In such instances, it is normal for the liquid to penetrate well
below the appliance in question and to give a burn far below that which is possible for the
appliance itself. Gas appliances and fireplaces will usually lead to fire above the base level
of the normal combustion, not to a region under the floor that supports it. Burns below this
level of support may be traceable to a failure or defect of the gas supply or firebox or to
improper installation (as seen in Figure 6-4c). In some instances, there may be penetration
of the floor under a fireplace or stove/heater due to the dropping of hot materials onto the
combustible floor through gaps in the firebox. Burning clear to the floor level is not normal
for a fire escaping from an appliance unless there is very obvious and severe damage to the
unit itself (see Figure 6-4a for an example). Naturally, the investigator will need to check
carefully that no such abnormality exists in the appliance. Misuse of any heat-producing
appliance can cause a fire, and any fire in the vicinity should be considered suspect and
investigated with extraordinary care. A portable X-ray system, such as those used by bomb
squads, can be used to examine appliances and connectors at the scene (see Chapter 15).
Such examinations may reveal useful information without destructive dismantling.
One cause of accidental fires that is often related to such appliances is the formation of
“self-heating charcoal” by prolonged exposure of a wood surface to heat, as discussed in
Chapter 6. The investigator should be aware of such processes wherever wood is exposed
to even moderate temperatures [over 105°C (220°F), or so)] in conditions where ventilation
is excluded or very limited. It must be remembered that the lower the temperature, the
longer it takes for a sufficient mass of such charring to form, and the less likely its forma-
tion will occur at all. Wood exposed to temperatures below 300°C (600°F) for such long
terms is characteristically dull- or flat-finished, with a sunken appearance. Char that has
penetrated completely through the wood structures involved and is much deeper than the
known circumstances of the fire itself would warrant should prompt further examination.
This phenomenon has been thought to turn even large wooden structural members completely
to charcoal over a period of months or years before finally igniting them (see Chapter 6).
Post-suppression smoldering can, of course, produce deep destruction if prolonged.

TRASH
Trash accumulations that are apparent as a portion of the fire scene will ordinarily be scru-
tinized very carefully, because they provide a suitable place for igniting a fire, especially when
312 Chapter 7 Structure Fires and Their Investigation
doused with a liquid accelerant. However, they also provide an inherent fire risk from acci-
dental causes. Both possibilities must be carefully weighed before considering that a trash
pile was used by an arsonist. If liquid flammables were poured on the trash, some of the
liquid may be recovered and identified, in which case, strong evidence of arson is available.
When a trash pile appears to be an origin or possible origin of the fire, it is highly impor-
tant to search for ignitable liquids, as well as ignition sources. It is important to search for
means of ignition that could be accidental in nature. Common sense sometimes does not
dictate what accumulations of trash some people keep, even in their kitchens and bath-
rooms. Even the most elementary safety measures of protecting such trash from open
flames such as candles, fireplaces, or room heaters are not taken by some people. It must
not be taken for granted that “no one would have that much trash around” or “no one
would light a candle near all that junk,” because time and again, people act unsafely.

DETECTION SYSTEMS MAPPING

In a large compartment or a building protected by fire alarm or sprinkler systems, the
detectors or sprinkler heads that were triggered during the fire should be mapped out on
a plot of the building. This map can reveal the most energetic (if not the first) stages of
the fire. When the times of their responses are included, a sequence of the fire’s spread
may be reconstructed. Many alarm systems have a data chip that can be read after the fire
that reveals the time of functioning of connected systems and the sequence of various
alarm functions. Alarm companies should always be contacted for information about
alarm settings, functions, and history.
The sprinkler system and its connections should be examined and documented. Fire
investigators should be familiar with system components, such as the one in Figure 7-79,
and signs of tampering or inoperability.

FIGURE 7-79 Typical

sprinkler system compo-
nents—OS&Y valves
(with chains and locks)
backflow preventer, PIV
(post-indicator valve),
and water-flow alarm
connection. Courtesy of
John D. DeHaan.

Chapter 7 Structure Fires and Their Investigation 313

 INTERIOR FIRES FROM EXTERIOR SOURCES
Fires within structures need not invariably have their origin in the interior. Buildings are
frequently consumed as a result of a fire that is initially outside the structure. Such sources
of fire may be roughly classified as follows:
■ fire from an adjacent building spread by radiant heat or burning debris
■ grass or brush fires exterior to the building
■ trash fires that are out of control
■ arson by means of trash or accelerant ignition against the side of the structure,
under porches, around doors, and the like
When the fire is initially outside the building, it may involve exterior surfaces and never
penetrate to the interior. Such a fire pattern is readily evident and the cause simple to deter-
mine. However, special circumstances may exist. For example, an open door or window,
combined with wind and burning fragments from a nearby fire, may start a fire on the inte-
rior with no burning of the exterior structure. This fire would have an interior origin that
would be traced in the conventional manner. Radiation from a large nearby fire can ignite
the interior of a building through windows and doors before the exterior is ignited. The
determination of cause in such an instance will have to rest on demonstrating the existence
of a nearby fire, the wind, the opening, and the absence of any other reasonable interior
cause. While such instances are not common, they can be quite troublesome to the investi-
gator, since they prompt reports of multiple fire origins from witnesses and firefighters.

ROOF AND ATTIC FIRES

The most common type of transmission of fire from an exterior source is by burning
material that settles on a combustible roof. The resulting fire is not initially an interior
fire. This is the mode of fire transmission that is so feared in residential areas surrounded
by brush or timber. Burning embers as large as whole branches are carried aloft from
major wildland fires to land on cedar or redwood shake roofs, quickly igniting them. Such
fires caused heavy destruction in San Diego, San Bernardino, Orange, and Los Angeles
counties in California in recent years, as well as being one vector of fire spread for the
destructive Oakland Hills fire that destroyed 2,500 homes and 500 apartment units in a
single day in 1991. So many variations in the mode of transmission of fire from point to
point (depending on the ambient conditions) are encountered that care and thoroughness
on the part of the investigator are essential. All possible factors must be considered and
each weighed comparatively and in light of the established environmental conditions. (See
Chapter 6 for an example of wind-drawn urban fire.)
In residences, especially, it is not uncommon for a roof to burn away or for a com-
bined attic-roof fire to occur. For this reason, attic and roof fires are often confused, and
this can lead to misinterpretation. Actually, a roof-started fire is totally different in behav-
ior from an attic-started fire. In both instances, the absence of fire on a lower level will
define the fire as being one or the other, but often it is difficult to determine which.
An attic fire is started under the roof and will usually involve all the roof that is acces-
sible to the spread of flames beneath it. Thus, it is likely that the entire roof is either
burned away or that destruction is very widespread. Widespread burning in the attic is
always indicative of this type of fire, as in Figure 6-13.
A true roof fire, in contrast, must have been ignited from the exterior, otherwise it
would have developed as an attic fire. Sparks from chimneys, or similar causes, may ignite
leaf debris, shingles, or shakes on a roof, resulting in an exterior fire. This fire will follow
the usual inverted-cone pattern on an inclined roof and will ultimately penetrate the roof
into the attic below. It is at this point that the main distinction between an attic-started
and a roof-started fire becomes apparent. The first hole that burns through the roof ventilates
the fire, drawing air upward through the hole, which generally confines the fire to the
314 Chapter 7 Structure Fires and Their Investigation
 FIGURE 7-80 Ceiling
joists exhibit typical “top-
down” burning in this
attic fire. Note relative
absence of damage to
worn contents below.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

exterior and to parts of the roof immediately adjacent to the hole. Here, air is being
drawn from the attic to the fire, which in some instances suffices to prevent a secondary
spread as an attic fire. The major burn is usually on the outside of the roof, not under it,
although in some steep roofs fire can spread internally up along the inside pitch of the
roof. If enough of the roof is involved, of course, structural collapse can spread the fire
down into the living space, eventually causing a total burnout.
In an attic fire, there is almost always a great deal of burning under all parts of the roof
before the latter is finally breached. Because upward ventilation is limited, spread of the fire
is inevitable. Fire spread on the underside of a combustible roof assembly (plywood or wood
lattice) can be very rapid and difficult to extinguish. After a hole is chopped in the roof to
ventilate the fire under it, that fire can then be controlled much more easily because the
flames are funneled toward the hole that ventilates them. Figure 7-80 illustrates the
“burned-from-the-top-down” appearance of ceiling joists after a typical attic fire.
Although the destruction may not be significantly different in the two cases, the inves-
tigation is greatly affected, because the roof fire is caused by external ignition, while the
attic fire is ignited within the structure.

TIMELINES
Every fire has a life cycle from the time a heat source first comes into contact with the
first fuel to its extinguishment. The timeline for this cycle is defined by several time fac-
tors: ignition time, incipient fire duration, response time, extinguishment time. As we saw
in Chapter 3, the nature of the ignition source and the nature of the first fuel control the
ignition time and incipient fire. The fire growth is controlled by the fuel, its arrangement,
and the ventilation conditions. These factors can be estimated from test data (from exist-
ing research or specialized tests) or case history as part of the hypothesis testing. As seen
on the modeling data form (Figure 7-7), establishing a timeline includes both “hard”
times and “soft” times. “Hard” times are linked directly or indirectly to an accurate and
Chapter 7 Structure Fires and Their Investigation 315
 reliable clock or timing device. They include the time someone was last in the vicinity,
alarm set or activation times, time of detection by witnesses, time of 9-1-1 call, arrival
time, and sometimes time of control or extinguishment. “Soft” times are estimates of
clock time or duration by witnesses or engineering analysis of known fire behaviors. Such
timelines can be used to test and exclude some scenarios (such as smoky self-heating
processes or “instantaneous” ignition by discarded cigarette) and suggest others. The suc-
cessful use of a timeline depends on the investigator’s knowledge of time factors in all
manner of fires. Timelines can be reconstructed from fire or police dispatch recordings,
9-1-1 calls, or videos of events. Security and surveillance cameras are everywhere today.
Videos from cell phones or police, fire, or public transport “windshield” cameras should
not be overlooked.

Collection and Preservation of Evidence

The contents of a complete fire scene evidence kit are detailed in Appendix D. Briefly, an
evidence kit has a number of containers for various kinds of physical evidence and the
tools that may be required to recover them. Evidence recognition, recovery, and preserva-
tion all follow rules of plain common sense. Preservation of evidence means just that—
preserve it until it reaches the laboratory or the court. Preservation, via proper collection
and packaging, means the evidence is not allowed to change by evaporating, breaking,
spoiling, or being contaminated. Think about the limitations of the evidence being picked
up and package it accordingly. If it is a volatile liquid you suspect, you do not want it to
evaporate, so package it in an airtight can or jar and keep it cool and sealed. If it is an
organic—blood, tissue, soil, or hair—you do not want to let it decompose by mildewing
or developing mold. If you allow such material to air dry at room temperature, package
it in clean paper containers and keep it dry and cool, it will stay intact. With fragile objects,
avoid further breakage by packing in suitable containers and hand-carry to the lab.
Minute trace evidence, which is easily lost or contaminated, is best preserved by placing
in a clean container of suitable size and sealing it. Comparison samples of any floor cov-
erings involved in the fire’s origin should be taken (preferably from areas not damaged by
fire) not only to aid laboratory analysis for volatiles but also to help reconstruct the
potential contributions the carpet and pad could make to the fuel load in the room and
the spread of fire.
The procedures for other kinds of evidence are just as logical and straightforward.
Table 7-3 lists some common evidence types and collection and preservation methods.
Guidelines for evidence collection are the subject of ASTM E1188 and E1459.58

DEBRIS SUSPECTED OF CONTAINING VOLATILES

The most commonly submitted form of arson evidence is solid debris suspected of contain-
ing the residue of ignitable liquids. The major concern here is to prevent the evaporation
of volatile liquids. Collection of all physical evidence must be carried out with full realiza-
tion of the chances of contamination. Due to the sensitivity of laboratory methods, invisi-
ble traces of volatiles transferred by the tools, gloves, or even the boots of the investigator
can compromise the value of such trace evidence. Carefully clean tools using detergent and
water between scenes and between collection sites at the same scene. When collecting debris,
you must wear clean disposable plastic or latex gloves and then discard them before col-
lecting other samples. Clean trowels, scoops, brooms, and brushes in clean running water
and wipe them with a clean paper towel between samples. Always use a container similar
in size to the amount of debris and collect more than you think the laboratory will need.
No matter how strongly it smells of accelerant, more debris is always preferred.
Such debris can best be preserved in clean metal friction-lid paint cans, which are rea-
sonably airtight and unbreakable. The cans come in a variety of sizes, they are easily opened
316 Chapter 7 Structure Fires and Their Investigation
 TABLE 7-3 Physical Evidence Collection and Preservation

SAMPLE AMOUNT
EVIDENCE TYPE DESIRED PACKAGING SPECIAL HANDLING
Checks, letters, and All Place in manila envelope; Do not handle with bare hands.
other documents do not fold; wrap securely.
Fire debris with All Seal in clean empty metal paint Label as to origin and collecting
suspected can, Kapak (by Ampac) or nylon investigator; keep cool see text.
accelerants bag, or glass jar; fill container
no more than 3/4 full.
Flammable liquids All (up to 500 ml; Seal in glass bottle, jar with Do not use rubber stoppers or
1 pt) Bakelite or metal top with jars with rubber seals; do not
Teflon liner, or metal can. use plastic bottles.
Charred or burned All Pack loosely on soft cotton; Pack in rigid container; mark
paper or hand-carry to laboratory. FRAGILE; do not treat with
cardboard lacquer or other coating.
Clothing All Mark directly on clothing in If clothing is wet with water or
waistband, pocket, or coat blood, hang to air dry before
collar with initials and date; packaging to prevent
pack in clean paper bags; putrefaction; fold neatly with
wrap each item separately in clean paper between folds, and
paper; do not use plastic bags; refrigerate if possible. If there
on outside of container indicate are residues of ignitable liquids,
nature of item, date obtained, seal in metal can or Kapak bag.
and investigator’s name. Freeze if possible.
Hairs and fibers Any loose hairs Pill box, paper envelope or On outside of container: label as
or fibers. bindle, cellophane or plastic to type of material, where
Comparisons: bag; seal carefully. found, date, investigator’s
clumps of 10 to name.
20 closely cut
pubic or head
hairs from
various areas
Paint chips At least 100 mm2 Pill box, paper bindle, or Obtain control sample from
(1/2 in.2) of envelope; seal carefully. outside transfer area.
flake; entire
object if small
Soil At least 25 g (1 oz) Plastic bag or jar. If suspected Collect several samples from area.
of containing volatiles, seal in
glass jar or metal can and freeze.
Tools All Wrap each tool separately to Secure envelope or folded sheet of
prevent shifting and loss of paper carefully over end of tool
trace evidence to prevent damage and loss of
adhering paints, etc.
Tool marks Send in tool if After marks have been Cover marks with soft paper, tape
found; submit protected, wrap with in place; keep from rusting.
all items having wrapping paper and place
tool mark in envelope or box.
scratches
Dry blood stains As much as Use paper bags, boxes, bindles, On outside of container: type of
possible; submit or envelopes; sealed to specimen, date secured,
object if possible prevent loss of scrapings; investigator’s name and where
keep dry; do not use plastic found.
bags; refrigerate if possible.

(continued )

Chapter 7 Structure Fires and Their Investigation 317

 TABLE 7-3 Physical Evidence Collection and Preservation (continued )

SAMPLE AMOUNT
EVIDENCE TYPE DESIRED PACKAGING SPECIAL HANDLING
Glass fragments As much as possible Box, paper bags, bindle, or On outside of container: type of
envelope; pack and seal to specimen, date, investigator’s
avoid movement in container. name and number. Keep evidence
separated from control sample.
Firearms All Place revolvers and automatics Unload: note cylinder position; if
in cardboard box or manila automatic, do not handle side of
envelopes; tag rifles; magazine—possible fingerprints
hand-carry to laboratory. in this location; note serial number
and make of weapon.
Latent fingerprints All Secure in rigid container so Allow to air-dry if damp. Keep
surfaces do not rub on immersed in water if recovered
packaging. Clean paper from water.
envelope OK for porous
objects. Keep dry and cool.

and resealed, and the top can easily be punched for insertion of a sampling syringe. They
can be lined or unlined, but unlined cans tend to rust very quickly, so lined cans (used for
water-base paints) are preferable. Today’s linings contain no significant volatile residues.
The second choice for containing debris suspected of containing volatiles is the familiar
Mason jar, which is airtight and transparent, permitting ready visual examination of the
contents, and has a metal top for easy sampling. The third choice is a glass jar or bottle
with a metal or Bakelite cap. Plastic caps, rubber seals, or stoppers should never be used
when a liquid sample of suspected material is submitted. Gasolines, paint thinners, and
other common volatile hydrocarbons will attack and dissolve such closures. Do not fill any
container more than three-quarters full of debris to leave a headspace for sampling.
Two types of polymeric bags have been found suitable for packaging most volatile
evidence. Bags made of nylon film (sold by Grand River) or special polyester film (Kapak
sold by Ampac) have been found to retain even the lightest hydrocarbons for moderate
lengths of time. They will allow the loss of methyl and ethyl alcohols, however, and
debris suspected of containing these unusual accelerants should be sealed in metal cans
or glass jars. For most debris, however, such bags provide a convenient, low-cost, and
unbreakable alternative container.59 The polyester bags are heat-sealable and provide
even greater integrity to the evidence container than a tape seal.60 Any polymeric bag
can be penetrated by sharp debris, so care must be used.61 Because of a minute chance
of contamination during manufacturing of some cans and bags, a sample container from
each batch should be tested for the presence of trace hydrocarbons before they are issued
for use.62
A common household plastic container should be used as a last resort only.
Polypropylene jars are permissible, but never polystyrene, which is readily soluble in
gasoline. Polyethylene plastic bags, caps for coffee cans, and wrapping films are perme-
able to some hydrocarbons and allow them to escape (or cross-contamination to occur)
before analysis can be attempted and should not be used for arson debris.63 Paper and
polyethylene bags also allow cross-contamination, as external volatiles can permeate
them and be absorbed by evidence within. For large, awkward items, large plastic bags
may be the only packaging alternative. In such cases, the evidence should be double-
bagged and taken to the laboratory as soon as possible, and be kept as cool as possible
in the interim.
318 Chapter 7 Structure Fires and Their Investigation
 Refrigeration and even freezing is recommended for all arson fire debris except for glass
jars containing wet debris that can expand and shatter the container if frozen. Freezing is
especially important when soil is collected for ignitable liquid residues, as the bacteria in
garden soil can decompose petroleum products if stored at room temperatures.64

OTHER SOLID EVIDENCE

Chemicals from clandestine drug laboratories and most chemical incendiary mixtures are
corrosive and care must be taken to ensure they are not packaged in metal cans, which
can be destroyed by the chemical action. When such corrosives are suspected, the debris
may be packaged in glass jars (then kept upright to avoid contact between the lid and
debris) or in nylon or polyester bags, which are then sealed in cans.
Nonporous materials such as glass do not retain volatile petroleum distillates, but
wrappers, labels, and associated debris may offer absorbent surfaces to retain enough for
characterization in the lab. If bottle fragments are found from a suspected Molotov cock-
tail, locate the neck and base of the bottle and package all of it to prevent further break-
age. The neck often retains portions of the cloth wick, the base bears manufacturer’s data
to identify the bottle, and the sides and label may bear usable latent fingerprints. The neck
may also bear saliva with identifiable DNA.
Newly developed techniques using lasers, chemicals, and physical developers are now
making it possible to recover identifiable latent prints from objects that have been exposed
to gasoline, water, or even fire. Unless an object has been heavily damaged by direct expo-
sure to fire, there is a chance that latents can be recovered. The crime laboratory should
be contacted for advice before dismissing the possibility of prints, because prints have
been identified on glass bottles, plastic jugs, and metal cans even after fire exposure.65 It
has recently been reported that DNA-typable materials were recovered from unburned
towel fragments from the neck of a broken and burned Molotov cocktail.66 More discus-
sion is presented in Chapter 14.

LIQUIDS
Uncontaminated flammable liquids from the scene are the best kind of evidence for lab-
oratory analysis. If the liquid is still in its original container, as long as that container can
be properly sealed, it may be left there. Otherwise, transfer liquids by eyedropper (pipette)
or syringe to a clean sealable container. Do not use a bottle with a rubber stopper or a jar
with a rubber sealing ring for any liquid hydrocarbon. It will dissolve the rubber and leak
out. A small quantity of liquid in debris may be recovered directly with an eyedropper or
by absorbing it into a pad of clean absorbent cloth, gauze, or cotton wool, and then seal-
ing it in a suitable airtight container.
When flammable liquids are suspected in concrete, samples can be collected from
cracks or expansion joints, but often, a better method is to wet-broom the suspected area
and then spread a quantity of absorbent material. The best absorbent is clay-type kitty lit-
ter, although diatomaceous earth or even flour can be used. This layer is allowed to stand
for 20 to 30 minutes and then recovered into a clean metal can using a clean putty knife
and brush for laboratory analysis.67 A control sample of the unused absorbent must be
sealed in a separate can. The cleanliness of the absorbent (and any collection tools used)
is critical to the value of this evidence. The use of hazardous material recovery pads is not
recommended due to the possible contamination of such pads by exposure to volatile
materials in the environment. Chemically aggressive absorbents such as activated charcoal
are not recommended due to their susceptibility for external contamination.

TESTING OF HANDS
Liquid hydrocarbon fuels can be absorbed into the skin after even brief exposure, but the
warmth of living tissue causes rapid desorption of such volatiles. Over the years, a number
Chapter 7 Structure Fires and Their Investigation 319
 of methods have been suggested for testing of hands of suspects in arson investigations.
Canines have often alerted to hands contaminated with gasoline even after some time has
elapsed, but the trace amounts present make recovery for laboratory analysis very difficult.
Methods such as swabbing with a dry or solvent-wetted cotton gauze are useless for recov-
ering such traces. The best system is to place a clean disposable PVC or latex glove on each
of the suspect’s hands or to secure a polyester or nylon fire debris bag around the hand
with a rubber band at the wrist, enclosing an activated carbon recovery device (such as a
carbon strip or solid-phase microextraction (SPME) device) in the bag for 15–20 minutes.
The warmth of the skin will cause the absorbed volatiles to evaporate into the glove or bag
(assisted by exposure to a heating lamp), where they are sampled by the activated carbon.
After the subject removes his or her hand, the glove or bag is sealed with the sampling
device inside and forwarded to the lab for analysis. Tests indicate that for this testing to
yield positive results, it must be done within an hour or two of contact with gasoline.68
Swabbing of the hands of suspects in explosives cases using dry cotton-tipped swabs
[such as those in gunshot residue (GSR) kits] or ones wetted with distilled water or methyl
or isopropyl alcohol (but not with the acid supplied with the GSR kit) can capture traces
of either organic or inorganic explosives. Fingernail scrapings using clean flat wooden
toothpicks are collected and sealed in clean glass or plastic vials. This technique is partic-
ularly valuable for soft organic explosives such as dynamite or C-4.

TESTING OF CLOTHING
Rapid pouring of an ignitable liquid onto a hard surface often causes spattering onto
nearby shoes and pants cuffs. Arsonists may also step into the liquid after it is poured.
The gas chromatographic profile of such transfers will be different from that from vapor
contact alone (see Chapter 14). Because such volatiles are lost within a few hours (1–6) if
the garment is worn, residues on clothing can help identify the perpetrator. Clothing of a
suspect should be seized as soon as possible and packaged separately in appropriate
sealed containers until tested by the lab.
A New Zealand study illustrated the desirability of analyzing clothing and shoes of sus-
pected arsonists for the presence of ignitable liquids. The study found that detectable
amounts of petrol were transferred to the clothing and shoes of a person during the action of
pouring it around the room, with various pouring heights and floor surfaces. Results of the
study showed that petrol was always transferred to the shoes and often transferred to both
the upper and lower clothing but that the seizure and proper packaging must take place
quickly.69 In tests using a carbon strip extraction method, it was found that 10 ml (0.34 oz
or 2 tsp) of gasoline on clothing worn continuously evaporated much more quickly than on
the same clothing left on the lab bench at 20°C–25°C (68°F–77°F) and was barely detectable
after only 4 hours of wear.70 Clothing must be recovered and packaged soon after exposure.
A second New Zealand study further examined a potential defense argument that the
gasoline residues located on a suspect’s clothing came from a legitimate source. The study
showed that it is extremely unlikely to have detectable levels of gasoline on clothing and
shoes following the activity of filling a vehicle with fuel, without the spillage of any notice-
able amount of fuel onto the clothing and shoes. The study found it possible for gasoline
to be detected on the shoes of people after using a gasoline-powered mower to cut their
lawns. The study further underscored that the time delays between commission of the
crime and seizure of the suspect’s clothing need to be taken into account when considering
chain of evidence the possibility of evaporation of any volatile residues from the suspect’s clothing.71
(chain of custody) ■
Written documentation
of possession of items CHAIN OF EVIDENCE
of physical evidence
from their recovery to
No matter what kind of material is recovered or how carefully it is preserved, it is value-
their submission in less as evidence if it has not been maintained and documented completely. The chain of
court. evidence (chain of custody) is a record of the existence and the whereabouts of the item
320 Chapter 7 Structure Fires and Their Investigation
from the time it is discovered at the scene to the moment it is yielded to the court. This
documentation of an item is in the form of a written record, usually on the sealed evidence
package, that lists every person who has had custody or possession of the item. The list
must include the dates and times of transfer and the signature of anyone involved in that
transfer (see ASTM E1459). The documentation starts with the photography of the item
in place at the scene and records all people who handle the evidence or its sealed container.
The number of people handling evidence should be kept to a minimum to minimize
chances for damage and to reduce the number of people who must be ready to testify in
court as to the existence of the evidence, for each person is then a link in this crucial chain.
An evidence log is an accurate, complete list of all items of evidence recovered from
a scene or location. Many agencies prefer to use a preprinted form such as that in
Appendix I (or see NFPA 921). Note that the form provides a checklist for recording what
was recovered, from where, and by whom, and also that the item was documented by
photograph and diagram before recovery. If more than one person is recovering evidence
at a scene, it is preferable that the evidence log for a scene be maintained by one person
to avoid duplicate or conflicting numbering. It is the responsibility of the person recover-
ing the item to ensure that each item is properly packaged, labeled, or tagged, and that
the photos and diagram entries have been made. Such a log is a useful supplement to the
custody list on the item itself.
The chain of custody is much like a ship’s anchor chain. If the anchor chain is broken,
it cannot be recovered, and the ship will probably be lost. If the chain of custody is bro-
ken, it cannot be restored, the evidence is valueless to the court, and the case may well be
lost. A complete, well-documented chain allows the court to admit evidence that can be
verified, allowing the investigator to say, “Yes, this is the device I recovered from the
scene. I know that because here is where I marked it, here is the container in which I
sealed it.” The importance of the chain of evidence to a successful fire investigation cannot
be overemphasized.

Analysis and Hypothesis Testing

The investigator must be aware of any tendency to prejudge the origin and cause of any
fire. The fire scene investigation is carried out to identify the origin of the fire—where it
started—and its cause—what burned at the area of origin, and what caused the fuel to
ignite. Careful evaluation of the indicators should result in an accurate determination of at
least an area, if not a point, of origin and a hypothesis of the direction, intensity, and dura-
tion of the fire. Without at least a definable general area or origin, no reliable evaluation
of ignition sources can be carried out. In the vicinity of the origin, the presence of the first
fuel ignited and a possible ignition source should have been detected. The investigator must
be able to define and defend the chain of events from the event that brought the first fuel
and ignition source together to the spread of the fire, based on knowledge of the location
and size of major fuel packages and relevant fire dynamics. The amount of damage visible
must correspond to the heat release rate that would be expected from those packages.
The scientific method demands that all reasonably possible alternative hypotheses be
evaluated, tested (against test data, scene evidence, or witness information), and eliminated
to demonstrate the reliability of the final conclusion. To be sure all data critical to such
testing are collected, a form such as the one shown in Figure 7-7 is strongly suggested.
If there are indications of a natural or accidental cause, do all the other indicators
agree? Accidental sources include lightning, trash fires, shorted or overheated electrical
connections, smoking materials, cooking accidents, chemical reactions, faulty appliances,
or even children playing with matches. Is there identifiable contact between a possible
heat source and a susceptible fuel, and is that contact enough to promote ignition? During
the course of the investigation, any indication of incendiary origin found will prompt
other inquiries—Who set it and why? What other signs of criminal activity such as burglary,
Chapter 7 Structure Fires and Their Investigation 321
 robbery, fraud, or murder are present? Even in the absence of an incendiary device, the
crime of arson can be proven in the absence of all logically possible accidental and natu-
ral causes at the point of origin. This methodology is sometimes called a negative corpus,
since a complete positive corpus for the crime of arson includes proof of burning, proof
of incendiary cause, and proof of absence of natural and accidental causes. The term
negative corpus has no recognized definition but implies that the corpus of the crime (or
event) has not been proven. This is not to say it is permissible to conclude a fire is an
arson simply because “too much damage was done” or because “no accidental source
could be identified.” While some investigators treat negative corpus as a “catchall,” it is
a very difficult case to prove and should be relied upon only in the most special circum-
stances. Negative corpus cases are accepted in some jurisdictions and not in others, so the
investigator should confer with local prosecutors before pursuing such cases.
The thorough investigator also considers all possible alternative explanations for the
fire and evaluates the evidence (whether physical or testimonial) in deciding whether those
alternatives are excluded. Diligent application of the ignition matrix concept described in
Chapter 6 will result in consideration of likely “first fires” and ignition sources. Such con-
sideration and elimination of alternative scenarios is a primary element of the scientific
method of fire investigation. Careful engineering analysis (including fire modeling) may
be useful in testing and eliminating hypotheses. If alternative explanations cannot be
excluded, additional information either from the scene or from research or testing can
throw light on the energies, times, or conditions needed for each. Based on these new
data, it may be possible to exclude all the other possible hypotheses and make it easier to
defend the final conclusion. (This process gives rise to the term affirmative exclusion to
be used instead of negative corpus.) Such reasoning can be applied in cases where the first
fuel ignited and source of ignition cannot be positively demonstrated if the scientific rea-
soning and assessment can be shown to be sound. If all possible explanations cannot be
excluded, then the investigator must admit so. It is not possible to identify the cause of
every fire. Depending on the extent of damage to the structure, there may not be enough
information to reliably conclude the actual cause. The professional investigator must
realize his or her limitations, as well as the limitations of the evidence, and know when
to admit that the cause is “undetermined.”

322 Chapter 7 Structure Fires and Their Investigation

 CHAPTER REVIEW

Summary
Before reaching a final decision, it is always a good known, although they may have little bearing on the
idea to step back, mentally if not physically, and present theory or interpretation of the total fire.
reassess what has been revealed. This process involves As has been previously discussed, the accuracy of
a resurvey of the entire site in view of the presumed the determination of origin and cause depend upon the
finding of the point of origin, source of fuel, ventila- observational and analytical skills of the investigator.
tion, and ignition mechanism, including heat release With so much depending on this determination, it
rates of the fuel packages involved. This is an impor- should be obvious that the investigator must carefully
tant portion of the investigation, because if there are assess the indications and “proof” available before
discrepancies in the theory that is presently offered, reaching a final conclusion about a fire. It is not a dis-
they can be found and reconciled, new data collected, honor to conclude that no origin or cause can be deter-
or the theory modified. Such disparities can be used by mined if there is any doubt in the mind of the expert
the opposing attorney in any subsequent trial that may investigator. If that trained expert investigator cannot
occur; failure to resolve all the contradictions can be identify an origin or cause, a less-qualified person should
highly embarrassing. This is the process of testing the not. In accordance with the scientific method, if new
final hypothesis. data comes to light or the unreliability of previous
It is wise to note specifically any types of building information is shown, the professional investigator
material that may have enhanced the development of must reassess the conclusions previously reached. This
the fire and to make certain that one clearly under- reevaluation may result in a modification of the earlier
stands the plan of the structure. The type and location conclusion. Certainly, in light of the destructive force
of ventilation sources, as well as those of fuel pack- of the event itself, it should not be surprising that the
ages, must also be fully considered. Although these destruction of some structures is so complete that
details may have all been noted in the earlier investiga- no determination can be made in good conscience.
tion, failure to fix them in the memory and the docu- No matter how skilled an investigator may be, in the
mentation can later raise problems that are better absence of adequate indicators of fire behavior or
avoided. It is also possible that this review of the divine intervention, some fire causes simply cannot be
inspection will turn up unclear details that should be reliably identified.

Review Questions
1. What is overhaul and why is it critical that it be 8. What is a heat horizon and how does it help estab-
kept to a minimum? lish direction of development of a structure fire?
2. What basic fire behaviors are revealed by post- 9. Name three ways ventilation openings can affect
fire indicators? fire patterns in a room fire.
3. Name three ways suppression efforts can affect 10. What types of fuels produce the deepest burn
fire spread and post-fire indicators. pattrns on floors?
4. Name five types of information that can be gath- 11. How does the radiant heat flux on a wood surface
ered from firefighters. affect the char rate of the wood?
5. What is the most common search pattern for 12. How does window glass provide useful information
structure fires and what are four of its advan- to a fire investigator?
tages? 13. Why is a contents inventory important in a fire
6. Why is the documentation and reconstruction of investigation?
fuel load so important? 14. Name four ways a fire scene should be documented.
7. What does the angle of a burn pattern on a wall 15. What is the chain of custody and why is it
or furniture reveal about the fire’s development? important?
Chapter 7 Structure Fires and Their Investigation 323
References
1. D. J. Icove and J. D. DeHaan, Forensic Fire Scene 17. ASTM E1459: Standard Guide for Physical Evidence
Reconstruction, 2nd ed. (Upper Saddle River, NJ: Labeling and Related Documentation (West
Pearson Education/Brady, 2009), 22–27. Conshohocken, PA: ASTM, 2005).
2. NFPA 921: Guide for Fire and Explosion Investigation 18. D. I. Icove and J. D. DeHaan, “Hourglass Patterns”
(Quincy, MA: NFPA, 2011); NIJ Guide for Fire and in Proceedings of IFSI Conference, Cincinnati, OH,
Arson Scene Evidence (Washington, DC: Department May 2006.
of Justice, 2000); IFSTA, Fire Investigator, 2nd ed. 19. NFPA 921: Guide for Fire and Explosion Investigations
(Oklahoma State University, Stillwater, OK: Fire (Quincy, MA: NFPA, 2008 and 2011).
Protection Publications, 2010). 20. D. Madrzykowski and S. I. Kerber, “Fire Fighting
3. M. Wallace and J. D. DeHaan “Overhauling for Tactics under Wind Driven Conditions: Laboratory
Successful Fire Investigation,” Fire Engineering Experiments,” NIST Technical Note 1618, 413 pp.
(December 2000), 73–75. (January 2009); S. I. Kerber and D. Madrzykowski,
4. Ibid. “Fire Fighting Tactics under Wind Driven Fire
5. D. Smith, “Firefighter’s Role and Responsibility in Conditions: 7-Story Building Experiments,” NIST, TN
Preserving the Fire Scene and Physical Evidence,” The 1629: NIST Technical Note 1629, 593 pp. (April, 2009).
Times: Risk Management for Emergency Medical 21. J. D. DeHaan, “The Reconstruction of Fires Involving
Services 3 (September 1995). Highly Flammable Hydrocarbon Liquids,” Ph.D.
6. Federal Bureau of Investigation, “Definitions and dissertation, Strathclyde University, Glasgow, Scotland,
Guidelines for the Use of Imaging Technologies in July 1995.
the Criminal Justice System,” Forensic Science 22. Ibid.
Communications 1, no. 3 (October 1999). 23. A. Putorti, J. A. McElroy, and D. Madrzykowski,
7. J. D. DeHaan, “Advanced Tools for Use in Forensic Flammable and Combustible Liquid Spill/Burn Patterns,
Fire Scene Investigation, Reconstruction and NIJ Report 604-00 (Washington, DC: NIJ, 2001).
Documentation” in Proceedings Interflam 2004 24. J. D. DeHaan, “The Consumption of Animal
(London: Interscience Communications, 2004), Carcasses and Its Implication for the Combustion
1079–1100; NFPA 921: Guide for Fire and Explosion of Human Bodies in Fire,” Science and Justice
Investigations (Quincy, MA: NFPA, 2008 and 2011). (January 1999).
8. Better Homes and Gardens, Chief Architect, Chief 25. J. D. DeHaan, “Are Localized Burns Proof of
Architect, Inc., Coeur d’Alene, ID; retrieved November Flammable Liquid Accelerants?” Fire and Arson
21, 2010, from http://www.chiefarchitect.com/company/ Investigator 38 (September 1987); J. C. Gudmann J. C.,
location.html. and D. Dillon, “Multiple Seats of Fire: The Hot Gas
9. The Crime Zone (Beaverton, OR: The CAD Zone Layer,” Fire and Arson Investigator 38, no. 3 (1988):
Inc.). 61–62; J. Albers, “Pour Pattern or Product of
10. T. G. Gersbeck, “Advancing the Process of Post- Combustion?” California Fire/Arson Investigator (July
Investigation,” paper presented at AAFS, Washington 2002): 5–10.
DC, February 2008. See also: M. G. Haag M. G. and 26. R. J. Powell, “Testimony Tested by Fire, “Fire and
T. Grissim, “Technical Overview and Application of Arson Investigator (September 1992): 42–51; J. J.
3-D Laser Scanning for Shooting Reconstruction and Lentini, “The Lime Street Fire: Another Perspective,”
Fire Scene Reconstruction,” paper presented at AAFS, Fire and Arson Investigator (September 1992): 52–54;
Washington DC, February 2008. J. D. DeHaan, “Fire: Fatal Intensity,” Fire and Arson
11. Icove and DeHaan, Forensic Fire Scene Reconstruction. Investigator (September 1992): 55–59.
12. ASTM E620: Reporting Opinions of Scientific or 27. D. D. Drysdale, An Introduction to Fire Dynamics,
Technical Experts (West Conshohocken, PA: ASTM, 2nd ed. (Chichester, UK: Wiley, 1999), 189; V.
2004). Babrauskas, “Charring Rate of Wood as a Tool for
13. ASTM E678: Evaluation of Technical Data (West Fire Investigators” in Proceedings Interflam 2004
Conshohocken, PA: ASTM, 2007). (London: Interscience Communications), 1155–70.
14. ASTM E860: Standard Practice for Examining and 28. S. Carman, “Progressive Burn Pattern Development in
Testing Items That Are or May Become Involved in Post-Flashover Fires” in Proceedings Fire and Materials,
Litigation (West Conshohocken PA: ASTM, 2006). 789–800, San Francisco, CA, 2009.
15. ASTM E1020: Reporting Incidents (West Conshohocken, 29. B. V. Ettling, “The Significance of Alligatoring of
PA: ASTM, 2006). Wood Char,” Fire and Arson Investigator 41
16. ASTM E1188: Collection and Presentation of (December 1990).
Information and Physical Items by a Technical 30. M. J. Spearpoint and J. G. Quintiere, “Predicting the
Investigator (West Conshohocken, PA: ASTM, Ignition of Wood in the Cone Calorimeter,” Fire Safety
2005). Journal 36, no. 4 (2001): 391–415.

324 Chapter 7 Structure Fires and Their Investigation

31. J. L. Sanderson, “Tests Add Further Doubt to Concrete 49. B. Béland, C. Ray, and M. Tremblay, “Copper-
Spalling Theories,” Fire Findings 3 (Fall 1995); F. P. Aluminum Interaction in Fire Environments,” Fire and
Smith, “Concrete Spalling: Controlled Fire Tests and Arson Investigator 37 (March 1987). Reprinted from
Review,” Fire and Arson Investigator 44 (September Fire Technology 19 (February 1983): 22–30.
1993): 43; C. R. Midkiff, “Spalling of Concrete as an 50. R. W. Henderson, “Thermodynamics of Ferrous
Indicator of Arson,” Fire and Arson Investigator 41 Metals,” National Fire and Arson Report 4 (1986).
(December 1990): 42. 51. Dennis Farrell, Minneapolis, MN, personal communi-
32. D. V. Canfield, “Causes of Spalling of Concrete at cation, 1996.
Elevated Temperatures,” Fire and Arson Investigator 52. A. R. Earls, “It’s Not Lightweight Construction. It’s
34 (June, 1984). What Happens When Lightweight Construction Meets
33. B. Béland, “Spalling of Concrete,” Fire and Arson Fire,” NFPA Journal (July–August 2009): 38–45.
Investigator 44 (September 1993): 26. 53. KTW Corp., Newark, CA, personal correspondence,
34. DeHaan, “The Reconstruction of Fires”; J. Novak, 2008.
personal communication, 2008. 54. J. D. DeHaan, “Canine Accelerant Detection Teams:
35. R. N. Thaman, “Laboratory Can Help Investigator Validation and Certification,” CAC News ( Summer
Pluck Arson Evidence Out of Debris,” Fire Engineering, 1994); M. E. Kurz et al., “Evaluation of Canines for
131, no. 8 (August 197): 48–50, 52. Accelerant Detection at Fire Scenes,” Journal of
36. N. Saitoh and S. Takeuchi, “Floor Scene Imaging of Forensic Sciences 39 (November 1994); R. Tindall and
Petroleum Accelerants by Time-Resolved Spectroscopy K. Lothridge, “An Evaluation of 42 Accelerant
with a Pulsed Nd-YAG Laser,” Forensic Science Detection Canine Teams,” Journal of Forensic Sciences
International 163 (2006): 38–50. 40 (July 1995); D. M. Gialamas, “Enhancement of
37. D. C. Mann and N. D. Pataansuu, “Studies of the Fire Scene Investigations Using Accelerant Detection
Dehydration/Calcination of Gypsum Wallboard” in Canines,” Science and Justice 36, no.1 (1996): 51–54.
Proceedings Fire & Materials 2009 (London: 55. D. M. Gialamas, “Is It Gasoline or Insecticide?” CAC
Interscience Communications, January 2009), 827–38. News (Spring 1994): 16–20.
38. J. E. Posey and E. P. Posey “Using Calcination of Gypsum 56. J. Lentini, C. Cherry, and J. Dolan, “Petroleum Laced
Wallboard to Reveal Burn Patterns,” Fire and Arson Background,” Journal of Forensic Sciences 45, no. 5
Investigator 33 (March 1983). (2000): 965–89; J. D. DeHaan, “Our Changing World:
39. P. M. Kennedy, K. C. Kennedy, and R. L. Hopkins, Part 3 Is More Sensitive Necessarily More Better?” Fire
“Depth of Calcination Measurement in Fire Origin and Arson Investigator 53, no. 2 (January 2003).
Analysis” in Proceedings Fire and Materials 2003 57. J. L. Sanderson, “Science Reconstruction,” Fire Findings
(London: Interscience Communications, 2003). 9, no. 1 (Winter 2001): 12; DeHaan, “Are Localized
40. W. A. Tobin and K. L. Monson, “Collapsed Springs in Burns Proof of Flammable Liquid Accelerants?” 90.
Arson Investigations: A Critical Metallurgical 58. ASTM E1188; ASTM E1459.
Evaluation,” Fire Technology 25 (November 1989). 59. J. D. DeHaan and F. A. Skalsky, “Evaluation of
41. J. J. Lentini, D. M. Smith, and R. W. Henderson, Kapak Plastic Pouches,” Arson Analysis Newsletter 5,
“Baseline Characteristics of Residential Structures no. 1 (1981).
Which Have Burned to Completion: The Oakland 60. D. C. Mann, “In Search of the Perfect Container for
Experience,” Fire Technology 28 (August, 1992). Fire Debris Evidence,” Fire and Arson Investigator
42. M. J. Skelly, “An Experimental Investigation of Glass (April 2000): 21–25.
Breakage in Compartment Fires,” NIST-GCR-90-578, 61. J. Demers-Kohl et al., “An Evaluation of Different
U.S. Department of Commerce, 1990. Evidence Bags Used for Sampling and Storage of Fire
43. P. J. Pagni and A. A. Joshi, “Glass Breaking in Fires” Debris,” Canadian Society of Forensic Science Journal
in Proceedings Third International Symposium on Fire 27, no. 3 (1994).
Safety Science (London: Elsevier Applied Science, 62. W. D. Kinard and D. R. Midkiff, “Arson Evidence
1991), 791–802. Container Evaluation II: Kapak Bags, A New
44. N. Nic Daéid, “The ENFSI Fire and Explosion Generation,” Journal of Forensic Sciences
Investigation Working Group and the European Live (November 1991).
Burn Tests at Cardington,” Science and Justice 43, no. 1 63. R. E. Tontarski, “Evaluation of Polyethylene Containers
(January–March 2003). Used to Collect Evidence for Accelerant Detection,”
45. J. J. Lentini, “Behavior of Glass at Elevated Journal of Forensic Sciences 28, no. 2 (1983).
Temperatures,” Journal of Forensic Sciences 37 64. D. C. Mann and W. R. Gresham, “Microbial
(September 1992). Degeneration of Gasoline in Soil,” Journal of Forensic
46. Ibid. Sciences 35, no. 4 (1990).
47. R. J. Gilchrist, “Die Casting,” The Ohio Engineer, 65. J. Deans, “Recovery of Fingerprints from Fire
(April 1937). Scenes and Associated Evidence,” Science and
48. Lentini, Smith, and Henderson, “Baseline Characteristics.” Justice 40, no. 3 (2006): 153–68; N. Nic Daéid,

Chapter 7 Structure Fires and Their Investigation 325

 Fire Investigation (London: Taylor and Francis and Persistence of Gasoline on Hands,” paper presented
and CRC Press, 2004). at EAFS, Cracow, Poland, September 2000.
66. “Testing Molotov Cocktails for DNA,” Fire Findings 69. S. A. Coulson and R. K. Morgan-Smith, “The Transfer
12, no. 4 (Fall 2004): 14. of Petrol onto Clothing and Shoes While Pouring
67. R. E. Tontarski, “Using Absorbents to Collect Petrol around a Room,” Forensic Science International
Hydrocarbon Accelerants from Concrete,” Journal of 112, no. 2 (August 14, 2000): 135–41.
Forensic Sciences 30, no. 4 (1985); D. Mann and N. 70. R. Morgan-Smith, “Persistence of Petrol on Clothing,”
Putaansuu, “Alternative Sampling Methods to Collect paper presented at ANZFSS Symposium of Forensic
Ignitable Liquid Residues from Non-porous Areas Sciences, Gold Coast, Queensland, Australia, March
Such as Concrete,” Fire and Arson Investigator 57, 2000.
no. 1 (July 2006): 43–46. 71. S. Coulson et al., “An Investigation into the Presence
68. J. R. Almirall et al., “The Detection and Analysis of of Petrol on the Clothing and Shoes of Members of
Ignitable Liquid Residues Extracted from Human Skin the Public,” Forensic Science International 175 (2008):
Using SPME/GC,” Journal of Forensic Sciences 45, no. 2 44–54.
(March 2000): 453–61; M. Darrer et al., “Collection

326 Chapter 7 Structure Fires and Their Investigation

 CHAPTER
8
Wildland Fires and Their
Investigation

KEY TERMS

area of transition, p. 337 flanking, p. 330 rekindle, p. 345

backing, p. 330 hot sets, p. 351 spot fires, p. 342
crowning, p. 330 ladder fuels, p. 329

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify the common causes of wildland fires.
■ Explain the basic concepts of fire behavior and spread of wildland and grass
fires over uniform ground cover with varying degrees of slope and wind
conditions.
■ Describe the investigative methodology for determining the origin of wildland
fires.
■ Explain how to conduct a fire scene search of a typical “V” pattern on a hillside.
■ List the typical burn patterns seen at wildland fire scenes.
■ Explain several techniques used to document wildland fire investigations.
■ List several examples of common sources of ignition for wildland fires.
■ Understand the concepts of the collection and preservation of physical evidence,
including documentation.

327
 For additional review materials, appendices, and suggested readings,
visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

F
ires in open land covered with grass, brush, or timber are often termed wildland fires.
Although they are often terrifying in their destructive power and intimidating in their
coverage, they begin, like almost every other fire, with suitable fuel and a small, local-
ized source of ignition. In some respects, the investigation of a wildland fire is simpler than
that of a structure fire, because the fuels involved are generally limited to naturally occur-
ring vegetation. They are ignited by some finite source of heat—natural, such as lightning;
accidental, such as a discarded match; or incendiary, with an intentional ignition device.
Some evidence of the source of ignition will remain unless, as in a fire started with a lighter,
the source has been removed from the scene.
The causes of wildland fires are more varied than for structure fires. The California
Department of Forestry and Fire Protection (CDF), which annually responds to an average of
over 5,500 wildland fires in its direct protection area, reported 11 major categories for iden-
tified causes, as shown in Table 8-1.1

TABLE 8-1 Wildland Fire Causes in California, 2006–2008

PERCENT
CAUSE 2006 2007 2008 (3-yr)
Arson 319 227 220 6.4
Campfire 113 41 23 1.5
Debris burning 455 490 431 11.5
Equipment use 1,237 489 401 17.7
Lightning 237 126 332 5.8
Miscellaneous 627 1,061 1,115 23.3
Playing with fire 40 12 8 0.5
Power line 130 27 20 1.5
Railroad 19 0 4 0.2
Smoking 87 84 62 1.9
Undetermined 986 969 908 23.8
Vehicle 555 84 69 5.9
Total 4,805 3,610 3,593 100.0

Source: 2006–2008 Historical Wildfire Activity Statistics (Redbooks) (Sacramento, CA: California Department of
Forestry and Fire Protection, 2008).

328 Chapter 8 Wildland Fires and Their Investigation

 Unlike with structure fires, however, there are complications in retracing the progres-
sion of the fire by variable environmental conditions—fuels, wind, weather, and terrain. The
large scale of many wildland fires results in major firefighting efforts that can disturb or
destroy delicate evidence. Because of the fire, an area of interest may remain inaccessible
to the investigator for days, during which time the evidence is exposed to the elements,
further diminishing its value. All fire investigations require thorough and systematic exami-
nation of the suspected area of origin and logical and analytical assessment of the evidence
found, and wildland fires are no exception. The investigator who understands fuels,
fire behavior, and the effects of environmental conditions is in a better position to interpret
the subtle and sometimes delicate signs of fire patterns in wildland fires, and therefore is
better able to identify the origin and cause, no matter what type of fire is involved.

Fire Spread
The spread of a wildland fire (and thereby the means of retracing its development back
to a point of origin) involves consideration of many more factors than the typical struc-
ture fire. These factors include the following:
■ Topography
■ Slope (and the local winds slopes can engender) and
■ Aspect (fuels on south-facing slopes in the Northern Hemisphere are exposed to
more sun than those on the north-facing slopes and are therefore drier and able
to support larger, faster-moving fires)
■ Terrain (natural noncombustible barriers of rock or bare soil versus chimneys
(defiles) that can draw fires faster upward
■ Weather
The weather obviously plays a much more important role than in most structure
fires, both before a fire in drying out fuels and during the fire. Weather conditions
include wind direction, speed, and variability, as well as temperatures and relative
humidity (as discussed in Chapter 3)2. Since these conditions affect both ignition
and spread, they must be documented for times before and during the fires as thor-
oughly as possible
■ Fuels
Fuels can be extremely varied and must be documented (both in the fire area and in
adjoining areas). While the names can vary, the general categories are displayed in
Table 8-2.
ladder fuels ■ Inter-
Each category supports a different type of fire. The duff and surface litter most often
mediate height fuels
burns as a slow smoldering fire, only contributing when strong winds can kick up burn- between the ground
ing embers to produce airborne fire brands. Grasses, shrubs, and slash can all support litter and the crowns
fast-moving flaming fires that can ignite the ladder fuels.3 of trees overhead.
The crown, being a porous array of fine fuel elements surrounded by lots of air, can
support extremely fast burning, intense, flaming fires. The canopy can be nonhomoge-
nous, as well. Ladder fuels constitute the bridge that can connect fires started in ground
litter to the canopy or connect a crown fire to the next susceptible fuel array. The oncom-
ing fire first evaporates (by radiant or convective transfer) the water in the exposed
foliage, then evaporates the volatiles present, and then pyrolyzes the hemicellulose, cellu-
lose, and lignin components into combustible gases.4 The seasonal stage of the foliage
(green or live, mature, cured or dead) has a significant effect on the fuel moisture and
thereby on the fuel’s ignitability and rate of spread of the fire. The volumes and arrange-
ments of fuels (continuous, intermittent, or isolated) also are important variables that the
investigator must consider. Rothermel has offered guidance regarding the interpretive
Chapter 8 Wildland Fires and Their Investigation 329
 TABLE 8-2 Fuel Categories in Wildland Fires

CATEGORY EXAMPLE
Duff Decomposed organic layer (humus) on the soil
Ground or surface litter Dropped leaves
Dropped branches
Needles and twigs
Ladder fuels Intermediate-height grasses
Shrubs [less than 2 m (6 ft) in height]
Seedlings
Small trees
Lower branches
Loose bark
Crown Canopy of foliage leaves, needles, and fine twigs (more than 2 m
above ground)
Slash Branches and foliage left as heavy ground cover after
logging operationsa
a
From R. C. Rothermel, How to Predict the Spread and Intensity of Forest and Range Fires, Gen. Tech. Rep. INT-143
(Ogden, UT: U.S. Dept. of Agriculture, National Wildfire Coordinating Group, June 1983), 10–12.

complexities of these factors and has developed nomograms for predicting the effects of
numerous such variables on fire behavior.5 Anderson has published a photographic guide
to identify the fuel model of various fuel arrays.6
The wildland fire investigator must be prepared to take all these factors into account
when evaluating the spread of a large fire with the intent of backtracking it to its origin
and then evaluating various ignition mechanisms there. The proper investigation there-
fore entails the determination and documentation of as many of these factors as possible.

Fuels
By definition, wildland fires consume grass, brush, and timber but will also involve struc-
tures and even vehicles and equipment as subsidiary fuels. Although the basic fuels are all
cellulosic in nature, they can vary considerably in their fire behavior because of their vol-
ume, density, arrangement, and moisture content. Short dry grasses, for instance, tend to
flash over once, contributing little to the fire (except to carry it into denser fuels), while
backing ■ Slow exten- short green grasses may not burn at all during their first exposure to the fire but will
sion of a fire down-
slope or into wind in
dry out, ready to provide fuel for a reburn of the same area. Long dry grasses may be
the opposite direction burned off at their tops if the fire is passing quickly across dense growth. The remaining
of its main spread. grass stems then provide fuel for a second or “following” burn. If a slow-moving fire
flanking ■ Lateral
moves into tall grass, it is more likely to burn the stems away near the ground, allowing
spread of fire in a the tops to fall. This behavior is most often seen in areas where the fire is backing into
direction at right the wind or down a hill or flanking or moving laterally at roughly right angles to its major
angles to the main direction of progression.
direction of fire Brush and heavy timber may also burn in the same ways. Fire may flash across and
growth.
upward through the leaves and finer twigs (called crowning), leaving heavier portions to
crowning ■ Rapid ignite later as fire conditions change. This condition is especially notable downwind and
extension of fire uphill from a fast-moving fire, in the area sometimes called the head for obvious reasons.
through the porous
array of leaves,
Radiant and convective heat from an intense nearby fire can raise the temperature of an
needles, and fine fuels entire bush or tree or its volatile resins to its ignition temperature, whereupon it
above 2 m. “explodes” in flames.
330 Chapter 8 Wildland Fires and Their Investigation
FIRE SPREAD
The nature of the fuels will have considerable effect on the spread of the fire and its inten-
sity. The higher resin content of firs and pines causes them to burn faster and more fiercely
than many hardwoods. Few trees are resistant to the intense flames of a fully developed
forest fire. Some, like the firs, will die when the cambium, the living layer of cells just
beneath the bark, is breached even though damage to the rest of the tree may be slight.
The critical temperature for cambium death is about 60°C (140°F), so as a general rule,
the thicker the bark, the greater the tree’s resistance to fire death.7 A few trees, like the
giant redwoods of the Pacific Coast, resist even serious fires with a thick, almost noncom-
bustible bark. (Some species actually require a fire to open their cones to allow the sow-
ing of new seedlings.)
Finely divided cellulosic fuels, such as grasses with a high surface-to-mass ratio, will,
of course, ignite easily when dry and allow the spread of fire by fast flashing across the
tops or by slow, creeping, smoldering spread along the ground. In semiarid regions, the
brush (sometimes generically termed chaparral) protects its living cells from death via
evaporation by developing a high oil or grease content in the wood. The high oil or grease
content shields the cells from the hot, dry environment allowing the cells to maintain a
normal moisture level. The oils or greases in this brush, however, provide an exception-
ally high-quality fuel for a fire. A fire in this type of fuel develops extremely high temper-
atures that cause the oils of nearby plants to volatilize at a rapid rate, generating clouds
of flammable vapors. The passage of a fire through such brush is awesome in its ferocity.

MOISTURE CONTENT
The general response of wildland fuels to changes in relative humidity is categorized in
terms of the time required to equilibrate. Small-diameter fuels (grasses, foliage, twigs) are
considered to be “⬍1-hour” fuels. Larger diameter branches are “1–10 hour” fuels.
Timber is rated 10–100 hours.
The moisture content of fuels is of more interest to the wildland fire investigator than
to the structure fire specialist, because it can vary considerably with the environment. The
effects of surface moisture and internal moisture content were treated in a previous chap-
ter on solid fuels. Fuels having excessive surface moisture will, of course, be more resist-
ant to ignition on the first pass of a fire. The heat generated can evaporate the moisture,
however, preparing the fuel for more ready ignition should the fire alter its course and
reburn the same area. The same holds true to a lesser extent for internal moisture con-
tent. This is water trapped within the plant (fuel moisture) that reaches equilibrium with
the humidity of the surrounding air through the life processes of the plant’s living cells.
Dramatic changes in the temperature and humidity in the surrounding air can affect this
cellular water, only at a much slower rate than the surface moisture.
The passage of fire or even hot gases can, however, reduce the moisture content of
leaves, brush, and needles to materially increase their combustibility in a short time. The
relationship of atmospheric humidity to fuel moisture content was examined with regard
to nonliving fuels in Chapter 3. The relationship is more complex with living cellulosic
materials, since wind and sun exposure can accelerate moisture losses. Fuels that are dif-
ficult to ignite with “weak” sources such as cigarettes in still air at moderate temperatures
and humidity become much more susceptible with increases in atmospheric temperature
and wind speed or decreases in humidity.8 Weather and exposure conditions in the area
of origin just prior to the start of the fire are therefore very important data to have.

INTENSITY OF WILDLAND FIRE

The intensity of a wildland fire is often measured in kilowatts per meter of a fire line. This
intensity can range from 40 kW/m for a smoldering fire in duff to 100–800 kW/m for
backing fires in surface or ladder fuels, 200–15,000 kW/m for advancing “head” fires,
Chapter 8 Wildland Fires and Their Investigation 331
 and up to 30,000 kW/m for a single-front crown fire in timber.9 Crown fires have been
measured to produce peak incident heat fluxes of 40–150 kW/m2 at distances of 10 m
(33 ft) from experimental fires in stands of 10-m tall jack pine–black spruce trees.10 The
durations of such peak exposures were typically 30 to 90 seconds.

Fire Behavior
Virtually all wildland and grass fires begin as a small source of flame or smoldering fire—
match, cigarette, electrical arcs, hot fragments or embers, lightning, and the like—that
kindles easily ignitable fuels in the area. As the small fire aimlessly explores sources of fuel
its growth and direction of travel will depend on availability of fuel, direction and
strength of wind, and the slope of the surrounding terrain. A small fire burning through
uniformly distributed fuel on level ground will grow slowly in every direction in which
fuel is available if there is no external wind, as illustrated in Figure 8-1a. Hot gases gen-
erated by the fire rise at the center of the fire and draw in air from all directions along the
ground (as shown in Chapter 3). The fire must then “back” into surrounding fuel against
its own induced draft, and its growth rate is limited as a result.
If this same small fire and fuel are arranged on a slope, convective entrainment draws
more air from the bottom of the incline, and the spread of the fire is then enhanced by both
this draft and the exposure of more fuel to hot gases and flames on the uphill side of the
fire. The fire progresses more rapidly up the slope and travels predominantly in this direc-
tion, following the available fuel. This gives the fire direction, and the head of this advanc-
ing flame front grows in energy as it moves into more and more fuel. The draft caused by
the fire grows in strength as the fire grows, thus perpetuating the development in a chain
reaction. Fire winds are those induced by the buoyancy or convective entrainment of the
fire plume. In a large wildland fire, this draft or entrainment effect is great, and a very
strong wind is created by the fire itself. In a fully developed fire, this can become a
firestorm of superheated gases and flames igniting all fuels in its path. The fire not only
spreads uphill in a fan-shaped pattern as a result of these dynamics, it can progress down-
hill as well (by radiant heating or direct flame contact). If sufficient suitable fuel is avail-
able, the fire will back slowly down the slope, even against a fire-caused draft. Burning

Head
Advancing
Wind 30 km/h

Flank Flank
Lateral
Upslope

Wind 45 km/h

Backing

Heel

(a) (b) (c)

FIGURE 8-1 (a) Typical fire spread in uniform ground cover with no wind on level ground. (b) Fire spread in
uniform fuel on moderate slope with no prevailing wind. Although some spot fires may occur downslope from
the origin, they will predominantly be upslope. (c) The higher the wind velocity, the more elongated the fire area
becomes and the closer the origin to the upwind heel.

332 Chapter 8 Wildland Fires and Their Investigation

embers, falling timber, and loose, burning debris can also fall downhill, starting secondary
fires, sometimes at a significant distance from the actual fire (see Figure 8-1b).

EFFECT OF WIND
Natural winds driven by high- and low-pressure atmospheric areas are called
meteorological winds. Any prevailing atmospheric wind will affect the progress of the
fire, particularly during its erratic, early development stages. It will force the flames and
hot gases in one direction or spread an uphill fire sideways across the slope, modifying its
fan-shaped pattern of general development. The higher the wind speed, the more narrow
the elliptical burn pattern becomes, with the origin becoming closer to the downwind end
(as shown in Figure 8-1c).11 Very strong winds can force a fire downhill or cause a fire to
reburn previously flashed-over areas where unburned fuel remains. An external wind will
have the greatest effect on the downwind edges and tops of the forest, as it is quickly dis-
sipated by the porous canopy.12
Warm, dry winds flowing across a mountain ridge can produce very strong downslope
winds on the leeward side (especially in valleys and passes). These are sometimes called
foehn winds (referred to locally as Chinook or Santa Ana winds) and have been responsible
for many major firestorms.13 Diurnal winds are caused by daytime solar heating and night-
time cooling. They flow upslope during sunny days and downslope after sunset.

EFFECT OF TALL FUELS

The growth of fires in tall fuels (⬎2 m; 7 ft) becomes more complex than growth in struc-
tures fires because the fire grows through a porous array of small fuel elements—leaves,
needles, twigs—rather than across a simple surface. As the buoyant (upward) flow and
wind-driven flow carries flames through the foliage the fire can quickly grow to enormous
intensity, the process called crowning. Fully developed crown fires in timber can reach
heat release rates of 30 MW/m of fire line.14 This intensity causes remote ignition by radi-
ant heating and generation of wind-driven flaming brands.
■ These slow-moving (essentially backing) fires are low in energy and often leave a lot
of unburned or slightly burned material on the ground and on vertically arranged
fuels above the average flame heights. These lightly burned materials are sometimes
referred to as color. This phenomenon is commonly seen in an origin area, with the
fire starting with low intensity at ground level and then transitioning into higher
fuels as it moves away from its origin.

OTHER EFFECTS
■ Ambient wind will modify the pattern by adding an additional spreading compo-
nent, so that the fan-shaped pattern on the hillside will deflect to one direction or
the other, and a predominant direction of travel will be created on level ground.
■ In the circumstance of a strong downhill wind, the fire will burn downhill only to the
degree that the ambient wind can overcome the fire’s own tendency to burn uphill.
■ In a fire having an extended perimeter, the direction of burning may vary locally in
almost any direction, depending on the interdependence of the fuel, topography of
the terrain, the air currents created by the fire itself, and the ambient wind.
■ High winds tend to produce long, narrow burn patterns.

Determination of Origin
The basic premise of wildland fire investigation is the same as that for structure fires—
determine the area of origin first, then establish what fuel was present and what source
of heat ignited it. Once these factors are identified, the investigation can then focus on
Chapter 8 Wildland Fires and Their Investigation 333
 what actions and series of events brought the ignition source and first fuel together to
cause the fire and who the responsible party was (i.e., was the fire human caused?).
Because terrain, weather conditions, and fuel distribution all play important roles in wild-
land fire behavior, those must be documented as part of the investigation. The determina-
tion of area of origin is generally carried out by examining the fire scene for indicators of
direction of fire travel. Once a pattern of these directional indicators can be established,
it can be used to point back toward the area of origin. Such interpretive signs should be
used strictly as indicators—none are foolproof, none are fail-safe. With the complexity of
exterior fire behavior, modified as it is by weather, fuel distribution, terrain, and suppres-
sion activities, it is the predominant pattern of indicators that must be sought out, not a
single isolated one here or there.

INVESTIGATION METHODOLOGY
Just as with structure fire examination, wildland fire scene examination proceeds in
accordance with a methodology that follows the principles of the scientific method.
■ Receive the assignment: know role and responsibility (usually origin and cause with
report); be aware of civil and criminal outcomes and type of report needed.
■ Prepare for the investigation.
■ After initial survey of scene and interviews, establish most suitable investigation
plan.
■ Obtain necessary personnel and equipment.
■ Conduct the examination.
■ Interview witnesses.
■ Examine scene to collect data.
■ Document investigation via photographs, diagrams, maps, evidence collection.
■ Collect and preserve necessary physical evidence and get analysis as needed.
■ Evaluate and analyze available data, test possible hypotheses, and establish final
conclusions. These conclusions will usually entail establishing origin, cause, mecha-
nism(s) of fire spread, and responsibility for the fire’s ignition.
The most widely accepted protocol for wildland fire investigation today is described
in the NWCG manual Wildfire Origin and Cause Determination Handbook.15 It is the
foundation for the FI 210 course, “Wildfire Origin and Cause Determination,” offered
nationally by a wide variety of state and federal agencies in the United States, Canada,
Australia, and other countries.
Prior to the start of the scene examination, the investigator must be prepared with
knowledge of the local area including topographical maps, people with reliable local
knowledge, and even fire histories of previous events in the area. Prior to scene arrival,
the investigator can record observations (weather conditions, fire and smoke conditions,
equipment, vehicles, and people in the area) and assess available manpower and expertise.
On arrival, the investigator must assure that the suspected area of origin is protected
(from vehicles, foot traffic, retardant or water drops, water runoff). Interviews with wit-
nesses are a critical first step to minimize wasted search time.

FIRST EVALUATION
It may seem to be a daunting task to seek out a small area of origin, say of a few square
meters in a whole forest of burned trees and brush, which may be hundreds or thousands
of hectares (acres) in size. The wildlands fire investigator is no more interested in the
damage wrought by the later stages of a major fire than is the structure fire specialist
concerned about buildings or vehicles burned by later stages of a large structure fire.
Almost every wildland fire will have been observed in the early stages of its development
by someone—camper, sightseer, hiker, ranger, pilot, or firefighter. Although these people
334 Chapter 8 Wildland Fires and Their Investigation
may not be easy to find, it will be much easier than spending days trekking over the later-
burned areas spotting indicators of fire direction. All subsequent investigation is then
limited to the general locality that witnesses indicate was first burned regardless of how
far it spread later. The investigation should follow a logical scheme that allows thorough
and accurate determination while preserving the evidence remaining (not to mention the
investigator’s time). Several of the first firefighters present should be questioned as to
wind direction and weather conditions, as well as to the extent, location, and behavior
of the fire. First-in fire companies are often aware of where the fire appears to have
begun. They are often trained to protect such areas (which may have already burned out)
with flags or barrier tape.16 Interviews with local residents or law enforcement witnesses
may help establish an area of origin or identify unusual vehicular or pedestrian traffic or
equipment use in the area.

OTHER SOURCES OF INFORMATION

Infrared photography or video imaging from aircraft has been used to map wildland fires
and detect hot spots within a fire area. If these are taken over a time interval, the progres-
sion of the fire can be retraced via these images. Images from earth-resources satellites can
show smoke movement as well as thermal images. These images, when combined with
topographic and weather maps of the fire area, may permit retracing the fire’s develop-
ment back toward its origin and eliminating certain mechanisms of fire spread.
Establishing multiple origins that are separate in place but coincidental in time and that
could not be the result of natural fire progression by airborne embers would go a long
way to demonstrating an incendiary origin for those fires. In the Oakland Hills fire of
1991, infrared imaging from a C130 aircraft and mapping via global positioning satellite
(GPS) and transponders allowed the rapid assessment of the fire’s extension (through fog,
smoke, and darkness) and its damage.17 American Defense Support Program (DSP) satel-
lites today offer a 10-second “revisit” rate of a target area, allowing for minute-by-minute
detection and monitoring of wildland fire events.18 The use of remote-controlled surveil-
lance drones is also reportedly being evaluated by DSP researchers.

THE SCENE SEARCH

The scene search involves a perimeter search and then a more focused search of areas
closer to the apparent area of origin.
Perimeter Search
First, the general area of the first-observed fire is surveyed to determine the outer perime-
ter of the search area. The general shape of the burn pattern often helps investigators
focus on a likely area of origin. Depending on the nature of the terrain and the extent of
the fire, a fan-shaped or V-shaped pattern may be visible from a short distance away. The
apex of this pattern can then be selected as the focal point for the search pattern (see
Figure 8-2). A barricade of rope, twine, or plastic banner is arranged to control access and
limit further damage to the physical evidence from personnel, vehicles, or the heavy
equipment so common to wildland fires. The general area is mapped out on a rough
sketch; a topographic map showing hills, valleys, and similar features of the area may be
very helpful at this time. This general origin area can be identified through witness state-
ments, fire behavior observations, or gross indicators of direction. The outer portions of
the secured area (away from the apparent origin) are searched by whatever systematic
pattern best suits the terrain. These early sweeps are for general indicators, observing the
larger, more obvious signs (sometimes called macro-scale indicators, to be described
later), if possible, working from one side or flank of the burn pattern to the other.
Indicators can then be mapped onto a drawing of the scene as they are found. It is
recommended that the perimeter of this general area of origin be searched twice from
opposite directions.
Chapter 8 Wildland Fires and Their Investigation 335
 FIGURE 8-2 Typical “V”
pattern of burn on hill-
side with suggested Later developing
search pattern. This
pattern starts from out-
side the margins of the
burn pattern and works
progressively toward the
earliest portions of the
fire.

Specific area of origin

(transition area)

Search lane markers

This perimeter search should focus on the macro indicators and examination of adja-
cent unburned areas to establish pre-fire fuel type and distribution. Unless the scene is
very large, this search should be done on foot, but always slowly, methodically, observing
and recording. Photographs should be taken of all significant indicators (macro and
micro). It may be helpful to keep track of this searching process by using visible markers
or flags (color-coded to reflect advancing fire progression, lateral progression, flanking,
backing, and items of evidence). A standardized color flag system has been developed to
help denote fire movement and evidence location, as seen in Table 8-3.
These markers can help keep track of the fire spread and provide a visual reference
of the overall fire scene.19 Keeping a written fire-spread map that corresponds to this flag-
ging system is also a good idea. Remember that the direction determinations you make
are based on the indicators you have found, not on the flags added.
Focused Search
The search can then progress to areas of interest closer to the apparent area of origin.
As the investigator’s search approaches the specific area of origin, there is often a change
in the burn pattern indicators being examined, from the larger-scale indicators to the indi-
cators (called micro-scale indicators) characteristic of the origin area. Note that the focus
is not on a single spot but on a small area of several square feet or yards. Wherever fires
of multiple origin merged, the areas of erratic fire travel around each will remain. These
areas contain the smallest and most delicate indicators and, of course, any remaining phys-
ical evidence. It is always best to minimize all traffic in these areas until an inch-by-inch
examination can be carried out. Burn indicators on larger brush and trees will usually not
be usable unless the fire took on a direction and some intensity. In the area surrounding

TABLE 8-3 Standardized Color Flag System

FLAG COLOR INDICATOR

Red Advancing fire spread
Yellow Lateral (flanking) fire spread
Blue Backing fire spread
Green, white, other Physical evidence

336 Chapter 8 Wildland Fires and Their Investigation

the point of origin, the smaller indicators in this specific origin area may show consider-
able variation as the fire sought fuel and responded to the conditions of wind and topog-
raphy. For this reason, once the area of origin has been located, it is usually necessary for
the examination to be carried out on hands and knees looking at individual pebbles, leaves,
and twigs for the resulting signs, and, ultimately, for the cause of the original ignition.
As the area of transition (to directional spread) is detected, the investigator should area of transition ■
proceed slowly around this area, but doing so from outside the burn itself or the suspected Mixture of fire direc-
tional indicators in a
area of origin. It is best to do the hands-and-knees search working in from the advancing
wildlands fire.
fire area, continuing to read the micro-scale indicators. The most common method is to
grid or lane this area off, using twine on stakes. Because the transition from random to
directional (driven by wind, fuel or terrain) may be very subtle, it is easier to overrun the
origin if it is approached from the backing or heel area; approaching from the advancing
area is recommended. Each lane or grid can then be searched visually, with magnification
and a magnet. The detail search is often carried out using a straightedge (ruler) as in
Figure 8-3 to help focus the visual examination. A handheld or headband magnifier or
2X–3X magnifying glasses are often useful at this point. The debris is then swept with a
handheld magnet in a plastic bag to collect iron-containing debris. A sensitive metal detec-
tor may be swept over the suspected area to locate metal vehicle or tool components, or
ignition devices. A small kitchen baster, photographer’s bulb, or blowing by mouth can
be used to direct small puffs of air to move surface debris and light ash. Heavier covering
debris can be moved using a soft paint brush, tweezers, or long forceps. If the ash layer
is deep, it will need to be removed by gently blowing it away after the first search, possi-
bly revealing additional indicators beneath.

FIGURE 8-3 Detail

search for micro-scale
indicators using a
straightedge. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

Chapter 8 Wildland Fires and Their Investigation 337

 FIGURE 8-4 Char depth
will be deeper on the side
(with uniform fuel) facing
the fire.

Fire travel
Char depth
deeper on side
facing fire

Remember that suppression and control efforts can vastly complicate the reconstruc-
tion of a wildland fire. Bulldozers can build firebreaks to divert “normal” progress.
Tanker air drops can cool off areas, allowing nearby areas to burn. Most important, sec-
ondary “backfires” or “burnouts” may be set deliberately to deprive the main fire of fuel.
All these can dramatically affect the appearance of burn indicators. Wind and rain may
quickly compromise many useful indicators, so the investigation should be carried as soon
as it can safely begin.

BURN INDICATORS
Because the effects of the nature of fuels and fire conditions in a wildlands fire are fairly
predictable, there are a number of fairly reliable indicators that can be used to determine
the direction of fire spread past a given point.
■ Charring of a tree trunk, plant stem, or fence post is deeper on the side facing
the oncoming fire. Char depth can be checked with a knife blade, pencil point, ice pick,
or screwdriver. It is strictly a comparative indicator, so absolute depth is of little conse-
quence (Figure 8-4).
■ Destruction of a bush or tree may be more extensive on the side facing the fire
(Figure 8-5). Fire will progress upward as it moves forward through the foliage, leaving
an interface between burned and unburned portions as the buoyancy of hot gases
causes the flame front to rise as it moves forward.
■ The beveling effect of a fast-moving fire may influence the appearance of
branches or twigs remaining upright. Twigs and branches facing the fire may have flat
or rounded ends, while those facing away from it (downwind) will be tapered or
pointed, as shown in Figure 8-6.

Remaining leaves or needles may be “frozen”

in position in the direction of the prevailing wind

FIGURE 8-5 Partial

combustion of trees and
bushes can reveal the fire
spread. Fire will progress
upward as it moves for-
ward through the foliage,
leaving an angled inter-
Fire spread Fire spread
face between burned and
unburned portions. Bare branches
Remaining leaves or
needles may be frozen in
position in the direction
of the prevailing wind. or less commonly

338 Chapter 8 Wildland Fires and Their Investigation

Fire spread
Fire spread Fire spread

FIGURE 8-7 In a fast-moving fire, the fire pattern on a tree will start
FIGURE 8-6 Beveling effects of oncoming fire lower on the trunk side facing the oncoming fire, even on a slope.
blunt the ends of branches facing the fire and The charred area may wrap completely around and extend up the
taper those away from it. “back” side. (See Figure 8-8).

■ A fast-moving fire creates a draft around large objects, which creates an angled
pattern around tree trunks, posts, plant stems, and the like (Figure 8-7). On foliage crowns,
this angle of fire pattern may also be evident when the tree or burn is viewed from the side.
It may reflect the fire behavior of the flames “coming in low and exiting high.”
■ Fast-moving (either fire- or wind-driven) fires moving past large-diameter vertical
tree trunks or power poles create a low-pressure area on the downwind side. This low-
pressure area draws the passing flames from local fuels into a vertical vortex, creating a
laminar flame. The laminar flame lasts longer than the passing flame front and can induce
significant fire damage to the downwind side, extending several times the height of the
flame front.20 (See Figure 8-8.) A very slow moving fire, especially one that is “backing”

FIGURE 8-8 This fire

(moving from right to
left) induced a low-
pressure zone on the
leeward side of the center
tree trunk which caused
the formation of a lami-
nar flame entraining the
flames in the ground
cover at its base. This
flame lasted longer than
the passing flame front
and produced much
higher burn damage on
the downwind side.
Used with permission of
Donna Deaton, Deaton
Investigation, LLC.

Chapter 8 Wildland Fires and Their Investigation 339

 Fire spread

Fire

FIGURE 8-9A In a slow-moving or “backing” fire, the burn pattern

will be roughly parallel to the ground. Wind-driven patterns often FIGURE 8-9B Observed low-intensity test fire moving from
exhibit a rising char pattern on the leeward side. Accumulations of right to left in light ground fuel with a light (5 mph) wind. Char
ground litter may produce char patterns on the upslope side of tree pattern is higher on the leeward (downwind) side of the trunks.
trunks and utility poles. Courtesy of Marion Matthews U.S. Forest Service.

against the wind or down a slope, will create a burn pattern approximately parallel to the
ground (as depicted in Figure 8-9a). Note that such patterns can be influenced by local fuel
load such as needles, leaves, and debris around the base of the tree, leaving a pattern higher
on the side facing the oncoming fire or the upslope side (which catches such debris). This
may also be the result of momentum flow of the fire against the vertical surface. Wind-
driven (either from fire plume convection or atmospheric wind) fires often produce a rising
char pattern on the leeward side of the trunk or pole, as in Figure 8-9b.
■ If the vertical stem is burned away, the remaining stump will be beveled or
cupped on the side facing the fire, as shown in Figure 8-10. This will be true even for
stems of weeds. When the back of the hand is brushed lightly against such stubble, the
beveled tops allow the skin to pass smoothly across in the same direction as the fire but
resist with sharp points a hand passing in the opposite direction.

FIGURE 8-10 Beveling Fire spread

or cupping on vertical
stems or stumps will
point back in the direc-
tion of the fire’s origin.

340 Chapter 8 Wildland Fires and Their Investigation

FIGURE 8-11A Deposits left on upper left sides of large rocks as FIGURE 8-11B Micro-scale indicator deposits on small stone amid
a result of observed fire spread (through coarse grass) from top to burned grass confirm fire was moving from right to left. Used with
bottom. Courtesy of Marion Matthews U.S. Forest Service. permission of Donna Deaton, Deaton Investigations, LLC.

■ Rocks, signs, and other noncombustibles will show greater heat discoloration (stain-
ing) on the side facing the fire, as pictured in Figures 8-11a and b. Lichen, moss, and close-
growing grass may survive on the side away from the fire. Extensive exposure may burn all
soot off (leaving a light-colored or white surface) on the side facing the oncoming fire.
■ Large rocks, walls, and large objects provide a barrier to the flow of the flames.
The area downwind of the obstacle may escape damage completely or be burned later
as the fire changes direction (Figure 8-12). Intense flame contact may cause spalling of
stone or concrete on the upwind or facing side.
■ Tall weeds and grasses, when burned by a slow-moving fire (particularly as it
spreads laterally along the flanks or backs downslope or against the wind), will be
undercut by the fire moving along the ground. The stems, if vertical, will then fall
toward the fire with the heads of grass stems pointing back toward the fire origin, as in
Figures 8-13a and b. This effect is highly dependent on the wind conditions prior to
and during the fire. Tall grass that is already matted down pointing away from the fire
will fall in that direction, so there may be conflicting directions from such indicators.
■ The height of remaining stems and grasses is roughly proportional to the speed
of the fire. The effect is dependent on the moisture content of the plants and will not be
visible in areas that have reburned.
■ In the immediate vicinity of the origin, the fire will not have developed any par-
ticular direction, and so the indicators will be variable or contradictory. Fuels in the

Fire spread

Rock
wall
FIGURE 8-12 The flow
of a fast-moving fire
around an obstacle often
leaves protected vegeta-
tion on the downwind or
“lee” side of the obstacle.
Unburned “protected” grasses

Chapter 8 Wildland Fires and Their Investigation 341

 Fire travel

FIGURE 8-13A Tall stems of grass or weeds

point predominantly back toward the source of
the fire in the absence of a strong prevailing FIGURE 8-13B Fire is backing to left (into wind). Note the long stems have burned off
wind. at the base and stems have fallen to the right. Courtesy of John D. DeHaan.

form of grasses or weeds may still be upright or only partially burned. Large fuel may
be burned only at its base with fire extending into upper branches only as the fire
moves away from its origin.
spot fires ■ Fires ■ Extension of a fire beyond a barrier, for example, a road or river, will cause the
started by airborne
embers some distance
appearance of a brand new origin. This is particularly true when the “new” fire is
away from the main started by airborne embers or burning debris (brands) from an established fire. (Such
body of the fire. new fires are sometimes called spot fires; see Figure 8-14.)
■ By their quantity and distribution, ashes can indicate the direction of wind and
the amount and nature of fuels present.
■ If the fire is very recent, the white ash formed by the fire’s impact on the facing
side (with more complete combustion) can be compared with the darker area of less
intense fire on the leeward side. Viewing the overall fire scene from the origin may reveal
the white ash deposits on exposed surfaces. Wind or rain will remove such deposits.
It should be noted that most of these indicators (particularly the micro-scale ones)
require careful systematic comparison of one side of an object with the other, not an
assessment of absolute condition. The investigator should also be aware of ventilation

Fire travel

River
FIGURE 8-14 Remote
spot fires caused by air-
borne embers across nat-
ural barriers. Wind Remote spot fires caused by airborne
embers across natural barriers

342 Chapter 8 Wildland Fires and Their Investigation

 FIGURE 8-15 Diagram
of fire scene search show-
ing indicators for fire
spread and evidence.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

and venturi effects in high-wind conditions or around/under objects not in full contact
with the ground, as they may produce contradictory indicators. Because “wind” condi-
tions change as the fire grows or responds to slope and terrain, the investigator must con-
sider the totality of indicators observed before concluding direction of fire spread.
Reliance on a single indicator is not good practice, as it may be misleading.
Once the area has been thoroughly examined, with direction indicators carefully
flagged and noted on the sketches, as shown in Figure 8-15, it is often beneficial for the
investigator to stop for a few moments and reevaluate what has been found. Questions
such as the following can be helpful:
■ What patterns do the indicators reflect?
■ What features of the terrain, weather, or the fuel load could affect the development
of those patterns?
■ Do there appear to be multiple origins that grew into one large fire?
■ What area of origin appears to be consistent with those patterns?
■ What firefighting tactics were used in the area that could have produced patterns of
their own—backfires, or retardant or water drops.
A few minutes’ reflection may reveal disparities in the indicators that were not appar-
ent during the actual search. Resolving those disparities at this time will avoid having to
repeat tiresome, painstaking, inch-by-inch examination on more than one indicated area
of origin. Remember, until the fire develops direction as a result of wind, fuel, or terrain,
it will often extend erratically in several directions, producing a pattern of contradictory
direction indicators. This specific origin area is worthy of close examination as a possible
point of origin. Once the questions have been resolved, one can then proceed to careful
examination of the fire indicators on a micro scale in the vicinity of the actual ignition
source. This allows documentation (by sketching and photography) of even the smallest
details while minimizing the risk of destroying such details by kneeling or stepping on
them during the search. Ultimately, the search of the specific area of origin is carried out
for one major purpose—to discover the means of ignition of the first small fire.
Chapter 8 Wildland Fires and Their Investigation 343
 Once the point of origin has been located and the first fuel identified (and possibly an
ignition source detected), the origin needs to be thoroughly documented with photographs. It
is recommended that a sampling of the specific indicators by class and category be selected at
various intervals throughout the general area of origin. These indicators are photographed,
documented in the photo log, and their location measured from a previously selected main
reference point. This information can then be used to draw a to-scale scene diagram that cor-
responds with the photographic record and re-creates the origin determination process. (Refer
to Chapter 7 and Appendix F for information on surveying and documenting large scenes.)

DOCUMENTATION
The recommended minimum documentation for wildland fire investigation should include
a comprehensive narrative report, photographs of burn pattern indicators used, maps,
diagrams, witness statements, weather information, and forensic lab analysis reports (if appli-
cable). Photographs, photographic logs, and diagrams are discussed next.
Photographs
Photographs provide the reader of the report a clear understanding of what was seen, done,
and used, and preserve critical data for review later by the investigator.
Photograph log
A photo log is essential for describing the subject of each photo and direction in which the
photo was taken. This may be in the form of a scene diagram with symbols (numbered to
show frame [image] number, or negative/roll number) with an arrow to show direction, or
a numbered list.
Diagrams
Diagrams show the location of indicators used, witnesses, and evidence collected, as illus-
trated in Figure 8-15. They must include date, location, compass, initials of preparer, and
documentation of dimensions and topographic features.21 Measurements can be in one of
three formats, as shown in Figure 8-16. GPS coordinates can be used to establish locations

FIGURE 8-16A Right-angle transect: FIGURE 8-16B Azimuth/baseline:

baselines at right angles with end- single baseline (in N–S or E–W orien-
points set by GPS and compass. tation) (set via compass and/or GPS).
Baselines are marked by string, paint, Locations noted by angle and distance FIGURE 8-16C Intersecting arcs from two fixed
or metal tape. from baseline or reference point. reference points.

FIGURE 8-16 Methods for baseline reference at wildland fire scenes.

344 Chapter 8 Wildland Fires and Their Investigation

of evidence and possible areas of origin. These coordinates can be used to create reports and
diagrams and to confirm jurisdiction.
There are several computer-aided drawing systems designed for use in wildland fire
investigations. These include standardized symbols and plotting methods.

Sources of Ignition
The next step in the investigation of the wildland fire is to determine the cause when pos-
sible. In some instances this may be impossible because the evidence has been destroyed.
It is always necessary to make a reasonable search for the physical remains of the ignition
source and to establish a defensible cause. Ascribing such cause to “unknown factors”
should be only a last resort. The causes of outdoor fires are well known in general. The
National Wildfire Coordinating Group (NWCG) maintains records regarding fire cause
for all wildland fires in the United States. Statistically, human-caused fires account for
approximately 80 percent of all wildfires annually.22 The California Department of
Forestry reports that in over 12,000 wildland fires over a three-year period, in the aver-
age year, the leading identified causes (in descending order of occurrence) were equipment
use (tractors, chain saws, weed cutters, etc.), debris burning, arson, vehicles (trucks and
cars), lightning, smoking, campfire, power line, playing with fire, and railroad.23 (See
Table 8-1.) The following are examples of some of the possible causes:
■ Unextinguished fires (including discarded unextinguished coals) left by campers,
hunters, and others.
■ Carelessness with smoking materials, including burning tobacco and matches.
■ Trash burning and, in some areas, controlled grass and brush burning (including
airborne firebrands).
■ Sparks (hot exhaust particles) from vehicles, especially locomotives or other motor-
driven equipment.
■ Lightning, a major cause of timber fires.
■ Power transmission lines and equipment, including birds and mammals that may come
into contact with the power, short-circuit it, and catch fire, falling into nearby fuel.
■ Arsonists, who may use a variety of devices but generally use a lighter or match.
■ Firearms, which under certain circumstances may blow sparks of burning powder
into dry vegetation, and steel-core and tracer bullets.
■ Spontaneous combustion, limited to very specific types of circumstances. (See
Chapter 6.)
■ Overheated machinery, which may be in contact with combustibles.
■ Sparks from any source, such as impacts of metal with rock.
■ Discharges of static electricity.
■ Rekindling of smoldering embers from previous controlled fires, whether they were rekindle ■ Reignition
barbecues, campfires, or previously burned timber slash or debris piles, or small of a fire due to latent
(possibly unreported) brush fires. These are sometimes referred to as holdovers. heat, sparks, or embers.
These have been reported to have time frames of up to and exceeding 12 months.24
■ Sleepers, which are old trees or stumps containing much rotted wood internally in
which a smoldering fire has been previously induced. Because of their unique burn-
ing characteristics, fire may persist in them for long periods, documented in some
instances to be as much as nine months.
■ Miscellaneous objects sometimes found in trash, including clear glass (cylindrical or
lens-shaped in cross section) or concave reflective surfaces that can focus the sun’s
rays. (See Chapter 6.) These are rarely responsible for wildland fires but should be
considered.
The examination of ground cover for ignition sources can be very tedious, but it is
essential. The searcher is encouraged to keep an open mind, look for things that do not
Chapter 8 Wildland Fires and Their Investigation 345
 belong, and take frequent breaks. Circumstances and proximity of roads, trails, railroads,
or signs of recent equipment use (freshly cut weeds, recent gouges in road or stone sur-
faces) are useful in some cases. A second search from a different direction or a second
examination is recommended.
The mechanisms of most of these sources were discussed in Chapters 5 and 6, and the
reader is directed there for further information.
The nature of the fuel at the origin and the relevant characteristics of the proposed
ignition source must be evaluated together. Because most wildland fires involve cellulosic
fuels, either open flames or hot (glowing) sources must be considered as competent.

POWER LINES
Power lines generate fires often enough so that the mere presence of such a line within or
close to a fire area renders it suspect as the cause of the fire. There are several methods
by which power lines may initiate a fire, including the following:
■ Transformer short-circuits and malfunctions—perhaps one of the more common
and dangerous situations, causing overheated oil or molten metal droplets to fall
into ignitable fuel
■ Leakage over dirty insulators and supports when moist, giving rise to “pole top” fires
■ Fallen wires that arc with the ground or objects on it
■ Arcing between conductors that accidentally come into contact (typically in high
winds)
■ Grounding of in-place conductors by external objects such as fallen trees, branches,
or birds
■ Unprotected fuses dropping hot metal
■ Tree limbs’ coming into contact with conductors
■ Animals’ coming into contact with equipment and shorting it out (see Figure 6-34)
The mere presence of a power line in the neighborhood of a fire should never be taken
as the cause without a thorough investigation. It is often easier to blame the obvious than
to prove the matter. Despite the inherent hazard of power lines as a source of ignition,
some fires may have been blamed on them when careful analytical investigation would
have revealed the fire to have been started by other means.
The wildland fire investigator should be familiar with basic power pole equipment
such as transformers, fuses, breakers, and switches, as shown in Figure 8-17, and how they
fail. High-voltage conductors that come into contact with soil can produce glassy fulgu-
rites, which are produced by the high-temperature fusion of minerals in the soil (see Figure
8-18). They will be found just below the surface of the soil by detail search and sifting.

LIGHTNING
Lightning is another common cause of timber and grass fires. Lightning strikes can shat-
ter or explode trees or limbs, peel bark from trees, split rock, or fuse soil or sand into
glassy fulgurites at the point of contact (similar to those in Figure 8-18 but much larger
in size.) Such signs should be sought out during the search for the fire origin. In addition,
evidence of lightning activity may be available from witnesses. The mechanism of light-
ning was discussed in full in Chapter 6. The tremendous heat generated by the passage of
lightning through wood can ignite even massive timbers under some circumstances.
Reports of lightning in an area can be compared against lightning data accumulated by
the National Oceanographic and Atmospheric Administration (NOAA), which operates
a National Lightning Detection Network reporting system. There is also a private firm,
Vaisala Inc., that employs more than 100 sensing stations in the lower 48 United States
and a global positioning system. This network can locate lightning strikes (to the earth)
to an accuracy of 500 m (1,500 ft). A report recording the time, intensity, polarity, and
346 Chapter 8 Wildland Fires and Their Investigation
 FIGURE 8-17 Investi-
gators must be familiar
with pole-mounted elec-
trical equipment that can
start fires or yield clues as
to the start, such as this
transformer with light-
ning arresters (arrows).
Courtesy Pacific Gas & Electric.

FIGURE 8-18 Small ful-

gurites of soil fused into a
glassy mass by a ground
strike of a high-voltage
power line. Larger ones
can be produced by light-
ning strikes. Courtesy of
Marion Matthews, U.S. Forest
Service.

Chapter 8 Wildland Fires and Their Investigation 347

FIGURE 8-19 StrikeNet
map of lightning strikes in
area, color-coded to dif-
ferentiate those between
0200-0300 and 0300-
0400. Courtesy of Vaisala,
Inc., Tucson, AZ.

location of all strikes for a particular location for a requested time period is available
online for a fee.25 See Figure 8-19 for an example. This service is reportedly 90 percent
accurate (multiple reports can sometimes reveal a strike not reported in a single report).
Lightning can also strike wire or metal fences, buried pipes or cables, and underground
storage tanks, causing direct ignition or ignition remote from the actual strike. See Figure
8-20 for an example of a lightning strike on a tree.

FIGURE 8-20 Lightning

striking this tree stripped
its bark and charred sur-
face grass along its roots.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

348 Chapter 8 Wildland Fires and Their Investigation

BURNING OR HOT FRAGMENTS
Fragments from a disintegrating catalytic converter are small and may escape detection, but
they may have a characteristic ceramic “honeycomb” appearance in which the cells may be
square, rectangular, or hexagonal in shape. Small spherical pellets are also often used. The
rare metals used in them, such as platinum or iridium, can be readily identified in the labo-
ratory. Fragments or “sparks” of burning carbon from the eductor tubes of diesel locomo-
tive engines are nondescript in appearance but tend to be dense and may “burrow” through
layers of dry leaves or grass before igniting them. Some exhaust fragments (especially those
from gasoline engines) contain iron residues that can permit their recovery using a strong
permanent magnet.26 Fragments of automobile or railroad equipment and welding slag will
also be recovered with a magnet. See Chapter 6 for a full discussion of such sources.

CAMPFIRES
Campfires can ignite wildland fires through direct flame contact, aerial sparks (embers),
improper disposal or covering of ashes and coals, or even direct ignition of hidden pock-
ets of combustible duff or buried roots. Ignition is usually close to the campfire site, and
proof of mechanism and improper care or extinguishment will be an important element
of the case. Shoe prints or food containers may reveal recent use. Hot wood or charcoal
coals can retain their heat for extended periods of time (1–2 days) and may be buried or
discarded away from the main campsite.

CIGARETTES
Cigarettes discarded while still smoldering represent a competent ignition source for
grasses under some conditions. The relative humidity has to be low (⬍22 percent) and
ambient temperature over 26°C (80°F) to produce the low fuel moisture necessary.27 The
lower the humidity, the lower the fuel moisture and the more likely that ignition will
occur. The fuel has to be loose enough to permit rapid burrowing of the burning cigarette
and good contact between the exposed coal and suitable fuel. Surrounding fuel has to be
sufficiently dense to sustain the fire spread. Ignition is more likely if the cigarette lands
coal-end down or facing into the wind.

INCENDIARY FIRES
Incendiarism cannot be overlooked as a significant cause of wildland fires, especially in
rural and urban-wildland interface areas. As in the case of structure fires, all possible
causes must be considered. During a thorough search through a fire scene, obvious signs
of recent activity should be noted. Cans, bottles, tire and shoe impressions, weapons, and
mechanical delay devices may all survive even the most awesome of conflagrations. An
eye should be kept open for these during all phases of a scene search. Ignition devices are
probably more common in wildland arsons, where one may need many more minutes to
flee the area than in urban structure arsons. Unless the fire has been set with a pocket
lighter, the remains of the ignition device often are recoverable from the point of origin.
Even single paper matches have been found by diligent, trained searchers. Cigarettes have
rarely been seen to ignite grasses or brush when the relative humidity of the surrounding
air was above 22 percent. Relative humidity lower than 22 percent can produce fuel mois-
tures below 14 to 15 percent, at which the low heat release rate of a smoldering cigarette
can ignite fine fuels such as grasses and needles (duff).28 (See Chapter 6 for the relation-
ship of humidity to fuel moisture.) Weather data for several days prior to the fire are crit-
ical for proper testing of hypothesis involving any wildland ignition source.
Multiple points of origin are certainly strong indicators of incendiarism, although
there are accidental circumstances that could account for such fires. An overheated
catalytic converter spitting glowing remnants of its matrix or a stuck brake shoe or failed
Chapter 8 Wildland Fires and Their Investigation 349
 bearing on a railroad car spewing bits of hot metal over some distances are examples.
Glowing particles of carbon from diesel locomotive engines can be thrown some distance
from their moving source to ignite multiple fires along the right-of-way. Witnesses may
see such small fires along the right-of-way following the passage of railroad equipment.
Because readily ignitable fuels are present in such quantities in wildland, the use of a
flammable liquid such as gasoline or kerosene as an accelerant is uncommon. Unusual odors
or the appearance of trailers between points of origin should always be carefully examined.
Devices may be used to ignite remote spot fires, to delay detection and suppression,
or to delay ignition of the set itself. The most effective remote ignition sets consist of a
bundle of matches wired or taped to a cigarette. A rock or steel washer may be added for
additional weight if the device is to be thrown any distance. Once ignited, the device can
be thrown from a passing car deep into a brush-covered ravine or similar inaccessible
location. Many such devices can be set in a short period of time. Some will not land on
suitable fuel, but those that do can ignite serious fires with little evidence with which to
link a suspect to the crime. Searches of nearby roadsides in the vicinity of a fire have
revealed similar “serial” devices that were unsuccessful and therefore undamaged by
ensuing fire. Such devices may bear fingerprints and trace evidence that can be associated
with a suspect. Shoe prints and tire prints from such locations have successfully linked a
suspect to a general fire area. Wrapping a rock with paper, tissues, or cloth and throwing
the ignited bundle into brush can also kindle fires with little traceable evidence remaining.
Incendiary devices are used by fire services for starting backfires. These can range
from commercially made pyrotechnic devices with a cardboard cylinder and a fuse to
table-tennis balls filled with gasoline/diesel mixture. They have occasionally been stolen or
misappropriated from fire agencies. Kerosene-fueled drip torches (as shown in Figure 8-21)
are often used to start intentional backfires but have also been misused.

FIGURE 8-21 Hand-

carried kerosene drip
torch as used by wildland
fire fighters to start back-
fires as a firefighting
tactic. Courtesy of John
D. DeHaan.

350 Chapter 8 Wildland Fires and Their Investigation

 FIGURE 8-22B Fragile remains of such a set after the passage of
FIGURE 8-22A Cigarette and the match delay of a type often the fire should be carefully photographed before any attempt is
found in wildland fires. Courtesy of Lamont “Monty” McGill, Fire/Explosion made to move them. Courtesy of Lamont “Monty” McGill, Fire/Explosion
Investigator. Investigator.

Elaborate time-delay devices are very rare in wildland fire arsons. Concave shaving
mirrors have been used to start remote fires, sometimes with more than a year’s pass-
ing before the focused rays by chance fell onto suitable fuel. Other time delays may
have been used and have escaped detection and analysis, just as in the case of structural
arson sets. Arsonists will generally stay with a device for an entire season if it proves
successful, thus creating a signature for linking incidents together and eliminating copy-
cat scenes.
As mentioned previously, the most common time-delay device is a bundle of matches
or a matchbook affixed to a burning cigarette, as shown in Figures 8-22a and b. When
such a device is found, it is vital that it not be disturbed until it has been fully documented
via appropriate sketches and photographs. The appearance and length of ash can indicate
if a cigarette was being smoked and discarded, whether it was burned in place by an
encroaching fire, or if it was ignited and left deliberately (with malice). The brand of cig-
arette and the conditions of burning have a great effect on burning time. Tests by the
California Department of Forestry demonstrated burn rates of 4–8 mm/min (0.16–0.33
in./min) for various cigarettes burning horizontally in still air. These rates were faster
when the cigarette was burning vertically upward and when burning in the direction of
an external wind. Total burn times were on the order of 10 to 15 minutes.29 The brand
of cigarette can often be determined if the butt end (with or without a filter) remains
unburned. Bourhill’s manual for such identifications remains a unique resource.30 The
introduction of “reduced ignition propensity” cigarettes is likely to reduce the number of
fires ignited accidentally or deliberately with cigarettes but not to eliminate them from
consideration.
Devices that do not provide time-delay ignition are sometimes called hot sets. The hot set ■ Direct ignition
direct application of flame from a match or lighter to available fuels is almost certainly of available fuels with
an open flame (match
responsible for a large percentage of wildland arsons. Needless to say, when the source of
or lighter).
flame is removed by the arsonist, establishing the cause of the fire will depend on thor-
ough elimination of all other possible causes and extremely careful searching and docu-
mentation (including documentation of sources that were considered and eliminated).31
Hot sets that may leave identifiable remains include dropped matches and cigarettes, “res-
cue” flares, flaming arrows, or lighted road flares or fusees. The remains of matches and
Chapter 8 Wildland Fires and Their Investigation 351
FIGURE 8-23A Top:
Intact debris from a road
flare (fusee). The hollow
cardboard end cap is at
lower right; the plastic
cap is melted and burned.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

FIGURE 8-23B Lower:

After being stepped on,
residues are fragmented
but still distinctive.
Courtesy of John D. DeHaan.

cigarettes tend to stay on top of other ash and debris but are easily destroyed by weather or
persons searching the scene. Road flares leave a cakey, white, or pastel-colored residue of
characteristic appearance (and chemical composition) and usually a wooden end plug and
plastic striker cap. The plug and cap often survive even a major fire if they are protected,
and the length of the original flare can be estimated from the size of the plug and the amount
of residue left behind. Figures 8-23a and b show the characteristically bright-white and
pastel residue of a common emergency road flare used to start a grass fire. The compo-
nents of the flare can vary with manufacturer and age, as pictured in Figures 8-24a and b.
352 Chapter 8 Wildland Fires and Their Investigation
 FIGURE 8-24A Com-
ponents of two varieties
of flares (fusees). The top
one has paper compo-
nents with a bottle cap
for striker support, card-
board end plugs, and a
wooden plug (inside).
Courtesy of John D. DeHaan.

FIGURE 8-24B The

lower one has a plastic
cap and striker (low-
melting-point plastic is
easily destroyed by fire).
Courtesy of John D. DeHaan.

Figure 8-25 shows a suspected ignition device surrounded by barrier tape until it can be
properly documented and recovered.

MODELING
Computer models have been developed for predicting growth and intensity of wildland
fires given known fuel, weather, and topography information. The most commonly used
programs—Behave, Behave Plus, and FARSITE—can generate predictions of large-scale
time lines and behavior patterns.32 Like computer fire models for room fires, they have
been used for hypothesis testing, but they should not be counted on to provide “proof”
Chapter 8 Wildland Fires and Their Investigation 353
FIGURE 8-25 Suspected
ignition device and sus-
pected origin protected
by barrier tape until they
can be photographed.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

of how and where a particular fire was ignited. They await validation for forensic appli-
cation. Given the complexity of wildland fire dynamics, as described earlier, it is not sur-
prising that accurate modeling has been a difficult challenge. Some programs for
two-dimensional, wind-driven spread on grasslands have been validated by fire tests.33
The advent of ever-more powerful computers has aided the development of limited mod-
els of crown fires.34 NIST is also evaluating the Fire Dynamics Simulator (FDS) for use in
predicting fire spread in wildland-urban interface fires where trees and shrubs present the
initial fire-spread mechanism, and structures are ignited by radiant heat from those
sources.35 Sullivan has compiled a comprehensive account of wildland fire modeling in
the years 1990–2007, and the reader is referred there for more detailed information.36

Collection and Preservation of Physical Evidence

The documentation of the appearance of the wildland fire scene and its related evidence
by sketching and photography is every bit as important as it is in structure fires.
Considering the fragile nature of some important evidence and the vagaries of the envi-
ronment, it is all the more surprising that some agencies do not routinely document their
fire evidence in this way. The same general principles outlined in Chapter 7 apply to
sketching and photography of wildland fires, as do the guidelines for collecting and pre-
serving the physical evidence found. Some situations peculiar to wildland fire evidence
should be reemphasized, however.

CIGARETTES, MATCHBOOKS, AND OTHER FRAGILE EVIDENCE

When recovering cigarettes or matchbooks, remember that spraying them with lacquer or
other preservative will generally make it impossible for the laboratory to test for latent
fingerprints or saliva-borne DNA typing results, or to conduct brand identification or
physical comparison tests. It is best to photograph the object in place and then package
it between layers of loosely fluffed cotton wool in a rigid box for hand transmittal to the
laboratory.
One recommended technique for collecting fragile evidence, such as cigarettes or
matchbooks, that permits recovery of parts of evidence that are buried in the ash beneath
is to dig into the ash and underlying soil using a sharp-edged, flat-bladed garden trowel
and loosen the soil under the object in a rectangular area. Once this is done, a clean rec-
tangle of sheet metal (such as sheet metal flashing from the hardware store) is worked
354 Chapter 8 Wildland Fires and Their Investigation
carefully through the soil without disturbing the surface debris. The sheet metal is worked
all the way under the soil platform and then used to lift the soil, ash, and object together.
The entire mass is then transferred to a properly sized and padded evidence box. If the
soil is very hard, water can be introduced to the trenches around the target and allowed
to stand to soften the soil. On rocky surfaces, it may be possible to lift only the ash layer
with the sheet metal.

SHOE AND TIRE IMPRESSIONS

Shoe and tire impressions should be photographed using low-angle oblique lighting and
a ruler or scale near (not in) the impression. (Placing a scale in or across an impression
can obscure critical details.) If at all possible, the impression should then be cast using
plaster, dental stone, paraffin, or other suitable casting agent.

CHARRED MATCHES
Matches that are completely charred are probably not useful as evidence (except to
demonstrate their presence), but if any unburned cardboard remains, an attempt can be
made to match it to a matchbook.

DEBRIS SUSPECTED OF CONTAINING VOLATILES

Debris suspected of containing volatiles should be sealed in appropriately sized cans, jars,
or polyester or nylon bags. Polyethylene plastic or paper bags must not be used.

CONTAINERS
All containers much be securely sealed to prevent loss or contamination, and the chain of
evidence must be recorded on it in ink. The date, time, and location of recovery and the
name of the person recovering the evidence must be on the package as well.

WEATHER DATA
A critical item of evidence that is not found at the scene itself and sometimes overlooked
is weather-related data. This should be obtained as soon as possible, either from first-
on-scene fire suppression personnel, websites such as Weather Underground (www.
wunderground.com) or NOAA, nearby Remote Automated Weather Stations (RAWS), or
collected firsthand using either a handheld digital weather recorder or a belt weather kit
(shown in Figure 8-26). The RAWS system uses remote sensor stations that automatically
send weather data via satellite every hour to the National Weather Service’s Weather

FIGURE 8-26 “Belt”

weather kit: Pitot-type
wind gauge (anemometer),
compass, water bottle,
psychrometer (wet/dry
bulb humidity gauge),
and case. Courtesy of U.S.
Forest Service (photo by
author).

Chapter 8 Wildland Fires and Their Investigation 355

 Information Management System. From there it is available for spot analysis by meteorol-
ogists. Currently, the website www.met.utah.edu/mesowest, operated by the University of
Utah, can give investigators acess to RAWS data.
Weather data should be taken from outside the burn but as close to the suspected ori-
gin as possible and as close to the estimated/known ignition time as possible. Two hand-
held devices that can be used to record temperature, relative (air) humidity, and wind
speed are shown in Figure 8-27. Weather data can provide a critical source of informa-
tion with which to assess fire behavior, rate of spread, and fire intensity through the use
of nomogram fire calculations or computer models. Such analysis can support burn pat-
tern interpretation and scientifically corroborate or exclude possible ignition sources.

(a)

(b)

FIGURE 8-27 Pocket weather tools. (a) Digital thermometer and humidity gauge. (b) Wind-speed indicator
(anemometer). Courtesy of John D. DeHaan.

356 Chapter 8 Wildland Fires and Their Investigation

 CHAPTER REVIEW

Summary
The careful investigator who attempts to locate the other reasonably potential causes excluded or satisfac-
exact cause of a fire in wildland areas must, above all, torily addressed. It is the expert opinion of the investi-
avoid assuming that because some hazard existed in gator on which all further proceedings regarding this
the general area of origin, it was necessarily the cause. fire will depend. If there is any doubt as to the cause
Just as with structure fires, the investigator must be of the fire, it should not be acted on. In civil cases, of
sure that the fuel involved was capable of being ignited course, the standard of proof is a preponderance of the
by the proposed or suspected ignition source. evidence, as opposed to the “beyond a reasonable
Although a fire along a roadside suggests that a care- doubt” standard in criminal cases.
less smoker may have caused it, without finding the Every wildland fire investigator must remember
residue of the match or cigarette, it would be unwise that identifying the origin of the fire is only the first
to do more than consider this as a possible cause, step, especially if a human-made cause is found. The
along with the possibility that sparks from a vehicle person who, through neglect or by an intentional act,
exhaust, deliberate hot set, or other source could also was responsible for a destructive fire must be brought
have played a part. Conversely, careful search at the to justice. Grass and wildlands are a valuable and irre-
origin has sometimes located a partially burned match placeable resource, just as important as any structure,
or the remains of a cigarette that turns out to be the and their loss to fire must be reduced. This can happen
device used by an arsonist. only if the responsible person or agency is identified
It is important that the investigator be satisfied and civil penalties are assessed for negligence, or crim-
that the cause of the fire has been identified and all inal charges are pressed for arson.

Review Questions

1. What are four of the major factors that affect the 6. Why is a magnet useful in examining suspected
spread of a wildland fire? areas of origin?
2. Why are weather conditions both before and 7. What kinds of physical evidence can be found in
during a wildland fire very important to the wildland fires?
investigator? 8. What is the critical fuel moisture content for cig-
3. Name three macro-scale indicators and three arette ignition of grasses and other fine fuels?
micro-scale indicators. 9. What is a hot set in a wildland fire?
4. What is the most common accidental cause and 10. Name four ways power lines can ignite wildland
most common natural cause for wildland fires? fires.
5. What is the area of transition and how is it use-
ful to an investigator?

References
1. California Dept. of Forestry, “2006–2008 Historical 3. R. C. Rothermel, How to Predict the Spread and
Wildfire Activity Statistics” (Redbooks) (Sacramento, Intensity of Forest and Range Fires, Gen. Tech. Rep.
CA: California Dept. of Forestry and Protection, INT-143 (Ogden, UT: U.S. Dept. of Agriculture,
2008). National Wildfire Coordinating Group, June 1983),
2. National Wildfire Coordinating Group, Wildfire 10–12.
Origin and Cause Determination Handbook (Boise ID: 4. E. A. Johnson, Fire and Vegetation Dynamics
NWCG, May 2005), 8–10. (Cambridge: Cambridge University Press, 1992), 40.
Chapter 8 Wildland Fires and Their Investigation 357
 5. Rothermel, How to Predict the Spread and Intensity of 25. “Lightning Detection Service,” Fire Findings 5, no. 3
Forest and Range Fires; National Wildfire (Summer 1997): 13–14.
Coordinating Group, Fire Behavior Nomograms 26. L. U. Bernardo, Fire Start Potential of Railroad
(Boise, ID: U.S. Dept. of Agriculture, March 1992). Equipment (San Dimas, CA: USDA Forest Services,
6. H. E. Anderson, “Aids to Determining Fuel Models for 1979).
Estimating Fire Behavior” (Boise, ID: National 27. C. Countryman, Ignition of Grass Fuels by Cigarettes
Wildfire Coordinating Group, April 1982). (San Dimas, CA: USDA Forest Services, 1983); P.
7. Johnson, Fire and Vegetation Dynamics, 56–57. Steensland, Cigarettes as a Wildland Fire Cause,
8. National Wildfire Coordinating Group, Wildfire Wildland Fire Origin and Cause Determination
Origin and Cause, 70–71. FI-210, NFES 2816, instructor guide/student handout
9. Johnson, Fire and Vegetation Dynamics, 45. material (Boise, ID: National Wildfire Coordinating
10. B. J. Stocks et al., “Crown Fire Behaviour in a Group).
Northern Jack Pine–Black Spruce Forest,” Canadian 28. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire
J. Forestry Research 34 (2004): 1548–60; J. D. Cohen, Science Publishers), 2003, 302, 839. See also C.
“Relating Flame Radiation to Home Ignition Using Countryman, “Ignition of Grass Fuels by Cigarette,”
Modeling and Experimental Crown Fires,” Canadian Fire Management Notes 44, 3, 1983, 3–7.
J. Forestry Research 34 (2004): 1616–26. 29. D. A. Eichmann, “Cigarette Burning Rates,” Fire and
11. Johnson, Fire and Vegetation Dynamics, 36. Arson Investigator, January–March 1980.
12. Rothermel, How to Predict the Spread and Intensity, 30. R. Bourhill, Cigarette Butt Identification Aid, 16th ed.
25–34. (Colorado Springs, CO: Forensic Science Foundation,
13. Ibid., 26. 1995).
14. Johnson, Fire and Vegetation Dynamics, 45. 31. R. T. Ford, “Proving Hot Start Arson in Wildfires,”
15. National Wildfire Coordinating Group, Wildfire Fire and Arson Investigator (April 2002): 42–43.
Origin and Cause, 6. 32. P. L. Andrews and C. D. Bevins, “BEHAVE Fire
16. National Wildfire Coordinating Group, NWCG Modeling System: Redesign and Expansion,” Fire
Course FI-110: Wildfire Observations and Origin Management Notes 59 (1999): 16–19; M. A. Finney
Scene Protection for First Responders, Boise, ID, and P. L. Andrews, “FARSITE: A Program for Fire
March 2005. Growth Simulation,” Fire Management Notes 59
17. “GPS Technology Aids Oakland Hills Fire Efforts,” (1999): 13–15; http://fire.org.
Public Works Journal (May 1992). Reprinted in 33. W. Mell et al., “Numerical Simulations of Grassland
Arizona Arson Investigator (August 1993). Fire Behavior from the LANL-FIRETEC and NIST-
18. D. W. Pack et al., “Civilian Class of Surveillance WFDS Models” in Proceeding East Fire Conference,
Satellites,” Crosslink (El Segundo, CA: Aerospace Fairfax, VA, May 2005; W. Mell, et al., “A Physics-
Corp., January 2000): 2–8. Based Approach to Modeling Grassland Fires”
19. National Wildfire Coordinating Group, Wildfire (Gaithersburg, MD: BFRL, NIST, May 2006).
Origin and Cause. 34. R. N. Meroney, “Fires in Porous Media: Natural and
20. S. L. Gutsel and E. A. Johnson, “How Fire Scars Are Urban Canopies,” prepared for NATO Advanced
Formed: Coupling a Disturbance Process to Its Study Institute: “Flow and Transport Processes in
Ecological Effect,” Canadian J. Forestry Research 26, Complex Obstructed Geometries,” Kiev, Ukraine,
no. 2 (1996): 166–74. May 5–15, 2004.
21. G. L. White, “Fire Scene Diagramming for Wildland 35. D. D. Evans et al., “Physics-Based Modeling of
Fire Investigations,” Fire and Arson Investigator Community Fires” in Proceedings Interflam 2004
(October 2004): 31–33. (London: Interscience Communications), 1065–76;
22. National Wildfire Coordinating Group, Boise, W. Mell et al., “The Wildland-Urban Interface Fire
ID, 2005. Problem: Current Approaches and Research Needs,”
23. California Department of Forestry and Fire Protection. International J. Wildland Fire (June 2009).
“2006–2008 Fires by Cause Statewide,” Sacramento, 36. A. L. Sullivan, “Wildland Surface Fire Spread
CA, 2006–2008. Modelling, 1990–2007: Pt. 1, “Physical and Quasi-
24. P. Steensland, “Long-Term Thermal Residency in physical Models,” 349–68; Pt. 2, “Empirical and
Woody Debris Piles,” paper presented at International Quasi-empirical Models,” 369–86; Pt. 3, “Simulation
Association of Arson Investigators Annual Training and Mathematical Analogue Models,” 387–403.
Conference, Denver, CO, May 2008. International J. Wildland Fire 18 (2009).

358 Chapter 8 Wildland Fires and Their Investigation

 CHAPTER 9
Automobile, Motor Vehicle,
and Marine Fires

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

KEY TERMS

Hot-plate ignition (hot surface ignition),

p. 362

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify the major sources of fires in automobiles and motor vehicles.
■ List the flash points and ignition temperatures of fluids found in vehicles.
■ Explain the systematic process of examining a vehicle for fire pattern damage and
other fire-related factors.
■ Identify the major problems associated with investigating fires in motor homes,
mobile homes, and recreational vehicles.
■ Describe the differences between fire spread in ships and in buildings.
■ List the motives for vehicle arson.

For additional review materials, appendices, and suggested readings,

visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

359
 A
ll types of motor vehicles—cars, trucks, buses, motorhomes, and the like—are sub-
ject to fires of either accidental or incendiary origin. They all contain substantial
quantities of ignitable liquids for fuels, electrical and mechanical systems to pro-
vide ignition sources, and combustible plastic and metal components and cargo to provide
fuel loads. Vehicle fires can start from a variety of accidental causes, and arsonists will often
try to simulate such accidental fires to escape detection. It is important for the fire investi-
gator to be familiar with the fuels and sources of ignition present in vehicles so that all such
fires can be properly assessed. The NFPA estimated that U.S. fire departments respond to an
average of over 270,000 fires in highway vehicles per year (2003–2008), which cause some
$1.1 billion in direct property damage each year1 (representing a slow decline in the num-
ber of vehicle fires since 1997) and an estimated 365 deaths. Vehicle fire investigators must
be familiar with fuels, ignition sources, and dynamics as encountered in vehicles if they are
to carry out correct investigations. We shall describe the fuels, fuel systems, and ignition
sources and then discuss the investigation.

Automobiles and Motor Vehicles

FUEL TANKS
The fuel tanks of most autos or light trucks contain a large reservoir of flammable and
potentially explosive gasoline and gasoline vapors (diesel, biodiesel, LPG, or CNG are, of
course, also encountered). Extensive firefighting experience and laboratory tests reveal
that contrary to expectations, fuel tanks very rarely explode during a vehicle fire.2 Despite
movie and television scenes, it is nearly impossible to ignite the gasoline inside. In part,
this is because its vapors are much heavier than air and tend to fill the tank completely to
the top. Except at extremely low ambient temperatures, the vapor pressure of gasoline is
such that it tends to form a mixture with residual air in the tank that is well above the
explosive limit and cannot be ignited. It follows that fire will not ordinarily invade an
open gasoline tank but will be limited to the outside of any openings where there is suf-
ficient air for combustion. Gasoline vapors outside an open fuel filler neck can usually be
ignited with an open flame, to create a steady, torch-like flame longer than a foot (30 cm)
or so in length. Propagation into the tank to cause an explosion is almost impossible
unless the tank has so little fuel in it that an equilibrium vapor pressure is not reached.
(The dynamics of such fires will be explored in more detail in Chapter 12.) A closed vapor
recovery system, when exposed to a fire, can release vapor under pressure, creating a
torch-like effect, sometimes in locations remote from the tank itself.
When exposed to a general fire, the soldered seams and fittings of a metal tank may
melt and separate, releasing fuel (sometimes under considerable pressure) to fuel the
fire. Pressure built up by vaporizing fuel can split the tank or blow the filler cap off (of
either metal or polymeric tanks). Accidents can, and frequently do, lead to rupture of
the fuel tank by mechanical force, producing a very hazardous spill. Since most large
trucks use diesel engines, their fuel presents less of a hazard when spilled at ordinary
temperatures, but in a general fire, diesel fuel can vaporize to form explosive vapor
mixtures. In addition, some vehicles are being fitted with LPG (liquefied petroleum gas)
or CNG (compressed natural gas) fuel systems, and a number of fires have been caused
by even small leaks or poor filling practices that resulted in explosive vapor mixtures
in or around the vehicle.
To minimize fabrication costs and vehicle weight, manufacturers are using polymeric
tanks in more and more vehicles today. They are strong and puncture-resistant, but unlike
metal tanks, they are subject to cracking and can melt or burn through in an enveloping
fire, releasing their contents quickly, giving the appearance of an accelerated fire. Under
360 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
rare conditions, polymeric tanks can fail explosively as a result of a boiling-liquid,
expanding-vapor explosion (BLEVE).

FUEL TANK CONNECTIONS

Because regular fuel tanks are mechanically strong and usually well protected, the tanks
themselves are rarely involved in accidental fires (except for severe crashes). Their filling
and fuel line connections are a different matter. The filler spouts of many tanks have flex-
ible rubber connections that can crack or loosen and allow fuel to spill into trunks and
passenger compartments.
CNG vehicles may have a glass-fiber-wrapped polymeric tank that can fail explo-
sively in a general vehicle fire. Topping off the tank often leads to overfilling, creating
hazardous spills beneath the vehicle. (In older vehicles, fuel filler leaks could introduce
fuel into the trunk or even the passenger or engine compartment.) Overfilling may
cause excess fuel to be siphoned into the vapor recovery canister in the engine compart-
ment. This canister, often filled with charcoal, is designed to trap fuel vapors, which are
then drawn into the engine to be burned. Liquid fuel in this canister can pose a fire haz-
ard. The flexible connectors (hoses) on metal fuel lines are subject to splitting or crack-
ing with age or loosening due to abuse or improper service repairs, causing a fuel spill
hazard. Many manufacturers have adopted the use of nylon fuel lines and polymeric
connectors. Physical damage, exposure to severe heat conditions, or improper service
procedures can result in these cracking, splitting, or leaking, which can result in a fire
hazard. It was observed that a polymeric fuel line, ignited by accidental contact with an
exhaust header, burned the length of the car to the fuel tank, with the fuel in the line
supporting the fire.

FUEL PUMPS, FUEL LINES, AND CARBURETORS

The fuel pumps on many older cars are mechanical ones, mounted on the side of the
engine, that function only when the engine is turning over. Most imported cars and
newer domestic cars are fitted with high-capacity electric pumps that are pumping when-
ever the ignition is on. Because most electric pumps “push” better than they “pull,” they
are most often mounted at or near the tank (and many are enclosed in the tank itself).
Without a proper interlock (to oil pressure or engine-speed sensors) (or with failure of
the interlock safety), such pumps can pump a great deal of fuel out through a loose or
ruptured fuel line even though the engine may not be running. Some new vehicles have
two pumps—one in the tank and a second one closer to the engine to produce high pres-
sure. If a fuel line is compromised by external fire or mechanical impact (collision),
gasoline may siphon out of the tank by gravity even if the fuel pump is inoperative.
Although the amount of gasoline in the carburetor itself or in the fuel injection lines
(rails) may be small, the fuel lines of some cars are positioned over electrical components
or the exhaust manifold, and loose connections pose a hazard for a large fire. If a fire is
limited to the fuel in the carburetor or fuel rail, the fire will generally be limited to mod-
erate damage of the air cleaner, hood, and engine wiring and rarely be transmitted to
other parts of the vehicle.

FUEL INJECTION SYSTEMS

Most fuel injection systems are electronically controlled and are found on nearly all newer
gasoline-fueled vehicles. These systems provide additional fire risks because their fuel deliv-
ery lines carry gasoline at pressures from 25 to 100 psi (150 to 700 kPa). Even a pinhole
leak in a line or minor crack in a fitting allows gasoline to be atomized and sprayed around
the engine compartment at high pressures, quickly producing an explosive vapor mixture
requiring only a suitable electrical arc to ignite. The fuel rail is the metal tubing that dis-
tributes the fuel to individual injectors on each cylinder (or, in older vehicles, directly to the
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 361
 TABLE 9-1 Flash Points and Ignition Temperatures of Vehicle Fluids

FLUID FLASH POINT IGNITION TEMPERATURE

280°C–456°C (536°F–853°F)
Gasoline –43°C–38°C (–45°F–36°F) (varies with grade)
Ethanol 13°C (55°F) 363°C (685°F)
Diesel fuel 52°C (125°F) 260°C (500°F)
Ethylene glycol (100%) 111°C (232°F) 398°C (748°F)
Propylene glycol (100%) 99°C (210°F) 371°C (700°F)
Windshield washer fluid (methanol) 11°C–15°C (52°F–55°F) 464°C–484°C (867°F–903°F)
Hydraulic brake fluid
DOT 3 (Normal auto service) 190°C (375°F)a
DOT 5 (Synthetic or silicone) 260°C (500°F)b
Lubricating oil (including power 121°C–205°C 310°C (590°F)
steering and automatic (249°F–402°F)
transmission fluid)
Compressed natural gas (CNG) N/A 632°C–650°C (1169°F–1202°F)
Propane (LPG) N/A 450°C–493°C (842°F–919°F)

Sources: NFPA, Fire Protection Guide to Hazardous Materials, 12th ed. (Quincy, MA: NFPA, 2001).
a
MSDS: Champion 1400 Series, DOT3, rev. 1/13/09.
b
MSDS Maxima Racing Oils, rev. Sept. 2009

throttle body). It usually has a high-pressure (delivery) side and a low-pressure (return)
line. Such systems can leak at the gaskets or seals at the injectors or the connectors.

VEHICLE FUELS
Today, more vehicles are being seen with gaseous fuels rather than liquid ones. Propane
(sometimes called LPG) is stored in a vehicle as a liquid in a pressurized metal tank at
pressures of approximately 75 psi. The fuel is delivered to the motor as a liquid, which
evaporates quickly to a clean-burning gaseous fuel. Natural gas may be encountered
(especially in large commercial or passenger vehicles) as compressed gas (CNG) with tank
pressures of 2,400–3,600 psi (16,500–24,800 kPa). These tanks may be of steel or glass-
fiber composites.
Gasoline is the most abundant fuel in most automobiles and the one with the lowest
flash point (see Table 9-1) and is therefore often thought to be the first fuel ignited.
However, its properties render it somewhat less likely to ignite upon contact with a hot
hot-plate ignition (hot surface (referred to as hot-plate ignition) than are other heavier combustible liquids such
surface ignition) ■ as hydraulic fluids and lubricating oils. Tests performed by auto manufacturers and fire
Temperature of a metal
investigators have shown that gasoline seldom ignites at hot-plate temperatures near the
test plate at which liquid
fuel ignites on contact. listed autoignition temperature (AIT) and, in the under-hood environment, requires sur-
face temperatures between 590°C and 730°C (1,100°F to 1,350°F) to reliably ignite.3
These temperatures may be attained at the exhaust manifolds but only after extreme and
sustained engine load or severe mechanical malfunction (timing and mixture) or on the
surfaces of turbochargers or catalytic converters. Textures of hot surfaces (including
crevices and cracks) may reduce the required temperature (see Chapter 6). The rough
undersides of hot cast exhaust manifolds may offer better chances of ignition. Leaks of
burning combustion gases around loose spark plugs or faulty manifold gaskets offer a
chance of flame ignition.
362 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
OTHER COMBUSTIBLE LIQUIDS
The flammability properties of some of the other combustible liquid fuels in the engine
compartment are shown in Table 9-1. As we saw in Chapter 4, the residence or contact
time between a fuel and a hot surface is critical, as are the nature of the surface and the
quantity of fuel. Hot-surface ignition of transmission oil (460°C-600°C) (860°F-1112°F)
or steering or brake fluid (540°C-570°C) (1000°F-1060°F) is much more likely than
“open” hot-surface ignition of gasoline. Of course, in a severe collision, all these fluids may
be released. Antifreeze is not often thought to be combustible, but pure ethylene glycol has
a flash point of 111°C (232°F) and an AIT of 398°C (748°F) (and the values for propylene
glycol are even lower).4 If antifreeze (often, a 50/50 mix of ethylene glycol and water) is
allowed to pool on top of an engine, the water can boil away until the pure glycol can form
a vapor ignitable by heated surfaces (530°C-570°C) (986°F-1058°F) or by arcing within
the distributor, alternator, faulty spark plug leads, fan motor, or other electrical
equipment.5 The surface temperatures of exhaust manifolds, turbochargers, catalytic con-
verters, and the like may be far in excess of the 398°C (748°F) required to ignite ethylene
glycol, and the high boiling point of the glycols slows their evaporation to give more time
for energy transfer and mixing). Ignition of a 50/50 ethylene glycol–water mix sprayed
directly onto a hot exhaust manifold has been reported, although documented cases of
such ignition are very rare.6
Brake fluid (which is usually a heavy glycol mix) and automatic transmission fluid
(ATF) and power steering fluid (which are light lubricating oil–grade petroleum distil-
lates) also represent potential fuels ignitable by hot surfaces.7 Such fluids may be present
in reservoirs (including polymeric ones that can leak, fracture, or melt) or in pressurized
lines. If these lines suffer even a small puncture or leak, the fluid is sprayed out in an eas-
ily ignitable aerosol. These fluids have low ignition temperatures and can be ignited if
sprayed onto hot exhaust components or disc brake rotors, or if they come into contact
with electrical arc sources. Gasoline, by contrast, is not readily ignited by contact with the
exterior of a hot exhaust manifold, due in part to its volatility (but other more viscous,
less volatile fuels with similar AITs, such as diesel fuel or motor oil, are ignited). As seen
in Figure 9-1, the normal maximum temperature for an exhaust manifold of 550°C

Behavioral properties of vehicle fluids in

relation to temperatures and ignition sources
Engine Brake
oil fluid Diesel Petrol LPG Temp (°C ) Ignition source
1500
Light
1400
filaments
1300
Sparks FIGURE 9-1 Repre-
1200
Definite ignition sentative ignition temper-
1100
Catalytic atures of vehicle fluids,
1000
converter temperatures of vehicle
900
interior fluids, and temperatures
800
Cigarette of ignition sources.
700
(maximum) Courtesy of Fire, October
600
1991. Reprinted in March,
500 Exhaust
Ignition G., “An Investigation and
400 manifold
may Evaluation of Fire and
300
occur No ignition Normal Explosion Hazards Resulting
200
Flash engine from Modern Developments
Flash 100
point exhaust in Vehicle Manufacture.”
Flash 0
point Ambient Newcastle-upon-Tyne, United
point Flash –100
air Kingdom. Tyne and Wear Fire
point Flash
point temperature Brigade, Newcastle, United
Kingdom, 1993.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 363

 TABLE 9-2 Typical Operating Pressures for Vehicle Fluids

Power steering with rack and pinion 50–300 psi (345–2,070 kPa)
(maximum relief pressures can be over 1,000 psi)
Power steering with conventional steering gear 600–1,200 psi (4,140–8,200 kPa)
(maximum relief pressures can be over 1,500 psi)
Automatic transmissions 75–400 psi (518–2,760 kPa)
Fuel pump for carburetors 3–9 psi (21–62 kPa)
Fuel pump for fuel injection 40–120 psi (276–828 kPa)
Engine oil pressure 15–75 psi (103–518 kPa)
Cooling systems 12–17 psi (83–117 kPa)
Hydraulic ride or active ride control systems 125–1,800 psi (862–12,420 kPa)
Operating pressures of brake system non-ABS 0–1,200 psi (0–8,280 kPa)
Operating pressure of ABS brake system 0–2,000 psi (0–13,800 kPa)
Diesel fuel in-line injection or rotary pump
(into combustion chamber) 3,000–20,000 psi (20,700–138,000 kPa)
(low-pressure side) 40–80 psi (276–552 kPa)

Sources: R. Parsons, Wright Group, Uxbridge, MA; G. Barnett, Automotive Fire Analysis, 2nd ed. (Tucson, AZ:
Lawyers & Judges Publishing, 2007).

(1,020°F) is not adequate for hot-surface ignition of gasoline (petrol), while temperatures
above 395°C (743°F) can ignite DOT 3 brake fluid or automatic transmission fluid
dripped onto its exterior.8
Many fluids in vehicles are at elevated pressures (see Table 9-2). The release of any
ignitable fluid under pressure produces an aerosol mist that is much more readily ignitable
than the bulk liquid. Operating fluids are usually at elevated temperatures as well, mak-
ing them easier to ignite upon release.
Transmission or torque converter failure can produce internal heat sufficient to vapor-
ize transmission fluid. This vaporization may produce internal pressures that result in fluid
being expelled from vents, seals, or dipstick tubes. When the fluid (or its aerosol) contacts
heated exhaust components, it can ignite. Post-fire indicators of this event include low
transmission fluid level combined with a darkened oil color, particulate contamination, or
a strong burned odor. Heat transfer patterns from transmission fluid fires tend to be at the
bottom of the compartment, near the transmission tunnel or at the base of the radiator
(where the fluid can leak from cooler connections). Confirmation of the failure may require
removal and disassembly of both the transmission and the torque converter.9

ENGINE FUEL SYSTEM FIRES

Engine fuel-system fires tend to show heat effects in the upper part of the engine com-
partment, and particularly on the hood, where a series of rings or halos of pyrolyzed
paint and metal oxidized to different degrees may indicate a localized fire from such a
fuel source. Such fires also tend to produce white oxidation (high-temperature) patterns
on the firewall, inner fenders, or strut towers. Whenever the fuel system is pressurized,
leaks elsewhere in the delivery lines may allow fuel to spill beneath the vehicle to be
ignited there. Leaks may occur from loosened connections and failures of the lines due
to corrosion, aging, splitting, vibration, or mechanical damage such as crushing or con-
tact against moving parts, as illustrated in Figure 9-2.10 Many connectors today are of
364 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
FIGURE 9-2 A rubber fuel line, connecting a fuel injector to the fuel rail, cracked from age and
sprayed gasoline on the top of an engine. Courtesy of Bloom Fire Investigation, Grants Pass, OR.

plastic, and some modern vehicles have fuel lines that are entirely of plastic. These can
burn away completely, so they may not be detectable post-fire. In-line fuel filters (metal,
plastic, or glass and metal) can fracture or separate to produce fuel leaks. O-ring seals
in fuel injection systems degrade from heat, pressure, or fuel additives (or improper
installation) and release fuel under pressure. Fires on the top of the engine are most likely
to be fuel fires, while fires near the firewall (bulkhead) or wheel wells are more likely to
be brake fluid or automatic transmission fluid fires. Lines can be cut and even ignited by
contact with moving parts such as pulleys, belts, or drive shafts. Because of the complex
routing of their hoses, power steering and coolant systems cannot be generalized as to
origin location. The trained investigator who is very familiar with the various systems,
the fluids they contain, and the variations in construction and routing (model, year, and
manufacturer) can sort out the relevant damage patterns.
Once ignited, the myriad hoses, vacuum lines, and plastic-insulated wires of mod-
ern vehicles provide a great deal of fuel to support a fire. (Very few of the polymeric
materials in engine compartments are fire-resistant.) Many vacuum system and vapor
recovery hoses use plastic connectors that can fracture, melt, or simply add to the fuel
load. The modern vehicle has many fresh-air, heating, and air-conditioning ducts of
synthetic materials that originate in the engine compartment and penetrate the fire-
wall, as shown in Figure 9-3. These are, in turn, connected to polymeric fan enclo-
sures. While they are not ignited by glowing ignition sources, all these can support
flaming ignition once ignited and allow flames to readily penetrate into the passenger
compartment. Fires in these enclosures and ducts may be ignited by failure of fan
motors or by fires in the engine compartment.11 Fire entering the interior through the
large openings created by failure of polymeric enclosures will impinge on plastic dash-
board components and spread readily into the interior. If the fire progresses to the
point where window glass is exposed to fire, side window glass, being tempered, is
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 365
FIGURE 9-3 Pene-
trations of the ”firewall”
in modern vehicle.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

especially vulnerable to thermal-shock failure, as depicted in Figure 9-4. If the window

glass breaks, there can be sufficient ventilation to allow the fire to eventually involve
all combustible interior components. Fire thus transmitted into the passenger compart-
ment may be revealed by fire damage coming from the defroster vents between the
windshield and top of the dashboard. Engine compartment, tire, or brake fluid fires
can ignite the polymeric fender liners used on many cars, and these fires, in turn, can
extend under front fenders and attack side windows, door seals, and the lower front
corners of windshields.12

FIGURE 9-4 Tempered

side window glass shat-
ters to cubes. Laminated
safety glass breaks in
curves. Courtesy John D.
DeHaan.

366 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

ELECTRICAL SYSTEMS
The general principles of electrical fire causation will be discussed in Chapter 10, and
they are just as applicable in vehicle circuits as in structural ones. Although the operat-
ing voltage of most autos and light trucks today is only 12 V, and the currents (in most
fused circuits) are limited to a few amps, electrical causes of vehicle fires cannot be dis-
missed out of hand. Most power on vehicles is direct current, so arcing, once estab-
lished will not self-extinguish, as often seen in alternating current events. Trucks and
commercial vehicles commonly have 24-V systems for starting and charging systems.
Most vehicles today have a single energized conductor connected to the positive battery
terminal, with the return path provided by the metal vehicle frame and components.
This is called a negative-ground system, since the negative post of the battery is con-
nected to the vehicle’s frame (American cars before 1955 often used a 6-V positive-
ground system.) This frame-ground system means that an energized conductor coming
into contact with any metal component can cause current flow. The high-capacity
cables fitted to the battery connections can carry sufficient amperage to start fires of
the cable insulation, starter, and related electrical fittings. In short circuits of these or
high-current circuits of the horn, cooling fans, cigarette lighter, or lighting equipment,
the temperatures developed will cause smoking or even flaming combustion of the insu-
lation and any connectors or terminals. Such circuits are usually energized even when
the key is in its off position, but all but the starter are often protected by a fusible link—
a special wire connector. There is no overcurrent protection on the starter circuit. If this
circuit shorts to start a fire near combustible components (fender underliners, splash
shields, ducts, hoses, etc.), a large fire can develop quickly. Attempts to deliberately
cause such electrical fires in the protected wiring in a series of modern vehicles, how-
ever, invariably caused the overcurrent devices in the vehicles to interrupt the power
before ignition could occur.13 Relays control many of the higher-current accessories
(like blower motors) in newer cars, so there is less chance for a high-current fault to
occur to initiate a fire.
Ignition switches (which have unfused “hot” wires and ground connections in close
proximity) have been linked to fires started when carbon tracking allowed current to
flow continuously between contacts in the switch and eventually to degrade and ignite
the plastic housing.14 Some recent vehicle fires have been linked to corrosion of fittings
associated with speed control of activation switches on master brake cylinders.15 When
large, energized, unfused conductors can come into contact with grounded metal com-
ponents, a massive current flow can result, with the accompanying increased likelihood
of ignition, as illustrated in Figures 9-5a and b. In one author’s experience, a battery
cable that came into contact with an exhaust pipe caused a direct short when the insu-
lation melted. The resulting massive current melted and charred the insulation on the
cable and interconnected wires within seconds. One form of grounded metal compo-
nent that is easily overlooked is the steel braid wrapping in high-pressure hydraulic and
brake lines. Chafing of these lines against metal edges (such as starter cable terminals)
can abrade the protective rubber coating and allow contact. The hot electric arc pro-
duced quickly chars the rubber liner, allowing a leak of the ignitable fluid and the
potential for a very destructive fire fueled first by the fluid, then by the rubber lines.
As with any other kind of fire investigation, the first fuel ignited has to be identified,
and the chain of reactions has to be demonstrated. Just having a source of heat, no mat-
ter how intense, will not result in a fire unless an adequate supply of a suitable fuel is
present and the mechanism for enough heat transfer from that source to that fuel can
be shown.
The newest generation of theft protection devices includes keyless entry and starting
(with an encoded transponder), remote locking and unlocking, and specially coded igni-
tion keys. Their contribution to fires and investigations is unknown at this time.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 367

 FIGURE 9-5A The vibra-
tions of a farm tractor
caused chafing to the
battery cables, resulting
in a dead short. Courtesy of
Tim Yandell, Public Agency
Training Council.

FIGURE 9-5B The main

alternator connector on
this vehicle failed, causing
a significant engine fire.
Courtesy of Tim Yandell, Public
Agency Training Council.

Although the current standard for automotive electrical systems is 12 VDC, negative
ground, recently developed technology suggests more efficient higher-voltage systems
will be implemented on most automobiles in the near future. These systems will likely
employ a 42 V alternator/starter motor combination that will be installed in the flywheel
area. Currently, many trucks and other diesel-powered vehicles use 24 V systems.
Electric- and hybrid electric-gas-powered vehicles are already in service. The high
wattage capabilities of such electrical systems (including batteries that can keep mechan-
ical systems operational) can create fire hazards after crashes. The power for electric
propulsion motors is 600 VDC with high-current batteries. This power is supplied in
two conductors (with brightly colored insulation—blue, orange, or green) separate from
the vehicle’s lighting and control systems.16
Many modified vehicles use high-current accessories (such as audio systems or
hydraulic pumps) in an unsafe fashion, and wiring may be exposed to all manner of
crushing and chafing hazards that can compromise the insulation. The same care
should be exercised in examining such vehicles as one would employ in examining an
older house for dangerous modifications. It is obvious that modern vehicles are often
fitted with power accessories such as power seats, power windows, seat heaters, tele-
phones, and even portable computers and video systems. All these add to the complex-
ity of the fire scene even if they do not usually add to the fire hazards when properly
installed.
One additional source of electrical fire hazard is the battery. In normal use, it pres-
ents no fire hazard, but when it is receiving a charging current, considerable volumes
368 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
of explosive hydrogen gas are generated. An explosion will result from an arc (even
between the cells of the battery) or other source of ignition, and some fire may follow
if the charging continues. Explosions unaccompanied by fire are the rule, and even
these are rare. Symptoms of malfunction will usually be apparent to the operator.
Extra batteries may be concealed in the vehicle to power hydraulic lift systems or high-
power sound systems in modified vehicles. Some vehicles have their operational bat-
teries in trunks or beneath cargo floors. Some “accessories” require the use of
inverters to convert the 12 VDC of the vehicle to 120 VAC. Such inverters can fail, to
cause fires.

MISCELLANEOUS CAUSES
The NFPA estimates that some 50 percent of highway vehicle fires are caused by mechan-
ical failures or malfunctions and 24 percent are caused by electrical failures or malfunc-
tions. Other identified causes include misuse (11 percent), intentional ignition (8 percent),
and collision or overturn (3 percent).17

Turbochargers
Many vehicles (both gasoline- and diesel-fueled cars and trucks) are fitted with tur-
bochargers that redirect the exhaust gases to a housing to spin an impeller. This impeller
then helps push air into the intake system. Such turbocharger housings operate at very
high external temperatures and also, because of their very high rotational speeds, are sub-
ject to bearing failures that can produce hot-fragment ignition as the bearing, housing,
and rotor fail. Bearing failure can also result in the ignition of the lubricating oil supplied
to the center bearing under pressure.

Catalytic Converters
Catalytic converters have been fitted to domestic (U.S.) cars since 1975 and to light trucks
since 1984. These units convert hydrocarbons, carbon monoxide (CO), and nitrogen
oxides (NOx) to less-toxic compounds by chemical conversion using a metallic catalyst
like platinum or palladium in a porous ceramic foam or in loose beads. Under normal
operating conditions, these converters operate at internal temperatures on the order of
600°C to 650°C (1,100°F to 1,200°F) with external surface temperatures of approxi-
mately 300°C (575°F). If the converter is misused or the vehicle is not maintained prop-
erly, the converter can develop internal temperatures that can exceed the melting point of
the ceramic matrix, 1,250°C (2,300°F), and surface temperatures on the order of 500°C
(900°F).18 Such radiative sources can ignite combustible materials under the vehicle or in
contact with the floorpan above them. Some vehicles have been completely burned when
carelessly parked in dry brush or leaves with the hot converter in contact with the com-
bustible litter. Prolonged idling of some converter-equipped vehicles, especially where air
circulation under the car is limited, can produce sufficiently high temperatures that floor
coverings and upholstery within the vehicle can be ignited by radiant or conducted heat.
The newer generation of catalytic converters use smaller, often multiple, reactors whose
surface temperatures are much lower than those of the 1980s. They also tend to be
shielded better and much less likely to ignite fires inside or below the vehicle by contact
with ignitable fuels.

Block and Other Heaters

In cold climates built-in block heaters or accessory block heaters (OEM and aftermarket)
have been known to fail and ignite combustibles. Improvised heaters have included hair
dryers, electric blankets, and trouble (service) lamps. Considering the damp conditions
that may be present, each of these presents new ignition risks that should be considered
(see Figures 9-6a and b).
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 369
FIGURE 9-6A Fire in engine compartment could have had several FIGURE 9-6B Engine block heater (above oil filter) failed and
possible causes. This one was a challenge because the vehicle was caused the fire. Courtesy of Peter Brosz, Brosz & Associates Forensic Services,
parked and off at the time of the fire. Courtesy of Peter Brosz, Brosz & Inc.; Professor Helmut Brosz, Institute of Forensic Electro-Pathology.
Associates Forensic Services, Inc.; Professor Helmut Brosz, Institute of Forensic
Electro-Pathology.

Rags
Other types of common carelessness cannot be ignored as fire causes. Rags have been left
in the engine compartment or on top of exhaust pipes and mufflers during servicing, only
to ignite when the exhaust system reached its operating temperature (as much as 550°C;
1,000°F). The burning rag then spreads the fire to other combustible materials. It matters
little whether the rag is oily or not because it is the rag that will be ignited by contact with
the hot surface. Improperly installed hydraulic lines can come into contact with moving
engine parts to cause penetration and loss of ignitable fluid, as shown in Figure 9-7.
Overflows from leaking overfilled or overheated transmissions, power steering, or brake
systems can flow or spray onto hot surfaces, causing ignitions.

FIGURE 9-7 An improp-

erly installed power steer-
ing line rubbed against
an engine fan belt and
sprayed its contents,
causing an engine com-
partment fire. Courtesy of
Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

370 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

Tires
Drivers can be almost unbelievably insensitive to the condition of their tires, especially
those on trailers. A deflated tire, when run against the road surface mile after mile, devel-
ops great frictional heat. This heat can build up sufficiently to ignite the tire, the hydraulic
system of the brakes, and even the vehicle or its cargo, sometimes with disastrous results.
Such a situation is most serious. The movement of the vehicle usually aids in dissipating
the heat being developed, but when the driver finally stops to inspect the problem, the
heat can ignite the tires to flaming combustion. Although tires under these conditions
require a great deal of heat to ignite, once ignited they are extremely difficult to extin-
guish. Once burning, the oily char of rubber repels water, and the insulative properties of
the rubber prevent cooling by conduction. Even a driver prepared with a portable fire
extinguisher may be unable to prevent a disastrous fire. Single large tires on industrial
(earth-moving) equipment have been measured to produce maximum fires of 3 MW (at
90 minutes after ignition by a diesel fuel fire).19

CONSIDERATIONS FOR FIRE INVESTIGATION

One interesting observation is that many accidental vehicle fires occur in the few minutes
after being shut off. When the circulation of air, oil, and coolant is stopped, the external
temperatures of many engine components (exhaust manifolds and systems, turbocharger
housings, and the like) rise. Furthermore, with the cessation of movement and fan circu-
lation, concentrations of ignitable vapors, which had been dispersed or diluted to safe
concentrations, can accumulate to above their lower explosive limit. The combination of
higher temperatures and higher concentrations of vapors, including pyrolysis products
from smoldering plastics, can result in ignition. Such situations may be preceded by
unusual odors, noises, mists, or electrical fluctuations, which can paralyze an engine man-
agement computer and cause stalling, so the driver should be asked specifically about any
peculiarities that may have indicated such problems. These fires normally begin with a
smoldering ignition that then goes on to flaming combustion to engulf the car, so the time
frame for development is some minutes.
Electrical shorts developing in the maze of wiring and myriad components have a rich
source of fuel available once temperatures are reached at which the plastics and resins begin
to pyrolyze and ignite. These initial stages take many seconds or minutes to develop.
Depending on the initial fuel and ventilation conditions, an accidental fire may not be visi-
ble for many minutes to someone outside the vehicle. A fire accelerated with a flammable
liquid would be expected to be visible as open flames almost immediately after ignition,
especially if a door, window, hatchback, or sunroof was open. However, a fire ignited only
by a direct flame in a closed car may not be visible from a distance for some minutes.
The investigator should exercise great care in deciding what amount of damage is
consistent with an accelerated fire versus that from an accidental fire. Normally, the time
factor of fire development is usually a significant diagnostic sign of accidental origin. If a
reliable witness observed the progress of the fire and can establish that the vehicle was
totally involved in flames within 5 minutes or so of its arrival or last occupancy, it is a
very strong indication that the fire was set with the use of an accelerant. Accidental fires
usually require considerably longer to start, develop into flaming combustion, and spread
than do accelerated fires. In the absence of a reliable eyewitness or videotape, it is very
dangerous to estimate speed of development.
If the vehicle was being driven (or had just been parked) when it caught fire, the
driver and any passengers should be interviewed to gain information about possible
mechanical problems. Questions should include the following: How long was the car
driven just prior to the fire? When was it last fueled? When and where was it last serv-
iced? Were there any problems with starting, steering, or shifting? Were there any unusual
smells? Did the car overheat recently? Were there any noises just prior to the fire? Were
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 371
 there any electrical events—lights coming on, instrument readings surging, accessories or
controls not working properly, etc.? The owner should also be interviewed about modifi-
cations, special equipment, or recent repairs.

COMBUSTIBLE MATERIALS
The upholstery of seats, carpeting, headliners, and side panels represent a significant fuel
load in every car or light truck. Modern cars make extensive use of vinyl plastics,
polyurethane foam rubber, nylon carpeting, and polyester and cotton seat fabrics. All these
products are combustible to a greater or lesser extent and thus provide fuel for fires started
by direct accidental ignition with any open flame, as well as for fires started in other por-
tions of the vehicle. It is estimated that some 20 percent (by weight) of a modern auto con-
sists of combustible fabrics, plastics, polymers, and rubber. This constitutes on the order of
228 to 273 kg (500–600 lb) of fuel load, which is responsible for the fire damage to the vehi-
cle, no matter how it is first ignited.20 Although no recent published data are available, it is
clear that today’s vehicles use considerably more combustible plastic and synthetic compo-
nents than these data reflect. Most of the interior (including seat backs, instrument panels,
dashboards, and side panels) are readily combustible. Unlike in earlier cars, polymeric mate-
rials are used throughout—including exterior body panels, structural components, and even
suspension and drivetrain—not just in the upholstery. Before an investigator concludes that
a component has been stripped off a car before it was burned, he or she must make sure
that it has not simply burned up with the rest of the fuel.

Motor Vehicle Safety Standard 302

It is a common misconception that the plastic upholstery materials in automobiles have to be
flame retardant or pass ignition tests. In fact, the only applicable fire performance test is U.S.
Motor Vehicle Safety Standard No. 302 (MVSS-302), which applies to all nonmetal parts of
vehicle interiors.21 It is an open-flame, horizontal-spread test that is very easy to pass, as it
was designed in response to vehicle fires in which the headliners ignited and dripped flaming
residues onto occupants. It has been in effect since 1972 without change and has been found
to have had no measurable effect on the incidence or severity of fire injuries to passengers.
The MVSS-302 test does not accurately reflect the fire performance of fabrics in real fires in
vehicles, and replacement tests are being evaluated.22 There are no tests for flammability of
seat padding, which is almost universally polyurethane foam. Tests by the authors have
demonstrated that even a modest open flame applied to the underside of a vehicle seat cush-
ion will result in rapid involvement of the seat and potentially full involvement of the inte-
rior of the vehicle.23 It has been suggested that materials used in higher-value cars today
include more fire-resistant compounds, but this has not been confirmed by testing.24

MISCELLANEOUS IGNITION MECHANISMS

Accidental fires today rarely start from cigarettes lying in the crevices of seats due to the
almost universal use of thermoplastic fabrics and urethane foam padding. (Of course, accu-
mulations of cellulosic trash—facial tissues, newspapers, food wrappers—pose the same
risks of smoldering ignition as they do in homes.) Although such trash fires are usually lim-
ited to smoldering, they can occasionally produce (if cellulosic fuels are present and if prop-
erly ventilated) flaming combustion that spreads quickly through the rest of the car. Pillows,
blankets, and other nonstandard accessories provide the same risks to vehicles as they do in
dwellings, because they may be subject to both flaming and smoldering ignition sources.25
Foam rubber in later-model cars is more fire resistant to smoldering (versus open-flame) igni-
tion than the latex foam and cotton padding used in earlier cars. Because of the variation in
fire susceptibility, if there is reason to believe that a fire was started accidentally, it may be
necessary to obtain a duplicate car from a wrecking yard and test various means of ignition
directly. Virtually all synthetic materials used in modern car interiors are susceptible to flame
372 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
ignition (e.g., a match flame). Most are thermoplastics, which means they melt as they burn,
and the burning droplets of material then fall onto other ignitable surfaces and spread the
fire very quickly. The flame temperatures and heat fluxes of such plastic fires exceed those
of a gasoline fire and can mimic many of the same effects. Studies by Hirschler and Hoffman
showed that nearly all plastic vehicle components in passenger compartments had short igni-
tion times and high heat release rates when tested using cone calorimetry.26

Vehicle Arson
The NFPA estimated that incendiary fires accounted for 9 percent of all reported highway
vehicle fires.27 (It should be noted that many U.S. jurisdictions do not investigate vehicle
fires, so this percentage should be considered very conservative.) Until recently, vehicular
arson was quite rare, because a predominant motive, insurance fraud, was of minimal
economic benefit to the arsonist. Unlike real property that can be traded back and forth
to inflate its insured value prior to the fire, most motor vehicles are insured on the basis
of their actual cash value—that is, their replacement cost less a substantial amount for
depreciation. If a vehicle is burned, the payoff on the policy is usually based on Kelley Blue
Book figures for that make and model, with credits added for accessories and reductions
made for high mileage. Because these values are often quite low, compared with the actual
resale value of the car or its components, it does not often pay the arsonist to burn his or
her own vehicle. Motives are discussed in more detail in a later section of this chapter.

CONSIDERATIONS FOR FIRE INVESTIGATION

Setting fire to a modern vehicle does not require the use of an ignitable liquid accelerant.
Any direct flame source of moderate size (a burning crumpled sheet of newspaper) under
the seat or under the instrument panel can result in a fast-spreading, destructive fire inside
a modern vehicle if there is ventilation. Whether an accelerant is used or not, ventilation
is critical. A modern vehicle is sufficiently airtight that if the windows and doors are all
closed, a fire in the passenger compartment will usually self-extinguish unless a window
(or sunroof) fails from thermal shock early in the fire. As in structure fires, the investiga-
tor must be willing to treat every vehicle fire as a possible arson from the standpoint of
evidence preservation and documentation, not from investigative bias or preconception.

Protocol for Vehicle Examination

The general protocol for vehicle examination follows the same logical pattern as for struc-
ture fires: documentation and examination of exterior, interior, and specialized evidence
features. The procedure is outlined in Figure 9-8 as a general checklist protocol.

SAFETY
Safety is an important consideration in vehicle examinations, as it is in structure fires! Before
beginning the examination, the investigator must be sure that all batteries are disconnected,
air bags (steering wheel, passenger dash, side curtain, side impact, and seat backs) are dis-
abled or discharged, pressurized fuel containers (LP or CNG) are sealed, and hydraulic
cylinders (shock absorbers, bumpers, tailgate assists) are cooled. There are also risks from
leaking fuel, broken glass, torn metal, and toxic gases from continuing smoldering. More
components are made of carbon fiber, and burned carbon fiber has been linked to inhala- Stay Safe

tion injuries.28 If the vehicle is lifted to allow examination of its underside, it must be
Safety is as important a
securely based on proper jackstands (or service rack) before anyone gets underneath. consideration in vehicle
Although it may be necessary to use a forklift to position a vehicle for examination, it is vio- examinations as it is in
lation of OSHA standards to have any person under a vehicle being supported by a forklift. structure fires!

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 373

 AGENCY FILE NUMBER
Motor Vehicle Fires
VEHICLE DESCRIPTION
Year Make Model License # ST Expiration

Color(s) VIN#

OWNER/OCCUPANT
Owner’s Name Owner Address Owner Phone#

Operator Name/License# Operator Address Operator Phone#

EXTERIOR
Prior Damage Fire Damage

Tread Depth Wheel Type Lug Nuts Present and Number

Condition of Glass

Tires/Wheels (Missing, Match, Condition)

Parts Missing

FUEL SYSTEM
Prior Damage Fire Damage

Type Fuel Condition of Tank Filler Cap Condition Fuel Line Condition

ENGINE COMPARTMENT
Prior Damage Fire Damage

Fluid Levels OIL TRANSMISSION RADIATOR OTHER

Parts Missing

INTERIOR
Prior Damage Fire Damage

AirBags Present Sound systems and accessory

Ignition system Keys in Ignition Yes or No

Personal Content Missing

Accessories Missing

Odometer Reading Service Sticker Information

VEHICLE SECURITY
Alarm Doors and Trunk Locked Window Position

ORIGIN/IGNITION SEQUENCE
Area

Heat Source

Material Ignitied

Ignition Factor

Odometer Reading Service Sticker Information

Field Fire Investigation Report Forms v6-10

FIGURE 9-8 Vehicle fire inspection data sheet. Courtesy of Det. Mike Dalton, Know County, TN, Sheriff's Office.

374 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

 AGENCY FILE NUMBER
Motor Vehicle Fires
Items found in Vehicle
Personal Effects
Trunk/Cargo Area Contents
Aftermarket Items
Airbags In Place/Missing
Personal Effects

NOTES

DIAGRAM

Passenger Vehicles Compact–Full size Pickup Truck SUV Van Mini - Fullsize

Field Fire Investigation Report Forms v6-10

FIGURE 9-8 (Continued)

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 375

FIGURE 9-9A Vehicle with windshield and mirror blown out by FIGURE 9-9B Partially burned air bag indicates it functioned during
explosion. Courtesy John D. DeHaan. the fire. Courtesy of Jamie Novak, Novak Investigations, and St. Paul Fire Dept.

PHOTOGRAPHY AND SKETCHES

The vehicle should be thoroughly photographed inside and out before any disturbance.
Photos should capture front, rear and both sides, plus quarter views. It is most preferable
to initially examine the vehicle at the fire scene if at all possible. There, the distribution
of glass can be measured and documented. Air bag explosions are very unlikely to throw
glass fragments more than a few feet even to the rear of the car, but the directionality of
fragment distribution from such devices should be considered (see Figure 9-9). Glass frag-
ments found more than 2 m (6 ft) away are most likely to have been thrown there by a
gas or vapor explosion within the vehicle. Smoke explosions are very rare in vehicle fires,
so such glass will be soot-free, indicating an explosion prior to onset of fire, so the con-
dition of the glass should be documented. Components that have fallen off from fire dam-
age or fire suppression (or rescue) can be documented in situ.
The rectangular outline of most vehicles, with two or three “separate” areas of con-
cern, lends itself to dividing the vehicle into quadrants, as shown in Figure 9-10: four in
the engine compartment, four to six in the passenger compartment, and one in the trunk
(and one for the underside!). Depending on the vehicle, separate quadrants can be con-
sidered for the console and dash/instrument panel. A photo showing number tags or a
plan-view diagram showing the quadrant designations will make examination and note-
taking easier.
Be sure to provide full documentation of the appearance of the vehicle by means of
both photographs and sketches at all stages of all the scene examination. If the vehicle has
not been gutted, document which portions(s) is more heavily damaged.
376 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-10 “Quadrant”
1 2 5 6 system for examining and
9 documenting vehicles:
4 3 8 7 four in engine compart-
ment, four to six in pas-
Engine Passenger Trunk senger compartment, one
or two in trunk, one for
the underside. Quadrants
can be added for console
1 2 5 6 7 and instrument panel.
11 12
4 3 10 9 8

Engine Dash / Passenger Trunk

Console

IMPORTANCE OF SCENE PRESERVATION

It is always much more preferable to examine a burned car or truck before it is moved or
disturbed in any way. Valuable evidence is often destroyed or relationships lost when the
vehicle is towed to another location for examination (see Figure 9-11b). Because most
vehicle arsons take place, conveniently enough, in lonely, deserted areas protected from
witnesses by trees or heavy brush, leaving the vehicle in situ is rarely a problem of public
safety or even inconvenience. Figure 9-11c is a good example of the value of scene preser-
vation. Had the utility trailer (foreground) been moved, the origin and cause of the much

FIGURES 9-11A AND B Even if vehicle has been moved, the scene of the fire should be examined and docu-
mented. Evidence of arson may also be found there. Courtesy of Greg Lampkin, Knox County, TN, Fire Investigation Unit.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 377

FIGURE 9-11C
Destructive fire in
motorhome caused by
radiant heating from fire
in utility trailer parked
alongside. Notice contigu-
ous “V” pattern on both
vehicles. Courtesy Chris J.
Bloom, CJB Fire Consultant,
White City, OR.

more serious and costly fire in the motorhome would have been very difficult to establish.
Access to the vehicle should be tightly restricted until the exterior examination is com-
pleted. Shoe prints of the perpetrator(s) (and sometimes those of the owner!) may be
found around the car if there is soft soil. Another vehicle may have been present to
retrieve the arsonists or to tow or push the vehicle to be burned to its final resting place.
Tire impressions may be invaluable in reconstructing the crime or identifying the second
vehicle and should be preserved by photography and casting, if possible.

EXTERIOR EXAMINATION
Portions of the vehicle that may fall off during any fire—gas cap, door handles, side mir-
rors, grillwork, headlamps, and the like—should be present in the debris near or under
the car if the vehicle was not stripped prior to the fire. It is important to remember that
in the interest of saving weight, substantial portions of the exterior trim and major body
panels of current vehicles are being made from synthetic materials, which can burn almost
completely away. Cans or jars used to carry an accelerant or the remains of initiating
devices such as safety flares may be found under the car or nearby. All such evidence must
be photographed in place before it is recovered. Latent fingerprints can also survive on
the smoother surfaces of jars and cans, and this evidence should not be handled exces-
sively before such an examination is conducted. Plastic side mirrors and door handles may
melt off early and should be considered for latent print examination.

EVIDENCE OF STRIPPING
From the outside of the vehicle it should be possible to determine whether it was stripped
prior to being burned. Most vehicle arsonists cannot resist removing the more salable
items from a car or truck before burning it, thinking all signs of their removal will be
destroyed anyway. Of course, many vehicles are burned to destroy the physical evidence
relating to the original crime of theft and stripping. Are all mechanical components pres-
ent and properly installed? Are the wheels and tires correct for the car? Are all the lug
nuts present and tight? (Most people will not bother to replace all the lug nuts when put-
ting “junk” tires and wheels on a car.) Even when the tires are badly burned, the portion
378 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-12 In spite of
considerable fire damage
to a vehicle, the tread
pattern and amount of
wear is still determinable
from the pads in contact
with the ground. Courtesy
of Greg Lampkin, Knox
County, TN, Fire Investigation
Unit.

of the tread in contact with the road will not be burned, as shown in Figure 9-12. The
tread pattern and depth can be established from these unburned “pads.” Do they match
the tires that the owner says were on the car?

CONSIDERATIONS FOR FIRE INVESTIGATION

An accidental fire in the engine compartment of a front-engine car is less likely to travel
into the passenger compartment than an intentional fire (although it can happen if there
are multiple duct penetrations of the firewall or sufficient combustible materials in con-
tact with the firewall).29 Recent changes in the design of automobiles have resulted in less
separation between the luggage and passenger compartments. As a result, fire established
in one area is more likely to spread to the other than in earlier cars, no matter what the
cause of the fire.
Determine whether the windows were rolled up or down (being aware that position
of electric windows may be impossible to establish). (Is the greatest amount of broken
window glass down inside the door or inside or outside the car?) Examine the nature of
char patterns on the outside of the body. If an accelerant was poured over a vehicle, splash
and drip patterns like those in structures may be detected on exterior painted surfaces as
discoloration, scorching, checkering, or blistering of the paint, as shown in Figure 9-13
(and Figure 9-4). Weather stripping and rubber seals are usually ignited.
Examination of the vehicle’s underside is important. The paint or undercoating on the
underside of a footwell floor pan is likely to be blistered or charred if a fire was ignited
on the floor with a flammable liquid in that area, as shown in Figure 9-14.30
Fire tests have revealed that plastic and pot-metal fuel door latches can fail at rela-
tively low temperatures, allowing spring-loaded fuel doors to pop open. Filler caps are
often made of plastic today; these can melt and fall into the filler neck or burn off com-
pletely. Internal pressure can push a heat-softened cap off the filler tube. The investigator
must be careful to examine the filler tube and tank before concluding the cap was
removed deliberately.
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 379
FIGURE 9-13 Gasoline
poured on the back of
the vehicle and then
ignited scorched and
discolored paint and trim.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

FIGURE 9-14 Gasoline

burning in a footwell
often produces localized
charring or blistering on
the underside of the body
pan. Courtesy of Tim Yandell,
Public Agency Training
Council.

Char discoloration and oxidization patterns on the painted metal surfaces of the body
and fracture patterns to the windshield glass may help reveal the direction of fire spread.
If a fire starts in the engine compartment, heat conducted through body panels causes
damage to paint layers that decreases in a radial pattern as the heat moves away from its
original source. Fire progressing from an engine compartment into the passenger com-
partment will usually cause the windshield to fail at its lower edge first, with the most
intense fracturing and delamination occurring there, as illustrated in Figure 9-15. Fire
starting in the passenger compartment will tend to form a hot gas layer near the roof that
380 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-15 Engine
compartment fire pattern.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

attacks the windshield from the top and causes patterns of damage that radiate out into
adjoining body panels, as shown in Figures 9-16a and b.
During a car fire, electrical connections can be made by the charring of switches,
relays, and insulation. With a battery in place, headlights, wipers, horn, gas door releases,
and even the starter have been seen to operate. If the vehicle is left in gear, the starter may
move the vehicle some distance while burning, unless the handbrake is engaged.31 Fires
ignited in pressurized fluids can produce localized areas of damage from fire plume
impact. Fuel leaks from hoses or fuel rails may cause “‘hot spots” on engines, firewalls,
or frame components. Transmission fluid fires will often be concentrated around the dip-
stick/filler tube or the housing below. (There may be oily residues adhering to nearby
metal surfaces from such fires.) Many petroleum products fluoresce under ultraviolet
(UV) or short-wavelength visible light, so examination with a forensic light source may
be useful in spotting localized deposits.

FIGURE 9-16A Concentric passenger compartment fire pattern on FIGURE 9-16B Concentric passenger compartment fire pattern on
hood. Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. hood. Courtesy of Tim Yandell, Public Agency Training Council.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 381

 FIGURE 9-17 Extensive destruction of light-alloy and plastic components in a nonaccelerated vehicle fire.
Courtesy of Greg Lampkin, Knox County, TN, Fire Investigation Unit.

Because of the large quantity of combustible fuels present in the modern car, the dura-
tion and intensity of an extensive fire cannot be related back to the presence or absence
of a flammable liquid accelerant. The accelerant serves mainly to ensure ignition of the
substantial “natural” fuel load.32 Figure 9-17 illustrates the extensive combustion of com-
bustible automotive components—aluminum deck lids, plastic trim, composite panels,
and light-steel-alloy hoods—in nonaccelerated fires in modern vehicles. Pouring gasoline
on the outside of a closed modern car produces surprisingly little damage. Paint will
scorch or blister, plastic trim will melt or char, and weather stripping may be ignited. Tests
of modern vehicles conducted in 1992, however, showed that even a modest flaming igni-
tion source beneath a seat would result in full involvement of the interior within 2 min-
utes of ignition, whether the windows were open or not.33 The interior of a vehicle can
be thought of as a compartment that will undergo the same development as the room fire
described in Chapter 3. If provided adequate ventilation, fires inside vehicles with their
rich fuel load can progress to intense, full post-flashover fires with the same high temper-
atures (over 1,000°C) as a well-furnished room.34 Because a vehicle is dimensionally
smaller, it will undergo the process even more quickly than a typical room. The maximum
heat release rate for a fully involved modern automobile ranges from 1.5 MW to 8 MW,
with newer cars exhibiting larger heat release rates.35 Partial enclosures such as canopies,
roofs, or parking garages produce larger fires that can readily spread to adjacent vehicles
in as little as 8 minutes.
Direct exposure to fire can cause not only tires to burst but also shock absorbers, gas-
filled suspension and hatchback struts, driveshafts, and bumper (collision) impact absorbers
to explode, sometimes with dangerous force. Fire-weakened assemblies have also been
known to explode as they cool after a fire. Such explosions can inflict damage that could be
misinterpreted by an investigator, as well as pose a safety risk to fire suppression personnel
and investigators.
Vehicle Identification Number
Before proceeding to the interior examination, the vehicle identification number (VIN)
plate should be located and the identity and ownership of the vehicle determined. It is not
beyond the realm of imagination that one vehicle was substituted for another to defraud
an insurance carrier. The National Insurance Crime Bureau (NICB) and the Motor
Vehicle Manufacturers Association (MVMA) produce guides to the location and interpre-
tation of vehicle identification data plates. The NICB or the Canadian Automobile Theft
Bureau (CATB) can also assist qualified members with locating “confidential” identifica-
tion numbers hidden on every modern vehicle if the VIN plate has been destroyed.
382 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
FIGURE 9-18A Despite extensive fire damage, the base of the FIGURE 9-18B The base of the steering column air bag may
passenger air bag (dashboard) is readily visible. Courtesy of Jamie be on the wheel or on the seat or floor below. Courtesy of John
Novak, Novak Investigations and St. Paul Fire Dept. D. DeHaan.

Interior Examination of the Vehicle

As previously noted, the vehicle can best be examined in quadrants—four in the engine
compartment and at least four in the passenger compartment. Each quadrant can then be
evaluated and documented for fire damage, fuel sources, and ignition sources using the
ignition matrix concept described in Chapter 6.36 Fire damage can include total or par-
tial destruction as well as direction and intensity indicators. Because stolen vehicles are
often at least partially stripped before they are burned, the most marketable components
and accessories should be accounted for in the passenger compartment, including the
stereo, CB, CD/DVD player, seats, navigation system, radar detector, air bags, and the
like. They may be melted and badly damaged, but some remains should be on the floor
beneath the mounting brackets. The seats are very valuable and a frequent target of strip-
and-burn auto thieves. The air bag(s) will be triggered by the heat of the fire, but some
remains of the deployed bag or its base plate should remain, as shown in Figures 9-18a
and b. Because air bags are very expensive, they are often stolen for resale. Air bags func-
tioning during a fire may give bystanders the impression that an explosive device was
present. The gas generated in an air bag is nonflammable nitrogen. The remains of an air
bag may be found outside the car (see Figure 9-9b).
If the owner is claiming a loss of valuable personal property with the vehicle—tools,
clothing, fishing tackle, and so on—the metal pieces of such items should survive even the
most complete conflagration. The jack and the spare tire (or at least its metal bead wires,
belting, and valve) should still be identifiable. The ignition lock should be located even if
it has melted and fallen from its mounting hole. Sometimes an “arranged” theft and fire
will have been provided a convenient key, and this will often be found still in the ignition
switch (as in Figure 9-19). Signs of tampering with, forcible damage to, or removal of the
ignition switch and steering lock should be eliminated.

Upholstery
If the seat upholstery has not burned completely, it will behave in much the same way as
household furniture, and its examination for relative consumption and direction of burn-
ing may be carried out in the same way. Most vehicles use a pad of foam rubber over a
network of metal springs. This pad is covered in turn by a vinyl, leather, or polyester fab-
ric covering. The foam rubber can vary considerably in its flammability characteristics,
and if none remains unburned, it may be necessary to secure a control sample from a dupli-
cate vehicle for laboratory testing. It tends to burn completely away in a well-ventilated
fire in 10 to 30 minutes, no matter how the fire was started.
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 383
FIGURE 9-19 The
remains of the ignition
key are visible in the lock.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

Accelerants and Other Liquids

If a liquid accelerant was used in excess, unburned liquid may be found soaked into the
upholstery, carpets, and carpet padding or even pooled in low spots of the body pan.37 (If
the doors and windows are fully closed, the fire may self-extinguish, leaving raw
unburned accelerant, as pictured in Figure 9-20.) One study reported that the thick soot
on the interior of window glass contained identifiable residues of the ignitable liquids
used to burn two vehicles in test burns (but the duration and maximum temperatures
reached were not cited).38 It is often the case that fire went undetected and the vehicle
burned to completion, with all combustibles within the vehicle completely consumed,
including any accelerants. In such cases, flammable liquid that leaked through and soaked
into the soil under the vehicle may yet be detectable. When the vehicle is moved, the
debris and soil under it should be examined carefully, using a hydrocarbon detector if pos-
sible. Samples should be collected and sealed in airtight metal cans or glass jars if there
are any indications of flammable liquids. Remember that a leaking fuel tank can simulate
such conditions, and a control sample of the fuel should be taken in most instances.
Electric fuel pumps may continue to pump as long as the ignition is on and will empty the
entire tank under the car if the electrical system remains intact and there has been no
severe mechanical impact that would have triggered the interrupter fitted to some cars.
(In many vehicles there is an interlock, so the loss of oil pressure or electrical signal from
a stopped engine will shut the fuel pump off as well.)

Interior Debris
Portions of timing and ignition devices may be discovered in the interior debris. Safety
flares make excellent ignition devices (and they are innocuously present in most vehicles)
but a careful search can reveal their wooden end plugs and white or gray lumpy residues
(see Figures 8-19 and 8-20). Paper matches survive even substantial fires if they are rest-
ing on a noncombustible surface and are worth looking for. Crumpled paper makes a
good, fast initial fire. Setting fire to a small plastic bag of potato chips provides more than
384 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-20 Windows
and doors were closed
after gasoline was ignited
in front seat. Fire self-
extinguished, leaving raw
gasoline on the rear seat.
Courtesy of Tim Yandell, Public
Agency Training Council.

enough heat to ignite the upholstery in a modern car and yet burns away to an amor-
phous, unrecognizable carbonaceous ash. The oils in the chips produce a prolonged,
intense fire because they are distributed on a thin, porous, combustible wick. The residue
is easily overlooked and, even if discovered, may be dismissed as accidental. The vegetable
oils present, however, may be identifiable in the carpet if it is analyzed by gas chromatog-
raphy/mass spectrometry (GC/MS) by the same method as used for self-heating oils (see
description in Chapter 14.) Fires under seats can also be accidental—from hot work done
recently, rags left on exhaust pipes, overheated converters, or service lights left on inside
the vehicle. Sometimes, an attempt is made to simulate an accidental fire by loosening fuel
lines or drain plugs or disconnecting them completely. All such plugs and connectors
should be accounted for and examined for the presence of loosening or fresh toolmarks
indicating recent tampering. One multiple-vehicle arson involved the removal of fuel tank
caps from a line of diesel tractor-trailer trucks. The fuel tanks were connected by means
of fuel-soaked rag trailers, and a series of the tanks were ignited by floating a plastic cup
with a small firecracker and blackpowder in each. The resulting fires destroyed a number
of the vehicles, despite the failure of most of the rag trailers.

Mechanical Condition
If it is possible, the mechanical condition of the vehicle prior to the fire should be estab-
lished. Mechanical problems may provide a motive for “selling” the vehicle to the insur-
ance company, as well as a possible cause of an accidental fire. Observations should
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 385
 include whether the battery is in place and connected and whether the engine’s electrical
connections are still intact. Even if the wires melt or burn away, the metal connectors gen-
erally remain. Are all the engine components and accessories present? Do the water, oil,
and transmission fluid levels indicate the vehicle was operational at the time of the fire?

Mechanical Examination
Disassembly of the engine or gearbox may reveal the operational status, which can then
be compared against the owner’s statements. Engine oil and transmission fluid should be
drained. Simple observations for condition and contaminants can be carried out by eye,
but such fluids can also be laboratory tested to reveal mechanical malfunctions prior to
the fire. If the owner claims the vehicle to be complete and operational, a check of the
engine compartment may reveal missing, disconnected, or out-of-place components that
would render the vehicle inoperative.39 Accidental fires involving the automatic transmis-
sion or power steering unit during operation usually produce severe damage inside the
unit as fluid is expelled. External fires do not do the same kind of internal damage.
Analysis of the oil or fluid may reveal the mechanical state of the unit.

Theft and Other Non-fire Damage

Non-fire-related body damage should be documented and compared against the owner’s
description of the condition of the vehicle. Previous damage may be detectable from part
numbers or inventory. Body damage that is costly to repair but makes the car unsalable
may be reason for destroying the car by an “accidental” fire. As in the investigation of
structural fires, it is important that the investigator look for anything that seems unusual
and out of place. No matter how clever the vehicle arsonist is, it is very difficult to arrange
a destructive fire that can pass for an accident without modifying the vehicle or adding
something to it to help it burn.
If the vehicle was found burned and the owner claims the vehicle had been stolen,
there are additional avenues of inquiry. The scene where the vehicle was found (even if
the vehicle has already been removed) should be examined as a potential crime scene for
shoe and tire impressions, tools, accelerant containers, and means of ignition (match-
books, lighters, flare striker caps). The area should be canvassed for possible witnesses
who can help establish time factors or information about people or vehicles in the area at
the time. The person who found the vehicle (or reported the fire) should be interviewed
as to how they came to be in that location. The owner should be interviewed about the
circumstances of the loss (Who last drove the vehicle and when? Was the vehicle locked?
Was the alarm set? What were the vehicle contents? At what time and location was the
vehicle last seen? At what time was the theft discovered? At what time was it reported?).
The status of keys for the vehicle and access by others should be established. Modern theft
or GPS/navigational systems may provide data about the movements of the vehicle prior
to the fire. Some vehicles are also equipped with event data (black box) recorders that
capture vehicle operational data before a crash! These recorders require specialized equip-
ment and expertise for access. A new ASTM monograph describes the examination of
vehicle data (black box) recorders from accident vehicles.40

Interviews
Wendt has published a comprehensive outline for interviewing the victim of a vehicle theft
and fire.41 Interviews may include the vehicle’s owner, driver, occupants, and mechanic.
Responding police and fire personnel are also important sources of information.

NHTSA Information
The National Highway Traffic Safety Administration (NHTSA) should be consulted for
recalls or problems associated with the type of vehicle involved. One must be very careful
not to assume that a recall proves the fire in question was caused by the failure reported!
Such recalls must be included as a testable hypothesis in complete investigation.42
386 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
Vehicle Fire Tests
Considerable experience has been gained from laboratory tests involving the burning of
complete automobiles under various conditions. These tests reveal that, for the most part,
the amount of structural damage done to the exterior of the vehicle does not bear a direct
relationship to whether the fire is of incendiary or accidental origin. Some tests have
revealed that temperatures near the roof of a burning car reach as high as 1,000°C
(1,860°F) at some stage of a fire no matter what the means of ignition.43 This is more
than sufficient to warp sheet metal and melt glass and aluminum trim. As a result, body
and roof panels buckle and warp, window glass melts and icicles, and seat springs lose
their temper at temperatures developed in slow-burning accidental fires as well as in fast,
accelerant-fueled fires. Similarly, the color, density, and appearance of soot and pyrolysis
products on remaining windows depend largely on the upholstery materials present and
the degree of ventilation present during the fire. Although the amount of structural dam-
age will, of course, be generally related to the intensity and duration of the fire, those fea-
tures are not necessarily determined by whether or not the fire was incendiary. One series
of tests indicated that, in fact, more complete damage can result from a slow-developing
accidental fire that has more time to bring more of the entire mass of the vehicle up to
high temperatures than from a fast-moving gasoline fire.44

Motorhomes and Other Recreational Vehicles

A recreational vehicle is defined as a motor vehicle containing living quarters for recre-
ation, camping, travel, or seasonal use. Recreational vehicles come in different styles,
shapes, sizes, and with different accessories. They range from a simple tent trailer to a
bus-chassis motorhome. The basic designs are the camping trailer, truck camper, travel
trailer, fifth wheel, park trailer, and the motorhome.

CHARACTERISTICS OF MOTORHOMES
Most motorhomes are constructed and assembled from four distinct and important compo-
nents, often made by different manufacturers: the unit’s engine; the transmission; the chassis;
and the remaining portions of the unit, including the assembly and installation of appliances
and furnishings. It is not uncommon for there to be several fire investigators examining the
remaining debris, each one representing the individual interests of each manufacturer involved.
Motorhomes are categorized into three distinct classes: Class A, Class B, and Class
C. Class A motorhomes are the largest units and are designed and built from the ground
up. Class B motorhomes are van conversions that have been altered to accept appliances
and furnishings for the living space. Class C motorhomes are units that incorporate a van
cab portion, with an attached body, usually with an over-the-cab sleeping area.

FIRE RISK
Motorhomes are subject to a combination of the risks of accidental fires of vehicular ori-
gin with those peculiar to the heating, cooking, and living conditions found in dwellings.
Because the interiors of these vehicles are used for day-to-day living, the hazards of cook-
ing and smoking materials are as present as they are in a stationary dwelling. Many such
vehicles have moderate-capacity electrical systems of 12 V, so electrical fires, while not
completely impossible, are quite rare. Some vehicles have large-capacity, complex electri-
cal systems of 24 V, which provide higher risks. In addition, many of them are fitted with
a parallel 120 VAC system to operate lights, heaters, radios, and the like when they are
connected to an outside source of AC power or to a portable generator mounted within
the vehicle. Many of the same appliances found in the home will be found in the
motorhome. High-current devices such as televisions, toasters, and microwave ovens are
common, as well as their attendant risks of misuse.
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 387
 PROPANE TANKS
Most such vehicles also include a propane (or LPG) tank to provide fuel for the heater,
stove, refrigerator, and even generator with the attendant hazards of explosions from
spills and leaks. If factory equipped, the recreational vehicle will be fitted with either
DOT or ASME tanks, steel and/or copper piping, rubber hose connections with brass fit-
tings, valves, and the appliances they serve, including a refrigerator, furnace, stove, and
even a generator. The size of the propane tanks can vary from 5-gallon tanks on trailers
to upward of 80-gallon tanks on the more expensive motorhomes and bus conversions.
There is no guarantee that propane systems will be properly installed in amateur conver-
sions. One explosion, costing 12 lives, involved a modified motorhome. The cause was
traced to an improperly installed LP tank that leaked its contents into the vehicle until it
was ignited by an electric arc or open flame as the vehicle proceeded down the highway.45
All propane tanks are equipped with some type of a pressure-relief valve, which is
designed to vent the contents of the tank at a certain pressure. When the propane tank
heats up due to flame impingement, the pressure inside increases. When the pressure
inside becomes greater than the pressure-relief valve setting, the valve opens and rapidly
expels the contents away from the motorhome and tank. This expulsion will continue
until either (1) the pressure inside the tank is less than the pressure-relief valve setting, at
which time the valve will close again, or (2) the valve expels all the contents of the tank
during the fire. See Chapter 4 for a full discussion of propane tanks.
Many motorhomes and travel trailers today are fitted with propane detectors. The
location of these must be carefully assessed, for thermal currents in the vehicle may move
the fuel/air vapor away from the detector. Recall that the density of an ignitable propane/
air mixture is very close to that of air, so it will not settle like pure propane.
Manufacturers make great use of plastics, synthetic fabrics, foam rubber, and parti-
cleboard to save both cost and weight. The structural elements are most often a skeleton
of aluminum tubing to which is riveted an exterior aluminum or fiberglass skin. Plywood
or synthetic paneling is fitted to the inside, and thermal and sound insulation is provided
by fiberglass or polystyrene or polyurethane foam sandwiched between those layers. Once
the main fabric of such a vehicle is ignited, there is little to prevent it from being com-
pletely destroyed. While the vehicle is burning, formaldehyde and various cyanogens are
being released from the interior finishes, sometimes with fatal results to any occupants.
Although most reputable manufacturers strive to make their vehicles as safe as possible,
there is no code applicable to custom conversion facilities, and the fire investigator must
be alert to the possibility of finding patently hazardous practices and products in such
vehicles. The competition in this area is fierce, and the most cost-effective use of materi-
als, even to taking shortcuts, is the key to survival.

CONSIDERATIONS FOR FIRE INVESTIGATION

Almost every fire involving a recreational vehicle will develop very quickly due to the
lightweight construction utilizing foams, resins, plastics, and other highly combustible
materials. The federal standards for allowable flame spread rate are on a scale from 25
to 200, with 200 being the maximum allowed. Due to the lightweight construction meth-
ods and materials employed, a flame spread rating of 200 is allowed for recreational vehi-
cle construction. Because of the materials involved, when a small fire occurs, it develops
rapidly and quickly involves the gasoline or diesel fuel tanks and the propane system. It
is not uncommon for a small fire to completely consume a recreational vehicle down to
the frame within 12 minutes.46
When investigating fires in such vehicles, the investigator should expect substantial
destruction of their structural elements, with consequent loss of evidence that would nor-
mally permit a reconstruction of the fire sequence. Within the limitations of the construc-
tion materials involved, diagnostic signs of fires caused by electrical failures, smoking
388 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-21 Remains
of a Class A motorhome
ready for examination.
Courtesy of Greg Lampkin,
Knox County, TN, Fire
Investigation Unit.

materials, heating systems, and arson remain largely as they are described elsewhere in
this book. The investigator must be aware of the extensive fire damage resulting from the
construction of such vehicles, as shown in Figures 9-21 and 9-22.
Photographs should be taken at every step prior to, during, and after the investigation
(including the general scene). This is to allow documentation and reconstruction of the
scene as found. Valuable information that was not readily apparent during the initial
examination could easily become visible after the photographs are developed. When exam-
ining the remains of a motorhome, it is also important to examine the scene where the
fire occurred, because critical evidence may have dropped off and away from the vehicle

FIGURE 9-22 The

remains of a large
motorhome being
properly transported on
a flatbed carrier. Total
destruction resulted from
an accidental origin.
Courtesy of Joe and Chris
J. Bloom, Bloom Fire
Investigation, Grants Pass, OR.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 389

FIGURE 9-23 A
passerby recorded this
accidental fire in a Class
C motorhome, which
destroyed the vehicle.
Such recordings preserve
indicators such as direc-
tion of fire travel, time
involved, suppression
attempts, and wind
direction. Courtesy of Joe
and Chris J. Bloom, Bloom Fire
Investigation, Grants Pass, OR.

during the fire and its suppression. It is most important when inspecting the unit, unless it
is absolutely necessary to move them, to leave the remains undisturbed. If items or debris
need to be moved to allow for an examination, only professionally accepted practices for
the documentation and removal of the debris should be employed. When the unit needs to
be moved from the scene location to a storage facility, the unit should be moved via a flat-
bed towing truck and tarp-wrapped, if at all possible, to minimize potential loss or dam-
age to components. The reader is referred to specialist references for guidance.47 Because
of the enormous variety of styles fittings and interior finishes, it is often helpful to locate
and inspect an exemplar unit of the same type, floor plan, and age. Photos or videos taken
by owners or passers-by may capture vital data, as illustrated in Figure 9-23.

Mobile Homes (Manufactured Housing)

CONSTRUCTION AND MATERIALS
In many respects, the structure and materials of mobile or manufactured homes (some-
times called modular homes), the causes of fires within them, and the problems of inves-
tigating such fires are similar to those just discussed in regard to recreational vehicle fires.
Structurally, manufactured homes are most often built on a heavy steel frame supported
on three axles. A wooden floor structure of plywood or particleboard provides a base for
the erection of a wooden or metal framework upon which the exterior aluminum or
wooden sheathing is installed. The interior walls are often framed (as are exterior walls)
with 2-in. × 3-in. wood studs and then covered with plywood, wood paneling, or more
recently, gypsum board specially designed to resist the flexing and vibration induced by
moving. Insulation is often fiberglass batting in the roof and exterior walls. The interior
finish in many such mobile homes is similar to the paneling of motorhomes and
thus poses a high risk of rapid fire development. The occurrence of fully involved
(post-flashover) fires seems to be much more common in mobile homes, and the time
to flashover is much reduced over that seen in most standard buildings. The small-
dimension wood members burn quickly and allow more rapid structural collapse than in
a structure of standard construction. With the advent of NFPA 501B: Manufactured
390 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
Housing (imposed by the U.S. Department of Housing and Urban Development on man-
ufactured homes starting in 1976), the speed and severity of development of fires has been
notably reduced.48 The date of manufacture of the structure being investigated is there-
fore critical to estimating what the materials and construction would be (as well as smoke
detectors and the like). The recent development of gypsum wall finishes suitable for man-
ufactured housing has made fire performance of such housing more closely parallel to that
of fixed structures.
The tubular nature of the interior structure uninterrupted by firestops provides for
rapid fire spread throughout, and combustible wall finishes and low ceilings (typically 7 ft)
provide for extremely rapid buildup and flashover conditions. Davie has reported meas-
uring temperatures at floor level of a mobile home equal to those at the ceiling—750°C
to 1,000°C (1,400°F to 1,800°F). This is due to the size and shape of the rooms, the com-
bustible nature of wall and flooring materials used, and the reflective properties of the
outer walls and roof.49
The development of fires in such structures to post-flashover follows the same pattern
as in fixed structures but more quickly. The location, thickness, and spacing of the struc-
tural members and the installation of electrical and heating equipment are considerably
more variable in mobile homes than in conventional structures. (Walls are often framed
and erected after carpet is laid across the entire floor area, for instance.) The motion pro-
duced when the vehicle is moved or when fans, ventilators, or air-conditioning units are
operated can chafe wire insulation or loosen connections far more than would be expected
in fixed structures.

CONSIDERATIONS FOR FIRE INVESTIGATION

It is thought that the small rooms and narrow hallways typically found in these structures
contribute to the rapid development of flashover and unusual burn patterns, such as those
observed in Figure 9-24, where the most intense fire was to the floor and wall to the right
of the bed and not at the origin. The minimal insulation and conductive roof covering
(typically aluminum or steel sheathing) may allow fire damage to the roof to record the
development of fire in the rooms beneath. An investigator should be wary of placing too

FIGURE 9-24 Fire in

the bedroom of this
mobile home started near
the floor to left of bed.
The dynamics of air
entrained from the
entry door during post-
flashover burning caused
a false “V” of severe burn
to the right of the bed.
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 391

FIGURE 9-25 Despite a
relatively short burn time
of 20 minutes, this fur-
nished mobile home was
nearly totally destroyed.
Burn patterns on the
floor are due to radiant
heat. Fire ignition was by
means of a small wood
crib on the kitchen
counter behind and to
right of the photogra-
pher. Low burns, con-
sumption of the sofa
against the far wall, and
penetrations of the floor
are the result of fire
spread. Courtesy of John
J. Lentini, Applied Technical
Services Inc., Marietta, GA.

much importance on the presence of complex burn patterns on the floors of a heavily
damaged mobile home. The combustible floor materials and wall coverings and high fuel
loads combine to produce intense post-flashover fires that in some cases can penetrate the
floors under furniture as well as in exposed floor areas. (Pre-fire degradation of floors
from water incursion will accelerate floor combustion and failure.) Rapid failure of single-
glazed windows adds ventilation sources. Post-flashover burning (especially when synthetic
carpet and pad are involved) will produce irregular burn patterns on floors, as shown in
Figure 9-25. Although the fire in a mobile home behaves in a manner consistent with
basic scientific and engineering principles, the investigator must be careful not to use the
same criteria for evaluating indicators of time and estimating fuel effects as used in eval-
uating fires in buildings of standard construction.
When investigating fires in mobile homes, the investigator should expect substantial
destruction of their structural elements, with consequent loss of evidence that would nor-
mally permit a reconstruction of the fire sequence. Within the limitations of the construc-
tion materials involved, diagnostic signs of fires caused by electrical failures, smoking
materials, heating systems, and arson remain largely as they are described elsewhere in
this book.

Heavy Equipment
Motorized equipment such as tractors, loaders, earthmovers, forklifts, harvesters, and the
like can pose special fire risks. If not cleaned properly, agricultural equipment can accu-
mulate thick layers of cellulosic dusts that can ignite from hot surfaces (as shown in
Figure 9-26). Excess dusts and fibers can accumulate in moving parts to cause frictional
heating in belts, bearings, and gearsets. Improperly maintained hydraulic and oil systems
cause special problems, as they operate at very high pressures when in use. The extremely
fine aerosol mists that are created at such leaks are readily ignited by hot surfaces or elec-
tric arcs. Vibration causes loosening of fittings and fracturing of rigid pipes—all leading
to loss of fluids under pressure. Fires started from hydraulic or other fluid leaks can
spread very quickly to engulf the equipment. Fires started in solid dusts grow more slowly
but can be ignited for some time after the equipment is shut off and left unattended. Either
392 Chapter 9 Automobile, Motor Vehicle, and Marine Fires
 FIGURE 9-26 Fire in
harvester caused by
accumulation of com-
bustible dust. Courtesy of
Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

ignition can lead to a major fire involving expensive equipment. The size of combustible
components is a factor in determining the size of the resulting fire. Heat release rates as
high as 3 MW have been observed in tests of single earthmover tires (ignited by a pool of
diesel fuel beneath the tire.)50 Large tanks of diesel fuel or hydraulic oil (or propane cylin-
ders to aid in cold-weather starting) can also contribute to the fire’s intensity and dura-
tion. The investigator must be very familiar with the specialized equipment used in such
vehicles before attempting to determine the origin and cause of major fires.

Boats and Ships

Pleasure boats are more like automobiles in their relative fire hazards and susceptibility
than dwellings or larger ships. The boat and the automobile most often have a gasoline-
powered motor, gasoline tank, and electrically driven accessories. There is one major dif-
ference that makes the boat fire more destructive and much more difficult to investigate
than the auto fire, namely, all the flammables are enclosed in a more-or-less tub-shaped
container—the hull. Leaking gasoline or its vapors will settle into the bottom of the hull,
where they remain unless circulated by drafts from the engine fan or bilge pump.
Comparable leakage in an automobile is generally to an open or well-ventilated exterior.
This is the basic reason for the greater danger and likelihood of explosion of fuel vapors
in a boat. In addition, many boats are made of wood or fiberglass, while the automobile
does not usually consist of such a combustible “container.”
A gasoline fire in a wooden or fiberglass container is a far more formidable occur-
rence than the same fire in a steel conveyance. Wooden decks and cabins are likely to be
burned completely, and the hull is likely to be burned down to the waterline. Fire patterns
are therefore obscured or destroyed, and points of origin are often not identifiable.
Smaller rooms and the presence of all-wood interiors will produce more intense, faster
growing fires than normally found in the larger, gypsum-wallboard finished compart-
ments of buildings. With larger wooden vessels that have considerable secondary interior
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 393
 structure, patterns will be found that are in every way comparable to those in any other
similar wood-framed structure. Such fires can be investigated in accordance with the prin-
ciples that are detailed throughout this book.
In addition to sources of gasoline vapors, boats may have heaters or stoves that are
fueled by kerosene, alcohol, or LP gas. Each of these poses risks of accidental fires and
must be considered when investigating these fires; LP gas can settle into the hull like gaso-
line vapors, and kerosene or alcohol stoves can spill or leak, producing pools of ignitable
fuel. The electrical systems may include separate systems for 120 VAC and 12 or 28 VDC
power. On newer vessels, these electrical systems must meet the requirements of the U.S.
Federal Boat Safety Act of 1971, but earlier ones need meet no requirements, and modifi-
cations may or may not have been done correctly. This same act requires flame arrestors,
appropriate fuel lines, and other safety equipment. Minimizing the risk of fire and loss of
life on such crafts is the focus of NFPA 302: Fire Protection Standard for Pleasure and
Commercial Craft.51 Like standards for structures, such guides can only describe recom-
mended (or required) practices, not capture the content and fire risks as seen in real-world
occupancies. The investigation of boat fires involves the evaluation not only of the pres-
ence of such features but their operability and effectiveness (e.g., flame arrestors may be
present but are compromised by a hole punched through them by a careless mechanic).
One survey of small-vessel fires reported that 55 percent were caused by wiring and appli-
ances (both AC and DC), 24 percent were caused by overheating of engines or transmis-
sions, 8 percent by fuel leaks, and 1 percent by stoves.52 Some vessels, such as the
houseboat shown in Figure 9-27, combine entirely combustible (exposed wood) structure
and linings with large windows and numerous household ignition sources and corroded
and contaminated electrical fittings, leading to a high risk of very intense, destructive fires.

SHIPS
Although large steel-framed ships would not be expected to burn, it is fortunate that most
fire investigators are not called on to examine such fire scenes frequently, because such

FIGURE 9-27 Houseboat

gutted by accidental fire.
All-wood structure on
concrete “bathtub” hull.
Large windows assured
excellent ventilation as
they failed in the fire.
Courtesy of John D. DeHaan.

394 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

ships and their flammable contents can burn with incredible ferocity and massive destruc-
tion. This destructiveness is based largely on three factors: fuel load, spread of fire by con-
duction, and inhibited access to firefighters, which are all peculiar to large steel ships.
Although the steel structure itself is not readily ignitable, a considerable fuel load is pres-
ent in the ship’s interior finishes, furnishings, cargo, and fuel. The interior finishes and fur-
nishings represent the same degree of combustibility hazard as comparable furnishings in
a dwelling. A large cargo of freight, however, imposes its own flammability characteristics
on the ship. In bulk shipping of products like wood chips, grain, coal, and natural fibers,
smoldering fires are common and are very difficult to extinguish. Water tends to wet only
the top layers and not penetrate, and extinguishing gases are expensive. In most cases, the
holds with the smoldering cargo are sealed and left, hoping the fire will self-extinguish.
Unfortunately, due to the low oxygen limits for smoldering combustion, considerable com-
bustion can occur, creating heat and combustible gases that are trapped in the hold. When
fresh air is admitted, a backdraft or smoke explosion can result with disastrous effect. Self-
heating of bulk cargoes is also a problem, particularly in hot climates. Steam, nitrogen, or
carbon dioxide gas injection is sometimes used to suppress cargo compartment fires. The
cargo itself can contain flammable or explosive vapors, as illustrated in Figure 9-28.
Explosive dust suspensions are encountered during loading and unloading of flour,
sulfur, and similar finely powdered products. Such products as ammonium nitrate, used
for chemical and fertilizer production, are well recognized for their hazardous properties,
and special precautions are required.53

TANKERS
Tankers, often loaded with very large quantities of fuel, generally of petroleum origin,
might be expected to be subject to enormously increased hazards. The main cargo of a
tanker while contained is no more dangerous than is the gasoline tank on an automobile
and for the same reason. Bulk cargo may be contained in freestanding containers or in
open holds. Many tanks are provided with inert gas (such as nitrogen) systems to exclude

FIGURE 9-28 Explosion

in container midvoyage
caused by leakage from
used gasoline tanks being
shipped for recycling.
Courtesy of Dennis Field, Fire
Cause Analysis.

Chapter 9 Automobile, Motor Vehicle, and Marine Fires 395

 air from the ullage spaces. However, the danger of a fuel spill above board, where the
vapors can mix with air, creates a hazard similar to that produced by a fuel spill in the
exterior in any other environment. Ignitable liquid cargoes require that special precau-
tions be taken. Far more serious hazards are associated with a rupture in the hull of a
tanker that releases great quantities of fuel or with tanks that have been unloaded, where
air can mix with the fuel vapors, forming an explosive mixture.

SHIP CONSTRUCTION AND FIREFIGHTING TECHNIQUES

Once a fire has started, by whatever cause, extinguishment is very difficult and hazardous.
The modern welded-steel ship can be thought of as a huge box girder with the steel plates
of the sides, bottom, and main deck forming the four sides. This box is strengthened by
steel girders forming the keel or “backbone” and the frames or “ribs.” It is separated into
many smaller compartments by the lower decks and transverse bulkheads.54 These ele-
ments are part of the strengthening substructure and are also all steel (or aluminum in
some ships). Unlike a fire in a frame or masonry structure, fire in one compartment of a
ship is not thermally insulated from neighboring compartments. The steel plates conduct
heat quickly and efficiently to other adjacent or overhead compartments. Combustibles
in these compartments quickly reach their ignition points and contribute to the fire
spread. Damage may be more extensive in compartments above or alongside the point of
origin than in the actual compartment where the fire started. Tunnels and vertical chases
for electrical and hydraulic systems provide chimneys for the spread of fire by convection
or even direct flame impingement.

TERMINOLOGY
Certain terms peculiar to vessels are helpful to the fire investigator in interviews (see Figure 9-29).
Some of these terms include the following:

Hull: Main body of a vessel—may be steel, aluminum, wood fiberglass, concrete

Deck: Horizontal element separating levels
Bulkhead: Vertical partition separating compartments
Bridge: Location from which the vessel is steered and controlled
Cockpit: In small boats, the opening in the deck from which the vessel is steered
Ladder: Vertical access between decks
Passageway: Horizontal access between compartments (hallway)
Porthole: Transparent windows from a compartment through hull
Galley: Kitchen area
Boat: Smaller water vehicle (sometimes defined as capable of being carried by a ship)
Ship: Larger, oceangoing vessel

Source: F. Herrera, “Investigation of Boat Fires,” California Fire/Arson Investigator (January 2005): 10–11.

Superstructure Bridge
Port Starboard or Cabin (in boats)

Gunwale
Stern
Main deck (Aft)
Hull
Keel
Rudder
From rear Bow/Stem (Forward) Amidships Hold Engine bay

FIGURE 9-29 Terminology for ship construction.

396 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

 Standard firefighting practice in dwellings often calls for getting above the fire to ven-
tilate it and thereby to cool it. In a steel-compartmented ship, access above the fire is often
restricted, and entry may not be achieved by anything less than an oxyacetylene torch. In
such conditions, the spread of fire upward through a ship may not be preventable, and
efforts will be directed at removing cargo and reducing loss of life.
Fires aboard very large modern cruise ships combine firefighting difficulties with huge
life safety concerns in which over 5,000 passengers may be at risk.55 Fires on such ships
have been started by discarded smoking materials, welding repairs, laundry fires, and
arson. NFPA 301 and the International Maritime Organization (IMO) Convention for
Safety of Life at Sea ((SOLAS) 1992) are the predominant codes for such ships.56
Idle vessels are often moored side by side or berthed alongside disused piers, where
fire can spread from pier to ship or from ship to ship by radiant heat, flaming debris, or
direct flame contact. The investigator must be aware of the mobility of such vessels
(before and after the fire) and that the fire may not have been an isolated occurrence. The
original location of the vessel must be determined as part of the investigation. Because of
the unusual nature of the fire conditions aboard such vessels, the investigator is directed
to seek the assistance of specialists in this area.57

Motives for Vehicle and Marine Arson

As in any other kind of fire investigation, the intimation of a possible motive for arson
may provide a significant turning point for any vehicle fire investigation. Although proof
of motivation is not necessary for the conviction of an arsonist, it can pull all the pieces
of an investigation together. Many of the same motives discussed in Chapter 16 will apply
to vehicle arson. Insurance fraud, either involving the owner directly or the financial insti-
tution that is carrying the loan, is certainly a consideration. The fuel crisis of the 1970s
and 2006–2009 prompted many owners of large, gasoline-thirsty vehicles to dispose of
them. Government-imposed requirements for mandatory emission-control modifications
may also force owners to abandon vehicles. Depressions in the general economy in the
1980s (and currently) forced people to acknowledge that they had paid too much for a
vehicle or that the monthly payments were too high to maintain. With a substantial
increase in the number of leased cars; changes in the tax deduction bases for cars, trucks,
and other vehicles; and changes in the marketability of expensive or fuel-inefficient cars
of recent vintage, vehicular arson is more common today than ever before. Faced with a
substantial payoff due on a leased car or truck or when stuck with a relatively new vehi-
cle whose resale value has plummeted, an individual or a company may see “selling” it to
the insurance or leasing company as a convenient way out. Large diesel tractors routinely
cost over $150,000 today, and a major engine rebuild can cost over $20,000. If a com-
pany or private trucker is operating on a thin margin, the loss of the vehicle to a conven-
ient fire may be the solution to financial problems. Vehicles are also subject to arson
motivated by a number of other reasons—spite or hatred, intimidation, vandalism, con-
cealment of a crime, or even pyromania. Vehicles stolen for joyriding, for parts, or for the
commission of other crimes are often burned to destroy fingerprints or other associative
evidence.
Inquiries as to the purchase price of the vehicle, standing loan balance, and monthly
payments should be accompanied by a determination of the fiscal well-being of the owner.
Are there indications of reduced income, unemployment, business losses, or gambling
debts? If a vehicle is mechanically unreliable or inoperative, it is more likely to be a can-
didate for arson. The same holds true if the vehicle is no longer used or has outlived its
usefulness. Domestic problems are sometimes behind the destruction of an automobile,
recreational vehicle, or pleasure boat. A vehicle that has sentimental value to one partner
may be destroyed by the other as a gesture of spite or anger. Any indications of marital
difficulties should prompt a closer examination of the circumstances of the fire. As with
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 397
 other fires, the investigator must always be careful to keep motive and proof of cause sep-
arate and distinct determinations.
With fires in large commercial ships, any of the same reasons that prompt the burn-
ing of land-based properties can hold true. Business-related motivations can include
intimidation, elimination of competition, labor unrest, or the concealment of other
crimes, particularly theft, pilferage of cargo, and smuggling of goods or drugs. The finan-
cial motivations are even stronger when entire fleets are idled by international trade or
monetary reversals. The thorough investigator must consider all possibilities when the
values of vehicles and their cargoes reach into the thousands or millions of dollars, thus
providing rich rewards to those clever enough to escape detection. Fires have been set
aboard military vessels to delay deployments.58
At the same time, the investigator must remember that a motive alone does not
change a fire of undetermined cause into an arson. The origin and cause of a fire is estab-
lished from careful data collection and proper analysis of the evidence at the scene. There
are many ignition sources in the average vehicle or vessel, rich fuel loads, and many ways
for accidental fires to ignite. The investigator has to be just as diligent in establishing the
first fuel ignited (suitable in both type and quantity), a competent heat source, and the
mechanism for contact in these fires as in a structure fire before concluding the cause of
the fire. If not suppressed shortly after ignition, these fires can result in such complete
destruction of the vehicle or vessel and evidence of the fire that an origin and cause cannot
be reliably established.

398 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

 CHAPTER REVIEW

Summary

With their combustible interiors, ignitable liquid heat hypotheses, the investigator must be aware of the igni-
sources, moving parts, and active electrical systems, it tion properties of a range of fuels commonly found in
is not surprising that motor vehicles catch fire. Intense, vehicles and the dynamics of fire spread and effects on
rapidly developing fires can result if sufficient ventila- materials. The investigator must also be very familiar
tion is available, and they can bring about nearly total with the wide variety of complex electrical and
destruction in a short time. (Peak heat release rates of hydraulic systems and accessories in modern vehicles
modern vehicle fires have been measured at between and how they can ignite and burn. As with structure
1.7 and 7.5 MW.)59 To carry out a scientific fire inves- fires, total destruction and rapid fire development are
tigation and properly create and test a full range of not proof of an incendiary fire.

Review Questions

1. What are the most readily ignitable liquids found an accelerated fire in a modern vehicle? Why or
in a modern auto? why not?
2. How do the boiling points of liquid fuels affect 7. Name five of the items most commonly removed
their hot-surface ignitability? before a stolen vehicle is set afire.
3. Which produces the higher internal temperatures 8. Why is it nearly impossible to cause an explosion
in a modern vehicle fire—a flammable liquid of a gas tank?
arson or direct ignition of the upholstery—and 9. Why do recreational vehicle fires develop
why? quickly?
4. Why is it important to examine the scene of a 10. What makes extinguishment of a ship fire so
vehicle fire before the vehicle is removed? difficult?
5. How can transmission fluid become ignited?
6. Does extensive combustion of body panels and
structural and suspension components indicate

References
1. M. J. Karter, “Fire Loss in the U.S, 2007,” NFPA 4. NFPA, Fire Protection Guide to Hazardous Materials,
Journal (September–October 2008): 46–51. 2001 ed. (Quincy, MA: National Fire Protection
2. R. J. Hrynchuk, “A Study of Automobile Fires,” Association, 2001).
Canadian Society of Forensic Science Journal 11 5. M. Higgins, “Vehicle Fires: A Practical Approach,”
(1978): 15–22; K. Allen et al., “A Study of Vehicle Fire and Arson Investigator (Spring 1993).
Fires of Known Ignition Source,” Fire and Arson 6. J. H. Potter, Ford Motor Company Specialty Products
Investigator 34 (June 1984): 33–43, 35; (September Engineering, personal communication, 1988.
1984), 32–44. 7. C. Cope, “Vehicle Engine Compartment Fires” in
3. American Petroleum Institute, “Ignition Risk of Proceedings Fire & Materials 2009, (London:
Hydrocarbon Vapors by Hot Surfaces in the Open Interscience Communications, 2009), 561–79.
Air,” API Publication 2216, 2nd ed., (Washington, DC: 8. G. March, An Investigation and Evaluation of Fire
API, January 1991); N. R. LaPointe, C. T. Adams, and and Explosion Hazards Resulting from Modern
J. Washington, “Autoignition of Gasoline on Hot Developments in Vehicle Manufacture, Newcastle,
Surfaces,” Fire and Arson Investigator (October 2005): Tyne and Wear Fire Brigade, UK, 1993.
18–21; J. D. Colwell and A. Reza, “Hot Surface 9. F. J. Suefert, “Automobile Engine Fires,” Fire and
Ignition of Automotive and Aviation Fluids,” Fire Arson Investigator 45, no. 1 (June 1995): 23–26;
Technology 41 (2005): 105–23.
Chapter 9 Automobile, Motor Vehicle, and Marine Fires 399
 “Automobile Engine Fires (Part 2),” 46, no. 1 36. Morrill, “Analysis of a Ford Speed Control
(September 1995): 23–26. Deactivation Switch Fire.”
10. Ibid. 37. A. F. Service and R. J. Lewis, “The Forensic
11. Ibid. Examination of a Fire Damaged Vehicle,” Journal of
12. R. K. Marley and J. D. DeHaan, CCAI Vehicle Fire Forensic Sciences 46 no. 4 (2001): 950–53.
Tests: A Photographic Report (Sacramento, CA: 38. D. Sutherland and K. Byers, “Vehicle Test Burns,” Fire
California Criminalistics Institute, 1994). and Arson Investigator 48 (March 1998): 23–25.
13. Ibid. 39. F. J. Suefert, “Steal and Burn,” Fire and Arson
14. J. Bloom and C. Bloom, “Ford Ignition Switch Investigator (December 1993): 18–23.
Overheating and Fires,” Fire and Arson Investigator 40. ASTM, Black Box Data from Accident Vehicles:
47 (March 1997): 10–11. Methods of Retrieval, Translation, and Interpretation,
15. J. Morrill, “Analysis of a Ford Speed Control monograph (W. Conshohocken PA: ASTM, 2009).
Deactivation Switch Fire,” Fire and Arson Investigator 41. G. A. Wendt, “Investigating the ‘Black Hole’ Motor
(July 2006): 22–27. Vehicle Fire,” Fire and Arson Investigator 49 (October
16. NFPA 921: Guide to Fire and Explosion Investigation 1998): 11–13.
(Quincy, MA: National Fire Protection Association, 42. www.nhtsa.gov.
2008), 223; (2011, 257). 43. Hrynchuk, “A Study of Automobile Fires”; California
17. NFPA, “U.S. Vehicle Fire Trends and Patterns” (Quincy, Department of Forestry, Vehicle Fire Experiments
MA: National Fire Protection Association, July 2008). (December 1979); see also R. Hrynchuk, “CAFI Vehicle
18. Universal Oil Products, personal communication, 1979. Fire Investigation Techniques: A Report on Vehicle Fire
19. H. Ingason and R. Hammarstrom, “3MW from a Single Tests,” Canadian Association of Fire Investigators
Tyre,” Brand Posten, SP, Boras, Sweden #39 2008, p. 6. Seminar, Toronto, Ontario, September 30, 1998.
20. Allen et al., “A Study of Vehicle Fires.” 44. Nicol and Overley, “Combustibility of Automobiles.”
21. J. F. Krasny, W. J. Parker, and V. Babrauskas, Fire 45. L. McGill, personal communication, 1995.
Behavior of Upholstered Furniture and Mattresses 46. J. Bloom and C. Bloom, personal communication, 1995.
(Norwich, NY: William Andrew, 2000), 145. 47. J. Bloom and C. Bloom, Recreational Vehicle Fire
22. M. Spearpoint et al., “Ignition Performance of New and Investigation, 2nd ed. (DVD), (Grants Pass, OR:
Used Motor Vehicle Upholstery Fabrics,” Fire and Bloom Fire Investigation, 2007).
Materials 29, no. 5 (September–October 2005): 265–82. 48. A. R. Earls, “Manufactured Homes Revisited,” NFPA
23. Marley and DeHaan, CCAI Vehicle Fire Tests. Journal (March–April 2001): 49–51.
24. G. Barnett, Automotive Fire Analysis, 2nd ed. (Tucson, 49. B. Davie, “The Investigation of Mobile Home Fires,”
AZ: Lawyers & Judges Publishing, 2007), 96–100 . Fire and Arson Investigator 39 (September 1988): 51.
25. J. D. DeHaan, “A Close Call for a Hot Hudson,” 50. R. Hammarstrom, “Joint Development of Fire Fighting
Arizona Arson Investigator (1993). Systems for Construction Machines,” Brand Posten,
26. M. M. Hirschler et al., “Rate of Heat Release for SP, Boras Sweden, #39, 2008, pp. 4–6.
Plastic Materials from Car Interiors,” BCC Flame 51. NFPA 302: Fire Protection Standard for Pleasure and
Retardancy Conference Proceedings, June 2002. Commercial Craft (Quincy, MA: National Fire
27. NFPA, “U.S. Vehicle Fire Trends and Patterns.” Protection Association, 2010).
28. V. L. Bell, “The Potential for Damage from the 52. F. Herrera, “Investigation of Boat Fires,” California
Accidental Release of Conductive Carbon Fibers from Fire/Arson Investigator (January 2005): 10–11.
Aircraft Composites,” National Aeronautics and Space 53. R. Hadden, F. Jervis, and G. Rein, “Investigation of
Administration, Langley Research Center, Hampton, the Fertilizer Fire aboard the Ostedijk,” Fire Safety
VA, April 1980. Science 9 (2009): 1091–1101.
29. Marley and DeHaan, CCAI Vehicle Fire Tests. 54. W. E. Swanson, Modern Shipfitters Handbook (New
30. S. Riggs, personal communication, July 23, 2009. York: Cornell Maritime Press, 1941).
31. Marley and DeHaan, CCAI Vehicle Fire Tests. 55. J. Nicholson, “Cruise Ship Fires,” NFPA Journal
32. Ibid.; J. D. Nicol and L. Overley, “Combustibility of (January–February 2007): 37–42.
Automobiles: Results of Total Burning, Journal of 56. NFPA 301: Code for Safety to Life from Fire on
Criminal Law, Criminology and Police Science 54 Merchant Vessels (Quincy, MA: National Fire
(1963): 366–68. Protection Association), 2008.
33. Marley and DeHaan, CCAI Vehicle Fire Tests. 57. D. V. Stuart, “Shipboard Fire Investigation,” Fire and
34. J. D. DeHaan and F. L. Fisher, “The Reconstruction of Arson Investigator 29 (January–March 1979): 55–58.
a Fatal Fire in a Parked Motor Vehicle,” Fire and 58. A. D. Sapp et al., “Arson Aboard Naval Ships:
Arson Investigator (January 2003): 42–46. Characteristics of Offenses and Offenders,” NCAVC,
35. J. Mangs, “On the Fire Dynamics of Vehicles and 1995.
Electrical Equipment,” VTT Publication, 521 Helsinki, 59. J. Mangs, VTT Building & Transport, April 2004,
2004. pp. 33–35.

400 Chapter 9 Automobile, Motor Vehicle, and Marine Fires

 CHAPTER
10
Electrical Causes of Fires*

Courtesy of Det. Mike Dalton, Knox

County Sheriff’s Office, Knoxville, TN.

KEY TERMS

ampacity, p. 413 eutectic, p. 437 service conductors, p. 420

circuit breaker, p. 411 ground, p. 402 service drop, p. 437
conductor, p. 402 open neutral, p. 422 thermal protector, p. 419
duty, p. 438 receptacle, p. 411 voltage, nominal, p. 417

OBJECTIVES
After reading this chapter, you should be able to:
■ Identify the most common causes of residential electrical fires.
■ Explain the basic concepts of basic electricity: voltage, resistance, current, and
Ohm’s law.
■ Describe the investigative methodology for identifying ignition of fires by electrical
means.
■ Explain how to identify wiring systems, current-carrying capacity, and overcurrent
protection.
■ Explain several techniques used to investigate electricity-related fires.
■ Understand the concepts of arc-fault mapping.
■ List and describe several post-fire indicators used when tracing an electrical system
in a burned structure.

*
The authors are deeply indebted to Chris W. Korinek, P.E., president and founder of Synergy Technologies
LLC, for conducting many demonstrations and contributing much of the material in this chapter. Contributions
for the lighting section were made by Richard E. Korinek, P.E., of Synergy Technologies LLC. The material on
arc-fault was generously provided by Dr. Robert Svare, Mark Svare, and Mathew Dubbin of Svare Professional
Engineering and Chris W. Korinek, P.E. of Synergy Technologies LLC.

401
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

A
s we saw in Chapter 6, almost any kind of energy can be responsible for igniting a
flame. One form of energy—electricity—is both familiar to us because it is in use
constantly around us and mysterious because its origins and properties are not well
understood, even by many fire investigators. In 2006, an estimated 52,500 home structure
fires reported to U.S. fire departments were attributed to some type of electrical failure or
malfunction. Such fires resulted in 340 deaths and 1,400 civilian injuries (13 percent of all
reported residential fires).1 Many fires are attributed to electrical causes, fairly or unfairly,
and it is therefore important that the investigator have a working knowledge of the basics
of electricity, its control, uses, and measurement, so that its effects on fuels in the fire envi-
ronment can be accurately assessed. That will be the function of this chapter. The post-fire
examination of wires, connectors, fuses, and appliances will also be described.
conductor ■ Any Electricity creates heat as it moves through materials, typically called conductors. The
material capable of term conductors refers to anything that conducts electric current and includes wires, termi-
permitting the flow of
electrons: (1) bare—a nals, connector lugs, bus bars, metallic enclosures, and plasma (such as in an arc and a
conductor having no flame). A plasma is a hot ionized gas that can conduct electric current and will be discussed
covering or electrical
insulation whatsoever;
later in more detail. Even materials normally not thought of as conductors such as charred
(2) covered—a conduc- plastic, electrolytic fluids, and human tissue can act as conductors. Our nerve cells and mus-
tor encased within a cles conduct electricity required for life. Conductors may carry electric current during nor-
material whose compo-
sition or thickness is mal operation (such as the hot and neutral wires and electric device load) and during an
not considered insula- electrical fault (such as a short to a metallic enclosure or ground wire). Electric current may
tive; (3) insulated—a
conductor encased
produce heat as a primary design function, as in the case of heaters, furnaces, range tops,
within a material rec- ovens, toasters, and the like. These pose ignition risks when, under normal operating con-
ognized by the electri- ditions, they are brought into contact with susceptible materials or are misused in such a
cal code as insulation.
way that their normal heat production causes ignition. Additionally, in a much larger number
ground ■ A conducting of situations heat is an unintended by-product of the passage of electric current where the
connection, whether
intentional or accidental, desired result is the operation of lights, appliances, tools, equipment, or simply the trans-
between an electrical mission of current from one place to another. In these cases, the current is flowing through
circuit or equipment,
and the earth, or to
its intended or design path, and the heat created is supposed to be dissipated safely so that
some conducting body the device does not degrade or fail. These items pose a risk when they are operated in such
that serves in place of a way that the heat created by their normal use cannot be safely dissipated. The user
the earth.
usually is not aware of the heat buildup until it causes failure of the device or ignition of
the device or nearby combustibles. There are also situations where the current does not flow
through its intended path but instead flows through an unintended path, causing severe
localized heating and ignition of the material. When current flows along unintended paths
402 Chapter 10 Electrical Causes of Fires
(across “insulating” surfaces or through carbonized materials, for example) the flow is
sometimes called arc tracking. This flow of current may occur as a result of surface contam-
ination by conductive materials or thermal degradation. We shall examine examples of all
three types of electrical fire causation, keeping in mind that it is not simply the creation of
heat that brings about ignition but the transfer of sufficient heat to a susceptible first fuel
to cause sustained combustion that is the concern of the fire investigator.

Basic Electricity
As we discussed in Chapter 2, all matter is composed of molecules that, in turn, are com-
posed of atoms. The atom is the building block of matter. An atom may be thought of as
a positive nucleus surrounded by negative electrons that rotate around the nucleus of pos-
itive charges in the same manner that the planets revolve around the sun in our solar sys-
tem. Electrons and the ease with which they can move about in a material determine the
electrical properties of that material.

STATIC ELECTRICITY
Static electricity is an electric charge produced by an actual transfer of electrons from
some atoms onto others. If a plastic comb is rubbed with a woolen cloth, electrons will
be rubbed from one onto the other, causing extra electrons to be on one piece and fewer
electrons on the other. Atoms are so constructed that there are normally the same num-
ber of positive charges in the nucleus as there are electrons revolving around them. In
other words, atoms are neutral—uncharged. If an atom gains electrons, it becomes neg-
atively charged. If it loses electrons, it becomes positively charged. Static electricity is
simply an electric charge due to the loss or gain of electrons. When we refer to static
electricity, we mean electricity that is not in motion but at rest. Such electricity is not
“visible”; since it is not doing any work, it represents only a potential source of energy.
Static electricity is observable when it is discharged, such as the arc seen and felt when
one walks across wool carpeting on a dry day and touches a metal object. In compar-
ison, the effects of electricity in motion are the production of heat and light or mag-
netic forces, such as those used to drive electric motors. Static electricity has also been
known as frictional electricity, because it is often generated by the rubbing or merely
contact and separation of the surfaces of two bodies of dissimilar materials (rubbing is
not necessary).
When static charge accumulates, one body takes a positive charge and the other a
negative charge. The charges remain on the outer surface of the bodies unless they
come in contact with or near a less-charged, uncharged, or oppositely charged body.
At such time, the charge may pass from one body to another and is neutralized. When
sufficient charge accumulates to break down the resistance of the insulating atmos-
phere, an arc of very short duration (an electrical discharge) occurs. (Because this dis-
charge is not sustained, it is often referred to as an electric spark. No hot particles are
created. It is this resulting arc that is capable of igniting suitable flammable materials
such as gases, vapors, or dusts. (An arc is defined here as the passage of electric cur-
rent through a gaseous medium, however brief its duration.) (Under special circum-
stances passage of current through liquids, plastics, or across a surface can be considered
to be an arc.)
The discharge of static electricity contributes to fire and explosion hazards in many
industries. Often, it is generated in places where it is not noticed and under circumstances
in which it may not ordinarily cause trouble. Charge built up from clothing and dis-
charged to a grounded object is usually harmless, while lightning is often harmless but can
be dangerous. It should be recognized, however, that static electricity is an important
source of potential ignition, particularly in those processes that involve the handling of
Chapter 10 Electrical Causes of Fires 403
 flammable liquids, gases, and combustible dusts. Industries in which the hazard of static
electricity is relatively common are the petroleum, chemical, textile, plastics, printing,
agricultural, powdered metals, mining industries, and service industries such as dry clean-
ers. Many industries recognize the static electricity hazard and regularly attempt to con-
trol it, while others have not yet appreciated its prevalence and danger.
Static electricity is generated in moving parts of equipment such as power transmis-
sion belts and by moving columns of dusts, smoke, powders, and even nonconductive flu-
ids in processing and storage equipment. Static electricity arcs can be discharged from
such equipment, and proper grounding (as described in the National Electrical Code, or
NEC) is required to prevent any accumulation of charge.
Any process that involves the storage and handling of flammable gases and liquids,
combustible fibers and dusts, and similar easily ignitable materials can be subject to the
fire hazard of static electricity. Although there is no generation of static electricity in the
actual storage of flammable liquids, static electricity charges are produced during their
flow or through turbulence in mixing or discharging of liquids into or from tanks or
other containers. Movement of clothing (friction or separation) can create static charge.
The key to the creation of static charge is that the materials are poor conductors of elec-
tricity. Thus, the better the insulative properties, the easier it is for charge to build up to
discharge levels. In dry air, the human body can store enough charge to produce a poten-
tial of 5–25 kV, which releases about 20 mJ of energy.2 Discharge of static electricity
built up on a vehicle or a person has been identified as the cause of numerous vehicle
fires during refueling.3
Despite the fact that humid air is actually a better insulator than dry air, arcs from
static charges occur more frequently during dry winter weather than in the hot humid
months of summer because the moisture accumulated in garments and on surfaces tends
to allow the charge to leak away as quickly as it is produced. Lightning, as a special case
of static electrical discharge, was described in Chapter 6.

CURRENT ELECTRICITY
Electricity must flow to create heat and do work. Current is a flow of electric charge
(electrons or charged particles, called ions), through a conductor.
Current electricity may be generated in various ways. One way is through the use of
an electric generator. In this device, a coil of wire is brought into a magnetic field. As long
as the magnetic field is changing—that is, if either the coil or the magnet is in motion—a
current will be induced in the coil winding. Another way of setting electricity in motion
is by the use of chemical energies in an electric cell (a group of cells is called a battery).
For the electric energy supplied by an electric cell (or other source) to do useful work, an
appropriate electrical path—a circuit—must be provided for the current to flow through.
In this respect, an electric circuit is similar to a water circuit, in which water must have a
path through which it can pass (a pipe) to perform intended tasks. The pressure applied
to the water causes it to flow through the pipe, and the rate at which the flow occurs is
measured in gallons (or liters) per minute. Voltage can be thought of as the electrical pres-
sure forcing the electrons through the electric circuit, and current as the electron flow
rate, as illustrated in Figure 10-1. Current will flow only when the circuit is complete. A
circuit can be defined as a complete path that allows current to flow from the source
through a load and return to the source. The electrons passing through a conductor con-
stantly collide with other electrons. These collisions causes heat to be produced in the
same way a restriction in a pipe causes resistance to the flow of water. The greater the cur-
rent, the more heat produced. If the conductor offers enough resistance, it can become
very hot and thus become a possible ignition source.
Whenever a current of electricity is flowing, as through a wire, a magnetic field
is also generated. Winding the wire into a coil concentrates the magnetic field. This
404 Chapter 10 Electrical Causes of Fires
 FIGURE 10-1 Compo-
nents of an electric circuit
(a) and (b) have equiva-
lents in an analogous
water circuit (c). A battery
provides electromotive
force (voltage); a lightbulb
offers resistance, and a
current flows. Here an
applied voltage of 12 V
produces a current flow of
4 A through a resistance
of 3 . The load (a light-
bulb) dissipates 48 W of
power. Courtesy of Chris
W. Korinek, P.E., Synergy
(a) Technologies LLC, Cedarburg,
WI.

High pressure
Water flow

Valve
(open)


V  12 VDC

Restriction
I4A Low pressure

R3
 P  48 W

(b) (c) Pump

phenomenon is the basis of an electromagnet. As stated earlier, whenever a conductor

is moved through a magnetic field, an electric current will be generated in the conduc-
tor. An electric generator works on the principle that a coil of wire moving through a
magnetic field generates current. An electric motor uses electricity to produce motion.
There is a coil of wire inside a magnetic field in the motor. When a current is sent through
the coil, the direction of the magnetic field produced by the coil is opposite that of the
magnetic field normally present in the motor. This opposition of fields causes the motor
to rotate.
A circuit can be designed to make one of these effects greater than the others, but
none of them can be completely eliminated (except in the special case of superconductiv-
ity). For example, the purpose of electric heaters is to convert electric energy into as much
heat as possible. The magnetism that is produced serves no useful purpose, even though
it is present around the heating coils and connecting wires. In contrast, in an electric
motor, it is the magnetism generated that produces the final desired product—torque; any
heat produced is undesirable.
Features of electron current and ionic current are discussed next.
Electron Current
Metallic copper is electrically conductive because its electrons are free to move. If one
atom at one end of a copper wire gains electrons, the entire wire becomes negatively
charged. As electrons are “pushed” into one end of the wire, electrons migrate from atom
to atom, and the charge is found at the other end because electrons pass through the cop-
per wire easily. However, if electrons were pushed into one end of a piece of dry string,
they would not be available at the other end; electrons do not migrate easily through
string, because it is an electrical insulator.
Chapter 10 Electrical Causes of Fires 405
 TABLE 10-1 Materials in Order of Decreasing Electrical Conductivity

Gold
Silver
Copper Conductors
Aluminum
Steel
Copper oxide
Nichrome
Carbon
Charred wood
Wet skin
Dry skin Semiconductors
Aluminum oxide
Salt water
Paints
Petroleum products
Dry wood
Pure water (deionized)
Most plastics Insulators
Glass
Porcelain
Air
Vacuum

Some materials are better conductors of electricity than others. In other words, elec-
trons flow with much less resistance in conductors than in insulators. Objects in which
static electricity can be induced are usually poor conductors of current electricity. Although
these objects can gain or lose electrons and become charged, they remain charged. The
current does not readily flow through them because their atoms are arranged in a man-
ner that does not permit the electrons to easily migrate among them.
Metals are generally considered to be good conductors of electricity. At the other
extreme, materials such as glass, stone, plastics, and synthetic textiles resist the flow of
electricity, being poor conductors (or nonconductors), and are used as insulators. Another
variety of materials can be described as semiconductors; they are poor conductors at low
voltages but will conduct at higher applied voltages, with higher resistances than “nor-
mal” conductors. See Table 10-1 for a list of such materials.
Ionic Current
Ionic current is a flow of charged particles, called ions, that occurs only through a con-
ductive liquid or paste or plasma. Ionic current is important in batteries such as the
lithium ion type, in current conduction through flames, in wet arc tracking, and in nerve
cells and tissue. This phenomenon is more complex than electron current (a flow of elec-
trons) and will not be discussed further in this text, except to describe how it is involved
with wet arc tracking.4
406 Chapter 10 Electrical Causes of Fires
DIRECT AND ALTERNATING CURRENT (DC AND AC)
A direct current (DC) is a flow of electrons in one direction only. A battery is a good example
of a DC power source. An alternating current (AC) is created by an alternating voltage
that reverses itself in a periodic manner, rising from zero to maximum strength (voltage),
returning to zero, and then going through a similar variation in the opposite direction.
One of these complete changes is called a cycle. The power supplied to homes in the
United States (at 120 or 240 V) varies at 60 cycles per second, often denoted in hertz (Hz)
(60 cycles per second  60 Hz). This means that the cycle—in which the current goes in
one direction, back to zero, then in the opposite direction, and then back to zero again—
is repeated 60 times every second. In many other countries, the power is supplied at 50 Hz
(and at different voltages), which is why appliances from one country require a converter
to work properly in another country. Currently, electronic devices such as laptops and cell
phones are being sold with “dual-mode” power supplies that will accept 100–250 V at
50–60 Hz. An important reason for using alternating current to supply power to home
and industry is the ease with which its voltage can be varied.
A transformer is a device used to convert one voltage to another (higher or lower).
However, a transformer operates on alternating current only; it will not work on direct
current. AC voltage is often measured as its root-mean-square (rms) value, not its peak
voltage. The rms voltage is a numerical “average” of the sinusoidal waveforms and is
about 70 percent of the peak voltage.

ELECTRICAL UNITS
The basic units of electrical measurement are the volt, which measures differences in
potential energy, the ampere, which measures the quantity of electrical charge flowing),
and the ohm, which measures resistance. Actually, the total resistance to current flow in a
circuit is called the impedance, and it is a sum of the simple (often called real) resistance
and more complex components of resistance. Simple resistance will be most commonly
used in this text when resistance is discussed. Additional units are watts or joules per sec-
ond (which measure power). Each unit will be further defined and explained.

Volt
Voltage (V) is the difference in potential energy between two points of an electrical circuit.
This difference provides a pressure or force (sometimes called electromotive force and
represented by the symbol E), which causes electrons or ions to flow between the two
points. This force can be likened to the pressure in the water circuit analogy that forces
the water to move through the pipe.

Ampere
The ampere (or amp; A) is the measure of the quantity of electricity or charge flowing
past a point of a circuit in a given period of time. Current is represented by the symbol I.
Because current is a measure of the number of electrons flowing per second, it can be
compared to the gallons of water per minute flowing in a pipe circuit. Connecting an
ammeter in series into the electrical circuit forces current to pass through the meter, which
allows the quantity of electron flow to be measured. An ammeter is analogous to a water
flow meter reading in gallons per minute. An ampere is the flow of approximately 6 quin-
tillion (6 billion times a billion, or 6  1018) electrons per second.

Ohm
In the water analogy resistance can be considered as the friction of the water flowing
through the pipe. Electrical resistance is measured in ohms (). Mechanical friction pro-
duces heat. For instance, if a piece of sandpaper is rubbed on a block of wood, both get
warm. When an electric current is sent through a resistance, the resisting element will also
get warm. This heat is caused by the “friction” encountered by the electric current as it
Chapter 10 Electrical Causes of Fires 407
 flows through a conductor and associated connection points. As we saw in Table 10-1,
certain materials are better conductors than others. Electrons can “pass” through copper
wire quite freely without bumping into too many other electrons. Thus, copper is a good
conductor. In contrast, electrons have more difficulty in passing through carbon because
the resistance is greater, and therefore more heat is produced. Analogously, water being
run through a smooth-walled pipe will encounter less resistance than if it is run through
a pipe having rough walls.
Resistance can also be thought of in terms of a pipe with a reduced cross section in
the water analogy, because a higher flow (more gallons or liters per minute) can move
through a large-diameter pipe than through a small-diameter tube (at the same starting
pressure). Thus, a larger conductor offers lower resistance, and therefore more current
will flow through it (at the same applied force of voltage) than through a smaller conduc-
tor with a higher resistance. An insulator is a material through which electric current will
flow very poorly, since it offers very high resistance. It is the opposite of a conductor.
Insulators are necessary in electricity and electronics to keep the electric current where it
belongs. Most resistances are “positive,” meaning that if the voltage is decreased, the cur-
rent decreases. For certain phases of an arc, the resistance actually is “negative,” meaning
that when the voltage decreases, the current increases. This helps explain why arcs can be
explosive and dangerous.5
Impedance
Impedance, or the complex resistance to current flow in an AC circuit, is due to the effects
of capacitance or inductance. This complex resistance may be inherent in a component
such as a capacitor or inductor or may even be a property of straight wires. Capacitance
is a property of a circuit or component where charge energy is stored. A capacitor typi-
cally consists of two parallel plates separated by an insulator. Inductance is the property
of a circuit or component where magnetic field energy is stored. An inductor typically
consists of a coil of wire.

Watts and Joules

Power is the rate at which electrical energy is dissipated (or work is done). It is measured
in watts (W) and is the same unit (J/s) as used in heat release rate measurements. Electric
motors are often rated in horsepower (hp), where 1 hp  746 W.

ELECTRICAL CALCULATIONS
Ohm’s law is the basic law of electricity and is very useful in making energy calculations.
Ohm’s law states that the voltage in any circuit is equivalent to the amperage multiplied
by the resistance. These relationships can be expressed as a series of formulas:

SYMBOL UNIT DEFINITION

V  IR V Volt (V) Potential (analogous to water pressure)
I  VR I Ampere (A) Current (analogous to water flow rate)
R  VI R Ohm () Resistance (analogous to water flow restriction)

In the circuit in Figures 10-1a and b, the applied voltage is 12 V, and the lightbulb
(load) has a resistance of 3 . Using Ohm’s law, I  V> R, we can solve for current: I 
12> 3  4 A. In the same way, but using the formula in a different form, we can solve
for either the resistance or the voltage, depending on which two of these quantities is
known.
408 Chapter 10 Electrical Causes of Fires
 SYMBOL UNIT DEFINITION
2
W  VI  I R W Watt Power (or, using W  V  I  12  4  48 W)
at steady state
WH  V  I  hr WH Watt-hour Energy used
kWH  WH/1,000 kWH Kilowatt-hour Energy used
JWs J Joule Energy used

Power (with simple resistances), is equal to the voltage multiplied by the amperes (or
the resistance times the square of the current, expressed as I2R). For a resistive heater,
power can be calculated from I2R, as the resistances are simple. A motor, however, has
complex impedances, so the I2R calculation cannot be used to calculate power without
involving other factors.
The lightbulb shown in Figure 10-1 would then dissipate (or use) W  I2  R 
2
(4)  3  48 W. The surface temperature of such a lightbulb operating in open air will
stabilize at approximately 175°C (347°F) (as noted in Figure 6-27).
The watt-hour (WH) is a unit of energy (total work done). It is equal to the watts
times the time in hours that an appliance is on. The watt-hour is a relatively small unit,
so the kilowatt-hour (kWH; 1,000 WH) is generally used. This is the unit used by utility
companies in selling electric energy.
If we substitute a 24-W light in this circuit, it will have a design resistance of 6 , so
with a 12-V power supply, the increased resistance causes a decrease in the current to
2 A. The surface temperature of the bulb [about 90°C (194°F) in open air] will then be
cooler than that of the 48-W bulb, per Figure 6-27. The temperature of a surface that is
radiating heat is a function of the wattage and the area over which the power is gener-
ated. This value is most often expressed as “watt density” in units of W/cm2. In the pre-
ceding lightbulb example, the temperature of a higher–watt density bulb is higher than
for a lower–watt density bulb.
For a 120-VAC outlet supplying power to a 48-W lightbulb, the circuit looks like
Figure 10-2. In comparison with Figure 10-1, note that although the wattage output has
stayed at 48 W, the voltage has increased by a factor of 10 (from 12 V to 120 V), the cur-
rent has decreased by a factor of 10 (from 4 A to 0.4 A), and the resistance has increased
by a factor of 100. Note that the power calculation I2R still yields 48 W.

FIGURES 10-2A–B With

120 V (AC in this case),
V  120 VAC

the 48-W light circuit

current drops to 0.4 A,
I  0.4 A while the lightbulb dissi-
R  300 
P  48 W pates the same power.
Courtesy of Chris Korinek,
P.E., Synergy Technologies LLC,
Cedarburg, WI.
(a) (b)

Chapter 10 Electrical Causes of Fires 409

 FIGURE 10-3 Ohm’s
law wheel gives all the V2 V
combinations of the R R
2R P
relationships of voltage, I
V
resistance, current, and 

power for resistive loads VI P
P I BR
such as incandescent
lights and heaters. The
expressions in the outer 
 V R V
C PR
ring equal the quantities I
in the inner circle area. P V2
All the values on each I P P
IR
segment must be known I2
to calculate the inner V  Voltage (volts)
quantity. R  Resistance (ohms)
I  Current (amps)
P  Power (watts)

A convenient way to show all the relationships for current, voltage, resistance, and
power is the Ohm’s law wheel, illustrated in Figure 10-3. All the expressions in each sector
of the outer ring are equal to the value in the inner segment. The values in the outer segment
must be known to establish the value in the inner sector.

SERIES AND PARALLEL CIRCUITS

Many circuits have more than one load. These loads can be connected in two basic config-
urations: series and parallel. In a series circuit, the loads are connected so that the same
amount of electric charge (current) flows through all of them simultaneously. Figure 10-4
shows a series circuit with two loads (lights). Note that it does not matter whether the
switch comes before or after the load in the circuit. No current can flow until the switch
contacts close. When the switch closes, the current flows from the battery through each of
the loads before returning to the battery. To analyze this circuit, the total load resistance is
calculated (3  3  6 ). With a total load resistance of 6  and a 12-V source, the cur-
rent flow will be 2 A. The water analogy shows that both water restrictions are plumbed
into the same line, reducing the total flow rate in the same way two resistances (loads) in
series reduce the current flow through both loads. Note that each light uses (dissipates)
only 12 W, for a total of 24 W. Series circuits are important when considering modes of
electrical overheating such as a poor connection (resistance heating), open/lost neutral,
series parting arc, and others. In summary, all resistances in series add up to the total load.

High pressure Medium pressure

Water flow

R3 Series restrictions

V  12 VDC
V  12 VDC

I2A
Low pressure


 R3

Pump
(a) (b)

FIGURES 10-4A–B Series electrical resistances are connected in-line, just like the series water resistances shown.
Series resistances add to the total circuit resistance. Courtesy of Chris Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

410 Chapter 10 Electrical Causes of Fires

 High pressure Medium pressure
Water flow

Parallel
restrictions
V 12 VDC

I8A

I4A
I4A

Low pressure
Pump
(a) (b)

FIGURES 10-5A–B Two parallel electrical resistances are connected such that the load will experience the same
voltage, and the two water resistances in parallel will experience the same pressure drop. The two parallel
resistors of 3  each reduce the total circuit resistance to 1.5  (resulting in a total current flow of 8A).
Courtesy of Chris Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

In a parallel circuit, seen in Figure 10-5, the loads are connected so that all of them
experience the full supply voltage (pressure). The total current will be the sum of all the
currents in given portions of the circuit. This is the way in which outlets are wired in nor-
mal household systems—one side of each receptacle is connected to the “hot” or “live” receptacle ■ A contact
(pressure) supply and the other to the neutral (return side). With parallel circuits it is pos- device installed as the
outlet for the connec-
sible for the current to increase to an amount greater than the rating of an overcurrent
tion of an appliance
protection device (OCPD) such as a fuse or circuit breaker. The OCPD should interrupt by means of a plug.
the current flow if the total load current exceeds the rating of the OCPD. In a parallel
circuit, each added new resistance (load) adds a new path for current without decreasing circuit breaker ■ A
the current through other loads. In such a circuit, components are connected so that the device designed to open
a circuit automatically
total current (flow) can split, allowing a partial flow through more than one conductive at a predetermined
path. Note that switches can now control individual loads. overcurrent without
The greatest current will be through the path with the least resistance. This is a very injury to itself when
important concept to remember. Given the opportunity, a tremendous current will flow properly applied within
where a path is made available with minimal resistance. Most of the current (but not all of its ratings.
it) will take the shortcut in resistance back to the source instead of going through the higher
resistance load (such a short circuit is an unintended current path). This short circuit can
produce current flows of hundreds or thousands of amps. In the water analogy, this is like
connecting a very large diameter hose between the pressure and return sides. Such a short
circuit is shown in Figure 10-6. A wire may have a resistance of only 0.1 , so a parallel cir-
cuit like this will create a current flow of 120 A through the short path, dissipating 1,440 W
(at a voltage of 12 V). Obviously, this will cause all the wires in the circuit between the bat-
tery and the short to overheat tremendously as long as the current flows (without OCPD
protection, small-diameter wires may reach fusing or melting temperatures). Parallel circuits
are used in most building and vehicle wiring. They help us understand overheating modes
such as short circuits, current leakage, excessive current heating, and arc tracking. Note that
a current of 2 A continues to flow in the intended branch of the circuit (with resistance of
5 ), stabilizing the short-circuit flow in the other branch.
I  122 A

I2A

FIGURE 10-6 A short-circuiting wire connected

V  12 VDC

 
across the legs of this circuit (in parallel with the
I  120 A

 R  0.1  load) has a very low resistance, and as a result, I 


P  1440 W R6 V> R  120 A flows through the short-circuit wire,
P  24 W dissipating I2R  1,440 W. The resistance of 0.1 

represents the short-circuit wire. Courtesy of Chris W.
Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

Chapter 10 Electrical Causes of Fires 411

 Electrical Systems
Vehicle wiring systems (12 to 24 VDC) often use the metal frame of the vehicle as the
return path for current. Voltage is supplied to one “hot” or “live” side of lights and acces-
sories, and the other side is connected to a common “ground,” (usually denoted by the
series of parallel lines in an arrowhead formation on diagrams). In a building, the frame-
work is not (sufficiently) conductive, so wires are provided for a return path, called the
neutral. The grounding system of building wiring systems extends through the fixed wiring
and into the appliance as a non-current-carrying conductor. For a residential 120> 240-VAC
system, the ground system is intended primarily for occupant safety against shock hazards
and does not normally carry current. The grounding conductor of most electrical systems
is bonded (firmly connected) between the main service panel and a copper rod or metal
water pipe in contact with the earth. The neutral (return) side of the system is usually
bonded in turn to the ground at the main service panel, so that there is only one “hot”/
“live” conductor to a 120-V circuit. (At subpanels, the ground and neutral are split.)
Figure 10-7 shows the typical North American wiring system for 120> 240-VAC, 60-Hz,
single-phase residential and commercial service. Two 120-V loads are shown, each ener-
gized by 120 V potential with respect to the neutral conductor. One 240-V load (like a range
or dryer) is shown energized by 240 V from one hot line to the other. The neutral is held
basically at 0 V by being tied to ground and carries the current only from unbalanced loads
on the 120-V legs. For simplicity, no ground wires are shown and only the ground symbol
is shown (on the neutral). In Europe and areas of Asia, the typical single-phase system is 230 V
(; 10 percent), 50 Hz power, as shown in Figure 10-8. Note that 120 V is not available.

CONDUCTORS AND INSULATORS

Three factors determine the resistance of a conductor: the material from which it is made,
its length, and its cross-sectional area. The heat produced is dependent on the resistance,
current flow, and its operating temperature. The size requirements of a conductor are gen-
erally specified by NFPA codes and standards that take into account the following factors:
■ How high a temperature can be tolerated by the insulation or the conductor. This
is dependent on the rate of heat generation and the rate at which that heat can be
dissipated. The temperature must remain below the point at which the conductor
will melt or be otherwise made ineffective. It also must be below the temperature
at which the insulation will be damaged. Asbestos insulation was once frequently used
on heater and appliance cords. Asbestos has been replaced with higher-temperature-

Normal 120/240 V, 60 Hz Wiring

Hot Normal 220/240 V, 50 Hz Wiring
Hot

120 VAC
120/240 VAC

Neutral
220/240 VAC

240 V 220/
220/240 V 240 VAC
120 VAC Load(s)

Hot Neutral

FIGURE 10-7 Typical normal North American FIGURE 10-8 Typical normal European and
wiring system with a grounded neutral Asian wiring system without a grounded
conductor network allows 120-V operation neutral conductor network allows only 220-
of many appliances as well as 240-V for to 240-V operation of appliances. Courtesy of
larger ones. Courtesy of Chris W. Korinek, P.E., Chris W. Korinek, P.E., Synergy Technologies LLC,
Synergy Technologies LLC, Cedarburg, WI. Cedarburg, WI.

412 Chapter 10 Electrical Causes of Fires

 resistant fiberglass or synthetics, which can operate acceptably at higher temperatures
and can safely carry greater current than conventionally insulated cords. The resistance
of a conductor changes with its temperature, so its normal operating condition must
be considered. (For most metals, the higher its temperature, the higher its resistance.)
■ The conductor must be of sufficient size to be mechanically sound. A small appliance
may require only a very small current, but a very small wire may be too fragile and
break.
■ The conductor must be of sufficient size that the voltage drop is not excessive.
Generally, in North America the size must be adequate to supply the load without
exceeding a 2 percent drop in voltage.
Take, for example, a correctly sized wire supplying power to a pump motor located
remotely to a dwelling. The wire must be large enough to (1) carry the required current
without overheating, (2) withstand mechanical stress such as wind and rain, and (3) be
large enough so that the voltage at the pump motor is not lowered significantly.
Metallic conductors are often made of copper or aluminum. Solid wire is used in the
smaller sizes and for purposes not requiring flexibility. Larger sizes and wires needing
flexibility are constructed of multiple strands of smaller wires. Insulation quality is basi-
cally dependent on the type of material and its thickness. In addition, important consid-
erations are the ability of the insulation to withstand elevated temperatures and moisture.
Wire insulations usually encountered are rubber, plastic, and asbestos. Other good insu-
lating materials are glass fiber, ceramic, and varnish.
Insulation is rated by voltage that it can withstand or suppress. Conductors with insu-
lation are rated by current and voltage. The current rating is determined by the current
that can be carried without a temperature rise above the point at which the insulation
could be degraded. The voltage is the maximum voltage that can be applied between the
conductor and the outside surface of the insulation without significant current flow.
Wires in U.S. household circuits today usually have thermoplastic insulation that is
generally colored according to the following convention:

Black Line (“hot” or “live”)

White or gray Neutral
Green (or bare) Grounding (earth) conductor

Other colors are additional line (hot) conductors. They will be found in #10, #12, and
#14 gauges in most residences.
Electrical insulation between the hot and ground conductors, as shown in Figure 10-9,
is generally extremely high resistance and can be thought of as infinite (at the usual rating
voltage of 600 V). In this circuit, the resistance is 10,000,000  (10 M), and the current
flow through it is only 0.00001 A. For this reason, electrical insulation resistance is usu-
ally not shown on a schematic. If the insulation becomes degraded or contaminated, the
level of resistance can become low. As the resistance drops, the insulation can offer a path
for higher current flow, and heat production will result.

CURRENT-CARRYING CAPABILITY (AMPACITY)

Ampacity is defined as the amount of current (amperes) that a conductor can continuously ampacity ■ Current-
carry without overheating (degradation) of its insulation. As with the water circuit, a larger carrying capacity of
conductor is required to pass a larger current, because forcing a large current through a electric conductors
(expressed in amperes).
small conductor will cause it to overheat. The ampacity is dependent on how high a
temperature can be tolerated by the insulation and the wire’s ability to dissipate heat into
Chapter 10 Electrical Causes of Fires 413
 FIGURE 10-9 Schematic of good

I  14.00001 A

I0.00001 A
insulation between hot and ground R  10,000,000 
 

V  120 VAC
conductors. In this case, insulation is P  0.001 W
10 M, and current flow through it

I  14 A
is negligible at this voltage. Courtesy of 
Chris W. Korinek, P.E., Synergy Technologies
R  8.4 
P  1650 W
LLC, Cedarburg, WI.

the surroundings. The ratings are usually for a wire at normal ambient temperatures in
air. Obviously, packing the wire in building insulation will reduce its ability to dissipate
heat and may accelerate the degradation of the wire’s insulation. Electrical codes list the
maximum allowable current for various sizes of conductors and types of insulation. The
generally acceptable current ratings are listed in Table 10-2 for a common grade of insu-
lation in North America but will vary due to special circumstances (high ambient temper-
atures) or different insulation types.
The ampacity of wire is usually conservatively rated, meaning that the wire generally
will not overheat if it carries a small additional current above its rated current capacity,
especially for a short amount of time. Over an extended period of time, it is possible for
the insulation to degrade if the current is slightly in excess and the wire is thermally well
insulated, so the listed rating should be adhered to. The labeled “ratings” of cables and
wires are based on what currents can safely be carried under particular conditions, such
that the insulation is not compromised by melting or degradation. Flowing twice the rated
current, for example, will significantly raise the temperature of the wire and its insulation,
and this elevated temperature may degrade the insulation over time.

TABLE 10-2 Allowable Ampacities of Common Copper Wire Sizes [for TW

Insulation, Moisture-Resistant Thermoplastic at 60°C (140°F)]

WIRE SIZE (AWG) DIAMETER (In./mm) AMPACITY (A)

18 0.040 (1.02) 6a
16 0.051 (1.30) 8a
14 0.064 (1.62) 15b
12 0.081 (2.05) 20b
10 0.102 (2.60) 30b
8 0.129 (3.28) 40
6 0.162 (4.11) 70
2 0.258 (6.55) 95
1> 0 (0) 0.302 (8.26) 125
2> 0 (00) 0.365 (9.27) 145
4> 0 (0000) 0.460 (11.684) 195c

Source: National Electrical Code (NEC), (Quincy, MA: National Fire Protection Association, 2008), table 310-16,
pp. 70–148. Allowable ampacities of insulated conductors rated 0 through 2,000 V, 60°C through 90°C, not more
than three current-carrying conductors in raceway, cable, or earth (directly buried) based on ambient temperature
of 30°C.
AWG  American Wire Gauge
a
NEC 2008, table 402-5 (fixture wires).
b
NEC 2008, 240.4(D).
c
At 90°C.

414 Chapter 10 Electrical Causes of Fires

 V  2.5 VAC FIGURE 10-10 Schematic of a nor-
mally operating circuit with a 1,650-W

V  120 VAC
I  14 A
R  0.18  heater operating at 120 V. In this case,

P  35 W 20 m of NM cable also offers a small
Ground conductors resistance of 0.18 , which dissipates
R  8.4  V  117.5 VDC 35 W. Courtesy of Chris W. Korinek, P.E.,
P  1650 W Synergy Technologies LLC, Cedarburg, WI.

The fusing current of copper wire is the current at which the wire will melt in a given
amount of time and is many times the normal current rating of the wire. For example, if
you wanted to fuse (protect) a large-gauge wire by de-energizing it after it carried 3,000 A
for 0.1 second, a #12 AWG (American Wire Gauge) solid wire would most likely accom-
plish this task by melting after 0.1 second at this high current.6 A current of 3,000 A is
150 times the normal current rating of 20 A for a #12 solid wire.
Also, it should be realized that heat generated in a conductor by current flowing
through it is uniform per unit length, provided that the entire conductor is the same size
throughout the circuit. Where overcurrent is suspected, examination of the wiring on the
same circuit between the suspect area and the service in an unburned location can frequently
offer useful information. If the wire insulation in the unburned area of the wire is in good
condition, it would demonstrate that the current was not excessive in the questioned area.
The circuit diagram in Figure 10-10 shows the effect of resistance heating in a 20-m
(66-ft) length of #14 AWG NM (nonmetallic) cable carrying power to a 1,650-W (1.65- kW)
electric heater. Note that the small series resistance of the wire causes a voltage drop, and
the resistance heating generates some 35 W. When this cable is out in the open, it allows
heat to safely dissipate, but if the cable is buried under thermal insulation or tightly coiled,
heat cannot safely dissipate and temperatures can increase to cause the breakdown of the
insulation and possibly a fire (as in Figure 10-11).

FIGURE 10-11 Extension

cord with 1200-W heater
plugged in and buried
under a pile of clothing
failed and ignited a fatal
fire. Note folded loops
of cord at center front.
Courtesy of John D. DeHaan.

Chapter 10 Electrical Causes of Fires 415

 Protection—Overcurrent and Short Circuit
Overcurrent protection devices (OCPDs) (fuses and circuit breakers) have two important
functions. One is to guard against overheating of conductors from overcurrent. (This pro-
tection may be to protect the conductors of the building or to protect appliances and their
power cords from overheating.) The other is to provide protection from a short circuit
(a condition caused by direct contact of the circuit conductors, as shown in Figure 10-6),
sometimes called a dead short. (A bolted fault refers to a short where the conductors are
physically restrained so they cannot separate.) A dead short creates the highest current
flow and triggers the fastest operation of the typical overcurrent device.
Some equipment, however, requires an initial or starting current that is considerably
higher than its operating current. Some motors, for example, draw on the order of
600 percent of rated current during their starting phase. If the overcurrent device trig-
gered immediately any time there was a significant draw, the motor would never start.
The protective device on this type of circuit has to be designed to accommodate this brief
overload and still provide the needed protection once the motor has started.

FUSES
A fuse basically consists of a short section with a metal of such composition and size that it
will melt at a predetermined current flow. In practice, however, a fuse opens when its tem-
perature reaches the melting point of the fuse material. This temperature can be reached by
current flow or from the surrounding ambient temperature. A fuse located in the hot sun
will therefore function at a lower current, while one located in icy surroundings will pass
more current. Two considerations should be noted in regard to fuse functioning:
■ Fuses are designed to carry current at their rated current continuously.
■ Fuses will carry overcurrents for a time period depending upon the degree of overload
(magnitude of the current).
Fuses are made in a variety of forms (as seen in Figures 10-12 and 10-13), varying
from a simple piece of fuse wire to the more complicated forms designed to extend the
period of time required for the circuit to open. Some fuses have a “slow-blow” feature to
allow them to pass starting or inrush currents to motors and other large loads for short
periods of time. All fuses, regardless of size and type, have a specified current rating, voltage

(a) (b)

FIGURE 10-12 Overcurrent protection devices (OCPDs). Left: circuit breakers; Right: fuses. Courtesy of Chris
W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

416 Chapter 10 Electrical Causes of Fires

 FIGURE 10-13 Appliance
protection in the form of a
fuse (lower left), automatic-
reset thermostat (upper
left), manual-reset thermo-
stat (lower right), and TCO
or fusible link (upper right).
This level of protection is
usually more sensitive to
appliance overheating
conditions than the circuit
breaker protecting the
building wiring to the out-
let feeding the appliance.
Courtesy of Chris W. Korinek,
P.E., Synergy Technologies LLC,
Cedarburg, WI.

rating, and thermal characteristics. The correct selection for safe and trouble-free circuit
protection relies on thoroughly understanding and considering these factors.

Voltage Rating
The voltage rating of a fuse should be equal to or greater than the voltage of the circuit.
This rating is not a measure of its ability to withstand a specified voltage while carrying
current. Instead, it is the ability of the fuse to prevent the open-circuit voltage of its system
from “restriking” and establishing an arc once the fuse link has parted. North American
standard fuse voltage ratings are 600, 300, 250, and 125 V and are usually stamped on the
fuse. Fuses are devices that are sensitive to changes in current, not voltage, and remain
functional at any voltage from zero to their maximum rating. It is not until the fuse reaches
melting temperature and fusing occurs that the circuit voltage and available power influ-
ence the fuse performance and determine the safe interruption of the circuit. A fuse may
be used at any voltage less than its voltage rating without detriment to its fusing charac-
teristics. A fuse used in a circuit at a voltage far in excess of its rating is potentially subject
to a restrike, that is, a renewed flow of current between the fused elements.

Continuous Current Rating

The continuous current rating of a fuse should be equal to or slightly less than the current-
carrying capacity of the circuit it protects. Only in special cases in which the connected
equipment has unusual characteristics or where the ambient temperature is quite high
should the fuse rating be greater than the current-carrying capacity of the circuit.

Examination of Fuses
Examination of functioned fuses can frequently provide some useful information. If the
entire link melted, it was very likely caused by external heat. If the fuse link melted only
at the center (smallest area), it was probably caused by overcurrent. If the fuse link vapor-
ized and coated the inside of the fuse housing with evaporated material, the cause was
very likely an enormous overcurrent (a short circuit). (See Chapter 14 regarding labora-
tory examination of fuses.)

voltage, nominal ■ A
value assigned to a
circuit or system for
the purpose of conve-
Nominal voltage is a value assigned to a circuit or system for the purpose of conveniently designating niently designating its
its voltage class (120240, 480Y/277, 600, etc.). voltage class (120/240,
480Y/277, 600, etc.).

Chapter 10 Electrical Causes of Fires 417

 FIGURE 10-14 Schematic Circuit Breaker Schematic
of a grounded circuit with  Circuit breaker
a circuit breaker sensing
the current through the
one energized conductor.
If this current exceeds the
rating of the OCPD, the
OCPD will trip. Courtesy Load
of Chris W. Korinek, P.E., Short-
 circuit
Synergy Technologies LLC,
current
Cedarburg, WI.
 Ground system connected
to grounded enclosure

CIRCUIT BREAKERS
Another common form of overcurrent protection is the circuit breaker (as shown in
Figure 10-12). The circuit breaker in its simple form consists of a mechanical means of
opening or switching the circuit. Its contacts are “normally open” and are closed by
mechanical pressure on an external lever acting against a spring. The contacts are kept
closed by a thermally sensitive element. When this element reaches its rated maximum
temperature (as a result of overcurrent), it does not melt, as in a fuse, but releases the
spring-loaded contacts, opening the circuit, as shown in Figure 10-14. The modern circuit
breaker is a dual-function device with either thermal or electromagnetic triggers that allow
it to respond to modest overcurrents slowly or to massive overcurrents (short circuits) very
quickly. External heating (from fire exposure) can cause circuit breakers to trip.
Circuit breakers are designed to respond in a time-temperature-dependent manner.
That is, they can carry 6 to 10 times their rated current for up to four cycles (1/15 sec-
ond) or twice their rated current for 2 minutes. In contrast, some electronic equipment
cannot tolerate any appreciable overcurrent, so the protective device must be a “fast-
blow” design to trip immediately. Some varieties of circuit breakers were shown in Figure
10-12. Circuit breakers can be actuated by magnetic or thermal mechanisms and can also be
tripped by external heating to temperatures of 135°C to 150°C (275°F to 300°F). A
breaker panel exposed to substantial direct flame impingement, internal or external, can
be expected to have one or more breakers in the tripped (open) position. European minia-
ture circuit breakers (MCBs) are made to the same European specification. They are clas-
sified as types B, C, and D, with C being normally used for discharge lighting and D for
motor circuits. These circuit breakers do not have a tripped position; therefore, it is not
possible during a scene examination to determine if the breaker was off prior to the fire
or tripped during the fire. All European specification circuit breakers incorporate thermal
and magnetic devices to protect circuits from overload and short circuit.
Circuit breakers, as opposed to fuses, are mechanical devices and therefore subject to
conditions of dirt and corrosion, which can affect their operation. It is possible to perform
tests on a suspected breaker, but if laboratory tests are to be made, any attempts to trip the
breaker may invalidate any further tests, because a single hand trip could loosen a possible
frozen/seized mechanism. X-rays can be taken prior to any testing to verify its condition.
Modern North American and European circuit breakers will function even if the han-
dle is mechanically held, as would be the situation if the handle was secured with a piece
of tape. This is not true with some older North American circuit breakers. This also can
be determined by laboratory tests on the questioned breaker. Some circuit breakers can
also be made to trip by mechanical shock (impact).
Fuses and circuit breakers, although meant for preventing overheating from overload-
ing in an electric circuit, are arc-producing devices and may be hazardous themselves
unless designed, installed, and maintained so that the arcs produced cannot ignite flam-
mable atmospheres.

418 Chapter 10 Electrical Causes of Fires

THERMAL PROTECTORS
In modern electronics, other devices called thermal protectors provide various thermal, thermal protector ■
overcurrent, and short-circuit protection. These include manual- and automatic-resetting An inherent device
thermostatic controls (used as high-limit controls) and replaceable thermal cutoffs or against overheating
that is responsive to
cutouts (TCOs) (see Figure 10-13.) A typical TCO consists of a pellet of nonconductor that temperature or current
holds two spring-loaded contacts together. When the designed temperature is reached, the and will protect the
pellet melts and allows the contacts to separate and interrupt the current flow. They are equipment against
not reusable or resettable and must be replaced if they function. TCOs have been known overheating due to
to fail to operate, and two are often connected in series in heat-producing appliances like overload or failure
to start.
coffeemakers. When TCOs are encountered during a fire investigation, they should be
carefully preserved for examination by a qualified specialist, such as an electrical engineer.

SURGE PROTECTION DEVICES

With the advent of sophisticated electronic devices such as computers and their peripheral
devices came a need to protect these devices from surges or spikes in the power delivered by
ordinary wiring systems. “Surge protectors” are available in a number of forms but the most
common device for consumer use is based on a “metal oxide varistor” or MOV. This is a small
disk of zinc oxide encased in epoxy, with a connector wire affixed to each side. When a nor-
mal-service steady (clamping) voltage is applied, the MOV acts as a high-resistance element.
When a surge of current or a spike in voltage occurs, the extra energy is dissipated through the
zinc oxide. If repeated, such surges can cause the MOV to fail either as a continuous, low-
resistance short; an open device (which no longer protects the connected device); or as a high-
resistance current path (which heats up as a current flows through it). This heating can melt
adjacent plastics, degrade them, and even ignite them, as pictured in Figure 10-15.7
Some manufacturers of better-quality power strips bond a TCO to the MOV disk so
that when the MOV overheats, the circuit opens. Others use a fine-wire fuse trace or a
thermal fuse spring that quickly opens when an overcurrent occurs. Failed MOV disks are
rough or melted in appearance, while serviceable disks are smooth-finished and uniform
in texture. Failures of MOVs have been linked to fires starting in multiple power tap
extension cords (power strips) with plastic cases.8

FIGURE 10-15 Failed

MOV in this plug strip
caused ignition. Courtesy
of Steve Riggs, Public Agency
Training Council.

Chapter 10 Electrical Causes of Fires 419

 OVERCURRENT DEVICES AND FIRE INVESTIGATION
The overcurrent device can be of importance from a fire investigation standpoint and should
be carefully examined and its condition documented, along with the fixed wiring, appliances,
receptacles, and other components. The OCPD is well named the “safety valve” of the system.
The rating of North American fuses and the setting of circuit breakers should not
exceed that permitted by the National Electrical Code (NEC). Similar guides also apply to
other countries. When an overcurrent device operates, it ordinarily indicates an overload
or trouble on the circuit. This condition may be due to a momentary overload. However,
if the OCPD continues to operate, the cause should be determined and remedied.
Repeated tripping of OCPDs may be a clue to the cause of a fire and should be established
by interviews of employees or residents.
In North America, as a general rule, the NEC requires that every piece of equipment
be protected by some form of overcurrent device. Normally, the rating or setting of the
overcurrent device should be as low as is practical for the particular equipment con-
cerned. In an electric range, for instance, it should be just large enough to carry the load
continuously, but with a large motor, especially one with a variable load, a specialized
motor controller is required. Smaller items, such as lamps, and household and office
appliances, are grouped under the protection of one overcurrent protective device. On the
typical residential circuit used for lighting, overcurrent protection is rated at 15 or 20 A.
The mere finding of a slightly larger fuse or breaker, such as a 20-A fuse when a 15-A fuse
is proper is most likely not a fire cause, since the extra current flow it will permit will not
have a significant effect on the temperatures generated.
The maintenance of good, clean contacts at fuse terminals is important to safety and
should be considered by the fire investigator. Overheating at this point from poor or dirty
contacts generally results in needless blowing of fuses and, frequently, replacement with
larger ones as a stopgap repair.
The following hazards most likely to be found in overcurrent protective devices:
■ Fuses of significantly too large a capacity or circuit breakers with too high a rating
for the intended load.
■ Circuit breakers made inoperative by blocking or taping of the tripping element
(old type).
■ Plug fuses that have blown and in which pennies have been inserted or where the
metal of the fuse holder has been cut back or in which a wire has been inserted.
■ Cartridge fuses that have blown and in which nails, wires, or other types of metal
have been inserted.
■ Refillable or rewirable fuses in which additional strips or wires have been placed.
■ Fuses or circuit breakers in poor mechanical condition or that have not been moved
and reset in many years.
service conductors ■ ■ The service conductors from the utility transformer to the meter and panel have no
The supply conductors overcurrent protection, other than a fuse on the primary side of the transformer, so
that extend from the
contact with trees, ladders, or tools can draw very high currents.
street main or trans-
■ Fuses, without enclosures, in the vicinity of combustibles.
former to the equip-
ment of the premises ■ Hazards created by doors or covers of fuse cabinets or breaker panel boxes off or
served. opened.
■ Corrosion or contamination of fuses, circuit breakers, or their connectors or sockets.

(Note that all but the last three items on this list can result in very high currents being
drawn that can cause fires by a variety of mechanisms, all without tripping the OCPDs.)

GROUND FAULT INTERRUPTERS

Ground fault interrupters (GFIs—also known as ground fault circuit interrupters, GFCIs)
are installed to minimize the risk of electric shock rather than to prevent fire-causing situ-
420 Chapter 10 Electrical Causes of Fires
 GFCI and AFCI Schematic
 GFCI sensing coil


Leakage current
GFCI trips at 0.005 A
AFCI trips at 0.030 A
Leakage
current

Ground system connected
to grounded enclosure

FIGURE 10-16A Schematic of a grounded circuit with a FIGURE 10-16B GFCI Outlet. “Push to test” button actually cre-
GFCI or AFCI sensing the current imbalance through both the ates a 0.005-A (5-mA) imbalance and tests the ability of the GFCI
hot and neutral conductors. If this current imbalance exceeds to sense the imbalance and trip. Courtesy of Chris W. Korinek, P.E.,
the leakage current trip level, the GFCI or AFCI will trip. Courtesy Synergy Technologies LLC, Cedarburg, WI.
of Chris W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

ations. They operate on the balancing principle that the current in the line (hot) wire must
be equal to and opposite that in the neutral. Any imbalance causes the device to interrupt
the flow (see Figure 10-16a.) GFCIs will trip on an imbalance of about 0.005 A (5 mA),
which is considered a safe level of current for passage through a human body. For house-
hold or light industrial use, GFCIs are commonly in the form of a modified receptacle with
manual “test” and “reset” buttons on the faceplate (Figure 10-16b). The permitted cur-
rents involved in such devices are generally infinitesimal as compared with that required
for ignition of common materials. Some outlets (receptacles) are protected by remote GFCI
breakers installed in the distribution panel or by another GFCI outlet “upstream” (toward
the power source) from that being examined. Note that GFCIs are designed to prevent elec-
trocution by detecting an imbalance between the hot and neutral, not to prevent short cir-
cuits or overcurrent. Fires can be started on connections to a protected circuit. Water
contamination in GFCIs has also been linked to fires.9 In the United Kingdom, GFCIs are
called residual current devices (RCDs). The most popular trip rating for general use is 30
mA. A 2009 change in the UK wiring regulations resulted in RCDs’ being installed into all
new circuits where the wiring is buried (enclosed) in the wall.

ARC-FAULT CIRCUIT INTERRUPTERS

Arc-fault circuit interrupters (AFCIs) are a relatively new type of device that protects conduc-
tors from overheating due to arcing faults. As of 2008, AFCIs are required by the NEC to be
Chapter 10 Electrical Causes of Fires 421
 FIGURE 10-17 Schematic for float- Open-Neutral 120/240 V, 60 Hz Wiring
ing (or open) neutral condition. Note Hot
that with the neutral conductor open
(not connected to the service ground),
Open
the two loads are now in series, 180 VAC R1

120/240 VAC
neutral
instead of being (essentially) in paral-
lel, potentially causing the voltage to 240 V R3
be higher than 120 V. Such an over-
voltage can overheat some appli- 60 VAC R2
ances. Courtesy of Chris W. Korinek, P.E.,
Synergy Technologies LLC, Cedarburg, WI.

Hot

used in all residential bedroom installations.10 They will be seen in most living areas of new
homes as well (per NEC 2008).11 The operation of the AFCI is similar to that of the GFCI,
with the hot and neutral wires connected through a sensing coil (as shown in Figure 10-16a),
but the AFCI will detect leakage to the ground (as opposed to imbalances only between the
hot and neutral currents). If a current imbalance of more than 30 mA (0.030 A) occurs, the
AFCI will trip, de-energizing the circuit. The AFCI is also designed to sense deviations from
the sine wave of normal AC current and trip if the voltage curve does not match the “design”
curve. Short-duration arcing occurs when normal switches are thrown or a plug is pulled
while an appliance is energized, but AFCIs are designed not to trip on such events. AFCIs
should detect many arc faults, but the mere presence of an AFCI should not be taken as evi-
dence that a fire could not have been started on a circuit protected by one. Fires linked to
electrical arcing on AFCI-protected circuits have been reported.

OPEN NEUTRAL
open neutral ■ In an If an open neutral condition exists in a North American electrical household system, the
American single-phase, normal return (neutral) leg is missing, and the two legs (each with its normal load) are now
120/240 V residential
in series rather than parallel, as shown in Figure 10-17. The load R1 expects to “see” 120 V
system, the neutral
return leg is not presented to it, but the voltage it sees can be anywhere from nearly zero to nearly 240 V
connected to the depending on the balance of loads at R2 and R3. This higher voltage can cause overheat-
service ground. ing by excessive current.12 If the homeowner noticed that some lights were abnormally
bright and others dim, or other appliance abnormalities that could indicate high or low
voltage, the neutral system should be inspected for continuity from the service through the
branch circuits. (Such variations can occur on branch circuits that share a common neu-
tral.) Sometimes, the service neutral is interrupted, which will cause current to find what-
ever path it can to return to neutral or ground, producing unintended path heating. The
presence or absence of a ground (to driven rod or water pipe) at the service panel is criti-
cal in determining whether a severed service conductor will cause a voltage imbalance.
In the United Kingdom a lost or open neutral condition will affect only circuits from
a 400-V three-phase supply. The majority of UK household electrical supplies are 230 V
single-phase, so the lost or open neutral fault would need to be at the substation trans-
former or a link junction box that supplies a large number of properties. Therefore, lost
or open neutral faults on supply substation transformers tend to affect a large number of
properties simultaneously.

Electrical Service Distribution

Electricity is delivered to homes and businesses at much higher voltages than are used, for
reasons of cost efficiency. In North America, power is typically generated at 12 to 25 kilo-
volts (kV), transmitted over long distances at 100 to 1,000 kV, and distributed locally at
3 to 20 kV. The voltages are reduced by surface “pad-mounted,” underground, or utility
pole transformers to 240 V for delivery to residences (as shown in Figures 10-18a–d) and
422 Chapter 10 Electrical Causes of Fires
FIGURE 10-18A Overhead service FIGURE 10-18B Typical pole-top transformer
entrance. Courtesy of John D. DeHaan. for 240-V residential service. Courtesy of John
D. DeHaan.

FIGURE 10-18C Lateral (underground) FIGURE 10-18D Pad-mounted transformer.

service entrance. Courtesy of Greg Lampkin, Courtesy of Steve Riggs, Public Agency Training
Knox County TN Fire Investigation Unit. Council.

240/480 V for businesses. Note that voltages over 350 V are high enough for arcing to
initiate through air prior to direct contact with the conductor. (See the work of Pachen,
as discussed in Babrauskas’s Ignition Handbook).

SERVICE ENTRANCE
The service entrance, the point where connection to the utility company is made, usually
contains the electric meter, disconnecting means, and overcurrent protection devices. In
addition, the system ground is made at this point. It is important for the fire investigator
to become familiar with the basic fundamentals of wiring methods and especially the serv-
ice entrance, because this is the basic electricity distribution center.
An example of a typical North American dwelling service entrance is shown in the
block diagram in Figure 10-19. The breakers (or fuses) are identified by size. The required
wire sizes are also shown. Each breaker is selected to protect the circuit it supplies from
overloads. Therefore, 50-A breakers must be wired to #6 copper conductors (or larger),
20-A to #12, and so on. The usual appliance circuit is wired with #12 wire to a 20-A
Chapter 10 Electrical Causes of Fires 423
 FIGURE 10-19 Typical
residential service #2
entrance electrical
connections with breaker 100
ratings and wire sizes.

50 20 15 20 50

#6 #12 #14 #12 #6

20 20 15 15 20 15 30

#12 #12 #14 #14 #12 #14 #10

breaker. Branch circuits for bedrooms and similar low-load areas are often wired using
#14 wires or cable. Small appliances and lamps are frequently supplied with #18 power
cords that are rated at 5 A. It should be noted that when failure occurs in these cords,
they are not protected by a 20-A breaker, since four times the cord’s rated current must
pass (sufficient to cause ignition by overcurrent) before the breaker is even passing its nor-
mal rated current.
Frequently, breakers or fuses are found to have operated after a fire. This finding can
serve as an indicator of possible electrical fault, but it is just one indicator. Breakers and
fuses can be tripped (opened) by fire damage to the electrical system in the absence of any
pre-fire fault. Also, fire-causing electrical faults need not draw sufficient current to cause
the OCPD to open the circuit.
The service entry may be from an overhead drop from a power pole or via an under-
ground connection (called a lateral), as shown in Figure 10-18. The main distribution
panel and meter is located at or near the point where power enters the building for max-
imum safety. This is because any conductor upstream of the main breaker or fuse block
(denoted by the 100-A block in Figure 10-19) is capable of delivering the full current of
the utility transformer (up to 2,000 A). The fire investigator must identify and document
all service entries to a building, and there is often more than one in a commercial or indus-
trial site.

RECEPTACLES
The termination of most fixed wiring in residences familiar to users is the receptacle or
outlet into which lamp or appliance plugs are inserted. Prior to 1955, electrical plugs were
two-pronged for 120-VAC 15-A use (hot and neutral only). Since then they have a third
D-shaped opening for a ground. The arrangement of “slots” in the receptacle face is
linked to their voltage and current rating (along with specialized outlets for hospital
equipment). In typical applications, several outlets will be connected in parallel on a sin-
gle branch circuit. Receptacles may be “daisy-chained,” where the contacts are used to
connect “downstream” receptacles. This means that connections in one receptacle will be
carrying current even when nothing is plugged into that receptacle. Connections may be
made via screw terminals on the sides of each receptacle, or wires may be inserted through
holes in the back of each receptacle (push-in connections) (see Figures 10-20a and b).
The security of the connections is a major focus of fire investigators because loose con-
nections cause “high-resistance” connections. The heat dissipated by such high-resistance
connections may be on the order of 4–12 W in 120-VAC 15-A service, but that is enough
to degrade plastic housings or wire insulation over a prolonged period of use. Such
degradation can lead to a flow of leakage current between the hot and neutral sides, even
424 Chapter 10 Electrical Causes of Fires
FIGURE 10-20A Screw-terminal duplex receptacle. Note FIGURE 10-20B ”Push-in” wiring receptacle. Note stubs of degraded
glowing wire and degradation due to loose connection. aluminum wiring broken off that caused a fire in the box. Courtesy
Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. of Peter Brosz, Brosz & Associates Forensic Services Inc.: Professor Helmut Brosz,
Institute of Forensic Electro-Pathology.

when nothing is plugged in. This can, in turn lead to fires in receptacle boxes (today often
made of plastic instead of steel) and wall cavities.13

Ignition by Electrical Means

When evaluating the possible ignition of a fire by electrical means, it is important to con-
sider just how much electrical energy would have been necessary to cause sustained igni-
tion of the first fuel involved in a particular environment. Small areas of overheated wire
or small arcs are of little concern in a solid material of high ignition temperature. The
smallest discernible arc, however, is a serious hazard in a combustible gas or vapor. As
discussed in Chapter 6, ignition occurs only when enough energy is transferred to the fuel
to initiate a self-sustaining combustion. It is not the total energy but the concentration of
the rate of transfer that determines whether an ongoing reaction will be initiated or will
simply cause localized damage such as a char or scorch.
Ignition by electrical means is caused by a number of mechanisms. We have seen how
misuse of heat-producing appliances can ignite fires, but here we are going to consider
overheating and arc production as a result of failures.
Conduction heating can occur as current travels through a conductor (with the resist-
ance of the conducting medium being the controlling factor) or as it encounters contact
resistance (with aluminum wiring causing special problems in that area). Arcs as sources
of extremely high temperatures have already been discussed. The degradation of electri-
cal insulation can, of course, contribute to either mechanism. Each of these problems will
be examined in turn in the following sections.
Chapter 10 Electrical Causes of Fires 425
 CONDUCTION HEATING
The heating of properly installed and used conductors used to carry current is negligible
under ordinary circumstances. In North America the NEC dictates the maximum current
a conductor shall carry and therefore the amount of heat generated. This maximum cur-
rent rating depends on the size of the conductor, its composition, and the type of insula-
tion covering it. Wherever these specified currents are exceeded (where a conductor is
overloaded), the generation of heat becomes a hazard.
The temperature of an electrical device (including cables and cords) is a function of
its temperature rise and the ambient temperature, which is the temperature of the sur-
roundings of the device when it is not carrying current. Many conductors and appliances
have a temperature rating recommended by the manufacturer to be the maximum the
device can safely withstand. Use above this temperature may result in degradation and fire
hazard. For example, nonmetallic (NM) cable (such as Romex) insulation has an NEC
rating of 90°C (194°F). If #14 AWG NM cable inside R38 thermal house insulation car-
ries a current of 14 A, its temperature rise will be on the order of 33°C (60°F). If the cable
is in a hot attic with an ambient temperature of 40°C (104°F), the cable temperature will
be about 73°C (163°F). If that same 14-gauge NM cable is used to carry 28 A, the power
dissipated will be very high, and the cable temperature will exceed its safe operating tem-
perature due to a high watt density. If thermally insulated, the cable can reach tempera-
tures of 172°C (342°F). Note that this is because the current flow has doubled, and since
power dissipation goes up as the square of the current (W  I2R), the temperature rise
increases by a factor of four [from 33°C to 132°C (60°F to 237°F)].
Overheating can be defined as the permanently destructive heating that has the poten-
tial to cause a fire. Overheating degrades the insulation, conductor, surrounding materi-
als, or all three. Overheating causes of fire can be separated into four categories: (1)
excessive current, (2) poor connection, (3) insulation breakdown, and (4) induction
(rarely seen in nonindustrial applications).

OVERHEATING BY EXCESSIVE CURRENT

On a current-protected circuit, a small overload has only a minor effect on overheating
because of the built-in safety factor. For instance, if a 20-A fuse is used instead of a 15-A
size, there is a 33 percent overprotection, but even with this amount of overload, some time
will be required to cause a significant temperature rise in the conductor, and even then the
maximum service temperature may not be exceeded if the conductor is in open air.
Note that the current flow needed to cause a typical conductor to reach its melting
(fusing) temperature is very high. For example, a #14 AWG copper conductor in open air
will withstand over 100 A before fusing.
Consider the situation in which a 1,200-W appliance is plugged into a 120-V appli-
ance circuit protected at 20 A using a typical #18 AWG extension cord. In case of a short
circuit in the extension cord, possibly from frayed or cracked insulation, the excess cur-
rent will be expected to trip the breaker quickly. In the absence of a short circuit, the
normal current draw (10 A) will not cause the breaker to trip (typically at more than 20 A),
and the circuit feeding the receptacle will not be overheated because, if properly
installed, it will be #12 AWG. The current through the #18 cord, however, is far beyond
its rated ampacity (#18 AWG is rated at 5 A), and this will cause overheating of the
extension cord even in open air. This is an example of an excessive current overheating
mechanism.
When an electric current flows through a single series circuit path, the current (meas-
ured in amperes) is the same in all parts of the circuit. The temperature at any point is
dependent on the electrical resistance at that point and also on the rate at which heat can be
transferred away. Referring to the previous example, the #18 AWG wire has a higher resist-
ance per unit length than the #12 AWG. Therefore, it will operate at a higher temperature
426 Chapter 10 Electrical Causes of Fires
 (a) (b) (c)

FIGURES 10-21A–C (a) Lab-created short circuit of a battery circuit where a #18 AWG wire carrying current
of 120 A degrades as shown by the smoking wire insulation. The light-colored wire is the short-circuit wire.
(b and c) Before-and-after photos of cable insulation degraded by overcurrent. Courtesy of Chris W. Korinek, P.E.,
Synergy Technologies LLC, Cedarburg, WI.

even if the same current is flowing through both. Note that the entire length of the extension
cord will be heated and so the insulation along its length may soften, melt, and perhaps
creep to bring conductors into contact with each other, as shown in Figure 10-21a. At any
location along the extension cord where there is thermal insulation (a rug, pillow, blan-
ket), the convective losses will be minimized, and the temperatures will be much higher.
These higher temperatures can result in significantly more thermal damage (or faster fail-
ure) at those locations. Heating may also produce pyrolysis and charring where the insu-
lation becomes a semiconductor (PVC insulation when heated to 200°C (400°F) begins to
conduct current because of its inorganic fillers.)14 (See Figures 10-21b and c.) Once the
insulation breaks down, the current flow through it can heat it to its ignition conditions,
with localized arcing able to ignite the pyrolysis vapors. The same process can apply to
combustibles in contact with the overheated wire, especially where their contact is imped-
ing heat losses. In a short circuit, where the currents can be very high before the OCPD
functions, the conductor may glow along its entire length (if the OCPD is rated at more
than 10 times the rating of the wire).

OVERHEATING BY POOR CONNECTION

Arcing can occur as either parallel or series arcs, as illustrated in Figure 10-22. In a par-
allel arc, the current flows from the hot conductor to ground, causing a short circuit. This
short circuit tends to be a brief, high-current event. A series arc takes place between two
ends of a broken conductor or through a poor connection. Such arcs can be continuous
with a low current flow in the circuit. When current tries to pass through a poor or loose
connection (one with high resistance), it can generate heat. In these cases, the heat and its
effects on nearby materials will be localized around the poor connection. These failures
can occur anywhere two wires connect (splices) or where single wires connect to appli-
ances, power sources, meters, and on the like. A good example of such a failure is where
a wire is wrapped around a screw terminal. If the screw is not properly tightened, loosens
by expansion and contraction from heating (especially for aluminum conductors), or

Load FIGURE 10-22 (a) Series

arc-load may “flicker” or
“operate” not alternate.
(b) Parallel arc: load will
Series arc—load may Parallel arc falter due to insufficient
“flicker” or alternate
current.
(a) (b)
Chapter 10 Electrical Causes of Fires 427
 V  2.5 VAC
Ohm’s law: V  I  R

I  14 A
2.5 V  14 A  0.18 
R  0.18 

V  120 VAC
 Power: P  I 2  R
P  35 W
Ground conductors 35 W  (14 A) 2  0.18 

R  8.4  V  117.5 VDC Heat flow: 35 W

P  1650 W
Watt density: 35 W/(0.00006 m 2)  550,000 W/m2

Temperature: over 400˚C (750˚F)

(a) (b)

FIGURE 10-23 (a) Schematic and (b) circuit heat flow analyses of one example of overheating by a poor
connection, a glowing screw caused by a loose connection. Courtesy of Chris W. Korinek, P.E., Synergy Technologies,
Cedarburg, WI.

loosens from vibration, the contact can make and break. The resistance added by the poor
connection may be only a fraction of an ohm, but if the circuit is carrying the current
to serve a high-wattage appliance, that can mean the creation of considerable heat. In
Figure 10-23, a situation is depicted where a poor connection creates a localized resist-
ance of 0.18 . Even at a fraction of an ohm, this resistance is many times the resistance
of a good, clean, tight connection, hence the term high-resistance connection. In a 120 VAC
circuit serving a 1,650-W appliance, this means that nearly 14 A is trying to pass through
that connection (with the contact equivalent of 6mm2), creating 35 W of heat. Although
this is the same amount of heat created by 20 m (66 ft) of NM cable, the heat is localized
at the poor connection. This condition highlights the concept of watt (energy transfer)
density. For the same reason that high heat fluxes cause more rapid heating and more likely
ignition, high watt densities create high temperatures that degrade nearby insulators and
combustibles. Figures 10-24a and b show some of the localized heating effects (sometimes
called glowing connections for obvious reasons). Ignition occurs when the arc ignites the
“cloud” of gaseous pyrolysis products. This ignition may be in the form of a puff of
flame, which, if sustained, can ignite the solid fuels to a sustained flame.

(a) (b)

FIGURES 10-24A AND B A lab-created condition in which a 120-V receptacle with a loose connection carries a
current of 14 A. (a) Initially, a series parting arc is seen when the screw and wire are moved. (b) As the heating
continues, the screw and wire glow as current travels through the oxides formed and produces a relatively high-
resistance connection. Courtesy of Chris W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

428 Chapter 10 Electrical Causes of Fires

 These mechanisms often create series parting arcs as the contact between the wire and
the connector is made and broken by mechanical movement or vibration. The intermittent
arc (current passing through the air gap as the conductors touch and separate) causes very
high temperatures at the tip of the wire. This accelerates the localized heating and degra-
dation and ignites the gases and vapors from pyrolysis.15 The receptacle in Figure 10-25a
was ignited by a poor connection fault. The vibration from the washer and dryer caused
the screw connector to loosen, and the high current for the dryer caused localized heat-
ing of the wire, connectors, and insulating materials. The heated insulation eventually
degraded and was ignited by the arcing within the receptacle. The points of high resist-
ance can be where a portion of the conductor size is smaller (as discussed previously),
where the conductor material is different (as in an electric heater), and at connection
points (ends) where poor electrical contacts or splices are made. See Figure 10-25b for an
example. The NEC and UK wiring regulations specify that all wiring connections be made
in protective enclosures (junction boxes).
To minimize heat buildup at connections, the full cross-sectional area of the conduc-
tor must be maintained, and positive contact between the conductors must be ensured.
This means that no oxide or other films or foreign material can appear between the
conductors. These failures are being seen where volatile sulfide contaminants in Chinese-
made wallboard are corroding copper wires and connectors inside walls.16 The total
resistance of a circuit determines the current flow (at a certain voltage) and thus the total
rate of heat production (watts). The distribution of the resistance throughout the circuit
determines where the heat will be concentrated. An example of the first case would be the
use of the inexpensive “zip” cord plugs that are installed by merely pushing the cord into
a slot in the opened plug and then pushing the plug back together. The connection to the
wire is made by pushing (stabbing) needlelike points through the insulation and into the
conductor. The cross-sectional area of contact is diminished at this point, and if such
plugs are used for anything other than perhaps a clock or small lamp, excessive heating
can occur, as shown in Figure 10-26. Flexing of a zip cord can cause strands to separate
and arc, as illustrated in Figures 10-27a and b. This heating can cause ignition of the insu-
lation of the cord or the plug. Frequently, draperies cover wall outlets and can be ignited
by the small fire supported by the plug and wire.

INSULATION BREAKDOWN (DEGRADATION)—CARBON TRACKING

With the exception of glass, ceramic, mica, and asbestos insulators, most electrical insu-
lation materials are organic compounds containing carbon. Therefore, degradation of

FIGURE 10-25A Vibration

from the clothes dryer
plugged into this 240-V
receptacle caused a loose
connection. The #10 AWG
wire involved is sputtered
away to a fine point. The
phenolic housing pyrolyzed
away until it ignited. The
resulting flames consumed
the exposed wire insula-
tion and the adjacent
plug. The fire was confined
by a metal receptacle
box. Courtesy of John D.
DeHaan.

Chapter 10 Electrical Causes of Fires 429

 FIGURE 10-25B Ignition
of plastic receptacle caused
by loose connection
shown in Figure 10-20a.
Courtesy of Jamie Novak,
Novak Investigations and St.
Paul Fire Dept.

FIGURE 10-26 Zip cord

plugs provide minimal
contact area. Courtesy of
John D. DeHaan.

430 Chapter 10 Electrical Causes of Fires

FIGURE 10-27A Cord flexing at strain relief caused shorting. FIGURE 10-27B Cord flexing caused a series arc connection failure.
Courtesy of John D. DeHaan. Courtesy of John D. DeHaan.

such insulators by heat produces a carbon char, which is an electrical semiconductor.

When this pyrolysis occurs over a large area, it is called carbonization. Breakdown can be
more subtle, because it applies wherever the insulation has lost its insulating capacity,
completely or in part, thereby allowing the current to follow an unintended path. A cir-
cuit must be energized for overheating to occur from insulation breakdown; however, the
equipment does not need to be operating. The circuit can offer an unintended path for
current to flow from hot (energized) to either ground or neutral via conductive contami-
nants or pyrolysis products by three routes: (1) through the insulating materials, (2)
across the surface of the insulating materials, or (3) through air. When current flows in a
limited area between conductors, it is often called arc tracking, arcing through char, or
carbon tracking. The result of this carbon tracking is that something that is supposed to
be an insulator becomes a semiconductor. This transformation often occurs over an
extended time and at such a slow rate that it is not readily detectable until the conductive
path it creates can conduct so much current that massive heating can occur. The time
frame can be as long as months or years prior to the ignition of a fire.17 For instance, if
an insulator absorbs moisture, a small current can flow, producing some heat. This heat
can slowly degrade organic insulation in the vicinity and turn some of it to carbon. Once
enough carbon is formed between the conductors, more current can flow along the car-
bon path, providing localized heating and further degradation. Note that since the cur-
rent is flowing along a path of limited area, even the passage of 1 A or so can create a
very competent ignition source without tripping the circuit breaker protection. As the
process continues, the current progressively increases as more and more carbon is formed.
Finally, unless a breaker or fuse functions, an arc may be struck, igniting the carbonized
insulation or volatile pyrolysis products, possibly resulting in a fire.

Leakage Current
This development of an alternative path is sometimes called leakage current. If the sur-
face of an insulator is contaminated, some of the current leaks across the trail of contam-
inant and begins the process just described. Figure 10-28 shows leakage current heating
the surface of a contaminated plastic plug. Note that the plug is energized but not con-
nected to an appliance. The current being drawn by this flow was 0.25 A or less, but at
120 V, this caused some 30 W of heat to be produced in a very small area and created
such high temperatures that glowing combustion was visible. This situation is more
prevalent in inexpensive electrical appliances in which low-quality insulation is used. This
Chapter 10 Electrical Causes of Fires 431
FIGURE 10-28 Lab-
created leakage current
across the face of a con-
taminated plastic plug.
Courtesy of Chris W. Korinek,
P.E., Synergy Technologies LLC,
Cedarburg, WI.

occurrence has been reported in appliance switches, from the ignition of the insulation
supporting the contacts, and has been seen in many devices—televisions, clothes dryers, cof-
feemakers, smoke alarms, molded plugs, and the like. Materials used for insulation, even in
high-quality appliances, are selected on the basis of the thermal conditions expected under
normal use. If an otherwise safe power cord is used in an abnormally hot environment or
even buried under carpets or blankets, the heat being generated cannot escape. The heat
can then build up, possibly beyond the design limits of the insulation, and the insulation
begins to degrade. It should be remembered that overheating conductors first cause degra-
dation of the insulation surrounding them, followed by degradation of plugs, sockets, and
other hardware. It is only then that heat from an overheated wire will affect other fuels.
Once flame is established, other fuels, whether cellulosic or thermoplastic, can be quickly
involved. Figure 10-29 shows the plastic receptacle plate and plug in flames as the result
of a simulated insulation degradation failure of the receptacle.
This degradation mechanism has been linked to causation of fires in a variety of
assemblies and circumstances: insulators on fluorescent lamp connectors, electrical cords
exposed to external heat, plugs, and receptacles. Just as heat exposure causes chemical
and physical changes in wood, similar changes can take place when plastic insulators are
exposed to heat. This heat can be from high-resistance connections, sustained high cur-
rent flows, or external heating. Loose connections in household (15 A) circuits have been
found to produce connector temperatures between 150°C and 200°C (300° and 400°F).
Such high temperatures (even if cyclic) will cause PVC and urea formaldehyde insulators
to degrade and develop tracking capability. Once current begins to flow through the
degraded material, localized heating and ignition will follow.18

Water and Carbon Tracking

Abuse of an appliance such as soaking it repeatedly with impure water or allowing con-
ductive dusts to accumulate can trigger such carbon tracking. The higher the voltage, the
more likely such tracking can take place. It is less likely in low-voltage (24 V) circuits
but becomes more likely at higher voltages.19 Figure 10-30a shows arc tracking between
two terminals energized with a 12-kV neon sign transformer that was triggered by spraying
distilled water on the wood between the terminals. The water leached enough materials
432 Chapter 10 Electrical Causes of Fires
 FIGURE 10-29 Simulated
residential fire after insu-
lation degradation in an
outlet produces small
flames from receptacle
and ignites nearby
combustibles such as
draperies or upholstery
fabrics. Courtesy of Chris
W. Korinek, P.E., Synergy
Technologies LLC, Cedarburg,
WI.

out of the wood to support leakage current. The arcing produced as the wood pyrolyzed
resulted in the ignition of the wood. Figure 10-30b shows the effect of rainwater that
eventually caused a short in a neon sign connector that nearly ignited nearby wood surfaces.
While pure water is an insulator, ordinary tap water contains enough dissolved salts
to be somewhat conductive, and the greater the salt content, the better the water con-
ducts. Small amounts of impure water, then, can play a role in triggering arc tracking on
insulators, but they can also cause corrosion that can result in high-resistance, heat-
producing connections. These connections create localized heat and pyrolysis of their
insulative supports, and additional current flows as the char develops. When water enters
electrical equipment, it can provide additional paths for the current in parts of the equipment

FIGURE 10-30A Lab-created arc tracking occurring across wood FIGURE 10-30B Neon sign (PK) connector, which had been
after water exposure between two terminals with a potential differ- exposed to rain, shorted, heating surrounding combustibles, and
ence of 12 kV, a typical neon sign transformer voltage. Courtesy of caused extensive damage to an outdoor sign. Courtesy of Joe Bloom,
Chris W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI. Bloom Fire Investigation, Grants Pass, OR, and Chris J. Bloom, CJB Fire
Consultant, White City, OR.

Chapter 10 Electrical Causes of Fires 433

 (a) (b)

(c) (d)

FIGURES 10-31A–D (a) Lab-created flaming ignition of a switch housing after the switch was exposed to water
containing an impurity. The voltage was 120 V, and current did not exceed 1.3 A. (b and c) Plastic between the
terminals of this microswitch has been severely degraded, allowing current to flow between “open” switch
terminals. (d) New switch. Courtesy of Chris W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.

that are not meant to carry such currents. In addition to providing a potential electric
shock danger, the presence of impure water has been seen to cause fires by allowing cur-
rent flows along these unplanned paths. In one case, a storm broke the utility drop con-
nectors and bent the entrance conduit. When repairs were made, the entrance conduit was
not repaired, and it had broken loose at the roof, allowing a path for rain to flow directly
down into the wall and into the electrical distribution system over time. The water con-
tamination ultimately resulted in ignition and fire destruction.
Fire codes list specific types of electrical equipment for wet or damp locations. When
regular equipment is exposed to these conditions, disaster can occur. In the test shown in
Figures 10-31a–d, water with an impurity was applied to an electrical switch. Sizzling,
smoking, and flaming occurred as the plastic in the switch overheated. Note that the light
was in series with the fault and limited the current flow in the circuit. The OCPD was not
tripped because the current did not exceed 1.3 A. Wall receptacles under towel dispensers
in public restrooms are often affected by water corrosion.

Heating and Carbon Tracking

Heating of an insulated cable (or insulated wires inside a conduit or appliance) will cause
the insulation to degrade, sometimes over a substantial area. Energized conductors within
can then provide current that flows through the charred insulation (arcing through char)
to either the ground or the neutral conductor or the grounded conduit. With conduit, this
can “blow” a hole through the metal in the vicinity of the arc (as seen in Figure 10-32).
Note that such a hole is almost always a result of a preexisting fire and not an initial
cause. The localized heating then can pyrolyze and ignite the insulation. Molten metal
particles can also be ejected to fall onto other combustibles below. Very high current
434 Chapter 10 Electrical Causes of Fires
 FIGURE 10-32 Lab-
created arcing through
char (caused by the heat
of an external fire) that
melted holes in the steel
conduit enclosing the
wire in a high-current
circuit. Courtesy of Chris
W. Korinek, P.E., Synergy
Technologies LLC,
Cedarburg, WI.

single-phase or three-phase circuits will allow enough current flow between a damaged
insulated cable and a grounded conduit (EMT) or flexible armored cable (BX) to melt a
hole through conduit to cause a fire, as shown in Figure 10-33.
A very different mechanism is illustrated in Figure 10-34. There the metal flexible
(BX) cable housing became energized due to an open neutral and became a heating ele-
ment that ignited the adjacent wood joists.

ARCS AND SPARKS

As we saw in Chapter 6, an arc can be defined as the flow of current through a gas (usu-
ally air). It can range from a brief discharge (spark) built up by walking on a dry carpet
to a bolt of lightning. If an arc is sustained, the atmospheric gases through which the cur-
rent is passing can be supplemented by ionized (plasma) vapors of the conductors

FIGURE 10-33 Conduit

from service mast
involved in fire showing
arcing and penetration
from within. Current
would be limited here
only by the line trans-
former (2,000–3,000 A).
Courtesy of E. R. Baker, Fire
Investigator, Fresno, CA.

Chapter 10 Electrical Causes of Fires 435

FIGURE 10-34 Armored
cable became energized
when open neutral
occurred in meter box
and acted as a heating
element to ignite ceiling
joists. Courtesy of Detective
Mike Dalton, Knox County
Sheriff’s Office, Knoxville, TN.

involved (or, in the case of welding, by the vapors of a flux or specialized gas). For an arc
to establish itself in air between two conductors requires either a high voltage or a very
small gap between the conductors, because air is a good insulator. Once an arc is estab-
lished, however, it can be maintained as the conductors separate via the conductive path
offered by the vaporizing metal and the ionized air molecules. This is especially true with
DC currents, whose voltage does not cycle through zero periodically as does AC. As
described previously, an arc produces very high temperatures, on the order of thousands
of degrees, but they are very localized, so a fuel has to be in almost immediate contact
with the arc for it to ignite. Gases and vapors, including those from active pyrolysis of
insulators, are much more readily ignited to flames than are solid fuels. Unless a solid fuel
is very finely divided or thin (like paper), it will resist ignition by all but the most ener-
getic and prolonged arcs. Thin paper, lint, sawdust, and flammable gases and vapors are
all ignitable by 120/240-V arcs in AC circuits.
When an excessive current is passed through a metallic conductor, its temperature
rises to the melting point of that metal. As the first melting occurs, the wire severs and a
parting arc occurs at this point (which may be sustained at high enough voltages). This
arc creates very high temperatures locally, and melted metal is spattered onto cooler
nearby surfaces, where it recondenses as droplets. While these sparks (often incandescent
particles) are very hot, they do not have sufficient thermal capacity to ignite massive solid
fuels nearby, except under the most unusual conditions. But recalling the information on
pyrolysis from Chapters 3 and 5, we realize that sustained heat breaks down most plas-
tic or cellulosic insulators into combustible gases and vapors and conductive char. It is
often the gases and vapors that are first ignited by the arc itself.
The breakdown voltage of dry air is about 30,000 V/cm (75 kV/in.), but the mini-
mum voltage at the minimum spacing (苲0.01 mm) is still 340 V (DC or RMS AC – 240
VAC peak  340 V RMS).20 So, if two bare conductors are kept separate in air by even
a very small amount, an arc cannot be started. If the two conductors touch and then are
separated, the ionized gases and metallic residues can maintain the arc at some distances
until the voltage differential goes to zero. In a 120-V 60-Hz AC circuit, arcing lasts typi-
cally 0.5 to 4 cycles (0.008 to 0.05 s).
436 Chapter 10 Electrical Causes of Fires
 An AC arc tends to self-extinguish as its cycle goes through zero volts in the AC volt-
age sine wave. A DC arc can be maintained for as long as there is power. The duration of
the arc is critical to its capacity to ignite nearby fuels. While the temperature of
gases/plasma in the arc is a function of current and is thousands of degrees Celsius, the
duration of the typical arc in an overcurrent-protected circuit will usually be so brief that
ignition of solid fuels is nearly impossible, except, for example, when the fuel is finely
divided (cotton waste or fine sawdust, for example) and in direct contact with the arc or
when the overcurrent protection fails to interrupt the circuit. Flammable gases and
vapors, of course, can be ignited by all but the shortest arc due to their low ignition ener-
gies and direct contact with the superheated gases of the arc.

ALUMINUM WIRING
Aluminum is sometimes encountered as a substitute for copper because of its lighter
weight and lower cost. Aluminum was used for a period of years in homes and mobile
homes for all internal wiring; however, today it is mostly used for high-voltage distribu-
tion lines and in subpanels. Aluminum cable is routinely used for service drops (from pole service drop ■ The
to weather head). Because aluminum is a poorer conductor than copper, to carry the same overhead service con-
ductors from the pole,
current as a copper conductor safely, the aluminum conductor must be larger. However,
transformer, or other
because aluminum is lighter, less conductor weight is required. Because aluminum costs aerial support to the
considerably less per pound than copper, aluminum is cheaper to use overall. service entrance equip-
The major problem with aluminum is that of making mechanical connections. Most ment on the structure,
difficulties with aluminum wiring are found not in the wire span but at the ends where including any splices.
the connections are made. Failures have resulted from directly substituting aluminum
wire for copper; for example, using the fittings, receptacles, and switches that were
designed for copper wire. The quality of the installer’s work is critical to avoiding prob-
lems that result from aluminum’s peculiar physical and chemical properties. Even the best
electrician cannot install small-diameter aluminum wiring to be robust and maintain a
good connection indefinitely. The following are some points to keep in mind.
First, the electrical conductivity of aluminum is somewhat poorer than that of cop-
per. As a general principle, aluminum wire should be two sizes larger (two AWG numbers
smaller) than its copper equivalent. Wire tables should be consulted for exact size require-
ments. This requirement may be overlooked by some installers.
Second, aluminum is very chemically active. Exposed surfaces always have an oxide
coating. This oxide film acts as an insulator and is a major problem in making good elec-
trical connections. Chemical preservatives (antioxide pastes) are usually applied when
making connections to minimize this oxide film formation.
Third, the thermal expansion coefficient of aluminum is different from that of cop-
per, and for any combination of these two metals, allowances must be made for this dif-
ference. Connections that are tight at one temperature may be loose at another. Also, the
continued tightening and loosening due to changing temperature can result in a perma-
nently loose connection.
Finally, another problem, mainly with the early products, was cold flow. Aluminum
creeps like putty when subjected to pressure, and it flows away from the pressure points,
resulting in a poor connection.
The basic problems with aluminum wiring have largely been corrected by new alu-
minum alloy connectors. The connectors, receptacles, and switches appear similar to their eutectic ■ An alloy of
old counterparts, but the selection of materials and their design has, for the most part, two materials having
corrected the problem. The devices acceptable for aluminum are so marked (AL/CU). special physical or
chemical properties,
Aluminum wiring connection failures can still occur even with AL/CU devices.
typically having the
Aluminum, having a lower melting temperature (660°C; 1,220°F), than copper lowest melting point
(1,083°C; 1,981°F), will usually be melted by most fires and therefore may not provide of any combination of
useful information after a substantial fire. As it melts, it can form eutectic alloys with the two.

Chapter 10 Electrical Causes of Fires 437

 copper conductors and cause them to melt away, sometimes giving the appearance of a
massive failure of the copper conductor. The presence of a gold “brass”-colored “stain”
on the remaining copper is a clue that there may have been aluminum alloying.

ELECTRIC TRANSFORMERS AND MOTORS

Electric transformers and motors are similar in that they contain copper wire that is
wound on an iron core that creates magnetic fields. Both can overheat from excessive cur-
rent flow, and this excess heat can cause insulation destruction that, in the absence of
proper overload protection, is capable of causing ignition.
Electric Transformers
Transformers contain several windings with some windings that have wires connected
to the midpoint of the windings (called mid-winding taps). They are used for raising or
lowering the voltage (and current) in an AC circuit. Large transformers frequently are
oil-filled to insulate them and to provide a cooling bath for the heat normally produced.
Large power transformers of this type have a potential fire hazard from the high volt-
ages and the presence of certain types of combustible oils. However, modern transform-
ers are sometimes filled with noncombustible oil or are dry-type devices (air-cooled
with no oil).
Fire problems with transformers are similar to those in motors; however, the mechan-
ical and high starting current conditions of motors are not applicable. Some electronic
equipment containing transformers has been known to start fires because of the absence
of fuses or improper fuse selection by the manufacturer. High-current devices may be able
to induce heating by magnetic induction, but this is beyond the scope of the investigations
covered by this text.
Electric Motors
In addition to the windings, motors have moving parts. Some have centrifugal switches
that operate when the motor attains 75 percent of its normal operating speed. Other
motors require an electrical connection to the moving part (armature or rotor), accom-
plished by brushes riding on a moving contact (commutator or slip ring).
Electrical motors present numerous possibilities for ignition, including the following:
■ Overheating because of excessive current
■ Defective windings or excessive mechanical load, which prevents attainment of
proper operational speed
■ Arcs from the centrifugal switch or brushes
■ Sparks from frictional contact with moving parts
duty ■ Conditions of ■ Overheating due to excessive use beyond its rated duty cycle
use in electrical service.
These hazards may be overcome by construction of the motor frame so that electri-
cal parts are suitably enclosed, or by keeping combustible material away. The starting
equipment of a heavy-duty motor presents hazards in the forms of arcing of current-
breaking contacts and the heat produced in some forms of starting equipment. The haz-
ards of overloading a motor may be overcome by proper overcurrent protection.
Fire hazards common to motors include the following:
■ Location close to combustible material that may be ignited by arcs or sparks.
■ Location in a damp place or where subject to corrosive vapors.
■ Linty or dusty condition of environment.
■ Burning out (failure of insulation on windings), which may be due to overload,
stalling of the rotor, or, in some cases, too low a voltage at the motor terminals.
■ Improper overcurrent protection. Except for manually started motors of 1 hp or
less, every motor should be protected by a suitable overcurrent device. (Such
devices may be internal or external.)
438 Chapter 10 Electrical Causes of Fires
■ Starting equipment that produces arcs and that is not located away from combustible
material.
■ A three-phase motor protected by fuses in which one fuse blows or one leg of the
circuit opens while the motor is started or operated.

Post-fire Examination of Motors

Examination of motor remains can frequently provide useful information. When electric
motors are found to be in a fire area near a suspected point of origin, a thorough evaluation
should be made to determine whether the fire was deep seated in the windings or whether
damage resulted from heat conducted to windings from fire outside the motor housing.
Figure 10-35b is an example of a failed winding in a motor that ignited the insulating varnish
and led to a fire. If the shaft is frozen (will not turn), it may be an indication that the heat
was internal. Because small motors have much less mechanical inertia, this symptom is not
reliable. If the motor was running during the fire, melted bits of solder thrown about the inte-
rior of the motor housing may sometimes be found in motors having a commutator (a rotary
electrical connector with soldered leads). Friction-caused overloading is indicated if the motor
belt is most heavily damaged where it passes over pulleys. Fire from other sources damages
motor belts much more between the pulleys and often does less damage to belts protected by
the pulleys. If immediately after a fire the motor housing is too hot to touch, but iron or steel
of similar size in the same area is relatively cool, it indicates internal heating of the motor. A
stalled motor, or one that is not running up to speed, will overheat from excessive current,
which can be caused by a defective motor, failure of the winding’s insulation, tight bearings,
too small a motor for the purpose, frozen or tight load, and the like.
Frequently, when an electric motor from a fire is completely gutted inside, it does not
prove that the motor was faulty. It can also indicate an overload condition or possibly
that something jammed the motor and prevented it from turning. If the motor was able
to operate with the load partially jammed, a fire could have been started from friction of
the belts or pulleys or an external interference (as in Figure 10-36). Fires from electric
motors can also be caused by bearings that were not lubricated, faulty starting mecha-
nisms, or excessive dirt in the motor. Excessive dust in a motor is often found where the
motor was used to power shop tools. Dusts can be subject to hot-surface ignition if
allowed to accumulate on hot motors.
Old motors are generally physically larger, for a given horsepower rating, than mod-
ern motors. In addition, modern motors are frequently designed to operate at a higher

FIGURE 10-35A Motor failure when interlock shorted out to FIGURE 10-35B Close-up of motor winding failure. Courtesy of
armature. Courtesy of Peter Brosz, Brosz & Associates Forensic Services Inc.: Jean-Claude Martin, Institut de Police Scientifique et Criminologie (IPSC), Lausanne,
Professor Helmut Brosz, Institute of Forensic Electro-Pathology. Switzerland.

Chapter 10 Electrical Causes of Fires 439

 FIGURE 10-36 Chafing
of the motor shaft against
a piece of clothing
wedged in the appliance
caused a dryer fire.
Courtesy of Joe Bloom,
Bloom Fire Investigation,
Grants Pass, OR.

temperature than old motors because of the use of modern high-temperature insulation.
The nameplate or manufacturer’s literature should be consulted before a modern motor
is suspected of overheating.

FIXED HEATERS
Electric radiant heaters that are fixed installations and hardwired into the electrical sys-
tem of a building should not be overlooked as potential fire causes. They can be located
in walls, in floors, and even in the ceilings of structures. They are designed to operate
under certain conditions of extended operation and limited heat dissipation, but because
they may be concealed, thought is often not given to the combustibles or the insulation
packed too close, beneath, or over them. The Consumer Product Safety Commission
(CPSC) estimated that fixed (gas and electrical) heaters and furnaces were responsible for
4,500 fires, 8 percent of all 56,500 accidental residential fires in 2004–2006.21

APPLIANCES
It must be remembered that electrical equipment can be divided into two general cate-
gories: the fixed conductors, switches, protective devices, and outlets of the distribution
system; and all portable, movable lamps, tools, and appliances that can be connected
into that system. The fixed distribution system must (in most areas) meet fairly strin-
gent guidelines and is not likely to be changed by the average owner or tenant, so it usu-
ally poses less likelihood of unwanted heating. Appliances and all other removable
440 Chapter 10 Electrical Causes of Fires
FIGURE 10-37A Coffeemaker ignited by runaway heating when FIGURE 10-37B Post-fire remains show heating elements, power
TCOs were disabled. Courtesy of Steve Riggs, Public Agency Training Council. cord, and TCOs. Courtesy of Steve Riggs, Public Agency Training Council.

electric utensils are subject to much more misuse and can be added together to pose all
manner of risks. By far the largest percentage of fires are ignited by the appliances, but
these fires are also more easily detected and extinguished by the user than are the fires
hidden in walls and ceilings. See Figures 10-37 and 10-38 for examples of small-appliance
fires.
Electric appliances are found in nearly all structures, and whether they produce
quantities of heat by design or by failure or simply represent a convenient source for
that all-important arc, their possible contribution to fire causation should not be over-
looked. Faults with power cords and plugs may result from poor design or misuse by
the consumer. Because power cords and plugs are subject to misuse, faults in them
probably represent the majority of fire causes from appliances. The symptoms of poor
contact resistance and inadequate conductor sizes were discussed previously in this
chapter. A new problem source is Chinese-made appliances, extension cords, or power
strips that are made with conductors that are far too small to carry normal currents.
These small conductors can overheat quickly and cause ignition of adjacent plastic
components.
Some appliances, however, because of their high power load, high voltages, or typically
unsafe locations, deserve special mention.

FIGURE 10-38A Runaway clothes iron (TCO FIGURE 10-38B Plastic components support FIGURE 10-38C Post-fire remains show
disabled). Courtesy of Steve Riggs, Public Agency large flames. Courtesy of Steve Riggs, Public Agency melted aluminum base and heating element.
Training Council. Training Council. Courtesy of Steve Riggs, Public Agency Training Council.

Chapter 10 Electrical Causes of Fires 441

FIGURE 10-39A Pattern on walls indicates origin in dryer. Courtesy FIGURE 10-39B Interior fire burned clothing in drum. Courtesy of
of Det. Mike Dalton, Knox County Sheriff’s Office, Knoxville, TN. Det. Mike Dalton, Knox County Sheriff’s Office, Knoxville, TN.

Clothes Dryers
High power usage, high temperatures produced even in normal use, poor locations with dust
and dampness common, and poor maintenance (such as not cleaning lint filters regularly) are
associated with fires in clothes dryers. Lint also can accumulate around motors and heating
elements (especially if the dryer is not properly vented) (see Figures 10-39a and b.) Clothes
dryers are responsible for an estimated 11.7 percent of all residential fires.22

Portable Electric Heaters

High power usage, high temperatures, and careless use (blocked by furniture or in con-
tact with flammable furnishings or clothing, for example) are associated with fires involv-
ing portable electric heaters. These devices are responsible for an estimated 2.8 percent of
all residential fires.23

Electric Ranges, Electric Ovens, and Small Cooking Appliances

High power usage, high temperatures produced, and proximity of fuels (such as foods,
utensils, or clothing) are associated with fires involving electric ranges and ovens and
small cooking appliances. The CPSC data indicate that these appliances are responsible
for an estimated 37.3 percent of all residential fires; however, these are often cooking
fires. Small cooking appliances can be dangerous due to the fuel load represented in the
food. Figure 10-40 shows a cooking-oil fire in a frying pan. CPSC estimated that cooking
equipment was responsible for 7.5 percent of fire deaths in 2004–2006.24

Dishwashers
Even though dishwashers use water inside to clean dishes, they have a number of electri-
cal overheating failure modes. These have been known to include overheating of drying
elements, lack of prevention of wiring damage during opening and closing, and egress of
moisture and other conductive fluids into area where wiring is present, causing arc track-
ing and ignition of nearby combustibles.25
Television Sets
High voltages and poor design (including the frequent use of flammable materials for cab-
inets) are associated with television fires. According to the CPSC, there was an extremely
high incidence of fires in sets made in 1970–1974. In 1979 the CPSC estimated that
approximately 2 percent of all residential fires, some 11,000 incidents causing 160 deaths,
resulted from television sets. Since 1974 the frequency has been decreasing because of the
442 Chapter 10 Electrical Causes of Fires
 FIGURE 10-40 A cooking-
oil fire in this electric skillet
involved the cabinets above
by convective flame plume.
The fire spread from this
point was very rapid.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion Investigator.

use of newer sets that have better wiring, lower operating temperatures, and cabinets
made from fire-resistant materials.26 By 2006, that figure had dropped to 700 fires,
0.2 percent of residential fires, with no recorded deaths (including radios and phono-
graphs).27 Except for televisions sold in North America and Japan, the plastic cabinets are
made from non-flame-resistant styrene or similar material. In the UK there have been a
large number of TV fires started by unattended candles.
It should be noted that it is possible for fires of both accidental and incendiary origin
to destroy all combustible elements within a modern TV. In such cases, even highly expe-
rienced laboratory examiners may find it impossible to determine whether the fire started
within the set or elsewhere because of the destruction of the diagnostic signs necessary for
such a determination.

ELECTRIC LIGHTING
Lighting fixtures and related issues of particular interest to the fire investigator include
magnetic and electronic fluorescent light ballasts, overheating in fluorescent lam-
pholders, lamp failures at the end of life, and other lamp considerations, as discussed
next. During the period 2003–2006, the NFPA estimated that annually, 2,800 fires in
nonresidential structures were attributable to the general category of lighting.28 The
NFPA also estimated that some 5,770 home structure fires were linked to lamps, light
fixtures, or lightbulbs, which caused 76 deaths and $181 million in damages in
2006.29
Chapter 10 Electrical Causes of Fires 443
 Magnetic Ballasts
Ballasts used in fluorescent lighting fixtures create a high voltage to initiate a discharge
(current) through the gas inside and to limit the current once the arc is established.30
While the majority of new ballasts put into service today are electronic, there are still
millions of magnetic ballasts in use, especially in fixtures that utilize the older 3.81-cm
(1.5-in.) diameter T12 lamps. In Europe, magnetic ballasts are still frequently used in
new fluorescent light fittings. Magnetic ballasts generate a considerable amount of heat
when in operation. In North America, ballasts manufactured for new fixtures since
1968, and all ballasts manufactured since 1984 have been equipped with autoresetting
thermal protectors, which greatly reduce the potential for fires caused by overheating
ballasts.31 However, magnetic ballasts are not failure-proof. If the protector fails to sense
an overtemperature condition or its contacts become welded shut, an overheating situa-
tion could be allowed to continue. Under ideal conditions, such a ballast may reach tem-
peratures high enough to ignite combustibles above the fixture, such as low-density
ceiling tiles or cellulose thermal insulation. Continued overheating can lead to an arcing
event that is energetic enough to penetrate both the ballast case and the enclosure of the
fixture (as shown in Figure 10-41). Molten metal and burning gasses from the vaporized
potting compound ejected out of the fixture as the result of such an event may have
sufficient energy to ignite combustible materials nearby.32 Most fixtures have metal cov-
ers to prevent such contact, but covers may be missing due to poor maintenance. The
NFPA statistics for 2003–2006 indicate that 26 percent of the fires in nonresidential
structures attributable to lighting were caused by fluorescent fixtures or ballasts.33 (See
Figure 10-41.)
Older fluorescent light fixtures with magnetic ballasts can cause fires from a break-
down of the ballast transformers, which contain a flammable potting compound. Some
older ballasts have a starting voltage as high as 700 to 900 V and operate at about 400 V;
newer units operate at lower voltages. They are designed to operate continuously at over
40°C (100°F), but the temperature can go much higher. This heat sometimes causes melt-
ing and vaporizing of the combustible contents, sometimes igniting combustible ceiling
material. The vapors can collect and ignite in the housing.

FIGURE 10-41 A field

example of an arcing
event inside a magnetic
ballast that breached the
metal fixture body.
Courtesy of David W. Powell,
SYTEK Consultants, East
Syracuse, NY.

444 Chapter 10 Electrical Causes of Fires

 Early ballasts had no thermal protection. Current or newer manufactured units contain
a thermal overload switch that is intended to disable the appliance in the event of overheating.
These thermal protectors are typically designed to open at 110°C  5°C (230°F).

Electronic Instant-Start Ballasts

Electronic ballasts generate much less internal heat and operate at considerably lower
temperatures than their magnetic counterparts. As a rule, they are fused internally or are
inherently thermally protected. As a result, fire causation due to electronic ballast failure
is extremely rare.
Instant-start electronic ballasts use a high voltage (5–600 V) to start the lamps. If the
lamp is inserted improperly, arcing in the lampholder, between the lamp pin(s) and the
lampholder contacts can occur due to the high voltages present. The extremely high tem-
peratures resulting from a prolonged arc can result in degradation of the lamp holder into
combustible materials, melting of the brass contacts, and possible ignition, as shown in
Figures 10-42a–c. Because the lamps in the fixture can remain lit while the arcing is occur-
ring in the lampholder, the overheating event can progress, possibly to ignition, before
occupants are aware of the problem.34

FIGURE 10-42A A lab-created arcing event in progress within a FIGURE 10-42B The result of a brief arcing event at the
lampholder in a fluorescent light with an instant-start electronic lampholder contact as a result of improper insertion of the lamp.
ballast. Courtesy of Richard E. Korinek, P.E., Synergy Technologies LLC, Arbor Courtesy of Richard E. Korinek, P.E., Synergy Technologies LLC, Arbor Vitae, WI.
Vitae, WI.

FIGURE 10-42C A field

example of arcing at a
lampholder contact
(second from left) that
resulted in a fire. Courtesy
of David W. Powell, SYTEK
Consultants, East Syracuse, NY.

Chapter 10 Electrical Causes of Fires 445

 Fixtures for 8-Ft (Commercial) Lamps
Some large lampholders incorporate spring-loaded contacts in which the springs conduct
the lamp current. A number of fires have been attributed to thermal relaxation of the
springs in the lampholders.35 Aging and thermal stress can result in poor connection in
the lampholder. Arcing can occur, degrading the lampholder while the lamps remain lit.
In another variation, improper spacing or mounting of the lampholders on the fixture
housing can cause poor contact pressure in the lampholders, resulting in arcing and over-
heating in the lampholders.
Compact Fluorescent Lamps (CFLs)
Compact fluorescent lamps are becoming very common, and heat-producing failures in
the electronics concealed in their plastic bases are being seen, as depicted in Figure 10-43.
These lamps are intended for base-down use, but many household lighting fixtures
require horizontal or even base-up installation. Some CFLs have already been the subject
of CPSC recall, prompting more reported failures.36 Such lamps can also overheat if used
with a dimmer switch.37
T5 Fluorescent Lamps
Lighting systems are now available that utilize new thinner fluorescent lamps, such as T5
lamps, with 15-mm (5/8-in.) diameters. At the end of a lamp’s life, overheating at the lamp
ends can occur, resulting in cracking and melting of the glass envelope, and overheating of
the lamp end.38 Temperatures can approach 700°C (1,335°F), which is high enough to ignite
combustibles nearby. Such overheating may also damage the lampholder, resulting in poor
connections between the lamp pins and the lampholder, leading to further overheating.39
In any situation where overheating occurs at the end of the fluorescent lamps, the pos-
sibility exists for such an occurrence to result in flaming ignition if there are combustible
materials nearby. For example, many fixture diffusers are made of acrylic plastic, which has
a relatively low ignition temperature in the range of 195°C–289°C (383°F–552°F).40

FIGURE 10-43 Electro-

nics in the plastic base of
a CFL may fail, perforating
housing, and cause an
external fire. Courtesy of
Steve Riggs, Public Agency
Training Council.

446 Chapter 10 Electrical Causes of Fires

Aquarium Lights
The moist environment near aquariums can result in corrosion of electrical connections
in the lampholders. Overheating at a poor connection has been shown to generate heat to
ignite the lampholder and then spread to body of the fixture, which is often made of a
combustible plastic. Insufficient contact pressure may be a contributing factor in these
fires.41 An example is shown in Figures 10-44a–c.

Metal Halide Lighting

Metal halide (MH) lighting is a type of high-intensity discharge (HID) lighting, along with
mercury vapor and high-pressure sodium types. Inside the thin outer glass shell of the MH
lamp is an inner arc tube made of fused quartz. The arc tube operates at extremely high tem-
peratures, up to 1,100°C (2,000°F) and pressures ranging from 73 to 441 psi (5–30 atm).42
The majority of MH lamps fail passively at their end of life. In some cases, the inner
arc tube can break due to the weakening of the quartz arc tube. The breakage can be a
violent event, rupturing the outer glass envelope and scattering 1,100°C (2,000°F) particles
of quartz glass out of the fixture. Hot quartz particles from such an event have been found
to ignite combustible components of the light fixture, materials below, and even wooden
gym floors.43
Incandescent Lamps
As we saw in Chapter 6, in normal (base-down) use, most incandescent lamps do not pro-
duce surface temperatures sufficiently high to ignite most combustible materials in open
air. There is no guarantee to the fire investigator, however, that such lamps will have been
used properly. Normal use for most lamps entails free circulation of cooling air around
the lamp. Any enclosure that restricts the flow of air around a lamp must be of nonflam-
mable materials, and the investigator must determine that no combustible materials came
into contact with the surface of the bulb while incandescent. Lamps that are portable by
their design—drop cords and inspection lamps—are more likely to be covered by rags,
clothes, or wood shavings during their use. If left on in this covered condition, they can
readily produce surface temperatures in excess of the ignition temperature of cloth or
wood (as discussed in Chapter 6). Lamps and light fixtures were associated with some 4.7
percent of residential fires in 2006 according to the CPSC.44

FIGURE 10-44A A field example of an FIGURE 10-44B Corrosion at the end of a FIGURE 10-44C A field example of a lamp-
aquarium light in which an overheating lamp- lamp that was in an aquarium light that holder in which an overheating connection
holder was the cause of a brief residential caused a structure fire (field example). The at the connector on the right side of the
fire. Courtesy of Richard E. Korinek, P.E., Synergy fire occurred at the other end of the lamp. lampholder caused the fire in an aquarium
Technologies LLC, Arbor Vitae, WI. Courtesy of Richard E. Korinek, P.E., Synergy light. Courtesy of Richard E. Korinek, P.E., Synergy
Technologies LLC, Arbor Vitae, WI. Technologies LLC, Arbor Vitae, WI.

Chapter 10 Electrical Causes of Fires 447

 High-wattage incandescent lamps (over 150 W), due to their higher operating tem-
peratures, can pose an ignition risk if ordinary combustibles come into direct contact with
their outer envelopes. As we saw in Chapter 6, halogen lamps have extremely high sur-
face temperatures at their quartz envelope.45 The lamps can produce high temperatures
at their metal shades, and their intensity can melt light diffusers and other plastics set too
close in front of them.

ELECTRIC BLANKETS
Electric blankets have been responsible for some fires. The cause of these fires can fre-
quently be attributed to misuse. Some blankets are controlled by a thermostat located at
the edge, and when blankets are partially covered with additional blankets or other mate-
rials, it is possible for other portions to become overheated. If the main control switch is
left on “high” in a cold room, the current can remain on continuously and cause the cov-
ered section of the blanket to overheat. Blankets have been known to cause fires when
tucked between a mattress and padded box spring, wedged between the bed and a wall,
or folded at the foot of the bed. It can be difficult to determine whether a bed fire was
caused by a blanket or other cause such as smoking in bed, unless the fire was detected
early and the char path is traceable. Sometimes the thermostat and control will remain
after a fire, and these can provide information as to whether the blanket was turned on.
Some blankets have multiple TCOs embedded in the blanket, so that if even a portion of
the blanket reaches a high temperature, the current is interrupted. The location and dis-
tribution of the fine heating wires can reveal the use (or abuse) of the blanket at the time
of the fire.
Electric blankets have evolved from the continuous-resistance-wire heating elements
protected by numerous TCOs to positive temperature coefficient (PTC) cable. This cable,
whose resistance varies with its temperature (the hotter the cable, the higher its resistance),
has two conductors, each wrapped around a polyester core and then embedded in insulat-
ing layers. There is typically an external thermostatic control and a safety circuit board
sewn into the blanket. Failures can occur at the connectors or the safety circuit board, or
broken PTC wires can cause arcing and arcing through char as the cord degrades.46

EXTENSION CORDS
Although they are not part of an appliance, extension cords are a movable element of the
electrical system. They are involved in a large number of fires in North America and other
countries because of the variety of ways they can be abused. Most often, their multiple out-
lets offer the chance to plug in so many appliances that their ampacity (typically 5 to 15 A)
is readily exceeded. The nature of their use results in their being hidden under furniture
or carpets, with the result that their heat dissipation is severely limited. The insulation
quickly melts and chars in overcurrent situations, and the char provides a semiconductive
path between the conductors that can create even more heat without drawing enough
current to trip the protective devices. They can also be used in hidden spaces (behind base-
boards, for example) or suffer so much damage during the fire and suppression that they
can be overlooked during post-fire examination.

HEAT TAPES AND HEAT CABLE

Buildings in cold climates may have electrical-resistance heaters in the form of tape strips
that are designed to be wrapped around water or sewer pipes that might be susceptible to
freezing, or laid in gutters to reduce ice/snow loading. The tape may then be wrapped in
layers of combustible or noncombustible insulation, concealing it from view. Heat tape
typically uses a resistive wire heating element either wrapped helically around a core fil-
ament or zigzagging between two bus wires, with an integral thermostat. Heat cable con-
tains two parallel conductors straddling a conductive, carbonaceous matrix whose
448 Chapter 10 Electrical Causes of Fires
resistance is dependent on its temperature. At warm temperatures, the heat cable matrix
has a high resistance, so no current flows; at lower temperatures, its resistance decreases,
allowing current to flow and provide heat.47
Maintenance of these installations is often neglected, and the tape insulation can fail
(along its length as well as at its connections) from a variety of causes after prolonged use.
As the insulation fails, arc tracking can cause prolonged external heating while not exceed-
ing overcurrent protection limits. Such tapes have been linked to fires in many locales.48

BATTERIES
Lithium-ion (Li-ion) and lithium ion–polymer (Li-Po) batteries have the potential to over-
heat and ignite due to their high energy density, flammable internal materials, and suscep-
tibility to generate oxygen.49 Failure modes include short circuits and overcharging. There
have been numerous major recalls of lithium batteries by laptop computer companies in
recent years. Lead-acid batteries have been known to explode due to short circuits inside
the batteries or when loose external connections ignite hydrogen gas evolved during the
charging process.50 Some battery chargers can develop poor connections, and charging
non-rechargeable batteries can cause overheating and ignition.

Investigation of Electricity-Related Fires

When a fire occurs, suspicion of its cause often falls on any electrical equipment present,
in part because of the basic fact that electricity can cause heat, and this heat can cause a
fire. Because most structures contain electrical wiring or electrical apparatus of some sort,
electricity-related causes are often suspected. However, in a notable study in which a num-
ber of fires were reexamined by fully qualified scientific investigators, the 20 percent of fire
causation attributed to electrical fires was estimated to be on the order of only 9 percent
upon further scrutiny.51 Some investigations focus on electrical causation for no better rea-
son than suspicion based on misunderstanding and mythology. Basically, when consider-
ing an electrical cause, the investigation should address the following possibilities:
■ Was the installation or appliance designed to produce heat? If so, how did the heat
transfer from the unit to ignite the first fuel ignited? (Without knowing what the
first fuel was, how can one realistically evaluate the possibilities?) The same analysis
applies when heat is a waste product (such as with incandescent lamps).
■ If the unit or appliance was not designed to create heat, how could electricity
escape from its intended path to produce heating and thereby the ignition?
■ Was thermal protection present, and if so, why did it fail to protect the appliance?
When electrical components or wires are found melted after a fire, there is immediate
suspicion that the fire and the melting were directly due to an electrical failure of some
sort. The following questions should be considered at this point:
■ Was the melting caused by the fire and not by an electrical fault?
■ Is there any indication that electrical failure occurred? (It is not recommended that an
item of evidence be plugged in or energized. This can destroy evidence and should be
done only after extensive examination and after the approval of all parties.)
■ When an electrical failure is detected, was the failure caused by the fire, or was the
fire caused by the failure?
■ Is enough energy released by the electrical source to ignite the first fuel in the form
in which it was presented to the ignition source?
Fires can be ignited by electrical equipment by one (or more) of the following events:
■ Design errors—The heat shielding on a resistance heater is not designed correctly and
the back of the heater can become dangerously hot, or electrical insulation between
Chapter 10 Electrical Causes of Fires 449
 two conductors is not designed properly and the unit can flex, bringing the two con-
ductors into contact. Thermal protection devices are absent or poorly located.
■ Manufacturing defect—The unit was designed properly, but someone forgot to
install the heat shielding; the wrong wire or insulation was used when the unit was
built; electrical connections were left loose, and the high resistance caused by the
poor connection caused the insulator to pyrolyze and ultimately arc track, with
resulting ignition of the plastic insulator housing.
■ Improper installation or use—A heater is used inside a closet where its housing
comes into contact with combustibles, or it is installed in an insulated enclosure
where the convective cooling needed to maintain safe external temperatures is lost.
■ Use of approved equipment for an unapproved application—Equipment is used in
a high-humidity or high-ambient-temperature environment for which it was not
designed; a thermostatically controlled heater is used in an explosive environment.
■ Accidents—A heater is properly designed, built, and installed, but rags drop onto
its wire shield and ignite; a can of lacquer thinner is used to clean tile spills, and
vapors come into contact with thermostat contacts as they open.
If properly designed, installed, and maintained, electrical systems for lighting, power,
heating, and other purposes are convenient and safe; but if not installed, used, and main-
tained in accordance with well-established standards, they may introduce both fire and
personal injury hazards. This fact has been recognized since electric power first came into
general commercial use. Unfortunately, however, electricity has historically received the
blame by fire investigators when they could not determine any other logical method of
ignition. There are methods by which useful information can be obtained from electrical
debris, and interpretation of this information by a person knowledgeable in the field can
very frequently provide valuable testimony.
Over the years, a variety of indicators or guidelines for determining an electrical cause
of fire have evolved. Some are useful, but unfortunately, some are not as reliable as gen-
erally believed. Little work has been done to substantiate these indicators because they
have been based primarily on what appeared to be logical explanations. In fact, some
experimental work has shown that some much-used indicators are produced by unex-
pected mechanisms.52 When there has been doubt as to the ignition source, reports have
named electricity as the sacrificial lamb by noting that “since no other source was deter-
mined, the cause probably was electrical.” Such speculative comments should never be
made, and any opinions should be offered only with sound indications.
Knowledge of the basic principles of fuels and their ignition and a basic knowledge
of the principles of electricity are the essential tools of the investigator approaching a fire
scene of possible electrical origin. This, of course, includes the majority of instances,
because few structures today lack the convenience of electric energy. Electric energy in
some form must be included in thorough hypothesis testing.

POST-FIRE INDICATORS
Post-fire indicators to be considered during a fire investigation are numerous. In the case
of possible electrical fires, these indicators involve damage to wiring, receptacles and
switches, electronic appliances, conductors, heat melting, and electrical failures, as
discussed next.

Wiring
Tracing an electrical system in a burned structure can be a tedious and time-consuming
task. While any portion of the remaining wiring might yield useful information, the most
fruitful places are where connections are made. This includes the service entrance, load
center box and other fuse and breaker boxes, switches, receptacles, and fixtures.
Particular attention should be given to high-current circuits such as heaters, stoves, and
450 Chapter 10 Electrical Causes of Fires
dryers. When part of the structure is undamaged, an examination of this wiring will tend
to show the general condition of the wiring before the fire.
Wire inside a metal conduit or box will rarely short-circuit except when the insula-
tion is destroyed by external heat (as in Figure 10-32) or animal damage. If the wire does
short, the arcing is generally retained inside, where it cannot cause ignition of thick mate-
rials. Thin paper, lint, or sawdust in the vicinity of the arc has been shown to be ignitable
by incandescent metal sparks created. Of course, if the arcing in high-current circuits
burns through the conduit, it can then cause ignition of nearby combustibles by dropping
molten metal sparks, not by the arcing itself.
The following hazards are most likely to be found in the wiring itself:
■ Damage to wire insulation caused by chafing against sharp surfaces when pulled
through conduit or boxes
■ Conductor insulation deteriorated from age or exposure to heat, moisture, or
chemical vapors
■ Mechanical damage to plastic-insulated cable caused by rodents (with subsequent
contamination), sharp edges, or nails or staples driven through the insulator (not by
crushing externally)
■ Conductors intended for a particular service load being overloaded (via substitution
of a significantly larger-capacity overcurrent device)
■ Connections between conductors not properly soldered or wire-nutted, allowing a
high-resistance connection to create localized heating with subsequent degradation
of plastic components
■ Wiring (sometimes 18-gauge lamp cord) installed for “temporary” use and never
replaced
■ Faults in the ground or neutral side of distribution systems that prompt current deliv-
ered to an appliance or outlet, instead of returning via the intended conductor path,
to return via an alternative path through conduits, flexible metal sheathing, and some-
times even through the plumbing, heating ducts, or wire stucco mesh of the house;53
covers of outlet or junction boxes removed, allowing dirt or combustibles inside
■ Conductors of “open wiring” (knob-and-tube) systems separated from their supports,
in contact with each other or with pipes or ductwork
■ Cables and power cords carrying substantial but “safe” currents being buried under
insulation, carpets, or bedding; coiled in a large mass; or exposed to high ambient
temperature conditions, all of which produce “unsafe conditions” with regard to
insulation degradation

Receptacles and Switches

In explosive environments, the chief ignition hazard of receptacles and switches is the arc-
ing produced when the switching device is operated or the plug is pulled while current is
still flowing (creating a parting arc). Many switches are either so designed or so enclosed
as to allow the arc to occur with minimal damage or effects. Ignition of ordinary com-
bustibles by switches is most likely due to one of the following:
■ Loose contacts between wires and terminals of switch or receptacle, allowing a
high-resistance connection and resulting heat production. These can occur with
push-in connectors, in units exposed to mechanical vibration (see Figure 10-25),
or with aluminum wires on units not designed for them.
■ Overheating due to overload.
■ Poor mechanical condition due to extended use or abuse, allowing contact between
conductors.
■ Enclosures removed or not effective. (This is especially true with the plastic receptacle
and switch boxes in common use today.)
■ Corrosion or erosion of internal conductors, producing high-resistance connections.
Chapter 10 Electrical Causes of Fires 451
 Electronic Appliances
Early electronic equipment contained heat-producing tubes that operated at lightbulb
temperatures and could cause ignition if in contact with combustible materials. Old elec-
tronic equipment, unless properly maintained, can exhibit frayed or worn-out power
cords, improper fuses, and missing cover plates that were intended to retain failures that
could cause a fire. Modern electronic equipment uses solid-state devices that generally
function at lower temperatures, lower currents, and lower internal voltages.
Some manufacturers have historically been lax in providing proper overcurrent pro-
tection. Rectifiers and power transistors have been known to fail and cause power trans-
formers to overheat, thereby causing the insulation to ignite for lack of a proper
protective fuse or breaker. Speaker cones have been known to ignite from excess current
passing through the voice coil because of a defective power transistor lacking fuse protec-
tion, or by being overdriven by a powerful amplifier. Most pieces of household electronic
equipment are considered low-power apparatuses, but situations do occur that have not
been anticipated by the designer.

Conductors
Over the years, some investigators have erroneously accepted the presence of a conduc-
tor with balled or “beaded” ends as a prima facie indication of electrical arcing. Actual
burning tests have demonstrated that balled ends can occur from flame exposure without
the passage of electric current. Therefore, fire investigators should show extreme caution
in reporting the presence of any balled or beaded conductors as symptoms of electrical
arcing. The various mechanisms involved in the melting of conductors will be discussed
in the following sections.

Heat Melting
Wires with ends showing drop-shaped, balled, and pointed formations are frequently the
result of localized heating where some locations along the wire have attained a higher
temperature than others (as shown in Figures 10-45a–e). The melting temperature of pure

(a) (b)

(c) (d)

FIGURES 10-45A–D Melted and offset wire segments resulting from fire exposure. Courtesy of Chris W. Korinek,
P.E., Synergy Technologies LLC, Cedarburg, WI.

452 Chapter 10 Electrical Causes of Fires

 FIGURE 10-45E Microscopic view of the
end of a copper wire melted by localized
fire heating. This shows the bulging and
uniformly oxidized surface of molten and
wire globules. Courtesy of Lamont “Monty” McGill,
Fire/Explosion Investigator.

copper is listed as 1,083°C (1,981°F), which is somewhat above the average temperature
of a typical fire. In other words, the copper wire during a typical structure fire is usually
below its melting temperature. A small localized area of higher temperature such as that
caused by an electrical arc, a concentrated fuel load, or possibly optimal combustion con-
ditions (post-flashover fire with extra ventilation), results in localized melting of the cop-
per wire, as previously characterized.
It has been shown that molten aluminum coming into contact with copper conduc-
tors can erode them at surprisingly low temperatures. Aluminum has a melting point of
660°C (1,220°F) and copper 1,083°C (1,981°F), but an alloy of the two metals, called a
eutectic, has a melting point as low as 548°C (1,081°F). Aluminum melted from connec-
tors, housings, conductors, or even household goods can literally dissolve copper conduc-
tors even without the interaction of electricity.54

Electrical Failures
When energized wires make contact and the contact is not a secure connection, such as
when two bare wires touch without any mechanical pressure, an electric arc is usually
produced. Its magnitude and duration will depend on the overcurrent protection (fuses or
breakers) and the physical conditions under which the wires contact each other. The pas-
sage of current produces extremely high localized temperatures (as seen in Figure 10-46).
If the resulting arc is not confined inside a suitable enclosure, the high temperature of the
arc may ignite very thin or finely divided combustible materials in contact. This situation
is sometimes referred to as primary arcing. Japanese researchers have reported “micro-
deformation” of stranded conductors in short circuits where the high current flows cause
the strands to separate and protrude.55
Concurrent with the general origin and cause investigation, the electrical system
should be examined to aid in the determination of the area of origin. The electrical sys-
tem’s copper wiring often survives the heat of the fire, leaving points of damage (pitting,
beading, or severing) caused by electric arcing as the fire attacks insulation around energized

35,000

30,000
Temperature (K)

25,000

FIGURE 10-46 Graph

20,000 of temperature of arcs in
ambient-pressure air
(experimental data).
15,000 Courtesy of Vytenis
Babrauskus, Fire Science and
10,000 Technology Inc., Issaquah,
1 10 100 1,000 10,000 WA., reprinted from the
Ignition Handbook.
Arc current (A)
Chapter 10 Electrical Causes of Fires 453
FIGURE 10-47 Micros-
copic close-up of wire
showing localized glob-
ules of molten copper
on surface of conductor
caused by electric arc.
Striations of wire drawing
are still visible just to the
left of the globule.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

conductors. The location pattern of the damage (sometimes called anomalies) left on the
electrical system is often more time-stable and less fragile than the fire patterns left by
burning materials. Analysis of this pattern of damage can provide valuable three-dimen-
sional information about the fire’s progression and can be carried out on exposed power
cords as well as fixed wiring as long as they were energized at the time they were exposed
to fire. Arcing can result in either very localized melting on continuous wires, welding the
conductors together; severing of the conductors (and spattering droplets of molten cop-
per) (as illustrated in Figures 10-47 and 10-48); and/or blowing of the fuse or circuit
breaker protecting the conductors.

MAPPING OF ARC FAULTS

If the wiring in the suspected area of fire origin is traced, and the location of arcing is
mapped out on a diagram of the building, there may be an indication of the origin and
development of the fire in the building. As a general principle, the arcing found farthest
from the electrical source can help identify the area of origin. As a fire grows, the circuit

FIGURE 10-48A Bright

copper ball (bead) at tip
of this oxidized wire and
small adhering globules
of molten copper are
consistent with this wire’s
having been exposed to
an electric arc. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

454 Chapter 10 Electrical Causes of Fires

 FIGURE 10-48B Multi-
stranded wire fused
together in limited areas
as a result of arcing
between conductors.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

may be interrupted when the fire attacks the nearest wiring (either by tripping the OCPD
or physically severing the conductors). If the wiring is severed, this means that as the fire
progresses, the more distant portions of the circuits will already be de-energized and show
no signs of arcing. If the OCPD trips, there will be no additional evidence of arcing.
Arc-fault mapping is initially not a part of cause determination but is often useful in
determining a possible area of origin of any fire (structural or vehicular). Figure 10-49a
shows fire attacking an insulated power cord. The source of electric current in the cord is
at the left (plug) end of the wire. The cord is lying on the floor and it feeds a heater at the
right side of the room. If a fire starts on the floor in the center of the room by the exten-
sion cord (fire #1), it will burn through the cord’s insulation, and an arcing fault will
occur as the cord’s insulation chars and degrades. Depending on the resistance of the
degraded insulation, this current flow can range from very low to very high (as high as
150 A has been measured). Because the conductors are not bonded together, they are free
to move, so intermittent current flows can delay tripping of an OCPD. The result of this
arcing fault is beading left on the conductors (as seen in Figure 10-48b). The beads
formed in this case severed both of the conductors but did not blow the circuit breaker.
As fire #1 expands into fire area #2, it attacks the cord “downstream” from the original
point of arcing and produces no further beading, since the electrical supply has been cut off
to this point (by severing a conductor or tripping the OCPD). Subsequent spread of the fire
attack on the cord at point #3, closer to the source, may result in beading of the source, since
the cord has not been de-energized by the circuit breaker. The investigator can then determine
where the fire first encountered the energized extension cord by looking for the point of bead-
ing and severing farthest from its plug, which is the source of its electrical energy.
If the fire started at the heater, there would be a tendency for its line cord to fault and
bead in the area of the heater. Conversely, failure in the power cord’s plug would leave
residue of arcing (beading) in the plug’s components but would leave no beading down-
stream in the line cord or by the heater if the wire was severed. If the fire started in the
building wiring, the investigator would not expect to find faulting inside the extension
cord’s plug or in any of the downstream electrical cords if the fixed wiring was severed.
The process can be described using the water system analogy. This analogy centers on
the use of the puddle pattern of a cut and severed garden hose to demonstrate how the
point of puddling farthest from the source (faucet) is the most likely point where the hose
was first severed. In Figure 10-49b, water will flow from the first cut, leaving a puddle
residue. A subsequent second cut downstream will not leave a puddle, since the water
source has been interrupted at the first cut. Subsequent cuts upstream from the first cut
will leave puddles until the faucet is turned off. If the hose and puddles are examined after
the hose has been damaged, the puddle farthest from the faucet is the point where it was
first cut and severed. The puddles here are analogous to the arc-fault residue (beading)
created in the electrical example.
Chapter 10 Electrical Causes of Fires 455
FIGURE 10-49A Fire
attacking an energized
line cord, resulting in a
pattern of arcing.

Possible arc Arc No arc

1
3 2

Residue of arcing (beading)

FIGURE 10-49B The

water-hose analogy to
3rd cut
Figure 10-49a. 2nd cut
1st cut

No puddle
Possible puddle Puddle

This method of using the arcing damage produced on both single- and three-phase
electrical systems has been taught to over 1,000 investigators at the ATF National
Academy and is based on hundreds of tests on electrical wiring and components in both
laboratory and structural burn situations since the 1980s.56 It has been accepted by the
courts and has also been addressed by NFPA 921, Ch. 17. Arc-fault mapping has been
experimentally verified as reliable in both North American (120-V branch main distribu-
tion) and UK (230-V ring main distribution) systems.57 It has been extensively tested in
exposed surface cable wiring exposed to full-scale compartment fires and found to be
quite reliable in indicating a fire’s area of origin.58

Height of Origin
A “third dimension” of the origin can be determined by examining the vertical and hor-
izontal runs of the electrical wiring. For example (Figure 10-50), the wall receptacle fed
from an in-wall building branch circuit feeds a line cord that has signs of beading and sev-
ering. This indicates that the early fire was in the room and not in the wall, because an
internal wall fire would de-energize the receptacle, so no faulting could occur in the line
cord. Conversely, if beading and severing had been found only in the wiring above the
ceiling, this would indicate an early fire in the attic. If beading and severing had been
found in both the line cord and above the ceiling, then it would indicate that a room fire
first attacked the line cord and then progressed into the attic before the branch circuit
breaker had time to open. Plotting these points on a three-dimensional diagram that
includes the layout of the pertinent sections of the electrical wiring can provide informa-
tion for the investigator to use in eliminating potential fire causes.

Further Considerations of Arc-Fault Interpretation

The investigator should also be aware that the farthest downstream arcing may not
be nearest the origin, because of the effects of wiring placement in relation to other objects
in the fire scene and the nonelectrical damage to the wiring caused by the fire and other
activities. Some factors to be considered are uneven shielding of the wiring from the fire by
456 Chapter 10 Electrical Causes of Fires
 Branch circuit Ceiling FIGURE 10-50 Plotting
source Attic Ceiling the pattern of arc faulting
(involved later) in three dimensions.

Arc bead
In-wall
electrical
wires

Floor

Arc bead

Appliance Fault points

drywall, types of electrical insulation, available fault current, the size of the OCPDs, and
the number and location of sources of electric energy. In larger or multistory structures
where there has been significant collapse or remodel, it is often impossible to accurately
trace each conductor back to its source, making arc-fault mapping of minimal value. An
electrician with experience in wiring can be of great help in considering these factors. This
method may also be fuel load–dependent. For example, a very small fire starting in one
corner of a room may not have enough energy to melt the insulation on overhead conduc-
tors, but a more massive fuel load in the other corner of the room may cause faulting in
the conductors in its corner. In the case of a building collapse or explosion, the physical
movement or fracturing of structural members, electrical conduit, or other electrical enclo-
sures can expose the conductors inside to arcing due to contact or heating faster than in
areas where the conductors are better protected. Such events can then result in arc damage
even if they are downstream of more protected areas closer to the area of origin.
Some of the points of electrical faulting may be modified by the subsequent melting
and chemical effects of some fires. The investigator should take this into account during
the examination. As a general guide, there must be a clear demarcation area between the
localized melting (purported arc) and the undamaged area of the conductor. Courtroom
battles have been centered on whether an alleged point of beading is really only thermal
melting due to the effects of the fire. Experienced metallurgists may be called upon to pro-
vide opinions on the source of a bead. As we shall see, the discrimination can be very dif-
ficult. The impact of this technique is on location of first major fire, not necessarily on
whether the fire was caused by an electrical fault.
The lower melting temperature of aluminum wiring limits its use in this method, since
it will undergo extensive gross melting in large fires. However, if gross melting is accounted
for, this method of pattern analysis may be useful. Finneran has recently published a thor-
ough description of the approach and its application.59
The origin(s) determined by examination of the electrical system should correlate
with the origin(s) determined by the other standard investigative methods. It is important
that the electrical investigator and the origin and cause investigator reach a consensus as
to the point(s) of origin before the cause is conclusively established. The possibility of an
electrical cause should then be investigated as part of the continuing examination.

ARCING THROUGH CHAR

Very frequently, signs of electric arcing are not indicative of electrical ignition and are
instead a consequence of the fire. Because most wire insulation materials are com-
bustible, arcing is often the result of current transmitted through charred insulation or
contact between wires as the insulation was burned away completely. When the condi-
tion of the wire indicates arcing has taken place, the arcing could actually have been the
Chapter 10 Electrical Causes of Fires 457
 result of fire-caused insulation degradation. [Such fire-caused faulting (arcing through
char) is sometimes called secondary arcing.] This arcing may be between conductors or
between an energized conductor and a grounded conduit, box, or appliance housing. The
arcing can progress up the wire (toward the power source in single-phase systems) for
some distance, producing sustained exterior effects and showers of molten metal, as in
Figure 10-32. Such arcing through char can occur and be sustained even in circuits pro-
tected by fuses or breakers, since the resistance of the char “connection” may limit the
current so that an overcurrent is not produced. The conductors are also free to move
away from each other (unlike a “bolted” fault). This means that the arc may be struck
and extinguished numerous times in time frames too short for an OCPD to be triggered.60
Microscopic examination of questioned wires/conductors may provide useful informa-
tion. The conditions of melting from external fire heat are different from those of electric
arcing. When wires melt from external fire heat, due to high thermal conductivity, there is
usually only a small temperature difference between the parts that have melted and the
unmelted sections. In this situation, the wire section has been raised to the fire tempera-
ture, and only a few additional degrees (from an area of optimized heating) are needed to
produce the melt and its balled or pointed symptoms (see Figure 10-45). By contrast, when
wires are subjected to arcing (whether that arcing is the cause of the fire or the result of it),
the wire itself is at ambient temperature while the arc is at many thousands of degrees. The
result is often an abrupt change from undamaged wire to a melted globule, as was shown
in Figures 10-47 and 10-48. There may be minute beads of melted copper spattered onto
nearby surfaces that will not be present when the melting is caused by thermal damage
alone with no current flow. The two different situations are usually discernible by exami-
nation. Visual inspection is usually sufficient to detect arcing effects, but running the wire
through a wad of cotton wool or even between fingertips may be more effective in detect-
ing very small defects (see next section). Of course, there are gray areas where both situa-
tions occurred and also where high temperatures have left the wire in such poor condition
that the indications have been minimized or destroyed. The difficult task is to assess what
arcing occurred prior to the fire and thereby created an ignition source. Remember that it
is not enough just to have an ignition source; there must also be fuel in a suitable physical
state in contact with that source for a sufficient length of time to have ignition. The inves-
tigator must continuously ask, “Do I have all the elements here, or am I blaming electric-
ity just because there is electrical wiring in the vicinity?”

LABORATORY EXAMINATION
As described more fully in Chapter 14, laboratory examination can assist the investigation
in a number of ways. As discussed next, laboratory examination of wiring, insulation,
switches, fuses, circuit breakers, and appliances can particularly assist in the investigation
of fires of possible electrical origin.

Laboratory Tools and Procedures

In many cases, the diagnostic signs of electrical fires are so subtle that they require labo-
ratory examination. Changes occur as copper wire is fire heated. First, oxidation pro-
duces discoloration. As the melting temperature is reached, blistering and bubbling may
occur from degassing. During melting, the striations caused by the manufacturing process
flow together, and there is a general flow of the molten copper, resulting in balling and
areas of erosion with a rough surface. When the wires are oriented in a vertical position,
the molten copper can run down to a ball and then drop off, leaving a pointed end. When
the ball remains, the portion immediately above it shows a reduced diameter from
where the metal flowed. There is no sharp line of demarcation showing points of melting
and nonmelting. Stranded wires melt and flow together and form a solid mass. Sometimes
the strands are discernible but have been fused together and are not separable. Sometimes
several wires inside a common conduit will fuse together. These symptoms will be seen
458 Chapter 10 Electrical Causes of Fires
whenever wire has reached its melting temperature. Usually, fusing is localized, because
the melting temperature of copper wire is above average fire temperatures, so sufficient
generalized heating is uncommon.
Electric arcing, by contrast, does not usually result in random melting, erosion,
generalized fusion of strands, or other symptoms indicative of flame-exposure heating.
The characteristic of electric arcing is the formation of a sharp demarcation line between
the arcing area and the unmelted wire, as was shown in Figures 10-47 and 10-48. In arc-
ing situations, stranded wire sometimes exhibits individual strands that have melted
together only at the point where the arc occurred, as was shown in Figure 10-48a.
Elemental analysis of such spots by scanning electron microscopy/energy dispersive
X-ray spectrometry (SEM/EDS) may be of value in establishing contact between conduc-
tors of different metals (copper wire against a zinc-plated steel receptacle box, for
instance)61 alloying with aluminum (see Figure 10-51). (See Chapter 14 for descriptions
of SEM/EDS technology and its applications to fire investigation. SEM alone can
improve the quality of visual examination and allow examination of very fine details,
such as the “draw marks” on conductors that disappear when the conductor approaches
its melting point.)
SEM/EDS may help confirm a particular accidental ignition mechanism. The presence
of metallic contaminants on insulators in switches and relays has been linked to arc track-
ing and electrical equipment fires. SEM/EDS can help identify the nature and origin of
such contamination.
Metallography is the study of a cross section of a metal showing details such
as grains, impurities, voids, and other important details which are important in the
material’s history which may help to show if the material was involved with a fire
cause or was a victim of the fire. Some examples of metallographic features are shown
in Figure 10-51.

FIGURE 10-51
Comparison of a copper
wire (heated) alloyed with
aluminum (on left) and a
copper wire exposed to
an electric arc (on right).
Top: Color photos.
Middle: False-color
SEM/EDS elemental maps.
On left, green represents
copper and red represents
aluminum. On right,
green represents
copper and red is oxygen.
Bottom: Metallographic
elemental maps. On
left, you can see that
aluminum (red) has pene-
trated into the copper as
well as on the surface. On
the right, the transferred
copper is on the surface
with a slight amount of
copper oxide due to the
arc’s heat.Courtesy of
Chris W. Korinek, P.E.,
Synergy Technologies LLC,
Cedarburg, WI.

Chapter 10 Electrical Causes of Fires 459

 It has been suggested that the identification of compounds trapped within the cop-
per as it melts and reacts with its environment may permit one to conclusively establish
whether the copper failed in an atmosphere of oxygen (open air) or chlorine (from vinyl
chloride insulation). In fact, double-blind tests of the Auger electron spectroscopy (AES)
method have concluded that the technique does not discriminate between arc beads cre-
ated during fire exposure and those that initiated a fire.62 In his very thorough review of
the AES analysis issue, Babrauskas concluded that such discrimination is unlikely, since
the two processes—electrical and fire—do not produce categorically different residues.63
Because most post-fire indicators are the summation of the fire’s effects, and fire envi-
ronments are so complex, it would appear that none of the metallurgical or instrumen-
tal techniques examined to date can reliably discriminate between copper arc bead
residues created as part of the ignition process and those created later in the fire by ther-
mal exposure.
Wiring and Enclosures
Conduction heating (due to excessive current flow for the size of the conductor) heats
the full length of the wire all the way back to the junction box or to the supply source.
Extreme overheating can cause the wires to reach fusion temperature, and when the
first part of the wire melts away, a parting arc can form. If the insulation is destroyed
first, the carbonized residue may provide a conduction path, or the bare wires them-
selves may touch and arc. When these things occur inside metal boxes or conduit, the
failures are usually trapped and of little concern. However, if the failure has sufficient
energy to burn a hole in the metal enclosure, a fire can result (such penetrations are rare
with protected 120/240-V circuits). Figure 10-52 shows the failure of a meter service
connection that came loose and grounded against the meter base. Frequently, the exam-
ination of the wiring in an unburned area can provide useful information, as it can indi-
cate the age and general condition of the wiring. If it has been poorly maintained and
the insulation has become brittle and cracked, electrical problems should be considered
more likely.
If a section of undamaged wire between the suspected area and the supply can be
found, the condition of the conductor can provide some useful information. If the insula-

FIGURE 10-52 Arcing

through the meter case
caused when power lead
came loose and shorted
to inside of case. Courtesy
of Greg Lampkin, Knox County,
TN, Fire Investigation Unit.

460 Chapter 10 Electrical Causes of Fires

FIGURE 10-53A The underside of a steel screw head after a FIGURE 10-53B Comparison of wire and screw from normal
glowing connection was lab created on a duplex outlet with a solid (tight) connection to oxidized and pitted screw and wire from loose
#12 AWG wire. Note the melting and discoloration at the bead at (high resistance) connection. Courtesy of Chris W. Korinek, P.E., Synergy
the rim of the screw head. Courtesy of Chris W. Korinek, P.E., Synergy Technologies LLC, Cedarburg, WI.
Technologies LLC, Cedarburg, WI.

tion and wire are in a “like-new” condition with no tarnishing of the conductor, it is
unlikely that overcurrent occurred to the degree that ignition took place. Electric arcing,
or arcing through char, however, with its local heating effects, would still have been pos-
sible under these conditions. Although failures in mid-run have occurred where the con-
ductor itself was not continuous (causing series arcing), by far the most common failures
occur at connections.

Connector Screws
Connector screws can show extensive effects of heat, including arc transfers on the edge of
the screw, as shown in Figures 10-53a and b. Prolonged heating in a poor connection con-
dition can cause heating of the entire length of the screw. When a screw is exposed to exter-
nal heating, the portion in the connector is protected unless fire exposure is very prolonged.
Similar localized heating can occur in any connector improperly used—wire nuts, twisted
wire ends, “push-in” connectors in receptacles, or stab connectors in plugs. Figure 10-54
shows the heating and erosion of a loose “split-bolt” connector that ignited the adjacent
timber.

Insulation
Very old wire was usually insulated with rubber and/or cloth material. Modern insulation
for wires is a thermoplastic (polyethylene or PVC being the most common material).
Insulation condition can sometimes be used as a guideline to indicate whether the heat
was produced internally (overcurrent) or externally (by the fire). Rubber is destroyed by
heat. When overcurrent is passed through rubber-covered wire, the internal heat can
destroy the rubber nearest the wire, leaving the exterior nearly intact. Afterward, the insu-
lation may remain loose and free to slide back and forth over the wire (sleeving effect).
The situation is different with the plastic insulation that is more common today. On an
overheated wire, the insulation softens and melts when heated and then rehardens when
Chapter 10 Electrical Causes of Fires 461
 FIGURE 10-54 Close-up of split bolt connector showing erosion and arcing to clamp. Courtesy of Chris J. Bloom,
CJB Fire Consultant, White City, OR.

cooled. Due to its thermal inertia, plastic will discolor or pyrolyze on the surface exposed
to the most heat, so this may reveal whether the insulation was heated from the inside
(heat from the wire) or from the outside (from an exterior fire). However, any heating is
generally accompanied by deformation, sagging, and discoloration of the insulation.64
Sometimes bubbling also occurs from gas formation as the plastic decomposes from the
heat. Bubbling is not as noticeable in rubber insulation because of its porosity, which min-
imizes the entrapment of gases.
When there is doubt, the laboratory can test a section of undamaged wire by passing
a current through it and observing its effects on that particular insulation. Remember,
heating from the wire is not necessarily electrical. Heating can result from thermal con-
duction in a wire exposed to a direct flame some distance away.

Examination of Switches, Fuses, Breakers, and Appliances

The laboratory examination of switches, fuses, breakers, and appliances will be dis-
cussed in Chapter 14. X-ray examinations are particularly valuable, since they can
reveal internal details without destructive examination.65 Even modest heating can
melt, scorch, or char an electric appliance beyond visual examination (see Figures 10-55a
and b). Sometimes the laboratory analysis will address issues other than electrical fail-
ure. Interviews with appliance users can sometimes save a lot of investigative effort
by suggesting a particular testable hypothesis. The shop vacuum in Figure 10-56 was
the cause of the fire, but it was unplugged at the time of the fire. It had been used 5
hours previously to clean up a quantity of dust from a floor-sanding project.
Laboratory analysis confirmed that the floor finish on the wood dust was susceptible
to self-heating.
462 Chapter 10 Electrical Causes of Fires
FIGURE 10-55A Burned smoke detector, which is difficult to FIGURE 10-55B X-ray shows battery connections in place.
evaluate nondestructively. Courtesy of John D. DeHaan. Courtesy of Dr. David Icove, University of Tennessee.

FIGURE 10-56 This

shop vacuum was the
cause of a fire in a utility
closet, but close examina-
tion showed it was not
plugged in at the time of
the fire. Lab analyses
confirmed that the bag
(completely destroyed)
contained dust from a
floor-sanding operation.
The floor finish/sanding
dust self-heated to igni-
tion. Courtesy of Investigator
Jeff Weber, San Jose, CA, Fire
Dept.

Chapter 10 Electrical Causes of Fires 463

 CHAPTER REVIEW

Summary
If wires, appliances, and other electrical equipment have ■ Open circuits causing fires. There must be a flow
not been excessively damaged by a fire, a considerable of current for ignition. Voltage alone on wiring
amount of information can be gleaned from them by (energized) will not cause ignition—there must be
careful examination in accordance with the principles a current flow (amperes). This flow can take many
outlined previously. The evaluation of electrical fire forms—short circuit, overload, leakage current,
causes includes establishing the means by which electric insulation degradation, and so on. Energized
current could have produced the necessary heat—heat conductors can cause electrical effects during the
by design, heat by failure or misuse, or heat from unin- fire, however.
tended current flow. The analysis must also include the ■ Arcing—cause or result. Did the arcing cause the
amount of heat produced, the rate at which it was pro- fire, or was the arcing the result of wires that
duced, and the area over which it was produced. The touched because their insulation was burned off
effect of heat on the first fuel and the mechanism of by the fire? Could an arc ignite nearby materials?
actual ignition also need to be evaluated. As in other Was there fire exposure that could cause arcing
fields, over the years fire investigators have used certain through char later in the fire?
criteria for judging the evidence of electrical fire symp- ■ Energy. Was there sufficient electric energy pres-
toms. Unfortunately, some are not as reliable or as invi- ent to cause ignition of the surrounding materi-
olate as once believed. These include the following: als? The mere presence of some electric energy is
not sufficient proof that it could have ignited the
■ Ignition due to slightly oversized fuse or breaker. fuels in question. Electrical events must be included
When a slightly higher rated fuse or breaker is in a complete scheme of hypothesis testing for
found than the rated capacity of the wire, it nearly all fire investigations because they indicate
probably is not the fire cause. Current-carrying there was some energy available. The critical test
capacities are conservatively rated. remains: Could this source have caused ignition
■ Balled wire ends. This condition can be the result of that fuel under the conditions present?
of fire melting as well as from electrical arcing, ■ Remember that, as with other forms of physical
neither of which may have anything to do with evidence, it is possible that there was sufficient
the fire’s ignition. fire damage to preclude conclusive analysis. When
■ Loose wire insulation (sleeving). If the insulation even a thorough and careful laboratory examina-
is loose, it probably is from overcurrent. If it is tion fails to reveal the cause or mechanism of the
not loose, however, overcurrent should not be fire, it should not be viewed as motive to invent
ruled out. The particular type of insulation should an electrical origin.
be tested to determine its thermal characteristics.

Review Questions
1. Describe Ohm’s law and state its basic formula. 7. Name three fuel types ignitable by discharge of
2. How can a poor connection cause a fire? static electricity.
3. How does a fuse function as an overcurrent 8. Can arc beads on copper wire that resulted from
protection device? fire exposure be distinguished from those that
4. How can electric current cause a fire when the were part of the cause of the fire?
circuit is protected by an OCPD? 9. How does aluminum wiring differ from copper
5. What is arc tracking and how does it contribute wiring in its conductance, mechanical properties,
to fire ignition? and melting point?
6. How does arc mapping aid in determination of 10. Give three examples of how electric motors cause
area of origin? fires.
464 Chapter 10 Electrical Causes of Fires
References
1. J. R. Hall, “Home Electrical Fires” (Quincy, MA: NFPA 16. CPSC, “Corrosion in Homes and Connections to
Fire Analysis and Research Division, March 2009), 5. Chinese Drywall,” press statement (Washington DC:
2. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire CPSC, November 23, 2009); see also “CPSC Alert to
Science Publishers, 2003), 548. Fire Safety Professionals,” www.cpsc.gov/info/drywall,
3. V. Babrauskas, “Ignition Events during Fueling of April 2010.
Motor Vehicles at Filling Stations,” California Fire and 17. Babrauskas, Ignition Handbook, 312–15, 333–34, 734,
Arson Investigator (April 2005): 16; J. L. Pharr and T. 774–77, 804–6; Hagimoto et al., “Short-Circuit Faults.”
T. Jonas, “Static Electricity Generation and Ignition at 18. Hagimoto et al., “Short-Circuit Faults”; Y. Hagimoto,
Fuel Delivery Systems,” North Carolina IAAI “Fire Causes in Electrical Plugs” in Proceedings
Newsletter, 2007; (www.nciaai.com/articles). Interflam 2001 (London: Interscience Communications,
4. R. C. Lee, E. G. Cravalho, and J. F. Burke, Electrical 2001); J. Zicherman, “The Fire Performance of
Trauma: The Pathophysiology, Manifestations and Electrical Insulating Materials Used in Fluorescent Light
Clinical Management (Cambridge: Cambridge Fixtures,” California Fire and Arson Investigator (July
University Press, 1992), 33–36; J. Webster, Medical 2004).
Instrumentation, Application and Design, 3rd ed. 19. Babrauskas, Ignition Handbook.
(New York: Wiley, 1998), 183–86; D. Fink and W. 20. V. Babrauskas, “What Voltage Is Required to Cause
Beaty, Standard Handbook for Electrical Engineers, Arcs and Sparks?” Fire Findings 12, no. 1 (Winter
13th ed. (New York: McGraw-Hill, 1993), chap. 11, 2004): 1–5.
p. 59; T. L. Brown and E. H. LeMay Jr., Chemistry: 21. D. Miller, R. Chowdury, and M. Greene, “2004–2006
The Central Science (Upper Saddle River, NJ: Prentice- Residential Fire Loss Estimates,” (Washington, DC:
Hall, 1977), 356. Consumer Product Safety Commission, October 2009).
5. V. Babrauskas, Ignition Handbook, 540–48; International 22. Ibid.
Institute of Welding, The Physics of Welding, ed. by J. F. 23. Ibid.
Lancaster (Oxford: Pergamon, 1984), 109–20. 24. Ibid.
6. Fink and Beaty, Standard Handbook for Electrical 25. Babrauskas, Ignition Handbook, 732; Consumer
Engineers, 4-74 – 4-79. Product Safety Commission, “Recall,” press release
7. J. Pharr, “Surge Suppressor Fires,” ESD Journal #07-094 (Washington, DC: CPSC, February 1, 2007).
(Moore, SC: Fowler Associates), www.esdjournal.com. 26. B. Harwood, Fire Incidence in Television Receivers
8. A. Chiste and D. Kladar, “Electronic Systems Protection (Washington, DC: CPSC, 1979).
via Advanced Surge Protection Devices,” Institute of 27. Miller, et al., 5.
Electrical and Electronic Engineers (IEEE) Industry 28. J. Flynn, “Lamp, Bulb or Lighting Fires in
Applications Conference, IEEE 13 (2002): 2244–47. Nonresidential Structures: 2003–2006 Annual
9. B. VanderPas and K. Lewis, “Corroded GFCIs Create Estimates” (Quincy, MA: NFPA Fire Analysis and
Fire Hazard,” Fire and Arson Investigator (January Research Division, January, 2010).
1999): 10–13. 29. Hall, “Home Electrical Fires,” 67.
10. G. Gregory, “AFCIs Target Residential Electrical 30. D. Powell, “Fluorescent Lighting,” Fire Findings 13,
Fires,” NFPA Journal (March–April 2000): 69–71. no. 3 (Summer 2005), and 13, no. 4 (Fall 2005): 8.
11. A. Manche and A. Haun, “Why AFCIs?” NFPA 31. NFPA 70, National Electric Code (Quincy, MA:
Journal (September–October 2008): 74–77. National Fire Protection Association, 2008).
12. V. Babrauskas, “How Do Electrical Wiring Faults Lead 32. Powell, “Fluorescent Lighting,” 7–8.
to Structure Ignitions?” in Proceedings Fire and 33. Flynn, “Lamp, Bulb or Lighting Fires.”
Materials 2001 (London: Interscience 34. Zicherman, “The Fire Performance of Electrical
Communications, 2001), 39–51. Insulating Materials.”
13. N. P. Dwyer, “Can a High Resistance Connection Lead 35. J. Zicherman, “Technical Note: The Fire Performance
to a Fire Even after the Load is Removed?” Fire of Electrical Insulating Materials Used in Fluorescent
Findings 16, no. 3 (Summer 2008): 12–13. Lighting Fixtures,” Fire and Materials 25 (2001):
14. Y. Hagimoto, K. Okamoto, and N. Watanabe, “Short- 209–31.
Circuit Faults in Electrical Cables and Cords Exposed to 36. CPSC, “Technical Consumer Products Inc. Recall of
Radiant Heat” in Proceedings Fire and Materials 2003 Fluorescent Light Bulbs” (Washington, DC: U.S.
(London: Interscience Communications, 2003), 215–26. Consumer Product Safety Commission, December
15. K. Okamoto, N. Watanabe, and Y. Hagimoto, “Fire 2004).
Hazard Caused by Thermal Degradation of Organic 37. N. P. Dwyer, “Compact Fluorescent Light Bulbs,” Fire
Insulating Materials at Plug and Receptacle Findings 16, no. 2 (Spring 2008): 4.
Connections” in Proceedings Fire and Materials 2003 38. Osram Sylvania, Lamp and Ballast Catalog, 2008,
(London: Interscience Communications, 2003), 203–14. fn. 74 (Danvers, MA: Osram Sylvania), 163.

Chapter 10 Electrical Causes of Fires 465

39. David Powell, personal communication, November (1980): 252–56; B. Béland, “Comments on Fire
2009. Investigation Procedures,” Journal of Forensic Sciences
40. Babrauskas, Ignition Handbook, 1075. 29, no. 1 (1984): 190–97; R. J. Svare, “Determining Fire
41. P. A. Lynch and C. Bloom, “Slippery Issues on Point-of-Origin and Progression by Examination of
Aquarium Fire Losses,” California Fire/Arson Damage in the Single Phase, Alternating Current
Investigator (July 2006): 4–10. Electrical System,” invited paper, Ministry of Public
42. NEMA, LSD 25-2008, Best Practices for Metal Halide Security, Beijing, People’s Republic of China. Presented
Lighting Systems, Plus “Q&A about Lamp Ruptures in Shanghai, PRC, April 1988, Journal of People to
in Metal Halide Lighting Systems,” (Rosslyn, VA: People, International Arson Investigation Delegation to
National Electrical Manufacturers Association, March the People’s Republic of China and Hong Kong, 1988;
10, 2009), 2; NEMA; David Powell, personal commu- M. Deplace and E. Vos, “Electric Short Circuits Help the
nication, November 2009. Investigator Determine Where the Fire Started,” Fire
43. Babrauskas, Ignition Handbook, 750. Technology, 19, no. 3 (1983): 185–91; B. Béland,
44. Miller et al., “2004–2006 Residential Fire Loss “Copper Behavior under Fire Conditions,” Fire and
Estimates”; see also Hall, “Home Electrical Fires,” 14. Arson Investigator (June 1994): 40–43; B. V. Ettling,
45. R. E. Lowe and J. A. Lowe, “Halogen Lamps II,” Fire “Arc Marks and Gouges in Wires and Heating at
and Arson Investigator (April 2001): 39 – 42. Gouges,” Fire and Arson Investigator 20, no. 4 (1996):
46. W. Hull and D. L. Pippen, “Electric Blanket Fires,” 19–27; J. M. Finneran, “Arc Mapping,” Fire Findings
Part I: Fire Findings 11, no. 1 (Winter 2003), 7–11; 13, no. 2 (Spring 2005): 7–11.
Part II: Fire Findings 11, no. 2 (Spring 2003): 1–3. 57. N. Carey, “Powerful Techniques: Arc Fault Mapping,”
47. “Heat Tape and Heat Cable Are Not the Same,” Fire Fire Prevention (March 2002): 47–50.
Findings 16, no. 4 (Fall 2008): 5. 58. N. J. Carey, Developing a Reliable Systematic Analysis
48. B. Béland, “Heat Tape Fires,” Fire and Arson for Arc Fault Mapping, PhD dissertation, University of
Investigator, September 1994. Strathclyde, September 2009; W. Johnson and L. Rich,
49. Brown and LeMay, Chemistry, 356. “Arc Mapping Follow-Up,” Fire Findings 15, no. 3
50. K. Rosenhan, “Acid Tests, Fire Risk Management (Summer 2007): 7–10. Also see Johnson and Rich,
(October 2009): 48–51; (also Fire & Arson Investigator Proceedings: Fire and Materials (London: Interscience
(April 2009); V. Babrauskas, Ignition Handbook, 742. Communications, January 2007).
51. J. D. Twibell, “Reporting of Electrical Fires Hides True 59. Finneran, “Arc Mapping.”
Picture,” Fire Prevention 282 (September 1995): 11. 60. Carey, “Developing a Reliable Systematic Analysis for
52. B. V. Ettling, “The Overdriven Staple as a Fire Cause,” Arc Fault Mapping.”
Fire and Arson Investigator (March 1994). 61. Ibid.
53. L. Weintraub, “Open Electrical Neutral and Pyrophoric 62. D. G. Howitt, “The Chemical Composition of Copper
Carbon Phenomenon,” Fire and Arson Investigator Arc Beads: A Red Herring for the Fire Investigator,” Fire
(December 1993). and Arson Investigator (March 1998): 34–39; R.
54. B. Béland, C. Roy, and M. Tremblay, “Copper-Aluminum Henderson, C. Manning, and S. Barnhill, “Questions
Interaction in Fire Environments,” Fire Technology 10 Concerning the Use of Carbon Content to Identify
(February 1983). ‘Cause’ vs. ‘Result’ Beads in Fire Investigations,” Fire
55. S. Miyoshi, “Micro-Deformation of Stranded Electrical and Arson Investigator (March 1998): 26–27; B. Béland,
Conductors Caused by Short Circuits” paper presented “Examination of Arc Beads by Auger Spectroscopy,”
at IAFS (International Association of Forensic Fire and Arson Investigator (April 2004): 41– 42.
Sciences) 2008, New Orleans, LA, July 2008. 63. V. Babrauskas, “Fires Due to Electrical Arcing: Can
56. L. I. Rothschild, “Some Fundamental Electrical Concepts ‘Cause’ Beads Be Distinguished from ‘Victim’ Beads by
in Locating the Cause and Origin of a Fire,” Journal of Physical or Chemical Testing?” in Proceedings Fire and
National Academy of Forensic Engineers (NAFE) 3, no. Materials 2003, (London: Interscience Communications,
2 (December 1986): 37–44; R. H. Kaufman and J. C. 2003), 189–201.
Page, “Arcing Fault Protection for the Low-Voltage 64. B. Béland and J. Tremblay, “On the Sleeving of Electrical
Power Distribution Systems: Nature of the Problem,” Insulation,” Fire and Arson Investigator 43, no. 4 (June
Institute of Electrical and Electronics Engineers (IEEE) 1993).
Transactions on Power and Apparatus Society (1960): 65. J. D. Twibell and S. C. Lomas, “The Examination of
160–67; J. R. Dunki-Jacobs, “The Escalating Arcing Fire-Damaged Electrical Switches,” Science and Justice
Ground-Fault Phenomenon,” IEEE Transactions on 35, no. 2 (April 1995): 113–16.
Industry Applications Society 1A-22, no. 2 (1986): 66. J. D. Twibell and C. C. Christie, “The Forensic
1156–61; B. Béland, “Examination of Electrical Examination of Fuses,” Science and Justice 35, no. 2
Conductors Following a Fire,” Fire Technology 16, no. 4 (April 1995): 141– 49.

466 Chapter 10 Electrical Causes of Fires

 CHAPTER
11
Clothing and Fabric Fires

Courtesy of John Jerome and Jim Albers.

KEY TERMS

elastomers, p 470 natural fibers, p. 469 synthetic fibers, p. 469

furniture calorimeter, p. 478

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify the common causes for fires in fabrics, upholstered furniture, and mattresses.
■ Identify the common types of cloth.
■ Explain the basic concepts of the Flammable Fabrics Act.
■ Explain how furniture testing is conducted.
■ List and explain common tests used to measure cigarette ignition, flame ignition,
heat release rate, smoke, and toxicity.
■ List and separate describe several flammability tests (examples: clothing, textiles, plastic
films, carpets and rugs, mattresses and pads, children’s sleepwear, upholstered furniture).
■ Search the Internet for current and proposed governmental policies and legislation.

For additional review materials, appendices, and suggested readings,

visit www.bradybooks.com and follow the MyBradyKit link to register
for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

467
 A
lthough the total number or residential fires in the United States continues to
decrease each year, fires in which some fabric or fabric-covered household material
was the first fuel to be ignited accounted for 14.3 percent of all residential fires in
the United States for the period 2004 – 2006 but were involved in approximately 40 percent
of all residential fire deaths. This means that upholstered furniture, bedding, mattresses,
garments, and draperies accounted for an estimated 26,900 residential fires (out of 384,100
total) and 1,030 deaths each year (2004–2006).1
The NFPA estimates that upholstered furniture alone was the item first ignited in
an average of 7,630 home structure fires per year, causing an annual average of 600 civilian
fire deaths, 920 civilian fire injuries, and $309 million in direct property damage (2002–2005).2
Fires beginning with upholstered furniture accounted for only 2 percent of reported
home fires but 21 per cent of home fire deaths. During the same time frame, a mattress or
bedding was the first fuel ignited in an average of 11,520 home fires. Such fires accounted
for 3 percent of the fires but 13 percent of home fire deaths (378 deaths, 1,286 civilian
injuries, and $357 million in direct property damage).3 According to the CPSC, ignition of
upholstered furniture was caused by cigarettes or other smoking materials in an estimated
29 percent of home fires and by open flames in about 16 per cent. Ignition of mattresses
and bedding was caused in about 22 percent of home fires by smoking materials and
in 26 percent by open flames.4 The NFPA estimated (in 2008) that 14 percent of mattress/
bedding home fires and 12 percent of upholstered furniture fires were ignited by smoking
materials.5
Although garment fires may not frequently go on to cause structural damage, the NFPA
estimated that in the years 1999–2000, 120 people died and 150 people were injured (per
year) because their clothing caught fire.6 A more recent study by Hoebel et al. used burn
injury data from the National Electronic Injury Surveillance System (NEISS) and death data
from the National Center for Health Statistics (NCHS) of the Centers for Disease Control and
Prevention (CDC). This study revealed that a great many clothing fires are not reported to
the local fire departments (and therefore do not appear in the NFIRS or NFPA data). This
analysis revealed that nearly 4,300 injuries and 120 deaths per year were caused (1997–2006)
when peoples’ clothes caught fire.7
The overall number of upholstered furniture, mattress, and bedding fires in homes has
decreased dramatically since the 1980s as fire-retardant materials have become more com-
monly used in some furnishings and as public awareness of the hazards has improved.8 In
spite of this progress, fires in which fabrics are the first materials to be ignited are still a very
common occurrence. The fire investigator should be familiar with the nature of common
fabrics and upholstery materials and their contributions to both ignition hazard and fuel
load in today’s studies.

Types of Fabric
Fabrics differ greatly in their properties of combustion, ranging from some that are intrin-
sically very hazardous to others that cannot be burned at all. A few extremely flammable
fabrics are prohibited in the United States by the 1953 Flammable Fabrics Act (16 CFR
1610), however, fabrics that “pass” may ignite and burn easily. Fabrics that are the most
fire resistant are, in general, used only for special purposes such as fire-resistant clothing
for pilots, firefighters, and drivers of racing cars. The more common types of fabrics made
from natural and synthetic fibers and their properties are outlined next.
468 Chapter 11 Clothing and Fabric Fires
NATURAL FIBERS
Common natural fibers include wool, cotton, linen, kapok, and silk. natural fibers ■
Fibers of animal or veg-
Wool etable origin without
Wool is used in many fabrics utilized for clothing as well as for bedding, upholstery, and chemical manipulation.
miscellaneous purposes. Its fire hazard is minimal because of its high nitrogen and pro-
tein content. It can be burned but with considerable difficulty, because it has a high igni-
tion temperature and low heat of combustion. It will not normally sustain a flame. Thus,
wool blankets are used to wrap a person whose clothing is on fire and is being tested as
a barrier layer in furniture. Exposing wool cloth to a flame will char a hole in it, but it
will self-extinguish when the outside heat source is removed. When it does burn, wool
generates toxic hydrogen cyanide fumes in addition to other combustion products.
Cotton
Cotton is used extensively in cloth for garments, bedding, and upholstery. It is possibly
the most combustible of the common natural fibers. It is cellulosic in composition, like
wood and paper. Having a very large surface-to-volume ratio, that is, a large surface for
combustion, cotton is ignited with relative ease and will sustain a fire or allow it to increase
greatly after ignition as long as it receives adequate ventilation. Cotton’s ease of burning
is conditioned by several factors, which will be discussed later. Cotton batting was once
used universally as padding in garments and furniture and also as filler in cushions. It may
still be found in some modern furniture, but it has largely been supplanted by polyester
fiberfill or other synthetics.

Linen
Linen is being used much less than formerly, especially in clothing, and only occasionally
is it seriously involved in major fires. Its cellulosic fibers are derived from stems (bast)
rather than seed pods, as is cotton. Because it is a bast fiber, linen is much more variable
than cotton, and this difference is expected to have some influence on combustion rates.
Cloth from predominantly fine fibers is expected to burn more freely than that from
coarser fibers. Like cotton, linen can be ignited by a smoldering cigarette or contact with
a hot surface to support a smoldering or flaming fire.

Kapok
Kapok has been used as a filling for pillows, bedding, and insulated clothing but is not
normally used as a cloth fiber. Kapok is involved in fires with sufficient frequency to
make its fire hazards important. Like cotton, kapok is a cellulosic fiber, but because of
its cell structure and its fineness, it exceeds cotton significantly in its flammability. It
will ignite with ease, and when ventilation is available, kapok will burn with great
intensity. It will also support smoldering fire very well. It is fortunate that kapok’s uses
are limited, so that it is not often involved in clothing fires. It can, however, pose a sig-
nificant hazard in industrial areas, such as life preserver and pillow manufacture, where
it is utilized. California’s strict upholstered furniture flammability laws do not permit
the use of kapok.

Silk
Silk is made by unwinding the cocoon of the silkworm and spinning the filament obtained
into thread. It is a proteinaceous material like wool and although not easily ignited in
bulk, it is often woven into a thin or sheer fabric that can burn quickly once ignited. It
generates toxic gases, including hydrogen cyanide, upon combustion.

PETROLEUM-BASED SYNTHETIC FIBERS

synthetic fibers ■
Synthetic fibers, because of their versatility, constitute the largest percentage of fabrics in Manmade fibers of
upholstery and clothing in the United States today. These fabrics, which bear names such chemical origin.

Chapter 11 Clothing and Fabric Fires 469

 FOAM RUBBER
Butyl, nitrile polybutadiene, and polyurethane rubbers are examples of common synthetic materials used
as coatings, padding, and insulation in household goods. They will resist ignition from a glowing source,
but without fire-retardant additives they are easily often ignited with a small open flame and will burn
with considerable energy.
Polyurethane foam rubber, because of its widespread use in upholstery (and occasionally clothing),
is the most commonly encountered synthetic with a substantial fire hazard. (Today, nearly all upholstered
furniture sold in the United States contains some polyurethane rubber padding). More and more of the
foam being used today is of the flame-resistant type. There are many types of flame-retardant foams;
most of them are resistant only to small open flames and will burn fiercely when ignited by a larger
flame. Others resist ignition even by much larger flame sources. It has been estimated that the percent-
age of upholstered furniture using flame-retardant grades of flexible urethane foams rose from 1 per-
cent in 1970 to 50 percent in 1994 as those products were improved in strength, durability, and cost.9
Flame-retardant grades of foam still remain very underused in residential furniture.10
elastomers ■ As noted in Chapter 5, not all “foam rubber” is polyurethane. There are other elastomers such as
Fibers or materials hav- isoprene isobutylene or polyester foams. These may have different ignition and flammability properties
ing elastic properties. than polyurethane and usually require lab identification. Most elastomers do not melt but decompose
when heated.

as nylon, Dacron, Orlon, Dynel, and Acrilan, can vary considerably in their ignition prop-
erties and burning behavior. Some of these variations are the effect of the chemistry of the
fiber itself, but they usually are the result of additives and modifiers used to improve par-
ticular properties of the fibers. Most synthetic fibers are made from chemicals that origi-
nate from petroleum products. As a class, they tend to be much more susceptible to
ignition by open flame than by a hot surface or glowing ember. Nylon, polyester, polyeth-
ylene, polystyrene, and polypropylene will ignite when in contact with a flame, especially
when in the form of thin fabrics or coatings. Fluffy polyester fiberfill (as seen in Figure
5-21) is very commonly found in cushions, pillows, and comforters.
In addition to being flammable, most (but not all) synthetic fabrics have a tendency
to melt and flow as they burn. When this occurs, burning droplets may fall into other
fuel and start secondary fires. If the fabric is in the form of a garment, the melting
material may fall away rapidly with only burning droplets clinging to the skin of the
wearer, with their difficult-to-extinguish flames inflicting localized burns.11 Interest-
ingly, the fabric of choice for the past 35 years for children’s sleepwear has been 100%
polyester. Its melting and shrinking behavior actually prevents the propagation of small
flames. This permits such fabrics to pass the CPSC sleepwear requirements of 16 CFR
1615 and 16 CFR 1616. When exposed to large flames, polyester will ignite and melt,
possibly causing serious burns. The ignition and burning properties of synthetics
change when they are blended with natural fibers. For instance, cotton-polyester
blended fabrics cause more injuries than polyester alone because the cotton keeps the
burned fabric together to support a rapidly spreading flame, and the fabric remains in
contact with the wearer.12 Pure thermoplastic synthetic fibers do not smolder, and burn
only in flames.

Nomex
One synthetic fiber, Nomex (an aromatic polyamide), resists all types of ignition and is
used to weave fabrics for fire protective gear for firefighters, aircraft pilots, and drivers of
racing cars. A single layer of Nomex cloth will protect human tissue from burning in con-
tact with a flame of 600°C (1,200°F) for up to 30 seconds. Other fire-retardant fabrics
available today include modacrylics, viscose rayon, polybenzimidazole (PBI), Kevlar,
Kynol, and some PVC (polyvinyl chloride) fabrics.13
470 Chapter 11 Clothing and Fabric Fires
NON-PETROLEUM-BASED SYNTHETIC FIBERS
Other synthetic fibers do exist that are not petroleum based—for example, rayon and
cellulose acetate. Rayon (regenerated cellulose) is chemically very nearly the same as cot-
ton because it is made by dissolving cotton or other cellulosic material and regenerating
the cellulose in the form of an extruded fiber by forcing the solution through a spinneret
into the regenerating bath. Because of the similarity to cotton, the main difference in fire
hazard is due to the different fabric weave and garment configuration and possibly size
of the fibers.
Acetate (acetylated cellulose) is similar to rayon in its manufacture and appearance,
but the chemical modification of the cellulose diminishes its flammability. It melts rather
easily but burns with some difficulty, except in lightweight fabrics, where it will burn rela-
tively rapidly.
Other synthetic fibers may be encountered in some unusual fabrics. Glass-fiber cloths
are manufactured and used mostly for nonapparel purposes, although clothing has been
made from them. Drapery is one of the common uses in which their obvious nonflamma-
bility and fire resistance would be of importance. Nearly all draperies sold for residential
use, however, are readily ignitable and will support an energetic fire. Curtains for public
occupancies are fire resistant and are often required to comply with NFPA 701.
Asbestos, a naturally fibrous mineral, has been used in special applications such as
theater curtains because of its great fire resistance and absolute nonflammability. It is
unlikely to be found in clothing or household textiles; it is no longer permitted because
of its potential carcinogenicity.

Fire Hazards
It will be noted that few of the fabrics encountered present an extreme fire hazard, espe-
cially those found in garments. Most common fabrics are readily ignitable and can sup-
port significant fires. Of those that remain, it appears that one of the most common fibers,
cotton, is also the most hazardous. The susceptibility of a fabric depends not only on
what fibers are used to make it but on how the fabric is woven and the design of the gar-
ment in which it is used. Yarn or thread that is tightly twisted will burn with less inten-
sity than loosely spun yarns. In the manufacture of some cloths, the threads comprise
tightly twisted fibers; in general, these are the fabrics that are most tightly woven as well.
The degree of tightness in spinning is a definite factor in determining the flammability of
the fabric when finished. The weight of the fabric (density) is also a key factor with regard
to its ease of ignition. Generally, the heavier a fabric is, the more it will resist ignition by
flaming sources. However, while heavier fabrics may be harder to ignite, once on fire, the
larger fuel load can sustain a fire more likely to cause injuries if the garment cannot be
removed or extinguished.

INFLUENCE OF WEAVE AND FIBER

The weave and finish influence the burning behavior of a fabric because they control the
amount of its fuel-air interface. A loosely woven cloth, with spacing between the threads,
will always burn much more rapidly than a tightly woven fabric of the same thread. For
example, a cotton canvas burns relatively poorly, while cotton gauze is rapidly consumed
(but canvas is more easily ignited by a cigarette), although both are made from the same
fiber. A tightly woven, heavy fabric like denim will ignite and burn much more slowly
than a sheer, lightweight cotton fabric such as the broadcloth used in shirts. A fabric with
a nap, such as cotton flannel, has air spaces between the loose fibers on the surface and
will ignite much more quickly than cotton denim, which is smooth surfaced. Cotton flan-
nel and terry-cloth fabrics have been shown to be particularly susceptible to ignition. A
fluffy high-pile fabric such as that used in some sweaters will ignite and burn faster than
Chapter 11 Clothing and Fabric Fires 471
 a close-knit low-pile fabric of the same fiber. Recent tests conducted to replicate fatal fire
cases have revealed the rapidity with which cotton shirt fabrics and synthetic windbreakers
(anoraks) can become fully involved from a single point of ignition with a flame.14 The
design of the finished garment has a bearing on its fire hazard because of the increased
likelihood of accidental ignition of loose-fitting garments such as robes, blouses, and
nightgowns. Garments with ruffles, pleats, loose and flowing sleeves, and similar features
are more prone to come into contact with a heat source during the course of the wearer’s
everyday activities. Excesses of material in such loose garments enhance the fuel load as
well as the fuel-air ratio and promote rapid burning. From a fire-safety standpoint, the
ease with which a burning garment can be removed is also a key design feature that influ-
ences the severity of possible injuries.

CLOTHING IGNITION
Ignition of clothing may originate in a variety of ways, most of which are identical with
the manner of kindling any other type of fire. Cigarette lighters in homes, as well as utility
or fireplace lighters, are some of the more dangerous items. Matches may lose a flaming
head when struck, and burning tobacco may drop on a dress, but research has shown that
it is unlikely that most garment fabrics will be ignited by a smoldering cigarette.15 One of
the more common and dangerous sources of ignition is a heating appliance against which
a flaring skirt may be pressed without the wearer’s knowledge, or a burning candle that
may readily ignite the loose sleeve at the dinner table. Again, the main item that predis-
poses to ready ignition is the design and the weight of the fabric, rather than the type of
cloth, so long as the cloth itself is reasonably ignitable.

Regulation of Flammable Fabrics

The flammability of flammable fabrics is regulated in the United States by the Consumer
Product Safety Commission (CPSC) on the federal level and also by states. Many states
and regulatory and control agencies have separate and broader-ranging regulations. This
is especially true with respect to carpets, wall coverings, and furnishings to be used in
hotels, theaters, hospitals, and other public buildings. Some local building codes establish
strict requirements for flammability of products in such high-risk occupancies. Of the
state regulations, those promulgated by California tend to be the most rigorous. Because
manufacturers who sell in California also sell in other states, the California regulations
function as de facto national regulations.
The federal and state regulations are based on specific test methods. Descriptions of
the relevant test methods are described in the Flammability Testing section.

REGULATION OF FLAMMABLE FABRICS

Overview of Federal Jurisdiction
The Flammable Fabrics Act (FFA) was passed in 1953 to regulate the manufacture of all
highly flammable wearing apparel and fabrics when such materials were to be offered for
sale on an interstate basis. Under the FFA, regulation rested with the U.S. Department of
Commerce (DOC).
The Consumer Product Safety Act of 1972 transferred all enforcement of the FFA to
the CPSC, which has broad jurisdiction over household products and their safety and
now promulgates standards and enforces them. Since then, a number of important regu-
lations have been enacted with regard to flammable fabric standards. Since 1972 the
Department of Health and Human Services (previously the Department of Health,
Education, and Welfare) has set standards for interior finishes and upholstery fabrics used
in hospitals and nursing homes. The fabrics and cushioning materials used in the interi-
ors of motor vehicles have had to pass certain tests prescribed by the Department of
472 Chapter 11 Clothing and Fabric Fires
Transportation (DOT) since 1972 (U.S. Motor Vehicle Safety Standard No. 302). (See the
discussion in Chapter 9 on MVSS 302.) DOT regulations on the flammability of aircraft
interior finishes were first passed in 1968 as FAA Title 14. The current tests for aircraft
upholstery deal with their resistance to radiant heat from a large nearby fire (e.g., a jet
fuel spill) and smoke emission.
Clothing Flammability
The first standard published under the FFA was CS-191-53, which measured flame spread
on fabric after direct flame ignition. This Standard for the Flammability of Clothing
Textiles is now known as 16 CFR 1610. It should be noted that the CPSC General
Wearing Apparel Standard 16 CFR 1610 is a very weak requirement (involving the appli-
cation of a needlelike flame to the surface of the test specimen for one second and observ-
ing fire spread.) Because newspaper will pass that test, the standard will exclude only the
most dangerous fabrics. Passing this test does not mean the fabric is fire safe! Tests 16
CFR 1615 and 1616 require the application of a larger flame to the exposed edge of the
test specimen for 3 seconds. In any such tests, only a fabric specimen is tested. The fin-
ished garment may well pose higher ignition risks. In 1967, the act was amended and
expanded to include textiles intended for nonclothing uses, such as interior finishes and
furnishings. Enforcement responsibilities were divided between DOC and the Federal
Trade Commission (FTC).
Flammability of Carpets and Rugs
The first standard created under the expanded FFA regulated the flammability of large
carpets and rugs. This standard, which is still mandatory for carpets and rugs sold in the
United States, measures horizontal flame spread upon ignition by methenamine tablet
(a reproducible ignition source). This Standard for the Surface Flammability of Carpets and
Rugs (DOC FF 1-70) is currently known as 16 CFR 1630. The flammability of small area
carpets and rugs was addressed in 1970 in DOC FF 2-70 (now known as 16 CFR 1631).
Flammability of Children’s Sleepwear
Children’s sleepwear was the next concern. In 1971, the smaller sizes (0 to 6X) of sleep-
wear sold in the United States came under the control of DOC FF 3-71 (now known as
16 CFR 1615). The larger sizes (7 to 14) were incorporated into the regulations under
DOC FF 5-74 (now 16 CFR 1616). Both standards are fairly stringent, dealing with igni-
tion of vertical fabric surfaces by open flames.
Mattress Flammability
Because of the frequency and severity of fires started in furniture and mattresses, the most
important and most controversial standards have to do with the flammability of mat-
tresses and upholstered furniture sold in the United States. Since 1973 there has been a
national standard for the cigarette ignition resistance of all mattresses, futons, and mat-
tress pads sold in the United States (DOC FF 4-72, now 16 CFR 1632, revised in 1984).
This standard mandates that mattresses and mattress pads must not be ignitable by ordi-
nary cigarettes placed directly on the mattress or between sheets covering that mattress.
Prior to 1973, the standard mattress consisted of coil springs covered with a layer of insu-
lating felt covered by thick batting, protected by a fabric covering or ticking. Although the
cotton batting was susceptible to smoldering ignition, it was difficult to ignite by open flame.
In compliance with the 1973 law, most manufacturers began using a non-fire-retardant
polyurethane pad as a supplement or a complete replacement for the cotton batting.
Although the mattresses made between 1973 and 2007 are, indeed, resistant to ignition
from cigarettes, they are much more likely to ignite from open flames and to continue to
burn than are their predecessors. Some manufacturers have produced mattresses made with
a fire-retardant-treated cotton or flame-resistant polyurethane pad to optimize all forms
of fire resistance. These materials are marginally more expensive than the unsafe prod-
ucts, however, and cost-conscious manufacturers have been slow to adopt such materials.
Chapter 11 Clothing and Fabric Fires 473
FIGURE 11-1 Foam
mattress topper burned
readily when ignited by a
match flame at its edge.
Courtesy of John Jerome and
Jim Albers.

Mattresses made for hospitals and prisons generally must meet more stringent standards
(discussed in a following section). Since 1977 the state of California has required all mat-
tresses that do not meet both smoldering and open-flame standards to bear a consumer
warning label stating that the mattress is susceptible to open-flame ignition.
In August 2001 the California legislature enacted a law (Assembly Bill 603) requir-
ing all mattresses, futons, and mattress sets manufactured on or after January 1, 2005,
and sold in California (including those for residential use) to be resistant to a large open
flame. The goal of the legislation was to reduce the incidence of fire deaths, injuries, and
property losses associated with bedroom fires.
California Technical Bulletin (TB) 603: Standard for Mattresses, Box Springs and
Futons required that all mattresses, box springs, and futons comply with new flame-
initiated testing procedures. The test was designed to simulate real-world flaming ignition
sources and measure the heat released from such items as they burn.16 This test has been
supplanted by CPSC test 16 CFR 1633, which is described later. Compliance with such
standards is almost exclusively achieved by the use of a variety of fire-blocking layers or
barriers rather than chemical retardants. These layers involve PVC cloth, metal films, and
other passive blocking systems.
Unfortunately, flammable bedding materials such as aftermarket foam bed-toppers,
comforters, pillows, and duvets are not covered under any current fire safety specifica-
tion. Such materials are readily ignited by small flames (as shown in Figures 11-1 and
5-20) and continue to be a threat to life safety despite strenuous efforts to include them
in fire standards. The California Bureau of Home Furnishings in 2003 published a draft
test standard for such items (denoted as TB 604). The CPSC in 2007 published a notice
of proposed regulation for bedclothes flammability (16 CFR 1634), but it has not pro-
gressed further. It would have mandated that such products pass a flame test such as that
in Figure 11-2. It has recently been shelved.
Upholstered Furniture
Upholstered furniture is subject, to a lesser degree, to the same sorts of regulations.
Although there are national industry standards regarding the flammability of upholstered
furniture, these are voluntary and aimed at smoldering ignition sources only. Since 1975
474 Chapter 11 Clothing and Fabric Fires
 FIGURE 11-2 Comforter
built to meet California’s
proposed safety standard
(TB 604) resists flaming
ignition, and fire spread
is dramatically reduced.
Compare with typical
comforter shown in
Figure 5-20. Courtesy of
Bureau of Electronics &
Appliance Repair, Home
Furnishings.

the state of California has required all upholstered furniture (and its component stuffing
materials) sold in the state to pass a standard test for flame and smoldering resistance
(California TB 116 and 117). The only exception is furniture whose intended use is solely
outdoors or recreational equipment whose use would preclude exposure to fire hazards
(exercise benches and the like). Furniture must bear consumer warning labels as to which
standards it has met.
TB 116: Flame Retardance of Upholstered Funiture is a voluntary standard that
requires that finished articles of furniture be resistant to cigarette ignition. TB 117 is a
mandatory requirement for furniture sold in California that specifies that all furniture
components must comply with a series of small-scale fire tests.17 These tests were designed
to measure the resistance of the furniture’s cushioning or padding to both flaming and
smoldering ignition sources.
In recent years California has developed rigorous standards and flammability test
procedures for mattresses and upholstered furniture for use in public buildings such as
jails, hospitals, nursing homes, hotels and motels, college dormitories, public auditoriums,
and so on.18 In 1980, California TB 121: Flammability Test Procedure for Mattresses for
Use in High Risk Occupancies, developed by the California Bureau of Home Furnishings,
was adopted by the California Board of Corrections as a mandatory requirement for all
California detention facilities.19
In 1984 the flammability test procedure for seating furniture for use in public occu-
pancies was first developed by the California Bureau of Home Furnishings and Thermal
Insulation.20 It was then published in 1988 as the Flammability Test Procedure for Seating
Furniture for Office and Public Occupancies, which is now known as California Technical
Bulletin 133 (TB 133).21 TB 133 became a mandatory law in California in 1992 and has
also been adopted by several other states and a number of local jurisdictions, including
Illinois, Ohio, Minnesota, Massachusetts, the city of Boston, and the New York–New
Jersey Port Authority. A generic version has been adopted as ASTM E 1537: Standard Test
Method for Fire Testing of Upholstered Furniture.22 Finally, TB 129: Flammability Testing
Procedures for Mattresses for Use in Public Buildings was published in 1992. TB 129 has
also been widely adopted, and a generic version has been published as ASTM E 1590.23
Chapter 11 Clothing and Fabric Fires 475
 The California Department of Consumer Affairs maintains a nationally recognized
testing laboratory as part of its Bureau of Home Furnishings and Thermal Insulation.
This laboratory is responsible for testing fabrics, apparel, and home furnishings to ensure
they comply with the stringent state standards. This laboratory reports that there is no
way to predict the time frame of smoldering versus flaming combustion in modern uphol-
stered furniture. In its tests, ignition by cigarette may lead to flaming ignition in times
ranging from less than 1 hour to 9 or more hours.24 This variability has been confirmed
in a number of full-scale furniture tests conducted by the authors.
The earliest flame-retardant urethane foams suffered reduced flame retardance as
the foam aged, especially at high temperatures. The additives being used today for the
most part are stable under normal living conditions, but a portion of the California law
requires that foams meet flammability tests before and after aging. The additives used in
some foams to make them more resistant to flaming ignition actually made them more
susceptible to smoldering ignition. It appears that manufacturers must be very careful in
their choice of flame retardants, adhesives, and even mold release agents to ensure the
best fire retardance properties in the finished product. Several flexible foams have recently
been made that have excellent fire resistance. Combustion-modified, high-resiliency
(CMHR) foams supplied by several domestic polyurethane foam manufacturers have
exceptional fire resistance but may be less flexible and of much higher density than stan-
dard foam. There is also a so-called low-smoke neoprene foam that has excellent flame
resistance. The CMHR foam has been useful in high-risk facilities such as jails and other
special applications, but until the advent of melamine urethanes (manufactured by
BASF) its physical limitation made it less desirable for regular use. The newest foams
promise excellent fire resistance and long life along with light weight, flexibility, and a
cost advantage over other fire-retardant foams.25 A variety of highly fire resistant foams
have been introduced, but except for specialty applications, they have not proven viable
in the consumer market.

Furniture Testing
Laboratory testing of furnishings can provide some general guidance to investigators as
to the ignition and flame-spread characteristics involved, but as Krasny et al. describe,
there are many approaches to testing and regulations.26 Ignition sources vary from radi-
ant sources such as cigarettes or glowing electric heating elements to flaming sources
from matches and methenamine pills to crumpled paper to gas burners of various heat
release rates and configurations, depending on the severity of the test. Information can
be gathered from rate of heat release (calorimetry), total heat released over a fixed time
(typically 180 seconds), rate of mass loss, total mass loss (absolute or percentage), visi-
ble charring, smoke production, rate of flame spread, or time to transition to flame.
(Pass-fail criteria can be based on any one or more of those factors.) The test samples
may be individual components, small-scale mock-ups, full-scale mock-ups, prototypes,
or real furniture sampled from stores. The small-scale mock-ups have to be tested to
demonstrate their correlation to full-scale performance before they can be used in regu-
latory testing. During the early 1990s, the European Union (EU) sponsored the largest-
ever research project on the fire behavior of upholstered furniture [published as
Combustion Behavior of Upholstered Furniture (CBUF)]. Some of the better-known or
more widely used tests for upholstered furniture are summarized in Table 11-1. Of par-
ticular interest are the tests that include flame ignition, such as BS 5852 and BS 6807
(UK), TB 133 and TB 129, and some of the ASTM tests. The International Organization
for Standardization (ISO) 9705 corner room configuration has also been studied for the
information it can provide to modeling real-world performance.27 In the United States,
the Upholstered Furniture Advisory Council (UFAC) has acted as an adviser to the indus-
try on flammability standards.
476 Chapter 11 Clothing and Fabric Fires
TABLE 11-1 Summary of Important Furniture Item Flammability Tests

(This item omitted from WebBook edition)

Chapter 11 Clothing and Fabric Fires 477

 FIGURE 11-3 Furniture Calorimeter.
Combustion products are drawn into Smoke transmission path
the exhaust hood, where the flow rate,
concentration of oxygen, carbon dioxide, Exhaust fan
and carbon monoxide are measured.
The heat produced is then calculated Gas sampling
Thermocouple
from the rate of oxygen consumption ring probe
based on the value 13 MJ of heat per Velocity probe
1 kg of oxygen. Source: Reprinted with Water-cooled
permission from the Society of Fire Protection skirt
Engineers, 2nd edition, National Fire Protection
Association, Quincy, MA, 1996. This reprinted Waste basket
material is not the complete and official position simulator
Gardon gauge
of the National Fire Protection Association on burner
the referenced subject, which is represented
only by the book in its entirety.
Weighing
platform

The heat release rate of furniture can be measured by bench-scale testing of models or
mock-ups of the furniture; however, more accurate results are generally produced by testing
in room calorimeters (such as those used in the California Technical Bulletin methods) or an
furniture calorimeter ■ oxygen-depletion furniture calorimeter of the type shown in Figure 11-3. The maximum
An instrumental system heat release rates measured in such devices for upholstered chairs ranged from 370 to 2,500
for measuring the kW; for sofas, 2,500 to 3,000 kW; for curtains, 150 to 600 kW; and for pillows from 16 kW
amount of heat pro-
duced by the combus-
(feathers) to 35 to 43 kW (urethane foam) to 117 kW (latex foam).28 As we saw in Chapter 3,
tion of a moderate- the heat release rates of common household furnishings are critical to evaluating the poten-
sized fuel package, and tial for development of the fire. It should be remembered that the rates as measured in a fur-
the rate at which the niture calorimeter are not directly transferable to predicting their performance in a room fire.
heat is produced. The maximum heat release rate of any such fuel is dependent to some extent on the radiant
heat that fuel is receiving from the fire environment. Heat feedback in a room fire may well
increase the maximum heat release rate produced over what is recorded in a calorimeter test.
At the same time, the furniture calorimeter allows maximum ventilation, while a fire in a
room may well reach ventilation limits that reduce the maximum heat release rate.

Flammability Testing
The testing of upholstered furniture and apparel fabrics is fairly complex and will not be
treated in great depth here. The American Society for Testing and Materials (ASTM
International) and various regulatory agencies have developed tests that address particu-
lar fabrics under reproducible laboratory conditions. Needless to say, variations in con-
struction, use, and the nature of fire exposure have a great influence on the flammability
of a particular fabric item. These variations cannot be duplicated completely in labora-
tory tests, and so, in real life, the fire hazards of such items may not be fully realized from
evaluating these tests. Some generalities can be offered, however, as in the case of open-
weave versus tight-weave fabrics. Apparel fabrics, because of their thinness, do not retain
heat very well and are therefore less susceptible to ignition by a glowing heat source such
as a cigarette than are heavy upholstery fabrics and stuffings, which retain heat until it
builds up to flaming ignition. The standardized tests reflect these general tendencies, and
some typical examples are outlined here.

FLAMMABILITY TESTS FOR FEDERAL REGULATIONS

Flammability of Clothing Textiles (Title 16 CFR 1610—U.S.)
Standard 16 CFR 1610 requires that a 10-in. (25-cm) piece of fabric that is placed in a
holder at a 45° angle (see Figure 11-4) and whose surface is contacted by a small flame
478 Chapter 11 Clothing and Fabric Fires
 FIGURE 11-4 16 CFR
1610 fabric test: 10-in.
(25-cm) strip of fabric
is held in a metal frame
at 45°. Flame is applied
to fabric at bottom for
1 second. Courtesy of
Salvatore Messina, CEO, The
Govmark Organization, Inc.

for 1 second not ignite and spread flame up the length of the sample in less than 3.5 seconds
for smooth fabrics or 4.0 seconds for napped fabrics.29
Flammability of Vinyl Plastic Film (Title 16 CFR 1611—U.S.)
Standard 16 CFR 1611 requires that vinyl plastic film (for wearing apparel) placed in a
holder at a 45° angle and ignited not burn at a rate exceeding 1.2 in. (3.0 cm) per second.
Flammability of Carpets and Rugs [Title 16 CFR 1630
(Large Carpets) and CFR 1631 (Small Carpets)]
Standards 16 CFR 1630 and 16 CFR 1631 require that a burning methenamine tablet
placed in the center of a specimen not cause a char more than 3 in. (7.6 cm) in any direction.
(The burning tablet or pill is a reproducible equivalent of a burning match, that is, ~50W
for 1 minute, as shown in Figure 11-5.)
Flammability of Mattresses and Pads (16 CFR 1632—U.S.)
A minimum of nine regular tobacco cigarettes are burned at various locations on the bare
mattress—quilted and smooth portions, tape edge, tufted pockets, and so forth. The char
length of the mattress surface must not be more than 2 in. (3 cm) in any direction from
any cigarette. The test is repeated with the cigarettes placed between two sheets covering
the mattress.
Flammability of Mattresses and Box Springs (16 CFR 1633—U.S.)
The top and side of the mattress or mattress set are exposed to two 18-kW gas burners
for a period of time (as shown in Figure 11-6). Failure of the test can be either (1) a peak
Chapter 11 Clothing and Fabric Fires 479
FIGURE 11-5 Methen-
amine pill on carpet test
(16 CFR 1630). Courtesy
of John D. DeHaan.

heat release rate of 200 kW or greater within 30 minutes of ignition or (2) a total heat
release of 15 MJ or greater within 10 minutes of ignition. As of July 1, 2007, all mattresses
sold for residential use in the United States must have passed the 16 CFR 1633 test.

Flammability of Children’s Sleepwear (16 CFR 1615 and 1616—U.S.)

Each of five 3.5 ⫻ 10-in. (8 ⫻ 25-cm) specimens is suspended vertically in holders in a cabi-
net and exposed to a small gas flame along its bottom edge for 3 seconds (as shown in

FIGURE 11-6 Mattress

and box springs built to
meet California’s safety
standard (TB 603) or
16 CFR 1633 are subjected
to test burner flames for
70 seconds but meet
stringent limits on heat
release rate and total heat
for 30 minutes. Compare
with fire in a typical
mattress in Figure 5-20.
Courtesy of Bureau of
Electronics & Appliance Repair,
Home Furnishings.

480 Chapter 11 Clothing and Fabric Fires

 FIGURE 11-7 Test appa-
ratus for 16 CFR 1615
(children’s sleepwear):
10 ⫻ 3.5-in. (25 ⫻ 8.0-cm)
Fabric strip hung vertically.
strip Flame is applied to bottom
Steel
cabinet edge.

Glass
door
Bunsen
burner

Figure 11-7). The specimens cannot have an average char length of more than 7 in. (18 cm);
no single specimen can have a char length of 10 in. (25 cm) (full burn); and no single
sample can have flaming material on the bottom of the cabinet 10 seconds after the igni-
tion source is removed. This test is required for finished items (as produced or after one
washing and drying) and after the items have been washed and dried 50 times.

FLAMMABILITY TESTS FOR CALIFORNIA REGULATIONS

Flame Retardance of Upholstered Furniture
(State of California TB 116 and 117)
For testing under TB 116, the article to be tested must be the finished product ready for
sale to the consumer or a prototype mock-up of actual components that duplicate the
design and structure of the finished product. Three regular tobacco cigarettes are ignited
and placed along each of the following structural features: smooth surface, welting, quilted
and tufted locations, crevices, and tops of arms and backs (as illustrated in Figure 11-8).

FIGURE 11-8 TB 116 9

(cigarette ignition) test:
Cigarettes are placed at
various locations on the
backrest, arms, and
cushions. Courtesy of Bureau
of Electronics & Appliance
Repair, Home Furnishings.

Chapter 11 Clothing and Fabric Fires 481

 All cigarettes are covered with a single layer of cotton sheeting. Furniture fails if obvious
flaming combustion occurs or if a char develops more than 2 in. (5 cm) in any direction
from a test cigarette.30 Filling materials are tested separately (as per TB 117) by inserting
a vertical specimen into the flame of a gas burner for 12 seconds and observing the length
of time for which flame and glowing combustion are maintained after the ignition flame is
extinguished and the length of the resulting char pattern.31 TB 117 requires both open-
flame and cigarette tests to be performed on all furniture fillings.

Flame Retardance of Mattresses and Seating Furniture

for Public Buildings (State of California TB 121 and 129)
Prison mattresses are tested using TB 121, a severe fire test using 10 double sheets of
newspaper in a waste container ignited directly beneath the penal institution mattress or
bed pad. The test is conducted in a specially designed full-scale test facility. The perform-
ance of the mattress is measured using temperature, smoke opacity, carbon monoxide
production, and heat release rate.32
Mattresses intended for health care facilities, convalescent homes, college dormito-
ries, and other multiple occupancies are tested using the protocol described in TB 129. It
is a full-scale test in which complete mattresses or other bedding components are tested
in a specially designed test facility. Ignition is by direct application of a T-shaped propane
burner (producing ~17 kW) to the bottom edge of the mattress for 180 seconds. Test cri-
teria are based on heat release rates achieved.33
Seating furniture for use in public buildings is tested under the protocol described in
TB 133. The test is carried out in a specially designed room calorimeter (like TB 121 and
TB 129). Ignition is induced by applying a large, calibrated square propane burner to the
seat surface of the furniture for 80 seconds. Failure is dependent on the maximum heat
release rate and total heat release observed.34 TB 133 tests can be performed by about a
dozen fire test labs in the United States. The concept of TB 133 testing requirements has
been promoted by the National Association of Fire Marshals, the International Association
of Fire Fighters, and several furniture manufacturers’ associations.35

California Technical Bulletin 603

The TB 603 test protocol, performed in a full-scale fire test facility capable of calorimet-
ric measurements, applies two large gas-jet flame sources (18 kW), ignited simultaneously,
to the bed set (mattress with a box spring or foundation if normally used with one), one
on the top surface for 70 seconds and a second flame on the lower edge for 50 seconds.
Sample burning is observed for 30 minutes, unless flashover appears inevitable or the
sample self-extinguishes. If the peak heat release rate in the test room exceeds 200 kW
during this period or the total heat release exceeds 25 MJ in the first 10 minutes, the prod-
uct fails the standard.36 TB 603 has now been supplanted by the CPSC national standard,
16 CFR 1633, which became a national requirement on July 1, 2007.
Mattresses meeting TB 603/16 CFR 1633 should exhibit a slower level of fire growth
when subjected to open-flame fires than those currently on the market and should thus pro-
vide a widened window of escape time for occupants before untenable conditions are
reached. As can be seen in Figure 11-6, there is a dramatic improvement in fire performance
in TB 603/16 CFR 1633–compliant mattresses. In addition, the federal mattress standard,
16 CFR 1632, requiring mattresses, futons, and mattress pads to be cigarette resistant, con-
tinues as law in all 50 states. Compare comforter fire tests in Figures 5-20 and Figures 11-2
to see what role such bedding materials can play in fires. As shown in Figure 11-1, aftermar-
ket mattress toppers are easily ignited with a match and can produce an intense, sustained fire.

GENERAL OBSERVATIONS
As part of the extensive testing conducted since 1975 by the California Bureau of Home
Furnishings and Thermal Insulation (BHFTI), some general observations have been made
482 Chapter 11 Clothing and Fabric Fires
regarding cigarette ignition of furniture and mattresses. The key factor in predicting the
smoldering propensity of upholstered furniture is the type of fibers used in the exterior
cover fabric. Furniture covered in fabrics made predominantly of cellulosic fibers is much
more likely to smolder than that covered in thermoplastic-fiber fabrics. The tendency to
smolder increases with increasing weight of cellulosic fabric. Furniture covered with ther-
moplastic fabrics, particularly with heavier fabrics, tends to be very resistant to ignition
by cigarette, even if the stuffing is cotton batting.37 In the BHFTI testing, cigarette-
induced furniture ignition required from 40 minutes to a number of hours before flaming
combustion begins. In some cases, open flaming never began. In others, flaming combus-
tion proceeded so quickly that a typical article of furniture, once flaming, could create
flashover conditions in a normal-size room in as little as 2 to 3 minutes.38 Extensive test-
ing by the Center for Fire Research at the National Institute of Standards and Technology
(NIST) has revealed that the transition to flame from a smoldering cigarette fire is quite
variable. In one series of tests on non-flame-retardant cotton fabric over cotton and ure-
thane foam padding, 22 chairs were tested; 10 never went to flame (self-extinguished or
smoldered to completion). The remaining 12 identical chairs took from 29 to 63 minutes
to transition to flame (with an average of 44 minutes). In various NIST tests, smolder-to-
flame in upholstered chairs required as little as 22 minutes to as much as 306 minutes.39
This range tallies with the authors’ experience, where the shortest time for flames to be
observed was 22 minutes when the cigarette was placed between intact upholstered sur-
faces of the armrest and seat cushion. When the outer covering is torn and the cigarette
is placed directly in contact with non-flame-retardant cotton padding, ignition has been
observed in 18 minutes. On other occasions, flaming ignition has taken up to 3 hours.
Cigarette ignition of kapok stuffing and latex foam rubber has also been created (with
ignition about 1 hour after insertion of the cigarette).

CONSIDERATIONS FOR FIRE INVESTIGATORS

Many variables contribute to susceptibility and time factors. These include fuel moisture,
ambient temperature, position of the cigarette, and nature of the outer fabric covering(s)
(thin or thick, cellulosic or synthetic, presence of a barrier layer, etc.). The presence of
external air currents can also affect flame transition, because a draft can increase the com-
bustion rate. As mentioned in the TB 116 test, a single layer of cloth over a smoldering
cigarette can reduce losses by convection and radiation and make ignition more likely.
Holleyhead published an extensive review of the furniture fire test literature and explored
a number of the furniture-related variables.40 The investigator needs to be aware of the
variables related to both the furniture and the scene, and what effects they can have on
ignitability. It should be noted that tests are conducted with full cigarettes, and most acci-
dentally dropped cigarettes are likely to be only partial ones. Ignition requires sufficient
contact, which may not occur with partial cigarettes or dropped embers. The bottom line
is that cigarettes and open flames are competent ignition sources for furniture depending
on the nature of the fabrics involved. Open flames will ignite modern fabrics almost on
contact (3–5 seconds), whereas a smoldering cigarette will require 20 minutes or more,
and then only if the other factors are favorable. Ignition from open flames is a more sub-
stantial risk today than just 10 years ago due to the popularity of candle usage (see
Chapter 6). Home decorating schemes and “aromatherapy” uses mean candles are in use
for prolonged periods around furnishings and bedding. These sources have to be consid-
ered in addition to “traditional” risks from matches and lighters.
Thermoplastic fabrics tend to melt and shrink when exposed to the radiant heat of a
growing fire, which can expose the underlying padding to flames and radiant heat, pro-
ducing very fast fire growth. Polypropylene (polyolefin) and polyvinyl fabrics pose partic-
ular hazards in this regard. The inclusion of fire-blocking layers (fiberglass, PBI, coated
fabrics, and even neoprene foam bonded to a textile fabric) can dramatically reduce the
Chapter 11 Clothing and Fabric Fires 483
 maximum heat release rate demonstrated by these furnishings.41 One recent observation
by NIST researchers is that as polyurethane foams burn they decompose and “melt” into
a semiliquid state that pools under the chair, sofa, or mattress to support a very intense
pool fire. Using inert fillers to inhibit the flow of such products resulted in a significantly
reduced maximum heat release rate.42 When carpets are involved, the types of face yarn
and backing are major factors. Because of its very low melting point, polypropylene fiber
carpet melts very readily to expose combustible padding beneath.
Fire investigators today are realizing that the types of fibers, fabrics, and padding
involved in furniture, mattresses, clothing, and carpet all can affect ignition, fire spread,
fire intensity, and generation of toxic gases. As a result, these items need to be sampled
and preserved for possible identification. In the future, fire investigators will have to
determine whether the furnishings have met appropriate fire standards. Meeting stan-
dards can be accomplished either by design (such as barrier layers), chemical structure, or
treatment with fire retardants. The mere presence of a tag or label does not prove the
product was compliant with those standards. CPSC and the California Deptartment of
Consumer Affairs report numerous failures of consumer products, some complete with
false or forged tags.

484 Chapter 11 Clothing and Fabric Fires

 CHAPTER REVIEW

Summary
In general, it should be remembered that tests applied to as part of the inquiry into the Stardust Disco disas-
fabrics and finished household goods reflect only the ter (1981) was that even when individual materials
ignitability of the product—its susceptibility to being and products pass appropriate flammability tests,
accidentally ignited. These tests do not determine they may interact in complex ways to create extraor-
whether a fabric will burn and contribute to a room’s dinary fire conditions. In the Stardust Disco fire, the
fuel load in a given room fire. All fabrics will burn if the combination of furnishings and wall coverings
temperature, ventilation, and duration of the fire are produced radiative feedback, drop down of flaming
sufficient; even highly fire-resistant fabrics of Nomex or residues, and piloted ignition of smoldering uphol-
PBI will burn under the right conditions. These com- stery that spread a fire throughout an open seating
ments on fabrics and their fire tests are offered to assist area much more quickly than anyone could have
the investigator in assessing the logical consequences of predicted. It was only through extensive full-scale
ignition from small flames or glowing embers that come simulations that the fire could be reenacted. The fire
in contact with apparel or upholstered furniture. investigator must keep those lessons in mind when
One lesson from the extensive tests conducted evaluating the contributions of individual fuels in
by the Fire Research Station in the United Kingdom any fire. 43

Review Questions
1. What is the chemical difference between natural 6. What is the California flame resistance standard
fibers and synthetic fibers? Name three examples (TB 603) and why is it so different from previous
of each. standards?
2. What fire behavior is peculiar to cellulosic fibers? 7. Describe three of the fabric flammability tests used
Why does it not occur with most synthetics? today.
3. Name three of the U.S. government regulations 8. What is the methenamine pill test and what is it
on flammability of fabrics and describe each used for?
application. 9. Why is polypropylene carpet such a special
4. What is a typical time frame for cigarette ignition consideration in fire investigations?
of upholstered furniture? 10. What is the USMV SS 302 and what are its
5. What does an oxygen-depletion calorimeter do? limitations?

References
1. D. Miller, R. Chowdhury, and M. Green, “2004–2006 (Quincy, MA: Fire Analysis and Research Division,
Residential Fire Loss Estimates,” (Washington, DC: National Fire Protection Association, 2005.
U.S. Consumer Product Safety Commission, October 7. J. F. Hoebel, G. H. Damant, S. M. Spivak, and G. N.
2009), 9–10. Berlin, “Clothing Related Burn Casualties: An Over-
2. M. Ahrens, “Home Fires that Began with Upholstered looked Problem,” Fire Technology 46(3), July 2010,
Furniture” (Quincy, MA: Fire Analysis and Research 629–649.
Division, National Fire Protection Association, 2008). 8. J. R. Hall, Jr. “Targeting Upholstered Furniture Fires,”
3. Ibid. NFPA Journal (March–April 2001): 57–60; J. F. Krasny,
4. Miller et al., “2004–2006 Residential Fire Loss Estimates.” W. J. Parker, and V. Babrauskas, Fire Behavior of
5. J. R. Hall, “The U.S. Smoking Material Fire Problem” Upholstered Furniture and Mattresses (Norwich NY:
(Quincy, MA: Fire Analysis and Research Division, William Andrew, 2000), 7.
National Fire Protection Association, 2008). 9. G. H. Damant, “The Flexible Polyurethane Foam
6. K. D. Rohr, “Selections for Products First Ignited in Industry’s Response to Tough Fire Safety
U.S. Home Fires: Soft Good and Wearing Apparel” Regulations,” Journal of Fire Sciences 4 (1986).
Chapter 11 Clothing and Fabric Fires 485
 Revised by G. H. Damant, personal communication, J. F. Mikami, “Flaming Combustion of Upholstered
January 1996. Furniture Ignited by Smoldering Cigarettes” (North
10. R. P. Foster and J. B. Zicherman, “Is There a Time Highlands, CA: California Department of Consumer
Bomb in the Sofa?” Trial (November 2005): 58–63. Affairs), International Conference on Fire Safety, San
11. A. K. Mehta and F. Wong, “Hazards of Burn Injuries Francisco, January 1986.
from Apparel Fabrics,” NTIS COM-73–10960, Fuels 25. Damant, “The Flexible Polyurethane Foam Industry’s
Research Laboratory, MIT, February 1983. Response”; Damant, “Recent United States Develop-
12. J. Holmes, “Some Clothes May Play with Fire,” Insight ments in Tests and Materials.”
(June 9, 1986): 46. 26. Krasny et al., Fire Behavior of Upholstered Furniture
13. V. M. Bhatnagar, ed., Advances in Fire Retardant Textiles and Mattresses.
(Westport, CT: Technomic, 1975); S. Backer, 27. B. Sundstrom, ed., Fire Safety of Upholstered Furniture:
G. C. Tesoro, T. Y. Yoong, and N. A. Moussa, Textile The Full Report of the European Commission Research
Fabric Flammability (Cambridge, MA: MIT Press, 1976). Programme (CBUF) (London: Interscience Communi-
14. R. McRae, Calgary Police, personal communication, cations, 1995).
June 2000; J. W. Munday, personal communication, 28. Krasny et al., Fire Behavior of Upholstered Furniture
March 2001. and Mattresses.
15. Mehta and Wong, “Hazards of Burn Injuries from 29. Guide to Fabric Flammability, U.S. Consumer Product
Apparel Fabrics”; J. L. Sanderson, “Cigarette Fires in Safety Commission, June 1976, 1–5.
Fabrics,” Fire Findings 15, no. 3 (Summer 2007): 1–3. 30. Technical Bulletin 116: Requirements, Test Procedure
16. Technical Bulletin 603: Requirements and Test Procedures and Apparatus for Testing the Flame Retardance
for Resistance of a Mattress/Box Spring Set to a Large of Upholstered Furniture (North Highlands, CA:
Open Flame (North Highlands, CA: California California Department of Consumer Affairs,
Department of Consumer Affairs, January 2002). January 1980).
17. Technical Bulletin 117: Requirements, Test Procedure 31. California Department of Consumer Affairs, Technical
and Apparatus for Testing the Flame Retardance of Bulletin 117.
Resilient Filling Materials Used in Upholstered Furniture 32. California Department of Consumer Affairs, Technical
(North Highlands, CA: California Department of Bulletin 121.
Consumer Affairs, January 1980). 33. California Department of Consumer Affairs, Technical
18. G. H. Damant, “Recent United States Developments in Bulletin 129.
Tests and Materials for the Flammability of Furnishings,” 34. G. H. Damant et al., “The California TB 133 Test:
Journal of the Textile Institute 85, no. 4 (1994a). Some Background and Experience,” Fire Safety Journal,
19. Technical Bulletin 121: Flammability Test Procedure 16 (1990): 1–12.
for Mattresses for Use in High Risk Occupancies 35. Krasny et al., Fire Behavior of Upholstered Furniture
(North Highlands, CA: California Department of and Mattresses.
Consumer Affairs, April 1980). 36. California Department of Consumer Affairs, Technical
20. P. S. Wortman, S. S. Williams, and G. H. Damant, Bulletin 603.
“Development of a Fire Test for Furniture for High 37. G. H. Damant, “Cigarette Ignition of Upholstered
Risk and Public Occupancies” in Proceedings of the Furniture,” Inter-City Testing and Consulting,
International Conference on Fire Safety (Sunnyvale Sacramento, CA, 1994; also in Journal of Fire Sciences
CA: Product Safety Commission, 1984), 55–67. (September–October) 1995.
21. Technical Bulletin 133: Flammability Test Procedure for 38. Krasny et al., Fire Behavior of Upholstered Furniture
Seating Furniture for Use in Public Occupancies (North and Mattresses.
Highlands, CA: California Department of Consumer 39. Ibid.
Affairs, 1988). 40. R. Holleyhead, “Ignition of Solid Materials and
22. Damant, “Recent United States Developments in Tests Furniture by Lighted Cigarettes: A Review,” Science
and Materials.” and Justice 39, no. 2 (1999): 75–102.
23. Technical Bulletin 129: Flammability Testing Procedures 41. Damant, “Recent United States Developments in Tests
for Mattresses for Use in Public Buildings (North and Materials.”
Highlands, CA: California Department of Consumer 42. bfrl.nist.gov.
Affairs, 1992). 43. P. T. Pigott, “The Fire at the Stardust, Dublin: The
24. V. Babrauskas and J. F. Krasny, “Upholstered Public Inquiry and Its Findings,” Fire Safety Journal 7
Furniture Transition from Smoldering to Flaming,” (1984): 207–12; W. A. Morris, “Stardust Disco
Journal of Forensic Sciences 42 (1997): 1029–31; Investigation: Some Observations on the Full-Scale Fire
J. A. McCormack, G. H. Damant, S. S. Williams, and Tests,” Fire Safety Journal 7 (1984): 255–65.

486 Chapter 11 Clothing and Fabric Fires

 CHAPTER
12
Explosions and Explosive
Combustion

Courtesy of Det. Richard Edwards

(retired), Los Angeles County Sheriff’s
Department, Whittier, CA.

KEY TERMS

backdraft, p. 502 explosive, p. 489 Molotov cocktail, p. 519

brisance, p. 511 fragmentation, p. 522 seat of explosion, p. 520
deflagration, p. 489 high explosive, p. 507 stoichiometric mixture, p. 495
detonation, p. 489 low explosive, p. 507

OBJECTIVES

After reading this chapter, you should be able to:

■ Explain the basic concepts and definitions of explosions.
■ Identify the common causes of diffuse-phase explosions by gases, vapors, and dust
suspensions.
■ Identify the common types of dense-phase explosives, their uses, and whether they
are high or low explosives.
■ List the proper methodology for conducting an explosion scene investigation.
■ Describe several techniques used during scene examination, evidence recovery, and
laboratory analysis.
■ Locate the latest explosives statistics published by the U.S. Bomb Data Center.

487
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

F
ires and explosions so frequently accompany each other that no treatise on fire and
its investigation would be complete without some special consideration of the nature
and role of explosions. Explosions range in violence from the diffused type of rolling,
progressive flame resulting from the combustion of a rich mixture of flammable gases or
vapors in air, to the violent, almost instantaneous detonation of condensed-phase explo-
sives. The destructive effects of these explosions vary widely and may involve broken
windows and dislodged bricks from a diffuse gas explosion or the shattering blow of a high-
explosive detonation. In the most general terms, an explosion may be defined as the
sudden conversion of potential chemical or mechanical energy to kinetic physical energy in
the form of pressure, heat, and sometimes light as gases are generated very rapidly or
released. There are four basic types of explosions that the fire investigator may encounter:
chemical, mechanical, electrical, and nuclear. By far the most common will be chemical or
mechanical, and chemical explosions will be the focus of most of this chapter. Mechanical
explosions involve the sudden release of a gas or liquid under extreme pressure as its con-
tainer, vessel, or pipe bursts. Electrical explosions occur when large currents suddenly flow
through air, oil (in transformers or switch gear), or undersized conductors. The extremely
rapid heating causes the expansion of surrounding gases, the boiling, and expansion of the
oil, or evaporation of the conductor. This expansion occurs so rapidly (as in the “crack” of a
nearby lightning strike) that explosive effects can be created, especially if the event is con-
fined. Such events can also trigger deflagrating chemical explosions. Nuclear explosions
cause massive blast effects by extremely rapid heating of surrounding gases and evaporat-
ing solids. (It is most unlikely that the reader will be involved in such a scene!) All the fore-
going processes involve the sudden release or production of gases that exert force and inflict
damage to nearby materials.

Chemical Explosions
As you will see, the conversion of chemical energy to kinetic energy can take on a variety
of forms, and the energy released can exhibit a range of effects. The processes involved
and the speed with which the reaction progresses depend not only on the chemistry of the
materials involved but on their physical form and environment (i.e., containment). For the
purposes of this book, explosive materials can be thought of as falling into two general
categories—diffuse explosive mixtures and condensed-phase or concentrated explosives—
and each will be examined in turn.
488 Chapter 12 Explosions and Explosive Combustion
 While diffuse explosive mixtures are most often combinations of a combustible gas
with air, the condensed-phase explosion is not dependent on oxygen from the air but
results from a combustive oxidation of the explosive. An explosive, then, can be defined explosive ■ Any mate-
as any material capable of undergoing an almost instantaneous conversion from one form rial that can undergo a
sudden conversion of
to another, producing gases and heat as it does so. An explosive can be a material that
physical form to a gas
contains a mixture of both a fuel for the combustion and an internal oxidizer, or one that with a release of energy.
is unstable and capable of rearranging its chemical structure when stimulated, with highly
exothermic effects. Black powder is an example of the first “mixture” type, while dyna-
mites, TNT, nitroglycerine, PETN, and lead azide are examples of the second (single com-
pound). All contain in themselves all the essential components for almost instantaneous
conversion to gases with great kinetic energy (heat). In this chapter, the nature of explo-
sive combustion will be explored and the processes, effects, and diagnostic signs of vari-
ous mechanisms will be discussed.

Key Terms and Concepts

Oxidation reactions can proceed at various rates, from the processes of rusting or decay,
which may progress slowly and go unnoticed, to the extremely fast detonation of some
explosives, which proceed so quickly that they cannot be directly observed. It is when oxi-
dation proceeds at a sufficiently rapid rate to generate conveniently measurable amounts
of heat and light that it is called combustion. One can actually accelerate the rusting
process of steel by heating steel wool in an oxygen-rich environment; it is then oxidized
with a flash of flame, undergoing true combustion.
Because some compounds and elements can undergo an oxidation reaction without
oxygen itself being present, the term oxidation must be reduced to its most basic meaning—
the loss of electrons from a reactant (to be gained by another reactant). Therefore, oxidation
can be defined as the combination of one compound or element with another that involves
the loss of electrons from its atomic structure.
Turner and McCreery offer some excellent definitions of the various classes of oxida-
tion reactions (chemical explosions):1

Combustion A rapid oxidation with the evolution of heat and light.

Deflagration A very rapid oxidation with the evolution of heat and light as well as the deflagration ■ A very
generation of a very low intensity pressure wave of moving gases. Note this rapid oxidation with the
is a low-energy pressure wave and not a shock wave. evolution of heat and
light and the generation
Explosion A very rapid oxidation with the evolution of considerable heat accompanied
of a low-energy pressure
by a disruptive effect (the disruptive effect being a result of pressure and
wave that can accom-
confinement). All reaction products are gaseous.
plish damage. The
Detonation An extremely rapid, almost instantaneous oxidation reaction (oxidation here reaction proceeds
meaning the loss of electrons, not necessarily the combination with elemental between fuel elements
oxygen) with the evolution of considerable heat accompanied by a very violent at subsonic speeds.
disruptive effect and intense, high-speed shock wave. The detonation shock
detonation ■ An
wave is supersonic (faster than the speed of sound) and requires no confine-
extremely rapid reaction
ment for its propagation. All reaction products are gaseous.
that generates very high
temperatures and an
intense pressure/shock
wave that produces
Note that these classes of reactions are not separate and discrete categories but por- violently disruptive
tions of a continuum of reaction rates—from the very slow to the ultrafast. Because the term effects. It propagates
explosion is used to describe the entire continuum from rapid combustion to detonation, through the material
at supersonic speeds.
it has been suggested that to minimize confusion, the categories be redefined as combustion,
deflagration, and detonation, as follows:
Chapter 12 Explosions and Explosive Combustion 489
 Combustion Rapid oxidation that generates heat and light but does not generate sufficient
gases to produce a pressure wave.
Deflagration A very rapid oxidation producing heat, light, and a pressure wave that can
have a disruptive effect on the surroundings. This is an oxidation process that
takes place at the surface of the reacting fuel (whether vapor or solid). This
process can proceed at speeds up to 1,000 m/s (3,280 ft/s).
Detonation An extremely rapid reaction that generates very high temperatures and an
intense pressure/shock wave that produces violently disruptive effects. The
process occurs through the reacting material at speeds in excess of 1,000 m/s
(3,280 ft/s)* as the shock wave is transmitted at supersonic speeds.2
*
Sometimes rounded to 3,000 ft/s. The speed of sound in air at 20°C (68°F) and sea level is 334 m/s (1,100 ft/s), but in
denser materials it is much higher. Because the “average” density of high explosives is about 1.5 g/cc, the “average”
speed of sound in those materials is about 1,000 m/s (3,280 ft/s).

These are, in fact, the definitions used by the Institute of Makers of Explosives and
the Bureau of Alcohol, Tobacco, Firearms and Explosives. Because there is less confusion
with this scheme and its transition criteria (pressure and speed) are fairly clearly defined,
it will be used as the basis for the discussions in this chapter. One can see from these def-
initions that each category produces a set of physical effects that will leave distinctive
diagnostic signs of pressure and speed in a structure or vehicle. A careful evaluation of
these diagnostic signs at a fire-explosion scene will allow the analytical investigator to
estimate the nature of the explosion that occurred.
Virtually all chemical explosions create gases and heat by their functioning. The post-
combustion gases usually occupy a larger volume than the prereaction components (a
great deal larger in the case of solid explosives), and the creation of great volumes of
gaseous products accounts for some gaseous phase effects. The heat created, however,
causes the majority of the expansion and thereby the majority of the blast effects.
Explosions are characterized by both the pressures produced and the rates at which
those pressures rise. Harris demonstrated that for hydrocarbon fuels (methane, ethane,
propane, etc.), the number of moles of precombustion gases is about the same as the num-
ber of moles of postcombustion gases.3 This means that the expansion of gases is caused
primarily by the thermal expansion caused by the heat produced. Under ideal conditions,
this heating [from 27°C (80°F; 300 K) to 2,027°C (3,680°F; 2,300 K), for example] could
cause overpressures of up to 8.3 bar, 827 kPa, or 120 psi. The pressures produced by real-
world deflagrations of hydrocarbon fuels are usually limited by the failure pressure of the
confining vessel or structure. Let us first examine the most common type of explosive
combustion in structure fires—that of a diffuse-phase gas, vapor, or dust explosion.

Diffuse-Phase Explosions
Diffuse-phase explosions are the rapid combustion of a fuel (gas, vapor, mist, or dust) pre-
mixed with air or oxygen and then ignited. The reaction proceeds as a flame front extends
from the ignition source into the premixed fuel/air mixture and then grows larger in volume
and surface area with time. Its effects are dependent on its type of fuel, concentration,
ignition mechanism, and degree of confinement. Each will be discussed in turn.

GASES
Perhaps the most common example of an accidental diffuse-phase explosion is that of
natural (or LP gas) escaping in quantity into a confined space and mixing with air and
then being ignited. Depending on the amount of gas escaping, the amount of air in the
490 Chapter 12 Explosions and Explosive Combustion
space, and the amount of leakage and mixing that takes place, the reaction mixture can
be within its explosive range (see Chapter 4), too rich, or too lean. As we saw in
Chapter 4, when a volume of natural gas is uniformly mixed with about 10 volumes of
air, a mixture close to its ideal explosive ratio, 苲9 percent, results. If this mixture fills
a room and is ignited, the combustion is a deflagration that progresses radially in all
directions away from the ignition source and through the room at a rate of up to 3 m/s
(10 ft/s). (The rate is pressure and temperature dependent and will increase as the igni-
tion proceeds and the flame front expands, so it can become considerably higher.) The
combustion produces a great amount of heat very suddenly, and the combustion gases
(and remaining atmospheric gases—nitrogen, etc.) are quickly heated and expand vio-
lently. This expansion is, of course, resisted by the walls, floor, and ceiling of the room.
The resulting overpressure, which is typically a few pounds per square inch (psi), sev-
eral kilopascals (kPa), or fraction of a bar* pushes the walls, floor, and ceiling, or
breakable portions like windows and doors, outward. If enough gas is premixed with
room air, the destruction can be extensive, as shown in Figures 12-1a–m.

ON SCENE
A natural gas explosion destroyed this four-unit apartment building when the gas shutoff valve was
removed from the 19-mm (3/4-in.) delivery pipe serving the gas range in a vacant unit. Over an esti-
mated 45-minute interval, gas filled the apartment until it was ignited in the vicinity of the wall furnace
in the living room. The resulting deflagration lifted the roof of the entire building and scattered large
and small pieces in all directions. The blast took out two of the walls of the living room and inflicted
serious damage to the living room of the adjoining apartment, injuring its occupant. The wall where the
wall heater was located was left intact, as was the longitudinal wall that separated the living room and
the kitchen. Spot fires were ignited in the furniture by burning debris (cellulose ceiling insulation) from
the attic. The plume of gas released from the open gas line charred the back of the stove and the side
of the adjoining cabinet unit as it burned post-blast (see Figures 12-1a–m).

FIGURE 12-1A This overhead photo shows FIGURE 12-1B Overhead, looking down FIGURE 12-1C A view showing the direction
the debris field extending more than 100 ft in onto the top floor units. The source of gas and intensity of the damage. The living room
all directions. Note the car in the street struck was in the kitchen (upper left). The source of the vacant unit had the highest concentra-
by a large portion of the front wall of the of the ignition was in/near the wall heater in tion of gas. The blast propagated into an
structure. Wall debris crushed the rear of the the living room (upper center of photo). occupied apartment (left) and dissipated its
convertible top, sparing the mother and child Courtesy of Det. Richard Edwards (retired) Los energy as it moved out. Fire damage to the
inside any injury. Courtesy of Det. Richard Edwards Angeles County Sheriff’s Department, Whittier, CA. ceiling joists resulted from post-blast fires in
(retired) Los Angeles County Sheriff’s Department, the furniture. The blast propagated into the
Whittier, CA. small hallway and then into the bedrooms,
propelling the walls out toward the street.
Courtesy of Det. Richard Edwards (retired) Los
Angeles County Sheriff’s Department, Whittier, CA.

*
1 atmosphere  14.7 psi  1.013 bar  101.32 kPa, so 1 psi  6.89 kPa  0.07 bar  70 mbar  27.67 in. w.c.
Chapter 12 Explosions and Explosive Combustion 491
FIGURE 12-1D A view into the kitchen FIGURE 12-1E A view from behind the stove FIGURE 12-1F Another view of the kitchen.
showing the water heater (center). If the water showing the plume damage to the side of The stove had been moved away from the wall
heater had been the ignition source, there the cabinet and the flex line with no shutoff and the burners and top moved before the fire
would have been more damage to the rooms valve. The open gas line is in the wall on the (in an attempt to remove the pressure regulator
to the left of the main dividing wall. Courtesy right. Courtesy of Det. Richard Edwards (retired) Los from the inside). The unvented hood and
of Det. Richard Edwards (retired) Los Angeles County Angeles County Sheriff’s Department, Whittier, CA. adjacent cabinets aimed the flow of venting gas
Sheriff’s Department, Whittier, CA. (from behind the stove) toward the living room
door (on right). Courtesy of Det. Richard Edwards
(retired) Los Angeles County Sheriff’s Department,
Whittier, CA.

FIGURE 12-1G Damage to the ceiling of the FIGURE 12-1H Localized fire damage from FIGURE 12-1I Localized fire damage to the
ground-floor apartment (below the origin) burning ceiling insulation. Courtesy of Det. Richard occupied apartment from a sofa ignited by
shows the downward deflection of joists. Edwards (retired) Los Angeles County Sheriff’s burning debris from the attic. Courtesy of
Courtesy of Det. Richard Edwards (retired) Los Angeles Department, Whittier, CA. Det. Richard Edwards (retired) Los Angeles County
County Sheriff’s Department, Whittier, CA. Sheriff’s Department, Whittier, CA.

492 Chapter 12 Explosions and Explosive Combustion

FIGURE 12-1J Fresh tool marks on the stove FIGURE 12-1K Documentation included FIGURE 12-1L Undamaged gas line and
pressure regulator showed the unsuccessful status of gas service to building. Courtesy of burner for wall furnace (most probable igni-
attempt at removal. Courtesy of Det. Richard Det. Richard Edwards (retired) Los Angeles County tion source). Note the undamaged louvers
Edwards (retired) Los Angeles County Sheriff’s Sheriff’s Department, Whittier, CA. (rear). Courtesy of Det. Richard Edwards (retired) Los
Department, Whittier, CA. Angeles County Sheriff’s Department, Whittier, CA.

Bedroom Bedroom

CL CL

CL

Kitchen WH Wall heater

Living room
(vacant)
Gas
line

CL

CL

Kitchen
Living room
(occupied)

CL
Bath CL

FIGURE 12-1M The

Bedroom 2 Bedroom 1
floor plan showing the
directions of blast
propagation through
the top floor.

Chapter 12 Explosions and Explosive Combustion 493

 FIGURE 12-2A Broken gas pipe filled this cabin from FIGURE 12-2B TV remote embedded in ceiling when
below the floor where it was ignited. Courtesy of Lamont floor and TV were lifted by explosion. Courtesy of Lamont
“Monty” McGill, Fire/Explosion Investigator. “Monty” McGill, Fire/Explosion Investigator.

This explosion process meets our definition of a deflagration because of the speed
of propagation and moderate pressure that develop. With the heat generated in such a
short time, items that are most readily ignited, charred, or damaged will be affected.
Eyebrows and scalp hair may be singed, but flesh will not be deeply burned, because a
longer time is needed for heat to penetrate and damage skin than hair. Clothing and
thin draperies may be scorched or set on fire, but furnishings will be little affected
because finer, less massive fabrics ignite more readily than upholstery and wood. The
pressures produced may move walls outward, break doors and windows, and even lift
the roof or floors (see Figures 12-2a and b). However, a person in the room (especially
one close to the source of ignition) may not be seriously injured by the blast itself. The
compressive pressure wave near the origin is usually slight enough not to cause perma-
nent injury and is moving away from the ignition source, and the flash fire may be tran-
sitory enough that an occupant may not be seriously burned, and combustibles may not
be ignited. The pressures are not such that a localized “seat” of shattered or pulverized
material is produced.
The combustion rate grows as the surface area of the flame front expands. As a result,
anything that increases the surface area by turbulent mixing can increase the combustion
rate. This mixing can be caused by furniture or by passage of the flame front through door-
ways or across barriers. Expansion of gases in the room of origin may also pressurize and
heat the mixture in adjoining rooms. Since higher pressures and temperatures increase the

494 Chapter 12 Explosions and Explosive Combustion

intensity of the combustion and its propagation rate, the reaction accelerates. While a
gas/air explosion originates at its ignition source, the greatest destruction will as a result
occur at distances from the ignition source (while still within the premixed volume).

Ideal Mixture
Every fuel has a ratio with its oxidizer at which its production of heat and reaction
velocity are at their maximum. This is called the ideal or stoichiometric mixture. It stoichiometric
is often close to the midpoint between the lean limit (LEL) and rich limit (UEL). The mixture ■ A reaction
total damage done by a gas or vapor/air mixture will be maximized, since the mixture in which the
reactants and products
amount of gaseous product and heat produced and the rate will all be at their respec- are chemically balanced.
tive maxima.

Rich Mixture
Consider now the “rich” mixture, where the escaping natural gas has mixed with the
room air to form a mixture of 15 percent gas in air, or consider that the mixture of gas
and room air is not uniform and there are pockets in which the gas concentration is
much higher than in other parts of the room. When the mixture is ignited, there is an
explosive “thump,” but there is insufficient air (oxygen) available for immediate and
total combustion. Much of the natural gas is not burned, and the speed and force of the
reaction are diminished. There is fuel left over and it has been heated. The combustion
products and remaining fuel expand when heated but cool immediately afterward, and
air is drawn into the partial vacuum that results. This mixing provides more oxygen to
combine with the remaining hot gas, and a rolling fire results as the turbulence caused
by the initial combustion mixes the remaining fuel and air. Note that because the reac-
tion takes place over a longer period of time than in the case of an ideal mixture, a sig-
nificant pressure wave may not develop, and the reaction may more accurately be
classified as a combustion than as a deflagration. A victim may generally describe the
event as, “I heard a ‘whoosh’ and was on fire.” Interestingly, a concussive “thump” will
be more apparent to a listener outside the room than to someone within. Destruction by
rich-mixture ignitions may far exceed that caused by theoretically perfect mixtures of gas
and air, because such explosions are followed by fire, and the fire may consume a build-
ing that otherwise would have been subjected only to a sudden expansion that might
move its walls but not destroy its contents.
It should be noted that the preceding discussion centered around natural gas (methane/
ethane) or LP gas. Other gases behave quite differently. Hydrogen and acetylene, for
instance, burn at much higher speeds than methane and exert much higher pressures.
Acetylene can also detonate under some conditions. The chemical properties of any poten-
tial fuel gas must be carefully considered when assessing a blast scene. (Detonations of
gaseous fuels will be discussed later.)

Lean Mixture
The “lean” mixture combusts differently than the others in an explosion. Because the
propagation rate for a gas/air mixture is greatest when the mixture is near its ideal or
stoichiometric ratio, the lean-mixture explosion may be heard or felt as sharper than
in the rich-mixture case. The force of a lean-mixture deflagration may be considerable,
and its noise is very startling. However, with insufficient fuel to develop maximum
speed, heat, and pressure during the blast or lingering, rolling post-blast fire, the dam-
age from a lean-mixture deflagration may be minor compared with that of the other
mixtures. No following fire is expected from a lean-mixture deflagration except for
filmy materials like sheer curtains that may be ignited (and then go on to set other

Chapter 12 Explosions and Explosive Combustion 495

 fuels alight). (See Figures 12-3a–g for an example.) Generally, although not in every
instance, an explosion of this type produces purely physical forces and not an estab-
lished fire.

FIGURE 12-3A Wood-frame residence looks undamaged at first FIGURE 12-3B Bottom of wall is bowed outward by
sight. Note the dead shrub and dying tree (despite drip irrigation on more than 12 cm (5 in.) at base. Note that windows
both) in foreground. This was an indication of possible underground (small, double-glazed, aluminum-framed) are still intact.
migration of natural gas. Courtesy of John D. DeHaan. Courtesy of John D. DeHaan.

FIGURE 12-3C Interior view of same wall shows the sill plate split FIGURE 12-3D Wall of bedroom opposite main explosion shows
at the bolt, and drywall and molding damaged. Courtesy of John D. outward deflection of nearly 7 cm (3 in.) at its center. No thermal
DeHaan. effects were noted in this bedroom. Courtesy of John D. DeHaan.

496 Chapter 12 Explosions and Explosive Combustion

FIGURE 12-3E Closet doors were blown off their tracks. Courtesy FIGURE 12-3F This interior wall was blown apart when
of John D. DeHaan. the light switch (far right) was thrown, as were other two
walls surrounding the shower stall pan (at far left). The
occupant was standing in the bathroom doorway (extreme
right) when she turned off the switch, and was unhurt.
Courtesy of John D. DeHaan.

FIGURE 12-3G Gas had migrated under concrete slab (apparently

following an abandoned sewer connection), percolated under the
pan, and filled all three adjacent walls. Ignition propagated through
all three walls, blowing drywall off both sides. Note loose soil under
shower pan and scorching on nearest studs. Courtesy of John D. DeHaan.

Chapter 12 Explosions and Explosive Combustion 497

 VAPORS AND VAPOR DENSITY
Not all explosions involve fuel gases. Many involve the combustion of vapors generated
by the evaporation of flammable liquids.

Vapor Density
As we discussed in Chapter 4, explosions that result from the ignition of vapors from
liquid fuels are much more likely to produce post-explosion fires than will those of gases.
The explosion itself produces heat to volatilize remaining liquid fuel and provides an
ignition source for the “new” vapors being generated. Gaseous fuels will be largely con-
sumed or physically displaced, leaving little for continuous flaming fire afterward (unless
there is a continuous source of new gaseous fuel, such as a broken supply pipe) (as seen
in Figure 12-4).
The vapor density of organic liquid vapors must also be considered in connection with
special circumstances of the environment. For example, machines are sometimes degreased
with hydrocarbon solvents. Not only is there possibly a high concentration of flammable
vapors in the area, but drains and service pits provide lower regions into which the vapors
can sink. If such vapors accumulate in such a depression, it is there that the hazardous con-
dition is first established. Any arc or flame in this vicinity can cause an explosion, even
though it may be separated from the working area by some distance. Drains, sewers, heat-
ing ducts, elevator shafts, tunnels, and utility shafts are all possible areas for such remote
ignition. The density of flammable liquid vapors can produce rapid flame spread through
the layer where the premixed fuel/air mixture is in the flammable range.
The previous discussion centered on natural gas deflagrations because natural gas is
lighter than air and mixes freely with it. Thus, it is a good example of the effects of var-
ious concentrations of fuel/air mixtures. Natural gas, however, is one of a few gases
(hydrogen, methane, carbon monoxide) that are lighter than air. It is important to note

FIGURE 12-4 Flame

plume from broken gas
line charred the interior
of the displaced wall and
nearby fence after a mas-
sive natural gas explosion.
Courtesy of John D. DeHaan.

498 Chapter 12 Explosions and Explosive Combustion

that all vapors generated by evaporation of ignitable liquids at ordinary temperatures are
heavier than air (at the same temperature). Pentane vapors, for example, are about 2.5
times as heavy as air. Most petroleum distillates are less volatile than pentane, and their
vapors are even heavier. Fresh automotive gasoline contains up to 8 percent n-butane and
its related isomers, as well as pentane, so its initial evaporation and ignition properties are
dominated by those of butane and pentane.4

Solvents
Most solvents (lacquer thinners, petroleum ether, acetone, and the like) have vapors like
those of gasoline that are heavier than air; their mixing properties in air must be thor-
oughly understood by the fire investigator. Dense vapors like these can be visualized as
having fluid properties like maple syrup. When released, the vapors form a pool a few cen-
timeters deep and then spread in a thick layer, seeping under doors, flowing down stairs,
and collecting in pools in low spots of the room or structure (see Figure 4-7). These vapors
are heavy enough to concentrate at the bottom of such low spots and there form a very
rich mixture. The vapors then mix slowly into overlying still air by diffusion, creating a
gradient of concentration from the too rich to the too lean. The combustion of such a gra-
dient will usually progress quickly, but a rolling fire may be sustained as the rich layers are
mixed by the turbulent combustion of the overlying layers. Ignition of gasoline vapors in
the house shown in Figures 12-5a–e produced enough force to blow window assemblies
from the building, but the following fire was limited to one room due to an excess of fuel
vapors and insufficient ventilation elsewhere (producing a “too rich” mixture).
Mechanical Mixing
If sufficient mixing of gases or vapors with air occurs before ignition, there can be a mas-
sive deflagration, as shown in Figure 12-2. An idealized layer-cake condition will exist
only, of course, in a space where there is no mechanical mixing of the vapor/air layers. This
mixing can be caused by open windows, doors, chimneys, machinery, or even people walk-
ing about. Such mechanical mixing will complicate the interpretation of the resulting fire.
Propane releases are very sensitive to thermal mixing. Solar heating onto or into a com-
partment can cause enough circulation in the room to mix a propane/air mixture through-
out. This may serve to dilute small amounts to concentrations below the LEL or create
ignitable mixtures throughout a room, despite the vapor density of pure propane.
In one case examined by one of the authors, the reconstruction of an explosion and
fire aboard a fishing vessel showed that the LP gases, which had leaked from a stove and

FIGURE 12-5A A gasoline vapor explosion severely damaged this FIGURE 12-5B Only the room where windows had blown out
abandoned house. Courtesy of Jamie Novak, Novak Investigations and St. Paul suffered fire damage. Courtesy of Jamie Novak, Novak Investigations and
Fire Dept. St. Paul Fire Dept.

Chapter 12 Explosions and Explosive Combustion 499

FIGURE 12-5C Gasoline had been poured throughout the house, FIGURE 12-5D Furnace ductwork was badly damaged by over-
but there was inadequate oxygen to allow full propagation. Courtesy pressure within. Courtesy of Jamie Novak, Novak Investigations and St. Paul
of Jamie Novak, Novak Investigations and St. Paul Fire Dept. Fire Dept.

FIGURE 12-5E In the

basement, the mixture
was so rich that raw
gasoline was still detected.
Gasoline had pooled
around the furnace and
water heater (possibly
ignited by the furnace).
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

accumulated in the bottom of the boat, were ignited only when the boat’s owner walked
about in them, stirring them up. Residues of the gas were detectable in the victim’s cloth-
ing (the pants of which were torn and scorched only below the knees) and in tissue sam-
ples from the lungs.
Draft Conditions and Pilot Lights
Draft conditions must be considered whenever remote points of ignition play a part in
the ignition of flammable vapors. The propagation of a layer of vapor along a horizon-
tal surface from a container or spill of flammable liquid can be predicted to occur at a
particular rate [on the order of 2–5 cm/s (0.8–2 in./s) for pentane at 20°C (68°F)] in still
air depending on the volatility of the liquid and the temperature of the surroundings.5
Even small drafts caused by fans, mechanical equipment, or the pilot light of a stove or
heater can cause a positive flow of vapors toward an ignition source at a much faster rate
than expected. A pilot light creates small volumes of hot gas, which rise because of the
reduced density of hot gases; cool air then flows into the base of the flame, giving rise to
a slow but definite flow of air toward the flame. If heavy vapors are being released, even
at some distance, this slow air flow will move the vapors toward the pilot and may result
in ignition. As discussed in Chapter 3, the layer of vapors lies close to the floor, and
if the combustion chamber is elevated, the vapors will not accumulate to an ignitable
500 Chapter 12 Explosions and Explosive Combustion
concentration. Pilot lights are particularly treacherous to the arsonist pouring flammable
liquids. They not only provide an open-flame ignition source but also draw the explosive
vapors toward it. Ignition of the main burner (or start-up of the fan) causes the flow to
occur much more quickly.

DEFLAGRATIONS
Although fires involving the vapors of flammable liquids can fall into the general descrip-
tive class of combustion—with rolling flame fronts and minimal pressures produced—
depending on circumstances of fuel, mixing, concentration, and confinement, they can
readily achieve the next, more violent class of oxidation, deflagration. For a pressure
effect to be produced, however, the deflagrating gas or vapor/air mixture must be con-
fined to some extent. This confinement normally takes the form of a rigid container or
even a room or structure. If very large quantities of gaseous materials are involved, suffi-
cient confinement may be produced by the inertia of the gas cloud itself. Although small
quantities of a vapor/air mixture may deflagrate when confined in a rigid container, very
large masses of the mixture must be present to provide a confinement effect in open air
to produce what is sometimes called an unconfined vapor cloud explosion (UVCE). This
open-air confinement effect is noticeable only in large clouds of airborne vapor above
very large spills of volatile liquids. This is a more common occurrence in refinery or indus-
trial premises where a large quantity of volatile liquid fuel is released quickly (often at
high temperature and pressure). This quickly produces a very large vapor cloud that can
then be ignited by any appropriate source in contact with it. This was the mechanism, for
instance, for the huge UVCE at Flixborough, UK (1974) when a large quantity of cyclo-
hexane was released from a failed pipe joint. The explosion destroyed the large plant and
killed 28 employees.6
One of the means of discriminating a deflagration from other violent forms of com-
bustion is based on the time required for it to occur. A deflagration may take from sev-
eral milliseconds to as much as a second to develop.7 As we shall see, the time scales for
detonations are much, much shorter. Similarly, the peak pressures developed by deflagra-
tions are many times lower than those developed by detonations. Naturally, these lower
pressures and slower rates of pressure rise have a different effect on their surroundings
than those from detonations. We shall treat these differences later on when the diagnos-
tic features of both mechanisms are discussed.
Gaseous fuels are susceptible to transitioning from combustion to deflagration to det-
onation. This is especially true if they are heated and pressurized in contact with enough
oxygen to support combustion. Heat and pressurization extend the explosive range, so
what is not an explosive mixture at room temperature may be explosive under the condi-
tions of actual use. As gases expand during combustion there is often a “pressure-piling”
action that raises pressures and temperatures ahead of the actual reaction front. The
geometry of the confining vessel is also critical to deflagration/detonation transitions. If
the mixture is confined in a pipe (or even long corridor) whose length (l) is significantly
greater than its width or diameter (D) (typically l> D  100; for some gases, it is as low as
l> D  10), the flame front stretches out due to frictional drag at its sides. This increases
the surface area of the flame front and increases the combustion rate. When this acceler-
ation reaches a certain point, the reaction transitions to a detonation.8 The detonation
velocities of some gas mixtures at room temperature and atmospheric pressure are shown
in Table 12-1. Note that all of them are above the 1,000 m/s (3,280 ft/s) threshold
described previously, even though their densities are close to that of air.
Interestingly, the detonation limits of such mixtures are not the same as the flamma-
bility/explosive limits discussed earlier, and the maximum velocities are not necessarily
achieved at a “central” or ideal concentration. Detonation velocities ranging from 1,500 m/s
to 2,000 m/s (between 4.2 and 50 percent) (4,920 to 6,560 ft/s) for acetylene in air were
Chapter 12 Explosions and Explosive Combustion 501
 TABLE 12-1 Detonation Velocities of Gas/Vapor Mixtures

GAS/VAPOR MIXTURE DETONATION VELOCITY

Acetylene/air 2,414 m/s (7,920 ft/s)
Acetylene/oxygen 2,716 m/s (8,911 ft/s)
Hydrogen/oxygen 2,821 m/s (9,255 ft/s)
Methane/oxygen 2,146 m/s (7,041 ft/s)
Methane/air 1,880 m/s (6,168 ft/s)
Pentane/air 1,680 m/s (5,512 ft/s)
Pure acetylene (98%) at 5 kg/cm2 pressure had a reported detonation velocity of 1,050 m/s (3,445 ft/s)

Source: B. Lewis and G. Von Elbe, Combustion, Flames and Explosions of Gases, 2nd ed. (New York: Academic Press,
1961), 534.

measured by Breton, depending on concentration.9 Pure (98 percent) acetylene under

pressure (5 kg/cm2) detonates at 1,050 m/s (3,445 ft/s). Hydrogen detonation limits in air
are reported to be 18.3 to 59 percent; carbon monoxide and hydrogen in air: 19 to 59
percent; and acetylene/air: 4.2 to 50 percent.10
The strength of the initiating source also can play a role in whether the reaction will
be a deflagration or detonation. The stronger the source, the more likely a detonation will
take place in a susceptible explosive or mixture.
backdraft ■ A defla-
grative explosion of
Backdraft
gases and smoke from One special case of deflagrations that many investigators do not always consider is that
an established fire that of backdraft or smoke explosions. The gases, vapors, and solids produced by a smoky fire
has depleted the oxy- constitute a vapor-state fuel. If supplied with enough oxygen and a source of ignition,
gen content of a room accumulations of smoke can deflagrate with sufficient energy to break windows, move
or structure, most often
initiated by introducing
light walls, and injure firefighters, producing pressures of 0.03 to 0.1 bar (0.5 to 1.5 psi).11
oxygen through ventila- The mechanisms responsible for smoke explosions are better understood today thanks to
tion or structure failure. recent research.12 See Figures 12-6a and b for an illustration of the explosive potential of

(a) (b)

FIGURES 12-6A AND B Front windows and side wall of garage blown out by smoke explosion observed by
arriving firefighters. Fire ignited against wood siding outside front door. Courtesy of Jamie Novak, Novak Investigations
and St. Paul Fire Dept.

502 Chapter 12 Explosions and Explosive Combustion

a backdraft. Depending on the combustion rate of the accumulated smoke, there may not
be sufficient overpressure to inflict physical damage. The expansion of the burning smoke
layer, however, may fill portions of the building that were clear moments before and force
smoke out door and window openings.

Dust Suspensions
The special case of a solid fuel in a diffuse-phase explosion must be mentioned—that is,
the suspension of dust or powdered fuel in air. Coal dust in mines, flour dust in granaries,
metallic dust in machine shops, sawdust in lumber mills, or any fuel in powdered form
can be dispersed in a cloud to form an explosive fuel/air mixture. Although we do not
normally think of a solid fuel deflagrating in a diffuse-phase mixture, it can happen. As
in gas- or vapor-based mixtures, there are ranges of concentration (typically measured in
grams of fuel per cubic meter of air) for each fuel that will support combustion. The
behavior of the ignited solid/air mixture parallels that described for rich, lean, and ideal
mixtures of gases and vapors. Unlike gases, however, there are restrictions on the particle
size for each fuel that will sustain combustion. Generally, the finer the dust, the more
likely it is to both (1) remain airborne in suitable concentrations in air and (2) combust
readily to contribute to a self-sustaining reaction. Organic dusts such as cornstarch, soy
flour, sugar and wheat starch are capable of creating maximum pressures of 8.5–9.9 bar
(850–990 kPa or 123–144 psi) under ideal conditions (fine particles, suitable concentra-
tions, strong ignition source).13 One major difference between gas/air deflagrations and
dust explosions is that accumulations of dust on surfaces are agitated by the initial defla-
gration, and this additional fuel will propagate the deflagration much longer, often gain-
ing considerable energy in later stages.14 More energetic ignition sources are frequently
required for dust suspensions than for gas/air mixtures.
It is thought that airborne dusts of pyrotechnic mixtures such as aluminum powder
and potassium perchlorate flash powders are responsible for the massive destruction
produced when illicit “fireworks” factories explode.
Turbulence caused by the initial combustion of fireworks can aerate such mixtures,
allowing almost instantaneous combustion throughout (because flame will then come
into contact with the fuel/oxidizer particles throughout the dust cloud). Such unstable
materials are capable (like acetylene and ethylene oxide) of transition to a detonation even
when not airborne dusts.

IGNITION
The hazard of open flames with regard to flammable fuel/air mixtures has been touched
on, but naturally other sources of heat can initiate such mixtures. It is important to
remember that no matter what other concentrations of vapors are present in the area,
the fuel mixture must be present at combustible concentrations in the immediate area
of the heat source to be ignited. With a fire, continued heating of a pyrolyzable fuel

CHEMICAL SAFETY BOARD (CSB)

The U.S. Chemical Safety and Hazard Investigation Board (called the CSB) is a nonregulatory, independent
federal agency charged with investigating industrial chemical accidents. Headquartered in Washington,
DC, the agency’s board members are appointed by the president and confirmed by the Senate.
The CSB conducts origin and cause investigations of chemical accidents, such as dust, flammable
vapor, gas, fire, and chemical explosions at fixed industrial facilities. The CSB does not issue fines or
citations but makes recommendations as to prevention measures to plants, regulatory agencies such as
the Occupational Safety and Health Administration (OSHA), the Environmental Protection Agency (EPA),
industry organizations, and labor groups.15

Chapter 12 Explosions and Explosive Combustion 503

 ILLICIT EXPLOSIVE DEVICES
The term fireworks is a serious misnomer when referring to “M-80” type devices (as shown in Figure 12-7).
With their explosive filler—1 to 20 g of highly reactive flash powder—and cardboard containers, they
constitute a significant risk from thermal and blast effects. The solid end plugs produce missile and frag-
mentation effects capable of causing injury and even death. They are dangerous and are controlled as
illicit explosive devices under federal law in the United States.

FIGURE 12-7 Illicit explosive devices

such as these “M-80s” are commonly
available. Courtesy of John D. DeHaan.

will create the conditions for the fire but not for the explosion. Any hot spot or region,
regardless of its size, can initiate an explosion of vapors if it meets the following criteria:

■ There has to be contact between the ignition source and the fuel.
■ At some point, the temperature of the source is well above the autoignition
temperature of the flammable material involved (i.e., it has enough energy).
■ At the ignition point (in time and space), the fuel/air mixture is within its explosive
range.
■ These conditions of contact last long enough for enough heat to be transferred to
the fuel to trigger a self-sustaining combustion.

These criteria apply to all flames; electric arcs, including those created by static elec-
tric discharge; and hot solids such as heated metal and incandescent sparks. Interestingly,
smoldering combustion, as on the tip of a lighted cigarette, is generally incapable of ini-
tiating an explosion in a fuel vapor/air mixture, as discussed in Chapter 6.
Unlike an open flame that produces a continuous source of high temperature, an electric
arc may be only of short duration. Either the vapors in contact with the arc are in their explo-
sive range and explode, or they are not, in which case nothing happens as a result of the arc.
A similar effect is produced when a small charge of high explosive is used to disperse gaso-
line from a plastic or glass container. The high explosive generates gases at very high temper-
atures (thousands of degrees), but they are of short duration, and the explosion pushes the
gasoline (fuel) and oxygen mixture away from the ignition source. More often than not, no
ignition occurs. If a metal container is used for the gasoline, however, the metal fragments
(shrapnel) are superheated by the initiating blast and frictional tearing and will often act as
point sources of ignition as they retain their heat longer than do the gases of the reaction.
504 Chapter 12 Explosions and Explosive Combustion
Condensed-Phase Explosions
Condensed-phase explosions involve solid (or sometimes liquid or gel) explosives in which
either the fuel and oxidizer are mechanically mixed or the monomolecular material itself
is capable of undergoing extremely rapid conversion. Explosions of higher energy than the
deflagrations described previously usually involve a condensed-phase or dense-phase
chemical, because detonation of a diffuse-phase system is very rare. (Such detonations may
occur in high-pressure or high-temperature gas systems or in systems involving especially
sensitive compounds such as acetylene or ethylene oxide.)
Oxidation reactions that fall into the detonation category differ greatly from combustion
and deflagration reactions in a number of ways, as graphically represented in Figure 12-8.
There are major differences, of course, in the reaction time and in the reaction volume of the
explosive reaction.
Compared with the milliseconds or seconds required for a deflagration, detonations
occur throughout the charge within microseconds. As will become apparent, the forces
generated and their effects on the surroundings will be different more because of this time
relation than because of the total force generated. The pre-blast reaction volume of a stick
of dynamite in a room is the small volume occupied by that stick, virtually a point source
from which all the explosive forces are generated as the post-blast gases occupy some
1,500 times the pre-blast volume. If the same room is filled with an explosive mixture of
vapors and air, the explosive volume is that of the entire room, which, upon combustion,
tries to occupy up to 8 times its original volume. In physical terms, this is the chief distinction
between diffuse- and condensed-phase explosions. (In chemical terms, of course, there are
many other differences, discussed later.) Even though the total forces produced (i.e., total
work done) may be similar, the specific effects of the two types of explosions will be quite
different. While the energy released by the complete deflagration of a quantity of gasoline
can be compared to that released by a detonation quantity of high explosives, the dam-
age potential of the two events is vastly different due to the time frames and dynamics
involved. The energy released by TNT detonating is on the order of 4.56 MJ/kg, while the
heat of combustion of methane is 55 MJ/kg and that of gasoline is 46 MJ/kg.16
Most condensed-phase high explosives require an initiating mechanical shock that
triggers a cascade effect of reactions throughout the reaction volume. It is thought that

104
Detonation

103
Pressure at origin (psi)

102

Deflagration
10

1 FIGURE 12-8 Phase

Combustion diagram showing ranges
of pressure and speed of
propagation for general
0
0 0.3 3 30 300 3,000 m/s states of explosive com-
0 1 10 100 1,000 10,000 ft/s bustion, deflagration,
and detonation.
Speed of propagation

Chapter 12 Explosions and Explosive Combustion 505

 materials that detonate have chemical bonds sensitive to energy applied in the form of
pressure and are disrupted by the transmittal of pressure or shock throughout the mass
of the explosive charge.
The characteristic feature of detonation is the shock wave that propagates the reaction
moving through the explosive at a speed greater than the speed of sound in that matrix.
Unlike the high-speed oxidations of gaseous explosive mixtures, detonations of a condensed-
phase explosive do not require confinement except when small quantities of material are
involved. Many sensitive materials will still detonate when unconfined, but not very effi-
ciently. The pressures produced will be low, and a considerable amount of the material will
remain undetonated in what is called a low-yield or nonideal explosion. Some materials will
detonate when heated to a critical temperature (which is often pressure dependent).
All explosives have a critical diameter, however. If a mass of explosive is too thin or too
small in cross section there will be no propagation. Such critical diameters are taken into
account when commercial preparations such as dynamite cartridges, grenades, or high-
explosive boosters are being designed but are ignored in many improvised criminal devices.

CHEMICAL AND PHYSICAL PROPERTIES

As you might expect, materials that explode or detonate have a different chemistry from
those that require the addition of an external source of oxygen to enable the reaction. The
molecules of materials that can detonate have the capability of shattering into fragments
when properly stimulated. As those molecular fragments recombine, they release large
quantities of heat in a fraction of the time it takes for combustion to take place.
Nitrogen compounds are probably the most common types of chemicals used in the
formulation of explosive materials. This is because many nitrogen compounds are basically
unstable from the standpoint of their compositional (potential) chemical energy. Nitrogen
is an element that resists chemical combination with other elements. This means that the
compounds of nitrogen tend suddenly to rearrange themselves to yield elemental nitrogen.
In doing so, they release their potential chemical energy as heat and light in an explosively
sudden manner. Oxygen is often combined with nitrogen in various carbon-containing
organic compounds in common explosives. In the explosion, nitrogen is released from its
chemical structure along with a great deal of energy, and the oxygen combines with carbon
compounds during the reaction to produce even more energy.
Materials that undergo explosive decomposition almost always contain one or more
characteristic chemical structures that are responsible for the sudden releases of great
energy. These structures include the following groups:
■ —NO2 and —NO3 in organic and inorganic nitrates
■ —N3 in inorganic and organic azides
■ —NX2, where X is a halogen
■ —ON—C — in fulminates
■ —OClO2 and —OClO3 in inorganic and organic chlorates and perchlorates,
respectively
■ —O—O— and —O—O—O— in inorganic and organic peroxides and ozonides,
respectively
■ —C ——
— C— in acetylene and metal acetylides
■ [M—C], a metal atom bonded with carbon in some organometallic compounds17

In addition to compounds that contain the preceding groupings, there is a wide variety of
mixtures containing both fuel and oxidizer that are subject to deflagration. These produce
generally low-energy deflagrating explosions but can produce severe damage when
encountered in improvised explosive devices. For instance, black powder is a mixture of
carbon, sulfur, and potassium nitrate (as the oxidizer), and flash powder is commonly a
mixture of aluminum powder and sulfur (as fuels) and a perchlorate as the oxidizer.
506 Chapter 12 Explosions and Explosive Combustion
 In addition to their similar chemistries, materials that explode or detonate have several
performance properties in common—they all produce large amounts of heat and when
properly initiated convert nearly completely to gaseous products (with the rare exception
of the metal acetylides that produce carbon and metal and no gaseous products). These
gases (and atmospheric gases nearby) in turn expand rapidly because of the heat generated.
Most explosive materials are sensitive to heat in some degree. Raising the temperature
will cause explosion or detonation, but unlike combustible vapors, explosive materials,
due to their chemistry, do not have precise ignition temperatures.18 The ignition tempera-
ture measured depends on the rate of rise of the temperature. The faster the temperature
rises, the lower the autoignition temperature appears to be.

Types and Characteristics of Explosives

In addition to being sensitive to heat, most explosives are also sensitive, to a greater or lesser
degree, to initiation by mechanical shock. Thus, the sensitivity of an explosive material to
both such effects is of interest to the fire and safety investigator. Based on the sensitivities
and oxidation processes of these materials, explosives are often classified as propellants or
low explosives or as high explosives. Explosions may also be characterized as high order low explosive ■ Any
or low order, as will be discussed. material designed to
function by deflagration.
Common explosives and their compositions are summarized in Table 12-2. A full
treatment of the formulation, use, and hazards of these high explosives is beyond the scope high explosive ■ Any
of this book. On the rare occasions when the fire investigator encounters such materials, material designed to
several texts in this area can be recommended.19 function by, and capable
of, detonation.
PROPELLANTS OR LOW EXPLOSIVES
Propellants or low explosives are combustible materials that carry both fuel and the nec-
essary oxygen for their combustion. Although both single-base smokeless powder (nitro-
cellulose) and black powder fall into this category, their reactions to heat and confinement
are somewhat different. Nitrocellulose propellants ignite at approximately 190°C (374°F)
and combust when burned in small quantities at normal atmospheric pressures.20 This oxi-
dation rate is pressure (confinement) dependent. Modest confining pressures elevate it to a
deflagration, which becomes extremely fast at the pressure developed inside a cartridge
when confined by a bullet. For example, the burning rate of Bullseye smokeless powder at
normal atmospheric pressure is 0.23 mm/s (0.009 in./s), at 34.5 bar (3.45 MPa; 500 psi)
its burning rate is 6 mm/s (0.24 in./s), while at 2,410 bar (241 MPa; 35,000 psi) (a typical
breech pressure inside a weapon), its burning rate is 300 mm/s (11.8 in./s).21 Double-
base smokeless powders contain nitroglycerin in addition to nitrocellulose and can be sen-
sitive to sympathetic (pressure) initiation and detonation if nitroglycerin is present in high
enough concentration. They do not detonate under most conditions. Black powder, on the
other hand, deflagrates even when burned in small quantities at atmospheric pressure and
can produce high-intensity blast effects when confined by only a moderate pressure (such
as the inertia of the surrounding charge itself). Black powder has a burning rate of 17 mm/s
(0.67 in./s) at atmospheric pressure, 30 mm/s (1.2 in./s) at 34.5 bar (3.45 MPa; 500 psi),
and only 62 mm/s (2.44 in./s) at 2,410 bar (241 MPa; 35,000 psi).22 It is not subject to
sympathetic (pressure) initiation and does not detonate at any time.
Both smokeless powder and black powder function as small-arms propellants when
they are ignited as thin flakes or small granules (as illustrated in Figure 12-9). The finer the
size, the faster they will burn to completion. Deflagrations can produce pressures of up to
2,760 bar (276 MPa; 40,000 psi) (solid explosives) with reaction velocities of up to 1,000 m/s
(3,280 ft/s) (depending on chemistry, confinement, and manner of initiation). Depending on
their means of ignition, geometry, and confinement, some materials, like double-base smoke-
less powder and potassium perchlorate/aluminum (KClO4/Al) flash powders, can detonate.
Under some conditions, then, they can be considered to be high explosives.
Chapter 12 Explosions and Explosive Combustion 507
 TABLE 12-2 Common Explosives

TYPE COMPOSITION CHARACTERISTICS

High Explosives
Primary/Main Charge
Straight dynamite Nitroglycerin (NG) Water resistant
Ethylene glycol dinitrate (EGDN) High strength
Sodium nitrate
Wood pulp
Gelatin dynamite NG/EGDN Water resistant
Nitrocellulose (gelling agent) High strength
Wood pulp
Sulfur (in some cases)
Ammonia dynamite NG/EGDN Less expensive
Ammonium nitrate (NH4NO3) Reduced water resistance
Sodium nitrate (NaNO3)
Wood pulp
Sulfur
Nitrostarch dynamite Nitrostarch Similar to nitrocellulose
Aluminum flakes No NG headaches
Does not separate or decompose
Military dynamite RDX Insensitive to shock, friction
Trinitrotoluene (TNT) Stable in storage
Cornstarch
Oil
TNT Trinitrotoluene Military uses
High brisance
Initiators
PETN Pentaerythritol tetranitrate High brisance
Also used in detonating cord and Semtex
RDX Cyclotrimethylenetrinitramine High brisance
Also used in detonating cord and Semtex
Blasting Agents Used in bulk or prepackaged form Main charge only
Slurries Ammonium nitrate or sodium nitrate, often Require booster
with fuels such as carbon or aluminum Can detonate
Water gels Ammonium nitrate Require booster
Sodium nitrate Can detonate
Sensitized with organic nitrates (e.g.,
monomethylamine nitrate)
In emulsion form
Low Explosives Commonly used in improvised explosive Deflagrate, but can produce detonation-
devices. Also used in explosive hardware like effects if adequately confined
and tool applications
Black powder Potassium nitrate Easily ignitable
Sulfur Filler in safety fuse
Charcoal Firearm propellant
Single-base smokeless Nitrocellulose Firearm propellant
powder
Double-base smokeless Nitrocellulose Firearm propellant
powder Nitroglycerin

508 Chapter 12 Explosions and Explosive Combustion

 FIGURE 12-9 Common
pipe bomb fillers. Upper
left: various smokeless
powders. Upper right:
flash powder. Bottom:
black powder. Courtesy
of Wayne Moorehead.

HIGH EXPLOSIVES
High explosive materials are intended to function by detonation and will detonate when
properly initiated by heat or shock. It is thought that the localized heating produced by
mechanical shock is sufficient to rupture some of their chemical bonds and initiate a chain
reaction in which heat from one bond disruption is quickly transmitted to surrounding
bonds in a cascade, which results in the explosion of the entire mass within a few
microseconds. As shown in Table 12-3, high explosive detonations have velocities of
1,000 to 9,000 m/s (3,280 to 29,500 ft/s) with maximum pressures (at the surface of the
explosive charge) of more than 68.9 bar (6,890 MPa; 1,000,000 psi). Such localized high
pressures result in a “seat” of most intense damage—often shattering or pulverizing nearby
surfaces, as shown in Figure 12-10.
Mohanty reported the densities and detonation velocities of common high explosives
to be as listed in Table 12-3.

Properties of Straight Dynamite

At one time, probably the most common high explosive was ordinary nitroglycerin or
straight dynamite. Straight dynamite is compounded with nitroglycerin (NG) and

TABLE 12-3 Densities and Detonation Velocities

EXPLOSIVE DENSITY (g/cm3) DETONATION VELOCITY

ANFO 0.84 4,600 m/s (15,090 ft/s)
Emulsion 1.20 5,200 m/s (17,060 ft/s)
Dynamite 1.45 5,400 m/s (17,720 ft/s)
TNT 1.64 6,930 m/s (22,740 ft/s)
PETN 1.67 8,000 m/s (26,250 ft/s)
RDX 1.77 8,700 m/s (28,540 ft/s)

Source: A. Beveridge, ed., Forensic Investigation of Explosions (London: Taylor & Francis, 1998), 22.

Chapter 12 Explosions and Explosive Combustion 509

FIGURE 12-10 Seat of
explosion for device using
1 gallon and 4 ounces
of C-4. Notice minimal
charring in the vicinity but
shattering and splintering
of nearby structural mate-
rials. See Figure 7-46 for
damage to the exterior
of the building. Courtesy
of Lamont “Monty” McGill,
Fire/Explosion Investigator.

ethylene glycol dinitrate (EGDN) soaked into a solid absorbent or filler such as saw-
dust. Such dynamites are very sensitive to shock, friction, degradation, separation
(weeping), and thermal ignition and are very rarely used today except for “ditching”
applications, where water resistance and sympathetic initiation are desirable features.
The strength of a straight dynamite is based on the actual percentage of NG present
(by weight).

Properties of Gelatin Dynamite

The next form of dynamite is gelatin dynamite, which uses a combination of NG dis-
solved in nitrocellulose. These formulations are water resistant and very energetic but are
expensive to make and prone to separation, “weeping” liquid nitroglycerin. These expen-
sive components (NG and nitrocellulose) are sometimes replaced (all or in part) with
ammonium nitrate or sodium nitrate. Such combinations are still effective high explosives
but are sensitive to water, which can render them unusable. When ammonium nitrate is
used, the product is called an ammonia or semigelatin dynamite. These ammonia dyna-
mites are rated in strength equivalent to a particular percentage of NG (i.e., 40 percent
ammonia dynamite does not contain 40 percent NG but contains the energy equal to
40 percent–strength straight dynamite).
Alternative High Explosives
Due to the expense and danger of making and storing nitroglycerin dynamites, they have
been largely replaced by water-gel dynamites that use organic nitrates as sensitizing agents
and ammonium nitrate as the main explosive ingredient. Today, nitroglycerin dynamites
are made only in small quantities for special applications. Some dynamites are made with
nitrostarch as the explosive agent. This dynamite contains no nitroglycerin, and the atten-
dant risks are absent. The water-gel explosives are based on formulations of ammonium
nitrate and organic nitrates such as monomethylamine nitrate or monoethanolamine in a
water-based gel or guar gum or similar agent. There are also slurry-type blasting agents
that contain ammonium or sodium nitrates and aluminum or coal tar powders as fuel.
Such materials are more stable than nitroglycerin but less sensitive, requiring a substan-
tial detonation to trigger them. These explosives degrade to inert compounds rather than
to more dangerous forms (e.g., straight dynamite “weeping” nitroglycerin). There are lim-
itations, however. Due to their physical form as an emulsion, suspension, or slurry con-
taining water, they are less sensitive at cold temperatures, are rendered unusable by
freezing, and cannot be used in wet applications.
510 Chapter 12 Explosions and Explosive Combustion
HIGH EXPLOSIVE CATEGORIES
High explosives may be further categorized as primary or initiating explosives and main
charge explosives. Initiators like mercury fulminate, lead azide, picric acid salts, and oth-
ers are highly sensitive to mechanical shock, friction, heat, and electricity and are com-
monly used in blasting caps and primers to initiate a larger explosion. Main charge
explosives include dynamite, TNT, PETN, RDX, and the like. When dispersed, main
charge explosives often simply burn, but in any mass at all (more than a few grams) these
materials can detonate. Their actual performance may depend on the size and shape of
the charge and the manner of initiation. High explosives are rated by factors related to
their intended application. Generally, commercial explosives are used in “pushing” or
“heaving” applications, while military explosives are usually needed for their “shatter-
ing” power. The effectiveness of such explosives is rated by their brisance or shattering brisance ■ The amount
power. Commercial high explosives are formulated in dilutions or mixtures of compo- of shattering effect that
can be produced by a
nents to control not only the energy released but this shatter strength or brisance, because
high explosive.
different applications require different explosive properties.

Blasting Agents
Blasting agents are explosive materials that generally require a high-explosive booster to
initiate a detonation. Ammonium nitrate/fuel oil (ANFO) and its variants are the most
common blasting agents. They are often sold in bulk rather than as individual units. See
Appendix G for a list of the materials considered by the U.S. Treasury Department; U.S.
Department of Justice; and the Bureau of Alcohol, Tobacco, Firearms and Explosives to
be explosives and blasting agents under federal law.

Binaries
There are a small number of high explosives called binaries made up of two components
that by themselves are not explosives. When mixed together binaries form a cap-sensitive
high explosive. Mixtures of nitromethane and ammonium nitrate or nitromethane and
organic amines have been marketed as binary explosives.23 Tannerite, a filler for explod-
ing firearms targets, is sold as a binary today. It consists of ammonium nitrate and ammo-
nium perchlorate mixture as one component and fine aluminum powder as the second.
Mixed together, they constitute a powerful, impact-sensitive high explosive.

Improvised Explosives
A number of high explosives can be synthesized by fairly simple chemical processes to
be used in improvised explosive devices (IEDs). These include nitrogen triiodide (NI3),
triacetone triperoxide (TATP), hexamethylene triperoxide diamine (HMTD), hydroxyl
ammonium nitrate (HAN), ammonium dinitro amide (ADN), and metallic azides, ful-
minates, and acetylides. These are powerful and sensitive primary explosives (initiated
by heat, friction, shock, or even static electricity) that would leave minimal detectable
residues in post-blast debris if properly made and functioned (high order).24 Other high
explosive blasting agents that have been easily synthesized and used in criminal bomb-
ings include ammonium nitrate/sugar and urea nitrate. These mixtures were estimated
to have detonation velocities of 3,730 m/s for urea nitrate and 3,560 m/s for ammo-
nium nitrate/sugar.25 Silver acetylide [Ag—C — — C—Ag] is an example of a rare category
—
of explosive that produces only solid residues (carbon and metallic silver) and no gas
during its detonation. The enormous heat produced, however, heats the surrounding
air to several thousand degrees, and the resulting expansion of gases exerts explosive
force. Despite its unstable and sensitive nature, TATP has become the most widely
used improvised explosive. It can be formulated using acetone, hydrogen peroxide, and
acid from easily obtainable consumer products (e.g., nail polish remover, hair bleaches,
and batteries). The processes are simple and can be carried out with a few jars, filters,
and ice.
Chapter 12 Explosions and Explosive Combustion 511
 COMPONENTS
High explosives usually require the use of initiators, time fuse, and boosters.
Blasting Caps
Since high-explosive main charges often require a detonation to initiate them, a small det-
onation source can be readily contained in a blasting cap, as shown in Figure 12-11. It
consists of an igniter compound, time delay, primary explosive and, typically, a high
explosive. An electric blasting cap has two leg wires and a bridge wire embedded in the
ignitor. Applying electric current causes the bridge wire to overheat quickly, igniting the
deflagrating material. A delay is often built in to increase the time between application of
current and initiation of the primary explosive (typically from 10 to 200 milliseconds).
The primary charge detonates, triggering the high explosive filler, which in turn, provides
the shock needed to cause the main charge to detonate. All this is contained within a
metal (often aluminum) sleeve or tube, about 5 mm (3/16 in.) in diameter and 40–65 mm
(1-1/2–3 in.) in length. Nonelectric blasting caps have a hollow end that is crimped over
the open end of a traditional safety fuse. There are today a variety of “shock tube” caps
that can be triggered only by the detonation of a fine coating of high explosive on the
inside of a small-diameter plastic tube. Such caps are safer to use than traditional caps,
since they cannot be triggered by stray electric currents or flame. The propagation of the
detonating coating is confined to the interior of the plastic tube.
Boosters
Boosters are small, specially formed high explosive charges (which may be “ice cube,” finger-
sized sleeves, or cylindrical in shape) made of TNT, PETN, or RDX. They are used as inter-
mediates between blasting caps and main charge explosives. High explosive detonating
cord is sometimes threaded through boosters or even used as substitute booster by knotting
it around or threading it through a main charge high explosive. Detonating cord is filled
with PETN or RDX high explosive and itself requires a blasting cap to initiate it.
Time Fuse
A variety of fuse types may be encountered by fire and explosion investigators. A com-
mon type is pyrotechnic or hobby fuse, in which a core of pyrotechnic powder is wrapped
with windings of fine string and coated with nitrocellulose lacquer. When ignited, it burns
with a visible exposed flame. Although intended to burn at rates of 12–25 mm/s (1/2–1
in./s), it can burn faster when crimped or crushed.
Safety fuse consists of a core of finely ground black powder protected by wrappings
of string, tar, and plastic. Such a fuse is ignited at one end, and a small spurt of flame is
emitted at the other end. There is no exposed flame anywhere along its length. It burns
at 80–100 cm/min (30–40 in./min), and due to its waterproofing layers, it can burn
underwater.

Aluminum alloy shell Ignition charge Priming charge

Open end for

insertion of Base charge
safety fuse PETN

DuPont No. 6 Commercial Nonelectric Blasting Cap

Rubber plug Ignition mix Bridgewire

FIGURE 12-11 Typical

blasting caps. Top: non- Shunt Leg wires Shell Primer charge Base charge
electric. Bottom: electric
DuPont No. 6 Commercial Electric Blasting Cap

512 Chapter 12 Explosions and Explosive Combustion

 Igniter cord consists of a perchlorate-based incendiary filler with a multiple spiral wrap
of fine nickel-chrome-iron wire. It burns very quickly and is rarely seen in improvised devices.
Improvised fuses can include string or paper impregnated with black powder or
pyrotechnic powders. Fuse materials removed from commercial fireworks such as quick
match (black powder core wrapped in layers of paper) are also encountered.

Firing Trains
The combination of initial heat source, fuse, blasting cap, booster, and main charge is often
called the firing train or “fuzing” system. High explosives generally require a small detona-
tion to initiate them but will not be initiated by a simple flame or electric arc. Time delays,
blasting caps, and boosters may not be present, depending on the explosive involved and the
intended use. It is often useful for the investigator to consider what evidence is present (or
should be sought) to fulfill the necessary roles: ignition source, delay, ignition, detonation.
The accidental explosion of a munitions ship in December 1917 in the harbor of Halifax
(Nova Scotia) is a good example of an accident with a complete firing train. Collision with
another ship caused the release and ignition of drums of benzene (as deck cargo). The fire
triggered the detonation of picric acid explosive, which in turn caused the detonation of the
main cargo of TNT. A total of 2,925 tons of high explosive disintegrated the ship and flat-
tened the center of Halifax, causing some 2,000 deaths and injuring 9,000.26

HIGH-ORDER/LOW-ORDER EXPLOSIONS
When one is describing the mechanism and effectiveness of an explosive process, there is
considerable confusion regarding the terms high order and low order. Only high explo-
sives can achieve detonation, the term referring to the propagation of a supersonic reac-
tion wave within the explosive charge. This shock wave actually causes the rupture of the
chemical bonds holding the explosive together. The entire mass of explosive converts to
gas and heat at supersonic speeds, causing extremely high temperatures and pressures as
well as shock, all of which bring about the shattering effect commonly associated with
such explosives. All other explosive mechanisms are deflagrations (by our definition),
since they propagate at subsonic speeds. The energy and heat produced in the initial
charge are transferred within the charge at lower speeds (often by particle-to-particle
transfer of thermal energy), and combustion of the individual fuel elements then proceeds
at the surface of each particle. Heat and gases are produced at a slower rate, and lower
pressures are produced. The external effects are less shattering than with a detonation,
but they can be substantial in terms of the total kinetic energy produced.
Confusion often arises when someone concludes that a high explosive used means
that a high-order (or high-yield) explosion took place. High explosives need not initiate
and propagate perfectly each time. This is particularly true with improvised explosive
devices. Commercial explosive charges and military munitions are carefully designed to
convert as completely to kinetic energy as possible so that the explosive effects are pre-
dictable. Such complete conversion is referred to as ideal, high-yield, or high-order deto-
nation. Improvised devices contain shapes and sizes of charges that cannot detonate
completely, even when high explosives are used. (Explosives have characteristic minimum
diameters below which detonation is incomplete.) If the initiating charge or device is inad-
equate in size (energy) or poorly placed in the main charge, detonation may not take place
at all. For this reason, even when traditional high explosives (TNT, nitroglycerine, PETN,
dynamite, RDX) are used, they may not detonate completely or at all. This is referred to
as low-yield, nonideal, or low-order detonation. Premixed gas or vapor fuel/air mixtures
at their ideal or stoichiometric mixture can be said to sometimes propagate high order
because they function at maximum efficiency with no residual starting material.
Low explosives (black powder, nitrocellulose) do not detonate even when confined
and initiated under ideal circumstances and, because of the inefficiency or ineffectiveness
of the thermal processes involved, do not explode “high order.” Double-base smokeless
Chapter 12 Explosions and Explosive Combustion 513
 powder, because it contains both a low explosive (nitrocellulose) and a high explosive
(nitroglycerin), can act as either low or high explosive, depending on its initiation and
confinement. When a powder such as Alliant Red Dot or Bullseye with a high nitroglycerin
content is used and is initiated with a suitable blasting cap, it can detonate, especially if
an improvised container of high bursting strength (such as a gas cylinder) is used, but will
rarely reach ideal or high-order conditions. In most improvised devices, the containment
is weak or the initiation is inadequate, and the resulting blast, while terribly destructive,
is not ideal (high order) and there will be intact, unreacted explosive scattered about. So
when asked, “Did such-and-such a material explode high order?” the correct answer is
usually, “It depends.” That determination often depends on careful assessment of thermal
and mechanical effects (indicators) at both macro and microscopic scale.

Mechanical Explosions
In a mechanical explosion, a container, vessel, or pipe bursts when internal gas or liquid
pressures exceed the tensile strength of the container. The resulting catastrophic release of
high-pressure gas meets our general definition of explosion and can produce blast effects
(pressure, shock, and fragmentation), which can damage or destroy vehicles and build-
ings. Because an exothermic chemical reaction is not involved, heat is not released (except
in the tearing or shattering of metals). If liquid is released at temperatures or pressures far
above ambient conditions, it vaporizes nearly instantaneously, and the rapidly expanding
vapors exert their destructive forces on nearby surfaces. An example of a massive mechan-
ical explosion involving steam is shown in Figure 12-12. Such mechanical explosions may
occur by themselves or before, during, or after fires. When associated with a fire, they
may be the cause of the fire, or simply add to its destruction. Once the general nature of
the explosion has been deduced, the search for pieces of the explosive device or the acci-
dental mechanism responsible, or for residues of the chemicals involved (if any), can be
undertaken more efficiently.
Mechanical explosions can produce blast effects that may be difficult to distinguish
from those of chemical explosions. In either case, fluids or gases at very high pressures

FIGURE 12-12 A massive

Northern-class steam
locomotive lost its boiler
assembly due to a
mechanical (steam)
explosion, probably
caused by flooding cold
water into a “dry” hot
boiler. The remains of the
boiler are on the far side
of the track. Courtesy of
John D. DeHaan.

514 Chapter 12 Explosions and Explosive Combustion

strike surfaces, distorting and sometimes shattering them to produce energetic secondary
missiles, or shrapnel. Cylinders of compressed gases, aerosol cans, storage tanks, hydraulic
lines, pipelines, and similar systems can be overpressurized due to mechanical failures of
compressors, relief valves, or regulators. Such containers may also be exposed to heat with
a resulting increase in internal pressure. When the pressures developed within the line or
vessel exceed its bursting strength, it will fail, and the pressurized fluid will be released,
potentially with catastrophic results. The escaping gases inflict shock or pressure effects
on nearby surfaces and may propel resulting fragments at high velocities. Mechanical
explosions can sometimes produce a localized seat of damage, similar to that produced by
condensed-phase explosives. Note that they are usually gaseous phenomena, but that changes
in chemical or physical state are not involved, and that heat is not produced (except heat
created by the rapid tearing of a metal container). The absence of fire or other thermal
damage is often a reliable indicator that a mechanical explosion occurred. When external
heating has been involved, it will be seen that heated metals have lower tensile strengths
than do those same metals at ordinary temperatures. An externally heated vessel, then, is
most likely to fail where it is heated. See Figure 12-13 for an example of a mechanical
explosion involving a medical oxygen cylinder that failed (where it was exposed to flames)
with tremendous concussive force, and see Figure 12-14 for an exploded fire extinguisher.
Propane cylinders often fail with a typical “fish-mouth” hole where exposed to fire that
produces little explosive effect. The location and orientation of that failure will direct the
“flow” of released gases, sometimes causing the container itself to become a projectile.

ACID, GAS, OR BOTTLE BOMBS

A variety of improvised explosive devices function in general by creating a gas, usually
non-flammable, by chemical reaction that causes the mechanical explosion of a confining

FIGURE 12-13 Medical

oxygen tank (aluminum
alloy) failed where it was
heated in an enveloping
fire. Courtesy of Detective
Mark Allen, Tempe Police
Department, Tempe, AZ.

Chapter 12 Explosions and Explosive Combustion 515

 FIGURE 12-14 Fire extinguisher casing after mechanical explosion caused by exposure to sustained fire.
Courtesy of Vic Massenkoff, Contra Costa County Fire Protection District.

vessel, most commonly a plastic soft-drink bottle. The reactions may produce CO2 with
little evolution of heat or may be strongly exothermic and produce steam as well as gas.
Such devices usually involve common household products, so they are a favorite of ado-
lescents. They can produce high pressures and toxic or caustic gases capable of inflicting
serious injuries as well as mechanical destruction.27 Some of the more common combina-
tions are shown in Table 12-4.
The mixtures that produce flammable gas (hydrogen or acetylene) can have a double
explosive effect—the mechanical failure of the confining container followed by the
deflagration of the turbulently mixed fuel/air mixture. This deflagration may be nearly

TABLE 12-4 Common Ingredients Found in Acid, Gas, and Bottle Bombs

COMBINATION PRODUCT OF REACTION

Dry ice and water Carbon dioxide
Alka-Seltzer and vinegar Carbon dioxide
Calcium carbide and water Acetylene gas
Sugar, dry bleach, and water Steam, bleach
Aluminum foil and Drano Hydrogen
Metal foil and hydrochloric acid Hydrogen
Toilet, pool, or spa treatment and alcohol/ammonia Chlorine gas or bromine gasa
or hydrogen peroxide
a
(Newest products contain chlorinated or brominated Also nitrogen trichloride or nitrogen tribro
hydantoins or chlorinated isocyanuric acids) mide explosives

516 Chapter 12 Explosions and Explosive Combustion

FIGURE 12-15 Melted and heat-distorted 2-L soda bottle is the remains of an acid–metal bottle bomb.
Residues of unreacted hydrochloric acid and metal are present. Courtesy of Wayne Moorehead, Forensic Consultant.

instantaneous if an open-flame ignition source (such as a lighted candle) is placed nearby,

or delayed if the mixture has to spread to find an ignition source.
The pressures produced by the failure of a soft-drink bottle can be on the order of
200 psi. Chemical hazards from residual acid, base, or bleach can cause injuries to persons
nearby. Due to the hazards of explosive, incendiary, and chemical injuries, states now con-
sider these devices to be explosive devices for prosecution.
The post-blast solid residues may be very limited in amount and subject to degra-
dation from exposure to water. As seen in Figure 12-15, the aluminum (or other metal
used) and unreacted acid or base are often found. The bottle used will have character-
istic deformation due to the heat produced and internal pressure. Dry-ice devices fail
with extreme cold fractures and no visible residues. Due to the unpredictable time inter-
vals between mixing and explosion, the maker of the device is sometimes among the
injured.

BLEVEs
If the original contents are liquid rather than gas, heating causes the liquid to reach a
temperature far above its normal boiling point, since the liquid would ordinarily shed
its excess heat as vapors to maintain its boiling point temperature. A failure of the ves-
sel results in a BLEVE (boiling-liquid, expanding-vapor explosion). The gases initially
involved are supplemented by the superheated liquid that turns to vapor immediately
on contact with normal atmospheric pressure. Such explosions can be extremely
destructive because of the energy contained in the liquid, even when the liquid and its
vapor are noncombustible. (BLEVEs involving 40-gallon water heaters have been
known to demolish houses.) BLEVEs involving superheated water or steam explosions
are notable for their absence of fire/thermal effects. When the contents are flammable,
the ensuing fires can be awesome (Big Loud Explosion, Very Exciting). Every explosion
investigator must be aware of the possibility of mechanical explosions and examine the
Chapter 12 Explosions and Explosive Combustion 517
 scene for remnants of gas cylinders, boilers, containers, or pipelines that could be the
origin. When a liquid storage vessel is exposed to an external fire, the liquid inside in
contact with the heated vessel wall absorbs the heat by conduction, while the wall in
contact with only gas or vapor is not cooled. This may result in a tide mark on the exte-
rior of the vessel that can tell the investigator what the level of liquid was in the vessel
during the fire or the position of the vessel during fire exposure. This also means a vessel
engulfed in fire will usually fail first where it is not in contact with liquid contents,
because that portion will reach its tensile failure temperature before the portion that is
in contact with liquid.28

Electrical Explosions
The passage of high currents through air, insulating oil, or undersized conductors can
cause very rapid heating, which, in turn, causes extremely rapid expansion of surround-
ing gases or liquids. The ultimate example of the explosive effects of arc-heated air is the
lightning bolt, which can involve current flows of over 100,000 A. In the case of oil (as
in oil-filled transformers or switchgear) the oil boils and turns to a vapor upon failure of
the vessel. This vapor may be ignited by the continuing electric arc to produce an exten-
sive fireball. Conductors can vaporize, producing a gaseous vapor moving at high speeds.
Such failures can cause flash burns as well as blast and pressure damage, as illustrated in
Figures 12-16a–c.

FIGURE 12-16A Electrical explosion caused when service FIGURE 12-16B Fire-damaged tools; door of cabinet on
person shorted out 10 kV at switch gear. Blast pattern on floor around corner. Courtesy of Paul Spencer, London Fire
wall. Courtesy of Paul Spencer, London Fire Brigade. Brigade.

518 Chapter 12 Explosions and Explosive Combustion

 FIGURE 12-16C Flash
pattern on insulators.
Serviceman fatally injured
by arc blast. Courtesy of
Paul Spencer, London Fire
Brigade.

Investigation of Explosions
Explosions can occur as the result of accidental mechanisms in materials in industrial
applications or processes, household mishaps, or the deliberate criminal act of setting off
a bomb. In rare instances, a bomb may be military ordnance that is being misused, but
much more often, a bombing involves an improvised explosive device (IED). The ATF
reported that there were an average of 576 bombings and attempted bombings reported
in the United States in 2004–2007,29 with some 59 percent being pipe, tubing, or gas
cylinder containers (pipe bombs). Smokeless powder, black powder, and flash powder
were the dominant explosive fillers used.30 (From 1993 to 1997, some 34 percent of the
38 devices used in explosive incidents were pipe bombs with various types of low explo-
sive fillers.)31 This trend does not appear to have changed in the years since, because such
materials are so readily available. A vast majority of the remaining devices were flamma-
ble liquid fillers in bottles (Molotov cocktails). Molotov cocktail ■
It is the responsibility of the investigator not to presume that every event is an acci- A breakable container
dent (or a bomb) but to treat each scene as a potential crime scene from the standpoint filled with flammable
liquid, usually thrown.
of scene security, preservation, and documentation. Evidence that an explosion is the It may be ignited by a
result of a particular accidental cause can be just as important as evidence that proves it flaming wick or by
was a criminal act, and requires the same standard of care. chemical means.
The examination of an explosion scene is generally carried out in the same manner
as a thorough and proper fire investigation and, in fact, is most often carried out as part
of the examination of the fire of which the explosion was a part. However, unlike at a
fire scene, it is most important that evidence or structural members not be moved prior
to careful examination, because the relationship between the various pieces may be crit-
ical to reconstructing the cause of the explosion, much more so than in fire evidence.
Chapter 12 Explosions and Explosive Combustion 519
 Witness interviews are especially important in explosion investigations to reveal what
someone was doing or where they were when the event occurred and what was observed.
Surveillance cameras are becoming more commonplace, and they have captured critical
events just before and during explosions.

THE SCENE SEARCH

A preliminary search of the scene will reveal the extent and direction of the spread of frag-
seat of explosion ■ ments (blast pattern) of the building, vehicle, device, or victim, as well as an apparent seat of
The area of most explosion. Scene overview (and photographic documentation) can best be accomplished
intense physical damage
from a fire department aerial platform. Figures 12-17a–d show how much more informa-
caused by high explosive
pressures and shock tion can be gained from an aerial view of a structure explosion compared with a ground-
waves in the vicinity level view of the same scene. The distance from the apparent seat to the farthest
of a solid or liquid discovered fragment should be multiplied by at least 1.5 (150 percent) to establish a pre-
explosive. liminary perimeter. This perimeter may have to be expanded as the search progresses and
reveals a specific directionality to the spread (as with pipe bombs, whose end caps may
be launched hundreds of meters away) but will suffice to protect most of the evidence.

FIGURE 12-17A A view of initial propagation of a structural propane FIGURE 12-17B Massive propagation (fireball) outside structure.
explosion from ground level. Courtesy of Jamie Novak, Novak Investigations Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.
and St. Paul Fire Dept.

FIGURE 12-17C High-energy shattering of large structure illustrates FIGURE 12-17D Aerial view of same scene records the distribution
the power of a deflagration at near-ideal concentration. Courtesy of of fragments (note cruciform distribution perpendicular to the four
Jamie Novak, Novak Investigations and St. Paul Fire Dept. main walls). Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept.

520 Chapter 12 Explosions and Explosive Combustion

 (a) (b)

FIGURES 12-18A AND B Survey photos document debris field of this bombing. The rear doors of the ambulance
were found nearly 65 m (200 ft) from the vehicle (at the wire fence in a). Courtesy of John D. DeHaan.

Security of an explosion scene is paramount. No one but the searchers should be allowed
access during the early stages. Figures 12-18a and b show an example of a large debris field
from a vehicle bombing. Unlike evidence at fire scenes or ordinary crime scenes, where
the evidence is readily visible, explosion evidence may be tiny fragments of paper, wire,
metal, or wood that are easily overlooked and may stick, unnoticed, to the shoes of any-
one in the scene and thus be carried off and lost forever.
The perimeter should be clearly marked with banner tape, rope, or barricades and
closely monitored to ensure that only authorized personnel essential to the search are
admitted. (A scene entry log such as that in Appendix I may be useful in controlling the
number of people wandering through a scene.) The perimeter can be expanded if search-
ing and evaluation indicates there is some directionality to the distribution of fragments.
Deflagrations, especially of diffuse-phase mixtures, tend to produce omnidirectional dis-
tribution, but devices like pipe bombs can produce directional effects, as pipe caps are
often expelled great distances along the axis of the device.
Documentation includes thorough photographic coverage of the scene as well as dia-
grams (plan views) showing locations and distances of debris. This was done in the past
with tape measures and protractors, and more recently with Total Station surveying systems.
The advent of laser scanning systems that can accurately document locations up to 500 m
(1,600 ft) in day or night conditions is likely to set new standards of accuracy and ease of
use.32 See Figures 12-19a and b for an example of the usefulness of such scanning technology.
The size of the search sectors will be determined by the size of the scene and the number
of searchers. They should be small [less than 0.4 m2 (4 ft2)] near the seat, becoming larger
Chapter 12 Explosions and Explosive Combustion 521
 (a) (b)

FIGURES 12-19A AND B Documentation of an explosion scene using a Leica C10 laser scanner allows
production of aerial and eye-level views. Courtesy of Leica Geosystems and Precision Simulations.

[up to 10 m2 (100 ft2)] toward the margins of the search area where the debris is lightly
scattered. The sectors can be marked with chalk, string, or rope with stakes, or with tape
stuck on the ground and must be numbered for later reference.
Once the scene is secured and the sectors marked, the search can begin, with a
perimeter search and photographic documentation, just as in a fire scene. The size, shape,
and conditions of the scene and the number of qualified searchers available will determine
the organization of the search. Grid, sector, and spiral search patterns have all been used
with success. The search should always start from the outside perimeter and work inward.
This allows better preservation of the critical explosion seat and improved protection
against souvenir hunters. The primary goal of the scene search and documentation is the
identification of displaced debris or fragments and measurement of their displacement
(distance from starting position).

The Four Rs
The scene search can be thought of as the first step in the investigative process, which may
be remembered as the “Four Rs”: Recognition, Recovery, Reassembly, and Reconstruction.
Recognition at the scene includes assessment of the blast damage (location, intensity,
direction) as well as recognition of pieces of physical evidence to be used in the later
stages. Physical evidence of an accidental situation or mechanism that brought about an
explosion can be just as important as the evidence of a deliberate act of bombing. The
training and experience of examiners or searchers is critical to the investigation, because
important evidence can be small and appear inconsequential. Such evidence is easily over-
looked by untrained searchers.
Recovery involves not only the physical process of picking up evidence but its docu-
mentation and preservation as well. This process must take into account the fact that
fragmentation ■ The explosions produce blast, pressure, fragmentation, and thermal effects. Reassembly of
fast-moving solid pieces the device or mechanism may take place within the laboratory or at the scene but is a nec-
created by an explosion.
essary test to see whether the entire chain of circumstances is accounted for. Finally, the
Primary fragmentation
is that of the explosive scenario must be reconstructed. What was present? What was the initiator? How did it
container itself; second- happen? Who was present?
ary fragmentation is
that of the target shat- The Four Cs
tered by an explosion. Just establishing that there was an explosion of a particular kind of material (gas, liquid,
or solid) is usually not adequate. One way to remember all the necessary elements is the
522 Chapter 12 Explosions and Explosive Combustion
“Four Cs”: Container, Concealment, Content, and Connections. While these are most appli-
cable to criminal bombings, they are useful even for assessing accidental explosions.
Container What evidence is there for the means of containment of gases, vapors, or low
explosives sufficient for the blast effects we see? This may be fragments of pipe, furnace,
wall, or vehicle. Fragments of the device are called primary fragmentation; fragments of
everything else shattered by the blast are called secondary fragmentation. High explo-
sives, of course, require no container to develop their maximum explosive effect. The
absence of any container, then, may be construed to imply that a high explosive was used.
Blast effects must corroborate this hypothesis.
Concealment Bombs are often concealed in some manner to prevent their premature
discovery. Are there fragments of wrapping, container, or other mode of concealment? Is
the location of the blast seat such that an explosive device could have been concealed
there? Is there an intended target from whom the device had to be concealed? The con-
cealing container does not form the actual containment of the explosive charge and can
be anything from a paper bag or suitcase to a vehicle. The question is: Did this item
belong at the scene naturally or was it brought here?
Content What was the actual explosive material? In improvised devices or accidental
explosions, the explosive is never totally consumed. Residues may be visible to the naked
eye and be easily found or they may require sophisticated laboratory analysis, but they
can be found if the investigator thinks to recover debris that was close to the initiating
explosive charge. In accidental explosions, gases, vapors, or solids can accumulate from
manufacturing processes or poor housekeeping, or inadequate safety measures may bring
about destruction of normal stock. Interviews and review of fire inspections or previous
accidents may reveal the content of potential explosive products.
Connections Finally, connections must be examined—not just electrical wires, but any
physical arrangement or sequence of events that could initiate an explosion. How did it
start? In bombings, such considerations include where the device was placed (vehicle,
dwelling, business), how it was triggered (movement, timer, on-command, etc.), and what
the sequence was from power source to initiator to main charge (sometimes called the fir-
ing train). The answers to such questions help reconstruct the entire event, including the
intent of the bomber—to destroy, to intimidate, or to kill? In accidental situations, the
same questions help establish cause and effect so that proper liability can be determined,
and more important, similar accidents can be avoided in the future. In accidental electri-
cal explosions, an appropriate source of very high amperage electrical power must be
identified, along with a conductive path. It must be remembered that some portions of the
path may be obliterated by movement, fire damage, or the passage of current itself.

SPEED AND FORCE OF REACTION

When it comes to interpreting explosion damage at a fire scene, it is important to remem-
ber the differences between deflagrations and detonations as they have been described here.
The differences in speed of reaction and in the forces and temperatures generated are
responsible for characteristic effects on a structure or vehicle. In determining whether a com-
bustion, deflagration, or a detonation took place, one guideline to keep in mind is the smaller
the pieces and the farther they are thrown, the higher the energy of explosion—combustion,
deflagration, detonation. The time required for a deflagrative explosion is on the order of
hundredths of a second as compared with that of a detonation, which is on the order of
thousandths or millionths of a second. The relative slowness of a deflagration is important
in distinguishing its effects on targets at close range, because they will be subjected to a more
or less progressive pressure pulse that is subsonic with respect to the target material. This
gives the material a chance to spread the pressure out, resulting in less shattering and more
pushing of larger pieces. The abrupt pressure or shock wave from a detonation is supersonic
Chapter 12 Explosions and Explosive Combustion 523
FIGURE 12-20A Localized shattering of wooden FIGURE 12-20B Results of deflagration of 1 gallon of gasoline (vaporized by a
cabinets by 230 g (0.5) lb of military dynamite. Courtesy small explosive charge). The roof was lifted 1.2 m (4 ft) into the air, and unsupported
of Lamont “Monty“ McGill, Fire/Explosion Investigator. walls were flattened. Note the surface scorching of the painted wall on right.
Courtesy of Lamont “Monty” McGill, Fire/Explosion Investigator.

at close range; the material cannot respond and is shattered rather than pushed. This intense
shattering effect gives rise to the localized damage called the “seat.” As energy dissipates,
the damage becomes less severe. The shock wave from a detonation compresses air into a
very thin layer that does a great deal of the shattering. It is followed by a high-pressure front
that does a great deal of the physical movement. The effects of an explosion may be cate-
gorized as blast wave, overpressure, missile, and thermal.
Figures 12-20a and b show the difference in effect between a deflagration of gasoline
vapor and the detonation of military dynamite in a wood-frame structure. The pressures
produced by a deflagration are very low in the vicinity of the ignition source and grow as
the flame front expands through the premixed fuel/air mixture. The lower, more uniform
pressures of a deflagration usually do not produce a localized seat of more intense dam-
age. Published tests by DeHaan have shown that pressures produced by a deflagration of
a shallow hexane vapor layer (0.1–0.2 m deep) (4–8 in.) on the floor of a compartment
(3.6  2.4  2.4 m in size) (10  8  8 ft) equilibrate at all surfaces of the compartment
within a few milliseconds of one another. This pressure was seen to grow over a period
of about 100 milliseconds until the failure pressure of the vent panel (5 to 6 kPa; 0.75 psi)
was reached.33 Such equalized pressure would cause the failure of the weakest part of the
compartment (in this case, the vent panel). In a real-world building this weakest part may
be the roof, the windows, the bottoms of walls weakened by decay, or other components.
The stronger the structure, the greater the confinement, and the higher the pressures that
can be developed by a deflagration, from less than 7 kPa (1 psi) to 700 kPa (100 psi) or
more, depending on the chemistry and quantity of fuel gas/air mixture present.34
524 Chapter 12 Explosions and Explosive Combustion
 (a) (b)

FIGURES 12-21A AND B Natural gas explosion (see also Figure 4-19) in an apartment tower block (second
floor from top). The blast wave reflected between facing tower blocks. Notice the broken windows on alternate
floors of both buildings. Courtesy of Fire Investigation Unit, Metropolitan Police, London, England.

Figure 12-3 showed the result of a natural gas explosion in a wood-frame residence where
the toeplates were split and walls deflected out at their bases 4 to 12 cm (2 to 5 in.), but the
small double-glazed, aluminum-framed windows remained intact. In a detonation, the confine-
ment of the structure plays no significant role in the pressures developed. Any pressure pulses
can be reflected from large flat surfaces and bounce back to induce damage around corners or
over or under barriers. The more energetic the explosion, the more pronounced such effects can
be. Reflected pressure pulses can add to the originating pulse and create even more damage in
the room of origin. Figure 12-21 is an example of a pressure pulse reflected back and forth
between two high-rise buildings that broke windows on alternate floors.
Note that when the pressure or shock wave is transmitted through air, it drops off in
a predictably rapid fashion, as reflected in Figure 12-22. Finally, the temperatures pro-
duced in a deflagration would not be expected to exceed 1,500°C (2,700°F) but can reach
several thousand degrees in a detonation. As one would expect, these differences in speed,
pressure, and temperature produce different effects on the target materials. As we saw in
Figure 12-17, a near-stoichiometric mixture of propane ignited below grade vented
upward, completely shattering the building above. An analyst familiar with the differ-
ences can accurately estimate the type of explosion, and often the location of the ignition,
that took place by examining fragments of the target and their distribution. Beveridge and
others have published excellent descriptions of the characteristic features.35
Explosion Damage
The pressures produced by a deflagration are usually fairly low and do not generally shatter
or pulverize, but they can cause serious structural damage to a building. Exceptions are
shown in Figures 12-17c and 12-23. The pressure wave moved at subsonic speed and
Chapter 12 Explosions and Explosive Combustion 525
 FIGURE 12-22 Pressures 105
as a function of distance
for the detonation of
100 lb of explosive and a
typical large deflagration. Detonation
Courtesy of Dr. J. H. Burgoyne
and Partners, London, England.
104

Pressure (psi)
103

102

Deflagration

0
0 10 100
Distance (ft)

FIGURE 12-23 Massive deflagration of near-ideal mixture of LP gas and air produced extensive fragmentation
and high velocities. Courtesy of Denise DeMars, Streich DeMars Inc., and Jamie Novak, Novak Investigations and St. Paul Fire Dept.

526 Chapter 12 Explosions and Explosive Combustion

retained its destructive force longer as it moved away from the source, where an ideal
mixture of LP gas and air was achieved. Because the rate of reaction of a deflagrating
fuel/air mixture increases with the expanding surface area of the expanding flame front,
the force exerted increases as the distance from the ignition source increases. For this rea-
son, the area of most destruction of a fuel/air deflagration will usually not be in the vicin-
ity of the ignition source but rather some distance away from it. This pattern makes
establishing the actual ignition source very difficult. Once the limit of the premixed vol-
ume is reached, the energy falls off with distance. The mixing and turbulence caused by the
passage of the flame front around furniture and through doorways can dramatically
increase the surface area of the flame front. Pre-ignition heating and pressurization of the
fuel/air mixture in adjoining rooms can also result in more energetic deflagration than in
the room of ignition.36 This in turn increases the damage done by the extensions of the
blast into areas away from the ignition source. The deflagration pressure wave can be
vented by the loss of doors or windows in a structure, minimizing damage to the walls.
The normal appearance of such a scene, then, is substantial destruction over a wide area
with large fragments and projectiles. (See Table 12-5 for typical blast effects.)
In contrast, a detonation, with its peak pressures in the tens of thousands of pounds
per square inch, shatters and pulverizes materials in its immediate vicinity and creates a
shock wave with a massive positive-pressure front with a negative-pressure wave or zone
right behind it. The negative-pressure wave creates a partial vacuum that is filled by air
rushing back in toward the explosion seat. This current can be strong enough to move light
debris or even topple weakened masonry walls in a direction opposite the original blast.
Windows can be “sucked out” of a wall as a blast wave moves parallel to it. The negative-
pressure pulse that follows the positive-pressure wave from a confined deflagration is low
in energy. The hexane vapor tests described earlier revealed a negative-pressure pulse about
one third as strong as the positive-pressure peak.37 This may be enough to move draperies
back through broken windows, pull pictures off walls, pull unsupported doors into a
room, or cause further collapse of structural members weakened by the positive pressure.
The detonation pressure wave initially moves at supersonic speeds but very quickly dis-
sipates, making its damage highly localized. Because of its supersonic speed, the pressure
wave is not subject to venting and can destroy a nearby wall whether it has windows and
doors or not. The normal appearance of a detonation scene, then, is highly localized destruc-
tion with pulverization, shattering, and a large number of very small pieces, some of which
will be projected large distances away. Keep in mind that all explosive detonations produce
what becomes a low-energy pressure wave at large distances from the source.
The distance that window glass fragments are thrown from their origin is directly
proportional to the pressure on the glass pane at the time of failure. Plain glass is thrown
50–60 m (180–200 ft) by a pressure of 27.6–34.5 kPa (4–5 psi), 30–35 m (100–120 ft)
by pressures on the order of 13.8 kPa (2 psi), and 12–15 m (40–50 ft) by overpressures
of 4.83–6.89 kPa (0.7–1 psi).38 Wire-reinforced or plastic-film-laminated glass, which has
higher failure pressures, will be thrown shorter distances by the same pressures. High
explosives have been used to vaporize flammable liquids in combination devices used as
antipersonnel weapons. Such scenes bear post-fire evidence of both types of explosions.
In the immediate vicinity of the device, the shattering effects of the high explosive pre-
dominate, as shown in Figure 12-23. Elsewhere in the structure the effects of the defla-
grating gasoline vapor are the most apparent, as was seen in Figure 7-46. The pressures
produced by detonations often pulverize nearby surfaces, creating a crater or seat, whose
depth and dimensions reflect the size and sometimes the shape of the explosive charge.
Such seats are not produced by diffuse-phase explosions.39

Pre-fire and Trans-fire Explosions

Because explosions can occur before a fire (as the means of ignition) or during a fire, the
presence or absence of pyrolysis products on debris ejected from a building can reveal
Chapter 12 Explosions and Explosive Combustion 527
 TABLE 12-5 Structural Damage Produced by Blast

PRESSURE (psi) DAMAGE

0.02 Annoying noise (137 dB), of low frequency (10–15 cps)
0.03 Occasional breaking of large glass windows already under strain
0.04 Loud noise (143 dB); sonic-boom glass failure
0.1 Breakage of windows (small) under strain
0.15 Typical pressure for glass failure
0.3 “Safe distance” (no serious damage below this value with a probability of .05)
Missile limit
Some damage to house ceilings; 10 percent of window glass broken
0.4 Limited minor structural damage
0.5–1.0 Large and small windows usually shattered; occasional damage to window frames
0.7 Minor damage to house structures
1.0 Partial demolition of houses; made uninhabitable
1–2 Corrugated asbestos shattered
Corrugated steel or aluminum panels and fastenings fail, followed by buckling
Wood panels (standard housing) and fastenings fail; panels blown in
Persons knocked down
1.3 Steel frame of clad building slightly distorted
2 Partial collapse of walls and roofs of houses
2–3 Concrete or cinderblock walls, not reinforced, shattered
2.3 Lower limit of serious structural damage
2.5 50 percent destruction of brickwork of house
3 Heavy machines (weight: 1,400 kg; 3,000 lb) in industrial buildings suffer little damage
Steel-frame buildings distorted and pulled away from foundations
3–4 Frameless, self-framing steel panel buildings demolished
Rupture of oil storage tanks
4 Cladding of light industrial buildings ruptured
5 Wooden utility poles (telegraph, etc.) snapped
Tall hydraulic press (weight: 18,000 kg; 40,000 lb) in building slightly damaged
5–7 Nearly complete destruction of houses
5.1–14.5 Eardrum rupture
7 Loaded railroad cars overturned
5.8–8.7 Reinforced concrete structures severely damaged
7–8 Brick panels, 20–30 cm (8–12 in.) thick, not reinforced, fail by shearing or flexure
9 Loaded railroad boxcars completely demolished
10 Probable total destruction of buildings
Heavy (3,200 kg; 7,000 lb) machine tools moved and badly damaged
Very heavy (5,500 kg; 12,000 lb) machine tools survive
29–72.5 Lung damage
102–218 Lethality
300 Limit of crater lip

Sources: V. J. Clancey, “Explosion Damage,” paper presented at Sixth International Meeting of Forensic Sciences, Edinburgh, 1972; G. Kinney
and K. Graham, Explosive Shocks in Air (New York: Springer-Verlag, 1985).

528 Chapter 12 Explosions and Explosive Combustion

FIGURE 12-24 The underside of plywood roofing panels (foreground) bear “flame-washed” effect from
fuel-rich natural gas explosion in the attic of the structure. Courtesy of John D. DeHaan.

whether the fire was underway before the explosion. Smoke explosions or BLEVEs
involving ignitable liquids occurring during a fire will often produce ejected debris that
show thermal effects or combustion residues, while debris from deflagrations or detona-
tions that initiate a fire usually have little or no such residues. If it is energetic enough
to disrupt the ceiling or roof of the confining structure, a natural gas explosion will cause
the combustion and release of the accumulated gas (as was seen in Figure 12-2). This
results in minimal post-blast burning except in the vicinity of the source of the gas.
Although the pressures of a deflagration inside a structure equilibrate at the speed
of sound and, as a result, usually produce no localized blast effects, the pre-explosion
distribution of fuel gas or vapor may produce localized thermal effects as the overrich
layers or pockets of fuel burn off. Scorching or “flame-washed” effects may be caused
by fuel/air deflagrations in wood structures (see Figure 12-24.) The distribution of melt-
ing or scorching of materials (particularly thin, low-mass materials like paper or plastic)
may reveal whether the pre-blast accumulation layer of gas or vapor was lighter or heavier
than air (see Figure 4-9). As debris is documented and collected, its appearance should
be noted.
When flammable liquid vapors in a closed room are ignited, the flame propagates
only through the vapor layer near floor level, as observed numerous times during tests by
DeHaan.40 Ide demonstrated extension of flames out the lower portion of a door open-
ing when gasoline vapors were ignited in the adjoining room. These flames charred the
lower legs of trousers worn by a mannequin placed in the doorway.41 Observation and
documentation of the distribution of thermal effects are critical to evaluation of the pre-
blast distribution of vapors or gases. See Figures 12-25a and b for an example of thermal
effects from a vapor explosion where liquid LPG was released into a room.
Chapter 12 Explosions and Explosive Combustion 529
 (a) (b)

FIGURES 12-25A AND B Ignition of commercial LP gas released into this structure produced floor-level
charring of vinyl tile and walls. Notice pronounced irregular charred areas and limited extension to baseboards
and some walls. Courtesy of Senior Special Agent Steven Bauer, ATF (retired).

SCENE EVALUATION AND HYPOTHESIS FORMATION

Once the preliminary evaluation of the explosion debris and thermal patterns is com-
pleted, the investigator will be able to direct the search in certain areas with specific types
of evidence in mind. The search of an explosion scene can be inward spiral, sector search,
strip search, or grid search. Searches of larger scenes will be aided by laying out grid lines
with string or rope and processing grid sectors in sequence. This is especially useful if
search personnel are limited in number.
A detailed sketch of the scene is most useful at this stage of the investigation. Because
items of evidence can be described on the sketch, along with measurements of their dis-
tance from a reference point and a reconstruction of their line of flight, such a sketch can
make relationships between those items apparent. If properly done, the sketch will show
how the paths tend to intersect at one area. This indicates that the seat of the explosion
(or the point of ignition of a deflagrating fuel/air mixture) was probably in that area and
allows the investigator to focus the search for pieces of the device (or ignition source)
itself in that room or area. The distribution, distance, and trajectory of window glass frag-
ments are especially important, since windows are usually broken in an explosion (see
Figure 12-26). A sample of the glass should be preserved or at least documented as to
thickness and type (safety, plate, plain, wired, etc.). If the distance, thickness, and type are
known, the pressure produced inside the structure can be estimated.

EVIDENCE RECOVERY
In spite of the ultrahigh temperatures and pressures developed in a high-order detonation,
some pieces of the device, however small, will remain in an identifiable form. Safety fuse
is particularly resistant to blast effects and will survive, sometimes even with a remainder
of the blasting cap still attached. Due to the frequency with which pipe bombs are used,
the examination of a scene should include an examination of the debris for any broken
bottles, pressure cylinders, or fragments of pipe or pipe cap found in the debris. The frag-
mentation of pipe, when used as a pipe bomb, follows a very predictable pattern. The
larger the total number of fragments and the smaller the fragments, the higher the energy
of the explosive filler.42 This information can be useful in quickly establishing the nature
of the device and explosive filler and the type of other evidentiary materials that should be
searched for.
530 Chapter 12 Explosions and Explosive Combustion
 FIGURE 12-26 Distribu-
tion of glass fragments
indicates the gasoline
vapor deflagration was
initiated in the room in
the lower right corner
Sidewalk of the structure.
Door Patio

Window Door
(closed)

Window

Small fragments of metal or plastic pipe, wire, or cardboard may easily be overlooked
by inexperienced searchers or compromised by being stepped on. Explosion scenes should
be searched grid-by-grid (preferably by two different searchers), and any item or debris
that looks out of place photographed and recovered for later evaluation. Some examiners
prefer to have each grid swept up after the visual search and all materials bagged for later
examination under controlled conditions. Sifting screens can be used to sift for fragments,
but unlike fire scene evidence, the debris must be sifted dry, since many of the post-blast
residues of low explosives are water soluble and will be lost if the debris is wet sifted. The
clothing and the bodies of any victims (living or deceased) must be carefully examined for
the presence of fragments (either primary or secondary). The patterns of thermal or
impact injury to the clothing and skin may also be critical in reconstructing the event
(whether the explosion was an accident or a bombing) and must be documented by pho-
tograph and diagram.43 Medical treatment can quickly alter the damage patterns of
injuries, so documentation and recovery of evidence from victims must take place as soon
as possible. X-rays are irreplaceable in assisting the search for fragments, but it should be
remembered that PVC and other plastics have the same physical texture and X-ray den-
sity as cartilage or bone in the body and can very easily be overlooked during X-ray or
visual examination of the wounds. (Whitaker et al. reported velocities of PVC pipe nip-
ple fragments to be as high as 290 km/h (465 mi/h; 80 m/s) and capable of inducing
injuries despite their low mass.44
Low explosives like black powder or smokeless powder are not completely con-
sumed when used in improvised devices (unless there has been a significant post-blast
fire), and traces will often adhere to larger pieces of the containers (such as the nipple of
a pipe bomb) or to large, cool surfaces near the exploding device or buried into nearby
wood surfaces. Organic explosives can be absorbed into plastic surfaces such as wire
insulation, vinyl upholstery or luggage, or plastic bottles. It is much better to recover any
item suspected of bearing explosive resides for laboratory examination and extraction
than to attempt to swab the contaminated surface and submit just the swabs. For this
reason, all loose debris from the seat or crater should always be collected and preserved
in clean new cans or jars. Polyethylene plastic bags and paper envelopes will allow
volatile organic explosives to migrate into and through the container with the possibil-
ity of generating cross-contamination concerns and should be avoided for the blast seat
debris or bulk organic explosives, if possible. Polyethylene (Ziploc) bags are suitable for
small components and small items where high concentrations of organic explosive
Chapter 12 Explosions and Explosive Combustion 531
 residues are not expected. Special plastic bags such as nylon or Kapak (by Ampac) are
suitable for explosive evidence, but ragged, sharp-edged debris can puncture them, so
careful handling is required.
Large immobile objects such as buildings will require swabbing with clean cotton gauze
or cotton swabs wetted with methyl or isopropyl alcohol, distilled water, or acetone.45 The
swabs then need to be sealed in an appropriate clean glass or plastic vial. The hands and fin-
gernails of persons suspected of handling explosives can be sampled with swabs, and finger-
nail scrapings recovered with clean flat wooden toothpicks. As with fire debris evidence,
maintaining the collected evidence in cool, dark surroundings until submitted to the lab will
minimize degradation or evaporation. Any comparison materials or reference specimens
(bulk explosives) should, of course, be kept isolated from scene debris.

LABORATORY ANALYSIS
Although high explosives like nitroglycerin or TNT can detonate with nearly 100 percent
efficiency, traces of the unexploded material can usually be detected on portions of the device,
its container, or nearby surfaces, especially with the sensitivity of today’s laboratory tech-
niques. These tests require solvent extraction of selected fragments of debris and analysis
of the solution by very sensitive laboratory methods including thin-layer chromatography,
capillary electrophoresis, gas chromatography, high-pressure liquid chromatography
(HPLC), ion chromatography, and mass spectrometry. Forensic scientists who have extensive
experience in explosives casework [such as those at the U.S. Department of Justice—Bureau
of Alcohol, Tobacco, Firearms and Explosives, and the Defense Explosives Research Agency
(DERA) in the United Kingdom] have developed effective analytical schemes for the
detection and identification of traces of high explosives and the more commonly used low
explosives. The methods used in such tests are beyond the scope of this text, and the reader
is referred to the specialist literature.46
Military high explosives in their original packaging and many slurry- and water-gel
blasting agents often leave virtually no detectable chemical residues if properly initiated.
Highly sensitive analytical methods such as ion mobility spectrometry and liquid chro-
matography/mass spectrometry have been shown to be useful in identifying such traces.47
Even residues of peroxide-based explosives have been identified post-blast.48 Portions of
the packaging such as the plastic wrappers, the heavy metal wire staples crimped on the
ends of some cartridges, and even the metal end panels from TNT charges have been
known to survive. Metallurgical tests on fragments may reveal the speed of reactions to
which a container was subjected, even when the explosive itself cannot be identified.49
Such tests are of particular value in examining debris from aircraft incidents, in which the
mechanical impact can shatter metal components but not at the velocities created by det-
onating high explosives.
Portions of the timing or triggering mechanism of IEDs will survive if they are not in
immediate contact with the explosive charge. Fragments of clockwork or electronic timers
and the batteries used to power timers and electric blasting caps are identifiable by expe-
rienced laboratory analysts in many instances. Much of the laboratory reconstruction and
identification of bomb components is visual, so the more complete the photographic doc-
umentation of the scene (with close-ups of the evidence in place before recovery) and the
more thorough the search and recovery process, the higher the chances of success will be.
Recently, more techniques have been developed for examining melted components
such as electrical or duct tape, and characterizing unburned smokeless powder.50 DNA
has been identified on post-blast bomb fragments. New analytical methods are being
introduced that make it possible to identify “batches” or sources of explosives on the
basis of ion ratios.52
The identification of explosive residues and the reconstruction of explosive devices
are specialized talents not found in all forensic laboratories. In addition to the extensive
532 Chapter 12 Explosions and Explosive Combustion
training required of personnel, the correct and comprehensive analysis of explosive-
related evidence requires specialized equipment and techniques and extensive reference
collections of such esoterica as pipe fittings, clocks, batteries, blasting caps, and fuses. The
investigator is encouraged to contact nearby laboratories to determine what services are
available before submitting evidence to them.
Gas chromatography/mass spectrometry using vapor concentration recovery methods
(described in Chapter 14) has been successfully applied to identify gaseous fuels (propane,
butane, pentane) trapped in clothing of victims (also in lung tissue and skin of decedents).53
Clothing, including shoes of explosion victims (living or deceased), should be collected as
soon as possible and stored separately in airtight metal cans or glass jars for analysis.
Porous materials such as upholstery if exposed to high concentrations of gases or vapors
and not exposed to extensive fire afterward may contain identifiable residue.

CASE STUDY
In a recent case, the primary victim of an explosion and fire in a parts painting operation died without regaining
consciousness. A coworker reported seeing the victim turning away from a 209-L (55-gal) steel drum where he
was in the habit of discharging cleaning solvent (xylene) from his electrostatic paint sprayer. The witness saw
a small flame at the bung hole on the top of the drum, but before he could warn his coworker, the flame pro-
gressed into the drum, causing it to explode from within. The drum separated at its bottom seam and was
propelled into the roof structure, where it deformed a steel truss, spewing many gallons of waste xylene and paint
onto the work area and the victim. Under most conditions there would not be propagation because of xylene’s
high flash point (ca. 30°C; 85°F). This incident occurred in midsummer in a non-air-conditioned factory where the
air temperature was reported to be in excess of 32°C (90°F) at the time. The aerosol discharge of the xylene from
the spray gun created an ignitable vapor at the mouth of the drum. Manufacturer data indicated that the high-
voltage discharge tip of the spray gun operated at a high voltage (15 kV) when the trigger was pulled. Bringing
the tip close to the metal drum during discharge apparently allowed an arc discharge to occur to cause ignition
of the xylene vapors.54

INCIDENT ANALYSIS
The analysis of explosion incidents follows the same scientific method described in
Chapter 1—gathering data, formulating and testing hypotheses, and reaching a testable,
defensible conclusion. Recent studies have focused on the damage and deformation of
road signs, power poles, and adjacent vehicles as a means of reconstructing the amounts
of blasting agent in large car bombs.55 Some of the analytical tools useful in fire recon-
struction (such as computer models) do not work on diffuse combustion explosion
events due to the very different dynamics of gas movement, pressure, and heat produc-
tion. See Figure 12-23 for an example of a very destructive LP gas deflagration. It is eas-
ier to analyze a condensed-phase explosion because it is a more predictable
single-pressure event. Combustion explosions can be very complex, with multiple prop-
agations dependent on the geometry of the building, distribution of fuel, failure mecha-
nisms, and the location of the ignition. (In tests observed by one of the authors, ignition
of a uniform natural gas/air mixture produced minimal flame extension out the vent/win-
dow and modest pressures when the mixture was ignited from the center of the room.
When a similar mixture was ignited from the rear wall, a mass of unburned fuel was
pushed through the window to create a very large, energetic fireball outside the compart-
ment. The pressure from this external deflagration stripped drywall lining off the walls
of the test compartment. Mniszewski and Campbell have offered some examples of ana-
lytical tools that can be of use, including some CFD computer models of explosions.56
Models of fuel concentration and distribution are suggested as ways of testing hypothe-
ses about fuel sources and ignition mechanisms. Recently, the CFD software FLACS has
Chapter 12 Explosions and Explosive Combustion 533
 been shown to be a very powerful analytical tool for modeling gas or vapor dispersion
and deflagration dynamics.57

Bomb Arson Tracking System (BATS)

Since 2003 the U.S. Bomb Data Center (USBDC) and ATF have developed a remarkable
Web-based data system for tracking fire, explosion, and explosives-related incidents of all
types occurring across the United States. The Bomb Arson Tracking System (BATS), which
is provided free of charge, allows any police or fire agency to enter incident information,
including location, description, area of origin, scene details, names of victims, witnesses, or
suspects for inclusion in a database searchable by any other law enforcement or fire inves-
tigation agency. Each agency is responsible for administering its own BATS account.
BATS is part of the USBDC’s mission to collect, analyze, and disseminate timely and
relevant information for use by U.S. federal, state, local, and tribal agencies. BATS has
been designated by the U.S. Attorney General as the sole repository for information per-
taining to explosives- and arson-related incidents. It combines data on IEDs and suspected
bomb events, stolen and recovered explosives, and fire investigations (structural, vehicle,
and wildland) so that trends and possible links can be identified. The system automatically
encodes all locations into GIS coordinates and offers mapping options. (See Appendix J
for a sample of BATS entry screens.)
BATS is capable of generating and printing out standardized reports for consistency
throughout the fire and explosion investigation community. Printed BATS reports are
customized with each participating agency’s header and seals. It can function as a case
management tool as well. BATS has drop-down menus for fire descriptors and scene details,
as well as provisions for narratives and paste-ins of photos, diagrams, log sheets, and sup-
plemental reports. It has provisions for “read-only” or “read/write” access designated by
the user to collaborate with other agencies on joint investigations. BATS can function as
a bomb/arson unit management system as well, because time spent on cases, court, training,
travel, and maintenance can be tracked individually and as a unit.
Information can be designated as “Restricted” or “Unrestricted.” If a case is desig-
nated as “Restricted,” only contact information for the investigator logging in the case is
provided. The USBDC provides two options to public safety agencies for access to BATS:
Law Enforcement and Non-Law Enforcement. Users from an agency with an NCIC-issued
ORI number have full access from any computer with Internet access. (Fire department
investigators gain full access through cooperative agreements with local police agencies or
after receiving an ORI from their state CJIS coordinator.) As of summer 2009, they can
also choose the “Non-Law Enforcement” role, which limits their access to their own and
other non-LE agency information. Data are released only to agencies authorized specifi-
cally by the user.
Because BATS has been designated as the sole reporting system by Congress and the
attorney general, regular reporting provides investigators with a powerful tool for access
to critical incident information. Arson investigators can compare local case specifics in
real time against a jurisdictional database without the limitations that NFIRS reporting
includes. (NFIRS is intended for documenting fire department responses to fire incidents,
not for documenting investigations. Its information is subject to unrestricted access by all
fire department personnel. According to the U.S. Fire Administration, NFIRS is not
intended as a case management system).
In September 2009 the House passed a bill (H.R. 1727: Managing Arson Through
Criminal History; the MATCH Act) establishing a National Arsonist & Bomber Registry.
The legislation would authorize the attorney general to establish and maintain a database
within ATF that would be built into BATS. The registry would require convicted arson-
ists and bombers to report where they live, work, and go to school and would include
descriptions of vehicles they own, an up-to-date photograph, and a photocopied I.D. In
addition to that information, the database would include finger and palm prints. This
534 Chapter 12 Explosions and Explosive Combustion
legislation will provide an important tool to law enforcement by enabling investigators to
track arsonists and bombers, regardless of where they live, across jurisdictions. By query-
ing a zip code to locate convicted arsonists living or working in the vicinity of a fire, or
by querying the components database for previous incendiary or bomb components that
match those used in an unsolved arson or bombing investigation, the process of narrow-
ing down suspects would be expedited. The bill, H.R. 1727, passed the House on
September 30, 2009, and on October 2009, was read in the Senate and referred to the
Committee on the Judiciary. The bill was allowed to die without action.
BATS is being used by police and fire agencies across the United States, and it promises
to dramatically improve the quality of information sharing and comparisons for the fire and
explosion investigation community. For more information, public investigators are encour-
aged to contact the U.S. Bomb Data Center ([email protected]) or BATS (www.BATS.gov).

Chapter 12 Explosions and Explosive Combustion 535

 CHAPTER REVIEW

Summary
Explosions may occur as part of the progress of a nor- and identity of the explosive material, (2) a competent
mal fire, as the manner of initiation of a fire, or by source of ignition for that explosive, and (3) the man-
themselves with no accompanying fire. Like fires, they ner or sequence of events that led to the initiation. The
may be the result of accident or criminal intent. same attention to small details and methodical
Although the destructive power of an explosion is approach are required for both types of scenes. The
released in a much shorter time than that of a fire, it reader is referred to the excellent outline of scene
should now be apparent that many similarities exist investigation offered in the NIJ Guide for Explosion
between the two processes. These similarities should and Bombing Scene Investigation58 and Thurman’s
make it easier for the careful investigator to recon- Practical Bomb Scene Investigation.59 In spite of the
struct the sequence of events for an explosion in the awesome destruction of some explosions, there
same manner as for a normal fire. The investigator always remains some evidence for the skillful investi-
should be able to demonstrate or prove (1) the nature gator to recover and interpret.

Review Questions
1. Name the four categories of “explosion” and 6. What are primary high explosives?
describe the features of each. 7. How do low explosives function?
2. What is a mechanical explosion? 8. What are the four Rs of an explosion scene
3. Name three major differences between a condensed- investigation?
or solid-phase explosion and a diffuse-(gas/vapor) 9. Why is it important to map the distance and
phase explosion. direction of travel of debris at an explosion
4. What is a blast propagation diagram and how is scene?
it useful in an explosion investigation? 10. What are the four Cs of an explosion device
5. Name five examples of high explosives. reconstruction?

References
1. C. F. Turner and J. W. McCreery, The Chemistry of Fire Protection Association, 2008), vol. 1, chap. 8, sec. 2;
and Hazardous Materials (Boston: Allyn and Bacon, J. D. DeHaan et al., “Deflagrations Involving Stratified
1981), 92. Heavier-than-Air Vapor/Air Mixtures,” Fire Safety
2. DuPont de Nemours, Inc., The Blasters Handbook, Journal 36 (2001): 693–710.
175th anniversary ed. (Wilmington, DE: DuPont de 8. B. Lewis and G. Von Elbe, Combustion, Flames and
Nemours Inc., 1978), 33. Explosions of Gases, 2nd ed. (New York: Academic
3. R. J. Harris, The Investigation and Control of Gas Press, 1961).
Explosions in Buildings and Heating Plant (New York: 9. Ibid., 533.
E & FN Spon, 1983), 126. 10. Ibid., 533.
4. J. D. DeHaan, “The Reconstruction of Fires Involving 11. Harris, The Investigation and Control of Gas Explosions.
Highly Flammable Hydrocarbon Fuels,” Ph.D. disser- 12. C. M. Fleischmann, “Research on Backdraft
tation, Strathclyde University, 1995. (Also, University Phenomena,” Fire Engineers Journal, 57, no. 186
Microfilms, Ann Arbor, Michigan, 1996.) (January 1997): 9–13; R. W. Bukowski, “Modelling a
5. Ibid. Backdraft Incident,” NFPA Journal
6. J. E. S. Venart, “Flixbourgh (sic): A Final Resolution” (November–December 1995): 85–89; see also Fire
in Proceedings Interflam 1999 (London: Interscience Engineers Journal 56, no. 185 (November 1996):
Communications), 257–72. 14–17; R. Chitty, “Management Summary of Survey
7. R. Zalosh, “Explosions,” chap. 8 in Fire Protection of Backdraught,” Fire Engineers Journal 56, no. 18
Handbook, 20th ed. (Quincy, MA: National Fire (November 1996): 12–13.
536 Chapter 12 Explosions and Explosive Combustion
13. R. G. Zalosh, “Explosion Protection,” Chaps. 3–15 in 31. U.S. Treasury, 1997 Arson and Explosive Incidents
SFPE Handbook of Fire Protection Engineering, 4th Report (Washington, DC: BATF, 1999).
ed. (Quincy, MA: NFPA, 2008). 32. T. G. Gersbeck, “Advancing the Process of Post-Blast
14. A. B. Spencer, “Dust: When a Nuisance Becomes Investigation,” paper presented at AAFS, Washington
Deadly,” NFPA Journal (November–December 2008): DC, February 2008; see also M. G. Haag and T.
57–61. Grissim, “Technical Overview and Application of 3-D
15. U.S. Chemical Safety and Hazard Investigation Board, Laser Scanning for Shooting Reconstruction and Crime
2175 K Street NW, Suite 400, Washington, DC 20037, Scene Reconstruction,” paper presented at AAFS,
http://www.chemsafety.gov; S. J. Selk, “Explosion Washington DC, February 2008.
Investigation and Reconstruction Using Multidisciplinary 33. DeHaan et al., “Deflagrations Involving Stratified
Methods,” paper presented at American Academy of Heavier-than-Air Vapor/Air Mixtures.”
Forensic Sciences, February 2008. 34. Harris, The Investigation and Control of Gas
16. Kirk-Othmer, Encyclopedia of Chemical Technology, Explosions.
vols. 1– 27, 4th ed. (New York: Wiley, 1998). 35. A. Beveridge, ed., Forensic Investigation of Explosions
17. T. Urbanski, Chemistry and Technology of Explosives, (London: Taylor & Francis, 1998); H. J. Yallop,
vol. 1 (New York: Pergamon, 1964). Explosion Investigation (Harrogate, England: Forensic
18. H. Henkin and R. McHill, “Rates of Explosive Science Society Press, 1980).
Decomposition of Explosives,” Industrial Engineering 36. R. Pape, K. R. Mniszewski, and A. Longinow,
Chemistry 44 (1952): 1391. “Explosion Phenomena and Effects of Explosions on
19. Urbanski, Chemistry and Technology of Explosives; Structures,” Practice Periodical on Structural Design
DuPont de Nemours Inc., The Blasters Handbook; and Construction, American Society of Civil Engineers,
J. J. Manon, “Explosives: Their Classification and 2009 (in press).
Characteristics,” Engineering and Mining Journal 37. DeHaan et al., “Deflagrations Involving Stratified
(October 1976); “The Chemistry and Physics of Heavier-than-Air Vapor/Air Mixtures.”
Explosives,” Engineering and Mining Journal 38. Harris, The Investigation and Control of Gas
(December 1976); “Explosives: The Mechanics of Explosions.
Detonation,” Engineering and Mining Journal 39. G. Kinney and K. Grahm, “Explosive Shocks in Air”
(January 1977); “Commercial Applications of (New York: Springer-Verlag, 1985).
Explosives,” Engineering and Mining Journal 40. DeHaan et al., “Deflagrations Involving Stratified
(February 1977). Heavier-than-Air Vapor/Air Mixtures.”
20. T. L. Davis, The Chemistry of Powder and Explosives 41. R. H. Ide, “Petrol Vapour Explosion: A
(Hollywood, CA: Angriff Press, 1943). Reconstruction” in Proceedings Interflam 1999,
21. P. W. Cooper and S. R. Kurowski, Introduction to the (London: Interscience Communications), 757–63.
Technology of Explosives (New York: Wiley-VCH, 42. J. C. Oxley et al., “Improvised Explosive Devices: Pipe
1996), 43. Bombs,” Journal of Forensic Sciences 46, no. 3 (2001):
22. Ibid., 43. 510–34; K. M. Whitaker et al, “Coming Apart at the
23. Ibid., 43. Seams: The Anatomy of a Pipe Bomb Explosion,”
24. D. Muller et al., “Improved Method for the Detection paper presented at AAFS, Denver, CO, February 2009.
of TATP after Explosion,” Journal of Forensic Sciences 43. J. D. DeHaan and W. R. Dietz, “Bomb and Explosion
49, no. 5 (September 2004): 935–40. Investigations,” National Association of Medical
25. S. A. Phillips et al., “Physical and Chemical Evidence Examiners, September 1987.
Remaining after the Explosion of Large Improvised 44. Whitaker, “Coming Apart at the Seams.”
Bombs, Journal of Forensic Science 45, no. 2 (2000): 45. Phillips, “Physical and Chemical Evidence Remaining.”
324–32. 46. C. R. Midkiff and R. E. Tontarski, “Detection and
26. D. B. Flemming, “Explosion in Halifax Harbor” Characterization of Explosives and Explosive Residue:
(Halifax, NS: Formac, 2004). A Review,” paper presented at 11th International
27. K. McDonald and M. Shaw, “Identification of INTERPOL Forensic Science Symposium, Lyon, France,
Household Chemicals Used in Small Bombs via Analysis November 1995; Beveridge, Forensic Investigation of
of Residual Materials,” paper presented at American Explosions; The Working Group on Fires and
Academy of Forensic Sciences, February 2009. Explosions, Guidelines for Explosives Analysis
28. P. K. Raj, “Exposure of a Liquefied Gas Container to (University of Central Florida, Orlando, FL: TWGFEX,
an External Fire,” J. Hazardous Materials A122 (2005): November 2004); J. Yinon and S. Zitrin, Modern
37– 49. Methods and Applications in Analysis of Explosives
29. U.S. Bomb Data Center, ATF, Washington DC, 2009. (Chichester, UK: Wiley, 1993).
30. U.S. Department of Justice, 2003 Annual Statistical 47. J. Cochran et al., “Optimization of Explosives Analysis
Report (Washington, DC: U.S. Bomb Data Center, by Gas Chromatography and Liquid Chromatography,”
ATF, March 2004). paper presented at AAFS, Denver, CO, February 2009;

Chapter 12 Explosions and Explosive Combustion 537

 M. N. Bottegal, B. R. McCord, and L. Adams, “Analysis paper presented at AAFS, Denver, CO, February 2009;
of Post-Blast Residues of Black Powder Substitutes by I. Corbin, M. Blas, and B. R. McCord, “Analysis of
Chromatography and Capillary Electrophoresis,” paper Smokeless Powder Components by Capillary
presented at AAFS, San Antonio, TX, February 2007; Chromatography–TOF Mass Spectrometry,” paper
A. Gapeev, J. Yinon, and M. E. Sigman, “LC/MS of presented at AAFS, Denver, CO, February 2009; W.
Explosives: RDX Characterization through Impurity Moorehead, A. Tibbetts, and A. Basconcillo, “The
Profiles,” paper presented at AAFS, Dallas, TX, Characterization of Reloading Smokeless Powders
February 2004); M. E. Sigman et al, “Matrix Effects on toward Brand Identification,” paper presented at
Explosives Recovery and Detection,” paper presented AAFS, Seattle, WA, February 2006.
at AAFS, Dallas, TX, February 2004. C. H. Wu and 51. K. L. Sewell et al., “Analysis of Nuclear DNA from
R. A. Yost, “Evaluation of Different Atmospheric Exploded Bomb Fragments and Spent Cartridge
Pressure Ionization (API) Sources for Use in Explosives Casings,” paper presented at AAFS, San Antonio,
Detection,” paper presented at AAFS, Denver, CO, TX, February 2007.
February 2009; M. E. Sigman, K. L. Steele, and C. D. 52. J. Ehleringer and M. Lott, “Stable Isotopes of Explosives
Clark, “Discrimination of C-4 Plastic Explosives by Provide Useful Forensic Information,” paper presented at
GC/MS Analysis of Impurities Associated with the AAFS, San Antonio, TX, February 2007.
Manufacturing Process of RDX,” paper presented at 53. K. Kimura et al., “Gasoline and Kerosene Components
AAFS, San Antonio, TX, February 2007. in Blood: A Forensic Analysis,” Human Toxicology 7
48. M. N. Bottegal et al., “Analysis of Trace Hydrogen (1988); J. Schuberth, “Post-Mortem Test for Low
Peroxides by HPLC-ED and HPLC-FD,” paper pre- Boiling Arson Residues of Gasoline by GC/MS,”
sented at AAFS, San Antonio, TX, February 2007; Journal of Chromatography 662 (1994).
M. E. Sigman and D. Clark, “Optimized Analysis of 54. DeHaan, J., unpublished case report, 2005.
Triperoxide by GC-MS,” paper presented at AAFS, 55. Phillips, “Physical and Chemical Evidence Remaining.”
San Antonio, TX, February 2007. 56. K. R. Mniszewski and J. A. Campbell, “Analytical
49. L. S. Chumbley and F. C. Lacks, “Analysis of Explosive Tools for Gas Explosion Investigation” in Proceedings
Damage in Metals Using Orientation Imaging Interflam 1998 (London: Interscience Communications).
Microscopy,” Journal of Forensic Sciences 50, no. 1 57. S. G. Davis et al., “Advanced Methods for Determining
(January 2005): 104–10. the Origin of Vapor Cloud Explosions. Case Study: 2006
50. E. M. Briley and J. V. Goodpaster, “Practical Danvers Explosion Investigation,” Proceedings ISFI
Applications of Pattern Recognition to the Post-Blast 2010 (Sarasota, FL: International Symposium on Fire
Analysis of Black Electrical Tape,” paper presented at Investigation Science and Technology, 2010), 197–207.
AAFS San Antonio, TX, February 2007; A. L. Hobbs 58. NIJ Technical Working Group for Bombing Scene
et al., “Use of a Database for Significance Assessment Investigation, A Guide for Explosion and Bombing
and Sourcing of Duct Tapes,” paper presented at Scene Investigation (Washington, DC: National
AAFS Seattle, WA, February 2006; J. G. Rankin, Institute of Justice, June 2000).
J. A. Meyers, and J. T. Harris, “GCIR as a Tool for 59. J. T. Thurman, Practical Bomb Scene Investigation
Analysis of Smokeless Powder Residues from IEDs,” (Boca Raton, FL: CRC/Taylor & Francis, 2006).

538 Chapter 12 Explosions and Explosive Combustion

 CHAPTER 13
Chemical Fires and Hazardous
Materials

Courtesy of U.S. Department of Justice,

Drug Enforcement Administration.

KEY TERMS

nonflammable, p. 557

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify the common causes of fires involving chemicals and hazardous materials.
■ List the typical flammable gases found in residential and commercial structures.
■ List the typical types of hazardous liquids found in residential and commercial
structures.
■ List the typical types of solid fuels found in residential and commercial structures
that can form incendiary mixtures.
■ Describe the chemical compounds associated with the production of improvised
high explosives.
■ Explain the explosion and fire hazards associated with clandestine drug laboratories
and marijuana cultivation.
■ Understand the concepts of the warning systems used to label and transport
hazardous materials.

539
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

N
ot every building or vehicle contains the same predictable fuels of wood, paper, tex-
tiles, and petroleum products. If they did, the fire investigator’s task would be fairly
simple—the physical properties, flammability, and fire behavior would all fall into
fairly narrow categories for easy analysis and reconstruction. Unfortunately, this is not the
case; many fires involve and, in fact, begin because of hazardous materials of various types.
Not only can these materials make the ignition of a fire much more likely than with “normal”
fuels, they can make the fire much more intense, dramatically increase the rate of fire
spread, and make the suppression of the fire more hazardous by producing toxic gases. It is
beyond the scope of this volume to discuss all possible hazardous materials—solids, liquids,
and gases. The reader is referred to specialized texts in this area.1 Instead, the characteris-
tics of the materials most commonly encountered will be discussed, along with the effects
they may be expected to have on a fire scene.

Gases
Hazardous gases are most commonly found in commercial structures like welding and
automotive shops, hospitals, laboratories, and fertilizer-manufacturing plants. They may
be considered hazardous because of their inherent toxic properties or flammability haz-
ards or simply because even inert gases, when compressed into cylinders for storage, can
produce explosive pressures when exposed to an enveloping fire. For the purposes of this
book, the toxic gases will not be considered. The reader is referred to other references in
that area. (See NPFA 497, 2008 ed.: Recommended Practice for the Classification of
Flammable Liquids, Gases, or Vapors and of Hazardous (Classified) Locations for
Electrical Installations in Chemical Process Areas.)
Compressed gases are usually packaged in metal bottles at pressures up to 20,000 kPa
(3,000 psi) for shipment and storage. The containers are normally checked periodically to
pressures at several times that level, but when heated, the contents can exceed those pres-
sures until the container fails. The resulting mechanical explosion causes damage to
nearby structures similar to that produced by a low-order explosion (see Chapter 12). The
damage may often be unidirectional because the container will generally fail in only one
direction—along a longitudinal seam or at a welded top or valve, for instance. Even if the
gas is inert, the rapidly expanding volume can do serious mechanical damage. If it is flam-
mable, the fuel made available to the fire may have a dramatic effect on the extent of the
fire and its rate of spread. A somewhat different problem is encountered when the gas is
subject to explosive or detonating decomposition when heated under pressure. Under
those conditions, the result is equivalent to the detonation of a large quantity of a high
explosive. When such a detonation occurs, massive destruction of the structure surrounding
540 Chapter 13 Chemical Fires and Hazardous Materials
the exploding tank is usually the result along with the creation of an explosion “seat.”
Fortunately, gases that undergo such explosive decomposition are fairly rare. The chemi-
cal properties of some of the more commonly encountered flammable and explosive gases
were included in Table 4-1 but will be discussed in more detail here.

HYDROCARBONS
Hydrocarbon gases include methane, ethane, propane, butane, ethylene, and acetylene. They
are discussed next, in order of chemical structure from the simplest to the more complex.
Methane (CH4)
Methane is a colorless, odorless gas also known as swamp gas, marsh gas, or sewer gas
because it is produced by the decomposition of organic materials in swamps, marshes, sew-
ers, and landfill dumps. Methane is the major constituent of natural gas for heating and
power generation. It burns readily in air and can explode if pressurized when heated. It is
much lighter than air, having a vapor density of 0.55. Methane has a fairly narrow explo-
sive range, from 5 to 14 percent by volume in air. Methane was once made by heating coal
or coke to make town gas or producer gas for illumination and cooking. This process
resulted in high concentrations of carbon monoxide (CO) along with the methane. Today
it is most often recovered during oil drilling, but methane is also being made as a commer-
cial fuel by the controlled fermentation of agricultural waste products in biogas conversion
facilities. The product is odorized with sulfur or mercaptan compounds to ensure leaks are
detected; however, under some circumstances, the odorant may be “scrubbed” out of the
gas, resulting in an undetected or odorless leak. Methane from sewers or swamps has a
strong odor due to the sulfur compounds (chiefly hydrogen sulfide) that accompany it.
Ethane (C2H6)
This colorless, odorless gas is the other major constituent of natural gas fuel. It has a
vapor density of 1.03, so it mixes readily with air to form an explosive mixture at con-
centrations of 3 to 12.5 percent by volume.
Propane (C3H8)
Propane is one of the most common hydrocarbon fuels, since it is a very flammable fuel
that is readily liquefied for storage in a variety of containers for vehicular, heating, cook-
ing, and industrial applications. Propane has a boiling point of -42°C (-44°F) and a vapor
density of 1.52. Its flammable mixture range is between 2.2 and 9.6 percent in air. It is con-
sidered to pose a severe explosive hazard when exposed to flame in suitable concentration.
Butane (C4H10)
Butane is a colorless gas that is most often encountered as a component of liquefied petro-
leum gas used as fuel for heating, cooking, or industrial uses. It is very flammable and will
ignite at concentrations between 1.9 and 8.5 percent in air. The gas itself has a negligible
odor but is usually (but not always) odorized with methyl mercaptan or sulfur com-
pounds to allow detection of leakage. The gas has a vapor density of about 2.0 and tends
to form a low-lying cloud along the ground. Under some circumstances it may be encoun-
tered as a liquid, since it has a boiling point of -0.5°C (30°F). Butane from hand-torch
cylinders or lighter-refill containers is often used as an extraction solvent in clandestine
drug labs or in the production of marijuana “oil” or “hash oil.”
Acetylene (C2H2)
Acetylene is a flammable gas used extensively in welding operations. It is particularly haz-
ardous in an enveloping fire because it has a very wide explosive range and a vapor density
of 0.9, allowing it to mix readily with air. Acetylene forms an explosive mixture in concen-
trations between 2.5 percent and 81 percent by volume, but pure acetylene can detonate
under certain conditions of heat and pressure. Even when simply burning in air, acetylene
Chapter 13 Chemical Fires and Hazardous Materials 541
 can create considerable damage because of the extremely high temperatures developed. In
most commercial applications, the acetylene is dissolved in acetone in a cylinder containing
a porous cement-like filler. If the contents are released while the cylinder is on its side, liq-
uid acetone will be released along with the gas. Heating of acetylene cylinders can trigger
decomposition that can cause explosions some time later, even after the cylinders are cooled.
Ethylene (C2H4)
Ethylene is a colorless, pleasant-smelling gas that burns or explodes very readily. It has a
vapor density of 0.96, so it mixes readily with air. Ethylene forms an explosive mixture
in air at concentrations between 3 percent and 30 percent. It can readily react with oxi-
dizing materials and has been known to explode its compressed gas cylinder when
exposed to heat and mechanical shock.

OTHER GASES
Nonhydrocarbons of particular interest to the fire invetsigator include hydrogen, oxygen,
ammonia, and ethylene oxide.
Hydrogen (H2)
Hydrogen is a colorless, odorless gas that is used in a variety of chemical, medical, and
industrial processes. It is formed as a by-product of many processes including battery
charging and electroplating. Hydrogen is the lightest gas, with a vapor density of 0.07,
and dissipates very quickly in open air. It has a wide range of explosive concentrations—
from 4 to 74 percent. Because of its flammability and explosive limits, hydrogen is con-
sidered by some to be the most hazardous of the compressed gases. Hydrogen gas may
also be liberated by the dissociation of water as certain metals, such as aluminum, mag-
nesium, or titanium, burn. For this reason, firefighters should exercise extreme caution
when considering the application of water to burning metal fires.
Oxygen (O2)
While it is not a flammable fuel or toxic gas, oxygen in pressurized tanks can have consid-
erable effect on a fire. Oxygen may be present in industrial applications as a compressed
gas or as a supercold liquid (boiling point -183°C; -297°F). When present as a gas,
oxygen can dramatically increase the ignitability of ordinary combustibles, the temperatures
developed by burning materials, and their rates of combustion. Organic greases in com-
pressed oxygen atmospheres become so readily ignitable (by friction or electrical discharge)
that they are forbidden. Many ordinary combustibles when exposed to liquid oxygen
become so readily oxidizable that they become virtually explosive. Even materials such as
synthetic fabrics and polymers, which are not considered flammable, become hazardous
when exposed to liquid oxygen or saturated with gaseous oxygen. Sacks of carbon black
saturated with liquid oxygen were used as a cap-sensitive blasting agent for many years.
Ammonia (NH3)
Ammonia is a colorless, sharp-smelling gas that is used in very large quantities as a fertil-
izer in agriculture and as a refrigerant in industrial plants. Anhydrous ammonia is the liq-
uefied form of the gas. While not often thought of as an explosive gas, ammonia can
explode when mixed with air. It has a vapor density of 0.6, so it behaves like natural gas
(methane) when released. Its explosive range is 16 to 25 percent by volume. Ammonia can
be corrosive to metal valves, especially brass. It is also extremely toxic, especially because
ammonia readily dissolves in water, forming the caustic ammonium hydroxide, which
attacks the mucous membranes of the eyes, mouth, and respiratory tract.
Ethylene Oxide (CH2OCH2)
Technically a flammable liquid, but with a boiling point of 11°C (51°F), ethylene oxide is
usually encountered in gaseous form. It is water soluble, and even dilute (5 percent) solu-
tions in water may be flammable. Like acetylene, ethylene oxide is ignitable over a wide
542 Chapter 13 Chemical Fires and Hazardous Materials
flammability range (3.6 to 100 percent) and can detonate in pure form. It has a vapor
density (1.52) similar to that of propane. Ethylene oxide can ignite spontaneously on con-
tact with pure oxides of iron or aluminum, or with alkali metal hydroxides (sodium, cal-
cium, potassium).

Liquids
There are many more hazardous liquids of concern to fire investigators than there are
gases, because they are found in more places and because many liquids volatilize to form
explosive or combustible vapors at temperatures developed in structure fires. Of course,
virtually all the petroleum distillates of concern to fire investigators fall into the category
of hazardous liquids because of their flammability. The general properties of these fuels
were discussed in Chapter 4 and will not be repeated here. Because of extreme fire haz-
ards and the frequency of occurrence of some particular liquids in industrial fires and
clandestine drug laboratories, they will be described in detail here. With very few excep-
tions, the liquids of interest are clear and colorless (to light amber in color) with various
viscosities. Some have characteristic odors, and it is suggested that, because of the some-
what subjective nature of smell, the fire investigator acquire samples of some of the more
commonly encountered hazardous liquids and lightly sniff them firsthand to learn to rec-
ognize particular fluids by their distinctive smells. The general chemical and physical
properties of some commonly encountered liquids and their vapors were described in
Chapter 4 but are summarized in Table 13-1.

SOLVENTS
Solvents of particular interest to the fire investigator, grouped by chemical structure,
include methyl alcohol, ethyl alcohol, acetone, ethyl ether, methyl ethyl ketone, rubbing
alcohol, carbon disulfide, aromatics, bromobenzene, cyclohexanone, nitroethane and
nitromethane, and petroleum products.
Methyl Alcohol
Methyl alcohol is a clear, colorless liquid used in many chemical manufacturing processes
including drug, plastic, paint, varnish, and glue manufacture. It is also used as a fuel for
high-performance auto and boat engines and is also known as methanol, wood alcohol,
or wood spirits. Methyl alcohol’s low flash point, wide flammability limits, and a vapor
density near that of air make it a very dangerous liquid. Because methyl alcohol burns
leaving no carbon residues, its flame is nearly transparent and blue in color. It can be
burning fiercely and not be visible until the flames impinge on another combustible that pro-
duces normal luminous flames. Methyl alcohol’s flames have a very high average tempera-
ture. It burns to form carbon dioxide and water vapor, but if combustion is incomplete, it
produces formaldehyde, a highly toxic material.
Ethyl Alcohol
Although it is best known as a beverage, ethyl alcohol (ethanol or grain alcohol) is not
encountered in its pure form except as a laboratory reagent. It is normally stored in water
solutions of various concentrations. Its flammability properties depend on its concentra-
tion, sometimes referred to (especially in beverages) as “proof” (50 proof  25 percent
alcohol, 100 proof  50 percent alcohol, and so on). Grain alcohol intended for industrial
purposes is most often at 95 percent concentration (190 proof) and is denatured by the
addition of a small quantity of a poisonous substance to make it undrinkable. This denat-
urant may be methyl alcohol, sulfuric acid, benzene, gasoline, or other toxic materials.
Although ethyl alcohol will burn at higher concentrations, it does not produce as
much heat per unit mass as do similar volatile petroleum products. It, too, burns with a
clear blue flame with little smoke. Because of the relatively low emissivity of burning ethyl
Chapter 13 Chemical Fires and Hazardous Materials 543
 TABLE 13-1 Properties of Hazardous Liquids

FLAMMABLE AUTOIGNITION
HAZARDOUS FLASH POINT LIMITS (%) VAPOR TEMPERATURE
LIQUIDS (°C) (°F) LOWER UPPER DENSITY (°C) (°F)
Methyl alcohol (methanol) 11 52 6.7 36 1.1 385 725
Ethyl alcohol (ethanol) 13 55 3.3 19 1.6 365 689
Acetone 18 0 2.6 12.8 2.0 465 869
Ethyl ether 45 49 1.9 36 2.6 160 320
(diethyl ether) (ether)
Methyl ethyl ketone (MEK) 6 21 1.8 10 2.5 515 960
(butanone)
Isopropyl alcohol (rubbing alcohol) 12 53 2.0 12.7 2.1 399 750
Carbon disulfide (carbon bisulfide) 30 22 1.3 50 2.6 90 194
Benzene (benzol) 11 12 1.3 7.1 2.8 560 1,040
Toluene (toluol) 4 40 1.2 7.1 3.1 480 896
p-Xylene 27 81 1.1 7.0 3.7 530 986
o-Xylene 32 90 1.0 6.0 3.7 465 869
m-Xylene 29 84 1.1 7.0 3.7 530 986
Gasoline (100 octane) 36 38 1.4 7.6 NA 456 853
Styrene (monomer) (vinyl benzene) 32 90 1.1 6.1 3.6 490 914
Camp stove fuel (white gas, naphtha) 28 18 1.5 7.4 NA 335 635
Paint thinner (various) 苲40 苲100 苲0.8 苲6 NA 苲210 苲410
Lacquer thinner (see toluene)
Petroleum ether (benzine) 18 0 1.1 5.9 2.5 288 550
Cyclohexanone 44 111 1.1 @ 212°C — 3.4 420 788
Nitroethane 28 82 3.4 — 2.6 414 778
Nitromethane 35 95 7.3 — 2.1 418 785
Diethylamine 18 0 1.8 10.1 2.5 312 594
Gum turpentine 35 95 0.8 UNK — 253 488

Source: NFPA, Fire Protection Guide to Hazardous Materials, 2001 ed. (Quincy, MA: NFPA, 2001).
NA  not applicable; UNK  unknown.

alcohol, it is not a very good radiator of heat, and fires ignited with it grow more slowly
than those involving petroleum distillates.

Acetone
A very common solvent in the manufacture of plastics, fabric dyes, lacquers, paints, oils,
waxes, and other products, acetone may be encountered in clandestine drug and explo-
sive laboratories. It is a colorless, watery liquid with a mintlike odor, and burns with a
yellow flame.

Ethyl Ether
Also known as diethyl ether or ether, ethyl ether is used as a solvent in pharmaceutical for-
mulations and is very common in clandestine drug laboratories. It has been used for many
years as an inhalant anesthetic in medicine, but ethyl ether also has wide use as a solvent for
544 Chapter 13 Chemical Fires and Hazardous Materials
fats, greases, and resins. Because of its extreme volatility, wide range of flammable concen-
trations, and low ignition temperature, ethyl ether is extremely dangerous as a fresh liquid.
Upon aging, it can form unstable peroxides, especially when exposed to air. These perox-
ides are sensitive to initiation by heat or shock and can cause a detonation. Burning vapors
of ethyl ether may be difficult to detect, as the flame produced is nearly transparent.
Methyl Ethyl Ketone (MEK)
Also called butanone, MEK is used in paints, varnishes, lacquers, glues, and general sol-
vents. It may be found as an intermediate solvent in clandestine drug laboratories. It is
similar in its physical and combustion properties to acetone.
Cyclohexanone
Cyclohexanone is another ketone like acetone and MEK, with similar physical and chem-
ical properties. Cyclohexanone is a moderate fire hazard that is used in clandestine drug
laboratories in conjunction with other flammable solvents.
Rubbing Alcohol (Isopropyl Alcohol)
Usually sold in water solutions (70 percent), isopropyl alcohol is a flammable liquid read-
ily available to consumers. It burns with a yellow, sootless flame.
Carbon Disulfide
A very common solvent in industrial processes, carbon disulfide is highly dangerous in the
presence of heat, flame, sparks, or even friction. At normal temperatures it is a toxic material
that affects the central nervous system like a narcotic or anesthetic. When heated, carbon
disulfide gives off highly toxic sulfur oxides as it decomposes. Because of its volatility and
extremely low autoignition temperature (100°C; 212°F), carbon disulfide is considered
one of the most hazardous materials known. It can explode readily at vapor concentrations
of around 5 percent but has a wide flammability range of 1.3 to 50 percent.
Aromatics
The aromatics—benzene, toluene, and the xylenes—are often found in combinations of
one another because of their similar chemical properties. All of them produce toxic
vapors, and all have been linked to some extent with causing cancer in animals chronically
exposed to aromatics. Benzene has the lowest flash point and greatest volatility; the prop-
erties of an aromatic mixture will largely be determined by the benzene concentration.
Benzene is common in the plastic-, resin-, wax-, and rubber-manufacturing industries.
Toluene and the xylenes may also be found as solvents for paints, lacquers, dyes, and
petroleum products. They are used in making explosives and high-performance aviation
fuel. All aromatics burn with a yellow, very sooty flame.
Bromobenzene
Bromobenzene is a low fire and explosion hazard because of its relatively high boiling
point (156.2°C; 313°F) and low vapor pressure. It is common in clandestine drug labora-
tories and can generate toxic bromine-compound fumes when heated to decomposition.
Nitroethane and Nitromethane
Like other nitrated materials, these colorless, slightly oily liquids pose a potential fire and,
particularly, explosion hazard. Both have relatively low flash points and ignition temper-
atures and will generate poisonous nitrogen oxides when heated. When they are contam-
inated with virtually any organic compound, particularly organic amines, they become
potent shock- and heat-sensitive explosives.
Nitroethane and nitromethane pose a potential fire and, particularly, explosion hazard.
Although nitromethane is used in large quantities as a lacquer and varnish solvent, when
sensitized it becomes an explosive with 95 percent the strength of TNT. Nitromethane is
often found in smaller quantities as a fuel for model cars and airplanes and as a fuel addi-
tive for gasoline engines.
Chapter 13 Chemical Fires and Hazardous Materials 545
 Petroleum Products
Because virtually all petroleum products of significance in accidental or incendiary fires
are a mixture of many separate chemical compounds that constitute a boiling point range
or “cut,” it should be kept in mind that the flammability hazard of a simple mixture of
chemicals is largely determined by the properties of the lightest, most volatile component.
For example, although automotive gasoline contains more than 100 compounds, many of
which have flash points above 50°C (122°F), the flash point (FP) of gasoline is close to
that of its most volatile major component, n-pentane (FP: -49°C; -56°F). Petroleum ether
constitutes a narrow range of hydrocarbons that distill from petroleum at temperatures
between 70°C and 90°C (158°F–194°F). Petroleum ether contains n-pentane, n-hexane,
and n-heptane, and hydrocarbons of similar properties. It may be called benzine, benzin,
or even naphtha depending on its components and its intended use. It may be considered
to have the same hazardous properties as gasoline. Because of its excellent solvent prop-
erties, high volatility, and negligible residue, petroleum ether is widely used in pharma-
ceuticals and is often encountered in clandestine drug operations.

MISCELLANEOUS LIQUIDS
Other liquids of particular interest to fire investigators include organic amines, turpen-
tine, and glycerin.
Organic Amines
Diethylamine and dimethylamine, like most organic amines, have a characteristic fishy odor.
While dimethylamine is a gas at room temperature (boiling point: 7°C; 44°F) it is most com-
monly encountered as a water solution that is considered to be a flammable liquid.
Dimethylamine has an ignition temperature of 400°C (725°F). Diethylamine is a liquid at
room temperature (boiling point: 57°C; 134°F) but has a very low flash point (-23°C; -9°F)
and an ignition temperature of only 312°C (594°F). Both diethylamine and dimethylamine
are extremely toxic and are important precursors in the illicit manufacturing of LSD and
methadone and may be encountered in fires involving clandestine laboratories.
Turpentine
Gum turpentine, not to be confused with mineral spirits or mineral turps, is an oil obtained
by steam distillation of the resins of various pine and fir woods. It consists largely of the
terpenes and other oleoresins found in the wood of those trees.2 Because it is a natural
product, turpentine will vary depending on the variety of woods from which it is made
and on the gums or oils added to it for particular properties. Turpentine is considered a
flammable liquid, although its high flash point (35°C; 95°F) makes it difficult to ignite at
room temperature in the absence of a wick. Turpentines of various types are widely used
in the paint, rubber, and plastics industries.
Glycerin
Glycerin (also glycerine and glycol) is a combustible liquid with a very high flash point
that is completely soluble in water. Because it is safe for foods, it is widely used in food
wrappings and cosmetics as a plasticizer. It is also used in paints, cigarettes, and the man-
ufacture of nitroglycerin explosives. It is becoming widely used as an antifreeze additive
in domestic fire sprinkler systems. Unlike ethylene glycol, glycerin is compatible with plastic
sprinkler piping and cements.

Solids
Most fuels in a structure fire are solids—wood, cloth, carpet, paper, and the like—
converted to combustible vapors by pyrolytic action of the flames. There are some solids
that, by virtue of their chemical properties, pose special hazards in terms of their flamma-
bility or speed of combustion. These solids will most often be encountered as part of the
546 Chapter 13 Chemical Fires and Hazardous Materials
fuel load of commercial or manufacturing structures when they burn. As such, they can
contribute significantly to the speed of and intensity of the fire. If the fire investigator is
not aware of the potential damage that can be caused by these compounds, he or she is
liable to misinterpret the diagnostic signs left by their combustion. In addition, it is pre-
cisely because of the ease of ignition or the intensity of reaction that certain materials are
used to initiate incendiary fires in all types of targets. It is because of their unusual fire
properties that certain hazardous solids will be examined here.

INCENDIARY MIXTURES
The most common incendiary solids are simple mechanical mixtures of various components—
usually a fuel and an oxidizer, sometimes with initiators, catalysts, and the like, added
to increase the efficiency of the reaction. Some of these mixtures include black powder,
flash powder, matches, safety flares or fusees, thermite, Cadweld, and aluminium powder/
plaster.
Black Powder
An explosive used in blasting, early firearms, and fireworks, black powder, or gunpow-
der, is a mixture of potassium nitrate, charcoal, and sulfur, usually in ratios of about
75:15:10. (The ratio is changed to affect the rate of burning.) When dry, black powder
can be ignited by a small spark, flame, or friction. It has an approximate AIT of 350°C
(662°F). It can be rendered inert by absorbing water from the atmosphere. Temperatures
on the order of 2,000°C (3,600°F) are produced during its combustion. When uncon-
fined, black powder deflagrates and can be used as a trailer to transmit flame from one
point to another. (The chemistry of explosive combustion of materials was treated in more
depth in Chapter 12.) It forms the central flammable core in safety fuse. Black powder is
now made in the United States at only one plant in Pennsylvania for military, antique
firearms, and fireworks users.3
Flash Powder
Flash powder, which is a mixture of a readily oxidizable metal such as powdered alu-
minum or magnesium with a strong oxidizer such as potassium perchlorate or barium
nitrate, will produce an intensely hot flame in an almost instantaneous deflagration.
(Sulfur is usually present to promote uniform burning.) These mixtures require minimal
energy in the form of spark, heat, or friction to ignite and can produce temperatures in
excess of 3,000°C (5,400°F). They are most often encountered in small explosive devices
(including those sold as cherry bombs or M-80s) and their larger, more destructive variants—
military artillery simulators. Because of their sensitivity and destructive potential, devices
using them should be opened or dismantled only by qualified EOD personnel, even for
evidentiary analysis. As discussed previously, the illicit factories set up to make devices
using these fillers pose serious explosion hazards because of airborne dusts, thin films on
working surfaces, and product stored in bulk. Some powders include a sulfur and potas-
sium chlorate mixture. The self-ignition propensity of this mixture has been well known
since the 1890s. It has been banned from commercial fireworks by most Western coun-
tries but is occasionally found in devices made in China or Mexico.4
Matches
Although innocuous enough in single use, the ignitable heads of common matches can be
combined into substantial incendiary devices. Safety matches contain sulfur (fuel), potas-
sium chlorate (oxidizer), and a variety of glues, binders, and inorganic fillers. When ignited
by flame or spark, this mixture produces considerable heat and evolves sufficient gases to
permit match heads to be used as propellants in improvised firearms. Such matches nor-
mally require frictional contact with a striker containing red phosphorus and powdered
glass in a binder. By contrast, “strike-anywhere” matches usually consist of two layers of
material with different flammability characteristics. Both contain potassium chlorate,
Chapter 13 Chemical Fires and Hazardous Materials 547
 phosphorus sesquisulfide (P4S3), rosin, and powdered glass, but the upper “igniter” button
contains a higher percentage of P4S3 and glass, while the lower contains paraffin and
sulfur.5 This combination allows the tip to ignite when struck against any hard surface,
while the lower part has good flame characteristics but poor frictional sensitivity. Match
strikers in bulk are extracted with a volatile solvent to provide phosphorus for metham-
phetamine production.
Safety Flares or Fusees
Safety flares and fusees contain a “fuel” of sulfur, wax, and sawdust and an oxidizer of
strontium nitrate (for burning with a red flame) and potassium perchlorate. These devices
burn with a very hot (in excess of 2,000°C; 3,600°F) flame and leave a white or gray
lumpy residue consisting almost entirely of oxides of strontium. The mixture requires a
high-heat source for ignition, and an igniter button is built into the end of each flare to
provide it (see Figures 8-23 and 8-24).
Thermite (or Thermit)
A mixture of powdered aluminum and iron oxide (Fe2O3), thermite, once ignited, will
burn with extreme violence, spattering molten metal and producing temperatures in
excess of 2,500°C (4,500°F)—that is, sufficient heat to melt iron, steel, and even concrete.
Fortunately, the reaction requires considerable heat input to begin. Usually, a magnesium
strip is ignited as an initiator. Once ignited, a thermite fire is very difficult to extinguish
because it supplies its own oxygen.
Cadweld
A mixture of powdered aluminum–copper alloy and iron powder that functions in the same
way as thermite, Cadweld produces a pool of molten copper as a residue and is used to weld
large copper buss bars in heavy-duty electrical equipment. The temperature produced is on
the order of 2,000°C (3,600°F). The mixture requires a high-temperature igniter, but the
commercially prepackaged containers include a small quantity of magnesium powder so
that the whole sample can be ignited with a match.
Aluminum Powder/Plaster
A mixture of fine aluminum powder, mixed with plaster of Paris and cast into ice cube–size
blocks with a magnesium ribbon igniter was used as an improvised incendiary device in
World War II and has much more recently been suspected as an ignition source in an arson
case. It burns similarly to thermite, with a very hot, intense flame.

OXIDIZING SALTS
Although they will not be treated individually here, all compounds containing chlorate or
perchlorate salts must be considered potentially hazardous materials because of their
potency as oxidizing agents. These materials in combination with virtually any oxidizable
substance can produce deflagrations with potential for serious structural damage. The same
holds true for nitrate salts of inorganics, or nitrate esters of organics (like cellulose).
Virtually all nitrogen-containing compounds are unstable to some extent, and nitrates are
generally even more unstable as very active oxidizers. Some inorganic nitrates such as
sodium nitrate are used as a substitute for more expensive explosives in dynamites and blast-
ing agents. Although ammonium nitrate in its pure form will explode when heated, when it
is contaminated with any organic material (such as charred paper or oil) to act as a fuel, it
explodes with less stimulation and considerably more force.6
This phenomenon is responsible for the frequent use of ammonium nitrate–fuel oil
(ANFO) as a blasting agent for both legitimate and criminal purposes. For many years
ammonium nitrate fertilizer was thought to be so stable that dynamite was used to break
up masses of it in damp warehouses. It was when these same piles of fertilizer were contam-
inated with oil that they detonated in sympathy with the smaller dynamite charges.
Exposure to fire can also cause detonation. This mechanism was the cause of the Texas City,
548 Chapter 13 Chemical Fires and Hazardous Materials
Texas, explosion of 1947.7 Sodium chlorate recovered from home-hobbyist welding torches
or from weed killers has been used as an oxidizer in improvised explosive and incendiary
devices.

REACTIVE METALS
Reactive metals include sodium metal, potassium, phosphorus, and magnesium.
Sodium Metal
Sodium metal is used in drug synthesis and related organic chemistry reactions. It is a soft
metal that looks like silvery white putty. It is normally stored in a kerosene or oil bath because
it reacts strongly with any water available, even that in the air or on human skin. When
exposed to dry heat, sodium metal gives off very toxic oxides and in a general fire will flame
spontaneously in air. It has an ignition temperature of about 115°C (240°F). In water, it gen-
erates large quantities of explosive hydrogen gas and caustic sodium hydroxide (lye) in a vio-
lently exothermic reaction, which spews flame and spatters hot lye and molten sodium.
Potassium
Potassium is used less frequently in chemical synthesis than sodium metal but presents the
same serious hazards as sodium. In its pure form it is a silvery metallic crystal. Potassium
generates hydrogen gas, potassium hydroxide, and a great deal of heat on contact with
water. In addition, potassium can react with atmospheric oxygen to form explosive
peroxides and a superoxide that form a yellowish layer on the metal. These oxides can
detonate when initiated by friction. In moist air, potassium can develop sufficient heat to
ignite spontaneously.
Phosphorus
Phosphorus occurs in two forms of waxlike metal: one white (or yellow) and the other
red. The two forms have quite different properties. White phosphorus is one of the few
known materials that will burst into flames when exposed to air, so it is stored under
water or a petroleum distillate. It has an ignition temperature of 30°C (86°F) and will
react with most oxidizers, producing very toxic phosphorus oxides. In spite of its flam-
mability, white phosphorus is not often encountered in accidental or incendiary fires.8 It
has been used in the form of a dispersion of finely powdered white phosphorus in a
volatile solvent, called “Fenian fire.” When this substance is thrown onto a target, the sol-
vent quickly evaporates, and the phosphorus bursts into flame.
If exposed to sunlight, white phosphorus can crystallize into the more stable red phos-
phorus, which does not burn when exposed to air and has an ignition temperature of 360°C
(500°F). It will burn, however, and react violently, even explode, on contact with strong oxi-
dizers. It is used as an igniter in frictional ignition devices, including those on matches and
safety flares. It has found increasing favor in illicit methamphetamine manufacturing, where
a mixture of red phosphorus and other precursors is often allowed to react unmonitored.
Magnesium
When people imagine the hottest flame possible, they often think of burning magnesium.
Although not readily ignitable in large masses with small surface areas, magnesium shav-
ings, powder, or ribbon can readily be lit with a common match. Magnesium fires can reach
temperatures of 3,000°C (5,400°F) and produce toxic oxides, most often a dense, white,
fluffy powder. When heated, magnesium, can react with water or steam and produce flam-
mable hydrogen gas as well as heat. See Figure 5-26 for an example of a magnesium fire.

Clandestine Laboratories
As used here, the term clandestine laboratories refers to clandestine drug labs, marijuana
cultivation, and clandestine explosives laboratories.
Chapter 13 Chemical Fires and Hazardous Materials 549
 CLANDESTINE DRUG LABORATORIES
A significant number of structure fires are caused every year by accidental ignition of the
solvents or vapors associated with the illicit manufacturing of drugs. The Drug Enforcement
Administration (DEA) reported that seizures of clandestine drug labs and dumpsites have
decreased dramatically since 2005, when 12,619 were seized, down to 5,910 in 2007, and
6,783 in 2008. The highest numbers are no longer recorded in California (286 – 470 in
that period); they are now reported in larger numbers in Missouri, Indiana, Illinois, and
Tennessee.9 Some 97 percent of the labs seized were reportedly producing methampheta-
mine, but amphetamine, phencyclidine (PCP), methadone, methaqualone (very rare), and
hallucinogens such as lysergic acid diethylamide (LSD), and methylene dioxyamphetamine
(MDA) or methylene dioxymethamphetamine (MDMA) are also found.10
The most common method of large-scale (10 lb or more) illicit manufacture of
methamphetamine uses hydriodic acid (HI) and red phosphorus. Hydriodic acid is
extremely corrosive, and red phosphorus can be converted to the more reactive white (or
yellow) forms by the application of heat. Large glass flasks containing these reactants read-
ily overheat and boil over, starting fires in all manner of buildings. Small batches of
methamphetamine can be made in a bathroom sink with just a few glass jars (as illustrated
in Figures 13-1a and b). The reaction uses anhydrous ammonia, metallic (elemental)
lithium (or sodium), and iodine. The lithium may be purchased or extracted from lithium
batteries; the iodine may be purchased in solid form or extracted from medical iodine solu-
tions (tinctures), and the phosphorus extracted from the striker strips of matchbooks. The
starting materials themselves are often extracted from large quantities of over-the-counter
cold tablets. Anhydrous ammonia, used extensively as an agricultural fertilizer, has been
stolen in quantities from gallons to tractor-trailer loads. Sometimes the anhydrous ammo-
nia is transferred to empty propane tanks. It is corrosive to the tank fittings and will cause
them to leak badly, as shown in Figure 13-2.
Until recently, Freon was used as the extraction solvent, but with the increasing scarcity
of fluorocarbons, flammable solvents are used. If bulk anhydrous ammonia is not available,
ammonium nitrate fertilizer can be converted to anhydrous ammonia using household lye.
Empty lye containers or fertilizer bags would be significant evidence of the use of this process.

(a) (b)

FIGURES 13-1A AND B Before-and-after (fire) photos of a mock-up of a common small-batch methamphet-
amine lab. Jars, cans, and tubing could easily be overlooked as evidence if the fire damage were more extensive.
Courtesy of John D. DeHaan.

550 Chapter 13 Chemical Fires and Hazardous Materials

 FIGURE 13-2 LP bottle
used to transport anhy-
drous ammonia with
cross-threaded “adapter.”
Green/blue corrosion on
brass valve will eventually
cause leakage of valve.
Note red phosphorus
reaction in glass jar
behind bottle. Courtesy
of the U.S. Department of
Justice, Drug Enforcement
Administration.

All these extraction and manufacturing processes involve the use of large quantities of
flammable liquids—methyl alcohol, butane, acetone, lacquer thinner, ethyl ether, toluene,
petroleum ether, camping fuel, and even gasoline—as solvents (which are then often evap-
orated by heating on hotplates or cookstoves). Many of the intermediate reactions take
place in such solutions, and almost invariably the final product is “salted out” of a solu-
tion using hydrogen chloride (HCl) gas or even table salt. The resulting mixture must then
be allowed to dry, and the room or building then fills with the vapors of evaporating solvent—
often to explosive concentrations. The explosions are initiated by the electric arc of a
switch or motor, the overheating of reactant materials, the flames of a cookstove, or the
match of an anxious observer. The ensuing fire is fueled by the large volumes of available
flammable liquids on hand. Because of the potential profits, the large-scale laboratories can
be surprisingly sophisticated, well equipped, and capable of turning out large quantities of
fairly pure dangerous drugs. (See Figure 13-3, for example. An example of a typical “back-
yard” meth lab is shown in Figure 13-4.)
The newest twist in clandestine methamphetamine manufacture is called the “one
pot” method. Here, pseudoephedrine from cold medications is mixed with ammonium
nitrate, lithium metal (from lithium batteries), lye (sodium hydroxide), water, and a flam-
mable solvent such as ethyl ether or white gas in a single container such as a 2-L soft drink
bottle, or similar jar or metal can. The reaction produces heat and a great deal of hydrogen
gas. If the container is not vented repeatedly, the pressure can cause a mechanical explo-
sion of the container. This explosion releases hydrogen, hot flammable liquid, hot lithium
metal, and caustic hydroxides. In about 20 minutes, the chemical reaction produces
methamphetamine base (oil) in the solvent, which then must be “salted out” with HCl
gas. This gas is often generated in a separate jar with table salt and battery acid, or metal
strips with hydrochloric acid. The final result, once filtered and washed, is a relatively
pure meth with no telltale odors of bulk chemicals.11

Chapter 13 Chemical Fires and Hazardous Materials 551

FIGURE 13-3 Clandestine
drug laboratories can
include reactive and toxic
materials in elaborate and
expensive glassware and
other preparation equip-
ment. Investigators should
be cautious so as to avoid
disturbing any reactions
in progress. Courtesy of
the U.S. Department of
Justice, Drug Enforcement
Administration.

To avoid detection, these laboratories are usually located in remote houses, garages,
shacks, barns, or even trailers—anywhere electric power can be made available. More
frequently, clandestine labs are appearing in motorhomes and camping trailers as isolated,
self-contained plants with their own power generators or batteries and closed water-cooling
systems. Clandestine drug labs have also been found in motels, rented houses, or apart-
ments in urban areas, sometimes close to police and fire facilities.

FIGURE 13-4 Typical

“backyard” meth lab with
hot plate, mixers, extrac-
tion solvents and extrac-
tion/reaction vessels (soda
bottles). Courtesy of the U.S.
Department of Justice, Drug
Enforcement Administration.

552 Chapter 13 Chemical Fires and Hazardous Materials

 FIGURE 13-5 Pseudo-
ephedrine extraction
of powdered diet tablets
using automotive fuel
additive. Lithium will be
extracted from the
batteries. Courtesy of the
U.S. Department of Justice,
Drug Enforcement
Administration.

The finding of one or more of the following at a fire scene should warn the fire inves-
tigator that he or she is dealing with a potential clandestine drug lab:
■ Large numbers of empty containers for the following:
■ Cold/allergy medicines or diet aids with pseudoephedrine (see Figure 13-5)
■ Lithium batteries (e.g., for cameras, watches, or calculators)
■ Butane (lighter refills)
■ Lye (drain cleaner)
■ Camping fuel or automotive fuel additive (see Figure 13-5)
■ Muriatic acid (hydrochloric acid)
■ Iodine solution
■ Dimethyl sulfone (MSM) (used as a diluent or cutting agent)
■ Cylinders or tanks of the following:
■ Anhydrous ammonia
■ HCl (hydrogen chloride)
■ H2 (hydrogen)
■ LP gases (modified or corroded)
■ Matchbooks without striker strips (in very large quantities)
■ Multiple hotplates, electric skillets, or camp stoves
■ Large quantities of alcohol (methyl, ethyl, or isopropyl), acetone, or MEK
■ Ice-water baths in sinks, pans, or beer coolers with miscellaneous jars or tubing
■ Residues dumped outside that look like oatmeal
Because of the complexity of the chemistries involved and the nature of the hazardous Stay Safe
materials present—flammable solvents; organic amines; methylamine; strong acids [sulfuric
(H2SO4), hydrochloric (HCl), and especially hydriodic (HI)]; caustic bases; dangerous gases Flammable, caustic, and
such as H2, HI, and HCl; and extremely reactive materials such as anhydrides, phosphorus, highly toxic chemicals
in gaseous, liquid and
metallic lithium, sodium, and potassium (not to mention the toxic intermediate products
solid forms may be
including iodine vapors and phosphine gas)—these scenes are very hazardous to investigate. encountered in such
At the first appearance of stored chemicals, glassware, laboratory equipment, pumps, ovens, labs. Proper PPE is
and related apparatus, the assistance of a qualified forensic drug chemist is essential, and the essential.

Chapter 13 Chemical Fires and Hazardous Materials 553

 agency should seal off the scene and notify the appropriate narcotics enforcement team or
DEA. In the interim, personnel should not turn on equipment that is off, should not turn off
equipment that is on, and should not remove bottles, jars, or flasks from ice-water baths, if
present. Many of the synthesis reactions are strongly exothermic and are being cooled by
immersion or circulation to prevent a runaway buildup of heat that can cause the reactant
mass to explode. All such scenes should be approached with extreme caution because of the
hazards of the solids, liquids, and even gaseous materials present.

MARIJUANA CULTIVATION
“Grow-ops” for the concealed cultivation of marijuana have been found in every type of
structure in nearly every locale in North America—private residences, commercial prem-
ises, apartments, storage rental units, underground mines, even motorhomes and trailers.
Except for the flammable solvents used for extraction of “hashish” oils, their main fire risk
is due to the massive power needs involved in the high-intensity lamps used to promote
rapid growth. Power connections (often illicit bypasses) can be haphazard or very sophis-
ticated. Electrical fires caused by these connections are the most common means by which
these “grow-ops” are discovered (as in Figures 13-6a–c). Fans, pumps (for water or hydro-
ponic solutions), timers, and the lamps all provide ignition sources.12 The high-intensity

FIGURE 13-6 Marijuana “grow operations” are often detected when illicit power connections (meter bypasses)
fail. Courtesy of Vic Massenkoff, Contra Costa County Fire Dept.

554 Chapter 13 Chemical Fires and Hazardous Materials

lamps used (often metal halide) can also ignite structural materials or growing plants by
direct contact, or ignite nearby wood by radiant or conducted heat.

CLANDESTINE EXPLOSIVES LABORATORIES

As discussed in Chapter 12, improvised explosive mixtures are becoming more common.
These typically involve hydrogen peroxide solutions and acetone (from consumer prod- Stay Safe
ucts). The intermediate and final products are all hazards due to their extreme sensitivity
and explosive power. The volatile solvents pose the usual fire risks. EOD units should Even innocuous-looking
solids and liquids can
handle all suspected explosive evidence and decontamination. be highly sensitive
explosives at such
scenes. If in doubt, call
Warnings for EOD assistance.
There are two different warning systems in common use in the United States to warn
those handling or transporting hazardous materials of the types and severities of dangers
posed if a material is exposed to heat or fire. These systems are the NFPA 704 system and
the Department of Transportation Hazardous Materials Transportation System.

NFPA 704 SYSTEM

NFPA 704, Standard System for the Identification of the Hazards of Materials for Emergency
Response, is used on storage tanks, exteriors of buildings, entry doors, pipelines, and the like.
This system warns of the hazards relative to health, flammability, and chemical reactivity
with a numbering system in which 0 represents no significant hazard, and 4 represents the
highest degree of that hazard. The NFPA label is in the form of a diamond-shaped sign, as
shown in Figure 13-7.
The topmost quadrant (often color-coded red) is the fire hazard. The left-hand quad-
rant is the health hazard (often color-coded blue), and the right-hand quadrant is the
chemical reactivity hazard (susceptible to release of energy) and is usually color-coded yellow.
The bottom quadrant is for special warnings when appropriate: the three-bladed pro-
peller for radiation hazard; the letter W with a horizontal line through it to designate
materials that are reactive to water (and therefore water should not be applied during fire
emergencies). The letters OXY in the lower quadrant signify that the material is an oxi-
dizer, while the letter P designates materials that can undergo polymerization (an exother-
mic reaction). The numeric designations are graded as follows:
A. Health Hazard
4: Materials that can cause death or major injury upon very short exposure even if
medical aid is given immediately
3: Materials that can cause serious temporary or residual injury on short exposure
2: Materials that could cause temporary incapacitation upon intense or repeated expo-
sure unless medical aid is offered immediately

Flammability
Health
hazard

Chemical FIGURE 13-7 NFPA 704 system for

reactivity labeling hazardous material containers.
hazard Courtesy of Meyer, E. E., Chemistry of
Special warning Hazardous Materials, 5th ed. Upper Saddle
(water reactive, River, NJ: Pearson/Brady, 2010.
oxidizer, radiation, etc.)

Chapter 13 Chemical Fires and Hazardous Materials 555

 1: Materials that can cause irritation only if no medical aid given
0: Materials that offer no hazard beyond that of any combustible material
B. Flammability
4: Materials that will evaporate or vaporize completely on exposure to air at normal
temperature and pressure, will mix with air, and will burn readily
3: Liquids and solids that can be ignited under almost all ambient conditions
2: Materials that will ignite if heated moderately or exposed to high ambient temperatures
1: Materials that must be preheated before they will ignite
0: Materials that will not burn
C. Reactivity
4: Materials that are capable of detonation or explosive decomposition at normal
temperatures
3: Materials that are capable of detonation or explosive reaction but require a strong
initiator, or heating under confinement, or water reactivity
2: Materials that become unstable at elevated temperatures and pressures, may react
violently with water, or form potentially explosive mixtures with water
1: Materials that can become unstable at elevated temperatures or pressures or react
with water
0: Materials that are stable even under fire exposure conditions, and are not reactive
with water

FEDERAL HAZARDOUS MATERIALS TRANSPORTATION SYSTEM

There are also designations for hazardous materials under the U.S. Dept. of Transporta-
tion under the Federal Hazardous Materials Transportation Law (49 C.F.R. Parts
171–179). Under this system, each material classified as hazardous has a three- or four-
digit number preceded by the prefix UN for United Nations (international shipping) or
NA for North America (United States Canada, and Mexico). DOT labels are diamond-
shaped and color-coded to a specific hazard class and usually include a warning picto-
graph (Figure 13-8). They are affixed to the outside of shipping containers (boxes, drums,

FIGURE 13-8 DOT

placard on fuel truck
(Biodiesel). Courtesy
of John D. DeHaan.

556 Chapter 13 Chemical Fires and Hazardous Materials

etc.) and in some cases to be on placards on a vehicle transporting them. The classifica-
tions are as follows:
Class 1: Explosives, with subdivisions reflecting sensitivity: 1.1 for detonating explo-
sives, to 1.6 for materials with a negligible probability of accidental initiation
Class 2: Gases (flammable, nonflammable, poisonous) nonflammable ■ Mate-
Class 3: Flammable liquids (materials with a flash point of less than 60.5°C (141°F) rial that will not burn
under most conditions.
Class 4: Flammable solids, spontaneously combustible materials, or materials that
are dangerous when wet
Class 5: Oxidizers and organic peroxides
Class 6: Poisonous materials and infectious materials
Class 7: Radioactive materials
Class 8: Corrosive materials
Class 9: Miscellaneous hazardous materials
The reader is encouraged to review the emergency response literature for more spe-
cific information. Note that under this classification system, liquids with a flash point of
less than 60°C (141°F) are classified as flammable liquids (as opposed to the 38°C
(100°F) classification used by NFPA.

Chapter 13 Chemical Fires and Hazardous Materials 557

 CHAPTER REVIEW

Summary
Hazardous materials of the types discussed in this change the behavior of a fire. If a fire appears to vio-
chapter are one of the problem areas facing the fire late the rules, the investigator must consider these
investigator. They can be the cause of the fire or explo- materials whenever the circumstances indicate. Like
sion (as the first fuel ignited or source of heat) or involved all other fuels, however, they follow basic principles of
in the fire as an additional fuel. Because of their reactivity combustion.
or intensity of reaction, their presence can significantly

Review Questions
1. Name three of the most hazardous flammable 6. Name four of the chemicals used in illicitly man-
hydrocarbon gases. ufacturing methamphetamine.
2. Name three flammable nonhydrocarbon gases. 7. What solvents have been found in clandestine
3. Why are ethyl ether and carbon disulfide so dan- drug labs?
gerous from a flammability standpoint? 8. List five of the waste materials or discards that
4. What is thermite (or thermit) and what residues indicate a clandestine meth lab is (or was) present.
does it leave behind? 9. Why are reactive metals so dangerous?
5. Name three metal–oxidizer mixtures that can be 10. Why do nitromethane and nitroethane pose a
used as incendiary materials. special safety risk?

References
1. C. F. Turner and J. W. McCreery, Chemistry of Fire 29, 1988),” FEMA, USFA National Fire Data
and Hazardous Materials (Boston: Allyn and Bacon, Center.
1980); DOT Hazardous Materials Emergency 7. H. W. Stephens, The Texas City Disaster—1947
Response Guidebook (Washington, DC: U.S. (Austin, TX: University of Texas Press, 1997).
Government Printing Office, 2008) (see http://www. 8. E. Meyer, Chemistry of Hazardous Materials, 5th ed.
phmsa.dot.gov/hazmat/erg for English and Spanish (Upper Saddle River, NJ: Pearson/Brady, 2009).
editions); NFPA, Fire Protection Guide to Hazardous 9. U.S. Department of Justice, Drug Enforcement
Materials, 2001 ed. (Quincy, MA: National Fire Administration, www.justice.gov/dea/index.htm,
Protection Association, 2001); N. I. Sax, Dangerous March 2009.
Properties of Industrial Materials, 10th ed. (New 10. U.S. Department of Justice, Drug Enforcement
York: Van Nostrand Reinhold, 2000). Administration, EPIC, September 2000.
2. M. A. Trimpe, “What the Arson Investigator Should 11. Forensic Applications Consulting Technologies Inc.,
Know about Turpentine,” Fire and Arson Investigator Bailey, CO, personal communications November 24,
44, no. 1 (September 1993): 53–55. 2009; D. Hunt, S. Kuck, and L. Truitt, Methamphetamine
3. J. A. Conkling, “Chemistry of Fireworks,” Chemical Use: Lessons Learned, final report to the National
and Engineering News, June 29, 1981, 24–32. Institute of Justice (NCJ 209730), available at www.
4. H. G. Tanner, “Instability of Sulfur–Potassium ncjrs.gov/pdffiles1/nij/grants/209730.pdf (Cambridge,
Chlorate Mixtures,” Journal of Chemical Education MA: Abt Associates, January 31, 2006).
36, no. 2 (February 1959): 58. 12. C. Reed and A. Reed, “Fire Investigation of
5. H. Ellern, Modern Pyrotechnics (New York: Chemical Clandestine Marijuana Grow Operations” in
Publishing, 1961). Proceedings Fire and Materials 2005 (London:
6. USFA, “Six Firefighter Fatalities in Construction Interscience Communications), 365–78.
Site Explosion, Kansas City, Missouri (November
558 Chapter 13 Chemical Fires and Hazardous Materials
 CHAPTER
14
Laboratory Services

Courtesy of John D. DeHaan.

KEY TERMS

adsorption, p. 570 fingerprints, p. 596 spoliation, p. 595

chromatography, p. 562 mass spectrometry, p. 566

OBJECTIVES

After reading this chapter, you should be able to:

■ Understand the role of laboratory services in fire and explosion investigations.
■ Describe the various types of services and tests available from forensic and fire
testing laboratories.
■ List and identify the common types of evidence found at fire scenes.
■ Identify examples of how failure analysis by forensic engineers can be used to
help determine fire causation.
■ Explain the basic concepts of how volatile accelerants are identified using gas
chromatography and mass spectrometry.
■ List the various types of techniques used to isolate and identify volatile residues.
■ Identify the common types of materials found in chemical incendiaries.
■ Describe investigative methodology for identification and preservation of non-fire-
related criminal evidence.
■ Search the Internet for the appropriate and up-to-date guidance on standards used
by public and private forensic laboratories.

559
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

E
ven an experienced fire investigator often requires an analysis or interpretation of
the physical evidence relating to fire causation or spread that entails the services of
a laboratory specialist. In fact, it is the wise and experienced investigators who recog-
nize the limitations of their own knowledge, beyond which they must take advantage of
other experts to ensure that the investigation is complete and accurate. The less experi-
enced may try to overextend their expertise and offer erroneous or misleading interpreta-
tions of evidence, which can embarrass them and their clients.
Today, courts expect fire investigators to rely on properly conducted tests for critical
data. It is well worth the time and effort for the investigator to anticipate the need for such
experts and determine their availability before they are needed in a crucial case.*

Availability of Laboratory Services

FORENSIC LABORATORIES
Most states in the United States have at least one state criminalistics laboratory that can
provide most of the analytical services a fire investigator will need. In fact, many states
have a system of regional laboratories to minimize travel time and expense. These labo-
ratories, operating under the auspices of the state’s department of justice, public safety,
state prosecutor, state police, or similar law enforcement authority, work in cooperation
with the city or county laboratories found in many areas to provide a wide range of serv-
ices. On a national level, the laboratories of the Federal Bureau of Investigation and
Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) are well known. In Canada,
fire debris analysis is provided by the Centre of Forensic Sciences in Toronto and Sault
St. Marie, the Laboratoire de sciences judiciaires et de médecine légale in Montreal, and
through the Forensic Science and Investigation Service of the Royal Canadian Mounted
Police, which operates laboratories in six cities across the country. With some exceptions,
the services from these public laboratories are available at no cost to personnel from pub-
lic fire or police agencies and to investigators from some semiprivate agencies such as
transit authorities, public utilities, and railroads. Due to limited laboratory resources,
there may be restrictions on which agencies they are permitted to serve. In addition, there
may be a case priority scheme in operation that relegates fire-related evidence to low-
priority status. Because the priority is often set on the basis of whether there were deaths
or injuries or on a court date’s having been set, that information should be provided to the

*Many such experts may be described as forensic scientists, forensic chemists, forensic engineers, and the like.
The term forensic means having to do with a court of law, so such terms refer to the application of a scientific
discipline to problems having to do with the law.

560 Chapter 14 Laboratory Services

lab when the evidence is brought in to ensure fastest possible service time. The investigator
should contact the nearest laboratory before its services are needed, so that time is not
wasted just in finding a laboratory.
Private investigators, whether retained by public agencies such as public defenders or
by private or corporate clients in civil or criminal cases, have a more difficult and costly
problem. Although a few public laboratories will service public defenders and even private-
sector investigators on a contractual basis, for the most part, private-sector investigators
must turn to private laboratories. Many large insurance and engineering firms, in addi-
tion to providing laboratory services to their own investigators, offer such services to out-
side clients on a fee basis. There are also small, private laboratories that can perform
appropriate tests on a per-case, per-analysis, or hourly fee basis and provide expert testi-
mony as to their results. Finally, there are qualified consultants who will evaluate the evi-
dence and provide interpretation and expert testimony while contracting with others for
the actual laboratory work. There are many more qualified, private-sector facilities avail-
able now than just a few years ago. Many also do work (on a fee basis) for public-sector
agencies whose own labs cannot provide timely services.

FIRE TESTING LABORATORIES

There are also many tests that forensic laboratories cannot provide. These include heat
release rate testing, flame spread, ignitability, and room (even multiroom) fire tests. Public
agencies in the United States may qualify for assistance from the Fire Research Laboratory
operated by ATF in Annandale, Maryland. Private laboratories such as Omega Point
(Elmendorf, Texas), Southwestern Research Institute (San Antonio, TX), MDE (Seattle,
WA), Underwriters Laboratories (Northbrook, IL), Chilworth Pacific Fire Testing, (Kelso,
WA), and Western Fire Center (Longview, WA) have provided qualified testing and analysis
services. Some of the ASTM tests such labs can provide have been described in previous
chapters.

EXPERT QUALIFICATIONS
The most important consideration is not whether a lab has the necessary equipment but
whether it has the experience and specialist training necessary to provide accurate inter-
pretation. The effects of fire, evaporation, exposure, and even sample collection and stor-
age on physical evidence from fires must be well understood by the forensic laboratory
analyst if critical errors in interpretation are to be avoided. Simply having a chart or
numerical reading does not do the investigator any good. Only a qualified analyst will be
able to correctly interpret that data and guide the investigator in assaying the contribu-
tion of the evidence to the case. Unfortunately, judging someone’s qualifications is much
more difficult than dialing a phone number, and there have been instances of malfeasance
and even fraud on the part of some laboratories.
The International Association of Arson Investigators (IAAI) first addressed the issue
of laboratory qualifications by establishing a Forensic Science Committee and developing
consensus guidelines on the analysis and interpretation of fire-related evidence.1 The
American Society for Testing and Materials (ASTM) used that IAAI document as the
foundation for what is now a series of test methods, guides, and practices covering many
of the techniques used in fire debris analysis, including ASTM E1618: Standard Test
Method for Ignitable Liquid Residues in Extracts from Fire Debris Samples by Gas
Chromatography-Mass Spectrometry2 and ASTM E1492: Practice for Receiving,
Documenting, Storing, and Retrieving Evidence in a Forensic Science Laboratory.3
(Additional ASTM methods will be described in following sections.) These methods were
developed by consensus of the scientists involved and they can be used by any analyst.
Adherence to them indicates a professional attitude on the part of the analyst. The
American Society of Crime Laboratory Directors (ASCLD) has established a protocol for
Chapter 14 Laboratory Services 561
 accreditation of forensic laboratories that calls for use of such standardized methods for
analysis and interpretation, as well as record keeping. Finally, the American Board of
Criminalistics (ABC) has established a certification program for individuals involved in
fire debris analysis. Scientists who pass a written specialty knowledge exam and who par-
ticipate in a quality assurance testing program may be designated to be a Fellow of the
ABC (F-ABC) in that specialty.
In addition, active membership in one of the regional or national forensic science
organizations is a hallmark of a professional forensic scientist. Fire testing laboratory per-
sonnel will often have membership in ASTM International (formerly known as the
American Society for Testing and Materials), the IAAI, the National Fire Protection
Association (NFPA), or the Society of Fire Protection Engineers (SFPE) as part of their
professional qualifications. These qualifications should aid the fire investigator in finding
a suitable laboratory and a qualified analyst.
The selection of a qualified lab is not easy and there are too many variables of
expertise, qualifications, analysis time, and costs to provide specific guidelines here. It is
clear that courts expect some sort of scientific data to be offered in support of the inves-
tigator’s conclusions and that the data be demonstrated to be accurate and reliable. With
the information offered here on types of evidence, possible analyses, and strengths and
weaknesses of evidence and techniques, the investigator should be able to confer knowl-
edgeably with various experts and determine their suitability for the particular problem
at hand.

Identification of Volatile Accelerants

By far the predominant service provided by laboratories to fire investigators is the analy-
sis of fire debris for suspected volatile accelerants. In one laboratory study, ignitable liq-
uids (flammable and combustible liquids) were found in 49 percent of all arson cases
submitted to a major crime laboratory over a 3-year period.4 Data from a more recent
survey verify the same percentages.5 Although fire databases do not include information
on the frequency of ignitable liquid use in setting fires, it is the consensus of fire and arson
investigators in the United States that a large percentage (more than 50 percent) of
detected arsons involve their use. The use of ignitable liquids, most commonly petroleum
products, has been common throughout the history of twentieth-century incendiarism.
Also, ignitable liquids are the first fuel ignited in an estimated 30 percent of all acciden-
tal building fires.6 Many ignitable liquid accelerants are used in excess, and their residues
remain for ready detection by instrumental analysis. Unfortunately, they can be used in
moderation or on combustible substrates like paper or plastic, and the resulting fire can
consume or evaporate all traces of the accelerant. One of the authors participated in a
number of structure fire experiments in which a petroleum product such as gasoline had
been used to ignite a fire, and the debris, even though collected directly after the fire,
packaged properly, and analyzed by the best available methods, failed to reveal any accel-
erant. A negative laboratory finding is not proof that an accelerant was not used; it merely
fails to establish its presence after the fire. On the other hand, petroleum products are
detectable in very low concentrations in a wide variety of consumer products and build-
ing materials, so the interpretation of “trace” levels in debris must take such background
chromatography ■
Chemical procedure
volatiles into consideration (to be discussed further). In addition, petroleum products
that allows the separa- change in their physical and chemical properties as they burn or evaporate during (and
tion of compounds after) a fire, and their specific identification can be a challenge even to experienced labo-
based on differences in ratory analysts.
their chemical affinities
for two materials in GAS CHROMATOGRAPHY
different physical
states, e.g., gas/liquid, Gas chromatography (GC) as an accepted laboratory procedure has been in use for the
liquid/solid. last 50 years or so. Gas chromatography is used both for the screening of samples to
562 Chapter 14 Laboratory Services
determine which contain adequate volatile ignitable liquids for identification and for the
identification itself once the volatile has been isolated. Although gas chromatographs
(called GCs) vary considerably in their complexity, their basic principles of operation are
quite simple. All types of chromatography are a means of separating mixtures of materi-
als on the basis of slight differences in their physical or chemical properties.
As represented in Figure 14-1, gas chromatography uses a stream of gas (nitrogen or
helium) as a carrier to move a mixture of gaseous materials along a long column or tube
filled or coated with a separating compound. The components in the mixture interact
with the separating compound by alternately dissolving in it and then volatilizing to be
swept farther along the column by the carrier. To ensure that all the compounds in the
unknown mixture remain in the gaseous state, the whole column is kept in an oven to
maintain it at a precisely controlled temperature. The mixture is injected onto the column
at one end, and at the other there is a detector, usually a small flame whose electrical
properties, monitored by electronic circuits, change whenever a compound comes off the
column [called flame ionization detection (FID)].
The whole process is analogous to a stream of vehicle traffic routed through a forest.
The motorcycles speed right through, zigzagging around the trees and come out the other
side of the forest first to encounter a radar speed trap. The small cars have more trouble
finding a path and come out shortly afterward to encounter the same radar. The big cars
take a long time to find spaces big enough to allow passage, and they exit after a longer
delay, followed by the trucks that take the longest of all.
In the case of volatile hydrocarbon analysis, the “traffic” would be the unknown
hydrocarbon mixture, and the “forest” would be the GC column. The analyst can adjust
the spacing of the “trees” by choosing the correct chemicals for the separating compound
and carefully controlling the temperature and carrier flow so that each type of “traffic”

Heated injection port

Capillary separation
Syringe column (silica tube)
w/sample

Temperature-controlled oven
2.5
2.0
Detector: Flame ionization
MCounts

1.5
Supply of carrier gas
Mass spectrometer 1.0
(nitrogen or helium) 0.5
Specialized detector 0.0
3 4 5 6 7 8 9
Minutes

Data system
and display

FIGURE 14-1 Schematic of gas chromatograph used for fire debris analysis.

Chapter 14 Laboratory Services 563

 comes out in a neat, readily identifiable group. To extend the “forest-traffic” analogy fur-
ther, the proper choice of GC column and conditions for the hydrocarbons found in fire
debris allow the lab to identify the type of “traffic” introduced by analysis of the groups
of “vehicles” that result. From simple, insulated box ovens that kept the column at a set
temperature, ovens have evolved to models that are programmed to provide steady or
increasing temperatures at preset time intervals to maximize the separation of similar
compounds. The technology of detectors has progressed from simple thermal detectors to
flame detectors that can be made sensitive only to compounds containing specific elements,
or to mass spectrometers.
Although the column can be selected to separate a mixture of compounds on the basis
of their chemical properties, most separations in volatile accelerant analysis are made on
the basis of differences in the boiling points of the various components. Nearly all the com-
ponents of petroleum distillates of interest are either aliphatic (straight-chain) or aromatic
(ring) in structure and thus will have very similar chemical properties within those two
groups. A typical separation of a series of aliphatic hydrocarbons is shown in Figure 14-2.
In this typical chromatogram, one can see that the heavier a molecule is, the less volatile
it is, and the later it will appear on the chromatogram. Columns for GC analysis of sus-
pected arson residues are usually set up to offer the best separation of components hav-
ing volatilities between those of pentane (C5H12) and triacontane (C30H62), since that
range includes all the commonly encountered petroleum products. Although the mainstay
of columns was for many years a packed tubular column, averaging 3 mm (1> 8 in. in
diameter and 3 m (10 ft) in length, it has now been supplanted by a new generation of
very narrow [0.1 to 0.25 mm (0.004–0.01 in.) in diameter] capillary columns of various
lengths.7 Early capillary columns required very long lengths [50 m (167 ft) or more] to
achieve good separations, but these extreme lengths usually required long analysis times.
Such columns were used primarily for research or development, in which short analysis
times could be sacrificed for sensitivity and selectivity. The newest columns provide
extremely high performance on short columns (10 to 25 m, or roughly 33 to 82 ft) using
high temperature-programming rates. The result is analysis times that are shorter than
those possible with packed columns and faster turnaround times. Capillary columns also

Scan Range: 1 –1760 Time Range: 0.00 –14.99 min.

4 ASTM (E1387 ) calibration mixture

A Toluene RIC all astm.sms
B Ethylbenzene
C p-Xylene, m-Xylene C20
D o-Xylene
E 3-Ethyltoluene
F 2-Ethyltoluene C16
3 G 1,2,4-Trimethylbenzene C18
Even-numbered n-alkanes C8–C20 C14
C10
C12
MCounts

C8 G
2
F
FIGURE 14-2 Capillary A E
GC performance standard, C
including C8 to C20 even
n-alkanes: (A) toluene; 1
(B) ethylbenzene;
D
(C) p-xylene, m-xylene; B
(D) o-xylene; (E) 3-ethyl-
toluene; (F) 2-ethyltoluene;
(G) 1,2,4-trimethylbenzene. 0
Courtesy of David Brien, EFI
3 4 5 6 7 8 9
Global, Rocklin, CA.
Minutes

564 Chapter 14 Laboratory Services

 Scan Range: 1 –1760 Time Range: 0.00 –14.98 min. FIGURE 14-3 Gas chro-
matogram of paint thinner.
RIC all paintthinner.sms
2.5 Courtesy of David Brien, EFI
Global, Rocklin, CA.

2.0

1.5
MCounts

1.0

0.5

0.0
3 4 5 6 7 8 9
Minutes

use smaller sample sizes than were ever possible with packed columns, an advantage of
much merit in forensic work, where the sample size is often critically small. Sensitivity is
now in the nanoliter range (10–9 liter).8 Examples of capillary gas chromatograms are
seen in Figures 14-3–14-5. Since gas chromatography is a separation technique, the result
of such GC-FID analysis is not a specific identification of each compound but a pattern
of peaks that the analyst attempts to match to a reference chromatogram to make the
identification. This approach is described in ASTM E1387: Standard Test Method for
Ignitable Liquid Residues in Extracts of Samples of Fire Debris by Gas Chromatograph.9
Due to the limitations of the GC-FID method, ASTM E1387 was withdrawn in 2008 and

Scan Range: 1 –1760 Time Range: 0.00 –14.98 min.

RIC all kerosene.sms

1.5
MCounts

1.0

0.5

FIGURE 14-4 Gas chro-

0.0 matogram of kerosene.
Courtesy of David Brien, EFI
3 4 5 6 7 8 9
Global, Rocklin, CA.
Minutes

Chapter 14 Laboratory Services 565

FIGURE 14-5 Gas chro- Scan Range: 1–1760 Time Range: 0.00 –14.98 min.
matogram of gasoline RIC all gasoline001.sms
(fresh unevaporated)
(TIC). Courtesy of David
Brien, EFI Global, Rocklin, CA.

1.5

MCounts 1.0

0.5

0.0
3 4 5 6 7 8 9
Minutes

was replaced by the GC-mass spectrometry (GC/MS) method ASTM E1618, described in
the next section. GC-FID is still used for special applications such as samples of very low
molecular weight.
In addition, progress in computer technology has dramatically improved the data col-
lection and handling in gas chromatography. Runs (and irreplaceable samples) are no
longer lost when the recording conditions are not set just right, since the data system col-
lects all the data and then displays whatever portion the analyst specifies. The displayed
data can expand or condense portions of the chromatogram to make identification eas-
ier. Collected chromatograms are also filed on disks and can be compared against a
library of stored chromatograms of known products. Such libraries can be swapped
between instruments and even between labs, permitting ready comparison.

GAS CHROMATOGRAPHY/MASS SPECTROMETRY (GC/MS)

Flame ionization detectors (FIDs) are sensitive and stable detectors, but they yield little
mass spectrometry ■ chemical information about the compounds that pass through to produce a signal. Mass
Analysis of organic mol- spectrometry allows the analyst to break apart each compound into small submolecular
ecules by fragmenting pieces and, by counting those pieces, establish the chemical structure of the original mol-
them and separating
them by size.
ecule. Mass spectrometry alone is an analytical technique that needs to be fed one com-
pound at a time, while gas chromatography is a separation technique that is good at
separating things into groups. Because most of the volatile residues in fire debris involve
complex mixtures, coupling the two together gives the best tool. While GC/MS has been
available for more than 30 years, only recently has it been made small, inexpensive, and
user-friendly enough to make it convenient for routine fire debris analysis. It has become
the “baseline” technique for fire debris analysis because it can yield so much more infor-
mation than GC-FID.10 In addition to displaying a mass spectrum for each peak, it can
be asked to scan for selected ions, those that are characteristic for a particular chemical
species of interest, such as the ion having a “weight” of 91, which is characteristic for
toluene. This means that the GC/MS can display a chromatogram just for aromatics, for
instance, ignoring all the peaks that do not yield that specific ion. This can be accom-
plished by displaying one ion at a time or by “summing” three or four ions that are char-
acteristic for the species of interest.11 A typical set of ion profiles for gasoline is shown in
566 Chapter 14 Laboratory Services
Figure 14-6, where the ions for alkanes (ion “weights” of 57, 71, 85, and 99), cycloparaffins/
alkenes (ions 55, 69, 83), simple aromatics (ions 91, 105, 119), and naphthalenes (ions
128, 142, 156) are shown separately. Note that the species responsible for the major
peaks in the gasoline chromatogram (Figure 14-5) can now be identified. The pattern
of peaks, their proportions, and the relative concentrations (higher concentrations of

Scan Range: 1 –1760 Time Range: 0.00 –14.98 min.

Ions: 57+71+85+99 all gasoline001.sms
(alkanes)

100

75
kCounts

50

25

0
3 4 5 6 7 8 9
Minutes
(a)

Scan Range: 1 –1760 Time Range: 0.00 –14.98 min.

Ions: 55+69+83 all gasoline001.sms

80 (alkenes and cycloparaffins)

70

60

50
kCounts

40

30

20

10

0
3 4 5 6 7 8 9
Minutes
(b)

FIGURE 14-6 Typical combined ion profile analyses of gasoline showing alkanes, cycloparaffins, aromatics,
and naphthalenes as separate chromatograms. Courtesy of David Brien, EFI Global, Rocklin, CA.

Chapter 14 Laboratory Services 567

 Scan Range: 1 –1760 Time Range: 0.00 –14.98 min.

Ions: 91+105+119 all gasoline001.sms

800 (aromatics)

700

600

500
kCounts
400

300

200

100

0
3 4 5 6 7 8 9
Minutes
(c)

Scan Range: 1-1760 Time Range: 0.00-14.98 min.

Ions: 128+142+156 all gasoline001.sms

(naphthenics)
17.5

15.0

12.5
kCounts

10.0

7.5

5.0

2.5

0.0
3 4 5 6 7 8 9
Minutes
(d)

FIGURE 14-6 (Continued)

aromatics than other species) yields very discriminatory information. This can make it
possible to pick out characteristic patterns even amid complex chromatograms with
high backgrounds. The “counts” of separate ions also can be compiled into a single chro-
matogram called a total ion chromatogram or TIC, which captures the same information
as a GC-FID analysis and displays it as a single chromatogram for pattern recognition.
Some particularly difficult samples may require target compound analysis (TCA), in
568 Chapter 14 Laboratory Services
which the detection of a “set” of particular compounds permits identification of a petro-
leum product among complex background peaks.12 TCA has also been used to compare
gasolines with suspected sources. Some particularly difficult samples may require
GC/MS/MS, or tandem mass spectrometry. This technique takes the molecular species
created by the ionization process of MS and breaks them into even smaller pieces, yield-
ing more specific information. Higher sensitivity is also claimed, but the method has
found more application in explosives analysis than in routine fire debris analysis.13
Because of its sensitivity and analytical power (resolution), GC/MS has become the
standard forensic technique for fire debris analysis.14
Currently, most forensic GC analyses for suspected accelerants are carried out using
a capillary column of moderate length (10–25 m; 31–76 ft), using dimethyl silicone,
phenyl silicone, or an equivalent nonpolar bonded phase) in a programmed temperature
range of 50°C to 250°C (122°F to 482°F), and employing a mass spectrometer.
Because the MS “detector” has to be turned off for the time period when the extraction
solvent used is passing through, volatiles such as methanol, ethanol, hexane, acetone, or
toluene will not be detected. If such a product is suspected, the lab should be advised so that
a GC-FID analysis can be conducted or an appropriate alternative isolation protocol followed
[such as heated head space or solid-phase microextraction (SPME)], which uses no solvent.

SAMPLE HANDLING AND ISOLATION OF VOLATILE RESIDUES

As described in Chapter 7, the best containers for debris suspected of containing volatile
ignitable liquids are clean, lined metal paint cans, sealable glass jars, or bags of nylon or
a suitable copolymer (e.g., nylon/acrylonitrile/methacrylate made by Ampac). The first
step in the normal examination protocol of such debris is to briefly open the container (in
the case of bags or jars, this will usually not be necessary) and to inspect the contents. The
nature of the contents (wood, carpet, soil, etc.) may provide guidance in selecting an ana-
lytical technique or predicting what interferences are likely to be encountered. The con-
tents should be checked against the labeled or indicated contents to ensure there have
been no errors in packaging or labeling that could compromise the value of the evidence.
The examination must be brief to minimize losses of volatiles that may be present only at
trace levels. Very strong odors of a volatile detected during this examination may indicate
that a particular test protocol should be used. The major problem with fire debris analy-
sis is having a small quantity of product mixed with a great deal of solids or liquids that
may provide contamination. Each of the techniques described here has advantages and
disadvantages that must be evaluated in light of the character of the sample.
Heated Headspace
Since any volatiles present will reach vapor equilibrium in a sealed container, advantage
can be taken of the vapors in the air cavity (headspace) of the container. Heating the sam-
ple container to 50°C to 60°C (120°F to 140°F) will ensure at least a partial vaporization
of heavier hydrocarbons prior to sampling. Some analysts prefer heating the container
above 100°C (212°F) to force even higher boiling materials (such as fuel oil) into the
vapor stage. It is the experience of most other analysts, however, that all but the heavy
ends of kerosene will be detectable in most samples heated to moderate temperatures
(60°C; 140°F) and, more important, that excessive temperatures (⬎80°C; 176°F) encour-
age the degradation of synthetics (carpet, upholstery, etc.) in the debris. These degrada-
tion products add to the complexity of the chromatogram, possibly masking low levels of
real accelerants while not contributing significantly to the sensitivity of the method. If a
heavy product is suspected, isolation using solvent extraction or dynamic headspace will
result in more complete recovery and analysis.
Once the sample container has equilibrated at a moderate temperature, it is punctured
and a small sample (0.1 to 3 mL) of the headspace vapor is drawn off using a gas-tight
syringe. This vapor is then injected directly into the gas chromatograph. For many mixtures,
Chapter 14 Laboratory Services 569
 positive results of this injection may be adequate to permit characterization of the volatiles
present (as described in ASTM E1388: Standard Practice for Sampling of Headspace
Vapors from Fire Debris Samples).15 Positive results also signal the presence of very
volatile residues that may be lost by careless handling, or alcohols or ketones that require
special handling. Because this technique is not as sensitive as others described, a negative
result here merely indicates that no high concentration of volatile exists and that another
isolation technique is suggested. The heated headspace technique is fast, convenient, and
requires virtually no sample handling, while providing some useful information.
Passive Headspace Diffusion (Charcoal Sampling)
Another technique that capitalizes on the vapor equilibrium in a sealed container while con-
adsorption ■ Trapping centrating some of the vapors in a more condensed form is passive adsorption. This method
of gaseous materials on includes several similar techniques that rely on the adsorptivity of activated charcoal. Originally
the surface of a solid
developed by forensic scientists in England using a charcoal-coated wire, this technique
substrate.
requires no manipulation of the sample or its original container.16 As originally described,
the method called for a coated wire to be inserted into the airspace of the sample container
(as shown in Figure 14-7) and allowed to equilibrate for at least 2 hours. The wire was then
removed from the container and inserted into the pyrolysis unit of a GC, where it was raised
instantly to a very high temperature, driving the adsorbed hydrocarbons off into the column.
This technique has been adapted in a variety of methods described in ASTM E1412:
Practice for Separation and Concentration of Ignitable Liquid Residues from Fire Debris
Samples by Passive Headspace Concentration.17 These adaptations include the insertion of
a plastic or glass bead coated with activated charcoal, or a sampling tube or bag containing
charcoal into the airspace of the container.18 The container is resealed and allowed to equil-
ibrate, usually at room temperature, for 12 to 15 hours or at 60°C to 80°C (140°F to 176°F)
for 2 hours. The sampling device is removed, and adsorbed volatiles are extracted from the
charcoal with a small quantity of solvent such as pentane, diethyl ether, or carbon disulfide
for injection into the GC. Today a carbon strip is used. These techniques are very sensitive
to all types of volatiles, require no heating or manipulation of the samples, and can be
repeated again and again because they take up only a small fraction of the volatiles each
time. They are suitable for all kinds of debris and are especially valuable for debris that will
be examined for trace evidence or latent fingerprints.
A variant of this technique has become the single most widely used sampling and con-
centration method for fire debris. In this form, a small strip of activated carbon fiber mat
(sometimes called a C-strip) is suspended in the container, at room temperature or at
moderately elevated temperatures (60°C–80°C). It is then extracted with carbon disulfide
or diethyl ether. The technique has excellent sensitivity [0.1 microliter (␮L) under some
conditions] and no interferences. It is described in full in ASTM E1412.
Solid-Phase Microextraction
A variant of passive activated-carbon techniques, called solid phase microextraction (SPME),
uses a solid-phase adsorbent bonded to a fiber that is inserted into the can of fire debris
through a hollow needle, as shown in Figure 14-8. After the fiber has been exposed to the
heated contents of the can for 5 to 15 minutes, it is withdrawn and then inserted directly
into the injection port of a GC. The heat of the injection port then thermally desorbs the
volatiles directly into the GC column. The technique offers extreme simplicity (with no
manipulation of the sample or any extract) and low cost, since it involves no solvents and
no modifications to existing gas chromatographs.19 SPME been found suitable for fire
debris analysis, and ASTM E2154: Standard Practice for Separation and Concentration
of Ignitable Liquid Residues from Fire Debris Samples by Passive Headspace
Concentration with Solid Phase Microextraction (SPME), applies.20 This technique
appears to be particularly well suited for isolating petroleum product residues from water
and skin tissue.21 The small volume of adsorbent can be overwhelmed by high
concentrations of petroleum distillates, causing distortion of the peak pattern in the GC
570 Chapter 14 Laboratory Services
 Syringe

Protective sheath (hollow needle)

Charcoal strip
(or charcoal-coated wire) Septum or top of container

Fiber of solid-phase adsorbent

Debris

Can with debris

FIGURE 14-7 Passive charcoal strip sampling of FIGURE 14-8 Solid-phase microextraction
fire debris. uses a thin fiber of adsorbent exposed to the
headspace of a debris container. The fiber is
then placed directly into the injection port of a
GC for desorption.

analysis. SPME has the advantage of using no extraction solvent, so all volatiles are read-
ily detectable in a single run. Its limited sample capacity means it is easily overloaded by
extended exposure or high concentration, which can lead to distortion of its GC profile.22
Dynamic Headspace (Swept Headspace)
This technique, sometimes called the charcoal trap method, employs a charcoal or poly-
meric matrix through which the air from the evidence container is drawn. It was adapted
from methods developed for trapping ultralow concentrations of hydrocarbons for envi-
ronmental monitoring.23 The sample container is fitted with a modified lid or a septum
that allows for the introduction of a heated carrier gas (purified room air or nitrogen). The
container is warmed while vapors are drawn off by a low vacuum through a cartridge fil-
ter containing activated charcoal or a molecular trapping agent such as Tenax (see Figure
14-9). Once extraction is complete (20 to 45 minutes), the charcoal can be washed with
carbon disulfide or a like solvent, and the resulting solution is ready for GC analysis. Tenax
or charcoal may be thermally desorbed to avoid the handling of toxic solvents, contami-
nation, or any further dilution. The technique requires the use of traps (commercially pre-
pared or laboratory prepared) and specially built apparatus, described in ASTM E1413:
Standard Practice for Separation and Concentration of Ignitable Liquid Residues from Fire
Debris Samples by Dynamic Headspace Concentration.24 It is very useful because of its
sensitivity (less than 1 ␮L) and its applicability to all ignitable liquid residues (from gaso-
line to fuel oil) including alcohols and ketones. The GC analysis of Tenax traps, which are
thermally desorbed, has been automated, making it more suitable for laboratories with a
large caseload. The technique does have the drawback that all the volatiles in a sample can
be swept out (and pulled through the trap) if the time and temperature conditions are not
monitored closely and are too severe for the sample size. Excessive extraction causes loss
of the lighter fractions, and the process can generally be carried out only once on a sample.
Steam Distillation
The oldest separation technique, originating from classical chemistry, requires only mod-
est but specialized distillation equipment. The specimen is removed from its container and
boiled with water in a glass flask (as illustrated in Figure 14-10). The steam, carrying the
volatiles with it, is condensed and trapped. It is suitable for volatiles that form an
azeotropic (nonmixing) layer with water; therefore, it cannot be used to recover alcohol
or acetone. The technique is suitable for petroleum distillates as heavy as paint thinner;
heavier products (e.g., fuel oil) require the use of ethylene glycol, which boils at a higher
temperature.25 Light hydrocarbons may be lost in recovery. Detection limits are on the
Chapter 14 Laboratory Services 571
 Condenser

Distillation
receiver
with condenser

Preheated To vacuum
carrier gas source

Vapor trap
with charcoal
or other adsorbent Water

Debris
Debris

Heat source Heating mantle

FIGURE 14-9 Swept headspace technique is the most efficient FIGURE 14-10 Steam distillation of arson debris
means of collecting volatile hydrocarbons but requires a source of requires only water, a boiling flask, a heating
carrier gas, a heat source, and a vapor trap of charcoal or other mantle, a condenser, and some form of receiver
adsorbent. It is not considered a nondestructive isolation technique. to collect distillates.

order of 50 ␮L (for gasoline), so it is not as sensitive as charcoal-trapping methods, but it

has the advantage of providing a neat liquid sample with no extraneous solvents.26 The
method is described in ASTM E1385: Standard Practice for Separation and Concentration
of Ignitable Liquid Residues from Fire Debris Samples by Steam Distillation.27

Solvent Extraction
This very direct and simple method for extracting volatiles has been in use for many years.
Debris is extracted or washed with a small quantity of n-pentane, n-hexane, dichloromethane,
or carbon disulfide that has been checked to ensure its purity. The liquid extract is filtered
and then concentrated by evaporation in a stream of warm air until a small quantity
remains for testing, as shown in Figure 14-11. Solvent extraction is especially suitable for

FIGURE 14-11 Solvent extraction

is not very sensitive but may be
used to extract trace volatiles from Liquid solvent
nonabsorbent materials like metal, Filter paper
glass, or stone.
Debris

Evaporating dish

572 Chapter 14 Laboratory Services

treating small, nonabsorbent specimens (glass, rock, metal) or washing the interior of
containers used to transport ignitable liquid accelerants. It provides good recovery for
many petroleum products, especially the heavier ends of kerosene and diesel fuel, but it is
not suitable for isolating the lighter petroleum distillates due to evaporative losses. Its
overall sensitivity is about the same as for steam distillation with the main drawback that
it will extract residues of partially pyrolyzed carpet or foam rubber, which often interfere with
the detection of low levels of hydrocarbons of evidentiary significance.28 This technique
requires only simple glassware and high-purity solvents. It is described in detail in ASTM
E1386: Standard Practice for Separation and Concentration of Ignitable Liquid Residues
from Fire Debris Samples by Solvent Extraction.29

IDENTIFICATION OF VOLATILE RESIDUES

Once the suspected ignitable liquid residues have been isolated, attention can be turned
toward identifying them as precisely as possible. Because recoveries are typically on the
order of several microliters, gas chromatography/mass spectrometry will be the technique
of first choice because of the quantity of information it yields while consuming
1 ␮L or less of the sample. The identification of the volatiles (usually a petroleum distillate
or product) present is then usually based on a comparison of the pattern of peaks gener-
ated in the chromatogram of the unknown against a library of chromatograms of known
materials made under the same conditions. A standardized approach to the classification
and characterization of volatiles is described in ASTM E1618. Classification of the volatile
into of the basic volatility categories—light, medium, or heavy—is based on the presence
of a particular range or sequence of hydrocarbons. The products are further categorized as
to their chemical nature—gasoline, true distillates, de-aromatized distillates, isoparaffins,
aromatics, naphthenic-paraffinics, n-alkanes, oxygenates, and other miscellaneous
volatiles—based on the MS data collected. This classification scheme is intended to make
description and comparisons easier for the laboratory analyst, but because consumer prod-
ucts may represent different classes of products, and a single product may have more than
one consumer designation, the system is not always well understood by the investigator.
For this reason, it is recommended that when an analyst describes an ignitable liquid
residue as a particular classification, examples of some common consumer products should
be included, as listed in Table 14-1, which is based on ASTM E1618.
A growing number of ignitable liquids are not petroleum distillates and therefore do
not lend themselves to a simple distillate classification scheme based on peak pattern
alone; for instance, there are aromatic specialty solvents used in insecticides, enamel
reducers, and pavement sealers, isoparaffinic mixtures used in consumer products such as
“odorless” paint thinners, and naphthenic–paraffinic petroleum products used in lamp
oils (and sold as “kerosene” for portable heaters). The GC-FID patterns of these products
are not sufficient to permit their identification, and GC/MS is required.
Table 14-1 is considerably revised from previous versions because of the influx of
petroleum products in various household materials such as naphthenic-paraffinic prod-
ucts that are not traditional petroleum distillates. Their “peak patterns” in some cases are
not readily differentiated from the patterns generated by some pyrolysis products. Thus,
the use of GC/MS is necessary in a larger percentage of cases, because the GC/MS can
yield information about what specific chemical species are present. For instance, isoparaf-
finic hydrocarbon products are being used in many consumer products, from copier ton-
ers to insect sprays. Because they are not true petroleum distillates, they should not be
classified as such, although in some instances they display a boiling point range that fits
light or medium distillates. When such products are identified, a list of the consumer
products that may contain them should be offered in the report to give the investigator a
clue as to what product may have been used to start the fire or that a product normally
at the scene may have become involved accidentally in the fire. Because the contents of
many products are not always described accurately on the labels, comparison samples
Chapter 14 Laboratory Services 573
 TABLE 14-1 Classification of Ignitable Liquids in Fire Debris

CLASS LIGHT (C4–C9) MEDIUM (C 8–C13 ) HEAVY (C 8–C20+ )

Gasoline (auto) Fresh gasoline in range C4 to C12
Petroleum distillate Petroleum ether Some charcoal starters Kerosene
Some camp fuels Some paint thinners Diesel fuel
De-aromatized distillates Some camp fuels Some charcoal starters Some charcoal starters
Some paint thinners Odorless kerosenes
Isoparaffinics Aviation gas Some copier toners Specialty solvents
Special solvents Some paint thinners Some copier toners
Aromatic products Xylene/toluene Some insecticide solvents Cleaning solvents
Some parts cleaners Parts cleaners
Fuel additives
Naphthenics/paraffinics Solvents Some charcoal starters Some lamp oils
Lamp oils Some insecticide solvents
Industrial solvents
n-Alkane products Solvents Candle oils Copier toners
Pentane/hexane Copier toners NCR papers (carbonless
forms)
Candle oils
Oxygenated solvents Alcohols/ketones Some lacquer thinners
Surfacers Industrial solvents
Some lacquer thinners Turpentines (gum)
Fuel additives
Other: Miscellaneous Single-component products,
blended products
Adapted from ASTM Standard E1618, Standard Test Method for Ignitable Liquid Residues in Extracts from Fire Debris Samples by Gas
Chromatography-Mass Spectrometry, copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken, PA 19428. A copy of the com-
plete active standard may be obtained from ASTM, www.astm.org.

(preferably in their original containers) of any product in the vicinity of the fire’s origin
should be collected and submitted for analysis if there is a chance they could have been
involved in the fire. Because the analyst is expected to match the GC/MS data from the
unknown to that of a known (reference) material, having immediate access to a compar-
ison sample from the scene is very important. This is especially true when the material is
an unusual proprietary or specialty solvent.
This classification scheme reflects changing consumer uses and products and a much-
changed petroleum refining process, as well as improved sensitivity and discrimination of
the lab methods used. There is also a desire to provide the investigator with the most
accurate and meaningful (useful) information. The investigator must strive to understand
what these product designations mean and use that information to evaluate various acci-
dental and intentional sources of such residues in fire debris.
Special detectors may also be used to gain more information about a volatile. Each
requires injection on a separate GC column with a specific detector in place of the FID, but
some GCs have the ability to split a single injection onto two separate columns, each with its
own detector. Photoionization detection (PID), for example, can be used in parallel with FID
because it can be made specific to aromatic, olefinic, or aliphatic hydrocarbons. Another
detector, originally intended for the detection of nitrogen- and phosphorus-containing com-
pounds, has been modified to specifically detect the oxygenated volatiles (ketones, alcohols,
and ethers) that have become widely used as fuel additives to minimize air pollution.30
When distillation is used to isolate a neat volatile, infrared (IR) spectroscopy can yield
considerable information about its chemical structure. About 20 ␮L of sample is needed
574 Chapter 14 Laboratory Services
for conventional liquid-cell spectrophotometry, but a specialized FTIR microcell has been
designed for use as a detector on a GC column when smaller quantities are recovered.
While the sensitivity of the method is limited, it can aid in the identification of unusual
liquids, complex pyrolysis products, and some natural products (terpenes).31
If sufficient quantities of a suspected ignitable liquid are recovered, physical properties
like flash point can be measured. The criminal charges against a suspected arsonist may
depend on the flammability of the accelerant in his or her possession, and a flash point
determination may be required. A precise determination requires a considerable sample and
specialized equipment, but a rough estimate can be made by placing several drops of the
liquid on a watch glass and placing it in a freezer. On removal from the freezer, the glass
will warm up gradually, and a small ignition source passed over the specimen will cause a
flash when the flash point is exceeded. If the liquid has a flash point above ambient (usually
25°C; 77°F), the glass can be placed on a hot plate and the sample tested in the same man-
ner. In rare circumstances, steam distillation has been able to recover enough volatile accel-
erant to permit flash point determination by Setaflash tester (which requires 2 ml of
sample).32 Ultraviolet (UV)/visible fluorescence techniques permit extensive characterization
of petroleum distillates, and these techniques, along with high-pressure liquid chromatogra-
phy (HPLC), may be useful in special circumstances because they provide information that
supplements gas chromatography (especially for heavy petroleum products).33
Evaporation and Degradation
Petroleum products behave in a fairly complex but predictable manner when burned or
simply exposed to air to evaporate. Due to their higher vapor pressures, the most volatile
components evaporate more quickly than those less volatile, and this changes the peak
profile of the product when it is subjected to gas chromatography. As we can see from
Figure 14-12, a gasoline profile changes dramatically as it evaporates, complicating its
identification. When the gasoline is burned, this same process occurs, only much more
rapidly. It is not possible to determine whether a petroleum product burned or merely

2.5 RIC all gasoline001.sms

2.0
MCounts

1.5
1.0
0.5
0
2.5 RIC all 50%wxgasoline.sms
2.0
MCounts

1.5
1.0
0.5
0
2.5 RIC all 90%wxgasoline.sms
2.0
MCounts

1.5
1.0
0.5
0
3 4 5 6 7 8 9
Minutes

FIGURE 14-12 GC/MS comparison of fresh, 50 percent evaporated and 90 percent evap-
orated gasoline. Note decreased amounts of “light” products and relative increases in
heavier components as the gasoline evaporates. Courtesy of David Brien, EFI Global, Rocklin, CA.

Chapter 14 Laboratory Services 575

 evaporated at room temperature by examination of its chromatogram.34 In the case of
gasoline, characteristic aromatic compounds are found in residues of gasoline that has
evaporated down to less than 3 percent of its original volume. Therefore, evaporated or
“weathered” gasoline can still be identified if those key or target compounds are identi-
fied in the debris, as illustrated in Figures 14-13 and 14-14. Unfortunately, this is not true

Scan Range: 1–1760 Time Range: 0.00–14.99 min.

Ions: 57+71+85+99 all 50% wxgasoline.sms

250

200
kCounts

150

100

50

0
3 4 5 6 7 8 9
Minutes
(a)

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 55+69+83 all 50% wxgasoline.sms

100

75
kCounts

50

25

0
3 4 5 6 7 8 9
Minutes
(b)

FIGURES 14-13A–D Combined ion chromatograms of 50 percent evaporated gasoline: (a) alkanes, (b)
alkenes and cycloparaffins, (c) aromatics, and (d) naphthalenes. Courtesy of David Brien, EFI Global, Rocklin, CA.

576 Chapter 14 Laboratory Services

 Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 91+105+119 all 50% wxgasoline.sms

1.25

1.00
MCounts

0.75

0.50

0.25

0.0
3 4 5 6 7 8 9
Minutes
(c)

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 128+142+156 all 50% wxgasoline.sms

70

60

50
kCounts

40

30

20

10

0
3 4 5 6 7 8 9
Minutes
(d)

FIGURE 14-13 (Continued)

for simple petroleum distillates like mineral spirits (paint thinners). As these substances
evaporate, their residues assume more and more of the characteristics of the next heavier
class of petroleum distillate. Thus, it is not always possible to conclusively establish the
exact nature of the original product.
In addition, petroleum products exposed to moist garden soil can undergo microbi-
ological degradation. This degradation involves the consumption of both aliphatic and
Chapter 14 Laboratory Services 577
 aromatic compounds, so the peak profile of gasoline, for instance, can be significantly
changed.35 This degradation is minimized if the samples are kept frozen until analysis.
Evaporation and degradation effects combined with contamination from pyrolysis
products in the debris can make the identification of a specific petroleum product from
post-fire debris difficult if not impossible. The increased use of non-distillate petroleum

Scan Range: 1-1760 Time Range: 0.00-14.99 min.

Ions: 57+71+85+99 all 90% wxgasoline.sms

200

150
kCounts

100

50

0
3 4 5 6 7 8 9
Minutes
(a)

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 55+69+83 all 90% wxgasoline.sms

50

40
kCounts

30

20

10

0
3 4 5 6 7 8 9
Minutes
(b)

FIGURES 14-14A–D Combined ion chromatograms of heavily (90 percent) evaporated gasoline: (a) alkanes,
(b) alkenes and cycloparaffins, (c) aromatics, and (d) naphthalenes. Courtesy of David Brien, EFI Global, Rocklin, CA.

578 Chapter 14 Laboratory Services

products (see Chapter 4) has resulted in GC chromatograms that are not so easily clas-
sified or identified. Examples in Figure 14-15 show the complex nature of these prod-
ucts, making GC/MS analysis, and comparison standards essential. In most cases, the
laboratory can be expected to characterize the product or its general family—for exam-
ple, light (petroleum ether), medium (paint thinners), or heavy (kerosene/diesel) petro-
leum distillates; lacquer thinners; and so on—from residues left in the debris unless the

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 91+105+119 all 90% wxgasoline.sms

1.25

1.00
MCounts

0.75

0.50

0.25

0.00
3 4 5 6 7 8 9
Minutes
(c)

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

Ions: 128+142+156 all 90% wxgasoline.sms

70

60

50
kCounts

40

30

20

10

0
3 4 5 6 7 8 9
Minutes
(d)

FIGURE 14-14 (Continued)

Chapter 14 Laboratory Services 579

 Scan Range: 1–1760 Time Range: 0.00–14.99 min.

RIC all lampoilmedallion.sms

MCounts 3

0
3 4 5 6 7 8 9
Minutes
(a)

Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

RIC all turpatine.sms

6

4
MCounts

0
3 4 5 6 7 8 9
Minutes
(b)

FIGURES 14-15A–C
Total ion chromatograms of petroleum products:
(a) lamp oil, (b) gum turpentine, and (c) pyrolyzed polyethylene.

580 Chapter 14 Laboratory Services

 Scan Range: 1–1760 Time Range: 0.00 –14.99 min.

RIC all polyethylene.sms

700

600

500
kCounts

400

300

200

100

0
3 4 5 6 7 8 9
Minutes
(c)

FIGURE 14-15 (Continued)

liquid fuel has been nearly or totally consumed. Any petroleum distillate, when used as
an accelerant, can be burned away completely, leaving no readily detectable traces if the
fire is hot enough and long enough, and the accelerant is exposed directly to its effects.
Fortunately, in most incendiary-caused fires, accelerants are used in excess, and the
resulting fires do not burn with sufficient intensity to completely destroy them.
Laboratories are often asked to determine the presence of petroleum distillate accel-
erants from the soot accumulated on window glass taken from the scene. As we discussed
in Chapter 7, the appearance of the soot is most directly linked to the state of combus-
tion in the room and the amount of ventilation. Fuel-rich fires produce heavy soot; lean
or well-ventilated fires produce little or none. The synthetic fabrics and rubbers that con-
stitute a large percentage of the fuel load in modern structures produce an oily soot very
similar chemically to that produced by petroleum distillate accelerants. It has been sug-
gested that there should be chemical species in the soot produced by gasoline that are
characteristic for that fuel and are not produced by any “normal” fuel. It was once
thought that the combination of bromine and lead in leaded automotive gasoline would
produce soot with a unique elemental profile. Unfortunately, some carpets contain both
elements, so there is a potential for background interference.36 More important, lead and
its related additives have been phased out of modern fuels, making them very unlikely to
be detected. There are several organic molecules, chiefly polynuclear aromatic hydrocar-
bons that appear to be produced by burning gasoline but are not produced by burning
synthetic materials.37 The most extensive work done in this area is that of Pinorini et al.,
who used GC combined with elemental analysis, X-ray diffraction, and scanning electron
microscopy (SEM) to characterize soot from various ignitable liquids and from synthetic
fibers and materials.38 They discovered that under controlled conditions, soot from the
liquids could be distinguished from that produced by solid fuels using multiple tech-
niques. Unfortunately, the presence of water from suppression produced nonreproducible
condensation of the soot and limited the discrimination value of even the combination of
techniques. Some analysts have reported that freshly made carbon soot will adsorb petro-
leum distillate fumes from the room air just as activated charcoal adsorbs them in laboratory
Chapter 14 Laboratory Services 581
 extractions. If a fire occurs and a too-rich mixture of fuel is present, the soot formed dur-
ing the early stages may retain enough of the accelerant to permit identification upon
extraction of the soot. Occasional analysis of soot on glass from accelerated vehicle fires
has been observed, but no systematic study has been reported.
Despite partial evaporation, most common petroleum distillates have a distinctive GC/MS
pattern that can be compared with the patterns of known compounds. The experienced ana-
lyst will know what the effects of fire and evaporation can be and will take them into account
when comparing chromatograms. All petroleum distillates are overlapping “cuts” or fractions
of a continuous spectrum of compounds. Therefore, it is possible for a complex hydrocarbon
mixture to be sufficiently affected by fire that its chromatogram cannot readily be identified.
It could, for instance, represent a heavily evaporated product such as kerosene, or a relatively
undamaged product like a heavier solvent. Today, the introduction of biofuels or synthetic
fuels further complicates the interpretation issue. Some of these have unusual gas chromato-
graphic profiles that do not readily fall into the existing classification schemes.39

Source Identification
As to the identification of a specific source for a product, the range of techniques avail-
able now enables the analyst to rule out many potential sources and thereby shorten the
list of “possibles.” The sophistication of GC/MS and today’s data handling systems make
it possible to recognize delicate differences in the patterns of GC peaks between products
that were thought to be indistinguishable. Determining the significance of these differ-
ences is entirely another matter, requiring extensive field sampling and validation of the
reproducibility of patterns detected. As with many mass-produced materials that start
from different feedstocks in different manufacturing facilities, changes can be produced
by many factors. To complicate matters, when gasoline, for instance, is distributed, it
mixes with gasoline from other sources already in storage and transportation facilities,
and its composition changes. Under some circumstances, such as having a liquid gasoline
sample that has not undergone significant burning or evaporation, and a small number of
potential sources, it is possible to state with some confidence that certain sources can be
completely excluded and that another represents a source that is indistinguishable by any
of its individual features.40 Rather than give the investigator a false lead with a possible
identification, the laboratory will usually suggest several alternative identifications in its
report. The investigator can evaluate the suggestions and determine which, if any, fit the
circumstances best. If the investigator does not clearly understand the identification or
characterization, the problem should be discussed with the laboratory before proceeding.

INTERPRETATION OF GC RESULTS
It should be noted that most of the newer isolation techniques, especially when combined
with capillary column GC/MS, are more sensitive than the techniques used less than a
decade ago. In many cases they are more sensitive than the human nose, often relied upon
to screen for the best debris samples at the scene. This means that sample collection may
depend not only on the detection of unusual odors but also on the presence of suspicious
burn patterns and on aids such as canine detection. Comparison samples of floor cover-
ings or other fuels involved are much more desirable now than ever before because of the
contributions made by the pyrolysis of synthetic materials, binders, and adhesives, which
can interfere with the analyis.41 Recent work has been published by Stauffer exploring the
chemical mechanisms of polymer pyrolysis and the contributions each type can make to
the GC profile.42 This work has improved the discrimination of pyrolysis products and
petroleum products substantially (see Figures 14-15a–c).

Sensitivity
As the minimum detection level of GC/MS has decreased, analysts have begun to see just
what a petroleum-based world we live in. From the volatile products in roofing paper, copy
582 Chapter 14 Laboratory Services
paper, carbonless forms, magazines, and newspaper print, the solvents in glues, and adhe-
sives used in floor coverings and footwear, to residues of dry cleaning solvents, insecticides,
and even cleaning agents (like the limonene used in “green” grease removers), petroleum
products can be detected in all manner of consumer goods today.43 Without specific iden-
tification of the product and comparison against a reference sample, the mere presence of
traces of a volatile petroleum product in fire debris may well be meaningless. Much of the
sensitivity issue came to light when canine detection was being debated. A properly trained
canine nose can detect petroleum products at very low levels with a high degree of accu-
racy, sometimes at concentrations below even what a qualified lab can properly identify,
so there may be samples in which a canine correctly detects traces of gasoline that the lab
cannot confirm if it adheres to the accepted standard practices. But the issue is not simply
sensitivity; there are far too many other possible (innocent) sources of petroleum products.
It is the accuracy of the identification of the actual product present that is essential. A
canine can signal only yes or no; it cannot identify whether it is a trace of the insecticide
(with its aromatic blend solvent) the owner used the week before or gasoline that an arson-
ist used to set the fire. Only the laboratory with its verified and accepted methods and cri-
teria can distinguish those products. The canine cannot discriminate between the residues
of carpet glue used to repair a section of carpet from the residues of isoparaffinic solvent
used as an accelerant. Despite the advice of the IAAI Forensic Science Committee and the
NFPA Technical Committee on Fire Investigations (NFPA 921), a number of courts have
admitted evidence of canine alerts as indications of the presence of ignitable liquid accel-
erants in the absence of laboratory identifications.44 Studies have shown that even the best-
trained canine teams cannot discriminate all possible pyrolysis products and background
contaminants from accelerants and will alert to a percentage of targets that really do not
contain ignitable liquid accelerants.45 Without a specific verifiable identification of exactly
what is present, the investigator cannot decide the significance of such positive alerts. It is
wrong to assert the absolute reliability of a canine alert.
Another issue of extreme sensitivity is erroneous reconstruction. If someone steps in
a puddle of gasoline leaking from an electrical generator or spilled from a fuel can out-
side the scene and then walks through the scene, traces theoretically might be found in
several adjacent locations. A study by Armstrong et al. demonstrated that detectable gaso-
line residues could be tracked onto clean substrates for a step or two, but not beyond.46
Similar results have been recently reported by Kravtsova and Bagby.47 Transfers by
absorption onto clothing of gasoline vapors in a room have also been shown to be
remotely possible but very weak.48 It has also been suggested that exhaust gases from
gasoline-powered air fans at fire scenes could contaminate exposed charred material.
Tests have revealed that although such contamination at trace levels is possible, it is very
unlikely to affect waterlogged debris.49 A blind reliance on the mere presence of trace
residues could lead to the erroneous conclusion that gasoline was used in several places
throughout the scene.
The investigator must ask: Why was this found in these concentrations, and how else
might it have gotten there? Fire investigators and forensic analysts need to ask: What is
the meaning of this “positive” at the concentration found here? With today’s sensitive
techniques, the soil gathered from the side of a busy highway will yield identifiable gaso-
line residues at parts-per-billion levels, just from the traces accumulated from leaks and
spills from passing vehicles. Is that finding of any significance? Does that mean that some-
one poured gasoline along the side of the road and set fire to it to commit arson? Lentini
demonstrated that floor-coating solvents (stain, varnish, and oil) were readily detectable
in interior woods for more than 2 years after application.50 Hetzel and Moss detected
waterproofing solvent after more than 2 weeks of severe exterior exposure.51 A recent
case showed that the gasoline used as a solvent for interior floor varnish some 15 years
before the fire was still identifiable when the wood was heated and the volatiles trapped
and analyzed by GC/MS. Scientists and investigators of 25 years ago only dreamed about
Chapter 14 Laboratory Services 583
 the sensitivity that is available today, but intelligence must be applied to interpret findings
accurately. The mere presence of a volatile is no longer sufficient; rather, it is the concen-
tration at which it becomes significant that is critical to its interpretation.52

Comparison Samples
With the sensitivity of today’s detection methods (including canines) and the complexity of
today’s world of petroleum products, it is essential the investigator recognize the critical
importance of comparison samples. If debris is collected at a scene for lab analysis, unless there
is a broken Molotov cocktail or a can of gasoline lying next to the burned area, comparison
samples should be collected (using clean gloves and cleaned tools). These samples should
include the floor covering, any padding or underlay, and the floor beneath, not only near the
point of origin but also at another location in the same room (preferably one protected under
furniture). If an ignitable liquid is suspected, comparison samples should also be collected (in
the original containers, if possible) of any solvents, cleaners, or insecticides that may have been
stored or used in the vicinity. This may involve interviewing the owner, tenant, or maintenance
staff to ascertain what might have been used in the course of business prior to the fire.
If any material is used to absorb suspected liquid residues (such as cotton, flour, Fuller’s
earth, diatomaceous earth, or kitty litter), a comparison sample of the absorbent must be
submitted in a separate, clean, sealed container. Any item or product used for absorbing sus-
pected ignitable liquids must be tested and shown not to contribute any volatiles that would
interfere with GC or GC/MS identification. One commercial product, ILA, was introduced
in 2003 but was later found to degrade upon extended storage and contribute volatile
breakdown products to the GC analysis.53 Contamination is the major concern regarding
use of activated charcoal or hazmat absorption mats. They are such aggressive absorption
media that they are easily contaminated by exposure to vehicle exhaust or other atmos-
pheric sources of hydrocarbons. If foams, wetting agents, or other additives are used during
suppression, they can affect the chromatographic data obtained from the sample but have
not been reported to prevent accurate GC/MS detection and identification.54 The fire inves-
tigator should note the use of any additives in suppression or post-fire hazard control and
at least inform the lab. Getting a comparison sample is strongly recommended.

Detection on Hands
Various methods have been suggested over the years for the recovery of volatile products
from human skin, particularly from the hands of suspects. Dry swabbing and solvent
swabbing have been tried with no success. Volatiles on living skin are rapidly driven off
by evaporation or soak into the epidermal and dermal tissues, preventing “surface”
residues. There have been reports of successful detection of ignitable liquid residues by
enclosing the hand in a PVC plastic or vinyl glove or polymeric bag with a charcoal strip
or SPME collection device and warming the hand gently with a radiant heat lamp (or sim-
ply enclosing the hand in a PVC glove and using the glove as the absorbent).55 For this
technique to be successful, there has to have been a considerable quantity of fuel on the
hand originally, and the collection has to be attempted within 1 to 2 hours of the contact.
This short time frame precludes useful testing of suspects taken into custody several hours
after contact. Similar testing has been more successful on tissues excised from a body dur-
ing postmortem and tested in small vials using C-strips. Gasoline and similar volatiles can
be detected on shoes and clothing for up to 6 hours if the garments are worn, and longer
if they are removed and piled together.56
The very nature of fire, fire debris, collection, and analysis means that no one can actu-
ally quantify the results of fire debris analysis, but from experience, experiments, and test
fires the analyst can establish a feel for the levels detected. The question to be answered then
becomes: Is this material significant at this concentration? The sensitivity and discrimina-
tion power of today’s fire debris methods is a great tool for the investigator, but it is one that
must be used carefully and judiciously if the investigations are to be accurate and just.
584 Chapter 14 Laboratory Services
Chemical Incendiaries
Volatile accelerants are not the only arson evidence that is amenable to laboratory test-
ing. A small but significant percentage of all detected arsons use a chemical incendiary as
an initiator or primary accelerant.57 Fortunately, most chemical arson sets leave residues
with distinctive chemical properties, although physically the residues may be nondescript
and easily overlooked.
Safety flares or fusees are the most commonly used chemical incendiary, presumably
because they are effective, predictable, and readily available. Red-burning flares contain
strontium nitrate, potassium perchlorate, sulfur, wax, and sawdust.58 Their post-fire
residue consists of a lumpy inert mass, white, gray, or greenish-white in color (see Figure
8.23.). It is not soluble in water, but exposure to water turns it to a pasty mush.
The residue contains oxides of strontium and various sulfides and sulfates. Although
strontium is found in all natural formations of calcium, only flare residues contain stron-
tium salts in nearly pure form. Elemental analysis by emission spectroscopy, atomic
absorption, or X-ray analysis will readily reveal these high concentrations of strontium in
suspected debris. The igniter-striker button on the cap and the wooden plug in the base
of the stick are resistant to burning and may be found in the fire debris and their manu-
facturer may be identified by their physical features.59

IMPROVISED MIXTURES
The solid chlorine tablets or powder used for swimming pool purification will react with
organic liquids such as the glycols in automotive brake fluid or some hair dressings to produce
a very hot flame that lasts for several seconds. This reaction occurs after a time delay that
varies with the concentration of the chlorine-based components and the degree of mixing
between the solid and liquid phases. The reaction has been studied and a mechanism for the
production of ethylene, acetaldehyde, and formaldehyde by the action of the hypochlorite on
the ethylene glycol has been suggested and verified.60 The production of a cloud of readily
ignitable vapors would certainly account for the dramatic yellow-orange fireball of flames that
occurs in these reactions, as seen in Figure 16.11. The starting solid compounds are usually
not completely consumed and are not immediately water soluble and may be visibly detected
after the fire. The presence of hypochlorite salts can be detected by infrared (IR) spectrometry
or chemical spot tests to confirm the identity of the oxidizer. Debris containing residues of such
chlorine-based products may give off the odor of household bleach when wetted.
Potassium permanganate is used in a variety of chemical incendiary mixtures, includ-
ing a hypergolic (self-igniting) mixture with glycerin. It is soluble in water and can be
washed away. Even a weak water solution of it, however, will have a pronounced red,
green, or brown color, depending on the oxidation state of the manganese ion. Such
chemical ions are readily identified by chemical test or by IR spectrometry if even a
minute chip of the solid chemical is recovered.
Flash powders contain a fine metallic powder, usually aluminum, and an oxidizer
such as potassium perchlorate or barium nitrate. Elemental analysis of the powdery
residue left after ignition will reveal the aluminum, chlorine, potassium, or barium. The
perchlorate or nitrate residues are detectable by micro chemical test if any trace of the
mixture remains unreacted.61 Mixtures of oxidizers (chlorates, nitrates, perchlorates, or
sulfates) have been shown to produce extremely hot fires with metallic aluminum or mag-
nesium powders as fuels. With a suitable binder to control combustion rates and prevent
aerosolization, these materials will burn progressively and have been used for decades as
solid rocket fuels (propellants). In most rocket fuel applications the binder is a rubber, but
improvised mixtures have used diesel fuel as a binder. These mixtures have sometimes
been referred to as “high-temperature accelerants.” Tests have demonstrated that such
mixtures produce extremely large volumes of hot gases, very high heat fluxes, and
Chapter 14 Laboratory Services 585
 extremely high temperatures (of both flame and gases). Even large compartments have
been brought to flashover conditions in 2 to 3 minutes by the combustion of about 110 kg
(250 lb) of an improvised mix of aluminum and magnesium with potassium nitrate and
diesel fuel.62 Sulfates (of calcium or barium) will also act as oxidizers.
Potassium chlorate is easily identified by chemical or instrumental tests if it can be
recovered in its unreacted state. When it reacts in an incendiary mixture, however, its
residues are almost entirely chloride salts, which are innocuous and which if recovered
from debris, could not be readily identified as having come from a device.
Sugar reacts with sulfuric acid (H2SO4) to leave a puffy brown or black ash of ele-
mental carbon. It may have a faint smell of burning marshmallows, but for the most part
it is innocuous in appearance and is easily overlooked. The acid present may char wood,
paper, or cloth in its concentrated form. It is very soluble in water, but mineral acids like
H2SO4 do not evaporate, and they remain active and corrosive for some time. A burning
sensation in the skin on contact with debris indicates that the debris should be checked
for acids. The pH (acidity) can easily be measured, and chemical tests for sulfate or nitrate
ions carried out. Reactive metals such as potassium or sodium can be used as an ignition
device because they generate great amounts of heat and gaseous hydrogen on contact with
water. The residues will contain potassium hydroxide or sodium hydroxide (lye)—both
very strong caustics that can cause skin irritation and burns. The pH of water containing
such residues will be very alkaline (basic), and chemical or spectrophotometric tests will
reveal the metal present. It should be noted that the residues from these and similar
devices are very corrosive to metal. Debris suspected of containing such corrosives should
be placed in glass jars with plastic or phenolic lids, or in nylon or polyester/polyolefin
bags—not in metal cans, paper bags, or polyethylene bags.
Because white phosphorus ignites on contact with air, it has been used as an incendiary
device in the past. (Fenian fire is a suspension of crushed white phosphorus in an organic
solvent. When the container is smashed, the solvent evaporates, exposing the phosphorus to
air with immediate ignition.) Elemental analysis of the debris may reveal an elevated con-
centration of phosphorus, and there are chemical spot tests, although it tends to be a diffi-
cult element to detect and confirm. A common source of phosphorus in incendiary mixtures
is the household match. Characterization of the components in match heads can be carried
out by elemental analysis alone or combined with microscopic analysis using SEM.63
As described in Chapter 12, a variety of heat- and gas-producing chemical reactions can
occur by accidental or intentional combination of household products. Many of these pro-
duce only enough heat to evaporate water to create a steam explosion, with no fire or flame.
Typical of these are lye (drain cleaner), aluminum foil and water, dry bleach, sugar and water,
or acid and metal. Still others create flammable gases like acetylene (calcium carbide and
water) or hydrogen (mineral acid and zinc) that if properly ignited, can cause fires as well as
explosions. Residues include unreacted starting materials, metallic oxides, chlorides, or sul-
fates. The heavy soot of acetylene/air combustion may be visible on adjacent target surfaces.
Many so-called bottle bombs or acid bombs involve such materials. They are capable of pro-
ducing high temperatures and high-pressure mechanical explosions. They can cause fragmen-
tation injuries, chemical and thermal burns, as well as property damage. Many states and the
federal authorities include such chemical mixtures in their explosive device criminal codes
today. Fingerprints may survive on the typical 2-L plastic bottles used (Figure 14-16).

LABORATORY METHODS
Recent advances in high-pressure liquid chromatography (HPLC) and the related tech-
nique of ion chromatography have made it possible to identify numerous inorganic ions
in a water extract of post-fire (or post-blast) debris. As with GC, the characterization of
an ion of interest is based on its detection by a suitable detector at a particular retention
time as it flows through a separatory column. Single-column analyses are now capable of
586 Chapter 14 Laboratory Services
 FIGURE 14-16 Bottle
bomb remains may bear
fingerprints and DNA.
Courtesy of Wayne
Moorehead, Forensic
Consultant.

separating and detecting all the significant anions: Cl-, ClO3-, ClO4-, NO2-, NO3-,
SO42-, etc. Others can separate the cations Na+, K+, NH4+, and even sugars, commonly
found in explosive and incendiary materials.64
Many improvised chemical incendiaries are described in the literature (too many to
describe here).65 Whenever such an incendiary is suspected, samples of nearby materials
should be collected for comparison. Sometimes, it is not the presence of a particular element
or ion but an elevated concentration with respect to that found in the background that
betrays an incendiary. Residues of most chemical incendiaries (sulfates, nitrates, chlorides)
are widely dispersed and innocuous in appearance, so selection of “negative control” areas
is not as straightforward as selection of such areas for liquid accelerants. Random selection
may be necessary, with chemical analysis revealing areas where no residues are found. Many
residues are water soluble, so exposure to rain or suppression water may compromise their
detection. Aluminum metal–fueled mixtures have been seen to produce microscopic spheres
of aluminum and aluminum oxide, which are visually unlike the residues of unaccelerated
combustion (oxidation) of aluminum. SEM may assist identification in such cases.66

General Fire Evidence

Considerations related to laboratory analysis of fire evidence include the following:
■ Identification of charred or burned materials
■ Burned documents
■ Failure analysis by forensic engineers
■ Evaluation of appliances and wiring
■ Miscellaneous laboratory tests
■ Spoliation

IDENTIFICATION OF CHARRED OR BURNED MATERIALS

One of the most common questions a fire investigator asks is, What is (or was) this stuff—
and does it belong at the scene? In earlier chapters we discussed the importance of look-
ing for things that appear to be out of place. In a fire, many items, both trivial and
Chapter 14 Laboratory Services 587
 FIGURE 14-17 A sus-
pected incendiary device
can be reassembled to
help establish the nature
of the container and aid
in search for fingerprint
and trace evidence.
Courtesy of the State of
California, Department of
Justice, Bureau of Forensic
Services.

significant, can be sufficiently damaged that a casual examination will not result in their
identification. Plastic, rubber, or metal objects having a relatively low melting point are
typical problems. The identification may be based on the shape of the remaining portion,
X-rays, portions of labels or details cast into the object that survive, or a determination
of the type of material present. Even though the plastic or metal can be chemically iden-
tified, if only minute amounts remain unburned, this information may not be of use
in establishing the specific item. Some materials, like polyethylene, polystyrene, or zinc
diecast, seem to be universal in their applications. Some plastic bottles may be identified
by their shape or casting marks, but the knowledge that a puddle of debris is melted
polyethylene is of little use in figuring out which of thousands of polyethylene objects it
might represent. Glass objects, whether broken by mechanical or thermal shock, can be
pieced back together (as shown in Figure 14-17) if there has not been excessive melting
of the pieces.
Laboratory identification may consist of visual examination, reassembly, cleaning, or
microscopic, chemical, or instrumental tests of considerable complexity, depending on the
nature of the material, the quantity recovered, and the extent of damage. By verifying the
composition of even partial remains, these services may be crucial in detecting the substi-
tution of less valuable items for heavily insured objects like jewelry, art, or clothes in cases
of insurance fraud, or even in confirming their complete removal in cases of fraud or
theft. If an ignition or time-delay device is suspected, the laboratory may be able to recon-
struct the device. The nature or specific details of the device may then be of use in iden-
tifying the perpetrator, linking him or her with the scene, or in interrogating suspects.
Figure 14-18 is an example of a common household object that careful lab examination
revealed to have been used as a time-delay device (similar to Figure 16-8).

BURNED DOCUMENTS
The reconstruction and identification services described earlier can extend to fire-damaged
documents. Writing paper has an ignition temperature of approximately 250°C (480°F),
and single sheets of paper exposed to fire in air will readily char and burn almost completely.
588 Chapter 14 Laboratory Services
 FIGURE 14-18 Plastic
milk bottle: Was it a
device or innocent
“victim”? (Compare
with Figure 16-8.) Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

However, most documents in files, stacks, or books will not burn completely but will char
and if intact, may be identified. Identification in the laboratory may employ visual exami-
nation under visible, UV, or IR light; photographic techniques; or treatment with glycerin,
mineral oil, or organic solvents to improve the contrast between the paper and the inks.
Some clay-coated papers and writing papers with a high percentage of rag content are more
resistant to fire and are more easily restored than cheaper wood-pulp papers like newsprint.
The most critical problem is the fragility of the charred papers. Even fragmentary
remains can sometimes be identified, but the more intact the document, the better the
chances of identification. When charred paper is to be recovered, the investigator should
take every precaution against physical destruction of the remains. Documents should be
disturbed as little as possible and carefully eased, using a piece of stiff paper or thin card-
board, onto a cushion of loosely fluffed cotton wool. This cotton should be placed into a
rigid box of suitable size and a layer of fluffed cotton wool placed over the documents to
keep them in place. (Rigid plastic containers like sandwich boxes or those used for cot-
tage cheese are excellent for small items like matchbooks.) The box should then be hand-
carried to the laboratory.
Reconstruction of burned documents is best done under controlled laboratory condi-
tions by experienced personnel with proper photographic support in case characters are
visible only fleetingly. For these reasons, the investigator should not attempt to perform
any tests in the field. Virtually anything done to a document in the way of coating, spray-
ing, or treating it with solvent will interfere with laboratory tests. Many general crime
laboratories and qualified document examiners have the expertise and equipment needed
to perform such tests properly.

FAILURE ANALYSIS BY FORENSIC ENGINEERS

If a mechanical system failed and caused a fire, it may be necessary to determine the cau-
sation of the original mechanical fault. A registered mechanical or materials engineer is
often able to tell whether a driveshaft or bearing failed as a result of misuse, overloading,
poor maintenance, or design error. The failure of hydraulic systems may produce leaks of
combustible hydraulic fluid that can contribute to the initiation or spread of a fire, and the
Chapter 14 Laboratory Services 589
 experience of a hydraulic engineer may be required. Because the civil liability of the man-
ufacturer or user of a mechanical system may be in the millions of dollars, it is very impor-
tant to have an accurate estimation of the condition of that system. Such determinations
are beyond the capabilities of most crime laboratory personnel, because they require spe-
cialized knowledge of the materials or processes involved in the product’s manufacture. A
registry of qualified professional engineers is offered in many states to assist investigators.

EVALUATION OF APPLIANCES AND WIRING

The guidelines offered in Chapter 10 will assist the investigator in making an accurate
assessment of the condition of appliances or wiring and the contribution they may have
made to a fire. Many diagnostic signs, however, require microscopic examinations, met-
allurgical tests, or continuity and conductance tests that require laboratory examination
and interpretation by a specialist. For instance, whether a switch was on or off at the
time of the fire requires careful dismantling and examination of the contacts, housing,
and switch mechanism. Due to fire damage, such dissection may easily result in move-
ment of the components and destruction of their precise relationships. X-ray analysis
may be used if the contacts and housing are arranged to permit cross-sectional viewing.
Twibell and Lomas described a novel resin-casting technique that allows cross-sectioning
of the switch unit while ensuring that the components do not move.67 Soft X-rays (of the
energies used for medical procedures, 75 to 85 kV) have also been demonstrated to be
of value in examining thermal cutoffs (TCOs) used to protect coffeemakers and other
similar appliances from overheating, and cartridge-type fuses. Such X-rays can be used
to determine their continuity and often to establish the mechanism of their operation or
cause of failure. The causes for electrical product failures are sometimes revealed by
X-ray examination (as seen in Figure 14-19b). In the case of fuses, those that failed due
to moderate overcurrents were often differentiable from those that failed in a short cir-
cuit.68 X-rays are also useful in establishing the operability of fire-damaged smoke detec-
tors, as shown in Figure 10-55b.
Under some conditions, whether a copper wire was melted by overheating due to
excessive current or by the heat of the fire may be determined by the degree of crystalliza-
tion within the conductor, because the entire cross section of a conductor will be heated
by an overcurrent, while the surface of a conductor is likely to be affected to a greater
degree than the core if the heat is applied from the exterior.
Some years ago, it was suggested that elemental analysis of the surface of some fire-
damaged copper conductors could yield information to determine whether the melting was

FIGURE 14-19A Fire-damaged switch: on/off or defective? FIGURE 14-19B X-ray of switch reveals internal connections.
Courtesy of Helmut Brosz and Peter Brosz, Brosz and Associates. Courtesy of Helmut Brosz and Peter Brosz, Brosz and Associates.

590 Chapter 14 Laboratory Services

the cause or result of the fire. Other researchers, however, suggested that the absorption of
gases from the ambient atmosphere while the metal was molten do not necessarily reflect
the composition of fire gases in the vicinity.69 Others contended that the composition of
fire gases, whether they are absorbed or not, do not discriminate between arcing that
occurs to cause the fire and arcs that occur as a result of external charring.70 Because most
post-fire indicators are the summation of the fire’s effects, testing would have to examine
a wide variety of fire exposures. One landmark paper discussing double-blind tests
revealed that diagnosis of “fire cause” versus “fire effect” using the Auger emission spec-
troscopy (AES) elemental profile method was no better than random chance.71 Currently,
there is no 100 percent reliable means of establishing whether melting of copper conduc-
tors was the result of fire exposure or electric overcurrent unless there are very small glob-
ules created (as shown in Figure 10-47).72 As a general rule, the finer the droplet size, the
higher the energy (and temperature) of the molten copper. Unless there was electric arcing,
the conditions needed to produce aerosol-type beads do not exist in normally fueled fires.
Even then, proof of electrical activity does not prove the fire was started by an electrical
fault (only that the conductor was energized).
Metallurgists can conduct crystallographic tests, but, unfortunately, elemental
analysis requires expertise and equipment available only in fairly sophisticated mate-
rials- or surface-science laboratories. Elemental analysis by scanning electron microscopy/
energy dispersive X-ray spectrometry (SEM/EDS) may be of value in establishing con-
tact between conductors of different metals (copper wire against a zinc-plated steel
receptacle box, for instance) (as shown in Figure 10-51). This finding may help con-
firm a particular accidental ignition mechanism.73 The presence of metallic contami-
nants on insulators in switches and relays has been linked to arc tracking and electrical
equipment fires. SEM/EDS can help identify the nature and origin of such contamina-
tion. Recent work by Carey and Howitt on arc faults between copper conductors in
fire environments has demonstrated the potential value of microscopic and elemental
features.74 Failures such as stress cracking, contamination, microorganism-induced
corrosion (MIC), or erosion require analysis by a combination of microscopic, metal-
lurgical, and chemical techniques. These, too, require specialized resources and knowl-
edge not found in public forensic labs.
It should be kept in mind that although the laboratory specialist can perform tests far
beyond the capability of the investigator, there are practical limits. It is possible for a fire
to be sufficiently hot and prolonged to destroy enough of the diagnostic signs that may
have once been present that the required determination cannot be made with any degree
of certainty. Every laboratory has its quota of miracles that it can perform!

MISCELLANEOUS LABORATORY TESTS

There are a number of other determinations useful in the assessment of a fire that the lab-
oratory is well suited to perform. Usually, very high localized temperatures produced
during the fire may give an indication of the use of an incendiary ignition device or an
unusually high fuel load (if post-flashover conditions are considered). These tempera-
tures may be indicated by the melting (or melted state) of various materials in the vicin-
ity. Because small variations in composition may result in considerable differences in
melting points, the objects of interest should be identified in the laboratory and their
melting points measured directly if at all possible. SEM/EDS has been shown to be of
value in establishing the nature of interactions between aluminum and copper in form-
ing eutectic alloys during a fire.75 SEM can show differences in crystal structure, and
when teamed with elemental analysis (SEM/EDS), it can demonstrate mixing of various
elements (as was seen in Figure 10-51). This finding may be important when melting (of
switch or thermostat contacts) or alloying is found. Such melting or alloying may be the
result of normal fire exposure or the failure mechanism that caused the fire.
Chapter 14 Laboratory Services 591
 Flash Point
The flash points of materials found at the fire scene, either as part of the scene itself or as
part of an incendiary device, should also be determined in a laboratory. Flash point (FP)
determinations are made using one of several types of apparatus:
■ Tagliabue (Tag) Closed Cup (FP below 175°F) (80°C)
■ Pensky-Martens Closed Cup (FP above 175°F) (80°C)
■ Tag Open Cup
■ Cleveland Open Cup
The ASTM has established standard methods for the use of these devices (previously
described), and the reader is referred there for more details.76 Unfortunately, many labo-
ratories cannot measure the flash point of a flammable liquid if that point is below ambi-
ent (room) temperature. In addition, the flash point apparatus most commonly found in
analytical labs requires at least 50 mL (2 fl oz) for each determination. This is far more
than is normally recovered as residue from an incendiary device and precludes such analy-
sis. However, a flash point apparatus called the Setaflash Tester has found acceptance in
forensic laboratories because it will accept samples as small as 2 mL (0.08 fl oz) as well
as permit subambient measurements. It is an approved ASTM test method.77
Melting or Softening Points
The temperatures at which plastics melt, soften, decompose, or char may be of great value
in scene evaluation (see Figure 14-20). Lab methods [including thermogravimetric analy-
sis (TGA) or differential scanning calorimetry (DSC)] are available in many material test-
ing or engineering labs. The melting points of even very small samples can be tested using
a Mettler Thermometric Hot Stage Microscope.
Mechanical Condition
In reconstructing the scene, visual examination is often adequate to ascertain whether a
door lock mechanism was locked or unlocked at the time of the fire (see Figure 14-21).

FIGURE 14-20 The soft-

ening temperature of the
plastic face of this wall
clock, if determined by
lab analysis, could reveal
the temperature of the
smoke layer or amount
of radiant heat to which
it was exposed. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

592 Chapter 14 Laboratory Services

 FIGURE 14-21 Door
lock assembly. Courtesy of
Det. Mike Dalton, Knox County
Sheriff’s Dept., Knoxville, TN.

The lock mechanisms are usually of brass or zinc die cast, both of which have relatively
low melting points. As a result, they will often seize at temperatures produced in normal
structure fires. When recovered at a scene, they should be photographed in place and
then worked to clearly indicate their orientation and photographed in the second posi-
tion. LP-fueled microtorches have been used to melt the internal components of door
locks to permit forcible entry while leaving the external housing intact. Internal (lab)
examination of critical door locks should be considered before assuming that the lock
was normally inoperative.
If a door was not exposed to too much direct fire, it is possible to ascertain whether
it was open or closed at the time of the fire by examining its hinges. In a closed door, the
plates of a typical butt hinge are protected from fire by the door jamb and the edge of the
door. The spine of the hinge will be exposed to the fire and will therefore be more heav-
ily damaged than the more protected plates, as shown in Figures 14-22a and b. In con-
trast, the spine and plates of the hinges of an open door will be exposed to roughly the
same amount of heat, and damage will be uniform on both. If the fire was not too
extreme, discoloration of the metal and remnants of paint on the hinge may indicate the
relative positions of the two plates. It is possible for the fire damage to be severe enough
to erase the diagnostic signs for this determination. Because such fire damage is more
likely near the ceiling of a structure, if it is important to know the position of a door, all
(usually three) hinges from a door should be recovered and labeled as top, bottom, and
middle. Photos of the hinges in place may also be helpful (along with photos of any protected
Chapter 14 Laboratory Services 593
 (a) (b)

FIGURES 14-22A–B As this door hinge is moved, the differential in damage proves it was in the closed position
when exposed to the fire. Courtesy of Greg Lampkin, Knox County, TN, Fire Investigation Unit.

area on the floor or carpet). These determinations, like the others in this section, are
within the capabilities of many experienced laboratory analysts.
Techniques for evaluating soot deposits on the components of smoke detectors have
recently been published.78 The mechanical vibration of an operating detector horn causes
characteristic patterns of soot deposits. These can be used to establish whether the alarm
was sounding during the fire. Soot patterns can also reveal if the battery inside was con-
nected or missing.79
Fire and Smoke Hazards
In accidental fires, the specialist fire laboratory can be of assistance in assessing the con-
tribution made by various materials to the fuel load. Carpets, particularly modern ones
with polypropylene-blend face yarns, can represent a hazard because of their flammabil-
ity (flame spread) properties. They have been shown to be ignitable by small localized
fires and then to spread fire across an entire room over a period of hours. Although fur-
nishings sold in some states must pass flammability tests before they can be sold, this is
not the case nationwide. Furnishings with either synthetic or cotton coverings and cellu-
losic, latex, or polyurethane fillings can therefore present a possible starting point for an
accidental fire, depending on the nature of the ignition source. Furniture with synthetic
594 Chapter 14 Laboratory Services
fabrics and padding can spread fire very quickly, and so their identification may be an
issue in determining why a fire grew so fast that escape was impaired. If any of the cov-
erings or fillings remain unburned, it may be possible to identify them, duplicate them,
and test the combination for susceptibility to accidental ignition (see Chapter 11). Such
testing, although it can be done informally by most laboratories, requires considerable
experience if the results are to be used in civil litigation. Some states have agencies, like
California’s Department of Consumer Affairs, Bureau of Home Furnishings and Thermal
Insulation, that will test such hazards on a contract basis; otherwise, a private consumer
or materials laboratory will have to be contacted. The simple identification of the fiber
constituents of carpets or upholstery, or the rubber (elastomer) in padding may be useful
in assessing the contributions those products might make to ignition or flame spread.
Such identifications can be accomplished by many forensic labs using pyrolysis gas chro-
matography (pyro GC) or Fourier transform infrared spectrometry (FTIR).
As part of a reconstruction of the cause and manner of death in fatal fires it may be nec-
essary to determine the origin of toxic gases that caused death, or of heavy smoke that pre-
vented the escape of still-conscious victims from a small fire. Although the production of
toxic gases and dense smoke is largely a variable of the temperatures and ventilation during
the actual fire, some laboratory tests can be carried out to evaluate potential hazards. Work
of this type has been done to test upholstery materials for closed environments such as air-
plane interiors, but it is costly and time consuming and may not be available to most fire
investigators. The NFPA has developed standard tests for evaluating the fire hazards of inte-
rior finishes, and the reader is referred to its publications for specific guidance.80
Self-Heating Mechanisms
Self-heating of vegetable oils, fish oils, and many polymers as they dry has been known to
lead to ignition of the materials involved (as discussed in Chapter 6). The identification of
vegetable oils from fire debris was a major challenge for laboratory analysts due to the
thermal destruction that often occurs in the vicinity of the ignition source. In addition, the
vegetable oils involved do not have sufficient vapor pressure to be recovered by heating iso-
lation methods. The chemistry of the materials and processes involved have recently been
explained by Stauffer.81 Sensitive new debris analysis techniques have been developed in
which the suspected self-heating oils are extracted with a solvent, concentrated, and chem-
ically derivatized. This process causes their fatty acid components to be rendered suscepti-
ble to GC/MS characterization by the same methods used for ignitable liquid residues
described previously. These techniques have been published by Stauffer, Coulombe,
Reardon, and others and are now used in many forensic labs.82 This same technique may
aid the detection of vegetable oils from potato chips if they are used as an accelerant.
Dryer fire investigations often must take into consideration the possibility of a spon-
taneous fire caused by the self-heating of kitchen oils in poorly laundered towels. Corn,
safflower, and soy oils can all self-heat to the point of open flame when the process is ini-
tiated by the temperatures of a dry cycle. Analysis of the rinse water of the washing
machine that is paired with the dryer is recommended when the dryer contents are sus-
pected of self-heating. While the dryer drum contents may be nothing more than a
charred mass, the washing machine rinse water most likely survived the fire intact, pre-
serving traces of the vegetable oil from the original laundered load. Residues in clothes
washer pumps and drains reportedly have also been identified.83
As described in Chapter 6, differential scanning calorimetry or critical temperature
testing may reveal self-heating propensities of particular combinations of materials. These
spoliation ■ The
tests, however, will require bulk quantities for testing rather than residual amounts. destruction or material
alteration of, or failure
SPOLIATION to save, evidence that
could have been used
Before any testing is conducted, however, the investigator needs to consider what effect a by another in future
test will have on the condition of the evidence. The term spoliation refers to any test or litigation.

Chapter 14 Laboratory Services 595

 examination that will change the condition of the evidence such that other tests cannot
be performed by other concerned parties. In criminal cases, the concept of nondestructive
testing has been well established. An examiner is obligated to retain a sample of evidence
if testing is going to be destructive, document the condition of the evidence completely
beforehand, or ask that a representative of the opposition be present to observe the test.
This principle extends to testing of evidence in civil cases as well. When testing of evi-
dence is likely to change that evidence such that the opposing party (or other interested
party) cannot test it properly, testing should be delayed until all interested parties are noti-
fied and have an opportunity to nominate a representative to take part in the testing.
Failure to notify interested parties may expose the investigator and the examiner to civil
penalties, including denial of admissibility of the results of the testing. Spoliation issues
will be examined in Chapter 17.

Non-Fire-Related Physical Evidence

The most frequent mistake made by fire investigators is to have a single-minded focus on
the fire itself and its origin and ignore other types of physical evidence. It is up to the fire
investigator to construct a complete scenario of a fire setter’s activities at the scene. These
activities are not necessarily limited to the setting of the fire itself. Because arson may be
used to destroy evidence of other criminal acts, the careful investigator must always be
alert to the existence of evidence of “normal” criminal activity.
In particular, the fire investigator needs a basic understanding of fingerprints, blood
evidence, impression evidence, physical matches, and trace evidence. Note that some of
these types of evidence—particularly impression evidence and physical matches—can play
a role in determining origin and cause for accidental as well as incendiary fires.

FINGERPRINTS
fingerprints ■ Unique Many investigators assume, as do arsonists, that a fire destroys all fingerprints. That is
friction ridge patterns not true. While it is true that many fingerprints will be lost during a fire, many will sur-
on the palmar surfaces
vive, even on ignition sources, as illustrated in Figure 14-23. Plastic (three-dimensional)
of the hands and fin-
gers (or their impres- impressions in window putty or patent (visible) impressions in paint or blood may remain
sions). even after direct exposure to fire. In fact, they may be permanently fixed in place by the
heat. Plastic containers used to carry or hold gasoline in an arson fire have been seen to
be softened sufficiently by the gasoline to allow the plastic to take on the fingerprint
impressions of the person holding the container. If such a container is not exposed directly
to the flames, these plastic impressions may survive. The plastic window attacked with a
butane mini-torch shown in Figure 14-24 bore both plastic impressions in the softened
plastic and patent impressions in the soot.
Latent fingerprints, those requiring some sort of physical or chemical treatment to
make them visible, consist basically of five components: water, skin oils, proteins, salts,
and contaminants. The water evaporates rapidly or soaks into absorbent surfaces and is
usually of little value. Skin oils are the residues normally detected by dusting with a soft
brush and fingerprint powder. These oils will remain on hard surfaces for some time or
will soak into absorbent materials like paper. They can be evaporated by elevated temper-
atures, and such “dried out” prints may not respond to dusting. The soot produced by a
smoky fire (such as a gasoline pool) may condense on a glass or metal surface to protect
the latent prints, however, needing only to be gently brushed or washed in a stream of tap
water to reveal the impressions. Exposure to heat may fuse the soot to the ridge details of
the print so the print becomes visible when the excess soot is washed away.
In the last 25 years, dozens of techniques have been developed for the detection of
latent prints that may aid the inspection of fire-related exhibits. Examination using lasers
and high-intensity tunable light sources (including UV and IR wavelengths) makes it pos-
sible to penetrate contaminants and make prints visible against a variety of backgrounds.84
596 Chapter 14 Laboratory Services
FIGURE 14-23 Palm print on soda can “developed” by exposure FIGURE 14-24 A plastic window in a school was attacked with a
to fire. Courtesy of John D. DeHaan. butane mini-torch to permit access. Plastic fingerprint impressions
in the softened plastic and patent prints (in soot) were discovered.
Courtesy of Joe Konefal, Arson/Bomb Investigator, Sacramento, CA.

Cyanoacrylate ester fuming and a variety of fluorescent dye stains and powders further
expand the possibilities of recovery of latent prints.85 All these methods permit the devel-
opment of latent fingerprints on many surfaces thought previously to be incapable of
retaining such impressions. These include surfaces such as leather, wood, vinyl plastics,
smooth fabrics, and even charred paper and metal.
Proteins degrade to amino acids that are detected by reaction with the chemical nin-
hydrin. They are easiest to detect when the latent impressions are on clean, light-colored,
porous surfaces like paper or cardboard. The proteins can be denatured by high temper-
atures so that they no longer react, but if the paper has not been charred by the fire, it
may be worth testing. However, exposure to water (from suppression, from condensation
of water vapor from the fire, or from wet weather) will dissolve the proteins and amino
acids and blur the fingerprints, but the fatty deposits in the latents (skin oils and cosmet-
ics) will not be affected by the water. They can be detected by the physical developer
method. Nonporous surfaces that are or have been wet can be examined using small par-
ticle reagent (SPR). The salts are the most fire resistant of all latent fingerprint residues,
resisting very high temperatures. Like amino acids, however, they are susceptible to water
Chapter 14 Laboratory Services 597
 damage. Salts may react with the inks or coatings on paper or with the metal surfaces of
cans. In doing so, they may leave an identifiable patent (visible) print in corrosion. This
effect may be enhanced by the heat and moisture of the fire, leaving a visible print requir-
ing no further treatment. Other contaminants that can produce patent prints include cos-
metics, food residues, grease, motor oil, and blood.
Fingerprint expert Jack Deans in cooperation with Gardiner Associates (UK) recently
conducted a series of fingerprint tests on objects subjected to observed room fires (most
of which had gone to full room involvement) Deans has successfully recovered latent
prints on plastic bottles (by dusting and by cyanoacrylate fuming), on newspaper (used as
a torch) and matchboxes by physical developer, on the paper used as a wick in a Molotov
cocktail by ninhydrin, and on other plastic surfaces by vacuum metal deposition and
forensic light techniques, and has developed latent prints on plastic and glass bottles with
SPR. He has also reported success in enhancing patent prints in blood and dirt by amido
black treatment and forensic light techniques on a variety of surfaces.86 Air temperatures
in each of these fire tests were measured at 200°C to 500°C (400°F to 1,000°F) at floor
level. Suppression of each fire was by normal water hose stream. Any surface that has
been protected from direct, prolonged flame contact should be considered as a possible
bearer of identifiable prints.87 See Figures 14-25 and 14-26a and b for examples. In addi-
tion, Deans recently reported that the powder from dry chemical fire extinguishers will
develop latents on smooth surfaces.88
Latent prints have even been recovered from human skin. Recovering such prints
requires appropriate skin conditions (smooth, dry, free of hair), and the contact has to

FIGURE 14-25
Enlargement of palm
print detail developed on
plastic trash bag (post-
fire) using cyanoacrylate.
Courtesy of Jack Deans, FFS,
MFSSoc, New Scotland Yard
(Ret.), Fingerprint Consultant,
Gardiner Associates, Ltd.

598 Chapter 14 Laboratory Services

FIGURE 14-26A Cardboard matchbox amid burned debris on FIGURE 14-26B Fingerprints developed by physical developer on
floor. Courtesy of Jack Deans, FFS, MFSSoc, New Scotland Yard (Ret.), inner tray of the same box. Courtesy of Jack Deans, FFS, MFSSoc, New
Fingerprint Consultant, Gardiner Associates, Ltd. Scotland Yard (Ret.), Fingerprint Consultant, Gardiner Associates. Ltd.

have been of adequate duration. Such prints lose detail after a few hours even if the vic-
tim is deceased. If a fire victim has not been dead more than a few hours and the skin was
not damaged by the fire, a search for such latents should be considered.89 There is no
more conclusive proof of identity than a fingerprint, making it worth the time and effort
to search for it. Because of the rapidity with which these new techniques are being put
into service, the investigator should never conclude that latents are impossible to recover
before consulting with a knowledgeable specialist in the field.

BLOOD
Bloodstains left at a crime scene have much greater evidentiary value now than ever
before because of great strides made in analysis of DNA in body fluids, hairs, and tissue.
Above 200°C (392°F), the stain degrades and probably cannot be identified even as
blood. If the proteins in a blood or semen stain (or tissue fragment) have not degraded,
the evidence can be subjected to DNA typing. DNA analysis has become the primary stan-
dard for analysis of blood, tissue, hair, saliva, and other body fluids. Techniques have
been developed for the analysis of increasingly smaller quantities, and databases have
been assembled for convicted felons, arrested suspects, and cold cases. Mitochondrial
DNA allows testing of hairs, bone, and tissue of great antiquity. This testing has become
a major factor in the solution of crimes, as it can dramatically increase the identifiability
of fire victims and fragmentary remains (proven at the Waco and World Trade Center
scenes), as well as the link between evidence at a scene or on a victim and a suspect (even
when no blood is shed).90 Advances in extraction and polymerase chain reaction (PCR)
and short tandem repeat (STR) DNA analysis methods have led to significant improve-
ment in the DNA analysis of even burned bone.91 It is always worth submitting samples
to the lab if there is any question of identity or source. Most forensic labs are equipped
or prepared to conduct such analyses, and databases are rapidly improving for compar-
ing a scene stain against data from large populations. DNA-typable material is much
more resistant to thermal and biological degradation than are the enzymes and protein
factors once used. DNA-typable material (from saliva) was recently reportedly recovered
from the cloth wick from a glass soft-drink bottle subsequently used for a Molotov cock-
tail.92 DNA does not always result in a unique identification, but it does offer a means of
associating a stain or tissue fragment with an individual with considerable certainty. If
Chapter 14 Laboratory Services 599
 there is any doubt about a suspected stain, it should be collected, kept dry and as cool as
possible, and submitted to the laboratory as quickly as possible.
Even in the absence of analysis, bloodstains can be of importance in the reconstruc-
tion of crimes (such as murder or assault) and other fire-related activity (such as escape
attempts). When dried bloodstains are coated with soot and smoke in an ensuing fire,
they are easy to overlook and can be very difficult to interpret even if detected. A num-
ber of chemicals have been used to enhance the visibility of faint or tiny bloodstains at
crime scenes. Luminol, for instance, requires complete darkness and continual respraying
to make bloodstains visible as glowing images (which are very difficult to photograph).
One chemical more recently introduced, called leuco crystal violet (LCV), reacts with
dried blood to form a deep blue-purple stain that is permanent and easily photographed
in ordinary room light. The colorless reagent is applied as a wash or a spray (using appro-
priate safety equipment). The reaction is nearly instantaneous, and the solvent action
washes much of the adhering soot away, enhancing the visibility of the stain pattern.
Blood spatters, shoeprints, and even fingerprints in blood can be recovered by such
enhancement methods. In one case, LCV was used and revealed blood spatters on the
walls of a building more than a year after a murder had taken place and after two subse-
quent fires in the structure.

IMPRESSION EVIDENCE
Often overlooked, impressions of tools on windows or doors can confirm the fact of forced
entry (see Figures 14-27 and 14-28), identify points of entry (or attempted entry), and if
tools are recovered, permit identification of the tools responsible. Such tool marks may
also be found on desk drawers, filing cabinets, locks, or chains at a scene, confirming that
a burglary took place before the fire. The investigator should learn to recognize signs of

FIGURE 14-27 Tool

impressions on the edge
of this “jimmied” door
include both impressed
and striated marks,
making it easier to iden-
tify the tool responsible.
Courtesy of John D. DeHaan.

600 Chapter 14 Laboratory Services

 FIGURE 14-28 Doors
should also be carefully
examined for signs of
forcible entry not consis-
tent with the means of
entry used by firefighters
on the scene, such as this
pried and shattered door
frame. Courtesy of Greg
Lampkin, Knox County TN Fire
Investigation Unit.

forcible entry and be able to account for all such damage at a fire scene. The best kind of
tool-mark evidence is the impression itself. After it has been sketched and photographed,
the object (or suitable portion) bearing the impression should be collected. If its size makes
recovery impractical, a replica may be cast in silicone rubber (Mikrosil) or dental impres-
sion material. Photos of a striated impression are of no use in a laboratory comparison
other than to establish the location and orientation of the mark. The submitted tools are
tested on a variety of surfaces to duplicate the manner of use and the qualities of the mark-
bearing material as closely as possible without risking damage to the tool. The evidence
and test marks are examined under 40⫻ to 400⫻ magnification to try to match the stria-
tions or contours present. See Figure 14-29 for a typical comparison. Tool impressions can

FIGURE 14-29 Tool-

mark comparisons
provide the potential of
linking a suspect and the
tools directly with a crime
scene. Courtesy of the State
of California, Department of
Justice, Bureau of Forensic
Services.

Chapter 14 Laboratory Services 601

FIGURE 14-30A This sooted gas can looks normal. Courtesy of John FIGURE 14-30B The underside reveals numerous penetrations in
D. DeHaan. the bottom made using a knife. These holes made it possible to use
the can as an effective means of spreading the gasoline inside. Tool
marks on the cuts may bear striated features to identify the knife
used. Courtesy of John D. DeHaan.

reveal the nature of the tool used to create or modify incendiary devices, and the manner
of their use, which can be used to prove intent, as illustrated in Figures 14-30a and b.
Impressions of footwear and vehicle tires may be found on the periphery of the fire
scene. Shoe prints have also been linked to forcible entry, as shown in Figure 14-31. The
visibility of shoe prints left as a transfer of dust or light-colored soil may be enhanced by

FIGURE 14-31 A shoe print on the door reveals forced entry despite the modest scorching of wood. Courtesy
of Greg Lampkin, Knox County, TN, Fire Investigation Unit.

602 Chapter 14 Laboratory Services

FIGURE 14-32 Good-quality photograph of a shoe impression at a crime scene includes a scale; low-angle,
oblique illumination; and carefully focused close-up photography. Courtesy of Greg Lampkin, Knox County, TN, Fire
Investigation Unit.

the scorching or charring of walls or doors. Although prints impressed in soil cannot gen-
erally be collected themselves, they are worth photographing in place and possibly casting
with plaster of Paris. Photographs must be taken using low-angle (oblique) lighting to high-
light the three-dimensional details; a scale or ruler must be included in the photo, near (not
on!) and preferably parallel to the impression; and the film plane of the camera must be
parallel to the ground to prevent distortion. See Figure 14-32 for a good example.
Although particular tires can sometimes be matched to their impressions, it is more com-
mon to establish only a correspondence of size and tread pattern. Footwear, due to its char-
acteristic details and slower rate of change, is more amenable to identification and offers
a means of placing the subject at the fire scene.

PHYSICAL MATCHES
One of the few conclusive identifications possible with typical evidence is the physical com-
parison of torn, cut, or broken edges or surfaces with each other to establish a “jigsaw” fit
between them. Pieces of glass from shoes or garments have been matched back to win-
dows, vehicle lamps, or bottles to connect a victim or suspect with a scene. Wrappings
from incendiary or explosive devices have been matched to source materials in the posses-
sion of the suspect, as have pieces of tape, wire, wood, and rope. The evidentiary value of
such positive identifications is very high, and the investigator must be aware of the possi-
bilities when searching a scene for anything that looks out of place or when searching a
suspect’s property or vehicle. Close cooperation and communication between the lab and
the investigator has produced some excellent evidence (see Figure 14-33 for an example).

TRACE EVIDENCE
Trace or transfer evidence such as paint, fibers, soil, or glass can be used to link a suspect
with a crime scene even in arson cases. Paint can be transferred to clothing or tools when
Chapter 14 Laboratory Services 603
 FIGURE 14-33 A physical match showing a jigsaw fit that conclusively links these two pieces of vinyl electrical
tape as having once been a single piece. One piece is from an incendiary device; the other is from a roll of
tape in the suspect’s possession. Courtesy of John D. DeHaan.

a forcible entry is attempted. Flakes as small as 1 mm (0.04 in.) square are adequate for
color or layer sequence comparisons or complete analysis by IR spectroscopy, spectrogra-
phy, or X-ray analysis. Such properties are generally class characteristics; that is, even a
complete correspondence between the chemical properties of known and questioned
paints does not imply that the two paints must have come from the same object. Instead,
it confirms that they came from the same class of objects painted with the same paint.
Multilayered paint chips, however, may be linked to a particular source if they have
enough layers to demonstrate a unique origin.
Glass is found in nearly all structures and can reveal a good deal about the events of
the crime. It can be broken by either mechanical force or thermally induced stress. Curved
fracture ridges on the edges of glass fragments reveal that the breaking force was mechan-
ical. The shape of these conchoidal fracture lines can help establish from which side of the
window the force came. The absence of these lines usually means the fracture was caused
by thermal stress. The presence or absence of soot, char, or fire debris on the broken edges
can establish whether breakage occurred before or after the fire (see examples in Chapter 7).
When a glass window is broken, minute chips of glass are scattered as much as 3 m (10
ft) away, especially back in the direction of the breaking force. If a subject is standing
nearby, minute glass chips may be found in his or her hair, clothes, hat, pockets, or cuffs.
Even very small (less than 1 mm; 0.04 in.) fragments can be compared with a suspected
origin, such as a bottle or window. Once again, this usually results in a class characteris-
tics comparison. Glass is made in large batches that vary little in their physical or chemi-
cal properties. But such transfer evidence does not require direct or prolonged contact with
the source to make it valuable to the investigator.
Fibers can be transferred from the furnishings of the fire scene to the subject or from
the subject’s clothes to the scene, particularly at the point of entry. Such two-way trans-
fers can be very useful in linking suspect and scene, even though fibers are usually another
type of class characteristic evidence. Fibers are compared by microscopic, chemical, ele-
mental, or spectrophotometric tests. Even the most sophisticated analyses must deal with
the fact that synthetic fibers are produced in such large quantities that they are used
mainly as corroborative or circumstantial evidence.
604 Chapter 14 Laboratory Services
 In a similar way, soils from shoes or clothing can be compared with soil from a fire
scene. This may be especially helpful in a wildland fire where a distinctive or unusual type
of natural soil is found in the vicinity of the fire’s origin or along an access road. In urban
environments, artificial contaminants such as slag, cinders, metal filings, or paint chips
may contribute to the uniqueness of a soil. Contact with the soil near a workshop or
industrial site may result in the transfer of soils that are unique and very much different
from the natural soils nearby. Normally, soil comparisons are of marginal evidential
value, but in instances such as those presented here—individual soils, especially artificial
ones—do occur and should not be overlooked.

Chapter 14 Laboratory Services 605

 CHAPTER REVIEW

Summary
The forensic scientists or criminalists who perform the presented by the fire investigator. It is in the role of
examinations discussed in this chapter are a breed suggesting alternative hypotheses and providing data
apart from scientists in research or industrial laborato- to support or refute those alternatives that the labora-
ries. Although they may share the same science or tory can be of maximum value. Forensic scientists can-
engineering backgrounds as the others, forensic scien- not work in a vacuum, however. To do their job
tists relate their work directly to the real world and its properly, they must interact with the fire investigators,
criminal activities. They are uniquely capable of apply- exchanging information and testing ideas and
ing scientific tests to evidence, interpreting the results, hypotheses. Investigators should encourage this kind
and reconstructing a sequence of events that led to the of communication; only in this way can better evi-
production of that evidence. The forensic scientist pro- dence be obtained through fuller use of all that the sci-
vides an independent evaluation of the evidence that entific expert witness has to offer.
should challenge the interpretation of the evidence as

Review Questions

1. Name four sources of forensic laboratory 7. Name three different kinds of chemical incendi-
assistance available to fire investigators. aries.
2. What can elemental analysis offer to a fire 8. Under what conditions can latent prints be devel-
investigator? oped on objects that have been in a fire?
3. What does the gas chromatograph do? 9. What are three ways DNA analysis can assist in
4. What is the difference between GC/FID and an arson investigation?
GC/mass spectrometry? 10. Name three kinds of impression evidence and
5. Which isolation technique is the most widely describe how each can assist in an arson investi-
used for preparing extracts of fire debris for gation.
volatile analysis in forensic labs?
6. Name three of the complications that can arise
when a lab attempts an identification of a
volatile product in fire debris.

References
1. IAAI Forensic Science and Engineering Committee, 5. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire
“Guidelines for Laboratories Performing Chemical and Science Publishers, 2003).
Instrumental Analysis of Fire Debris Samples,” Fire 6. California State Fire Marshal, CFIRS Report,
and Arson Investigator 38, no. 4 (June 1988). Sacramento, CA, March 1995.
2. ASTM E 1618: Standard Test Method for Ignitable 7. A. T. Armstrong and R. S. Wittkower, “Identification
Liquid Residues in Extracts from Samples of Fire of Accelerants in Fire Residues by Capillary Column
Debris by Gas Chromatography–Mass Spectrometry Gas Chromatography,” Journal of Forensic Sciences
(Conshohocken, PA: ASTM, 2006). 23 (October 1978); D. Willson, “A Unified Scheme
3. ASTM E 1492: Practice for Receiving, Documenting, for the Analysis of Light Petroleum Products Used as
Storing, and Retrieving Evidence in a Forensic Science Fire Accelerants,” Forensic Science 10 (1977):
Laboratory (Conshohocken, PA: ASTM, 2005). 243–52.
4. J. D. DeHaan, “Laboratory Aspects of Arson: 8. W. Jennings, Gas Chromatography with Glass
Accelerants, Devices, and Targets,” Fire and Arson Capillary Columns (New York: Academic Press,
Investigator 29 (January–March 1979): 39–46. 1980.
606 Chapter 14 Laboratory Services
 9. ASTM E1387: Standard Test Method for Ignitable 20. ASTM E 2154: Standard Practice for Separation and
Liquid Residues in Extracts of Samples of Fire Debris Concentration of Ignitable Liquid Residues from Fire
by Gas Chromatography (West Conshohocken, PA: Debris Samples by Passive Headspace Concentration
ASTM) (withdrawn 2008). with Solid Phase Microextraction (SPME) (West
10. J. Nowicki, “An Accelerant Classification Scheme Conshohocken, PA: ASTM, 2008).
Based on Analysis by Gas Chromatography/Mass 21. J. R. Almirall, J. Bruna, and K. G. Furton, “The
Spectrometry,” Journal of Forensic Sciences 35, no. 5 Recovery of Accelerants in Aqueous Samples from Fire
(September 1990); ASTM E 1618. Debris Using Solid-Phase Microextraction (SPME),”
11. M. W. Gilbert, “The Use of Individual Extracted Ion Science and Justice 36, no. 4 (1996): 283–87; J. R.
Profiles versus Summed Extracted Ion Profiles in Fire Almirall, J. Wang, K. Lothridge, and K. G. Furton,
Debris Analysis,” Journal of Forensic Sciences 43 “The Detection and Analysis of Ignitable Liquid
(1998): 871–76. Residues Extracted from Human Skin Using
12. P. Sandercock and E. D. Pasquier, “Chemical SPME/GC,” Journal of Forensic Sciences 45, no. 2
Fingerprinting of Unevaporated Gasoline Samples.” (2000): 453–61.
Forensic Science International, 134 (2003): 1–10; J. G. 22. H. Yoshida, T. Kaneko, and S. Suzuki, “A Solid-Phase
Rankin and M. Fletcher, “Target Compound Analysis Microextraction Method for the Detection of Ignitable
for the Individualization of Gasolines,” paper pre- Liquids in Fire Debris,” Journal of Forensic Sciences
sented at AAFS, Washington DC, February 2007; 53, no. 3 (May 2008): 668–76; J. A. Lloyd and
M. R. Williams, M. E. Sigman and R. G. Ivy, P. L. Edmiston, “Preferential Extraction of
“Individualization of Gasoline by GC/MS and Hydrocarbons from Fire Debris Samples by Solid
Covariance Mapping,” paper presented at AAFS, Phase Microextraction,” Journal of Forensic
Washington DC, February 2007. Sciences 48, no.1 (January 2003): 130–36.
13. D. A. Sutherland and K. C. Penderell, “GC/MS/MS: An 23. J. E. Chrostowski and R. N. Holmes, “Collection and
Important Development in Fire Debris Analysis,” Fire Determination of Accelerant Vapors from Arson
and Arson Investigator (October 2000): 21–26; B. J. de Debris,” Arson Analysis Newsletter 3, no. 5 (1979):
Vos et al., “Detection of Petrol (Gasoline) in Fire Debris 1–17.
by Gas Chromatography/Mass Spectrometry/Mass 24. ASTM E1413: Standard Practice for Separation and
Spectrometry (GC/MS/MS),” Journal of Forensic Concentration of Ignitable Liquid Residues from Fire
Sciences 47, no. 4 (July 2002): 736–42. Debris Samples by Dynamic Headspace Concentration
14. E. Stauffer, J. Dolan, and R. Newman, Fire Debris (West Conshohocken, PA: ASTM, 2007.
Analysis (New York: Elsevier/Academic Press, 2008). 25. J. Brackett, “Separation of Flammable Material of
15. ASTM E1388: Standard Practice for Sampling of Petroleum Origin from Evidence Submitted in Cases
Headspace Vapors from Fire Debris Samples (West Involving Fires and Suspected Arson,” Journal of
Conshohocken, PA: ASTM, 2008). Criminal Law, Criminology and Police Science 46
16. J. D. Twibell and J. M. Home, “A Novel Method for (1955): 554–58.
the Selective Adsorption of Hydrocarbons from the 26. S. Woycheshin and J. D. DeHaan, “An Evaluation of
Headspace of Arson Residues,” Nature 268, 1977. Some Arson Distillation Techniques,” Arson Analysis
17. ASTM E1412: Standard Practice for Separation and Newsletter 2 (September 1978).
Concentration of Ignitable Liquid Residues from Fire 27. ASTM E1385: Standard Practice for Separation and
Debris Samples by Passive Headspace Concentration Concentration of Ignitable Liquid Residues from Fire
with Activated Charcoal (West Conshohocken, PA: Debris Samples by Steam Distillation (West
ASTM, 2007). Conshohocken, PA: ASTM, 2009).
18. J. Juhala, “A Method of Absorption of Flammable 28. Woycheshin and DeHaan, “An Evaluation of Some
Vapors by Direct Insertion of Activated Charcoal into Arson Distillation Techniques.”
Debris” (Bridgeport, MI: Bridgeport Crime Lab); 29. ASTM E1386: Standard Practice for Separation and
reprinted in Arson Accelerant Detection Course Concentration of Ignitable Liquid Residues from Fire
Manual (Rockville, MD: U.S. Treasury Department, Debris Samples by Solvent Extraction (West
1980); J. D. Twibell, J. M. Home, and K. W. Conshohocken, PA: ASTM, 2005).
Smalldon, “A Comparison of Relative Sensitivities of 30. P. L. Patterson, “Oxygenate Fingerprints of
the Adsorption Wire and Other Methods for the Gasolines.” DET Report No. 18, October 1990, 8–12.
Detection of Accelerant Residues in Fire Debris,” 31. S. E. Hipes et al., “Evaluation of the GC-FTIR for the
Journal of the Forensic Science Society 22, no. 2 (April Analysis of Accelerants in the Presence of Background
1982): 155–59. Matrix Materials,” MAFS Newsletter 20, no. 1
19. K. G. Furton, J. R. Almirall, and J. C. Bruna, “A Novel (1995): 48–76.
Method for the Analysis of Gasoline from Fire Debris 32. ASTM D3278: Standard Test Methods for Flash Point
Using Headspace Solid-Phase Microextraction,” Journal of Flammable Liquids by Setaflash Closed-Cup Tester
of Forensic Sciences 41, no. 1 (January 1996): 12–22. (West Conshohocken, PA: ASTM, 1996).

Chapter 14 Laboratory Services 607

33. J. Alexander et al., “Fluorescence of Petroleum Fire and Explosion Investigations (Quincy, MA:
Products II: Three-Dimensional Fluorescence Plots of National Fire Protection Association, 2008), 138.
Gasolines,” Journal of Forensic Sciences 32 (January 45. S. R. Katz and C. R. Midkiff, “Unconfirmed Canine
1987): 72–86. Accelerant Detection: A Reliability Issue in Court,”
34. D. C. Mann, “Comparison of Automotive Gasolines Journal of Forensic Sciences 43, no. 2 (1998): 329–33;
Using Capillary Gas Chromatography (I and II),” R. Everett et al., “False Positive and False Negative
Journal of Forensic Sciences 32 (May 1987): 616–28. Rates for Canine Detection of Ignitable Liquids,”
35. D. C. Mann and W. R. Gresham, “Microbial paper presented at AAFS, Seattle, WA, February 2006;
Degradation of Gasoline in Soil,” Journal of Forensic J. D. DeHaan, “Canine Accelerant Detection
Sciences 35, no. 4 (July 1990): 913–23; K. P. Validation,” CAC News, Spring 1994.
Kirkbride, et al., “Microbial Degradation of Petroleum 46. A. Armstrong et al., “The Evaluation of the Extent of
Hydrocarbons: Implications for Arson Residue Transporting or ‘Tracking’ an Identifiable Ignitable
Analysis,” Journal of Forensic Sciences 37, no. 6 Liquid (Gasoline) throughout Fire Scenes during the
(November 1992, 1585–99; D. A. Turner and J. V. Investigative Process,” Journal of Forensic Sciences 49,
Goodpaster, “A Comprehensive Study of Degradation no. 4 (July 2004): 741–46.
of Ignitable Liquids,” paper presented at AAFS, 47. Y. Kravtsova and D. Bagby, “Transfer of Gasoline
Denver, CO, February 2009; Journal of Forensic from Footwear to Flooring Materials: Can This Occur
Sciences (in press). at a Fire Scene?” paper presented at AAFS,
36. J. Andrasko, et al., “Practical Experiences with Washington DC, February 2007.
Scanning Electron Microscopy in a Forensic Science 48. R. J. Kuk and E. C. Diamond, “Transference and
Laboratory,” Linkoping University, NCJRS— Adherence of Gasoline Vapors onto Clothing in an
Microfiche Program, Report No. 2, 1979. Enclosed Room,” paper presented at AAFS, New
37. J. Andrasko, “Analysis of Polycyclic Aromatic Orleans, LA, February 2005.
Hydrocarbons in Soils and Its Application to Forensic 49. T. L. Lang et al., Detection of Exhaust Components
Science,” Linkoping University, NCJRS—Michofiche from Ventilation Fans (Toronto: Centre of Forensic
Program, Report No. 4, 1979. Sciences, 2002).
38. M. T. Pinorini et al., “Soot as an Indicator in Fire 50. J. J. Lentini, “Persistence of Floor Coating Solvents.”
Investigations: Physical and Chemical Analysis,” Journal of Forensic Sciences 46, 6, December 2001,
Journal of Forensic Sciences 39, no. 4 (July 1994): 1470–73.
33–73. 51. S. S. Hetzel and R. D. Moss, “How Long After
39. R. J. Kuk and M. V. Spagnola, “Alternate Fuels and Waterproofing a Deck Can You Still Isolate an
their Impact on Fire Debris Analysis,” Journal of Ignitable Liquid?” Journal of Forensic Sciences 50,
Forensic Sciences, 53, no. 5 (September 2008): no. 2 (March 2005): 369–72.
1123–29. 52. J. D. DeHaan, “Our Changing World: Part 3—
40. J. G. Rankin, J. Wintz, and R. Everette, “Progress in Detection Limits: Is More Sensitive Necessarily More
the Individualization of Gasoline Residues,” paper pre- Better?” Fire and Arson Investigator 52, no. 4 (July
sented at AAFS, New Orleans, February 2005; N. D. 2002).
K. Petraco, “Statistical Discrimination of Liquid 53. D. E. Byron, “An Introduction to the New Ignitable
Gasoline Samples from Casework,” paper presented at Liquid Absorbent (ILA),” Fire and Arson Investigator
AAFS, Washington DC, February 2008; A. T. Barnes (January 2004): 31–32; D. Mann and N. Putaansuu,
et al., “Comparison of Gasolines Using Gas “Alternative Sampling Methods to Collect Ignitable
Chromatography/Mass Spectrometry on Target Ion Liquid Residues from Non-porous Areas Such as
Response,” Journal of Forensic Sciences 49, no. 5 Concrete,” Fire and Arson Investigator 57, no 1 (July
(September 2004): 1018–27. 2006): 43–46; M. Nowlan et al., “Use of a Solid
41. J. D. DeHaan and K. Bonarius, “Pyrolysis Products of Absorbent and an Accelerant Detection Canine for the
Structure Fires,” Journal of the Forensic Science Detection of Ignitable Liquids Burned in a Structure
Society 28, no. 5 (September 1988): 299–309. Fire,” Journal of Forensic Sciences 52, no. 3 (May
42. E. Stauffer, “Concept of Pyrolysis for Fire Debris 2007): 643–48.
Analysts,” Science and Justice 43 (2003): 29–40. 54. E. McGee and T. L. Lang, “A Study of the Effects of a
43. J. J. Lentini, J. A. Dolan, and C. Cherry, “The Micelle Encapsulator Fire Suppression Agent on
Petroleum-Laced Background,” Journal of Forensic Dynamic Headspace Analysis of Fire Debris Samples,”
Sciences 45, no. 5 (2000): 968–89; S. B. Wells, “The Journal of Forensic Sciences 47, no. 2 (March 2002):
Identification of Isopar H in Vinyl Flooring,” Journal 267–74; S. A. Coulson, R. K. Morgan-Smith, and D.
of Forensic Sciences 61, no. 4 (July 2005): 865–72. Noble, “The Effect of Compressed Air Foam on the
44. IAAI Forensic Science Committee, “Position on Canine Detection of Hydrocarbon Fuels in Fire Debris
Accelerant Detection,” Fire and Arson Investigator 45, Samples,” Science and Justice 40, no. 4 (2000):
no. 1 (September 1994): 22–23; NFPA 921: Guide for 257–60; B. S. Geraci, G. Hine, and W. Shaw, “CAFS

608 Chapter 14 Laboratory Services

 and Its Impact in Fire Scene Investigations,” retrieved ‘Cause’ vs. ‘Result’ Beads in Fire Investigations,” Fire
from www.firehouse.com, 2008. and Arson Investigator 48, no. 3 (July 1998): 26–27.
55. J. R. Almirall et al., “The Detection and Analysis of 71. B. Béland, “Examination of Arc Beads by Auger
Ignitable Liquid Residues Extracted from Human Skin Spectroscopy,” Fire and Arson Investigator 44, no. 4
Using SPME/GC,” Journal of Forensic Sciences 45, no. (June 2004): 20–22.
2 (March 2000): 453–61.; I. Montani, S. Comment, 72. V. Babrauskas, “Fires Due to Electric Arcing: Can
and O. Delemont, “The Sampling of Ignitable Liquids ‘Cause’ Beads Be Distinguished from ‘Victim’ Beads by
on Suspects’ Hands,” Forensic Science International Physical or Chemical Testing?” in Proceedings Fire and
194 (2010): 15–124; M. Darrer, J. Jacquemet- Materials 2003 (London: Interscience Communications,
Popiloud, and O. Delemont, “Gasoline on Hands: 2003).
Preliminary Study on Collection and Persistence,” 73. M. J. McVicar, “The Perforation of a Copper Pipe dur-
Forensic Science International, 175 (2008): 171–78. ing a Fire,” The/Le Journal, Canadian Association of
56. P. J. Loscalzo, P. R. DeForest, and J. M. Chao, “A Fire Investigators (June 1991): 14–16.
Study to Determine the Limit of Detectability of 74. N. Carey, PhD dissertation, Strathclyde University,
Gasoline Vapor from Simulated Arson Residues,” Glasgow, November 2009; D. G. Howitt and S. Pugh,
Journal of Forensic Sciences 25, no. 1 (January 1980): “Direct Observation of Arcing through Char in
162–67; Kuk and Diamond, “Transference and Copper Wires,” paper presented at AAFS, Washington
Adherence of Gasoline Vapors.” DC, February 2007.
57. DeHaan, “Laboratory Aspects of Arson.” 75. E. C. Buc, D. J. Hoffman, and J. Finch, “Failure
58. H. Ellern, Modern Pyrotechnics (New York: Chemical Analysis of Brass Connectors Exposed to Fire” in
Publishing, 1961). Proceedings Interflam 2004 (London: Interscience
59. W. L. Dean, “Examination of Fire Debris for Flare Communications), 731–39.
(Fusee) Residues by Energy Dispersive X-Ray 76. O. M. Slye Jr., “Flammable and Combustible Liquids,”
Spectrometry,” Arson Analysis Newsletter 7, 1984. chap. 5 in Fire Protection Handbook, 20th ed.
60. K. P. Kirkbride and H. J. Kobus, “The Explosive (Quincy, MA: NFPA, 2008), sec. 3.
Reaction between Swimming Pool Chlorine and Brake 77. ASTM D 3278.
Fluid,” Journal of Forensic Sciences 36, no. 3 (May 78. C. Worrell et al., “Enhanced Deposition, Acoustic
1991): 902–7. Agglomeration and Chladni Figures in Smoke
61. R. E. Meyers, “A Systematic Approach to the Forensic Detectors,” Fire Technology 37, no. 4 (2001): 343–62;
Examination of Flash Powders,” Journal of Forensic K. C. Kennedy, et al., “Fire Analysis Tool: Revisited
Sciences 23 (October 1978): 66–73. Acoustic Soot Agglomeration in Residential Smoke
62. N. R. Keltner et al., “Investigation of High Alarms” in Proceedings Interflam 2004 (London:
Temperature Accelerant Fires” in Proceedings Interscience Communications), 719–24.
Interflam 1993 (London: Interscience 79. J. Colwell and A. Reza, “Use of Soot Patterns to
Communications), 607–20. Evaluate Smoke Detector Operability,” Fire and Arson
63. J. Andrasko, “Identification of Burnt Matches by Investigator (July 2003): 42–45.
Scanning Electron Microscopy,” Journal of Forensic 80. N. R. DeHaan, “Interior Finish,” chap. 4-1 in Fire
Sciences 23, no. 4 (October 1978): 637–42. Protection Handbook, 16th ed. (Quincy, MA: NFPA,
64. J. M. Doyle, et al., “A Multicomponent Mobile Phase 1991), sec. 6, pp. 6-37–6-47; R. G. Gann and N. P.
for Ion Chromatography Applied to Separation of Bryner, “Combustion Products and Their Effects on
Anions from the Residues of Low Explosives,” Life Safety,” vol. 1, chap. 2 in Fire Protection
Analytical Chemistry 72, no. 10 (2000): 2302–7. Handbook, 20th ed. (Quincy, MA: NFPA), sec. 6.
65. U.S. Army, Unconventional Warfare Devices and 81. E. Stauffer, “A Review of the Analysis of Vegetable Oil
Techniques, TM 3–201–1, May 1966. Residues from Fire Debris Samples: Spontaneous
66. D. Mann, unpublished test data, 2005. Ignition, Vegetable Oils, and the Forensic Approach,”
67. J. D. Twibell and S. C. Lomas, “The Examination of Journal of Forensic Sciences 50, no. 5 (September
Fire-Damaged Electrical Switches,” Science and Justice 2005): 1091–1100.
35, no. 2 (1995): 113–16. 82. E. Stauffer, “A Review of the Analysis of Vegetable Oil
68. J. D. Twibell and C. C. Christie, “The Forensic Residues from Fire Debris Samples: Analytical Scheme,
Examination of Fuses,” Science and Justice 35, no. 2 Interpretation of Results and Future Needs,” Journal
(1995): 141–49. of Forensic Sciences, 51, no. 5 (September 2006):
69. D. G. Howitt, “The Chemical Composition of Copper 1016–32; R. Coulombe, “Chemical Analysis of
Arc Beads: A Red Herring for the Fire Investigator,” Vegetable Oils Following Spontaneous Ignition,”
Fire and Arson Investigator 48, no. 3 (March 1998): Journal of Forensic Sciences 47, no. 1 (January 2002):
34–39. 195–201; A. K. Gambrel and M. R. Reardon,
70. R. Henderson, C. Manning, and S. Barnhill, “Questions “Extraction, Derivatization and Analysis of Vegetable
Concerning the Use of Carbon Content to Identify Oils from Fire Debris,” Journal of Forensic Sciences, 5,

Chapter 14 Laboratory Services 609

 no. 6 (November 2008): 1372–80; L. M. Schewenk 89. D. Graham, “Some Technical Aspects of the
and M. R. Reardon, “Practical Aspects of Analyzing Demonstration and Visualization of Fingerprints on
Vegetable Oils in Fire Debris,” Journal of Forensic Human Skin,” Journal of Forensic Sciences 14 (1969):
Sciences 54, no. 4 (July 2009): 874–80. 1–12; M. Trapecar, “Lifting Techniques for Finger
83. D. Mann, “Washing Machine Effluent,” Fire Findings Marks on Human Skin: Previous Enhancement by
7, no. 4 (1999): 4. Swedish Black Powder,” Science and Justice 49, no. 4
84. P. Margot and C. J. Lennard, Techniques for Latent (December 2009): 292–95.
Print Development (Lausanne, Switzerland: University 90. J. Di Zinno et al., “The Waco, Texas, Incident: The
of Lausanne, 1991). Use of DNA to Identify Human Remains,” Fifth
85. H. C. Lee and R. E. Gaensslen, “Cyanoacrylate ‘Super International Symposium on Human Identification,
Glue’ Fuming for Latent Fingerprints,” The London, England, 1995; A. Z. Mundorff, E. J.
Identification Officer, Spring 1985; E. R. Menzel and Bartelink, and E. Mar-Casslh, “DNA Preservation in
J. M. Duff, “Laser Detection of Latent Fingerprints: Skeletal Elements from the World Trade Center
Treatment with Fluorescers,” Journal of Forensic Disaster,” Journal of Forensic Sciences 54, no. 4 (July
Sciences 24 (1979): 96–100. 2009): 739–45.
86. J. Deans, “Recovery of Fingerprints from Fire Scenes 91. J. Ye et al., “A Simple and Efficient Method for
and Associated Evidence,” Science and Justice 46, no. Extracting DNA from Old and Burned Bone,” Journal
3 (July 2006): 153–68. of Forensic Sciences 49, no. 4 (July 2004): 754–60.
87. S. M. Bleay, G. Bradshaw, and J. E. Moore, 92. “Testing Molotov Cocktails for DNA,” Fire Findings
“Fingerprinting Development and Imaging Newsletter; 12, no. 4 (Fall 2004): 14; J. Mann, N. Nic Daéid, and
Special Edition: Fire Scenes,” Publ. 26/06, Home A. Linacre, “An Investigation into the Persistence of
Office Scientific Development Branch, UK, April 2006. DNA on Petrol Bombs” in Proceedings European
88. J. Deans, personal communication, September 2009. Academy of Forensic Sciences, Istanbul, 2003.

610 Chapter 14 Laboratory Services

 CHAPTER
15
Fire-Related Deaths
and Injuries

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify the common causes (or contributory factors) for fire-related deaths and
injuries.
■ Explain the basic concepts of the team approach used when investigating a fire death.
■ List the six common questions to be asked when remains (human or suspected
human) are discovered at fire scenes.
■ Explain several techniques involved in pathological and toxicological examinations
in fire deaths.
■ Explain the fire conditions and effects associated with the destruction of a body.
■ Understand the role of carbon monoxide in causing deaths in fires.
■ List several examples of other toxic gases produced by fires.
■ Search the Internet for the autopsy protocol published by the U.S. Fire Administration
that is used in firefighter line-of-duty deaths.

For additional review materials, appendices, and suggested

readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

611
 D
eaths and injuries frequently accompany fires. Structure fires, whether accidental or
incendiary, have a high potential for causing human fatalities. Fire is a leading
cause of accidental death in the United States, estimated to have claimed more
than 3,400 civilian lives in 2007 and 2008 each (with 16,700 injured in 2008). Approximately
83 percent of all fire deaths occur in the home (representing a fairly constant percentage in
recent years), with approximately 4 percent occurring in commercial or industrial buildings
and 10 to 11 percent in vehicles.1 Arson is used as both an agent of murder and to destroy
evidence of a previously committed crime.
Every fire investigator must be prepared to deal with fire scenes where there has been
a loss of human life. Figures 15-1 through 15-3 illustrate a fire that was first reported as an
accidental (smoking) fire but turned out to be set to conceal a natural death (for insurance
fraud) that progressed to cause the deaths of a family of four in the apartment above.

FIGURE 15-1 Exterior

of apartment building.
Family of four died in
the upstairs apartment.
Courtesy of Randy Crim, Fire
Marshal, Lake Jackson, TX.

FIGURE 15-2 Interior of

ground-floor apartment
identified as room of
origin. Courtesy of Randy
Crim, Fire Marshal, Lake
Jackson, TX.

612 Chapter 15 Fire-Related Deaths and Injuries

 FIGURE 15-3 Close-up
of recliner with body of
occupant. Area of origin
of fire, ignited by dropped
smoking materials.
Suspected accidental fire
was later determined to
have been set by the wife
of the decedent (who had
died of natural causes) for
insurance fraud. Courtesy
of Randy Crim, Fire Marshal,
Lake Jackson, TX.

Most fatalities in fires are not directly the result of the effects of the flames but rather
of the asphyxiation caused by replacement of breathable air by toxic gases. In fact, about
three times as many victims die from asphyxiation as from the thermal or physical impact of
fire or explosion-caused injuries.2 Subsequent exposure to fire then causes destruction of
the body. No matter what the actual cause of death, the finding of a victim amid the ashes
initiates a new series of investigative steps. Deaths from fires are not always instantaneous;
they may occur hours, days, or weeks after the fire. For this reason, every fire that produces
a serious injury to an occupant or emergency responder (serious enough to warrant hospi-
talization) should be considered a potential fatal fire and should be treated accordingly.

The Team Effort

Every fire investigation involving a death becomes a team effort, and properly so. The fire
investigator, pathologist, toxicologist, radiologist, and odontologist all play vital roles in
the investigation, each to a greater or lesser degree depending on the exact circumstances.
Without their combined expertise, the cause and manner of death cannot be determined,
and the failure of a fire investigator to appreciate the value of forensic medical contributions
will seriously compromise the success of the entire investigation.
Whenever human, or suspected human, remains are discovered at a fire scene, the
investigator must consider six questions:
1. Are the remains human?
2. Who is the victim?
3. What was the cause of death?
4. What was the manner of death?
5. Was the person alive and conscious at the time of the fire? If so, why did he or she
not escape?
6. Was the death due to the fire or only associated with it?
We shall examine each of these points in order.
Chapter 15 Fire-Related Deaths and Injuries 613
 SPECIES OF REMAINS
Unless the remains can be conclusively identified as those of an animal, they should be pre-
sumed to be human, and appropriate personnel (coroner, medical examiner, or sheriff)
should be contacted before anything more is disturbed. It is better to err on the side of cau-
tion. One investigator who used a pocket knife to slice into what was thought to be a pig
carcass in a farm building fire was horrified to learn later that the remains were, indeed,
human. The reverse is not unusual. Seriously burned remains of pigs, deer, even those of large
dogs, have been mistaken for human remains. In many areas, human remains are the sole
province of the coroner or medical examiner by statute, and it is a violation of the law to
disturb them. Crime laboratory personnel can be called upon to make a preliminary assess-
ment, but a physical or forensic anthropologist from a local college or university may be
required to identify badly charred or fragmentary remains. Even fragmentary remains can be
identified as to species if any unburned tissue is present to allow serological tests in the crime
laboratory. DNA has been found to be quite resistant to degradation. Even badly burned
bones and teeth have yielded some DNA of use in establishing species. The investigator
should also consider having the remains of any pets found in the fire examined. The pres-
ence of injuries or bullets or the presence of carbon monoxide (CO) or drugs in the body of
the pet has sometimes been an important clue in the reconstruction of equivocal fire deaths.

IDENTITY OF THE VICTIM

If the remains are human, it is important to determine the identity of the victim. Because
the identification will usually have great bearing on the motive or opportunity in an
incendiary fire, such an identification is essential to any thorough fire investigation. It is
also possible that remains have been substituted for the purposes of insurance fraud or
homicide. In instances of homicide, there must be a positive identification in order to
prosecute the responsible party. The identification offered by personal effects such as jewelry
or documents such as a driver’s license or credit cards should not be relied on except
under special circumstances (such as airplane crashes where the identities of a fixed number
of possible victims are already suspected) and then only if there are no disparities between
known facts and the remains. Although in some rare cases, visual personal identification
can correctly be made by friends or relatives of the decedent, the effects of fire and putre-
faction usually make such “identifications” of very limited reliability. Edema can bloat
faces, and heat shrinks skin, tightening facial features and sometimes taking years off the
estimated age. Heat and soot can also dramatically change the appearance of hair color
and skin color, making misidentifications of even basic factors like age, race, or hair color
more likely. The emotional trauma of dealing with a relative killed in a fire makes visual
personal identification even more unreliable.
For the purpose of identification then, the services of a qualified forensic pathologist,
forensic odontologist, and radiologist may be required. Identifying characteristics may
include the pathological signs left by illness or previously suffered injuries, unique tattoos,
surgical scars, or reconstruction of height and weight by measurement of the long bones
of the limbs.
The best form of conclusive identification is often provided by comparison of the
teeth of the remains with dental charts and X-rays. The teeth, fillings, replacements (pros-
thetics), bridges, and similar features can survive even extensive incineration and offer
conclusive identification in most cases if there are adequate antemortem records (such as
X-rays, impressions, and photographs).3 Porcelain crowns can resist temperatures up to
1,100°C (2,000°F), and dental amalgam resists temperatures up to 870°C (1,600°F).
Some of the newer generations of restorative dental materials degrade at temperatures of
300°C to 500°C (575°F to 930°F).4 Composite resins have been identified as to manufac-
turer by scanning electron microscopy/energy dispersive X-ray spectrometry (SEM/EDS)
even after thermal damage.5 Resins and metal fillings can be damaged by direct fire-exposure
614 Chapter 15 Fire-Related Deaths and Injuries
heat, but the teeth themselves will withstand more severe fire exposure (eventually shat-
tering from failure of the dentin).
Other evidence revealed by the X-rays may lead to an identification, since pacemakers,
orthopedic plates, screws, pins, artificial joints, implants, and heart valves may be custom-
made for a patient and may be traceable. False teeth may be identified by a numbered tag,
microchip, or other feature imbedded in the resin when the plate is made. Previous frac-
tures or unusual skeletal features revealed by radiography have been used to substantiate
identifications of seriously burned bodies. The internal structural features of sinus cavi-
ties, roots of teeth, and other bones have been used to make identifications if the jaws and
exposed teeth are missing and if antemortem X-rays are available.6
In addition to providing information of value in the identification of the victim, the
radiologist is essential to a fire death investigation, for he or she is able to see beyond the
char and destruction and detect both injuries and weapons. Fires may be set to conceal
evidence of foul play, so a careful X-ray examination for the presence of injuries, bullets,
broken bones, or even knife blades is essential. More than one investigator has been con-
vinced that he or she was dealing with an accidental death until moving the body onto
the stretcher revealed a knife still protruding from its fatal wound, or X-rays revealed
shotgun pellets in the torso.
Fingerprints should not be discounted as a means of identification if any of the skin
tissue of the fingers remains. If the identity of the victim is suspected, even a small por-
tion of fingertip skin may bear sufficient detail for comparison. The epidermis of finger
skin can be released by heat-caused skin slippage and may be found in the debris adjacent
to the body. The analysis of remaining tissue for its DNA profile is becoming a practical
reality for identification, particularly when immediate family members are available to
provide blood or tissue samples for comparison with the unknown body. DNA analysis
has been shown to provide accurate identifications even after extreme fire exposure and
post-fire decomposition.7 DNA identifications were made even from fragmentary remains
after the 2001 World Trade Center attack.8

CAUSE OF DEATH
The establishment of cause and manner of death is almost always the province of the med-
ical examiner, coroner, or equivalent official.9 The cause of death is the actual event that
brought death upon the victim. It may be further defined as the event, injury, or disease
responsible for initiating the sequence of events that ultimately produced the fatal result.
Aside from the entire range of homicidal and accidental causes of death, causes can be
quite varied in a fire and include the following:
■ Asphyxiation by carbon monoxide or other toxic gases (the most common)
■ Burns (incineration)
■ Anoxia due to oxygen depletion
■ Internal edema caused by the inhalation of hot gases
■ Hyperthermia from exposure to high temperatures alone
■ Falls
■ Electrocution
■ Trauma from being struck by falling structural members of the building

Because of the importance of asphyxia and burns, these topics will be treated in greater
depth later in this chapter.
In general terms, fire-related causes of death can be linked to the time elapsed
between the fire and time of death. If death ensues very rapidly (seconds to minutes to a
few hours), the cause is often related to heat, smoke, or carbon monoxide. Deaths in the
first day or so are more often associated with shock, fluid loss, or electrolyte imbalance,
while those occurring many days or weeks after the fire are usually precipitated by infections
Chapter 15 Fire-Related Deaths and Injuries 615
 or organ failures. It is in this last set of cases that the link between the death and the actual
cause may be lost. Here the “cause” of death may be listed as an infection, kidney fail-
ure, respiratory failure, toxemia, or even cardiac arrest, but these are mechanisms of
death, not causes. The mechanism of death is the immediate medical event that brought
about the cessation of life. The erroneous listing of one of these mechanisms as the cause
of death obscures the involvement of fire as the actual causal event.

MANNER OF DEATH
The manner of death may be defined as the way or circumstances in which the cause of
death occurred. This is determined by pathological and toxicological findings supported
by a reconstruction of the fire scene and the victim’s activities there. This reconstruction
requires the fullest cooperation between the fire investigator and the pathologist (and
sometimes the criminalist). It is under these circumstances that teamwork is of the highest
necessity. There are five generally accepted manners of death in the United States: natural,
homicidal, suicidal, accidental, and undetermined. (In some jurisdictions, a finding of
death by misadventure or medical intervention is accepted.) Although the largest percent-
age of deaths at fire scenes will be accidental, investigators should never prejudge the
manner of death any more than they should prejudge the cause of fire itself.

VICTIM STATUS AT TIME OF DEATH

Carbon monoxide levels, indications of asphyxiation and the degree of inhalation of flam-
ing gases, and the presence of toxic combustion products in the blood or tissues and soot
in the airways all play important parts in establishing the manner of death, because they
can help reconstruct the circumstances at the time of the fire as well as help answer the
question: Was the victim alive and conscious during the fire? If the victim was conscious,
what circumstances prevented escape? Barred windows, inadequately marked exits, men-
tal confusion due to drugs, alcohol, age, or partial asphyxiation; incapacitation by irritant
smoke; obscuration of exits by heavy smoke; physical restraints; or physical disability—
all are possibilities that must be considered by the investigator. The clothing worn by the
victim and the position and location of the body may also be clues to his or her activities
at the time of the fire. Items found near the body—a fire extinguisher, flashlight,
telephone, purse, car keys, and the like—may provide a clue as to what the victim was
doing or attempting to do when overcome.

DEATH DUE TO FIRE VERSUS DEATH ASSOCIATED WITH FIRE

Only a comprehensive investigation including a full pathological and toxicological exam-
ination can determine if the death was due to the fire, as we will see shortly. But fire can
be part of a ritual or retaliative crime, in which case the victim is already dead due to
other causes, or it may be the result of a completely accidental sequence of events. The
cause of the fire and the cause of death may or may not be related. The relationship
between the victim and the fire may be determined by assessing both in terms of the
spread of fire and its heat and smoke. Accidental fires may occur whether the death is
accidental, suicidal, natural, or homicidal. Similarly, an arson fire may trigger an acciden-
tal or natural death or be part of a homicide. Factors such as smoke detector activation
(or, usually, lack thereof) may play a major role in the reconstruction.10

Pathological and Toxicological Examination

GENERAL CONSIDERATIONS
Although even partial destruction of a body by fire complicates the work of the patholo-
gist, it is a very rare fire that precludes the discovery of at least some information useful
to the investigator in answering the questions outlined earlier. In short, every body recovered
616 Chapter 15 Fire-Related Deaths and Injuries
from a fire, no matter how badly burned it is, should be X-rayed and given a proper foren-
sic postmortem examination. Precautions should be taken at the scene to ensure that all
remaining relevant evidence is preserved. The body should be thoroughly photographed
and sketched in position before any attempt is made to move it. Once it has been ascer-
tained that the victim is indeed dead, there is no reason to rush removal (unless continuing
fire or collapse would further compromise the body). If the normal protocol for on-scene
medicolegal processing is to take the internal temperature of the decedent, that should be
done with a fire victim. It has recently been found that torso temperatures of an adult are
very slow to rise, requiring fire exposure of 30 to 60 minutes.11 While the internal temper-
ature cannot be used in the traditional death interval formula, a very low temperature can
reveal that the body was dead for a considerable period of time prior to the fire.
Skin tissue, clothing, and even bones are turned delicately brittle by fire, so extreme
care must be exercised in moving the body and its associated debris; the skin, clothing,
and related documents in wallets or pockets may all have vital information to offer. It is
often useful to wrap the head of the victim with plastic cling-wrap before it is moved. This
helps retain the smallest fragments, such as those of teeth and bone embrittled by heat, in
their relative positions until they can be documented and examined in the autopsy.
When a fire death is being investigated, the team should include both a fire investiga-
tor (for whom the body and its clothing represents a source of fire patterns, often an ele-
ment of the fuel load, and a potential cause or even point of origin) and a homicide
investigator to whom the elements of violent death scenes are more familiar ground. The
primary investigators should together conduct the search of debris in the immediate area
of the body by careful sectoring and layering (examining small areas in careful sequence
from top to bottom). This is where much of the most critical evidence is found and should
not be delegated except under extreme circumstances. A fire death scene that has been grid-
ded off with string is shown in Figure 15-4. The examination should include wet sifting of
the debris through screens of various sizes to ensure maximum recovery. Pope warns, how-
ever, that heavily calcined bone turns the consistency of wet chalk when wetted, so dry sift-
ing by hand is suggested when searching the area where small bone fragments (fingers,
teeth, skull fragments) might be found.12 Sustained exposure to direct fire can vitrify cal-
cined bones to glassy “fossils” that are not affected by water. Debris adhering to the body
should be sent with the body where it can be examined under the best possible conditions.
The pathologist should examine the remnants of clothing and other fire debris collected
with the body for evidence that may help establish identity, cause, or manner of death. If
not, the investigator should personally carry out such visual examination.

FIGURE 15-4 A fire

death scene gridded
off with string to allow
comprehensive search
and reconstruction.
Courtesy of Lamont “Monty”
McGill, Fire/Explosion
Investigator.

Chapter 15 Fire-Related Deaths and Injuries 617

 FIGURE 15-5 Sieving
of the debris on the bed
under the victim yielded
jewelry, bullets, and hard-
ware from furniture that
had been piled atop him
to add fuel. Courtesy of
Lamont “Monty” McGill,
Fire/Explosion Investigator.

Once the body and associated debris have been removed, the remaining ashes and
fire debris in the area should be carefully examined, preferably again by sifting through
large-mesh screens and using a metal detector to ensure that all jewelry and other evi-
dence are recovered. Bullet casings and other types of evidence of homicide may also be
uncovered in this fashion, as shown in Figure 15-5. The debris within 1 m (3 ft) of the
body should be carefully searched for physical evidence to help reconstruct the victim’s
last actions. Are the remains of a flashlight, car or house keys, dog leash, or keepsakes
found nearby?
Clothing is not often completely destroyed by a fire. What the victim was wearing
may indicate his or her activity just prior to the fire. Consideration must be given here to
the time of the fire relative to the state of dress. Was the victim normally fully clothed at
3 A.M. or in pajamas until noon? If the fire occurred at 3 P.M. and the victim was dressed
for bed, was he or she ill or incapacitated prior to the fire? More traditional time of death
markers such as rigor mortis and body temperature are seriously skewed by fire exposure,
and environmental indicators take on additional importance.
The postmortem examination should be carried out as soon as practical. The body
must not be embalmed in the interim, because any such treatment will interfere with tox-
icological tests vital to the investigation. The postmortem examination is nearly always
preceded by a full set of X-rays of the entire body. X-rays are critical because they can
reveal bullet fragments, knife blades, broken bones, or other evidence of assault or murder
that are not visible from an external examination as well as assist in the identification by
the methods described earlier. The sex, approximate age, and height are determined from
examination of the skull, pelvic bones, and major limb bones, while skin, hair, or other
soft tissues remaining may be used to determine race and complexion. [White hair is
turned yellow by temperatures on the order of 140°C (290°F), but other colors of hair
remain unchanged up to the temperature at which the hair itself begins to char (260°C;
500°F) according to one author.]13 Observations by one of the authors indicate that dark
hair may lighten and turn reddish with some fire exposure. Artificial dyes can react dif-
ferently than natural pigments. (This may be due to ventilation effects, with an oxidizing
atmosphere having one effect, while a fuel-rich reducing atmosphere has quite another.)
618 Chapter 15 Fire-Related Deaths and Injuries
In severely charred bodies, computed tomography (CT) and magnetic resonance imaging
(MRI) have been used successfully to document injuries caused by burns, as well as vital
reactions in soft tissues caused by mechanical trauma.14

DESTRUCTION OF THE BODY

One problem the authors have encountered in fire death investigations is a presumption
that the condition of the body as recovered from the fire is the same condition it was in
at the time of death. This is obviously illogical. The body represents a fuel package and
will undergo changes postmortem if it continues to be exposed to fire. It is very rare to
encounter heat of sufficient intensity and duration that it destroys all anatomical features.
Complete legal incineration (cremation) of an adult human body requires temperatures
on the order of 950°C to 1,100°C (1,800°F to 2,100°F) for destruction to be completed
within 1 to 2 hours.15 Even then, identifiable portions of the skull, pelvic bones, and den-
tition usually survive. Legal cremation occurs in a gas-fired furnace in which the body is
fully exposed to intense, well-ventilated flames and radiant heat from the firebrick oven
lining. Lower temperatures and less complete exposure would be expected to produce less
complete destruction or require longer times. The few recorded experiments on burning
bodies and observations of open-air cremations in India indicate that a well-ventilated
wood or coal fire with sufficient fuel can destroy most of an adult body in a few hours.
A fully involved, wood-frame structure fire lasting one hour resulted in reduction of an
adult male body to large bone fragments.16 Once again, identifiable portions of the skele-
tal structures still remained. Case and experimental studies indicate that repeated and sus-
tained burnings of an adult human body with mechanical crushing can result in remains
that may be unidentifiable, as shown in Figure 15-6.17 The skeletal remains of children,
however, are much more readily destroyed due to their smaller size, lower mass, and less
complete calcification. Such remains may be completely consumed in accidental fires.18
Under most circumstances, the flame temperatures and durations reached in a normally
fueled and ventilated fire in a frame dwelling will not be sufficient to destroy the skeletal
remains of an adult or even all the soft tissues of the torso because of the significant
percentage of water in the tissues and large mass. A body that is supported, even partially,

FIGURE 15-6 Sieved

bone fragments from a
cadaver burned in a
wood-fueled “burn barrel”
for 8 hours and then
crushed with a wood
2 ⫻ 4. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

Chapter 15 Fire-Related Deaths and Injuries 619

 on metal furniture or seat springs provides for better exposure to surrounding flames and
may be consumed to a much larger degree than might be expected. This may account for
the extensive consumption of bodies in vehicle fires, even of fairly short duration. Here
the body may be well exposed to a high-temperature, energetic fire fueled by the uphol-
stery materials and plastic components. In one case in DeHaan’s experience, the body of
an adult was burned to less than half its weight with one leg, one arm, and half the
torso reduced to bone in an accidental fire (of some 30 minutes’ estimated duration) in a
camping van.19
Under some conditions of fuel and ventilation, however, there can be substantial incin-
eration of the body, even without destruction of the dwelling or even other fuels in the vicin-
ity. There have been a number of fire deaths described in the literature where the victims
were found with most of the torso consumed, and the fire had not extended beyond fuels in
immediate contact with the body.20 A common scenario is a “cremated” body found in a
room whose furnishings are largely undamaged by fire but that bear smoke and soot
deposits and often signs of radiant heating. Such cases, fancifully labeled “spontaneous
human combustion,” are most often described in popular literature with only marginal sci-
entific assessments of the incidents. Nickell and Fischer have published reviews of the his-
torical cases and noted that most of the reports date from the eighteenth and nineteenth
centuries, when much less was known about the chemistry and pathology involved.21
A careful examination of the well-documented cases reveals that there are certain ele-
ments common to them. First, the victims are usually elderly, with indications of partial
incapacitation due to alcohol, medication, or infirmity. Many of the victims (but not all)
are overweight. There is almost always a room heater, fireplace, or stove present as a
source of ignition and steady heat input. Smoking materials are often present, sometimes
with a history of previous accidental ignition of bedding or garments (and often with
physical evidence of cigarette burns in clothing and furniture to support that history).
There is usually a long time (5–12 hours) between the time the victim was last seen and
discovery. Most significantly, there are almost always furnishings, bedding, or carpets
involved. Such materials would provide not only a continuous source of fuel for an exter-
nal fire but also a supply of char to act as a wick. With this combination of features, the
investigator can appreciate the basics—fuel, in the form of clothing or bedding as the first
fuel ignited and then furnishings as well as the body fat to feed later stages; an ignition
source—smoking materials or cooking or heating appliances; and finally, the dynamics of
heat, fuel, and ventilation to promote a slow, steady flaming fire that may generate only
small open flames and insufficient radiant heat to encourage fire growth. In such circum-
stances, the fat rendered from a burning body will act in the same manner as the fuel in
an oil lamp or candle. If the body is positioned so that oils rendered from it can seep into
a porous wicking material or drip or drain onto an ignition source, it will continue to fuel
the flames. This effect is dramatically enhanced if there are combustible fuels—carpet
padding, bedding, upholstery stuffing—that form a rigid, carbonaceous char that can
absorb the rendered fat and act as a wick. These observations are supported by experi-
ments reported by Ettling,22 Gee,23 and DeHaan et al.24
This phenomenon, termed the candle effect or wick effect, fits with what was dis-
cussed in earlier chapters on fuels and fire dynamics. Clothing or bedding is ignited acci-
dentally by an appliance or smoking material; the victim is quickly overcome by smoke
or carbon monoxide (alone or in combination with drugs or alcohol) and cannot escape.
(Note that the human body is chemically incapable of spontaneous combustion. The igni-
tion source and the initial fuel are always external to the body.) The fire progresses
through the normal fuel as a slow, low heat release fire that burns the skin, causes it to
split and shrink after some minutes, and begins to render the body fat into fuel. This fuel
is essentially animal fat, which has a total heat of combustion on the order of 39 MJ/kg.25
Figures 15-7 and 15-8 illustrate two such cases. This fat does not burn cleanly, however,
and has a water content of some 10 to 20 percent by weight, so its effective heat of
620 Chapter 15 Fire-Related Deaths and Injuries
FIGURE 15-7A Lower torso and legs of elderly FIGURE 15-7B Unburned combustible boxes,
homicide victim are almost completely destroyed, clothing, and wallpaper in the same room as the
along with the mattress stuffing. The blanket wrapped victim are only smoke stained. The corner of the
around the upper body conceals the remains of the bed, where the victim was found, is in the left
shoulders and head. The victim was not discovered for foreground. Fire had self-extinguished prior to
at least 12 hours after his death, which was caused by discovery. Courtesy of St. Martin Parish Sheriff’s
bludgeoning and asphyxiation. Courtesy of St. Martin Department, St. Martinsville, LA.
Parish Sheriff’s Department, St. Martinsville, LA.

FIGURE 15-8 The body of a female

murder victim was set afire in an
isolated woodland. Over a period
of some 5 hours, the body was
consumed between the mid-chest
and knees. The head (not visible),
shoulders, left arm, right hand, and
lower legs remained. Charred leaves
beneath the body provided the needed
“wick.” Courtesy of the Oregon State
Police, Medford, OR.

Chapter 15 Fire-Related Deaths and Injuries 621

 combustion is reduced, and the unburned pyrolysis products condense on nearby surfaces
as a brown or black oily soot (often noted at such scenes). The heat of combustion is cer-
tainly high enough to sustain a modest flame with localized temperatures high enough to
destroy tissue and degrade bone over a period of hours. One case in which the body was
found while still burning yielded graphic proof of the mechanism. Photographs of the
burning body showed low flames, less than 0.4 m (15 in.) high issuing from the torso
where body fat was sustaining a hot but localized fire. This flame height correlates to a
heat output of less than 40 kW.26 The extensive destruction of the body in this case is
shown in Figure 15-8. Clearly, such a small fire would be enough to raise the temperature
of a closed room but not enough to bring the temperature of the room up to ignition tem-
perature, nor even produce enough radiant heat to ignite close-by combustibles.
Note that such situations are difficult to create intentionally because they require a
subtle balance of the rate at which fuel is made available and the heat release rate of
the resulting flames. DeHaan conducted a test in which a 95-kg (210-lb) pig carcass
was used to simulate the fuel load of a normal human body. The carcass was wrapped
in a cotton blanket on a carpeted wood floor. The process was initiated by igniting 1 L
(1 qt) of gasoline poured across the shoulders of the carcass. The gasoline fire burned
itself out in less than 3 minutes. The synthetic carpet surrounding the carcass burned
for an additional hour and 20 minutes. At the conclusion of the gasoline-fed fire, there
appeared to be only surface scorching of the skin. After about 10 to 15 minutes, as the
blanket and carpet continued to burn, the skin shrank and split (with the expansion of
the fat and tissue beneath), exposing the subcutaneous body fat, which melted, to be
absorbed by the carbonaceous char of the blanket and carpet. It could then support a
flaming fire. This rendering process was sustained for some 6.5 hours before extin-
guishment. During that time about half the body of the pig had been burned to ashes,
and the exposed bones had turned light gray and white in color, as seen in Figure 15-9.

FIGURE 15-9 The

remains of a 210-lb
pig carcass after self-
supporting combustion
for 6.5 hours. About half
the mass of the body was
consumed by a fire of
some 60 kW. The bones
are friable and light gray
to white in color. Courtesy
of John D. DeHaan.

622 Chapter 15 Fire-Related Deaths and Injuries

 FIGURE 15-10 The
room at the end of the
test shows brown-black
soot/smoke staining in
the upper half of the
room. Plastic objects in
the hot-smoke layer have
begun to melt, while
plastic-coated paper
poster near the carcass
fire is undamaged.
Courtesy of John D. DeHaan.

The exposed bones were very friable and tended to shatter to ash and small fragments
when handled. Flame temperatures during this process were measured at 812°C to
913°C (1,500°F to 1,675°F), while the heat release rate was a steady 50 to 60 kW dur-
ing the body combustion. Such a small fire requires only a modest supply of oxygen, so
the air leaking into the average room could keep it burning, even with doors and win-
dows closed.27 The small, smoky but prolonged fire produced limited effects in its sur-
roundings, as seen in Figure 15-10. The results of this test were supported by testing in
a cone calorimeter, which established the effective heat of combustion of both human
and pork fat to be about the same, ~32 kJ/g. Note that this was not a smoldering fire.
To support combustion, the body fat has to be presented on a porous material that can
act as a wick to support a continuous flame. The range of fire sizes that might be
expected from this process range from 20 to 130 kW depending on the surface area of
the wick involved.28 The amount of fuel (body fat) available determines the length of
time the process will continue before it self-extinguishes for lack of fuel. Recent tests
with dressed human cadavers (ignited only with direct flame) confirmed that the torso
will sustain the longest-burning flames due to its larger supply of subcutaneous fat. This
can cause more damage to the torso in sustained fires (6 to 7 hours) where the rendered
fat can burn under the body.29 See Figure 15-11 for an example of a body burned nearly
to completion (including the torso).
The self-fueling process can be initiated only by a fire external to the body that is
sustained long enough to cause the skin to shrink and split. When very badly damaged
bodies are found, the immediate thought is that an accelerant had to be present. If an
accelerant is used, the initial flames are usually intense enough to ignite other fuels, the
room may reach a very high temperature, and some exposed fuels burn much more
quickly. A pool of flammable liquid burns only for such a short period of time, how-
ever, that the fire is sustained only long enough to burn the clothing and some of the
Chapter 15 Fire-Related Deaths and Injuries 623
FIGURE 15-11 Cadaver
burned on a box spring
with only cotton clothing
and a cotton blanket.
Ignited with a match,
this fire burned for over
6 hours with extensive
destruction of the torso.
Courtesy of John D. DeHaan.

exposed skin and surface tissue away, but not nearly long enough to destroy the body
itself or even initiate a self-sustaining fire. Figures 15-12a and b illustrates a test where
gasoline was burned on the back of a cadaver with minimal effect to the skin. Experiments
by one of the authors revealed that a pour of 1.9 L (0.5 gal) of a flammable camping fuel
on a carpeted surface produced flame plume temperatures on the order of 775°C to 950°C
(1,430°F to 1,740°F), but only for a period of 2 minutes when ignited in a medium-sized
room, at which time the flames self-extinguished as the camping fuel was consumed.30

FIGURE 15-12A Gasoline pool burning on the back of a cadaver. FIGURE 15-12B Flames nearly self-extinguished. Some scorching
Flames lasted less than 1 minute. Courtesy of Jamie Novak, Novak of skin visible, but no deep damage. Courtesy of Jamie Novak, Novak
Investigations and St. Paul Fire Dept. Investigations and St. Paul Fire Dept.

624 Chapter 15 Fire-Related Deaths and Injuries

 This is not nearly long enough to begin the process just described. Most of the cases
involving badly burned bodies are ignited with open flame onto clothing or furnishings. In
the absence of furnishings to provide the extra fuel needed to maintain the combustion
process, it is possible that incidents may be attributed to the flammable nature of some
clothing fabrics, with an accidental ignition by contact with a match or candle flame. (See
Chapter 11 for a discussion of clothing fire injuries.) The fuel load offered by cellulose
acetate, nylon, or other synthetic fibers (in cotton-blend fabrics) in the form of a com-
plete garment could produce severe charring of the body of an unfortunate wearer. Such gar-
ments do not, however, provide enough fuel to support extensive, long-term combustion
or enough rigid, porous char needed to sustain a fat-fueled fire. The investigator must care-
fully assess the fuel load present (including the body fat), the nature of any possible “wick”
material under the body, potential sources of ignition, the time elements involved, and the
extent of damage (or lack thereof) to the surroundings and balance the fire dynamic equa-
tion. Such considerations would certainly explain most, if not all, of the incidents described
in the literature as “spontaneous human combustion.” In a recent case, DeHaan examined
a scene where the body of a homicide victim clothed in a wool overcoat and other winter
garments burned almost to skeletal remains on a concrete floor. The fire plume never
reached the low wooden ceiling overhead. The fire had self-extinguished before discovery.
In summary, the rate at which a body is consumed by fire is dependent on the size of
the fire and the extent of exposure of the body to the fire. If the fire is a small, localized
fire supported only by the combustion of the subcutaneous body fat and other suscepti-
ble tissues, in which the heat release rate of such fires has been measured to range from
20 to 130 kW, the mass loss rate of such fires has been measured to be 3.6 to 14 kg/hr (8
to 30 lb/hr).31 (In contrast, a fully involved post-flashover room fire of 10 MW produces
flame temperatures in excess of 1,000°C (1,850°F) and radiant heat fluxes of 120 to 150
kW/m2 with commensurately rapid combustion of the body if it is exposed.) This fuel
consumption rate has been used to estimate duration of the time the body was burning if
the conditions of fire exposure can be documented (i.e., nature of external fuel, maximum
fire size, etc.).
With regard to the destruction of the corpse, it must be kept in mind that quite dif-
ferent conditions apply in industrial fires. Fires involving fuels such as magnesium, ther-
mite, or similar incendiary materials can involve temperatures of thousands of degrees,
resulting in complete reduction of all the bones to calcium and silicon salts if the fire is of
sufficient duration. Accidental, homicidal, or even suicidal incidents involving steel smelt-
ing, cast metal foundries, or cement making have resulted in total destruction of bodies.

EFFECTS OF FIRE
In most structure and wildlands fires, however, the pathologist has at least the skull and
torso of the body with which to work, and the effects of the fire are much more manage-
able. Fire (heat) causes the muscles and tendons to contract, flexing the joints, particu-
larly of the limbs. The resulting “pugilistic pose” (see Figure 15-13) or fighter’s stance has
been mistaken by some investigators for a last-minute self-defense posture on the part of
the victim. This is not the case, as it is a commonly found artifact that has nothing to do
with the physical activity of the victim just before death. In fact, the flexor muscles of the
human body are usually stronger than the opposing extensor muscles. When heated, the
shrinkage causes flexion at all affected joints. The final position of the body may have
changed from its pre-fire position, as shown in Figures 15-14 and 15-15. Physical restraint
of the limbs (binding or position) can prevent heat-induced flexion. Flexion at major
joints can cause the burning body to shift position. This shift can cause bodies to fall off
chairs or beds. Heating by fire exposure requires considerable time before the tempera-
ture of deep internal organs is affected. Even so, a low liver temperature of a fire victim
may reveal an extended time interval between death and fire exposure.
Chapter 15 Fire-Related Deaths and Injuries 625
FIGURE 15-13 Pugilistic
posture (flexing of joints
from heat exposure) of a
fire victim. Courtesy of Greg
Lampkin, Knox County, TN,
Fire Investigation Unit.

Skin shrinkage and combustion exposes underlying bone to direct flame contact. This
exposure occurs first where there is minimal thickness of overlying tissue—skull, elbows,
knees, and wrists.32 Fire exposure burns away the organic components of bone and then
calcines the mineral portion, which causes the color of the bone to change from tan to
charred black to gray to white. Extensive exposure can cause shrinkage and fracturing of
calcined bones, which can be mistaken for antemortem injuries. The appearance of frac-
ture edges in the vicinity allows the distinction to be made between postmortem and ante-
mortem fractures by an experienced pathologist or anthropologist. Fire can also cause
fractures of the skull, with separations sometimes occurring along the suture lines.
Exposure to a sustained fire can cause the skull to calcine and shatter in a variety of fracture

FIGURE 15-14A Pre-fire position of cadaver: on back, arms FIGURE 15-14B Post-fire position due to heat-induced posturing
extended by sides, legs on bed. Courtesy of Jamie Novak, Novak and partial collapse of mattress: right arm and right leg off bed, left
Investigations and St. Paul Fire Dept. leg flexed. Courtesy of Jamie Novak Investigations and St. Paul Fire Dept.

626 Chapter 15 Fire-Related Deaths and Injuries

 FIGURE 15-15A Near
start of the fire: arm
extended at side. Fingers
are starting to flex toward
the fire in blanket. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

FIGURE 15-15B During

the fire (mattress and
body well alight): arm
raised, hand flexed down.
Note pool fire under bed
as rendered body fats
burns on noncombustible
floor. Courtesy of Jamie
Novak , Novak Investigations
and St. Paul Fire Dept.

FIGURE 15-15C Late

in the fire: arm has
collapsed away from
body. Note rendered fat
on floor under wrist and
hand. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

Chapter 15 Fire-Related Deaths and Injuries 627

FIGURE 15-16 Skull
exposed to prolonged
fire exhibits charring,
heat induced fractures,
calcination, delamination
(exposing the diploe
between the inner and
outer tables of the skull).
Courtesy of Jamie Novak,
Novak Investigations and
St. Paul Fire Dept.

patterns as shown in Figure 15-16. The expansion of tissues, steam, or gases created by
the action of the fire within the skull then may cause the fractures to separate and the tis-
sue to exude from the skull. In controlled tests by Pope, no head was seen to “explode”
as a result of fire exposure. Fire exposure caused the skull to become very friable and eas-
ily broken by post-fire movement or impact by falling debris.33 It had been suggested that
preexisting trauma (such as bullet holes or fractures) prevented such “exploding,” so its
absence would prove the absence of such trauma. Pope’s experiments showed preexisting
trauma made no difference. Heat fracturing is usually distinguishable from the depressed
skull fractures caused by blunt-force trauma or beveled edges caused by bullet entry or
exit. Microscopic analysis of the margins of the fractured bones may determine whether
the bone was broken by mechanical or thermal insult.34
Bone consists of collagen for tensile strength and hydroxyapatite (calcium phosphate)
mineral crystals that provide compressive strength and hardness.35 Extreme heat expo-
sure can cause the shrinkage of some bones, reportedly up to 25 percent in large bones at
temperatures of more than 700°C (1,300°F).36 Such changes may cause errors in the esti-
mation of height. Bone can crack and split during severe heating. This cracking varies
with the pre-fire state of the bones, with fresh bones with moisture still present cracking
differently than those that have desiccated due to long-term exposure (such as archeolog-
ical remains). There are indications that spalling of the fresh bone surfaces (similar to that
on concrete) occurs as a result of expansion of trapped moisture. Most flames are also
intermittent and turbulent, which cause the surface temperatures to rise and fall quickly,
inducing mechanical stresses.37 Under some sustained fire conditions the calcined bones
can vitrify or turn to a glassy green-gray material that is not water soluble, as shown in
Figure 15-17.
Tests by Christensen have revealed that bones of the elderly suffering from osteoporo-
sis are much more easily fragmented by fire exposure than are bones of normal density.38
This may explain why some victims of body-fueled fires have such extensive destruction
of the exposed bones (not because the fire was particularly hot or intense).
Pope and Smith have described several means by which mechanical and thermal
trauma to skulls can be differentiated.39 Heating can cause partial delamination of the skull
at a fracture, simulating the beveling effect of bullet exits. They recommend reassembly
628 Chapter 15 Fire-Related Deaths and Injuries
 FIGURE 15-17 Bone
fragments recovered from
a cadaver ”burn barrel”
test show charring calci-
nation and vitrification as
a result of sustained fire
exposure. Courtesy of Jamie
Novak, Novak Investigations
and St. Paul Fire Dept.

of fractured pieces and evaluating fracture patterns and differential heat effects on adjoin-
ing pieces to distinguish between fractures caused by thermal stress from those of mechani-
cal origin.
Conversely, damage by firearms projectiles or blast debris may be mistaken for fire
damage by the inexperienced. See Figure 15-18a for an example. Fire damage is very
rarely localized, so severe tissue and bone disruption in a single, isolated area must be
considered in light of damage to surrounding tissues and the circumstances of the fire
before it is dismissed as fire-caused. The skin’s vital response to the application of heat,
to redden and to blister, ceases near the time of death, but fluid-filled blisters can be
caused by postmortem heating. Blisters and first- and even second-degree burns can also
be caused by contact with gasoline and caustic chemicals, while reddening can be simu-
lated by the desiccation of underlying muscle even after death.40 The reddening of tissue
caused by carbon monoxide inhalation can be obscured by sooting, burning, and other
fire effects. Wounds from small-caliber weapons can be overlooked if fire damage is
severe, so X-rays are very important (see Figures 15-18b–d).
Small amounts of blood may be forced from the internal tissues by heat effects,
which will seep from the nose, ears, or vagina of a burned body. Significant external
bleeding is not caused by post-fire seepage, so the presence of pooled liquid blood under

FIGURE 15-18A This X-ray of the upper arm of a victim shows FIGURE 15-18B Bomb squad portable X-ray in use at fatal fire with
shattered bone and missing pieces as a result of a gunshot injury badly incinerated body. Courtesy of Jamie Novak, Novak Investigations and
prior to the fire. Courtesy of Lamont “Monty” McGill, Fire/Explosion St. Paul Fire Dept.
Investigator.

Chapter 15 Fire-Related Deaths and Injuries 629

FIGURE 15-18C Positioning the head with minimal FIGURE 15-18D X-ray showing bullet fragments in head (small-
disturbance. Courtesy of Jamie Novak, Novak Investigations caliber entry wound concealed by fire damage). Courtesy of Jamie
and St. Paul Fire Dept. Novak, Novak Investigations and St. Paul Fire Dept.

a body is a strong indication of antemortem violence. Fire exposure sears or cauterizes

the tissue as it is exposed, thus minimizing blood loss. Internal bleeding into the
extradural or epidural space of the skull (between the dura mater and the skull), how-
ever, can be caused by fire exposure. An accumulation of blood between the dura and
the brain (subdural hematoma), however, should be evaluated as possible evidence of
physical trauma. An examination of the burn injuries alone may not reveal whether they
were caused immediately before or after death, and the success of the inquiry depends
on the extent of cooperation and interaction between the pathologist, homicide investi-
gator, and fire investigator.
Short-term exposure to high temperatures can cause slippage or “gloving” of the
exposed skin (epidermis). The underlying dermis will tend to be red and moist-looking.
Human skin consists of three major layers—the epidermis or exterior layer of hardened
“dead” skin cells, the dermis, and subcutaneous tissue. The epidermis overlays the dermal
layer, which includes the germinative layer and nerve endings, glands, and blood supply.
Under that is a layer of connective tissue and subcutaneous fat (see Figure 15-19). First-
degree burns are reddened areas of the epidermis, as seen in Figure 15-20. Second-degree
or partial-thickness burns that expose the dermal layer can be very painful because the
underlying pain sensors in the dermis are nearly exposed. Deeper burns destroy the dermis
and the nerves (and the capacity for the skin to regenerate). Thermal exposure of skin
well after death tends to produce a wrinkled, dry, yellow or brown skin surface, and gas-
filled blisters can be caused by fire-induced pyrolysis long after death. Fire-induced shrink-
age of the skin and underlying tissues may result in splitting or tearing, which allows the
630 Chapter 15 Fire-Related Deaths and Injuries
 FIGURE 15-19 Cross
section of human skin:
thin (0.5 mm; 0.02 in.)
epidermal layer, overlying
a thicker (1.5 mm; 0.06 in.)
dermal layer. The epidermis
is joined to the dermis by
the dermal papillae (pegs)
that bear the growing skin
cells that replace those
Epidermis shed from the exterior.

Dermal papillae with

germinative layer

Dermis
(blood vessels, nerve endings,
glands)

Connective tissue/
subcutaneous fat
Vein Nerve endings

Muscle

protrusion of internal organs from the abdomen or the tongue from the mouth. The
margins of such torn tissues will be irregular, and various layers will tear differently
as opposed to an incised wound, where the edges are sliced uniformly. Such splitting or
tearing must not be confused with antemortem cutting or blunt-force trauma injuries. If the
victim lived under hospital care for some time prior to death, normal surgical procedures

FIGURE 15-20 Protected

area shows where shirt
was in place when heat
injury (first-degree burn)
occurred. Courtesy of Greg
Lampkin, Knox County, TN,
Fire Investigation Unit.

Chapter 15 Fire-Related Deaths and Injuries 631

 such as cutting the burned tissue (eschar) to allow circulation, expansion and motion,
debridement, or even harvesting of skin from unburned areas may appear to be ante-
mortem injuries. Burn injuries are discussed in a later section of this chapter.

OTHER PATHOLOGICAL FINDINGS

There are other pathological findings that may be of significance to the fire investigator.
Particles of solid soot and carbon from a smoky fire will be inhaled and will cling to the
moist surfaces in the throat and lungs. The finding of such contaminants in the lungs,
windpipe, or larynx during the postmortem is a strong indication that the victim was alive
at the time of the fire. Soot found adhering just to the inside of the mouth or nose may
be the result of soot settling on exposed surfaces (onto the face of a victim lying face up
in the room, for instance). It may not be an indication that the victim was alive and
breathing. The position of the victim must be taken into account. A body lying face down
in a lightly smoke-charged room may not show significant soot in the airways even if the
victim was breathing.
Inhalation of extremely hot gases can cause a reflexive constriction of the pharynx
(sometimes referred to as laryngeal spasm) and cessation of breathing sometimes called
vagal inhibition. In such cases, the other pathological signs will be consistent with those
of strangulation, but some pathologists will not offer much of an opinion because the
spasm (if it occurs) relaxes on death, so there is no positive pathological finding.
Inhalation of flame or very hot gases can also produce a rapidly developing edema
(swelling) of the tongue and pharynx that produces the same symptoms as asphyxiation.
In both such causes of death, the carbon monoxide level in the victim’s blood may be
quite low, especially if the victim was exposed to a sudden flash of flame early in the stage
of the fire. A victim who suddenly opens a door from a cool, smoke-free room into a hall-
way completely enveloped in flame is typical of such fatalities. Very hot dry gases are rap-
idly cooled by the evaporation of moisture from the mouth and throat, and fire damage
(desiccation or edema) will not normally extend below the larynx. If superheated steam
or moisture-laden fire gases are inhaled, there is no evaporative cooling, and thermal
damage can penetrate all the way to the lungs. Such damage may occur in flash fires of
hydrocarbon fuels, in which the combustion products are carbon dioxide and water vapor
at very high temperatures. When such a premixed fuel/air mixture deflagrates in a room,
the atmosphere becomes deficient in oxygen (nearing zero) and rich in carbon dioxide.
Both these factors contribute to the collapse and death of a person in the room, and yet
neither oxygen saturation nor carbon dioxide saturation can be measured in a post-
mortem sample because the complexes they form with the blood are comparatively unsta-
ble and begin to change as soon as death occurs.41
Although instances of it are quite rare, the possibility of suicide by fire should not be
discounted by the investigator. Such suicides are often (but not always) accomplished by
pouring flammable liquid over oneself and then igniting it with a match. In these cases, low
carbon monoxide levels are typical, because death is often caused by shock or asphyxiation
as a result of inhaling the flames or superheated gases. The remnants of the clothing found
on the body and nearby floor coverings and furnishings may bear residues of the ignitable
liquid used and should be put into separate clean metal cans, Kapak (by Ampac) bags, or
sealable glass jars to preserve them for gas chromatographic identification of the liquid.
In the absence of ignitable liquids, consideration should be given to the flame-spread
characteristics of the clothing worn by the victim (as described in Chapter 11) (in both
accidental and intentional fire deaths). Figures 15-21a and b show the burns to clothing
and skin to an elderly man who ignited the chair in which he was sitting. He collapsed
after rising from the chair. His wife died in the adjacent chair of smoke inhalation. There
have been several cases in which bodies were recovered from the closed trunk of a burned
car, and other evidence (non-fire) indicated the fire was set deliberately in the interior of
632 Chapter 15 Fire-Related Deaths and Injuries
 FIGURE 15.21A The
chair and pant leg of this
elderly fire victim were
ignited by discarded
smoking materials. The
victim arose from the
burning chair to collapse
nearby. Courtesy of Chiefs
Mike Oakes and Tony Hudson,
Clallam County Fire
Investigation Team, Port
Angeles, WA.

FIGURE 15.21B Outline

of body on carpet pro-
duced by soot (generated
by the burning chair).
Courtesy of Chiefs Mike Oakes
and Tony Hudson, Clallam
County Fire Investigation
Team, Port Angeles, WA.

the car with suicidal intent.42 Suicide cases have also been recorded in which the decedent
crawled or jumped into an operating incinerator or industrial furnace, or into an established
fire in a structure or vehicle.
Samples of blood and undamaged organs are mandatory. Even in badly burned bodies
there may remain some liquid blood and undamaged internal organs (kidney, liver, gall-
bladder, heart). The blood will be used for toxicological tests to determine the presence
of carbon monoxide, cyanide, alcohol, drugs, or poisons, as well as serological or DNA
tests to establish the identity of the victim. Even if no liquid blood is present, tissue can
be typed as long as it is still intact and not carbonized. (See Chapter 14 for a discussion of
Chapter 15 Fire-Related Deaths and Injuries 633
 TABLE 15-1 Summary of Postmortem Tests Desirable in Fire Death Cases

TEST OR ACTIVITY SUBJECT TESTED FOR

Blood tests Major blood vessels Carbon monoxide
HCN
Heart chamber Blood alcohol level
Not from body cavity Drugs (therapeutic and abuse)
Tissue tests Brain Drugs
Kidney Poisons
Liver Volatile hydrocarbons
Lung
Skins near burns Vital chemical or cellular response
to burns
Chemical assay Stomach contents Activities prior to death
Possible time of death
Airway tests Full longitudinal transection Edema
of airways from mouth to Scorching/dehydration
lungs Soot
See figures
Internal body temperature Liver or rectal temperature May be low due to pre-fire death
May be elevated due to hyper-
thermia antemortem
X-rays (not fluoroscope) Full body Unusual features such as fractures,
Debris in body bag surgical appliances, implants,
prostheses
Details of teeth
Clothing* Volatile accelerants
Photographs (in color, with Major burn damage
scale) Any visible trauma
Attend postmortem†
*Collect and air dry clothing, package separately. If volatile accelerants are suspected, seal immediately in clean
metal cans or nylon debris bags
†
It is very useful for the fire investigator to be present when the postmortem is conducted, not only to ensure that
appropriate observations are made but also to be on hand to answer any questions that arise during the examina-
tion. Few pathologists have extensive knowledge of fire chemistry or fire dynamics, so the investigator is in a good
position to advise the pathologist as to fire conditions in the vicinity of the body.

forensic blood analysis.) The organs should be examined for signs of pathology due to
disease both as an aid to identification and in determining the cause of death. Ignitable
liquid accelerants may be absorbed by the tissues and may be detected in the blood, sub-
cutaneous fat, or internal organs. To ensure the adequacy and accuracy of the investiga-
tion, it is very important that organ and tissue specimens be tested for toxic materials and
drugs. The presence and level of drugs, particularly alcohol, will often have great bearing
on the determination of cause and manner of death. Alcohol is often associated with fire
deaths, particularly in accidental fires where the victim, incapacitated by alcohol, suc-
cumbs to burns or asphyxiates as a result of even minor fires. In larger fires, the influence
of drugs or alcohol may be increased by smoke and toxins, producing confusion in the
victim, thereby preventing escape from the fire. See Table 15-1 for a list of postmortem
tests recommended. Failure to obtain a postmortem and toxicology exam precludes reli-
able conclusions as to the real causes of both the death and the fire.
Asphyxiation by carbon monoxide or other toxic combustion gases causes most fire
deaths. Direct fire exposure is rarely the significant agent. In the Dupont Plaza Hotel fire
(San Juan, Puerto Rico, 1986) and in the Stardust Disco fire (Dublin, Ireland, 1981), a
large percentage of the victims had low CO and cyanide levels inconsistent with smoke
634 Chapter 15 Fire-Related Deaths and Injuries
inhalation deaths.43 Victims caught in flashover situations in those fires were seen to catch
fire by radiant heat as they tried to escape. In the Dupont Plaza fire, the fatalities in the
hotel away from the fire had high COHb saturations.44
Because only carbon-based fuels are involved in virtually all structure fires, one
would expect carbon dioxide (CO2) to be the predominant gaseous product if combus-
tion is complete. Fires rarely produce flames in which combustion is complete, and the
ventilation (amount of oxygen available) and flame temperatures determine the produc-
tion ratios of CO, CO2, and other species.45 In nearly all fires, the inefficiencies of com-
bustion, especially of carbon-based solid fuels, will mean that CO will be present and will
play a part in fire deaths. Depending on the composition of the building’s contents and
furnishings, other toxic gases will be produced that are also of great significance to the
fire investigator. We shall discuss those separately.

Carbon Monoxide Asphyxiation

Carbon monoxide (CO) is probably the single most important cause of death from fires.
Carbon monoxide, like oxygen in the air, combines with the hemoglobin of the blood.
The blood hemoglobin transports all the oxygen to the tissues of the body and is there-
fore a vital link in the chain that supports oxidative metabolism and life itself. Unfortunately,
CO combines with hemoglobin to form a complex called carboxyhemoglobin (COHb) with
about 200 times the avidity or stability of the oxygen-hemoglobin complex (OHb, called
oxyhemoglobin).46 As more and more of the hemoglobin is tied up with CO, there is less
available to carry oxygen to the brain and muscles. The amount of CO present in whole
blood is normally expressed as a percentage saturation (COHb satn), that is, the percent-
age of the hemoglobin that is occupied by CO.
The intrusion of a material that displaces the oxygen from oxyhemoglobin and pro-
duces a more stable and useless hemoglobin combination serves to starve the cells of
needed oxygen, and an internal suffocation results. In addition, CO interferes with the
basic energy production process in the body’s cells [the adenosine triphosphate (ATP)
cycle], so it poisons cells and reduces their viability. CO also bonds with the myoglobin
in red muscles (such as the heart) and affects their functioning.47 In addition to replacing
oxygen (asphyxiant), CO also seems to affect the transfer of signals in the nerve cells and
can have permanent or delayed neural effects.48
These effects are fortunately reversible in a victim given oxygen or fresh air. The O2
will slowly displace the CO from the carboxyhemoglobin, and the CO content of the
latter will diminish. Thus, if a person who was exposed to a dangerous concentration of
CO is placed in clean air or, better, given oxygen, he or she will generally recover as a
result of this displacement.49 The COHb level in a living person will begin to decrease as
soon as the CO exposure is stopped and fresh air or oxygen is made available. The COHb
level will decrease by 25 percent of the initial level after 13⁄4 to 2 hours in fresh air and
after about 40 minutes breathing 100 percent oxygen. It will decrease by 50 percent after
about 4–5 hours in air and about 11⁄2 hours of 100 percent oxygen.50 Elimination will be
even faster if the victim is treated in a hyperbaric chamber, which increases the pressure
of O2 and hastens elimination of CO. If a victim received medical treatment before death,
such effects must be taken into account when evaluating the postmortem COHb level.
Based on the relative stability of COHb and OHb, it is relatively simple to calculate
the carbon monoxide content of the air that may be dangerous. If the concentration of
oxygen in air was just 200 times as high as the carbon monoxide, extended breathing of
this mixture would result in a mixture of COHb and OHb that would be about equal,
and half the effective oxygen-carrying capacity of the blood would be lost. Because air
contains about 20 percent oxygen, 1/200 of this value would be 0.1 percent. Thus, breath-
ing a concentration of carbon monoxide of 0.1 percent [1,000 parts per million (ppm)]
would eventually result in a potentially fatal 50 percent COHb level. Experience shows that
Chapter 15 Fire-Related Deaths and Injuries 635
 TABLE 15-2 Accumulation of Carbon Monoxide in Blood

CO CONCENTRATION (%) TIME OF EXPOSURE PERCENT BLOOD SATURATION

0.02–0.03 5–6 hours 23–30
0.04–0.06 4 –5 hours 36–44
0.07–0.10 3– 4 hours 47–53
0.11–0.15 11⁄2 –3 hours 55–60
0.16–0.20 1–1 ⁄2 hours
1
61– 64
0.20–0.30 1/2–3/4 hours 64 –68
0.30–0.50 20–30 minutes 68–73
0.50–1.0 2–15 minutes 73–76

Reprinted, with permission, from the Journal of Forensic Sciences, Vol. 7, No. 4, copyright ASTM International, 100
Barr Harbor Drive, West Conshohocken, PA 19428.

the dangerous limits are actually around 0.05 percent, and that a concentration of carbon
monoxide of 0.06 percent can be breathed only for a period of 11⁄2 hours without becom-
ing critically dangerous. See Tables 15-2 and 15-3 for concentration and exposure times.
The Stewart equation relates the CO concentration in air, the breathing rate of the
victim, and the COHb saturation in the blood.51 It is usually given as
% COHb ⫽ (3.317 ⫻ 10 ⫺5 )(ppm CO)1.036 (RMV)(t)
Where:

%COHb ⫽ percent saturation

ppm CO ⫽ concentration of CO in the air being inhaled
t ⫽ time of exposure
RMV ⫽ respiratory minute volume (liters per minute breathing rate)
RMV is dependent on age, sex, and physical activity52 (see Table 15-4).
Substituting into the equation the values COHb ⫽ 50%, RMV ⫽ 20 L/min, ppm CO ⫽
100,000 ppm (10%), we obtain
t ⫽ (3.015 ⫻ 104 ) (50)>(1 ⫻ 105 )(20) ⫽ 0.75 min

TABLE 15-3 Time to Reach 30 Percent Concentration of Carbon Monoxide

in Hemoglobina

PERCENT OF CO IN BREATHED AIR TIME

0.01 (100 ppm) —
0.025 (250 ppm) —
0.05 (500 ppm) 2–3 hours
0.1 (1000 ppm) 15–30 minutes
0.2 (2000 ppm) 5–10 minutes
1.0 (10,000 ppm) 1–5 minutes

Reprinted, with permission, from the Journal of Forensic Sciences, Vol. 7, No. 4, copyright ASTM International, 100
Barr Harbor Drive, West Conshohocken, PA 19428.
a
This concentration produces confusion and collapse in some healthy adult individuals.

636 Chapter 15 Fire-Related Deaths and Injuries

 TABLE 15-4 Respiratory Minute Volume (L/min)

ADULT MALE ADULT FEMALE CHILD INFANT

Resting 7.5 6 4.8 1.5

Light activity 20 19 13 4.2

Source: Health Canada, “Investigating Human Exposure to Contaminants in the Environment” (Ottawa, Ontario:
Ministry of National Health and Welfare, 1995); www.hc-sc.gc.ca.

This means that in moderate activity (or stress), an adult male can inhale a fatal CO
amount in less than 1 minute or in a few breaths if the CO concentration is very high in
the air he is breathing (as from a smoldering fire). Note that heavy activity can accelerate
the RMV to 50 to 60 L/min for an adult male. This can mean even faster accumulation
of CO to fatal concentrations. The “general” effects of CO are shown in Table 15-5.

THE CARBON MONOXIDE HAZARD

COHb is normally found at low concentrations in human blood, as it is a product of the
normal breakdown of hemoglobin. Healthy individuals have a COHb saturation of 0.5
to 1 percent. Some diseases can cause COHb levels as high as 3 percent.53 People exposed
to CO in their environments—auto mechanics or anyone using gasoline- or propane-
powered equipment inside a building—can have levels of 5 to 15 percent. Smokers will
have elevated levels depending on how frequently they smoke. Heavy smokers can have
levels as high as 10 percent, and even a nonsmoker sitting for any time in a smoke-filled
room can have a COHb level of 3 to 5 percent. For this reason, many laboratories will
not report COHb levels below 10 percent. In some instances even low levels would be of
evidentiary significance. In a case in which a young, healthy woman, nonsmoker, who
lived in a clean, semirural environment was found dead in a burned-out vehicle, a 3 percent
COHb saturation was considered as significant indication of brief exposure to the fire
before dying.

TABLE 15-5 Physiological Effects of Carbon Monoxide

(for Healthy Adults)

PERCENT SATURATION IN BLOOD PHYSICAL EFFECTS

0–10 Slight loss of mental acuity
10–20 Slight headache, dilation of skin vessels
20–30 Severe headache, throbbing
30–40 Severe headache, weakness, dizziness, confusion, nausea,
vomiting, collapse
40–50 Fainting, rapid heartbeat, collapse, death in some individuals
50–60 Fainting, rapid breathing, possible coma, convulsions,
respiratory irregularity
60–70 Convulsions, depressed heart action, death
70–80 and higher Weak pulse, respiratory failure, death

Source: P. E. Besant-Matthews, “Investigation of Fire-Related Deaths,” in Arson Investigation, vvi–8 (Sacramento,

CA: California District Attorneys Association, 1980).

Chapter 15 Fire-Related Deaths and Injuries 637

 Regardless of the exposure of the ordinary citizen to carbon monoxide in quantities
that appear distinctly in the blood, such exposure is not expected to result in fatalities.
Although a 5 percent saturation level of COHb may be found in many ordinary citizens,
it usually requires about 20 percent saturation to produce serious symptoms such as
headache, nausea, disorientation, and dizziness in a healthy subject (see Table 15-5). At
about 40 percent saturation, unconsciousness is expected, although the results may be
more or less extreme. Victims determined to have been killed by carbon monoxide
asphyxiation have been found to have levels below 30 percent; however, this is unusual.
When the individual has preexisting heart or respiratory conditions that compromise the
body’s oxygen-delivery system or heart muscle, the loss of even 20 to 30 percent of the
capacity may be fatal (when combined with CO’s effects on heart muscle and nerves).
Other toxic gases such as HCN may also play a role, as it is impossible to separate CO
from other toxic gases in fire exposure. A low COHb level in a fire victim may indicate
death resulted early in a fire caused by massive incineration, inhalation of flame-hot
gases, or mechanical trauma (crushing, burial in debris, explosion trauma).
Hyperthermia or anoxia can also kill even in areas not directly affected by products of
combustion. A negligible COHb saturation does not prove the victim died before the
fire. Carbon monoxide is very stable in postmortem samples and tissues. Unlike ethanol,
it is not produced endogenously in the body. Carbon monoxide is not absorbed from a
fire environment postmortem; it is absorbed only by inhalation during life.
Cyanide is present in many combustions and is considered to be the most toxic gas
in fire environments. It can kill at concentrations as low as 200 ppm in inhaled air, pro-
ducing 2 to 3 mg/L in blood (normal blood CN levels are on the order of 0.25 mg/L).54
Unlike COHb, CN concentrations can change with decomposition and storage condi-
tions. It is always found in combination with CO, but its effects seem to be simply
additive.
For example, in the Happy Land Social Club (New York, 1990), where the fire was
limited to a gasoline-fueled fire in the ground-level entry, the COHb concentrations in vic-
tims ranged from 37 to 93 percent with a mean of 76.5 percent. Cyanide blood concen-
trations ranged from 0 to 5.5 mg/L; nine had zero cyanide detected. CN concentrations
did not correlate with amount of smoke inhalation, so its role in causing deaths was
doubtful. Burn injuries were relatively minor and apparently played no role in causing the
deaths.55 It is generally agreed that a level of 50 percent saturation or greater is incom-
patible with life, despite the finding in dead bodies of both lower and higher degrees of
saturation. Data in Table 15-6 from Gettler and Freimuth indicate the COHb percentage
in a total of 65 cases of fatal asphyxiation.56 It will be noted that they found some deaths

TABLE 15-6 Fatal Cases of Carbon Monoxide Asphyxiation Classified

according to Carbon Monoxide Saturation of the Blood

TOTAL CASES (%)a CO (SATURATION OF BLOOD, %)

4.5 30–40
4.5 40–50
14.5 50–60
28.0 60–70
48.5 70–80

Source: A. O. Gettler and H. C. Freimuth, “Carbon Monoxide in Blood: A Simple and Rapid Estimation,” American
Journal of Clinical and Pathological Technology, Supplement 7, no. 79 (1943).
a
Total of 65 cases.

638 Chapter 15 Fire-Related Deaths and Injuries

below 40 percent and none over 90 percent. One of the authors has encountered labora-
tory values of over 90 percent in a very few instances (usually when the victim was most
likely asleep during the entire exposure).

EFFECT OF RATE OF ABSORPTION

It is well recognized that all individuals show different tolerances to the influence of toxic
gases and other environmental influences. Certainly, such variations must exist with car-
bon monoxide as well. It remains to be explained how the death of one person from car-
bon monoxide at 40 percent saturation can be consistent with that of another who had
accumulated a 90 percent saturation. The relevant factors appear to be as follows:
■ Rate of inhalation of the gas
■ Activity or its lack, which changes the requirements for oxygen
■ Individual variations in susceptibility

The first two factors are firmly linked. The more physical activity, the faster the
breathing will be, and the faster a lethal level will be reached. Physical activity increases
the demand for oxygen from the voluntary muscles, and oxygen starvation of those issues
will tend to incapacitate a person and reduce his or her chances of escape. Case histories
reveal that persons at rest (asleep or unconscious) tend to have the highest COHb satura-
tions because they make no demands on their heart. The role of CO interaction with heart
muscle may be the strongest factor in human variability. If a person is relatively inactive,
the need for oxygen is diminished, the breathing is slower and more shallow, and carbon
monoxide in the air will accumulate more slowly than if breathing is faster and deeper.
Carbon monoxide in the air will rapidly be accumulated if breathing is deep and at a high
rate, as would be required in performing physical work. Thus, it can be stated that if a
person suspects that carbon monoxide is present in significant amounts in the ambient air,
he or she is better off resting quietly, breathing as little and as shallowly as possible, and
avoiding all activity that will increase the breathing rate. Such considerations are often
important for firefighters who must exert themselves in fighting a fire, rescuing trapped
persons and the like, because the high degree of activity will inevitably increase the haz-
ard considerably. This is one of the reasons firefighters use SCBA during interior firefight-
ing efforts. Physical activity also increases the body’s demand for oxygen, and a sudden
increase in demand when 40 or 50 percent of the blood is already saturated may bring
about collapse.57 Variations in individual tolerance lie in the area of physiological differ-
ences, concentrations of hemoglobin in the blood, and related factors that are generally
not under the control of the individual. Much of the toxicological studies cited previously
dealt with healthy adult males as test subjects. Effects on the elderly and on small chil-
dren have not been studied, but it would be expected that CO would have a negative
effect on these victims at lower levels than for healthy adults. Elderly victims, whose heart
and lung functions may already be impaired as a result of age or illness, are most proba-
bly more likely to die when less than 40 percent of their blood is bound up with CO.
Infants have a faster respiration rate, and their uptake of CO may be faster than that evi-
denced in adults. There are also indications that the effects of CO are combined with or
even enhanced by the effects of other toxic gases (such as CN or formaldehyde) from the
fire, and by the CO2 and low O2 levels present.58
Fires are complex events, and their environments are not uniform, so the exposure of
a person is not very predictable, being highly dependent on what the victim was doing
during the fire. Death in a fire may be brought about by a combination of negative factors—
heat, flames, oxygen deficiency (hypoxia plus enriched carbon dioxide concentrations),
carbon monoxide, and other toxic gases. Each of these factors has what is called a
fractional incapacitation dose. If a victim is exposed to half of a fatal (incapacitating) dose
of CO (developing a level of 30 percent, for instance), oxygen of 12 percent (just able to
Chapter 15 Fire-Related Deaths and Injuries 639
 sustain consciousness), and some heat, the total incapacitation dose may exceed 100 per-
cent, and the person collapses to die from structural collapse or flame contact (incinera-
tion). No one of the factors was enough to cause death, but together they created a
fatality. It is important to remember how many things can kill someone in a fire beyond
CO alone. The fire environment and the nature of the victim’s interaction with it must be
considered in evaluating the toxicology findings.

SOURCES OF CARBON MONOXIDE

It was stated earlier that virtually all carbon monoxide is the product of incomplete com-
bustion of carbon compounds. Almost every instance of asphyxiation from this gas results
from one of the following:
■ Fires involving structures, clothing, or furnishings
■ Defective heating equipment (including charcoal grills used indoors), the primary
cause of nonfire fatalities involving CO
■ Automobile or other internal combustion motor exhaust
■ Industrial processes in which CO is intentionally generated as a fuel or as part of a
manufacturing process or in which it is produced as an unintentional product by
incomplete combustion of some fuel

Structural Fires
In every structure fire, carbon monoxide is generated and is the cause of most fire deaths.
The rate at which a fire produces CO is controlled by the size of the fire, the efficiency of
its combustion, and the availability of oxygen. As we saw previously, smoldering, under-
ventilated, or post-flashover fires produce CO concentrations of 1 to 10 percent in its by-
products, while well-ventilated, free-burning fires produce less than 200 ppm (0.02 percent).
It is also produced in forest and wildlands fires, but the ventilation and exposure conditions
of such fires make CO fatalities very rare. Fires in large buildings such as hotels produce
vast quantities of combustion gases that may “channel” through halls, elevator shafts, and
ventilation ducts and collect at areas remote from the actual fire. Deaths may occur as a
result of the fire even though no fire is in the vicinity. Flash fires involving premixed fuel/air
mixtures (which produce very little CO until the predominant fuel becomes the pool of
residual liquid or solid fuels ignited by the flash fire) or conflagrations in enclosed structures
can consume enough of the oxygen in the room, replacing it with CO and CO2, that the lit-
tle remaining oxygen is incapable of supporting life. Even if the victim avoids inhaling the
hot gases, there may not be enough oxygen to maintain consciousness (hypoxia), and the
victim collapses.59 Oxygen concentrations between 15 and 21 percent have no effect on
healthy individuals; atmospheric oxygen concentrations below 15 percent can cause dizzi-
ness, confusion, and loss of motor skills. Concentrations below 10 percent will lead to
unconsciousness and death. In a fire, the carbon dioxide produced dilutes the available oxy-
gen and increases the breathing rate, making the situation even more critical.

Heating Equipment
Of the 480 deaths each year in the United States from CO from nonfire sources, the
Consumer Product Safety Commission (CPSC) estimates that about 33 percent are from CO
emissions from consumer products and that some 47 percent of those incidents involve heat-
ing appliances (gas, oil, or wood).60 Heating equipment is always suspect when an asphyxi-
ated person is found in a residence with no fire damage and no automobile. Some general
principles of the operation of nonelectric heating systems are important to understand. For a
dangerous CO level to be created in an environment, two factors are essential:
■ The fire (combustion) being maintained in the equipment is generating carbon
monoxide in dangerous quantities.
■ The gas being generated is escaping into the room air instead of being vented.
640 Chapter 15 Fire-Related Deaths and Injuries
 FIGURE 15.22 Typical
venturi operation for gas
burner under various

No turbulence
configurations.

Normal Rich Lean

Turbulent flow
Angular Obstructed Deflection

These two factors are so important that they must never be overlooked if errors are to be
avoided. Most gas appliance burners generate some CO, and the color of the flame is one
clue to its efficiency, since an orange or yellow flame is probably producing significant CO in
addition to the carbon that is making the flame yellow. A blue flame of natural gas properly
adjusted, for example, reflects almost complete conversion of the fuel to CO2 and H2O and
generates very little CO. Such a flame may not be a hazard even if all its combustion gases
are escaping directly into the room (except for producing CO2 and replacing the oxygen in
the room). Venting of gas appliances would not be necessary if the flames were perfectly
adjusted at all times and adequate replacement air was provided (see Figure 15-22). Because
flames do not always stay perfectly adjusted and there is no way to ensure replacement air,
there is a strong and valid argument for the adequate venting that is required by the codes of
most jurisdictions. (See Chapter 6 for an example of a malfunctioning water heater.)
Once the cause of generation of excessive amounts of carbon monoxide is established, it
is then necessary to locate the reason that the venting of the appliance did not harmlessly
remove the gas but rather delivered it into the air that was breathed by the victim. This reason
always has to do with the system by which combustion gases are vented or kept from enter-
ing the environment of the persons utilizing the appliance. Examination of water heaters, gas
stoves, fireplaces, and their vents is beyond the scope of this text (except for the related causes
of fire) and will not be dealt with here. However, visual examination of the vents may reveal
readily identifiable blockages such as birds nests, trash, or accumulated debris.
“Heating equipment” sometimes includes fireplaces, wood stoves, and even charcoal
barbecue grills. All these produce CO (especially the charcoal), so if their products are not
properly vented to the outside, a fatal concentration of CO can quickly develop in a
room. Charcoal grills being used for indoor cooking as well as heating are responsible for
a large number of non-fire CO deaths. Portable LP gas heaters and salamanders can also
produce CO if used in confined spaces or if improperly adjusted. Gasoline-powered elec-
tric generators and extractor fans can produce lethal CO concentrations if used in con-
fined spaces for a suitable length of time.

Automobile Exhaust/Industrial Processes

Deaths involving CO from non-fire sources such as motor vehicles or industrial processes
will not often concern the fire investigator and will not be dealt with in depth here. The
Chapter 15 Fire-Related Deaths and Injuries 641
 investigator should remember that whether there is a fire or not, a source must be identi-
fied and a mechanism for its transport confirmed before it is blamed.
For many years, a certain means of committing suicide was to connect the exhaust pipe
of an automobile with the interior of the vehicle by means of a hose. With today’s emis-
sion-controlled engines, the CO content of the exhaust is on the order of 0.1 to 1.0 per-
cent at the tailpipe (in warm engines). Dilution of this low concentration by introducing it
into a car produces a level that may be too low to be fatal. The would-be suicide may be
waiting for hours for death while the car, whose catalytic converter was not designed for
long periods of stationary idling, ignites a fire inside the vehicle by overheating the exhaust
system. The victim then succumbs to the fire gases. In tightly sealed garages, the running
modern engine may consume so much oxygen that the victim succumbs to carbon dioxide
or oxygen depletion before a lethal dose of CO is accumulated.61
For many years, the piped gas for heating and appliances in most cities was “town
gas,” “coal gas,” or “producer gas,” which was made by heating coal to yield methane,
hydrogen, and carbon monoxide. Inhalation of this product could be fatal in a very short
time. Today nearly all such systems use natural gas, which contains no carbon monoxide.
Would-be kitchen suicides find themselves inhaling large quantities of gas with no fatal
results, unless the flammable mixture thus produced is ignited by open flame or electric
arc, with spectacular, if tragic, deflagrations resulting.

INVESTIGATION OF CARBON MONOXIDE ASPHYXIATIONS

Make Sure the Victim Actually Died from Carbon Monoxide
Carbon monoxide often produces a bright pink or red color in the victim’s skin if present in
concentrations greater than 30 percent COHb.62 This coloration should be used only as a
preliminary indicator because there are other conditions that can produce such coloration—
for example, death by hypothermia and cyanide poisoning. The pink coloration may not be
detectable in all people (except in the gums, nailbeds, or internal organs). It may also be
masked by decomposition or fire effects (smoke and thermal effects). The postmortem
report must, in such cases, include a COHb value (preferably a numerical value). If there
was fatal CO asphyxiation, the value will generally be over 40 percent but may occasion-
ally be lower. If much lower, there will be real doubt as to the cause of death, and the inves-
tigation may take a different direction. (The multiple contributions fires can make toward
death were discussed earlier in this chapter.) The measurement of CO level in blood is a rel-
atively straightforward test that can be accomplished by chemical assay or gas chromato-
graphic analysis.63 Levels below 10 percent may be reported by the laboratory as a negative
because of the expected “background” levels in smokers. Request an exact percentage. The
percentage of smokers and nonsmokers exposed to cigarette smoke is considerably lower
today than 20 years ago. In living subjects, Suchard advises that standard pulse oximeters
used in emergency rooms do not detect or accurately measure COHb and OHb levels,
because COHb maintains a bright red color that other devices interpret as regular OHb.64
Determine a Probable Origin
The circumstances of the asphyxiation will normally throw heavy suspicion on a partic-
ular source. In a house that did not burn, gas- or oil-fired appliances are first-choice
offenders. If a fire occurred in the house, the fire is the most probable source. Such prob-
able sources are generally apparent to the first person on the scene, and they require lit-
tle effort to identify. CO deaths have occasionally been documented in open air, when a
vehicle trapped in mud continued to run, and when swimmers in the water behind a run-
ning motorboat or sitting on the shelf behind the transom of a ski boat were killed by
CO.65 Such cases might erroneously be dismissed as simple natural deaths or drownings.
Structural Fires
Because deaths can occur from CO both in the vicinity of the fire and remote from it, a care-
ful and complete inquiry must be made into every death at a fire scene. A complete postmortem
642 Chapter 15 Fire-Related Deaths and Injuries
is required to ensure that the true cause of death is established. Contributions to the manner
of death from CO, smoke, toxic gases, drugs, alcohol, and illness must be taken into
account. As in every type of investigation, all the factors must balance out. All the pieces
must fit. The investigator must be careful not to jump to a convenient conclusion. Computer
modeling techniques (such as NIST’s Fire Dynamics Simulator—FDS) allow prediction of
species concentration in smoke at any position in the developing fire. Such predictions have
been used to predict times to incapacitation of occupants or firefighters.66
Carbon monoxide detectors are becoming more common in residences in the United
States. They have been credited with reducing deaths from all three major sources of
CO—undetected smoldering fires, malfunctioning gas appliances, or exhaust from run-
ning vehicles seeping into attached houses. It must be remembered that carbon monoxide
has a very wide flammable range (12 to 74 percent) and a vapor density of nearly 1.0.
This means that if it is at the same temperature as the air, it will not settle or rise in an
enclosure and will diffuse very readily throughout the space to form a deadly mixture
from both the life safety and fire standpoints.

Other Toxic Gases

As discussed in previous chapters, many compounds found in building products, furnish-
ings, and household products produce significant quantities of toxic gases as they burn.
The actual chemistries of their production are very complex. The final products are
dependent not only on the fuel but also on the temperatures involved and the amount of
ventilation available (pyrolysis versus oxidation). Considerable work is underway in fire
toxicology to enable one to predict the toxic effects of various products.67

HYDROGEN CYANIDE AND OTHER TOXIC GASES

Aside from CO, the predominant toxic gas hazard is probably the hydrogen cyanide
(HCN) produced by many common polymers that contain nitrogen. Urethanes are the
most common because they are present in many coatings, paints, and varnishes; flexible
polyurethane foam rubbers are found in nearly all upholstered furniture; and rigid ure-
thane foams are used in ceiling and wall insulation. Other materials that contain nitrogen
in their molecular structure will produce HCN as well as nitrogen dioxide (NO2) as a
combustion by-product—hair, wool, nylon, silk, and urea melamine and acrylonitrile
polymers. Because these products are nearly universal in modern structures, they must be
considered a threat to life in all fires. In deaths where the carbon monoxide level of
the blood is found to be low and there appears to be no other cause of death, an assay of
the blood or tissues should be carried out for HCN. Exposure to levels of 0.3 percent in
air is rapidly fatal, so firefighters and police personnel involved in the rescue operations
must exercise due caution.68 Hydrogen cyanide remains in water-soaked debris for long
periods of time, and fire investigators who rely on their sense of smell to detect flamma-
ble liquids should be aware that low-level exposure to HCN can reduce the sensitivity of
the human nose, so that lethal concentrations cannot be smelled.
Other toxic gases such as nitrous oxide (N2O) and nitric oxide (NO) are produced by
the same nitrogen-containing polymers described earlier. Polymers containing chlorine,
predominately polyvinyl chloride (PVC), can produce large amounts of hydrogen chloride
(HCl), a very toxic gas that turns into corrosive hydrochloric acid on contact with water,
as well as elementary chlorine. PVC pipe is widely used in water and sewer connections
in homes, mobile homes, and some light industrial structures and poses a widespread risk
both to firefighters and to fire investigators. If Freon refrigerants or related halogen-based
fire extinguisher gases are released into a hot fire, hydrochloric acid (HCl) and hydroflu-
oric acid (HF) may be formed, further adding to inhalant dangers. These products (and
the acrolein produced by wood combustion) are strong irritants that cause watering eyes
and convulsive coughing that can interfere with escape.
Chapter 15 Fire-Related Deaths and Injuries 643
 TOXIC GASES FROM SULFUR-CONTAINING POLYMERS
Sulfur-containing polymers—polyether sulfone and vulcanized rubber among them—pro-
duce sulfur dioxide (SO2 ), hydrogen sulfide (H2S), and carbonyl sulfide (COS). Carbonyl
sulfide is considered to be more toxic than carbon monoxide.69 Polystyrene, most often
found in packaging materials, household products, and light fixture diffusers, produces
styrene monomer on combustion, another toxic compound. All polymers and petroleum
products produce formaldehyde, acrolein, and similar volatile toxic materials when burn-
ing unconfined. In addition to their direct toxic effects, HCl and acrolein are very active
irritants. In even modest concentrations, they cause irritation of the mucous membranes
of the esophagus and lungs, which can interfere with breathing. It has also been suggested
that chemical compounds called free radicals produced in flames, if inhaled, interfere
directly with the transport of oxygen from the air to the bloodstream. This causes anoxia,
which can lead to unconsciousness in 10 to 15 seconds.70 Such an effect could bring about
death in spite of other nonfatal indicators such as low CO levels, minimal toxins in the
body, or even burns or soot in the throat and lungs of the deceased. Finally, a method has
been reported in which gasoline and kerosene-type hydrocarbons have been detected in
the blood of subjects exposed to petroleum distillate vapors. The levels detected by gas
chromatography/mass spectrometry (GC/MS) could be produced by breathing an explo-
sive vapor atmosphere.71 In addition, very sensitive GC/MS methods have been reported
that are capable of detecting other volatile traces in blood or tissue samples, which can
help in the reconstruction of fire fatality incidents. For instance, the detection of normal
alkanes and methyl t-butyl ether (MTBE) can reveal the presence of automotive gasoline,
while the detection of acetone (found in some paint thinners) can confirm the involvement
of such a solvent.72 Volatiles have reportedly been extracted from human skin and tissues
using solid-phase microextraction (SPME) and activated carbon strip.73

Other Mechanisms
The human body is able to adapt to hot environments by using blood circulation and
evaporative cooling (from the skin and from the mucosal tissues of the mouth, throat, and
lungs) to maintain a constant body temperature. If the body core temperature exceeds
43°C (109°F), the brain stops functioning and the person loses consciousness and will die.
If exposed to dry air at high temperatures 120°C (250°F) or higher for more than a few
minutes, the body loses the ability to maintain its core temperature, with heatstroke
occurring at a core body temperature of 40°C (105°F), leading to a lethal cascade of ther-
moregulatory and physiological events. Heatstroke can occur sooner or at lower temper-
atures if the air is moisture-laden (which prevents evaporative cooling).74
Blunt trauma can also kill people during a fire. Such trauma can result from struc-
tural collapse, explosion fragmentation, from falls, or from crashing into door frames or
walls while attempting to escape the smoke. The nature of the impact and the location in
which the victim was found may suggest the origin of such injuries. Bloodstains on walls
and floors may reveal contact with hands, fingers, shoes, and help reconstruct the impact
area and the attempted path of escape.
While the gaseous components of smoke are normally considered to be the danger-
ous components of the smoke from a fire, the soot itself may also play a role. Soot is car-
bon condensed into solid particles of various sizes. Hot particles of soot are not readily
cooled by evaporative cooling when they are inhaled, so they can retain their heat until
they lodge in the tissue itself, causing burns and edema. The carbon can also carry chem-
ically active and toxic compounds from the combustion (being activated carbon). When
inhaled or ingested, these toxins will be transferred directly to the mucosal tissues and
absorbed into the bloodstream. Soot can also be so thick in the smoke that it completely
obscures light and any view of an escape route. In extreme cases, the soot is inhaled in
such quantities that it physically blocks the airways and causes asphyxiation.
644 Chapter 15 Fire-Related Deaths and Injuries
BURN INJURIES
Even when they do not result in death, fires are responsible for an enormous toll in human
injuries. As such they may come to the attention of the fire investigator. It is a sad obser-
vation, but it is generally true in the United States that fire fatalities are much better doc-
umented than fires in which the victims are only injured, yet the civil and criminal
involvement can be more substantial in injury fires than in fatal fires. The burn patterns
on flesh and clothing are just as important in injury cases as in death cases, and yet, unfor-
tunately, the clothing of a burn victim is likely to be discarded in the emergency room and
there will be no attempt to document the injuries until after treatment is well underway,
sometimes only after weeks of healing. The investigator must realize that serious burns
may result in death after weeks of hospitalization, and that a simple household fire may
become a fire death case. That makes it mandatory to take the same steps to preserve
physical evidence in injury cases as in immediate-death fires to ensure that clothing, shoes,
and other items associated with the victim are preserved for presence of accelerants as
well as reconstruction of burn patterns and to see that injuries are documented accurately
by means of photographs as well as by notes and diagrams. The notes made by physicians
on admission are often sketchy, as the physician is naturally more interested in beginning
treatment rather than documenting exactly where the burns are located. Reports of burn
injuries often include two values that are sometimes misunderstood by the investigator.
First, the depth of injury to the tissue is often reported in degrees:
■ First degree: superficial, reddened skin (like ordinary sunburn)
■ Second degree: deeper damage with blisters formed and sloughing of epidermis;
partial-thickness burns that will heal by cell growth from surviving underlying
tissue (dermis)
■ Third degree: severe damage to full thickness of skin to underlying tissue; full-
thickness burns that can heal only from the edges and generally require skin grafts
■ Fourth degree: very severe including underlying tissue with charring
■ Fifth degree: underlying muscle and bone involved

Second, the predominant means of estimating the amount of body (skin) surface area
that is involved is the “Rule of Nines,” whereby major body areas are assigned various
values (for adults) as follows:
■ 9 percent for the head (front, back, and sides)
■ 9 percent for each arm
■ 9 percent for the back of each leg
■ 9 percent for the front of each leg
■ 18 percent for the back of the torso
■ 18 percent for the front of the torso
■ 1 percent for the genitalia

Different percentages are used for infants and children because of their different body
proportions.
Burns, particularly first, second, and third degree, can be created by exposure to radi-
ant or convected heat, or heat conducted from hot surfaces or liquids, or by chemical
action. Skin can blister after about 11 seconds of exposure to radiant heat flux of 10 kW/m2
or after about 5 seconds at 20 kW/m2.75 A full-thickness burn can be produced by immer-
sion in hot water for 30 seconds at 54°C (130°F) or 1 second at 70°C (158°F) in adults.
A child’s skin is more fragile, sustaining full-thickness burns after immersion in hot water
for 10 seconds at 54°C (130°F) or 0.5 second at 65°C (149°F).76 Because pain acts as a
deterrent for someone seeking an escape route in a fire, a threshold heat flux of 2.5 kW/m2
or sustained exposure to hot gases in excess of 120°C (250°F) is considered to be the
“incapacitation” level.77
Chapter 15 Fire-Related Deaths and Injuries 645
FIGURE 15.23 The
wearer of these denim
pants suffered only local-
ized second-degree burns
(in the vicinity of the tear
in the upper left leg)
despite the fact that
gasoline had been
splashed across her legs.
The tight denim pants
over knit thermal under-
wear conducted enough
heat to produce a burn
under the tear. Courtesy of
John D. DeHaan.

The flammability properties of fabrics were treated in Chapter 11. While injuries are
usually matchable to burned or melted portions of the garments worn at the time of the
fire, burn injuries may not tally with the clothing damage. Clothing of natural fibers can
conduct heat sufficiently so that second-degree burns can be produced through cloth
even if the cloth remains intact. Because many synthetic cloth fibers are capable of trans-
mitting infrared radiation, burns can be induced by radiant heat through single layers of
synthetic cloth. One interesting observation is that some blue denim fabric turns red
upon exposure to temperatures in excess of about 60°C (130°F), whereas scorching of
the cotton fiber or even melting of the acrylic fiber requires much higher temperatures.
All these observations mean that fire injuries to the body can be correlated in many cases
with damage to the clothing as long as the physical and chemical properties of the fab-
rics are appreciated and fire behavior is well understood (see Figure 15-23). A sustained
fire supported by heavy clothing can cause third-degree burns and skin splitting, as seen
in Figure 15-24.
The thermal inertia of extraskeletal human tissue is not unlike that of polyethylene
plastic or even wood, so the length of time the body is exposed will establish how deep
the damage will be. A flash fire involving flammable liquid vapors (or any bright radiant
heat exposure) is of such short duration that only the epidermis may be affected. The epi-
dermis may start to separate from the underlying dermis, raising blisters (similar to those
formed when wallpaper or paint is heated on a wall). The blisters have no tissue in con-
tact beneath and are quickly heated and burned away, sometimes leaving a spotted pat-
tern on the skin of blackened skin surrounding oval patches of freshly exposed dermis (on
living or postmortem skin), as seen in Figure 15-25. If an ignitable liquid is burned on a
horizontal skin surface, the damage may resemble the halo pattern described earlier for
carpet and floor. As seen in Figure 15-26, a ring or halo of blister was raised around the
central portion where the evaporating liquid protected the skin. The hair was singed off
in both areas. If the skin surface is vertical, the burning liquid may pour off so quickly
that the skin is only reddened, unless retained by garments. Due to the mass of the torso,
it takes a long time for normal external fire exposure to have an effect on the deep tissues
646 Chapter 15 Fire-Related Deaths and Injuries
 FIGURE 15.24 Sustained
fire from gasoline-
dampened clothing
induced skin splits and
charring. Compare with
Figure 15-12B. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

and organs. Tests monitoring internal body temperature demonstrate that 60 to 120 minutes
of exposure to a generalized low-intensity fire is necessary to raise the core body temper-
ature of a cadaver.78
Documentation (by diagram, as in Figure 15-27, and by photography) should be car-
ried out on all burn victims (living and deceased) as soon after the fire as possible. This
ensures accurate assessment before medical attention and healing can distort the appear-
ance of this transitory evidence. The burn patterns can then be compared with the burn
patterns of clothing and other surfaces at the scene.

FIGURE 15.25 Radiant

heat caused postmortem
“leopard spot” blisters
on this cadaver. Courtesy
of Jamie Novak, Novak
Investigations and St. Paul
Fire Dept.

Chapter 15 Fire-Related Deaths and Injuries 647

 FIGURE 15.26 Cigarette
lighter fuel poured on the
back of the hand (hori-
zontal surface) and
ignited produced a halo
pattern blister around the
pool. The liquid pool pro-
tected the skin under the
liquid. Courtesy of Dr. Roger
Berrett, Weybridge, Surrey,
UK.

Body diagram
Front Back

FIGURE 15.27 Simple

body diagram useful for
recording extent and Name
depth of burn injuries.
Examined by Date

648 Chapter 15 Fire-Related Deaths and Injuries

MANNER OF DEATH
The determination of manner of death (accidental, suicidal, homicidal, etc.) is usually made
by medicolegal authorities in consultation with the fire investigators, since it is the set of
circumstances under which the fire injuries (cause of death) were suffered that establishes
manner. For the reasons described in the early pages of this chapter, it is clear that the fire
death investigation is a three-pronged affair—cause of death, cause of fire, and conditions
under which both occurred. The investigator should not presume an accidental death
occurred any more than he or she should presume the fire was accidental. A homicidal
death or a natural death may result in an accidental fire. An intentionally set fire may
result in unintended death. A suicide death may result in an accidental fire or may be the
result of an intentionally set fire. Suicides by fire are rare (estimates are on the order of
1 percent of all identified suicides). Most that are identified are the result (sometimes
observed) of pouring and igniting an ignitable liquid over oneself, but there are variations.
One of the authors knows of one case in which the individual set fire to a pile of wood
using kerosene and sat amid the burning material for some time. The decedent apparently
changed her mind, since the body was ultimately found several feet from the fire site after
the fire had self-extinguished. In another incident that was filmed by a passerby, a man
ignited the interior of his parked car, waited until it was fully involved, and then jumped
through the passenger window into the flames just as firefighters arrived. Other suicides
are less direct, involving setting fire to a structure or a vehicle and then waiting in the
smoke-filled enclosure (away from the fire). One case was published by investigators from
Montreal (structure fire).79 DeHaan investigated a structure fire in a residence that
resulted in smoke inhalation deaths of four children and the near death of the firesetter
and an older child, who survived in the smoke-filled house for several hours.80 There was
also a case in which the individual, despondent over a pending jail term, set fire to his
vehicle and then crawled into the trunk of the vehicle, where he was found. He broke a
rear window before the fire to ensure adequate ventilation for the fire. A similar case was
described by Adair.81 The full circumstances of events preceding the fire, physical evidence,
postmortem findings, and fire engineering analysis of heat and toxic gas (CO) production
and routes of incapacitation are needed to reach a reliable conclusion in such challenging
cases.

Chapter 15 Fire-Related Deaths and Injuries 649

 CHAPTER REVIEW

Summary
There is a temptation, especially for the pathologist or 9. All burns and physical trauma on the body should
death investigator, to regard a fire as a single event (like be documented by photos, diagrams, and notes.
a gunshot) and assess the physical and toxicological 10. The carbon monoxide level in the blood should
effects seen on the body as the result of exposure to that be determined in all fire fatalities and, if possible,
event. Those familiar with fire know that fire is not uni- in living victims rescued from the fire.
form in its intensity or the distribution of its toxic gases 11. The presence of alcohol, drugs, and toxic materials
and that humans can be exposed to it in a tremendous should be determined in all cases where it is
variety of ways over a period of time. The hot and toxic deemed warranted.
gases in many room fires will be almost completely con- 12. The body represents an important part of the
centrated in the hot smoke layer, so a person reclining fuel load of many fires and often bears burn
on a bed or the floor of a room may not be exposed to patterns that must be interpreted as part of the
any of it until he or she sits or stands up. Fire can kill by burn patterns in the rest of the room. The fire
any number of means, and people can react to it in a dynamics and ignition analyses may include the
wide variety of ways—from ignoring it, to going to body and its clothing.
investigate, to fleeing, to calling for assistance, or escap- 13. Clothing bears important burn patterns in both
ing and then returning to rescue people, pets, or senti- fatality and injury cases and must be preserved.
mental items. When one or more persons die in a fire Burn injuries suffered by surviving victims must
and others escape there is always the question: Why did be preserved by photographs as soon as possible
someone survive when others did not? after the fire. Residues of ignitable liquids may
In summary, the following are the primary issues be detected both in clothing and skin.
facing the fire investigator in dealing with a fire-related 14. Fires and people’s interactions with them are
death: often complex and prolonged. The cause of death
is rarely a single event or condition. Combinations
1. Assistance of specialized personnel should be
of effects such as shock, CO, hypoxia, hyperther-
obtained as soon as possible.
mia, and toxic gases may result in death when
2. These personnel should meet (preferably ahead
no single factor would be fatal.
of time) to set a protocol for who does what at
15. The cause of the fire and the cause of death may
these scenes.
or may not be related. The relationship between
3. Access to the scene and in particular the area
the victim and the fire may be determined by
around the body must be tightly controlled.
assessing both in terms of the spread of fire and
4. The body should be fully documented with
its heat and smoke. Accidental fires may occur
photos and diagrams before it is moved if
whether the death is accidental, suicidal, natural,
circumstances permit.
or homicidal. Similarly, an arson fire may trigger
5. Extreme care must be used in the removal of the
an accidental or natural death or be part of a
body and associated debris. The debris around/
homicide. This consideration should extend to
under the body should be sifted and checked
vehicle fires. Both accidental and incendiary fires
with a metal detector and X-rays. Gridding and
have been linked to numerous deaths.82
layering may be necessary.
6. The area of the body should be photographed The fire investigator, homicide investigator, and
before, during, and after debris removal (by pathologist must be prepared to exchange all informa-
layering). tion fully and to work in close cooperation with one
7. The body should be completely X-rayed prior another to expedite the identification of the victim and
to examination. determination of the cause and manner of death. Such
8. A complete forensic autopsy should be carried cooperation will maximize the information available
out in all fire deaths to determine the cause of so as to accurately reconstruct the sequence of events
death. at a fire death scene.

650 Chapter 15 Fire-Related Deaths and Injuries

Review Questions
1. List the six major questions that must be answered 6. What is the relationship between CO concentration,
when burned remains are discovered in a fire scene. breathing rate, time, and COHb saturation?
2. Who are the members of the ideal fire death 7. What is the range of potentially fatal COHb
investigation team? saturations and why is it so large?
3. What is the “wick effect” and how does it 8. Describe the “degree” system of burn injuries
contribute to the destruction of a body? and relate it to human skin.
4. How hot is a commercial crematorium and how 9. Name three other gases that can contribute to
long does it take to cremate a normal adult human the death of a fire victim.
body? 10. What are the three elements of a fire death
5. Name five effects heat and fire exposure can have investigation?
on a human body.

References
1. M. J. Karter Jr., “Fire Loss in the United States 2008” 8. A. Z. Mundorff, E. J. Bartelink, and E. Mar-Cash,
(Quincy, MA: National Fire Protection Association, “DNA Preservation in Skeletal Remains from the
Fire Analysis and Research Division, August 2009); see World Trade Center Disaster,” Journal of Forensic
also M. J. Karter Jr., “U.S. Fire Loss for 2008,” NFPA Sciences 54, no. 4 (July 2009): 739– 45; National
Journal (September–October 2009). Institute of Justice, Mass Fatality Incidents: A Guide
2. D. A. Purser, “Toxicity Assessment of Combustion for Human Forensic Identification (Washington DC:
Products,” Chap. 2–6 in SFPE Handbook of Fire NIJ, 2005).
Protection Engineering, 3rd ed. (Bethesda, MD: 9. M. J. Hickman et al., “Medical Examiners’ and
Society of Fire Protection Engineers, 2002), 2–83, 84. Coroners’ Offices, 2004” (Washington, DC: U.S.
3. V. F. Delattre, “Burned Beyond Recognition: Systematic Department of Justice, Bureau of Justice Statistics,
Approach to the Dental Identification of Charred June 2007).
Human Remains,” Journal of Forensic Sciences 45, 10. J. M. Fleming, “Smoke Detector Technology and the
no. 3 (July 2000): 589–96. Investigation of Fatal Fires,” Fire and Arson Investigator
4. F. G. Robinson, F. A. Rueggeberg, and P. E. Lockwood, (April 2000): 35– 40.
“Thermal Stability of Direct Dental Esthetic Restorative 11. J. D. DeHaan and E. J. Pope, “Sustained Combustion
Materials at Elevated Temperatures,” Journal of of Bodies: Some Observations,” American Association
Forensic Sciences 43, no. 6 (November 1998): for Forensic Science, Denver, February 2009 (in
1163–67. press).
5. M. A. Bush, P. J. Bush, and R. G. Miller, “Detection of 12. E. J. Pope, O. C. Smith, and T. G. Huff, “Exploding
Classification of Composite Resins in Incinerated Teeth Skulls and Other Myths about How the Human Body
for Forensic Purposes,” Journal of Forensic Sciences Burns,” Fire and Arson Investigator (April 2004):
51, no. 3 (May 2006): 636– 42. 23–28.
6. N. J. Kirk et al., “Skeletal Identification Using the 13. T. Watanabe, Atlas of Legal Medicine (Philadelphia:
Frontal Sinus Region,” Journal of Forensic Sciences 47, J. B. Lippincott, 1968), 136.
no. 2 (March 2002): 318–23; see also C. W. Schmidt, 14. M. J. Thali et al., “Charred Body: Virtual Autopsy
“The Recovery and Study of Burned Human Teeth” with Multi-Slice Computed Tomography and Magnetic
in The Analysis of Burned Human Remains, ed. C. W. Resonance Imaging,” Journal of Forensic Sciences 47,
Schmidt and S. A. Symes, 55–74 (New York: Academic no. 1 (March 2002): 1326–32.
Press, 2008). 15. W. G. Eckert, S. James, and S. Katchis, “Investigation
7. J. di Zinno et al., “The Waco Texas Incident: The of Cremations and Severely Burned Bodies,” American
Use of DNA Analysis in the Identification of Human Journal of Forensic Medicine and Pathology 9, no. 3
Remains,” paper presented at Fifth International (1988): 188–200; M. Bohnert, T. Rost, and S. Pollak,
Symposium on Human Identification, London, England, “The Degree of Destruction of Human Bodies in
1994; see also D. W. Owsley et al., “The Role of Relation to Duration of the Fire,” Forensic Science
Forensic Anthropology in the Recovery and Analysis International 95, no. 1 (July 1998): 11–21.
of Branch Davidian Compound Victims: Techniques 16. E. J. Pope, “Burning Bodies,” paper presented at
of Analysis,” Journal of Forensic Sciences 40, no. 3 American Academy of Forensic Sciences, New Orleans,
(May 1995): 341– 48. LA, February 2005.
Chapter 15 Fire-Related Deaths and Injuries 651
17. DeHaan and Pope “Sustained Combustion of Bodies.” 33. Pope et al., “Exploding Skulls and Other Myths.”
18. C. S. Petty, “Fire Death Investigation and Pathology,” 34. N. P. Herrmann and J. L. Bennett, “The Differentiation
Fire and Arson Investigator (April–June 1968). of Traumatic and Heat-Related Fractures in Burned
19. J. D. DeHaan and F. Fisher, “Reconstruction of a Bone,” Journal of Forensic Sciences 44, no. 3
Vehicle Fire Fatality,” paper presented at the First (May 1999): 461–69; E. J. Pope and O. C. Smith,
Fire Investigation and Research Conference, Fire “Identification of Traumatic Injury in Burned Cranial
Research Station, Garston, UK, October 1995; see Bone: An Experimental Approach,” Journal of
also “Reconstruction of a Fatal Fire in a Parked Motor Forensic Sciences 49, no. 3 (May 2004): 431–37;
Vehicle,” Fire and Arson Investigator (January 2003): Symes et al., “Patterned Thermal Destruction of
42– 46. Human Remains.”
20. C. Palmiere et al., “Ignition of a Human Body by a 35. Hermann and Bennett, “The Differentiation of
Modest External Source: A Case Report,” Forensic Traumatic and Heat-Related Fractures.”
Science International 188, no. 1 (July 2009): 1–3 36. Eckert, James, and Katchis, “Investigation of
(e17–e19). Cremations”; T. J. O. Thompson, “Heat Induced
21. J. Nickell and J. F. Fischer, “Spontaneous Human Dimensional Changes in Bone and their Consequences
Combustion,” Fire and Arson Investigator 34 (March for Forensic Anthropology,” Journal of Forensic
1984): 4–11; J. Nickell and J. F. Fischer, “Incredible Sciences 50, no. 5 (September 2005): 1008–15.
Cremations: Investigating Spontaneous Combustion 37. Symes et al., “Patterned Thermal Destruction of Human
Deaths,” The Skeptical Inquirer 11 (Summer 1987): Remains.”
352–57. 38. A. M. Christensen, “Experiments in the Combustibility
22. B. V. Ettling, “Consumption of an Animal Carcass of the Human Body,” Journal of Forensic Sciences 47,
in a Fire,” Fire and Arson Investigator (April–June no. 3 (May 2002): 466–70.
1968). 39. Pope and Smith, “Identification of Traumatic Injury
23. D. J. Gee, “A Case of Spontaneous Combustion,” in Burned Cranial Bone.”
Medicine, Science and Law 5 (January 1965): 40. L. A. Simpson and C. W. Cruse, “Gasoline Immersion
37–38. Injury,” Plastic and Reconstructive Surgery” 671
24. J. D. DeHaan, S. Nurbakhsh, and S. J. Campbell, (January 1981): 54–57; J. F. Hansbrough et al.,
“The Combustion of Animal Remains and Its “Hydrocarbon Contact Injuries,” Journal of Trauma
Implications for the Consumption of Human Bodies 25, no. 3 (1985): 250–52.
in Fire,” Science and Justice 39, no. 1 (January 1999): 41. DeHaan, “The Dynamics of Flash Fires.”
27–38. 42. T. W. Adair et al., “Suicide by Fire in a Car Trunk:
25. NFPA. Fire Protection Handbook, 20th ed. (Quincy, A Case with Potential Pitfalls,” Journal of Forensic
MA: National Fire Protection Association, 2008), Sciences 48, no. 3 (September 2003): 1113–17; J. D.
app. A. DeHaan, unpublished case report.
26. J. D. DeHaan and M. Scanlan, “A Case of Not-So- 43. National Fire Protection Association, Report on
Spontaneous Human Combustion,” CAC News, the Fire at the Dupont Plaza Hotel and Casino,
1996; also in Fire and Arson Investigator (June 1997): December 31, 1986 (Quincy, MA: NFPA, 1987);
14–16. W. D. Wooley et al., “The Stardust Disco Fire,
27. J. D. DeHaan and S. Nurbakhsh, “The Sustained Dublin 1981: Studies of Combustion Product during
Combustion of an Animal Carcass and Its Implications Simulation Experiments,” Fire Safety Journal 7
for the Consumption of Human Bodies in Fires,” Journal (1984): 267–83.
of Forensic Sciences 46, no. 5 (September 2001): 44. B. C. Levin et al., “Analysis of Carboxyhemoglobin
1076–81. and Cyanide in Blood from Victims of the Dupont
28. DeHaan et al., “The Combustion of Animal Remains.” Plaza Hotal Fire in Puerto Rico,” Journal of Forensic
29. DeHaan and Pope, “Sustained Combustion of Sciences 35, no. 1 (1990): 151–68.
Bodies.” 45. D. G. Penney, ed., Carbon Monoxide Toxicity (Boca
30. J. D. DeHaan, “The Dynamics of Flash Fires Involving Raton, FL: CRC Press, 2000).
Flammable Hydrocarbon Liquids,” American Journal 46. P. B. Hawk, B. L. Oser, and W. H. Summerson, Practical
of Forensic Medicine and Pathology 17, no. 1 (March Physiological Chemistry, 12th ed. (Philadelphia:
1996): 24–31. Blakiston, 1947).
31. DeHaan and Nurbakhsh, “The Sustained Combustion 47. J. M. Feld, “The Physiology and Biochemistry of
of an Animal Carcass.” Combustion Toxicology” in Proceedings of Fire Risk
32. S. A. Symes et al., “Patterned Thermal Destruction and Hazard Assessment: Research Applications
of Human Remains in a Forensic Setting” in The Symposium (Quincy, MA: Fire Protection Research
Analysis of Burned Human Remains, ed. C. W. Schmidt Foundation, 2002); H. J. Vreman, R. J. Wong, and
and S. A. Symes (New YorK: Academic Press, 2008), D. K. Stevenson, “Carbon Monoxide in Breath, Blood,
15–54. and Other Tissues,” Chap. 2 in Carbon Monoxide

652 Chapter 15 Fire-Related Deaths and Injuries

 Toxicity, ed. D. G. Penney (Boca Raton, FL: CRC Carbon Monoxide Deaths Immediately at the Death
Press, 2000). Scene?” Journal of Forensic Sciences 40, no. 4
48. A. C. Berent et al., “Carbon Monoxide Toxicity: A (1995): 596–98.
Case Series,” Journal of Veterinary Emergency and 63. Vreman, “Carbon Monoxide in Breath, Blood, and
Critical Care 15, no. 2 (2005): 128–35. Other Tissues,” 35–42.
49. A. Stolman and C. P. Stewart, Toxicology, Mechanisms, 64. Suchard, “Motor-Vehicle Related Toxicology,” 77.
and Analytical Method, vol. 1 (New York: Academic 65. M. I. Jumbelic, “Open Air Carbon Monoxide
Press, 1960). Poisoning,” Journal of Forensic Sciences 43, no. 1
50. Vreman et al., “Carbon Monoxide in Breath, Blood (1998): 228–30.
and Other Tissues.” 66. A. M. Christensen and D. J. Icove, “The Application
51. D. J. Icove and J. D. DeHaan, Forensic Fire Scene of NIST’s Fire Dynamics Simulator to the Investigation
Reconstruction, 2nd ed. (Upper Saddle River, NJ: of Carbon Monoxide Exposure in the Deaths of Three
Brady/Prentice Hall, 2009), 311–12; see also D. A. Pittsburgh Fire Fighters,” Journal of Forensic Sciences
Purser, “Interactions among Carbon Monoxide, 49, no. 1 (January 2004): 104–10.
Hydrogen Cyanide, Low Oxygen Hypoxia, Carbon 67. Purser, “Interactions among Carbon Monoxide”;
Dioxide, and Inhaled Irritant Gases,” Chap. 7 in W. D. Woolley and P. J. Fardell, “Basic Aspects of
Carbon Monoxide Toxicity, ed. D. G. Penney, 164 Combustion Toxicology,” Fire Safety Journal 5 (1982):
(Boca Raton, FL: CRC Press, 2000). 29– 48; K. J. Vorhees and R. Tsao, “Smoke Aerosol
52. Health Canada, “Investigating Human Exposure to Analysis by Pyrolysis—Mass Spectrometry/Pattern
Contaminants in the Environment” (Ottawa, Ontario: Recognition for Assessment of Fuels Involved in
Ministry of National Health and Welfare, 1995); Flaming Combustion,” Analytical Chemistry 57,
www.hc-sc.gc.ca. no. 8 (1985): 1630–36; J. Rio et al., “An Investigation
53. Vreman et al., “Carbon Monoxide in Breath, Blood, of the Toxic Effects of Combustion Products—Analysis
and Other Tissues.” of Smoke Components,” Journal of Analytical
54. D. A. Purser, “Assessment of Hazards to Occupants Toxicology 12, no. 5 (September–October 1988):
from Smoke, Toxic Gases, and Heat” in SFPE 274–78.
Handbook of Fire Protection Engineering, 3rd ed. 68. NFPA. Fire Protection Handbook, 20th ed. (Quincy,
(Quincy, MA: National Fire Protection Association, MA: National Fire Protection Association, 2008),
1995), 2–96; see also 172. Chaps. 3–5.
55. J. R. Gill, L. Goldfeder, and M. Stajic, “The Happyland 69. C. J. Hilado and C. M. Machado, “Effect of Char
Homicides: 87 Deaths Due to Smoke Inhalation,” Journal Yield and Specific Toxicants on Toxicity of Pyrolysis
of Forensic Sciences 48, no. 1 (January 2003): 161–66. Gases from Synthetic Polymers,” Fire Technology 16,
56. A. O. Gettler and H. C. Freimuth, “Carbon Monoxide no. 59 (February 1979).
in Blood: A Simple and Rapid Estimation,” American 70. S. Hattangadi, “A Chemical Clue to Mystery Deaths,”
Journal of Clinical and Pathological Technology Industrial Chemist, October 1987, 12–15; W. T.
Supplement 7, no. 79 (1943). Lowry et al.,” Free Radical Production from Controlled
57. M. J. Hazucha, “Effect of Carbon Monoxide on Work Low-Energy Fires: Toxicity Considerations,” Journal
and Exercise Capacity in Homes,” Chap. 5 in Carbon of Forensic Sciences 30, no. 1 (1985): 73–85.
Monoxide Toxicity, ed. D. G. Penney, 102–7 (Boca 71. K. Kimura et al., “Gasoline and Kerosene Components
Raton, FL: CRC Press, 2000). in Blood—A Forensic Analysis,” Human Toxicology 7
58. D. A. Purser, “Interactions among Carbon Monoxide, (1988).
Hydrogen Cyanide, Low Oxygen Hypoxia, Carbon 72. J. Schuberth, “Post-Mortem Test for Low-Boiling Arson
Dioxide, and Inhaled Irritant Gases.” Residues of Gasoline by Gas Chromatography—Ion
59. DeHaan, “The Dynamics of Flash Fires.” Trap Mass Spectrometry,” Journal of Chromatography
60. J. Flynn, “CO Deaths,” NFPA Journal (January–February B 662, no. 1 (December 1994): 113–17.
2008): 33–35; see also J. C. Mah, “Non-Fire Carbon 73. J. R. Almirall et al., “The Detection and Analysis
Monoxide Deaths and Injuries Associated with the of Ignitable Liquid Residues Extracted from Human
Use of Consumer Products,” (Bethesda, MD: CPSC, Skin Using SPME/GC, Journal of Forensic Sciences 45,
October 2000), 57. no. 2 (2000): 453–61; I. Montani, S. Comment, and
61. S. J. deRoux, “Suicidal Asphyxiation of Automobile O. Delemont, “The Sampling of Ignitable Liquids on
Emission without Carbon Monoxide Poisoning,” Suspects Hands,” Forensic Science International 194
Journal of Forensic Sciences 51, no. 5 (September (2010): 115–24.
2006): 1158–59; J. R. Suchard, “Motor-Vehicle 74. J. L. DeJong and T. Adams, “Entrapment in Small
Related Toxicology,” Topics in Emergency Medicine Enclosed Spaces: A Case Report and Points to
28, no. 1 (2006): 76–84. Consider Regarding the Mechanism of Death,”
62. D. Risser, A. Bonsch, and B. Schneider, “Should Journal of Forensic Sciences 46, no. 3 (2001):
Coroners Be Able to Recognize Unintentional 708–13.

Chapter 15 Fire-Related Deaths and Injuries 653

75. Purser, “Assessment of Hazards to Occupants,” Journal of Forensic Sciences 50, no. 4 (July 2005):
2–115. 937–38.
76. P. E. Besant-Matthews, “Deaths Associated with 80. DeHaan, unpublished case.
Burns, Fire, and Arson,” Dallas, TX, 1993. 81. Adair et al., “Suicide by Fire in a Car Trunk.”
77. Icove and DeHaan, Forensic Fire Scene Reconstruction. 82. Karter, “Fire Loss in the United States 2008.”
78. DeHaan and Pope, “Sustained Combustion of Bodies.”
79. A. Savageou, A. Racette, and R. Yesovitch, “Suicide by
Inhalation of Carbon Monoxide in a Residential Fire,”

654 Chapter 15 Fire-Related Deaths and Injuries

 CHAPTER
16
Arson as a Crime

Courtesy of Jamie Novak, Novak

Investigations and St. Paul Fire Dept.

KEY TERMS

arson set, p. 671 fraud, p. 662 pyromania, p. 670

corpus delicti, p. 658 motive, p. 661 torch, p. 662

OBJECTIVES

After reading this chapter, you should be able to:

■ Explain the basic concepts of arson as a crime and list its three elements.
■ Identify and list the motive classifications for arson as cited by the FBI’s National
Center for the Analysis of Violent Crime (NCAVC).
■ List and describe the typical findings for the location, arrangement, types of fuel,
and methods of initiation used by arsonists in four of the classifications.
■ List the investigative methodology and deductions for the interpretation of
evidence.
■ Search the Internet for actual case examples detailing arson cases and motives.

655
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

A
rson, in its most general terms, may be defined as the willful and malicious burn-
ing of a person’s property. It must be regarded as a possible, if not probable,
cause of every fire investigated. It was estimated at one time that arson fires
accounted for up to 31 percent of all reported structure fires.1 Informal surveys of expe-
rienced U.S. fire investigators support that up to 40 percent of all urban structure fires are
incendiary (intentionally set). Arson is a serious threat to public safety and well-being in
urban, suburban, and even rural areas. Based on data reported by U.S. fire departments
in its annual survey, the NFPA estimated there were 30,500 intentionally set structure fires
in 2008. This represents a decrease of 6.2 percent from 2007 totals. These estimates (about
6 percent of all structure fires) are the lowest in the 27 years studied. The authors note
that recent National Fire Protection Association (NFPA) estimates, however, are based on
United States Fire Administration/National Fire Incident Reporting System (USFA/NFIRS)
5.0 data, whose criteria were changed in 2002 to exclude so-called suspicious fires and
unknown cause fires. Based on the methodology used in making these estimates and
changes in reporting criteria, comparisons with data from previous years are not valid.
The reported “intentionally set” fires resulted in an estimated 315 civilian deaths and
about $866 million in direct property damage.2 In 2008, there were an estimated 17,500
intentional fires in vehicles, with property damages of around $139 million. By compari-
son, in 2002, there were an estimated 68,000 intentional fires in structures, costing $1.9
billion in direct property damage, 630 civilian deaths, and over 2,000 civilians injured.3
There were also an estimated 30,500 vehicle arsons. Because many fire departments in the
United States release many “suspected arson” cases to police authorities, their arson sta-
tistics may be inaccurate, and some state agencies do not forward fire statistics to NFIRS
at all. As a result, NFIRS incendiary estimates are very conservative.
The U.S. Department of Justice, Federal Bureau of Investigation (FBI), operates and
maintains the national collection of statistics on arson crimes known as the Uniform Crime
Reporting (UCR) Program. In 2008, based upon reporting from 13,980 law enforcement
agencies (representing an estimated population of 250,243,947 persons), there were 56,972
documented arsons in the United States. Of these incidents, 24,750 (43.4 percent) were
structures, 16,454 (28.9 percent) were mobile, and 15,768 (27.7 percent) other properties
were targets of arsonists. The average monetary loss per property was $16,015, with 18 per-
cent of the cases cleared by an arrest or other exceptional means. Juvenile offenders under

656 Chapter 16 Arson as a Crime

 TABLE 16-1 Arson by Type of Property, 2008 (13,980 agencies; 2008 estimated
population 250,243,947)

NUMBER PERCENT PERCENT OF PERCENT OF

PROPERTY OF ARSON PERCENT NOT AVERAGE TOTAL ARSONS CLEARANCES
CLASSIFICATION OFFENSES DISTRIBUTIONa IN USE DAMAGE CLEARANCES CLEAREDb UNDER 18
Total 56,972 100.0 $16,015 10,277 18.0 37.4
Total structure: 24,750 43.4 19.0 29,701 5,520 22.3 35.2
Single occupancy
residential 11,435 20.1 21.4 28,788 2,389 20.9 25.7
Other residential 4,062 7.1 13.6 29,343 983 24.2 25.0
Storage 1,629 2.9 20.7 18,021 298 18.3 48.7
Industrial/
manufacturing 251 0.4 22.3 212,388 48 19.1 41.7
Other
commercial 2,331 4.1 15.7 55,531 450 19.3 28.0
Community/
public 2,667 4.7 16.2 15,799 857 32.1 65.3
Other structure 2,375 4.2 21.4 13,668 495 20.8 47.1
Total mobile: 16,454 28.9 8,766 1,461 8.9 20.3
Motor vehicles 15,572 27.3 8,186 1,314 8.4 18.7
Other mobile 882 1.5 18,991 147 16.7 34.7
Other 15,768 27.7 2,099 3,296 20.9 48.7

Source: Uniform Crime Report 2008 (Washington DC: U.S. Department of Justice, FBI, 2009); www.fbi.gov.ucr.
a
Because of rounding, the percentages may not add to 100.0.
b
Includes arsons cleared by arrest or exceptional means.

18 years of age were responsible in 37.4 percent of the case clearances. Table 16-1 summa-
rizes the FBI’s UCR reported statistics for arson during 2008.4
The Department on Communities and Local Government in the United Kingdom (UK)
estimated that 17.8 percent (9,400) of 52,700 reported dwelling fires in 2007 were deliber-
ate, while 38.7 percent of 31,000 “other building” fires were incendiary.5 The Fire Protection
Association reported that in 2007 to 2008 in the United Kingdom, some 30 percent of all
serious structure fires [those involving over £250,000 ($415,000) loss each] were arson. (Over
the period 2000 to 2003, arson fires constituted 30 to 35.7 percent of major fires) From
December 2007 to November 2008, of 120 reported serious fires 36 (30 percent) were delib-
erate, causing over £26 million ($43,000,000) in direct damages.6 The Office of the Deputy
Prime Minister estimated that the number of arson structure fires increased about 17 per-
cent over the decade 1994–2003. Including the costs of investigation, prosecution, commu-
nity losses, and preventive measures, the total cost of arson fires in the United Kingdom is
estimated to be £2.8 billion per year.7
Arson is a difficult crime to investigate and prosecute, for the following four reasons.
First, the scene must be carefully investigated before it can be determined whether a crime
actually took place. This is unlike finding a dead body and a bloody knife, which are strong
indications that a murder may have taken place. Such indicators gain the immediate atten-
tion of investigators and warn them to proceed carefully so as not to compromise any evi-
dence present, unlike a common dwelling fire that may be investigated cursorily, if at all.

Chapter 16 Arson as a Crime 657

 Therefore, it is vitally important that every fire scene be treated as a potential arson scene
(from the standpoint of security, preservation, and evidence) until clear proof of natural or
accidental cause is discovered.
Second, the crime itself, if successful, destroys the physical evidence at its origin. In some
cases, proving the crime of arson may be likened to a homicide investigation where the vic-
tim’s body has turned to dust. Evidence of its existence is still there, but it requires careful
and methodical analysis. Arson is the one crime that destroys evidence rather than creates
it as it progresses.
Third, the official investigation of arson seems to fall into a gap between the area of
responsibility of the fire department and that of the police. Fire department personnel may
consider their primary responsibility to be public safety through extinguishment of the fire.
They may consider the warrants, interviews, and fieldwork necessary for a criminal investi-
gation to be outside their normal authority and more appropriate for the police. However,
police investigators realize that the physical evidence of the crime—that is, fire behavior,
diagnostic signs, and charred debris—is beyond their training and may decline to deal with
it. As a result of this schism, valuable evidence of one type or another is often overlooked.
Fourth, statistical data on arson are likely underreported. In addition to underreporting
in the NFIRS program, law enforcement data and trends are also spotty. It has long been
known that unless fire officials in concert with police submit proven arson cases to the FBI’s
Uniform Crime Reporting (UCR) arson system, case clearance and arrests do not appear in
national tracking of law enforcement trends.8

The Crime of Arson

The crime of arson comprises the following three elements:
1. There has been a burning of property. This must be shown to the court to be
actual destruction, at least in part, not just scorching or sooting (although some
states include any physical or visible impairment of any surface). As used here,
“burning” includes destruction by explosion.
2. The burning is incendiary in origin. Proof of the existence of an effective incendi-
ary device, no matter how simple it may be, is adequate. Proof must be accom-
plished by showing specifically how all possible natural or accidental causes have
been considered and ruled out.
3. The burning is shown to be started with malice, that is, with the intent of destroy-
ing property (i.e., starts a fire or causes an explosion with the purpose of destroy-
ing a building of another or one’s own with fire).9 The “degree” of the arson
charge in most jurisdictions has to do with the occupancy—first degree correspon-
ding to an occupied structure, second to an unoccupied structure, and third to other
property.
corpus delicti ■ The investigator must keep these three elements of proof, the corpus delicti of the
Literally, the body of crime, in mind as the scene is examined. To prove a criminal case, of course, there has to
the crime. The funda- be proof that the accused did in fact set the fire in question (or arranged to have it set).
mental facts necessary
to prove the commis-
Although the mechanics of the examination are carried out in the same fashion as any
sion of a crime. other thorough fire investigation, evidence to support any and all of these elements must
be constantly sought at a possible arson scene. The detection of abnormal fire behavior or
unusual fire conditions (as revealed by the diagnostic signs discussed earlier), incendiary
devices, fuels, or simply things that are out of place for some reason, should sound an
alarm in the examiner’s brain that warns of criminal activity. Because the circumstances of
658 Chapter 16 Arson as a Crime
arsons are so varied, they cannot all be treated here. We shall touch on characteristics of
fire behavior and means of ignition, as well as the motivation of fire setters as they relate
to the detection of arson. The motives, devices, intents, and targets for arson are all inter-
related to some degree. These relationships will be discussed as well, because the discovery
of one factor can be used to suggest or deduce others. Intent can be described as the pur-
pose or deliberateness of someone’s action (or omission). Evidence linking the suspect to
the arson can include proof of access and opportunity, physical evidence such as finger-
prints, video or documentary records, or even conduct consistent with his/her involvement.
The crime of arson is not a specific-intent crime. That means the prosecution need
prove only that the arsonist meant to cause damage or bodily harm by setting the fire.
This absence of specific intent may have an impact on a search for a motive. Arson-related
charges may include considerations of intent to cause the specific damage incurred, intent
to cause some damage, or even the foreseeability of damage from the act (setting a fire
under conditions that would likely lead to severe damage).

Arson Law
English common law, on which most U.S. law is based, defined arson as the willful and
malicious burning of the dwelling house of another. This made arson a crime against the
security of habitation (because loss of habitation could well cost the lives of its dwellers by
exposure to weather or enemies). It required certain elements to be present—the structure
had to be a dwelling, it had to belong to another, and it had to be burned as a deliberate
or intentional act. As the common law concept of arson became inadequate, statutory law
(laws passed by governmental bodies) expanded the definition to include other buildings
and property. Omitting the dwelling or occupancy requirement means that other property
such as shops, factories, prisons, public buildings, forests, fields, boats, and cars are now
included in most arson statutes. By the 1950s most states had adopted an equivalent to the
Model Arson Law first proposed in 1940 or had statutes incorporating significant portions
of it (see Appendix E for the text of the Model Arson Law). This law brought about some
measure of uniformity and recognized the seriousness of the offense in terms of several
degrees of arson crime. The first degree held to the common law concept that the burning
of a dwelling or associated building is the most serious offense. Lesser degrees depend on
the nature of the property, extent of involvement of the accused in a conspiracy, or the
nature of unsuccessful attempts to set fire to property. Because the statutes defining the
property susceptible to arson and the liability of persons involved vary from state to state,
the reader is urged to consult local legal sources.10
For many years, arson was considered in the United States to be a minor crime involv-
ing only property. In 1979, arson was elevated to Part I crime category status by the FBI,
and incidents are now reported under the Uniform Crime Reports. The Arson Control Act
of 1982 made this change permanent and focused public and governmental attention on
arson as a violent crime directed against people as much as property. This increased
awareness has made funding and investigative resources more available in recent years,
but they are still woefully inadequate.

Elements of Proof
The finding that a fire has been caused by incendiary (deliberate) means is the starting point
of a sometimes frustrating legal prosecution to try and punish the person(s) responsible. Not
all incendiary fires are arson, for the crime of arson carries with it certain requisites of proof.
The elements of proof that must be established to prove any crime are called the corpus
delicti, the body of the crime. For arson, there are three elements that must be proven:
(1) there has been burning of the property in question, (2) the burning was the result of an
intentional act, and (3) it was done with intent, that is, a malicious (willful) criminal act.
Chapter 16 Arson as a Crime 659
 “Burning” does not require that the building, vehicle, or timber be completely con-
sumed. It is adequate that there be some destruction of the actual fiber of the material, how-
ever small. Material that is merely discolored or scorched does not qualify in most states;
that the material has been charred or similarly changed is adequate. Proof of such an ele-
ment may be carried out by photographic exhibits or the testimony of witnesses. Failure to
provide some sort of documentation of this damage will result in dismissal of charges.
The remaining elements of proof (malicious intent) require that two conditions in fact
be proven: (1) that the fire was intentionally set and (2) that it was set with malice, that
is, with the purpose of destroying property for criminal gain. Legally, there is always a
presumption of innocence, even as to the cause of the fire. A court will assume that a fire
was caused by accidental means or by natural means, that is, an act of God, unless it can
be proven beyond a reasonable doubt that no accidental or natural cause was present.11
The discovery of an actual incendiary device may be of secondary importance because (in
many jurisdictions) arson may be provable in the absence of an ignition device (sometimes
referred to as negative corpus) as long as every logically conceivable accidental and nat-
ural cause has been considered and ruled out. Even when a device is found, the investiga-
tor should be careful to rule out other means of ignition as part of the investigation. This
proof will defuse any attempt by the opposition counsel to claim that no accidental or
natural source was considered, thereby introducing doubt as to the reliability of the
expert’s conclusion.
Proof of intent is sometimes the most difficult proof to obtain. Note that motive
itself is not required, but the discovery of a valid motive may often lead to discovery of
intent; for example, a failing business needs money to sustain its operations (motive), but
to burn the warehouse to allow an insurance claim would provide needed capital
(intent). Intent is most often proven by means of circumstantial evidence—a concept
deserving some discussion. The natural and probable consequences of a person’s acts are
presumed to be the intended consequences. The word willful as used in the statutes
means intentional or with knowledge and purpose.C1* When a person unintentionally
sets fire to a building in the course of committing another felony, the felonious intent is
also attributed to the arson.C2
Although intent is an element of the crime of arson, motive is not. However, motive
may be relevant to helping identify the accused as the responsible party but it is not evi-
dence of the crime itself.C3

DIRECT AND CIRCUMSTANTIAL EVIDENCE

Evidence in legal proceedings is any means used to help establish or disprove an alleged
fact. Direct evidence may be in the form of testimony from eyewitnesses who saw some-
one, for instance, actually touch a match to fuel at the scene. Black’s Law Dictionary (7th
ed.) defines direct evidence as that which conclusively establishes a fact without inference
or presumption, such as a description of events or an identification of a perpetrator by an
eyewitness to the crime. Because fires in general and arson in particular are rarely wit-
nessed firsthand, direct evidence is rare in fire-related cases. Circumstantial evidence is the
means by which alleged facts are proven by inference or deduction from other facts that
can be testified to firsthand. Such circumstantial evidence may be considered by fiction
writers always to be imperfect if not erroneous, but a carefully constructed case with
properly interpreted circumstantial evidence is every bit as valid as any case based on eye-
witnesses. In fact, circumstantial evidence such as a crime scene reconstruction employ-
ing a crime laboratory evaluation of the evidence is usually much more reliable than an
eyewitness interpretation of what was seen. Confessions are considered to be a form of

*In this chapter, citation numbers that appear preceded by a superscript C refer to legal references that are listed
at the end of the chapter under “Court Citations.”

660 Chapter 16 Arson as a Crime

direct evidence. Many people believe them to be infallible, self-sufficient proof of guilt,
but a confession that is unsubstantiated by independent corroborating evidence is not
proof by any means. One need only review the dozens of confessions recorded in each of
the heinous crimes of the past century to realize their lack of reliability.
Arson may be proved by direct or circumstantial evidence. However, it is not often
that there is an eyewitness to the crime. Most often the evidence is largely circumstantial.
This means that the crime and guilt of the accused must be inferred or deduced from other
facts that can be established. There are limitations to the conclusions that are based on
inferences. That is, the incriminating facts must be consistent with guilt of the accused
and inconsistent with any other rational conclusion.C4,C5 Usually, incendiary origin of
fire is established by circumstantial evidence showing that a foreign flammable liquid, its
container, or a contrived ignition source was present, or by showing the improbability of
an accidental fire—for example, by proving three separate origins.C6

Motive
Motive is some inner drive or impulse that causes a person to do something or act in a motive ■ Inner drive
certain way. Basically it is the cause, reason, or incentive that induces or prompts specific or impulse that causes
behavior. In a legal context, motive explains why the offender committed the illegal act. a person to do some-
thing or act in a certain
Though motive, unlike intent, is not an essential element in criminal prosecution, it often way.
lends support to it.12 The investigator must be careful not to overemphasize the impor-
tance of a possible motive revealed during the investigation. Having a “motive” does not
prove criminal action. This caution regarding the distinction between intent and motive
is emphasized in NFPA 921: Guide for Fire and Explosion Investions.13
The investigator must, however, give consideration to possible motives for every fire,
because they may provide clues predicting certain behavior patterns that can lead to iden-
tifying and intercepting the culprit, particularly in cases of multiple arson.
Multiple offenses may be categorized in three ways:
■ Serial arson several fires set at different locations with an emotional “cooling-
off” period between fires (the cooling-off period being from several hours to several
weeks)
■ Spree several fires set in different locations close to one another in time (i.e., with
no cooling-off time)
■ Mass arson multiple fires set in the same place at one time14
Arson classification by motive is an intuitive attempt to make logical sense of the
crime of arson and to accurately categorize firesetters. Combining information about the
offender characteristics and/or offense features had developed over the years as an
accepted approach. Inciardi in 1970 was one of the first to focus primarily on motive for
fire setting, and he classified adult arsonists into six categories: excitement, revenge, insur-
ance fraud, vandalism, crime concealment, and institutionalized firesetters.15
In 1987, Icove and Estepp conducted research on arson from a law enforcement point
of view when they considered the motives of adult and juvenile arson offenders. In their
study, over a 4-year period fire investigators filled out protocols after each arrest.
Consistent with previous work by Inciardi, their analysis of these cases again reported the
main classification categories as vandalism, excitement, revenge, crime concealment, and
fraud. They also included a category of “other” to address instances in which motive
could not readily be determined.16
The FBI’s National Center for the Analysis of Violent Crime (NCAVC) has identified
six major motive classifications that apply (either singly or in combination) to the vast
majority of arsonists. These can apply to either a single instance of fire setting or multi-
ple offenses. These classifications are also referenced in Chapter 22 of NFPA 921 (2008
and 2011). The FBI studies on arson motives are offender-based; that is, they look at the
Chapter 16 Arson as a Crime 661
 relationship between the behavioral and the crime scene characteristics of the offender as
it relates to motive. A motive-based method of analysis can be used to identify personal
traits and characteristics exhibited by an unknown offender.
The NCAVC has classified these motives, which are also cited in the 2008 and 2011
editions of NFPA 921, as the following:
■ Profit (NFPA 921, Sec. 22.4.9.3.6)
■ Vandalism (NFPA 921, Sec. 22.4.9.3.2)
■ Excitement (NFPA 921, Sec. 22.4.9.3.3)
■ Revenge (NFPA 921, Sec. 22.4.9.3.4)
■ Crime concealment (NFPA 921, Sec. 22.4.9.3.5)
■ Extremism (NFPA 921, Sec. 22.4.9.3.7)
These motives are discussed in turn, as well as juvenile fire setting, irrational fire
setting, and mixed motives.

PROFIT
Arsonists in this category set fires with the expectation of profit, which may be direct mon-
etary gain or indirect gain from a goal other than financial. The direct gains could be from
insurance fraud, eliminating or intimidating business competition, extortion, improving
property values by removal of unwanted structures (either on the property or nearby to
eliminate an eyesore or even to improve a scenic view), escaping financial obligations, or
even gaining employment. The profit motive may be extended to include murder by fire of
family members with the intent of gaining insurance proceeds or inheritances.
One common, well-recognized motive for insurance fraud is poor financial condition
at and before the time of the loss and/or financial distress. That is, however, by no means
the only reason people intentionally set fires. While the motive sometimes is “financial
need” driven, sometimes the motivation can more aptly be described as “greed” driven.
Indeed, there are many, many well-documented cases in which the person responsible for
an intentionally set fire was of very substantial means but was motivated by greed, or
other motives. For example, a prominent example is the case of Constantine
Pappadopoulos in California—in which Mr. Pappadopoulos, a multimillionaire business-
man, was caught on an FBI wiretap plotting the fire of his luxurious mansion home with
torch ■ A professional the torch he hired to burn it down.
fire setter. In many communities the single most common motive is profit or fraud. It is a major
factor in residential and commercial fires. Economic hardships fall on private property
owners and businesspeople alike, and they all may seek relief from their problems by
burning them. Fraud may be defined as a “deception deliberately practiced in order to
secure unfair or unlawful gain.” Under common law, burning one’s own property was not
considered a crime even if one gained by it. This aspect has been remediated by statutory
laws in most states that prohibit the destruction of one’s own property with the intent of
defrauding another party.
Fraud is principally of two types: commercial and residential.
Commercial Fraud
fraud ■ Deception Commercial fraud fires may be set or arranged by the owner to destroy old or outmoded
deliberately practiced stock or equipment, destroy records to avoid taxes or audits, clear an unsalable building
in order to secure
from marketable real estate, or gain insurance money to rebuild a better building at a bet-
unfair or unlawful gain.
ter location. Fires may also be set by a competitor to gain a market advantage or by agents
of organized crime for purposes of extortion, insurance fraud, protection rackets, or
intimidation. Labor problems are a frequent cause, particularly in the building trades.
During construction, buildings are especially susceptible to fires set by pro-union or anti-
union agitators. When property values escalate, a business may be sold repeatedly among
members of an arson “ring” to inflate its value, with huge insurance settlements on the
662 Chapter 16 Arson as a Crime
business when it finally burns. Some rings operate by buying small-to-midsize businesses
that may be marginal or even successful, draining the assets, selling off stock, replacing
furniture and fittings, and virtually stripping the business physically as well as financially.
The remains are then burned, and the insurance due on a “thriving business” is paid to
the members of the ring.
Signs of a commercial fraud fire can include the following:
■ Substitution of cheap, old furniture, equipment, or stock for new
■ Open drawers in file cabinets or desks to encourage destruction of records
■ Remodeling underway with minimal or poor-quality materials
■ Disabling of fire detection or sprinkler systems
The background investigation should determine recent changes in ownership, value, or
tenancy; the state of the business—debts, back orders, financing, and mortgages; the
nature of any inventory—contents new or old, seasonal, outmoded, obsolete, or unsalable
because of changes in governmental regulations; and the extent of competition. Contact
with the insurance agent and adjuster will reveal the amount of insurance, multiple poli-
cies, or recent changes in valuation or coverage. All these factors may reveal the motive
and intent of a would-be arsonist. Extensive documentation of business conditions may
permit a qualified forensic accountant to identify failing businesses, even to the point of
calculating the money being lost every month. This can be telling evidence to dispute a
claim for generous business interruption benefits.

Residential Fraud
Residential fires may be set by owners with the intent of defrauding the insurance carrier
or by tenants wishing to defraud the owner or a welfare agency. The burning of residen-
tial properties has been particularly popular since the implementation of Fair Access to
Insurance Requirements (FAIR) plans in many states. These government-backed plans
guarantee substantial insurance coverage to property owners, in many cases with minimal
investigation of the building’s value or the owner’s prior history. These plans have come
under considerable abuse by both individual landlords and organized rings who buy
housing units, insure them heavily, and then profit from convenient fires. Many states
have insurance statutes that preclude coverage for intentional acts.
Increasing tax rates, physical deterioration of structures (accompanied by legally
mandated repairs at great cost), pressure from condominium developers, and statutory
rent-control measures all combine to “force” multiple-dwelling landlords to seek relief by
destroying the now burdensome structure.
Welfare fraud plays a significant part in residential fires in many urban areas. When
large numbers of tenants were being displaced from their housing by arsons, it was
decided that welfare agencies should reimburse them for their unplanned costs of tempo-
rary housing and new furniture, appliances, and clothes. It was not long before such ben-
efits were being abused. If a move was already being contemplated, who could refuse a
free “grubstake” to buy new clothes or furniture? The usable belongings could be moved
out or sold off, and up to $3,000 would be provided when the apartment was gutted by
a convenient fire.
In addition to the substitutions described previously with respect to commercial fires,
there are some diagnostic signs of arson fires peculiar to dwellings. Items having senti-
mental value (photographs, antiques, family Bibles) or high monetary value (cash, stocks,
deeds, insurance policies, stamp or coin collections) are frequently removed before the fire
is set. Clothing or furnishings may be removed or substituted with cheap goods. The scene
should be inventoried by the investigator for his or her own protection as well as that of
the owner. The victim should be asked to provide a complete list of the contents of the
destroyed areas. Substitutions may be detected by checking for noncombustible remains
like hangers or buckles from clothes and metal frames or springs from furniture. As noted
Chapter 16 Arson as a Crime 663
 in Chapter 14, a laboratory may be able to identify the remains of suspected valuables
and confirm that they have not been replaced with junk. Removal of portable valuables
may not entail removal from the property, and a check of the outbuildings may reveal
their location. The investigator should determine whether family pets (dogs, cats, birds,
horses) survived and if so, whether they were moved or released prior to the fire or
escaped during it. Convenient “vacations” may take the responsibles themselves out of
the area. In the Pappadopoulos case mentioned previously, the wife was convicted of con-
spiring with her husband to have their 929 m2 (10,000 ft2) home burned while they were
vacationing overseas. Mr. Pappadopoulos remains a fugitive in Greece.
Insurance policies should be thoroughly checked with the carrier agent and adjuster
for any overinsurance, recent increases in coverage, multiple coverages, or the nature of
any particularly valuable items inventoried separately for coverage. The investigation of
the victim should include the status of finances, job changes (move or termination), sta-
bility of marriage, illnesses, injuries, or lawsuits pending. The history of the victim should
be checked for previous fires on or near the property [see Chapter 17 for information on
NICB (National Insurance Crime Bureau) resources]. Once a formula is found successful,
such fire setters often cannot resist the greedy temptation to cash in again.
Other Profit Motives
In addition to direct insurance or welfare/public assistance fraud, other possibilities for
financial gain need to be checked and eliminated. Fires have been set to secure employ-
ment in a variety of situations. Firefighters, either seasonal, part-time, or paid volunteers
(or full-time firefighters who are paid overtime for off-duty responses), have been found
responsible for setting fires to increase their income. Unemployed security guards have
been known to set fires to prove their value to the threatened business, with the intent of
promoting employment. Wildland fires have been set to improve game-hunting conditions
(within a historical context). Construction workers have burned partially completed
buildings to ensure continued employment, and heavy-equipment agents have set wild-
land fires to provide a market for the rental of their equipment. Fires have been set to gain
freedom from underpaid contractual and even military obligations.17 Tragically, there
have been cases in which family members (particularly young children) have been killed
by set fires, sometimes to gain directly from insurance but also to gain freedom from fam-
ily obligations or to improve the conditions for another relationship.18

VANDALISM
Malicious or mischievous fire setting that damages property, seemingly at random and
with no identifiable purpose, is often considered vandalism. It is usually a motivation
related to juveniles and adolescents, and targets are often schools or school properties.
Abandoned buildings, vehicles, and wildland (vegetation) are also frequently targets. Fires
are set when the opportunity arises, often after school or work or on weekends.
Boredom and frustration among youths sometimes lead to a peer-group challenge to
create some excitement. It may be either a group effort of the gang or a solo trial as an
initiation rite for a new member. This is a group Dr. Dian Williams calls “delinquent” fire
setters, since they often have records for other forms of lawbreaking.19 According to Dr.
Williams, such fires may be set as part of (or to conceal) common vandalism (breaking
windows, graffiti) or burglary, or as a demonstration of “gang” behavior. This gang
behavior may be part of an initiation or to impress others in the gang to gain entry or sta-
tus. Common vandalism, petty theft, or burglary may accompany such fire setting.
Devices, rarely used, will be very simple, with paper trailers or Molotov cocktails being
the most complex. Fuels or accelerants may be brought in, but fuels incidental to the scene
are usually employed, since the attacks are usually not planned. These fires may be set as
a convenient weapon against someone (e.g., revenge motive). Such fire setters can be dan-
gerous because they generally lack any regard for the safety of others. Apprehension is
664 Chapter 16 Arson as a Crime
difficult because the fire setter’s behavior is unpredictable and may follow no set pattern.
Vandalism by setting fires is a very different motive from “fire play” fire setting by young-
sters, which will be considered in a later section.

JUVENILE FIRE SETTING

Fire setting by juveniles can extend far beyond vandalism, since it can occur with children
too young to form intent or understand the consequences of losing control of a fire or
flame (often as a result of the curiosity phase of fire experimentation, typically with boys
3 to 10 years old). Juveniles may also set fires as a solitary activity to express anger or
seek attention. In such cases, the fires generally will be small—in closets, beds, waste cans,
or garages—but often get out of control and destroy much property and even cost lives.
Such fires may also be set as a cry for help because of neglect or an abusive (physical, sex-
ual, or psychological) family situation or in deliberate response to a threatened departure
(as in divorce) or unwanted arrival (new step-parent, child, or lover). Such fire setting is
often foreshadowed by emotional upheavals in the family (death, divorce, unemployment,
alcoholism, even moving), an absent or inattentive parent, or even school or social prob-
lems. In one recent case, numerous small household items were set on fire in kitchen
drawers and the refrigerator by a troubled 13-year-old girl. Fortunately, the fires self-
extinguished and were discovered only later by family members or investigators.
Juvenile fire setting has been seriously underestimated in years past, most probably
because many incidents were not reported by families or social welfare workers. It is now
clear that the arsons cleared by arrests each year in the United States involve a large per-
centage of suspects under the age of 18. The investigator should be careful not to assume
that young juveniles are not capable of forming the intent for arson. It is estimated that
nearly 7 percent of those arrested for arson in the United States in 1994 were under the
age of 10.20
Traditionally, about half of all arson arrests each year in the United States are juve-
niles (probably because they are the easier cases in which a suspect can be identified). FBI
Uniform Crime Reporting (UCR) statistics show that for the most recent 10-year period,
the percentage of persons arrested for arson who were under the age of 18 dropped
13 percent, from 55 percent in 1999 to 48 percent in 2008.21 Of the 10,784 suspects
arrested in 2008, 36.8 percent were under 18 years, 18.7 percent were under 15 years,
and 1 percent were under 10 years of age. In the past, arson was considered a crime tra-
ditionally committed by males; however, FBI UCR statistics show that of the 10,784 total
suspects arrested in 2008, 84.1 percent were male and 15.9 percent were female.22
Bob Wood of the United Kingdom has explored some of the complex issues that dis-
tinguish between juvenile “fire play” and deliberate fire setting.23 True curiosity or “fire
play” fire setting is considered by many to be a natural part of learning about one’s envi-
ronment. Such fire setters are boys between ages 3 and 10, there is no expressed intent to
do damage, and there is rarely repeat behavior (beyond two fires). Such fire setting
appears to be opportunistic and spontaneous when matches or a lighter are available and
there is no adult supervision.24 It is clear that both caregivers and fire investigators have
to learn to recognize the critical patterns and features so that the problem can be
addressed. Through local counseling programs (through police, and public health and
community aid agencies) great progress has been made in addressing this national prob-
lem.25 Investigators must appreciate the delicate legal ground involved in interviewing
juveniles as suspects or witnesses and should remember that the parents will lie to protect
their offspring, even more than the subject will.

EXCITEMENT AND THRILL SEEKING

Fires may be set by those seeking excitement, attention, recognition, and, very rarely, sex-
ual stimulation. Excitement or thrill seeking as a motive for fire setting is closely related
Chapter 16 Arson as a Crime 665
 to the human psychological need for attention, with the difference largely one of intent—
that is, did the fire setter realize that he or she was doing it out of a specific search for
excitement, just to gain the attention of someone, or out of anger or desire for revenge?
Such anger/revenge motives will be discussed later. Fires have been set as attention-getting
devices by persons to call attention to themselves or a crisis situation.
Adults and children as young as 10 have been identified as setting fires to make oth-
ers aware of physical, emotional, or sexual abuse conditions. These fires may be very
small, set in readily ignitable household materials—books, towels, drawer contents,
even clothing. Fires set for excitement can range from small trash and dumpster fires to
fires involving occupied buildings. The typical excitement-seeking fire setter will set
many fires during his or her career, often with the severity (amount of damage and risk
to life) increasing with time.26 The fires are usually set in a neighborhood familiar to
him or her (from school, work, recreation, etc.). The arsonist will sometimes stay in the
area to watch, possibly even photographing or videotaping the action. Devices are very
rarely used; matches or lighters and available combustibles are typical. Although the
excitement usually relates to the fire itself, some may be more focused on the fire appa-
ratus, sirens, and lights of the suppression efforts. In these cases, the arsonist begins to
set fires just to participate, actively or vicariously, in the suppression. (This has
occurred with firefighters seeking the “adrenaline rush” of emergency responses.) This
brings us to consider a subclass of the excitement-seeking fire setter—the hero/vanity
arsonist.
The excitement inherent in a large fire and its suppression efforts offers many pos-
sible motives for would-be arsonists. People seeking attention or respect may set fires
to gain that attention, either by being in attendance (part of the crowd), taking part in
rescue and suppression efforts, or even by taking responsibility for the fire. Such fire
setters are often described as “hero/vanity” arsonists. Firefighters may wish to impress
their families or their superiors with their performance and set fires to give themselves
more opportunity to do so. Those who get extra pay for responding to calls (as volun-
teers or as bonus-paid, full-time firefighters) have been known to set nuisance, structure,
or wildland fires for the income.27 Security guards may set fires simply to relieve bore-
dom, force the company to hire additional guards, or prove their value to the firm by
discovering the fire and assisting in its extinguishment. Their increased value may
improve their chances of a pay raise or forestall a threatened layoff. According to the
2003 USFA report, firefighters who set fires often do so for the chance to be a hero.
This heroic role can be in the form of first detecting or reporting fire, rescuing occu-
pants, or acting prominently in its suppression. The ultimate such case could well be
that of John Orr, a captain with the Glendale Fire Department. Over a period of years,
he is suspected of having set scores of fires (both wildland and structures) where he
could intercede as the chief investigator to “solve” many difficult fires.28 He is thought
to have been primarily a vanity fire setter who enjoyed a considerable thrill from work-
ing alongside the people looking for the arsonist. His motivation would be difficult to
classify.29 He was convicted in 1992 and 1998 of numerous arsons in commercial pre-
mises and wildlands.
The hero/vanity fire setter often has an overriding self-image as a superhero or “big
shot,” which is mostly bravado. These people often carry a weapon because it is an exten-
sion of their power. Attacks are generally unplanned, using materials and circumstances
as they are found. He sets fires to provide opportunities to show his courage or dedica-
tion or to gain public attention or acclaim.
The investigation of this type of fire setter should include a background check of cur-
rent personnel, especially those involved with discovering or extinguishing the fire.
Because attention seeking may be related to frustrated ambitions to be a police officer or
firefighter, recently rejected applicants for these positions should be considered potential
suspects.
666 Chapter 16 Arson as a Crime
REVENGE, RETALIATION, OR SPITE
For adult fire setters, revenge or spite is the most common motive. Fires may be set in ret-
ribution for some real or imagined insult or hostile behavior. Because of the complex psy-
chological factors, revenge may be part of other motive categories. Targets of revenge fires
may be a person, an institution, the values or symbols of society in general, or a group.

Persons
An attack on a person is often in retaliation for a personal grievance. Fires stemming from
hatred or spite between individuals can be a spontaneously conceived, one-time occurrence
or multiple events with deliberately and carefully preplanned fire setting. The target may
be the person’s personal possessions, vehicle, or home, and the fires are often accompanied
by vandalism or physical destruction prior to the fire. Such revenge-motivated fires can
often turn deadly when the “target” or innocent persons are trapped by the fire. The loca-
tion of the fire is often a significant indicator, because burning the bedroom or the clothes
closet is considered to be worse punishment of the victim than just burning the living room
or den. Bedding and clothes may be cut or torn up, or paint, ink, or glue may be spread
on them and the furnishings before the fire is set. Photographs or memorabilia are some-
times smashed. The attack may be directed at the target of the hate or a substitute. Because
the attack is not rationally prepared, it is typically disorganized and unsophisticated. The
arsonist is usually reacting to an extreme emotional state. Fuels or accelerants are usually
whatever is conveniently at hand at the scene. Time-delay devices are rare, because the per-
petrator is not rationally considering escape and an alibi at the time.
One case in the authors’ experience defied both of the last two such guidelines, how-
ever. An estranged wife used gasoline to soak clothing and furnishings of the husband’s
vacation home and then attempted to set a candle-and-paper delay device. Unfortunately,
she had been alternating her destruction with heavy drinking, and the resulting delay
allowed an explosive mixture of gasoline vapor to permeate the house. When she
attempted to light the candle, the house exploded in flames, seriously burning her as she
fled. In such premeditated cases, fuels, trailers, and delay devices may be found.
All types of interpersonal relationships can result in spite fires. Check marital rela-
tionships, friendships, or job-related situations (in revenge for loss of a job or promotion).
Such spite fires may be encountered in either the gay or straight community. The tensions
released when a heterosexual or homosexual love relationship fails can result in disas-
trous and tragic consequences. Such fires may be set against either the new lover or
the unfaithful former partner. The psychological complexities of gay relationships and
the strained or even hostile relationships between the gay and the straight communities
make an investigation of such a fire even more challenging.

Institutions
A person who feels he or she has been wronged by an institution such as government,
medicine, courts, military services, or even by employers, banks, or other representatives
of authority may attack (usually in serial attacks) the buildings housing these agencies.
The executives of those agencies may also be selected as targets. The fires can range in
severity from nuisance fires in trash containers to planned fires of great destructiveness,
although they tend to be mass arsons (multiple sets in the same building). The same tar-
get may be attacked repeatedly over a long period.30 Because the grievance against the
institution may be real or imagined, suspects would not necessarily be limited to former
employees or disgruntled customers. The Oklahoma City Murrah Federal Building bomb-
ing is an example of an attack on an institution. To Timothy McVeigh, the Federal Court
House represented the government that, among other things, had attacked the religious
compound at Waco, Texas, two years previously. Ted Kaczynski, the Unabomber, had a
similar motive, but his targets covered the spectrum of imaginary grievances—against air-
lines, computer technology, psychologists, and forestry.
Chapter 16 Arson as a Crime 667
 Society
Retaliation against society is the motive of someone who, feeling abused, persecuted, and
powerless, strikes out against that society in a series of often-random attacks. Such fires
are set in a variety of targets, typically using available materials and matches or a lighter
for ignition, although they are often premeditated. Interestingly, the societal retaliation
fire setter rarely contacts the media to claim responsibility, apparently satisfying him- or
herself that something has been done to even the score.31 Interestingly, Dr. Dian Williams
links many terrorism acts with fire or explosions to revenge motivation somewhere in the
perpetrator’s background.32
Groups
These fires are often called hate crimes because they are the result of hate or anger that
the subject has against a group rather than a person. The targets for these fires may be
religious, racial, or fraternal and may be a building or other symbol of the group rather
than an individual. Churches or synagogues are common targets, but the meeting places
of clubs, organizations, and (recently) rival urban gangs are also struck. In some cases, an
emblem (cross, star, logo) may be attacked, regardless of its location.33 Graffiti and other
vandalism may accompany the fire. Premeditation and planning are encountered, and
devices (particularly Molotov cocktails) are often used.
Church arsons across the southeastern United States were the impetus for a national
task force. One series of arson attacks was carried out by a lone transient with a grudge
against religion. Others have been racially motivated or have targeted Jewish houses of
worship. The National Church Arson Task Force (NCATF)—which included the FBI, the
Bureau of Alcohol, Tobacco and Firearms personnel, U.S. Attorneys, the U.S. Marshals
Service, and state and local fire and police personnel—was formed in 1996 in response to
a rash of fires at houses of worship, mostly in the South. The task force made significant
progress in the fight against such arsons. In a report issued in September 2000, the task
force said it opened 945 investigations into arson and attempted arson at houses of wor-
ship between January 1995 and August 2000 and arrested 431 suspects in 342 of the
cases. This translates into a 36.2 percent arrest rate, which is more than double the
16 percent rate of all arsons as reported by the FBI.34 As reflected by the convictions from
the NCATF efforts, the arsons—at both African American and other houses of worship—
were motivated by multiple factors, including racism and religious hatred. Arsonists have
burned churches for other reasons, including opportunistic and random vandalism, pyro-
mania, mental health disturbances, feuding with ministers, retribution against religious
authorities, parking or neighborhood disputes, cover-up of burglaries, and financial
profit. In some cases, the arsonists claimed they believed the church to be an abandoned
building.

CONCEALMENT OF ANOTHER CRIME

Because of the destruction brought on by a major structure fire, many people believe that
any fire destroys all physical evidence of any criminal activity. Although it is true that
even a moderate fire will obscure and render partially useless some types of evidence,
blood and physiological fluids will retain much of their evidential value, fingerprints can
survive under a variety of conditions, and weapons like knives, guns, or bullets require a
great deal of fire exposure before they are rendered completely worthless. See Chapter 14
for further information on physical evidence.
While fire can obliterate fingerprints and facial features on which normal identifica-
tions are based, it requires significant combustion over a period of hours before the body
can no longer tell its tale of cause and manner of death. (The reader is referred to Chapter 15
for more information on the evidence related to fire deaths.) Nevertheless, fires may be
set to impair identification of a victim, destroy evidence of foul play, or simulate an acci-
dental death in the hope that authorities will not examine it too closely.
668 Chapter 16 Arson as a Crime
 It must be remembered that fire is one of the human’s oldest weapons against fellow
humans and, as such, may be used to kill or assault another. In a number of recent cases a
flammable liquid was poured or thrown on someone and then ignited intentionally. Such
fires have been used as a means of murder of spouses and other family members, recalci-
trant lovers, and even of robbery. Such cases usually involve premeditation (to procure the
flammable liquid), but fires may occur in the midst of a heated verbal argument or some
time afterward. Sadly, some of these incidents are written off as household accidents.
Although fires that occur during or after the commission of another crime are usually
thought to be intended to destroy evidence of the crime, it has been suggested that
they may be part of an anger-retaliation crime. In murder cases, the subject may want not
only to kill but also to obliterate or destroy the victim and will set fire to the body (or
clothing or piled furniture) to achieve that end. Because such behavior may increase the
thrill or satisfaction of such crimes, when they are part of a series they may include care-
ful planning; complex, even ritualized preparation; and specialized tools. The cunning
and organization extend to security arrangements (remote or private locations).
According to psychologist Richard Walter, such fire setting tends to be episodic (serial)
but follows no set time pattern. It is so much an instrument of aggression that the fire set-
ter will often keep a souvenir from each scene as a trophy of the triumph.35
The use of fire as crime concealment is not limited to crimes of personal assault. Fire
may be set to obliterate signs of theft, pilferage, burglary, or even embezzlement and
fraud. In such cases the manner of preparation and ignition are often afterthoughts to the
main criminal activity. Sophisticated devices are rare and accelerants used, if any, are usu-
ally those incidental to the scene. Lack of complete ignition, absence of proper fuel or ven-
tilation, or a general appearance of things “out of place” are diagnostic signs associated
with this kind of fire. When fuels are moved about to ensure destruction, they are usually
the ones involved in the theft—racks of clothing, boxes of stock, or drawers of financial
records are positioned with the intent of promoting the highest destruction in their imme-
diate vicinity. Administrative or correctional records may be targeted in educational,
military, or court office to prevent detection or punishment.

EXTREMISM (SOCIAL PROTEST AND TERRORISM)

Fire has been used as a weapon of social, political, or religious protest since revolutions
first began. If one could not attack an oppressor directly, then one could destroy the
oppressor’s castle, counting house, or crops. Fire is considered by some political minori-
ties as a legitimate expression of dissent. During the 1960s and 1970s, wholesale arson,
often during general riot conditions, was employed against the white “establishment” to
express the hostility of racial minorities and to force urban renewal into the then-
destroyed areas.36
These fires may be started by individuals or by groups and are usually accompanied
by ransacking, vandalism, and looting. Even though a fire may have occurred during a
period of unrest, it should not be a foregone conclusion that it was started by rioters.
After the major urban riots of the 1960s and 1990s, it was discovered that a number of
store owners burned their own property for fraudulent purposes, using the riots as cover.
(A similar pattern develops whenever there is a serial arsonist operating in an area, with
property owners hoping that every fire will be written off as the work of the fire setter.)
Devices in riot or unrest cases are usually limited to what is at hand or the ubiquitous
Molotov cocktail. Urban street gangs have recently taken up Molotov cocktails as a
weapon of intimidation, thus accounting in part for a substantial increase in the number
of such weapons reported by police agencies.
Many of the thousands of political movements in the world have spawned waves of
fire and explosive incidents as statements of political protest or outright terrorism. The
differences between protest and terrorism may be only in the degree of organized planning.
Chapter 16 Arson as a Crime 669
 Although an organized terrorist campaign often involves random targets of innocent per-
sons to maintain an aura of anarchy, terrorists, like protesters, often select targets of polit-
ical or economic significance. Political targets include government offices, newspapers,
universities, or political party headquarters. Economic targets were very popular for ter-
rorists in the 1970s and included the offices of big business (auto, oil, computers) and the
offices or distribution facilities of utilities, banks, or companies that were thought to have
adverse effects on the environment. In the 1980s, political terrorism by fire and explosion
in the United States drew such diverse targets as animal welfare (protesting use of animals
in lab research), neighborhood protection (stores or industries not deemed suitable for the
area), and abortion clinics (which were hit by an extensive series of explosive and incen-
diary devices in many states). Today, so-called eco-terrorists are targeting lumber compa-
nies, automobile dealers, commercial livestock facilities (poultry and fur), property
developers, ski lodges, and even university facilities engaged in agricultural, animal,
or stem cell (reproductive) research with incendiary device attacks. Terrorists, because
of their planning and financing, make more frequent use of sophisticated timing and
ignition devices: tamper-proof electronic timers, chemical incendiary mixtures, radio-
controlled detonators, as well as large-scale explosive charges. The World Trade Center
(New York City, 1993) and the Murrah Federal Building in Oklahoma City were both
targets of terrorist vehicle bombs, each estimated to involve over 500 kg (1,100 lb) of
blasting agent. The September 2001 attack on the World Trade Center accomplished com-
plete destruction by causing structural collapse by means of massive fires from fuel-laden
airliners.37 Retired ATF Special Agent John Malooly outlined many of the targets in the
United States that terrorists could effectively target using simple fire sets. Targets could
include major freeways and bridges, commercial or financial centers, places of public
assembly, schools, and public transport.38
The “Provisional” wing of the Irish Republican Army must be considered an excel-
lent example of a well-financed, experienced terrorist force. It selected a mixture of mili-
tary and civilian police emplacements and personnel for its main targets (using both
explosive and incendiary devices) and added an assortment of random civilian targets to
maximize the confusion and fear of the general populace. One device used with devastat-
ing effectiveness on occupied civilian targets was a quantity of gasoline dispersed and
ignited by a small explosive charge. One or two 20-L (5-gal) jerry cans were placed in
close contact with a charge of 0.4 kg (1 lb) of military or commercial high explosive. The
device, placed in a window or doorway, produced an instantaneous fireball that filled the
room, killing nearly all occupants. The resulting flash fire sometimes caused only super-
ficial damage to the building, because the gasoline was completely consumed while in the
gaseous phase. Afterward, the room bore no smell of gasoline, and the obvious diagnos-
tic signs pointed toward a gas explosion rather than an incendiary device.

MIXED MOTIVES
Often, a set of mixed motives may be found in an arsonist. For example, an element of
revenge may be found in many vandalism fires. Many arsonists who set fires for excitement
may also profit from them. It may also be argued that many extremists have an underlying
revenge motivation in setting their fires. The point is that in dealing with human behavior
it is necessary to examine each case to determine motive, mixed or otherwise.

IRRATIONAL FIRE SETTING

As we have seen, the vast majority of incendiary fires fits into one of the rational motive
categories described here, where the fire setter “sees” some specific benefit is accruing to
pyromania ■ Uncon-
him or her as a result of setting the fire (even if an outsider cannot perceive any benefit).
trollable psychological Irrational or “motiveless” fire setting is sometimes called pyromania. Recent evaluation of
impulse to start fires. this term has shown that there is no consensus even among mental health professionals and
670 Chapter 16 Arson as a Crime
behavioral scientists as to what constitutes pyromania, and, indeed, whether there really
is such a disorder.39 Some analysts believe that what investigators consider irrational
motives for setting fires are perfectly rational to the perpetrator suffering from some men-
tal disorder. Moreover, the motives in the mind of the fire setter fit into the major cate-
gories described previously. Subjects may claim that the fires are set merely for the release
of tension or anxiety, and they simply “feel better” once the flames are started. The cur-
rent version of the American Psychiatric Association’s Diagnostic and Statistical Manual
of Mental Disorders-IV (DSM-IV-TR) describes “pyromania” as an impulse-control dis-
order characterized by “a fascination with, interest in, curiosity about, or attraction to
fire and its situational context.”40 The key element is lack of impulse control. If the fire
setting were truly uncontrollable, the subject would be setting fires in front of witnesses,
even police. If the impulse can be controlled such that the fire starting is away from
known observers, it is not “pyromania.” Extensive research at the NCAVC challenges
the concept of the motiveless fire setter. Analysts there are persuaded from extensive
interviews with arsonists that they all have a motive, but some lack the introspection or
communication skills to adequately describe their motivation.41
Although the preceding is the condition to which most people refer when they use the
term firebug, it is clear that only a small percentage of all arsonists fit the description of
motiveless fire setters. Even David Berkowitz, New York’s “Son of Sam” killer, who
claimed credit for setting some 2,000 fires in New York over a 3-year period and who is
considered by many to have been a pyromaniac, may have better fit the description of a
societal retaliation arsonist because he felt inadequate and lonely, and struck out with fire
(and ultimately with a gun) in response to those feelings.42
Such behavior usually starts with the setting of small trash fires (or grass fires in rural
areas), often escalating in target selection to fires in empty buildings, abandoned vehicles,
and, in extreme cases, to occupied apartment houses, sometimes with devastating and
fatal effects. Serial dumpster fires may be the first signs of the presence of a long-term
problem fire setter, who may be a simple excitement-motivated fire setter. While the fires
set by such subjects are often planned, they rarely, if ever, employ accelerants, trailers, or
time-delay devices. They are most often started in piles of paper or trash in out-of-the-
way corners, hallways, or alleys, which may delay their detection. The fires are almost
always set under cover of darkness, and the setter is nearly always a low-profile loner.
Although it was once a criterion, IQ cannot be used to classify such offenders, since sur-
veys have shown no reliable relationship. Most fire setters are in the normal range of IQ.
Fire setting has been the subject of research in the past; however, it is clear that although
concepts of fire setting and pyromania have evolved with new knowledge in recent years,
research interest is lacking.43 (Lewis and Yarnell’s “classic” study, for instance, has been
found to be based on very limited data of questionable reliability.) Dr. Dian Williams has
published an excellent study of arsonists—both adult and juvenile—and their “manage-
ment” in investigations.44 Because diagnosing someone as a pyromaniac is a clinical diag-
nosis to be made only by a mental health professional, fire investigators should be very
careful not to label the subject of their investigations as such.

The Arson Set

The arson set may be defined as the contrivance device or deliberate assembly of first fuel arson set ■ Device,
and heat source intended to ignite the fire. The professional and, to a lesser extent, the assembly, or con-
trivance used to ignite
amateur arsonist want to produce a maximum of destruction no matter how small the first
an incendiary fire.
ignition. To this end, most rational arsonists will be sure the fundamental requirements for
a large, spreading fire are present: ready availability of fuel, proper ventilation, and reli-
able ignition. Although much attention has been given to the ignition devices used, the
most important considerations to the success of the fire are sufficient fuel and access of air.
Only when delayed ignition is essential is there any need for unusual methods of igniting
Chapter 16 Arson as a Crime 671
 the fire. Many more arsonists choose the time and circumstances of the fire to minimize
their chances of being seen than rely on any unusual (and often undependable) igniting
device. Perhaps the preoccupation of some investigators with such devices results prima-
rily from their novelty, which adds much interest to their discovery. In one study by
DeHaan, only 12.3 percent of the arson cases examined involved the use of any type of
ignition (timing) device. By far the most common “set” was the direct pouring of an
ignitable liquid with ignition by open flame (match or lighter), which was found in 61.2
percent of the cases.45 This trend was found in cases examined in the author’s laboratory
during the 1980s and continues today. The extent and nature of preparations, such as
building or assembling a device, preparing the scene (providing kindling of a particular
type or arranged in a certain fashion), providing accelerants, ensuring fire spread, and the
like, are all indications of the organization characteristics of the fire setter. Such character-
istics (organized versus disorganized) can be used to focus on a possible motive and even
suggest the characteristics and traits of the arsonist (profile) for use in the investigation.46

ARRANGING THE FIRE—LOCATION

Once the target is selected, the arsonist must make several choices that are critical to the
success of the fire. If destruction is a specific goal, a point of origin must be selected so
that flames will rise into fresh fuel. Although flames will spread fastest across the under-
side of a ceiling or overhang, it is usually impractical to arrange a point of ignition high
up on a wall. The floor adjacent to a combustible wall is the likeliest point for initiating
the fire. Walls covered with gypsum sheetrock do not offer much fuel, and once the paint
and surface paper burn off, the fire will die back rapidly. Unless holes have been broken
or cut through the sheetrock to expose the combustible wall studs or ceiling joists before
the fire, such covering will seriously inhibit the spread of the fire. Wooden paneling prod-
ucts vary considerably in their fire resistance but eventually they will all add greatly to the
heat release rate (fuel load) of a room. Furniture may be used if present in the correct type
and sufficient quantity. Older furniture stuffed with horsehair and covered with wool fab-
rics is difficult to ignite and does not burn readily once ignited. Nearly all modern furni-
ture contains large quantities of polyurethane foam padding that can be readily ignited
with a match and burns with great speed and intensity. Lighter upholstery fabrics of cot-
ton and synthetics can be readily ignitable, adding further to the ease of ignition and the
fuel load. In a commercial or industrial building, stocks of finished goods, packing mate-
rials, or raw materials may be moved about to provide a large fuel load in a readily com-
bustible form. High rack storage provides ideal fuel distribution and concealment for the
fire setter. Unless in-rack sprinklers are fitted, a fire can become very large before detec-
tion. It must be remembered that knowledge of fire behavior, detection, and suppression
is not always familiar to the fire setter and so he or she may pick a location unaware that
the fire will not grow or will be readily detected or extinguished by a sprinkler head
immediately above it.
Within the structure, the point selected for the origin may be limited by what areas
of the building are readily accessible, and the “ideal” spot may be protected by a locked
door, gate, or elevator. The arsonist may then have to select a less perfect location.
Basements are very suitable because the walls and ceilings are often not covered, the entire
structure above the fire providing maximum fuel. A basement location is remote, provid-
ing the best concealment, and the utilities and trash often found in basements provide
many possibilities for an “accidental” fire. Any basement fire should be investigated with
special care, because of all possible locations, this one is generally most suited to the
arsonist’s purposes.
Attics, although they provide convenient fuel in exposed structural members, trash
accumulations, and points of possible “accidental” ignition from exposed wiring, are gen-
erally not selected as a point of origin. The availability of fuel is limited to the contents
672 Chapter 16 Arson as a Crime
and the roof and adjacent structure. Insurance fraud fires are sometimes started in the
attic, however, for the very reason that the ceiling and roof are destroyed, the building is
rendered unusable, and business is interrupted. When the insurance for these claims is col-
lected, repairs are begun, and the building, while being rebuilt, suffers another more com-
plete fire; the insurance on an upgraded building is then collected. One interesting variant
seen recently is to set fire to adjoining businesses knowing that the fire setter’s property is
likely to incur some roof or smoke damage from a large fire. The insurance proceeds are
then available while the investigation focuses on the adjacent property. Common or adja-
cent roofs also provide a convenient route of concealed entry and exit. A history of pre-
vious fires in adjacent properties may offer a link to the perpetrator.
Closets provide both concealment and some fuel in the form of stored clothing. Other
than nuisance, spite, or juvenile-set (fire play) fires, arson fires in closets are rare. A
knowledgeable arsonist realizes that the poor flammability of some clothing, the limited
air, and the difficulty of encouraging fire spread to other rooms make a closet a poor
choice for an area of origin. Clothes closets and beds, when confirmed as points of ori-
gin, may strongly indicate a revenge/spite motive rather than fraud. However, when
smoke damage is the desired end (for insurance fraud—to gain remodeling expenses,
stock replacement, etc.) rather than a lot of fire, a closet may be the best place to ignite
the fire because ventilation and spread can usually be controlled.
In an attempt to simulate accidental ignition, some arsonists ignite the fire in the
immediate vicinity of a fireplace, heater, dryer, water heater, or similar heat-producing
appliance. The arsonist hopes that a cursory inspection of the scene will place the point
of origin in the vicinity of such a “normal” cause, and the fire will be written off as an
accidental malfunction of the appliance. All such fires should be examined carefully, and
the investigator must not prejudge such fires and fall into the arsonist’s trap. With very
few exceptions, accidental fires associated with such appliances will start above floor level
and travel upward. A fire set with a flammable liquid will usually burn from the floor
upward. If the diagnostic signs of fire behavior indicate a very low burn, the entire area
should be examined carefully, as described in Chapter 7. Fire starting accidentally in con-
cealed floor spaces or post-flashover room fires, of course, can produce floor-level burns
and penetrations.

FUELS
Arson investigators find themselves investigating a fire set by a truly professional “torch”
only on rare occasions, probably because of two guidelines used by such professionals:
(1) the simpler the better, and (2) use fuels available at the scene. Both guidelines mini-
mize the materials that must be brought to the scene by the arsonist, and the fewer the
number of times the individual is seen in the area (especially carrying containers or odd
packages) the fewer the chances of being spotted by a witness. Simpler sets minimize the
chances that elaborate devices will fail to ignite while they maximize the chances that the
entire set will be destroyed by the fire or even normal suppression activities.
The use of accumulated trash as an ignition point obeys both guidelines. Trash piles
provide an inherent risk of ignition by accidental causes, and many of them are written
off as accidental by investigators without careful examination. In the experience of the
authors, a small quantity of ignitable liquid on a pile of loose paper and plastic cartons
will ensure complete ignition and yet will be completely undetectable in the post-fire
debris. Trash accumulations are frequent in basements, attics, storage rooms, and around
back doors (inside and out). A trash-filled utility room is ideal for producing a large
destructive fire. These can be difficult fires to establish as arson-caused because there is
nearly always some possible means of accidental ignition, such as defective electrical
wiring, furnace, water heater, or appliance; and loose paper or empty cardboard or plastic
foam boxes make excellent first fuels. Accidental fires involving these must be eliminated
Chapter 16 Arson as a Crime 673
FIGURE 16-1A A gasoline trailer ignited only the bed (left front FIGURE 16-1B This gasoline trailer ignited only the adjacent
corner) leaving minimal fire scorch marks, and a clear trailer carpet. Courtesy of Lamont “Monty” McGill, Fire/Explosion Investigator.
pattern of carpet protected by the gasoline. Courtesy of Lamont
“Monty” McGill, Fire/Explosion Investigator.

as causes to prove the crime of arson if the scientific method is being followed. Outside
trash piles are often used by the arsonist who does not have access to the inside of the struc-
ture. Because such fire setters usually do not intentionally provide a path for the fire to
spread to the structure, the fires do not always spread to the interior of the nearby building.
Ignitable Liquids
Many arsonists, especially amateur ones, do not rely on solid fuels available at the scene
and will heavily supplement them with ignitable liquid accelerants to ensure complete
ignition. As defined previously, an accelerant is as any fuel deliberately added to increase
the intensity or speed of spread of an intentional fire. This is often an ignitable liquid.
Although such fuels have the ability to produce a great deal of fire in any given location,
there is no automatic assurance that the flames so rapidly produced will find other fuel
in their path so that they can build into a conflagration (as shown in Figure 16-1). The
larger the quantity of ignitable liquid used, the larger the initial fire may be, but if a
volatile fuel is involved, the vapor/air mixture in an enclosed room is more likely to be
above the upper explosive limit (meaning no ignition). A very large flammable liquid fire
may also deplete the oxygen in the room or structure and result in self-extinguishment
before other fuels in the room can be ignited. It is difficult (and often ineffective) to trans-
port and disperse more fuel than can readily be carried by hand. Except in rare instances,
less than a few liters (gallons) will be used. In instances where large drums of volatile
hydrocarbons are readily available (such as in an industrial situation) or where the fuel
tanks or vehicles can be used as sources, massive quantities of flammable liquids can be
employed. Such unusually high fuel loads can be used to ignite a secondary source of fire
such as a large accumulation of automobile tires, which burn ferociously and are all but
impossible to extinguish, or the wood framing of the structure itself.
The most frequently used amount of an accelerant is that amount a person can carry
conveniently without attracting undue attention. A person can easily carry an amount suf-
ficient for igniting a very large fire, so this is not necessarily a serious limitation. The ama-
teur nearly always uses much more of a volatile accelerant than is necessary. By using it
judiciously on top of loosely piled combustibles, a person can start a very satisfactory fire
in a building using only a very small quantity of accelerant. If one were trying to burn a
large building in an urban area where detection and suppression would quickly follow
ignition, it might be desirable to employ a large reservoir of liquid fuel over a period of
time. A metal drum of a suitable fuel such as gasoline, paint thinner, or other petroleum
674 Chapter 16 Arson as a Crime
 FIGURE 16-2 A large
LPG cylinder placed inside
this car and set alight
caused massive explosive
damage. Courtesy of Det.
Sgt. Ken Legat, New Zealand
Police, Christchurch.

product might be placed at the desired point of origin with enough fuel under and around
it to start the fire and to heat the drum to the point of rupture. A well-filled drum is not
difficult to burst by heating, and the ruptured drum will then spray fuel into the fire for
a long time with virtually no possibility of being extinguished by conventional water or
fog sprays. There is always the possibility of an explosion from such a device when flam-
mable vapors are suddenly mixed into a large fire, but maximum damage is usually
desired by the arsonist. Such a drum placed at the origin of the fire kindled with addi-
tional fuel may not be suspected as an instrument of arson because there are many legit-
imate reasons for the storage of liquid flammables on most industrial or commercial
premises. Cylinders of compressed gases such as LPG cylinders may also be used, as shown
in Figure 16-2.
If the fire is of very limited duration and has not spread horizontally or progressed to
flashover, the approximate amount of flammable liquid used may be estimated from the sur-
face area of the “pool” of fuel. Due to its low viscosity, gasoline will flow freely on a
smooth, nonporous floor to a depth of 0.5–1 mm (0.04 in.), so a 1–L (1–qt) quantity will
cover a maximum area of about 1–2 m2 (21 ft2). If it is poured on a semiporous surface (not
just scattered or sprayed), such as smooth unpainted wood or concrete, its effective depth
will be about 2 to 3 mm (0.08–0.12 in.), so a liter (1 qt) of gasoline will saturate an area of
about 0.3 to 0.5 m2 (3.2–5.4 ft2). On carpet, the saturation depth is the thickness of the
carpet (plus that of an absorbent pad beneath), 10 to 20 mm (0.4 to 0.8 in.). The surface
area of a pool can be very small, with 1 L (1 qt) saturating 0.05 to 0.1 m2 (0.5–1 ft2).47
The amateur arsonist does not always place the liquid accelerant where it will pro-
vide the most effective ignition. Instead, most commonly, he or she pours accelerant over
carpeting, furniture, and open floor, a pile of paper or trash, or a combination of these.
Pouring a flammable liquid onto an absorbent material, such as soil, newspapers, or car-
pet, allows some fuel a chance to escape combustion. For example, a stack of newspapers
soaked in gasoline allows the fire to burn only on the outside of the stack, with the
absorbent papers feeding additional fuel to the fire in the same manner as the wick of a
kerosene lamp. Fuel near the center of the stack is not exposed to air or even to extreme tem-
peratures and remains for detection days later, especially if the debris has been water-soaked
Chapter 16 Arson as a Crime 675
 to minimize post-fire evaporation. A pool of paint thinner in a carpet may burn a distinct
ring, with the carpet pile in the center of the ring barely scorched and a substantial por-
tion of the fuel still present where it was not exposed to the flame–air interface. Fuel
soaked into the soil likewise is poorly consumed and then only at the surface. Loscalzo,
DeForest, and Chao poured gasoline on plywood, carpet, and soil, ignited it, and tested
detectability of the residue at various time intervals. Depending on the length of time the
gasoline was originally allowed to burn, gasoline residues were identifiable by gas chro-
matography as long as 24 hours after burning on carpet and as long as 162 hours after
on soil.48 More recent tests showed that small quantities of kerosene (unburned) were
detectable by standard forensic methods on denim jeans for only 24 hours and on shoes
for only 72 hours, stored indoors at 22°C (68°F). Unburned gasoline was detectable for
less than 24 hours on shoes and denim jeans under the same conditions but was still
detectable on carpet after 7 days of storage outside at 5°C to 15°C (40°F to 50°F).49
Recent tests using the most modern, more sensitive techniques but with smaller amounts
of gasoline showed detectable quantities only up to 4 hours on clothing.50
The splattering of liquid fuel on a floor or furnishings will often produce detectable
residues on shoes and pants cuffs, even more so if the arsonist steps into the liquid. The
gas chromatographic profile of such liquid transfers will readily be distinguishable from
the light volatiles transferred by vapor contact alone. The clothing items worn by a sus-
pect in an “accelerant” fire should always be collected as soon as possible and packaged
separately in appropriate cans or Kapak (made by Ampac) bags pending lab analysis.
Because of the relatively short “lifetime” of such traces, a positive lab finding is a useful
indicator of recent contact.
The sensitivity of today’s methods and the detection of background petroleum prod-
ucts in innocent clothing (especially athletic shoes) make it important to package items
separately and collect comparison samples (of liquids and clothing). A common explana-
tion for gasoline traces on hands is that gasoline was spilled while fueling a vehicle. Tests
have shown that this process does not produce detectable amounts on the hands if prop-
erly conducted. Even when a living subject’s hands are deliberately doused with 1 ml of
gasoline, gasoline is detectable by the most sensitive method for only 2 hours after con-
tact.51 (See Chapter 14.)
The pouring of a flammable liquid directly on the floor is probably the most com-
monly detected form of arson. It is readily detected because a considerable portion of the
fuel remains unburned, and what does burn can leave characteristic burn patterns or diag-
nostic signs. Poor ignition is due in part to the amateur arsonist’s misconceptions about
the properties of such fuels. On a smooth, nonabsorbent, tight floor, little or no damage
may result if the burning of the fuel did not raise the floor covering to its ignition point.
(As seen in Chapter 4, the liquid fuel tends to protect the floor covering by evaporation
rather than damaging it.) On a floor with cracks, holes, or joints into which the fuel can
penetrate, holes and lines will burn into the floor in a characteristic manner (see Figure
16-3). Remember that if the room subsequently goes to flashover, the radiant heat pro-
duced may obliterate many of the more subtle indicators on the floor. Sustained post-
flashover burning will quickly burn wood floors to complete penetration. (Many of the
misinterpretations of burn patterns as indicators of arson were discussed in Chapter 7.)
On ground-level floors, sometimes enough flammable liquid is poured out that it seeps
completely through the floor and into the soil beneath. In such cases, the wooden sub-
flooring, joists, and soil should be checked for the presence of volatile fuels. Another com-
mon mistake for the amateur is to dump the fuel in the center of the room without
ensuring that the flames will reach furniture, walls, or other fuels to keep the fire going
(as in Figure 16-1).
Not all sets or trailers involve ignitable liquids. As seen in Figure 16-4, crumpled
newspaper or wax paper can be used. Such materials may burn quickly and completely
to leave very little residue.
676 Chapter 16 Arson as a Crime
 FIGURE 16-3 Gasoline
pool fire on wood plank
floor after 1 minute.
Gasoline flashed off flat
areas and continued to
burn only in joints.
Courtesy of Jamie Novak,
Novak Investigations and St.
Paul Fire Dept.

FIGURE 16-4A Trailer of crumpled newspapers. Courtesy of FIGURE 16-4B After fire in hall: minimal residues. Courtesy of
Lamont “Monty” McGill, Fire/Explosion Investigator. Lamont “Monty” McGill, Fire/Explosion Investigator.

Chapter 16 Arson as a Crime 677

 Multiple Ignition Points
Multiple ignition points are common in arsons. Not satisfied to gamble on a single point of
ignition, the arsonist thinks that “more is better,” with the intent of producing a larger fire
that spreads more quickly and widely than that from one point. These multiple points are
usually connected by a trailer of flammable liquid, smokeless gunpowder, rags, twisted
ropes of newspaper, waxed paper, or wood packing material (excelsior). In a few cases, trail-
ers of flammable liquid were poured from electrical outlet to outlet along a baseboard,
apparently in hopes of simulating massive electrical fires. Although multiple sets can pro-
vide a more complete destruction of the target, they present many problems that the arson-
ist would just as soon avoid. They require additional materials and fuel, and there is also
the chance that one of the elements will fail, leaving intact sets to be discovered. It requires
more time to distribute fuel in several sites than one, and in the process of distribution,
vapors from the earlier sets may be forming an explosive mixture that may deflagrate,
involving the arsonist. The arsonist’s shoes and clothing are often splattered with the fuel
and may ignite. The hazards from these sources are likely to outweigh any advantage that
might be gained from multiple points of ignition, especially if the fuel has a high volatility,
like gasoline or lacquer thinner. The time required also prolongs the chances that the earlier
sets will be detected and the perpetrator will be spotted while still at work. A low-volatility
fuel such as kerosene may be suitable for such operations, but it requires a “wick” to pro-
vide an ignitable vapor, since it cannot be ignited as a pool on a nonporous surface.

FALSE MULTIPLE ORIGINS

It should be noted that multiple points of ignition are not necessarily proof of arson. Flames from a large
ordinary combustible fire can melt and ignite the polystyrene light diffusers used in ceiling lighting fix-
tures, draperies, or the asphalt from a collapsing roof. These flaming “liquids” can produce multiple,
apparently unconnected, points of origin throughout a large area. Such multiple points have been mis-
taken in the past for arson sets. The investigator should be careful in documenting the scene to establish
where such light fixtures, draperies, or roof vents were prior to the fire. (See Figures 16-5a–c for an example

FIGURE 16-5A Fire started with a candle: FIGURE 16-5B Burning draperies fall.
initial vertical spread. Courtesy of Jamie Novak, Courtesy of Jamie Novak, Novak Investigations and
Novak Investigations and St. Paul Fire Dept. St. Paul Fire Dept.

678 Chapter 16 Arson as a Crime

 FIGURE 16-5C Chair in
corner ignites to produce
a false second origin.
Courtesy of Jamie Novak,
Novak Investigations and St.
Paul Fire Dept.

of a candle fire spread by draperies to look like separate origins.) A sudden surge on the line power or a
lightning strike can induce a surge in line voltage that can cause ignition of multiple appliances (or
anything plugged in). That hypothesis has to be ruled out, but since most power line transformer– or
lightning-caused surges will affect multiple dwellings, neighboring properties should be checked.

Type of Accelerant
As we have seen previously, the type of flammable liquid used as an accelerant is deter-
mined in great part by the experience, motivation, and intent of the arsonist. In a premed-
itated structure fire, some thought will usually be given to selecting an appropriate fuel.
Gasoline is by far the most commonly detected accelerant because it is an efficient and
easily ignited flammable liquid and because it can be purchased and transported even in
large quantities without arousing the suspicions of the seller or authorities.
Paint thinners and kerosenes are the next most common fuels.52 They are effective
fuels and can be purchased in moderate quantities without suspicion because they have
legitimate uses even in the private home. Turpentines and lacquer thinners are almost
never encountered as premeditated fuels.
The circumstances or living conditions may dictate what fuel is most convenient—
kerosene or even camping fuel may be closer at hand than gasoline. The authors know of
one fire set that used gum turpentine because it was readily available as part of a remod-
eling project. Gum turpentine is very costly and not often used in bulk by private individ-
uals. The chemistries of new paints make the use of solvents more of an exception than
the rule it once was, since most paints sold today for household use are the water-based
latex type, which do not require ignitable liquids as thinning or cleaning agents.
Although most arsonists seem to give little thought to the accelerant’s volatility, it is
a property that has great bearing on the fuel’s suitability to the task. Extremely volatile
fuels such as gasoline, lacquer thinner, or alcohols may evaporate excessively at high tem-
peratures, interfering with the intended ignition and spread of the fire. The vapors gener-
ated can travel considerable distances and produce explosive mixtures in the vicinity of
distant “accidental” sources of ignition, most typically the pilot light on the water heater
or furnace. This makes it necessary to set the accelerant very quickly and to ignite it from
Chapter 16 Arson as a Crime 679
 a safe distance, thereby reducing the arsonist’s control of the set. Kerosenes, paint thinners,
and proprietary lantern fuels are safer to use because of their lower volatility. The varieties
of lantern and campstove fuels can vary considerably in their volatilities and flash points.
Because they may contain heavier hydrocarbons with commensurately higher boiling
points, however, they are more resistant to evaporation and will remain in the debris longer
as residues than will, say, lacquer thinners. Some fire setters use a 50/50 mix of diesel fuel
and gasoline, thinking the diesel fuel will make it “safer” to use. Unfortunately, the vapor
pressure of the gasoline dominates the ignitability of the vapor mixture. Until there is less
than 5 percent gasoline in the mix, the liquid will ignite just the same as gasoline.
At the opposite end of the petroleum distillate scale is fuel oil. It is very difficult to
burn except when it is dispersed into a mist and well mixed with air. When applied to
paper, cloth, or trash, it can burn reasonably well from the wick and thereby add to the
fuel load, but it will not be ignited as a pool.
Alcohols, both methyl (wood) and ethyl (grain), are as volatile as gasoline and carry
the same hazards. Because of their flammability and their reduced chances of detection
(due to their miscibility with water), they would appear to be good choices as flammable
accelerants. They are not, however, efficient fuels, and it takes a larger quantity of burn-
ing alcohol to generate sufficient heat to bring other nearby combustibles to their ignition
points. Alcohols will still burn when mixed with some water, but with a lower risk as well
as effectiveness. Potable liquors of 90 to 100 proof (45 to 50 percent alcohol) can be used
as accelerants, though not very effective ones. Alcohols may also be encountered in chaf-
ing dish heaters and campfire starting fuels (Figure 16-6). They produce very little dam-
age to the surface beneath and leave little if any residues post-fire.

FIGURE 16-6 Campfire

starters may be either
wax/sawdust sticks or
viscous ignitable liquids
such as the one here.
Courtesy of John D. DeHaan.

680 Chapter 16 Arson as a Crime

 The foregoing discussions deal with premeditated fire sets involving accelerants.
Depending on the motivation and opportunity of the arsonist, such fuels may not be
acquired beforehand. When it comes to fires set using materials already at the scene, inge-
nuity knows no bounds. Glues, dye solvents, rubber cements and their solvents, mop
cleaners, uncured polyester resin, paints, brush cleaners, and the like are common in
homes, schools, and businesses, and all have been used as effective arson accelerants.
Under these circumstances an inventory of the scene and a thorough interview with own-
ers, tenants, and maintenance people may be vital to the success of the inquiry. Although
it is commonly known that the duplicating fluid used in the old-style “spirit” duplicators
or mimeograph machines is highly flammable methyl alcohol, it is not commonly known
that the copier toner used in some older office “dry” copiers contains a petroleum distil-
late or isoparaffinic blend similar to paint thinners. Many offices have cases of this fluid
stored where it is easily accessible to burglars intent on erasing signs of their presence. The
same holds true for commercial insecticides that are often dissolved in a paint thinner–
like petroleum distillate or aromatic solvents. When such unusual fuels are suspected by
the investigator, these suspicions should be relayed to the laboratory examiner along with
control samples of all possible fluids observed at the scene.

METHOD OF INITIATION
The methods chosen by the arsonist to initiate the fire will be closely related to the type
of fuels used and sometimes to the manner selected to best avoid detection. If liquid accel-
erants have been used, an almost immediate ignition is desired because the vaporizing fuel
may be lost or form an undesirable explosive hazard. With less volatile materials, the time
of delay may still be considerable, but it is impossible to leave open-flame devices such as
a lighted candle. These will not delay the fire for long. If the primary fuel is not volatile,
a delayed-action igniter may be suitable, but in general, the development of the fire will
be less certain. The choice of primary ignition devices include flames, fuses, smoldering
materials, cigarettes, electric arcs, glowing wire, chemical ignition, black powder and
flashpowder, and improvised chemical delays.
Flames
Simplest of all ignition devices to produce, the flame of a match or lighter is easy to
obtain, to apply, and to eliminate as evidence. The flame is a certain means of starting the
fire because it is a fire already and requires only propagation. Whatever is left at the scene,
for example, a burnt match, is likely to be destroyed to the point of escaping detection.
One exception is the common candle, which often is not sufficiently destroyed to
remain undetected. Melted wax will persist in a fire if it accumulates on a noncombustible
surface even though the form of the candle itself will be lost. Remnants of the wick and
metal wick anchor tab will often survive a generalized fire.
Another source of a flame, often overlooked as being an accidental malfunction, is an
altered gas pilot light assembly or a broken gas line that is set to throw a blowtorch-like
flame on nearby combustibles. A special case is that of a flamethrower, weed burner, or
similar device that generates and maintains a large flame for a protracted time. The use of
such a device is not common but is highly effective even when a refractory fuel like fuel oil
is employed. It is possible to ignite even large timbers and to spread a fire widely through
a variety of combustibles with such a flame source. In one sawmill fire, it was determined
that a weed burner was the igniter because of extensive deep burns on the bases of heavy
timbers that were lying on soil where no other source of such intense flame was possible.
The fire was started in this fashion in a number of places so that overall development of
the fire was very fast and destruction extensive. The weed burner was found nearby.
Molotov Cocktails and Delay Devices
Of all the remote-ignition devices known, the Molotov cocktail is the most common. It
consists of a breakable glass container of a flammable liquid that depends on a lighted
Chapter 16 Arson as a Crime 681
FIGURE 16-7 Flame
plume from Molotov
cocktail (1 qt gasoline).
Courtesy of John D. DeHaan.

wick or a chemical incendiary device to provide ignition. Its effects are maximized when
it can be thrown against a hard surface (wood is often not hard enough) to break the
bottle and mechanically atomize the gasoline into an explosive mist. The ignition then
produces a fireball that can spread at very high speeds (up to 10 m/s; 40 ft/s). Although
a Molotov is usually ignited by its own flaming wick, hypergolic mixtures have been
used to avoid the use of an open flame. A packet containing sugar and potassium chlo-
rate is taped or tied to the outside of the bottle, and sulfuric acid is added to the gaso-
line inside. Upon breaking of the container, the acid reacts with the sugar and chlorate
to initiate a very hot flame that, in turn, ignites the gasoline. The burning time of the
gasoline from a thrown quart-size Molotov (as shown in Figure 16-7) is on the order of
approximately 10 seconds, so the target has to be ignited in that time, or only scorch
marks will result. Although the production of a thickened “napalm” has been recom-
mended in terrorist literature, the minimally enhanced burning effects are generally not
worth the extra time and effort required to mix the gasoline with soap flakes, poly-
styrene foam, or paraffin wax.
Although the use of such a device is convenient from the standpoint of the arsonist,
its easy detection at a fire scene makes it less than ideal (for him or her). A broken bot-
tle, often a broken window, and perhaps eyewitnesses accompany its use. Tracing it to the
arsonist may be difficult but often is not. The broken bottle is usually found at the bot-
tom of the fire, where it is exposed to the minimum amount of heat, so fire damage to the
fragments is minimal. The portions of the bottle bearing the most information about its
origin—the neck with its labeling and the base with its cast-in production data (date and
place of production and product code)—are the pieces most resistant to mechanical and
fire damage and usually survive intact. The investigator should endeavor to collect all or
nearly all the pieces to allow reconstruction of the device for identification or even a
search for fingerprints. (As noted in Chapter 14, the oily components of latent prints are
somewhat soluble in gasoline and may be washed off by the accelerant as it seeps from a
broken container, but new techniques now promise recovery of prints even in the pres-
ence of gasoline.) As described previously, DNA has been identified on the mouth and
wick of Molotov cocktails that have burned.
682 Chapter 16 Arson as a Crime
 (a) (b)

FIGURE 16-8 (a) Delay device—plastic container of gasoline with sponge wedged in handle. Sponge set
alight, flames melt top of bottle and ignite gasoline within. (b) Post-fire. Courtesy of Jamie Novak, Novak Investigations
and St. Paul Fire Dept.

Plastic bottles or containers filled with fuel have been encountered in a number of
fires (see Figures 16-8a and b). These may be placed over room heaters to soften and melt
but are more commonly surrounded by loose paper and trash that is set afire to provide
a means of softening them and igniting the spilled contents. (There also may be a timing
delay such as a cigarette-and-match device.) If properly supported by combustibles, these
containers can burn sufficiently to make identification impossible.
One variation on this is the use of a toy balloon or plastic bag containing gasoline
that is suspended by a string and set swinging over a candle. Once the oscillations grow
smaller, the balloon stops over the candle, which then melts a hole in the bag; the fuel
released quickly spreads the fire to the surrounding trash. Although effective if small
quantities of accelerant are used, the amateur will often use a large trash bag with lots of
fuel that is more likely to extinguish the candle as it rushes out than be ignited by it.
Fuse
A fuse must be considered as a means of timing delay, a trailer connecting several points
of origin, and a means of directing a flame to a remote location. There are three basic types
of fuse: (1) pyrotechnic, which employs a train of flammable powder packed around a
string core: (2) safety, which contains a train of black powder within layers of string and
Chapter 16 Arson as a Crime 683
 FIGURE 16-9 Pyro-
technic or rocket fuse
burning on plywood pro-
duces an external flame
that scorches the surface
beneath it. Courtesy of John
D. DeHaan.

asphalt wrappings and sometimes a continuous tube of water-resistant plastic; and

(3) igniter cord. All three types of fuses should be considered essentially a remote-control
match, a means of transmitting a live flame from one point to another at a predictable rate.
The pyrotechnic fuse has an exposed flame at all points along its length and will
scorch and even ignite combustibles upon which it rests (as shown in Figure 16-9).
Pyrotechnic fuse, also known as hobby or rocket fuse, is widely available in hobby stores,
while safety fuse is almost always sold through suppliers of commercial blasting supplies.
As time-delay devices, pyrotechnic fuses can vary considerably in their burning rates,
from 30 s/ft (2 ft/min; 0.6 m/min) to as quickly as 10 s/ft (6 ft/min; 1.8 m/min). Safety
fuse has an exposed flame only at its ends. Although the asphalt coating will melt and
smolder, the wrappings will remain in place, and damage to surroundings will be minimal
(as shown in Figure 16-10). Safety fuse is made to exacting standards to burn predictably
at 30 to 40 s/ft (0.5–0.6 in./s; 1.2–1.5 cm/s) at sea level.
Igniter cord has a core of pyrotechnic powder wrapped in wire and is used only in
special blasting operations where its high temperatures and quick burning times are desir-
able. After burning, a multiple spiral of fine wire is left behind.
Smoldering Materials
Although mattresses, sawdust piles, and other fuel materials with limited access to air may
smolder and ultimately develop into a large fire, they are not common ignition sources for

FIGURE 16-10 Burned

safety fuse leaves a trail
of melted wax on the
surface but scorches the
wood only at the end
where the flame exits
(lower left corner).
Courtesy of John D. DeHaan.

684 Chapter 16 Arson as a Crime

arson sets. Their major use is in delaying the outbreak of the fire, allowing time for escape.
Unfortunately, such fires are unreliable in their behavior, and the heavy smoke they pro-
duce increases the chances of detection before the “real” fire can take place. Slow-burning
materials like a cattail, incense, or punk have been used as delay devices when wrapped
with matches. The matches in turn are set to ignite a black-powder or pyrotechnic initiator
(as seen in Figure 16-6). Smoldering ignition sources alone will not ignite liquid acceler-
ants, which require open flame.
Cigarettes
Cigarettes are rather uncommon as the source of ignition for a successful structure fire. If
they are set to ignite liquid flammables, they will almost certainly fail. If they are used to
start a smoldering fire (typically in a couch, chair, or bed), the results will be highly uncer-
tain. Prank fires started with a cigarette are more likely than a planned and deliberate
arson, because the prankster may throw a lighted cigarette into appropriate fuel as a
whim or experiment. The lighted cigarette in the wastebasket is a source of accidental
fires from carelessness, but a fire can also result from a prank. When thrown into crum-
pled paper, a lighted cigarette tends to char its way through the papers it encounters with-
out providing a sufficient source of heat to initiate self-sustaining combustion.
Although the cigarette alone may be of little value in igniting fires, there are ways
of combining it with other flame sources to enhance its value. It is commonly found
in combination with matches. If the lighted cigarette is placed between two layers of
paper matches in a matchbook, when it burns down to the heads it will ignite them,
and the entire book will burst into flame. Another variation is to tie a number of
matches around the outside of a cigarette, which is then allowed to burn down to
them. Such a device is fairly resistant to burning to completion (especially the match-
book), and its remains can be found by careful examination of the point of origin (see
Figure 8-22). These are the most commonly detected delay devices in grass and wild-
land fires. Although they rarely burn so completely as to avoid detection, they are usu-
ally charred sufficiently so that fingerprints, physical comparisons, and saliva (DNA)
testing may not be possible. Improvements in DNA techniques and more extensive
DNA databases are making it more likely to identify suspects from cigarette-butt test-
ing, however.
Electric Arcs
As we saw in Chapter 10, electric arcs create a great deal of energy but usually for only
a short period of time. As a tool of the arsonist, the arc can be used to ignite only certain
types of fuels. Most such fuels are explosive in nature, either explosive fuel/air mixtures
or sensitive low explosives like black powder or flashpowder. A fire ignited by an arc,
then, will be characterized by an initial explosion. The only alternative to the explosive
fuel as an initial fuel is to convert the arc into a flame. This may be done by a powder
trail or a readily ignitable flashpowder. The use of solid fuels in conjunction with arcs is
uncertain and not an efficient method of kindling fires.
Because arcs can be generated by electrical means, delay devices using electric arcs
would seem to offer many advantages to the arsonist. However, without elaborate
arrangements of ignitable vapors, explosives, or similar active fuels, and corresponding
elaborate electrical devices that will, in part, survive the fire, it is difficult to ignite a fire
by such means. Any custom-made device that survives the flames increases the chances of
identifying the fire as incendiary in nature and provides possible clues with which to track
down the perpetrator. Such devices are therefore usually avoided by professional arson-
ists. Some ingenious devices have been employed that used the arc generated when a light
switch is turned on or a motor is started. For example, a short circuit can be set up
between two wires of a multiwire power cord (even by the simple expedient of a wire sta-
ple through the insulation). The short-circuited portion is then put in contact with read-
ily ignited fuel. Because the short circuit is minor it is difficult to trace, but the heat
Chapter 16 Arson as a Crime 685
 generated by even a brief arc can start a fire if nearby fuel is susceptible. Even the most
ingenious devices are unreliable, however, and are not often encountered.
Solar
There have been reports of focusing mirrors (see Chapter 6) or convex lenses set up with
ignitable fuel at their focal point. Only when the sun would be in the right position relative to
the set device would there be enough energy to cause ignition. The resulting residue—broken
glass and a charred framework—could easily be overlooked during the scene examination.
Candles
Candles have been used alone (burning down into combustibles) or in combination with
other devices for many years. Wax candles are useful timing delays (their combustion rate
is determined by their design and diameter—the smaller the diameter, the “faster” they
will burn down). They are, however, easily extinguished by drafts, and some can fall over
easily, making them less reliable as ignition sources. Residues of wax or unburned wick
may be impossible to find on heavily burned substrates like upholstery.
Glowing Wire
Although they are rare, devices using an electric current to heat a wire to incandescence
to ignite a combustible material in contact with it have been reported by some investiga-
tors. Such a device would allow the fire to be ignited by any electric circuit, which could
by completed by a timer, by remote radio control, or even by an unwitting bystander. The
elements used could come from a heater, toaster, lightbulb, or flashbulb. (The latter, of
course, includes its own limited supply of flammable fuel.) One device reported in the
underground literature involves drilling a hole in the base of an electric lightbulb, par-
tially filling it with black powder or incendiary mixture, and screwing it back in place.
When the light is turned on, the glowing filament ignites the powder, and the heat from
the combustion shatters the lamp, allowing the burning powder to fall onto flammable
fuels below. A more commonly encountered glowing-wire device is an electric heater ele-
ment connected to a power source through a timer. The heater is wrapped or packed with
paper or cloth to provide fuel for ignition. This fuel is often supplemented with gasoline
in plastic bags or bottles set on the cloth or paper.
Chemical Ignition
A wide variety of chemical igniter systems have been devised by arsonists. Fortunately,
they are sufficiently sophisticated and unpredictable that they are employed only rarely.
They can be as simple as metallic sodium or potassium brought into contact with water.
Small chunks of such materials are dropped into water that reacts with the metal, produc-
ing a flash of flame that can ignite volatile fuel/air mixtures under the correct conditions.
Incendiary mixtures can be more complex, such as thermite—a mixture of aluminum
oxide and iron filings that when ignited by magnesium “starter” produces an intense fire.
This fire can reach a temperature of thousands of degrees and burn through steel and con-
crete. It leaves a significant amount of residue that is chemically distinctive but may be
innocuous in appearance and easily overlooked. The magnesium igniter may be difficult
to obtain and initiate on its own, so the use of thermite devices is fortunately rare.
Black Powder and Flash Powder
Mixtures of aluminum powder and potassium perchlorate or potassium nitrate must be
considered effective chemical initiators. Black powder or flashpowder from small “fire-
crackers” can ignite exposed fuels or larger powder charges. The flashpowder in an M-
80-type explosive device (as in Figure 12-7) can shatter a glass bottle of gasoline and
ignite a sizable fireball in the manner of a well-thrown Molotov cocktail.
Improvised Chemical Delays
Dry chlorine of the type used for swimming pool purification can spontaneously ignite a
mixture with glycol-based fluids, such as brake fluid or hair dressing (see Figure 16-11).
686 Chapter 16 Arson as a Crime
FIGURE 16-11 Clear flame plume from glycerin–pool FIGURE 16-12 Post-fire residue of device in Figure 16-11 is not visu-
chlorine mixture is very hot but low in radiant heat ally distinctive; however, residues of the foam cup and some unreacted
and lasts only a few seconds. Courtesy of Keith Parker, pool chlorine are still present. Courtesy of Keith Parker, Marin County, Fire
Marin County Fire Department, Woodacre, CA. Department, Woodacre, CA.

Residues from such mixtures or from acid–chlorate mixtures previously described are not
remarkable in their post-fire appearance and are also easily mistaken for plaster debris (as
in Figure 16-12). A wide variety of other chemical systems are susceptible to spontaneous
ignition under proper circumstances, and the means are known for controlling the reac-
tion time.53 These means include physical ones where a physical barrier separating the
reactants is allowed to fail or where a chemical solvent, which keeps the reaction from
proceeding, is allowed to evaporate. Other timing factors include particle size, amount of
mixing, and the temperature of the environment or of the reactants. The influence of these
controlling factors serves to make the performance of such mixtures unpredictable. The
residues are variable depending on the original mixture and the completeness of the reac-
tion. They may be overlooked or misinterpreted by even the best scene investigators. If
there is reason to believe that a fire is arson and that such an “exotic” means of ignition
was used, then the investigator is wise to enlist the aid of a competent chemist or crimi-
nalist. Such improvised chemical mixtures can also be used as the main accelerant.
Mixtures of metallic aluminum or magnesium powder and strong oxidizers (chlorates,
perchlorates, nitrates, or even sulfates) have been demonstrated to be capable of generat-
ing very high temperatures and extremely high heat release rates. These incendiary mix-
tures are similar to flash powder formulations but contain a binder (wax, rubber, or diesel
fuel) to prevent aerosolization and explosion while acting as an additional fuel.

Deductions from the Interpretation of Evidence

All the preceding considerations are important to the arsonist deciding which procedure to
use in starting the fire. They are of equal consequence to the fire investigator who must cor-
relate the observed facts about the fire with the possibility that any fire may be deliberately
Chapter 16 Arson as a Crime 687
 set. They also serve as a guide to the thinking and experience of the perpetrator. Inept or
inexperienced arsonists may set themselves on fire, cause unnecessary or premature explo-
sions, or kindle a fire that is not self-sustaining or capable of developing into a large confla-
gration. Such occurrences as well as the availability of fuels at the scene and the means of
procuring other fuels or devices that can be traced all serve as indicators of the possible iden-
tity of the arsonist. Because arsonists are sometimes very ingenious, the investigator must
develop a comparable ingenuity in investigating cases as well as a highly developed obser-
vational ability.

CRIMINAL INVESTIGATIVE ANALYSIS OR PROFILING

One of the most interesting developments of recent years has been the improvement of
offender profiles, especially those dealing with serial crimes. These profiles offer inves-
tigative leads of the inductive type—predicting the age, race, sex, occupation, socioeco-
nomic status, and mental state of an offender—based on the data extracted from an
analysis of the crimes. Several textbooks address criminal investigative analysis and
offender profiling.54
Creating a profile involves an in-depth study of the crime scene and the victim as well
as psychological and sociological data. The FBI’s approach used at the NCAVC for creat-
ing a profile mirrors the scientific method in that it follows a series of five overlapping
stages that resemble a medical diagnosis. These stages consist of:
1. Collecting and assessing data
2. Reconstructing the situation
3. Formulating a hypothesis
4. Testing the hypothesis against known information
5. Reporting the results55
Determination of an apparent motive is the primary step in developing an offender
profile that is designed to focus on two fundamental questions: Who committed the
crime? and Where can they be apprehended? Data evaluated on incidents include police
and fire reports (including origin and cause determination), date, time, and description of
target. The areas involved need to be mapped, showing the locations of all fires (not just
those identified as arson-caused), and demographic data must be available. All incidents
are analyzed for trends or patterns in time or place of occurrence, sequence, target, or
means of ignition. (One of the possible values of the new BATS system described in
Chapter 12 is trend analysis linked to geographic mapping.) Predictions of date, time,
place, or target can be tested against newly reported incidents as a pattern of behavior is
elucidated. These models are sufficiently sophisticated to allow the prediction of the sub-
ject’s state of mind and behavior after the fire as well—immediately after and long term,
and even his or her reaction to police contact.56 Some studies have been based on care-
fully structured interviews with hundreds of convicted arsonists, both male and female.57
They have greatly expanded the knowledge base of fire setters and improved the accuracy
and precision of analytical and predictive methods.58
Geographic Profiling
Offender-based classification of arson extends beyond merely identifying the motive.
Investigators should also closely examine the geographic locations selected by the arson-
ist. An analysis of these sites may reveal much about the intended target, add insight into
the arsonist’s motive, and provide a potential surveillance schedule for attempting to iden-
tify and apprehend the arsonist. The results of a joint research effort on the geography of
serial violent crime illustrate the patterns found in serial arson cases. Results indicate
those criminal offenders who repeatedly set fires exhibit temporal, target-specific, and
spatial patterns that are often associated with their modus operandi (method of operation
or “m.o.”) and motive.59
688 Chapter 16 Arson as a Crime
 With geographic profiling, the analysis of times and locations of events has shown
great promise in evaluating criminal trends, allocating preventive or response resources,
or assisting investigators. Because the locations are linked to fire details, similarities in
modus operandi can be spotted, aiding in the identification of serial arsonists. Depending
on the motive classification described earlier, some serial fire setters will target selected
areas by convenience, accessibility, travel time, or target type. Geographic profiling has
been very useful in establishing such patterns, estimating the residence or base location of
the perpetrator, or predicting future targets.60
The two primary references in the discipline of geographical profiling are by two pio-
neers—Kim Rossmo in Canada and David Canter in the United Kingdom61 followed by
extensive work in Australia by Richard N. Kocsis.62 A Web-based commercial computer
program, CrimePoint Fire and Arson, allows any investigator to exploit this valuable
investigation tool.63
Work by Katarina Fritzon in the United Kingdom examined and modeled the spa-
tial relationship between the distance traveled and behavioral motivational aspects of
arsonists. This research confirmed that arsonists whose behavior contained a strong
emotional component tended to travel much shorter distances, on average, than other
arsonists. Wherever an individual was targeted in a revenge-motivated attack, the
arsonist was prepared to travel a greater distance. Also, arsonists who had recently
separated from a partner were more likely to travel greater distances, due to the target-
specific nature of the crime. Trends were also seen between the age of the arsonist and
the distance traveled.64
Tracking Serial Offenders
Work by Brett Martinez lays out the strategy for the identification and tracking of serial
arsonists. Using basic crime analysis techniques, his textbook assists an experienced fire
investigator in detecting and identifying temporal and geographic patterns of arson fires
as well as understanding the motivation behind these crimes. Martinez provides examples
of how to present information in a final report for dissemination to all personnel involved
in the investigation.65

ANALYTICAL REASONING
Although the physical evidence of an arson set and its laboratory analysis are often of
great value to the investigator in reconstructing the crime and identifying the perpetrator,
successful prosecution of the case is dependent almost entirely on the expertise of the
investigator and his or her skill in reliably interpreting the diagnostic signs of fire behav-
ior at the fire scene. Many cases have been brought to successful completion even in the
absence of conclusive laboratory results with respect to the presence of flammable or
incendiary materials. Analytical reasoning skills found useful to successful FBI profilers
consisted of experience in the investigation of crime, understanding human behavior and
motivations, and a capacity for logical and objective reasoning, and intuition.66
When an investigator considers motive, it must be remembered that in virtually any
in-depth inquiry, facts will be uncovered that may be construed as “motive.” Everyone
who has a mortgage and an insecure job has a potential motive for burning the house.
Anyone faced with expensive repair estimates on a vehicle is tempted to “sell it to the
insurance company.” Means and opportunity are, likewise, available to any person who
lives in the house or drives the car. It is not enough (in fact, it is malfeasance) to simply
establish a possible motive and use that to bootstrap a fire investigation that on its own
merit would be an “undetermined” fire into an “arson” and then pursue it through legal
channels. The determination of whether a fire is arson or accidental has to be established
on the strength of the information from the scene investigation, supported by interview
data regarding the fire event. If that call cannot be made at the conclusion of the active
investigation, then the cause of the fire should be considered “undetermined.”
Chapter 16 Arson as a Crime 689
 The investigator must be prepared to analyze each fire carefully for the basic elements
described previously in this volume. Is the identified area of origin specific enough to per-
mit reliable assessment of ignition sources and first fuels in the area? Is there enough fuel
normally to account for the destruction seen, or was additional fuel needed? Is there a rea-
sonable accidental mechanism for igniting the normal fuels present? Could the suspected
deliberate ignition device accomplish the fire spread observed? Does the arrangement of
fuels and contents appear normal or contrived? Are the residues of ignitable liquid
detected the result of normal room contents or proof of an intentionally added acceler-
ant? Was the direction and rate of fire spread normal for the kind of fuel and amount of
ventilation present? As we have seen, fires in today’s furnishings can grow very quickly,
achieve flashover, and create extremely high temperatures throughout a compartment
without the presence of any accelerant or even in the absence of a deliberate ignition.
The extensive data collected by fire researchers regarding ignition and fire spread are
readily available today. The development of computer models of fire growth and spread
continues every year. Both fairly simple and very sophisticated computer programs exist
today that allow the investigator to enter data on the fuel, room dimensions, ventilation,
and insulation and predict the time and intensity of the developing fire, as well as predict
the onset of flashover. Some aspects of these models as investigative tools will be
described in Chapter 17.

ELIMINATION OF ACCIDENTAL OR NATURAL CAUSES

Because one of the elements of proof of the crime of arson is the process of elimination of
all reasonably possible accidental or natural sources of ignition, the investigator should
not overlook such causes even in the most obvious fires. The process of elimination of such
possible causes should be carefully documented as part of the thorough investigation,
because they will be of great concern in later judicial proceedings. This is where the scien-
tific method offers the best means for evaluating the conclusion reached. The more alter-
native hypotheses that have been formed, tested, and rejected, the more reliable the final
hypothesis or conclusion. The process also prepares the expert for cross-examination.
Practices such as ASTM E620: Standard Practice for Reporting Opinions of Technical
Experts recommend that an expert report elucidate what other explanations for the facts
have been formulated and tested.67 It should be noted that this practice includes reporting
of data that helped rule out alternative hypotheses considered.
Careless smoking or waste disposal methods are still common causes of structure
fires. The presence of people in the wilderness with their campfires, matches, cigarettes,
heaters, trucks, and motorcycles presents an awesome risk to wildlands. Accidental fires
annually burn far more acreage than any natural cause. Sparks from machinery and hot
exhaust systems from vehicles must be considered. Overloaded electrical systems and
heating or electrical systems in violation of building codes must also be evaluated. All
appliances or equipment that uses electricity or gas must be considered as possible sources
of ignition. Decorations, particularly Christmas decorations, may involve candles, tempo-
rary wiring, and electric lamps. Cooking utensils and exhaust systems (particularly those
in restaurant kitchens) combine both a source of heat and readily combustible fuels (foods
and greases). It seems that common sense cannot be legislated, and people constantly defy
logic and common sense in their storage and handling of hazardous materials. Gasoline
is used to clean car parts or remove tar from clothes without regard for its fire hazard and
so may be found in kitchens and laundries. Aerosol products are found in profusion in
homes and commercial locations, and every can holds LP propellant gases (and may also
have an ignitable liquid solvent). They are easily punctured or overheated and can cause
significant fire danger. Even what appear to be deliberate sets may be found to be acci-
dental: electrical insulation destroyed by rodents or insects, matches gathered by pack
rats, or even sources of ignition knocked over into combustible household furnishings by
690 Chapter 16 Arson as a Crime
pets, with disastrous consequences. Natural means of ignition such as lightning, overheat-
ing by sunlight (more prevalent in these energy-conscious times), accumulations of static
electricity, disasters such as earthquakes, and wind also occasionally cause serious fires.
In spite of appearances to the contrary, man’s destructive hand does not play a part in all
fires. The use of a fuel–ignition matrix such as described in Chapter 6 can help ensure that
all likely ignition sources have been evaluated in the context of the scene at hand.68
While a comprehensive fire investigation may be able to eliminate all possible acci-
dental causes on the basis of physical data alone, other information (if reliable) can make
it feasible. Witnesses (or surveillance cameras) that saw the start of the fire and saw a per-
petrator in the act of producing a flame, for instance, would be difficult to overturn.
Witnesses who reported no fuel (or insufficient) fuel or no competent accidental ignition
sources should be considered. A proper fire engineering analysis (based on a complete and
reliable reconstruction of pre-fire conditions) can test various hypotheses regarding the
development of the fire and production of fire patterns. This analysis can reveal which
ignition scenarios can be reliably eliminated. NFPA 921, in discussing this analysis, sug-
gests that it is more often more practical to eliminate all accidental sources of ignition
except for the use of an open flame.69 Sources of open flame (match or lighter) can read-
ily be removed from the scene, leaving no physical evidence behind. The hypothesis being
tested is really: Can the “first fuel” be ignited by any source other than open flame? In
the absence of conclusive physical evidence of the deliberate cause of the fire—residues of
a foreign ignitable liquid, the remains of an incendiary device, and a clearly defined area
of origin uncompromised by post-flashover fire or structural collapse, the investigator
must be very cautious in concluding that arson took place.

Chapter 16 Arson as a Crime 691

 CHAPTER REVIEW

Summary
Arson is a crime of violence that inflicts terrible losses of Arson is the most difficult crime to solve and pros-
life, property, and prosperity on everyone in a commu- ecute (as national statistics prove) because of the
nity. Three elements must be proven to substantiate a destruction of evidence and the problems of linking a
charge of arson: a burning (destruction) of property; specific person to the crime even after it has been
the burning was intentional, not accidental in cause; established that a crime was committed. The investiga-
and the fire was set with malice. Motives for the crime tor must be cautious that establishing a motive alone
include vandalism, excitement, revenge/spite, profit, (or even motive, means, and opportunity together)
crime concealment, and extremism (terrorism). The does not constitute proof of guilt of the crime of arson.
selection of the target, as well as the preparations and The court will treat a fire (no matter how serious) as
manner of setting the fire, may yield useful clues as to the an accident until the elements of the crime are
motive, intent, and identity of the fire setter. Incendiary positively demonstrated by the investigator and prose-
devices can range from the simple flame of a match or cutor. One of the major elements of such proof is the
lighter to the use of accelerants to elaborate time delay exclusion of any reasonably accidental mechanism for
or remote ignition devices. A variety of heat-producing the fire. This means that the investigator must care-
sources can be employed: electric appliances, chemical fully apply the scientific method to all the information
reactions, cigarettes, fuses, even solar energy. Criminal gathered and create and evaluate as many reasonable
behavior analysis has been found useful in solving serial alternative (accidental) hypotheses before concluding
arson crimes, as has geographic mapping or “profiling.” the fire was deliberately set.

Review Questions
1. Define the three elements of the crime of arson. 7. Name four different sources of heat used to
2. Name the six general classifications of motive for ignite incendiary fires.
fire setting as defined by the NCAVC. 8. What is the difference between arson and an
3. Briefly describe three ways in which fraud for incendiary fire?
financial gain can be exhibited in an arson. 9. Describe geographic profiling and its application
4. Into what general categories do targets of to the investigation of serial arsonists.
revenge or retaliation fall? 10. What steps must a good investigator take to
5. What is the current status of irrational (motive- ensure that a sound case has been assembled
less) fire setting? to prove arson?
6. Name three different types of locations for fire
sets and the possible intent behind each.

Court Citations
C1. People v. Andrews, 234 Cal. App. 2d 69; 44, Cal. C4. People v. Lepkojes, 48 Cal. App. 654; 192 P. 160
Rptr. 94 (1965). (1920).
C2. People v. Fanshawe, 137 NY 68; 32 N.E. 1102 C5. People v. Kessler, 62 Cal. App. 2d 817; 145 P. 2d
(1893). 656 (1944).
C3. Pointer v. United States, 151, U.S. 396; 14 S. C6. People v. Saunders, 13 Cal. App. 743: 110 P. 825
Ct. 410 (1894). (1910).
692 Chapter 16 Arson as a Crime
References
1. J. F. Boudreau et al., “Arson and Arson Investigation: 20. R. Estrin, “Juveniles: Nation’s Top Firebugs,” Oakland
Survey and Assessment” (Washington, DC: National Tribune, January 20, 1996, A-4.
Institute of Law Enforcement and Criminal Justice 21. FBI Uniform Crime Reporting Program, Table 32,
(NILECJ), October 1977). Ten-Year Arrest Trends, 1999–2008, Washington,
2. M. J. Karter Jr., “Fire Loss in the United Stated during DC (http://www.fbi.gov/ucr/ucr.htm), 2009.
2008” (Quincy, MA: National Fire Protection 22. FBI Uniform Crime Reporting Program, Tables 39, 40
Association, Fire Analysis & Research Division, and 41, Arrests for Arson by Age and Sex, 2008,
August 2009); see also M. J. Karter Jr., “U.S. Fire Loss Washington, DC (http://www.fbi.gov/ucr/ucr.htm), 2009.
for 2008,” NFPA Journal (September–October, 2009). 23. B. Wood, “Children’s Fire Setting Behaviour,” Fire
3. M. J. Karter Jr., “Fire Loss in the United States during Engineers Journal (November 1995): 31–33.
2003,” NFPA Journal (October 2004). 24. Williams, Understanding the Arsonist.
4. FBI Uniform Crime Reporting Program, Table 2, Arson by 25. Factory Mutual Engineering Corp., Arson: Not Only
Type of Property in the United States, 2008, Washington, for Profit (Norwood, MA: Factory Mutual, 1985);
DC; retrieved from http://www.fbi.gov/ucr/ucr.htm, 2009. Williams, Understanding the Arsonist.
5. Communities and Local Government Fire Statistics 26. Sapp et al., A Motive-Based Offender Analysis.
(UK): 2007 (London); retrieved August 2009 from 27. USFA, “Special Report: Firefighter Arson,” USFA-
www.communities.gov.uk. TR-141 (Washington, DC: U.S. Fire Administration,
6. “Twelve-Month Analysis of £250,000 plus Fires in the January 2003).
UK,” Fire Risk Management Journal (November 2009): 6. 28. J. Wambaugh, Fire Lover (New York: HarperCollins,
7. Office of Deputy Prime Minister, Economic Cost of 2002).
Fire in England and Wales (London: ODPM, 2003). 29. Williams, Understanding the Arsonist.
8. P. G. Jackson, “Assessing the Validity of Official Data 30. Sapp et al., A Motive-Based Offender Analysis.
on Arson,” Criminology 26, no. 1 (1988): 181–95. 31. Ibid.
9. J. F. Decker and B. L. Ottley, Arson Law and Prosecution 32. Williams, Understanding the Arsonist.
(Durham, NC: Carolina Academic Press, 2009), 337–41. 33. Sapp et al., A Motive-Based Offender Analysis.
10. Anon. “Model Arson Law,” Encyclopedia of Crime 34. http://www.justice.gov/crt/church_arson/arson98.php, 2005;
and Justice, vol. 1 (2002); http://law.jrank.org. see also Independent Insurance Institute, iii.org, 2005).
11. For a detailed discussion on this topic, see J. Q. 35. R. D. Walter, “Arson: Who Burns and Destroys with
Whitman, The Origins of Reasonable Doubt: Design?” paper presented at the Tenth Australian
Theological Roots of the Criminal Trial (New Haven, International Forensic Sciences Meeting, Brisbane,
CT: Yale University Press, 2007). Australia, May 1988.
12. A. O. Rider, “The Firesetter: A Psychological Profile,” 36. P. C. Hoffer, Seven Fires (New York: Public Affairs,
FBI Law Enforcement Bulletin 49 (June–July 1980). 2006), chap. 5.
13. NFPA 921: Guide for Fire and Explosion 37. K. McGrattan and C. Bouldin, “Simulating the Fires in
Investigations (Quincy, MA: National Fire Protection the World Trade Center” in Proceedings Interflam 2004
Association, 2008, 2011), sec. 22.4.9.2.1.1, 188. (London: Interscience Communications), 999–1008;
14. A. D. Sapp et al., A Motive-Based Offender Analysis J. Quintiere, “A Predicted Timeline of Failure in the
of Serial Arsonists (Washington DC: Department of WTC Towers” in Proceedings Interflam 2004
Justice, 1994); see also D. J. Icove and M. H. Estepp, (London: Interscience Communications), 1009–22.
“Motive-Based Offender Profiles of Arson and Fire- 38. J. E. Malooly, “Fire as the Terrorists’ Weapon”
Related Crimes,” FBI Law Enforcement Bulletin 56, (Chicago, IL: American Academy of Forensic
no. 4 (April 1987): 17–23. Scientists, February 2003).
15. J. A. Inciardi, “The Adult Firesetter: A Typology,” 39. L. H. Gold, “Psychiatric Profile of the Firesetter,”
Criminology (August 1970): 145–55. Journal of Forensic Sciences 77 (October 1962); Sapp
16. D. J. Icove and M. H. Estepp, “Motive-Based Offender et al., A Motive-Based Offender Analysis; T. G. Huff,
Profiles of Arson and Fire-Related Crimes,” FBI Law G. P. Gary, and D. J. Icove, The Myth of Pyromania
Enforcement Bulletin 56, no. 4 (April 1987): 17–23. (Quantico, VA: National Center for the Analysis of
17. A. D. Sapp et al., “Arsons aboard Naval Ships” Violent Crime, 1997); M. Gardiner, Arson and the
(Washington, DC: National Center for the Analysis of Arsonist: A Need for Further Research, Project Report
Violent Crime, FBI, 1993). (London: Polytechnic of Central London, June 1992).
18. T. G. Huff, “Filicide by Fire: The Worst Crime,” Fire 40. American Psychiatric Association, Diagnostic and
and Arson Investigator (June 1993). Statistical Manual of Mental Disorders (DSM-IV-TR),
19. D. L. Williams, Understanding the Arsonist: From 4th ed. (Arlington, VA: American Psychiatric Publishing,
Assessment to Confession (Tucson, AZ: Lawyers & June 2000).
Judges Publishing, 2006). 41. Huff et al., The Myth of Pyromania.

Chapter 16 Arson as a Crime 693

42. A. O. Rider, “The Firesetter: Part II,” FBI Law International Association of Arson Investigators,
Enforcement Bulletin 49, July 1980. 1983); J. E. Douglas and A. E. Burgess, “Criminal
43. N. D. C. Lewis and H. Yarnell, Pathological Firesetting Profiling: A Viable Investigative Tool against Violent
(Pyromania), Nervous and Mental Disease Monograph Crime,” FBI Law Enforcement Bulletin, December
No. 82, (New York: Coolidge Foundation, 1951); J. James, 1986; D. J. Icove and M. H. Estepp, “Motive-Based
“Psychological Motives for Arson,” Popular Government Offender Profiles of Arson and Fire-Related Crimes,”
March 1965; R. G. Vreeland and M. B. Wilter, Psychology FBI Law Enforcement Bulletin, April 1987; D. J.
of Firesetting: Review and Appraisal (Washington, DC: Icove, “Automated Crime Profiling,” FBI Law
U.S. Government Printing Office, January 1979). Enforcement Bulletin, December 1986.
44. Williams, Understanding the Arsonist. 57. A. D. Sapp et al., A Report of Essential Findings from a
45. J. D. DeHaan, “Laboratory Aspects of Arson,” Fire Study of Serial Arsonists (Quantico, VA: National Center
and Arson Investigator 29, no. 26 (March 1979). for the Analysis of Violent Crime (NCAVC), 1993).
46. J. E. Douglas et al., Crime Classification Manual (New 58. Sapp, A Motive-Based Offender Analysis.
York: Lexington Books, 1991). 59. D. J. Icove, E. C. Escowitz, and T. J. Huff, “The
47. J. D. DeHaan, “The Reconstruction of Fires Involving Geography of Violent Crime: Serial Arsonists,” Agenda
Highly Flammable Hydrocarbon Liquids,” PhD disser- and Abstracts, 8th Annual Geographical Resources
tation, Strathclyde University, Glasgow, 1995. Analysis Support System, (GRASS) Users Conference,
48. P. J. Loscalzo, P. R. DeForest, and J. M. Chao, “A Study Reston, VA, March 1983; K. Fritzon, “An Examination
to Determine the Limit of Detectability of Gasoline of the Relationship between Distance Traveled and
Vapor from Simulated Arson Residues,” Journal of Motivational Aspects of Firesetting Behavior,” Journal of
Forensic Sciences 25, no. 1 (January 1980): 162–67. Environmental Psychology 21 (2001): 45–60.
49. T. E. Folkman et al., “Evaporation Rate of Gasoline 60. D. J. Icove and J. D. DeHaan, Forensic Fire Scene
from Shoes, Clothing Wood, and Carpet Materials and Reconstruction, 2nd ed. (Upper Saddle River, NJ:
Kerosene from Shoes and Clothing,” Canadian Society Pearson/Brady, 2009), 255–59; 262–70.
of Forensic Sciences Journal 23, nos. 2–3 61. D. K. Rossmo, Geographic Profiling (Boca Raton, FL:
(June–September 1990): 49–59. CRC Press, 2000); D. V. Canter and P. Larkin, “The
50. N. Sester et al., “Evaluation of Gasoline Traces Environmental Range of Serial Rapists,” Journal of
Transferred on Hands during Car Tank Filling” in Environmental Psychology 13 (1994): 63–69; D. Canter,
Proceedings European Academy of Forensic Science, Mapping Murder (London: Virgin Books, 2003).
Cracow, Poland, 2000. 62. R. N. Kocsis, “Psychological Profiling of Serial Arson
51. M. Darrer, J. Jacquenet-Papilloud, and O. Delemont, Offenses: An Assessment of Skills and Accuracy,”
“Gasoline on Hands: Preliminary Study on Collection Criminal Justice and Behavior 31, no. 3 (June 2004):
and Persistence,” Forensic Science International 175 341–61; R. N. Kocsis, “Offense Location Patterns:
(2008): 171–78. Geographic Profiles,” Chap. 10 in Criminal Profiling:
52. DeHaan, “Laboratory Aspects of Arson”; V. Principles and Practice (Totowa, NJ: Humana, 2006).
Babrauskas, Ignition Handbook (Issaquah, WA: Fire 63. Crime Point Fire and Arson, Crime Prevention
Science, 2003), 683. Analysis Lab, Burnaby, BC, Canada.
53. U.S. Army, Unconventional Warfare Devices and 64. K. Fritzon, “An Examination of the Relationship
Techniques, TM31-201-1, May 1986. Reprinted as between Distance Traveled and Motivational Aspects
Improvised Incendiary Mixtures by various publishers. of Firesetting Behavior,” Journal of Environmental
54. B. Turvey, Criminal Profiling: An Introduction to Psychology 21 (2001): 45–60.
Behavioral Evidence Analysis, 3rd ed. (New York: 65. B. Martinez, Multiple Fire Setters: The Process of
Academic Press, 2008. R. Girod, Profiling the Criminal Tracking and Identification (Tulsa, OK: PennWell Books,
Mind: Behavioral Science and Criminal Investigative 2002).
Analysis (Bloomington, IN: iUniverse, 2004); J. H. 66. R. R. Hazelwood et al., “Criminal Investigative
Campbell and D. Denevi, Profilers: Leading Investigators Analysis: An Overview,” in Practical Aspects of Rape
Take You Inside the Criminal Mind (Amherst, NY: Investigation: A Multidisciplinary Approach, 2nd ed.,
Prometheus Press, 2004); J. Esherick, Criminal Psychology ed. R. R. Hazelwood and A. W. Burgess, 115–26
and Personality Profiling (Broomall, PA: Mason Crest, (Boca Raton, FL: CRC Press, 1995).
2005); R. N. Kocsis, Criminal Profiling: Principles and 67. ASTM E620: Standard Practice for Reporting
Practice (Totowa, NJ: Humana Press, 2006). Opinions of Technical Experts (West Conshohocken,
55. J. Douglas and M. Olshaker, The Anatomy of Motive PA: ASTM, 2004).
(New York: Scribner, 1999). 68. L. Bilancia, “The Ignition Matrix,” paper presented at
56. P. Horbert, “Profiling the Arsonist,” paper presented California Conference of Arson Investigators Conference,
at the Kansas City Arson Task Force Seminar, Kansas March 2009.
City, Missouri, May 1988; D. J. Icove and M. O. 69. NFPA 921: Process of Elimination (Quincy, MA: National
Soliman, Arson Information Management System: Fire Protection Association, 2008), sec. 18.2, 156; and
User’s Guide and Documentation (St. Louis, MO: (2011), sec. 18.6.5, 174.
694 Chapter 16 Arson as a Crime
 CHAPTER
17
Other Investigative Topics

Used with permission of Retired Major

J. Ron McCardle, Florida State Fire
Marshal’s Office.

KEY TERMS

seat of fire, p. 713

OBJECTIVES

After reading this chapter, you should be able to:

■ Identify specialized tools that are relevant and useful in fire investigation.
■ List common safety hazards encountered at fire scenes.
■ Discuss the use of fire modeling and the questions to be asked when selecting and
using a model.
■ Select a particular fire model and discuss how it meets the various ASTM standards.
■ Discuss the elements of proof necessary in investigating and prosecuting arson
case investigations.
■ Inventory various sources of information available to fire investigators.
■ Explain the concept of spoliation as it applies to fire investigation.
■ Identify the elements necessary to include in a comprehensive fire investigation
report.
■ Discuss the requirements for expert witness testimony, scientific method, and
exhibits.
■ Search the Internet for expert witness requirements in your jurisdiction in both
state and federal court.

695
 For additional review materials, appendices, and suggested
readings, visit www.bradybooks.com and follow the MyBradyKit link
to register for book-specific resources.
Register for MyFireKit by following directions on the MyFireKit
student access card provided with this text. If there is no card, go to
www.bradybooks.com and follow the MyFireKit link to Buy Access
from there.

Safety and Health

It is important that the fire scene investigation be carried out thoroughly, professionally,
and safely. There are many dangers to a fire scene—structural collapse, electricity, toxic
fumes and dusts, water-filled holes, fire-weakened floors and stairs, falling glass, and other
penetration hazards are just a few typical ones. The preliminary scene assessment should
include a survey of the hazards the scene presents, since some of them may have to be con-
trolled before the detailed scene examination can even begin. This is a requirement cited
by NFPA 1033: Standard for Professional Qualifications for Fire Investigator.1 Steps may
have to be taken, for instance, to shore up or even demolish unsupported walls (particu-
larly masonry), roofs, or ceiling structures, remove large glass panels from upper floors,
and pump out standing water. The assessment can be carried out with the assistance of the
Stay Safe fire department safety officer, a county or city building engineer, a chemist with hazardous
materials experience, or even with experienced demolition contractors. The scene in Figure
The preliminary scene 17-1 illustrates multiple hazards typical of fires in older communities: a multistory
assessment should
include a survey of the
masonry building that has been reduced to a facade and partially collapsed interior walls
hazards the scene (all a collapse risk), severe ice and cold weather conditions, a multiuse building with unde-
presents. termined contents, and the usual risks of multiple electrical systems and raw sewage.

FIGURE 17-1 Ice-

covered remains of fire-
damaged multistory
building. Hazards include
structural collapse, ice,
cold, multiple electrical
systems, unknown build-
ing contents, and open
sewer lines. Courtesy of Don
Perkins, Fire Cause Analysis.

696 Chapter 17 Other Investigative Topics

 The investigator should come prepared to deal with the simplest safety problems by
having suitable protective clothing (simple overalls or synthetic fabric suits for hazardous
material exposure), gloves (leather for penetration protection, heavy rubber for chemical
exposure, latex for debris handling), safety helmet with face shield, boots with safety toes
and sole plates, and masks for both dusts and toxic fumes. Dusts (charcoal dust, asbestos,
plaster, etc.) are present at every dry scene and pose significant risks of inhalation. Simple,
disposable nuisance-dust masks should be used at all dry scenes or whenever there is a
risk of exposure to airborne biological pathogens (blood, dried tissue). Mask-type respi-
rators with appropriate filter cartridges can be used at scenes with modest levels of some
toxic fumes, but some scenes may require full self-contained breathing apparatus (SCBA)
such as shown in Figure 17-2.
Carbon-fiber composite materials, once solely found in aircraft components are now
found in a wide variety of automotive applications to reduce weight while increasing
strength. Combustion of such materials produces gaseous and aerosols of toxic pyrolysis
products from the polymers involved as well as airborne microscopic fibers that are small
enough to be inhaled or enter the skin. While the long-term health risks of inhaled fibers
are not clearly defined, fire investigators need to be aware of inhaling any potentially
toxic aerosol chemical and take appropriate steps to minimize their exposure.2
Specialized training is needed to properly deal with biological (sewage, infectious
waste) and chemical (toxics, corrosives) hazards, and the investigator should seek assis-
tance from qualified personnel before risking injury. Some fire departments will insist that
the carbon monoxide (CO) levels be checked in a building before an investigator is
allowed in. Clip-on personal carbon monoxide monitoring devices are useful at “fresh”
scenes when CO may still be accumulating from residual smoldering.3
Electrocution is a constant hazard in both buildings and even outdoors. In older Stay Safe
buildings with multiple occupancies, conversions, floating neutrals, and even “back fed”
power, it is very dangerous to assume that because the overhead service outside the front The investigator is
encouraged to assume
door has been disconnected, there is no danger inside. The investigator is encouraged to electrical lines are still
assume electrical lines are still “hot” until proven otherwise and to use tools such as volt- “hot” until proven
age probes (Figure 17-3) to check individual outlets and connections. otherwise.

FIGURE 17-2 Asbestos was one of the hazards identified in this FIGURE 17-3 Voltage probes help investigators establish whether
large scene. Full “hazmat” precautions were initiated, which required wires or equipment are energized. On left: Non-contact type, battery
all investigative agencies and support services to wear personal powered. On right: Contact-type AC or DC up to 600 V, no batteries
protection equipment. Fire investigators and fire engineers from needed. Courtesy of John D. DeHaan.
the Office of the Fire Marshal were on scene for five days. Courtesy
of Pierre Yelle, Office of the Fire Marshal, Ontario, Canada.

Chapter 17 Other Investigative Topics 697

 Large industrial or commercial sites may have large battery arrays to serve as uninter-
ruptible power supplies (UPSs) in the event of utility failures. A new danger today is the
increasing installation of solar cell power arrays. Such sources are often back-fed into the
line supply via a separate shutoff and can deliver significant electrical power to the system
of either a home, business, or industrial plant. Because that power may not be disabled when
the main breakers are thrown, there is a significant electrocution risk.
Scene examinations should not be carried out alone, and some means of emergency
communications should be at hand (cell phone, radio, or telephone). Severe weather, such
as ice or high winds, even darkness, makes scenes unduly dangerous as well as obscuring
critical evidence and is a justifiable reason for postponing a search. No scene is worth risk-
ing death or serious injury. A comprehensive guide to safety and health issues for fire inves-
tigators has been published by IFSTA.4 The Fire Protection Association (UK) has also
published an excellent guide to safety at fire scenes, and the reader is referred there for
more complete information.5 Safety considerations are also outlined in NFPA 921: Guide
for Fire and Explosion Investigation.6 It should be noted that if any governmental body
declares a fire scene to be a hazardous material site, all responders must follow
Occupational Safety and Health Administration (OSHA) regulations. OSHA has authority
Stay Safe
to enforce stringent safety regulations at a fire scene if employees are involved (including
No scene investigation mandatory annual training),7 and local building authorities in many jurisdictions have the
is worth risking death authority to prohibit examination until a scene can be made safe. The investigator is
or serious injury. encouraged to seek counsel on these matters before launching the examination.

Fire Modeling
Fire modeling can be in the form of bench-scale (miniature) physical models into which
are introduced dyed liquids, smoke, or hot gases to re-create the flow of smoke and heat
or in the form of mathematical analyses. Because the use of physical scale models is rare
in fire investigations, this text focuses on mathematical fire models.

MATHEMATICAL FIRE MODELING

Mathematical fire models help us understand complex fire processes, such as the relation-
ship of heat release rate to other factors, including ventilation and heat of vaporization.
Such methods also can help us relate post-fire indicators to fire events and behaviors and
can help us isolate and analyze unknown factors in fire, such as for example, the effects
of ventilation, ignition location, and fuel type.
Examining these variables with a fire model is much easier and more cost-effective
than constructing and burning life-size models. Further, such modeling is entirely consis-
tent with, and indeed at the heart of the scientific method—creating (or discerning) alter-
native hypothesis and testing them.
Fire modeling also helps satisfy the court’s demand that fire investigation, analysis,
and conclusions follow and utilize the scientific method. Mathematical fire modeling is at
least one way the “error rate”—or at least the effect that unknown or unknowable vari-
ables may have on estimations of fire processes such as size, rate of growth, likely path of
fire travel or spread, the likelihood of flashover, and other variables—can be measured
and explored and more effectively explained to the court and/or trier of fact.
Mathematical fire modeling can be as simple as using the algebraic equations for
flame height, radiative heat loss, heat transfer, and others (some of which were discussed
in this book). These simple calculations can be done on a pocket calculator with scientific
functions. Such calculations have also been codified in software packages such as
FPETOOL (developed at NIST) and ASKFRS (developed at the Fire Research Station,
UK) or Fire Dynamics Tools (NRC). They are PC compatible and available from those
agencies for download or purchase.
698 Chapter 17 Other Investigative Topics
 Since the late 1980s dozens of computer models have been introduced as a way of
understanding the fire environment. A survey by Factory Mutual Research in 1991 revealed
that there were 31 computer programs available to study fires in compartments and another
31 to predict special behaviors like smoke and fire travel, sprinkler activation, or evacua-
tion.8 Today, there are many more. (The reader is referred to www.firemodelsurvey.com for
current models.) Some of these models will run on a personal computer, others require a
mainframe or even a supercomputer.

ZONE MODELS
Many fire models in use today are zone models; that is, they predict behavior based on
movement of gases in the two large zones normally found in a developing compartment
fire: a hot gas layer (ceiling) and a normal air layer (floor). ASET-BX (Available Safe
Egress Time-Basic), FPETOOL, and CFAST (Consolidated Fire and Smoke Transport)
or FASTlite are some of the most commonly used zone models.9 (A more comprehen-
sive set of programs, HAZARD-I, is also occasionally used.) These models calculate
the temperature and depth (height) of the hot smoke layer interface in rooms (some-
times in a series of rooms) and are useful in assessing fire and smoke spread, especially
where occupant escape is a factor. The predicted temperature of the hot smoke layer
is used to predict when a room is untenable and when it is likely to progress to
flashover. CFAST has been recently upgraded by NIST (Version 6.1.1) to become a
very powerful and well-validated model. Many programs are available for free down-
load from the Internet.

FIELD MODELS
A few models like JASMINE (Analysis of Smoke Movement in Enclosures) or SOFIE
(Simulation of Fire in Enclosures) are field models, which look at the energy and mass
transfers occurring in numerous very small volumes of space in a room.10 The comput-
ing power needed to run these field models (sometimes called CFD, or Computational
Fluid Dynamics, models) often used to limit them to mainframe computers, but with
today’s powerful personal computers, they are more widely used. The Center for Fire
Research (now the Building and Fire Research Laboratories) at NIST developed a very
advanced CFD field model called the Fire Dynamics Simulator (FDS). FDS uses the
thermal and fire behavior properties of furnishings, ceiling and wall linings, and floor
coverings to calculate fire growth and spread. A second NIST-developed program,
called Smokeview, is a visualization program that displays the results of an FDS model
as either snapshots or as two- or three-dimensional animations.11 (See Figures 17-4 and
17-5 for examples.) Some newer zone models have been developed such as BRANZ-
FIRE, which predicts and evaluates performance and hazards of complex environments
such as combustible room linings, and FRS’s CRISP (Computation of Risk Indices by
Simulation Procedures), which evaluates multiroom fire scenes using a two-layer zone
model but incorporates complex models of human behavior and movement.12 Many of
these programs are valid only up to flashover because the massive turbulence created
by full involvement overwhelms the zone models. Some cannot deal with multiple
rooms or changes in ventilation, so users must be aware of limitations that may make
some programs unsuitable for studying some fires.13
Fire modeling has been used by NIST in numerous forensic engineering analyses and
reconstructions since the mid-1980s. For a more detailed review of these cases, the reader is
referred to Table 6-5 in Chapter 6, “Fire Modeling,” in Forensic Fire Scene Reconstruction.14
This textbook also addresses the other uses of fire modeling, such as in detector activa-
tion, witness observations, and the determination of the origins and causes of fires by
comparing visual fire pattern damage with predicted temperature and heat flux to
exposed compartment surfaces.
Chapter 17 Other Investigative Topics 699
FIGURE 17-4 FDS prediction of the temperature zones of The FIGURE 17-5 Smokeview (FDS) prediction of the fire and smoke
Station nightclub 90 seconds after ignition of combustible foam conditions inside The Station 60 seconds after ignition. (Top)
insulation on stage. (Top) Without sprinklers (prediction compared Without sprinklers. (Lower) With sprinklers. Courtesy of NIST.
against video recording of actual fire). (Lower) Same scenario with
normal sprinklers functioning. Courtesy of NIST.

MODELS FOR SPECIALIZED APPLICATIONS

Models for specialized applications include the following:
■ Post-flashover COMPF, SFIRE-4 (time–temperature history for energy, mass and
species—evaluating structural integrity in fire exposure in post-flashover rooms)
■ Thermal and structural response FIRES-T3, HEATING7, TASEF (finite element
calculations about structural fire endurance of a building or particular components
using known failure conditions such as loss of tensile strength as a function of
temperature)
■ Fire protection DETACT-QS, DETACT-T2, LAVENT (sprinkler and detector
response times for specific fires)
■ Smoke movement CONTAM96, Airnet, MFIRE (dispersion of smoke and gaseous
species)
■ Egress lsafe, buildingEXODUS, EESCAPE, ELVAC, EVACS, EXITT, EXIT 89,
EVACNET, WAYOUT, SIMULEX [probabilistic (stochastic) modeling of escape of
people from fire using predicted smoke conditions, and occupant and egress variables)]
700 Chapter 17 Other Investigative Topics
■ Glass breakage BREAK1
■ Wildland BEHAVE Plus, Prometheus, FARSITE

ASTM AND CRITICAL MODELING ISSUES

It is particularly important to point out to potential users of fire models that there are
existing ASTM guides that deal with critical modeling issues. Those most relevant to our
discussion here are the following:
■ ASTM E1355-05a: Standard Guide for Evaluating the Predictive Capability of
Deterministic Fire Models
■ ASTM E1591-07: Standard Guide for Obtaining Data for Deterministic Fire
Models
■ ASTM E1895-07: Standard Guide for Determining Uses and Limitations of
Deterministic Fire Models
■ ASTM E1472-07: Guide for Documenting Computer Software for Fire Models
All these ASTM standards contain vital information and recommendations for some-
one contemplating the use of a fire model. As they are published by ASTM International
(and developed by ASTM Committee E-05 on Fire Standards), they represent peer-reviewed
guides with which all fire model users should be familiar. They outline the documentation
(of both the scene and the model) necessary for demonstrating the reliability of models
used. They are intended for evaluation of zone models but are applicable to field models
as well.

WHAT SHOULD WE ASK ABOUT ANY MODEL WE USE?

Fire investigators should be able to answer the following questions before using any model,
by applying the criteria in the ASTM guides and evaluating the case at hand:
■ Is it applicable?
■ Is it the right tool for the job?
■ Does it give accurate results?
■ How often does it predict events that do not occur in real fires?
■ How sensitive is it to changes in input?
■ What is its error rate?
■ Has it been used to predict events in real fire tests?
■ Has it been validated?
■ Where did it come from?
■ Where was it published?
■ What supporting (or contradictory) data have been published?
If we use such models, we quickly realize that the input data they depend on are usu-
ally significantly more extensive than investigators are used to gathering. This paucity of
data is more often than not the result of careless or incomplete documentation of the
scene. This problem has recently been addressed by the inclusion of recommended data
collection forms in Appendix J of this text and in Forensic Fire Scene Reconstruction,15
and NFPA 921. It is a rare scene that is so completely destroyed that basic dimensions,
structural and finish materials, and furnishings (type and placement) cannot be estab-
lished by careful examination. Even in instances where there has been massive destruc-
tion, interviews, examination of nearby “exemplar” structures, or recovery of pre-fire
photos or videos, can often fill in many of the missing pieces.
NIST gives the following warning to users of fire models, particularly their Fire
Dynamics Simulator (FDS):
Users are warned that FDS is intended for use only by those competent in the fields of fluid
dynamics, thermodynamics, heat transfer, combustion, and fire science, and is intended
Chapter 17 Other Investigative Topics 701
 only to supplement the informed judgment of the qualified user. The software package is
a computer model that may or may not have predictive capability when applied to a spe-
cific set of factual circumstances. Lack of accurate predictions by the model could lead to
erroneous conclusions with regard to fire safety. All results should be evaluated by an
informed user.16

FIRE ASSESSMENT
As outlined in ASTM E1895-07, for instance, the user’s first step should be to define the
scope of the fire assessment and then determine whether fire modeling is an appropriate
tool.17 Then, the user should determine which models are available and suitable to run
on the available computer hardware considering the size and complexity of the problem.
For the models being considered, the available documentation should be acquired and
evaluated in terms of guidance offered in ASTM E1472-07: Guide for Documenting
Computer Software for Fire Models.18 The limitations of the candidate models must be
compared with the problem to be solved—one-room versus multiroom, pre-flashover versus
post-flashover.
Although it is possible to modify existing models to deal with particular problems,
any modifications must be made in cooperation with the original model developer and
then subjected to suitable validation as outlined in ASTM E1355-05a: Standard Guide for
Evaluating the Predictive Capability of Deterministic Fire Models.19 Other tools, such as
small- or large-scale fire tests or mathematical calculations, should be considered as well.
Once a model is selected, the following steps are recommended:
1. Verify the known limitations of the model—room dimensions, fire size, or ventilation.
2. Determine the underlying assumptions (two-layer zone or CFD/field model) and
assess their impact on the results.
3. Determine the characteristic variables.
4. Determine what input data is required and where it can be obtained.
5. Determine the rigor of the mathematics involved and check to see whether it will
give an answer given the constraints of the problem.
6. Determine the extent of validation to establish its appropriateness for the problem.
Validation processes are described in ASTM E1355.
7. If validation data are not available, conduct a sensitivity analysis to establish the
effect of changing critical variables.
8. Thoroughly document the model “run,” including all input data, all assumptions
made, and any and all modifications (including validation to support the accuracy
of those modifications).

DOCUMENTATION
The documentation for a computer fire model should include a technical guide or user’s
manual (as described in E1472). The source code for the model should also be made
available to any potential user. Some well-known programs, such as FPETOOL and FDS,
are available for downloading from NIST at no charge. Other programs, such as BRANZ-
FIRE, SIMULEX and ASKFRS, must be purchased as a “user license” for a given period
of time. The assumptions used, known numerical and physical limitations, and the phys-
ical and mathematical treatments used must also be made available to the user. An excel-
lent source of current information is the www.firemodelsurvey.com website, where
information about more than 150 models is available. It describes the abilities and limi-
tations of current fire models. A summary of this information was published by Olenick
and Carpenter.20
A good example of a documentation guide for a fire modeling package is FPETOOL:
Fire Protection Engineering Tools for Hazard Estimation, by H. E. Nelson.21 FPETOOL
is a collection of analytical tools about fire behavior and properties with simplifications
702 Chapter 17 Other Investigative Topics
(as assumptions) to make approximations rather than exact predictions using portable or
desktop computers. It consists of the following three main elements:
■ Fireform (Fire Formulas) A collection of fire safety calculations
■ Makefire A series of procedures to produce fire input data files for use with Fire
Simulator
■ Fire Sim An integrated set of equations (i.e., a model) designed to allow the user
to create a fire case study in a Lotus spreadsheet format with specifications of room
and vent dimensions; fuel characteristics; and ceiling, wall and floor materials; and
input fire to predict layer temperature, flashover, and tenability factors
The FPETOOL guide describes the main elements of the package, its hardware and soft-
ware requirements, mathematics, underlying assumptions, and comparisons of FPETOOL
(Fire Simulator routine) with fire test data. A separate user guide was published as FPETOOL
User’s Guide.22
The FIRM-QB zone model developed by Marc Janssens has a technical description,
program description, and user’s manual all included in his excellent book, An Introduction
to Mathematical Fire Modeling.23 The entire program code and supporting documentation
are included in a CD-ROM packaged with the book.

MODEL EVALUATION
According to ASTM E1355, the evaluation process consists of four steps:
1. Define the scenarios for which the evaluation is to be conducted.
2. Validate the theoretical basis and assumptions used in the model.
3. Verify the mathematical and numerical robustness of the model.
4. Evaluate/quantify the uncertainty and accuracy.
ASTM E1355 offers the following definitions regarding models.
Evaluation The process of quantifying the accuracy of chosen results from a model
when applied for a specific use.
Validation The process of determining the correctness of the assumptions and
governing equations implemented in a model when applied to the entire class
of problems addressed by the model.
Verification The process of determining the correctness of the solution of a system
of governing equations in a model. Verification does not imply the solution of
the correct set of governing equations, only that the given set of equations is
solved correctly.
The steps identified in ASTM E1355 are not isolated ones. As Janssens points out:
Step 4 is usually based on a comparison between model output and experimental data and
provides an indirect method for validation (Step 2) and verification (Step 3) of a model for
scenarios of interest (Step 1). It is generally assumed that the model equations are solved
correctly and the terms validation and evaluation are therefore often used interchangeably.
It is very rare for anyone but the model’s developer to spend the time necessary to carry
out steps 2 and 3, although an independent reviewer or researcher may. Portions of the
mathematics are sometimes compared to other analytical results. Step 4 can be carried out
by comparing a model’s predictive results to full-scale tests done specifically for that pur-
pose, tests done by others and published in the literature, results of standard room fire
tests done in accordance with ASTM E603-07, or even against observations or reconstruc-
tions of real fires (historical fire data).24

A good example of a comparison of a compartment fire test against the predictions

of mathematical calculations, zone models, and a field model was published by
Spearpoint, Mowrer, and McGrattan in 1999.25 It showed how accurately calculations
Chapter 17 Other Investigative Topics 703
 and zone models (Fire Simulator, FAST, and FIRST) agreed with test data. The fire model
used was ES3D (a predecessor to FDS). Its predictions (including early fire development)
were not as accurate as the calculations and indicated further development was needed.
In a more recent study conducted by Rein, Bar-Ilan, Fernandez-Pello, and Alvares in
2006, the researchers compared three models for the simulation of three accidental fire
scenarios in a single-family dwelling.26 The modeling approaches used were an analytical
model of fire growth, the CFAST zone model, and the FDS field model. The research
found that in spite of the differences in these three modeling approaches, the results were
found in relatively good agreement, particularly in the early stages of the fire.
Model uncertainty is based on repeated runs of similar data and sensitivity analyses
to identify critical data. For a complex model with many inputs, it is usually prohibitively
time-consuming to run repeated runs with different single inputs. Mathematical tech-
niques have been used to reduce the computer time needed for each run (such as linking
multiple personal computers to conduct parallel processing).
The accuracy with which FDS predicts temperatures and heat release rates has been
validated by large-scale fire tests. Testing has shown that FDS temperature predictions
were within 15 percent of the measured temperatures, and heat release rates were within
20 percent of measured values. Results, however, are often presented as ranges to account
for some uncertainty. FDS (Version 5) is the primary fire-modeling tool used by NIST
today and has been used in major fire investigations involving large losses and death.27
When model results are compared with a full-scale test fire, it is usually assumed that
the real-fire data are the gold standard. The complexities of the room fire environment
and the variables of turbulence make it impossible, however, to get exactly the same meas-
urements when the real fire is repeated. Estimates of the uncertainties inherent in some
measurements can be as high as ⫾30 percent.28
For example, work by Beard evaluating the Dalmarnock Tests in 2006 looked at
modeling a test fire started in a sofa in a two-bedroom apartment in a high-rise residen-
tial complex.29 Fire modeling users were given the details of the apartment, arrangement,
and materials but not the experimental results. The users were asked to predict the time
courses of variables such as the heat release rate, temperatures, and smoke obscuration.
Of the ten modelers, two used the CFAST zone model and eight selected the FDS field
model. Beard noted that generally the predictions fared poorly in comparison with the
actual experimental results.
However, recent work sponsored in part by the NRC examined the verification and
validation (V & V) of fire models, including CFAST and FDS.30 Although the NRC’s
study centered on fire hazards particular to nuclear power plants, it addressed many of
the concerns raised by those who think fire models are not accurate or appropriate for
forensic fire scene reconstruction. These concerns include the ability of the models to
accurately predict common features of fires, such as upper-layer temperatures and heat
fluxes. One of the features of this report was a comparison between actual fire test results
and predictions of hand calculations, zone models, and field models. As shown in Figure
17-6, there is generally good agreement among all the models and the variability of real-
world fires.
Experienced fire scientists realize the limitation of applying models to real-world fire
investigations. These models, in general, are designed to start with the ignition of a fire
under preset conditions and predict the time factors and conditions of growth and, some-
times, decay. They are not designed to re-create a particular fire by working backward
from a set of final observations to determine what the starting (or even intermediate) con-
ditions were. Models can be used for field investigations, but the process begins with what
amounts to an initial guess about the starting conditions (a scenario), and then variants
of that scenario (created by changing the input data) are analyzed until the results com-
pare favorably with the post-fire conditions observed. Application of a model results in
one set of conditions that could have produced the observed results but not necessarily
704 Chapter 17 Other Investigative Topics
 400
+13 %

Predicted HGL temperature rise (˚C)

300 -13 %

200

100 CFD model

Zone models
Hand calculation methods

0
0 100 200 300 400
Measured HGL temperature rise (˚C)

FIGURE 17-6 Comparison of hot gas layer temperatures measured in full-scale tests compared with predic-
tions of hand calculations (n), zone models (•), and FDS models (♦). A predicted ⫾13 percent variability range
is included. Note that the hand calculations tended to overpredict layer temperatures, whereas both zone and
field model predictions were generally within variability limits. Adapted from M. H. Salley, J. Dreisbach, K. Hill, R.
Kassawara, B. Najafi, F. Joglar, A. Hammins, K. McGrattan, R. Peacock, and B. Gautier. “Verification and Validation: How to
Determine the Accuracy of Fire Models.” Fire Protection Engineering (Spring 2007): 34–44.

the actual set of conditions that did exist. The investigator must remember that a model
is just that—a model—and not proof of how a particular fire happened. It is essential to
understand that any model, whether it is a simple calculation (such as the Fire Dynamics
Tools spreadsheets31), a simplified model with a few values to input (like FPETOOL or
ASKFRS), a predictive model of general fire behavior in a room, or a complex zone model
of fire spreading in multiple compartments requiring extensive data, if given faulty data
will give faulty results. Note that some of the more advanced models will accommodate
a change in ventilation (such as a door opening or a window breaking (note that this
means not simply cracking or “failing” but breaking out so as to materially affect the ven-
tilation conditions in the room).
Any of these models can be forced to give misleading results to fit the post-fire
appearance by manipulating the input values. This manipulation may not necessarily be
intentional but may be due to lack of knowledge and skill on the part of the user. Some
of the models make assumptions about the size, shape, ventilation, and other features
about the room to give a general prediction. Many will ask for specific data about the
room: its dimensions, its ventilation, contents, building materials, and other factors. The
investigator must be prepared to provide these data, which are rarely collected during a
scene investigation. A data collection form (a sample is shown in Appendix J) will help
the investigator gather critical data for later analysis and reconstruction, even if modeling
is not carried out. If these data are not supplied, many programs will assign a default
value, and the user must then be fully aware of what the default values are, what they
mean, and what effect they will have on the prediction if the actual fire scene values
are different.
Many of the simpler models cannot accommodate changes during fire development.
Because one rarely investigates a fire whose ventilation conditions do not change during
the fire (windows failing, doors opening), such models are of limited use in actually pre-
dicting the final result. Very few of the computer models will accurately predict the post-
flashover behavior of a fire because of the extremely complex behavior of the turbulent
gases, and the user must be cautious not to rely on predictions past flashover (even if the
Chapter 17 Other Investigative Topics 705
 computer model is willing to offer them!). Some of the programs have a built-in warning
that advises the user that conditions are likely to produce flashover and requests specific
acknowledgment to proceed. Few of the models will work on smoldering fires, even
though they are sometimes the most dangerous because of the toxic gases generated. Some
of the models require the user to estimate the size (HRR) of the initial fuel package. This
applies only to open-flame progression, not smoldering initial fires (in furniture, for
example). Fortunately a great deal of data are available regarding the heat release curves
for a variety of furniture and other first-fuel packages.32 Some models will not accommo-
date unusually intense fires (more than 2 or 3 MW), or ultra-fast-developing (accelerated)
fires as input. A few will default to a “standard time–temperature curve” model for the input
fire. This default is acceptable when one is trying to evaluate the influence of some factor
later in the fire, but very few real-room fires behave as a smoothly increasing curve.
Instead, they are a series of steps, with each step representing the HRR of another fuel
package as it ignites, decaying as that package burns away. As we saw in Chapter 3, if the
steps are large enough or occur quickly enough, the fire will grow to full involvement.

TESTING COMPLEX COMPUTER MODELS

One of the major differences between zone and field models is that field models often
include routines that calculate the growing fire based on first principles of thermal response,
heat flux, and flame spread. Of course, this requires that the initial and secondary fuels be
identified and their physical and thermal properties carefully defined as input data.
The limitations of this text are such that detailed descriptions of complex fire models
cannot be included here. The reader is referred to the references included here or to the
www.firemodelsurvey.com website. The selection of which model to use requires the ana-
lyst to ask the following questions about the model:
General
■ Appropriateness—is the model’s output useful and applicable?
■ Limitations (time, ventilation, output)
■ Resources—computer speed and capacity needed
■ Experience needed
■ Input data needed—what default conditions apply if input data are incomplete and
how do those presumed defaults affect the results?
■ Sensitivity—what happens to output when input data are changed?
Accuracy
■ Are the results/outputs realistic? Would events occur that way in a real fire?
■ Has the model been used to predict the outcome of a test burn—such as tenability,
temperatures, or time to flashover? How accurately did it predict the actual fire results?
■ Was the model “fine-tuned” to make its predictions more accurate? (Was the program
run multiple times with different data to slightly “tweak” the result?)
■ How were the data collected in the test burn? Direct observation, thermocouple
measurements, radiometers?
■ Where were the measurement/observation point(s)?
As noted previously, the fire environment can be so complex that temperature or
radiometric data collected in one location may not be representative of the entire room,
leading to possible variations of as much as ⫾30 percent.
Reproducibility
■ If the program is run with the same data by the same person, does it give the same
answer?
■ If the program is run with the same data by a different person, does it give the same
answer?
706 Chapter 17 Other Investigative Topics
 Robustness
■ Is the program applicable to different situations? Has it been tested and demon-
strated to give accurate results if starting conditions are very different?
■ Has it been shown to give reliable results for:
■ Small fire in a big room versus large fire in a small room?
■ Adequate ventilation versus underventilated?
■ “Ultra-fast” t2 fire versus “medium” t2 fire?
■ Was the testing done by “game playing” or thorough testing of alternative hypotheses?
Analysis
■ What features of fuel, starting conditions, initial fire size are required input data?
■ What assumptions have been made by the analyst based on personal judgment or
“bias”?
■ How do such assumptions affect the prediction?
Generally, if the models are used by knowledgeable users with good input data, they
are capable of assessing the probable events for a fire scenario. Although precise predic-
tions are not the intended end product of most fire modeling programs, they can be use-
ful for making general estimates about a given fire and for evaluating the effects of
what-if hypotheses.33 Very often in fire investigations, the time involved is the most critical
factor, and yet it is one of the most elusive. Models offer an opportunity to change one
variable at a time in a particular fire and see what effect it may have on time of develop-
ment. Evaluations currently being carried out to compare the results of computer model
predictions with real-world fires have demonstrated generally good agreement with full-
scale lab tests.

Critical Analysis of Cases

To ensure the appropriate application of computer modeling to fire scenario hypotheses
can best be approached using critical analysis or critical thinking, a field of study popu-
larized by Dr. Richard Paul and Dr. Linda Elder.34 Critical thinking improves the quality
of the thought process by providing structure and intellectual standards, thus enabling the
person conducting the fire model to make appropriate and rational decisions in correctly
choosing, applying, and interpreting the results.
Paul and Elder suggest that eight elements of reasoning be applied to any scientific
thinking process. As with the application of the scientific method’s iterative process, each
element assists in eliciting inferences that assist in generating or considering new evidence
necessary to answer questions or solve problems. The elements are summarized next with
illustrations of their application to fire modeling.
1. Purpose: Scientific reasoning has a purpose to its approach. State the purpose
of the chosen fire model, distinguish the basis of its selection, and periodically check to
ensure that it indeed was the best choice. (Example: The choice of a field model over a
zone model to evaluate smoke distribution and toxicity.)
2. Questions: Scientific reasoning attempts to either answer questions or
solve problems. Clearly and precisely state the question to be answered by the
model, break the question into subquestions, and determine whether the question
has one right answer or requires additional reasoning taking into consideration
other hypotheses or points of view. (Example: Determination of tenability based
upon subproblems of smoke, heat, and concentrations of noxious by-products of
combustion.)
3. Information: Scientific reasoning is based upon data, information, and evidence.
State and document what information or data are being used to feed the model, determine
whether sufficient data are available, and describe what experience you have had to
Chapter 17 Other Investigative Topics 707
 validate these data. Make sure you also search for data that oppose your opinion, and
develop alternative hypotheses. (Example: The selection of heat release rate curves based
upon the availability of past historical testing versus data generated through actual test
burns using similar materials.)
4. Inferences/Conclusions: Scientific reasoning uses inferences or interpretations
to draw conclusions and thus give new meaning to scientific data. Clearly describe
how and by what assumptions the conclusions were reached, check that the inferences
are consistent, and determine if there is another way to interpret the information or
arrive at the same conclusion. (Example: Comparison of burn patterns based upon
visual observations with patterns predicted from field modeling.)
5. Concepts: Scientific reasoning is expressed and shaped by concepts and theories.
Identify and clearly list the key scientific concepts that underpin your hypothesis and
model. Recheck that the scientific precision is bounded, and consider alternative concepts.
(Example: The knowledge of smoke distribution theory along with ample scientific testing,
on-scene observations, and witnessed tests.)
6. Assumptions: Scientific reasoning is often based upon assumptions. Clearly and
precisely identify your assumptions, and assess whether they are justifiable. Be aware
of how your assumptions shape your point of view. (Example: The positions of doors,
windows, and other paths of leakage out of a compartment in calculations of the minimum
heat release rate for flashover.)
7. Implications/Consequences: Scientific reasoning has both implications and con-
sequences that follow from data and reasoning. Consider all possible implications and
consequences, trace the reasoning paths, and search for both negative and positive
implications. (Example: The use of three separate fire models to test the hypothesis of
the actual activation time of a smoke detector.)
8. Points of View: Scientific reasoning is done from a chosen point of view.
Consider what your point of view is and whether it has scientific footings. Remember to
identify and consider other scientific points of view along with their respective strengths
and weaknesses. (Example: The determination that additional fuels were needed to cause
flashover in a near-vacant compartment and not taking into consideration fuels contained
in the walls, flooring, and ceiling.)
Examples of the use of critical thinking are well-founded in other fields.35 Its use in assessing,
choosing, and evaluating fire models and their results can clearly enhance the results of
their application.

Search and Seizure

Evidence to be used in fire investigation is subject to laws of search and seizure.
Evidence that was unlawfully seized cannot be used in court. By extension, any infor-
mation developed as a result of an illegal search must also be held inadmissible. This
principle is called the exclusionary rule, and it controls all evidence to be used in crim-
inal prosecutions. The legality of a search depends on the court’s interpretation of how
a fundamental principle of the U.S. Constitution, expressed in the Fourth Amendment,
applies to the case:
The right of the people to be secure in their persons, houses, papers, and effects against
unreasonable searches and seizures shall not be violated, and no warrants shall issue, but
upon probable cause, supported by oath or affirmation, and particularly describing the
place to be searched, and the persons or things to be seized.

It is essential that every investigator appreciate the requirements for lawful searches
of fire-damaged property and abide by them to ensure that the evidence gathered, ana-
lyzed, and presented will be held valid by the court. It was previously mentioned that it
is necessary for the fire service to maintain control and custody of the scene until the
708 Chapter 17 Other Investigative Topics
investigation is completed. In light of recent Supreme Court decisions, there are several
avenues to a permissible search:
■ The fire department has a public obligation to investigate all fires (not only to
identify arson but to prevent accidental fires). The investigation should begin promptly
when the fire is actually extinguished, but it need not be carried out immediately if con-
ditions such as darkness, smoke, or fumes would prevent a full examination. Its start
may be delayed for a reasonable amount of time (and be carried out for a reasonable
time) after extinguishment as long as custody and security of the scene are maintained.
The mere custody of the scene by a fire department, however, does not ensure that a
court will not find a delayed start search to be in violation of the Fourth Amendment.
■ The investigation can be carried out immediately in the absence of a search warrant if
there is a high probability that evidence of its causation will be lost (due to flames, water, or
deliberate destruction or removal) if the search is delayed (so-called exigent circumstances).
■ A signed consent can be obtained if the person legally responsible for the property
can be located. A prepared consent form such as that shown in Figure 17-7 is suggested.

FIGURE 17-7 Consent

to Search form. Courtesy
of Chicago Fire Department,
Office of Fire Investigations.

Chapter 17 Other Investigative Topics 709

 Permission must be given voluntarily; any coercion used may invalidate the search and its
evidence. Consent searches can be thought of as limited by the owner’s right to refuse, the
right to restrict the area being searched, and the right to withdraw permission at any time
during the search. Failure to recognize these limitations may make a “consent to search”
invalid (or at least restrict the admissibility of evidence seized).
■ It is permissible to search a property that has been abandoned by the owner, how-
ever, this circumstance may be virtually impossible to prove and is therefore rarely used.
■ Evidence may also be seized legally if it is in open view (from a point where the
public would have legal access), if it is in plain view of a police officer carrying out his or
her normal duties (including hot pursuit of an individual into his or her home), or if it is in
an area under the immediate control of, or accessed by, a person being placed under arrest.
■ In many jurisdictions, administrative or inspection warrants may be issued by
local authorities. Such a warrant is an order, in writing and signed by a judge, directed
to a state or local official, ordering that an inspection required or authorized by law or
regulations be carried out. An inspection warrant is usually issued only on cause sup-
ported by an affidavit, particularly describing the place to be searched and its purpose.
It is usually issued for a limited period of time, and the ordered inspection may be pro-
hibited during certain hours or without advance notice. In some cases, such a warrant is
considered to be purely administrative in intent, and if clear evidence of a criminal ori-
gin is seen in plain view as a result of such a search, that evidence may be seized, but a
criminal search warrant must be obtained before the search can proceed any further.
■ A full criminal search warrant may be sought. It is an order in writing, signed by
a magistrate, that directs a peace officer to search for particular property and bring it
before the court. In most jurisdictions, a search warrant requires that cause be shown,
supported by an affidavit of facts in hand. The affidavit must describe the person, prop-
erty, and place to be searched and include a description of the evidence to be sought. A
full description of search warrant procedures is offered in the literature.36
A search warrant should be obtained if the owner or his or her designated representative
has made significant efforts to secure the property (locking gates, erecting fences, board-
ing windows, etc.), all of which re-establishes a reasonable expectation of privacy. Although
insurance investigators have certain rights to examine property on behalf of the insurer,
an attempt to use the investigator as an agent of the police or fire authorities will invoke
Fourth Amendment protection.

SEARCH AND SEIZURE COURT DECISIONS

A series of landmark cases from the period 1967–2000 affect the admissibility of evidence
recovered at fire scenes. A brief commentary on each is offered; full details should be
obtained from the case citation.
Camara v. Municipal Court of San FranciscoC1*
In this case, a housing authority inspector conducting an annual safety inspection of a liv-
ing area was refused entry by the tenant. The California Supreme Court held that a war-
rant was required for entry no matter what statutory authority was present. An
“inspection warrant” that functions as an administrative search warrant can be issued
under such circumstances. The probable cause for the issuance of such a warrant can
include elapsed time since last inspection, a preset schedule of such inspections, or condi-
tions visible from the exterior of the dwelling.
See v. City of SeattleC2
A fire chief performing a routine fire inspection demanded entry to a commercial ware-
house. The chief was denied entry by the owner, who claimed this was an invasion of his
*
In this chapter, citation numbers that appear preceded by a superscript C refer to legal references that are listed
at the end of the chapter under “Court Citations.”

710 Chapter 17 Other Investigative Topics

expected rights of privacy. The Supreme Court held that due to considerations of public
safety, different standards were expected of commercial areas and that inspections prior
to or commensurate with licensing or granting a permit were authorized.
United States v. Biswell C3
A sporting goods/firearms dealer gave consent for a search that ultimately revealed the pres-
ence of sawed-off illegal weapons. He claimed that the consent search was not legitimate;
the U.S. Supreme Court held that the firearms business is closely regulated by the govern-
ment for the purposes of public safety, and a business conducting firearms trade must expect
such inspections, and that consent in this case was submitting to a lawful authority.
Marshall v. Barlow’s, Inc.C4
An OSHA inspector doing an inspection by statutory authority in Pocatello, Idaho, was
denied entry by a plant manager. The manager subsequently obtained an injunction pro-
hibiting reentry of the OSHA inspector on the grounds that he had certain expected stan-
dards of privacy. The Supreme Court held that a plumbing and electrical business such as
his was not closely regulated by the government and that his expected standards of pri-
vacy were higher than those for the firearms or liquor trades.
Michigan v. Tyler C5
The landmark 1978 decision that was expected to have the greatest impact on arson cases
is Michigan v. Tyler. The order of events can be summarized as follows: 2400 hours—fire
in progress; 0200 hours—fire chief arrives and is given containers of suspected flammables
recovered from the interior of the dwelling by the fire suppression crews; 0400—the fire
was extinguished. An attempt was made at 0400 hours to enter the building; however, due
to heat, smoke, and darkness, no suitable search could be carried out. At that time, all fire and
police personnel were secured from the scene. At 0800 hours the fire chief returned with a
fire inspector who examined the building with no results. At 0900 hours an assistant chief
and a police investigative team reexamined the building and found evidence of an arson
trailer. At that time, the Michigan State Police Arson Investigation Team was requested.
That team arrived later that day for examination, photography, and collection of evidence
at the scene during several subsequent searches. The U.S. Supreme Court held that the
returns to the crime scene at 0800 and 0900 were permissible, since they were essentially
a continuation of the examination from 0400 (which had to be discontinued due to poor
conditions). The evidence gathered by the Michigan State Police Arson Team, however,
was suppressed. The court held that the prosecution’s contention that all rights of privacy
by the owner were given up by committing a crime, that is, destroying that property, was
not substantiable. However, an investigation begun and continued for a reasonable amount
of time for the benefit of public safety at the scene was permissible. Prior to an extended
examination or repeated examinations over a long period of time, a warrant was required
to notify the owner of the intended search. The definition of “reasonable period of time”
is entirely flexible, depending on the type of building involved, the time and nature of
suppression efforts, the complexity of the scene, and the extent of emergency situations.
The Court reinforced the statement that it is the duty of fire officials to determine the cause
of fires. The circumstances defining “a reasonable length of time” for an investigator to
conduct a warrantless search must be considered by the investigator at the time of the inci-
dent. If there is any doubt, obtain a warrant.
Michigan v. Clifford C6
This 1984 case refined several of the issues not clearly defined in Tyler. The private resi-
dence involved was heavily damaged by fire during the night while unoccupied. At the
conclusion of firefighting, all fire and police personnel left the scene. A team of arson
investigators arrived at the scene some 5 hours later. In the interim, the property owners
had retained salvage crews to pump water from the basement and board up windows.
These efforts were underway when the investigators began the investigation without
Chapter 17 Other Investigative Topics 711
 consent or warrants. A suspected point of origin and evidence of an incendiary device
were found in the basement. The investigators then proceeded to search the remainder of
the house, finding that much of the furnishings and family possessions had been removed.
When charged with arson, the defendants argued that the searches were conducted with-
out consent, warrants, or emergency conditions requiring immediate examination and
that the evidence from those searches should be suppressed. The State argued that it
should be exempt from all warrants requirements in fire scene searches. The Supreme
Court held that the searches were unconstitutional on several grounds, which distin-
guished the case from Tyler. First, there was no continuous custody of the scene. The
search was not begun until several hours after police and fire units had left. The search in
Clifford was neither prompted by emergency conditions (impending loss or destruction of
evidence) nor was it a continuation of an earlier search. Further, the structure was a pri-
vate residence, and the Court felt that there was a higher expectation of privacy in such
a dwelling than in a business such as in Tyler. That expectation of privacy was expressed
in the steps the owners had taken in ordering salvage and security arrangements. The
search begun hours after the fire was extinguished required a consent, newly discovered
exigency, or an appropriate warrant. The Court also distinguished between the two
phases of the search. Even if the basement search had been conducted legally, it did not
justify the upstairs search without a criminal search warrant. The basement search
revealed that the fire could well have been of incendiary origin. At that point, the find-
ings could have been used as probable cause to justify seeking a criminal warrant. The
investigators failed to do this, so the upstairs search could not be undertaken legally with-
out a warrant. It would appear that no matter what the basis for conducting a fire inves-
tigation search—consent, emergency, or administrative warrant—once evidence is
discovered indicating the fire could have been criminally set, the investigator would be
well advised, before proceeding, to obtain a criminal search warrant based on the proba-
ble cause revealed.37
Mincey v. ArizonaC7
A drug arrest/raid ended in a shootout with an undercover police officer, several plain-
clothes police officers, Mincey, his girlfriend, and an associate shot. The scene was
searched at that time for additional confederates. A shooting team was called in to search
the crime scene because a police officer and several suspects had been shot. The subse-
quent crime scene search lasted 4 days with considerable destruction of the house and its
contents. The court held that evidence in plain view discovered during a search for other
confederates or in similar emergency situations is admissible. Once the emergency situa-
tion ends, however, the requirement for a warrant prior to further searching resumes.
Once again, the determination of “reasonable time” and “end of emergency” rests with
the investigative officer and superiors but is open to interpretation by the courts.
Romero v. Superior Court C8
In this case, the search of a burning building for a victim revealed weapons and dynamite.
A locker suspected of containing a large quantity of ammunition was opened by the fire
department. In emergency situations, Section 13875 of the California Health and Safety
Code grants peace officer duties to firefighters at the scene. Thereby, this search, which
was conducted for safety reasons, was legitimate.
Two additional points should be raised in the matter of crime scene searches. If a
fire official has obtained an administrative inspection warrant and is conducting a
search pursuant to that warrant, criminal evidence in plain view from where the inves-
tigator has a legitimate right to be—that is, where permitted to go under the protection
of that inspection warrant—is admissible in criminal proceedings. It is interesting to
note that insurance investigators are considered private citizens in the eyes of the law
while working on their own, and not subject to the same Fourth Amendment proscrip-
tions. If they are engaged by a public fire or police agent to work as associates in an
712 Chapter 17 Other Investigative Topics
investigation, they become (legally) deputy agents and are subject to the same laws as
police and other public officials.
Swan v. Superior Court C9
This decision helped define “reasonable rights” of expected privacy. After an April 4 fire,
a search conducted on April 24 of the boarded-up, burned premises was held to be ille-
gal and evidence gained therein was suppressed because the boarding up indicated an
expectation of privacy. An earlier search had also revealed evidence of arson, and the
court held that subsequent reentry was to gather evidence for the prosecution and not to
determine the cause and origin of the fire. In 1973, U.S. v. GreenC10 allowed the submis-
sion of 29 counterfeiting plates that were discovered at the point of origin of a fire by an
investigating fire marshal. The fire marshal was acting in line with his legislative author-
ity in conducting a search for the seat of the fire that happened to be coincidental with
the plates. U.S. v. DelgadoC11 held that a search of a desk drawer was permissible if the
evidence present might be susceptible to fire and water damage.
People v. Holloway C12
The Illinois Supreme Court suppressed evidence that was found during an origin and
cause investigation that was begun 2.5 hours after the fire had been extinguished and fire-
fighters had left the scene. It was held that the investigation should have begun promptly
and that the privilege to begin an investigation without a warrant does not survive a
departure and return to unprotected property in the absence of an emergency situation.
Passerin v. State of DelawareC13
The Delaware Supreme Court held that reentry of a scene 24 hours after the fire had been
extinguished and 21 hours after the first origin and cause investigation had ended without a
warrant was not permissible because no exigent circumstances were present. On remand, the
Court commented that because the first search had resulted in probable cause that a crime
had been committed, a proper search warrant was needed for any subsequent entries. Woods
and Wallace (1991)38 and Cabral (1995)39 have extensively examined Fourth Amendment
issues and the relevant court decisions, and the reader is referred there for more details.
Horton v. CaliforniaC14
Once firefighters enter a building, they may seize evidence that is in plain view.
U.S. v. FinniganC15
Discovery of explosives in a building in plain view of responding officers provided prob-
able cause to support a search warrant to explore the remainder of the premises for other
explosives.

Sources of Information
Because of the complexity and sophistication of many arson investigations, particularly
those of arson-for-profit schemes, the investigator should not hesitate to use all possible
sources of information. Many of these have been suggested in the literature.40 Particularly
valuable sources at the federal level include the ATF, FBI, U.S. Fire Administration
(Federal Emergency Management Agency, FEMA), and the U.S. Postal Service.41 The new
Bomb Arson Tracking System (BATS) from ATF (described in Chapter 12) promises to be
a very valuable aid and resource to fire and explosion investigators.42 State government
sources include the motor vehicle and driver licensing agencies, fire marshal, corrections
department, and the prosecutor, attorney general, or state police departments. On a local
level, sheriffs, police, probation, and district (or prosecuting) attorney’s departments or
the offices of recorder, county clerk, or fire prevention may be of value.
Task forces involving city, county, state, and federal agencies have proven their abil-
ity to solve complex serial arsons involving dozens of targets. These investigations have
successfully identified single arsonists setting dozens of fires across regions or multistate
Chapter 17 Other Investigative Topics 713
 areas, as well as organized arson-for-profit rings. They can bring together experience,
knowledge, and resources that no single agency could duplicate, but they do require the
mutual trust and understanding of not only the investigators directly involved but of the
command ranks of all the agencies represented.43 A sample agreement setting up a suc-
cessful interagency task force is included in Appendix L.
Private companies such as utilities (phone, gas, water, cable TV), Better Business
Bureaus, and real estate offices can offer some aid in the way of data on the financial well-
being of the business or individuals affected by the fire.44 In a more general sense, the
National Insurance Crime Bureau (NICB) provides resources on insurance fraud fires
(combining the Insurance Crime Prevention Institute with the National Auto Theft
Bureau). The NICB is a nonprofit, private organization set up to facilitate contact
between insurance investigators and law enforcement. All insurance companies can sub-
mit data on fire, fraud, stolen vehicles, and the like, both nationally and internationally,
to NICB and public agencies. Public agency investigators then access this data through the
regional representatives for NICB.
By far the richest sources of information regarding the financial motivators of fraud
fires are the insurance companies and their adjusters. Until recently, however, insurance
information could not be given to police agencies without risking severe civil damages for
violating the client’s right to confidentiality. In the last few years, all 50 states have
enacted some form of what has become known as arson reporting immunity law. These
laws generally have several important provisions. First, they require insurance companies
to notify the state fire marshal or other authorized agency when the insurer suspects that
a fire loss on property it insures was caused by incendiary means. Second, these laws grant
immunity from criminal prosecution or civil liability to these insurance companies for
releasing information about these fires. Third, the insurance company can be fined or
denied license to operate if it intentionally withholds this kind of information. While such
laws have generally made it easier to successfully detect and prosecute insurance fraud
cases, insurance companies are still very cautious about releasing information to police or
fire agencies, sometimes preferring to risk possible fines than to risk bad-faith civil judg-
ments. It is best to contact your state’s insurance board for specific guidance.
Many major insurance companies in the United States have established Special
Investigation Units (mandated by law in some states). These SIU investigators may be retired
police officers with extensive arson/fraud investigation experience, staff adjusters given spe-
cial training, or outside (consultant) employees with specialty experience and knowledge.
The insurance company representative with whom most investigators have contact is
the adjuster. There are several types of adjusters, each with some differences in authority.
Property adjusters (fire, flood, tornado, etc.) deal with first-party direct claims on dam-
age and determine the amount of monetary loss. Casualty adjusters deal with liability
issues on third-party claims. In a fire, the investigation is aimed at determining a cause
and origin and then providing an estimate of the repair or replacement of the damaged
structure. If warranted, the adjuster may be authorized to pay the insured for the loss.
Independent adjusters are employed on behalf of the insurance company and therefore
have no vested interest in the company or the insured. Staff adjusters are employees of
the insurers and therefore may have an interest in the well-being of the company and the
insured. Public adjusters, in contrast, are employed directly by the insureds to represent
them in the claim adjustment negotiations. They have no vested interest except in the
amount of the claim settlement, of which they normally take a percentage fee.

Spoliation
As defined in Chapter 14, spoliation of evidence is the destruction or material alteration
of, or failure to save, evidence that could have been used by another in future litigation
(or in litigation that is thought reasonably possible in future). The available remedies for
714 Chapter 17 Other Investigative Topics
spoliation of evidence continue to evolve but, depending on the jurisdiction, may include
criminal penalties, adverse inferences, sanctions, and/or tort liability. For a summary of
U.S. court case citations and opinions on the spoliation of evidence, see Table 4-4 in
Chapter 4 of Forensic Fire Scene Reconstruction.45
Criminal investigators and forensic scientists are familiar with the concept of preserv-
ing physical evidence or a portion thereof to permit retesting by opposition experts. It is the
responsibility, however, of anyone who handles or examines evidence to see that evidence
or documents are not lost, destroyed, or changed in such a way that they cannot be ana-
lyzed or interpreted by someone else, whether the evidence is in a criminal or civil case.46
The investigation of a fire or explosion will invariably involve moving or changing
some portions of the evidence even to make an initial evaluation. Because such changes
are unavoidable, the state of the evidence must be documented as it was found.

PUBLIC-SECTOR INVESTIGATORS AND SPOLIATION

The public investigator in most jurisdictions has the right to seize evidence of criminal
activity but not to remove evidence that clearly has no connection to the proof of arson
(unless it is needed to show that certain accidental ignition sources were considered and
eliminated). The public-sector investigator should not remove evidence of accidental igni-
tion for personal examination or training purposes. For the most part, public-sector inves-
tigators are protected from penalties in spoliation cases as long as they are carrying out
their official duties. Willful destruction of evidence or its removal from a scene may
expose them to sanctions (criminal or civil). Evidence that is seized for prosecution of an
arson case must be properly documented via photographs and diagrams, properly pack-
aged and tagged, and maintained with a proper chain of custody. Failure to protect evi-
dence in a criminal case may compromise the prosecution of the case and open the door
for civil spoliation claims against the investigator and his or her agency. Preservation of
evidence of fire cause is not limited to the material concerned with the arson set or acci-
dental cause itself. It may (depending on the circumstances) extend to other heat sources
in the area that have been examined and eliminated as not being responsible. They should
not be discarded or destroyed simply because they have been eliminated.

PRIVATE-SECTOR INVESTIGATORS AND SPOLIATION

The private-sector investigator or forensic expert does not have governmental immunity.
Before removing evidence from a scene, the investigator should make reasonable efforts
to notify known or possible defendants, plaintiffs, or other interested parties of the loss
and the intent to remove evidence. If the identity of possible defendants is not known,
documentation by photography (with complete photo log) and sketching is the best alter-
native. If evidence is retained for testing, known parties (manufacturer, installer, etc.)
should be notified in writing before any testing or changes are carried out, so that they
can arrange to examine the evidence before testing or arrange for their own experts to be
on hand during the examination. The recovery of exemplar or comparison samples of
undamaged materials from the scene is also suggested. All investigators must be certain
that all the parts of an item are collected (when collecting an appliance, the cord, plugs,
and receptacle involved should be collected).

CONSEQUENCES OF SPOLIATION
If an investigator (or someone else handling evidence) destroys, materially alters, or fails
to preserve such evidence, a number of adverse consequences may result. Those remedies
may include denying permission to introduce any testimony regarding the lost evidence,
or instructing the jury that the testing would have resulted in results favorable to the party
denied access (adverse inferences), or to civil and even criminal penalties against the person
responsible for the loss or destruction. Though quite a few jurisdictions have gone so
Chapter 17 Other Investigative Topics 715
 far as to make spoliation of evidence a criminal offense, such laws are only rarely
enforced. (See, e.g., Cal. Penal code §135 [misdemeanor]; 720 Ill.Comp.Stat.5/31-4 [class
4 felony]; N.Y. Penal Law §215.40 [class E felony].)
The most controversial, and as yet unsettled , remedy for spoliation of evidence has been
as an independent cause of action for either intentional or negligent spoliation of evidence.
Both remedies were first recognized in California. The California Court originated the tort of
“intentional spoliation of evidence” in 1984 in Smith v. Superior Court.C22 In Smith, the
plaintiff was blinded by injuries when, while she was driving a car, the wheel of a passing van
flew off and crashed through her windshield. The van was taken for repair to the company
that had installed the van’s custom wheels. The defendant repair company, however, later
destroyed certain automotive parts key to the plaintiff’s claim, despite an agreement to main-
tain those parts. The California Court of Appeal held that the plaintiff could bring an inde-
pendent lawsuit against the company for intentional spoliation of evidence.
Soon after Smith was decided, the state of California adopted a separate cause of
action for “negligent spoliation of evidence” in Velasco v. Commercial Bldng.
Maintenance Co. In Velasco, a maintenance worker at a law firm discarded an unmarked
paper bag sitting on an attorney’s desk.C17 The bag contained pieces of a glass bottle that
were evidence in a product liability claim brought by the attorney’s client, Pedro Velasco.
Though the court found the maintenance worker not liable (since he would have no way
of knowing the unmarked bag of glass was anything other than trash), it found that an
action for “negligent spoliation” could be brought under the right circumstances (in this
case against the lawyer, for failing adequately to mark or protect this important evidence).
Velasco was also notable because it involved the actions of a “third party” in spoliation,
that is, one who was not a party to the litigation.
Since Velasco more than 25 jurisdictions have addressed the issue of whether to rec-
ognize an independent tort for the spoliation of evidence. However, the “tort of spolia-
tion of evidence has not been widely adopted in other jurisdictions, nor has much
agreement emerged on its contours and limitations.”C18
In reality, an analysis of the trends in spoliation involves four separate possible torts:
(1) intentional spoliation by a party to underlying litigation; (2) negligent spoliation by a
party to underlying litigation; (3) intentional spoliation by a third party not a party to
underlying litigation; and (4) negligent spoliation by a third party not a party to the
underlying litigation.47
Significantly, in 1998 and 1999, the California Supreme Court rejected the Smith and
Velasco rulings; and the state that created these causes of action no longer allows inten-
tional or negligent spoliation of evidence as a separate cause of action for first- or third-
party conduct.C19 Cedars-Sinai Med. Ctr. v. Superior Court held that there is no tort cause
of action in California for first-party intentional spoliation of evidence. Temple Community
Hosp. v. Superior Court (1999) held that there is no tort cause of action against a third
party for intentional spoliation of evidence.C20 Farmers Ins. Exch. v. Superior Court (2000)
held that there is no tort cause of action for negligent spoliation of evidence.C21 Coprich
v. Superior Court (2000) also held that there is no tort cause of action for negligent
spoliation of evidence (but indicated that plaintiff might be able to assert a claim for
breach of a contractual duty to preserve evidence).C22 Penn v. Prestige Stations, Inc. (2000)
held that the ban on spoliation tort claims applies retroactively, reversing judgments
obtained prior to Cedars-Sinai.C23
Similarly, the trend in the great majority of jurisdictions in recent years has been
to curtail the actionability of separate causes of action for spoliation, especially in the
context of alleged spoliation by a party to the underlying litigation.C24 These courts have
reasoned that such an action would be superfluous, given the available remedies within
the matter at hand.
As of 2010, seven states recognize the tort of intentional spoliation as a tort applica-
ble to first parties, namely, Indiana, Ohio, Connecticut, Louisiana, Montana, New Jersey,
716 Chapter 17 Other Investigative Topics
and West Virginia. States that appear to recognize the tort of negligent spoliation are the
District of Columbia, Kansas, Pennsylvania, Illinois, and New Jersey.48 Florida recognizes
causes of action for both intentional and negligent first-party spoliation. The majority of
jurisdictions also do not recognize third-party spoliation as an independent claim. As one
court held,
the benefits of recognizing a tort cause of action, in order to deter third party spoliation
of evidence and compensate victims of such misconduct, are outweighed by the burden to
litigants, witnesses and the judicial system that would be imposed by potentially endless
litigation over a speculative loss, and by the cost to society of promoting onerous records
and evidence retention policies.C25
Likewise, the majority of jurisdictions also have rejected a cause of action for negli-
gent third-party spoliation. For example, in MetLife Auto & Home v. Joe Basil Chevrolet,
Inc., a home insured by MetLife sustained over $330,000 in damage from a fire that
began in an aftermarket electronic remote starter system installed in the dashboard of a
Chevy Tahoe parked in the garage.C26 Royal Insurance Company, which insured the
dealer who owned the car, took possession of the vehicle after indemnifying its insured,
the car dealer. Even though it voluntarily agreed to preserve the evidence for inspection
by others, Royal allowed the car to be sent to a junkyard and scrapped. The court held
that the societal costs of mandating the preservation of anything that might conceivably
be evidence may be enormous and concluded that it did not recognize a cause of action
against negligent third-party spoliators.
However, a minority of jurisdictions have recognized third-party spoliation as an
independent tort. For example, the West Virginia Supreme Court of Appeals, in Hannah
v. Heeter, ruled that both intentional and negligent spoliation by third parties are recog-
nized causes of action.C27 In so ruling, the court concluded that to rule otherwise would
be inconsistent with “our policy of providing a remedy for every wrong and compensat-
ing victims of tortious conduct.” The Hannah court further advised that “a duty to pre-
serve evidence for a pending or potential civil action may arise in a third party to a civil
action through a contract, agreement, statute, administrative rule, voluntary assumption
of duty by the third party, or other special circumstances.” The court set out six elements
for the bringing of a claim for negligent spoliation of evidence:
1. The existence of a pending or potential civil action;
2. Actual knowledge of the pending or potential civil action;
3. A duty to preserve evidence arising from a contract, agreement, statute, administra-
tive rule, voluntary assumption of duty, or other special circumstances;
4. Spoliation of the evidence;
5. The spoliated evidence was vital to a party’s ability to prevail in a pending or
potential civil damages action; and
6. Damages.
For intentional spoliation to be proven, the court required seven elements:
1. A pending or potential civil action;
2. Knowledge by the spoliator of the pending or potential civil action;
3. Willful destruction of evidence;
4. The spoliated evidence was vital to a party’s ability to prevail in the pending or
potential civil action;
5. The intent of the spoliator to defeat a party’s ability to prevail in the pending or
potential civil action;
6. The party’s inability to prevail in the civil action; and
7. Proof of damages.
As it stands now, the law is in a state of flux. Although most jurisdictions have rea-
soned that sufficient remedies are available without a separate cause of action for first-party
Chapter 17 Other Investigative Topics 717
 spoliation of evidence, eight jurisdictions still retain independent causes of action for inten-
tional and/or negligent spoliation, and this area of the law is by no means settled. Similarly,
although the majority of jurisdictions also do not recognize an independent cause of action
for negligent spoliation of evidence, some jurisdictions still do permit this separate remedy.C28
Firefighters and fire investigators, particularly investigators in the private sector, are
now being cautioned that they have a duty to preserve evidence.49 This duty to preserve
may be required by statute, agreement or contract (protocols, engagement letter, or insur-
ance agreement), court order, industry standards, or some other special relationship or
circumstance.
Forensic engineers involved with failure analysis investigations are cautioned regard-
ing the spoliation of evidence, and they are advised to become familiar with the require-
ments of ASTM E 860: Standard Practice for Examining and Preparing Items That Are
or May Become Involved in Criminal or Civil Litigation, for preservation of evidence and
notification of potential interested parties when destructive testing is required. The con-
sequences of not being familiar and complying with ASTM E 860 could subject the engi-
neer to personal liability for professional malpractice.50
Even without an independent cause of action for spoliation, however, several potent
remedies are available when material evidence is destroyed, disposed of, altered, or lost.
Thus, the proper and safest course of action should be carefully to safeguard evidence
until it can be confirmed that any potential claimants no longer require access.

Chain of Evidence
The documentation of the whereabouts of any item of physical evidence is as important as
the evidence itself. This documentation, called the chain of evidence or chain of custody,
was discussed in a previous section of evidence collection (Chapter 7). Suffice it to say that
the chain of evidence provides a record of every person who had custody of an item. The
evidence presented in court must be documented as being the exact item removed from its
place of discovery. If the investigator has never let the item out of his or her personal con-
trol, the matter is simple. The investigator can testify, “That is the very same item that I
recovered, I placed it in this container, and sealed it, and it is in the same condition as when
I first picked it up.” If that object has been turned over to others for storage, photography,
or laboratory examination, it is within the power of the court to call every person involved
into court and have him or her testify as to its identity and preservation. The chain of evi-
dence is a record of all such custody. Most often, it is accepted as proof of continuity, but
if the evidence appears to have been altered or substituted, the chain provides a record on
which to base any inquiries. The documentation begins, not with the physical recovery of
the item of evidence, but from its first discovery as something of interest, and includes
notes, sketches, and photos (see Appendixes F and I).
It is important to remember that the chain of evidence is to the case what an anchor
chain is to a ship. It is only as strong as its weakest link. If even one link breaks, the chain
is useless and the case (or the ship) can be lost. Extra care must be exercised by the inves-
tigator to keep a chain of custody as short as possible and to keep careful records of its
every step.

Report Writing
Thorough fire scene investigation and accurate and insightful reconstruction of the origin
and cause of a fire are of no value if the investigator cannot transmit the painstakingly
gathered information to others. Although some of this is done through verbal reports and
court testimony, by far the most critically important means of communicating is through
the written report. The preliminary report is often a brief summary of the what, who, and
when of a fire, sometimes even a checklist on a preprinted form. These reports get the
718 Chapter 17 Other Investigative Topics
basics of the case—names of property owners, dates, property, witnesses—but little else.
The investigation report, however, is an official record of the investigator’s activities and
findings. It must reflect all the important information gathered as to the fire, its circum-
stances, and the people involved, as well as present the conclusions with regard to origin
and cause. The report must be complete so that someone reading it later will have full
knowledge of what was found, even if the investigator is not there to interpret it. It will
represent the investigator’s efforts to police and fire personnel, lawyers, victims, court,
insurance companies, judges, and juries. The report must address not only the investiga-
tor’s final conclusion of the origin and cause of the fire but also the evidence that supports
that conclusion. It must also discuss the alternative explanations for the fire and what
data were evaluated to eliminate those alternatives. Although report content and format
may be constrained by company or agency policies or court rules, many investigators
today are following the suggested practices in ASTM E620-04: Standard Practice for
Reporting Opinions of Technical Experts.51 This information, in addition to that already
listed, includes the name of the person or organization requesting the report, date of
preparation, list of items examined, all pertinent facts upon which the opinion was based
(evaluated in accordance with ASTM E678: Standard Practice for Evaluation of Scientific
or Technical Data—the scientific method), expert opinions and the logic and reasoning
behind those opinions, and a signature of the reporting expert.52 One way to ensure the
completeness of a report is to think to oneself, “If I were to die tomorrow and this report were
the only information on this incident, how completely and accurately would it reflect
what I know about this fire?” It is surprising how many gaps one finds in a report when
it is regarded as the ultimate document. With completeness and accuracy as the primary
considerations, it is only necessary, then, to make it as clear, concise, and readable as possi-
ble. Reporting formats and styles will vary from agency to agency, but let us take a simple
report and illustrate some major considerations.

REPORT SUMMARY
The most readable reports begin with a short summary of the facts and conclusions to
be presented in the body of the report. In this way, the reader is given an idea of where
the report is heading without reading the whole document. This summary often includes
the following:
■ Date, time, and location of fire
■ Date, time and circumstances of scene examination
■ Nature of fire damage
■ Basic conclusions about the origin and cause of the fire.

THE SCENE
Valuable information about the fire cause can be obtained from a history of the building
or property involved. Information to be gathered from witnesses or documentary sources
and reflected in the report may include a description of the building or property prior to
the fire (age, construction, use, tenancy, owner, insurance, financial history, and the occur-
rence of other fires in area). The names and addresses of witnesses are included along with
their statements.
The absence of a history of previous problems—flickering lights, power failures, gas
smells—may be useful in eliminating particular accidental sources and may be useful in
establishing an arson corpus. Such information can be gathered from residents, owners,
neighbors, recent visitors, and utility company records.
The fire itself—time and manner of discovery and by whom, time and nature of
response, and extent of all firefighting operations—should be documented. Interviews
with fire personnel regarding their observations as to fire conditions and extent should
reflect unusual flames, smoke, odors, or the presence of vehicles, bystanders, or obvious
Chapter 17 Other Investigative Topics 719
 features of the building such as open doors and windows. Weather conditions at the time
of the fire (wind, rain, lightning, temperature) all may play a role in the ignition or spread of
a fire and should be included. The pre-fire condition of the building, its contents, and
operational systems (electrical, HVAC, heat/smoke detection, suppression, and security)
are essential to an accurate and reliable investigation.

THE INVESTIGATION
The observations of the investigator in retracing the progress of the fire must be recorded
to substantiate any conclusions reached as to origin and causation. It is not adequate to
simply state that the origin was identified as a particular location. The diagnostic signs—
char depth, heat horizons, low burns, unusual fuels, and so on—used to reach that con-
clusion must be listed. These observations lend credence to the investigator’s work and
allow one to convince a jury of the “reality” of the scene.
Once the origin has been described, the cause can be discussed. What fuel and source
of ignition appeared at the origin to start the fire? What signs were there of accidental,
natural, or incendiary origin? If it appears to have been accidental, subsequent examina-
tions of the appliance or device responsible may have been carried out. If it appears to
have been incendiary, what possible sources of natural or accidental ignition were pres-
ent, and how were they excluded? If an incendiary origin is suspected, how was it started,
and what other signs of criminal activity—broken windows, forced locks, ransacked
office, bloodstains, tool marks, tools, or weapons—were found? In short, the investiga-
tive report must reflect all evidence found and how that evidence was interpreted.

REPORT CONCLUSIONS
In concluding the investigative report, all follow-up and supplemental inquiries should be
cited along with recommended later courses of action. A list of evidence collected and a sum-
mary of what is to be done with it is an important part of this supplemental information.
In some cases, a separate report of all investigative activities may be required for judi-
cial processing. This is especially true in some complex arson cases in which documentary
evidence from financial, insurance, or governmental sources may be a part of the proof
of intent. This report should be complete in its scope and coverage and often will include
photos, sketches, diagrams, affidavits, laboratory reports, and other supporting evidence.
This report should be organized in such a way that the prosecutor handling an arson case
or a civil attorney handling a lawsuit can see all the elements of the case laid out in a clear,
well-organized manner. As in other phases of investigation, careful and thorough prepa-
ration of this final report will offer great rewards in terms of more successful prosecutions
or litigations. Reports from consultants or specialists involved to investigate or develop
portions of an investigation, of course, will not be as comprehensive. Those reports
should be as complete and comprehensible to any future user.
If a report is to represent the total case to be presented to a prosecutor for consider-
ation, a great deal more material will be included. Such a report should include compre-
hensive lists of defendants, victims and witnesses, evidence and lab reports, documentary
evidence, and interviews. Such reports may also include assessments of motives, m.o. and
financial status.53

REPORT WRITING BASICS

No matter to what use the report will be put, the following basics should be followed to
avoid problems:
■ Write simply and clearly. Others reading this report will not be firefighters; use
everyday terms and avoid jargon or code numbers. Use the “active” form of writing
rather than the passive.
720 Chapter 17 Other Investigative Topics
■ Do not report unsupported allegations or suppositions. Your report may be read to
a court verbatim, and such suppositions may be used to show bias.
■ Because the report is the product of a professional, it should look professional.
Errors in spelling or grammar are embarrassing and may give the impression that
all the work was sloppy. Keep a dictionary handy. Using simple active sentences will
keep grammatical errors to a minimum.
■ The report will be relied upon to tell the complete story of the fire by those who
refer to it later, so be sure it is complete and all facts reported are accurate.
■ Once the report is written, preserve all supporting notes, unless there is a specific
departmental protocol that requires notes to be destroyed. Destruction of notes as a
personal policy leaves questions in the minds of jurors and offers a line of chal-
lenges from the opposition lawyer.54

Courtroom Testimony
The final test for a fire investigation is its presentation in court. The fire investigator will
often be called on to offer testimony as to what evidence was found at a fire scene and
what conclusions were drawn from that evidence.
There are differences with respect to the admissibility of testimony given by eyewit-
nesses and experts. In general, experts are allowed to express opinions, and eyewitnesses
may testify only to facts or events observed. In some cases, testimony as to the cause of a
fire by an eyewitness may be acceptable as a statement of fact when the testimony does
not extend past the facts observed. Likewise, in some cases, opinions by the expert are
excluded when an opinion is given as to the ultimate cause or origin of the fire (holding
this to be an issue for the jury).

THE EXPERT WITNESS

Anyone with special knowledge or experience with respect to the subject in question may
qualify as an expert through a showing of training or experience to testify as an expert.
In the case of arson, the expert giving opinion testimony as to the cause or origin of a fire
may be a fire investigator, criminalist, or one of any number of specialists—chemist, engi-
neer, and so on—depending on the nature of the fire. Testimony by such an individual
stating that the fire is of “incendiary origin” or “deliberately set” is admissible whenever
it is determined that the cause or origin is not within the common knowledge of the
layperson. Opinion testimony tending to negate incendiary origin is admissible from
experts but usually not from a lay (nonexpert) witness.
Impact of Frye, Daubert, and Related Cases
The standards for admitting expert testimony have evolved significantly over the twenti-
eth century. In 1923, the Circuit Court of Appeal for the District of Columbia developed
the first test for determining the reliability and thus admissibility of expert testimony. In
Frye v. United States, the court held that for scientific evidence to be admissible, the party
proffering the evidence must establish that the method used to produce it has been gen-
erally accepted as reliable in the “scientific community.”C29
Since 1975, Federal Rule of Civil Procedure 702 has been the standard for admission
of expert testimony in federal (as well as many state) courts. Essentially, it required that
for a person to qualify as an expert, the court must determine the witness’s credentials and
methodology used by the expert on two bases: relevance and reliability. In 1993 the U.S.
Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc. explained that whenever
the expert’s factual basis, data, principles, method, or their application are challenged, the
court (trial judge), acting as a “gatekeeper,” must determine whether the expert testimony
has a reliable basis in the knowledge and experience of the relevant discipline.C30 The trial
judge also must determine whether the expert is proposing to testify to scientific knowledge
Chapter 17 Other Investigative Topics 721
 that will assist the court in understanding or determining a fact in issue. If an expert is to
testify on a scientific theory or technique, then it must be evaluated on four factors: the the-
ory or technique (1) has been or can be tested, (2) has been subject to peer review and
publication, (3) has a known or potential rate of error, and (4) has gained general accept-
ance in the relevant scientific community. Recognizing that those factors would
not apply in every case, the Court stated that those four factors do not constitute a defini-
tive checklist or test for all cases and that the trial court must gear its evaluation to the
particular facts of each case. The fire investigation community therefore could present other
factors for consideration by the court as to the admission of expert testimony in a fire case.
Although the Court explained that this decision was based on the 1975 Federal Rules,
many questioned whether the case applied outside of the medical scientific community.
U.S. v. Markum contended, for instance, that fire investigation for origin and cause deter-
mination is based on scientific knowledge, and therefore, a proposed expert should be
able to demonstrate the appropriate levels of understanding of the sciences involved.C31
The expert should also be able to demonstrate the accuracy and reliability of the conclu-
sions reached as part of the determination, as well as the data or information relied upon.
This argument was rejected in 1993 in Markum, in which it was reinforced that experi-
ence alone can qualify in fire causation.
The states accepting Daubert standards in state courts include Connecticut, Indiana,
Kentucky, Louisiana, Massachusetts, Missouri, New Mexico, Oklahoma, South Dakota,
Texas, and West Virginia. Frye states include Alaska, Arizona, California, Colorado, Florida,
Illinois, Kansas, Maryland, Michigan, Nebraska, New York, Pennsylvania, and Washington.
States with their own tests, or typically a Frye-plus test, include Arkansas, Delaware, Georgia,
Iowa, Minnesota, Montana, North Carolina, Oregon, Utah, Vermont, and Wyoming.55
In Michigan Millers Mut. Ins. Corp. v. Benfield, it was contended that a fire investi-
gator could offer an opinion that a fire was arson simply as a result of looking at the evi-
dence (with no samples, no tests, and no proof of excluding other possible causes).C32 It
was suggested that fire investigation was not a science and therefore should not be held
to the standards in Daubert. The U.S. Court of Appeals (11th Circuit) rejected that rea-
soning and remanded the case for trial, concluding that more substantiation was needed
than just the unsupported opinion of a fire expert. It said that there was no reason to
admit opinion evidence that was connected to existing data only by the ipse dixit of the
expert (ipse dixit—it is so because I say so—is defined as a bare assertion resting on the
authority of an individual).
In 1997, the Supreme Court clarified Daubert in General Electric Company v.
Joiner.C33 In Joiner, the Court reinforced the notion that the trial judge is the “gate-
keeper” and that the trial judge is supposed to draw his or her own conclusions about the
methodology and data employed by the expert as well as the ultimate conclusions
reached. Thus, after Joiner, under Daubert, the trial judge not only must decide whether
a witness is qualified to testify as an expert (as before Daubert and Joiner) but also must
scrutinize the methods employed by the expert, the application of those methods to the
facts of the case, and even the ultimate conclusions reached by the expert. All these must
meet the approval of the trial judge before they can ever be presented to the jury.
Joiner also clarified that the appropriate standard of appellate review of a district
court’s decision to exclude expert evidence is the abuse of discretion standard. Thus,
unless the trial court can be shown to have abused its (very broad) discretion, the trial
judge’s decision will not be disturbed on appeal.
In 1999, in Kumho Tire v. Carmichael, the Supreme Court, again referring to Rule
702, held that the standard for admission applies not only to scientific testimony but also
to any technical testimony or to that based on specialized knowledge.C34 Fire investiga-
tors potentially fall into all three categories. Further, the Court confirmed that just
because the expert declares he or she is an expert, the trial court must still examine the
expert’s credentials and methodology on a case-by-case basis.
722 Chapter 17 Other Investigative Topics
 In December 2000 Congress expanded the language of Rule 702 regarding experts in
an effort to put this issue to rest. The new language states that an expert witness may tes-
tify if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the prod-
uct of reliable principles and methods, and (3) the witness has applied the principles and
methods reliably to the facts of the case. There are several important conclusions that
might be drawn in applying the new Rule 702, Daubert, Joiner, and Kumho Tire to the
field of fire investigation, First, in all jurisdictions applying the Daubert rule, the expert
testimony of fire investigators will be subject to challenge on the reliability of the investi-
gator’s methods, techniques, and conclusions. Second, the trial judge will determine the
reliability of the expert’s methods and conclusions, using a framework of criteria that may
include some or all of the Daubert criteria, or other different criteria appropriate to the
particular case. This evaluation will take place before expert is permitted to give any tes-
timony before the jury. Third, it will be the fire expert’s burden to show that the methods
he or she employed in the fire scene analysis are reliable, and the conclusions drawn also
are reliable. The judge must be convinced of the investigator’s methods and conclusions
before the expert’s testimony will be allowed at trial. Fourth, once the trial judge makes
his or her decision, that ruling rarely will be overturned on appeal. Fifth, to convince the
judge of the reliability of his or her methods and conclusions, the investigator must be
prepared to validate and verify every facet of the investigation. The investigator must
show why a particular method was used and how the method has been shown to be reli-
able. Similarly, the investigator will need to show why the conclusions reached are valid,
based upon those methods, with documentation of the validity of those conclusions.
Some state jurisdictions still rely on Frye v. United States rather than Daubert (or FRE
702). Under Frye, the court essentially examines whether the technique has been gener-
ally accepted within a relevant expert community (the fourth prong of Daubert). One way
to demonstrate that the methods and techniques employed by the investigator are “gen-
erally accepted” is to show the court that they are consistent with the standards, recom-
mended practices, texts, and guides that are presented as a variety of authoritative
treatises in the discipline of fire investigation. Among the treatises on which an expert
may rely are this textbook, NFPA 921 (and 1033), Forensic Fire Scene Reconstruction
(Icove & DeHaan) ), NFPA standards and guides, ASTM practices and methods, and
other publications that are recognized by the fire investigation community.
In Weisgram v. Marley Co., a North Dakota court struck the testimony of three fire
experts.C35 In that case, an electrical engineer and a metallurgist were disqualified because
the court concluded their opinions were based on speculation and conjecture rather than
on provable facts and evidence relevant to the case and on a reliable methodology. The
testimony of a fire captain was partially admitted as to origin of the fire but excluded as
to cause where he expanded his opinion to electrical and chemical issues for which he had
no substantial knowledge, training, or experience.56
Normally, the weight of evidence as to the cause of a fire is determined by the same
tests applied to other evidence. In other words, the jury is not bound to accept it. Also
experts may be allowed to introduce evidence of experiments to sustain a theory of how
the fire started if the conditions surrounding the experiment are not too dissimilar from
the actual event (i.e., relevance) must be demonstrated.C36
In Pride v. Bic Corp., a man mysteriously caught fire while inspecting a pipe behind
his house.C37 In an ensuing product liability action, the widow’s theory was that her
deceased husband’s butane lighter first failed to extinguish, igniting the man’s clothing,
then exploded, dousing the man with isobutane and fueling a conflagration that ulti-
mately caused his death. The widow offered three experts: (1) a mechanical engineer who
had testified in numerous product liability suits, on subjects ranging “from car seat belts
to manure spreaders”; (2) a firefighter who had previously testified in Bic lighter cases on
causes and origins of fires; and (3) an analytical chemist. The engineer opined, based on
his inspection of the lighter, that the exploding-lighter scenario was the most likely cause
Chapter 17 Other Investigative Topics 723
 of the fire and that the mishap was caused by a manufacturing defect and also by failure
of the lighter’s design to incorporate redundant safety features. The firefighter opined that
the lighter was the most likely cause of the fire based on elimination of other plausible
causes as well as information suggesting that the fire had started in the victim’s breast
pocket. The chemist opined, based on information regarding the condition of plastic from
the lighter, that the lighter had exploded. After a Daubert hearing, the trial court excluded
the preferred testimony of all three experts. The appellate court upheld the trial court’s
exclusions of the evidence on the basis that none of widow’s experts conducted replicable
laboratory tests showing that explosion of the lighter was consistent with a product
defect. The engineer’s testimony regarding manufacturing defects was contradicted by the
widow’s other witnesses and by defense experts’ lab tests. The firefighter admitted he was
not an engineer, had performed no tests, and was not an expert in lighters. The chemist
admitted his lack of expertise in fire investigations and that he did not personally exam-
ine the lighter. The chemist also designed a lab experiment to test his hypothesis but said
he “chickened out and shut the experiment down.”57

Impact of MagneTek
A 2004 decision of the U.S. Court of Appeals for the Tenth Circuit has been the subject
of great discussion for its discussion of both legal issues and fire science. This case under-
scored the importance of properly evaluating and presenting expert testimony in fire liti-
gation cases. Truck Insurance Exchange v. MagneTek, Inc. not only demonstrated the
application of the Daubert standard for the admissibility of expert testimony but effec-
tively contradicted a long-standing principle of fire science that had been used in a signif-
icant number of cases to prove the cause of a fire.C38
The MagneTek case involved a subrogation action filed by Truck Insurance Exchange
against the manufacturer of a fluorescent light ballast that was alleged to have caused a
fire that destroyed a restaurant in Lakewood, Colorado. Investigations by both the local
fire protection district and a private fire investigation firm hired by the insurer determined
that the fire had started in a void space between the basement storage room ceiling and
the kitchen floor. In the basement, the investigators found the remains of a fluorescent
light fixture that had been mounted to the ceiling of the storage room. They determined
the light fixture had been located in the area of origin of the fire and concluded that the
fire had been caused by an apparent failure of the ballast in the light fixture.
The investigators theorized that the heat from the ballast had caused “pyrolysis” to
occur in the adjacent wood structure of the ceiling, causing “pyrophoric carbon” to form
over a prolonged time, which would be capable of ignition at temperatures substantially
below the normal ignition range of 204°C (400°F) or more for fresh whole wood. The
phenomenon of pyrolysis in the formation of “pyrophoric carbon” has been the subject
of a number of studies, reviews, and articles by fire investigators and fire scientists. It has
been cited as the cause of a number of fires having no other apparent explanation, often
linked to heated pipes in structures within walls, ceilings, and floor areas. As the
MagneTek court case noted, however, the validity of this phenomenon has been discussed
and debated by fire investigators and fire scientists for a number of years.
Pyrolysis and the formation of “pyrophoric carbon” was the foundation of the plain-
tiff’s case against MagneTek. The ballast in the light fixture showed no signs of failure in
the thermal protector, which would have limited the heat generated by the ballast to
about 111°C (232°F). Even with the exemplar ballast whose thermal protector failed to
perform as it had been designed, the temperatures generated did not exceed 171°C
(340°F). The investigators admitted the ignition temperature of fresh whole wood is
typically at least 204°C (400°F), and the temperatures from the ballast alone would not
have been sufficient to cause ignition. Their theory that the ballast had caused the fire
depended on the concept of pyrolysis to allow ignition to occur at a lower temperature
within the range of the temperatures generated by the ballast.
724 Chapter 17 Other Investigative Topics
 The court noted that under NFPA 921 an investigator offering the hypothesis of an
appliance fire must first determine the ignition temperature of the available fuel in the
area and then must determine the ability of the appliance or device to generate tempera-
tures at or above the ignition temperature of the fuel. In that regard, the experts failed on
both counts. The ignition temperature of the fuel (wood) exposed to the fixture in this
case could not be scientifically proved to be below the 204°C (400°F) threshold for the
ignition of most types of wood, and the tests of the ballasts had shown that even with a
failed thermal protector, a ballast could not generate temperatures anywhere near that
range. Their hypothesis would have to be based on either an unreliable theory (pyrolysis)
or unsubstantiated assumptions and speculations about the temperature of the ballast
that contradicted their own test results. As such, their testimony could not be admitted.
This decision has significant implications for the litigation of fire cases everywhere. It
demonstrates the importance of developing sound and scientifically verifiable theories for
proving the cause of a fire as required by the standards of the Daubert decision.
Moreover, it provides a compelling example of how a theory that has not been validated
and generally recognized by others in the scientific community may not withstand a
Daubert challenge in court.
A scientific analysis and assessment of the MagneTek case decision indicates that the
court may not have understood several important aspects of fire science, including pyrol-
ysis, which is not newly discovered nor scientifically disputed. In the MagneTek case, the
Court rejected under Daubert the “pyrolysis theory” of long-term, low-temperature igni-
tion as being unreliable. Fire investigators who wish to explain this phenomenon in court
must be prepared to make clear and convincing arguments using authoritative treatises
and clearly defined definitions.
In the Ignition Handbook, pyrolysis is clearly defined as “the chemical degradation
of a substance by the action of heat.”58 This definition includes both oxidative and
nonoxidative pyrolysis, taking into account the cases where oxygen plays a role. Pyrolysis
is a process that has been studied for many decades. Without pyrolysis to break down
their molecular structures, most solid fuels will not burn. Unfortunately, the published
court decision declares “pyrolysis” to be inadequately studied and documented (presum-
ably intending to refer to “pyrophoric process”). It is unclear what effect this misstate-
ment of the court will have on future deliberations.
Even more problematically, the court effectively denied the existence of any ignition
due to self-heating, because it declared that ignitable substances have a “handbook”
ignition temperature, and a hot object below that temperature will not be a competent
source of ignition. In actual fires, self-heating substances do not have a unique ignition
temperature. Furthermore, if a substance (e.g., wood) can ignite owing to external,
short-term heating or to a prolonged process involving self-heating, the minimum tem-
perature for the latter has no relation to the published ignition temperatures for the for-
mer. In this respect, the citation from NFPA 921 regarding ignition temperatures is in
error. A full discussion of the legal and scientific issues is included in Forensic Fire Scene
Reconstruction.59

Other Considerations for Experts

The responsibilities of the expert witness are much greater than those of a factual or eye-
witness. The success of the case will very often depend on how well the fire investigator
performs even under the duress of intensive cross-examination. Whether the investigator
is there as only one of several witnesses testifying as to case findings or whether he or she
is acting as the investigating officer, guiding and assisting the attorney, the conduct and
testimony offered must be of the highest caliber. The investigator should be aware of (and
it is hoped, be able to meet) the qualifications of knowledge, skills, and abilities for a pro-
fessional fire investigator as laid out in NFPA 1033: Standard for Professional
Qualifications for Fire Investigator.59 It does little for an expert witness’s qualifications
Chapter 17 Other Investigative Topics 725
 when the expert seems to know less about the professional requirements for his or her
profession than does the opposing counsel.
In a recent series of articles about the Daubert, Kumho, and Benfield cases,
Professor Vincent Brannigan claimed that merely employing the scientific method to
test hypotheses about past events does not make it science and that the critical ability
of a scientific endeavor is to allow the prediction of future events by creating testable
hypotheses.60 Barring the exceptional cases in which a full-scale re-creation of a fire
event is carried out, the hypotheses tested in nearly all fire cases must be based on data
from past experiences as to what effects one predicts that certain processes or events
will have. This does not necessarily make a fire inquiry nonscientific, any more than not
being able to re-create the birth of the universe does not make hypotheses about the
process any less scientific, as long as relevant and accurate methods are used to gather
data and test pieces of the puzzle. A great deal of reliable scientific data have already
been published regarding basic fire processes and fire growth (especially since 1985),
and much more about the effects of flammable liquid fires on rooms, furnishings, and
people.61 A great deal of these data have debunked some of the mythology used for so
long in fire investigation (see Chapter 7) and confirmed the reliability of other indica-
tors. It is these data that investigators today can use to verify their conclusions about a
fire, if the time is taken to do so. Research is continuing in many different venues, in
both the public and private sectors.

PRETRIAL PREPARATION
It is essential that the investigator be prepared for the testimony. Notes and reports must
be reviewed and any questions or problems discussed with counsel. Any errors or discrep-
ancies discovered during this review should be brought to the attention of counsel and
resolved, if possible, before testimony. If errors are discovered, it will be better for all con-
cerned if they are freely admitted to upon direct examination rather than during a poten-
tially hostile cross-examination.
An effective expert witness is often in the role of a teacher in the courtroom. Visual
aids are very valuable in teaching the jury about fire patterns, physical evidence, and fire
behavior. “Before-and-after” illustrations can be very effective. Showing the jury what
the building was like before the fire via photos, computer-generated overheads, cut-
aways, and “walk-throughs” can be helpful. Some examples of such aids were published
recently.62 Digital projectors and “Elmo” overhead projectors can be of great value, but
only if the illustrations are well thought out and prepared beforehand. Throwing a few
photos on a projector is not adequate. The “CSI effect” has been to heighten the expec-
tations of judges and juries about the quality and quantity of “scientific” information
presented. Figures 17-4 and 17-5 are examples of NIST’s FDS/Smokeview modeling
showing the dramatic difference in predicted conditions in the Station Nightclub fire
(February 2003) with and without sprinklers. The visual and measurement data that can
be collected today by laser scanners and scanning digital cameras can streamline the
process and vastly improve the quality of the presentation (as described in Chapter 7 and
Appendix M).
Charts, diagrams, or other visual aids must be planned to ensure their completeness
and accuracy. A pretrial conference is very important, so that both the attorney and the
witness know the strengths and weaknesses of the testimony before it is presented. Make
a list of your qualifications. Study it and provide a copy to counsel before court. Such
preparations are equally important when preparing for depositions in a civil case because,
to a large extent, they take the place of the preliminary hearing in a criminal case.
To give yourself the best mental “edge” before testimony, allow extra time for travel
to court and try to arrive a few minutes early to provide a breather beforehand. Dress
conservatively and in a manner suitable to the venue (a three-piece suit is fine for big cities
726 Chapter 17 Other Investigative Topics
but may be overdressing for a small rural court). Avoid flamboyant clothes or distracting
jewelry. The jury has to be able to concentrate on what is said.

TESTIMONY
Testify in a clear, firm voice. Avoid slang, professional jargon, or profanity (unless quoting
what someone else said). Consider every question and its answer carefully before respond-
ing. Listen carefully to each question. Answering in a rapid, offhand manner can be a risky
error. As an expert witness you have a right (subject to review by the court) to answer every
question in full and explain your answers. Speak to the jury (or the court if there is no
jury), and watch the reactions of your audience. You are there to explain things about
which they will have no other information. If the jury cannot hear you or is confused by
technical language, your main value as a source of irreplaceable evidence will be lost.
Keep calm and give the same careful consideration to every question no matter which
side is asking it. Above all, tell the truth. A skilled cross-examiner can discredit any wit-
ness who exaggerates testimony, embellishes it, or fills in unreliable information. Being
caught in a web of even such “little” lies will impugn your credibility and may discredit
your whole case in the eyes of the jury. Remember that the opposition counsel’s job is
specifically to test your knowledge, probe the accuracy of your information, and reveal
the errors or weaknesses of your conclusions. The professional investigator does not
become angry with the opposition or try to engage in an argument. A professional comes
thoroughly prepared, testifies to the limits of the information, explains and defends con-
clusions, and provides a source of reliable information to the trier of facts. It is only in
this way that the fire investigator sees that justice is served in the case at hand.

Scientific Method
In the opening chapter of this book, the concept of the “scientific method of fire investi-
gation” was introduced. As you will recall, this means to apply the scientific method of
logical inquiry to the investigation. It is an approach most good fire investigators already
apply. It simply means the investigation follows a series of logical steps:
1. Observe an event (such as a fire scene).
2. Define the problem (usually to determine how the fire started).
3. Collect data (information about the event—witness interviews, scene examination,
fuel load, videos or photos of the fire in progress).
4. Formulate a hypothesis (an estimate of what the first fuel was, what ignition source
was present, how long the fire took to grow, etc.).
5. Test the hypothesis against all available data and collect new information (this
includes testing alternative hypotheses (possibilities).
6. Revise the main or working hypothesis as needed to fit the available data, possibly
collecting more data and doing further analysis (a feedback loop).
7. Reach a final hypothesis and review it against all data and ensure that other possi-
bilities are excluded.
8. Report a final conclusion (along with other possible explanations that cannot be
excluded based on the information available).
It is step 7 that is so crucial to the fire investigation process and one that triers of fact will
be looking for in the future as a result of Daubert and Kumho. The scientific method goes
beyond just collecting information and interpreting data, it is a way of using those inter-
pretations to predict other future events. It is the accuracy with which future events can
be predicted that is the real test of a “final” hypothesis as new data are gathered from
these later events. If applied to mechanical systems, the scientific method, yields immedi-
ately testable “final” hypotheses: My car breaks down; I apply some basic data gathering
Chapter 17 Other Investigative Topics 727
 and hypothesis testing: is it ignition, is it fuel, is it fuel pump, fuel filter, etc.? I can test
most of those hypotheses with a wrench or pliers by the side of the road. I find the fuel
filter blocked as part of my testing, flush it out, and reinstall it, with my “final” conclu-
sion being that this is what stopped my car. How do I test it? By getting in the car, start-
ing it, and driving off. That test data affirm what I predicted from previous experience,
that unblocking a plugged fuel filter will allow me to go on. But what if that was not the
whole (final) solution? The car dies again a few miles later, and more testing reveals mud
in the gas tank. To test that I have to drain the tank, flush it, and try to drive it again.
With fire investigation, testing of the “final” hypothesis is more difficult, because the
investigator rarely has other data (such as a surveillance video that shows the fire start-
ing in exactly the way he or she concluded). Occasionally, however, off-site surveillance
video has proven to be the key or confirming evidence of causation.63 It is very rare that
one has the opportunity or the resources to rebuild a room and set fire to it to see if it
really does burn the way one predicts it would. It is nearly impossible to duplicate the
fuels and precise conditions that may affect the ignition and growth of some fires. The
analytical fire investigator is forced to verify the data (observations of indicators for
instance) by comparing them against the physics and chemistry of heat and combustion
and the results of controlled tests of similar scenarios. (Such information may be gathered
either from personal testing or published test results.) As an expert investigator, one has
to be able to demonstrate to the court that the data collected and the method by which
they were analyzed are reliable and relevant.
Another important step is implied in step 5, and that is the formulation and testing
of alternative scenarios or hypotheses. One way to test the conclusions one has reached
is to see if one can create another scenario that will yield exactly the same data. The more
alternatives one can create and test (meaning that they have to be provable hypotheses)
and show that they do not fit or can be excluded (because of a, b, or c data that do not
support them), the more confident one can be that one’s final hypothesis is the correct and
only answer (at least based on available data). For example, the investigator observes at
a fire scene that the floor of a room is very badly burned in an irregular fashion. Rather
than just concluding that flammable liquids were spread around generously, a good inves-
tigator knows that this irregular burning can happen in several ways, so there is a need
to test them all. Was there falldown of combustible ceiling or roof materials? If observa-
tions at the scene tell the investigator there were no falling decorations or collapsing ceil-
ing or roof, one can eliminate that possibility. What about flashover? Were there enough
ordinary fuels in the room and enough ventilation to allow flashover to occur? If so, that
possibility cannot be excluded unless more data are gathered. Did the firefighters see full
room involvement? Was there floor-to-ceiling damage? If not, flashover as a cause of these
low burns can be eliminated because published tests show what happens in post-flashover
room fires. What about a flammable-liquid hypothesis? That can and must be tested, too.
Were there residues detected by lab tests after the fire? Were there empty containers
strewn about? Were there odors detected? Even if all those data support the mechanism
that all the floor damage was the result of flammable liquids, in reaching a “final” con-
clusion, the good investigator also must test alternatives of cause: Was this deliberate or
accidental? Did those containers belong there? Was there a leak in a vehicle or storage
tank that introduced flammable liquids? Is there any physical proof that someone delib-
erately spread the liquids (toolmarks at forced entry points, fingerprints on cans, contain-
ers manually opened rather than by fire effects, etc.). Establishing the first fuel ignited is
only part of the fire investigation puzzle, and demonstrating intent versus stupidity or bad
luck often requires a great deal of additional inquiry and effort.
If the investigator cannot exclude all alternatives based on available data he or she
will have to report the fire as undetermined (possibly explaining what the various unex-
cluded possibilities are). Guesswork is not the same as hypothesis forming. Guesses can-
not be verified, whereas good, sound hypotheses can be tested, verified, or excluded. It is
728 Chapter 17 Other Investigative Topics
 FIGURE 17-8 The
50-year-old building, a
single-story block con-
struction with common
framed roof constructed
of fiberglass shingles. The
flooring of the main sanc-
tuary was an above-grade
wood frame with tongue-
and-groove heart pine
wood construction. Used
with permission of Retired
Major J. Ron McCardle, Florida
State Fire Marshal’s Office.

possible that new data (video, witnesses, weather conditions, etc.) will come along at a
later date that will resolve the issue. Some fire scenes provide so little reliable data, how-
ever, that several possibilities still exist, and no matter how clever, experienced, or well-
trained the investigator may be, those fires will remain undetermined. The case shown in
Figures 17-8–17-11 is a good example of testing alternative hypotheses during a thorough
scene examination.

FIGURE 17-9 The most

heavily damaged area
was in the sanctuary
along the first row of
pews, where the entire
length of the pew and
the floor area beneath
was destroyed. Used with
permission of Retired Major
J. Ron McCardle, Florida State
Fire Marshal’s Office.

Chapter 17 Other Investigative Topics 729

FIGURE 17-10 After
many hours of reconstruct-
ing the scene, interviewing
witnesses including those
who reported seeing
someone in the area, uti-
lizing K-9s, and obtaining
assistance from LP gas
experts, a gas line was
found running beneath
the first row of pews.
Inspection of the floor
joists revealed fire patterns
with travel upward from
the ground level. During
the sifting and sampling of
debris beneath the floor
and testing the copper gas
lines for any loss of pres-
sure, a rupture was found
in the copper gas tubing.
Used with permission of Retired
Major J. Ron McCardle, Florida
State Fire Marshal’s Office.

FIGURE 17-11 The area was inspected from the bucket truck of the county electric company, and a chipped-
out section of a large pine tree was located. This appeared to have been hit by a lightning strike. The lightning
strike jumped to the top of an electrical pole at the point where the electrical ground was attached. It then
traveled the service drop to the meter box assembly, then eventually to the copper gas line, then onto grounded
support, where it breached the line below the first row of pews and ignited the gas. This lightning was con-
firmed by running a lightning strike report for that location and time. Used with permission of Retired Major J. Ron
McCardle, Florida State Fire Marshal’s Office.

730 Chapter 17 Other Investigative Topics

 FIGURE 17-12 An
analysis of the electrical
system of the church
revealed that the electrical
meter and meter base
appeared to have experi-
enced a surge of current
that traveled through the
grounding connection.
Used with Permission of Retired
Major J. Ron McCardle, Florida
State Fire Marshal’s Office.

CASE STUDY
A rural wood-frame church was found on fire at 7:34 P.M. one summer evening with fire venting from two win-
dows and the roof of the sanctuary (Figure 17-8). Interior fire damage included significant destruction of the first
pew and the wood floor beneath it (Figure 17-9). The primary hypothesis was that someone had forced entry by
one of the windows and set one pew on fire in an attempt to destroy the church. However, no residues of
ignitable liquid were detected. Excavation discovered a copper gas line running beneath the church floor with a
large rupture located beneath the floor under the burned pew. Examination of the electrical system revealed dam-
age to the electric meter, meter base, and grounding rod connected to the meter. When the the ground wire for
the meter was traced, the copper gas line was found lying atop the ground wire. Corrosion normal to the cop-
per pipe was cleaned off for several inches around the contact point (Figure 17-10). The electric company
responded with a “bucket” service truck. Seen from the bucket, a chipped-out section of a large pine was visi-
ble evidence of a recent lightning strike. The lightning jumped from the tree to the power pole, then via the serv-
ice drop to the meter box to the ground wire. The current then energized the copper gas line, which was
breached by arcing to a “grounded” support. That arcing breached the copper gas line and ignited the now-
escaping gas (Figure 17-11). The fire plume then burned through the church floor and onto the pew above. The
lightning strike hypothesis was confirmed by obtaining a StrikeNet report for that location and time.64

Chapter 17 Other Investigative Topics 731

 CHAPTER REVIEW

Summary
Every fire deserves careful and thorough investigation. Heat sources? Electrical lamps? Candles? Can the
Whether it is a small fire started by a faulty light cause of the fire be specifically ruled out as accidental?
switch, a commercial fire started by an arsonist, or a This “negative corpus” approach is very much
huge wildlands fire started by a careless camper, the dependent on the skill of the investigator in ruling out
origin and cause should be identified so that, possibly, all possible noncriminal sources of ignition. Was there
we can prevent such destruction from being repeated. more than one source of ignition at the scene? Use of
Incendiary fires inflict terrible damages on our society an ignition matrix such as that described in Chapter 6
via direct dollar losses, public costs, and personal will ensure that all combustions of ignition sources
damages and suffering. Absent positive proof of an and first fuels have been evaluated. In many cases, the
incendiary cause, the only way to be sure a fire was identification of several true points of origin is suffi-
incendiary in origin, is to be certain that all accidental cient indication of an arson-caused fire (as long as
and natural causes are ruled out using the ignition accidental causes, such as a power surge, are elimi-
matrix described in Chapter 7 and thorough applica- nated). A thorough knowledge of the physics and chem-
tion of hypothesis formation and testing. istry of fire, the behavior of materials when exposed to
With proper training, fire investigators can apply a fire, the heat release rates of furnishings and how they
methodical, analytical approach to all fires. The fire affect the fire’s development, sources of ignition, and
investigation can be broken down into two parts that the nature of electric appliances and electricity-related
are closely interrelated: the scene and its examination, phenomena are all essential to the accurate recon-
and the field or “legwork” investigation. The scene struction of the fire. Using recognized fire engineering
investigation is quite critical, because the physical evi- principles, the expert must be able to re-create the
dence present has a finite lifetime, and if the investiga- fire’s ignition and growth, accounting for as many of
tion is not begun soon enough, the materials of interest the observed indicators as possible, in a manner that
will simply disappear. A matter of hours is usually the is defensible from sound scientific and engineering
minimum delay involved; however, many investigations principles.
do not begin until days or weeks after the incident, thus Laboratory analysis can play an important part in
compromising vast quantities of information that could the origin and cause investigation. Comparison sam-
be obtained from a fresh fire scene. ples of floor coverings and other fuels from the scene
What does the scene investigator look for? One of can be identified by chemical or microscopic tests for
the major objectives is to locate the seat of the fire, evaluation of their ignition properties. Larger control
that is, the location where the fire itself began. samples (acquired from a manufacturer or retail
Numerous fire indicators such as char patterns, heat source but shown to be the same as those found at the
and smoke horizons, direction of fire spread, localized scene) can be subjected to various standard tests to
destruction of structural materials, and the like, are all demonstrate their response to ignition sources. Such
useful indicators to the trained investigator. These tests will often answer important questions like: What
must be thoroughly documented for later reference. temperatures and heat release rates are produced?
There are many new tools available to improve the What happens when the material is exposed to radiant
quality of documentation. heat rather than direct flame? Will these materials heat
Once at the seat or beginning of the fire, what spontaneously or degrade if exposed to particular
were the fuels in the vicinity? Which of these was the environmental conditions? What residues will be pro-
most likely to be ignited by virtue of its location or its duced (volatile and nonvolatile char) if such fuels
thermal properties? What kind of ignition source burn? Are such residues found in the fire debris? These
would be needed (flaming or glowing, large or small) are questions that the forensic or fire engineering lab-
to ensure ignition in the time frame indicated? What oratory can help answer.
sources were there in the vicinity? Could it have been The functions of photography and note-taking, as
accidental? Were there appliances? Electrical outlets? in any other type of crime scene, are, in fact, even more

732 Chapter 17 Other Investigative Topics

important on a fire scene. The physical relationships of that they can be reasonably relied upon as a basis for
objects and structures in fire scenes are often crucial. one’s expert opinion.C39 The condition of evidence has
These can be recorded only by proper photography and to be maintained so that opposing experts have an
crime scene sketching. opportunity to examine it.
Along with the scene investigation, the other In the absence of a thorough understanding of fire
equally important aspect of the investigation is field- evidence, many investigators will sample large por-
work. This is the area where police investigative talents tions of the scene entirely at random and submit large
are required. The arson investigator is required to inter- numbers of samples to the laboratory for screening.
view witnesses, possible suspects, and insurance inves- This “shotgun” approach is notoriously unsuccessful.
tigators and adjusters, and conduct background In many instances, the material of interest, such as a
investigation into the financial and psychological well- volatile flammable accelerant, is in hidden and remote
being of the individuals involved. Fundamental to the corners of the scene and not readily subject to random
proper conduct of a field investigation is the knowledge collection procedures. When canine or electronic
of what evidence is needed to substantiate a charge of hydrocarbon detectors are available to help screen
arson. In many cases this knowledge is lacking, and the samples for the ones most likely to yield identifiable
investigator cannot provide the prosecutor with the volatiles, they should be used. The investigator should
information he or she needs. Communications between be able to apply rules of common logic to the collec-
the fire investigators and the prosecutorial agencies are tion procedure before proceeding. Once the investiga-
the only means by which these gaps can be filled. It is tor is thoroughly familiarized with the properties of
helpful if the prosecutor has visited the scene, prefer- the materials with which he or she is dealing and the
ably during the fire or during the early stages of the complications that can be brought on by poor preser-
investigation. Many arson task forces that include vation, common sense will generally lead to the proper
prosecutors have provisions for calling out one of them packaging for any kind of evidence.
any time a fire involves major property losses or loss of The analysis of the material submitted is generally
life. The on-scene exposure improves the quality of the domain of the laboratory. The scientists in most
understanding among the team members, thereby crime laboratories are thoroughly trained and highly
improving the quality of the cases that may result. experienced in the application of all available scientific
Once the scene and field investigations are com- methodologies to the solution of analytical problems.
pleted, the physical evidence collected must be ana- Forensic science is the application of scientific proce-
lyzed in the laboratory. There are five aspects of dures to the analysis and interpretation of physical
physical evidence: documentation, collection, preser- events such as those encountered in fire scenes.
vation, analysis, and interpretation. The documenta- Without reliable interpretation of the bits and pieces of
tion of evidence before it is collected has to be information that were obtained from the fire scene, the
thorough so as to establish the chain of evidence con- investigator may still be left with unanswered ques-
clusively. Documentation should be complete, so that tions as to how, when, where, and by whom.
another qualified investigator can review it, establish The final test of any investigation is the courtroom—
agreement with the conclusions, and offer testimony whether civil or criminal. Today’s investigator needs to
regarding the scene. Although it is preferable that an have given the analysis and interpretation of his or her
investigator visit the scene before it is cleared, it is facts a great deal of consideration and be prepared to
sometimes not possible. In that case the quality of the demonstrate to the court how the conclusions were
documentation will determine the value and reliability reached, what the evidence was for each conclusion,
of later determinations. In one notable case in the and how other possible explanations were evaluated
authors’ experience, the fire investigator who con- and eliminated. The expert may be expected to prove
ducted the “interior” portion of a large scene exami- how the methods used to investigate the fire and reach
nation died before the case could be tried. An expert conclusions about it are soundly backed by acceptance
was brought in to testify as to the origin and cause of within the fire investigation community as well as sci-
the fire based upon the documentation of the previous entific data, and are not just opinions drawn from thin
examiner and verification by other investigators at the air (or hazy recollections). Whatever the investigator
scene. The court found that fire investigation experts can do to be prepared and improve communications
commonly rely on photographs, videotapes, diagrams, will be to everyone’s benefit. Only an accurately and
sketches, interview reports, and interviews with per- completely informed court can be expected to render
sons present at the scene to formulate an opinion and justice.

Chapter 17 Other Investigative Topics 733

Review Questions
1. What is the first step in ensuring personal safety 6. What was the major impact of Michigan v.
during a fire investigation? Clifford?
2. What are three of the basic types of fire modeling? 7. Name 10 sources of information an investigator
3. What are the major differences between zone can use in a fire investigation.
models and field (CFD) models? 8. Define spoliation as it applies to fire investigators—
4. What steps does the investigator need to take public and private.
before using a computer fire model? 9. What information should be contained in an
5. Name three ways a public investigator can gain acceptable fire expert’s report?
legal access to a property to conduct a search for 10. What are three of the major Supreme Court
evidence. decisions that affect fire investigators today?

Court Citations
C1. Camara v. Municipal Court of San Francisco, 387 C21. Farmers Ins. Exch. v. Superior Court, 79 Cal. App.
U.S. 523; 87 S. Ct. 1727 (1967). 4th 1400, (2000).
C2. See v. Seattle, 387 U.S. 541; 87 S. Ct. 1737 (1967). C22. Coprich v. Superior Court, 80 Cal. App. 4th 1081,
C3. United States v. Biswell, 406 U.S. 311; 92 S. Ct. 1593 (2000).
(1972). C23. Penn v. Prestige Stations, Inc. 83 Cal. App. 4th 336,
C4. Marshall v. Barlow’s, Inc., 436 U.S. 307; 98 S. Ct. (2000).
1816 (1978). C24. Trevino v. Ortega, 969 S.W.2d 950, 961 (Tex. 1998).
C5. Michigan v. Tyler, 436 U.S. 499; 98 S. Ct. 1942 C25. Temple Community Hosp. v. Superior Ct., 20 Cal.
(1978). 4th 464, 470 (1999).
C6. Michigan v. Clifford, 464 U.S. 287; 104 S. Ct. 641 C26. MetLife Auto & Home v. Joe Basil Chevrolet, Inc.,
(1984). 303 A.D.2d 30 (N.Y. App. Div. 2002).
C7. Mincey v. Arizona, 437 U.S. 385; 98 S. Ct. 2408 C27. Hannah v. Heeter, 584 S.E. 2d 560, 568–70 (W. Va.
(1978). 2003).
C8. Romero v. Superior Court, 266 Cal. App. 2d 714; 72 C28. Smith v. Atkinson, 771 So.2d 429 (Ala. 2000);
Cal. Rptr. 430 (1968). Holmes v. Amerex Rent-A-Car, 710 A.2d 846 (D.C.
C9. Swan v. Superior Court, 8 Cal. App. 3d 392; 87 Cal. 1998); Boyd v. Travelers Ins. Co., 652 N.E.2d 267
Rptr. 280 (1970). (Ill. 1995).
C10. United States v. Green, 474 F.2d 1385 (1973). C29. Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
C11. United States v. Delgado, 459 F.2d 471 (1973). C30. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509
C12. People v. Holloway, 86 Ill. 2d 78, 426 N.E.2d 871 U.S. 579; 113 S. Ct. 2786 (1993).
(1981). C31. United States v. Markum, 4 F.3d 891 (10th Cir.
C13. Passerin v. Delaware, 419 A.2d 917 (1980). 1993).
C14. Horton v. California, 496 U.S. 128, 110 S. Ct. 2301 C32. Michigan Millers Mut. Ins. Corp. v. Benfield, 140
(1981). F.3d 915 (11th Cir. 1998).
C15. United States v. Finnigin, 113 F.3d 1182, 1186 (10th C33. General Electric Company v. Joiner, 522 U.S. 136,
Cir. 1997). 188 S. Ct. 512 (1997).
C16. Smith v. Superior Ct., 151 Cal. App. 3d 491, 198 Cal. C34. Kumho Tire Co. v. Carmichael, 526 U.S. 137; 119 S.
Rptr. 829 (1984). Ct. 1167 (1999).
C17. Velasco v. Commercial Bldng. Maintenance Co., 169 C35. Weisgram v. Marley Co., 169 F.3d 514 (9th Cir.
Cal. App. 3d 874 215 Cal. Rptr. 504 (1985). 1999).
C18. Dowdle Butane Gas Co., Inc. v. Moore, 831 So.2d C36. People v. Sherman, 97 Cal. App. 2d 245; 217 P.2d
1124, 1128 (Miss. 2002). 715 (1950).
C19. Cedars-Sinai Med. Ctr. v. Superior Court 18 Cal. 4th C37. Pride v. Bic Corp., 218 F.3d 566 (6th Cir. 2000).
1, 74 Cal. Rptr. 2nd 248 (1998). C38. Truck Insurance Exchange v. Magnetek, 360 F3d
C20. Temple Community Hosp. v. Superior Court, 20 Cal. 1206 (2004).
4th 464, (1999). C39. United States v. Gardner, 211 F.3d 1049 (2000).

734 Chapter 17 Other Investigative Topics

References
1. NFPA 1033: Standard for Professional Qualifications 18. ASTM E1472-07: Guide for Documenting Computer
for Fire Investigator (Quincy, MA: National Fire Software for Fire Models (West Conshohocken, PA:
Protection Association, 2009), chap. 4, sec. 4.1.3. ASTM, 2007).
2. U.S. DOT, “Health Hazards of Combustion Products 19. ASTM E1355-05a: Standard Guide for Evaluating the
from Aircraft Composite Material,” DOT/FAA/AR- Predictive Capability of Deterministic Fire Models
98/34 (Washington, DC: Office of Aviation Research, (West Conshohocken, PA: ASTM, 2005).
September, 1998). 20. S. M. Olenick and D. J. Carpenter, “An Updated
3. KWJ Engineering, Inc., Newark, CA; http:www. International Survey of Computer Models for Fire and
kwjengineering.com. Smoke,” Fire Protection Engineering 13 (May 2003):
4. M. L. Donohue, Safety and Health Guidelines for Fire 87–110.
and Explosion Investigators (Stillwater, OK: Fire 21. H. E. Nelson, FPETOOL: Fire Protection Engineering
Protection Publications, 2002). Tools for Hazard Estimation, NISTIR 4380
5. J. W. Munday, Safety at Scenes of Fire and Related (Gaithersburg, MD: National Institute of Standards
Incidents (London: Fire Protection Association, 1994); and Technology, Center for Fire Research, October
available online. 1990).
6. NFPA 921: Guide for Fire and Explosion Investigations 22. H. E. Nelson, FPETOOL User’s Guide, NISTIR 4439
(Quincy, MA: National Fire Protection Association, (Gaithersburg, MD: U.S. Department of Commerce,
2008, 2011), chap. 12. National Institute of Standards and Technology,
7. D. Icove, V. Wherry, and R. Schroeder, Combating October 1990).
Arson for Profit, 2nd ed. (Columbus, OH: Battelle 23. M. L. Janssens, An Introduction to Mathematical Fire
Press, 1998). Modeling (Lancaster, PA: Technomic, 2000).
8. R. Friedman, Survey of Computer Models for Fire and 24. Ibid., 147–53.
Smoke, 2nd ed. (Norwood, MA: Factory Mutual 25. M. Spearpoint, F. W. Mowrer, and K. McGrattan,
Research Corp., 1991). “Simulation of a Compartment Flashover Using Hand
9. B. T. Hume, Fire Models: A Guide for Fire Prevention Calculations, Zone Models and a Field Model” in
Officers, Report No. 6/93 (London: Home Office Fire Proceedings ICFRE, Chicago, IL, 1999, 3–14.
Research and Development Group, 1993); H. E. 26. G. Rein et al., “A Comparison of Three Models for
Nelson, FPETOOL: Fire Protection Engineering Tools the Simulation of Accidental Fires,” Journal of Fire
for Hazard Estimation (Gaithersburg, MD: U.S. Protection Engineering, 16, no. 3 (2006): 183–209.
Department of Commerce, National Institute for 27. J. Sutula, “Applications of the Fire Dynamics
Standards and Technology, October 1990). Simulator in Fire Protection Engineering Consulting,”
10. B. Karlsson and J. G. Quintiere, Enclosure Fire Journal of Fire Protection Engineering 14 (Spring
Dynamics (Boca Raton, FL: CRC Press, 1999), 2002): 35–43.
chap. 10. 28. DeHaan, “Reliability of Fire Tests.”
11. D. Madrzykowski, “The Future of Fire Investigation,” 29. A. N. Beard, “Role Models,” Fire Risk Management
Fire Chief, October 1, 2000. Journal (August 2009): 21–22; www.frmjournal.com.
12. A. Brown, “A Review of Fire Modeling Computer 30. M. H. Salley et al., “Verification and Validation: How
Software: Branzfire V.99.005,” University of Central to Determine the Accuracy of Fire Models, Fire
Lancashire, UK, 2001. Protection Engineering 34 (Spring 2007): 34–44.
13. J. D. DeHaan, “Reliability of Fire Tests and Computer 31. N. Iqbal and M. H. Salley, Fire Dynamics Tools
Modeling in Fire Scene Reconstruction,” Part I: Fire (FDTs) (Washington, DC: U.S. Nuclear Regulatory
and Arson Investigator (January 2005): 40–45; Part II: Commission, November 2004).
(April 2005): 35–45. 32. S. Sardqvist, Initial Fires, Lund University, Lund,
14. D. J. Icove and J. D. DeHaan, Forensic Fire Scene Sweden, April 1993; J. F. Krasny, W. Parker, and
Reconstruction, 2nd ed. (Upper Saddle River, NJ: V. Babrauskas, Fire Behavior of Upholstered Furniture
Pearson/Brady, 2009). and Mattresses (Norwich, NY: William Andrews, 2001).
15. Ibid. 33. Icove and DeHaan, Forensic Fire Scene Reconstruction,
16. K. McGrattan et al., Fire Dynamics Simulator chap. 6.
(Version 5), Technical Reference Guide, vol. 1: 34. R. Paul and L. Elder, “The Guide to Scientific
Mathematical Model (Gaithersburg, MD: Fire Thinking” (Dillon Bay, CA: The Foundation for
Research Division, Building and Fire Research Critical Thinking, 2003).
Laboratory, April 6, 2010). 35. D. T. Moore, “Critical Thinking and Intelligence
17. ASTM E1895-07: Standard Guide for Determining Analysis,” Occasional Paper No. 14, (Washington,
Uses and Limitations of Deterministic Fire Models DC: Joint Military Intelligence College, May 2006),
(West Conshohocken, PA: ASTM, 2007). 8–10.

Chapter 17 Other Investigative Topics 735

36. Search Warrants (Sacramento, CA: California District 52. ASTM E678: Standard Practice for Evaluation of
Attorney’s Association, 1978); J. R. Davis, “The Scientific or Technical Data (West Conshohocken, PA:
Warrant Requirement in Crime Scene Searches,” FBI ASTM, 2007).
Law Enforcement Journal 47 (November–December 53. D. J. Icove and M. W. Dalton, “A Comprehensive
1978). Prosecution Report Format for Arson Cases” in
37. S. R. Rubenstein, “Michigan v. Clifford et al.,” Fire Proceedings ISFI International Fire Science, Cincinnati,
and Arson Investigator 35 (June 1985). OH, 2008; Icove, Wherry and Schroeder, Combating
38. E. K. Woods and D. H. Wallace, “Investigating Arson: Arson for Profit.”
Coping with Constitutional Constraints,” Security 54. M. Cabral, California Conference of Arson Investigators
Management, November 1991; reprinted in Arizona Seminar, November 2009.
Arson Investigator, February 1994. 55. R. D. Kolar, “Scientific and Other Expert Testimony:
39. M. J. Cabral, “The Legal Aspects of Arson Investigation: Understand It; Keep It Out; Get It In,” Appendix 1,
The Tyler/Clifford Legacy” in Proceedings International Federation of Defense and Corporate Council, FDCC
Symposium on the Forensic Aspects of Arson Quarterly (Spring 2007): 207–35.
Investigation, Washington, DC, August 1995. 56. E. Stauss and C. Hogan, “Three Experts Excluded:
40. J. O. Campane, “Chains, Wheels and the Single Court Examines Fire Expert Testimony under
Conspiracy,” FBI Law Enforcement Journal 50 Reliability Microscope,” Fire and Arson Investigator
(August 1981): 24–31; LEAA, Investigating Arson-for- (April 2001): 14–18.
Profit Schemes (Washington, DC: U.S. Department of 57. D. Wood, analysis of Pride v. Bic, private communication,
Justice; U.S. Government Printing Office, 1978). 2005.
41. FEMA, Arson Resource Directory (Emmitsburg, MD: 58. V. Babrauskas, Ignition Handbook (Issaquah, WA: Fire
U.S. Fire Administration, 1993). Science, 2003).
42. www.BATS.gov. 59. Icove and DeHaan, Forensic Fire Scene Reconstruction,
43. E. Strauss, “Paul Kenneth Keller,” National Fire and 49–53.
Arson Report 11, nos. 4–6 (1993); E. Comeau, 59. NFPA 1033, 2nd ed., 2009.
“Working Against Arson,” NFPA Journal 60. V. M. Brannigan, “Arson, Scientific Evidence and the
(January–February 2001): 65–68. Daubert Case,” Fire Chief (August 1998): 98–109;
44. Icove, Wherry, Schroeder, Combating Arson for Profit. V. M. Brannigan, “Kumho Tire Verdict: Daubert
45. Icove and DeHaan, Forensic Fire Scene Reconstruction, Applies to All Evidence,” Fire Chief (July 1999):
232. 98–109; V. M. Brannigan and J. Torero, “The Expert’s
46. P. A. Lynch, “After the Fire’s Out: Spoliation of New Clothes: Arson ‘Science’ after Kumho Tire,” Fire
Evidence and the Line Firefighter,” Fire Engineering, Chief (July 1999): 60–65.
January 1998. 61. J. D. DeHaan, “The Reconstruction of Fires Involving
47. B. S. Wilhoit, “Spoliation of Evidence: The Viability of Highly Flammable Hydrocarbon Liquids.” PhD disser-
Four Emerging Torts,” UCLA Law Review 46 (1998): tation, Strathclyde University, Glasgow, UK, 1995;
631. published by UMI, Ann Arbor MI, 1996; A. D. Putorti
48. 70 A.L.R. 4th 984 “Intentional Spoliation of Evidence, Jr., Flammable and Combustible Liquid Spill/Burn
Interfering with Prospective Civil Action, as Patterns (Washington, DC: National Institute of
Actionable” (and cases cited therein); 101 A.L.R. 5th Justice, Office of Justice Programs (NIJ/OJP), March
61 “Negligent Spoliation of Evidence, Interfering with 2001; A. D. Putorti Jr., Full-Scale Room Burn Pattern
Prospective Civil Action, as Actionable” (and cases Study (Washington, DC: NIJ/OJP), December 1997;
cited therein). J. H. Shanley, USFA Fire Burn Pattern Tests
49. Lynch, “After the Fire’s Out”; K. J. Clinkinbeard and (Washington, DC: U.S. Fire Administration, FEMA,
G. A. King, “Spoliation: Can the Investigator be Sued July 1997).
for Destruction of Evidence?” paper presented at ISFI 62. J. D. DeHaan, “Advanced Tools for Use in Forensic
2008: International Symposium on Fire Investigation Fire Scene Investigation, Reconstruction and
Science and Technology, Cincinnati, OH, 2008. Documentation” in Proceedings Interflam 2004
50. R. A. Dean, “Legal Issues Involved in Failure (London: Interscience Communications, 2004).
Analysis,” Journal of Failure Analysis and Prevention 63. R. S. Pepler, “An Atypical Cause of Fire: CCTV
4, no. 2 (April 2004): 7–16. Evidence to the Rescue” (Dallas, TX: AAFS, February
51. ASTM E620-04: Standard Practice for Reporting 2004).
Opinions of Technical Experts (West Conshohocken, 64. R. McCardle, Florida State Fire Marshal (retired), June
PA: ASTM, 2004). 2009.

736 Chapter 17 Other Investigative Topics

SUGGESTED READING

Chapter 1 Faraday, M. 1861. The Chemical History of a Candle.

Reprint, Atlanta: Cherokee, 1993.
Arson—The Neglected Epidemic: Report of the Senate Friedman, R. Principles of Fire Protection Chemistry and
Select Committee of Fire Services. California State Physics, 3rd ed. Sudbury, MA: Jones and Bartlett, 2008.
Senate, June 29–30, 1977. Kunzig, R. “Infernal Combustion.” Discover, January 2001,
Badger, S. G. “U.S. Multiple-Death Fires for 2007.” NFPA 35–36.
Journal (September–October 2008): 52–56. Meyer, Eugene. Chemistry of Hazardous Materials, 5th ed.
Bilancia, L. “The Scientific Method and Fire Investigation.” Upper Saddle River, NJ: Pearson/Brady, 2009.
The Fire Place, Washington IAAI (Spring 2005). Quintiere, J. G. Principles of Fire Behavior. Albany, NY:
Boudreau, J. F., G. Denault, and K. Quon. Arson and Delmar, 1998.
Arson Investigation: Survey and Assessment. Stauffer, E. J., A. Dolan, and R. Newman. Fire Debris
Washington, DC: Law Enforcement Assistance Analysis. Burlington MA: Academic Press, 2008.
Administration, October 1977.
DeHaan, J. D. “Arson: A Problem of Attitudes.” Washington Chapter 3
State Chapter IAAI Newsletter (April–May 1995).
Hazen, M. H., and R. M. Hazen. Keepers of the Flame: Babrauskas, V. “Estimating Room Flashover Potential.”
The Role of Fire in American Culture. Princeton, Fire Technology 16, no. 2 (May 1980).
NJ: Princeton University Press, 1992. Babrauskas, V. “HRR—The Role of Heat Release Rates
Heys, S., and A. B. Goodwin. The Winecoff Fire. Atlanta: in Describing Fires.” Fire and Arson Investigator
Longstreet Press, 1993. (June 1997): 54–57.
Home Office (UK). Safer Communities: Towards Effective Babrauskas, V., and D. Holmes. “Heat Release Rate.”
Arson Control. London, 1999. Part I: Fire Findings 6, no. 1 (Winter 1998): 7–11;
Icove, D. J. “Incendiary Fire Analysis and Investigation.” Part II: Fire Findings 6, no. 2 (Spring 1998): 2–11;
Open Learning Fire Science Program Course. Lexington, Part III: Fire Findings 6, no. 3 (Summer 1998).
MA: Ginn, 1983. Bales, R. F. The Great Chicago Fire and the Myth of
National Fire Academy. “Incendiary Fire Analysis and Mrs. O’Leary’s Cow. Jefferson, NC: McFarland, 2002.
Investigation, 2nd ed.” Emmitsburg, MD: FEMA, 1996. Browne, F. L. Theories of the Combustion of Wood
National Fire Protection Association. NFPA 921: Guide for Fire and Its Controls. Madison. WI: U.S. Forest Products
and Explosion Investigations. Quincy, MA: NFPA, 2008 Laboratory, Report 2136, U.S. Department of
and 2011. Agriculture, 1958. Available at http://www.fpl.fs.fed.us/.
National Institute of Justice. Fire and Arson Scene Evidence: Brannigan, F. L., and G. P. Corbett. Brannigan’s Building
A Guide for Public Safety Personnel. Washington, Construction for the Fire Service, 4th ed. Quincy, MA:
DC: U.S. Department of Justice, 2008 and 2011. National Fire Protection Association, 2007.
National Fire Protection Association. NFPA 1033: Standard Crombie, R. The Great Chicago Fire. Nashville, TN:
for Professional Qualifications for Fire Investigator Rutledge Hill, 1994.
Quincy, MA: NFPA, 2009. DeHaan, J. D. “Compartment Test Fires (1996–98).” CAC
Riggs, S. Quick Reference: Critical Tasks in Fire News (Spring 2001).
Investigations. Indianapolis, IN: Public Agency Drysdale, D. “The Flashover Phenomenon.” Fire Engineers
Training Council/Fire Science Training Institute. Journal (1998).
Rossotti, H. Fire: Technology, Symbolism, Ecology, Science, Fristrom, R. M. “The Mechanism of Combustion in
Hazard. Oxford: Oxford University Press, 1993. Flames.” Chemical and Engineering News 150
(October 14, 1963).
Gess, D., and W. Lutz. Firestorm at Peshtigo. New York:
Chapter 2
Henry Holt, 2002.
Babrauskas, V. Ignition Handbook. Issaquah, WA: Fire Gorbett, G. E., and J. L. Pharr. Fire Dynamics, Upper
Science Publishers, 2003. Saddle River, NJ: Pearson/Brady, 2011.
Brenton, J. R., and D. D. Drysdale. “Flames in Fires and Explo- Hamer, M. “The Night That Luck Ran Out.” New Scientist
sions.” Fire Engineers Journal 57, no. 189 (1997): 27–29. 7 (July 1988).
Drysdale, D. An Introduction to Fire Dynamics, 2nd ed. Karlsson, B., and J. G. Quintiere. Enclosure Fire Dynamics.
Chichester, UK: Wiley, 1999. Boca Raton, FL: CRC Press, 2000.
Drysdale, D. “Chemistry and Physics of Fire.” Chapter 2-1 Krasny, J. F., W. J. Parker, and V. Babrauskas. Fire Behavior
in Fire Protection Handbook, 19th ed. Quincy, MA: of Upholstered Furniture and Mattresses. Norwich,
NFPA, 2003. NY: William Andrew, 2000.

737
National Fire Protection Association. NFPA 555: Guide Debris Analysis. Burlington MA: Elsevier-Academic
on Methods for Evaluating Potential for Room Press, 2008.
Flashover. Quincy, MA: NFPA, 2000. U.S. Department of Agriculture. The Encyclopedia of
National Fire Protection Association. “Report on the Fire Wood. New York: Skyhorse, 2007.
at the Dupont Hotel and Casino—December 31, 1986.” White, R. H. “Wood-Based Paneling as Thermal Barriers.”
Quincy, MA: NFPA, 1987. Forest Products Laboratory Research Paper. FPL 408.
Society of Fire Protection Engineers. Predicting 1st and 2nd June 1982.
Degree Skin Burns from Thermal Radiation. Bethesda, “Wood and Wood-Based Products.” Chapter 8.3 in Fire
MD: SFPE, March 2000. Protection Handbook, 19th ed. Quincy, MA: NFPA,
2003.
Chapter 4
Chapter 6
Eaton, T. E. “Physical Properties of Gasoline.” Journal of
Applied Fire Science 1, no. 2 (1990–91): 175–83. Ackland, P., and A. E. Willey. Fire Investigation in
Ellis, W. H. “Solvent Flash Points, Expected and Commercial Kitchen Systems. Summerland, B.C.,
Unexpected.” Journal of Coatings Technology 48 Canada: Philip Ackland Holdings, 2005.
(March 1976): 614. ASTM International. ASTM E2187-02: Standard Test
Fried, V., H. F. Hameka, and U. Blukis. Physical Chemistry. Method for Measuring Ignition Strength of
New York: Macmillan, 1977. Cigarettes. West Conshohocken, PA: ASTM,
Glassman, I., and F. L. Dryer. “Flame Spreading across Liquid 2002.
Fuels.” Fire Safety Journal, 3 (1980–81): 123–38. Babrauskas, V. “Truck Insurance v. Magnetek: Lessons to
Harris, R. J. The Investigation and Control of Gas Be Learned Concerning Presentation of Scientific
Explosions in Buildings and Heating Plant. New York: Information.” California Fire/Arson Investigator
E & FN Spon, 1983. (October 2004): 6.
LaPointe, N. R., C. T. Adams, and J. Washington. Babrauskas, V., B. F. Gray, and M. L. Janssens. “Prudent
“Autoignition of Gasoline on Hot Surfaces.” Fire Practices for the Design and Installation of Heat-
and Arson Investigator (October 2005): 18–21. Producing Devices near Wood Materials.” Fire and
Leffler, W. L. Petroleum Refining in Nontechnical Language, Materials 31 (2007): 125–35.
3rd ed. Tulsa, OK: PennWell Corp., 2000. Babrauskas, V., and J. Krasny. Fire Behavior of Upholstered
National Fire Protection Association. NFPA 54: National Furniture. NBS Monograph 173. Gaithersburg, MD:
Fuel Gas Code. Quincy, MA: NFPA, 1999. U.S. Department of Commerce, National Bureau of
National Fire Protection Association. NFPA 58: NFPA Standards, 1985.
Liquefied Petroleum Gas Code. Quincy, MA: NFPA, Cuzzillo, B. R. Pyrophoria. Ph.D. dissertation, University
1998. of California Berkeley, 1997.
Damant, G. H. “Should Polyurethane Foam Be Banned?”
Chapter 5 Fire and Arson Investigator 42, no. 2 (December 1991):
48–52.
Colonna, G. P., and B. M. Cohn. “Plastics and Rubbers.” DeHaan, J. D. “Our Changing World: Part V—The Candle
Chap. 8.10 in Fire Protection Handbook, 19th ed. Threat.” Fire and Arson Investigator (January 2003):
Quincy, MA: NFPA, 2003. 28–39.
Fire, F. L. Combustibility of Plastics. New York: Van Flynn, B. “Trends in Home Heating and Cooking Fires.”
Nostrand Reinhold, 1991. NFPA Journal (March–April 2001): 63–65.
Fire, F. L. “Plastics and Fire Investigations.” Fire and Henderson, R. W., and G. R. Lightsey. “Safety Device for
Arson Investigator 36 (December 1985). Preventing Uncontrolled Flareup in Wick-Fed Liquid
Hilado, C. J. Flammability Handbook for Plastics, 5th ed. Fuel Burners.” U.S. Patent, 5,338,185, U.S. Patent
Lancaster, PA: Technomics, 1998. Office, August 16, 1994.
Marris, P. “Steel Determination.” Fire Engineering Journal Holleyhead, R. “Ignition of Solid Materials and Furniture
and Fire Protection (May 2005): 31–35. by Lighted Cigarettes: A Review.” Science and Justice
National Fire Protection Association. NFPA 705–Field 39, no. 2 (1999): 75–102.
Flame Test for Textiles and Films: Recommended Sanderson, J. L. “Kitchen Fires.” Fire Findings 5, no. 3
Practice. Quincy, MA: NFPA, 2009. (Summer 1997): 6.
Shipp, M., K. Shaw, and P. Morgan. “Fire Behavior of U.S. Fire Administration. Agricultural Storage Fires. Topical
Sandwich Panels Constructions.” Proceedings Interflam Fire Research Series, USFA 2, no. 7. Emmitsburg, MD:
1999. London: Interscience Communications, 1999. FEMA, January 2003.
Stamm, A. J. “Thermal Degradation of Wood and Cellulose.” U.S. Fire Administration. Heating Fires in Residential
Industrial and Engineering Chemistry 48 (1956). Structures. Topical Fire Research Series, USFA 2,
Stauffer, E., J. A. Dolan, and R. Newman. “Chemistry and no. 5. Emmitsburg, MD: FEMA, January
Physics of Fire and Liquid Fuels.” Chapter 4 in Fire 2001.

738 Suggested Reading

Chapter 7 Research Council of Canada, Division of Building
Research, NRCC 23065, Ottawa, Ontario, 1984.
Abrams, M. S. “Performance of Concrete Structures Lin, T. D., et al. “Fire Resistance of Reinforced Concrete
Exposed to Fires.” Skokie, IL: Portland Cement Columns.” Skokie, IL: Portland Cement Association,
Association, 1977. 1992.
Abrams, M. S., A. H. Gustfero, and E. A. B. Salse. “Fire Lyons, P. R. Techniques of Fire Photography, Quincy, MA:
Tests of Concrete Joist Floors and Roofs.” Skokie, IL: NFPA, 1978.
Portland Cement Association, 1971. Malhotra, H. L. “Spalling of Concrete in Fires.” Technical
ASTM International. ASTM E84: Standard Test Method Note 118. London: CIRIA, 1984.
for Surface Burning Characteristics of Building NIJ Technical Working Group on Fire/Arson Scene
Materials. Philadelphia, PA: ASTM, 2007. Investigation. Fire and Arson Scene Evidence: A
ASTM International. ASTM E860: Standard Practice for Guide for Public Safety Personnel. Washington, DC:
Examining and Testing of Items That Are or May National Institute of Justice, June 2000.
Become Involved in Product Liability Litigation. Portland Cement Association. Fundamentals of Concrete.
Philadelphia, PA: ASTM, 2007. Skokie, IL: Portland Cement Association, 2001.
Better Homes and Gardens. Home Designer Deluxe. Posey, E. P., and J. E. Posey. “Outline for Fire Scene
Coeur d’Alene, ID: Chief Architect, 2005. Documentation.” Fire and Arson Investigator 38
Buchanan, A. H. Structural Design for Fire Safety. (March 1988).
Chichester, UK: Wiley, 2002. Putorti, A. D., Jr. Flammable and Combustible Liquid
Chitty, R., and J. Fraser-Mitchell. Fire Safety Engineering: Spill/Burn Pattern. NIJ Report 604-00. Rockville, MD:
Reference Guide. London: IHS BRE Press, 2003. National Criminal Justice Reference Service, National
Cooke, R. A., and R. H. Ide. Principles of Fire Investigation. Law Enforcement and Corrections Technology Center,
Leicester, England: Institution of Fire Engineers, 1985. March 2001.
Corry, R. A. Pocket Guide to Accelerant Evidence Collection, Putorti, A. D., Jr. Full-Scale Room Burn Pattern Study.
2nd ed. Boston: Massachusetts Chapter, 2001. NIJ Report 601-97. Rockville, MD: National Criminal
Cramer, S. M., O. M. Friday, R. H. White, and G. Justice Reference Service, National Law Enforcement
Sriprutkiat. “Mechanical Properties of Gypsum Board and Corrections Technology Center, December 1997.
at Elevated Temperatures.” Proceedings: Fire and Sanderson, J. L. “Burn Pattern Testing.” Fire Findings 9,
Materials 2003. London: Interscience Communications, no. 3 (Summer 2001): 7–11.
pp. 33–42. Sanderson, J. L. “Floor-Level Burning: What Kinds of
Dietz, W. R. “Physical Evidence of Arson: Its Recognition, Damage Do Burning Accelerants Cause?” Fire Findings
Collection, and Packaging.” Fire and Arson Investigator 9, no. 1 (Winter 2001): 1–4.
(June 1991). Stensaas, J. P. “Determination of Fire Patterns Due to
Duncan, C. D. Advanced Crime Scene Photography. Boca Liquid Accelerants on Floor Covering.” Proceedings
Raton, FL: CRC Press, 2010. Interflam 2001. London: Interscience Communications,
Earls, A. R. “Manufactured Homes Revisited.” NFPA 2001, pp. 1491–96.
Journal (March–April 2001): 49–51.
Eaton, T. E. “Underfloor Fires.” Fire and Arson Investigator
Chapter 8
38, no. 1 (September 1987).
“High Rise Office Building Fires.” News about Fire. Quincy, Brown, D. J., Under a Flaming Sky: The Great Hinckley
MA: NFPA. Reprinted in Fire and Arson Investigator Firestorm of 1894. New York: Harper Perennial,
(September 1993). 2007.
Hutchinson, W., ed. Admission of Photographic Material. Cheney, P., and A. Sullivan. Grassfires: Fuel, Weather and
Elmhurst, IL: International Fire Photographers Fire Behavior. Collingwood, NSW 3066, Australia:
Association, 1981. CSIRO, 1997.
Ide, R. H. “Investigative Excavation at Fire Scenes.” Science Evans, D. D., R. G. Rehm, and E. S. Baker. Physics-Based
and Justice (Spring 1997): 210–11. Modeling for WUI Fire Spread: Simplified Model
Kitchen, A. “Concrete Ideas.” Fire Protection and Fire Algorithm for Ignition of Structures by Burning
Engineers Journal (October 2004): 40–41. Vegetation. NISTIR 7179. Gaithersburg, MD: NIST,
Kocisko, M. J. “Absorption of Ignitable Liquids into October 2004.
Polyethylene/Polyvinylidine Dichloride Bags.” Ford, R. T. “Fire Scene Investigation of Grassland and
Journal of Forensic Sciences 46, no. 2 (2001): Forest Fires.” Fire and Arson Investigator 38
356–62. (March 1988).
Lewko, J. “Camera Techniques at Fire Scenes.” Fire Ford, R. T. Investigation of Wildfires. Bend, OR: Maverick
and Arson Investigator (December 1983 and March Publications, 1995.
1984). Ford, R. T. (attributed to Jones, A. W.) “Investigation of
Lie, T. T., T. D. Lin, D. E. Allen, and M. S. Abrams. Fire Wildland Fires” in Arson Investigation. California
Resistance of Reinforced Concrete Columns, National District Attorneys Association, 1980.

Suggested Reading 739

Maclean, J. N. The Thirty Mile Fire. New York: Henry Fire and Materials 2005. London: Interscience
Holt, 2007. Communications, 2005, pp. 175–86.
Manzello, S. L., T. G. Cleary, J. R. Shields, and J. C. Yang. Cole, L. S. A Survey of Vehicle Fire Causes. Novato, CA:
“Ignition of Mulch and Grasses by Firebrands in Lee Books, 1988.
Wildland-Urban Interface Fires.” International J. Cole, L. S. Investigation of Motor Vehicle Fires, 4th ed.
Wildland Fire 15, no. 3 (2006): 427–31. San Anselmo, CA: Lee Books, 2001.
Manzello, S. L., S. H. Park, T. G. Cleary, and J. R. Shields. Cole, L. S., and R. C. Herold. The Investigation of
Developing Rapid Response Instrumentation Packages to Recreational Boat Fires. Novato, CA: Lee Books,
Quantify Structure Ignition Mechanisms in Wildland 1983.
Urban Interface (WUI) Fires. In Proceedings Fire and Davis, S., S. Kelly, and V. Somandepall. “Hot Surface
Materials 2009: 11th International Conference. London: Ignition of Performance Fuels.” Fire Technology 46
Interscience Communications, 2009, pp. 215–24. (April 2010): 363–74.
Maranghides, A., and W. Mell. “A Case Study of a Dodge, L., et al. “Development of a Dedicated Ethanol
Community Affected by the Witch and Guejito Fires.” Ultra Low Emission Vehicle (ULEV): Final Report.”
NIST Tech Note 1635, Gaithersburg, MD: NIST, Golden CO: National Renewable Energy Laboratory
April 2009. (NREL), September 1998.
Mitchell, J. W. “Power Lines and Catastrophic Wildland Gardiner, D., M. Bardon, and G. Pucher. “An
Fires in Southern California.” Proceedings Fire and Experimental and Modeling Study of the
Materials 2009. London: Interscience Flammability of Fuel Tank Headspace Vapors from
Communications, 2009, pp. 225–280. High Ethanol Content Fuels.” Golden, CO:
Rehm, R. G., Effects of Winds from Burning Structures National Renewable Energy Laboratory (NREL),
on Ground-Fire Propagation at the Wildland-Urban U.S. Department of Energy, October 2008.
Interface. Final Report. NIST GCR 06-892. Holleyhead, R. “Thermal Radiation and Ship Fires.”
Gaithersburg, MD: NIST, April 2006. Proceedings Ninth IAFS. Oxford, September 1984.
Steensland, P. “Long-Term Thermal Residency in Woody Merrall, S., and S. Chenery. “Vehicle Fires: Explaining the
Debris Piles.” Presented at International Association Rise in Vehicle Arson.” London: Office of Deputy
of Arson Investigators, Annual Training Conference, Prime Minister, February 2005.
May 2008. National Fire Protection Association. NFPA 556: Guide
TriData Corporation. “Urban Wildlands Fire.” Fire and on Methods for Evaluating Fire Hazard to Occupants
Arson Investigator 38 (March 1988). of Passenger Road Vehicles. Quincy, MA: NFPA (pro-
Tulloch, M. H. “Photonics Technology Assists Oakland posed for adoption 2010).
Firestorm Efforts.” Photonics Spectra (December National Fire Protection Association. NFPA 921: Guide
1991): 18. for Fire and Explosion Investigations. Quincy, MA:
U.S. Forest Service. Wildfire Cause and Origin NFPA, 2008 and 2011, Chap. 25.
Determination Handbook (FI-210 Course Manual). Shaw, A., W. Epling, C. McKenns, and B. Wichman.
Boise, ID: NWCG, May 2005. “Evaluation of the Ignition of Diesel Fuels on
Vaisala Global Atmospherics, 2705 E. Medina Rd., Hot Surfaces.” Fire Technology 46 (April 2010):
Suite 111, Tucson AZ 85706, (800) 283–4557, 407–23.
www.thunderstorm.vaisala.com. Thomas, G., and M. M. Witts. Shipwreck: The Strange
Wildfire Cause Determination Handbook: NFES 1874. Fate of the Morro Castle. New York: Dorset, 1972.
NWCG Handbook 1, PMS 412-1 (Including FI-110
and FI-210 course curricula), 2005. [National Chapter 10
Interagency Fire Center, Attn: Great Basin Cache
Supply Office, 3833 S. Development Ave., Boise, ID ASTM E2345-0: Standard Practice for Investigating
83705; Fax (208) 387-5573]. Electrical Incidents, Vol. 14.02. West Conshohocken,
PA: American Society for Testing and Materials, April
2004.
Chapter 9
Babrauskas, V. “How Do Electrical Wiring Faults Lead to
Bloom, J., and C. Bloom. Motorhome and Recreational Structure Ignitions?” In Proceedings Fire and Materials
Vehicle Fire Investigation, 2nd ed. Grants Pass, OR: Conference (San Francisco, 2001). London: Interscience
Bloom Fire Investigation, 2005. Communications, 2001, pp. 39–51.
Bourn, G., T. Callahan, L. Dodge, J. Mulik, D. Naegeli, K. Béland, B. “Electrical Damages—Cause or Consequence?”
Shouse, L. Smith, and K. Whitney. “Development of a Journal of Forensic Sciences 29 (July 1984): 747–61.
Dedicated Ethanol Ultra Low Emission Vehicle (ULEV): Béland, B. “Electricity as a Cause of Fires.” SFPE Technical
Final Report.” Golden CO: National Renewable Report 84–5, May 1989.
Energy Laboratory (NREL), September 1998. Béland, B., and C. Fortier. “Arc Tracking in Relation to
Brannigan, V., and E. Buc. “Testing the Fire Hazard of Fire Investigation.” Fire and Arson Investigator
Materials: Problem for Regulators.” Proceedings (March 1995).

740 Suggested Reading

Brannigan, F. L., R. G. Bright, and N. H. Jason, eds. Fire Furniture.” NIST Report NISTIR 4847. Gaithersburg,
Investigation Handbook. NBS 134, Washington, DC: MD: U.S. Department of Commerce, NIST, June 1992.
U.S. Department of Commerce. U.S. Government Damant, G. H., and S. Nurbakhsh, “Heat Release Testing
Printing Office, August 1980. of Furniture and Bedding Products.” Fifth Fire Safety
Brown, K. “Resistance Heating of Electrical Terminal Conference and Exhibition, Orlando, FL, February
Connections on Receptacles.” Berkeley, CA: Fire 1993.
Cause Analysis, 2000. McCormack, J. A., G. H. Damant, and C. J. Hilado.
Ettling, B. V. “Glowing Connections.” Fire and Arson “Flammability Studies of 700 Articles of Upholstered
Investigator 33 (March 1983): 3–5. Furniture.” Journal of Fire Sciences (September–
Gregory, G. D., and G. W. Scott. “The Arc-Fault Circuit October 1988).
Interrupter: An Emerging Product.” Institute of O’Malley, B. “Cigarettes and Sofas.” Mother Jones
Electrical and Electronics Engineers, Transactions of (July 1979): 56–63.
Industry Applications 34, no. 5 (September–October Sundstrom, B., ed. Fire Safety of Upholstered Furniture:
1998). The Full Report of the European Commission
Hagimoto, Y., K. Kinoshita, and T. Hagiwara. “Phenomenon Research Programme (CBUF). (London: Interscience
of Glow at the Electrical Contacts of Copper Wires.” Communications, 1995.
National Research Institute of Police Science Reports 41, Weiger, P. R. “Adapting Tests for Interior Finishes.” NFPA
no. 3 (August 1988). (Also at www.tcforensic.com.au). Journal (March–April) 2001.
The Investigation of Fires of Electrical Origin. Kent
County Constabulary, UK, 1992. Chapter 12
Korinek, C. W. “Classification of Electrical Overheating
Modes.” Fire and Arson Investigator (July 2001): Baker, W. E., et al. Explosion Hazards and Evaluation.
30–37. Amsterdam: Elsevier, 1983.
Lentini, J. J. “Examination of Televisions in Arson Fraud Béland, B., “An Unusual Explosion.” Fire and Arson
Investigations.” Fire and Arson Investigator 33 (June Investigator (October 2006).
1983): 15–22. Beveridge, A., ed. Forensic Investigation of Explosions.
Liebler, G. E. “Ground versus Earth: There Is a Difference.” London: Taylor & Francis, 1998.
Journal of Forensic Sciences 29, no. 2 (April 1984): Conkling, J. A. The Chemistry of Pyrotechnics. New York:
550–56. Marcel Dekker, 1985.
Malcolm, J. D. “Documenting the Electrical System.” Fire Explosives and Propellants from Commonly Available
and Arson Investigator 36 (December 1985). Materials. New York: Gordon, 1988.
Meese, W. J., and R. W. Beausoleil. Exploratory Study of FBI Bomb Data Center. 1997 Bombing Incidents.
Glowing Electrical Connections. Gaithersburg, MD: Washington, DC: U.S. Department of Justice, 1997.
National Bureau of Standards, October 1977. Harris, R. J. The Investigation and Control of Gas
National Fire Protection Association. National Electrical Explosions. New York: E & FN Spon, 1983.
Code. Quincy, MA: NFPA, 2008. Harris, R. J., M. R. Marshall, and D. J. Moppett. “The
National Fire Protection Association. NFPA 921. Quincy Response of Glass Windows to Explosion Pressures.”
MA: NFPA, 2008 and 2011, Chap. 8. Institute of Chemical Engineering Symposium, Series
Sanderson, J. L. “Exterior Electrical Exam.” Fire Findings No. 49, 1977.
10, no. 1 (Winter 2002): 1–3. Hough, E., “London: A Whole City Response.” Crisis/
Sanderson, J., and J. Finneran. “Floating Neutral Problems.” Response 1, no. 4 (2008): 9–17.
Fire Findings 3, no. 1 (Winter 1995): 9–11. Leslie, I. R. M., and A. M. Birk. “State of the Art Review
of Pressure Liquefied Gas Container Failure Modes
and Associated Projectile Hazards.” Hazardous
Chapter 11
Materials 28, no. 3 (1991): 326–65.
California Department of Consumer Affairs. Technical Martin, R. J., A. Reza, and L. W. Anderson. “What Is an
Bulletin 117: Requirements, Test Procedures and Explosive? A Case Study of an Investigation for the
Apparatus for Testing the Flame Retardance of Resilient Insurance Industry.” Journal of J. Loss Prevention in
Filling Materials Used in Upholstered Furniture. North the Process Industries 13 (2000): 491–97.
Highlands, CA: California Department of Consumer Meyer, R. Explosives. New York: Verlag Chimie-Weinheim,
Affairs, January 1980. 1977.
California Department of Consumer Affairs. Technical Midkiff, C. R., and R. E. Tontarski. “Detection and
Bulletin 133: Flammability Test Procedures for Characterization of Explosives and Explosive
Seating Furniture for Use in Public Occupancies. Residue: A Review.” 11th International INTERPOL
North Highlands, CA: California Department of Forensic Science Symposium, Lyon, France,
Consumer Affairs, 1988. November 1995.
Cleary, J., T. Ohlemiller, and K. Villa. “Influence of Ignition Newhouser, R. Introduction to Explosives. Gaithersburg,
Source on the Flame Fire Hazard of Upholstered MD: International Association of Chiefs of Police, 1972.

Suggested Reading 741

NIJ Technical Working Group for Bombing Scene Chapter 14
Investigation. A Guide for Explosion and Bombing
Scene Investigation. Washington, DC: National Institute Bertsch, W., Q. W. Zhang, and G. Holzer. “Using the Tools of
of Justice, June 2000. Chromatography, Mass Spectrometry and Automated
Quest’s Dynamic Fireball Radiation Model from the Quest Data Processing in the Detection of Arson.” Journal of
Quarterly. Issue 3. Norman, OK: Quest Consultants High Resolution Chromatography 13, no. 9 (September
Inc., Fall 1999. 1990): 597–605.
Smith, K. D., et al. “Detection of Smokeless Powder Residue Coulombe, R. “Chemical Markers in Weathered Gasoline.”
on Pipe Bombs by Micellar Electrokinetic Capillary Journal of Forensic Sciences 40, no. 5 (September
Electrophoresis.” Journal of Forensic Sciences 44, no. 4 1995): 867–73.
(1999): 789–94. Dolan, J. A., and E. Stauffer. “Aromatic Content in Medium
Stuart, D. “Shipboard Explosion and Fire.” Fire and Arson Range Distillate Products; Part I: An Examination of
Investigator (June 1994). Various Liquids.” Journal of Forensic Sciences 49, no. 5
Stull, D. R. Fundamentals of Fire and Explosion. New York: (September 2004): 992–1000.
American Institute of Chemical Engineers, 1977. Frontela, L., J. A. Pozas, and L. A. Picabea. “A
Thompson, R. Q., et al. “Aqueous Recovery from Cotton Comparison of Extraction and Adsorption Methods
Swabs of Organic Explosives Residues Followed by for the Recovery of Accelerants from Arson Debris.”
Solid Phase Extraction.” Journal of Forensic Sciences Forensic Science International 75 (1995), 11–23.
44, no. 4 (1999): 795–804. Frysinger, G. S. and R. B. Gaines.“Forensic Analysis of
Yallop, H. J. Explosion Investigation. Harrrogate, North Ignitable Liquids in Fire Debris by Comprehensive
Yorkshire, UK: Forensic Science Society Press, 1980. Two-Dimensional Gas Chromatography.” Journal
Yinon, J., and S. Zitrin. Modern Methods and Applications of Forensic Sciences 47, no. 3 (May 2002): 471–82.
on Analysis of Explosives. New York: Wiley, 1993. Fultz, M. L., and J. D. DeHaan, “Gas Chromatography
Zalosh, R. G. “Explosion Protection.” SFPE Handbook of in Arson and Explosives Analysis.” In Gas
Fire Protection Engineering, 4th ed. Quincy, MA: Chromatography in Forensic Science, edited by I.
NFPA, 2008, Chaps. 3–15. Tebbett. Chichester, UK: Ellis Horwood, 1992.
Zalosh, R., “Explosions.” Fire Protection Handbook, Gresham, J. “Investigators, Introducing the Mini Torch.”
20th ed. Quincy, MA: National Fire Protection Fire and Arson Investigator 47, no. 1 (September
Association, 2008, Vol. 1, Chap. 8, Sec. 2. 1996): 19–20.
Harris, A. C., and J. F. Wheeler. “GC-MS of Ignitable
Liquids Using Solvent-Desorbed SPME for Automated
Chapter 13 Analysis.” Journal of Forensic Sciences 48, no. 1
Christian, D. R. Forensic Investigation of Clandestine (January 2003): 41–46.
Laboratories. Boca Raton, FL: CRC Press, 2004. Kuk, R. J., and M. V. Spagnola. “Extraction of Alternative
Ellern, H. Modern Pyrotechnics. New York: Chemical Fuels from Fire Debris Samples.” Journal of Forensic
Publishing, 1961. Sciences 53, no. 5 (September 2008): 1123–29.
Federal Emergency Management Agency. “Six Firefighter Lennard, C. J., V. T. Rochaix, and P. Margot, P. “A GC-MS
Fatalities in Construction Site Explosion, Kansas City, Database of Target Compound Chromatograms for
Missouri (November 29, 1988)”, FEMA, USFA the Identification of Arson Accelerants.” Science and
National Fire Data Center. Justice 35, no. 1 (January 1995): 19–30.
Hildebrand, M. S., and G. G. Noll. Hazardous Materials Levinson, D. W. “Copper Metallurgy as a Diagnostic Tool
for Fire and Explosion Investigators. Chester, MD: for Analysis of the Origin of Building Fires.” Fire
Red Hat, 1998. Technology 13, no. 5 (August 1977): 211–22.
National Fire Protection Association. Fire Protection Guide Mat Desa, N. S., and N. Nic Daeid. “The Application of
to Hazardous Materials. Quincy, MA: NFPA, 2001. Multivariate Pattern Recognition for Petroleum-Based
National Fire Protection Association. NFPA 70: Standard Accelerant Identification and Classification.” Paper
System for the Identification of the Hazards of Materials presented at European Academy of Forensic Sciences,
for Emergency Response. Quincy, MA: NFPA, 2007. Glasgow, September 2009.
National Fire Protection Association. NPFA 497: McCord, B. R., K. A. Hargadon, K. E. Hall, and S. G.
Recommended Practice for the Classification of Burmeister. “Forensic Analysis of Explosives Using
Flammable Liquids, Gases, or Vapors and of Hazardous Ion Chromatographic Methods.” Analytica Chimica
(Classified) Locations for Electrical Installations in Acta 288, nos. 1–2 (March 1994): 43–56.
Chemical Process Areas. Quincy, MA: NPA, 2008. McDonald, K., and M. Shaw. “Identification of Household
Stephens, H. W. The Texas City Disaster––1947. Austin, Chemicals Used in Small Bombs via Analysis of Residual
TX: University of Texas Press, 1997. Materials.” Paper presented at American Academy of
U.S. Department of Health and Human Services. NIOSH Forensic Sciences, New Orleans, February 2005.
Pocket Guide to Chemical Hazards. Washington, DC: Midkiff, C. R. “Arson and Explosive Investigation.” In
U.S. Government Printing Office, June 1990. Forensic Science Handbook, 2nd ed., edited by R.

742 Suggested Reading

 Saferstein, 513–15. Upper Saddle River, NJ: Pearson Rho, Yong-Myun, M.D. “The Investigation of Fatal Fires.”
Education, 2002. FBI Law Enforcement Bulletin 55, no. 8 (August 1986).
Newman, R., M. Gilbert, and K. Lothridge. GC/MS Guide Rogers, D. “The Equivocal Fire Death.” Fire and Arson
to Ignitable Liquids. Boca Raton, FL: CRC Press, 1996. Investigator (September 1997): 14–16.
Phelps, J. L., C. E. Chasteen, and M. M. Render. Rossiter, W. D. “The Investigation of Fatal Fires.” Fire and
“Extraction and Analysis of Low Molecular Weight Arson Investigator ( January–March 1968).
Alcohols and Acetone from Fire Debris Using Passive Schmidt, C. W., and S. A. Symes. The Analysis of Burned
Headspace Concentration,” Journal of Forensic Human Remains. Amsterdam: Academic Press, 2008.
Sciences 39, no. 1 (January 1994): 194–206. Sopher, I. M. “The Role of the Forensic Pathologist in Arson
Sheff, L. M., and J. A. Siegel. “Fluorescence of Petroleum and Related Investigations.” FBI Law Enforcement
Products V: Three-Dimensional Fluorescence Bulletin (September 1972).
Spectroscopy and Capillary Gas Chromatography of Stuerwald, J. E. “Arson Homicide Investigation.” Fire and
Neat and Evaporated Gasoline Samples.” Journal of Arson Investigator (October–December 1970).
Forensic Sciences 39, no. 5 (September 1994):
1201–14. Chapter 16
Small, D. “Hot Evidence.” Investigative Practice Journal
(July 24, 2008): 27–29. Comeau, E., and R. F. Tunkel. “A Field Guide to Incident
Stauffer, E., and D. Byron, D. “Alternative Fuels in Fire Analysis.” National Fire and Rescue Magazine, May 2000.
Debris Analysis: Biodiesel Basics.” Journal of Forensic Doley, R. “Making Sense of Arson through Classification.”
Sciences 52, no. 2 (March 2007): 371–79. Psychiatry, Psychology and Law 10, no. 2 (2003): 346–52.
Geller, J. L., “Firesetting: A Burning Issue.” Chapter 9 in
Criminal Profiling: Principles and Practice. Totowa,
Chapter 15
NJ: Humana, 2006.
Artz, C. P., J. A. Moncrieff, and B. A. Pruitt Jr. Burns—A Huff, T. G. Fire-Setting Fire Fighters: Arsonists in the Fire
Team Approach. Philadelphia: W. B. Saunders, 1979. Department—Identification and Prevention. Quantico,
Bradtmiller, B., and J. E. Buikstra. “Effects of Burning on VA: National Center for the Analysis of Violent Crime,
Human Bone Microstructure: A Preliminary Study.” 1994. Published in On Scene, IAFC, August 1994.
Journal of Forensic Sciences 29, no. 2 (1984): 535–40. Icove, D. J. “Principles of Incendiary Fire Analysis.”
Brogan, R. “The Speed Street Fire—Was It Murder by or Ph.D. dissertation, College of Engineering, University
Prior to—the Fire?” Fire and Arson Investigator (July of Tennessee, Knoxville, 1979.
2000): 17–21. Icove, D. J., and M. W. Dalton. “A Comprehensive
Gamlin, L., and B. Price. “Bonfires and Brimstone.” New Prosecution Report Format for Arson Cases.” Paper
Scientist, November 5, 1988. presented at ISFI 2008, International Symposium on
Hill, I. P. “Immediate Causes of Death in Fires.” Medicine, Fire Investigation Science and Technology, Cincinnati,
Science and the Law 29, no. 4 (1989): 287–92. Ohio, May, 2008.
Landrock, A. H. Handbook of Plastics Flammability and Icove, D. J., V. B. Wherry, and J. D. Schroeder. Combating
Combustion Technology. Park Ridge, NJ: Noyes, 1983. Arson-for-Profit: Advanced Techniques for
Langlois, N. E. I., et al. “Gas Analysis Following the Investigators, 2nd ed. Columbus, OH: Battelle, 1988.
Diversion of Catalytically Converted Exhaust Fumes Kidd, S. “Arson: The Problem That Won’t Go Away.”
into the Interior of a Vehicle.” Forensic Science Fire Engineers Journal 58, no. 193 (March 1998):
International 74 (1995): 187–91. 12–17.
Leitch, D. A. Guide to Fatal Fire Investigations. Leicester, Little, C. M. “Recent Decisions Upholding Criminal
UK: Institution of Fire Engineers, 1993. Convictions for Use of Gasoline or Fire as a Deadly
Nickell, J., and J. F. Fisher. “Spontaneous Human Weapon.” Fire and Arson Investigator 50, no. 2
Combustion: The ‘Cinder Woman Mystery—Part II.’” (January 2000): 17–19.
Fire and Arson Investigator 34 (June 1984). Office of Deputy Prime Minister. The Burning Issue:
Penney, D. G., ed., Carbon Monoxide Toxicity. Boca Research and Strategies for Reducing Arson. London:
Raton, FL: CRC Press, 2000. ODPM, August 2002.
Pitts, W. M., E. L. Johnson, and N. P Bryner, “Carbon Rider, A. O. The Firesetter: A Psychological Profile, 2nd
Monoxide Formation in Fires by High Temperature ed. Edited by T. G. Huff. Washington, DC: U.S.
Anaerobic Wood Pyrolysis.” Twenty-Fifth Investigation Department of Justice, 1993.
Symposium on Combustion, The Combustion Institute, Sapp, A. D., and T. G. Huff. Arson Homicide: Findings
1994, 1455–-62. from a National Study. Washington, DC: National
Purser, D. A., “Assessment of Hazards to Occupants from Center for the Analysis of Violent Crime (NCAVC),
Smoke, Toxic Gases, and Heat.” Chap. 2–6 in SFPE U.S. Department of Justice, 1994.
Handbook of Fire Protection Engineering, 4th ed., Wood, B. “Arson Profiling—A Geographical, Demographic
Sec. 2. Quincy, MA: National Fire Protection and Motivational Perspective.” Fire Engineers Journal
Association, 2008. 60 (September 2000): 208.
Suggested Reading 743
Chapter 17 Hammett, T. M. “Cracking Down on Arson: Issues for
Investigators and Prosecutors.” Rockville, MD:
Arson Case Briefs. Washington, DC; U.S. Department National Criminal Justice Reference Service, 1987.
of the Treasury, Bureau of Alcohol, Tobacco, and Hewitt, T-D. “On the Hot Seat.” NFPA Journal
Firearms, November 1985. (January–February 2000): 43–46.
Arson Investigation and Prosecution, 2nd ed. Sacramento, Icove, D. J., and J. D. DeHaan. Forensic Fire Scene
CA: California District Attorney’s Association, 1995. Reconstruction, 2nd ed. Upper Saddle River, NJ:
Arson Prosecution—Issues and Strategies. Washington, DC: Pearson/Brady, 2009.
FEMA, U.S. Fire Administration (FA-78), May 1988. Iqbal, N., and M. H. Salley. Fire Dynamics Tools (FDTs).
Avato, S. J. “Scientific Methodology.” Fire and Arson NUREG 1805. Washington, DC: U.S. Government
Investigator (July 2000): 35–36. Printing Office, 2004.
Bukowski, R. W. “Modeling a Backdraft: The Fire at 62 Keierleber, J. A., and T. L. Bohan. “Ten Years After
Watts Street.” NFPA Journal 89, no. 6 (November– Daubert: The Status of the States.” Journal of Forensic
December 1995): 85–89. Sciences 50, no. 5 (September 2005): 1154–63.
Bukowski, R. W., and R. C. Spetzler. “Analysis of the Lentini, J. J. “A Calculated Arson.” Fire and Arson
Happyland Social Club Fire with Hazard I.” Fire and Investigator (April 1999): 20–25.
Arson Investigator 42, no. 3 (March 1992): 36–47. Little, C. M. “Recent Decisions Upholding Criminal
Decker, J. F., and B. L. Ottley. Arson Law and Prosecution. Convictions for Use of Gasoline or Fire as a Deadly
Durham, NC: Carolina Academic Press, 2009. Weapon, Dangerous Weapon or Dangerous Instrument.”
DeHaan, J. D. “Advanced Tools for Use in Forensic Fire Fire and Arson Investigator (January 2000): 17–19.
Scene Investigation.” In Proceedings Interflam 2004. Lynch, P. A. “Illinois Court Reverses Sanctions for
London: Interscience Communications, 2004, pp. Alleged Spoliation.” California Fire/ Arson
1079–1100. Investigator, July 2005.
Donahue, M. L. Safety and Health Guidelines for Fire McDowell, J. L., and G. C. Burton. “Computer Modeling
and Explosion Investigators. Stillwater, OK: OSU, and the Fire Investigator.” Fire and Arson Investigator
Fire Protection Publications, 2002. (March 1998): 10–12.
Drysdale, D. D. “Learning from Experience.” Fire McGrattan, K., et al. Fire Dynamics Simulator: Technical
Engineers Journal (November 2000): 34–39. Reference Guide. NISTIR 6467. Gaithersburg, MD:
Ellington, J. M. “Some Perspectives in Fire Modeling NIST, January 2000.
by a Fire Investigator.” Fire and Arson Investigator, Munday, J. W. Safety at Scenes of Fire and Related
June 1995. Incidents. Moreton-in-Marsh, UK: Fire Protection
Factory Mutual. A Pocket Guide to Arson and Fire Association, 1996.
Investigation. London, UK, Factory Mutual Nelson, H. E., and S. Deal. “Comparing Compartment
International and Fire Service College, 1994. Fires with Compartment Fire Models.” In Fire Safety
Factory Mutual Data Sheets, “Factory Mutual Global Science: Proceedings of the Third International
Property Loss Prevention Data Sheets,” To obtain Symposium, ed. G. Cox and B. Langford. London:
see: http://www.fmglobal.com/fmglobalregistration/, Elsevier Applied Science, 1991, 719–26.
Johnston, Rhode Island, last retrieved, March 4, 2011. Peacock, R. D., et al. CFAST—The Consolidated Model
Ferrall, R. T. “Arson Information: Who, What, Where?” of Fire Growth and Smoke Transport. NIST TN 1299.
FBI Law Enforcement Bulletin 50 (May 1981): 16–22. Gaithersburg, MD: National Institute of Standards
Fire Investigation Handbook. Palos Hills, IL: National and Technology, February 1993.
Insurance Crime Bureau, 1995. Quintiere, J. G. Principles of Fire Behavior. Albany, NY:
Giannelli, P. C., and E. J. Imwinkelreid. Scientific Evidence, Delmar, 1998.
4th ed. Newark, NJ: Lexis /Matthew Bender, 2008. Urbas, J. “Use of Modern Test Methods in Fire Engineering
Gorbett, G. E., “Computer Fire Models for Fire Investigation and Litigation.” Fire and Arson Investigator (December
and Reconstructon.” In Proceedings International 1997): 12–15.
Symposium on Fire Investigation Science and
Technology, Cincinnati, OH, May 2008, 23–34.

744 Suggested Reading

GLOSSARY

A B
absolute temperature Temperature above absolute backdraft A deflagrative explosion of gases and smoke
zero (in K or °R). from an established fire that has depleted the oxygen con-
accelerant A fuel (usually a flammable liquid) that is tent of a room or structure, most often initiated by intro-
used to initiate or increase the intensity or speed of ducing oxygen through ventilation or structural failure.
spread of fire. backing Slow extension of a fire downslope or into wind
accident Unplanned or unintentional event; an event in the opposite direction of its main spread.
occurring without design or intent. BLEVE Boiling-liquid, expanding-vapor explosion. A
accidental fire A fire occurring without design or mechanical explosion caused by the heating of a
intent. liquid in a sealed vessel to a temperature far above
adiabatic Conditions of equilibrium of temperature and its boiling point.
pressure. boiling point The (pressure-dependent) temperature at
adsorption Trapping of gaseous materials on the surface which a liquid changes to its gas phase.
of a solid substrate. branch circuit The circuit conductors between the out-
aliphatic Hydrocarbons with straight-chain structures of let(s) and the final overcurrent device protecting that
the carbon atoms; normal hydrocarbon. circuit.
alligatoring Rectangular patterns of char formed on brisance The amount of shattering effect that can be
burned wood. produced by a high explosive.
ambient Surrounding conditions. Btu British thermal unit. A standardized measure of heat,
ampacity Current-carrying capacity of electric conductors it is the heat energy required to raise the temperature
(expressed in amperes). of 1 pound of water 1 degree Fahrenheit.
ampere Quantity of electrical charge passing a point in buoyancy Tendency or ability to rise or float in air or
an electrical circuit per unit of time. liquid as a result of a difference in density.
annealing Loss of temper in metal caused by burn patterns Patterns created when applied heat fluxes
heating. are above the critical thresholds to scorch, melt, char
appliance Equipment, usually nonindustrial, that is or ignite a surface.
installed or connected as a unit to perform one or more
functions such as clothes washing, air conditioning,
food mixing, cooking, etc. Normally built in stan- C
dardized sizes and types.
arc Flow of current across an air gap (or another non- calcination Loss of water of crystallization caused by
conductor) between two conductors. heating.
area of transition Mixture of directional fire indicators cellulosic Plant-based material based on a natural poly-
in a wildlands fire. mer of sugars.
area of origin The general locale in which a fire was chain of evidence (chain of custody) Written docu-
ignited. mentation of possession of items of physical evidence
aromatic Hydrocarbon compound whose structure is from their recovery to their submission in court.
based on a benzene ring. char Carbonaceous remains of burned organic materials.
arson The intentional setting of a fire with intent to char depth The measurement of pyrolytic or combustion
damage or defraud. damage to a wood surface compared with its original
arson set Device, assembly, or contrivance used to ignite surface height.
an incendiary fire. chromatography Chemical procedure that allows the
atom Smallest unit of an element that still retains its separation of compounds based on differences in
properties. their chemical affinities for two materials in different
autoignition Ignition by surrounding temperature in the physical states, e.g., gas/liquid, liquid/solid.
absence of an external source of ignition; nonpiloted circuit breaker A device designed to open a circuit
ignition. automatically at a predetermined overcurrent without
autoignition temperature The temperature at injury to itself when properly applied within its ratings.
which a material will ignite in the absence of any clean burn An area of wall or ceiling where the charred
external pilot source of heat; spontaneous ignition organic residues have been burned away by direct flame
temperature. contact.

745
combustible A material that will ignite and burn when duty Conditions of use in electrical service: (1) continuous
sufficient heat is applied and when an appropriate duty—operation at substantially constant load for an
oxidizer is present. indefinitely long time; (2) intermittent duty—operation
combustible liquid A liquid having a flash point at or for alternate intervals of (a) load/no load; (b) load/rest; or
above 38°C (100°F). (c) load/no load/rest; (3) periodic duty—intermittent
combustion Oxidation that generates detectable heat operation in which the load conditions are regularly
and light. recurrent; (4) short-time duty—operation at substantially
concealed wiring Wiring rendered inaccessible by the constant load for a short and definitely specified time;
structure or finish of the building. Wiring in covered (5) varying duty—operation at loads, and for intervals of
raceways is considered concealed. time, both of which are subject to wide variation.
conduction Process of transferring heat through a material
or between materials by direct physical contact.
conductor Any material capable of permitting the flow E
of electrons: (1) bare—a conductor having no covering
or electrical insulation whatsoever; (2) covered—a elastomers Fibers or materials having elastic properties.
conductor encased within a material whose composition endothermic Absorbing heat during a chemical reaction.
or thickness is not considered insulative; (3) insulated— entrainment The mixing of two or more fluids as a
a conductor encased within a material recognized by result of laminar flow or movement.
the electrical code as insulation. eutectic An alloy of two materials having special physical
convection Process of transferring heat by movement of or chemical properties, typically having the lowest
a fluid. In convective flow, the warm fluid becomes less melting point of any combination of the two.
dense than the surrounding fluid and rises, inducing a exothermic Generating or giving off heat during a
circulation. chemical reaction.
corpus delicti Literally, the body of the crime. The explosion The sudden conversion of potential energy
fundamental facts necessary to prove the commission (chemical or mechanical) into kinetic energy with the
of a crime. production of heat, gases, and mechanical pressure.
crazing Stress cracks in glass as the result of rapid explosion-proof apparatus Apparatus enclosed in a case
cooling. that is capable of withstanding an explosion of a speci-
crowning Rapid extension of fire through the porous fied gas or vapor within it or designed to prevent the
array of leaves, needles, and fine fuels above 2 m. ignition of a specific gas or vapor surrounding it by
arcs, flashes, or explosion of fuels within, and that oper-
ates at such an external temperature that a surrounding
D flammable atmosphere will not be ignited by it.
explosive Any material that can undergo a sudden
deep-seated Fire that has gained headway and built up conversion of physical form to a gas with a release
sufficient heat in a structure to require great cooling of energy.
for extinguishment; fire that has burrowed deep into explosive limits (flammability limits) The lower and
combustible fuels (as opposed to a surface fire); deep upper concentrations of an air-gas or air-vapor mixture
charring of structural members. in which combustion or deflagration will be supported.
deflagration A very rapid oxidation with the evolution exposure Property that may be endangered by radiant
of heat and light and the generation of a low-energy heat from a fire in another structure or an outside fire.
pressure wave that can accomplish damage. The reac- Generally, property within 40 feet is considered an
tion proceeds between fuel elements at subsonic speeds. exposure risk, but larger fires can endanger property
detonation An extremely rapid reaction that generates much farther away.
very high temperatures and an intense pressure/
shock wave that produces violently disruptive effects.
It propagates through the material at supersonic F
speeds.
device Any chemical or mechanical contrivance or means fingerprints Unique friction ridge patterns on the palmar
used to start a fire or explosion. surfaces of the hands and fingers (or their impressions).
diatomic Molecules consisting of two atoms of an fire Rapid oxidation with the evolution of heat and light;
element. uncontrolled combustion.
direct attack Application of hose streams or other fire behavior The manner in which a fuel ignites, flame
extinguishing agents directly on a fire, rather than develops, and fire spreads. Unusual fire behavior may
attempting extinguishment by generating steam within reveal the presence of added fuel or accelerants.
a structure. fire load The total amount of fuel that might be involved
dropdown The collapse of burning material in a room in a fire, as measured by the amount of heat that would
that induces separate, low-level ignition; falldown. evolve from its combustion (expressed in units of heat).
746 Glossary
fire patterns Indicators of damage (or relative lack of G
damage) created by a combination of smoke, flames,
and heat effects. ghost marks Stained outlines of floor tiles produced by
fire-resistive A structure or assembly of materials built the dissolution and combustion of tile adhesive.
to provide a predetermined degree of fire resistance as glowing combustion The rapid oxidation of a solid
defined in building or fire prevention codes (calling fuel directly with atmospheric oxygen creating light
for 1-, 2-, or 4-hour fire resistance). and heat in the absence of flames.
firestorm Overwhelming progression of fire through ground A conducting connection, whether intentional or
structures or wildlands caused by a combination of accidental, between an electrical circuit or equipment,
convective and radiative processes. and the earth, or to some conducting body that serves
fire wall A solid wall of masonry or other noncombustible in place of the earth.
material capable of preventing passage of fire for a ground fault An interruption of the normal ground
prescribed time (usually extends through the roof return path of electricity in a structure that leads to
with parapets). unintended current flows.
flame A luminous cloud of burning gases.
flameover The flaming ignition of the hot gas layer in
a developing compartment fire. H
flame point Temperature at which a flame is sustained
by evaporation or pyrolysis of a fuel; fire point. heat horizon The demarcation (usually horizontal) of
flame resistant Material or surface that does not fire damage revealed by the charring, burning, or
maintain or propagate a flame once an outside source discoloration of paint or wall coverings.
of flame has been removed. heat of combustion The quantity of heat released from
flame spread The rate at which flames extend across the a fuel during combustion measured in kilojoules per
surface of a material (usually under specific conditions). gram (kJ/g) or Btu per pound (Btu/lb).
flaming combustion Rapid oxidation of gases and heat release rate The rate at which heat is generated
vapors that generates detectable heat and light. by a source, usually measured in watts, joules per
flammable (same as inflammable) A combustible mate- second, or Btu per second.
rial that ignites easily, burns intensely, or has a rapid heat transfer Spread of thermal energy by convection,
rate of flame spread. conduction, or radiation.
flammable liquid A liquid having a flash point below high explosive Any material designed to function by,
38°C (100°F). and capable of, detonation.
flanking Lateral spread of fire in a direction at right hot-plate ignition (hot-surface ignition) Temperature of
angles to the main direction of fire growth. a metal test plate at which liquid fuel ignites on contact.
flash point Temperature at which an ignitable vapor is hot set Direct ignition of available fuels with an open
first produced by a liquid fuel. flame (match or lighter).
flashback The ignition of a gas or vapor from an hydrocarbon Chemical compound containing only
ignition source back to a fuel source (often seen with hydrogen and carbon.
flammable liquids).
flashover The final stage of the process of fire growth;
when all exposed surfaces of combustible fuels I
within a compartment are ignited, the room is said
to have undergone flashover. ignitable liquid Classification for liquid fuels including
fragmentation The fast-moving solid pieces created by both flammable and combustible classes.
an explosion. Primary fragmentation is that of the ignition energy The quantity of energy that must be
explosive container itself; secondary fragmentation is transferred into a fuel/oxidizer combination to trigger
that of the target shattered by an explosion. a self-sustaining combustion.
fuel load All combustibles in a fire area, whether part ignition temperature (same as autoignition) The
of the structure, finish, or furnishings. minimum temperature to which a substance must be
fully involved The entire area of a fire building so heated in air to ignite independently of the heating
involved with heat, smoke, and flame that entry is source (i.e., in the absence of any other ignition
not possible until some measure of control has been source; nonpiloted ignition).
obtained with hose stream attacks. incendiary fire A deliberately set fire.
fraud Deception deliberately practiced in order to secure incipient Beginning stages of a fire.
unfair or unlawful gain. indicators Observable (and usually measurable) changes
furniture calorimeter An instrumental system for meas- in appearance caused by heat, flame, or smoke.
uring the amount of heat produced by the combustion indirect application A method of firefighting by applying
of a moderate-sized fuel package, and the rate at water fog into heated atmospheres to obtain heat
which the heat is produced. absorption and smothering action by generating steam.
Glossary 747
inorganic Containing elements other than carbon, oxygen, overhaul The firefighting operation of eliminating hidden
nitrogen, and hydrogen. flames, glowing embers, or sparks that may rekindle the
fire, usually accompanied by the removal of structural
contents.
J oxidation Combination of an element with oxygen;
chemical conversion involving a loss of electrons.
joule The SI unit of measurement of heat energy (1 J ⫽
0.238 cal ⫽ 0.00095 Btu).
P
L paraffinic Hydrocarbon compounds involving no double
or triple C—C bonds; alkanes; saturated aliphatic
ladder fuels Intermediate-height fuels between the
hydrocarbons.
ground litter and the crowns of trees overhead.
piloted ignition Ignition aided by the presence of a separate
low explosive Any material designed to function by
external ignition source such as a flame or electric arc.
deflagration.
piloted ignition temperature The minimum temperature
at which a fuel will sustain a flame when exposed to a
M pilot source.
plume The convective column of hot gases generated by
mass spectrometry Analysis of organic molecules by a flame.
fragmenting them and separating them by size. point of origin The specific location at which a fire was
Molotov cocktail A breakable container filled with ignited.
flammable liquid, usually thrown. It may be ignited pyrolysis The chemical decomposition of substances
by a flaming wick or by chemical means. through the action of heat, in the absence of oxygen.
motive Inner drive or impulse that causes a person to pyromania Uncontrollable psychological impulse to start
do something or act in a certain way. fires.
pyrophoric Capable of igniting on exposure to atmos-
pheric oxygen at normal temperatures.
N
naphthenics Class of hydrocarbons based on cycloalkanes. R
natural fibers Fibers of animal or vegetable origin with-
out chemical manipulation. raceway Any channel for holding wires, cables, or busbars
nonflammable Material that will not burn under most that is designed expressly for, and is used solely for,
conditions. this purpose.
normal hydrocarbons (n-alkanes) Hydrocarbons having radiation Transfer of heat by electromagnetic waves.
straight-chain structures with no side branching; aliphatics. receptacle A contact device installed as the outlet for
the connection of an appliance by means of a plug.
rekindle Reignition of a fire due to latent heat, sparks,
O or embers.
resistance Opposition to the passage of an electric current.
odorant Organic chemical (often a mercaptan) added to rollout Exiting of burner flames in an appliance due to
natural gas or LP gas prior to its distribution to make massive overfueling.
it detectable by a person with normal sense of smell at
a concentration less than 20% of its LEL.
olefinic Hydrocarbons containing double carbon-carbon S
bonds (denoted C “ C); nonsaturated hydrocarbons;
alkenes. salvage Procedures to reduce incidental losses from
open neutral In an American single-phase, 120/240 V smoke, water, and weather following fires, generally
residential system, the neutral return leg is not con- by removal or covering of contents.
nected to the service ground. saturated Hydrocarbons that have no double or triple
open wiring Uninsulated conductors or insulated C—C bonds.
conductors without grounded metallic sheaths or shields, seat of explosion The area of most intense physical
installed above ground but not inside enclosures or damage caused by high explosive pressures and shock
appliances; knob-and-tube wiring. waves in the vicinity of a solid or liquid explosive.
organic Compounds based on carbon. self-heating An exothermic chemical or biological
outlet A point on the wiring system at which current is process that can generate enough heat to become an
taken to supply equipment or appliances by means of ignition source; spontaneous combustion.
receptacles or direct connections. self-ignition See autoignition.
748 Glossary
service The conductors and equipment for delivering the equipment against overheating due to overload or
electricity from the supply system to the equipment failure to start.
of the premises served. thermoplastic Polymer that can undergo reversible
service conductors The supply conductors that extend melting without appreciable chemical change.
from the street main or transformer to the equipment thermosetting (resin) Polymer that decomposes or
of the premises served. degrades as it is heated rather than melts.
service drop The overhead service conductors from the pole, torch A professional fire setter.
transformer, or other aerial support to the service entrance trailers Long trails of fast-burning materials used to
equipment on the structure, including any splices. spread a fire throughout a structure.
short circuit Direct contact between a current-carrying
conductor and another conductor producing
unwanted current flow. V
smoke horizon Surface deposits that reveal the height at
which smoke and soot stained the walls and windows vapor The gaseous phase of a liquid or solid that can
of a room without thermal damage. be returned to that phase by the application of
smoldering combustion The direct combination of a pressure.
solid fuel with atmospheric oxygen to generate heat in vapor density The ratio of the weight of a given vol-
the absence of gaseous flames; see glowing combustion. ume of gas or vapor to that of an equal volume of
soot The carbon-based solid residue created by incomplete air.
combustion of carbon-based fuels. vented The fire has extended outside the structure or
spalling Crumbling or fracturing of a concrete or brick sur- compartment by destroying the windows or burning
face as a result of exposure to thermal or mechanical stress. an opening in the roof or walls.
spark Superheated, incandescent particle. ventilation A technique for opening a burning build-
spoliation The destruction or material alteration of, or ing to allow the escape of heated gases and smoke
failure to save, evidence that could have been used by to prevent explosive concentrations (smoke explo-
another in future litigation. sions or backdrafts) and to allow the advancement
spontaneous ignition Condition in which a chemical of hose lines into the structure; air supply to a room
or biological process generates sufficient heat to ignite fire.
the reacting materials. See self-heating. volatile A liquid having a low boiling point; one that is
spot fires Fires started by airborne embers some distance readily evaporated into the vapor state.
away from the main body of the fire. volt The basic unit of electromotive force.
stoichiometric mixture A reaction mixture in which the voltage, nominal A value assigned to a circuit or system
reactants and products are chemically balanced. for the purpose of conveniently designating its voltage
stoichiometry Balance of chemical reactants and products. class (120/240, 480Y/277, 600, etc.).
suspicious Fire cause has not been determined, but there
are indications that the fire was deliberately set and
all accidental fire causes have been eliminated. W
synthetic Material that is human-made, usually referring
to organic polymers. watt Unit of heat release, 1 watt ⫽ 1 joule/second; unit
synthetic fibers Manmade fibers of chemical origin. of power or work (in electrical circuits, equivalent to
voltage multiplied by amperes).
T
thermal protector A device against overheating that is
responsive to temperature or current and will protect

Glossary 749
This page intentionally left blank
INDEX

A Ammunition: fuels in, 673–81

exploding, 201 geographic profiling, 688–89
Abductive reasoning, 15 military, 200–201 ignitable liquids in, 674–81
Absorption, defined, 570 Ampacity, 413–15 ignition sources, 681–87
Accelerants: Ampere (amp), 407 incidence, 656–57
in arson, 679–81 Anhydrous ammonia, 550 information sources, 713–14
classification of, 573–74 Annealed furniture springs, 289–90, investigation, 687–91
gas chromatography flame 298 juvenile fire setting, 665
ionization detection (GC-FID), Annealing, defined, 289 laws about, 659
563–66 Antifreeze, 363 marine fires, 397–98
gas chromatography/mass Appliances: mass, 661
spectrometry (GS/MS), 566–69 condition of, 312 motive in, 661–71
on hands, 584 defined, 174 motor vehicle fire, 373
identification of, 562–84, 573–81 electrical, 186–87, 440–43 pyromania, 670–71
isolation of, 569–72 electronic, 452 serial, 661, 689, 713–14
in motor vehicle fire, 384 gas, 179–86 statistics on, 4–5, 656–58
source identification, 581–82 ignition source, 174–76 terrorism, 669–70
Accident, defined, 41 laboratory examination of, 590–91 tracking offenders, 689
Accidental fire: Aquarium lighting, 447 Arson set:
defined, 2 Arc: defined, 671
vs. arson, 690–91 defined, 174, 435 fuels, 673–81
Acetate, 471 electrical, 312, 403, 435–37, location, 672–73
Acetone, 544 453–54, 459, 685–86 Asbestos, 412–13, 471
Acetylated cellulose, 471 parting, 436 Asphalt tiles, 286–87
Acetylene, 25, 100, 541–42 spark, 174 ASTM, 659, 697
Acid bottle bombs, 514–17, 586 Arc-fault circuit interrupters (AFCIs), ASTM, fire modeling, 601–4,
Acrilan, 470 421–22 701
Activated charcoal, 132, 570–71 Arcing: ASTM E620, 690, 719
Adhesive, role of, 137 primary, 453 ASTM E678, 719
Adiabatic, defined, 104 secondary, 458 ASTM E860, 718
Adsorption, 570 through char, 431, 457–58 ASTM E1355, 702–3
Advective flow, 63 vs. alloying, 469 ASTM E1385, 572
Aerosol cans, 115–16 Arc mapping, 311–12, 454–57 ASTM E1386, 573
Affirmative exclusion, 322 Arc tracking, 431 (see also Carbon ASTM E1387, 565
Agricultural fertilizer, 550 tracking) ASTM E1412, 570
Aircraft imaging, 335 Area of origin, 250 ASTM E1413, 570
Alcohol fuels, 108, 680 Area of transition, 337 ASTM E1472, 702
Aliphatic compounds, 24 Aromatic compounds, 24 ASTM E1492, 561
Alligatoring, 281, 298 Aromatics, 545 ASTM E1537, 475
Alloying, vs. arcing, 469 Arrow patterns, 265 ASTM E1590, 475
Alternating current (AC), 407, 437 Arson, 655–92 (see also Incendiary ASTM E1618, 561, 565, 573
Aluminum, 159–60 fire) ASTM E1895, 702
melting point, 295 accelerant type, 679–81 ASTM E2154, 570
powder/plaster, 548 vs. accidental fire, 690–91 Atoms, 20–23
wiring, 437–38, 453 as crime, 6–8 Attic fires, 314–15
Ambient, defined, 35 crime concealment in, 668–69 Auger electron spectroscopy
American Board of Criminalistics defined, 4, 6, 656 (AES), 460
(ABC), 562 detectors for, 302–4 Autoignition temperature (AIT), 94
American Society for Testing and electric arcs in, 685–86 (see also Ignition temperature)
Materials (ASTM), 561 elements of proof, 658–60 of cooking oils, 185
Amines, 546 elimination of other ignition for plastics, 144–46
Ammonia, 542 sources, 310–13 of wood, 129
Ammonium nitrate–fuel oil (ANFO), evidence of, 298–304, 660–61 Automobile (see Motor vehicle)
548–49
751
Automobile exhaust, in carbon Buoyancy, defined, 35 flammability testing, 479
monoxide asphyxiation, 641–42 Buoyant flow, 63 as fuel, 152–53
Burn indicators: polypropylene, 154–55
in structure fires, 257–82 radiant heat ignition of, 270–71
B
in wildland fire investigation; 338–44 types of materials, 74
Babrauskas calculation, 60, 131 Burning fragments (see Sparks) Carpet pads, 74, 152
Backdraft, 41, 62, 502–3 Burn injuries, 645–48 Catalytic converters, 369
Back fires, intentional, 350 Burn patterns, 258–78 Cedars-Sinai Med. Ctr. v. Superior
Backsplash, 268–69 Butane, 24, 100, 541 Court, 716
Ballasts: in LP gas, 113–14 Ceiling construction, 74–76
electronic instant-start, 445 Button board, 78 Ceiling insulation, blown-in, 139–40
magnetic, 444–45 Ceiling jet, 63
Balloon frame construction, 71 Cellulose, 27, 140
C
Barbecues, 194, 640–41 acetylated, 471
Battery, 223 Calcination: Cellulosic:
defined, 404 defined, 288 building materials; 139–40
lead-acid, 449 gypsum board, 288–89 defined, 125
lithium, 449 California TB 116, 475, 481–482 Celotex, 139–40
motor vehicle, 368–69 California TB 117, 475, 481–482 16 CFR 1634 bedding, 474
Bedding: California TB 121, 475, 482 16 CFR 1630 carpets, 473, 479
cigarettes and, 203–5 California TB 129, 475, 482 16 CFR 1631 carpets, 473
flammability regulations, 474 California TB 133, 475, 482 16 CFR 1615 children’s sleepwear,
Benzene, 24, 544–45, 546 California TB 603, 474, 482 470, 473, 480–81
Benzine, 546 California TB 604, 474 16 CFR 1616 children’s sleepwear,
Bernoulli effect, 191 Calorimetry, 27 470, 473, 480–81
Beveling, 262, 264–65 Cadweld, 548 16 CFR 1610 clothing, 473, 478–79
in wildland fires, 338–40 Camara v. Municipal Court of San 16 CFR 1632 mattresses, 473–74, 479
Binary explosive, 511 Francisco, 710 16 CFR 1633 mattresses, 474,
Biodiesel, 109 Cambium, 331 479–80, 482
Biofuels, 108–9 Campfires, 349 16 CFR 1611 vinyl plastic film, 479
Birds, 223–24 Candle effect, 620 Chain of evidence, 320–21, 718
Black powder, 200, 506–8, 547, 686 Candle flame, 37 Chain reaction, 42
Blasting agents, 508, 511 Candles, 173 Chaparral, 331
Blasting caps, 512 Canine arson detector, 302–4 Char:
BLEVE, 120, 517–18 Carbohydrates, 26–28 arcing through, 431, 457–58
Blood analysis, 599–600, 633–34 Carbon dioxide molecule, 21–22 deep, 298
Blown-in ceiling insulation, 139–40 Carbon disulfide, 544–45 defined, 128
Blowtorches, 171–72 Carbonization, 431 depth, 279–81
Boat fires, 393–97 Carbon monoxide asphyxiation, laboratory examination of, 587–88
Boiling points: 635–43 surface, 281–82
defined, 98 absorption rates, 639–40 wood degradation to, 133–36
of liquid fuels, 98–99, 110 investigation of, 642–43 Charcoal, 137–38
Bolted fault, 416 physiological effects of, 635–37 activated, 132
Bomb Arson Tracking System sources of, 640–42 fire, 39
(BATS), 534–35 Carbon monoxide molecule, 22 from low-temperature ignition,
Bonfires, 197 Carbon tracking, 429–35 (see also 194–96
Boosters, 512 Arc tracking) sampling, 570
Bottle bombs, 514–17, 586 defined, 431 Char rates, for wood, 136–37
Brake fluid, 363 heating and, 434–35 Charring, in wildland fires, 338
Branch circuits, 422 leakage current, 431–32 Chemical explosions, 488–514
Brisance, defined, 511 water and, 432–34 Chemical fires, 539–57
Bromobenzene, 544–45 Carbonyl sulfide, 644 clandestine laboratories, 549–57
Btu, defined, 27 Carboxyhemoglobin, 635 gases, 540–43
Building construction, 69–81 Carburetors, 361 hazardous materials warnings,
fire resistance rating, 72–73 Carpet: 555–57
internal structure, 73–81 as fire hazard, 594 incendiary mixtures, 547–48
structural shell, 69–72 flammability regulations, 473 liquids, 543–46

752 Index
 oxidizing salts, 548–49 Combustible liquid, defined, 94 Consumer Safety Act, 472
reactive materials, 549 Combustion: Container:
solids, 546–49 of carbohydrates, 26–28 as arson evidence, 298–99
Chemical incendiaries, 58–87 carbon, 22 in explosion investigation, 523
Chemical reaction, as ignition conditions for, 33 plastic fuel, 90–93
source, 178 defined, 10, 489–90 in wildland fires, 355
Chemical Safety Board (CSB), 503 explosive, 41–42, 489–90, 505 Containers, for arson evidence, 326–29
Children’s sleepwear: flaming, 28 Content, in explosion investigation,
flammability regulations, 473 glowing, 38 523
flammability testing, 480 heat of, 27–28, 104–5 Continuous current rating, 417
synthetic, 470 hydrocarbons, 23–25 Convection, 46–47
Chimney fires, 190–94 oxidation reaction, 21 Convective flows, 46
Chipboard, 138–39 petroleum products, 25–26 Cooking oils, 185, 211
Christmas trees, 59, 68 of plastics, 143 Copper:
Chromatography, 532 (see also Gas pyrolysis, 28 electrical wire, 311, 453, 590–91
chromatography) role of atoms, 20–23 melting point, 295
Church arson, 668 smoldering, 33 Coprich v. Superior Court, 716
Cigarettes, 201–6 spontaneous (see Self-heating) Corpus delicti, 658–59
in arson, 685 spontaneous human, 620 Cotton, 469
bedding and furnishings, 203–5 turbulent, 38 Courtroom testimony, expert
burning characteristics, 201–3 of wood, 127–37 witness, 15–16, 721–26
in death investigation, 620 Commercial fraud, 662–63 Crazing:
flammable liquids and gases, Compact fluorescent lamps defined, 293
205–6 (CFLs), 446 of glass, 293, 297
in wildland fires, 349, 351–52 Compound, defined, 20 Crime, arson as, 6–8
Cigars, 206 Computer fire modeling (see Fire Crime concealment:
Circuit breaker, 411, 418 modeling) in arson, 668–69
laboratory examination of, Computer modeling: in explosion investigation, 523
462–63 structure fires, 698–707 Crime scene:
panel, 311 wildland fires, 353–54 possible, 6
Circuits: Concealment: security, 236, 238
branch, 422 in arson, 668–69 Crown fires, 126
parallel, 411 in explosion investigation, 523 Crowning, 330
series, 410 Conchoidal fracture lines, 290–91 Crown wildland fires, 329–30
Circumstantial evidence, 660 Concrete: Cryogenic liquid (CNG), 115–16
Clandestine laboratories, 549–57 floors, 73 Crystallography, 591
drug, 550–54 ghost marks, 286–88 Current:
explosives, 555 spalling, 283–86, 295–96 alternating, 407
marijuana cultivation, 554–55 Condensed-phase explosions, 505–7 defined, 404
Clean burn, defined, 281 chemical and physical properties, direct, 407
Clothing (see also Fabric): 506–7 electron, 405–6
children’s sleepwear, 470, 473, 480 vs. diffuse-phase, 505 ionic, 406
in death investigation, 618 Conduction, 45 leakage, 431–32
flammability regulations, 473 Conduction heating, 426–29 overheating by excessive, 426–27
flammability testing, 478–79 overheating by excessive current, Current-carrying capability, 413–15
ignition of, 472 426–27 Current electricity, 404–6
testing for residue, 320 overheating by poor connection, Custody, chain of, 320–21
Coal, 160 427–29 Cyanide poisoning, 638
fire, 39 Conductors: Cyclohexane, 25
self-heating of, 215 defined, 402 Cyclohexanone, 545
Coke, 137–38 electrical, 412–13 Cycloparaffins, 25
Color: in fire investigation, 452
of flames, 161–62 service, 420
D
of smoke, 162 Confidence, levels of, 15–16
of smoldering fire, 38–41 Connections, in explosion Dacron, 470
Color flag system, standardized, 336 investigation, 523 Daubert v. Merrell Dow
Combustible, defined, 28 Consent form, for search, 709 Pharmaceuticals, Inc., 721–23

Index 753
Dead short, 416 Digital photography, 239–40 Electrical circuit, 404–5
Death, 611–49 Dimethylamine, 546 Electrical circuits:
cause of, 615–16 Direct current (DC), 407, 437 parallel, 411
manner of, 616, 649 Direct evidence, 660 series, 410
Death investigation, 617–49 Dishwasher, 442 Electrical conductivity, 406
body destruction, 617–25 DNA analysis, 599–600, 615 Electrical conductors, 412–13
carbon monoxide asphyxiation, DOC FF 2-70, 473 Electrical conduits, 311
635–43 DOC FF 3-71, 473 Electrical explosions, 518
fire effects, 625–32 DOC FF 4-72, 473 Electrical fire investigation, 449–63
pathological examination, 616–35 DOC FF 5-74, 473 arc mapping, 454–57
species remains, 614 Documentation, 238–49 electric motors, 439–40
victim identification, 614–15 in burn injuries, 647 laboratory examination, 454–57
Decay stage of fire, 58, 62–63 of debris removal, 241–42 overcurrent protection devices
Deck materials, 139, 149–50 diagrams, 243–45 (OCPDs), 420
Decomposed wood, 130–31 in explosion investigation, 521 post-fire indicators, 450–54
Decomposition, liquid fuels, 112 fire modeling, 702–3 Electrical fires, 401–64
Deductive reasoning, 14 forms for, 247 Electrical fuses, 416–17
Deflagration, 501–3, 505 in motor vehicle fire investigation, Electrical ignition sources, 425–49
backdraft, 502–3 377–78 aluminum wiring, 437–38
damage due to, 525, 527 notes, 245, 247–49 appliances, 440–43
defined, 489–490, 501 photography, 238–43 arcs and sparks, 435–37
dust suspension, 503 report writing, 718–21 batteries, 449
explosion, 494 in wildland fire investigation, conduction heating, 426–29
Degradation: 338–44 electric blanket, 448
of electrical insulators, 429–35 Documents, laboratory examination electric motors, 438–40
of petroleum products, 575–78 of, 588–89 electric transformers, 438
wood to char, 133–36 Door hinges, 592–94 elimination of, 310–11
Delay devices, in arson, 681–82, Door locks, 592–94 extension cords, 448
686–87 Doors, 81 fixed heaters, 440
Dental records, in victim Downward spread, 251 heat tape or cable, 448–49
identification, 614–15 Draft conditions, in vapor insulation degradation, 429–35
Detection systems mapping, 313 explosions, 500–501 Electrical insulators, 412–13
Detonation, 505–7 Dropdown, defined, 264 degradation of, 429–35
damage due to, 527 Drug laboratories, 550–54 laboratory examination of,
defined, 488–90 Dryer: 461–62
high-yield, 513 electric, 442 Electrical resistance, 407–8
low-yield, 513 fire, 595 Electrical systems, 412–15
Detonation velocities, 501–2, 509 Drying oils, 209–13 ampacity, 413–15
Device, defined, 169 Drywall, 77–78 conductors, 412–13
Diagrams, 243–45 Duff wildland fires, 329–30 insulators, 412–13
in motor vehicle fire investigation, Dust explosions, 160–61 overcurrent protection devices
376–77 Dust suspension explosions, 503 (OCPDs), 416–22
in wildland fire investigation, 344–45 Duty, defined, 438 Electrical wiring, 311
Diatomic, defined, 20 Dynamic headspace, 571 Electric appliances, 440–43
Diesel fuel, 107 Dynamite, 507 laboratory examination of, 462–63
Diethylamine, 544, 546 gelatin, 508, 510 portable, 186–87
Differential diagnosis, 15 straight, 508–9 Electric blankets, 448
Differential scanning calorimetry Dynel, 470 Electric-gas powered motor vehicles,
(DSC), 592 368
Diffuse-phase explosions, 490–504 Electric heater:
E
deflagrations, 501–3 fixed, 440
gas, 490–97 Earthmover fire, 392–93 portable, 442
ignition, 503–4 Eco-terrorists, 670 Electricity:
vapors, 498–501 Eddy flow, 63 calculations for, 408–10
vs. condensed-phase, 505 Elastomers, 145 current, 404–6
Diffusion flames, 36 Electrical arcs, 312, 403, 427, frictional, 403
Diffusion rates, gas, 99–100 435–37, 459, 685–86 ignition source, 188

754 Index
 series circuits, 410 Expert witness, 15–16, 721–26 Federal Bureau of Investigation (FBI),
service distribution, 422–25 court cases about, 721–26 4, 6–7, 657
static, 403–4 pretrial preparation, 726–27 Federal Hazardous Materials
units of, 407–8 testimony, 727 Transportation System, 556–57
Electric lighting, 218–22, 443–48 Exploding ammunition, 201 Fenian fire, 549
Electric meter, 423 Explosion investigation, 519–35 Fiberboard, 215
Electric motors, 438–40 evidence, 530–32 Fiber evidence, 604
Electric outlet, 423–25 hypothesis, 530 Field models, 699–700
Electric sparks, 174, 403 incident analysis, 533–35 Fighter’s stance, 625
Electric transformers, 438 laboratory analysis, 532–33 Fingerprints, 596–99, 615
Electron current, 405–6 scene search, 520–23 Fire:
Electronic appliances, 452 speed and force of reaction, accidental, 2
Electronic arson detector, 302–4 523–29 defined, 2
Electronic instant-start ballasts, Explosions, 11, 487–535 flaming, 34–36
445 BLEVE, 120 incendiary, 3
Element, defined, 20 chemical, 488–514 smoldering, 38–41
Embers, 194 condensed-phase, 505–7 suspicious, 7
Energy density, 130 defined, 2, 489 Firearms:
Engine block heaters, 369–70 detonation velocities, 501–2 military ammunition, 200–201
Entrainment, 35, 52–53, 64 diffuse-phase, 490–504 residue, 200–201
Environmental conditions, effects of, dust, 160–61 Fireball, 47, 126
65–69 dust suspension, 503 Fire behavior, 10, 249–52, 257–96
Ethane, 99–100, 541 electrical, 518 annealed furniture springs, 289–90
Ethanol, 109 gas, 490–97 building construction, 69–81
Ethene, 24 high-order/low-order, 513–14 burn patterns, 258–65
Ethyl alcohol, 543–44 low-yield, 506 char depth, 279–81
Ethylene, 542 mechanical, 514–18 combustion, 33–34
Ethylene oxide, 542–43 pre-fire vs. trans-fire, 527, 529 defined, 10
Ethyl ether, 544–45 smoke, 62, 502–3 environmental conditions, 65–69
Eutectic, defined, 437 speed and force of reaction, explosive combustion, 41–42
Evaporation, of petroleum products, 523–29 flame plume, 51–54
575–80 vapors, 498–501 flame structure, 36–38
Evidence: Explosive: flaming fire, 34–36
of arson, 298–304, 687–91 binary, 511 floor burns, 276–79
chain of, 320–21, 718 defined, 489 gas flow, 63–64
circumstantial, 660 high, 507 ghost marks, 286–88
defined, 660 improvised, 511 glass, 290–95
direct, 660 low, 507 gypsum board calcination, 288–89
general fire, 587–96 Explosive combustion, 41–42 heat, 42–51
search and seizure, 708–13 Explosive limits, 88–91, 160 heat level, 265–68
spoliation, 595–96, 714–18 Explosives laboratory, 555 low burns, 268–76
trace, 603–5 Extension cords, 448 room fire sequence, 54–63
Evidence collection, 316–21 Extremism, 669–70 rules of, 250–52
in explosion investigation, smoke level, 268
530–32 smoldering fire, 38–41
F
liquids, 319 spalling, 283–86
preservation, 317–18 Fabric, 467–84 wall displacement, 282–83
solids, 319 flammability testing, 478–84 wildland fires, 332–33
testing of clothing, 320 furniture testing, 476–78 Firebug, 671
testing of hands, 319–20 natural, 469 Fire Dynamics Simulators (FDS),
volatile, 316, 318 non-petroleum-based synthetic, 471 701–2
in wildland fires, 354–56 petroleum-based synthetic, 469–70 Firefighter:
Excitement, in arson, 665–66 regulation of flammable, 472–76, interview of, 253–54
Exclusionary rule, 708–13 646 role in fire investigation, 236–37
Exothermic, defined, 21 Failure analysis, 589–90 Fire investigators:
Expert opinion, 15 Farmer’s Ins. Exch. v. Superior job performance requirements, 2–3
levels of confidence, 15–16 Court, 716 role of, 5–6

Index 755
Fire load, defined, 457 protected areas, 304–10 Flammable fabrics, regulation of,
Fire modeling, 698–707 protocol, 233–34 472–76
ASKFRS, 698, 703 purpose of, 232 Flammable Fabrics Act, 472
ASTM guides, 701 report writing, 718–21 Flammable liquids (see also Ignitable
BRANZFIRE, 699, 702 safety and health, 696–98 liquids):
CFAST, 699 scientifically based, 10–16 cigarettes and, 205–6
CFD, 699–702 search patterns, 255–57 defined, 36, 94
critical analysis, 707–8 timelines, 315–16 vapor density, 498–99
documentation, 702–3 wildland, 333–45 vapor explosions, 41, 498–501
evaluation of, 701–6 witness interviews, 254–55 Flashback, 41, 62
field models, 699–700 Firestorm, defined, 47 Flashover:
FPETool, 698–99, 702–3 Fire tetrahedron, 33–34 damage patterns, 75–76, 270–72,
HAZARD-I, 699 Fire triangle, 33 278
JASMINE, 699 Firewhirl, 47, 126 heat release rate to trigger, 60
mathematical, 698–99 Fireworks, 504 progression to, 298
specialized applications, 700–701 Firing trains, 513 stage of fire, 56–61
testing complex, 706–7 Flame: Flash point:
zone models, 699 defined, 21 of cooking oils, 185
Fire patterns, 249–50, 257–58 diffusion, 36 defined, 87
Fireplace: laminar, 37 determination of, 94, 592
in carbon monoxide asphyxiation, premixed, 37 of liquid fuels, 91, 93–95, 110
640–41 of smoldering fire, 40 vapor pressure and, 87
sparks from, 190–94 structure of, 36–38 Flash powder, 506, 547, 585, 686
Fire point: Flame color, solid fuels, 161–62 Floor:
defined, 94 Flame impingement, direct, construction, 73–74
liquid fuels, 94–95 50–51, 57 coverings, 74
Fire protection engineers, 15 Flame ionization detection (FID), gas displacement, 282–83
Fire resistance ratings, 72–73 chromatography, 563–66 irregular damage to, 297
Fire-resistive, 72–73 Flameover, defined, 56 Floor burns, 276–79
Fire scene investigation: Flame plume, 51–54 Floor coverings (see Carpet)
arc mapping, 311–12 Flame point: Flues, 133
arson, 671–91 defined, 94 Fluorescent lights:
carbon monoxide asphyxiation, of liquid fuels, 94–95 compact, 446
642–43 Flame resistant, defined, 156 as ignition source, 221
critical analysis, 707–8 Flame spread, defined, 40 T5, 446
death, 617–49 Flame temperatures: Forensic Fire Scene Reconstruction, 2
debris removal, 241–42 gases and liquids, 105 Forensic laboratory services,
documentation, 238–49 plastics, 143 560–61
electrical fires, 449–63 wood, 136 Forklift fire, 392–93
electric motors, 439–40 Flaming combustion, defined, 28 Fractional incapacitation dose, 639
evidence collection, 316–21 Flaming fire, 34–36 Fragmentation, defined, 522
explosions, 519–35 Flammability limits, 88–91 Frank-Kamenetski technique, 194
fire behavior, 249–52, 257–96 closed system, 89–90 Fraud, 662–63
firefighter interviews, 253–54 defined, 160 commercial, 662–63
heavy equipment, 392–93 open system, 88–89 residential, 663–64
hypothesis testing, 321–22 Flammability regulations: Free-burning stage of fire, 56,
incendiary indicators, 296–98 bedding, 474 270–71
information sources, 713–14 carpets and rugs, 473 Free radicals, defined, 23
legal opinions, 16, 242 children’s sleepwear, 473 Freon, 550
mobile home, 390–92 clothing, 473 Friction, 176
motorhome, 387–90 mattresses, 473–74 Frictional electricity, 403
motor vehicle, 371–87, 373–87 upholstered furniture, 474–76 Frye v. United States, 721
overcurrent protection devices Flammability testing, 478–84 Fuel:
(OCPDs), 420 California regulations, 481–82 alternative, 108–9
phase of, 234–36 children’s sleepwear, 480 in arson, 673–81
pre-fire conditions, 249 federal regulations, 478–81 ladder, 329–30
preservation of scene, 236–38 Flammable, 22 physical properties of, 87–104

756 Index
 properties of, 10 Gas chromatography/mass Glass-fiber fabric, 471
states of, 28–30 spectrometry (GC/MS), 566–69, Global positioning satellite (GPS),
tall, 333 573 335, 344–45
types of, 86–87 comparison samples, 584 Glowing combustion, defined, 38
in wildland fires, 329–32 interpretation of results, 582–84 Glowing connection, 428
Fuel containers, plastic, 90–93 petroleum distillates, 581 Glucose, 27
Fuel gas, sources of, 112–15 petroleum products, 575–80 Glycerin, 27, 546
Fuel injection system, 361–62 sensitivity, 582–83 Glycerol, 27
Fuel lines, 361 source identification, 582 Glycol, 546
Fuel load, 54, 251 toxic gases, 644 Grasses, self-heating of,
Fuel moisture content, 67–68, Gaseous fuels, 30 213–15
331 cigarettes and, 205–6 Ground, defined, 402
Fuel packages, 59–60 flammability limits, 88–91 Ground fault, defined, 420
Fuel pump, 361 hydrocarbon, 104–8 Ground fault interrupters (GFIs),
Fuel system: sources of, 112–15 420–21
fires, 364–66 vapor density, 99–104 Gum turpentine, 544, 546
motor vehicle, 360–62 Gas explosions, 490–97, 501–3 Gyp rock, 77–78
Fuel tank: Gas heater, 179–80 Gypsum board:
connections, 361 back fire, 184 calcination of, 288–89
motor vehicle, 360–61 plugged vent, 184 ceiling, 75
Fully involved, defined, 142 portable, 185–86 walls, 77–79
Furnishings: Gas law relationship, 29
changes over time, 150–56 Gas lines, 112–13
H
cigarettes and, 203–5 failure due to heat, 118–20
as fire hazard, 594–95 leakage, 115–17 Halogen lamps, as ignition source,
protected areas under, 304–10 mechanical failure, 117–18 221–22
Furniture, 59–60 Gasoline, 106 Hands:
flammability regulations for in arson, 679 ignitable liquids on, 584
upholstered, 474–76 in chemical fires, 544, 546 testing for residue, 319–20
testing upholstered, 476–78 evaporation and degradation of, Hannah v. Heeter, 716
Furniture calorimeter, 478 575–80 Harvester fire, 392–93
Furniture springs, annealed, 289–90, home storage of, 91 Hate crime, 668
298 motor vehicle, 362 Hay, self-heating of, 213–15
Fuse box, 311 trailer, 271 Hay clinkers, 214–15
Fuses, 416–17, 462–63, 548, 585, winter-blend, 26 Hazardous gases, 540–43
683–84 Gasoline vapors, 100 compressed, 540
Fuzing system, 513 in boat fire, 394 hydrocarbons, 541–42
in motor vehicle fire, 360 Hazardous liquids, 543–46 (see also
Gas water heater, 181–82 Flammable liquids)
G
GC-FID, 563–66 Hazardous materials, 555–57
Gas: GC/MS (see Gas Federal Hazardous Materials
defined, 29 chromatography/mass Transportation System,
density of, 63 spectrometry (GC/MS)) 556–57
diffusion rates, 99–100 Gelatin dynamite, 508, 510 NFPA 704 system for, 555–56
flow of hot, 63–64 General Electric Company v. Joiner, Headspace:
hazardous, 540–43 721–23 dynamic, 571
LP. (see LP gas) Geographic profiling, 688–89 heated, 569–70
Gas appliances, 179–86 Ghost marks, 286–88 passive diffusion, 570
altered gas nozzle, 184 Glass, 290–95 Heat, 42–51
direct contact, 184 crazing, 293, 297 gas line failure due to, 118–20
insulation, 183–84 as evidence, 604 horizon, 268
open flame, 185–86 fracture lines, 290–91 rate of reaction, 42–43
plugged vent, 184 fragments, 292 temperature and, 43
Gas bottle bombs, 514–17 melting point, 295 water, 179–83
Gas chromatography, 562–69 shards, 293 Heat cable, 448–49
Gas chromatography flame ionization tempered, 294 Heat capacity, 43–44
detection (GC-FID), 563–66 Glass construction, 71–72 Heated headspace, 569–70

Index 757
Heater, 133 final, 15 incinerators, 197
back fire, 184 testing of, 321–22 interior vs. exterior, 314
in carbon monoxide asphyxiation, working, 14–15 kerosene heaters, 187
640–41 lightning, 216–19
electric, 442 lighters, 171
I
fixed electrical, 440 low-temperature, 194–97
gas, 179–80 Ideal detonation, 513 oil storage, 187–88
kerosene, 187 Ideal gas law, 29 plantings, 206
on-demand, 182 Ideal reaction, 39 portable electric appliances,
Heat flux, 45, 48–49 Ignitability, 43 186–87
Heat horizon, 268 Ignitable liquids: portable heaters, 187–88
Heat of combustion, 27–28, 104–5 in arson, 300–302, 674–81 radiant heat, 177–78
Heat release: behavior of, 273–76 secondary, 173–78
flashover stage of fire, 59 classification of, 573–74 self-heating, 207–15
incipient stage of fire, 54–55 clothing test for residue, 320 sparks, 174, 188–94
Heat release rate, 43–44 defined, 272 torches, 171–72
flashover stage of fire, 59 detectors for, 302–4 Ignition switches, 367
ignition sources, 170 gas chromatography flame Ignition temperature (see also
incipient stage of fire, 54–55 ionization detection (GC-FID), Autoignition temperature
to trigger flashover, 60 563–66 (AIT)):
Heatstroke, 644 gas chromatography/mass of cooking oils, 185
Heat tape, 448–49 spectrometry (GS/MS), 566–69 defined, 94
Heat transfer, defined, 45 on hands, 584 of liquid fuels, 94–97, 110
Heavy equipment fire, 392–93 hand test for residue, 319–20 Impedance, 408
Heskestad equation, 52 identification of, 562–84, 573–81 Impingement, 50–51, 57
Hexane, 103 isolation of, 569–72 Impressions:
High explosive, 507–15 low burns, 272–76 shoe, 602–3
categories of, 511 in motor vehicle fire, 384 tools, 600–602
characteristics of, 508 source identification, 582 Improvised explosive devices
compounds of, 512–13 Ignition: (IEDs), 511
defined, 507 clothing, 472 Incandescent lighting, 218–21,
gelatin dynamite, 508, 510 defined, 168 447–48
straight dynamite, 508–9 matches, 169 Incendiary fire (see also Arson):
High-order detonation, 513 primary, 168–73 defined, 3
High-order/low-order explosions, spontaneous, 128 detection, 6–8
513–14 of wood, 128–31 devices, 673–87
High-pressure liquid Ignition energy: identification, 584–87
chromatography (HPLC), 586 defined, 97 myths about indicators, 296–98
High-resistance connection, 428 liquid fuels, 97–98 problems with estimating, 8–9
Hobby fuse, 512, 684 Ignition matrix, 225–26, 310–11 wildland, 349–53
Horton v. California, 713 Ignition sources, 167–226 Incendiary mixtures, 547–48,
Hot ashes, 194 animal interaction with, 223–24 584–86, 686–87
Hot sets, 351 in arson, 681–87 Incident analysis, 533–35
Houseboat fire, 394 batteries, 223 Incinerator fires, 197
Humidity effects, 66–67 bonfires, 197 Incipient, defined, 54
Hydriodic acid, 550 candles, 173 Incipient stage of fire, 54–55
Hydrocarbons, 23–25, 541–42 chemical reaction, 178 Inductive reasoning, 12, 14
defined, 23 cigarettes, 201–6 Infrared photography, 335
normal, 24 cooking oils, 185 Infrared radiation, 47
types of, 24–25, 541–42 in diffuse-phase explosions, 503–4 Infrared spectroscopy, 574, 585
Hydrochloric acid, 643 electrical, 188, 425–49 Inorganic, defined, 23
Hydrogen, 542 electric lighting, 218–22 Insulation, gas appliances, 183–84
Hydrogen chlorine, 643 firearms residue, 200–201 Insurance fraud, 662–63
Hydrogen cyanide, 643 friction, 176 Intent:
Hypothesis: gas appliances, 179–86 in arson, 659–60
defined, 14 hot metals, 197–98 malicious, 660
in explosion investigation, 530 hot surfaces, 174–76 Intentional back fires, 350

758 Index
International Association of Arson Latent fingerprints, 596–99 tank systems, 114–15
Investigators (IAAI), 561 Latex foam, 146–48 vapor density, 102–3
Interviews: Lead-acid batteries, 449 Lubricating oils, 107, 211
firefighter, 253–54 Leakage current, 431–32
in motor vehicle fire, 386 Lean mixture, gas, 495–96
M
witness, 254–55, 295 Legal opinions, 16
Intumescents, 157 Levels of confidence, 15–16 M-80, 504
Ionic current, 406 Lightbulb: Macro-scale indicators, 335
Ions, defined, 404 fire behavior and, 294–95 Magnesium, 159, 549
Iron powder, 158 as ignition source, 218–22 MagneTek court decision, 135,
Isobutane, 24 Lightning, 216–19, 346, 348–49 724–25
Isopropyl alcohol, 544–45 Lighters, 171, 681 Magnetic ballasts, 444–45
Lighting, 443–48 Malicious intent, 660
aquarium, 447 Manufactured housing, 390–91
J
commercial, 446 Marijuana cultivation, 554–55
Joule, 27, 408 electronic instant-start ballasts, Marine fires, 393–97
445 boats, 393–94
fluorescent, 446 cargo, 395
K
as ignition source, 218–22 ships, 394–95
Kapak, 317–18, 532, 632, 676 incandescent, 447–48 tankers, 395–96
Kapok, 469 magnetic ballasts, 444–45 Marshall v. Barlow’s, Inc., 711
Kerosene, 107 metal halide, 447 Masonite, 139–40
in arson, 679 Linen, 469 Masonry:
in boat fire, 394 Linoleic acid, 27 construction, 69, 71
heaters, 187 Liquefied petroleum gas (see spalling, 283–86
Kerosene-fueled drip torches, 350 LP gas) Mass arson, 661
Kirk’s Fire Investigation, 2 Liquid: Matches, 169
Kumho Tire v. Carmichael, 721–23 combustible, 94 in arson, 681
defined, 28 in chemical fires, 547–48
flammable, 36, 94 in wildland fires, 351–52, 355
L
hazardous, 543–46 Mathematical fire modeling, 698–99
Laboratories, clandestine, 549–57 Liquid fuels, 29–30, 86 Mattresses, 153–54
Laboratory burners, 185 boiling points, 98–99 flammability regulations, 473–74
Laboratory services, 559–605 combustion of, 109–12 flammability testing, 479–80
blood, 599–600 flame point, 94–95 McCaffrey/Quintiere/Harkleroad
chemical incendiaries, 584–87 flash point, 91, 93–94 (MQH) calculation, 60
chromatography, 532 fuel gas sources, 112–15 Mechanical explosions, 514–18
in electrical fire, 458–63 heat of combustion, 104–5 BLEVEs, 517–18
expert qualifications, 561–62 hydrocarbon, 104–8 bottle bombs, 514–17
in explosion investigation, 532–33 ignition energy, 97–98 Mechanical mixing, in vapor
failure analysis, 589–90 ignition temperature, 94–97 explosions, 499–500
fingerprints, 596–99 pyrolysis, 112 Mechanical sparks, 174, 198–99
fire testing, 561 vapors from, 86 Melting points, 295–96, 592
forensic laboratory services, Lithium batteries, 449 Metal construction, 71–72
560–61 Low burns, 268–76 Metal halide lighting, 447
gas chromatography, 562–69 Lower explosive limit (LEL), 88 Metallography, 459–60
general fire evidence, 587–96 Low explosive, 507–8 Metal oxide varistor (MOV), 419
impressions, 600–603 Low-order detonation, 513 Metals:
metallography, 459–60 Low-temperature ignition sources, as fuel, 158–60
physical matches, 603 194–97 as ignition sources, 197–98
spoliation, 595–96, 714–18 Low-yield explosions, 506 Methamphetamine laboratories,
trace evidence, 603–5 LP gas, 105 550–54
volatile accelerants, 562–84 in boat fire, 394 Methane, 23, 99–100, 541
Ladder fuels, 329–30 characteristics of, 113–15 Methyl alcohol, 543–44
Laminar flames, 37 explosion, 490–97 Methyl ethyl ketone (MEK), 544–45
Laryngeal spasm, 632 in motorhome fire, 388 Metlife Auto & Home v. Joe Basil
Laser scanning, 245–46, 521–22 pressurized containers, 115 Chevrolet, Inc., 716

Index 759
Mice, 223–24 National Fire Protection Association Origin:
Michigan Millers Mut. Ins. Corp. v. (NFPA), 4 area of, 250
Benfield, 721–23 National Highway Traffic Safety in fire scene investigation, 242
Michigan v. Clifford, 711–12 Administration (NHTSA), 386 Origin and cause, 2
Michigan v. Tyler, 711 National Incident Based Reporting in wildland fires, 333–45
Micro-scale indicators, 336–37 System (NIBRS), 4 Oven, electric, 442
Microtorches, 172 Natural fabrics, 469 Overcurrent protection devices
Mid-winding taps, 438 Natural gas, 100, 103–4, 113 (OCPDs), 416–22
Military ammunition, 200–201 Natural gas explosion, 490–97 circuit breakers, 418
Mincey v. Arizona, 712 Negative corpus, 322 fire scene investigation, 420
Mobile home fire, 390–92 NFPA: fuses, 416–17
Modular homes, 390–91 fire resistance ratings, 72–73 Overhaul, 40, 237–38
Modus operandi, 688 ignitable liquids classification, 94 Oxidation, defined, 21, 489
Moisture content, wildland fire NFPA 704 system for hazardous Oxidation reaction, 21, 489–90
fuels, 331 materials, 555–56 Oxidizing salts, 548–49
Molecule, defined, 20 NFPA 921, 2, 12, 15–16, 583, 662, Oxygen, 542
Molotov cocktail, 519, 669, 681–82 691, 701, 723 in explosions, 506
Molten metals: NFPA 1033, 2–3, 725–26 flaming fire and, 35–36
at fire scene, 295–96 Nitric oxide, 643 Oxygen cannulas, 276
as ignition sources, 197–98 Nitrocellulose, 507 Oxygen concentration effects, 68–69
Momentum flow, 63, 268–69 Nitroethane, 544–45 Oxygen masks, 276
Motive, in arson, 661–71 Nitrogen, 22 Oxyhemoglobin, 635
Motorhome fire, 387–90 Nitrogen compounds, in explosions,
Motors, electric, 438–40 506
P
Motor vehicle fire: Nitromethane, 544–45
arson, 373 Nitrous oxide, 643 Paint, 77, 156–58
combustible materials, 372 Nomex, 470 Panoscan, 245–46
electrical system, 367–69 Nominal voltage, 417 Paper:
fuels, 362–64 Nondistillates, 26 as fuel, 140–42
fuel system, 364–66 Nonideal detonation, 513 laboratory examination of,
Motor vehicle fire investigation, 373–87 Nonideal explosions, 506 588–89
arson, 373 Note taking, 245, 247–49 Paraffinic compounds, 24
checklist protocol for, 373–75 Nylon, 146, 470 Parallel circuits, 411
photography, 376–77 Particleboard, 138–39
scene preservation, 377–78 Parting arc, 436
O
sketches, 376–77 Passerin v. State of Delaware, 713
Motor vehicles, 360–87 Odorant, 113 Passive headspace diffusion, 570
CNG, 361 Ohm, 407–8 Penn v. Prestige Stations, Inc., 716
electrical system, 367–69 Ohm’s law, 408–10 Pentyne, 25
electric-gas powered, 368 Oils: People v. Holloway, 713
fuel system, 360–62 cooking, 185, 211 Petroleum, 106
heavy equipment, 392–93 drying, 209–13 Petroleum-based synthetic fabrics,
motorhome, 387–90 self-heating, 209–13 469–70
tempered glass, 294 Oil storage, 187–88 Petroleum distillates, 25–26, 110,
vehicle fluids, 362–64 On-demand heater, 182 579–81
Motor Vehicle Safety Standard 302, Open flame, gas appliance, 185–86 Petroleum gas:
372 Open neutral condition, 422 formula, 23
Mushrooming, 63, 268–69 Open wiring, 451 liquefied (see LP gas)
Opinion (see also Expert opinion), Petroleum products, 25–26,
defined, 15 110, 573
N
Organic, defined, 23 in chemical fires, 546
n-Alkanes, 24 Organic compounds, 23–28 evaporation and degradation of,
Naphtha, 544, 546 carbohydrates, 26–28 575–78
Naphthenics, defined, 26 hydrocarbons, 23–25 specialty, 108
National Arsonist & Bomber petroleum products, 25–26 Phosphorus, 549
Registry, 534–35 Oriented strand board (OSB), red, 550
National Electrical Code (NEC), 420 138–39 white, 586

760 Index
Photography, 238–43 Potassium permanganate, 585 Red phosphorus, 550
aircraft, 335 Power lines, in wildland fires, 346–47 Reducing conditions, 35
digital images, 239–40 Pre-fire conditions, reconstruction Rekindling, defined, 345
infrared, 335 of, 249 Remote Automated Weather Stations
in motor vehicle fire investigation, Premixed flames, 37 (RAWS), 355–56
376–77 Pride v. Bic Corp., 723 Report writing, 718–21
in wildland fire investigation, 335 Primary arcing, 453 Residence times, 175
Photoionization detection (PID), 574 Profiling, geographic, 688–89 Residential fraud, 663–64
Piloted ignition, defined, 185 Profit motive, in arson, 662–63 Resins:
Piloted ignition temperature, Propane, 100, 541 as fuel, 137
defined, 185 in LP gas, 113–14 thermosetting, 125
Pilot lights: in motorhome fire, 388 Resistance, defined, 169
in arson, 681 Propellants, 507–8 Retaliation, in arson, 667–68
in vapor explosions, 500–501 Propene, 25 Revenge, in arson, 667–68
Pipes, 206 Protected areas, 304–10 Rich mixture, gas, 495
Plantings, 206 Pseudoephedrine, 551 Road flares, 352–53, 548, 585
Plasterboard, 77–78 Pugilistic pose, 625 Rocks, in wildland fires, 341
Plastic building materials, 149–50 PVC, 145 Rodents, 223–24
Plastic fuel containers, 90–93 Pyrolysis: Rolling fire, 90
Plastics, 143–56 defined, 23, 28, 125 Rollout, 184
autoignition temperatures for, of liquid fuels, 112 Romero v. Superior Court, 712–13
144–46 of solid fuels, 125–26 Roof fires, 314–15
behavior of, 145–49 of wood, 28 Room fire sequence, 54–63
characteristics of, 143–45 Pyromania, 670–71 Rubber:
combustion of, 143 Pyrophoric, defined, 132 foam, 146–148
fire investigation considerations, Pyrophoric carbon, 132 vulcanized, 644
149–56 Pyrotechnic fuse, 512, 684 Rubbing alcohol, 544–45
in furnishings, 150–56 Rugs:
melting point, 295 Q flammability regulations, 473
in motor vehicle fire, 372 flammability testing, 479
Q10 value, 42
Plugged vent, heater, 184 Rule of nines, 645
Plume, defined, 35
R
Plywood, 79, 138
S
Polyester, 470 Radiant heat, 47–50
Polyether sulfone, 644 from fireplace, 190 Safety flares, 352–53, 548, 585
Polyethylene fabrics, 470 in floor burns, 276–79 Safety fuse, 512
Polymers (see Plastics) in floor covering ignition, 270–71 Salts, oxidizing, 548–49
Polypropylene: as ignition source, 177–78 Salvage, 236, 238
carpet, 154–55 Radiation, 47–50 Sandwich panel construction, 72
fabrics, 470 Rags, in motor vehicle fire, 370 Saturated compounds, 24
Polystyrene, 644 Railroads, 199, 349 Saturated vapor pressure, 87
fabrics, 470 Range, electric, 442 Scanning electron microscopy/energy
foams, 146–47 Rate of reaction, 42–43 dispersive X-ray spectrometry
Polyurethane foams, 146–47 Rats, 223–24 (SEM/EDS), 459, 581, 591
Polyvinyl chloride, 145 Rayon, 471 Scene:
Portable electric appliances, 186–87 Reaction, rate of, 42–43 in death investigation, 616–17
Portable gas heater, 185–86 Reactive materials, 549 in explosion investigation, 520–23
Portable heaters, 187–88 Reassembly, in explosion in motor vehicle fire investigation,
Portable stoves, 187–88 investigation, 522 377–78
Positive pressure ventilation (PPV), Receptacles, 411, 424–25, 451 safety and health at, 696–98
62, 251 Recognition, in explosion search patterns, 255–57
Post-flashover fire, 61 investigation, 522 security, 236, 238, 521
Post-flashover stage of fire, 58, Reconstruction, in explosion in wildland fire investigation, 335–38
61–62 investigation, 522 Scientific method, 11–15, 722–31
Postmortem examination, 618–19 Recovery, in explosion investigation, critical analysis, 707–8
Potassium, 549 522 defined, 11
Potassium chlorate, 586 Recreational vehicle fire, 387–90 steps in, 12–14

Index 761
Search and seizure, 708–13 Solid fuels, 29, 86–87 Stoichiometric reaction, 39
consent form, 709 coal, 160 Stoichiometry, defined, 39
court decisions, 710–13 dust explosions, 160–61 Stove:
Secondary arcing, 458 flame color, 161–62 electric, 442
Security, scene, 236, 238, 521 metals, 158–60 portable, 187–88
See v. City of Seattle, 710–11 paint, 156–58 Straight dynamite, 508–9
Self-heating, 207–15 paper, 140–42 Stripping, of vehicle, 378–79
characteristics of, 207–9 plastics, 143–56 Structural shell, 69–72
defined, 40 pyrolysis, 125–26 Suicide:
humidity and, 67 smoke production, 162–63 carbon monoxide asphyxiation,
oils, 209–13, 595 wood combustion, 127–37 642
of vegetation, 213–15 wood products, 138–40 by fire, 632–33, 649
Self-ignition (see Autoignition) Solid-phase microextraction (SPME), Sulfur-containing polymers, 644
Semiconductors, defined, 406 570–71 Sulfur dioxide molecule, 22
Serial arson, 661, 713–14 Solids: Sulfuric acid, 586
Series circuits, 410 in chemical fires, 546–49 Surface litter wildland fires, 329–30
Service, defined, 179 defined, 28 Surge protection device, 419
Service conductors, 420 thermal radiation effects on, 48 Swan v. Superior Court, 713
Service drop, defined, 437 Solvent extraction, 572 Swept headspace, 571
Sheetrock, 77–78 Solvents, 108 Synthetic, defined, 143
Ship fires, 394–97 types of, 543–46
arson, 397–98 in vapor explosions, 499
T
cargo, 395 Soot:
construction considerations, in death investigation, 635–43 Tall fuels, in wildland fires, 333
396 defined, 143 Tanker fire, 395–96
firefighting, 396–97 Spalling, 283–86, 295–96 Task forces, 713–14
Shoe impressions, 355, 602–3 Sparks: Television set, 442
Short circuit, 179, 416 bonfires, 197 Temperature:
Silk, 469 defined, 188 absolute, 29
Sketches, 243–45, 376–77 electric, 403 autoignition, 94
Skillet fires, 185 fireplace and chimneys, 190–94 effects on fire, 65–66
Slash wildland fires, 330 hot metals, 197–98 ignition and, 94–97
Sleepers, 345 ignition source, 174 of smoldering fire, 38–41
Slope considerations, in wildland as ignition source, 188–94 units of, 43
fires, 332 incinerators, 197 Temple Community Hosp. v.
Slurry-type blasting agents, mechanical, 67, 198–99 Superior Court, 716
508, 510 windblown, 189 Terrorism, 669–70
Smith v. Superior Court, 716 Specific gravity, of liquid fuels, 96 Testimony, courtroom, 15–16,
Smoke: Spoliation, 595–96, 714–18 721–76
black dense, 297–98 consequences of, 715–18 T5 fluorescent lamps, 446
explosions, 62, 502–3 private-sector investigations, 715 Theft, in motor vehicle fire, 386
horizon, 268 public-sector investigations, 715 Theft protection devices, 367
Smokeless powder: Spontaneous combustion (see Self- Thermal capacity, 43–44
double-base, 507–8 heating) Thermal conductivity, 43–46
single-base, 507–8 Spontaneous human combustion, 620 Thermal cutoffs, 179, 419
Smoking (see also Cigarettes): Spontaneous ignition, 128 Thermal inertia, 43–44, 168
cigars, 206 Spontaneous ignition temperature Thermal protector, defined, 419
as ignition source, 201–6 (SIT), 94 Thermal radiation, 48–49
pipes, 206 Spot fires, 342 Thermit, 548
Smoldering combustion, defined, 33 Spree, 661 Thermite, 548, 686
Smoldering fire, 38–41 Sprinkler systems mapping, 313 Thermogravimetric analysis
Smoldering phase, 40 Static electricity, 403–4 (TGA), 592
Social protest, 669–70 Steel wire springs, annealed, Thermoplastics, 145, 148–49
Sodium metal, 549 289–90, 298 fabrics, 483–84
Soil comparisons, 605 Stefan-Boltzmann relationship, 47–48 in motor vehicle fire, 373
Solar ignition, 686 Steam distillation, 571–72 Thermosetting plastics, 145
Solder, 198 Stewart equation, 636 Thermosetting resin, 125

762 Index
Thomas correlation, 60 Vapor pressure, 87–88 Wildland fires, 10–11, 327–56
Thrill seeking, in arson, 665–66 and flash point, 87 causes, 328–29, 345
Time fuse, 512 saturated, 87 fast-moving, 339
Tire impressions, 355 Vapors: fire behavior, 332–33
Tires, in motor vehicle fire, 371 defined, 21 fire spread, 329–31
Toluene, 24, 544–45 flammability limits, 88–91 fuels considerations, 329–32
Tools impressions, 600–602 from liquid fuel, 86 ignition sources, 345–54
Topography, in wildland fires, 329 “V” burn patterns, 259–63 intensity, 331–32
Torch, defined, 662 Vector analysis, 265 Windblown sparks, 189
Torches, 171–72 Vector patterns, 265 Wind effects:
Trace evidence, 603–5 Vegetation, self-heating of, 213–15 in fires, 68
Tractor fire, 392–93 Vehicle fire tests, 387 in wildland fires, 333
Trailer: Vehicle identification number Winter-blend gasoline, 26
as arson evidence, 298 (VIN), 382 Wire springs, annealed, 289–90,
defined, 271 Velasco v. Commercial Bldng. 298
gasoline, 271 Maintenance Co., 716 Wiring:
Trains, 199, 349 Veneer board, 138 aluminum, 437–38, 453
Transformers, electric, 438 Vent, heater plugged, 184 in arson, 686
Transmission fluid, 364 Vented, defined, 64 copper, 311, 453, 590–91
Trash fire, 306–8 Ventilation, 10, 38 in fire investigation, 450–51
Trash pile: Victim identification, 614–15 heat melting of, 452–53
as fuel, 673, 676–77 Vinyl plastic film, flammability laboratory examination of,
ignition source, 312–13 testing, 479 460–61, 590–91
Trauma, 644 Volatile, defined, 26 open, 451
Truck Insurance Exchange v. Voltage, 404, 407 Witness interviews, 254–55, 295
MagneTek, Inc., 135, 724–25 nominal, 417 Wood:
Turbochargers, 369 rating, 417 char depth, 279–81
Turbulent combustion, 38 Vulcanized rubber, 644 char rates for, 136–37
Turpentine, 544, 546 char surface, 281–82
W combustion of, 127–37
U components of, 127–28
Wall:
“U” burn patterns, 259, 261 decomposed, 130–31
construction, 77–81
Unconfined vapor cloud explosion degradation to char, 133–36
displacement, 282–83
(UVCE), 501 flame temperatures, 136
protected areas on, 308–9
Uniform Crime Reports (UCR), 6–7 ignition of, 128–31
Wallboard, 77
United Kingdom, fire statistics in, low-temperature ignition, 194–96
Wall factor, 52–53
4–5, 8 pyrolysis of, 28
Wallpaper, 77
United States, fire statistics in, 4–5 Wood construction, 71
Wastebasket fire, 306–8
United States v. Biswell, 711 Wood floors, 73–75
Water, and carbon tracking, 432–34
Upholstered furniture: Wood paneling, 78–80
Water gel dynamites, 508, 510
flammability regulations, 474–76 Wood products, 138–40
Water molecule, 21
testing, 476–78 Wood stoves, in carbon monoxide
Watts, 408
Upholstery, in motor vehicle fire, 383 asphyxiation, 640–41
Weather, in wildland fires, 329,
Upper explosive limit (UEL), 88 Wool, 469
355–56
Uranium, 158 Working hypothesis, 14–15
Weisgram v. Marley Co., 723
Urethane foam, flame-retardant, 476 Welfare fraud, 663 X
Urethanes, 643 White phosphorus, 586
Utilities, inspection of, 310 Wick effect, 620 X-ray analysis:
Wildland fire investigation, of appliances and wiring, 590–91
V in postmortem examination, 618
333–45
Vagal inhibition, 632 burn indicators, 338–44 in victim identification, 614–15
Valences, defined, 23 computer modeling, 353–54 Xylenes, 545
Vandalism, 664–665 documentation, 338–44
Z
Vapor density, 99–104 evidence collection, 354–56
Vapor explosions, 41, 498–501 first evaluation, 334–35 Zinc, melting point, 295
deflagration, 501–3 protocol, 334 Zinc die cast, 311
solvents, 499 scene search, 335–38 Zone models, 699

Index 763