Failure Analysis of Heat Treated Steel Components (2008) PDF

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_i.
pdf/Front Matter_Title-page/
12/8/2008 8:00PM Plate # 0
2008 ASM International. All Rights Reserved.

Failure Analysis of Heat Treated Steel Component (#05113G)
www.asminternational.org
Failure Analysis of
Heat Treated Steel Components
L.C.F. Canale
R.A. Mesquita
G.E. Totten
ASM International
Materials Park, Ohio 44073-0002
pg i
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_ii.pdf/Front Matter_Copyright/
12/8/2008 5:09PM Plate # 0

Copyright # 2008
by
ASM International
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any
form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the
written permission of the copyright owner.
First printing, September 2008
Great care is taken in the compilation and production of this book, but it should be made
clear that NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION,
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE
GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to
be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this
publication alone. This publication is intended for use by persons having technical skill, at their sole
discretion and risk. Since the conditions of product or material use are outside of ASMs control, ASM
assumes no liability or obligation in connection with any use of this information. No claim of any kind,
whether as to products or information in this publication, and whether or not based on negligence,
shall be greater in amount than the purchase price of this product or publication in respect of which
damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND
SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR
SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY
OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation
of the material under end-use conditions prior to specification is essential. Therefore, specific testing
under actual conditions is recommended.
Nothing contained in this book shall be construed as a grant of any right of manufacture,
sale, use, or reproduction, in connection with any method, process, apparatus, product, composition,
or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in
this book shall be construed as a defense against any alleged infringement of letters patent, copyright,
or trademark, or as a defense against liability for such infringement.
Comments, criticisms, and suggestions are invited, and should be forwarded to ASM
International.
Prepared under the direction of the ASM International Technical Book Committee
(20072008), Lichun L. Chen, Chair.
ASM International staff who worked on this project include Scott Henry, Senior Manager
of Product and Service Development; Steven Lampman, Technical Editor; Ann Britton, Editorial
Assistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production
Coordinator; Diane Grubbs, Production Coordinator; Patricia Conti, Production Coordinator; and
Kathryn Muldoon, Production Assistant
Library of Congress Control Number: 2008925435
ISBN-13: 978-0-87170-868-7
ISBN-10: 0-87170-868-X
SAN: 204-7586
ASM International
Materials Park, OH 44073-0002
Printed in the United States of America
pg ii
Name ///sr-nova/Dclabs_wip/Failure_analysis/5113_iii.pdf/Front Matter_Dedication Page/
19/8/2008 2:16PM Plate # 0

This book is dedicated to our families, without whose continued

support the completion of this work would not have been possible:
My husband, Antonio Carlos Canale,
and my children, Amanda, Sara, and Bruno
L.C.F.C.
To my lovely wife, Carla Mesquita, and my dear son, Rafael
R.A.M.
My wife, Alice
G.E.T.
pg iii
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/

20/8/2008 12:41PM Plate # 0
pg v
Contents
Preface ............................................................................................................................. ix
Component Design ............................................................................................................ 1

Mario Solari, Consultores de Tecnologa e Ingenera SRL
Pablo Bilmes, Universidad Nacional de La Plata
Introduction to Heat Treat Processing .................................................................................. 1
Important Design Aspects ..................................................................................................... 2
Techniques for Controlling Distortion ................................................................................ 16
Examples of Failures due to Heat Treatment ...................................................................... 18
Heat Treatment Design ........................................................................................................ 29
Modeling of Heat Treatment ............................................................................................... 31
Failure Aspects of Welded Components ............................................................................. 33
Heat Treatment Procedures Applied to Welded Components ............................................ 36
The Risk-Based Approach and Heat Treatments ................................................................ 40
Overview of the Mechanisms of Failure in Heat Treated Steel Components .................... 43

Scott MacKenzie, Houghton International, Inc.
General Sources of Failure .................................................................................................. 43
General Practice Conducting a Failure Analysis ................................................................ 47
Determination of the Fracture Mechanism ......................................................................... 51
Summary .............................................................................................................................. 83
Mechanisms and Causes of Failures in Heat Treated Steel Parts ....................................... 87

Debbie Aliya, Aliya Analytical, Inc.
Types of Damage and Failure ............................................................................................. 88
Factors Contributing to Poor Response from Heat Treatment ......................................... 101
Concluding Comments ...................................................................................................... 108
General Aspects of Failure Analysis ............................................................................... 111

Waldek Wladimir Bose-Filho, Universidade de Sao Paulo
Jose Ricardo Tarpani, Universidade de Sao Paulo
Marcelo Tadeu Milan, Instituto de Materiais Tecnologicos do Brasil Ltda.
General Guidelines of Failure Analysis ............................................................................ 111
Fracture .............................................................................................................................. 118
Distortion ........................................................................................................................... 127
Wear-Assisted Failure ....................................................................................................... 129
Environmentally Assisted Failure ..................................................................................... 131
v

20/8/2008 12:41PM Plate # 0
Failure in Steel Forging .................................................................................................. 133

Md. Maniruzzaman, Worcester Polytechnic Institute
Charlie Gure, Forging Consultant
Stephen R. Crosby, The Stanely Works
Richard D. Sisson, Jr., Worcester Polytechnic Institute
Forging Process Design ..................................................................................................... 134
Case Studies ....................................................................................................................... 138
Failures from the Casting Process .................................................................................. 151

Omar Maluf, Instituto de Materiais Tecnologicos do Brasil Ltda.
Luciana Sgarbi Rossino, Instituto de Materiais Tecnologicos do Brasil Ltda.
Camilo Bento Carletti, Centro de Caracterizacao e Desenvolvimento de Materiais
Celso Roberto Ribeiro, Centro de Caracterizacao e Desenvolvimento de Materiais
Clever Ricardo Chinaglia, Centro de Caracterizacao e Desenvolvimento de Materiais
Jose Eduardo Mya, Centro de Caracterizacao e Desenvolvimento de Materiais
Failures due to Improper Cast Design ............................................................................... 151
Effects due to Porosity ...................................................................................................... 154
Effects due to Decarburization during Microfusion ......................................................... 162
Effects due to Cold Joints ................................................................................................. 163
Inclusions ........................................................................................................................... 165
Sources of Failures in Carburized and Carbonitrided Components ................................ 177

Malgorzata Przylecka, Poznan University of Technology
Wojciech Gestwa, Poznan University of Technology
L.C.F. Canale, Universidade de Sao Paulo
Xin Yao, Portland State University
G.E. Totten, Associacao Instituto Internacional de Ciencia and Portland State University
Design ................................................................................................................................ 179
Steel Selection and Hardenability ..................................................................................... 181
Residual Stress ................................................................................................................... 196
Dimensional Stability ........................................................................................................ 200
Quenching and Grinding Cracks ....................................................................................... 204
Insufficient Case Hardness and Improper Core Hardness ................................................ 209
Influence of Surface Carbon Content ................................................................................ 211
Influence of Grain Size ...................................................................................................... 217
Internal Oxidation .............................................................................................................. 219
Carbides and Carbide Structure ........................................................................................ 222
Noncarbide Inclusions ....................................................................................................... 228
Micropiting ........................................................................................................................ 230
Contact Fatigue Piting (Macropiting) ............................................................................... 230
Case Crushing .................................................................................................................... 231
Pitting Corrosion ............................................................................................................... 232
Partial Melting ................................................................................................................... 233
Fatigue Fracture of Nitrided Layers ............................................................................... 241

Aleksander Nakonieczny, Institute of Precision Mechanics
Fatigue Resistance ............................................................................................................. 241
Fatigue Evaluation of Nitrided Steels ............................................................................... 244
Fatigue Property Characteristics after Carbonitriding ...................................................... 246
Summary ............................................................................................................................ 250
vi
pg vi

20/8/2008 12:41PM Plate # 0
pg vii
Steel Heat Treatment Failures due to Quenching ........................................................... 255

G.E. Totten, Associacao Instituto Internacional de Ciencia and Portland State University
Phase Transformation During Heating and Quenching .................................................... 255
Effect of Materials and Quench Process Design on Distortion ........................................ 263
Stress Raisers and Their Role in Quench Cracking .......................................................... 272
Case Studies in Quench Cracking ..................................................................................... 273
Steel Failures due to Tempering and Isothermal Heat Treatment ................................... 285
Jan Vatavuk, Universidade Mackenzie
Martensite .......................................................................................................................... 285
Tempering .......................................................................................................................... 289
Embrittlement .................................................................................................................... 293
Case Studies ....................................................................................................................... 303
Failure Analysis in Tool Steels ....................................................................................... 311
Rafael A. Mesquita, Villares Metals
Celso Antonio Barbosa, Villares Metals
Classification of Tool Steels .............................................................................................. 311
Heat Treating Failures of Cold Work Tools ..................................................................... 314
Heat Treating Failures of Hot Work Tools ....................................................................... 330
Conclusion ......................................................................................................................... 349
Case Studies of Steel Component Failures in Aerospace Applications ............................ 351
Failure Analysis of a Catapult Holdback Bar ................................................................... 351
Cracking in a Main Landing Gear Attach Pin .................................................................. 354
MLG Linear Actuating Rod and Cylinder ........................................................................ 355
Failure Analysis of AISI 420 Stainless Steel Roll Pin ...................................................... 359
Failure Analysis of a Main Landing Gear Lever .............................................................. 362
Failure Analysis of an Inboard Flap Hinge Bolt ............................................................... 364
Failure Analysis of a Nose Landing Gear Piston Axle ..................................................... 367
Multiple-Leg Aircraft-Handling Sling .............................................................................. 372
Failure Analysis of an Aircraft Hoist Sling during Static Test ......................................... 373
Failure Analysis of an Internal Spur Gear ........................................................................ 375
Main Landing Gear Axle ................................................................................................... 378
Nondestructive Testing and Failure Analysis of
Fin Attach Bolts after Full-Scale Fatigue Testing ........................................................ 380
Failure Analysis of Powder Metal Steel Components ..................................................... 395
S. Ashok, Sundram Fasteners Ltd.
Sundar Sriram, Sundram Fasteners Ltd.
Powder Metallurgy Process ............................................................................................... 395
Case Hardening ................................................................................................................. 397
Failure Analysis Techniques ............................................................................................. 399
Case Studies of PM Steel Failures .................................................................................... 401
Induction Hardening ..................................................................................................... 417
Janez Grum, University of Ljubljana
Steels for Surface Hardening ............................................................................................. 419
vii

20/8/2008 12:41PM Plate # 0
Main Features of Induction Heating .................................................................................

Induction Hardening of Machine Parts .............................................................................
Magnetic Flux Concentrators ............................................................................................
Conditions in Induction Heating and Quenching of Machine Parts .................................
Time-Temperature Dependence in Induction Heating .....................................................
Quenching Systems for Induction Hardening ...................................................................
Time Variation of Stresses and Residual Stresses ............................................................
Workpiece Distortion in Induction Surface Hardening ....................................................
Residual Stresses after Induction Surface Hardening and Finish Grinding .....................
Hardness Profiles in the Induction Surface-Hardened Layer ............................................
Fatigue Strength of Materials ............................................................................................
Stress Profiles in Machine Parts in the Loaded State ........................................................
Input and Output Control of Steel for Induction Surface Hardening of Gears ................
420
422
437
440
444
449
452
466
472
477
481
485
491
Failure Analysis of Steel Welds ...................................................................................... 503

J.H. Devletian, Portland State University
D. Van Dyke, MEI-Charlton, Inc.
Discontinuities in Steel Welds .......................................................................................... 503
Fatigue of Welded Joints ................................................................................................... 505
Hydrogen-Assisted Cracking Theory ................................................................................ 506
Types of Hydrogen-Assisted Cracking ............................................................................. 509
Stress-Corrosion Cracking of Steel ................................................................................... 513
Solidification Cracking of Steel ........................................................................................ 515
Appendix 1: Metric Conversion Guide ..........................................................................
Appendix 2: Temperature Conversion Table ..................................................................
Appendix 3: Steel Hardness Conversions .......................................................................
Appendix 4: Austenitizing Temperatures for Steels ........................................................
Appendix 5: Temper Colors for Steels ............................................................................
Appendix 6: Physical Properties of Carbon and Low-Alloy Steels ...................................
Appendix 7: AISI to Non-AISI Steel Cross Reference .....................................................
Appendix 8: Non-AISI to AISI Steel Cross Reference .....................................................
Appendix 9: Iron-Carbon Equilibrium Diagram .............................................................
Appendix 10: Isothermal Diagrams of Selected Steels ...................................................
Appendix 11: Continuous Cooling Diagrams of Selected Steels .....................................
521
525
529
537
539
541
551
563
585
587
601
Index ............................................................................................................................. 629
viii
pg viii
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_ix.pdf/Front Matter_Preface/

19/8/2008 3:39PM Plate # 0
pg ix
Preface
Material failures can lead to many potentially disasterous consequences, including poor product
quality, necessary repair or component or equipment replacement, production downtime losses,
environmental impact, and even loss of life. Furthermore, failures may arise from not one but various
causes, including design, material composition, and, in the case of metals such as steel, improper
thermal processing. Therefore, when failures do occur, it is critically necessary to not only identify
these failures but also to determine and correct their root cause. This is a primary objective of this
work.
There are many books, journals, and other references that focus on various aspects of failure
analysis. However, there are relatively few that focus on steel failures arising during thermal processing, such as forging, casting, heat treatment, welding, and others. A second objective of this book
is to provide a reasonably thorough reference detailing potential failures that may occur during
thermal processing and the identification of their root cause, even if it is not specifically the thermal
process being considered.
An important feature of Failure Analysis of Heat Treated Steel Components is that it not only
discusses various causes of a failure and its identification but also integrates this discussion with the
metallurgy of the process, thus providing one comprehensive resource. This book was developed as a
reference source for use by designers, practicing metallurgists, mechanical and materials engineers,
quality-control technicians, and heat treaters. This book also will serve as an important textbook for
various advanced undergraduate and graduate courses on either failure analysis or thermal processing
of steel.
The editors are indebted to the invaluable guidance of many persons in the development and
production of this text, including Prof. George Krauss (Colorado School of Mines), George Vander
Voort (Buehler Ltd., USA), N. Gopinath and V. Raghunathan (Fluidtherm Technology P. Ltd.), Ross
Blackwood (deceased), Larry Jarvis (Tenaxol Inc.), and many others. In addition, the editors are most
appreciative of Steve Lampman for his continued patience, guidance, and assistance during the
various stages of the preparation of this text. The editors are especially grateful for the support of the
chapter authors for the diligence, dedication, and patience involved in their vital contributions to this
work. Most of all, the editors are especially appreciative of the support and sacrifices made by their
spouses, Antonio Canale, Carla Mesquita, and Alice Totten, without which the preparation of this
book would not have been possible. We also express our gratitude to Villares Metals S.A. for their
continued and vital assistance and generosity throughout this project.
Lauralice C.F. Canale, Ph.D.
Sao Carlos, SP, Brazil
Rafael Agnelli Mesquita
Sumare, SP, Brazil
George E. Totten, Ph.D., FASM
Seattle, WA, USA
ix
ASM International is the society for materials
engineers and scientists, a worldwide network

dedicated to advancing industry, technology, and
applications of metals and materials.
ASM International, Materials Park, Ohio, USA
This publication is copyright ASM International. All rights reserved.
Publication title
Product code
Failure Analysis of Heat Treated Steel Components
05113G
To order products from ASM International:

Online Visit www.asminternational.org/bookstore
Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US)
Fax 1-440-338-4634
Mail
Customer Service, ASM International

9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
Email CustomerService@asminternational.org
American Technical Publishers Ltd.

27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX,
In Europe United Kingdom
Telephone: 01462 437933 (account holders), 01462 431525 (credit card)
www.ameritech.co.uk
Neutrino Inc.
In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan
Telephone: 81 (0) 424 84 5550
Terms of Use. This publication is being made available in PDF format as a benefit to members and
customers of ASM International. You may download and print a copy of this publication for your
personal use only. Other use and distribution is prohibited without the express written permission of
ASM International.
No warranties, express or implied, including, without limitation, warranties of merchantability or
fitness for a particular purpose, are given in connection with this publication. Although this
information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be
obtained from the use of this publication alone. This publication is intended for use by persons having
technical skill, at their sole discretion and risk. Since the conditions of product or material use are
outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this
information. As with any material, evaluation of the material under end-use conditions prior to
specification is essential. Therefore, specific testing under actual conditions is recommended.
Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale,
use, or reproduction, in connection with any method, process, apparatus, product, composition, or
system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this
publication shall be construed as a defense against any alleged infringement of letters patent,
copyright, or trademark, or as a defense against liability for such infringement.
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_1-42.pdf/Chap_01/

L.C.F. Canale, R.A. Mesquita, and G.E. Totten, editors, p 1-42
DOI: 10.1361/faht2008p001
18/8/2008 2:47PM Plate # 0
Copyright 2008 ASM International

All rights reserved.
Component Design
Mario Solari, CTI Consultores de Tecnologa e Ingeniera SRL
Pablo Bilmes, Universidad Nacional de La Plata
DESIGN involves different creative aspects:

planning, development, procedures, availability,
and fitness concerning the materials and processes used to manufacture the component.
Design is an iterative process, often based on
experience, to provide an assessment of the
performance of a component for a certain period
of time of expected or intended service life. The
design process culminates in a technical specification for the part or system and suitable
manufacturing processes. Another obvious aim
of design is to prevent failures throughout the
component lifetime cycle and avoid situations
resulting in severe failure.
Heat treating achieves the desired changes in
structure and properties, and various types of
heat treatments may be employed to meet design
requirements for mechanical strength, corrosion, wear, and so on. Heat treatments include
stress relieving, austenitizing, normalizing,
annealing, quenching, and tempering (Ref 1).
Heat treating may also involve chemical or
additional physical processes. A systematic
procedure for minimizing risks involved in heat
treated steel components requires a combination
of metallurgical failure analysis and fitness for
service with respect to safety and reliability
based on risk analysis. The effects of steel heat
treatment may include (Ref 1):

Control of microstructure formation

Increase of strength, toughness, or perhaps
creep resistance
Relief of residual stresses and prevention of
cracking
Control of hardness (and softness)
Improvement of machinability
Improvement of corrosion resistance or wear
resistance
Introduction to Heat Treat Processing

Material behavior related to heat treatment
can be analyzed by developing models that
involve a complex interrelationship of variables

associated with the material, manufacturing
processes, and service conditions (Ref 2).
The ability of ferrous materials to develop
required properties through heat treatment is
a broad concept that refers both to the ease
with which a material may be heat treated
and the resulting in-service fitness of the component.
The iron allotropic transformation between
more densely packed face-centered cubic iron,
nonmagnetic gamma (c) phase designated as
austenite, and the less densely packed bodycentered cubic iron, alpha (a) phase designated
as ferrite, is the basis for heat treatment of steels.
Austenite can dissolve up to approximately
2.0 wt% C and in most steels is not stable at
low temperature. On the other hand, the interstitial sites in ferrite are much smaller than in
austenite; therefore, ferrite can only dissolve
very small concentrations of carbon (0.025 wt%
maximum) and is relatively soft and stable at
room temperature.
The iron-carbon phase diagram shows the
compositional limits of the different transformational phases formed by a steel alloy that
exist during heating or cooling as a function
of temperature. In hypoeutectoid steels (those
with 50.80 wt% C), upon cooling two different
phases can exist, ferrite and austenite, each containing different amounts of carbon. Upon further cooling, the microstructure of these steels
exhibits ferrite grains in a pearlite island. Pearlite
is a metastable microstructure formed during
austenite decomposition. The pearlite structure
is an aggregate consisting of alternating lamellae
of ferrite and cementite that is formed on slow
cooling during the eutectoid reaction. Cementite
is a very hard and brittle compound of iron
and carbon (Fe3C). Depending on the thermal
history, cementite will appear as lamellae (with
ferrite), spheroids, or globules in a ferritic
matrix.
pg 1
18/8/2008 2:49PM Plate # 0
2 / Failure Analysis of Heat Treated Steel Components
Microstructures that are formed upon cooling and the proportions of each are dependent on
austenitization time, temperature, cooling history of the particular alloy, and the composition
of the alloy. The transformation products from
austenite decomposition change from a mixture
of ferrite and pearlite to bainite or martensite
with increasing cooling rate.
Bainite is a two-phase mixture of ferrite
and cementite consisting of fine lines of iron
carbide in acicular ferrite. Upper bainite has a
feathery appearance and forms just below the
temperature where fine pearlite is formed.
Lower bainite exhibits an acicular microstructure that is formed just above martensite,
which is produced at approximately 350 C
(660 F).
Martensite is a supersaturated solid solution of carbon in alpha iron (ferrite) that is
less densely packed than the c body-centered
tetragonal lattice and is a magnetic platelike
structure formed by a diffusionless shear type
of transformation of austenite below the martensite start (Ms) temperature. The amount of
transformation depends on the martensitic
temperature range (Ms to Mf). (Mf is the martensite finish temperature.) The three forms of
martensite are lath, plate, and tempered martensite.
Transformation from austenite to martensite
results in a volumetric expansion at the Ms
temperature. Dimensional changes are possible,
depending on the carbon content and microstructural transformation product formed. The
volume change (%) is [4.640.53 (%C)] for the
reaction from austenite to martensite.
The two most commonly used transformation diagrams are time-temperature transformation for isothermal transformation, and
continuous cooling transformation diagrams.
These diagrams can be used to predict steel
microstructures and hardness after heat treatment, or they may be used to design a heat
treatment process. Heat treating processes
include hardening, austenitization, annealing
(full annealing, intercritical annealing and subcritical annealing, recrystallization annealing,
isothermal annealing, soft annealing, diffusion
annealing), normalizing, stress relieving,
quenching and tempering, and austempering,
and are summarized in Table 1 (Ref 1).
Hardening and tempering are common
heat treatment processes. If steel is cooled sufficiently fast, without microstructural transformation, thermal stresses can develop. Under
these conditions, the surface of the part is initially cooled much more quickly than the core.
Therefore, the specific volume in the core is
greater than at the surface, and the reduction in
volume at the surface is resisted by the greater
volume in the core, resulting in the surface being
in tension and the core in compression. After the
cooling processes have been completed, the
residual-stress distribution between the surface
and core is obtained. If the surface stresses
exceed the hot yield strength of the material, it
plastically deforms, resulting in thermally
induced dimensional changes (Ref 3). When
steels that undergo transformational changes are
quenched, the possibility of the formation of
both thermal and transformational stresses must
be considered.
Steel parts are often tempered by reheating
after quench hardening to obtain specific
mechanical properties. The tempering process
involves heating hardened steel to some temperature below the eutectoid temperature for the
purpose of decreasing hardness and increasing ductility and toughness while relieving
quench stresses and ensuring dimensional stability. Tempering processes include tempering of martensite, transformation of retained
austenite to martensite, tempering of the decomposition products of martensite, and decomposition of retained austenite to martensite.
In addition, tempering may also lead to dimensional variation due to relaxation of residual
stress and plastic deformation, which is due
to the temperature dependence of yield
strength. Tempering may lead to an increase in
hardness if secondary hardening occurs, which
is due to precipitation of a compound or to
the formation of martensite or bainite from
retained austenite, decomposition during tempering, or destabilization during this process
and then transformation during subsequent
cooling.
Quenchant selection and quenching conditions are critically important parameters in
quench system design. For example, the dimensional changes after austenitizing and then
quenching in water are greater than quenching in
oil (Ref 3).
Important Design Aspects

The importance of good design cannot be
overemphasized. Poor design can cause or
pg 2
18/8/2008 2:49PM Plate # 0
Component Design / 3
Table 1 Heat treating process for carbon steels

Process
Characteristics
Austenitization:
Complete transformation to austenite by
heating the steel above the critical
temperature for austenitic formation
The optimal austenitization temperature is 3050 C (5590 F) above Ac3 for

hypoeutectoid steels and 3050 C (5590 F) above Ac1 for hypereutectoid steels.
Ac3 is the temperature at which the transformation of ferrite to austenite is completed
during heating. Ac1 is the temperature at which austenite begins to form during heating.
The heating rate must be limited and uniform to avoid cracking or warpage and to control
thermal stresses in the range of 250600 C (4801110 F).
The carbon equivalent controls the propensity for steel to crack.
The holding time is dependent on geometrical factors related to the furnace (emissivities,
temperature, and atmosphere composition) and load (type of steel and thermophysical
properties).
Annealing:
Heat treatment consisting of heating and
soaking at suitable temperature followed by
cooling under conditions such that, after
return to ambient temperature, the metal will
be in a structural state closer to that of equilibrium
Full annealing: Heat 3050 C (5590 F) above Ac3 for hypoeutectoid steels, then
furnace cool through the critical temperature range at a specified cooling rate. The aim
is to break the continuous carbide network of high-carbon steels. It improves
machinability.
Partial (intercritical) annealing: Heating within the critical temperature range (Ac1Ac3),
followed by slow furnace cooling. It improves machinability.
Subcritical annealing: Heating 1020 C (2035 F) below Ac1 followed by cooling in
still air. It can be used to temper bainitic or martensitic structures to produce softened
microstructures containing spheroidal carbides in ferrite. Improves the cold working
properties of low carbon steels (525% C) or softens high-carbon and alloy steel
Recrystallization annealing: Heat the steel for 30 min1 h at temperature above the
recrystallization temperature (TR=0.4 Tm), then the steel is cooled. The treatment
temperature depends on prior deformation, grain size, and holding time. The
recrystallization process produces strain-free grain nucleation, resulting in a ductile,
spheroidized microstructure.
Isothermal annealing: Heating the hypoeutectoid steel within the austenitic transformation
range above Ac3 for a time sufficient to complete the solution process, yielding a
completely austenitic microstructure. At this time, the steel is cooled rapidly at a
specific rate within the pearlite transformation range until the complete transformation
to ferrite plus pearlite occurs, and then it is cooled rapidly.
Spheroidizing (soft annealing): Involves the prolonged heating of steel at a temperature
near the lower critical temperature (Ac1), then furnace cooling
Diffusion (Homogenizing annealed): Heat the steel rapidly to 11001200 C
(20102190 F) for 816 h, furnace cool to 800850 C (14701560 F), and then
cool to room temperature in still air. It is performed on steel ingots and castings to
minimize chemical segregation.
The primary purpose of annealing is to soften

the steel to enhance its workability and
machinability. Also, it relieves internal
stresses, restores ductility and toughness,
refines grains, reduces gaseous content in the
steel, and improves homogenization of
alloying elements.
Normalizing:
The aim is to provide a uniform
microstructure of ferrite plus pearlite (small
grains and finer lamellae than in annealing).
Heat the steel to 4050 C (8090 F) above Ac3 for hypoeutectoid steels and 4050 C
(8090 F) above Acm for hypereutectoid steels. The holding time depends on the size,
and then the steel is cooled in still air. It produces grain refinement and improved
homogenization.
Stress relieving:
It is typically used to remove residual
stresses that have accumulated from prior
manufacturing processes. Stress relieving
results in a significant reduction of yield
strength in addition to reducing the residual
stresses to some safe value.
Heat to a temperature below Ac1for the required time to achieve the desired reduction in
residual stresses, and then the steel is cooled at a rate sufficiently slow to avoid the
formation of excessive thermal stresses. Below 300 C (570 F), faster cooling rates
can be used. No microstructural changes occur during stress-relief processing. The
recommended heating temperature range is 550700 C (10201290 F), depending on
the type of steel. These temperatures are above the recrystallization temperature. Little
or no stress relief occurs at temperatures 5260 C (500 F), and approximately 90% of
the stress is relieved at 540 C (1005 F). The maximum temperature for stress relief
is limited to 30 C (55 F) below the tempering temperature used after quenching.
The results of the stress-relief process are dependent on the temperature and time.
Hardenability:
Ability to develop hardness to a given
depth after having been austenitized and
quenched
The hardenability depends on the concentration of dissolved carbon in the austenitic

phase, alloying elements, austenitizing temperature, austenitic grain size at the moment
of quenching, size and shape of the cross section, and quenching conditions.
Quenching:
Quench severity is the ability of a
quenching medium to extract heat from a
hot steel workpiece.
Specific recommendations for quench media selection for use with various steel alloys are
provided by standards such as SAE AMS 2759. Quench media include water, brine,
aqueous polymer, gas or air quenching, and caustic quenching.
Tempering:
Tempering is the thermal treatment of
hardened and normalized steels to obtain
the desired mechanical properties, which
include improved toughness and ductility,
lower hardness, and improved dimensional
stability.
The tempering process involves heating steel to any temperature below the Ac1
temperature. During tempering, as-quenched martensite is transformed into tempered
martensite, which is composed of highly dispersed spheroids of cementite (carbides)
dispersed in a soft matrix of ferrite, resulting in reduced hardness and increased
toughness. The objective is to allow the hardness to decrease to the desired level and
then stop the carbide decomposition by cooling. The extent of the tempering effect
is determined by the temperature and time of the process.
Source: Ref 1
pg 3
18/8/2008 2:49PM Plate # 0
pg 4
promote heat treatment failures before the

component is put into service, or it may reduce
service life, sometimes dramatically. The
objective of proper design for heat treatment is
to provide the minimum engineering requirements, the desired material properties at the
lowest total cost, and, in particular, to minimize
the expense of scrap due to rework of parts that
may have undergone excessive distortion or
cracked.
The Heat-Transfer Theory Applied to Heat
Treatments. The laws that govern heat transmission are very important to the engineer in
heat treatment design. There are three different
types of heat transfer: conduction, convection,
and radiation (Ref 4). They have in common that
temperature difference (thermal gradient) must
exist and that the heat is always transferred in the
direction of decreasing temperature. When the
temperature profile does not change with time,
the fundamental relation for the unidirectional
steady flow of heat through a solid by conduction, Fouriers first law, can be expressed by:
Q=7l
T
x
(Eq 1)
where Q is the quantity of heat flowing through

the unit area of a wall per unit time in the
direction of the x-axis and is directly proportional to the thermal conductivity, l, and the
thermal gradient in x-direction. Thermal conductivity has a nearly linear dependence on
temperature.
During heat treatments, the temperature varies in time as well as in space; these processes
are called unsteady, nonstationary, or transient.
As the body heats, the temperature at each point
asymptotically approaches the temperature of
the medium. The temperature of points near the
surface of the body changes most rapidly. The
differential equation for one-dimensional transient heat conduction, Fouriers second law, in
the absence of inner heat sources is:
T
2 T
=a 2
t
x
(Eq 2)
where T is the temperature, t is the time, and a is

the thermal diffusivity of the metal and is:
a=
l
rCp
(Eq 3)
where l is the thermal conductivity, r is the

density, and Cp is the specific heat at constant
pressure of the material. The thermophysical
properties l, r, and Cp vary with temperature.

The differential equation of heat conduction
establishes the relation between the time and
space variation of temperature at any point
of the body in which conduction takes place.
The factor of proportionality thermal diffusivity,
a, represents a physical property of the material,
is essential for transient processes of heat flow,
and defines the rate of change of temperature.
If the thermal conductivity, l, is the ability of
a solid to conduct heat, thermal diffusivity is
the measure of a material thermal inertia. The
quantity rCp is the volumetric specific heat;
this product is approximately constant for solid
metals. So, in the case of austenitic stainless
steels, low thermal conductivities correspond
to low thermal diffusivities. In other words,
equalization of temperature at all points of
space will proceed at a lower rate in austenitic
stainless steels, with respect to ferritic steel, due
to its lower thermal diffusivity, and there are
difficulties in homogenizing temperature during
heat treatments.
Heat transfer by convection occurs between
the surface of the body and surrounding fluids;
for this type of heat transmission, the following
Newtons equation is in general use:
Qh =hDT
(Eq 4)
It simply states that an invariable temperature

difference, DT, between a surface and a fluid in
contact with it causes a steady heat flow of Qh.
The factor of proportionality, h, is called the
coefficient of heat transfer.
The third type of heat-transfer mode is
radiation. The heat flow by radiation is commonly written:

Q=es T14 7T04
(Eq 5)
where e is the emissivity (1 for a black body),

T is the temperature, and s is the Stefan
constant (Ref 5). The three different types of
heat transfer, conduction, convection, and
radiation, are present during the heat treatment
processes.
Perhaps the most important physical property
of steel to be considered in design is its coefficient of thermal expansion. Most heat treating
problems could be solved if this coefficient
could be controlled. Because it cannot, it is
necessary to learn to design with it (Ref 6).
Almost all solids expand on heating. As the
temperature is raised, the thermal vibration
pushes the atoms apart, increasing their mean
18/8/2008 2:49PM Plate # 0
pg 5
spacing. The effect is measured by the linear

expansion coefficient:
a=
1 dl
L dT
(Eq 6)
where L is a linear dimension of the body

(Ref 5).
The relationship between thermal conductivity and thermal expansion is important in
designing against thermal distortion. Thermal
gradients can cause a change of shape, which is a
distortion of the component. The strain is related
to temperature by:
e=aT0 7T
(Eq 7)
where T0 is ambient temperature, and a is thermal expansion (Ref 5).

The distortion is proportional to the gradient
of the strain, so it is proportional to the thermal
gradient. By Fouriers first law, the heat flow is
proportional to the thermal gradient through
the thermal conductivity, l. For a given geometry and heat flow, the distortion is minimized by selecting materials with large values
of l/a (Ref 5). For example, austenitic stainless steels have low thermal conductivity and
high thermal expansion, related to ferritic steel,
so distortion during welding becomes a problem.
Thermal expansion has a strong influence on
the development of residual stress. Whenever
the thermal expansion or contraction of a body is
prevented, thermal stresses appear; if large
enough, they cause yielding, fracture, or elastic
collapse (buckling). For axial constraint, the
stress, Ds, produced by a temperature change of
1 C or the stress per C caused by a sudden
change of surface temperature in one that is not
constrained is equal to aE, where a is the
expansion coefficient, and E is the elastic modulus of the material. For biaxial and triaxial
constraint, the stresses shall be multiplied by
(1n) and (12n) respectively, where n is
Poissons ratio. These stresses are large and can
cause a material to yield, crack, spall, or buckle
(Ref 5).
Linear thermal expansion (Table 2) in going
from room temperature to 700 C (1300 F) is
approximately:

10.3 mm/m (0.124 in./ft) for low-alloy steel

8.5 mm/m (0.102 in./ft) for martensitic
stainless steel (type 13Cr)
13.1 mm/m (0.157 in./ft) for austenitic
stainless steel (type 18Cr-8Ni).
Thermal expansion of austenite is larger than

that of ferrite. These thermal expansion values
are representative of solution-annealed material.
Subsequent precipitation hardening treatments
may affect thermal expansion (Ref 7).
In high-temperature components design,
incompatibility of thermal expansion becomes a
major problem. Choice of materials and designs
should take this into account. Bolts used to hold
high-temperature casings together must be
selected to have sufficient elevated-temperature
strength and make a good thermal expansion
match with the casing material. When rotor and
casing are made of ferritic steel, modified 12%
Cr bolts work well up to 565 C (1050 F), but
nickel-base superalloys are needed at 595 C
(1100 F) or higher.
The ability of a material to resist thermal
shock, due to a sudden immersion in a cold
medium, without cracking depends on its thermal expansion coefficient, a; tensile strength,
st, for metals; Youngs modulus, E; thermal
conductivity, l; and heat-transfer coefficient, h.
A temperature change of DT applied to a
constrained body or a sudden change DT of the
surface temperature of the unconstrained component induces a stress:
s=
EaDT
C
(Eq 8)
where C is equal to 1 for axial constraint, (1n)

for biaxial constraint, (12n) for triaxial constraint, and n is Poissons ratio. If this thermally
Table 2 Linear thermal expansion for ferrous
materials
Linear thermal expansion
in temperature range
from room temperature
to 704 C (1300 F)
Steel
mm/m
in./ft
Carbon and low-alloy steels:

C; C-Mn; C-Si; C-Mn-Si; C-1/2 Mo to
11/4 Cr-1/2 Mo; Mn-1/2 Mo-1/2 Ni
5Cr-1Mo and 29Cr-7Ni-2Mo-N steels
9Cr-1Mo steel
12Cr; 12Cr-1Al; 13Cr; and 13Cr-4Ni
steels
15Cr and 17Cr steels
27Cr steels
10.3
0.124
9.5
8.8
8.5
0.114
0.106
0.102
8
7.3
0.096
0.088
13.1
0.157
12.25
0.147
Austenitic stainless steels:

16Cr-12Ni-2Mo; 16Cr-12Ni-2Mo-N;
16Cr-12Ni-2Mo-Ti; 18Cr-8Ni
Austenitic stainless steels:
29Ni-20Cr-3Cu-2Mo; 20Cr-18Ni-6Mo;
22Cr-13Ni-5Mn
Source: Ref 7
18/8/2008 2:49PM Plate # 0
induced stress exceeds the local tensile strength

of the material, yielding (permanent plastic
deformation) or cracking results (Ref 5). This
plastic flow causes permanent shape change
(distortion) and impacts the magnitude and distribution of residual stresses. Water quenching
gives a high h, and then the values of DT calculated from the previous equation give an
approximate ranking of thermal shock resistance
(Ref 5).
However, when heat transfer at the surface is
poor and the thermal conductivity of the solid is
high, the thermal stress is less than that given by
the previous equations. A measure of the thermal
shock resistance that takes into account the finite
rate of heat transfer at the surface, a heat-transfer
coefficient that is never infinite, is given by:
BDT=
st
aE
(Eq 9)
where st, a, and E were defined previously,

B=C/A, where C also was defined previously;
and A is:
A=
sh=l
1+sh=l
(Eq 10)
where s is a typical dimension of the sample in

the direction of heat flow, h is the heat-transfer
coefficient, and l is the thermal conductivity
(Ref 5). The quantity Bi=sh/l is usually called
the Biot modulus. If Bi41, heat flow is limited
by conduction. For fast water quench of metals,
the heat-transfer coefficient, h, is high (h=
104 W/m2K), and the thermal conductivity is
also high, so the factor A approaches 1. On the
900
other hand, for fast air flow (h=102 W/m2K),

the factor A results are equal to 3 10 2 (for
section s=10 mm), and the thermal shock
resistance DT is larger by the factor 1/A (Ref 5).
As an example of the use of the aforementioned equations, Fig. 1 shows schematically the
effect of the thermal expansion coefficient (a)
and the heat-transfer coefficient (h) in thermal
shock resistance (DT) for a hypothetical steel
with 800 MPa (~120 ksi) tensile strength, biaxial constraint, and thermophysical properties
constant with temperature. Three cases were
analyzed:
a) In the first case, a ferritic steel with low
thermal expansion and a very high-heat
transfer coefficient (fast water quench,
h=104 W/m2K) was considered. The maximum temperature change (thermal shock
resistance) that induces stresses below the
tensile strength, avoiding yielding or cracking, is 240 C (470 F).
b) The second example similar to case (a) but
with high thermal expansion. The thermal
shock resistance results in temperatures
above 135 C (270 F) increasing the failure
risk.
c) The third example is similar to case (a) but
with a lower heat-transfer coefficient (air
flow, h=50 W/m2K). The thermal shock
resistance results in 280 C (535 F)
decreasing the risk of failure.
These factors (residual stresses and dimensional changes) have the greatest influence on
Tensile strength
Stress, MPa (ksi)
800
(120)
700
Induced thermal stress for very

high heat-transfer coefficient, h,
and low thermal expansion
600
(90)
500
Induced thermal stress for very

high heat-transfer coefficient, h,
and high thermal expansion
400
(60)
300
200
(30)
Induced stress for low heattransfer coefficient and low

thermal expansion
Thermal shock
resistance
100
0
0
50
(90)
100
(180)
150
(270)
200
(360)
250
(450)
300
(540)
Temperature change, C (F)
Fig. 1
Schematic representation of thermal stresses resulting from a sudden change, DT, of the surface temperature and thermal shock
resistance
pg 6
18/8/2008 2:49PM Plate # 0
the design process of a component. In addition to

thermal strains, many materials systems undergo
phase transformations as a function of temperature. Often, the new phase(s) that forms has
a different volume and different coefficient of
expansion as well as different mechanical
behavior(s) than the parent phase(s). For example, phase transformation from austenite to
martensite results in a volumetric expansion at
the martensite start (Ms) temperature. These
differences increase the complexity of understanding the effect of thermal gradients on the
strains produced and the resulting plastic
deformation (Ref 8). Thermal and transformation-induced strains can result in substantial
plastic deformation and residual stresses. The
total induced strain is the result of the sum
of the strain produced by thermal expansion
(aE DT=eth) of a piece with initial length (E)
and the transformation strain (etr). The total
induced strain must be accommodated through
either elastic (eel) or plastic (eep) strain, which
sums to the total strain et=eth +etr=
(eel +eep). In order to determine the accommodation strain values, Youngs modulus (E)
and the yield strength (sys) are required as a
function of phase and temperature. Most of the
plastic deformation occurs during the heat-up
and cool-down stages of the process (Ref 8).
Primary Stresses, Secondary Stresses, Peak
Stresses, and Residual Stresses. Primary
stress is a normal or shear stress developed by
the imposed loading that is necessary to satisfy
the laws of equilibrium of external and internal
forces and moments. The basic characteristic
of a primary stress is that it is not self-limiting.
Primary stress that considerably exceeds the
yield strength will result in failure or at least
in great distortion. Secondary stress is a normal
or shear stress developed by the constraint of
adjacent parts or by self-constraint of a structure.
The basic characteristic of the secondary stress
is that it is self-limiting. An example of secondary stress is a general thermal stress.
The elastic stresses calculated previously are
nominal values, that do not take into account
local discontinuities such as holes, notches, or
section changes. Even on a structure where
stress intensity has been limited by yield criteria,
there may exist highly localized regions where
peak stresses are several times higher than yield.
Maximum local stresses on a structure can be
determined by considering nominal stresses
multiplied by a stress-concentration factor and
can be estimated through a detailed stress ana-
lysis or by using approximate formulas that

account for the most common cases. Design
should be verified to confirm whether there are
stress-concentration points that may activate
failure mechanisms due to brittle fracture, corrosion, or fatigue.
Examples of peak stresses are thermal stresses
in the austenitic steel cladding of a carbon steel
vessel, thermal stresses in the wall of a vessel or
pipe caused by rapid change in temperature of
the contained fluid, and the stress at a local
structural discontinuity.
Residual stresses (Ref 9) can be defined as
those stresses that remain in a material or body
after being manufactured and processed in
the absence of external forces or thermal gradients. Residual stresses can be defined as either
macro- or microstresses, and both may be present in a component. Macroresidual stresses
vary within the body of the component over a
much larger range than the grain size. Microresidual stresses, which result from differences
within the microstructure of a material, operate
at the grain-size level or at the atomic level.
Microresidual stresses often result from the
presence of different phases or constituents in a
material.
Residual stresses develop during most
manufacturing processes involving material
deformation, heat treatment, machining, or
processing operations that transform the shape
or change the properties of a material. They arise
from a number of sources and can be present in
the unprocessed raw material, introduced during
manufacturing, or can arise from in-service
loading.
In heat treated parts, residual stresses may be
classified as those caused by a thermal gradient
alone or a thermal gradient in combination
with a microstructural change (phase transformation). When a steel part is quenched from
the austenitizing temperature to room temperature, a residual-stress pattern is established due
to a combination of a thermal gradient and a
local transformation-induced volume expansion. Thermal contraction develops nonuniform
thermal (or quenching) stress due to different
rates of cooling experienced by the surface
and interior of the steel part. Transformational
volume expansion induces transformation stress
arising from the transformation of austenite into
martensite or other transformation product
(Ref 10).
Residual stresses may be sufficiently large to
cause local yielding and plastic deformation,
pg 7
18/8/2008 2:49PM Plate # 0
both on microscopic and macroscopic levels,

and can severely affect component performance. Both the magnitude and distribution of
the residual stress can be critical to performance and should be considered in the design
of a component. In any free-standing body,
stress equilibrium must be maintained, which
means that the presence of a tensile residual
stress in the component will be balanced by a
compressive stress elsewhere in the body.
Tensile residual stresses in the surface of a
component are generally undesirable, since
they can contribute to, and are often the major
cause of, fatigue failure, quench cracking, and
stress-corrosion cracking. Compressive residual stresses in the surface layers are usually
beneficial, since they increase both fatigue
strength and resistance to stress-corrosion
cracking and increase the bending strength of
brittle ceramics and glass. In general, residual
stresses are beneficial when they operate in
the plane of the applied load and are opposite
to it (for example, a compressive residual stress
in a component subjected to an applied tensile
load).
The origins of residual stresses in a component may be classified as mechanical, thermal,
and chemical. Mechanically generated residual
stresses are often a result of manufacturing
processes that produce nonuniform plastic
deformation. They may develop naturally during
processing or treatment or may be introduced
deliberately to develop a particular stress profile
in a component. Examples of operations that
produce undesirable surface tensile stresses
or residual-stress gradients are rod or wire
drawing (deep deformation), welding, machining (turning, milling), and grinding (normal or
harsh conditions).
On a macroscopic level, thermally generated
residual stresses are often the consequence of
nonuniform heating or cooling operations.
These, together with the material constraints
in the bulk of a large component, can lead to
severe thermal gradients and the development
of large internal stresses. An example is the
quenching of steel (or aluminum alloys),
which leads to surface compressive stresses
balanced by tensile stresses in the bulk of the
component.
Chemically generated stresses can develop
due to volume changes associated with chemical
reactions, precipitation, or phase transformation.
Chemical surface treatments and coatings can
lead to the generation of substantial residual-
stress gradients in the surface layers of the

component.
The criterion applied to avoid plastic deformation states that the calculated stress intensity
or effective stress must be lower than the yield
and design life creep-rupture stresses of the
material. When effective stress is exceeded
somewhere within the component, it does not
necessarily indicate plastic collapse of the entire
structure. Primary stresses may locally exceed
yield, within certain limits, provided that there is
enough ductility to allow the material to yield
without cracking. Plastic collapse occurs when
primary stresses are uniform on the entire
structure and exceed effective stress. To prevent
an incremental collapse or thermal stress ratchet
in each loading cycle, the total elastic stressintensity range, considering residual and applied
stresses, should be limited to twice the yield
stress.
Factors Leading to Size and Shape Changes
in Heat Treated Components. Within a typical component manufacturing process, there are
seven major factors that lead to size and shape
changes and the development of residual stresses in heat treated components (Ref 8):

Variation in structure and material composition throughout the component, leading to

anisotropy in properties and transformation
behavior
Movement due to relief of residual stresses
from prior machining and forming operations
Creep of the part at elevated temperature
under its own weight or as a result of fixturing
Large differences in section size and asymmetric distribution of material, causing differential heating and cooling during
quenching
Volume changes caused by phase transformation
Nonuniform heat extraction from the part
during quenching
Thermal expansion
All of these factors, except relief of prior

residual stresses (second item) and creep at
elevated temperature (third item), can be
directly related to thermal and transformationinduced strains in the component (Ref 8).
A simple example of how this thermophysical
property affects heat treating is given in Fig. 2
(Ref 11). As the shaft is quenched, the corner
cools first, and as it shrinks, it mechanically
pg 8
18/8/2008 2:49PM Plate # 0
upsets the hot steel beneath it. As the quench

progresses, the entire shaft cools, but now,
because the end is hot upset, the diameter is too
small to accommodate the circumference. As a
result, the end is (usually) in a high state of
residual tensile stress, and if the steel is brittle,
quench cracks may develop.
The coefficient of expansion is a factor that
requires serious design consideration because it
affects a part during austenitization. With furnace heating, a part is heated to the austenitization temperature mainly by radiation (80 to
98%) and partly by convection (2 to 20%). By
radiantly heating particular portions of a part,
thin sections heat fastest, especially those that
expose a large surface area, such as a spline or
gear. A typical example is the gear on the left in
Fig. 3 (Ref 11). Because its thin sections are

relatively rapidly heated, this part could not be
made to the required tolerances. The redesign
shown on the right was an improvement, but the
necessary large access holes were still troublesome.
Several other factors at the design stage can
contribute to problems traceable to austenitization:

Combinations of components with widely

varying (nonuniform) section sizes
Designs requiring contact with furnace
hearths or placement near walls
Designs requiring processing that results in a
state of high residual stress before austenitization.
Parts that are very thin or long or parts that
are large in surface area, which are difficult
to heat treat because of distortion during
austenitization
Designs that are unsuitable for the type of
furnace equipment available
In designing a tool or die, various factors must

be considered. In practice, it is difficult to
separate the design stage from steel grade
selection because the two steps are interdependent. The choice of a certain grade of steel,
such as one that must be brine or water quenched, will affect all aspects of design and manufacture. In general, any steel grade that requires
liquid quenching demands very conservative,
careful design. Air-hardening grades tolerate
some design and manufacturing aspects that
could never be tolerated with a liquid quenching. The design must also be compatible
with the equipment available, for example,
heat treatment furnaces and surface-finishing
Fig. 2
Effect of coefficient of thermal expansion in heat

treating a shaft. Source: Ref 6
Fig. 3
Two gear designs showing the effect of coefficient of thermal expansion. At left is a widely used design, which is very
troublesome to heat-treat. A preferred design is shown at right. Source: Ref 11
pg 9
18/8/2008 2:49PM Plate # 0
devices. Designing tools and dies is more difficult than designing components made from
structural steels because of the difficulty in
predicting service stresses. Despite advances
made in design procedures, much of the design
work is still empirically based. Such experience
is primarily based on past failures; therefore, it is
important that the findings of the failure analyst
be incorporated into future work. Despite the
shortcomings of the empirical approach, there
is a vast body of common-sense engineering
Fig. 4(a, b)
Lathe tool bit of 1.45% C and 1.4% Cr steel

with acute-angled and rough-machined crosssectional transition that fractured during hardening. (a) Fracture.
Original magnification: 1. (b) View into angle. 2. Source:
Ref 13
Fig. 5
knowledge available for guidance.

Analysis of many tool and die failures shows
that two relatively simple design problems
cause the most failures. These design shortcomings are the presence of sharp corners and
the presence of extreme changes in section mass
(Ref 12). A sharp corner concentrates and
magnifies applied stresses, stresses that arise in
tool and die manufacturing (such as during
quenching), or stresses that occur during service.
In addition to promoting cracking during liquid
quenching, sharp corners promote buildup of
residual stresses that may not be fully relieved
by tempering and can therefore reduce service
life. The largest possible fillet should be used at
all sharp corners. Air-quenching grades of steel
are more tolerant of sharp corners than liquidquenching grades and are preferred when only
minimal fillets can be used. Changes in section
size can be the locus of premature failures.
Figures 4 to 8 (Ref 13) show failures caused
by design errors and selection of unsuitable
material.
Figure 4(a) shows the fracture of a lathe tool
bit made of steel with approximately 1.45% C
and 1.4% Cr that was hardened in oil at 870 C
(1600 F), which was at least 20 C (35 F) too
high. The fracture propagated from a rectangular
cross-sectional transition that was not properly
filleted and moreover was rough machined, as
shown by the grooves in Fig. 4(b) (thereby the
notch effect was further aggravated). Many
failures in service, especially those caused by
shock or cyclic loads, can be caused by such
design errors.
Figure 5 shows a bolt from a self-service
elevator that failed as a result of reverse-bending
fatigue. In this case, the fracture also propagated
from a sharp-edged cross-sectional transition.
To avoid further damage and prevent potential
Bolt of a self-service elevator that failed as a result of reverse-bending fatigue. Source: Ref 13
pg 10
18/8/2008 2:49PM Plate # 0
Fig. 6
Different parts of rock drills of different durability and made of a steel with 0.95% C, 1.2% Cr, and 0.25% Mo. (a) Broken drills
had sharp edges in the hexagonal shaft. Original magnification: 2. (b) Drills free of defects had well-rounded-off edges.
Original magnification: 2. (c) Fatigue fractures propagated from the sharp edges. Original magnification: 3. (d,e) Differences are
clearly seen in the cross section of the hexagonal shaft. Etching shows that the failed drills also were surface decarburized, which further
reduced the fatigue strength. Source: Ref 13
accidents, 24 other bolts that had not yet failed

were examined metallographically or in bending
tests. Eight of these proved to have incipient
fatigue cracks in the cross-sectional transitions.
The bolts were partially normalized and partially
heat treated. Their strength was determined from
Brinell hardness to be between 440 and
700 MPa. Cracks had occurred in the annealed
as well as the heat treated bolts, that is, in soft as
well as hard bolts. The higher strength of the heat
treated bolts was made ineffective by the unfavorable design.
Figures 6(ae) show different parts of rock
drills of different durability and made of a steel
with composition 0.95% C, 1.2% Cr, and 0.25%
Mo. They had failed after a short period of service in the hexagonal shaft, while others had
proved free of defects. At first glance, it could be
seen that the broken drills had sharp edges
(Fig. 6a), while those free of defects were
well rounded off (Fig. 6b). Fatigue fractures
propagated from the sharp edges. These, in turn,
led to catastrophic failures under shock loading
(Fig. 6c). The design differences could be
clearly seen in the cross section (Fig. 6d,e).
Etching showed that the failed drills also were
surface decarburized, which further reduced the
fatigue strength.
Figures 7(a,b) show a compressor transmission shaft with a fracture propagating from an
acute-angled keyway, and Fig. 8 shows a drive
shaft pinion with fatigue fractures propagating
from the acute-angular edge of the helical gear.
On the other hand, ignorance, carelessness,
and false economies in the selection of materials
cause many errors (Ref 12). Not every steel user

is in a position to select the most suitable
material for his purpose from the many varieties
available. When in doubt, consult the steel
manufacturer whose materials specialists possess the necessary knowledge of mechanical and
technological properties of the required materials. A close cooperation between the materials
specialist of the producer and the designer and
plant engineers of the user is the best formula for
success.
One of the overwhelming causes of steel
cracking and unacceptable distortion control
is part design (Fig. 911). Poor part design
promotes distortion, cracking, and nonsymmetrical heat transfer during heating and cooling.
Sometimes, designers make designs in which
combinations of parts intended to reduce costs
can actually increase cost due to problems
during austenitization. The classic example is
the gear-and-hub combination shown in Fig. 12
(Ref 11). As this part is heated, the thin extremities at the top of the hub heat faster than the
sections near the gear. Accordingly, this area has
a propensity to increase in size but is restrained
by the colder metal nearer the gear; therefore, it
upsets itself (yields in compression). Finally, the
entire part comes to the prescribed temperature. On cooling, however, and even without
quenching, the top end of the hub pinches in
because it has upset itself. This upsetting can
result in a serious taper condition in the bore. (If
the bore is broached before heat treatment, the
extremities of the hub will stretch and then close
pg 11
18/8/2008 2:49PM Plate # 0
in, thus causing additional taper.) The top

portion of the hub of the pinion in Fig. 12(a) is
used only as a spacer and need not be heat
treated. A much shorter hub with a steel tubing
spacer (Fig. 12b) would solve the problem in
Fig. 7(a, b)
austenitization, making forging easier and, in

most cases, reducing total cost. The bevel pinion
shown in Fig. 13 presents a similar problem,
although the hub extension is necessary. Here,
steel or, preferably, a heat-resistant alloy cap
Compressor transmission shaft with a fracture propagating from the acute-angled keyway. Source: Ref 13
Fig. 10
Design solutions to the distortion problem shown in

Fig. 9. Source: Ref 3
Fig. 11
Distortion often encountered when quenching a

notch. Source: Ref 3
Fig. 8
Drive shaft pinion with fatigue fractures propagating

from the acute-angular edge of the helical gear.
Source: Ref 13
Fig. 9
Schematic of a gear that is difficult to harden without

the distortion shown. Source: Ref 3
pg 12
18/8/2008 2:49PM Plate # 0
that will create mass can be put over the thin hub
before austenitization to retard the heating rate
during carburizing.
Because of the extremely rapid heating
rate, austenitization with high-frequency electric current can be likened to reverse rapid
quenching. Accordingly, design is of utmost
importance. As a general rule, steel exposed
to the flux of the inductor will heat fastest on
corners, around holes (as shown in Fig. 14), and
through thin sections. The bottoms of keyways
and the roots of gear teeth and splines are austenitized last, often mainly by conduction from
adjacent areas.
As in quenching, however, induction tooling

can be designed to concentrate flux by using
appropriate coil configuration and laminated
core material. The different frequencies available provide not only for various depths of
heating but also for the sharpness of the heating
effect, because the induction-heated layer is
often much thinner than the hardened depth of
austenitized steel. The extent of conduction
is a function of the differential between the
surface temperature and that of the core. Thus,
preheating, either in the induction coil with a
suitable delay or in a furnace, can be employed
to reduce heat transfer inward.
Fig. 12
Two designs for gear-and-hub combinations. (a) Difficult to heat treat without excessive taper in the bore. (b) A preferred
design. Source: Ref 11
Fig. 13
Redesign of a bevel pinion using electron beam welding that was impossible to heat treat in one piece. Source: Ref 11
pg 13
18/8/2008 2:50PM Plate # 0
Finally, part design recommendations to

avoid distortion and cracking problems (Fig. 15)
include:

Parts that exceed the following dimensions

often must be straightened or press quenched
to maintain dimensional stability: long and
thin parts, L=5d for water quenching and
L=8d for oil quenching (L is the length,
and d is the thickness or diameter); and parts
that possess large cross-sectional area (A)
and are thin (t), which are defined as A=50t.
Balance the areas of mass.
Avoid sharp corners and reentrant angles.
Avoid sharp corners between heavy and thin
sections.
Avoid single internal or external keys, keyways, or splines.
Provide adequate fillet or radius at the base
of gear teeth, splines, and serrations.
Do not have holes in direct line with the
sharp angles of cutouts.
Avoid sharp corners at the bottom of small
openings, such as in drawing or piercing
dies, because spalling or flaking is likely to
result at these points.
Keep hubs of gears, cutters, and so on as near
the same thickness as possible, because
dishing is likely to occur.
Order stock large enough to allow for
machining to remove decarburized surfaces
and surface imperfections, such as laps and
seams.
Do not drill screw holes closer than 6.35 mm
(0.25 in.) from the edges of die blocks
or large parts, where possible. Cracking
may be avoided by using steel that may be
hardened by using lower quench severity,
or, if possible, pack the bolt hole to
reduce thermal stresses arising due to
quenching.
Avoid blind holes, if possible.
Design all parts with round corners and
Fig. 14

fillets wherever possible.

Use air-hardening or high-carbon (oil- and
air-hardening) tool steel on unbalanced and
intricately shaped dies.
Add extra holes, if possible, on heavy,
unbalanced sections to allow for faster and
more uniform cooling when quenched.
Do not machine knife blades to a sharp cutting edge before hardening.
Avoid deep scratches and tool marks. The
insertion of identification marks on the
hardened component is recommended, preferably after hardening, with tools having
well-rounded edges and minimum deformation (shallow penetration depth) and at
positions far away from the high-stressconcentration zones (reentrant angles,
bends, and so on).
On long, delicate parallels, shafts, and so on,
rough out and have pieces annealed to
remove stresses before finish machining.
Always use the grade or composition of steel
most suitable for the work that the part has to
perform.
Design symmetry is also an important variable to minimize distortion.
A general rule for solving such quench distortion problems is that the short side is the
hot side, which means that the inside of the
bowed metal was quenched more slowly
than the opposite side.
Steel Grade and Condition. Although steel

cracking is most often due to nonuniform
heating and cooling, material problems may
be encountered. Some typical material problems include the compositional tolerances:
dirty steels, those containing greater than
0.05% S (such as SAE 1141 and SAE 1144),
are more prone to cracking. It is well known
that cracking propensity increases with carbon
content. Therefore, the carbon content of the
steel is one of the determining factors for
Section through a hole in a part following rapid heating in an induction coil, showing distortion that leads to cracking.
Source: Ref 11
pg 14
18/8/2008 2:50PM Plate # 0
quenchant selection. As a rule of thumb, plain

carbon steels with less than 0.35% C rarely
crack on hardening, even under severe quenching conditions. The carbon content of steel
should never be greater than necessary for
the specific application of the part. As a general
rule, steels with more than 0.35% C will
require oil quenching to avoid cracking. This
Fig. 15
means that higher-cost, higher-alloy steels are

required for adequate response to the slower
oil quench when carbon content exceeds
0.35%.
However, when strength alone (and/or
hardness), without the toughness of a quenched
and tempered microstructure, fulfills the minimum engineering requirements, the use of
Part design recommendations for minimal internal stresses. Source: Ref 3
pg 15
18/8/2008 2:50PM Plate # 0
cold-finished bars made with extra-heavy draft

or elevated-temperature drawing should receive
consideration. In fact, why heat treat when it is
unnecessary? Thus, SAE 1050, SAE 1140, or
SAE 1144 steels should be particularly attractive
to firms with no heat treating facilities or with no
commercial heat treater nearby. By adjustment
of their composition and the degree to which
they are cold (or warm) worked, these steels can
be made to have good machining characteristics
(Ref 6).
On the other hand, it is well known that
regions containing high concentrations of coarse
carbide microstructure as a result of improper
forging may become the initiation point for
subsequent quench cracking, particularly with
parts of complex shape. It is important to
provide a sufficient forging for microstructure to
become fine and uniform. Because part manufacture, such as gear production, often requires
machining, the condition of the steel that is
going to be machined is critically important.
Some workers have recommended normalized
and subcritical-annealed steels as the ideal
condition. The subcritical annealing process
reduces the carbon content and alloy carbide
content in the austenite, allowing the production
of more lath martensite in the microstructure,
which provides higher fracture toughness and
higher impact toughness.
Steel hardenability is determined by its
chemistry. The quench conditions required to
obtain the desired properties are a function
of the hardenability. Therefore, if the steel
chemistry is incorrect, the selected quench process conditions may, if too severe, lead to
cracking. Unfortunately, this problem is not
uncommon.
Techniques for Controlling Distortion

In applying one or more of the effective
methods of minimizing distortion, cost is usually
the major consideration. Therefore, in planning
manufacturing operations, it behooves the prudent processor to evaluate the costs of minimizing distortion against the alternatives (Ref
14). In almost any instance, there are at least
three alternatives:

Change to another heat treating process

Make allowances for stock removal in finishing operations to correct the distortion
Incorporate straightening operations as required
In considering the alternatives that relate to

minimization of distortion, it is assumed that the
grade of material is fixed, and no deviations are
allowed in this area. There are often instances,
especially for parts of complex design, where a
change in steel composition will permit a less
drastic quench and thereby reduce distortion.
Such changes are usually to steels with higher
hardenability.
In most instances, however, immediate
changes in workpiece composition are not feasible. Pros and cons of the three most likely
alternatives (listed previously) are discussed
separately in the paragraphs that follow.
Consider Change to Another Process. In
this area, there are sometimes two or three possibilities, such as changing from a throughhardening steel to a case-hardening type or
changing to one that does not require rapid
cooling, such as nitriding. One of the most likely
changes that is often made in this area is to the
use of localized heating, such as induction. For
example, a shaftlike member requires hardening
only in certain bearing areas. This can be
accomplished easily by induction and, in addition to eliminating distortion, is often more
economical for other reasons. Parts such as ring
gears represent other examples where a change
to induction hardening resulted in keeping distortion within acceptable limits.
Increase Stock Allowance. In many instances, allowance for stock removal in the
finishing operation (usually grinding) is the most
economical approach. Under these conditions,
some study is usually necessary to determine the
magnitude of distortion caused by heat treating
and thereby how much stock allowance is
required for cleanup. Frequently, it is necessary
to take reasonable precautions (perhaps some
special procedures) in heat treating and then
take further steps by increasing stock removal
in finishing.
Mechanical straightening, either during
processing or applied to heat treated parts, offers
a third approach for solving distortion problems.
Straightening is sometimes used as the sole
technique for correcting distortion, but more
often it is used in conjunction with systematic
stock removal.
Heat Treating Practices for Minimizing
Distortion. Positioning in the furnace may
have a marked influence on total distortion,
especially for parts having a relatively large
length-to-cross section ratio. For example, for
long, shaftlike parts (solid or tubular), the
pg 16
18/8/2008 2:50PM Plate # 0
poorest loading technique would be to pile them

horizontally and more or less indiscriminately
on the furnace hearth. Under these conditions,
distortion begins immediately and continues as
the parts heat up and lose strength. Parts at the
bottom of the pile will naturally distort the most,
because they are subject to the greatest stress
during heating.
In contrast to the poor technique described
previously, the best technique for such parts
would be to hang them (preferably with spaces
between each part) in a vertical furnace for
heating. As a rule, some further improvement
can be achieved by heating in molten salt as
opposed to a gaseous atmosphere. This is due to
the fact that some support is supplied by the
buoyancy effect of molten salt.
One possible disadvantage (relating to distortion) of heating in molten salt is the heating
rate. Parts are heated four or five times as fast in
a medium of molten salt compared to heating in
a gaseous atmosphere. Rapid heating sometimes
increases distortion, especially when various
section thicknesses are involved.
Position during quenching may also have a
marked effect on the total amount of distortion.
Parts that are hung vertically in the furnace to
minimize distortion should likewise be hung
vertically in the quenching tank; that is, deep
tanks are preferred for this type of work. In
most instances, minimum distortion of specific
workpieces is achieved when the quenching
medium is not agitated. To obtain full hardness,
however, agitation usually must be used. When
minimum distortion is required, if agitation is
used, the quenching medium should be agitated
with agitating force at the bottom of the tank.
Although this specific system is used for water,
the principle applies to any quenching medium.
The quenching medium should never be agitated
from the side in such a system when minimum
distortion is important.
Choice of quenching medium can affect
distortion. Typical quenching media listed in
approximate order of decreasing cooling power,
are as follows:

Water
Brine solutions (aqueous)
Caustic solutions (aqueous)
Polymer solutions
Oils
Molten salts
Molten metals
Gases, including still or moving

Fog or mists
Air
The higher rate of heat extraction (quenching

power) is obtained by agitated brine. Minimum
distortion would be obtained by vertical heating,
then cooling (quenching) by hanging in still air.
Of course, this technique would not usually be
practical, largely because the parts would have
to be made from air-hardening steels. There are
several factors that influence choice of quenching medium, but hardenability of the steel is
usually the key factor. Cooling rate thus has a
marked effect on the amount of distortion.
Consequently, the quenching speed should
never be faster than is required to attain the
required critical cooling rate, when distortion is
an important consideration.
Special quenching techniques may be
needed. Vertical heating and quenching one part
at a time could be construed as a special technique. However, commonly recognized special
techniques include:

Martempering
Press quenching
Cold-die quenching
All of these methods can be used effectively to

reduce distortion, but as a rule, they are all
relatively costly and will greatly increase the
total manufacturing cost.
Whenever the handling of parts individually
is involved, the cost of heat treating increases
rapidly. For example, tubes made from 52100
steel, 1.2 m (4 ft) long, 63.5 mm (2.5 in.) in
diameter with a 3.175 mm (0.12 in.) wall thickness were required to be hardened to a minimum
of 60 HRC with a maximum total indicator
reading of 0.75 mm (0.030 in.). No salt bath of
sufficient depth was available for quenching;
thus, martempering could not be considered.
Several different procedures were tried, but the
one that was ultimately successful consisted of
heating vertically (hanging) and quenching in
unagitated oil, one tube at a time. This procedure
cost more than six times the estimated cost for
bundling the tubes, then mass heating and
quenching. Therefore, the cost of such a practice
would normally be considered prohibitively
expensive. Straightening during processing
may have solved the problem, but it also is very
expensive.
Martempering can, under many conditions,
be used to effectively reduce distortion and
can still be applied to mass production. This
pg 17
18/8/2008 2:50PM Plate # 0
depends largely on workpiece shape and size.

However, for long, shaftlike or otherwise
unwieldy workpieces, martempering requires
handling each workpiece individually and is
thus a relatively expensive operation, particularly when simultaneous straightening is incorporated.
Press quenching is probably the oldest special
quenching technique and is still used to a considerable extent. The greatest use of press
quenching is for gears that cannot be heat treated
with sufficient dimensional accuracy by mass
quenching in baskets.
Selection of press quenching should be done
with the full knowledge that it is very expensive.
First, the presses are expensive machine tools.
Second, the dies are expensive, and the dies must
be tailored to the specific workpiece. Also, press
quenching is slow, tedious, and thus expensive.
Minimum distortion is achieved by press
quenching, but heat treating cost is high.
Dry die quenching is another tedious and
expensive process and should be considered
only for highly specialized applications.
Examples of Failures due

to Heat Treatment
Different types of errors and failures produced
in heat treatment (Ref 15) include:
Heating errors: heating too fast causing

stresses in outer zones; heating nonuniformly or locally overheating; heating at
too high a temperature or for too long a time
(distortion, cracking, residual stresses, decarburization, alloy depletion)
Temperature errors: overheating (scaling,
burning, internal oxidation, hot shortness,
grain coarsening, aging, phases precipitation)
Heat treating errors: improper thermal
cycle; too high a temperature; too low a
temperature; improper heating rate; improper cooling rate; improper soaking from
timing errors or nonuniform quenchant;
improper atmosphere control, which is critical in carburizing and nitriding; delay
between quench and temper; improper aging
treatment or postweld heat treatment (improper, unacceptable, or mixed structures and
microstructural features; temper embrittlement; sigma-phase embrittlement; 475 C
embrittlement; sensitization; carburization;
metal dusting; sulfidation; nitridation; disbonding of chromium-molybdenum steels

overlayed by austenitic stainless steels)
Influence of design, steel grade, and condition
are illustrated in the following examples.
Example 1, in Fig. 16 (Ref 16), shows two
AISI W1 carbon steel concrete roughers that
failed after a few minutes of service. Cracking
occurred at the change in section size due
to bending stresses. Although the section
change has a smooth, filleted surface, it is still
a very effective stress concentrator. Subsequent design changes involved a tapered
change in section at the cracked location and
later at the start of the wrench above the
cracked region.
Example 2 illustrates that holes placed too
close to the edges of components are a common
source of failure during heat treatment or in
service. Figures 17(a) and (b) (Ref 16) show an
AISI O1 tool steel die that cracked during oil
quenching. The die face contained numerous
fine cracks. The left side of the die broke off
during quenching. Figure 17(b) shows both sides
of the fracture. Temper color (arrow), typical of
the 205 C (400 F) temper used in this case, is
apparent. This indicates the depth of the crack
produced during quenching that was open during tempering. Coarse machining marks and
deep stamp marks were also present. Sharp,
unfilleted corners may also promote quench
cracking.
Example 3, Fig. 18 (Ref 16) shows a 76 by 87
by 64 mm (3 by 37/16 by 2.5 in.) AISI O1 tool
steel die that cracked during oil quenching. The
cracking pattern (emphasized using magnetic
particles) that emanates from the sharp corners is
visible. A few cracks are also associated with the
holes that are rather close to the edges. Temper
color was observed on the crack surfaces, indicating that the cracks were present before tempering.
Example 4, in Fig. 19 (Ref 16), shows another example of a quench crack initiated by a
sharp corner. This fixture was also made of AISI
O1 tool steel that was oil quenched. In this case,
the corner was filleted, but there was a nick in the
corner where cracking began. The shape of this
fixture is also poor for steel that must be oil
quenched. The thinner outer regions cool more
rapidly, forming martensite first, while the more
massive central region cools at a slower rate. An
air-hardenable steel would be a better choice for
this part.
pg 18
18/8/2008 2:50PM Plate # 0
Example 5 provides another case of a poor

design for liquid quenching, as shown in Fig. 20
(Ref 16). This 76 mm diameter by 76 mm long
(3 in. by 3 in.) threaded part made of AISI W2
carbon tool steel cracked in half at an undercut at
the base of the threads. Figure 20 shows the two
broken halves along with a cold-etched disk
taken from the hollow portion of the part. The
hardened outer case can be seen in the fracture
detail and in the cold-etched disk. Similar
parts, without the undercut, were successfully
hardened.
Example 6, in Fig. 21 (Ref 16), is a punch

made of AISI S7 tool steel that cracked during
quenching because of rough machining marks (a
common cause of quench cracking). Because of
the section size, the punch was oil quenched to
540 C (1005 F), then air cooled. The cracking
pattern has been emphasized with magnetic
particles. Temper color was observed on the
crack walls.
Example 7, Fig. 22(a,b) (Ref 16) show a
classic example of a failure due to improper
Fig. 17
Fig. 16
AISI W1 (0.85% C) tool steel concrete roughers that

failed after short service (2 min for S, 7 min for S11).
Failures occur at the change of sections. Source: Ref 16
(a) Front view of an AISI O1 tool steel die that

cracked during oil quenching. The die face contains
holes that are close to the edge for safe quenching. (b) Side view of
broken die halves showing the mating fracture surfaces and
temper color (arrow) on the crack surfaces. Source: Ref 16
pg 19
18/8/2008 2:50PM Plate # 0
Fig. 18
AISI O1 tool steel that cracked during oil quenching.

Note the cracks emanating from the sharp corners.
The four holes, which are close to the edge, also contribute to
cracking. Source: Ref 16
Fig. 19
Fixture made from AISI O1 tool that cracked during

oil quenching. The design is poor for liquid
quenching. Source: Ref 16
(a)
Fig. 20
electrical discharge machining (EDM) technique. Die cavities are often machined by EDM.
The technique has many advantages, but failures
have been frequently observed due to failure to
remove the as-cast surface region associated
with the as-quenched martensitic layer. Cavity
surfaces must be stoned or ground, then tempered to prevent such failures. Figure 22(a)
shows four 3.2 mm (0.125 in.) diameter EDM
holes in an AISI A4 tool steel primer cup plate.
The holes were finished by jig-bore grinding,
during which spalling was observed at many
of the holes (see upper-right hole). The surface
was swabbed with 10% aqueous nitric acid to
reveal regions affected by EDM. Figure 22(b)
shows the microstructure of these regions.
An as-cast region was present at the extreme
edge (approximately 35.5 HRC). Beneath this
layer was a region of as-quenched martensite
(approximately 63.5 HRC). Next was a backtempered region (approximately 56 HRC) and
then the base-unaffected interior (59 to 61
HRC). The brittle nature of the outer layers and
the associated residual-stress pattern caused the
spalling.
In many EDM-related failures, the as-cast
layer is not observed because of the technique
used or because of subsequent machining. In
these failures, however, an outer layer of brittle
as-quenched (white etching) martensite is present. Such a failure is shown in Fig. 23; this
failure occurred in a plastic mold die made from
AISI S7 tool steel. The crack followed the lower
recessed contour of the larger-diameter gear
teeth and extended to a depth of approximately
1.6 mm. Etching of the surface revealed an asquenched martensite surface layer (thin, white
layer), while the internal structure was grossly
(b)
Threaded part made from AISI W2 carbon tool steel that cracked during quenching at an undercut at the base of the threads.
Source: Ref 16
pg 20
18/8/2008 2:50PM Plate # 0
overaustenitized (note the retained austenite,

white, and coarse plate martensite). Both factors
led to cracking. If the EDM surface layer was not
present, poor service life would have resulted
anyway due to poor microstructural condition.
Example 8, in Fig. 24(a,b) (Ref 16), shows a
41 mm (1.6 in.) square 1.4 kg (3 lb) AISI S5
tool steel sledgehammer head that cracked during quenching. A disk cut from the head was
macroetched, revealing a heavily decarburized
surface (Fig. 24b). Such a condition promotes
quench cracking, particularly in liquid-quenching grades such as S5 (oil quenched), due to
differential surface stresses. A deep stamp mark
also helped promote cracking.
Example 9 presents a situation with stamp
marks, such as that shown in Fig. 25 (Ref 16)
that commonly promote quench cracks. This
was present on an air-quenched die made from
AISI S7 tool steel. In this case, the die was
not tempered, another prime cause of quench
cracking.
Example 10, Fig. 26 (Ref 17) shows SAE
4140 grade steel seamless tubing that
failed because of quench cracks. During production of hydraulic cylinder housings being
fabricated from this steel seamless tubing,
magnetic particle inspection indicated the presence of circumferential and longitudinal cracks
in a large number of cylinders. Figure 26(a) is a
cross section of the tube showing extensive
cracking revealed by dye-penetrant inspection.
Figure 26(b) is a scanning electron microscope
Fig. 21
Punch made of AISI S7 tool steel that cracked during

quenching because of rough machining marks
(a common cause of quench cracking). Source: Ref 16
(SEM) micrograph showing intergranular fracture at a crack origin. Figure 26(c) is an SEM
micrograph illustrating the brittle mode of
failure associated with the fracture. Figure 26(d)
is a micrograph showing the typical concentrations of nonmetallic stringers in the tube
material, and Fig. 26(e) is a micrograph showing a quench crack with a heavy oxide. Although
the steel met the compositional requirements
of SAE 4140, the sulfur level was 0.022%
and would account for the formation of the sulfide stringers observed. The combination of the
clustered, stringer-type inclusions and the
quenching conditions was too severe for this
component geometry. The result was a high
incidence of quench cracks that rendered the
parts useless.
Example 11 presents the case of six wrist
pins designed especially for a high-performance
six-cylinder automotive engine (Ref 18) that
failed after 4800 km (3000 mi) of normal
operation. The wrist pins were made of lowcarbon steel carburized on both the outer and
inner diameters. Two failed wrist pins were
submitted for examination. Sample 1 had fractured into three pieces (Fig. 27). Sample 2 had
not fractured but exhibited circumferential
cracks on the surface of the central zone. Some
of the cracks had progressed for most of the 360
of the pin. Both samples showed some evidence
of scoring on the outer diameter.
The fractured faces of sample 1 were battered
but showed a fairly smooth annular ring around
both the outer and inner diameters, with a ductile
and fibrous core. The condition of the fractured
faces did not permit the definite establishment of
a fatigue failure. Figure 28 shows the dimensions of a pin. The machining on the inside
diameter surface (indicated by B) was relatively rough. The inner diameter had a raised
central section with a small fillet on either side
(indicated by A).
The core (Fig. 29) had a banded microstructure of ferrite and pearlite and contained
some MnS inclusions. The case (Fig. 30) showed
a tempered martensite matrix with a nearly
continuous grain-boundary network of cementite. Other cracks started on the surface caused
by the cementite network (Fig. 31). All cracks
progressed inward from the carburized surface
and circumferentially around the pin. Hardness
tests were performed on the cross section of
the pin, and the depth of case penetration to a
value of 50 HRC was measured. Results indicated a case depth of 0.89 to 1.0 mm (0.035
pg 21
18/8/2008 2:50PM Plate # 0
Fig. 22
(a) Surface of an AISI A4 primer cup plate showing spalling at one of the 3.2 mm diameter holes made by electrical discharge
machining (EDM) Original magnification: 2.5 . (b) Microstructures associated with the spalled hole in (a) caused by
improper EDM technique. Source: Ref 16
pg 22
18/8/2008 2:50PM Plate # 0
pg 23
to 0.040 in.). Determination of case depth by

visual examination on a microspecimen was
0.89 mm (0.035 in.).
The principal causes of failure were inadequate heat treatment of the case and a design that
incorporated a raised central section of the inner
diameter, which acted as a stress raiser. Rough
machining of the inner diameter aggravated the
situation. The case with the cementite grainboundary network had not been heated to a high
enough temperature or long enough to take the
cementite into solution in the austenite. It was
suspected that after slow cooling from the carburizing temperature, pins were heated slightly
above the Ae1 prior to quenching and then given
a low temper. The case was refined, and the core
was unrefined.
Thus, poor heat treatment, resulting in a brittle
grain-boundary network of cementite, and a
design that formed locations of stress concentration in the inner diameter were the most
probable causes of failure.
The pins should be carburized to a double heat
treatment to refine both case and core and to
eliminate the brittle grain-boundary network of
cementite. The pin design should be changed to
eliminate the central raised section of the inner
diameter to avoid the fillets acting as stress raisers. The machining of the inner diameter should
be improved to avoid a rough surface. The depth
of the carburized case should be reduced to
approximately 0.38 mm (0.015 in.) to increase
pin toughness.
Example 12 features cracking of an alloy
steel bolt. A heat treated, cadmium-plated AISI
8740 steel bolt broke through the head-toshank fillet while being handled during assembly (Ref 19). Dimensions of the alloy steel bolt
(MSD 21250-10070) were 15.9 mm (0.625 in.)
(a)
Fig. 23
(b)
diameter, 111 mm (4.375 in.) grip length, and

134.5 mm (5.294 in.) overall length. It was heat
treated to a tensile strength of 1240 to 1380 MPa
(180 to 200 ksi) and a hardness of 39 to 43 HRC
and then cadmium plated per QQ-P-416 type II
class 2 (23 h bake).
The bolt fractured through the head-to-shank
fillet, a type of failure usually traceable to a
poorly controlled manufacturing process, such
as heat treating (quench cracking) or chemical
plating (hydrogen embrittlement). In this instance, delayed cracking caused by hydrogen
embrittlement was initially suspected, because
the bolt reportedly had passed a magnetic particle inspection.
The fracture surface (Fig. 32), has two distinct
zones. Zone 1 was covered with a thick layer of
baked-on scale. The scale was removed and the
area examined using an SEM. The fracture
topography shown in Fig. 33 is a combination of
tearing and intergranular hairline cracks, features often associated with both hydrogen
embrittlement and quench cracking. However,
the heavy adherent nature of the scale suggests
that it formed on the crack surface at high temperature, that is, during heat treating but before
conventional quenching and subsequent cadmium plating. Zone 2 is characterized by
equiaxed dimples, a common feature of ductile
tension overload (final stage of fracture). The
bolt head was cut through its centerline, and
the crack cross section was metallographically
examined in a zone 1 area to further explore
whether failure was caused by hydrogen embrittlement (step cracking, no decarburization)
or quench cracking (temper scale). The crack
had both heavy decarburization and temper scale
(Fig. 34), ruling out hydrogen embrittlement
cracking. However, because quench cracking
(c)
Plastic mold die made from AISI S7 tool steel that was found to be cracked before use. A crack followed the lower recessed
contour of the large gear teeth and had an average depth of 1.6 mm. Smaller cracks were also observed on the flat surfaces.
Source: Ref 16
18/8/2008 2:50PM Plate # 0
occurs at relatively low temperatures, decarburization of the surface could only have
occurred if the crack was present prior to heat
treating. When asked to clarify the situation, the
bolt manufacturer admitted that the part had
been quenched in water from high temperatures
to verify dimension integrity and was returned to
the production lot, instead of being scrapped. It
was assumed that this uncontrolled quench
between hot heading and heat treating caused the
bolt to crack. Decarburization and scaling
occurred during subsequent heat treating of the
cracked part.
The crack in the bolt occurred subsequent to
the hot heading operation prior to the production
run. The bolt was quenched in water, dimensionally inspected, and returned to the production lot instead of being scrapped. The heavy
decarburization layer on the crack surface supports this scenario. A schematic of the quench
crack formation is shown in Fig. 35.
Example 13 involves hydrogen embrittlement failure of several cadmium-plated carbon
steel socket head cap screws (Ref 20). The cap
screws were part of a slide valve assembly on a
regenerator line in a petrochemical plant. The
screws were exposed to Gulf Coast atmosphere,
with no exposure to a chemical process or significant temperatures. The cap screws failed
during initial loading, while maintenance was
being performed on the valve. One failed and
one unfailed cap screw were sent to a laboratory
for analysis.
The as-received cap screws (Fig. 36) were
visually examined. One of the two screws had
fractured at the head-to-shank radius and was
missing its head. Both screws had been sectioned in the threaded part of the shank (17th and
19th threads) approximately 25 mm (1.0 in.)
Fig. 24
(a) AISI S5 tool steel hammer head that cracked

during heat treatment. The fracture was caused by
quench cracking by the decarburized surface (b) and deep stamp
mark (arrows). Actual size. Source: Ref 16
Fig. 25
Quench crack promoted by the presence of a deep,

sharp stamp mark in a die made of AISI S7 tool steel.
Source: Ref 16
pg 24
18/8/2008 2:50PM Plate # 0
(a)
(b)
(d)
(c)
(e)
Fig. 26
A 4140 grade steel seamless tubing that failed because of quench cracks. (a) Cross section of tube showing extensive cracking
revealed by dye-penetrant inspection. (b) SEM micrograph showing intergranular fracture at a crack origin. Original magnification: 90. (c) SEM micrograph illustrating the brittle mode of failure associated with the fracture. Original magnification: 50. (d)
Micrograph showing the typical concentrations of nonmetallic stringers in the tube material. (e) Micrograph showing a quench crack.
Note the intergranular branching and heavy oxide. Original magnification: 400 . Source: Ref 17
Fig. 28
Fig. 27
Failed wrist pin (sample 1), showing fractured faces.

Source: Ref 18
Schematic of wrist pin. Note stress raisers at A and

the rough machining on surface B. Source: Ref 18
from the head of the screw. A crack was

observed in the first thread root below the
unthreaded part of the shank in the fractured
screw. The screws appeared to have been plated.
The crack in the fractured screw was opened to
reveal its fracture surfaces. Both the initial and
laboratory-opened fracture surfaces of the screw
were examined with a stereomicroscope at
magnifications of 7 to 45 and with an SEM at
much higher magnifications. The fracture surfaces displayed similar fracture modes. Figure 37
shows the overall fracture surface. Figure 38
pg 25
18/8/2008 2:50PM Plate # 0
shows a ductile fracture mode, which was

observed over the majority of the fracture surfaces. Figure 39 shows an intergranular fracture
mode, which was observed around the circumference of the screw, next to the plated surface.
The part of the shank containing the fracture
surfaces was metallurgically prepared in cross
section to look for secondary cracking and
pitting. No secondary cracking or pitting was
observed.
To determine whether the cap screws were

plated, a fracture surface and the outside surface
of the unthreaded part of the shank were cleaned
of oil and other deposits and analyzed using
energy-dispersive x-ray spectroscopy (EDS) in
conjunction with SEM examination. The EDS
Fig. 31
Fig. 29
Central longitudinal zone of sample 2, showing

banded structure of white ferrite and dark unresolved
pearlite with MnS inclusions (light gray). 2% nital etch. Original
magnification: 200 . Source: Ref 18
Fig. 30
Surface structure along a longitudinal axis of specimen 2. The dark matrix is tempered martensite; the
light-colored grain-boundary network is cementite. Nital etch.
Original magnification: 200 . Source: Ref 18
Macrograph of sample 2 taken along the longitudinal

axis, showing cracks emanating from both the inner
and outer diameters. Unetched. Original magnification: 15.
Source: Ref 18
Fig. 32
Ref 19
Close-up view of the bolt-shank fracture surface.

Note the heavy scale on the zone 1 surface. Source:
pg 26
18/8/2008 2:50PM Plate # 0
Fig. 33
SEM fractography of a field on the zone 1 surface (see

Fig. 32). Note the combination of tearing and
intergranular fracture. Source: Ref 19
Fig. 34
Optical micrograph of a portion of the crack along a

cross section of the fractured bolt head. Note the
decarburization at the surface of the crack. Source: Ref 19
Fig. 35
Schematic of quench crack formation. Source: Ref 19
Fig. 36
As-received socket head cap screws. Arrow indicates

a secondary crack in the screw thread root. Source:
Ref 20
Fig. 37
Fracture surface of the crack in the failed screw after

the crack was opened in the laboratory. L indicates
the laboratory-induced overload region. Original magnification:
20. Source: Ref 20
pg 27
18/8/2008 2:51PM Plate # 0
Fig. 38
SEM micrograph showing a ductile fracture mode

observed over the majority of both fracture surfaces.
Original magnification: 1000 . Source: Ref 20
Iron
20K
LT= 100 SECS

BASE METAL OF SCREW
COUNTS
15K
5000
Iron
Manganese
10K
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000
(a)
ENERGY
Cadmium
Cadmium
Cadmium
5000
Cadmium
COUNTS
10K
Cadmium
15K
keV
LT= 100 SECS

PLATING ON THE SCREWS SURFACE
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000
(b)
ENERGY
keV
Fig. 40
Energy-dispersive x-ray spectroscopy spectra of (a)

the base metal of the screw and (b) the plating on the
outside surface. Source: Ref 20
results are shown in Fig. 40. It was determined

that the screws were plated with cadmium.
Superficial Rockwell hardness measurements
were taken on the metallographic section. The
average hardness for the failed cap screw was
Fig. 39
SEM micrograph showing an intergranular fracture

mode, observed around the entire circumference at
both fractures in the screw. Structure at top is the base metal;
structure at bottom is cadmium plating. Original magnification:
1000 . Source: Ref 20
80.5 HRN, which converts to approximately

40 HRC.
The presence of a ductile fracture mode at the
core and an intergranular fracture mode at the
outer surface of a plated bolt is typical of
hydrogen embrittlement but could also be stresscorrosion cracking (SCC). However, SCC can
be eliminated, because the metallographic
results showed no evidence of secondary cracks
or other corrosion mechanisms, such as pitting.
Many hydrogen embrittlement mechanisms
have been proposed, but none is universally
accepted. However, the phenomenon of hydrogen embrittlement is widely known. The presence of hydrogen in steel reduces the ductility
of the steel and causes premature failure under a
static load. The time for failure depends on
the stress applied to the component and the
amount of hydrogen that has diffused into the
steel. A component may fail initially when put
under load or may fail several weeks after being
loaded. Because of this characteristic, hydrogen
embrittlement is sometimes called hydrogeninduced delayed failure.
Electroplating is a common cause of hydrogen embrittlement in bolts and screws, because
hydrogen is evolved (or liberated) during the
process. The screws had been plated, and because no other source of hydrogen was identified, it is likely that the plating process was the
source of the hydrogen.
After most plating processes, bolts and screws
are usually baked for several hours at 190 C
(375 F) to diffuse any hydrogen out of the steel.
However, when a bolt or screw is cadmium
plated, it requires a much longer time (approximately 24 h) for baking, because hydrogen
pg 28
18/8/2008 2:51PM Plate # 0
diffuses less readily through cadmium than other

electrodeposited metals. In the case of the cap
screws, the screws either were not baked at all or
were not baked for a sufficient period of time
and/or at a high enough temperature.
The cap screws failed because of hydrogen
embrittlement. The most probable root cause
was absence of baking or insufficient baking
of the cap screws after the cadmium plating
process.
To eliminate the possibility of future hydrogen embrittlement failures, the screws should
be baked at approximately 190 C (375 F)
for 24 h.
Heat Treatment Design

As was analyzed in previous sections, heat
treatments are a series of operations in the course
of which a solid ferrous product is totally or
partially exposed to thermal cycles to achieve
the desired change in structures and properties
(Ref 2). The chemical composition of the
material may possibly be modified during these
operations (thermochemical treatment).
In addition to the ability of the heat treatment
to achieve desired mechanical properties, heat
treatment also produces dimensional changes
and residual-stress patterns that, in some cases,
can lead to component cracking and distortions.
In the following section, a procedure is analyzed to improve the performance of the design
process. A typical component manufacturing
process includes metalworking, machining, or
other forming operations, followed by heat
treatment.
Different types of heat treatment of steels
are usually employed by industry: hardening,
austenitization, annealing, normalizing, stress
relieving, quenching and tempering, and austempering.
Heat treatment processes include component
heatup, holding at temperature for through-heat
solutionizing, or thermochemical treatments
such as carburizing or nitriding, quenching from
elevated temperature, postquench tempering, or
aging treatment. All the steps can influence
dimensional changes, residual-stress patterns,
and cracking in heat treated components.
The Process of Component Design
In order to avoid failures associated with
heat treatments, it is necessary to develop a
systematic component design process. The

typical phases of component design include
planning and requirements definition, concept
design, detail design, and test and validation.
The component is designed to provide a specific
mechanical, thermal, and chemical function
throughout its life cycle and is often limited by
space, cost, and safety considerations. The
selection of materials and manufacturing processes for a cost-effective component design is a
complex process and often involves iterative
decision making. The iterative nature of design
requires a continuous analysis and redesign
process.
Process design employs stress-analysis tools
with stress-concentration factors, design rules
based on experimental data, material property
databases, and mechanical properties resulting
from a broad range of heat treatment processes.
Computer modeling is a valuable design tool for
heat treated components. A computer process
simulation model allows a particular design to
be tested under a specific set of process conditions. The computer software can graphically
display not only the resulting residual stresses
and distortions in the component but also the
associated transient evolution of temperature,
metallurgical phases, volume changes, and
stresses.
Regardless of the procedure used for developing the design, at the end of the design process, the required quality of the product should
be defined and then described in a technical
document, definitive layout, or final design
intended to accomplish the product manufacturing. Only the minimum quality needed
for the product to perform the function intended
should be specified. Overspecifying and including restrictive features in the quality description can cause delays and increase costs to
the buyer.
The design process requires developing
operational definitions. Operational definitions
are, for example, a specific test of a piece of
material or criterion for judgment. Without
operational definitions, a specification is meaningless. A specification for heat treatment containing the clause Avoid long, thin sections
requires operational meaning of long, thin
section. This definition depends greatly on the
quenching media; however, any section length
greater than 15 times the diameter is almost
always characterized as such, and the slightest
nonuniformity in quench will cause it to distort.
As the quenching medium becomes more severe
pg 29
18/8/2008 2:51PM Plate # 0
(i.e., water, caustic quench), this criterion is

reduced to as low as 5 times the diameter. For
larger length-to-diameter ratios, consideration
should be given to fixture quenching or induction hardening.
The component design process can be
divided into two phases:

Phase 1, which corresponds to the basic

definition of the product (including concept
and detail design)
Phase 2, which corresponds to the design
review aiming to prevent failures and minimize risks
Phase 1 deals with the basic definition of the

product. As a first step, the generic type of the
material as well as its geometric configuration
should be selected. The criterion applied to
avoid plastic deformation states that the calculated effective stress must be lower than the
yield and design life creep-rupture stresses of
the material. If the product to be designed is an
element that must withstand not only tension
loads but also bending, torsional, and axialcompressive loads, then the combined effect of
the applied load type, shape, and size and the
material properties should be analyzed. Design
for those elements subjected to axial compression should be intended to avoid not only
plastic collapse but also elastic instabilities.
Elastic instabilities may cause Euler buckling
and local buckling. The occurrence of this
failure mode depends on the geometry of the
elements and on the Youngs modulus of the
material. In general, the relationship between
the Youngs modulus (E) and the density (r),
E/r, should be maximized in order to increase
stiffness. By doing so, yield occurs before
buckling, whereas by increasing specific
strength (sf/r), strength also increases; hence,
buckling occurs before yield. During the basic
design step, brittle fracture should be avoided.
Brittle fracture is associated with very little or
no plastic deformation. A material may fail in a
catastrophicbrittlemanner under stresses
even lower than the allowable design stresses
used to avoid ductile failures. The material
property that controls brittle fracture strength is
toughness. Other factors that have an effect on
brittle fracture are material thickness; local
stress level, including nominal stresses, residual stresses, and stress-concentration factors;
temperature; and loading rate. Carbon and lowalloy steels undergo a transition from ductile
failure mode to brittle failure mode at low
temperatures. If the material is heat treated

during fabrication, it will have adequate
toughness in its final condition. Design must be
reviewed in order to minimize the presence of
notches and defects that concentrate stresses.
When mechanical loads are an alternative,
there may be a risk of fatigue failures.
In fatigue failures, a crack grows in each
loading cycle until the remaining ligament fails
due to ductile or brittle fracture. This phenomenon can occur at stress levels lower than
the allowable stresses for static loads. It should
be emphasized that fatigue failures strongly
depend on design and manufacturing quality,
which is accomplished by increasing fatigue
strength and minimizing stress concentrators.
Technical requirements should be complied
with at the lowest cost. A phase 1 basic design
detailed analysis is beyond the scope of this
work.
Phase 2 design review has the purpose of
assuring that the basic design fulfills the
requirements and reviews the design to avoid
failures. This step verifies that the basic types of
failure modes have been properly controlled by
design and determines the types of damages
associated with each failure mode in order to
implement methods for detection.
Material behavior can be analyzed by
developing models that relate materials attributes, required functions, and manufacturing
processes. Due to the large number of aspects
involved, the problem can be simplified by
considering blocks of knowledge that correspond to specific mechanisms and functions.
Each block of knowledge represents a simplified
model that relates some properties to the
required functions, through the knowledge provided by materials science and engineering. The
use of state-of-the-art criteria, which, in some
cases, are based on practical experience, can
optimize the accomplishment of the analyzed
functions. The results of the analysis are synthesized in the definition of design.
Consider two groups of behavior models.
The first group relates materials attributes
generally known as propertiesthat are well
defined and individually determined. These
properties are component shape and size,
modulus of elasticity, Poissons ratio, ultimate
tensile strength, yield strength in tension, shear
strength, compressive strength, ductility, elongation, fracture toughness, hardness, thermal
conductivity, thermal diffusivity, thermal
expansion, specific heat, density, fracture
pg 30
18/8/2008 2:51PM Plate # 0
toughness, creep strength, creep-rupture

strength, and so on. For these properties, test
and criteria have accurate operational definitions.
The second group of behavior models includes attributes that involve the complex
interrelation of a number of variables associated
with materials, manufacturing (heat treatment)
processes, and service conditions. For this
second group of blocks of knowledge, physical metallurgy is intensely used, together
with the laws of mechanics and empiric knowledge.
Modeling of Heat Treatment

Modeling of heat treatment processes, like
other materials processes such as casting and
welding, is quite complex due to the tight
coupling of various metallurgical transformations and the associated changes in thermal
and mechanical states. A heat-transfer model,
coupling with a phase transformation model, a
thermomechanical model, and a thermochemical model (Fig. 41), is considered.
Fig. 41
Behavior model to analyze heat treatment design
Heat Transfer Model

Heat treatments are a series of operations in
the course of which a solid ferrous product is
exposed to thermal cycles. There are different
types of heat transfer: conduction, convection,
and radiation. During heat treatments, the temperature varies in time as well as in space; these
processes are called unsteady, nonstationary, or
transient. The factor of proportionality thermal
diffusivity, a, defines the rate of change of
temperature, and the heat-transfer coefficient, h,
controls the heat flow through the component
surface. The thermophysical properties, l, r,
and Cp, vary with the temperature. The geometric design will try to avoid mass asymmetries. Large differences in section size and
distribution of material causes differential
heating and cooling during heat treating. The
thermal cycle induces phase transformation and
thermomechanical stresses.
Solid-Phase Transformation Model
The thermal cycle induces solid-phase
transformation. Microstructures that are formed
upon cooling and the proportions of each are
pg 31
18/8/2008 2:51PM Plate # 0
pg 32
dependent on austenitization time, temperature,

cooling history of the particular alloy, and the
composition of the alloy. The transformation
products from austenite decomposition change
from a mixture of ferrite and pearlite to bainite
or martensite with increasing cooling rate.
Isothermal transformation (time-temperature
transformation, or TTT) and continuous cooling
transformation (CCT) diagrams can be used to
predict steel microstructures and hardness after
heat treatment. These diagrams are a set of
curves drawn in a semilogarithmic coordinate
system with logarithmic time/temperature
coordinates that define, in the case of the TTT
diagram, for each level of temperature, the
beginning and end of the transformation of
austenite under isothermal conditions, and, in
the case of the CCT diagram, define each variation in temperature as a function of time during
cooling, the temperature at which the austenite
begins and ends its transformation. Industrial
heat treatments consist of heating and soaking
at a suitable temperature, followed by cooling
at an appropriate rate in order to obtain a structural state closer to that equilibrium (annealing),
reduce the internal stresses without substantially
modifying the structure (stress relieving),
increase hardness by more or less complete
transformation of austenite to martensite and
possibly bainite (quench harden), or obtain a
uniform and fine-grained structure with pearlite
(normalizing). Other heat treatments are applied
to ferrous products after quench hardening
(tempering) or solution treatment (aging) to
bring the properties to the required level. The
process conditions that shall be taken into
account usually include the furnace atmosphere
(for example, temperature and carbon potential),
heating rates, and quench conditions.
Thermomechanical Modeling
Primary stress is a normal or shear stress
developed by the imposed loading that is
necessary to satisfy the laws of equilibrium of
external and internal forces and moments. Secondary stress is a normal or shear stress developed by the constraint of adjacent parts or by
self-constraint of a structure. The basic characteristic of the secondary stress is that it is selflimiting. An example of secondary stress is a
general thermal stress.
Thermal stresses are related to temperature
by the thermal expansion coefficient. The
deformation is proportional to the thermal
gradient through the thermal conductivity, l.

Thermal expansion has a strong influence on the
development of residual stress; whenever the
thermal expansion or contraction of a body is
prevented, thermal stresses appear. If large
enough, they cause yielding, fracture, or elastic
collapse (buckling).
The total induced strain is the result of the
sum of the strain produced by thermal (aE
DT=eth) and transformation (etr) strain due to
local transformation-induced volume expansion. The total induced strain must be accommodated through either elastic (eel) or plastic
(eep) strain, which sums to the total strain:
et =eth +etr = (eel +eep )
(Eq 11)
Thermal Shock. The ability of a material to

resist thermal shock, due to a sudden immersion
in a cold ambient, without cracking depends on
its thermal expansion coefficient, a; the tensile
strength, st, for metals; the Youngs modulus, E;
the thermal conductivity, l; and the heat-transfer
coefficient, h. A temperature change of DT
applied to a constrained body or a sudden
change, DT, of the surface temperature of the
unconstrained component induces a thermal
strain.
Stress Concentration. The elastic strain and
stresses calculated previously are nominal
values that do not take into account local discontinuities such as holes, notches, or section
changes. Even on a structure where stress
intensity has been limited by yield criteria, there
may exist highly localized regions where
peak stresses are several times higher than
yield. Maximum local stresses on a structure can
be determined considering nominal stresses
multiplied by a stress-concentration factor and
can be estimated through a detailed stress analysis or by using approximate formulas that
account for the most common cases. Design
should be verified to confirm whether there are
stress-concentration points that may activate
failure mechanisms due to brittle fracture, corrosion, or fatigue. A sharp corner concentrates
and magnifies applied stresses, stresses that arise
in tool and die manufacturing (such as during
quenching), or stresses that occur during service.
In addition to promoting cracking during liquid
quenching, sharp corners promote buildup of
residual stresses that may not be fully relieved
by tempering and can therefore reduce service
life. The largest possible fillet should be used
at all sharp corners. Air-quenching grades of
18/8/2008 2:51PM Plate # 0
steel are more tolerant of sharp corners than

liquid-quenching grades. Fatigue fractures propagate from the sharp edges.
Residual stresses can be defined as those
stresses that remain in a material or body after
being manufactured and processed in the
absence of external forces or thermal gradients.
After the cooling processes have been completed, the residual-stress distribution between
the surface and core is obtained. If the surface
stresses exceed the hot yield strength of the
material, it plastically deforms, resulting in
thermally induced dimensional changes and
thermal and transformational stresses. In heat
treated parts, residual stresses may be classified
as those caused by a thermal gradient alone and a
thermal gradient in combination with a microstructural change (phase transformation).
Distortion is any change in the shape and
original dimensions of a ferrous product occurring during heat treatment. For a given geometry
and heat flow, the distortion is minimized by
selecting materials with large values of l/a.
Shape and volume changes during heating and
cooling can be attributed to three fundamental
causes:

Residual stresses that can cause shape

change when they exceed material yield
strength
Stresses caused by differential expansion
due to thermal gradients
Volume changes due to transformational
phase change
Quench Cracking. If thermally induced

stress exceeds the local tensile strength of the
material, cracking results.
Thermochemical Model
Thermochemical treatments may be applied
to a ferrous product in the austenite state to
obtain a surface enrichment in carbon (carburizing), which is in solid solution in the austenite.
Carburizing can be done in gas atmosphere,
in solid medium, or in a bath of molten salt.
Other thermochemical treatments can be applied
to produce surface enrichment in nitrogen
(nitriding); in nitrogen and carbon (nitrocarburizing); in sulfur, carbon, and nitrogen
(sulfidizing); in silicon (siliconizing); in chromium (chromizing); in boron (boriding); and in
aluminum (aluminizing). By error in the heat
treatment, the surface can be decarburized.
Table 3 shows guidelines to avoid heat treatment failures during the design review (phase 2
design review); the recommendations come from
the examples presented in the previous sections.
Failure Aspects of Welded Components

Brittle
Fracture. Low-temperature/lowtoughness fracture is sudden failure of a structural component that is usually initiated at a
crack or defect. This is an unusual occurrence,
because design stresses are normally sufficiently
low to prevent such an occurrence. However,
Table 3 Phase 2 design review to avoid heat treatment failures

Characteristic of the material
prior to heat treatment
operation
Always use the grade or composition of steel most suitable for the work that the part has to perform.
The carbon content of the steel is one of the determining factors for quenchant selection; plain carbon steels
with less than 0.35% C rarely crack on hardening.
Avoid structural and compositional material heterogeneity.
Avoid dirty steels (clustered, stringer-type inclusions, high sulfur).
Geometry: shape and

dimensions
Avoid asymmetric design.

Avoid points of stress concentration (sharp corners, blind holes, reentrant angles, single internal or external
keys, deep keyways, splines, holes, grooves, very coarse machining marks, deep scratches, and tool
marks). Provide adequate fillet or radius at the base of gear teeth, splines, and serrations. Do not have
holes in direct line with the sharp angles of cutouts. Do not drill screw holes closer from edges.
Add extra holes, if possible, on heavy, unbalanced sections to allow for faster and more uniform cooling
when quenched.
Order stock large enough to allow for machining to remove decarburized surfaces and surface
imperfections, such as laps and seams, and to correct the distortion.
Heat treatment operation
Verify design suitability for the type of furnace equipment available.

Avoid thermal shock.
Avoid decarburized surface.
Consider contact with furnace hearths or placement near walls.
Avoid heat treatment errors.
Control positioning in the furnace and during quenching.
Incorporate straightening operations as required.
Avoid creep of the part at elevated temperature under its own weight or as a result of fixturing.
pg 33
18/8/2008 2:51PM Plate # 0
some older equipment with thick walls, equipment that may be subjected to low temperature
due to an upset, or equipment that may be
modified could be susceptible to varying degrees
of embrittlement.
Brittle fracture is associated with very little or
no plastic deformation and with a cleavage
fracture or intergranular surface, unlike ductile
fracture, which is associated with a fibrous surface. A material may fail in an unstable and
catastrophic brittle manner under stresses even
lower than the allowable design stresses used to
avoid ductile failures. This may occur with a
combination of material properties and applied
stress levels.
The material property that controls brittle
fracture strength is toughness. Other factors that
have an effect on brittle fracture are material
thickness; local stress level, including nominal
stresses, residual stresses, and stress-concentration factors; temperature; and loading rate.
Carbon and low-alloy steels undergo a transition
from ductile failure mode to brittle failure
mode at low temperatures. Resistance to crack
propagation is measured through fractomechanical tests. Crack propagation will occur when the
stress intensity at the crack tip, K, reaches a
critical value, Kc (MPa m1/2). Brittle materials
are those that remain elastic until breaking
(break occurs before yield).
According to API 581 (Ref 21) lowtemperature/low-toughness fracture of steel is
affected by:

Applied loads: Fracture is less likely at low

applied loads.
Materials specification: Some materials are
manufactured to exhibit good fracture
properties or toughness properties. Materials
are often qualified for use by performing an
impact test that measures the energy needed
to break a notched specimen. Fine-grained
structures, such as tempered martensite, with
low impurity content are associated with
a high degree of toughness. Other microstructural elements, such as precipitates,
second-phase particles, dislocations, and
solutes in a solid solution, contribute to
increased yield strength but reduce toughness.
Temperature: Many materials (especially
ferritic steels) become brittle at a temperature called the transition temperature. Brittle
fracture is typically not a concern above
300 C (570 F).

Residual stresses and postweld heat treatment

Thickness
Temper embrittlement is one of the main

causes of toughness degradation in ferritic steels
during high-temperature service. This degradation may lead to component failure during service. The problem arises when some types of
steels are exposed to temperatures between 345
and 565 C (650 and 1050 F). Typically, 21/4
Cr-1=2 Mo steels with a bainitic structure are
the most susceptible to this phenomenon.
Temper embrittlement can also occur in
C-1=2 Mo, 1Cr-1=2 Mo, 11=4 Cr-1=2 Mo, 3Cr-1Mo, and
5Cr-1=2 Mo steels. Conversely, 9Cr-1Mo steels
are less susceptible. Welded joints (weld metal
and heat-affected zone) are the most susceptible
zones.
In all cases, the solution to the problem lies in
alloy purity. Exposures within the critical temperature range may occur during temper or
postweld heat treatments or during service, and
these conditions should be avoided. However,
many components operate within the critical
temperature range.
The segregation of residual elements antimony, arsenic, phosphorus, and tin toward
austenitic grain boundaries is the main cause
of temper embrittlement. Also, manganese and
silicon play an important role in this segregation, and their content should be limited.
Both residual and alloy elements can segregate, but the former can be concentrated up to
300 times their average value in the material.
Segregation only occurs in ferrite within a 315
to 540 C (600 to 1005 F) temperature range
but never occurs during austenitization. In
addition to segregation in grain boundaries, a
fine precipitation can occur within the grains,
resulting in a strength increase (Mo2C precipitation).
The phenomenon associated with changes in
grain boundaries causes intergranular brittle
fracture. In general, ductility and rupture
strength are not affected; nevertheless, both
can be reduced under severe conditions.
Toughness is affected by up to a 100 C
(212 F) shift toward the right of the ductilebrittle transition curve, as evidenced from
impact testing.
When a material is exposed to a 370 to 565 C
(700 to 1050 F) temperature range, property
degradation may become irreversible; in this
case, temper embrittlement and creep operate
pg 34
18/8/2008 2:51PM Plate # 0
pg 35
simultaneously. During equipment operation

in a hydrogen service environment, hydrogen
may diffuse into the metal. During cooling
from the operating temperature (shutdown), the
material becomes oversaturated with hydrogen.
The combination of thermal stresses and
hydrogen oversaturation may lead to hydrogeninduced cracking. If the material toughness has
been reduced considerably due to temper
embrittlement, the risk of a catastrophic cracking is high.
Material degradation due to temper embrittlement may result not only in brittle fractures
with catastrophic consequences but also in a
reduction of the equipment useful life and in a
decrease in the equipment reliability and efficiency, since it may be necessary to operate at
lower temperatures to avoid temper embrittlement or to depressurize to avoid stresses when
the equipment is cold.
Temper embrittlement is reversible. Heat
treatment for a short period of time at temperatures above 565 C (1050 F), followed by
quick cooling, can restore the initial properties.
However, if material thickness is high and
cooling rates required by precipitation kinetics
are not achieved, embrittlement may reoccur.
Moreover, the component may crack during
heat treatment due to the effect of thermal
stresses.
Material thickness over 25 mm (0.98 in.) is
more susceptible to brittle fracture due to temper
embrittlement. Postweld heat treatments minimize the susceptibility to brittle fracture through
this mechanism.
Aging Tendency. The tendencies for aging
after cold working increase hardness and tensile
strength, with a simultaneous reduction in ductility. Therefore, the risk of embrittlement due to
welding in cold-worked areas increases. The
killed steels show enhanced resistance to aging
in normalized conditions. When the degree of
deformation is high, the material should be
thermally treated (normalized or stress relieved)
before the deformed zone is welded. The risk of
embrittlement due to welding in cold-worked
areas to be welded, without special requirements, is allowable under the following conditions. The relationship between the inner
bending radius (r) and the plate thickness (t)
shall be:
r/ti1.0 and tj4.0 mm (0.16 in.)
r/ti1.5 and tj8.0 mm (0.32 in.)
r/ti2.0 and tj12.0 mm (0.47 in.)

r/ti3.0 and tj24.0 mm (0.94 in.)

r/ti10.0 and all thicknesses
These conditions correspond to an elongation,

e=1/(2r/t+1), in the cold-worked areas of 33,
25, 20, 14, and 5%, respectively.
Hardening Tendency. The maximum hardness in the heat-affected zone (HAZ) depends
on the chemical composition (carbon equivalent) and the cooling rate during welding (t8/5).
Carbon most affects hardenability, and its
effect and that of other elements have been
included in carbon equivalent formulas. The
International Institute of Welding carbon equivalent formula, recommended for steel with
more than 0.18% C, is:
CEIIW =C+
Mn Cr+Mo+V Cu+Ni
+
+
6
5
15
(in wt%)
(Eq12)
When carbon is 50.18%, it is generally recommended that the percentage of cementite

(Pcm) formula be used:
Si Mn Cu Ni Cr Mo V
+
+ + + +
+
30
20
40 60 20
15
10
+5B (in wt%)
(Eq13)
PCM =C+
Hardenability of steel is not necessarily an

indicator of HAZ hardness. It is important to
control the maximum hardness in the HAZ in
order to avoid two main problems: cold
cracking (hydrogen-assisted cold cracking)
during welding fabrication, in which the limit
of maximum hardness is usually 350 HV; and
in-service cracking in hydrogen environments,
where the maximum hardness shall be less
than 220 HV. Maximum HAZ hardness can
be accurately calculated as a function of
chemical composition and the cooling time
from 800 to 500 C (1470 to 930 F) (t8/5)
(Ref 22).
The HAZ maximum hardness for low-alloy
steels can be calculated using the following
expressions by Duren (Ref 22). For 100% martensite (alloying elements in wt%):
HVM =802 C+305
(Eq 14)
For 100% bainite:

HVB =

Si Mn Cu Cr Ni Mo V
350 C+ +
+
+ + +
+
11
8
9
5
17
6
3
+101
(Eq15)
18/8/2008 2:51PM Plate # 0
For 05%martensite5100:
h
HVx =2019 C (170:5 log t8=5 )

Si Mn Cu Cr Ni Mo V
+0:3
+ +
+ + +
+
11
8
9
5
17
6
3
+66 (170:8 log t8=5 )
(Eq16)
where the hardness of martensite is HVM, bainite is HVB, the variable amount of martensite is
HVx, and t8/5 is the cooling time between 800
and 500 C (1470 and 930 F).
Cold cracking can occur after the welding
process if four factors are present: local hydrogen concentration, susceptible weld metal or
HAZ, local metal hardness, and a high level of
residual stress remaining after welding, which
may cause cracking at temperatures less than
100 C (212 F).
The causes of cold cracking are related to
many factors, including initial weld metal
hydrogen content, residual hydrogen content at
100 C, steel carbon equivalent, yield stress of
steel or weld metal, heat input, preheat temperature, material thickness, joint restraint
intensity, notch concentration factor, welding
process thermal efficiency, and others. Cold
cracking is a diffusion-controlled phenomenon
that requires days or weeks at room temperature
to develop cracks.
Hydrogen-induced cracking in the HAZ can
be parallel to the fusion boundary adjacent to a
fillet weld or in the form of toe cracks. Weld
metals are by no means immune when the steels
possess high yield strength. Thus, cold cracking
induces surface-connected cracking or subsurface cracking, which may provide initiation
points for further cracking by brittle fracture or
fatigue.
Segregation Tendency. Element (phosphorus, sulfur, carbon, etc.) segregation impairs
weldability. Chemical heterogeneity can contribute to localized increases of hardenability.
Thus, a normal chemical composition of the heat
of the steel may exhibit hardness higher than the
maximum allowable hardness in certain parts of
the HAZ, despite a normal chemical composition. From the point of view of segregation
behavior, semikilled and killed steels are better
than rimmed steels. If segregation zones are
involvedas in butt weldingcare should be
taken to limit penetration and hence minimize
weld metal dilution. In addition, suitable filler
metal and low-hydrogen basic electrodes should
be used. The annealing treatment relieves
internal stresses, restoring ductility and toughness, refining grains, reducing gaseous content
in the steel, and improving homogenization of
alloying elements.
Stress-corrosion cracking (SCC) of austenitic stainless steels can be caused by chlorides and polythionic acids, by hydrosulfuric acid
on carbon and low-alloy steels, and by caustic
corrosion on carbon steels. Stress-corrosion
cracking may arise when a susceptible material
is simultaneously combined with certain levels
of tensile stresses and a critical environment
within a specific temperature range. Tensile
residual stresses, resulting from manufacturing
processes such as welds, contribute to cause this
type of damage. Also, cold plastic deformation
causes hazardous residual stresses. Tensile
stresses must be reduced by controlling manufacturing processes and design. Stress relieving
postweld heat treatments can be used to minimize susceptibility to SCC in austenitic stainless
steel, attempting to avoid sensitization by using
stabilized or low-carbon grades.
Heat Treatment Procedures Applied

to Welded Components
Carbon and Low-Alloy Steels
Postweld heating (soaking) is an option for
carbon-manganese and low-alloy steels. Material of thick-walled low-alloy steel pressure
vessels is susceptible to hydrogen-induced
cracking, and there are many difficulties in
detecting small cracks in the HAZ of heavysection steel by conventional nondestructive
examination methods. Therefore, it is very
important to reduce to a minimum the risk of
hydrogen-induced cold cracking of weldment of
thick plate under the heavily restricted condition
during the entire welding process. One possible
solution is to apply both preheating and postheating (Ref 23). There is a temperature range
where weld hydrogen-induced cracking may
occur and a temperature above the upper limit
where no delay cracking occurs, although the
hydrogen content or the restraint intensity is
high. Therefore, if postheating is carried out
above this critical temperature, the discharge of
hydrogen is allowed; then, the material can
reach room temperature with a minimum risk of
hydrogen-induced cracking.
The purpose of soaking is to allow hydrogen
diffusion to avoid critical values in the
pg 36
18/8/2008 2:51PM Plate # 0
pg 37
weldments. Welding must be performed, maintaining preheating and interpass temperatures

that depend on the material, process, thickness,
type of joints, and heat input. The postheating
must be carried out immediately after welding,
not allowing the temperature to be lower than
120 C (250 F). The temperature for postheating generally is 300 C (570 F) for 4 to
6 h, depending on the steel type. Then, the
weldment is cooled to room temperature. Generally, the same devices recommended for preheating are used. This treatment does not
produce either stress relief or microstructural
changes.
Postweld heat treatment (stress relieving)
may be used for carbon-manganese and lowalloy steels. Postweld heat treatment (PWHT) is
a uniform heating of a weldment at a temperature below the critical range to relieve the major
part of the residual stresses, followed by uniform
cooling in still air. The PWHT is carried out to
fabricate vessels to increase fracture toughness
and minimize the levels of residual stress, which
confers resistance to brittle fractures, tempers
material structure, and removes the possibility
of SCC. The accrued benefits depend on the
material under consideration.
It is usually required by codes and customers
that manufacturers should heat treat all welds on
a thick-wall pressure vessel, including any repair
welds. The PWHT is conducted at a temperature
and for the period specified in the applicable
fabrication codes. There are two possibilities:
one is to stress relieve the completed vessel in
the furnace as a whole; the other is to stress
relieve the subassemblies separately in the furnace and then heat treat the final circumferential
seam locally. Techniques for PWHT are somewhat similar to those thermal methods in preheating and soaking. When local treatment is
carried out, care must be taken to ensure full
through-thickness heating, and similarly, the
temperature gradient is such that the length
of material on either side of the weld at a
temperature exceeding half the treatment temperature is at least 2.5(r t)1/2, where r is the bore
radius, and t is the material thickness. Problems
can arise where butt welds attaching pipes to
nozzles positioned close to nozzle/vessel welds
require separate heat treatment. In such cases,
it may be necessary to apply individual stress
analysis to verify the proposed conditions
(Ref 23).
For both local and furnace treatment techniques, it is very important to exercise control over
heating rates (particularly at temperatures below

300 C, or 570 F, with complex components),
time at soak temperature, soak temperature, and
cooling rate to avoid such undesirable events as
flame impingement, distortion, overheating, air
quenching, and reheat cracking. Very largediameter pressure vessels that have transport
difficulties may be erected on site and PWHTed
from inside by gas burners, using the vessel as its
own furnace (Ref 23).
It is well known that the mechanical properties
of material are degraded due to stress relieving in
some materials (Ref 23). Generally, in low-alloy
pressure vessel steels, the yield strength, tensile
strength, and toughness diminish, and elongation
and reduction of area increase as the temper
parameter (TP) is increased:
TP=T (log t+20) 103
(Eq 17)
where T is temperature (K), and t is time (h). The

PWHT is not considered to be a significant
variable for carbon and carbon-manganese steels
up to 50 mm (2 in.) (Ref 23).
Postweld Heat Treatment of Stainless Steels
Postweld heat treatment is used for stainless
steels (Ref 2426). Welding stainless steels in
thick sections, when the thickness exceeds
approximately 20 mm (0.8 in.), is a complex
operation. In addition to codes or engineering
specifications, which may impose definitive
procedures, it is important to have sufficient
metallurgical background to understand what
may happen during welding and the subsequent
heat treatment operation.
The stresses induced by welding often need
to be eliminated if dimensional stability of the
construction is to be guaranteed and if resistance
to SCC is mandatory. The properties of welded
joints of stainless steels in thick sections that
may be modified during a PWHT are principally
the corrosion resistance and the mechanical properties. The possibility of distortion
occurring during heat treatment must also be
considered.
For the purpose of analyzing PWHT in
stainless steels, the materials considered were
divided into four groups.
Chromium Steels. For both ferritic and
martensitic types, some codes recommend preheating in the temperature range from 150 to
400 C (300 to 750 F) to avoid problems with
hydrogen-induced cracking in welding these
steels.
18/8/2008 2:51PM Plate # 0
Martensitic stainless steels are normally

PWHTed between 600 and 800 C (1110 and
1470 F), whereas ferritic steels are sometimes
heated between 730 and 800 C (1345 and
1470 F), with rapid cooling in order to avoid
embrittlement. Treatments above 900 C
(1650 F) in these materials are usually intended
as homogenizing treatments in order to achieve
better properties after the tempering treatment
that follows. The mechanical properties are
usually improved after this double heat treatment. Low-temperature stress relieving should
not be applied to the straight chromium stainless
steels, since it may markedly affect the ductility
and toughness. The phenomenon is known as
475 C (885 F) embrittlement, and it is due to
the coherent precipitation of chromium-rich
ferrite, known as alpha prime, within the miscibility gap of the iron-chromium system. This
precipitation leads to a slow increase in hardness
accompanied by a corresponding loss of toughness. The alpha-prime phase also decreases the
corrosion resistance. This type of structural
change can be reversed by an annealing treatment at approximately 600 C (1110 F). In
17% Cr steels containing nickel and molybdenum, the toughness is increased by tempering at
630 to 650 C (1165 to 1200 F), below the
temperature where austenite or ferrite is formed.
In straight 17% Cr steels, the precipitation of
sigma phase can occur between 550 and 800 C
(1020 and 1470 F), and it is accompanied by a
loss of ductility. The sigma phase is formed only
after a very long time and may be eliminated by
heat treating above 800 C (1470 F).
Soft martensitic stainless steels have resulted in an increasingly worldwide use in
petrochemical and chemical plants or industries,
gas turbine engines, turbine blades, compressors
and discs, and in a variety of aircraft structural
and engine applications (Ref 27). They have
high proof strength and high toughness even in
very low temperatures or thick cross sections
(Ref 28, 29). If the 12% Cr stainless steels are
used as high-strength structural steels, they must
be weldable, formable, and have good impact
toughness (Ref 30). Hence, in soft martensitic
stainless steels, the carbon content is kept below
mass 0.1% to improve weldability by promoting
a structure with fewer tendencies for cold
cracking, better corrosion resistance, and better
toughness. Because of the lower carbon, the
addition of 4 to 6% Ni (the most powerful
austenite former after carbon and nitrogen)
is required to avoid delta ferrite, which is
deleterious to impact toughness. For enhanced

corrosion, temper embrittlement, and tempering
resistance, 0.5 to 2% Mo is added, depending on
the intended use. In order to develop the maximum strength and toughness, the steel must be
mostly martensitic after cooling, with limited
delta ferrite. The martensite must be tempered to
obtain good toughness, ductility, and stresscorrosion resistance. In the as-welded condition,
the microstructure consists of low-carbon martensite, some presence of delta ferrite, and
retained austenite in agreement with the nickel
content of the alloy. Postweld heat treatments
are necessary to satisfy the service mechanical
property requirements (Ref 31). If high impact
values are required, PWHTs such as solution
annealing plus tempering or double tempering
are necessary (Ref 32, 33). The aim of solution
annealing is the homogenization of the microstructure by dissolution of the delta ferrite,
which is a nonequilibrium solidification product.
The delta ferrite is harmful since it increases the
ductile-brittle transition temperature. On the
other hand, intercritical tempering at 600 C
(1110 F) or double tempering (hypercritical
plus intercritical) at 670+600 C (1240+
1110 F) produces tempered martensite with
finely dispersed austenite that is stable and not
transformable during cooling (Ref 33). It is
known that this austenite, which can be observed
only by scanning electron microscopy, increases
toughness sharply, although it slightly reduces
the strength. It has been argued that when
retained austenite is present near a propagating
crack, the concentrated strain at the crack tip
induces transformation into martensite. This
mechanically induced transformation would
absorb energy and thus increase the toughness.
The associated volumetric expansion of the
martensitic transformation would tend to close
the crack and relieve stresses at its tip. The
latter mechanism absorbs strain energy during
fracture and therefore limits crack extension
(Ref 34).
In many codes for austenitic chromiumnickel stainless steels with low ferrite content
and fully austenitic alloys, no PWHT is prescribed. In the case of austenitic chromiumnickel steels, low-temperature treatment (400 to
525 C, or 750 to 975 F) will help to achieve
dimensional stability of the construction by
reducing peak stresses, although this treatment is
not frequently used. Treatment in the temperature range from 550 to 1170 C (1020 to
2140 F) is a true stress-relieving treatment. The
pg 38
18/8/2008 2:51PM Plate # 0
highest part of the range (960 to 1170 C, or

1760 to 2140 F) involves solution treatment
and achieves maximum relief of stress; it may be
performed to dissolve most of the carbides and
sigma phase or delta ferrite. Treatments in the
range from 1040 to 1170 C (1905 to 2140 F),
followed by water quenching after annealing,
are applied to prevent intercrystalline corrosion
and SCC. When PWHT is recommended, it only
applies to heavy thickness, and a temperature
between 900 and 1000 C (1650 and 1830 F) is
chosen, followed by water quenching or air
cooling, depending on the thickness of the
component. Although it may seem difficult to
imagine such treatments being applied to heavy
components, there are some industrial examples.
The stress relieving of austenitic steels is not
usually applied except for very thick sections.
An exception is in the case of cladding, where
the stress-relieving temperature is chosen with
respect to the base material, and it is frequently
in the range of 540 to 700 C (1005 to 1290 F).
Austenitic stainless steels have a higher coefficient of thermal expansion and lower thermal
conductivity than ordinary ferritic steels, so
greater distortion of welded components must be
expected.
Unstabilized Austenitic Stainless Steels
(UNS S30400, S31600, S30403, and
S31603). These grades normally possess
excellent weldability, provided they are welded
with filler metals that yield an austenitic-ferritic
weld metal to avoid hot cracking during welding
(5 to 15 ferrite number). It is necessary to follow
certain procedures in order to achieve sufficient
corrosion resistance, cracking resistance, and
toughness. It is well known that austenitic
stainless steels are always subjected by the steel
manufacturer to a solution-annealing treatment,
normally in the range of 1050 to 1100 C (1920
to 2010 F). In the course of this heat treatment, carbide M23C6, sigma phase, and delta
ferrite are completely dissolved, and the annealing process produces a homogeneous, fully
austenitic structure. With a subsequent quenching treatment, this state is maintained up to
room temperature. If possible, the PWHT of
welded components should be avoided, with the
exception of a solution-annealing treatment.
However, if heat treatment cannot be avoided,
special attention must be paid to the influence of
carbide and phase precipitations on the corrosion-resistance and toughness properties of the
weld. The precipitation of sigma phase is the
most important of all precipitation phenomena,
apart from M23C6 precipitation, particularly

regarding mechanical properties. Due to the
highest chromium content of delta ferrite, the
weld metal containing delta ferrite is often more
precipitation prone than the base metal of similar
composition (because the solution-annealing
treatment normally applied to the base metal
dissolves any delta ferrite).
Stabilized Austenitic Stainless Steels. This
group of steels contains grades that are alloyed
with titanium or niobium in order to improve
their intergranular corrosion resistance. Titanium- and niobium-stabilized steels can be
welded using niobium-stabilized filler metals
with delta ferrite contents in the range of 7 to 15
ferrite number. Whenever possible, heat treatment after welding, for example, stress-relieving
treatment, should be avoided. If it cannot be
avoided, it is important to use special filler
metals with less delta ferrite content to avoid
sigma precipitation and detrimental effects on
intergranular corrosion resistance. Stabilized
steels are somewhat more susceptible to sigmaphase precipitation or to knife-line attack.
Fully Austenitic Stainless Steels. This
group of stainless steels has a stable austenitic
structure that must normally be welded with
fully austenitic filler metals that do not produce
any ferrite in the weld deposit. In the event of a
PWHT, the intergranular corrosion attack range
is relatively strongly influenced by the chromium and nitrogen content. The presence of a
small amount of sigma-phase or chi-phase precipitation is sufficient to give a marked drop
in the pitting resistance. A 475 C (890 F)
embrittlement does not occur in fully austenitic
weld metal, because this only occurs in the
presence of a ferritic structure.
Duplex austenitic-ferritic chromiumnickel stainless steels having delta content in
the range of 30 to 60% are considered. An alloy
containing 25% Cr and 5% Ni, which, in
metallurgical terms, is very close to the widely
used duplex stainless steel 22Cr-3Mo-5Ni,
solidifies completely to delta ferrite from melting. During further cooling, d-c transformation
starts at approximately 1200 C (2190 F) with
the precipitation of predominantly nodular austenite at the ferrite grain boundary. During further cooling to room temperature, there is only a
partial transformation to austenite. The structure
now contains some 60% primary delta ferrite
and approximately 40% secondary precipitated
austenite. If such an alloy is now subjected to a
solution-annealing treatment, this ratio can be
pg 39
18/8/2008 2:51PM Plate # 0
shifted to slightly higher austenite contents.

Preheating of duplex steels is not normally
required. With thick material, a preheating
between 100 and 150 C (212 and 300 F) may
be advantageous. In order to obtain high ductility in the welded joint, a solution-annealing
treatment, followed by water cooling of the
completed welded component, is not normally
necessary. If it is required, however, the temperature must be set according to manufacturers
specifications. With a solution-annealing treatment in the range of 1020 to 1100 C (1870 to
2010 F), very close to a metallurgical equilibrium, it is possible to reverse any harmful
structures in the HAZ that may have occurred
during welding. It should be noted that this may
lead to severe distortion.
Heat Treatment of Austenitic-Ferritic
Dissimilar Joints. The PWHT of austeniticferritic dissimilar joints or weldments should be
avoided whenever possible. However, heat
treatments such as annealing or stress relieving
may at times be unavoidable or even mandatory.
They should always be adapted to suit the
requirements of the low-alloy steel section of
the joint. Often, the steel in question is a lowalloy creep-resistant steel type used in boiler
and pressure vessels that demand a PWHT to
suit each particular steel grade. In austeniticferritic dissimilar joints, such heat treatments
may lead to the occurrence of the following
phenomena:

Carbon enrichment in the weld metal due to

the diffusion of carbon from the low-alloy
steel into the austenitic weld metal
As a consequence thereof, carbon depletion
in the HAZ of the low-alloy steel
Coarse grain formation in the HAZ of the
low-alloy steel due to recrystallization processes
Embrittlement of the austenitic weld metal
due to precipitation of brittle phases, for
example, sigma phase
All these processes are time and temperature

dependent. Considering that the normal annealing time is in the range of 2 to 10 h, eventual
damage will normally occur, depending on the
material combination, at temperatures above
600 C (1110 F). Nickel-base weld metals,
due to their high nickel content outside the range
of sigma-phase precipitation, show no signs
of embrittlement during heat treatments. The
use of nickel-base filler metals is recommended
if austenitic-ferritic dissimilar joints are subject
to PWHT at temperatures above 600 C

(1110 F). Nickel alloy filler metals, abundant
and versatile, are frequently used for piping and
pressure vessel applications in refineries, chemical plants, and power plants (Ref 35).
The Risk-Based Approach

and Heat Treatments
Risk analysis is a powerful tool to rationalize
the decision-making process and is applicable to
heat treated steel components. Increasing incidents led to regulatory action. Some regulations
require that recognized and generally accepted
good engineering practice (RAGAGEP) must be
followed. This also created the desirable objective for industry to document what RAGAGEP
was.
The concepts of risk-based design, inspection,
and maintenance have been developed and are
being implemented (Fig. 42). Risk is the combination of the probability (or frequency of
occurrence) and consequence (or severity) of a
hazard. Its scope is limited to a specific environment during a certain period of time. The
intent of risk-based initiatives is to use finite
resources and allocate these resources in a
manner that achieves the greatest overall
reduction in risk. Flaws that exceed the limits
permitted by codes may be found. After a defect
that is not acceptable is found via a flaw evaluation/fitness-for-service analysis, a repair or
replacement is required for continued operation.
Various types of damage may occur
throughout the different heat treatment stages.
These types of damage may lead to failures
in the product during its useful life. The term
failure refers to the inability of part of the
component or the entire component to perform
the functions it was designed for, leading to an
unprofitable and technically useless product.
Failure modes analysis is a procedure in
which each potential failure mode is analyzed to
determine its effects and the criticality of these
effects on the system and to rank each potential
failure according to its severity. The criticality
analysis involves the use of risk analysis techniques based on the assessment of the likelihood
of failure and its potential consequences.
Among the most widely used tools in the
analysis of design is failure mode and effects
analysis (FMEA), which is accepted by the
Occupational Safety and Health Administration
pg 40
18/8/2008 2:51PM Plate # 0
pg 41
Risk Based Initiatives

Economic
Factor
Risk-Based
Design
Risk-Based
Inspection
Risk Assessment,
HAZOP, FMEA, WI, FTA,
RCA, RCM, RBI,
Fitness-For-Service,
Life Extension
Material Science
FEA, Welding, etc.
FitnessFor-Service
Failure
Analysis
Human
Factor
Mechanical Integrity
Initiatives
Fig. 42
Strategies to minimize risk throughout the cycle life based on risk and mechanical integrity initiatives
(OSHA). These methodologies are based on

reliability and safety.
By using the FMEA methodology in the
analysis of a new design, it is possible to identify
single-point failures (a single-point failure refers
to an individual failure that may cause the entire
system to collapse) and redesign the product
to avoid them, thus eliminating them completely
or achieving a more robust redesign that is less
sensitive to failures. The FMEA has evolved
from an ad hoc technique, dependent on a
designers experience, to a formal and accepted
analysis technique. Failure modes can be
eliminated by removing their causes or at least
having their probabilities of failure reduced to
acceptable levels.
There are several methodologies to assist in
logical thinking to resolve undesirable events.
These tools include, among others, reliabilitycentered maintenance, risk-based inspection,
FMEA, modified FMEA, and root-cause
analysis.
Working into an integrated system with riskbased efforts, fractomechanical-based structural
integrity approaches, and failure analysis will
allow the development of new tools to assist

the designers and manufacturers in minimizing
failures related to heat treating operations.
REFERENCES
1. L. Campos Franeschini Canale, G. Totten,

and D. Pye, Heat-Treating Process Design,
Handbook of Metallurgical Process Design,
G. Totten, K. Funatani, and L. Xie, Ed.,
Marcel Dekker Inc., New York, 2004
2. M. Solari, Risk Based Design, Chapter 2,
Handbook of Mechanical Alloy Design, G.
Totten, K. Funatani, and L. Xie, Ed., Marcel
Dekker Inc., New York, 2003
3. G. Totten, M. Narazaki, R.R. Blackwood,
and L.M. Jarvis, Failures Related to Heat
Treating Operations, Failure Analysis and
Prevention, Vol 11, ASM Handbook, ASM
International, 2002, p 192223
4. M. Jacob and G.A. Hawkins, Elements of
Heat Transfer, John Wiley & Sons, Inc.,
1957
18/8/2008 2:51PM Plate # 0
5. M.F. Ashby, Material Selection in Mechanical Design, Butterworth Heinemann,

1999
6. R.F. Kern and M.E. Suess, Steel Selection,
John Wiley & Sons, 1979, p 35
7. ASME Boiler and Pressure Vessel Code,
Section 2 D, ASME International, 1999
8. W.E. Dowling, Jr. and N. Palle, Design for
Heat Treatment, Materials Selection and
Design, Vol 20, ASM Handbook, ASM
International, 1997
9. F.A. Kandil, J.D. Lord, A.T. Fry, and P.V.
Grant, A Review of Residual Stress Measurement MethodsA Guide to Technique
Selection, Report MATC(A)04 Project
CPM4.5, Measurement of Residual Stress in
Components, Materials Centre, Middlesex,
U.K., Feb 2001
10. R.W.K. Honeycombe and H.K.D.H. Bhadeshia, Steels: Microstructure and Properties, 2nd ed., Arnold, 1995
11. R.F. Kern, Selecting Steels and Designing
Parts for Heat Treatment, American
Society for Metals, 1969
12. K.E. Thelning, Steel and Heat Treatment,
Butterworth, London, The Institute of Material, 1993, p 637
13. F.R. Hutchings and P.M. Unterweiser,
Failure Analysis, American Society for
Metals, 1981, p 3542
14. H.E. Boyer, Quenching and Control of
Distortion, ASM International, 1988, p 245
15. Failure Analysis and Prevention, Vol 11,
ASM Handbook, ASM International, 2002
16. G. Vander Voort, Failure Analysis and
17. Failure Analysis and Prevention, Vol 11,
Metals Handbook, 9th ed., American
Society for Metals, 1986, p 335
18. W.B.F. Mackay, Failure of Wrist Pins in an
Automotive Engine, Handbook of Case
Histories in Failure Analysis, Vol 1, ASM
International, 1992
19. E. Levy, Cracking of an Alloy Steel Bolt,
Handbook of Case Histories in Failure
Analysis, ASM International, 1992
20. G.M. Tanner, Hydrogen Embrittlement
Failure of Socket Head Cap Screws,
Handbook of Case Histories in Failure
Analysis, ASM International, 1992, p 332
21. Risk-Based Inspection, 1st ed., API Publication 581, American Petroleum Institute,
Washington, D.C., 2000
22. C. Duren, Equations for the Prediction of
Cold Cracking in Field-Welding Large
Diameter Pipes, IIW Document IX-135685, The International Institute of Welding,
Cambridge, England, 1985
23. R.W. Nichols, Ed., Developments in Pressure Vessel Technology 3, Applied Science
Publishers LTD, London, 1980
24. E. Folkhard, Welding Metallurgy of Stainless Steels, Springer-Verlag, New York,
1994
25. A.W. Marshall, Document IIW-IX-H-42298, The International Institute of Welding,
26. J.C.M. Farrar, Document IIW-IX-H-42398, The International Institute of Welding,
27. P. Bilmes, C. Llorente, and M. Solari, Role
of the Retained Austenite on the Mechanical Properties of 13Cr-4NiMo Weld
Metals, Proc. the 20th ASM Heat Treating
Society International Conference and
Exposition, Oct 2000 (St. Louis, MO), ASM
International
28. P. Brezina, Escher Wyss News, Vol 12,
1980, p 218
29. H. Niederau, Stahl Eisen, Vol 98 (No. 8),
1978, p 385
30. F.B. Pickering, Physical Metallurgy and the
Design of Steels, Applied Science Publishers, London, 1978, p 165
31. T.G. Gooch, Weld. J., July 1995, p 213s
32. R.D. Kane, Corrosion, Vol 33 (No. 7),
1977, p 231
33. P. Bilmes, C. Llorente, and M. Solari, Effect
of Post Weld Heat Treatments on the
Microstructure and Mechanical Behaviour
of 13Cr-4NiMoL and 13Cr-6NiMoL Weld
Metals, Proc. the 18th ASM Heat Treating
Society Conference and Exposition, Oct
1998 (Chicago, IL), ASM International
34. P. Bilmes, C. Llorente, and M. Solari,
Characteristics and Effects of Retained
Austenite Resulting from Tempering of
13Cr-NiMo Weld Metals, Mater. Charact.,
Vol 46 (No. 4), April 2001, p 285296
35. H. Ebert, Nickel Alloy Filler Metal Review,
Weld. J., July 2004, p 6062
pg 42
18/8/2008 2:54PM Plate # 0

DOI: 10.1361/faht2008p043

Overview of the Mechanisms of Failure

in Heat Treated Steel Components
Primum non nocere First do no harm, attributed to the ancient Roman
physician Galen. Declare the past, diagnose the present, foretell the future;
practice these acts . . . make a habit of two things to help, or at least to do no
harm (Ref 1).
FAILURES IN STEEL components, like any

other material, may have various consequences,
such as:

Making the device or component completely

inoperable
Preventing an operable device from functioning satisfactorily
Making the device or component unsafe or
unreliable, with immediate removal from
service required
Many aspects may also be involved in tracing

back to the possible sources of failure of a
component. Some of these sources include:

Design
Material issues, such as improper materials
selection or material imperfections (laps,
seams, inclusions, porosity, etc.)
Fabrication and processing
Rework
Assembly
Inspection
Storage and shipment
Service conditions
Maintenance
Unanticipated service conditions
Many times, more than one factor contributes to

a part failure. Rarely is it only one factor.
General Sources of Failure

Design deficiencies are a common source of
component failure. Examples include the presence of a sharp notch in regions of high stress or
a fillet radii that is too sharp. Using a component
design for a new application can also lead to
unanticipated failures. Higher stresses or unanticipated service conditions can cause unforeseen failure because of complex or increased
stress fields. Stress concentrations may become
more critical because of the increase in loading
for the new application.
Insufficient design criteria can also be the
cause of unforeseen failures. Inadequate
knowledge of the stress state in the component
or inadequate stress calculations can contribute
to failures. Much higher stress states than initially assumed or improper stress assumptions can
result in premature service failures. Lack of
consideration of severe environmental, fatigue,
or impact conditions may contribute to failure.
Material issues can usually be attributed to
either selection of material or material imperfections rendering it unsuitable for service. Inadequate material data can also result in
conditions that may contribute to failure. For
example, adequate fatigue data, elevated-temperature tensile data, or creep or corrosion data
may not be available, and the designer may have
to extrapolate or estimate the effects or these
properties.
Other sources of failure can be attributed to
material imperfections. For wrought products,
this could be related to segregation, inclusions,
porosity, laps, and seams. For castings, these
imperfections could be cold shuts, inclusions,
shrinkage, voids, and porosity. Forgings can
have laps, seams, segregation, and anisotropy in
properties from forging flow lines.
In one example (Fig. 1), a large roll was heat
treated, and several large cracks were observed
after inspection. This was originally attributed to
quench cracking. On further examination, it was
determined that a lap was present in the forging,
pg 43
18/8/2008 2:54PM Plate # 0
indicated by the presence of high-temperature

oxides in the crack along the crack faces.
Manufacture and Processing. Processing
can have a large influence on properties and the
resulting residual stresses. Typically, this is
related to wrong procedures or improperly specified procedures. Ambiguous processes or
specifications can also contribute to failures due
to interpretation or application. Simple things
like improper selection of processing sequences
or procedures or specifications that were not
followed can also contribute to failure.
Cold forming, such as stretching or deep
drawing, can develop highly localized residual
stresses. Local changes in microstructure can
occur. Because of the changes in reduction, a
large anisotropy in material properties also results. Due to the drawing operation, cracks or
microcracking can occur. This could be due to
improper lubrication or improper die design.
The localized changes in ductility can also contribute to failure.
Machining and grinding can create high
residual stresses from either machining practice
(feeds and speeds) or improper cutting tool
selection, material, or geometry. Grinding, if
Fig. 1
abusive, can cause large temperature gradients

and localized overheating. This overheating can
cause changes in microstructureeither localized softening of the material or localized
transformation to martensite and other transformation productsresulting in hard spots.
In Fig. 2, a large gear was ground after heat
treatment. Because of abusive grinding, local
temperatures exceeded the austenitization temperature, and transformation to martensite
occurred upon cooling. This transformation and
the resulting residual stresses caused cracking of
the gear. Temper etch examination of the gear
using dilute nitric acid in water in the regions of
cracking showed evidence of localized abusive
grinding.
Identification of parts can also cause failure to
initiate. This is from localized impact or electroetching. Localized mechanical stress concentrations or changes in microstructure can occur.
This creates either a mechanical or microstructural notch or stress concentration.
Heat treatment can cause a variety of different
root causes for failures. Overheating, decarburization, quenching, tempering, annealing, and
other heat treatments can cause failure to occur.
A large roll was found to have cracks on the outer and inner surfaces of the forging. These cracks were found during final
inspection. During examination of metallographic sections taken from the roll, high-temperature oxides were found on the
crack faces, strongly suggesting forging laps.
pg 44
18/8/2008 2:54PM Plate # 0
Overview of the Mechanisms of Failure in Heat Treated Steel Components / 45
This could also include improper austenitization

temperatures and times. Decarburization is the
result of a low-carbon surface from improper
atmosphere control. Typically, there is a depleted carbon layer at the surface that, when
quenched, is softer than the core material. This
soft layer can be completely devoid of carbon
(complete decarburization) or only partially
depleted in carbon (partial decarburization).
This decarburized layer can contribute to premature fatigue failures, because the surface
material is different than the designer expected,
or failure can result from high residual stresses
created at the surface from the quenching
operation. The low-carbon surface area can also
result in distortionagain, high residual tensile
stresses at the surface with low surface hardness.
Carburization is similar to the effects of decarburization. In this case, there is a higher surface carbon than expected. High residual tensile
stresses can result as well as increased distortion.
Quenching can also contribute to high residual stresses or the formation of cracks or
microcracking. Transformation stresses from
quenching cause the high residual stresses.
These high residual tensile stresses can drastically reduce the fatigue strength or have other
ramifications in service.
Overheating can cause excessive grain
growth, with resulting increases in hardenability
and increased embrittlement. Underheating can
cause poor mechanical properties, because there
was an incomplete transformation to austenite
and therefore an incomplete transformation to
martensite. Poor mechanical properties, such as
low tensile and yield stress, and poor impact
properties may occur.
Fig. 2
There are also several embrittlement mechanisms caused by the use of improper tempering
temperatures. Temper embrittlement and blue
brittleness are just two of the common mechanisms that can occur from improper heat
treatment and tempering operations.
Cleaning, pickling, and electroplating operations can also cause potential failures or
contribute to them. Hydrogen charging of highstrength steels from the dissociation of hydrogen on the surface of high-strength steel can
occur from cleaning operations in acids. Charging of hydrogen from high current densities
in electroplating can cause hydrogen embrittlement unless proper baking procedures are used
to allow the hydrogen to diffuse out. Electroplating can also cause high residual tensile
stresses, which can contribute to crack initiation.
Welding can cause many different problems.
These problems can be cracks that are initiated
from improper welding procedures, high residual stresses, porosity from inadequately dried
weld rods, or dirty workpieces. Microstructural
notches or stress concentrations from the heataffected zone and the transition to the base
material can be the result of improper preheat
and postheat. Improper weld penetration, weld
geometry, and excessive weld current (undercutting) can also cause mechanical stress concentrations (Fig. 3).
The mast arm failure shown in Fig. 3 (Ref 2)
was the result of weld bead undercutting and
poor weldment design. Fatigue cracking initiated at the site of the weld toe undercut. This
location was a highly stressed area and the location of a large mechanical stress-concentration
factor because of the weld toe undercut. Typical
Large gear that cracked during grinding operations. Localized thermal gradients during grinding resulted in high residual
stresses and eventual cracking. Temper etching (dilute nitric acid in water) revealed the presence of abusive grinding.
pg 45
18/8/2008 2:54PM Plate # 0
Fig. 3
Failure of a mast arm due to fatigue that initiated at a weld toe undercut. Source: Ref 2
causes of undercutting include excessive weld

current.
The assembly of a group of components can
also cause eventual failure. Force-fitting a component creates high residual stresses or damage
and causes premature failure to occur. Incorrect
placement of a component or incorrect assembly
order can also cause high residual stresses or
failure to occur. Improper specifications or torque requirements can also cause premature
failure. Misalignment of components within the
assembly could also result in inadequate service
life, because the stresses are not what the
designer had anticipated.
Service conditions obviously can have a
large role in the failure of a component. The
service conditions could be normal operations
but unanticipated by the designer. It could also
be abnormal operations, such as speed, temperature (high or low), or a chemical environment, that were also unanticipated. The lack of
proper scheduled maintenance can be a major
contributor to premature failure. Maintenance
procedures are often reduced as a cost-savings
measure. Inadequate lubrication or improper
lubrication can also play a role in failure (Fig. 4).
In the case of Fig. 4 (Ref 3), the lubrication
schedule was extended to reduce aircraft
Fig. 4
The probable cause of this accident was a loss of airplane pitch control resulting from the in-flight failure of
the acme nut threads on the horizontal stabilizer trim system
jackscrew assembly. The thread failure was caused by excessive
wear resulting from insufficient lubrication of the jackscrew
assembly. Source: Ref 3
pg 46
18/8/2008 2:54PM Plate # 0
downtime. This, and other contributing factors,

resulted in the loss of 88 lives.
Stresses from startup can also contribute,
along with rapid temperature gradients and rapid
localized changes in the environment. Start-up
procedures and maintenance are critical for
intermittent operations. Shut-down procedures
and resulting stresses are just as critical as proper
startup. Inspection procedures to prevent failure
are also important. Failure to properly inspect
for problems or cracking can be catastrophic
(Fig. 5), (Ref 4). In this case, maintenance and
inspection personnel failed to detect a fatigue
crack in the compressor stage of an aircraft
engine. Upon application of power, the compressor stage ruptured, with shrapnel severing
fuel lines and igniting the fuel, ultimately leading to the loss of the aircraft.
General Practice Conducting

a Failure Analysis
The primary objective of any failure analysis
is to determine the primary root cause of failure
and to establish the appropriate corrective
action. There are several stages of an analysis,
which can proceed one after the other or occur at
the same time. There is no set fixed-in-stone
procedure, because it is highly dependent on the
part and procedures/capabilities of the specific
laboratory.
These stages of analysis are:

Collection of background data

Preliminary visual examination

Nondestructive testing
Selection and preservation of specimens
Mechanical testing
Macroexamination
Microexamination
Metallographic examination
Determination of the fracture mechanism
Chemical analysis (bulk and microanalysis)
Exemplar testing
Analysis and writing the report
These stages are described as follows, and additional information on failure analysis procedures is given in the chapter General Aspects of
Failure Analysis in this book.
Collection of Background Information
During the collection of background data, the
engineer is trying to gather an understanding of
the purpose of the part. The engineer is attempting to discern the design criteria, service
conditions, and failure conditions. In the background information, the operating details and
manufacturing history should be examined and
collected. This manufacturing history should
include all the mechanical processing, thermal
history or processing, and any chemical process
performed on the part.
The service history should include all the
maintenance records of the part. It should also
include the expected environment and loading at
the time of failure, as well as the normal environment and loading. Any quality records should be
examined for discrepancies. Unfortunately,
these records are not always available, and it is
often up to the experience of the engineer to
determine the quality of the part.
Preliminary Visual Examination
Fig. 5
The probable cause of this accident was the failure of

maintenance and inspection personnel to perform a
proper inspection of a seventh-stage high compressor disk, thus
allowing the detectable crack to grow to a length at which the disk
ruptured under normal operating conditions, propelling engine
fragments into the fuselage. The fragments severed the right
engine main fuel line, which resulted in a fire that rapidly
engulfed the cabin area. Source: Ref 4
Documenting the failure or fracture is extremely important. There can never be too many
drawings or photographs. The cost of photographs (especially digital) is cheap compared to
analysis. A high-quality camera with macrocapability is very important and is one of the
best tools that a failure analysis laboratory can
have. The use of gray cards to ensure proper
color rendition is also very important, because
the color of scale or oxides can often give an
indication of the temperatures that the part has
experienced.
Sample selection is also very important. All
associated debris should be collected and identified. Similar parts should also be collected for
pg 47
18/8/2008 2:54PM Plate # 0
comparison. In the case of a fastener failure, it is

important that the nut and washer be collected,
too. All mating pieces should be gathered for
subsequent analysis.
Any abnormal conditions should be observed
and compared with new and used components.
Any discoloration or debris should be noted and
collected. Any distortion of the part should be
noted, along with dimensions of the part.
Weather conditions at the time of failure should
be collected, as well as all bearing and lubrication conditions and records.
During the initial wreckage analysis, the
determination of all wreckage should be identified and located on a map or grid before any is
touched or moved. Photograph each piece of
wreckage and its surroundings. Inventory the
parts present or missing. Determine the operating conditions at time of failure. This should
include the position of control surfaces, power
settings, position of throttles, and any lights or
annunciations that occurred.
As best as possible during the initial examination of the wreckage, the sequence of failure
should be determined. This can be accomplished
by examining chevron markings and crack
order. The parts should then be closely examined
and reassembled. DO NOT allow the fracture
surfaces to touch each other, because this can
cause potential damage to the delicate surfaces.
This analysis can also help determine the
sequence of events leading up to failure. Preliminary examination of the part should note any
paint, debris, or deposits present. Always
remember to do no harm.
The visual examination should be detailed.
Fracture surface crack directions should be
noted, identified, and documented. Any abuse or
discoloration should be identified, and a general
assessment of the workmanship should be
determined. Document all findings with photography, with multiple photographs taken from
different directions. The incorporation of rulers
or scales is important to determine the size and
direction of fracture.
Nondestructive Testing
Nondestructive testing is very useful for
determining the extent of cracking. Magnetic
particle inspection is useful for ferrous alloys,
with dye-penetrant and ultrasonic inspection
as additional methods available for initial
inspection.
Magnetic particle inspection uses discontinuities in the magnetic field to identify cracks
or discontinuities. Fluorescent dyes with small
magnetic particles are used. These magnetic
particles gather at the discontinuities in the
magnetic field, indicating flaws or indications. It
is a common, sensitive, and reliable method that
is simple to learn and use. This method has no
limitation in part size but is limited to magnetic
materials. No elaborate precleaning of the surfaces is necessary. Detection is limited to the
surface of the part or section examined. Care
must be exercised to prevent local arcing.
The dye-penetrant method is useful for
examining surface flaws or cracks. It is used
primarily for nonferrous alloys but is used for
examining ferrous weldments for cracks and
porosity. In this method, a high-wetting liquid is
spread on the surface of the part. Excess liquid
is wiped off. A developer is applied to the part
surface. Any cracks, flaws, or other indications
will appear. Limitations of this method are the
necessity of cleaning the surface prior to and
after application of the indication fluid and
developer solution. Surface features may also
mask indications. It is simple to use, but an
understanding of the limitations must be
understood prior to application to a part.
Eddy-current methods depend on the principle that all metals conduct electricity. An alternating current is applied, and eddy currents
occur by electromagnetic induction. Cracks or
other flaws cause distortions in the electromagnetic fields, with a result of changing the
field impedance. The advantage of this method
is that subsurface discontinuities can be detected. No special skill is required to use this
method, and the method can be automated.
Probe contact with the part is not needed. Limitations of this method are that the depth penetration is limited, and the part must be capable of
conducting electricity. Reference standards are
needed for specific flaw sizes and materials.
Many things can influence readings, including
segregation, carburized layers, and changes in
profile.
Ultrasonic testing uses high-frequency sound
waves transmitted through a conducting medium. Any discontinuous boundary can cause a
deflection. This method is very sensitive and has
high penetration. It is possible to get accurate
measurements of flaw position and size, but
reference standards must be used. Shape and size
can cause errors in interpretation. Experienced
operators are required to properly interpret the
pg 48
18/8/2008 2:54PM Plate # 0
results of testing. Effects of grain size, porosity,

and inclusions can also hinder interpretation.
Radiography, using x-rays, neutrons, or
gamma rays, is also often used to examine
structures. Film or sensors (charge-coupled devices) pick up the emitting radiation, with the
intensity proportional to the density of the
sample. Light areas indicate a dense region, and
dark areas indicate a greater exposure or less
dense region. Advantages of radiography are the
detection of subsurface and internal features at
various depths and the documentation of these
features by film or other imaging techniques. The
primary disadvantage is that reference standards
must be used, and the area for testing must be
enclosed to prevent radiation from leaking out.
Mechanical Testing
Mechanical testing is useful to determine the
properties of the part and to verify that it meets
expected properties and specifications. There are
many types of mechanical testing available,
including hardness, tensile testing, and impact
fracture testing.
Hardness testing is probably the most versatile and widely used. It is often used to evaluate
heat treatment and can be used as an approximation for tensile strength. It can be used to
detect the presence of work hardening or softening and hardening or softening from localized
thermal events such as grinding. For the most
part, it is a nondestructive test. For microhardness testing, it is necessary to use a metallographic specimen.
Tensile testing is used more to establish conformance to specification. It is not necessary to
show inadequate ductility because of service
loads. Because of the size of the tensile specimen, it may not be possible to excise an appropriately sized sample from the part. Anisotropy
of properties can be expected to lower measured
tensile and yield strength properties.
Impact and fracture toughness testing is
typically used to determine conformance to
specifications. Charpy impact testing has a high
variability in results and may be temperature
related. Results must be taken with temperature
in mind and may not correlate with real results
because of size limitations. Fracture toughness
testing and the results from KIc testing can be
used in design, and the results are useful for
calculating critical flaw sizes. It can also be
used to examine estimated crack growth rates;
however, samples are difficult to prepare and
test. These methods also do not incorporate the

effects of residual stresses.
Selection and Preservation of Specimens
The selection and preservation of fracture
surfaces is vital to prevent the destruction of
evidence. Unprotected, the fracture surfaces or
parts can become mechanically or chemically
damaged. This damage can obliterate evidence
and make the determination of fracture difficult
or impossible. Both sides of the fracture must be
protected. This is in the event that if one surface
is damaged, the other side can be examined.
Protection of the specimens during shipment is
also very important, because evidence could be
destroyed. Avoid touching surfaces with the
hands, because the chemicals and acids present
can cause artifacts or destroy data. NEVER fit
surfaces together, because the delicate fracture
features can be destroyed. Since both surfaces
would be damaged, it could destroy the chances
for determining the fracture mechanism.
Cleaning of specimens is to be done only
when absolutely necessary. For the most part,
it is required to prepare the sample for the
scanning electron microscope (SEM). Dry air
blasts or soft artist brushes are typically all that
is needed. Rinsing in organic solvents then
evaporating the solvent with dry air is useful for
preparing specimens for the SEM.
Chemical cleaning is generally not recommended under any circumstance. Foreign substances such as scale or debris should be
preserved. Do not use rust inhibitors, because of
the inevitable damage to the part and fracture
surfaces. These rust inhibitors are also extremely
difficult to remove. Avoid washing the sample or
parts with water unless seawater or other chemical is present. In this case, gently wash with
distilled water and follow that with high-quality
alcohol or acetone. Allow to dry and place in a
dessicator.
Plastic replicas are useful in preserving
fracture surfaces and removing debris for further analysis. Softening replica tape (available
at transmission electron microscope supply
houses) with small amounts of acetone forms
plastic replicas. The softened tape is pressed
gently onto the fracture surface. Additional
layers of tape, softened with acetone, are applied
to the fracture surface. After multiple layers
have been applied, the entire replica is allowed
to dry and then is placed in a dessicator. When
the part is ready to be examined, the replica is
pg 49
18/8/2008 2:54PM Plate # 0
carefully removed using tweezers. Any debris

on the surface is also preserved for further
analysis in the replica. Multiple plastic replicas
can be used to clean a surface of a part. This can
be repeated as necessary.
Sectioning of Specimens
Sectioning is very important, because it captures the portion of the fracture surface for
examination or the appropriate metallographic
specimen. The biggest limitation is size. It is
important that the portion to be removed is documented by photographs and sketches, showing
the location of the specimen to be removed.
Preserve any fracture surface by plastic replicas
or other method to prevent damage or attack.
Regions adjacent to cracks are also to be preserved and protected. Cutting the specimens
should be done very carefully so as not to cause
any heat damage. Coolants are not recommended, unless the material cannot be cut without
heat generation. The use of plastic replicas is
useful for protecting surfaces and preserving any
debris present.
Opening secondary cracks is useful when the
primary fracture surface is damaged. These secondary cracks may provide better information,
because they are tightly closed, and the fracture
surfaces are not exposed to surface contaminants
and corrosion. Care must be taken not to damage
the primary fracture surface. Bending to open
the crack is preferable, to expose the crack face.
Often, the use of a sawcut to the back of the part
will reduce the amount of force necessary to
open the secondary crack. Another method is
to use a tensile machine to open the crack face.
The crack opening should be measured prior to
opening, and the crack opening displacement
can also be measured as the crack is slowly
opened and exposed. One technique is to immerse the specimen in liquid nitrogen and
impact the part so that the fracture surfaces are
rapidly opened. One problem with this method is
that it is very easy to damage the fracture surface
from a misapplied hammer hit.
information on the location of fracture origins,

direction of cracking, configuration of the stress
state, and the last region to fail (shear lip). The
presence of chevron marks can indicate the
direction of rapid crack growth, and the different
textures of the fracture can differentiate between
fast final fracture and the initiating mechanism
of fracture. Different textures from the region of
fast fracture can indicate a different mechanism,
such as fatigue, stress-corrosion cracking, or
hydrogen embrittlement.
Microscopic Examination
The microscopic examination is usually conducted with an SEM (Fig. 6). This instrument
is probably the most useful of all instruments
for determining the mechanism of failure. It is
capable of a large depth of field, with magnifications of 10 to 300,000 . It allows for direct
examination of specimens, and when coupled
with an energy-dispersive spectrometer, very
small regions can be examined and analyzed for
chemistry. It is very easy to use and requires very
little training to take quality images.
Interpretation of the images requires experience and understanding of the four basic modes
of failure: dimpled rupture, cleavage, brittle
intergranular, and fatigue. From these four basic
modes, the detailed mode can be examined, and
the failure mechanism is fit to the evidence.
A greater discussion of the mechanisms of failure is found later in this chapter and elsewhere in
this book.
Metallography
Metallography is a vital part of a failure
analysis investigation. It can examine crack
Macroscopic Examination
The macroscopic examination is conducted
by a detailed examination at 1 to 100 by eye or
binocular microscope. High-quality optics with
excellent depth of field are required to properly
examine the fracture surfaces. This detailed
macroscopic examination can reveal a wealth of
Fig. 6
Typical scanning electron microscope used for

microscopic analysis of a fracture surface
pg 50
18/8/2008 2:54PM Plate # 0
morphology and its relationship with the

microstructure present. It can help determine the
thermal history of a component or region of a
part and can show if work hardening was present. There can never be too many photographs
and metallographic sections. Metallographic
sections should be taken away from the crack
and near the determined origins of cracking.
Because this method is destructive, it is undertaken last. Typically, the crack face and edges
are protected from rounding by applying support. This support can be electroless nickel plate
or the use of alumina beads or steel shot in the
metallographic specimen, adjacent to the surface.
Metallographic specimens are prepared using
an epoxy or phenolic resin. The sample is placed
into a small press, and phenolic resin is poured
over the section. The press compacts the resin
and forms a small, round sample that is then
polished, etched, and examined under a metallographic microscope. When the specimen has
cooled, it is taken out of the press and ground
through a sequence of sandpapers. Typically, the
sequence is 240, 320, 400, and 600 grit. The
specimen is ground very flat before polishing.
During polishing, the metallographic specimen
is polished using a flat platen and 3 mm alumina
slurries. Final polish is accomplished using
0.15 mm alumina slurry. Other polishing agents
can be used, with diamond being a very common
polishing agent. A finished metallographic
sample used for the determination of the fracture
mechanism in a steel weldment is shown in
Fig. 7.
Examination of the metallographic specimen
reveals surface imperfections, inclusions, and
microstructural details. It can reveal the presence of decarburization and improper heat
treatment. It often provides the needed documentation and support for the fracture analysis
and determination of the root cause of failure.
Determination of the Fracture

Mechanism
Examination of the fracture surface and
metallography are used to determine the cause of
failure. First, it is necessary to determine the
fracture mode. Unfortunately, there is no clear
or logical classification of fracture. Generally,
classification is based on the crack growth mechanism (see also the chapter General Aspects of
Failure Analysis in this book).
Ductile Fracture
On a macroscopic scale, a ductile fracture is
accompanied by a relatively large amount of
plastic deformation before the part fails. After
failure, the cross section is reduced or distorted.
Shear lips are observed at the latter part of
the fracture and indicate the final failure of the
part. The fracture surface is dull, with a fibrous
appearance. Microscopically, ductile fracture
is characterized by several distinct stages
(Ref 58); an example is shown in Fig. 8. In this
case, an ISO 12.9 low-alloy bolt failed by ductile
torsional overload. The fracture was smooth,
with fracture initiating from the threads. The
fracture mode was microvoid coalescence
(Ref 9), which occurs by the following process:

Fig. 7
Typical metallographic specimen. This specimen was

used to examine microstructures in a failed weldment.
A free surface is created from a small particle. This particle can be a second-phase
particle, dispersoid, or inclusion. The separation of the metal matrix from the small
particle at the matrix/particle interface can
form this free surface, or the fracturing
of the small particle can form the free surface.
The free surface around the small particle
creates a void. This void grows by plastic
strain and hydrostatic stress.
Finally, the voids grow to a size that they join
or coalescence with adjacent voids.
This process of void formation, growth, and

coalescence is shown schematically in Fig. 9. If
the particles are well matched to the matrix and
form a strong interface between the matrix and
the particle, then the initial formation of voids is
the critical step. Fracture occurs shortly after
pg 51
18/8/2008 2:54PM Plate # 0
Fig. 8
Fracture of an ISO 12.9 bolt by ductile torsional overload. (a) Overall view of fracture. (b) Smooth and fibrous fracture as seen
through the SEM. (c) Microvoid coalescence (dimples)
Fig. 9
Schematic showing the formation of microvoid coalescence
void formation (Ref 10). If the interface between

the particles and the matrix is weak, then voids
form and grow readily. Substantial plastic
deformation occurs. Fracture occurs when the
voids reach a critical size. These voids substantially reduce the cross section, with the
resulting local plastic instability (Ref 11). These
voids coalesce to form a central crack perpendicular to the applied tensile stress. Depending
on the applied stresses, the shape and configuration of the dimple shape can be changed
(Fig. 10). This fact is important in determining
the type of loading during a postfracture investigation. Dimples are small and can only be
detected by using electron microscopy (Fig. 11).
The presence of inclusions in steel plays a
major role in the ductility of steel. As indicated
previously, the inclusions fracture and separate
from the matrix during decohesion. Therefore,
the deformability of these inclusions is important to determine the ductility of steel.
Nearly all steels have nonmetallic inclusions.
The size and frequency of these inclusions is
determined by the methods described in ASTM
E45 (Ref 12). The cleanliness of the steel is
Fig. 10
Schematic representation of the creation of dimples

in a loaded member by (a) simple tension, (b) shear
loading, and (c) tearing
Fig. 11
Microvoid coalescence as seen through the SEM
pg 52
18/8/2008 2:54PM Plate # 0
important to the ductility of the steel. All other

things being equal, the steel with the lower
inclusion size, shape, and frequency will have a
greater ductility than another steel with a greater
inclusion count. Modern steelmaking practices
generally produce low inclusion content. Often,
steels for aerospace applications require a frequency/severity determination of inclusions in
accordance with AMS 2300, AMS 2301, AMS
2303, or AMS 2304 (Ref 1316). A specificsized test specimen must be heat treated and
examined using magnetic particle inspection.
The procedures are outlined in the aforementioned specifications.
The inclusions found in steels have been
divided into five categories related to their
deformation behavior (Ref 17):

The inclusions Al2O3 and calcium aluminates are produced during deoxidation of
steel during the production of molten steel.
They are brittle at practically all temperatures.
Spinel-type oxides are not deformable up to
1200 C but may be deformed above this
temperature.
Silicates of calcium, manganese, iron, and
aluminum in various proportions are brittle
inclusions at room temperature but become
more deformable at higher temperatures.
The formability increases as the melting
temperature of the silicate decreases. Therefore, aluminum silicate has much less formability than the lower-melting manganese
silicates.
FeO and (FeMn)O are deformable at room
temperature but gradually become more
brittle at temperatures above 400 C.
Manganese sulfide (MnS) is the most common inclusion found in steel, and it is
increasingly deformable as the temperature
falls. The morphology of the MnS inclusions
changes, depending on how they were
formed.
Ductile failure can occur with any of the

types of inclusions. This is true whether it is
the brittle alumina-type inclusions or the more
ductile sulfide-type inclusions. Inclusions generally initiate ductile cracking above a critical
size. Coarser inclusion sizes tend to have a larger
local stress-concentration factor, which can
cause local decohesion and microcrack formation. Work by Maropoulos and Ridley (Ref 18)
has shown the effect of volume fraction of ironalumina on the ductility of steel. Increasing
amounts of inclusions reduce the ductility of the

steel. A reduction in the yield stress, due to the
stress concentrations around the inclusions, is
evident at low volume concentrations of inclusions.
The presence of inclusions in the size range of
1 to 30 mm reduces the energy absorbed during
ductile fracture. Fine dispersions of ductile
inclusions will delay the onset of cleavage-type
fracture by localized relaxation of stresses. At
the same time, the yield stress also increases.
During deformation, forming, or forging, the
ductile inclusion MnS has a marked effect on the
ductility of the final product. Types 1 and 2 MnS
inclusions will elongate on deformation, while
type 3 MnS inclusions will rotate into the rolling
plane. This will reduce toughness and ductility
in the transverse direction. Type 2 inclusions are
the most harmful to ductility and toughness, so
some effort is being made to eliminate these
inclusions by ladle additions of other strong
sulfide formers, such as titanium, zirconium, and
calcium.
Ductility is also influenced by the fact that
MnS contracts more than the iron matrix upon
cooling. The bond between the MnS inclusion
and the matrix is not strong enough to prevent
microvoid formation. Because MnS inclusions
tend to form as strings or stringers along the
rolling direction, the toughness and ductility
are strongly influenced in the rolling direction.
Transverse to the rolling direction, ductility and
toughness are much worse.
In a similar fashion to that of inclusions, the
distribution of carbides can also influence the
toughness and ductility of the steel. The strain
needed for void formation decreases with
increasing carbide volume fraction. Spheroidal
carbides will not crack at small strains and
exhibit decohesion. Spheroidized steel is much
more ductile than similar steel of the same
hardness containing only ferrite and pearlite.
Pearlite has a lower critical strain for void formation. In addition, when a crack or void forms
in a pearlitic matrix, it will tend to run along
the length of a pearlite lamella. Examining this
type of fracture under the SEM reveals that the
base of the dimples contain fractured pearlite
lamella.
Brittle Fracture
Very little plastic deformation and a shiny
fracture surface characterize brittle fractures.
Often, chevron patterns point back to the origin
pg 53
18/8/2008 2:54PM Plate # 0
of failure (Fig. 12) (Ref 19). It can occur at low

stress and propagate with rapidity, often at
speeds approaching the speed of sound in the
failed material.
Since the early 1940s, there has been tremendous growth in the number of large welded
structures. Many of these structures have failed
catastrophically in service, most notably the
Liberty ships (Ref 20) used to transport war
material during World War II. Analysis of the
fracture surfaces of the failures (Ref 21) indicated that they initiated at a notch and propagated with no plastic deformation. These
notches were of three types:
Design features: Structural members were
rigidly joined at angles less than 90 and
then welded.
Fabrication details: Procedures used during
the manufacture of the part caused the
formation of notches. Welding arc strikes,
gouges, and fitting procedures created
physical notches. Welding procedures and
In brittle fractures, limited energy is absorbed

by the fracture. Energy is absorbed through
regions of small plastic deformation. Individual
grains separate by cleavage along specific
crystallographic planes. This is shown in Fig. 13.
Visually, little or no plastic deformation or
distortion of the shape of the part characterizes
brittle fractures. The fracture is usually flat and
perpendicular to the stress axis. The fracture
surface is shiny, with a grainy appearance.
Failure occurs rapidly, often with a loud report.
Because the brittle cleavage is crystallographic
in nature, the fracture appearance is faceted.
Often, other features are present, such as river
patterns (Ref 23). These are shown schematically in Fig. 14.
There are three basic factors that contribute to
a cleavage type of fracture in steels. They are:

Fig. 12
Fig. 13
Chevron markings point back to the origin of failure

in brittle steels. Source: Ref 19
heat treatment caused metallurgical or

microstructural notches to occur from abrupt
changes in microstructure or the production
of microstructures that were brittle. Features
such as porosity from welding or casting also
caused brittle fracture initiation.
Material flaws: These flaws resulted from
melt practice at the mill and appeared as
large inclusions, internal oxidation, porosity,
or segregation.
Triaxial stress state that forms at a notch,

similar to that described previously
Low temperature
High strain rate or rapid loading rate
These three factors do not have to be present

for cleavage-type fracture to occur. Most brittle,
cleavage-type fractures occur when there is a
triaxial stress state and low temperature. This is
Cleavage fracture in a low-carbon steel, seen through an SEM. Cleavage fracture in a notched impact specimen of hot-rolled
1040 steel broken at 196 C ( 320 F), shown at three magnifications. The specimen was tilted at an angle of 40 to the
electron beam. The cleavage planes followed by the crack show various alignments, as influenced by the orientations of the individual
grains. Grain A, at center in fractograph (a), shows two sets of tongues (see arrowheads in fractograph b) as a result of local cleavage along
the {112} planes of microtwins created by plastic deformation at the tip of the main crack on {100} planes. Grain B and many other facets
show the cleavage steps of river patterns. The junctions of the steps point in the direction of crack propagation from grain A through grain
B, at approximately 22 to the horizontal plane. The details of these forks are clear in fractograph (c). Source: Ref 22
pg 54
18/8/2008 2:54PM Plate # 0
actuated by a high rate of loading. Many types of

tests have been developed to determine the
susceptibility of steels to brittle behavior. These
tests include the Charpy impact test (ASTM
E23) (Ref 24) and the fracture toughness test
(ASTM E399) (Ref 25). Others include the nilductility test (ASTM E208) (Ref 26) and
dynamic tear test (ASTM E604) (Ref 27).
The notch toughness of low- and mediumstrength steels is highly dependent on temperature. There is a transition from ductile fracture to
brittle fracture as the temperature decreases.
One criterion for the transition temperature is the
nil-ductility temperature (NDT). The NDT is the
temperature where fracture becomes 100%
cleavage, and there is essentially no plastic
deformation.
Fig. 14
Changes in the NDT can be produced by

changes in microstructure and chemistry. The
largest change can be effected by changes in the
amount of carbon and manganese. The NDT is
lowered by approximately 6 C (10 F) for
every 0.1% increase in the manganese concentration. Increasing the carbon content also
lowers the NDT. The manganese-carbon ratio
should be approximately 3 to 1 for good notch
toughness.
Decreasing the concentration of phosphorus
also decreases the NDT. Nitrogen causes the
NDT to increase (more brittle). However,
because of the interaction with other alloying
elements in steel, it is difficult to quantify the
increase of NDT with increasing nitrogen concentration.
Schematic of river patterns formed in brittle materials. (a) Tilt boundary. (b) Twist boundary. Source: Ref 23
pg 55
18/8/2008 2:55PM Plate # 0
Nickel is beneficial for increasing ductility.

Up to 2% Ni is effective in lowering the NDT.
Increasing concentrations of silicon have the
effect of increasing the NDT. Chromium has
nearly no effect, while molybdenum is extremely effective in increasing the ductility of
steels and drastically decreasing the NDT.
Oxygen strongly decreases the ductility. It can
also cause an increased propensity for intergranular fracture by creating brittle oxides at the
grain boundaries. Decreasing the grain size has a
strong effect on increasing the ductility and
notch toughness.
Section thickness can also influence ductile
and brittle behavior (Ref 28). The results showed
that there was considerable variation of toughness with the thickness of the specimen (Ref 29,
30). Further, at large thickness, the toughness
appeared to reach a constant value (Fig. 15)
(Ref 31). Within this curve, there are three
apparent regions. First, there is the region where
maximum toughness is obtained (thin sections).
Second, there is the region of intermediate
toughness, and lastly, a region with relatively
constant toughness (thick sections).
In the first region, the fracture appears to
consist entirely of a shear lip, or, in other words,
the fracture surface is inclined at an angle
of approximately 45 to the tensile axis. In this
situation, the stress in the direction of the
thickness of the specimen tends toward zero,
and a state of plane stress is achieved. As the
specimen is pulled, it experiences buckling. Because of this buckling, yielding occurs on the
through-thickness planes at an angle of 45 to
the tensile axis. Crack extension occurs by
sliding. This sliding motion is achieved by the
movement of a number of screw dislocations
(Ref 32, 33) on the 45 plane, as shown in
Fig. 16.
In the intermediate range, the fracture behavior is complicated. The fracture does not consist of entirely slant-type fracture, nor does it
contain entirely a flat plane-strain-type fracture.
Instead, the regions of flat and slant fracture are
approximately equal. At the thin end of the
thickness range, the slant ligaments on either
side of the testpiece carry most of the load. At
the thick end of the range, the side ligaments
carry a much smaller percentage of the load. The
amount of flat fracture increases. This is shown
schematically in Fig. 17. It has been found (Ref
28) that the amount of flat fracture depends only
on the thickness of the test specimen and was
independent of crack length.
In the third region, the fracture consists of
predominantly flat fracture. Some evidence of
very small shear lips may be present at the later
part of fracture. Fracture is catastrophic and
rapid. No plastic deformation is evident. In this
third region, any increase in the thickness of the
testpiece causes no further decrease in the
toughness.
These fracture patterns are useful in determining the state of stress within a failed component and can help to understand the
mechanism of failure.
One famous failure involving brittle fracture
was the Great Boston Molasses Disaster
Fig. 15
Fig. 16
Variation of toughness with thickness
Mode of separation in a thin sheet
pg 56
18/8/2008 2:55PM Plate # 0
(Ref 34). In this failure, the United States

Alcohol Company fabricated a large cast iron
molasses tank in Boston in December 1915. This
tank was 27 m (90 ft) wide and 17.7 m (58 ft)
tall, with a head of 15 m (49.5 ft) of molasses. It
was fabricated of cast iron plates riveted together. It held 8.7 106 L (2.3 million gal) of
molasses, ostensibly used for the fermentation of
ethanol used for liquor. The man who oversaw
construction could not read blueprints, nor did
he have any technical training. No engineers or

architects were consulted to ensure that the tank
was constructed safely. On January 15, 1919, the
tank exploded with great force, and the streets of
Boston were flooded with waves of molasses
from 2 to over 4 m (8 to 15 ft) tall (Fig. 18). This
great wall of molasses was reported to have
moved at speeds up to 35 miles (56 km) per
hour and devastated a large section of Boston
along Commercial Street between Copps Hill
and the playground of North End Park. Half-inch
steel plates were torn apart, and these plates
were thrown with enough force to cut girders of
the elevated railway. This explosion, and the
subsequent wave of molasses, resulted in 21
people killed, 150 people injured, many buildings destroyed, and an entire area devastated.
The elevated train trestles were knocked over.
Early accounts of the disaster included reports
that the tank was destroyed by anarchists. In a
trial, it was found that the company was liable
for $628,000 in damages (in 2007 dollars,
approximately $7,000,000). Investigation many
years later indicated that the probable cause was
brittle fracture of the tank at the rivets, with the
temperature below the ductile-to-brittle transition temperature. One interesting result of this
disaster was that Massachusetts and many other
states created laws to certify engineers and to
regulate construction. It also required stamped
drawings certifying that an engineer had
reviewed the plans. It was this failure that was
the origin of the professional engineers license
and stamp, as it is known today (2007). As a side
note, the 18th Amendment was ratified and
Prohibition signed into law on January 16, 1919.
In another example of brittle fracture, an AISI
4330V hook-point, used for the arrestment of
Fig. 17
Schematic of fracture in the intermediate range
Fig. 18
The Great Boston Molasses Disaster. Twenty-one people were killed and over 150 buildings destroyed as the result of
2.3 million gal of molasses flooding North Boston.
pg 57
18/8/2008 2:55PM Plate # 0
Fig. 19
Arresting gear hook-point, manufactured from AISI 4330V, that failed during landing. Failure occurred at the inner fillet
radius of the right-hand lug
naval aircraft on landing, failed during field

trials during the 13th arrestment. The landing
configuration was severe, with high aircraft sink
rates, high aircraft gross weight, and landing at a
large angle to the cable. The hook-point failed at
the inner fillet radius of the right-hand lug
(Fig. 19). The hook-point successfully engaged
the arrestment cable, with no other aircraft
damage. The part was forged, machined, heat
treated, and hard surfaced in the cable groove,
using a high-velocity oxyfuel coating for wear
resistance. Examination showed that the microstructure of the hook-point was quenched and
tempered martensite. Hardness measurements
showed that the hook-point had a substantially
higher hardness (HRC 54) than the specified
hardness of HRC 46 to 48. The chemistry of the
hook-point indicated that it was at the high side
of the specification, increasing the hardenability
of the steel and increasing the resistance to
tempering. Hydrogen measurements indicated
that the hydrogen content was 0.2 ppm. The
high strain rate during landing and the low
concentration of hydrogen precluded failure by
hydrogen embrittlement. An SEM examination
of the fracture surface showed that the fracture
contained microvoid coalescence and quasicleavage, suggestive of brittle failure (Fig. 20).
Charpy impact testing showed that the impact
toughness of the as-received part was significantly lower than a part of the same chemistry properly tempered to HRC 46. Finite
element analysis showed a high localized stress
concentration at the lug inside fillet radius.
It also showed that the stresses were highly
triaxial. Based on the analysis, it was determined
that the hook-point lug failed by quasi-cleavage,
Fig. 20
SEM fractographs showing (a) location of origin at the

inner fillet radius and (b) quasi-cleavage evident on
the fracture surface
and that the failure was aggravated by high local

stress concentration at the fillet radius, improper
heat treatment (making the material more
pg 58
18/8/2008 2:55PM Plate # 0
brittle), and extremely high dynamic loading. It

was recommended that the radius be made larger
to reduce the stress concentration and also to
retemper the hook-points to meet specification.

Intergranular Brittle Fracture
Another form of brittle fracture is called
intergranular cracking. In this fracture mechanism, failure occurs by decohesion along grain
boundaries and not on specific crystallographic
planes, such as in cleavage fracture. Intergranular cracking can have several different
causes. Typical causes of intergranular cracking
in steel alloys include:

Quench-age embrittlement: Cooling of carbon steels and low-alloy steels from subcritical temperatures can precipitate carbides
within the microstructure. The strength is
raised, but toughness is lost.
Quench cracking: During quenching, the
transformational and residual stresses
developed during quenching of steels can
cause cracking during heat treatment.
Tempered martensite embrittlement: Within
the range where blue-purple oxides can form
on steels (230 to 370 C, or 450 to 700 F),
precipitates can form that increase the tensile
strength and hardness while reducing the
ductility and toughness.
Temper embrittlement: Quenched steels
containing appreciable amounts of manganese, silicon, nickel, or chromium are susceptible to temper embrittlement if they
contain even trace amounts of antimony, tin,
or arsenic. Embrittlement of susceptible
steels can occur after heating in the range of
370 to 575 C (700 to 1070 F) but occurs
most rapidly at approximately 450 to 475 C
(840 to 885 F).
Graphitization: This happens when the
pearlite in steels begins to decompose into
ferrite and graphite following very long,
high-temperature service, for example, in
steam power stations. For these applications,
a few steels turn out to be satisfactory, while
many others are subject to graphitization.
Internal oxidation: This is one of the common failures in high-temperature, oxidizing
conditions.
Liquid metal embrittlement or solid metal
embrittlement: Intermetallic compounds
form at grain boundaries when low-meltingtemperature metals (cadmium, zinc, etc.)

penetrate by diffusion. An example would be

galvanized steel where the zinc has diffused
into the steel in the vicinity of 420 C
(787 F).
Hydrogen embrittlement: The presence of
hydrogen and static loads or a low strain rate
can result in hydrogen embrittlement.
Stress-corrosion cracking
Grain-boundary decohesion at elevated
temperatures (creep rupture)
The fracture surface appearance of intergranular cracking is generally shiny and faceted.
It has a rock-candy appearance. Often, when
the mechanism is from corrosion, the corrosion
product is present. This can dull the appearance
of the facets. The appearance of intergranular
fracture is most clearly seen in the electron
microscope, and an example is shown in Fig. 21.
Quench cracking is the limiting case of
excessive residual stresses exceeding the tensile
strength of the material. Two processes contribute to quench cracking, as well as distortion
and residual stresses. The first process is the
stress from the volume expansion of martensite
during transformation from austenite to martensite. The second source is from thermal stress
due to differential contraction due to different
cooling rates in the steel. The transformational
stress from the formation of martensite is primarily responsible for cracking during quenching, and thermal stresses from differential
cooling are usually from subcritical heat treatments such as annealing.
During quenching, the volume expands from
the close-packed face-centered cubic structure
of austenite to the body-centered tetragonal
structure of martensite. This volume expansion
Fig. 21
Intergranular fracture from hydrogen embrittlement,

as seen through the SEM
pg 59
18/8/2008 2:55PM Plate # 0
is approximately 4% and is related to the carbon

content of the steel. During quenching, the outer
surface of the part cools first and transforms to
martensite. There is an attendant volume
expansion at the surface, and the untransformed
and still hot interior surface usually has sufficient plasticity to accommodate the changes in
the part volume. The outside surface is in compression. Upon cooling, the interior of the part
also transforms to martensite but is constrained
by the hard outside surface layer of previously
transformed martensite. On the transformation
of the inner core, a volume expansion occurs in
the interior of the part, and the outer surface is
placed in tension. If quenching is severe, the
resulting tensile residual stresses can exceed the
ultimate tensile stress of the surface untempered
martensite. Cracking is intergranular and often
exhibits an oxide scale on the fracture surface. If
cracking occurred during quenching, remnants
of quench oil can be found on the surface of the
crack, and often, elevated-temperature scale is
apparent. Cracking can be delayed due to the
transformation of retained austenite. This is one
reason why it is recommended to temper parts
immediately after quenching. Should delayed
quench cracking occur, then the temper scale
is thinner and often shows the characteristic
temper colors, indicative of the temper temperature. High-carbon steels and steels with high
hardenability are the most prone to quench
cracking.
Surface features such as sharp radii, large
changes in section, or the presence of laps, burrs,
rough-machined surfaces, and other surface
discontinuities increase the constraint during
quenching and increase the propensity toward
quench cracking.
Quench cracking can be mitigated by
improved surface condition and the removal
of scale, burrs, and sharp edges. Geometry
changes, by increasing transitions from thin
to thick sections, and generous radii can also
help reduce quench cracking. The use of higherhardenability alloys will also reduce the propensity for cracking, because it will allow a
reduced quench rate to achieve the same properties. Reducing the austenitizing temperature
or reducing the temperature differential between
the austenitizing temperature and the quenchant
temperature will reduce the propensity for cracking. Often, the geometry is set, as is the alloy of
the part. In this case, the heat treater can reduce
the quench rate or use martempering to reduce
quench cracking.
Martempering is the process of using hightemperature quench oils and quench oil temperatures of 90 to approximately 200 C (200 to
400 F). The part is quenched into the hightemperature oil, and the parts are allowed to
equilibrate or at least minimize the temperature
gradient across the interior of the part. The part
is then removed from the oil and allowed to cool
in any convenient manner. This method has
proven to be very effective in reducing quench
cracking as well as distortion from quenching.
A long pinion gear failed in service near the
midlength of the shaft (Fig. 22). One gear tooth
fractured during service, resulting in the gear
being removed from service and sent to the
laboratory for failure analysis. Magnetic particle
inspection, using a fluorescent dye, revealed
the presence of multiple linear indications on
cracking of the gear tooth faces (Fig. 23).
Examination of the fracture surface showed a
discolored region at the origin of cracking
(Fig. 24). This discolored region was attributed
Fig. 22
As-received pinion gear that failed in service
Fig. 23
Magnetic particle inspection of the failed pinion gear

showed arc-shaped cracks on the gear tooth faces.
pg 60
18/8/2008 2:55PM Plate # 0
pg 61
Region of Cracking
Fig. 24
Overall view of the cracked pinion showing the location of the fracture and the presence of a discolored region
Smeared surface showing

secondary cracking
Region of Discoloration
Fracture
Origins
Region of Discoloration
Fig. 26
Rough machining at the surface of the tooth showing

smearing and tearing of the machined surface. This is
suggestive of abusive machining, due to dull cutting tools,
inadequate coolant, or excessive speeds and feeds.
Fig. 25
Closeup of the fracture region showing the discolored region. The color of the oxidation indicated
that the crack occurred after quenching and during the tempering
operation.
to oxidation that occurred during heat treatment. The coloration of the oxide scale suggested that the oxidation occurred during
tempering (Fig. 25, 26). If the crack was preexisting prior to heat treatment, it would be
darker and thicker.
Examination of the tooth faces showed secondary cracking at regions of tearing and
smearing along the tooth face (Fig. 27), suggestive of abusive machining practice, including
the use of a tool that was dull or excessive feeds
and cutting speeds.
18/8/2008 2:55PM Plate # 0
Fig. 27
Secondary cracking evident at regions of abusive

machining
Fig. 28
Metallographic specimen of the pinion showing

inadequate case at the root of the tooth. Etched in
0.5% nital
Fig. 29
Large gear that showed evidence of cracking. (a) As-received gear. (b) Crack evident on gear face. (c) Region after temper
etching showing evidence of abusive grinding
Metallography of the teeth showed no evidence of burning or excessive temperature. The

root of the tooth showed little evidence of proper
hardening or case (Fig. 28). The microstructure
in the root consisted of ferrite and pearlite, with
lightly tempered martensite, further suggesting
inadequate heat treatment. The tooth tip showed
a fine-grained martensitic structure. No evidence
of overheating was present. Examination of
the tooth surface showed tears and smearing.
Microhardness of the hardened regions of the
tooth showed a hardness of HRC 58, while the
root of the tooth was HRC 29, consistent with
the observed microstructure.
Investigation of the induction heat treating
conditions revealed that the concentration of the
quenchant used was approximately 5%, while
6 to 10% was specified. The concentration was
controlled solely by refractometer. Contamination of the quenchant was unknown.
Based on the evidence, it was determined that
fracture and failure of the pinion gear tooth was
caused by quench cracking, aggravated by

improper concentration control and inductionhardening parameters. The situation was further
aggravated by poor machining practice, creating
tearing and smearing at the surface.
Often, quench cracking can result not from
heat treating operations but from other sources,
such as abusive grinding (Fig. 29). In this case, a
large gear was found to be cracked. As is usually
the case, the heat treater was blamed. Temper
etching of the region of cracking showed a darkened region, suggesting overtempering of the
part in a localized region.
Localized overheating during service can also
result in quench cracking. A hook-point, used
for catching the large cable on an aircraft carrier,
showed evidence of cracking in the cable groove
(Fig. 30) after a carrier landing. The hook-point
was manufactured from AMS 6411 (AISI
4330V), heat treated, and a high-velocity oxyfuel (HVOF) coating was applied. Imprints of
the arresting cable were left in the cable groove.
pg 62
18/8/2008 2:55PM Plate # 0
The vertical cracks were exposed using liquid

nitrogen and impact loading. The crack faces
were discolored with a golden tint; they were
subsequently examined using the SEM
(Fig. 31). The cracks were intergranular along
Fig. 30
Arresting gear hook-point showing vertical cracking

in the cable groove and evidence of localized heating (from the temper colors in the cable groove)
Fig. 31
prior-austenite grains. The laboratory fracture

showed microvoid coalescence. The cracks
showed three distinct regions: incipient melting
at the surface, intergranular regions, and
laboratory-induced ductile fracture (Fig. 32).
A metallographic section (Fig. 33) through
the vertical cracks showed untempered martensite at the surface, a transition region of overtempered martensite, and finally, a region of
tempered martensite. No evidence of the HVOF
coating was observed at the crack initiation site.
Hardness in the core was KHN 460. Hardness in
the transition region was KHN 390, and the
surface had a hardness of KHN 620. The
microstructure is similar to a weldment heataffected zone and shows that a significant heat
event occurred.
Chemical analysis showed that the material
conformed to AMS 6411, with alloying elements at the top of the range increasing the
sensitivity to quench cracking. Hydrogen analysis indicated 0.8 ppm hydrogen. The levels of
hydrogen present and the high strain rate of
SEM examination of the vertical cracking. A indicates the presence of intergranular cracking along prior-austenite grain
boundaries. B indicates microvoid coalescence from the laboratory fracture during the exposure of the crack face.
pg 63
18/8/2008 2:55PM Plate # 0
loading precludes hydrogen embrittlement as a

possible failure mechanism.
Based on the analysis, it was determined that
the vertical cracking in the cable groove was the
Fig. 32
Fig. 33
result of transformational stresses from frictional heating during capture of the arresting
cable by the hook-point. The mechanism is
similar to quench cracking and was aggravated
Exposed crack face showing two distinct regions on the crack face. A, region of incipient melting. B, intergranular
fracture
Metallographic section through the vertical crack showing (from right to left) a lightly etching region of fine-grained
untempered martensite, a transition region of overtempered martensite, and a region of nominally tempered martensite.
Hardness in the untempered martensite was KHN 620. The transition region showed a hardness of 390 KHN, and the nominal core
hardness was KHN 460
pg 64
18/8/2008 2:55PM Plate # 0
by the higher-than-normal hardenability of the

alloy. Recommendations included the application of a different HVOF coating to better resist
the frictional heating of the cable during carrier
arrestments.
Tempered martensite embrittlement
(TME) may not be associated with impurity
atoms segregating to prior-austenite grain boundaries. The most common factor in TME is
the formation of cementite during tempering
(Ref 35). When a given steel has a low impurity
content, the source of TME is the decomposition
of retained austenite during the second stage of
tempering. Thomas first proposed this mechanism (Ref 36). This was found when transmission electron microscopy showed the presence
of thin regions of retained austenite between
martensite in as-quenched steels, which subsequently transformed to cementite on tempering
in the range of 230 to 470 C (450 to 700 F).
The presence of phosphorus also plays a role.
If two steels are compared, one containing a
higher concentration of phosphorus, the steel
with the higher phosphorus content will have
poorer impact properties than an identical steel
with a lower phosphorus level. This will remain
true through the entire range of tempering temperature up to approximately 500 C (932 F).
The fracture mode is intergranular along prioraustenite grain boundaries (Ref 37). It is likely
that phosphorus is present at the prior-austenite
grain boundaries. It is only after cementite precipitates in the tempered martensite that TME is
fully present. Often, the presence of molybdenum at concentrations up to approximately 0.5%
will reduce the effect of TME.
On June 19, 1974, during a cold start after a
long shutdown for repairing the Tennessee
Valley Authority Gallatin No. 2 unit, the intermediate-pressure/low-pressure rotor burst at
approximately 3400 rpm. The rotor had been in
operation for 106,000 h from its operational
start in May 1957 (Ref 38). The burst rotor was
forged from an air-melted ingot. This ingot was
produced by a large region of MnS segregation
zone that was present at the center of the ingot,
which was subsequently bored by machining
during fabrication of the rotor. The steam
temperature was 566 C (1050 F). Tempered
martensite embrittlement occurred over the long
period of operation and substantially reduced the
toughness of the rotor. The presence of the MnS
inclusions initiated fracture by creep-fatigue
interaction and was enhanced by the presence of
TME (Ref 39).
Rail steels have been documented to fail

because of TME (Ref 40). This was especially
true of older rails manufactured in open-hearth
furnaces with high phosphorus content. This
occurred because of slow cooling through the
500 C (930 F) range or from isothermal holding at 500 C. Figure 34 shows a representative
SEM fractograph of an Fe-0.26C-2.11Si2.27Mn-1.59Cr wt% carbide-free bainitic rail
steel that has been temper embrittled by heat
treatment at 500 C for 5 h (Ref 40).
Temper embrittlement is only now becoming understood with regard to its mechanism.
However, the conditions of temper embrittlement are well known (Ref 41, 42).
Steels must be heat treated or cooled through
the range of 375 to 575 C (706 to 1070 F) in
order to become temper embrittled. Temper embrittlement is typically detected by an increase
in the ductile-to-brittle transition temperature.
This is shown in Fig. 35 for AISI 3140 steel
temper embrittled by furnace cooling through
the critical range and holding at 550 C
(1020 F) (Ref 35). The embrittlement reaction
follows a typical C-curve, with the minimum in
embrittling time at approximately 1 h at 550 C
(1020 F) and several hundred hours at 375 C
(706 F) (Ref 43). By heating to approximately
575 C (1070 F), temper embrittlement is
reversible and can be eliminated after holding
for only a few minutes at temperature.
For temper embrittlement to occur, specific
embrittling impurities must be present. These
include antimony, phosphorus, tin, and arsenic.
Quantities of less than 0.01% are enough to
cause temper embrittlement. For the most part,
simple plain carbon steels are not considered
to be susceptible to temper embrittlement
Fig. 34
SEM fractograph of Fe-0.26C-2.11Si-2.27Mn1.59Cr wt% carbide-free bainitic rail steel that has
been temper embrittled by heat treatment at 500 C for 5 h
pg 65
18/8/2008 2:55PM Plate # 0
pg 66
as long as manganese concentrations are held

to below 0.5%. Alloy steels containing chromium and nickel are the most prone; however,
additions of molybdenum at a concentration
of up to 0.5% are effective in reducing the
susceptibility of these steels to temper embrittlement.
Large forgings have been prone to temper
embrittlement because of the slow cooling that
occurs during fabrication. These large forgings
are also prone because of the operating
temperatures applied, especially in large turbine
rotors.
Liquid Metal Embrittlement or Solid Metal
Embrittlement. Exposure of steels to liquid
metals has been observed to result in brittle
fracture along prior-austenite grain boundaries
(Ref 44). Steels may be embrittled by exposure
to any of the low-melting metals shown in
Table 1 (Ref 45). Embrittlement occurs by
wetting of the prior-austenite grain boundaries
with a thin film of the molten metal. Usually,
very low tensile stresses are required to fail parts
that are liquid metal embrittled.
In general, three conditions are necessary for
liquid metal embrittlement. First, the embrittling
metal must be present, either externally as a
coating or internally. Internal sources can
include lead used to enhance machinability.
Second, temperatures that the part is exposed to
Energy absorbed, ft-lbf (J)
120
(160)
100
(130)
80
(100)
60
(80)
40
(50)
20
(30)
must be high enough that the embrittling metal

can melt. Lastly, tensile stresses must be present
as externally applied or internal residual stresses. Should any of these conditions not be met,
then it is unlikely that the steel will fail by liquid
metal embrittlement.
Liquid metal embrittlement has been known
to embrittle gun tubes. In 1977, during the
manufacture of a 105 mm M68 gun tube, lead
was electroplated to the tube and used as a
lubricant during the autofrettage process. During
the postautofrettage thermal treatment, the lead
melted and embrittled the gun tube. A complete
transverse brittle failure occurred. The axial
tensile residual stresses from the autofrettage
process were adequate to completely fracture the
tube, even though the hoop stresses were much
greater (Ref 46).
In another example, an ISO 8.8 low-alloy
steel bolt that was electroplated with cadmium
was used for an extended time at an elevated
temperature of 230 C (455 F). The resulting
failure showed intergranular fracture, with cadmium penetration along grain boundaries. This
cadmium penetration was detected by x-ray
diffraction (Fig. 36) (Ref 9).
On April 28, 1997, United Flight 1210, a
Boeing 737222 equipped with Pratt and
Whitney JT8D-7B engines, experienced an uncontained failure of the No. 2 engine (right side,
facing forward) high-pressure compressor
stage disk during takeoff. Takeoff was aborted,
and the aircraft was evacuated. Only two passengers were slightly injured during evacuation.
Postincident examination of the engine revealed
that two-fragments (approximately 50 by
100 mm, or 2 by 4 in.) separated from the disk.
Examination of the disk revealed a 100 mm
(4 in.) circumferential fracture around the diameter, with three additional fractures emanating
diagonally outward toward the rim. Also, cracks
Table 1 Melting temperatures of metals known
to embrittle high-strength steels
Melting temperature
Fig. 35
Shift in ductile-brittle transition temperature curve to

a higher temperature for AISI 3140 steel by holding at
500
and continuous cooling through the temper embrittlement critical range. Source: Ref 35
C
Metal
Mercury
Gallium
Indium
Lithium
Cadmium
Tin
Lead
Zinc
Antimony
39
29
156
180
321
232
327
419
642
38
85
313
356
610
449
620
787
1187
Source: Ref 45
18/8/2008 2:55PM Plate # 0
emanated radially outward from two tie-rod

holes, one of which bisected the fracture at the
diameter. No cracks were detected on any of the
other high-pressure compressor disks in the
engine.
On further examination by the National
Transportation Safety Board, the fractures were
found to have large intergranular areas in the
steel compressor disk (Ref 47). Solid molten
cadmium was detected along the prior-austenite
grain boundaries, indicative of liquid metal
(cadmium) embrittlement. Inadequate nickel
plating prestrike thickness was observed at the
surface of the disk. The failed disk was plated by
a trainee and inadequately plated with nickel. It
was found that the nickel plating was approximately 0.003 mm (0.00012 in.) in thickness,
which is below the Pratt and Whitney specified
thickness of 0.015 to 0.02 mm (0.0006 to
0.0008 in.). This thickness was inadequate to
prevent migration of cadmium into the steel
grain boundaries.
This was not the first time that liquid metal
embrittlement occurred in a compressor disk. On
July 23, 1990, the crew of a JT8D-9-equipped
Boeing 737100 reported that they heard a
muffled explosion during climb, followed by
a loss of rpm of the No. 1 (left) engine. The crew
returned to Houston, Texas, without incident.
Engine examination revealed a failure of a
disk spacer due to liquid metal embrittlement
(Ref 48).
An AISI 4330V hook-point failed during field
trials. This hook-point is used to grab the cable
on aircraft carriers and arrest the forward
movement of the aircraft during landing. Previous hook-points failed because of excessive
hardness and high triaxial stresses during impact
loading. These hook-points were evaluated, and
Fig. 36
discrepant parts were segregated. A series of

parts were then retempered to the specified
hardness of HRC 46 from HRC 51. The lug
radius was enlarged, and a new HVOF coating
was applied. During field trials, multiple hookpoints were identified by magnetic particle
inspection as having cracks in the lug radius
(Fig. 37). No through cracks were found. The
cracks were exposed, and a narrow uniform
region of intergranular fracture (approximately
200 mm) was observed (Fig. 38). Metallography
indicated that the microstructure was quenched
and tempered martensite, and the hardness was
within the specification of 46 to 48 HRC.
Metallography revealed that no decarburization
or precipitates were found at the prior-austenite
grain boundaries. Hydrogen analysis showed a
concentration of 0.2 ppm of hydrogen. The low
concentration of hydrogen and the rapid rate of
loading eliminated hydrogen embrittlement as a
cause of cracking. Auger analysis of the grain
boundaries within the intergranular region
showed the presence of cadmium at the grain
boundaries. The concentration of cadmium also
decreased as the grain boundaries were ion
milled away. This analysis indicated that the
fracture occurred because of liquid metal
embrittlement or solid metal embrittlement.
Solid metal embrittlement is similar to liquid
metal embrittlement, except that temperatures
are not high enough to cause melting of the
cadmium. For cadmium, solid metal embrittlement can occur at temperatures above 230 C
(450 F). A review of the planning showed that
the work order release did not include removal
of the cadmium plating prior to retempering
of the hook-points. Tempering to bring the hookpoints to the proper hardness was above 320 C
(610 F). Based on this, it was determined that
Liquid metal embrittlement of a low-alloy bolt plated with cadmium that failed during service. Cadmium was found to have
penetrated at the grain boundaries due to service above 230 C. (a) Overall fracture surface. (b) SEM examination of fracture
showing intergranular fracture. (c) X-ray diffraction spectrum at grain boundaries showing cadmium penetration
pg 67
18/8/2008 2:55PM Plate # 0
Fig. 37
AISI 4330V cadmium-plated hook-point used to

arrest landings of naval aircraft. Overall view of the
part, showing location of cracks observed using nondestructive
testing
the cracking observed in the hook-points was

due to liquid metal embrittlement caused by
failure to remove the cadmium plating prior to
tempering.
Hydrogen embrittlement is a particularly
insidious form of failure. Often, failure is delayed for hours, months, and possibly years after
the component has been fabricated. The results
may be catastrophic and unpredictable. The
failure mode is typically intergranular along
prior-austenite grain boundaries (Fig. 21).
Hydrogen can come from either external or
internal sources. One common source of
hydrogen is the steelmaking process (Ref 49),

and it is a significant problem in large sections
(Ref 50), where hydrogen embrittlement is
observed as flakes or a reduction in ductility (Ref
51). These flakes or blisters are regions where
hydrogen collects, until a bubble of hydrogen is
adequate to deform the surrounding region.
External sources of hydrogen are from manufacturing processes such as pickling (Ref 52, 53)
and plating (Ref 5456). Additional sources of
hydrogen can be the result of galvanic coupling
in an aqueous medium, in a similar fashion to
electroplating.
One of the particularly serious characteristics
of hydrogen embrittlement is the incubation
time required for it to occur. As a general rule,
the higher the hydrogen concentration, the
shorter the time to failure. For a given hydrogen
concentration, as the stress is increased, the
incubation time is decreased.
In quenched and tempered steels, there are a
number of sites that can trap hydrogen. These
include martensite interlath interfaces, high
density of dislocations, and the carbide-matrix
interface. All of these sites can act as traps
for hydrogen (Ref 57). Once present, hydrogen
diffuses to traps, such as dislocation cores, and is
transported by dislocation motion (Ref 58).
Hydrogen can also collect at inclusions and
carbides, which are also good hydrogen traps.
The incubation time is dependent on the hydrogen diffusion rate in steel to the point of crack
initiation.
Quenched and tempered steels that have a
hardness above HRC 38 are generally given a
hydrogen embrittlement relief at 135 C
(275 F) for 24 h. This enables the hydrogen in
the part to diffuse out. This is based on the
study by Johnson, Morlet, and Troiano (Fig. 39)
(Ref 59). This hydrogen embrittlement relief is
usually mandated whenever parts are plated,
cleaned, or exposed in some fashion to aqueous
solutions such as coolants or acid. Alkaline
solutions are not generally prone to causing
Fasteners are prone to hydrogen embrittlement. In this example, an ISO 10.9 low-alloy
steel bolt grade that was zinc plated failed during
service. Multiple fracture initiation sites were
evident along the bolt head transition, with
intergranular fracture morphology and heavy
secondary cracking. A hydrogen source was
suggested from manufacturing (pickling stage)
and/or cathodic hydrogen charging due to anodic zinc plating (Fig. 40) (Ref 9).
pg 68
18/8/2008 2:56PM Plate # 0
pg 69
To verify that the baking process after chromium plating was adequate, a plating shop tested four chromium-plated 4340 notched tensile
Fig. 38
specimens. These test specimens were heat

treated to 1515 MPa (220 ksi) and chromium
plated. During a sustained load test, one of the
SEM examination of the hook-point showing a narrow region of intergranular fracture along prior-austenite grain boundaries
Applied Stress, 1000 psi (MPa)
Normal Notch Strength = 300,000 psi (2070 MPa)

(2070)
300
(1895)
275
(1725)
250
(1550)
225
(1380)
200
(1205)
175
(1035)
150
(860)
125
(690)
100
(520)
75
(345)
50
0.01
Uncharged
Bake 24 hr
Bake 18 hr
Bake 12 hr
Bake 17 hr
Bake 3 hr
Bake 0.5 hr
0.1
10
100
1000
Fracture Time, hrs
Fig. 39
Baking AISI 4340 steel at 300 F for different times, showing the effect of baking on the incubation of failure
18/8/2008 2:56PM Plate # 0
Fig. 40
Hydrogen embrittlement failure of an ISO 10.9 low-alloy steel bolt grade. (a) As-received bolt. (b) Multiple initiation sites
with secondary cracks evident. (c) Intergranular fracture along prior-austenite grain boundaries
Fig. 41
As-received notched tensile specimen showing

location of fracture. Tensile specimen was fabricated
from 4340 steel, heat treated to 1515 MPa (220 ksi), and chromium plated.
specimens failed prematurely (Fig. 41). The

fracture was located at the notch. The fracture
surface (Fig. 42) was examined, and the origin
showed a shiny, faceted surface. At the origin,
the fracture was intergranular, while away from
the origin, near the center of the fracture surface,
the fracture mechanism was microvoid coalescence. Hydrogen analysis on the notched
tensile specimen yielded an average hydrogen
concentration of 12 ppm hydrogen. This is
considered very high and is sufficient to cause
hydrogen embrittlement. Metallography of the
test specimen showed a normal quenched and
tempered microstructure, typical of a steel heat
treated to this hardness.
During a routine wheel and tire change, a new
jack pad for a military aircraft failed, causing an
aircraft to drop prematurely onto the new wheel.
No damage occurred to the aircraft. The jack pad
was machined from 300M steel that was heat
treated to HRC 54 to 55. The jack pad was
chromium plated.
The as-received jack pad (Fig. 43) was
examined, and two fracture surfaces were identified (Fig. 44). These were identified as origins
1 and 2. Ridges emanated away from a distinct
origin location on each of the fracture surfaces.
Evidence of light corrosion products was found

at the fracture origin of origin 2. SEM examination of each of the origins (Fig. 45, 46)
revealed that the fracture was intergranular. At a
distance away from the origin, the fracture
consisted of microvoid coalescence, consistent
with rapid ductile rupture.
A metallographic specimen was removed
from the largest origin location and examined
(Fig. 47). The microstructure of the steel was
quenched and tempered martensite, typical of
300M heat treated to HRC 54 to 55. Chromium
plating was found to be intact at the fracture
origin.
Hydrogen analysis conducted on the jack pad
showed hydrogen concentrations of 4 and
6 ppm, which is considered adequate hydrogen
to cause hydrogen embrittlement in 300M steel
heat treated to this hardness.
Based on this investigation, it was concluded
that the jack pad most likely failed from
Stress-corrosion cracking is the attack of a
material by the combined action of tensile stress
on a part, either externally from an applied force
or internally from residual stresses, and a specific corrosive environment. Common features
are brittle fracture with little ductility, localized
corrosive attack, and a specific environmentalalloy system. Failure by stress-corrosion cracking (SCC) is characterized by exposure to a
specific chemical environment and the simultaneous application of a tensile stress. Without one
or the other, SCC will not occur. Fine cracks can
penetrate deeply into the part without obvious
signs of attack. Impending failure can occur
without warning.
The applied tensile stresses can be from
the service environment or from any of the
pg 70
18/8/2008 2:56PM Plate # 0
(a)
Fig. 43
As-received jack pad showing the locations of the

two distinct origins on the inside bore of the hole for a
pressed-in pin
(b)
(c)
Fig. 42
Overall view of the fracture surface, showing location and results of SEM examinations. (a) Overall
fracture surface and location of origin. (b) Intergranular fracture at
the origin of cracking (location A). Original magnification:
1000 . (b) Microvoid coalescence at location B
Fig. 44
Fracture surfaces of the jack pad showing location of

the origins. Original magnification: 2
pg 71
18/8/2008 2:56PM Plate # 0
numerous sources of residual stresses from

manufacturing (thermal processing, machining,
grinding, surface finishing, fabrication, or
assembly). The tensile stress is important in
the rupture of any protective film during initiation and subsequent propagation of the crack.
There appears to be a threshold tensile stress
intensity, KISCC, below which SCC does not
occur. This stress intensity is dependent on
the alloy, the heat treated condition, and the

environment.
The site of initiation of SCC may be microscopic. This could be from local differences in
metal composition or stress concentrations. A
pre-existing mechanical flaw or discontinuity
may act as a stress raiser and serve as a site for
SCC initiation.
Stress corrosion cracking usually exhibits extensive branching and propagates in a direction
(a)
(a)
(b)
(b)
(c)
(c)
Fig. 45
Fig. 46
SEM examination of origin 1. (a) Location of the

fracture origin. Original magnification: 20 . (b)
Location A showing a region of intergranular fracture along prioraustenite grain boundaries. Original magnification: 1000 . (c)
Location B, at a distance away from origin 1, showing microvoid
coalescence. Original magnification: 2000
SEM examination of origin 2. (a) Location of the

fracture origin. Original magnification: 100 .
(b) Location C showing a region of intergranular fracture along
prior-austenite grain boundaries. Original magnification: 1000 .
(c) Location D, at a distance away from origin 2, showing
microvoid coalescence. Original magnification: 2000
pg 72
18/8/2008 2:56PM Plate # 0
perpendicular to the tensile stresses contributing

to propagation and initiation. However, this is
not always the case. Structural steels exposed to
agricultural ammonia may exhibit nonbranched
cracking.
Stress-corrosion cracking has several special
characteristics that differentiate it from other
forms of cracking:

Only certain specific environments for a

specific alloy system cause SCC. There is
no general pattern regarding the corroding
environments or alloy systems.
Pure metals are much less susceptible to
SCC.
Cathodic protection has been successful in
preventing the initiation of SCC or in stopping the propagation of SCC.
Addition of certain soluble salts effectively
can poison the environment and either reduce or stop the propagation of SCC cracks.
Certain metallurgical features, such as grain
size, can influence the susceptibility of an
alloy system to SCC attack.
Macroscopically, fractures produced by SCC

show little ductility and nearly always appear
brittle. The fracture surfaces usually contain regions that are identifiable as the crack initiation
site, slow crack propagation, and final failure.
The regions containing the slow propagation
often contain corrosion products or are
discolored. This region extends to the region of
final fast fracture. However, this can also
be misleading, because the fracture could have
corroded before inspection, or the environment
may not be conducive to straining the fracture.
Fig. 47
Micrograph showing quenched and tempered martensite, typical of 300M heat treated to HRC 54 to 55.
Note that the chromium plating is intact.
It is often difficult to differentiate between

SCC and hydrogen-induced damage solely from
the fracture surface. Fractures of both types
exhibit intergranular features and tend to follow
prior-austenite grain boundaries. Metallography
is important to determine if branched cracking
has occurred. Even so, the absence of branched
cracking may not preclude SCC. In general, the
environment that the part was exposed to can be
the deciding factor of whether it is SCC or
hydrogen embrittlement (Fig. 48).
Low-carbon steels generally become more
susceptible to SCC as the carbon concentration
increases. Decarburized steels and pure iron are
resistant to SCC. Microstructure plays a greater
role in susceptibility to SCC than does the
alloying elements. High-alloy steels in a variety
of environments show that the heat treated
strength of the alloy is more important than
strictly the concentration. Steels that have been
heat treated to 1240 MPa (180 ksi) or higher are
especially susceptible to SCC. Typical environments that can cause SCC in steels are shown
in Table 2.
Caustic cracking in boilers is a serious SCC
problem and has caused many failures in steam
boilers. These failures usually initiate in riveted
and welded structures, where small leaks allow
buildup of caustic soda and silica. Cracking is
usually intergranular. Failures of this type have
occurred with concentrations of NaOH as low as
5% in water. Failures take place when the
operating temperature is in the range of 200 to
250 C (390 to 480 F). The concentration of
NaOH needed to cause cracking initiation
decreases as the temperature is increased.
Cracking of low-carbon steels and low-alloy
steels in nitrate solutions occurs in tubing and
couplings in high-pressure condensate wells.
Cracking in nitrate solutions is intergranular,
following prior-austenite grain boundaries.
Generally, acidic solutions cause this type of
cracking. Raising the pH of the solution enhances resistance to SCC, while increasing the
concentration of nitrate-containing solutions
tends to increase the susceptibility to SCC.
Carbon steel tanks containing ammonia have
also developed leaks because of SCC. Both plain
carbon steels and quenched and tempered steel
plate have shown a susceptibility to SCC in
ammonia. Failures occurred in ammonia mixed
with air and carbon dioxide. The presence of
water vapor delayed cracking.
Halide-containing environments, such as
seawater, are particularly severe for alloy steels
pg 73
18/8/2008 2:56PM Plate # 0
Direction of advancing cracking into metal
H+ + e
++
++
M M+
Anodic Stress
Corrosion Cracking
Time to Cracking
Region of Anodic
Stress Corrosion Cracking
Region of
Immunity
Anodic Current
M
Fig. 48
M++ + 2e
Hydrogen
Embrittlement
Region of
Hydrogen Embrittlement
Cathodic Current
2e + 2H+ 2H
Schematic differentiation of anodic stress-corrosion cracking and cathodic hydrogen embrittlement
Table 2 Environments that produce

stress-corrosion failures in carbon and
low-alloy steels
Medium
Type of
fracture
(a)
Aqueous chloride
environments
I,T
Caustic solutions
Nitrates
I
I
HNO3
HCN
Seacoast and
industrial
environments
Water, humid air,
and gas
I
I
I
Comments
Prevalent in high-strength steels heat

treated to 1380 MPa (200 ksi) or
greater
Well known as caustic embrittlement
Examples of bridge cable failures in
ammonium nitrate or sodium
nitrate solutions
...
...
High-strength steels heat treated to
1380 MPa (200 ksi) or greater are
especially prone.
High-strength steels heat treated to
1380 MPa (200 ksi) or greater are
especially prone.
(a) I, intergranular failure; T, transgranular failure
heat treated to above 1380 MPa (200 ksi).

The use of cadmium plating, low-hydrogen
practices, and adequate baking are helpful
in preventing SCC in steels such as 300M

or 4340.
On August 22, 2003, an empty cargo tanker
pulled upto a tank containing anhydrous
ammonia. Approximately 1 hour after being
filled, the front head cracked open (Fig. 49)
and started to release anhydrous ammonia.
Approximately 100 workers were evacuated
from the building. Five people were treated
for inhalation injuries and released. The
cost to repair the trailer was approximately
$25,000.
Examination revealed a 40 cm (16 in.) long
through-wall crack next to the radial weld in the
front head at the 1 oclock position (Fig. 50).
Internal examination using magnetic particle
inspection found two additional cracks that had
not yet propagated through the wall of the tanker. SEM examination of the cracks (Ref 60)
found that the fracture was branched and intergranular, with extensive surface corrosion on the
crack faces.
pg 74
18/8/2008 2:56PM Plate # 0
In the 1950s, the Agricultural Ammonia

Institute determined that caustic cracking of
ammonia-containing tanks was the reason that a
number of carbon steel tanks had failed (Ref 61).
They further determined that the addition of
0.1% water to anhydrous ammonia inhibited
SCC in carbon steel. The committee recommended that at least 0.2% water be added to
inhibit cracking. Further cracking occurred in
the 1960s in quenched and tempered ASTM
A517 steel, because purity levels had increased
and water was no longer being added. In 1975,
the Department of Transportation adopted regulations (Title 49 Code of Federal Regulations,
Parts 171 to 180) that required cargo tanks
fabricated from quenched and tempered steel
should only be used for anhydrous ammonia
if the solution contained 0.2% water. The
Fig. 49
regulation further required tankers to be

placarded with signs indicating QT or NQT,
for quenched and tempered or not quenched and
tempered.
The National Transportation Safety Board
determined that the failure of the tank and the
subsequent release of anhydrous ammonia were
due to caustic cracking (SCC) of the tank from
the transport of anhydrous ammonia containing
less than 0.2% water.
A Boeing 7572008 was parked at a gate at
Copenhagen, Denmark, and boarding of passengers was nearly completed when the righthand main landing gear truck beam failed (Ref
62). As the beam failed, the right side of the
aircraft rested on the shock strut instead of on the
wheels. Figure 51 shows the failed truck beam
and the aircraft resting on the shock strut. A
Accident cargo tank with QT designation, which indicates quenched and tempered steel
pg 75
18/8/2008 2:56PM Plate # 0
sketch of the main landing gear assembly of a

B-757 is shown in Fig. 52. The fracture surface,
showing evidence of corrosion, is shown in
Fig. 53.
Metallurgical analysis indicated that the
fracture mode was due to SCC. Examination of
the finish on the inner diameter showed that the
plating on the inside diameter was thin or nonexistent, and that it did not receive the required
shot peening. Because the truck beam was
overhauled and the original plating was retained
and worn during service, it was likely that the
overhaul was inadequate or improper. This had
the result of minimal cadmium protection on the
inner diameter surface of the truck beam. Subsequent loss of the plating led to premature and
severe corrosion in service and eventual fracture
due to SCC.
Creep Rupture The effects of temperature
on mechanical properties and material behavior
are commonplace in everyday living. Examples

include pipes bursting in the middle of winter,
the expansion of a bridge in the middle of summer, and the sagging of a fireplace grate. Each of
these examples is an indication that properties
change with temperature. In addition, the previous discussion indicated that steels become
more brittle as the temperature is decreased.
There are many other effects of temperature that
have been cited (Ref 63). Even the concept of
elevated temperature is relative (Ref 64). What
is considered hot for one material may be considered cold for another; for example, gallium
has a melting point of 30 C, while tungsten has
a melting point of approximately 3400 C.
Creep is the continuous deformation of a
material as a function of time and temperature.
This topic is treated very thoroughly in Ref 65.
The creep of a material is shown in Fig. 54. It can
be seen from the figure that creep in a material
occurs in three stages:

Stage I, where a rapid creep rate is seen at the

onset of load application, then gradually
decreases
Stage II, where creep remains at a steadystate rate
Stage III, where the creep rate shows an
increasing rate until failure occurs
The behavior and creep rate are sensitive to

the temperature to which the material is
Fig. 50
Through-the-wall crack on accident tanker
Fig. 51
Boeing 757-2008 truck beam failure occurring on

Icelandic Air, aircraft registration TF-FIJ.
Fig. 52
Schematic of the assembly of a Boeing 757 main

landing gear showing the location of fracture
pg 76
18/8/2008 2:56PM Plate # 0
exposed, the surrounding atmosphere, and the

prior strain history. Andrade and Chalmers (Ref
66) were pioneers in the study of creep and
proposed that creep followed the equation:
e=e0 (1+bt1=3 )ekt
where b and k are material constants that can be

evaluated by several different methods (Ref 67).
A better fit for the creep of materials was
proposed by Garofalo (Ref 68). He indicated
that:
testing, the time to failure is measured at a constant stress and constant temperature. This test
has gained acceptance for elevated-temperature
testing of turbine blade materials in jet engines.
Using a tensile machine and high-temperature
furnace, the strain is measured in creep testing
by special extensometers suited for elevated
temperatures. In stress-rupture testing, a simple
apparatus such as a dial calipers is used, since
only the overall strain at constant time and
temperature is needed.
Fatigue
de
e=e0 +et (17e7n )+ t
dt
where de/dt is the steady-state creep rate, e0 is

the strain on loading, n is the ratio of the transient creep rate to the transient creep strain, and
et is the transient creep strain.
Very early, it was recognized that fractures
at elevated temperatures occurred along grain
facets (Ref 69). In stage III creep, intergranular
wedge cracks and cavities form. Wedge-shaped
cracks and creep cavities usually initiate at or
near grain-boundary triple points and propagate
along grain boundaries normal to the applied
tensile stress. Creep cavities form at higher
temperatures and lower working stresses. These
structural features are shown in Fig. 55.
Creep testing is usually performed for 1000
to 10,000 h with strains of up to 0.5%. Stressrupture testing, or testing to failure, uses much
higher loads and temperatures, and the test is
usually terminated after 1000 h. In stress-rupture
Fig. 53
Fracture surface of Boeing 757 main landing gear

truck beam on Icelandic Air aircraft TF-FIJ
Parts are subject to varying stresses during

service. These stresses are often in the form of
repeated or cyclic loading. After enough applications of load or stress, the components fail at
stresses significantly less than their yield
strength. Fatigue is a measure of the decrease in
resistance to repeated stresses.
Fatigue failures appear brittle, with no gross
deformation. The fracture surface is usually
normal to the main principal tensile stress. Fatigue failures are recognized by the appearance of
a smooth, rubbed type of surface, generally in a
semicircular pattern. The progress of the fracture
(and crack propagation) is generally suggested
by beach marks. This is illustrated in Fig. 56
and 57. The initiation site of fatigue failures is
generally at a stress-concentration site or stress
raiser. A typical fracture appearance is shown
schematically in Fig. 58.
Three factors are necessary for fatigue to
occur. First, the stress must be high enough that a
crack is initiated. Second, the variation in the
stress application must be large enough that the
crack can propagate. Third, the number of stress
applications must be sufficiently large that the
crack can propagate a significant distance. The
Fig. 54
Schematic representation of creep
pg 77
18/8/2008 2:57PM Plate # 0
fatigue life of a component is affected by a number of variables, including stress concentration,

corrosion, temperature, microstructure, residual
stresses, and combined stresses.
The structural features of fatigue failures are
generally divided into four distinct areas
(Ref 70):

Crack initiation, the early development of

fatigue damage
Slip band crack growth, the early stages of
crack propagation. This is often called stage
I crack growth.
Stable crack growth, which is usually normal
to the applied tensile stress. This is called
stage II crack growth.
Unstable crack growth, with final failure
from overload. This is called stage III crack
growth.
Fig. 55
Fig. 56
Creep cavities and creep wedges forming at grain

boundaries
Actual fatigue failure of a crankshaft showing

characteristic beach marks. Fatigue initiated at the
radius of the journal and exhibits classic bending fatigue.
Fatigue usually occurs at a free surface, with

the initial features of stage I growth, fatigue
cracks, being initiated at slip band extrusions and
intrusions (Ref 71, 72). Cottrell and Hull (Ref 73)
proposed a mechanism for the formation of these
extrusions and intrusions (shown schematically
in Fig. 59) that depends on the presence of slip,
with slip systems at 45 angles to each other
operating sequentially on loading and unloading.
Wood (Ref 74) suggested that the formation of
the intrusions and extrusions was the result of
fine slip and buildup of notches (Fig. 60). The
notch created on a microscopic scale would
be the initiation site of stable fatigue crack
growth.
In stage II, stable fatigue crack growth, striations (Fig. 61) often show the successive position of the crack front at each cycle of stress.
Fatigue striations are usually detected using
electron microscopy and are visual evidence that
fatigue occurred. However, the absence of fatigue striations does not preclude the occurrence
of fatigue.
Striations are formed by a plastic blunting
process (Ref 75). At the end of the stage I crack
tip, there exists sharp notches due to the presence
of slip. These sharp notches cause stress to be
concentrated at the crack tip. The application of
a tensile load opens the crack along slip planes
by plastic shearing, eventually blunting the
crack tip. When the load is released, the slip
direction reverses, and the crack tip is compressed and sharpened. This provides a sharp
notch at the new crack tip where propagation
Fig. 57
Fatigue failure of a fastener, with initiation of fatigue

occurring at the threads
pg 78
18/8/2008 2:57PM Plate # 0
Fig. 58
Schematic illustration of simple fatigue failures
pg 79
18/8/2008 2:57PM Plate # 0
can occur. This is shown schematically in

Fig. 62.
An alternative hypothesis on striation formation was presented by Forsyth and Ryder
(Ref 76). In their model, the triaxial stress state
at the crack tip forms a dimple ahead of the crack
front. The material between the crack tip and the
dimple contracts and eventually ruptures,
forming a fatigue striation. This is shown schematically in Fig. 63.
In mild steel, well-defined striations are
observed but not as well defined or as spectacular as in aluminum. This was first assumed
to be due to the crystal lattice structure, since
face-centered cubic austenitic steels show welldefined striations, and mild steels (basestructured) do not (Ref 77). Other alloys, such as
titanium alloys, with a hexagonal close-packed
crystal structure show very defined striations
(Ref 78). However, aluminum alloys (bodycentered cubic) show strongly defined striations
(Ref 79). Therefore, attributing defined striations to crystal lattice alone was discounted as
a viable theory.
Deformation and available slip systems were
presumed to be more significant (Ref 80).
However, this does not follow, because mild
carbon steels are more ductile than austenitic
steels. It is now generally accepted that
fatigue striations form by the plastic blunting

process.
It has also been found that the thicker the
testpiece, the faster the crack propagation rate
(Ref 81). It is likely that the propagation rates for
thicker pieces are due to increased plane-strain
conditions, with a small plastic zone at the crack
tip. Since there is a greater stress gradient for
a small plastic zone, a faster crack propagation
rate may be expected. Also, in thicker
panels there is a higher state of triaxial stress,
which would also tend to increase crack growth
rates.
Since fatigue failures usually begin at the
surface, the surface condition is very important.
Surface roughness is a primary factor influencing fatigue. Highly polished specimens exhibit
the longest fatigue life, with increasingly
rougher surfaces yielding decreased fatigue life.
Rough lathe or coarse grinding reduces the
fatigue strength by approximately 20% below
polished specimens (Ref 82). Electropolished
specimens have lower fatigue limits than
mechanically polished specimens, by up to 25%
(Ref 83). This reduction is due to the removal of
surface compressive residual layers induced
during mechanical finishing.
An example of a typical fatigue failure in an
ASTM B7 low-alloy steel bolt grade is shown in
Fig. 64 (Ref 9). Fracture initiation occurred
along the threads with typical and pronounced
beach marks (i.e., cyclic fracture propagation)
and transgranular fracture mode.
An example of a manufacturing effect on
fatigue is the following example of an arresting
gear hook shank (Fig. 65) used to slow down
aircraft when landing on aircraft carriers. In this
example, the hook failed after 1361 simulated
arrestments, which was below the lifetime of
2250 arrestments. The part is designed to last
Fig. 59
Schematic representation of the mechanism of fatigue intrusions and extrusions
Fig. 60
Mechanism of intrusions and extrusions. PSB, persistent slip bands. Source: Ref 74
pg 80
18/8/2008 2:57PM Plate # 0
two lifetimes, or 4500 arrestments, without

cracking (0.25 mm, or 0.010 in. detectible
flaw). The arresting shank was fatigue tested in a
fixture, with hydraulic cylinders providing loads
at the vertical damper and hook-point cable
groove. The maximum applied load was 90 mg
(200,000 lb). A schematic of the arresting hook
shank is shown in Fig. 65.
The arresting hook shank was fabricated from
an AerMet 100 rotary forging. It is rough turned
on a lathe on the outside, then gun drilled to
create a pilot hole down the length of the forging. The outer surface is turned to the final
Fig. 61
Typical fatigue striations in 7075 aluminum
Fig. 62
Mechanism for fatigue striation formation
diameter. The bore is then injection drilled to the

final dimensions. A follower supported the
injection drill. This is not a method that is
commonly used for final machining operations.
It is heat treated in vacuum to 1930 MPa
(280 ksi) ultimate tensile strength. The part is
inspected using dye-penetrant and magnetic
particle nondestructive testing methods. The
bore is visually inspected using a bore scope.
This is a difficult inspection because of the long
length and narrow bore.
Examination of the fracture surface showed
that cracking initiated at the hook-point side,
on the inner diameter, at a location approximately 26.5 cm (10.5 in.) aft of the uplock
retainer. The fracture had characteristics of
fatigue fracture, with multiple origins observed.
Surface roughness measurements varied across
the inner bore, from approximately 1 to 5 mm
(40 to 180 min.). The drawing requirement was
3 mm (125 min.). Circumferential machining
marks were found at the fracture origin
(Fig. 66). SEM examination (Fig. 67) showed
fatigue striations emanating away from the
identified origin. Cracking was found to have
initiated at circumferential machining marks.
Machining marks were observed at 4.3 mm
(0.17 in.) intervals. Many secondary cracks
were observed at the machining marks. Fatigue
was found to initiate subsurface to the inner
bore, adjacent to the machining marks. A welldefined surface layer was observed. This layer
had the appearance of mechanical working or
damage. This observed layer followed the feeds
and speeds of the injection drill. Metallography
showed that the material was quenched and
tempered martensite and was typical for this
material heat treated to this hardness. At the
Fig. 63
Striation formation from ductile dimple formation

ahead of a crack front
pg 81
18/8/2008 2:57PM Plate # 0
Fig. 64
ASTM B7 low-alloy steel bolt grade. Fracture initiated along threads, with typical and pronounced beach marks (i.e., cyclic
fracture propagation) and transgranular fracture mode. (a) Location of bolts in pump coupling. (b) Beach marks showing
asymmetrical bending with initiation at high stress-concentration factor at bolt threads. (c) Transgranular fracture morphology
Fig. 65
Fig. 66
Schematic of the failed arresting hook shank showing location of loads
Machining marks found on the inside of the bore, at

the origin of cracking
origin, the presence of well-defined subsurface

cracking was observed. This layer had the
appearance of smeared metal and base metal
pullout. Flat cracking, suggestive of fatigue
cracking, was observed to emanate from the
flaw (Fig. 68). The flaws were located at
4.3 mm (0.017 in.) intervals, identical to the
feed rate of the injection drilling process.
During the injection drilling process, three
cutters are used. Coolant is forced through a
central hole to cool the cutting tools and to flush
the chips. AerMet 100 tends to form long strands
of material during machining and does not want
to form chips. Hot chips can contact the freshly
machined surface. These chips or long strands
are under pressure at the cutter or follower
and can be forced onto the newly machined
surface by the follower. If the temperature and
pressures are high enough, solid-state welding of
the chips and bore surface can occur. As the
cutter boring bar moves, pullout can occur. The
pg 82
18/8/2008 2:57PM Plate # 0
Fig. 68
Metallography of the arresting hook shank. (a) Typical quenched and tempered martensite found. This is
typical for the hardness of the arresting hook shank. (b) Pulled
material at 4.3 mm (0.17 in.) intervals along the inner bore of the
arresting hook shank. Origin is to the left. (c) Secondary cracking
observed at the location of pulled material
Fig. 67
SEM examination of the fracture surface. (a) Fatigue

striations emanating from the fracture origin. (b)
Machining marks found on the surface of the inner bore. (c) Welldefined layer showing fatigue emanating from the damaged
material at the surface of the inner bore
examined flaws matched the machining feeds

and speeds.
The arresting hook failed by fatigue,
initiating at flaws created during the final
machining process. The defect morphology
suggested localized solid-state welding and
pullout from chip contact with the freshly
machined surface. The surface roughness and
finish of the inner bore did not meet drawing
requirements.
Summary
In this short overview of the possible
mechanisms of failure for steels, the following
were discussed:

Techniques for examining fractures

Ductile and brittle failures
Intergranular failure mechanisms
Fatigue
The previous discussion has shown that it is

important to look at not only the fracture surface
but at all the factors (manufacturing history,
service conditions, and loading). All the tools
available to the metallurgist should be used
these include photography, fractography, and
pg 83
18/8/2008 2:57PM Plate # 0
metallographyto understand the sources and

root cause of failure.
REFERENCES
1. Hippocrates, Epidemics, Book I, Section XI

2. D.S. MacKenzie, G. Chen, M. Barker,
C. Ramsay, J. Alderson, L. Dharani, and
J. Yu, Forensic Investigation of Failed Mast
Arms of Traffic Signal Supported Structures, Transport. Res. Record, Des. Struc.
2002, 2002, p 916
3. Loss of Control and Impact with Pacific
Ocean Alaska Airlines Flight 261 McDonnell Douglas MD-83, N963AS about 2.7
Miles North of Anacapa Island, California,
January 31, 2000, NTSB/AAR-02/01,
National Transportation Safety Board,
Adopted Dec 30, 2002, Notation 7263E
4. Uncontained Engine Failure/Fire Valujet
Airlines Flight 597 Douglas DC-9-32,
N908VJ, Atlanta, Georgia, June 8, 1995,
NTSB/AAR-96/03, National Transportation Safety Board, Adopted July 30, 1996,
Notation 6579A
5. W.M. Garrison and N.R. Moody, J. Phys.
Chem. Solids, Vol 48, 1987 p 1035
6. J.F. Knott, Advances in Fracture Research,
Proc. Seventh International Conference on
Fracture (Oxford, U.K.) 1989, p 125
7. J.F. Knott, Fracture 1977, Vol 1, Proc.
Fourth International Conference on Fracture, June 1977 (Waterloo Canada), p 61
8. H.G. Wilsforf, Mater. Sci. Eng., Vol 59,
1983, p 1
9. B. Bgner, G. Rrvik, and L. Marken, Bolt
FailuresCase Histories from the Norwegian Petroleum Industry, Microsc. Microanal., Vol 11 (Suppl. 2), 2005, p 16041605
10. T.L. Anderson, Fracture Mechanics, CRC
Press, New York, 1995
11. C. Ruggieri, Numerical Investigation of
Constraint Effects on Ductile Fracture in
Tensile Specimens, J. Braz. Soc. Mech. Sci.
Eng., Vol 26 (No. 2), 2004
12. Standard Test Methods for Determining
the Inclusion Content of Steel, E4505,
ASTM International
13. Premium Aircraft-Quality Steel Cleanliness: Magnetic Particle Inspection Procedure, AMS 2300
14. Aircraft-Quality
Steel
Cleanliness:
Magnetic Particle Inspection Procedure,
AMS 2301
15. Aircraft-Quality Steel Cleanliness: Martensitic Corrosion-Resistant Steels Magnetic

Particle Inspection Procedure, AMS 2303
16. Special Aircraft-Quality Steel Cleanliness:
Magnetic Particle Inspection Procedure,
AMS 2304
17. R. Kiessling and N. Lange, Non-Metallic
Inclusions in Steel, Part IIV, Iron and Steel
Institute, London, 1978
18. S. Maropoulos and N. Ridley, Inclusions
and Fracture Characteristics of HSLA Steel
Forgings, Mater. Sci. Eng. A, Vol 384 (No.
12), Oct 25, 2004, p 6469
19. Railway Investigation Report, Derailment
and Collision Canadian National Train
No. U-783-21-30 and Train No. M-306-3130 Mile 50.84, Saint-Hyacinthe Subdivision
Mont-Saint-Hilaire, Quebec, December 30
1999, Report R99H0010, Transportation
Safety Board of Canada
20. M.L. Williams, STP 158, ASTM, 1954,
p 1144
21. M.E. Shank, STP 158, ASTM, 1954, p
45110
22. Fractography, Vol 12, Metals Handbook,
9th ed., ASM International, 1987
23. J. McCall and P. French, Metallography in
Failure Analysis, Plenum, New York, 1978,
p6
24. Standard Test Methods for Notched
Bar Impact Testing of Metallic Methods,
E2388, American Society for Testing and
Materials
25. Standard Test Method for Plane Strain
Fracture Toughness, KIc, of Metallic Materials, E39974, American Society for
Testing and Materials
26. Standard Test Method for Conducting
Drop-Weight Test to Determine NilDuctility Transition Temperature of Ferritic
Steels, E20895A, American Society for
Testing and Materials, 1983
27. Standard Test Method for Dynamic Tear
Testing of Metallic Materials, E60483,
American Society for Testing and Materials, 1983
28. J.M. Kraft, A.M. Sullivan, and R.W. Boyle,
Proc. Symp. Crack Propagation, Cranfield,
U.K., 1961
29. W.F. Brown and J.E. Strawley, STP 381,
American Society for Testing and Materials, 1965, p 133
30. A.S. Tetelman and A.J. McEvily, Fracture
of Structural Materials, Wiley, New York,
1967
pg 84
18/8/2008 2:57PM Plate # 0
31. J.F. Knott, Fundamentals of Fracture

Mechanics, Wiley, New York, 1973
32. A.H. Cottrell, Proc. R. Soc. A, Vol 276
(No. 1), 1963
33. A.H. Knott, Mater. Sci. Eng., Vol 7 (No. 1),
1971
34. J.S. Colton, Class notes, Great Boston
Molasses Disaster, Jan 15, 1919, ME6222
Manufacturing Processes and Systems,
Georgia Institute of Technology
35. G. Krauss, Steels: Heat Treatment and
Processing Principles, ASM International,
1990, p 231, 237
36. G. Thomas, Retained Austenite and Tempered Martensite Embrittlement, Metall.
Trans. A, Vol 10, 1979, p 16431651
37. S. Banerji, C. McMahon, and H.
Feng, Intergranular Fracture in 4340Type Steels: Effects of Impurities and
Hydrogen, Metall. Trans. A, Vol 9, 1978,
p 237247
38. D. Kalderon, Steam Turbine Failure at
Hinkley Point A, Proc. Inst. Mech. Eng.,
Vol 186, 1972, p 341377
39. L.D. Kramer and D. Randolph, Analysis
of TVA Gallatin No. 2 Rotor Burst: Part
1Metallurgical Considerations, Proc.
1976 ASME-MPC Symposium on CreepFatigue Interaction, p 124
40. H.K.D.H. Bhadeshia, High Performance
Bainitic Steels, Mater. Sci. Forum, Vol 500,
2005, p 6374
41. I. Olefjord, Temper Embrittlement, Review 231, Int. Met. Rev., Vol 23, 1978, p
149163
42. B. Woodfine, Temper Brittleness, A Critical
Review of the Literature, JISI, Vol 173,
1953, p 229240
43. F. Carr, M. Golman, L. Jaffee, and D. Buffum, Isothermal Temper Embrittlement of
SAE 3140 Steel, Trans. TMS-AIME, Vol
197, 1953, p 998
44. Liquid Metal Embrittlement, Failure Analysis and Prevention, Vol 10, Metals
Handbook, 8th Ed., American Society for
Metals, 1975, p 228229
45. P. Fernandez, R. Clegg, and D. Jones,
Failure by Liquid Metal Induced Embrittlement, Eng. Fail. Anal., Vol 1 (No. 1),
1994, p 5163
46. G. Vilganlte, G. Troiano, and C. Mossey,
Liquid Metal Embrittlement of ASTM
A723 Gun Steel by Indium and Gallium,
ARCCB-TR-99011,
Army
Research,
Development and Development Center,
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
Benet Laboratories, Watervliet Arsenal,

June 1999
NTSB CHI97IA117, National Transportation Safety Board, adopted June 28,
1998
NTSB Safety Recommendation, in memo to
J. Garvy, Administrator A-97-83, FAA,
Aug 29, 1997
S.F. Carter, Effect of Melting Practice on
Hydrogen, J. Met. (AIME), Vol 188,
Jan 1950, p 3040 and J. Met. (AIME), Vol
188, Feb 1950, p 245255
F.B. Foley, Flakes and Cooling Cracks
in Forgings, Met. Alloys, Vol 12, 1940,
p 442445
F.J. Shortsleeve and A.R. Troiano, A Literature Survey on the Mechanism of Flake
Formation, Interim Technical Report 1,
Contract DA 33-019-ORD-897
C.A. Edwards, Pickling or the Action
of Acid Solution on Mild Steel and the
Diffusion of Hydrogen Through the Metal,
J. Iron Steel Inst., Vol 110, 1924, p 960
C.A. Zapffe and M.E. Haslem, Acid Composition, Concentration Temperature and
Pickling Time as Factors in the Hydrogen
Embrittlement of Mild Steel and Stainless
Steel Wire, Trans. ASM, Vol 39, 1947,
p 213236
J.R. Gustafson, Some of the Effects of
Cadmium, Zinc, and Tin Plating on Springs,
Proc. ASTM, Vol 47, 1947, p 782
K.B. Valentine, Stress Cracking of
Electroplated Lockwashers, Trans. ASM,
Vol 38, 1947, p 488504
C.A. Zapffe and M.E. Haslem, Measurement of Embrittlement during Chromium
and Cadmium Electroplating and the Nature
of Recovery of Plated Articles, Trans. ASM,
Vol 39, 1947, p 241258
G. Nagu and T. Namboodhiri, Effect of
Heat Treatments on the Hydrogen Embrittlement Susceptibility of API X-65 Grade
Line-Pipe Steel, Bull. Mater. Sci., Vol 26
(No. 4), June 2003, p 435439
A. Thompson and I. Berstein, Ed., Effect
of Hydrogen on Behavior of Materials,
TMS-AIME, 1976
H. Johnson, J. Morlet, and A. Troiano,
Hydrogen Crack Initiation and Delayed
Failure in Steel, Trans. TMS-AIME, Vol
212, 1958, p 528536
Release of Hazardous Materials from
Cargo Tank in Middletown, Ohio, on August
22, 2003, Accident DCA03MZ002, NTSB
pg 85
18/8/2008 2:57PM Plate # 0
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
Hazardous Materials Accident Brief,

Adopted July 22, 2004
A.W. Loginow, Stress-Corrosion Cracking
of Steel in Liquefied Ammonia Service,
Bull., Natl. Board Boiler Pressure Vessel
Insp., Vol 45 (No. 6), Oct 1988
Report by the Danish Accident Investigation Board into the Incident to Boeing 757,
TF-FIJ at Kastrup Airport, Denmark on
June 28, 2001, M-04001/AIG-09, Danish
Accident Review Board
J.E. Dorn, Ed., Mechanical Behavior of
Materials at Elevated Temperature,
McGraw-Hill, New York, 1961
R.W. Guard, Prod. Eng., Vol 27 (No. 10),
1956, p 160174
I. Finnie and W.R. Heller, Creep of Engineering Materials, McGraw-Hill, New
York, 1959
E.N. da C. Andrade and B. Chalmers, Proc.
R. Soc. (London) A, Vol 138, 1932, p 348
J.B. Conway, Trans. Metall. Soc. AIME,
Vol 223, 1965, p 2018
F. Garofalo, Properties of Crystalline
Solids, STP 283, ASTM, 1965
W. Rosenhahn and D. Ewen, J. Inst. Met.,
Vol 10, 1913, p 119
W.J. Plumbridge and D.A. Ryder, Metall.
Rev., Vol 14, 1969, p 136
P.J. Forsyth and C.A. Stubbington, J. Inst.
Met., Vol 83, 1955, p 395
W.A. Wood, Some Basic Studies of Fatigue
in Metals, Wiley, New York, 1959
73. A.H. Cottrel and D. Hull, Proc. R. Soc.

(London) A, Vol 242A, 1953, p 211
74. W.A. Wood, Bull. Inst. Met., Vol 3, 1955,
p5
75. C. Laird, Fatigue Crack Propagation, in
STP 415, American Society for Testing and
Materials, 1967, p 136
76. P.J. Forsyth and D.A. Ryder, Metallurgica,
Vol 63, 1961, p 117
77. G. Jacoby, Current Aeronautical Fatigue
Problems, J. Schijve, Ed., Pergamon, New
York, 1965, p 78
78. W.R.Warke and J.M. McCall, Fractography Using the Electron Microscope,
ASM Technical Report We-2-65, American
79. G. Jacoby, Fractographic Methods, Exp.
Mech., 1965, p 65
80. P.J. Forsyth, A Two Stage Process
of Fatigue Crack Growth, Symp. Crack
Propagation, Vol 2, Cranfield, U.K., 1961,
p 76
81. D. Broek and J. Schijve, The Influence
of Sheet Thickness in the Fatigue Crack
Propagation in 2024-T3 Alclad Sheet
Material, NLR Technical Report M2129,
Amsterdam, 1963
82. N.E. Frost, K.J. Marsh, and L.D. Pook,
Metal Fatigue, Oxford University Press,
London, 1974
83. T.T. Oberg and E.J. Wad, Technical Note
Report 56-289, Wright Air Development
Department, 1956
pg 86
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_87-109.pdf/Chap_02b/
21/8/2008 10:50AM Plate # 0

DOI: 10.1361/faht2008p087

Mechanisms and Causes of Failures

in Heat Treated Steel Parts
Debbie Aliya, Aliya Analytical, Inc.
THE TERM mechanism of failure means

different things to different people. One possible
definition refers to a particular product application, a particular type of component, and a
particular type of industry where certain environmental conditions are common. In many
cases, these mechanisms may apply only to a
limited type of material or an alloy family. In
order to understand particular failure mechanisms, it is important to understand the causes of
failure and categories of damage. This chapter
reviews various ways to classify failure categories. Information on mechanisms of damage,
in particular environment-material-type pairs, is
pro-vided toward the end of this chapter.
The term damage is often preferable to the
term failure, because damage is a technical term
that is very clear and has a specific meaning,
generally related to the physical condition of the
component. The term failure, on the other hand,
has many more philosophical connotations. It is
possible to see at once, perhaps, that the part is
damaged. Only at the end of the investigation
does one have a good chance of knowing with
a high degree of engineering certainty whether
the part itself failed, rather than the design, the
design system, the employee training system,
the procurement system, or the user-certification
system. Thus, it is important to avoid using the
term failed part until one has truly determined
that the part was the problem.
Often, one of the important goals in a failure
analysis of a heat treated part is to determine
whether the damage is the result of improper
heat treatment, that is, the heat treaters fault.
For many people in manufacturing, especially if
an independent or job shop heat treater is
involved, it is easy to blame the heat treater.
After all, he was the last one to touch the part
before assembly, and most people in the general
manufacturing arena understand heat treating
poorly, if at all.
If the part does not meet the specification for

mechanical or physical characteristics after heat
treating, it may be the heat treaters fault.
However, there may have been something
wrong with the raw material, or prior manufacturing processes that allowed a part that went
through the normal heat treating and inspection
process to have substandard properties, which
would not be the heat treaters fault after all.
This chapter gives some examples of lack of
conformance to specification that may at first
look like the heat treater did something wrong,
but where other contributing factors made it
difficult or impossible for the heat treater to meet
the specification.
This chapter also summarizes the basic types
of damage, with particular consideration given
to whether their likelihood can or cannot be
influenced by the heat treating process. The
classical organization for types of damage
(failures) is as follows: deformation, fracture,
wear, corrosion or other environmental damage,
and multiple or complex damage. Separately
from what the damage type is, one also should
look at the potential causes, sources, or factors
promoting the damage.
Failure analysts used to be taught to classify
failures as a result of defects or abuse. There is
still a large amount of literature that presents
failure causes in such binary terms. By limiting
the analysis to one of two possible causes,
opportunities may be missed for improving
the product. There are actually several
different ways to classify causes of failures and
damage.
This chapter also describes a process that can
be used to demonstrate likelihood that a product
was abused. Sometimes, the physical evidence
speaks clearly. If this is not the case, it may be
necessary to quantitatively prove, for example,
that the part was overloaded. This may be difficult to do, because exact service conditions are
pg 87
21/8/2008 10:51AM Plate # 0
frequently impossible to determine with any

degree of certainty.
Besides service or maintenance abuse, four
potential sources or origins of damage and failure relate to the product life cycle, as follows:

Design process omissions

Undesired raw material characteristics
Undesired component manufacturing characteristics (includes heat treating)
Improper service or maintenance conditions
In this chapter, a modified classification is

developed specifically for failure analysis of
heat treated parts where the heat treating is
suspected to be the cause of the failure. The first
logical potential source of poor component
performance is raw material characteristics.
Note the use of the word characteristics rather
than defects. This is an important thing to keep in
mind when performing this type of work. The
term defect has a specific legal definition that
does not necessarily apply to all cases of physical flaws or suboptimal material properties.
Thus, the use of the word defect may have
undesirable cascading consequences, especially
when personal injury or large financial losses
were a result of the component malfunction.
The second potential source of poor performance is undesirable component characteristics.
For example, castings, forgings, and machined
or molded components may have discontinuities
or microstructural features present that make it
appear to the casual observer that there was a
heat treating problem. These discontinuities can
contribute to a poor heat treating outcome, a
poor service outcome, or both. The third potential trouble spot is design characteristics, many
of which may fall under the subclassification of
inadequate attention to detail. The fourth category of causes of failure in this scheme is true
heat treating process problems. Note that in this
classification, these are, so far, all things that can
go wrong in the engineering and manufacturing of the component. Finally, service and/or
maintenance abuse can be considered.
cannot be created or facilitated as a result of a

problem in the heat treating process.
Deformation
Figure 1 shows the result of a major deformation event. Note the image of a cylindrical
structure. There is a dark sign-wave-shaped
band where the material is crumpled. The weight
of the structure above that area created the force
that caused the cylinder to suddenly deform.
There were also some tears or fractures as well,
but the main visible damage type is deformation.
There are different ways to categorize deformation. One way is to compare gradual to sudden deformation. Gradual deformation can
occur when something is loaded and the load is
sustained. Due to the sustained load, the structure can stretch or bend. This type of deformation can happen during heat treating. Imagine a
part that has a protrusion that is not supported by
an appropriate fixture. The part is heated to red
heat. Depending on the particular configuration
and the presence of residual stresses in the
component, the protruding feature may droop
due to gravity, or it may change its shape in some
other way to relieve residual stresses from earlier parts of the processing. It is important to
consider this type of deformation in a component that has a critical dimensional envelope. If
the part goes outside of the specified or required
envelope and is placed into service, the stresses
that are experienced by the dimensionally nonconforming component may be quite different
from the original design intent.
Gradual onset deformation may also happen
in parts that are loaded near the elastic limit.
It is important to note that the published elastic limit for different materials is a numerical
quantity that is determined by using a specific set
of test parameters, usually of relatively short
duration. Long times at a stress level that may be
Types of Damage and Failure

The four basic types of damage (deformation,
fracture, wear, and corrosion) are briefly reviewed in this section, and discussions address
whether each of the four types of damage can or
Fig. 1
Example of a sudden deformation event due to

buckling
pg 88
21/8/2008 10:51AM Plate # 0
Mechanisms and Causes of Failures in Heat Treated Steel Parts / 89
considered to be below the yield strength

may actually cause unacceptable permanent
deformation. Since the yield strength or elastic
limit can be affected by heat treating, it is possible, although probably not a frequent cause,
that this type of service-related gradual onset
deformation is also related to a heat treating
problem. Gradual onset deformation could also
be a result of loading the part beyond its design
intent or actual physical limitations.
The alternative to gradual onset of deformation is sudden onset. Examples of sudden onset
deformation are buckling instability in compression of columns or torsion of tubes. A roof
truss that collapses under a snow load is such an
example. Roof trusses are not usually heat
treated for strength, at least not separately from
the thermal element of the hot rolling. However,
it is possible that annealed steel may be used in
some structural components. Most sudden onset
damage is primarily related to the basic geometry and modulus of elasticity, which is not a
strong function of any heat treating process.
Thus, heat treating problems are generally a
minor or insignificant factor in most sudden
onset damage events.
Another way to classify deformation is by
level or degree: elastic (in other words, if the
load is removed, deformation is relieved) or
plastic (which is permanent deformation). A
possible example of an elastic deformation
failure is a spring that does not have the correct
spring constant. Imagine a coil spring that is
supposed to stretch out 0.1 mm when subjected
to 12 N of force. What if the spring stretches out
0.05 or 0.2 mm? Can that type of failure be due
to heat treating? What are the causes or the
factors that allow these two types and two levels
of deformation to happen?
As with sudden onset, the two main factors
that control elastic deformation are geometry
and modulus of elasticity. These are not factors
that are greatly influenced by the heat treater.
Thus, the heat treater is usually not at fault in the
case of sudden onset buckling of a column,
sudden onset buckling of a tube in torsion, or for
elastic deformation failures. In general, sudden
onset and elastic failures are a result of the
combination of the design or actual geometry
and the elastic modulus being insufficient to
sustain the loading conditions. Sudden onset and
elastic deformation damage is only secondarily
related to the yield strength. The modulus of
elasticity is, in general, a constant. However,
modulus is not totally a constant in wrought
materials, where differences can occur due to

preferred crystal orientation. This is less likely
in heat treated steel than in steel used in the asworked condition. In general, in most heat
treated steels, the modulus of elasticity will vary
little. In attempting to determine the predominant factor or predominant cause in a sudden
onset deformation, one should look first at the
geometry and the applied loading. Heat treating
issues would be considered in puzzling cases
that did not lend themselves to ready analysis.
For gradual onset deformation, the geometry
and the elastic modulus are still important, but
the heat treating can have a much more significant effect. If the yield strength is not as high
as it is supposed to be, unacceptable levels of
gradual onset deformation may occur. This
concludes the introduction on how to think about
deformation failures and whether the heat treater
even needs to be involved.
Figure 1 shows what was reported to be a
sudden onset event. This is actually an example
of a complex failure or damage mode, because
this structure was standing for quite a few
years, and then one day, the wind blew and
the damage happened. Investigators found evidence of long-term corrosion on the inside,
associated with significant wall thinning in some
areas.
As another example, if a hollow tube for a
truck drive shaft is not heat treated properly,
could it buckle more readily? If the damage is
buckling deformation, the primary factor would
be the wall thickness and the modulus of elasticity. However, there may also be some strength
issues. When people talk about buckling due to
instability (or Eulers buckling, or what is
referred to as sudden onset), the main factors are
the geometry and the modulus of elasticity.
However, with geometrical configurations that
are not exactly like the extreme examples that
Euler used to develop his theory, one can
appreciate that the equations become very complicated and actually do have factors based on
strength values. However, the predominant
factors, in general, are the geometry and the
modulus. In failures of heat treated structural
steel, the modulus will be 30 million psi, with
some variations. For stainless steel, the modulus
may be a little different. If the component
material is a heavily cold-worked steel with
oriented, heavily textured microstructure, then
the modulus of elasticity may be different
in the different directions. However, most of
this orientation may be eliminated by any
pg 89
21/8/2008 10:51AM Plate # 0
subsequent heat treating that included an austenitization phase. Note again that this discussion
is strictly speaking about deformation in the
absence of a primary fracture event.
Before leaving the subject of deformation,
consider a small spring clip that was designed to
be used at the top of the linear portion of the
stress-strain curve. Why would anyone design a
spring to be used at a stress almost at the nominal
yield strength? No experienced metallurgical
engineer would expect each one of a quarter of a
million such spring clips to sustain multiple load
cycles to the theoretical elastic limit and not
have any permanent shape change. These clips
could not even sustain the minimum 15 loadcycle requirement without excessive permanent
deformation. This was a case of a nominally
elastic spring application where deformation
failure was caused by the heat treatment, specifically by normal variations of the heat treating
process. It would be difficult to blame the heat
treater in the absence of very strict qualitycontrol specifications. In this case, the major
cause of the failure was the design engineers
unrealistic expectations of uniformity of heat
treated components or the design process in the
company in which the designer was working.
Fracture
How can fractures of heat treated steel parts
be examined to determine the existence of any
factors related to the steel itself or its heat
treating? Basically, there are two types of
questions to ask. One of the things that people
like to ask is, Why did this one break? That is a
useful question in the case of a part that has been
used successfully for many years. Maybe there is
one part that failed out of a half-million parts that
are in service. Proper examination and evaluation of the physical evidence can reveal much
to answer this question. Before evaluating the
effects of the heat treatment itself, one must first
examine the physical condition of the damaged
property. It must be understood how the loads
interacted with the component to create new
surface area where none used to exist. The visual
appearance in three dimensions can reveal a
large amount of information on issues related to
how the part was really loaded. The colorations
and surface texture of the newly formed (undesired) crack surfaces can indicate how the crack
happened and how long it took for the crack
to grow.
The other question that is frequently asked is,

Why did this one break at this specific time?
This is a different question, and to answer it,
fracture mechanics type of explanations and
theories must be explored. In such a case, one
assumes that every structure has some small
discontinuity and a related crack growth rate,
which is a function of the stress intensity (measured in terms of the mathematical product of the
nominal component stress and the crack size)
and the fracture toughness. These are functions
of service condition and material parameters.
Particularly for parts that appear to have broken
in fatigue, (consider a two-year service life when
no similar component had previously cracked in
under five years), it may be informative to look
at the microstructure and how that may have
impacted the fracture toughness. Microstructure
and fracture toughness could definitely be related to heat treating issues. Other chapters in
this book give more information on fracture
mechanics.
For a more conceptual, lower-math-content
methodology to understand why something
cracked or why it broke, start by reviewing the
stress and strength variations that are at work in
the component. Any place where the local stress
exceeds the local strength can initiate a crack.
There can even be single grains that are low
strength for some reason. If the local strength is
lower than the local stress, then it is possible to
initiate a crack. Once a crack exists, it may or
may not progress to complete component fragmentation. Sometimes, cracks do not propagate.
However, cracks generally do not heal themselves, and they often do propagate. It is actually
reasonably straightforward to learn to look at
a component, if one has an understanding of
the loading geometry and the heat treating,
and determine if there was something wrong
with the component in question. Someone
inspecting a component with a complex shape
or complicated loading history may benefit
from a good finite element analysis. However,
many components do lend themselves to ready
evaluation regarding the presence or absence of
an extraordinary factor promoting premature
fracture.
To use the stress/strength distribution concept
to analyze fractures, one should be familiar with
the six basic loading geometries, including
tension, compression, bending, torsion, contact
stress, and shear. A review of some of the
loading geometries is presented in Ref 1; however, the three-dimensional characteristics of the
pg 90
21/8/2008 10:51AM Plate # 0
fragments are not systematically presented.

Armed with information on these basic fundamental loading geometries, one can learn to
predict macroscale fracture appearance in ductile and brittle materials given different loading
conditions, along with the more well-known
fracture surface features, such as beach marks,
ratchet marks, and chevrons.
This is a much more powerful method of
visual examination than simply interpreting
surface texture features. With practice, one can
look at a broken component and obtain a realistic
idea of what loading conditions actually caused
the fracture. Sometimes, it can be shown that the
real loading was very different from the design
loading.
Before leaving this subject, a review is needed
on details that are often poorly understood
relating to how to distinguish ductile and brittle
cracks on the macroscale. This is clearly very
important for a heat treater to know to defend
against incorrect accusations of embrittlement.
There are some types of macroscale ductile
cracks that can easily be misinterpreted as brittle
cracks. In fact, much of the published literature

is unclear on this issue. To understand this more
clearly, see Fig. 2. The image is of a broken
tensile bar. Near the fracture, the material is
necked down. Because of the visible shape
change, this is an obvious ductile fracture. The
image in Fig. 3 is of a threaded fastener. Based
on the ridge patterns in evidence, the crack
started at the root of one of the threads and went
back into the page. There is very little indication
of any deformation visible at the macroscale.
This is correctly called a macroscale brittle
fracture.
Figure 4 has more challenging fragments for
fracture analysis. The image shows a chain link,
which did not originally have an open shape. The
two protruding ends were touching; there was no
gap. The lower portion of the link did not have a
curved arc shape as depicted. A significant
amount of deformation was clearly associated
with this fracture event. Does that make this a
Fig. 2
Fig. 3
Example of a macroscale ductile fracture in tensile

loading
Example of a macroscale brittle fracture in tensile

loading
pg 91
21/8/2008 10:51AM Plate # 0
ductile fracture, or could this be a brittle fracture? Imagine that only this one photograph is
available, and it must be determined whether
this was a ductile or brittle crack.
Clearly, the steel that was used to make the
link has the capability of being deformed. In
other words, the steel is ductile at the conditions
present at the time of bending. The analyst must
not be satisfied with that answer, though. The
proper analysis includes a determination of
when the deformation happened. Postfracture
deformation does not make the crack event itself
ductile. The macroscale deformation must occur
during and as an inseparable part of the fracture
process for the crack itself to be a ductile crack
event.
For those unfamiliar with this methodology,
imagining that the link broke suddenly while it
was under load can be helpful. If the crack were
ductile, one would likely see some necking at the
crack location, since chain links are generally
loaded in tension. No such localized deformation is observed in this case. On closer examination, one can see a tiny shear lip at the top
edge of the fracture surface. That makes this
crack, for the most part, a brittle crack at the
macroscale, despite the presence of available
ductility in the material. The visible deformation
near the center of the lower (originally straight)
portion of the link happened after the crack was
completely formed.
In doing fracture analysis, it is important
to distinguish the capability of ductility in the
material from the behavior at the time of
the crack event. Despite the material ductility,
the crack happened in a brittle way. Closer
examination of other views not shown provide
clear evidence that this was a fatigue crack.
Beach marks were visible. Fatigue cracks grew
below the yield strength of the component,
creating macroscale brittle features. To clarify
one other potential source of confusion, it is
important to remember that tensile refers to a
loading geometry. Fatigue is a type of crack
Fig. 4
Example of a macroscale brittle fracture in tensile

loading
path. In this case, a fatigue crack grew due to a

tensile load in the horizontal portions of the link
(as shown).
To add a few more details to this case study,
the chain was in service at a plant that processes
meat, and strong acids were used to clean the
conveyor systems. This crack actually initiated
at a corrosion pit on the inside surface. The
cleaner reached the inner surfaces, but the employees may not have rinsed the chain very well.
This allowed a corrosion pit to form, which then
allowed a fatigue crack to grow. Again, it is
important to understand that the material itself is
ductile; there is nothing wrong with the material.
People involved with failure analysis need to
keep in mind that material behavior is or at
least may be different from material capability.
People doing failure analysis work need to be
able to distinguish inherent capability and actual
behavior.
To underscore the importance of separating
the behavior from the capability, imagine the
potential corrections that may be considered if
someone found this to be a ductile overload
fracture. The cure may be to make it harder. In
the case of the acid cleaning, harder steels are
often more susceptible to stress corrosion than
softer steels. If the harder-is-the-answer theory were put into practice, an undesirably short
life may become a horribly short life. It is
important to be sure that a crack that is diagnosed as ductile is really ductile and one that is
diagnosed as brittle is really brittle. Finally,
returning to the heat treating issues, the fact that
someone misdiagnosed this crack as a ductile
fracture may lead to the heat treater being
blamed for overtempering or inadequate hardening. In fact, until now, the potential blame or
innocence of the heat treater has not been
investigated in any thorough manner. It is possible that poor heat treating or poor material
manufacture contributed to the ease of corrosion
attack, and the meat processing plant employees
were blamed incorrectly. Further examination of
the microstructure is required to reveal the root
physical cause of the fracture and its timing.
Finally, there is the shaft fragment shown in
Fig. 5. The shape is cylindrical. The image
shows one fracture face. No necking or reduction in area is visible at the fracture face location.
However, this is a ductile fracture. It is necessary
to know that this shaft broke in torsional
loading. In torsion, the shear stresses are in
the transverse orientation to the length of the
shaft. To best understand macroscale ductile and
pg 92
21/8/2008 10:51AM Plate # 0
brittle fracture, one must be familiar with normal

and shear stresses. On a simple level, normal
stresses cause macrobrittle cracks due to crack
opening forces created at the crack tip. Shear
stresses allow slip and are the basis of the
deformation that creates the ductile crack event.
So, even though there is no necking, this is a
ductile crack.
Often, it is relatively easy to see some evidence of twisting on the side of a ground shaft
that has failed from ductile fracture by torsional
forces, which would lend more credibility to this
diagnosis of ductile fracture. However, this shaft
was extremely smooth, and it required a long
etching time in a heated acid solution to reveal
permanent twisting on the original cylindrical
surface.
Another possible source of confirmation that
this is a ductile crack is the classical smear
features on the fracture face. It has been argued
that these smeared features could be a result of
postfracture damage. While this is a possibility
that should be considered, since the background
evaluation revealed that the shaft was loaded in
torsion, and there are no crack opening stresses
operating on the transverse planes, this must be
a macroductile crack. A macrobrittle crack in
torsion is helical.
In closing this section on fracture, note how
important it is to follow the advice of the many
authors and teachers who state that background
research is step one of a failure analysis. The
Fig. 5
Example of a macroscale ductile fracture in torsional

loading
possible loading geometries that could have

created the fracture must be reviewed to make a
proper determination of whether or not the crack
is macroscale ductile or brittle. That determination cannot be made without assessing what the
loading geometries may have been. If this fragment had been totally covered with red rust, it
would have been even more difficult to determine the basic ductile or brittle behavior of the
material without knowledge of the loading
geometries and expected fragment shape.
Stress versus Strength. Almost all real
loading geometries cause the stress to be highest
somewhere along the part surface. If the strength
is uniform, for example, if there is a piece of hot
rolled 1050 steel that does not have any decarburization or carburization and has not been
shot peened, the crack initiation is expected
somewhere at the original part surface. In the
presence of any type of bending or torsional
stresses, the highest stress will be at the surface
of the part. In the presence of pure tensile
loading, theoretically the crack could start anywhere in the cross section. Such pure tensile
loading is rare. Imagine the case of a hydraulic
cylinder rod. Even here, there must be a section
change, a fillet, at some point. The loading at the
fillet is not uniform; there is a stress concentration. Even a tensile test coupon that is forced to
break in an area of nominally uniform strength
and stress is not totally uniformly loaded. Most
tensile test coupons are tapered so that the stress
is slightly higher at the center of the gage length.
This brings the discussion to what is so useful
about heat treating steel. Many types of steels
and heat treatments create harder or stronger
layers at the surface. Heat treating allows
the strength to be increased where it is useful.
Figure 6 is from Ref 1with annotations. This
figure shows a subsurface crack initiation along
the boundary between the induction-hardened
case and the softer core. A second initiation
appears to be inside the induction-hardened
case. It must be recalled that cracks can happen
whenever the local stress is higher than the local
strength. Normally, the stress is expected to be
highest at the surface, and the stress decreases
toward the center of the part or the center of the
cross section. In this case, the ratio of the stress
to the strength was higher below the surface than
it was at the surface, so the crack initiated subsurface. It is important to note such an unusual
situation, where the location of the ratio of the
stress to the strength was highest at a subsurface
position.
pg 93
21/8/2008 10:51AM Plate # 0
Fig. 6
Fracture features of an induction-hardened shaft (1541 steel) after fatigue testing in rotary bending. A, B, fracture origins.
Adapted from Ref 1, with annotations by W.T. Becker
Fig. 7
Stress and strength as a function of position in a

cylindrical component loaded in torsion. Fracture initiation may be at either the surface or subsurface. Subsurface
initiation depends strongly on the hardness profile from surface to
center if loading is in bending or torsion.
Imagine a simpler case of a cylindrical component (Fig. 7). The surface of the part is shown
along the left side of the graph, and the centerline is shown at right. For such a cylindrical
component that is loaded in either bending or
torsion, the stress will be highest at the
outside surface, and at the centerline, or neutral
axis, it will be nominally zero. A carburized or
induction-hardened material is actually stronger
at the surface layers where the stress is highest.
The y-axis, instead of being the stress level, can
be conceptually viewed as either the hardness or
some other measure of the strength of the component. If the part has a heavy case, then the
strength follows the dashed line. In this case,
high strength levels go in deep toward the core.
At some point, the strength and hardness drop
off to a lower level. In this situation, if the solid
line represents the stress and the dashed line
represents the strength, this part should not have
a subsurface crack initiation. Everywhere, the
stress is lower than the strength. If the case is
too thin for the application in question, and
the strength drops off as the dotted line shows,
the stress is higher than the strength within a
subsurface band, which allows subsurface crack
initiation.
This figure shows a powerful technique for
specification of case depths, which has the
potential to complement the usual experiential
method of case depth specification. Anyone
doing fracture analysis on a case-hardened part
can also use this information to obtain an idea
about the appropriateness of the hardening specification.
It is important to realize that there is one
other case where a subsurface initiation may
occur that is not related to the heat treating or
specification quality. Imagine the presence of
a subsurface discontinuity, such as an inclusion,
a void, or a tiny crack. Even if the steel has
pg 94
21/8/2008 10:51AM Plate # 0
Stress and strength as a function of position in a

cylindrical component loaded in torsion with subsurface discontinuities. Surface conditions may include: inadvertent
decarburization, typically thin and may not be easy to find; deep
case from induction or carburization; nitrided, thin case, often
not more than 510 mils. Part of the case may be ground off in the
finishing operations. The defect could be a faceted inclusion
(nitride) in a low-ductility matrix. Nondeformable nitride causes
stress concentration in the matrix.
or other rolling elements and races.

Understanding that most loading geometries
create the highest stresses at the surface also
allows one to understand why decarburized
layers can be so damaging. A decarburized layer
is softer and lower in strength than the material
with the desired nominal amount of carbon.
Decarburization can even occur on carburized
steel. Figure 9 shows a metallographic cross
section of a piece of steel that is carburized and
has quite a bit of retained austenite. Note the
dark constituent at the surface (arrows). On
carbon steels in the medium-carbon range, decarburization usually looks white, but here it
looks dark due to the presence of pearlite. The
decarburization affected the hardenability as
well as the hardness in this case. There was not
enough carbon to form martensite at the surface
when the part was quenched in heat treatment.
Very fine pearlite was formed instead. The
pearlite structure is not as strong or fracture
resistant as the martensite structure that is
expected in the absence of the decarburization.
Thus, this part could be more susceptible to
fracture because of the decarburization during
heat treatment.
To complete the discussion of fracture, the
previous is summarized by emphasizing that the
macroscale features reveal the loading conditions. Fracture analysts must start with the
macroscale, or the big picture. Many people start
with the details, or the little picture, and move on
to the big picture, but this can be a problem that
facilitates mistakes on the part of the analyst.
Microscale Fracture Features. Scanning
electron microscopy can be used to reveal the
microscale fracture features. Figure 10 shows a
Fig. 9
Fig. 10
constant strength all the way through its cross

section, an inclusion that is big enough to locally
increase the stress above the fracture strength of
the part makes it possible to create a subsurface
initiation (Fig. 8).
Another potential cause of subsurface fracture
initiation is contact loading. A well-lubricated
bearing without any friction has the highest
stresses in its subsurface layers. Thus, fracture
initiation will be at a subsurface location where
contact loads are the predominant source of
stress. Inclusions can thus be very damaging in
contact-stress applications, such as bearing balls
Fig. 8
Decarburization of carburized steel. Each small scale

division is 2 mm.
Scanning electron micrograph of microvoid coalescence
pg 95
21/8/2008 10:51AM Plate # 0
ductile crack path, or what is often called ductile

dimples, microvoids, or microvoid coalescence
(MVC). The MVC is a characteristic fracture
morphology that indicates a component was
subjected to stresses in excess of the nominal
ultimate tensile strength of the component
material.
The MVC is often an indication that the heat
treated steel did not have a gross problem with
the heat treating that caused or contributed to an
embrittlement problem. Brittle fracture features,
including cleavage and (most) intergranular
features, are often indicative of a heat treating
problem. The arrows in Fig. 10 show the nonmetallic inclusions that initiated the void formation.
Brittle fractures are often unexpected and
occur suddenly without any prior warning.
Ductile fracture by MVC is typically accompanied by prior plastic deformation, which gives
advanced warning of the impending fracture
event. This prior warning makes MVC the preferred mechanism if fracture occurs.
While MVC is generally desirable, it can
indicate that the material is too soft if a highstrength material is in question. The MVC can
also reveal that the heat treater made a mistake,
such as no heat treatment, despite the often
preferred MVC fracture path.
Classical microscale brittle crack paths are
along the grain boundaries (intergranular),
(Fig. 11) or cause the grains themselves to split
(transgranular or cleavage). Many heat treating
and other processing problems can cause undesirable intergranular cracking at the microscale.
To have intergranular cracking, something may
either cause a low-strength condition at the grain
boundaries or cause the stresses at the grain
boundaries to be higher than in the core of the
grains. Refer to the previous concept regarding
the relationship between the local stress and the
local strength.
One mechanism where intergranular fracture
occurs is quench cracking. When a piece of steel
is quenched to harden it, martensite will generally start forming near the surface, because that
is where it cools off fastest. As anyone familiar
with heat treating of steel knows, each deeper
layer of grains will subsequently transform to
martensite. On transforming to martensite, the
material expands. As the material continues
to cool, it contracts. So, there are grains that
undergo expansion during transformation
while the grains next to them contract. Could
that create a shear stress at the grain-boundary
location? The situation is obviously more complicated than just described; obviously, the grainboundary strength is a function of temperature,
which is rapidly changing, but it is helpful to
think about what could be causing that intergranular crack and why quench cracks are generally intergranular.
Note also that carburized steel often will have
intergranular cracking in fatigue. Sometimes,
even the best metallography cannot show any
problem at the grain boundaries, that is, no
grain-boundary carbides, oxides, nitrides, porosity, and so on. It is true that there are many
heat treating problems that can facilitate, or be a
factor in, intergranular cracking at lower stress
levels than the part usually sustains. However, it
is important to note that just because there is an
intergranular crack, it does not mean there is
surely a heat treating problem.
Hydrogen embrittlement is often revealed in
part by its intergranular crack habit, particularly
in steels heat treated to high strength levels.
Hydrogen embrittlement is not always purely
intergranular; sometimes, there will be tiny,
shallow microvoids on the grain-boundary surfaces. Hydrogen embrittlement could be the heat
treaters fault (if the hardness is too high), and it
could be an issue with the plating (if records do
not show proper baking); it could be an interaction between these two factors.
The important point is not to confuse identification of the crack path with the cause. There
is a difference between identification of the
physical shape of the crack and the physical
characteristics relating to the crack event. These
do not automatically lead to the cause.
Fig. 11
Scanning electron micrograph of intergranular

cracking
pg 96
21/8/2008 10:51AM Plate # 0
The last microscale crack path is cleavage,

and Fig. 12 shows a classical view. Ferrite
cleaves readily at low temperatures. If a part is
not supposed to have any ferrite in it and there is
a large amount of cleavage, then that may be a
clue to look carefully for ferrite during microstructural analysis. The classical way to recognize cleavage is the presence of patterns that
look like riverbeds with multiple tributaries.
The arrows in Fig. 12 show these river line
features.
In closing this section, fatigue microscale
features in heat treated steels are often not very
interesting or classical. The experienced analyst
can recognize them, but they are difficult to
describe. They rarely have the textbook striation
features that are commonly shown for superalloys or aluminum alloys in published microfractographs. Striations may be visible in a
low-carbon annealed steel, particularly in steels
with ferrite as part of the microstructure.
Beginners must be careful to distinguish pearlite
platelets from striations. Pearlite stops at grain
boundaries, and striations may cross grain
boundaries. If there are questions, pearlite spacing can be examined on a cross section at a later
time in an effort to distinguish one from the
other. It is quite rare to find striations in any kind
of hardened steel. Beach marks are often visible
at the macroscale, but striations are very uncommon as microfractographic features.
Summarizing the differences between macroand microscale features: The macrofeatures
show the loading geometry. The microfeatures
show the result of the microstructural interaction
with the environment, and mechanical and chemical aspects may influence the way the crack
interacts with the microstructure.
Fig. 12
Wear
The original shape of the object shown in
Fig. 13 was a gear with normal-shaped teeth. It is
severely worn. No judgments can be made about
the cause with this one image.
Wear has many similarities to fracture and
deformation. Wear is basically deformation and
fracture going on at a microscale, and it can
continue until the point that macroscale damage
is present. Scanning electron microscopy (SEM)
is very helpful in understanding how wear
happens. Wear specialists have identified many
different wear mechanisms. However, even
without that specialized knowledge, the SEM
can reveal useful information for diagnosis and
prevention. In Fig. 14, there is smeared material.
This solid steel has now flaked and smeared to
the point that it is present as thin platelets, which
are breaking off and allowing material loss. This
is one example of a combination of deformation
and fracture.
Fretting is a common type of wear that is
almost never related to heat treating or any
problem with heat treating but rather is related to
the geometry of the assembled parts. Fretting in
steel will generally produce a reddish, iron oxide
powder, and it roughens the surface. Figure 15
is an SEM image that just barely reveals
the initiation of a crack. Fretting often produces
a crack in an area that is thought to have
low stress.
Contact forces can cause surface damage due
to the action of Hertzian stresses. Figure 16
Fig. 13
Scanning electron micrograph of cleavage cracking
Cross section of worn gear teeth. Approximate width

of steel segment shown is 23 mm (0.9 in.).
pg 97
21/8/2008 10:51AM Plate # 0
shows a bearing race and a bearing ball. The ball

is being pressed into the race. This figure is
a rough schematic, conceptually showing an
exaggerated view of the most highly stressed
area due to the elastic deformation of the two
components. Without the load, there is essentially a point. Under load, the contact area
becomes circular or elliptical; in other words,
there is a contact footprint. The actual highstress location is just below the surface. In the
presence of friction, the highest stresses are
moved toward the surface. An important part of
bearing wear failures is determining whether the
crack initiated at the surface, showing the possibility of a lubrication issue, or if it was truly a
subsurface initiation, in which case it may be a
microstructure problem, an inclusion, or another
anomaly. Bearings are generally loaded to very
high stress levels. The quality of the steel and, in
particular, minimization of inclusions are very
important in this type of application.
Another problem is that can cause surfaceinitiated cracks grinder burn, which can create
high tensile stresses at the surface. Regardless of
what the service load is, a stress field has been
created with a very high tensile stress at the
surface, the most undesirable location.
To close this section on mechanical damage,
it is often important to determine if there was a
service problem, such as abuse, or misuse, such
as using a screwdriver as a pry. Understanding
loading geometries and related fragment shapes
can shed light on this type of question.
Figure 17 shows a steel bar with a threaded
portion that is much smaller in diameter than the
rest. Note that the crack is in the large-diameter
portion. Could this have been caused by

the user?
To answer the question, one would have to do
a large amount of background information collection. The important point to realize is that the
crack location is totally unexpected, and it is
difficult to think of something that could have
happened in service to create a weak spot at this
hefty location.
If it cannot be qualitatively demonstrated that
somebody abused something, then it must be
quantitatively demonstrated. This is often difficult. It is important to know what the material
strength, fracture toughness, and other material
properties were at the time of the damage. It is
important to determine what the load or loading
geometry was that caused the damage. One must
Fig. 15
Scanning electron micrograph of fatigue crack initiating on worn carbonitrided steel. Original magnification: approximately 4000
Race
Bearing ball
Stress
concentrations
Fig. 14
Scanning electron micrograph of a worn piece of

hardened medium-carbon plate showing details of
the wear mechanism
Fig. 16
Rough schematic of stresses in contact loading of a

bearing ball on a race
pg 98
21/8/2008 10:51AM Plate # 0
consider how the suspected loading geometry

differs from the actual design intent. The analyst
again must remember that just because it is
broken does not mean it was abused. Even if it
meets the specification, it does not prove that the
user abused the product.
Corrosion and Environmental Damage
Figure 18 shows an example of a 300-series
stainless steel that was probably not heat treated
in the most ideal manner, since small precipitates can be seen along the grain boundaries.
Note the crack location, which seems to be
seeking the grain boundaries. To create a crack,
there must be a load or a stress. There is corrosion present too, so this may be a stresscorrosion crack. Did the nonideal heat treating
condition cause the crack? In this case, a chain
of reasoning cannot directly link the specific
heat treating problem with the crack, because

there is no information about the stress
levels. Stress-corrosion cracking, by definition,
requires a threshold level of corrosion and
stress. In this case, there is no evidence that the
crack was caused by bad heat treatment. This
seemingly subtle distinction may be very
important in the case of a catastrophic failure
event.
Figure 19 shows an example that involves
another stainless steel weld. Welding is a kind
of heat treatment, although not as controlled as
an intentional heat treatment. An acid substance, polythionic acid, was in contact with
the weldment. There are some cracks on one
side of the weld, while the other side is free of
cracks. What kind of cracks are these? The
crack path is intergranular (Fig. 20). The
micrograph in Fig. 21 was taken after an
ASTM International test, and it shows ditching
Fig. 18
Grain-boundary precipitates in a 300-series stainless

steel
Fig. 19
Stainless steel weldment
Fig. 17
Steel bar with crack in unexpected position. Originally shock loaded in compression. Threaded
portion diameter is approximately 2 cm.
pg 99
21/8/2008 10:51AM Plate # 0
characteristics. This indicates that whatever the

heat sequence from the weld, it did create this
condition that appears to have facilitated
intergranular corrosion. However, steel would
have been considered to pass the test for freedom from sensitization. Despite this, the thermal experience of the material can be directly
linked to the form of the crack. Although this is
not a heat treating example, this demonstrates
the line of reasoning that is required to determine cause. It is interesting to note that one
side cracked and the other side of the weld did
not. It is likely that the side that did not crack
was an L-series, a low-carbon series, specifically made to minimize the chance of cracking
in weld heat-affected zones in stainless
steels. The cracked material was probably not
an L-series. This is an example of a classical
Fig. 20
Scanning electron micrograph of stainless steel

weldment with intergranular cracking
Fig. 21
ASTM International sensitization test results showing

ditching characteristics
damage mechanism, stress-corrosion cracking.

When the regular 300-series stainless heated
up, the chromium and carbon combined and
precipitated along the grain boundaries, taking
the chromium out of solution and making the
material less corrosion resistant. Frequently,
such situations lead to pit formation, which
then allows the crack to propagate from the
stress concentration at the pit. This is an
example of a complex damage mechanism.
Another commonly named damage mechanism is corrosion fatigue. It must be understood that a corroded part that broke due to
repeated crack extension under load did not
necessarily experience the mechanism called
corrosion fatigue. Corrosion fatigue is a damage mechanism that is studied in the laboratory.
The named mechanism is invoked when it can
be demonstrated quantitatively that the crack is
growing much faster under the same loading
conditions than it would in the absence of the
corrosive substance. In a real component out in
the field, such as a heavy off-road vehicle
application, it is very difficult to obtain an
accurate service history day-by-day. Going
back to published research data for standard
test coupons and proving that a particular
situation is or is not corrosion fatigue will
likely prove very difficult. In Fig. 22, a shaft is
Fig. 22
Cracked shaft used in a corrosive environment.

Diameter of the shaft is approximately 10 cm (4 in.).
pg 100
21/8/2008 10:51AM Plate # 0
obviously corroded. The band highlighted

by the arrows was covered with a layer of red
rust. Figure 23 shows the crack created by the
corrosion and stress combination. Note that
there are also secondary cracks. The tip of this
crack looks like a witchs broom or the trails of
a sparkler. This is not a normal fatigue crack. A
normal fatigue crack does not branch out and
have multiple tips. In this case, the corrosion
definitely has some kind of significant impact
on the damage mechanism. In many cases,
examination of field failures is likely to leave a
question mark regarding the quantitative evaluation of how much faster the crack is growing
because of the corrosion.
To demonstrate that corrosion is the cause
or the fault of a failure, it is not enough simply
to say the component was in a corrosive environment and it cracked because of that. It must
be assured that the specific conditions, the specific material, and the specific process condition
of the material were in the realm that has been
demonstrated to be a problem for the damage
mechanism invoked.
For example, concentration of an aggressive
substance, the threshold stress level, and the
temperature may be required to be in a restricted
range before a particular mechanism can be
properly said to have been acting.
Fig. 23
Factors Contributing to Poor Response

from Heat Treatment
Raw Material Characteristics That
Can Contribute to Poor Response
from Heat Treatment
What are the raw material characteristics that
can contribute to poor heat treating outcome?
One very important characteristic is composition. There will be a range of values for each
type of atom that is specified for the grade in
question, as well as for unspecified elements.
A heat treater may receive material of an iron
matrix that could, with the same name, have a
very wide range of responses to the heat, heating
rate, heating dwell time, cooling rate, and so on.
Lean and rich alloy content can have a strong
influence on whether or not quench cracks occur.
Lean and rich compositions also strongly influence how readily the hardness specification is
attained. For a hardness specification that is
toward the upper limit of what can be reliably
obtained for a particular grade, it may be difficult to meet the specification for a given lot if all
the elements are on the lean end. If the part(s)
crack on quenching, an important task in the
troubleshooting process is to determine whether
it is due to a lean or rich (more likely) alloy
Crack profile and adjacent secondary crack tip. Original magnification: 50 . Inset is of a different secondary crack tip.
Original magnification: 500
pg 101
21/8/2008 10:51AM Plate # 0
composition that is facilitating martensitic transformation. This must be differentiated from the
situation where the major cause of the crack is
that the heat treater quenched the part(s) more
severely than usual.
It is difficult to distinguish quench cracks due
to extreme quench severity from those due to
rich composition. However, it is easier to determine whether the crack has an expected
geometry for a quench crack. In Fig. 24, the
crack is at the section change, which is a prime
location for a quench crack, as are sharp corners.
When examining a part to determine if rich
alloy/quench severity issues were at stake,
examine the crack surfaces for traces of temper
colors, including blues and browns. A dark
matte gray or black surface may indicate the
presence of an oxide-filled discontinuity that
simply opened due to thermal stresses rather
than a rich alloy/quench severity problem.
If temper colors are visible at the portion of
the crack surface nearest the part surface, it is
likely that one or more of the following factors
was present:

Quench was more severe than was appropriate

Alloy was richer than usual
Section change was more severe (for example, the fillet radius was too sharp)
To confirm the suspicion of a pure quench

severity/rich composition-related quench crack,
it is advisable to confirm that the microscale
crack path is intergranular (usually SEM is
required). If a seam had been found at the
quench-opened crack, it would not be correct to
blame the heat treatment. Figure 25 shows an
interesting crack. There is a very heavy oxide
layer revealed by the cross section. The inset
shows a higher-magnification view of the seam
detail. There is a rounded particle that is totally
covered by a heavy, rounded oxide layer. The
crack surface does not have blue or brown
temper colors but was found to be a dark charcoal black. This type of heavy oxide is unlikely
to have happened between when the part was
quenched (and presumably cracked) and when it
exited the temper furnace. Another unusual
feature of this crack is that it changed direction
multiple times (large arrows show crack growth
direction changes). The fact that these are not
predominantly intergranular cracks deals the
final blow to any theory stating that this crack
was due to a heat treating problem.
Some people wonder why it was not possible
to see the seam before it went into the heat
treating process. Many seams are tightly closed
or smeared over until the part experiences the
stresses of the heat treating operation. Limitations of nondestructive testing methodology
may also play a role. In the case of a part with a
quench crack in a location that would not be
expected to have high stresses in quenching due
to differential cooling rates, the experienced
Fig. 24
Quench crack with typical geometry
Fig. 25
Oxide layer along a seam most likely present in the raw material. Original magnification: 50 . Inset original magnification:
200
pg 102
21/8/2008 10:51AM Plate # 0
analyst looks for some type of discontinuity. The

discontinuity does not have to be very thick. This
example was particularly heavy, but even a very
thin oxide layer is enough to be the predominant
cause of the crack.
Macro- or microsegregation (otherwise
known as banding) are other raw material characteristics that can interfere with the expected
outcome of a heat treating process. Figure 26
is a longitudinal cross section from a mediumcarbon piece of steel. If the low-carbon layer is
right at the surface, it may be difficult to meet a
minimum hardness specification.
Banding, or microsegregation, is not always
bad. It can make it easier for the crack to grow in
Fig. 26
Longitudinal cross section showing microsegregation. Original steel segment shown is approximately
2 cm wide.
a particular direction, and that may be an

advantage for a particular application. Samurai
and Damascus swords from antiquity had basically banded microstructures with very desirable
characteristics. While banding is not necessarily
bad, it can cause some variation in the response
to the heat treatment.
If a material has intermittent coarse grains, it
may be easier to form martensite in the large
grains and pearlite, ferrite, or bainite in the
surrounding fine grains. Can the heat treater
create such a grain size distribution? This may
be possible by overheating.
However, it is also possible that the coarsening came from a subcritical amount of cold work
stored in the material. Normalizing the steel
prior to hardening may eliminate the nonuniform response to hardening, but the added cost
of normalization is often objectionable (Fig. 27).
Another type of raw material characteristic
that can cause problems in some applications is
heavy bands of stringers. Figure 28 shows a
piece of steel that has long sulfide stringers in it,
which can act like a seam.
Figure 29 is of a wire product that was used in
a coil spring. The material was subject to torsional loading. A longitudinal discontinuity in a
material that is subject to torsional loading can
create very high local shear stresses in the
longitudinal or radial directions. This is one
reason that seams are not allowed in critical
applications for spring wires. Decarburization is
not detected on the surface, but there is a heavy
decarburized layer on either side of an inclusion
of the shape characteristic of a seam. The heat
treatment is unlikely to create such a varying
thickness layer of oxide.
Fig. 27
A few coarse grains in the core of a fine-grained

material that has been carburized are the only portion of the core able to form martensite. Original magnification:
100
Fig. 28
Scanning electron micrograph of sulfide stringers in a

piece of bar stock
pg 103
21/8/2008 10:51AM Plate # 0
pg 104
Component Characteristics
Figure 30 shows a cast steel product that was
quite uniformly carburized, except for the white
script features. The pattern looks very similar to
microsegregation between the dendrites and the
matrix. It would be very difficult for the heat to
overcome the initial segregation in the raw
material. This component may not perform as
well as a component with a uniform case. These
Fig. 29
Optical micrograph of an oxide-lined seam in a

piece of steel wire
10
script features may make the case even more

brittle than usual. The heat treater can control the
carbon potential and the heating cycle. However, the heat treater cannot locally put fewer
carbon atoms into the steel at the locations that
already have too many. When performing failure analysis of steel components, analysts must
Fig. 30
20
Unusual crack
shape
Oxide filled crack
Evidence of
pre-existing
discontinuities
Fig. 31
Thread root of a steel fastener. Original magnification: 100
30
Cast steel after carburizing. Original magnification:

100
40
Dark etching
area possible
indication of
heat treating
problem
Note multiple
cracks in one
thread root
21/8/2008 10:51AM Plate # 0
always be on the lookout for the whole history of

the part that led up to the heat treating event.
Forging discontinuities are another situation
that could cause a problem. Figure 31 shows a
cross section of a threaded fastener with a locally
different compositional steel inclusion. Something unusual happened to create this feature.
Other problems relating to poor heat treating
or poor service outcome include many different
types of design details. Other chapters in this
book address these issues in greater detail. To
summarize, some areas of common oversight
include:

Materials selection
Heat treating process selection
Hardness level specification and range and
position on part for hardness test
Process details (batch or continuous
oven, etc.)
Heat treaters familiarity with the size and
complexity of the part and the quality level
needed
Distortion control
Fig. 32

Testing competence
Hardness scale selection
Frequency of part testing
Figure 32 shows two micrographs at the same

magnification from the same component,
25 mm (1 in.) apart from each other. One of
them is virtually all martensite. In the other
location, there are wide grains of pearlite interspersed between the martensite grains. There
was as much as 12 Rockwell C points difference
between the two microstructures. The designer
never indicated where to test the component!
This is a typical design issue.
Another common cause of disputes between
purchasers and providers of heat treating services is that designers specify Rockwell tests
when there is no way to perform anything but a
Knoop or Vickers test. There are still many
newly minted engineering prints with case
depth specifications that are very unclear. There
are standard methods for specifying carburized,
induction-hardened, or carbonitrided case
depths, and it is helpful to use an industry
Medium-carbon steel microstructures from the same component at two locations separated by approximately 25 mm (1 in.).
Each small scale division is 5 mm.
pg 105
21/8/2008 10:51AM Plate # 0
standard when possible. The automotive manufacturers have done a good job of providing a
range of methods at a range of ease of testing. If
a standard method is not being used, it is often
difficult to determine the designers intent.
The specifications for 400-series stainless
steel can be particularly difficult for the average
mechanical designer to write. Many designers
specify 400-series stainless steel because they
want stainless, but they want to be able to heat
treat it to obtain higher hardness than standard
300-series annealed bar stock. For a number of
the 400-series grades, one must determine in
advance whether maximum strength or maximum corrosion resistance is desired. Two
totally different heat treating processes attain
those goals. The heat treater has no means of
guessing which characteristic is required.
What Are the Things That Can Go Wrong
in the Heat Treating Process?
This chapter has attempted to cover all the
aspects for which the heat treater should not be
blamed. What are the aspects for which the heat
treater may or should bear responsibility?
One approach is to say that heat treaters must
take responsibility for those aspects that are
specific to their equipment. These are details that
Fig. 33
only the heat treater could know. Design engineers specify materials and thus need to know
how to specify the type of testing and evaluation
required for application. A design engineer
cannot know how fast a load of parts will be
heated in a particular companys individual
furnace. The design engineer cannot be expected
to know what type of fixturing may be necessary
to maintain required distortion levels in the
part. These are aspects the heat treaters must
know. The heat treaters must know what load
size can be treated in their own furnaces and how
the load should be distributed. The heat treaters
must understand the characteristics of the
interactions between their equipment and the
full range of part sizes and load sizes they are
processing.
Heat Treating Errors. Excessive heating
rate, excessive time at temperature, and excessive temperature can lead to excessive distortion. Excessive temperature can cause problems
with excessive autotempering. If a massive part
is heated to 75, 100, or more degrees hotter than
it needs to be to obtain the uniform austenite
required before quench, then all the extra
heat must be removed, which can make it difficult to create the desired martensite at all. Furthermore, the extra heat may act to partially
temper the martensite that is present. Excessive
Undissolved ferrite and martensite in improperly specified and improperly induction-hardened medium-carbon steel part.
Each small scale division is 5 mm.
pg 106
21/8/2008 10:51AM Plate # 0
heating can cause trouble in obtaining minimum

specified as-quenched hardness values. These
low as-quenched hardness characteristics may
or may not make a difference in the component
performance. However, when as-quenched
hardness tests are required, it is important to
know that excessive temperature may be a
cause.
Inadequate Heating Rate, Inadequate
Time at Temperature, and Inadequate Temperature. Figure 33 shows the microstructure
of a part that was supposed to be induction
hardened. There is some martensite. The light
constituent is ferrite that never went into solution. The initial microstructure was probably a
mixture of ferrite and pearlite. The pearlite
transformed into martensite, but little of the
ferrite did. On quenching, the material produced
islands of martensite with islands of ferrite. The
irregular shapes of the ferrite islands are classical undissolved ferrite. This is not a typical
shape of ferrite grains formed on cooling.
Because of the severity of the consequences,
most of the common problems in heat treating
are related to hardening. Annealing can also be
done incorrectly. Figure 34 shows the microstructure of a steel that was supposed to be
spheroidized annealed. Spheroidization is a process that may take 12, 15, or even 20 h at 600 to
700 C. Spheroidized annealed steel is generally
quite expensive. Its applications are usually
reserved for severe forming operations where
Fig. 34
Incompletely spheroidized annealed steel. Each

small scale division is 2 mm.
the added ductility is a necessity. The annealing

line management may be tempted to cut down
the process time to save money. The minimum
spheroidization that the material processor
believes will work is what is often provided.
In the case of varying incoming microstructures,
the part may not be as ductile as usual. Examine
the situation where the forming process creates a
crack that was undetected, and then the part is
heat treated for hardening. If the crack remains
undetected, there is now a part with a discontinuity due to the spheroidizing being done
poorly. This type of situation may be difficult to
figure out, especially if the failure happens some
years after the fact.
Insufficient time or temperature could apply
to formed parts requiring stress relief. Even at
relatively low hardness values, excessive residual stresses can make the part sensitive to
hydrogen embrittlement. A low-carbon steel
part that has been heavily deformed and
improperly stress relieved can crack after a very
short service life or even while sitting on a storage shelf. Stress relief is often used on weldments, and if it is not done properly, fatigue
cracks can initiate more readily at weld toes.
Machined parts that are improperly stress
relieved can distort or crack at a later time,
because the stresses are higher than one may
think.
Cooled Too Fast. The part that is cooled too
fast due to cold quenchant or excessive quenchant agitation may crack or suffer excessive
distortion. Undesired microstructures may also
result from excessively fast cooling. Bainite
may be desired, but the process formed martensite.
Cooled It Too Slowly or Cooled to the
Wrong Temperature. This can be a result of a
delay in moving the parts into the quench tank.
Alternatively, the composition of the polymer
quench tank may be improperly maintained.
Slow cooling can be a problem, because the
crack resistance of a microstructure with excess
ferrite may be lower than a properly hardened
and tempered martensitic structure.
Improper Atmosphere. Decarburization can
result from low carbon potential in the atmosphere surrounding the parts. Carburization
can occur if there is too much carbon in the
atmosphere. Retained austenite in undesirable
amounts can also result from excessively rich
carbon in the atmosphere. There are still heat
treaters in business who believe that it is good to
use some ammonia when carburizing, even if the
pg 107
21/8/2008 10:51AM Plate # 0
customer did not ask for carbonitriding. These

heat treaters may believe that the customer is
happy to get the lower price due to a faster
process that meets a surface hardness specification. The problem is that if the stress state
requires a certain level of strength at a certain
depth, the faster ammonia-enhanced process
may be inadequate.
Retained austenite in greater amounts than
normally found in parts that are straight
carburized can also be a problem. Porosity can
also be created in the case with nitrogen atmospheres. This can be a problem if the surface
hardening is desired for strength rather than just
scratch resistance.
Figure 35 shows an example of bad carburizing. There is a significant fraction of retained
austenite in this case-hardened part, as well as a
large, chunky, unusually shaped puzzle piece
carbide. This feature could be a problem for
some applications of carburized parts and may
be the result of excessive carbon potential.
Concluding Comments
For those who do failure analysis of heat
treated steel parts, it is important to understand
what microstructure is expected, given the heat
treating process that is specified. For example, in
10
20
a 1050 steel that is 100 mm in diameter and

100 mm long and water quenched, should it be
martensite? There probably will be some martensite near the surface, but it may not be a very
deep layer. The exact depth will depend on how
hot it was heated prior to quench, the details of
the quench tank design, and many other factors
that may not be readily apparent.
It is important to have a large amount of
reasonably deep knowledge to be able to make a
fair and correct determination of where there
may have been a problem in the entire process of
designing a part, procuring material, and making
a component. This knowledge base includes
failure analysis, fracture analysis, and microstructure analysis and interpretation in order to
read the process history. Simply checking the
hardness and the composition to see if they meet
the specification is not failure analysis. Failure
analysis includes a determination of the loading
geometry and background information, at the
very least, in addition to the basic certification
conformance tests.
Some people legitimately perform internal
process failure analysis on a part that never left
the door of the manufacturing plant. In failure
analysis of a field return, even from a non-enduser assembly problem, it is important to do
more than simply look at the composition and
the hardness. The type of damage needs to be
identified, as well as the possible sources of that
30
40
50
Fracture #2 has both large chunky and script carbides

and large amounts of retained austenite.
Fig. 35
Puzzle piece carbide microstructure in carburized steel, possibly due to excessively high carbon potential. Each small
scale division is 2 mm.
pg 108
21/8/2008 10:51AM Plate # 0
type of damage. A review of the comprehensiveness of the design process may be in order. If
recurrence prevention is a goal of the failure
analysis, the damage specialist may lend some
understanding to the design engineers to help
them clarify the requirements of the component
characteristics.
Failure analysis can be very routine, or it can
be extremely involved. This chapter has considered only a few categories of the analysis
procedures and some of the reasoning involved
in determining what went wrong and at what part
of the life cycle the problem initiated.
A decision must be made at the beginning
of the analysis about how detailed the project
will be. If there is a single component, especially, or a very limited number of failed
parts, inadequate planning can leave inadequate
specimen material for testing in the case of
unanswered questions at the end of the project. It

is difficult to overemphasize the importance of
spending enough time initially figuring out
exactly what the goals of the failure analysis
project are and how much detail is required.
ACKNOWLEDGMENTS
The author thanks Mrs. W.T. Becker for permission
to use the copyrighted material of William T. Becker
in Fig. 6 to 8.
REFERENCE
1. D. Wulpi, Understanding How Components

Fail, American Society for Metals, 1985
pg 109
18/8/2008 3:02PM Plate # 0

DOI: 10.1361/faht2008p111

General Aspects of Failure Analysis

Waldek Wladimir Bose-Filho and Jose Ricardo Tarpani,
Universidade de Sao Paulo
Marcelo Tadeu Milan, Instituto de Materiais Tecnologicos do
Brasil Ltda.
FAILURE ANALYSIS is the process of

collecting, examining, and interpreting damage
evidence. The objective is to understand the
possible conditions leading to a failure and
perhaps prevent similar failures in the future.
A failure analysis should provide a welldocumented chain of evidence that either
excludes or supports possible interpretation of
the damage evidence. Clear-cut conclusions
do not always occur, and the tendency of
developing preconceived interpretations should
be avoided.
Various publications (e.g., Ref 16) describe
the guidelines and methods of failure analysis,
and this chapter briefly outlines some of the
basic aspects of failure analysis. The first section
describes some of the basic steps and major
concerns in conducting a failure analysis. This
is followed by a brief review of failure types
from fracture, distortion, wear, and corrosion.
Fracture is a common damage feature, because
the vast majority of mechanical failures involve
crack propagationtypically classified as ductile, brittle, and fatigue, as briefly described
in more detail. Distortion, wear, and corrosion
also can be important damage factors in failure
analysis.
General Guidelines of Failure Analysis

For a complete evaluation, the sequence of
stages in the investigation and analysis of failure, as detailed in Ref 5, is as follows (Ref 2):
1. Collection of background data and selection
of samples
2. Preliminary examination of the failed part
3. Nondestructive and mechanical testing
4. Selection, identification, preservation, and/
or cleaning of specimens
5. Macroscopic examination and analysis and

photographic documentation
6. Microscopic examination and analysis
7. Selection, preparation, examination, and
analysis of metallographic specimens
8. Determination of failure mechanism
9. Chemical analysis
10. Fracture mechanics analysis
11. Testing under simulated service conditions
12. Analysis of all the evidence, formulation of
conclusions, and writing the report
These stages or steps are briefly outlined as
follows.
Collection of Background Data and
Selection of Samples. There are basically
three fundamental principles to be carefully
followed when collecting damage evidence
from a fractured material (Ref 2):
Locate the origin(s) of the fracture. The

whole fracture surface should be visually
inspected to identify the location of the
fracture-initiating site(s) and to isolate the
areas in the region of crack initiation that
will be most fruitful for further microanalysis. Where the size of the failed part
permits, visual examination should be conducted with a low-magnification wide-field
stereomicroscope having an oblique source
of illumination (Ref 3).
Do not put the mating pieces of a fracture
back together, except with considerable care
and protection. Protection of the surfaces is
particularly important if electron microscopic examination is to be part of the procedure (Ref 2). Appropriate packaging of
failed components for shipping is equally
important. Wrapping them directly into a
plastic bag, or placing pieces directly into
a plastic bottle or container, can introduce unwanted hydrocarbon contaminants.
pg 111
18/8/2008 3:03PM Plate # 0
Fingerprints on the failed surfaces can also

introduce contamination (Ref 4);
Do not conduct a destructive testing without
considerable thought. Alterations such as
cutting, drilling, and grinding can ruin
an investigation if performed prematurely.
Destructive testing must be performed only
after all possible information has been
extracted from the part in the original condition and after all significant features have
been carefully documented by photography
(Ref 2).
Preliminary Examination of the Failed

Part. In addition to locating the failure origin,
visual analysis is necessary to reveal stress concentrations, material imperfections, presence of
surface coatings, case-hardened regions, welds,
and other structural details that contribute to
cracking. A careful macroexamination is necessary to characterize the condition of the fracture
surface so that the subsequent microexamination
strategy can be determined. Corrodents often
do not penetrate the crack tip, and this region
remains relatively clean. The visual macroanalysis will often reveal secondary cracks that
have propagated only partially through a cracked member. These part-through cracks can be
opened in the laboratory and are often in much
better condition than the main fracture (Ref 3).
Nondestructive and Mechanical Testing.
A wide variety of nondestructive testing is
available, including dye penetrant, ultrasonics,
x-ray, and eddy current, which can help in
the failure analysis task in order to unveil
even subtle and/or internal defects in a part.
Mechanical property tests are also ready to use,
ranging from a sample hardness test to elevatedtemperature tensile and impact testing. These
tests are often used to determine if degradation is
related to fabrication or to the service environment. Sometimes, a standard test can be adapted
to simulate manufacturing or in-service conditions more closely (Ref 4).
Selection, Identification, Preservation,
and/or Cleaning of Specimens. Unless a
fracture is evaluated immediately after it is
produced, it should be preserved as soon as
possible to prevent attack from the environment.
The best way to preserve a fracture is to dry it
with a gentle stream of dry compressed air, then
store it in a desiccator, a vacuum storage vessel,
or a sealed plastic bag containing a desiccant.
However, such isolation of the fracture is often
not practical. Therefore, corrosion-preventive
surface coatings must be used to inhibit oxidation and corrosion of the fracture surface. The
primary disadvantage of using these surface
coatings is that fracture surface debris, which
often provides clues to the cause of fracture, may
be displaced during removal of the coating.
However, it is still possible to recover the surface debris from the solvent used to remove
these surface coatings by filtering the spent
solvent and capturing the residue. In regard to
cleaning techniques, fracture surfaces exposed
to various environments generally contain unwanted surface debris, corrosion or oxidation
products, and accumulated artifacts that must be
removed before meaningful fractography can
be performed. Before any cleaning procedures
begin, the fracture surface should be surveyed
with a low-power stereobinocular microscope,
and the results should be documented with appropriate sketches or photographs. Low-power
microscope viewing will also establish the
severity of the cleaning problem and should also
be used to monitor the effectiveness of each
subsequent cleaning step. It is important to
emphasize that the debris and deposits on the
fracture surface can contain information that is
vital to understanding the cause of fracture. The
most common techniques for cleaning fracture
surfaces, in order of increasing aggressiveness,
are (Ref 3):

Dry air blast or soft organic-fiber brush

cleaning
Replica stripping
Organic-solvent cleaning
Water-based detergent cleaning
Cathodic cleaning
Chemical-etch cleaning
Macroscopic Examination and Analysis

and Photographic Documentation. More
often than not, the investigation starts with a
low-magnification, if any, observation of the
failed part. This visual examination can often
quickly answer questions such as: What was the
mode of failure? Did it crack, or was there a
uniform or pitting corrosion failure? Did the
protective oxide film break down? Were the
welds visibly contaminated? A variable magnification stereoscope equipped with a ring light
and directional fiberoptic lighting is a powerful
tool for macroscopic visual examination.
Contemporary stereoscopes can operate over a
range of 2.5 to 50 (Ref 4).
Microscopic Examination and Analysis.
Once the area of interest is isolated, a smaller
pg 112
18/8/2008 3:03PM Plate # 0
General Aspects of Failure Analysis / 113
portion can be cut from the sample and mounted

for metallographic polishing and microscopic
examination. The microstructure of specimens
may be enhanced by a wide variety of metallographic techniques that include, for example,
heat tinting, stain etching, anodizing, and illumination by bright-field and polarizing light.
Optical microscopic examination generally begins at 50 magnification and continues through
1000 or even 1500 . Higher levels are best
supplemented by differential interference contrast lighting, which allows theoretical resolution of features as fine as one-third of a
micrometer. Features that are important to
recognize include the uniformity and size of the
grain structure, the size distribution and shape of
intermetallic particles, and inclusions. Scanning
electron microscopy (SEM) is most useful
where extreme depth of focus and high magnifications are needed. Fractures generally are
complex, undulating surfaces that are difficult to
image, and an optical microscope can only focus
on a very narrow region because of the very
shallow depth of field. However, the SEM excels
at imaging fracture surfaces, and it can be
operated in many different modes. The most
common mode is secondary electron imaging,
which provides a detailed, high-depth-focus
image that is easy to interpret. Backscattered
Z contrast is used to identify regions of
impurities within a matrix. High-atomic-number
species produce a light appearance, whereas
low-atomic-number species create a darker
appearance. The topographic backscattered
mode enhances the surface topography of the
sample and accentuates height or elevation differences on a fracture surface. The characteristic
x-rays can be detected and analyzed according to
their energy. This is called energy-dispersive
x-ray analysis. The x-ray wavelength corresponds to the presence of a specific element, and
its amplitude corresponds to the quantity of
such element. This technique allows quantitative characterization of elements within a given
phase. Bulk chemistry is typically analyzed
during failure analysis to verify conformance
with industry-accepted chemical limits. In the
case of reactive metals, light elements can
embrittle them due to improper processing or
service conditions (Ref 4).
Selection, Preparation, Examination, and
Analysis of Metallographic Specimens. One
of the worst things that can happen to the sample
is inadequate handling, examination, or packaging. It is imperative that the sample remains in
an undisturbed state prior to analysis, because

the culprit is often found in minute surface
features or traces of impurities. Fracture surfaces
must remain untouched so that high-magnification images can accurately determine the failure
mode. The sample must be removed carefully.
Important evidence can be destroyed by overheating or by allowing adjacent fracture surfaces
to fret or rub together during sectioning. The
ideal method would be to unbolt the component
or to provide adequate support so that a slowspeed saw can be used to cut out the component.
However, sawing lubricants can mask or
destroy residual chemicals or elements on the
failed surface, so precautions become extremely
necessary. If the component has failed in the
middle of a large area, more aggressive cutting/
sectioning techniques may be warranted, but
keep a good distance from the failed region
(Ref 4).
Determination of Failure Mechanism
(with Adapted Text from Ref 7). A thorough
investigation should ensure that all damage is
found and documented, because multiple modes
and mechanisms may be present in most realworld failure analyses. It is also important to
recognize that many unique mechanisms may be
driven by more than one environmental factor,
such as stress, temperature, corrosion, wear,
radiation, or electrical factors.
The term failure mechanism, or damage
mechanism, is meant to convey the specific
series of events that describe both how the
damage was incurred and the resulting consequences. Examples of damage mechanisms
include high-temperature creep, hydrogen
embrittlement, stress-corrosion cracking, and
sulfidation. A failure or damage mechanism
describes how damage came to be present.
This definition of failure mechanism also
should not be confused with the description of
the physical characteristics of damage observed.
For example, intergranular fracture, buckling,
transgranular beach marks, and pits can all be
thought of as damage modes. The term damage
mode or failure mode is best used to describe
what damage is present.
Much confusion has occurred because of
the tendency of engineers to use the terms
mechanism and mode interchangeably; in doing
so, it is unclear that two distinct characteristics
need to be assessed. Sometimes this occurs
because, within a given system, the same
wording is used to describe both the failure
mode and mechanism. For example, pitting
pg 113
18/8/2008 3:03PM Plate # 0
describes a damage mode because the surface of

a material is pitted. In certain systems, pitting is
also a possible damage mechanism. In boiler
tubing, for example, a pitting damage mechanism describes a specific localized corrosion
mechanism where pits form through dissolution
of metal either from low-pH or high-oxygen
conditions. The metal under the pit surfaces is
unaffected. In this system, pitting is a specific
damage mechanism, but many other damage
mechanisms also result in a pitting damage
mode in boiler tubing, including hydrogen
damage, phosphate corrosion, and caustic
gouging.
It is helpful to be as specific as possible in
differentiating damage mechanisms in a system.
For example, fatigue is often identified as both
a damage mode and a damage mechanism. A
fatigue damage mode is the observable damage
that occurs under fatigue loading cycles (e.g.,
the presence of beach marks). Classifying fatigue as a damage mechanism is not necessarily
complete because it does not point to the specific
environment that results in a fatigue damage
mode. Instead, specific mechanisms that can
result in a fatigue damage mode must be
examined. Examples include corrosion fatigue,
thermomechanical fatigue, creep-fatigue interaction, and mechanical fatigue.
Determination of damage mechanisms starts
by characterizing the component(s) being examined. It is impossible to know what is different about a failure without first understanding
what is expected from unfailed components.
In general, the analyst should obtain as much
information as possible about a part and its
background during the course of an investigation. Some key questions worth evaluating
include:

What was the part supposed to do? How was

it supposed to work?
How was the part made? What processes
were involved in its manufacture (e.g.,
forming, joining, and heat treatment)? What
properties were expected at the time of
manufacture?
What were the specified dimensions and
tolerances for the as-manufactured part?
How was the part installed?
To what service environment(s) was the part
exposed? Typical environments to examine
include operating temperatures, stresses
(steady state or slowly rising and cyclic),
oxidizing/corrosive environments, and wear

environments. What properties were required during service? How were properties
expected to change from service exposure?
How was the part inspected during service
intervals? What information was found
during these inspections?
What material characteristics were specified
for the part (e.g., composition, strength,
hardness, impact, and stress-rupture properties)? What specifications, industry standards, and contracts govern these properties?
What were the various ways the part could
fail?
The last item is a key question to repeatedly

ask throughout a failure investigation. The list of
various damage mechanisms by which a part can
fail can be narrowed down through two basic
concepts (Ref 7). Limiting conditions that refine
the scope of explanations for observed damage
can be defined by using the following two rules
of thumb:

When the impossible is eliminated, whatever

remains, however improbable, must be
considered (Sherlock Holmes rule).
When two or more explanations exist for a
sequence of events, the simple explanation is
more likely to be the correct one (Occams
razor).
Chemical Analysis. In a failure investigation, routine analysis of the material is usually

recommended. There are two main categories of
chemical analysis that are often used by failure
analysts:
Bulk composition evaluation: often performed in order to determine whether the

correct alloy was used in the subject component
Microchemical analysis: to find evidence of
contamination, to evaluate the composition
of microphases revealed on a metallographic
specimen, or to evaluate corrosion products
Often, chemical analysis is done last, because

an analysis usually involves destroying a certain
amount of material. There are instances where
the wrong material was used, under which conditions the material may be the major cause of
failure. In many cases, however, the difficulties
are caused by factors other than material composition.
Extreme care must be used in interpretation of
chemical analysis work performed as part of a
failure investigation. Minor deviations from
pg 114
18/8/2008 3:03PM Plate # 0
specified composition must not be interpreted as

the sole cause of a failure, without much additional supporting evidence. In most instances,
slight deviations from specified compositions
are not likely to be of major importance in failure
analysis. However, small deviations in aluminum content can lead to strain aging in steel, and
small quantities of impurities can lead to temper
embrittlement. In specific investigations, particularly where corrosion and stress corrosion
are involved, chemical analysis of any deposit,
scale, or corrosion product, or a substance with
which the affected material has been in contact,
is required to assist in establishing the primary
cause of failure.
Where analysis shows that the content of a
particular element is slightly greater than that
required in the specifications, it should not be
inferred that such deviation is responsible for the
failure. Often, it is doubtful whether such a
deviation has played even a contributory part in
the failure. For example, sulfur and phosphorus
in structural steels are limited to 0.04% in
many specifications, but rarely can a failure in
service be attributed to sulfur content slightly in
excess of 0.04%. Within limits, the distribution
of the microstructural constituents in a material
is of more importance than their exact proportions. An analysis (except a spectrographic
analysis restricted to a limited region of the
surface) is usually made on drillings representing a considerable volume of material and
therefore provides no indication of possible
local deviation due to segregation and similar
effects.
Also, certain gaseous elements, or interstitials, normally not reported in a chemical
analysis, have profound effects on the mechanical properties of metals. In steel, for example,
the effects of oxygen, nitrogen, and hydrogen
are of major importance. Oxygen and nitrogen
may give rise to strain aging and quench aging.
Hydrogen may induce brittleness, particularly
when absorbed during welding, cathodic cleaning, electroplating, or pickling. Hydrogen is
also responsible for the characteristic halos or
fisheyes on the fracture surfaces of welds in
steels, in which instance the presence of hydrogen often is due to the use of damp electrodes.
These halos are indications of local rupture
that has taken place under the bursting microstresses induced by the molecular hydrogen,
which diffuses through the metal in the atomic
state and collects under pressure in pores and
other discontinuities. Various effects due to gas
absorption are found in other metals and alloys.

For example, excessive levels of nitrogen in
superalloys can lead to brittle nitride phases that
cause failures of highly stressed parts.
Various analytical techniques can be used
to determine elemental concentrations and to
identify compounds in alloys, bulky deposits,
and samples of environmental fluids, lubricants,
and suspensions. Semiquantitative emission
spectrography, spectrophotometry, and atomicabsorption spectroscopy can be used to determine dissolved metals (as in analysis of an
alloy), with wet chemical methods used where
greater accuracy is needed to determine the
concentration of metals. Combustion methods
ordinarily are used for determining the concentration of carbon, sulfur, nitrogen, hydrogen,
and oxygen.
Wet chemical analysis methods may be
employed for determining the presence and
concentration of anions such as Cl , NO3 , and
S . These methods are very sensitive.
X-ray diffraction identifies crystalline compounds either on the metal surface or as a mass
of particles and can be used to analyze corrosion
products and other surface deposits. Minor and
trace elements capable of being dissolved can be
determined by atomic-absorption spectroscopy
of the solution. X-ray fluorescence spectrography can be used to analyze both crystalline
and amorphous solids, as well as liquids and
gases.
Stress Analysis and Fracture Mechanics
Analysis. When confronted with a cracked,
fractured, or deformed component, the failure
analyst will usually seek to answer some basic
questions:
Were the loads and stresses encountered by

the part at the level anticipated during
design? Or did some unexpected condition(s) contribute to the failure?
Was the material in the area of the cracking
or deformation capable of meeting the conditions anticipated during design? Was there
some deficiency or discontinuity that contributed to the failure, or was there a local
stress raiser at the critical location? Was this
taken into account by the designer?
In general, there are two types of conditions that

may lead to structural failure:
Net-section instability, where the overall

structural cross section can no longer support the applied load
pg 115
18/8/2008 3:03PM Plate # 0
The critical flaw size (ac) is exceeded

by some preexisting discontinuity or when
subcritical cracking mechanisms (for example, fatigue, stress-corrosion cracking, or
creep) reach the critical crack size
Failures due to net-section instability typically occur when a damage process such as
corrosion or wear reduces the thickness of a
structural section. This type of failure can be
evaluated by traditional stress analysis or finite
element analysis (FEA), which are effective
methods in evaluating the effects of loading and
geometric conditions on the distribution of stress
and strain in a body or structural system.
However, stress analyses by traditional
methods or FEA do not easily account for
crack propagation from preexisting cracks or
sharp discontinuities in the material. When a
preexisting crack or discontinuity is present,
the concentration of stresses at the crack tip
becomes asymptotic (infinite) when using the
conventional theory of elasticity. In this regard,
fracture mechanics is a useful tool, because it is a
method that quantifies stresses at a crack tip
in terms of a stress-intensity parameter (K).
The fracture mechanics of cracking from a discontinuity or crack in a statically loaded component has two possible situations:

The crack reaches a critical length with rapid

(brittle) separation.
The crack blunts, redistributing the stress
state, with continued loading creating a tear
zone (and sharpened crack-tip radius) in
front of the crack. In steels, this tear zone
can then cause the critical crack length to be
exceeded, such that unstable cleavage fracture occurs or unstable microscale ductile
fracture is induced.
Which event occurs depends on the temperature

and the loading rate, but in either event, crack
propagation is unstable (i.e., does not require an
increasing load after creation of the tear zone).
Fracture mechanics is a tool to help evaluate the
implications of preexisting discontinuities or
cracks.
Testing under Simulated Service Conditions. During the concluding stages of an
investigation, it may be necessary to conduct
tests that simulate the conditions under which
failure is believed to have occurred. Often,
simulated-service testing is not practical because elaborate equipment is required, and even
where practical it is possible that not all of the
service conditions are fully known or understood. Corrosion failures, for example, are
difficult to reproduce in a laboratory, and some
attempts to reproduce them have given misleading results. Serious errors can arise when
attempts are made to reduce the time required for
a test by artificially increasing the severity of one
of the factorssuch as the corrosive medium or
the operating temperature. Similar problems are
encountered in wear testing.
On the other hand, when its limitations are
clearly understood, the simulated testing and
statistical experimental design analysis of the
effects of certain selected variables encountered
in service may be helpful in planning corrective
action or, at least, may extend service life. Most
of the metallurgical phenomena involved in
failures can be satisfactorily reproduced on a
laboratory scale, and the information derived
from such experiments can be helpful to the
investigator, provided the limitations of the tests
are fully recognized.
Analysis of All the Evidence, Formulation
of Conclusions, and Writing the Report.
Before starting this final step, some questions
must already be answered:
Fracture surface:
a. What is the fracture mode?
b. Is the origin of the fracture visible?
c. What is the relation between the fracture
direction and the normal or expected fracture directions?
d. How many fracture origins are there?
e. Is there evidence of corrosion, paint, or
some other foreign material on the fracture
surface?
f. Was the stress unidirectional or was it
reversed in direction?
The surface of a part:
a. What is the contact pattern on the surface
of the part?
b. Has the surface of the part been deformed
by loading during service or by damage
after fracture?
c. Is there evidence of damage on the surface
of the part by manufacturing, assembling,
repairing, or service?
Geometry and design:
a. Are there any stress concentrations related
to the fracture?
b. Is the part intended to be relatively rigid,
or is it intended to be flexible, like a
spring?
c. Does the part have a basically flawless
design?
pg 116
18/8/2008 3:03PM Plate # 0
d. How does the partand its assembly

work?
e. Is the part dimensionally correct?
Manufacturing and processing:
a. Are there internal discontinuities or
stress concentrations that could cause a
problem?
b. If it is a wrought metal, does it contain
serious seams, inclusions, or forging problems, such as end grains, laps, or other
discontinuities, that could have an effect
on performance?
c. If it is a casting, does it contain shrinkage
cavities, cold shuts, gas porosity, or
other discontinuities, particularly near the
surface of the part?
d. If a weldment was involved, was the
fracture through the weld itself or through
the heat-affected zone in the parent metal
adjacent to the weld? If through the weld,
were these problems something like gas
porosity, undercutting, underbead cracking, or lack of penetration? If through the
heat-affected zone adjacent to the weld,
how were the parent metal properties
affected by the heat of welding?
e. If the part was heat treated, was the treatment properly performed?
Material properties:
a. Are the mechanical properties of the metal
within the specified range, if this can be
ascertained?
b. Are the properties of the metal suitable for
the application?
c. Residual and applied stress relationship.
The residual-stress system that was within
the part prior to fracture can have a powerful effectgood or badon the performance of a part.
d. What was the influence of adjacent parts
on the failed part?
e. Were fasteners tight?
Assembly:
a. Is there evidence of misalignment of the
assembly that could have had an effect on
the fractured part?
b. Is there evidence of inaccurate machining, forming, or accumulation of tolerances?
c. Did the assembly deflect excessively under
stress?
Service conditions: It is important to determine if there were any unusual occurrences,
such as strange noises, smells, fumes, or
other happenings, that could help explain the
problem. The following questions should

also be considered:
a. Is there evidence that the mechanism was
overspeeded or overloaded?
b. Is there evidence that the mechanism was
abused during service or used under conditions for which it was not intended?
c. Did the mechanism or structure receive
normal maintenance with the recommended materials?
d. What is the general condition of the
mechanism?
Environmental reactions: The problems
related to the environment can arise anywhere
in the history of the part: manufacturing,
shipping, storage, assembly, maintenance,
and service. None of these stages should be
overlooked in a thorough investigation that
asks:
a. What chemical reactions could have taken
place with the part during its history?
b. To what thermal conditions has the part
been subjected during its existence?
Report writing: Finally, the report analyzing
the failure should be written in a clear,
concise, logical manner. It should be clearly
structured with sections covering the following (Ref 6):
a. Description of the failed item
b. Conditions at the time of failure
c. Background history important to the
failure
d. Mechanical and metallurgical study of the
failure
e. Evaluation of the material quality
f. Discussion of any anomalies
g. Discussion of the mechanism or possible
mechanisms that caused the failure
h. Recommendations for the prevention of
future failures or for action to be taken with
similar pieces of equipment
Irrelevant data should be omitted, and,
depending on the nature of the problem and the
data, not every report will need full treatments
for every one of the sections listed previously.
Many times, the readership may include purchasing, operating, or accounting personnel
who are not technically trained. If this is the
situation, the report should be written so that it
is comprehensible to these persons. At least,
those sections of the report that bear on their
decision-making or information needs should be
written in language that is accessible to them.
Frequently, a cover letter summarizing the most
important findings and the suggested action is a
pg 117
18/8/2008 3:03PM Plate # 0
good vehicle for reaching top executives who

are not as interested in the technical specifics but
need key findings and recommendations as a
basis for decision making. Followup on the
recommendations is frequently a difficult task
but should be undertaken for the more critical
failures. Cooperation between the investigator,
the designer, the manufacturer, and the user is
critical in developing good, workable changes.
Fracture
The process of fracture, in general terms, can
be described in terms of the mechanisms of
crack initiation and/or crack extension (growth).
Different mechanisms may occur for crack
initiation and the subsequent process of crack
growth. For example, crack extension may occur
by the brittle mechanism of cleavage, even
though extensive elongation accompanied or
preceded crack initiation. The fracture may be
classified as either ductile or brittle, depending
on whether the mechanism is describing crack
initiation or crack growth, respectively. Likewise, the low-energy catastrophic fracture of a
high-strength aluminum alloy by microvoid
coalescence is also difficult to classify because,
although the fracture energy is low and failure
initiates by fracture or decohesion of brittle
particles, the growth and coalescence of the
microvoids occurs by plastic deformation.
Another difficulty is that cleavage fracture may
be initiated by dislocation interactions that, by
definition, involve plasticity. This is why fractures are sometimes difficult to logically classify
(Ref 5). Therefore, it is helpful to be clear
whether fracture mechanisms are describing the
process of crack initiation or extension. Crack
extension also can be multimode over time (e.g.,
fatigue crack growth followed by overload).
In terms of fracture appearances (or fracture
modes, defined earlier in the section Determination of Failure Mechanism in this chapter), a general summary of the visual and
microscopic aspects of fracture surfaces for
metallic materials is provided in Table 1 (Ref 8).
Several analytical procedures are available for
distinguishing among the various types of fracture. For example, the presence or absence of
plastic macrodeformation can be determined
with the unaided eye or by use of a steel scale, a
machinists micrometer, or a machinists or
measuring microscope. Differences in some
dimensional attribute of parts (such as width or
thickness) at and well away from the fracture can

serve to define macrodeformation after assurance that both points of measurement had the
same dimension before fracture.
Fracture-surface matching is also used to
determine the presence or absence of plastic
deformation. It is very important, however, to
resist the temptation to fit the matching fracture
surfaces together, because this almost always
destroys (smears) microscopic features. The
fracture surfaces should never actually touch
during fracture-surface matching.
The origin of a fracture may be indicated by a
discoloration or by the topography of the fracture surface. A discolored area on a fracture
surface may be produced by a preexisting crack
whose surfaces have been corroded or oxidized.
For example, the surfaces of a quench crack can
be oxidized during a subsequent tempering heat
treatment; the oxide film gives a bluish-black
color to the surfaces of the crack. Topographical
features that often reveal the origin of a fracture
are either chevron or river patterns or a set
of diverging ledges. If the fracture surface is
essentially featureless, the presence of a shear
lip can be used to locate, within limits, the origin
of a fracture. For example, a shear lip is not
formed at the origin of a stress-corrosion crack,
but when the crack begins to propagate rapidly, a
shear lip is formed wherever the crack front
exits from the interior to the free surface. Beach
marks, which are associated with fatigueinitiated fractures, also provide a definite indication of the crack origin; however, it should be
noted that fracture surfaces having an appearance similar to that of the beach-mark pattern
can be produced by stress corrosion.
Generally, cyclic loading produces only a
single crack, which is usually located at a site of
stress concentration or of a metallurgical defect,
whereas additional cracks, formed independently of the main crack and at a distance from it,
may be observed on the surface of a structural
or machine component subjected to corrosion
fatigue or stress corrosion.
On the microscopic level, striations on the
fracture surface are unique to fatigue, and the
crack path, although normally transgranular, can
be intergranular. For example, intergranular
fatigue cracking can occur in the case of a carburized steel or in a material that has a high
density of second-phase particles at the grain
boundaries.
Corrosion-fatigue and stress-corrosion cracks
may propagate transgranularly, intergranularly,
pg 118
18/8/2008 3:03PM Plate # 0
pg 119
Table 1 Fracture mode identification chart

Instantaneous failure mode(a)
Method
Visual, 1 to
50
(fracture
surface)
Ductile overload
Necking or
distortion in
direction
consistent with
applied loads
Dull, fibrous
fracture
Shear lips
Brittle overload
Progressive failure mode(b)

Fatigue
Corrosion
Wear
Creep
Multiple brittle Gouging,

Flat progressive General
Little or no
appearing fissures
abrasion,
wastage, roughzone with beach
distortion
External surface
ening, pitting, or polishing,
marks
Flat fracture
and internal
or erosion
trenching
Bright or coarse Overload zone
fissures contain
consistent with Stress-corrosion Galling or storing
texture,
reaction scale
in direction of
and hydrogen
applied loading
crystalline,
coatings
motion
damage may
direction
grainy

Fracture after
create multiple Roughened areas
Ratchet marks
Rays or
limited
with compacted
cracks that
where origins
chevrons
dimensional
powdered debris
appear brittle
join
point to origin
change
(fretting)
Smooth gradual
transitions in
wastage
Path of penetra- Wear debris and/or Multiple

Progressive
Cleavage or
Microvoids
Scanning
intergranular
abrasive can be
tion may be
zone: worn
intergranular
(dimples)
electron
characterized as to fissures covered
irregular,
appearance,
fracture
elongated
microscopy,
with reaction scale
morphology and
intergranular, or
flat, may show
Origin area may
in direction of
20 to
Grain faces may
a selective phase composition
striations at
contain an
loading
10,000
show porosity
Rolling-contact
attacked
magnification
imperfection
Single crack with
(fracture
fatigue appears
EDS may
above 500
or stress
no branching
surface)
like wear in early
help identify
Overload zone:
concentrator
Surface slip band
stages
corrodent(c)
may be either
emergence
ductile or brittle
Metallographic
inspection,
50 to 1000

(cross
section)
Contributing
factors
Grain distortion
and flow near
fracture
Irregular,
transgranular
fracture
Load exceeded the

strength of the part
Check for proper
alloy and processing by hardness
check or destructive testing,
chemical analysis
Loading direction
may show failure
was secondary
Short-term,
high-temperature,
high-stress rupture
has ductile
appearance
(see creep)
Microstructural
May show
General or
Little distortion Progressive
change typical of
localized surface localized
zone: usually
evident
overheating
distortion at
attack (pitting,
transgranular
Intergranular or
surface consistent Multiple intercracking)
with little
transgranular
granular cracks
with direction of
Selective phase
apparent
May relate to
Voids formed on
motion
attack
distortion
notches at
surface or brittle Overload zone: Thickness and Identify embedded grain boundaries
or wedge-shaped
particles
morphology of
phases internally may be either
cracks at grain
corrosion scales
ductile or brittle
triple points
Reaction scales
or internal
precipitation
Some cold flow
in last stages of
failure
Load exceeded
the dynamic
strength of the
part
Check for proper
alloy and
processing as
well as proper
toughness, grain
size
Loading
direction may
show failure was
secondary or
impact induced
Low
temperatures
Attack morphol- For gouging or

Cyclic stress
abrasive wear:
ogy and alloy
exceeded the
check source of
type must be
endurance limit
abrasives
evaluated
of the material
Evaluate effec Check for proper Severity of
tiveness of lubristrength, surface exposure
cants
finish, assembly, conditions may
Seals or filters may
be excessive;
and operation
have failed
Prior damage by check: pH,
Fretting induced
temperature,
mechanical or
by slight looseness
corrosion modes flow rate,
in clamped joints
dissolved
may have
subject to
oxidants, elecinitiated
vibration
trical current,
cracking
metal coupling, Bearing or materi Alignment,
als engineering
aggressive
vibration,
design may reduce
agents
balance
or eliminate
High cycle low Check bulk
composition and problem
stress: large
Water
contaminants
fatigue zone;
contamination
low cycle high
High velocities
stress: small
or uneven flow
fatigue zone
distribution,
cavitation
Mild overheating
and/or mild
overstressing at
elevated
temperature
Unstable microstructures and
small grain size
increase creep
rates
Ruptures occur
after long
exposure times
Verify proper
alloy
(a) Failure at the time of load application without prior weakening. (b) Failure after a period of time where the strength has degraded due to the formation of cracks, internal
defects, or wastage. (c) EDS, energy-dispersive spectroscopy. Compiled by C.R. Morin, S.L. Meiley, and Z.B. Flanders, Packer Engineering Associates, Inc.
18/8/2008 3:03PM Plate # 0
or by a combination of both modes. A distinguishing feature of stress corrosion is the

branching of the main crack. If corrosion pits
or corrosion products are found only on the
slow-growth region of a fracture surface, the
environment was in all probability sufficiently
corrosive to affect the fracture mechanism.
However, if evidence of corrosion is found on
both the slow-growth and fast-growth areas,
some corrosion took place subsequent to fracture, and the environment may or may not have
influenced fracture.
Ductile Fracture
Ductile fracture takes place when a material
capable of undergoing plastic deformation is
subjected to stresses that culminate in its rupture. Macroscopically, the ductile fracture process presents some peculiarities that allow it to
be identified immediately. The first feature is the
presence of plastic deformation that may be
accompanied by neck formation. In tensile
testpieces of ductile materials, besides necking,
the fracture surface presents a fibrous aspect and
a cup-cone geometry, as seen in Fig. 1.
Fig. 1
The fracture process begins in the center of

the testpiece with microvoid nucleation along
grain boundaries or from interfaces such as those
found in base metal/inclusions boundaries. As
the applied stress increases, microvoids grow
and coalesce, forming a crack in the center of the
part. This process, depicted in Fig. 2, ends up in
rapid crack propagation by shearing of the
remaining ligament of the neck region, at an
angle of 45 in relation to the loading direction.
It is important to emphasize that a cup-cone
geometry will depend on the geometry and
dimensions of the part and mechanical properties of the material. Thin sheets, for instance,
present neck formation and a fracture surface
oriented at an angle of 45 in relation to the
applied load, as observed in Fig. 3. Ductile
fracture takes place intergranularly, unless some
sort of mechanism weakens the grain boundaries. The microscopic aspect of the fracture
surface consists of several small elliptical
cavities, or microvoids, as depicted in Fig. 4.
Brittle Fracture
Brittle fracture occurs with little or no plastic
deformation. This type of fracture is often
Fig. 3
Thin sheet testpiece of a low-carbon steel after fracture
Fig. 4
Microvoids on the fracture surface of AA6061-T1

tensile testpiece
Ductile fracture showing the typical cup-cone geometry
Microvoids
Fig. 2
Schematic representation of the cup-cone geometry

formation during the ductile fracture process
pg 120
18/8/2008 3:03PM Plate # 0
associated with materials of high strength and

low ductility or materials that were subjected
to an embrittlement process. The crack, once
nucleated, propagates very quickly in a direction
perpendicular to the applied load. Figure 5 presents an example of a gray cast iron testpiece that
presented brittle fracture.
Besides the mechanical properties, several
other factors may result in a brittle behavior,
such as temperature, loading rate, presence of
stress concentrators, and dimensions. Low temperatures tend to reduce the ductility of metals,
especially those possessing a body-centered
cubic structure, resulting in a typically brittle
fracture. Figure 6 shows that as the temperature
drops, the brittle aspect on the fracture surface of
impact testpieces increases. The presence of
stress raisers or larger dimensions introduces a
more severe triaxial stress state within the
material, and thus, there is larger probability that
brittle fracture will occur. However, it is known
Fig. 5
that the superposition of high hydrostatic stresses on the material reduces the triaxiality levels,
increasing ductility. High applied loading rates
are likely to make plastic deformation more
difficult because shearing processes are timedependent, resulting in brittle behavior.
Crack propagation by brittle fracture can
occur across the grains (transgranular) or
along the grain boundaries (intergranular). In
the transgranular mode, the fracture process
takes place by cleavage along specific crystallographic planes. Figure 7 presents cleavage
regions in a microalloyed low-carbon steel,
which can be identified by flat regions on the
fracture surface. Additionally, it is worth mentioning that most parts of steels will present
alternate regions consisting of cleavage areas
and microvoids, evidencing a mixed mode of
crack propagation.
In another situation, fracture can take place
intergranularly, because the grain boundary is a
Tensile testpiece of gray cast iron presenting brittle

fracture
Fig. 7
Fig. 6
Fracture surfaces of SAE 4140 impact testpieces.

Tested at room temperature, right, and at 196 C, left
(a) Cleavage region observed in low-carbon steel.

(b) Magnification of the region delimited by the rectangle in (a) showing an inclusion in the center of the cleavage
region
pg 121
18/8/2008 3:03PM Plate # 0
pg 122
weaker path for crack propagation. Normally,

this fracture mode will occur when some
embrittlement process resulted in grain boundaries being more susceptible to crack propagation than the core of the grain, such as an
unsuitable heat treating or by environmental
factors. Figure 8 presents an example of intergranular brittle fracture in an austenitic stainless
steel SAE 316L, where grain boundaries can
clearly be observed on the fracture surface.
Fatigue Fracture
According to the definition given by ASTM
E1823, fatigue is the process of progressive
localized permanent structural change occurring
in a material subjected to conditions that produce fluctuating stresses and strains at some
point or points and that may culminate in cracks
or complete fracture after a sufficient number of
fluctuations. A material subjected to fatigue
can fracture at applied stresses much lower than
those necessary to fracture the same material
under monotonic conditions. The fluctuating
stresses can be originated from mechanical,
thermal, or vibration loading conditions, and the
phenomenon is responsible for more than 80%
of mechanical failures of components. For more
than 150 years, the study of metals fatigue
has involved engineers, physicists, chemists,
and mathematicians, and everyday this study
becomes more and more complex and important. The theory about fatigue is extremely vast,
and for each question answered, another one,
more instigating, appears, requiring a broad
knowledge of materials science. In the following
topics, a brief overview is given about the main
mechanisms and factors influencing the fatigue
life of a component during both the nucleation

and crack propagation phases.
Fatigue Crack Initiation. Generally, fatigue
cracks are initiated at free surfaces, where there
is no constraint to material deformation; however, in some cases, cracks may be initiated in
the interior of the material where interfaces are
present, such as the interface of a carburized
surface layer and the base metal or the interface
of an inclusion and the base metal, or from gas
bubbles. In other cases, subsurface cracks were
found to nucleate below the surface where high
compressive residual stresses were introduced
by shot peening or surface rolling.
One of the classic models of fatigue crack
nucleation considers that when a material is
under loading (monotonic or cyclic), slips occur
at the high-shear-stress planes, creating steps on
the material surface. Under cyclic loading,
the formation of intrusions and extrusions is
observed, as schematically represented in Fig. 9.
Slip band intrusions are excellent stress raisers
that can be sites of crack nucleation.
Besides the applied stress amplitude, DS/2,
several other factors are likely to affect the
nucleation of a fatigue crack, such as the mean
stress, Sm, or load ratio, R; geometry and surface
finishing of the part; mechanical properties; and
environment. Here, the R ratio is defined as the
ratio between the minimum and maximum loads
during the fatigue cycle.
A large proportion of fatigue data found in the
literature refers to tests conducted at Sm = 0,
that is, for a load ratio R = 1. However, in
many engineering situations, the fluctuating
stresses are superimposed to a static stress.
Larger mean stresses reduce the nucleation time
because they facilitate the plastic deformation
mechanism associated with this phenomenon. In
an S-N graph, this can be represented by curves
shifted to the left and down, as represented in
Fig. 10.
Intrusion
Extrusion
Metal Surface
Fig. 8
Fractograph of SAE 316L showing intergranular brittle

fracture
Fig. 9
Schematic representation of an intrusion formation on

the surface of a metallic material
18/8/2008 3:03PM Plate # 0
pg 123
The mechanism proposed in Fig. 9 is

adequate to explain the initiation of cracks
on polished testpieces or components without
the presence of geometric discontinuities.
However, in engineering components, there are
several stress concentrators, such as scratches,
notches, machining marks, corrosion pits, and
microconstituents such as grain boundaries,
triple points, and inclusions, that individually or
synergistically can reduce the initiation time.
Since the initiation depends essentially on
plastic deformation mechanisms, high-strength
materials normally present a higher resistance to
fatigue crack nucleation. In this sense, several
surface-hardening treatments are employed to
selectively reinforce the material, aiming to
retard crack initiation and therefore to increase
fatigue life.
The chemical composition and/or the microstructure of the surface can be modified by
thermochemical treatments, such as carburizing
or nitriding, or by cold deformation processes,
such as shot peening or surface rolling.
Mechanical parts that necessarily present stress
concentrators, such as crankshafts, gears, and
bolts, can be subjected to these treatments
to increase the fatigue limit of the material.
Figure 11 shows a micrograph of the transverse
section of a bolt, where the thread was cold
formed by surface rolling. As a consequence,
surface grains are flattened due to the mechanical deformation imposed. In this case, besides
increasing hardness and mechanical strength,
the process avoids the introduction of harmful
machining marks.
Surface treatments may also increase fatigue
life by the introduction of compressive residual
stresses on the surface of the material. As long as
the material remains in linear elastic conditions,
the principle of stress superposition can be
employed to describe the actual stress state in

materials containing residual stresses. Therefore, the effective stress, S0 , is given by the
sum of the applied stress, S, to the residual stress,
Sres:
S0 =S+Sres
(Eq 1)
Similarly, the effective minimum and maximum

stresses are defined, respectively, as:
S0max =Smax +Sres
(Eq 2)
S0min =Smin +Sres
(Eq 3)
Consequently, the effective stress amplitude,

mean stress, and load ratio are given, respectively, by:
DS0 S0max 7S0min (Smax +Sres )7(Smin +Sres )
=
=
2
2
2
Smax 7Smin DS
=
=
(Eq 4)
2
2
S0max +S0min (Smax +Sres )+(Smin +Sres )
=
2
2
Smax +Smin
+Sres =Sm +Sres
=
(Eq 5)
2
S0m =
R0 =
S0min
S +Sres
= min
S0max Smax +Sres
(Eq 6)
Therefore, the presence of a residual-stress

field does not affect the stress amplitude but
affects the mean stress and the load ratio.
A compressive residual stress reduces the mean
stress and the load ratio, increasing the number
Increasing
Sm
S/2
Nf
Fig. 10
Mean stress effect on S-N fatigue curves
Fig. 11
Optical micrograph of the transverse section of a

thread fillet machined by surface rolling. The material consists of duplex stainless steel
18/8/2008 3:03PM Plate # 0
of cycles for crack nucleation and vice versa.

In some situations, where high surface compressive residual stresses are found, such as in
materials subjected to surface-hardening treatments, a crack may initiate below the surface,
where the compressive residual-stress level is
lower. An example of subsurface crack nucleation is observed in Fig. 12 for a surface-rolled
ductile cast iron subjected to bending-rotating
fatigue.
Fatigue Crack Propagation. Basically, fatigue crack propagation can be divided into three
stages: stage I (short cracks), stage II (long
cracks), and stage III (final fracture).
A fatigue crack, once initiated, propagates
along high shear-stress planes (45 ), as schematically represented in Fig. 13. This is known
as stage I or the short crack growth propagation
stage. The crack propagates until it is decelerated by a microstructural barrier, such as a grain
boundary, inclusions, or pearlitic zones, that
cannot accommodate the initial crack growth
direction. Therefore, grain refinement is capable
of increasing fatigue strength of the material due
to the insertion of a large quantity of microstructural barriers, that is, grain boundaries, that
must be overcome in stage I of propagation.
Surface mechanical treatments, such as shot
Fig. 12
Probable subsurface crack nucleation site in a surface-rolled ductile cast iron testpiece tested under
bending-rotating conditions
Stage II
Surface
Stage I
Fig. 13
Stages I and II of fatigue crack propagation
peening and surface rolling, contribute to the

increase in the number of microstructural
barriers per unit of length due to the flattening of
the grains.
When the stress-intensity factor, K, increases
as a consequence of crack growth or higher
applied loads, slips start to occur in different
planes close to the crack tip, initiating stage II of
propagation. While stage I of propagation is
orientated 45 in relation to the applied load,
propagation in stage II is perpendicular to load
direction, as depicted in Fig. 13. An important
characteristic of stage II propagation is the
presence of ripples on the fracture surface,
known as striations, which are only visible with
the aid of a scanning electron microscope. Not
all engineering materials exhibit striations. They
are clearly seen in pure metals and many ductile
alloys, such as aluminum alloys. In steels, they
are frequently observed in cold-worked alloys.
Figure 14 shows examples of fatigue striations
in an interstitial-free steel and in aluminum
alloys. The most accepted mechanism for the
formation of striations on the fatigue fracture
surface of ductile metals (Ref 9) consists of
successive blunting and resharpening of the
crack tip, as represented in Fig. 15.
Finally, stage III is related to the unstable
crack growth as Kmax approaches KIc. At this
stage, crack growth is controlled by static modes
of failure and is very sensitive to the microstructure, load ratio, and stress state (planestress or plane-strain loading).
Macroscopically, the fatigue fracture surface
can be divided into two distinct regions, as
shown by Fig. 16. The first region corresponds to
the stable fatigue crack growth and presents
a smooth aspect due to the friction between
the crack-wake faces. Sometimes, concentric
marks, known as beach marks, can be seen on
the fatigue fracture surface as a result of successive arrests or decrease in the fatigue crack
growth rate due to a temporary load drop or to an
overload that introduces a compressive residualstress field ahead of the crack tip.
The other region corresponds to the final
fracture and presents a fibrous and irregular
aspect. In this region, the fracture can be either
brittle or ductile, depending on the mechanical
properties of the material, dimensions of the
part, and loading conditions. The exact fraction
of area of each region will depend on the applied
load level. High applied loads will result in a
small stable fatigue crack propagation area,
as depicted in Fig. 16(a). On the other hand,
pg 124
18/8/2008 3:03PM Plate # 0
pg 125
if lower loads are applied, the fatigue crack

will have to grow longer before the applied
stress-intensity factor, K, reaches the fracture
toughness value of the material, resulting in a
smaller area of fast fracture (Fig. 16b).
Ratcheting marks are another macroscopic
feature that can be observed in fatigue fracture
surfaces. These marks originate when multiple
(a)
(d)
(b)
(e)
(c)
Fig. 15
Proposed mechanisms of striation formation in stage

II of propagation. (a) No load. (b) Tensile load. (c)
Maximum tensile load. (d) Load reversion. (e) Compressive load.
Source: Ref 9
Fig. 14
Fatigue striations in (a) interstitial-free steel and (b)

aluminum alloy AA2024-T42. (c) Fatigue fracture
surface of a cast aluminum alloy where a fatigue crack was
nucleated from a casting defect, presenting solidification dendrites on the surface. Arrow at top right indicates fatigue striations.
Fig. 16
Fatigue fracture surface. (a) High applied load.

(b) Low applied load
18/8/2008 3:03PM Plate # 0
pg 126
cracks, nucleated at different points, join together, creating steps on the fracture surface.
Therefore, counting the number of ratchet marks
is a good indicator of the number of nucleation
sites. Figure 17 presents in detail some ratchet
marks found on the fracture surface of a large
SAE 1045 rotating shaft, fractured by fatigue.
Similar to the initiation phase, many factors
can affect long fatigue crack propagation rates.
Among them, special attention should be given
to the effects of load ratio and the presence of
residual stresses.
Increasing the load ratio has a tendency to
increase the long crack growth rates in all
regions of the fatigue crack growth rate versus
applied stress-intensity factor range curve, or
simply, da/dN versus applied DK curve. Generally, the effect of increasing load ratio is less
significant in the Paris regime than in nearthreshold and near-failure regions (Fig. 18).
Fig. 17
Near the threshold stress-intensity factor,

DKth, the effects of R ratio are mainly attributed
to crack closure effects, where crack faces come
in contact at an applied Kcl that is higher than the
minimum applied stress-intensity factor, Kmin.
Several different mechanisms may contribute
to premature crack closure. One of them consists
of plasticity-induced closure, represented in
Fig. 19(a). As the crack grows, the material that
has been previously permanently deformed
within the plastic zone now forms an envelope of
plastic zones in the wake of the crack front. This
leads to displacements normal to the crack surfaces as the restraint is relieved. This is no problem while the crack is open; however, as the
load decreases, the crack surfaces touch before
the minimum load is reached, shielding the
crack. This type of premature contact can also
occur due to the crack-wake roughness and
irregularities (Fig. 19b) or by the presence of
corrosion subproducts, such as oxides (Fig. 19c).
As observed in Fig. 20, the effect of closure
produces a reduction in the effective DK range
because of the increase in the effective Kmin,
reducing the driving force for fatigue crack
growth. The effect is more significant near the
threshold region because the crack tip opening
displacements are smaller and the crack faces
are closer to each other. Additionally, for the
same applied DK, higher R ratios increase the
applied values of Kmax and Kmin, increasing
DKeff.
For most materials, the Paris regime is considered closure-free and Kmax-independent, and
Ratcheting marks, indicated by the arrows, in an SAE

1045 shaft fractured by fatigue
Plastic deformation
envelope
Final failure
Plastic zone
Crack tip
(a)
da/dN
Paris regime
Increasing
R
(b)
Premature contact points
Near threshold
Oxides
K
(c)
Fig. 18
Schematic representation of the R ratio effect on

fatigue crack growth curves. The near-threshold,
Paris regime, and final failure regions are also indicated on the
curves.
Fig. 19
Crack closure mechanisms induced by (a) plasticity,

(b) roughness, and (c) oxide
18/8/2008 3:03PM Plate # 0
pg 127
0 j0, then:
If Kmin
the crack growth rates are generally very similar

for tests conducted under different R ratios. Near
the final failure, the effects of R ratio are related
to the higher monotonic fracture component as
Kmax approaches KIc. Therefore, for the same
applied DK, Kmax values are higher for tests
conducted under higher applied R ratios, and
consequently, da/dN values are higher.
The effects of residual stress on fatigue crack
growth are related to alterations in the R ratio
and in the applied DK. In other terms, the residual stresses affect the two parameters that
control the crack driving force, that is, Kmax and
DKeff. When a crack is introduced in a plate
subjected to a residual-stress field, a residual
stress-intensity factor, Kr, arises that can either
decrease or increase the crack driving force
parameters.
The superposition principle can also be
applied in terms of the stress-intensity factor,
provided that the material remains linearly
elastic. In this sense, Kr can be added to Kmax and
Kmin:
0
Kmax
=Kmax +Kr
(Eq 7)
0
=Kmin +Kr
Kmin
(Eq 8)
0
Kmin
K +Kr
= min
0
Kmax Kmax +Kr
(Eq 11)
0
=Kmax +Kr
DK 0 =Kmax
(Eq 12)
It is important to note that these equations

assume that the part of the fatigue cycle during
which the crack is closed at its tip (i.e., K050)
makes no contribution to crack growth.
Distortion
Distortion is the least serious mode of failure,
but it can lead a part to failure or a structure to
collapse. It is easy to recognize but very difficult
to prevent. This is due to the fact that distortion
does not involve the part itself but its use and
design. There are four reasons for distortion:
yielding, buckling, creep, and residual stresses.
Yielding. When a load is put on a part, and it
causes the part to be permanently distorted, it is
unable to perform the intended function and
therefore must be considered failed. In a welldesigned part, the stresses never exceed the yield
point, and the part deforms only elastically; that
is, when the load is released, the part returns to
its original dimensions.
In a good design, the part operates in the
elastic range, that is, below yielding point;
beyond this, the part will be permanently
deformed, and greater loads will cause the part
to actually break. This point is considered to be a
very basic point to design and applies when the
load on a part is applied in a quasi-static way,
such as the load on a building structure or the
stress in the legs of a desk. A ductile failure is
As a result, R0 and DK0 are defined as follows. If

0 40, then:
Kmin
R0 =
R0 =0
(Eq 9)
0
0
DK 0 =Kmax
7Kmin
=Kmax +Kr 7Kmin +Kr
=Kmax 7Kmin =DK
(Eq 10)
Kmax
Kmax
Kap=Keff
K
Keff
Kap
Kmin
Kcl
Kcl
Kmin
Time
Time
(a)
Fig. 20
(b)
Load ratio effect on DKeff in a fatigue cycle. (a) Kmin5Kcl. (b) Kmin4Kcl
18/8/2008 3:03PM Plate # 0
one where there is a great deal of distortion of the

failed part. Commonly, a ductile part fails when
it distorts and can no longer carry the needed
load. However, some ductile parts break into
two pieces and can be identified because there is
a great deal of distortion around the fracture
face, similar to what would happen if too much is
placed load on a low-carbon steel bolt.
Buckling. The failure of an engineering
component is not always caused by materials
fracture. In many occasions, the component
distortion may be sufficient to put it out of
function. The distortion can be elastic or plastic.
The elastic distortions are temporary; however,
they may be sufficient to cause interference on
the mobile parts. The plastic distortion is permanent and can be a result of an overload or
creep deformation. The overload causes permanent plastic deformation when the material
yield limit is overcome. This may happen in the
presence of stress concentrators, high temperature, inadequate heat treatment, or incorrect
materials selection for the component application. Compressive overloads may lead the
material to overcome the buckling strength
limit, such as the one shown in Fig. 21 for an
aluminum part. The buckling strength is essentially a design problem (not metallurgical), and
the load depends on the dimensions of the part
and the Youngs modulus of the material (the
only materials factors involved).
Creep is a time-dependent phenomenon that
causes a part failure if it is under both quasistatic load and temperatures higher than 0.3 Tm
(absolute melting temperature). Creep strain
may produce sufficiently large deformation or
distortion that a part can no longer perform its

intended function. The two general types of
creep processes are grain-boundary sliding and
voids at grain boundaries (cavitation creep).
The creep processes are easily identified by
the local ductility and large numbers of intergranular cracks that will depend on the temperature and strain rate imposed. In general, a
high strain rate combined with high temperature
results in ductile fracture, followed by a large
elongation and neck formation. Additionally,
the grains near the fracture surface tend to be
elongated. On the other side, the combination of
low strain rate and high temperature results in
intergranular brittle fracture, with low elongation or necking. Intergranular fracture in such
conditions normally initiates by grain-boundary
sliding from triple points or at grain-boundary
intersections with second-phase particles, causing cavities on the material microstructure, as
presented in Fig. 22.
Once the crack nucleates, it propagates by
grain boundaries, and given that some significant plastic deformation may take place,
the fracture surface tends to exhibit grains of
equiaxial shape. Therefore, to increase creep
strength, the material is normally heat treated to
increase the grain size, reducing the ratio
between the grain surface area and volume. In
turbines that work at very high temperatures, the
creep mechanism must be considered. In this
case, the component may be produced from
monocrystals that significantly increase the
creep resistance.
Most creep curves show three distinct stages
(Fig. 23). After the elastic strain, there is a region
of increasing plastic flow at decreasing rate (first
stage), followed by a region of approximately
constant strain rate (secondary stage), and finally
a region of intense increase in the strain rate,
which rapidly extends to fracture (third stage).
Cavities
(a)
Fig. 22
Fig. 21
Aluminum part that suffered buckling
(b)
Intergranular crack formation at high temperature by

grain-boundary sliding at (a) triple points and
(b) inclusions
pg 128
18/8/2008 3:03PM Plate # 0
Residual stresses can play a significant role

in explaining or preventing failure of a component. One example of residual stresses preventing failure is the use of shot peening processes
that increase the fatigue life of a component by
inducing surface compressive stresses.
Unfortunately, there are also processes or
processing errors that can induce excessive
tensile residual stresses in locations that may
promote failure of a component. The internal
state of stress is caused by thermal and/or
mechanical processing of the parts. Common
examples of these are bending, rolling, or
forging a part. Thermal residual stresses are
primarily due to differential expansion when a
metal is heated or cooled. Two control factors
are thermal treatment (heating or cooling)
and restraint. Both the thermal treatment and
restraint of the component must be present to
generate residual stresses. Residual stresses
can result in visible distortion of a component.
However, in the case of residual stresses, the
distortion can also be useful in estimating the
magnitude or direction of these stresses.
Wear-Assisted Failure
Wear may be defined as damage to a solid
surface caused by the removal or displacement
of material by the mechanical action of a contacting solid, liquid, or gas. It may cause significant surface damage, and the damage is
usually thought of as gradual deterioration.
While the terminology of wear is unresolved,
the following categories are commonly used:
adhesive wear, abrasive wear, erosive wear,
fretting, cavitation, rolling, contact fatigue, and
corrosive wear.
Adhesive wear has been commonly identified

by the terms galling or seizing. It is caused by
the material transference from one surface to
another during their relative movement due to a
solid-state welding process. Figure 24 shows a
schematic representation of this process. High
contact pressure among the surface roughness
results in local plastic deformation and points
of microwelding. The movement between the
surfaces causes the rupture of the junctions,
resulting in a rough peak in one surface and a
valley on the other. Eventually, the tip of a peak
may break, and an abrasive particle is formed.
Abrasive wear, or abrasion, is caused by the
displacement of material from a solid surface
due to hard particles or protuberances sliding
along the surface. The particles may be found
free between two surfaces or attached to one of
them, and the wear level depends on the relative
hardness between the particle and the surface
(Fig. 25). The abrasion may also happen due to
the protuberances or sharp asperities on one of
the surfaces in contact. The process of abrasive
erosion may be considered as abrasive wear.
Erosion, or erosive wear, is the loss of
material from a solid surface due to relative
motion in contact with a fluid that contains solid
particles. In this case, the particle is found to be
dispersed in a fluid or gas means, and it reaches
the surface under relatively high velocity
(Fig. 25d). Figure 26 shows the microstructure
of the transversal section of an H11 tool steel
that has been subject to abrasive erosion.
Fatigue wear can be characterized by the
formation of cracks superficially and/or subsuperficially and the removal of posterior
material due to cyclic loading of solid surfaces.
The sliding contact and/or rolling between solids
Adhesion
Fracture X
t = creep rate
Stage I
Stage II
Particle
Stage III
Time
Fig. 23
Schematic strain-time curve at constant load and

temperature showing the three stages of creep
Fig. 24
Transference mechanism of a material from one

surface to another and the formation of an abrasive
particle in the process of adhesive wear
pg 129
18/8/2008 3:03PM Plate # 0
or the repetitive impact of solids and/or liquids

in a surface are responsible for the superficial
fatigue. When two surfaces of this nature interact due to load application, the area effectively
in contact may be very small, resulting in high
compressive and shear stresses that may lead to
crack nucleation. If only rolling is present, the
maximum shear stress takes place just below the
surface, giving rise to cracks that propagate
parallel to the surface and emerge at the surface,
causing part of the material to separate from the
component, as shown in Fig. 27.
However, pure rolling is not found in inservice conditions. Normally, there is some
sliding between the two surfaces, which alters
the stress field due to an increase in the shear
component, displacing the resulting stress closer
to the surface. The cracks start to nucleate on
the component surface, propagating at a very
shallow angle, as shown in Fig. 28.
Fretting fatigue is considered a phenomenon

where the damage is introduced by a conjunction
of events consisting of adhesion, oscillatory
movement of very low amplitude, oxidation, and
abrasion. The small oscillatory movements may
cause points of adhesion on the surface that
eventually break, forming oxidized particles that
Fig. 27
(a)
(b)
(c)
(d)
Schematic representation of contact fatigue under

pure rolling between two surfaces
Fig. 25
Abrasive wear. (a) Free particle between two surfaces. (b) Particle attached to one of the surfaces.
(c) Sharp asperity. (d) Erosion
Fig. 28
Fig. 26
Fractography showing an H11 tool steel that has

suffered abrasive erosion
Damage by contact fatigue in rolling combined with

sliding conditions in gears produced from a quenched and tempered AISI 8620 carburized steel. (a) Transversal
section. (b) Frontal view from a formed cavity
pg 130
18/8/2008 3:03PM Plate # 0
act as abrasives on the surface, since the smallamplitude movements avoid their dispersion
apart from the source point. Figure 29 presents a
micrograph from a plasma nitrided Cr-Mo-V
steel, where a microcrack formed in the fretting
region.
More than one mechanism can be responsible
for the wear observed on a particular part. The
most critical function provided by lubricants is
to minimize friction and wear to extend equipment service life. Gear failures can be traced to
mechanical problems or lubricant failure.
Lubricant-related failures are usually traced to
contamination, oil film collapse, additive
depletion, and use of improper lubricant for the
application. The most common failures are due
to particle contamination of the lubricant. Dust
particles are highly abrasive and can penetrate
through the oil film, causing plowing wear or
ridging on metal surfaces. Water contamination
can cause rust on working surfaces of gears and
eventually destroy metal integrity. To prevent
premature failure, gear selection requires careful
consideration of the following: gear tooth geometry, tooth action, tooth pressures, construction materials and surface characteristics,
lubricant characteristics, and operating environment.
Environmentally Assisted Failure

Corrosion is chemically induced damage
to a material that results in deterioration of
the material and its properties. Corrosion
can seldom be totally prevented, but it can be
minimized or controlled by proper choice of
material, design, coatings, and occasionally
by changing the environment. Various types of
Fig. 29
Fretting fatigue at the surface of a Cr-Mo-V steel
metallic and nonmetallic coatings are regularly

used to protect metal parts from corrosion.
Corrosion may result in failure of the component. Several factors should be considered
during a failure analysis to determine the effect
corrosion played in a failure, such as type of
corrosion, corrosion rate, the extent of the corrosion, and the interaction between corrosion
and other failure mechanisms.
Uniform, pitting crevice, galvanic, and stresscorrosion cracking are the most common types
of corrosion. Uniform corrosion is characterized
by corrosive attack proceeding evenly over the
entire surface area or a large fraction of the total
area. General thinning takes place until failure.
On the basis of tonnage wasted, this is the most
important form of corrosion.
Stress-corrosion cracking necessitates a
tensile stress, which may be caused by residual
stresses, and a specific environment to cause
progressive fracture of a metal. Aluminum
and stainless steel are well known for stresscorrosion cracking problems. However, all
metals are susceptible to stress-corrosion cracking in the right environment.
Pitting corrosion is a localized form of corrosion by which cavities or holes are produced
in the material. Pitting is considered to be
more dangerous than uniform corrosion damage
because it is more difficult to detect, predict, and
design against. Corrosion products often cover
the pits. A small, narrow pit with minimal
overall metal loss can lead to the failure of
an entire engineering system. Pitting corrosion,
which, for example, is almost a common
denominator of all types of localized corrosion
attack, may assume different shapes.
Crevice corrosion is a localized form of
corrosion usually associated with a stagnant
solution on the microenvironmental level. Such
stagnant microenvironments tend to occur in
crevices (shielded areas) such as those formed
under gaskets, washers, insulation material,
fastener heads, surface deposits, disbonded
coatings, threads, lap joints, and clamps. Crevice
corrosion is initiated by changes in local chemistry within the crevice.
Galvanic corrosion (also called dissimilarmetal corrosion or, wrongly, electrolysis) refers
to corrosion damage induced when two dissimilar materials are coupled in a corrosive
electrolyte. It occurs when two (or more) dissimilar metals are brought into electrical contact
under water. When a galvanic couple forms, one
of the metals in the couple becomes the anode
pg 131
18/8/2008 3:03PM Plate # 0
and corrodes faster than it would all by itself,

while the other becomes the cathode and corrodes slower than it would alone.
6.
REFERENCES
1. D. Dennies, How to Organize a Failure

Investigation, ASM International, 2005
2. D.J. Wulpi, Chapter 1: Techniques of Failure
Analysis, Understanding How Components
Fail, 2nd ed., ASM International, 2000,
p 111
3. C.R. Brooks and A. Choudhury, Chapter 1:
Introduction, Metallurgical Failure Analysis,
McGraw-Hill, 1993, p 172
4. R. Graham, Strategies for Failure Analysis,
Adv. Mater. Process. Aug 2004, p 4550
5. D.A. Ryder, T.J. Davies, I. Brough, and F.R.
Hutchings, General Practice in Failure
7.
8.
9.
Analysis, Failure Analysis and Prevention,

Vol 11, Metals Handbook, 9th ed., American
G.F. Vander Voort, Conducting the Failure
Examination, Prac. Fail. Anal., Vol 1 (No 2),
April 2001, p 1446 and Failure Analysis and
International, 2002
A. Tanzer, Determination and Classification
of Damage, Failure Analysis and Prevention,
Vol 11, ASM Handbook, ASM International,
2002
G. Powell, Identification of Types of Failure,
Failure Analysis and Prevention, Vol 11,
Metals Handbook, 9th ed., American Society
for Metals, 1986, p 7581
C. Laird, The Influence of Metallurgical
Structure on the Mechanisms of Fatigue
Crack Propagation, Fatigue Crack Propagation, STP 415, ASTM, p 131168
pg 132
18/8/2008 3:05PM Plate # 0

DOI: 10.1361/faht2008p133

Failure in Steel Forging

Md. Maniruzzaman and Richard D. Sisson, Jr.,
Worcester Polytechnic Institute
Stephen R. Crosby, The Stanely Works
Charlie Gure (deceased)
IN-PROCESS OR SERVICE FAILURES of

forgings may occur for a variety of reasons. The
starting material may be of insufficient quality to
be adequately formed without cracking, or the
forging process may introduce various types of
discontinuities that cause failure during services.
For example, well-known forging-related discontinuities include:

Laps
Bursts
Flakes
Segregation
Cavity shrinkage
Centerline pipe
Parting-line grain flow
Inclusions
Forging discontinuities are discussed in more

detail in the texts on forging (Ref 14).
This article describes six case studies of
failures with steel forgings (summarized in
Table 1). The case studies illustrate difficulties
encountered in either cold forging or hot forging in terms of preforge factors and/or discontinuities generated by the forging process.
Tables 2 and 3 summarize these factors for
cold and hot forging, respectively. Supporting
topics that are discussed in the case studies

include:

Validity checks for buster and blocker

design
Lubrication and wear
Mechanical surface phenomenon
Forging process design
Forging tolerances
As case studies were being selected, each of

the aforementioned supporting topics was
reviewed for any impact that particular study
had on the case being examined. It is a wellknown fact that forging solutions have several
possible avenues to follow. There is no unique
theory in plasticity that leads to the solution.
Most of the work reported here was performed
using the minimum amount of energy to create
the particular product. Factors unrelated to the
deformation process, such as chemistry, microstructure, phase, grain size, segregation, and
prior strain history, are not addressed here.
Instead, factors directly related to the deformation process itself are presented in this abbreviated discussion.
Wear, plastic deformation processes, and
laws of friction are introduced as a group of
Table 1 Failure analysis of steel forgings and components

Case study
Crankshaft underfill
Tube bending
Spade bit
Trim tear
Upset forging
Flow-through laps
and avoidance
Defect
Solution
Unable to fill crankshaft flanges with existing press

capacity
Unable to control exterior wall thinning and interior wall
thickening
Unable to achieve center web thickness at programmed
force and sufficient flow to wings
Forge material tore at trimline when forging was
trimmed immediately following finish forging
Cracking at circumferential bulge after upset
Introduce creep stages for last increment of

displacements
Introduce induction heating and cooling to limit the
heated axial tube length prior to making the bend
Adjust the die angle to create more shear stress, enabling
full flow to the wings
Introduce a delay time after forge and prior to trim,
allowing the forge material to cool and gain strength
Re-examine the strain and strain rate and process map
for stable flow
Replace the input piece with a newly designed preform
piece, following the design procedures given in this
work
Material foldover at tops of rib and flange intersections

and cases of material flow under previously filled
flanges
pg 133
18/8/2008 3:05PM Plate # 0
subjects that have been considered in the case

studies. Added factors that were evaluated in the
case studies are:

Crankshaft underfill: induction coil inside

diameter and stock diameter, equivalent
current depth and subsequent time for conduction to reach a uniform stock temperature, total heating time for scale control,
transfer time to press, and forging force
applied
Tube bending: precise heat input, control of
temperature, and heated axial length of tube
Spade bit: direction of forging relative to
part shape and assessment of shear effect in
extended wings
Trim tear: trimmer tool tolerances, part
temperature, and process time
Upset forge: principal strains and equivalent
plastic strain
Flow through: strains going from round or
flattened piece to finish, and assessing the
need for a more generalized shape for input
to finish die. Preform designstreamline

shape (not the same for aluminum-, magnesium-, steel-, titanium-, or nickel-base
alloys)
Lubrication: select one that provides the
lowest coefficient of friction and other
acceptable properties
Forge process: total process for entire
manufacturing train, including heat treatment and product testing
Forge checking: fixture check for critical
dimensions
Forge tolerances: component to fit the customers assembly
Simulation of process: verify that laws of
plasticity are met
Forging Process Design

Forging process design requires the application of integrated engineering principles that
bring together factors such as:
Relationship between the important subsystem of a deformation system (Fig. 1)
Table 2 Factors in analysis of cold forging failures

Preforge factors
1. Raw materialchemistry, microstructure, mechanical properties, size, surface finish, and cleanliness
2. Shape sequencinggeneral nature of shape to be created; strain, strain rate, and load requirements
3. Forgingequilibrium forging temperature, strain and strain rate, workpiece volume control, forge equipment, loading and transfer
devices, lubrication, parts collection, inspection, and annealing
4. Trimming
Causes of defects during cold forging
1. CrackingThree factors combine to produce cracks: stress from thermal expansion and contraction, hydrogen, and a susceptible
microstructure.
2. Product underfillpoor flow, sufficient volume, and proper distribution
3. Unbalanced forceslaps/lap fillin, nonhomogeneous strain, strain rate, nonuniform microstructure, and work hardening
a. Seamsexternal and internalon or within a metal surface, an unwelded fold or lap that appears as a crack usually resulting from
a discontinuity
b. Inclusionsraw material; internal and external substance that is foreign and insoluble to the matrix; particles of a foreign material
in metallic matrix. Particles are usually compounds, such as oxides, sulfides, or silicates but may be of any substance that is foreign
and insoluble to the matrix.
c. Tearsoccur when the equivalent plastic strain exceeds the capability of the material
d. Entrapped scaleforged in contamination consisting primarily of oxides but can include other products left on metals
4. Strain hardeningincrease in hardness and strength of metals caused by plastic deformation at temperatures below the recrystallization
range; also known as work hardening
5. Flow through/push throughcondition at which excessive material is provided in the preform in error, such that as elements of the
shape are completely filled, such as flanges or rails, the central material continues to displace outward underneath the filled flanges
6. Porosity/voidssmall openings, interstices, or channels within a consolidated solid mass or agglomerate usually larger than atomic
or molecular dimensions
7. SegregationIn the casting process, the solidifying front moves away from the surface of the casting as a plane front, and
lower-melting-point constituents in the solidifying alloy are driven toward the center. This is called normal segregation.
8. Internal shearingThis effect can occur when material displacements cause excessive sliding of adjacent volumes of material.
9. Surface impuritiesany foreign substance deposited on the part unintentionally
10. Grain size structureThe number of grains per unit volume and the phase of the material dictate the forging response.
11. Flakes, blistersThese flaws typically result from the raw material or other processing steps but may show up when materials are
forged.
12. Residual stresses/distortionMost materials (especially steels) will have residual stresses after cold forging; distortion occurs when
the stresses are not symmetrical.
13. Lubricationdies and workpieceviscosity and flow, hydrodynamics of lubrication, friction, heat generation and power losses,
coefficients of friction
pg 134
18/8/2008 3:05PM Plate # 0
Failure in Steel Forging / 135
Table 3 Factors in analysis of hot forging failures

Preforge factors
1.
2.
3.
4.
Raw materialchemistry, microstructure, mechanical properties, size, surface finish, and cleanliness
Shape sequencingshape nature, temperature, strain and strain rate, upset tooling, fuller (roll), open-die tooling
Hot forgingtemperature, strain and strain rate, forge center cell, loading and transfer device, lubrication, parts collection, and inspection
Trimmingtrimmer unit and capacity, flash removal, temperature trace of product flashline
Causes of defects during hot forging

1. Crackingoccurs when the imposed equivalent plastic strain exceeds the material capability at the temperature of operationsurface
(hot tears), cooling (centerline cracking)
2. Product underfillunderachieved thickness goal, inadequate material displacements, poor 3-D flow, inability of input shape to
subsequent follow-on dies to satisfy local volume requirements, control of centroid path of newly created shapes
3. Unbalanced forceslaps/lap fillin, nonhomogeneous strain, strain rate, nonuniform and continuous microstructure
a. Seamsexternal and internalon or within a metal surface, an unwelded fold or lap that appears as a crack usually resulting from a
discontinuity
b. Inclusionsraw material; internal and external substance that is foreign and insoluble to the matrix; particles of a foreign material
in metallic matrix. The particles are usually compounds, such as oxides, sulfides, or silicates but may be of any substance that is
foreign and insoluble to the matrix.
c. Hot tearsoccur when the equivalent plastic strain exceeds the capability of the material at the temperature of operation
d. Entrapped scaleforged-in contamination consisting primarily of oxides but can include other products left on metals
4. Flow through/push throughcondition at which excessive material is provided in the preform in error, such that as elements of the
shape are completely filled, such as flanges or rails, the central material continues to displace outward underneath the filled flanges
5. Porosity and voidssmall openings, interstices, or channels within a consolidated solid mass or agglomerate usually larger than atomic
or molecular dimensions
6. SegregationIn the casting process, the solidifying front moves away from the surface of the casting as a plane front, and
lower-melting-point constituents in the solidifying alloy are driven toward the center. This is called normal segregation.
7. Internal shearingThis effect can occur when material displacements cause excessive sliding of adjacent volumes of material.
8. Surface impuritiesany foreign substance deposited on the part
9. Grain size structureThe number of grains per unit volume and the phase of the material dictate the forging response.
10. Flakes/blistersThese flaws typically result from the raw material or other processing steps but may show up when materials are
forged.
11. Residual stresses/distortionMost steel forgings will have inherently residual stresses and distortion due to cold straightening
or following quenching.
12. Lubricationdies and workpieceviscosity and flow, hydrodynamics of lubrication, friction, heat generation and power losses,
coefficients of friction
Constitutive
Equation
behavior to achieve stable deformation at

a specified rate and proper evolution of
microstructures and properties
Material
system
Equipment
or
ol
ntr
Co stem
Sy
ka
bil
ity
Forging Tolerances
Control
system
Fig. 1

Relationship between important subsystems of a deformation system. Source: Ref 5
Interdependence of forging process parameters (Fig. 2)

Forging process design task overview
(Fig. 3)
Relationship between process and machine
variables (Fig. 4)
Characteristics of forging machines
(Table 4)
Workability modeling (process maps showing zones of stable flow) of workpiece
The need for verification of the nominal

dimensions and application of forging tolerances is important for quality assurance. Tolerances are required on forged products to allow
for practical variations in die preparation, temperature effects during forging, equipment, and
distortion during and after heat treatment. Forging tolerance review is a basic requirement to
ensure that the part meets the multitude of design
features and tolerances. A listing of the more
important forging tolerances includes:

Dimensionallength, width, center-tocenter, and external-internal

Die
weargenerally
approximately
0.102 mm (0.004 in.)/surface
Die closurethickness of approximately
0.813 to 6.35 mm (0.032 to 0.250 in.) as a
function of plan form area at trimline
pg 135
18/8/2008 3:05PM Plate # 0

Matchalignment of the top and bottom

dies
Radiistrong influence on material displacements

Flash extension
Straightnesstaken as a separate feature
and then assessing its effect on the remainder
of other tolerances
Data on
Billet material
Ram velocity
Strain rate
Billet/Forging
Geometry,
Volume and thickness
Contact time
under pressure
Flow stress/
forgeability
Temperature
distribution in
forging
Die temperature,
cooling
Friction
Conditions and
coefficient
Interface
lubrication
Metal flow
Forging load
Forging energy
Fig. 2
Interdependence of forging process parameters
Fig. 3
Forging process design task overview
pg 136
18/8/2008 3:05PM Plate # 0
pg 137

Draft angle
Datum plane location for three-plane (x, y, x)
setup
Alternative machined tooling points
Finish allowance between forging and
machined part
Tolerance review is conducted in various

ways, and in the past, numerous forgings have
been rejected and held for material review until
some decision could be reached regarding their
disposition for rejection or alternative repair.
Even though this has typically been from review
of the part drawings, another useful way to
assess a completed forging is a fixture check. A
uniquely designed fixture in conjunction with a
dimensional inspector sets the forging into fixed
tooling point locations and proceeds with the
check go or no-go, determining whether
the part will or will not serve its function in the
assembly. In many cases, special fixtures and
Fig. 4
gages can confirm the accuracy of dimensions

that are critical to the function of the component
dimension. Large forgings are good candidates
for fixture checking.
Wear and Lubrication
Surface interactions of two materials are
influenced by small regions where contact is
made at the atomic level. The real area of contact
is determined by elastic and plastic deformation
under consideration of loading. Lubrication
reduces friction by introducing a viscous and
low-shear-strength layer at junctions. Surface
interactions can lead to wear, or the removal of
material as a result of mechanical action.
Wear types include:

Adhesion wear: particle transfers (pulled

off) from one and adheres to the other
Abrasive wear: a hard, rough surface plows
grooves into the softer one
Forging equipment characteristics; relationship between process and machine variables
Table 4 Characteristics of forging presses

Equipment type
Hydraulic press
Mechanical press
Screw press
Hammer
Deformation rate
Slow
Slow to medium
Moderate to high
High
Temperature loss
High
Moderate
Moderate
Low
Consistency
Production rate
Very good
Good
Fair to good
Fair to good
Low
Moderate to high
Moderate to high
Moderate
18/8/2008 3:05PM Plate # 0

Corrosive wear: mechanical action removes

a protective layer from a surface and exposes
it to corrosive attack
Surface fatigue wear: spalling occurs after
the formation of surface or subsurface cracks
Volume wear: proportional to the load and
distance traveled and inversely proportional
to the material hardness
During the early 1950s, the importance of

proper lubrication was recognized on the shop
floor. If an inexperienced oiler inadequately
applied lubricant (in spray or paste form), then
forging problems could occur even for an
acceptable preliminary workpiece (preform).
Alternatively, a questionable preform for the
first closed impression die would be proven
acceptable if an experienced oiler knew where
the lubricant should be applied over the die
impression and also when the die impression
needed additional heavy lubricant in given die
locations that appeared difficult to fill.
These anecdotes vanished quickly as more
science replaced art in forging. Presently, there
are numerous ways that lubricants are used in the
forging industry. Wrapping the workpiece during heating is an approach to prevent the formation of scale in the case of steel or thin metal
sheets or cloths with impregnated graphite, in
addition to the automatic spraying of lubricants.
Lubricants play an important part in forging by
minimizing the load required for maximizing
material flow, protecting the die surface finish
(critical for a specific lubricant), and assisting
the entire forging process.
Lubricant performance factors include:

Adequate lubricity
Stability in gas-fired and electric furnaces
Protect stock against atmospheric contaminants
Provide good surface finish
Act as a release agent
No buildup in die cavity
Ease of application and removal
Conform to Environmental Protection
Agency (EPA) and Occupational Safety and
Health Administration (OSHA) requirements
Acceptable cost
Compatible with die materials
Graphite products for forging lubrication

are:

GPCfor hot and warm forging

Die lubricantsGP series

GP 100low dilution ratios and spray

application
LSoil and water
Precoat workpiececontains graphite as a
lubricant pigment
Adhesion colloids are reliable for high

pressure and temperature. Types include:

Colloidaldispersions
Delta forge lubricantsfor hammer, press,
and upsetters
Deltaglazeprotective lubricants for billets
applicable to steel
Case Studies
Case Study 1: Crankshaft Underfill. There
are several large steel forging components, such
as ship crankshafts and airplane landing gear,
being manufactured successfully in the United
States and throughout the world today (2008).
Crankshaft forgings in the weight range of 2268
to 4536 kg (5000 to 10,000 lb) are products
made by a forging process creating a pair of
flanges and a pinion shaft diameter at one time.
The inboard and outboard flanges along with the
pinion diameter become integral parts of the
main shaft diameter.
The forging operation creates one set of
flanges by means of a working stroke in line with
the major shaft diameter, while a 90 off-set load
forges the pinion shaft between the flanges.
These operations are generally performed following one local heating of the starting bar
diameter for forging a set of crankshaft throws,
including the two flanges and an offset pin diameter. The forging process is repeated until all
of the flanges plus the offset pinion diameter are
created along the major diameter of the crankshaft. The nature of the ready-for-assembly finish forging design for the incrementally forged
crankshafts includes locations where material is
provided for machining along with selected
as-forged surfaces.
During the forging of the flanges, there had
been cases of small amounts of underfill at the
flange extremities, as shown in Fig. 5. That
extent of underfill has caused the entire component to be rejected.
A test run was planned to measure material
displacements while the flanges were being
forged at the prior selected process variables of
strain, strain rate, temperature of workpiece and
dies, and forging force exerted. The conclusion
pg 138
18/8/2008 3:05PM Plate # 0
Fig. 5
(a) Crankshaft flanges not filled. Main shaft diameter shown between flanges of adjacent throws. (b) Crankshaft flange with
left side filled and right side not filled. Pinion shaft diameter located between flanges of single throw
reached was that since the workpiece material

temperature at the end of the press working
stroke was still in the hot working range, an
extended length of time with the maximum force
applied would be helpful to displace the relatively small amount of missing material into the
remote regions of the flange dies. The thought
was that allowing material to creep would aid in
the final filling of the die cavities.
Creep is an example of viscous flow and is
defined as continuing flow at constant stress. At
characteristic stresses, the creep strain reaches a
steady state in which the rate of straining is
constant. This is called the steady-state creep
rate, e_ (or d e_ =dt). In hot working, the relationship between temperature (T), stress (s), and
strain rate (e_ ) in the steady-state condition is best
expressed as:
0
e_ =A(sinh as)n exp (7Q=RT)
where A, a, and n0 are temperature-independent

constants, Q is the activation energy, R is the
universal gas constant, and T is the temperature
in Kelvin.
At low stresses characteristic of creep (as5
0.8), this equation reduces to:
0
e_ =A0 sn exp (7Q=RT)
which describes the relationship among three

variables under creep conditions.
Evaluation of experimental data of the activation energy, Q, indicates that some metals in
hot working soften the recovery process of
repolygonization, and others soften by dynamic
recrystallization. Thus, there is a distinct correspondence between hot working and viscous
creep deformation.
Fig. 6
Both crankshaft flanges filled. Pinion shaft diameter

located between flanges of single throw
In this case study, several time elements (all

less than 60 s) were established in subsequent
trials where all process variables were monitored (including a lower-than-press-capacity
force), and flange fill results were measured.
Finally, the proper combination of the important
process variables, of which temperatures played
an important part, enabled the consistent filling
of the flange extremities, as shown in Fig. 6.
Case Study 2: Tube Bending. Bending
large-diameter tube (310 stainless steel) created
a 90 bend in the diameter range of 635 to
762 mm (25 to 30 in.) and at a nominal wall
thickness of 12.7 mm (0.50 in.) This offered a
major challenge to manufacture. The challenge
was to create a 90 angle bend without excess
thinning at the outer wall and excess thickening
at the inner wall.
Earlier efforts had centered on using gas
heaters around the circumference of the pipe
pg 139
18/8/2008 3:05PM Plate # 0
and, in some cases, heating inside the tube.

However, success was limited since the axial
length of the heated zone was excessively long
to permit control of the incremental bending
strain and strain rate.
As developments continued, a European
company (Cojafax) had designed and manufactured a large bending machine to handle a
762 mm (30 in.) diameter stainless steel as the
input and to impart a 90 bend with controlled
thinning/thickening of the tube walls. An
external axially thin induction coil supplied the
heat and enabled control of the heated zone
while a 90 bend was being made.
The approach in the manufacturing process
was to exert an axial force in line with the
straight tube axis and then push the tube for a
low strain rate through the induction coil until
the heat input was sufficient to cause material
displacements through the heated zone. In some
cases, cooling rings were added to control the
axial heated length, as shown in Fig. 7. This
process was used for a number of trials in an
attempt to achieve the goal of a 90 bend on a
762 mm (30 in.) diameter tube with 12.7 mm
(0.50 in.) initial wall thickness.
Fig. 7
During these trials, process data were generated to show the variables that were used, such as
machine force, bending moment, axial tube
length inside the induction coil, tube axial speed,
and amount of heat supplied, all of which were
used to control the shift of the tube neutral axis
and thus the thinning and thickening of the tube
walls.
Optimizing the process variables for a 90
bend on the 762 mm (30 in.) stainless steel tube
resulted in minimal thinning of the outer wall, as
shown in Fig. 8. The thinning of the outer wall of
several tubes met the initial program goal of
18%.
Case Study 3: Spade Bit. The cutting end of
a proposed wood boring spade bit (AISI 1000
series) consists of a central web connected to
angular extensions from each side of the web, as
shown in Fig. 9. The forging process for the
spade bit forge was designed so that the finish
shape could be cold forged in a continuous
line, starting out with wire, straighten, clean,
lubricate, room-temperature forge, trip flash,
and followed by heat treatment. The inherent
tooling design presented an opportunity to run a
series of numerical experiments using computer
Closeup of tube bending assembly. (a) Induction coil and water ring. (b) Induction coil and partially heated tube and water ring
pg 140
18/8/2008 3:05PM Plate # 0
Bent product tube with 90 bend and minimal

thinning at outer tube wall of ~18% of starting tube
wall thickness
Fig. 8
Fig. 9
(a) Spade bit drawing. (b) Photograph of each half
simulations varying the forging direction, so that

the direct compressive force applied by the dies
to the central web of the workpiece would also
then have components of shear acting to cause
material displacements in the extensions.
Several die rotations were attempted to create
the longest wing extensions with a minimal
central web and the lowest forging force. The
fundamental idea behind this approach was that
steels in shear are weaker than steels in direct
compression. Strain-limiting criteria for cold
forging low-alloy steels are shown in Fig. 10.
Initially, analyses were made to determine the
strains to be encountered when forging a round
bar to a flat central section with two attached
wings off at different angles. Shear stresses were
pg 141
18/8/2008 3:05PM Plate # 0
determined by Mohrs three-dimensional (3-D)

analysis to quantify the stresses and strains
throughout the central web and extended wings.
Forming simulation software Antares (from
UES-Software, now defunct) was used for the
deformation analysis, followed by a shop trial.
Strains in the wings as well as in the center
section were calculated along with the compressive force and stress on the center section,
before/after which simulations verified the analysis. Following a few shop trials supported by
simulations, an optimized process was determined that led to the minimum amount of energy
to be used to forge the spade bit center and
wings.
Case Study 4: Trim Tear. During the forging
sequence of a typical mechanic tool product
(AISI 4000 series), the process was running
satisfactorily, except that tearing occurred at the
flash trimline and then propagated into the forging proper. Trimming of the flash around the
perimeter of a finish forging has traditionally
been a very dependable operation by maintaining the proper clearance between the punch and
the trim tool blades, so that the deformation zone
between the punch, forging with flash, and the
trimmer sheared the flash with no bending of the
flash extension.
Fig. 10
Strain-limiting criteria for cold forging low-alloy steels
Excessive clearance encourages material

bending displacements, which lead to shearing
in the zone. Figure 11 depicts a schematic of the
effect of punch-to-die clearance on characteristics of edges of holes produced by piercing a
low-carbon steel. In any event, the flash shearing
is not clean and adjacent to the draft wall of the
forging. In some cases, the shear mechanism
initiates a crack on the edge of the forging and
flash. The case that is being reported here is one
where cracks initiated and propagated into the
forging proper. Micrographs of crack formation
at the flash edge are shown in Fig. 12.
During the search for a single or multiple
solutions, several conventional avenues of
attack were followed:

Trimmer tool setup for proper clearance at

the trimline, seating of the forging inside the
trip plates, and proper contact of the workpiece in the trim plate nest should be assured.
Raw material condition did show some evidence of banding and inclusions.
Temperature in the workpiece may be too
high due to high speed of production and
resulting lower strength at trim.
Measured clearance between punch and die
is reduced.
pg 142
18/8/2008 3:06PM Plate # 0
pg 143
The final analysis concluded that the temperature of the trimming operation (immediately
following forging) was excessive and triggered
the initiation of a small crack. During further
trials in the shop, the time between forge and
trim was increased. This made the workpiece
have a higher strength because of the lower
temperature, and no cracking occurred even
though there was evidence of material banding.
Other changes were made, such as decreasing
the clearance between the punch and die from
0.127 to 0.076 mm (0.005 to 0.003 in.), which
made the trimming operation cleaner and more
robust.
Case Study 5: Upset Forging. One of the
most important operations in the forging process
is the upsetting of a billet of material with ratios
of axial length to diameter of less than 3 to 1 and,
Edge characteristic
Fracture angle
Rollover(a)
Burnish(a)
Fracture
Burr
preferably in some cases, 2.5 to 1 using flat dies.

The material displacements are primarily radial,
extending out from the billet center and forming
the outside diameter. Material radial displacements come from the decreasing axial length of
the starting billet. Material displacements during
upsetting have been well analyzed in the past by
numerous investigators who have identified the
steel alloy; axial, radial, and tangential strain;
strain rate; temperature of the workpiece and
die; along with the frictional effect at the
workpiece-die interface. During an extended
working stroke, there is a period of time where
the entire reduction operation is considered as
one of nonsteady state. During this operation,
there is one plasticity law that states that the sum
of the principal strain at any time is equal to 0.
This has proven to be a significant benchmark to
Type 1
Type 2
Type 3
Type 4
Type 5
1416
1020% t
1020% t(b)
7080% t
Large, tensile plus
part distortion
811
810% t
1525% t
6075% t
Normal,
tensile only
711
68% t
2540% t
5060% t
Normal,
tensile only
611
47% t
3555% t(c)
3550% t(e)
Medium, tensile
plus compressive(g)
...
25% t
5070% t(d)
2545% t(f)
Large, tensile
plus compressive(g)
(a) Rollover plus burnish approximately equals punch penetration before fracture. (b) Burnish on edge of slug or blank may be small and irregular or even absent. (c) With
spotty secondary shear. (d) In two separate portions, alternating with fracture. (e) With rough surface. (f) In two separate portions, alternating with burnish. (g) Amount of
compressive burr depends on die sharpness.
Fig. 11
Piercing of low-carbon steels. Source: Ref 6
Fig. 12
Micrographs of crack formation at flash edge
18/8/2008 3:06PM Plate # 0
evaluate the effects of the lubrication at the diematerial interface.

Many forged circular components and also
circular components with off-center port bosses
begin as a billet of material of a given diameter
and are then forged between flat dies to a larger
diameter and a reduced axial dimension. Subsequent forging operations on the forged initial
preform create cylindrical walls along with
various configurations, depending on the final
forging design.
Thus, an initial flattening (or sometimes
referred to as pancaking) between flat dies is a
common initial forge operation for a number of
steel components, for example, missile cases
and bowl shapes. A typical steel upset forged
between flat dies made on a screw press is shown
in Fig. 13. However, during the upsetting
Fig. 13
Steel upset forged between flat dies made on a screw

press
operation, cracking on the barrel of the upset

piece often occurs because of higher strain levels
than the material can sustain and, in particular,
higher effective strain levels when each of the
principal strainsaxial, tangential, and radial
are considered.
The fix for a problem of this nature is to calculate the strain field in advance when plans are
made to create the upset piece. A numerical or
computer simulation of the process can provide
a go or no-go on the planned process. In the
past, there have been cases reported of increased
upset temperature within the workpiece, and
the solution to that problem is a reduction in the
strain field. Most of the cases fall into the
situation of an unanticipated strain-rate effect,
requiring the rate of forging to be reduced.
Table 5 (with Fig. 14) shows some selected
results obtained from an MSC superforge
simulation of forging a flattened disc (pancake)
between two flat dies (Fig. 15). The analyses are
based on plasticity laws. The damage variable is
defined as the ratio of the total cavity area over
the total area found in a representative volume
element. Therefore, the damage variable is a
dimensionless quantity between 0 and 1, where
D=0 describes the undamaged representative
volume element, and D=1 is the failure due to
rupture.
A critical value of damage for multiaxial
states of stresses may be defined as a quantity
that describes the occurrences of measurable
cracks in the material. When the critical value of
damage reaches a certain magnitude, one can
Table 5 Analysis and simulation of upset forging a disc between flat dies
Analytical analysis
Numerical simulation (superforge)
Starting billet size: 101.6 mm (4 in.) diameter and 101.6 mm (4 in.) length
Ending flattened disc size: 198.2 mm (7.8 in.) diameter and 26.67 mm (1.06 in.) length (axial thickness)
Strain:
Axial?ln [1.05/4.00]=1.337
Radial+tangential?2 (ln[7.8/4.0])=2 (0.668)=+1.336
Equivalent plastic strain:

1.389
1.389
Analysis: axial strain (1.337)+radial strain (+0.668)+tangential strain (+0.668) ~0

Plasticity law: The algebraic sum of the principal strains equals 0.
Z stress average (disc center-to-edge readings): determined by maximum force
divided by disc (after working stroke) plan form area (PVA)
1.894 108 Pa (27,469 psi)

1.575 108 Pa (22,843 psi)
1.240 108 Pa (17,984 psi)
Z stress=Forging force /PVA=1.081 106/47.78=1.56 108 Pa (22,624 psi)
(Disc thickness center-to-barrel-edge readings)
Ratio of the average stress exerted during the flattening operation to the yield
strength is 19,809 psi/16,215 psi=1.16, which is in the range of Z stresses
reported.
1.2554 108 Pa (18,207 psi)

1.211 108 Pa (17,563 psi)
1.225 108 Pa (17,766 psi)
Average of all Z stresses taken is 20,305 psi.

This ratio agrees closely with the graph of the average stress/yield strength when
plotting the dimensionless parameter of coefficient of friction disc diameter/part
thickness when forging a flat disc (Fig. 14)
(Disc center-to-barrel-edge readings)

1.162 108 Pa (16,853 psi)
8.247 108 Pa (11,961 psi)
4.466 108 Pa (6,477 psi)
Yield strength: 1.118 108 Pa (16,215 psi)
pg 144
18/8/2008 3:06PM Plate # 0
Fig. 14
Pressure multiplication factors for forging of thin panels. Source: Ref 7
Fig. 15
MSC superforge simulation. Disc upset forged

between flat dies, showing (a) start position and
(b) end position after 74.93 mm (2.95 in.) stroke
deduce that the material is irreversibly damaged.

A typical result plot is shown in Fig. 16.
Case Study 6: Avoidance of Flow Through,
Lap, and Crack. During the forging of an Hshaped cylinder of a flat web with projecting ribs
and flanges, material displacements are required
to turn 90 in the direction from the web within
the die cavity to fill the external/internal ribs and

flanges. There have been numerous cases in the
past where insufficient thickness of the material
front moving horizontally caused the front to
contact the die wall and upset on itself, enabling
the filling of the outer flange.
Following the filling of an outside flange, as
shown in Fig. 17, material displacements continue to move outward toward the flash opening
and underneath the filled flange. This combination of material displacements causes a flow
through at the base of the flange.
Another common case of underfill coupled
with lap formation in many forging designs
occurs at the intersection of outer flanges and
cross ribs, where material displacements are
primarily 3-D. Additional material is required to
fill the top of the flange and rib intersection
because of the volume required. That additional
volume is provided by a preform shape with
increased fillet radii or taper at the base of
the web.
The term lap describes a defect that forms
whenever material folds over itself during the
forging of a new shape, using a previously
designed preform as the input to follow on the
set of dies. Laps occur when both vertical and
pg 145
18/8/2008 3:06PM Plate # 0
Fig. 16
Typical state of strain in hot upset forging of steel showing fracture criteria in MSC superforge simulation
Fig. 17
Typical forging defect caused by excessive natural

flow through the forging of a rib (flange)-and-web
part. Flow through is the tendency of a metal to flow naturally past
the rib (flange) opening.
horizontal sections intersect. When this occurs,

it is an indication that a preliminary shape is
required as input to the next die in the forging sequence to provide material to fill intersecting elements. Also, these types of problems
are analyzed by a technique referred to as
streamlining. The series of shapes are tracked

backward from the finish shape to blocker and
preblocker shapes to the cogged, rolled, or upset
piece.
A series of part-way downs were forged
using aluminum alloy 7075 on a hydraulic press
at the temperature of 399 C (750 F) in a set of
preform (blocker) dies that had been designed to
show the material displacement field when
attempting to fill a flange from a web surface.
The samples were approximately 101.6 mm
(4.0 in.) in diameter and 203.3 mm (8.0 in.)
long to minimize the end effect. The strain rate
during the trials was selected so that the displacement field could also be applied to other
alloy systems, such as carbon steels, so that the
same level of forging pressure would be used
with similar results. The results are shown in
Fig. 18.
In the fifth and sixth part-way down, there is
evidence that the material front is moving away
from the entrance to the flange, and die contact is
lost until an additional stroke is applied. Then,
the outer die design adjacent to the flash gutter
creates sufficient back pressure to displace
material into the flange. Thus, a shape is made
that will be suitable as an input to the finish dies,
pg 146
18/8/2008 3:06PM Plate # 0
meeting the web thickness tolerances and filling

the flange.
The process is to reverse-integrate from the
finish design to determine a more generalized
shape with precise volume distribution along the
Fig. 18
three principal axes. Preforms in the series must

satisfy subsequent shapes of the finish product
design and local volume requirements, providing die materials for intersecting product features and other geometric attributesround
Creation of a streamlined preform serving as an input to a finish die rib (flange) and web design, avoiding flow through
pg 147
18/8/2008 3:06PM Plate # 0
(simple and compound), elliptical, and tapered

shapes must all be accommodated.
Admissible criteria and specific tasks for
preform designs include:

Specific alloy characteristics

Microstructure requirements of a finished
product in terms of percent reduction at each
forging operation
Match areas and volumes along principal
axes and location of centroids
Perform reverse-integration, streamlining
the finish shape to more generalized featureslower ribs and rails (flanges) coupled
with increased web thickness and connecting radii
Calculate principal strains when comparing
finish cross sections to preform cross sections
Generate an overall preform shape to obtain
uniform deformation in the finish die
Examine the nature of material displacements over the die contour for unsupported
material fronts as the working stroke progresses
Examine preform locations in the finish (or
subsequent) die and initial die contacts
Examine unsupported webs at die contact to
prevent buckling
Make short plots of the displacement field in
terms of the material contacting fixed die
boundaries and the change of shape of the
material front being generated as the deformation progresses
Follow the continuous trace of the displacement field in terms of the material
contacting fixed die boundaries and the
change of the shape of the materials front
being generated as the deformation progresses
Determine the amount of energy expended
for each preform evaluated and then the
entire shape sequence
REFERENCES
1. G.E. Dieter, H.A. Kuhn, and S.L. Semiatin,

Ed., Handbook of Workability and Process
Design, ASM International, 2003
2. Metalworking: Bulk Forming, Vol 14A, ASM
Handbook, ASM International, 2005
3. T. Altan, G. Ngaile, and G. Shen, Cold and
Hot Forging: Fundamentals and Applications, ASM International, 2005
4. J.E. Johnson, Ed., Forging Industry Handbook, Forging Industry Association, Cleveland, OH, 1966
5. H. Gegel, G. Huang, and S. Manna, Precision ForgingQualityProductivity
EquipmentA Technical Article, UES
Software Inc., Dayton, OH
6. Piercing of Low-Carbon Steel, Metalworking: Sheet Forming, Vol 14B, ASM Handbook, ASM International, 2006, p 159
7. M.D. Stone, The Design and Construction
of Large Forging and Extrusion Presses
for Light Metals, United Engineering and
Foundry, Pittsburg, PA
SELECTED REFERENCES
J. Burke and V. Weiss, Advances in Deformation Processing, Sagamore Army Materials Research, Vol 21, Army Materials and
Mechanics Research, Massachusetts and
Syracuse University, NY
H. Chandler, Metallurgy for the Non
Metallurgist, ASM International, 1998
G.E. Dieter, Mechanical Metallurgy, 3rd
ed., McGraw-Hill Book Co., 1986
D.D. Fuller, Theory and Practice of Lubrication for Engineers, John Wiley and Sons,
Inc., Chapman and Hall, Ltd., 1956
C.G. Johnson, Forging Practice, American
Technical Society Publisher, Chicago, IL,
1954
S. Kalpakjian, Manufacturing Processes of
Engineering Materials, 3rd ed., AddisonWesley, 1977
A. Kannappan, Wear in Forging DiesA
Technical Paper, Swedish Institute of
Product Engineering Research, Goteborg,
Sweden, 1969
C. Lipson, WearConsideration in
Design, Residual Stresses and Contact
StressesA Technical Paper, University of
Michigan, Ann Arbor, MI
F.A. McClintock and A.S. Argon, Ed., An
Introduction to the Mechanical Behavior of
Metals, School of Engineering, Massachusetts Institute of Technology, Cambridge,
MA, 1962
Research Report: Work at IIT Research
Institute, Committee of Hot Rolled and
Cold Finished Bar Products, American Iron
and Steel Institute, New York
A.M. Sabroff, F.W. Boulger, and H.J. Henning, Forging Materials and Practices,
pg 148
18/8/2008 3:06PM Plate # 0

Battelle Memorial Institute, Columbus, OH,

Reinhold Book Company, 1968
J.A. Schey, Introduction to Manufacturing
Processes, McGraw-Hill Book Company,
New York, 1977
J.A. Schey, Ed., Metal Deformation Process,
Marcel Dekker Inc., 1970
J.A. Schey and P.W. Wallace, Research
Report: Metal Flow in Closed Die Forging
of Steel, Part 2: Speed and Lubrication
Effects, American Iron and Steel Institute,
New York, 1966
J.A. Schey, P.W. Wallace, and F.A. Shunk,
Research Report: Metal Flow in Closed Die
Forging of Steel, Part 1: Fundamental
Study, American Iron and Steel Institute,

New York, 1966
T.M. Silva and T.A. Dear, Wear in Drop
Forging DiesA Technical Paper, Department of Mechanical Engineering, University of Birmingham, 1969
J.W. Spretnak, Technical Notes on Forging, Forging Industry Education and
Research Foundation, Cleveland, OH,
1976
Technical Notes: Mechanical and Physical
Properties of Ferrous Forging, Committee
of Hot Rolled and Cold Finished Bar Products, American Iron and Steel Institute,
New York
pg 149
18/8/2008 3:09PM Plate # 0

DOI: 10.1361/faht2008p151

Failures from the Casting Process

Omar Maluf and Luciana Sgarbi Rossino, Instituto de Materiais
Tecnologicos do Brasil Ltda.
Camilo Bento Carletti, Celso Roberto Ribeiro, Clever Ricardo
Chinaglia, and Jose Eduardo May, Universidade Federal de
Sao Carlos
THE HEAT TREATMENT of a steel component is often the last step or near the end of
a somewhat complex manufacturing process.
Finished products require attention to each step
of the long operation chain from raw material to
finished product.
Early in-service failures of components after
heat treatment may result from improper planning, lack of required equipment, nonqualified
personnel, not enough time to execute the
expected operations, or even a combination of
some or all of these deficiencies (Ref 1). However, most of the early failures that happen
during the heat treatment process are the result
of features generated in previous manufacturing
stages. A component lifetime basically depends
on the following factors:

Global component project

Materials selection
Material quality
Processing methods, such as casting and
machining operations prior to the heat
treatment
Heat treatment
Final finishing operations
Mechanical solicitation of the component
and the service environment
However, it is not a simple task to identify

which of these items is responsible for the early
failure of a component during heat treatment or
in service. The failure analyst uses these seven
items as a guide in a failure analysis. This
chapter deals specifically with improper casting
projects and those features that originated in
the casting process itself, including porosity
(generated by the presence of gas as well as by
shrinkage pores), decarburization, cold joint,
and inclusions. These features may not be
called defects because, according to ASTM
E1316-2005, Standard Terminology for Nondestructive Testing, components have defects

only when they fail to meet their specification
requirements. If a component has a large amount
of porosity, for example, it is not a defect unless
(1) an inspection porosity is specified, (2) its
amount exceeds the required acceptance criterion, or (3) the component fails because of this
porosity. This chapter describes cast steel features that may be identified or attributed to
component failure during heat treatment or
subsequent processing or service. As such, these
casting features are referred to as defects in this
chapter.
Failures due to Improper Cast Design

The engineers designing job may face
inevitable weak points due to some inherent
component characteristic in use. The engineer
should, however, try to overcome these inconveniences by looking for alternative solutions
and finding the middle ground between the
component functionality and the manufacturing
difficulties. The project aspects that should be
avoided at all costs in the cast component production are (Ref 27)

Sharp edges, sharp corners, and nonround

edges
Abrupt section changes
Holes, especially when located near the
external wall of the component
Sections with cross connections
Unfavorable length/width relationship
Rounding the corners, as shown in Fig. 1,

should always be performed in order to avoid the
stress concentration that can originate from
pg 151
18/8/2008 3:10PM Plate # 0
cracks formed during casting solidification

in the mold, during heat treatment, especially
quenching and tempering, or even during heating for austenitization (Ref 2).
The risks will be reduced if steel with
increased temperability is chosen, which requires a less severe quenching medium, such as
oil or air. Other strategies in the design of cast
tooling are to avoid creating components with
right angles or to machine the corners to make
them round. Another option is to quench and
temper the component and then remove the
exceeding material to give the component sharp
corners if they are required for its function. This
last strategy requires a steel with good temperability. Otherwise, when the exceeding material
is removed, that region in the component will
present a surface with lower hardness and less
resistance to abrasion than the previous one.
The quenching of components with abrupt
section variations in a liquid environment
always represents a serious problem due to the
associated stress concentrations, even if the
transitions are made using the apparently correct
concordance radius resource (Ref 4, 7). In this
case, the solution is to create the component
in different parts, treat them separately, and
assemble later on. However, if the component
must be made as one unit due to a functional
imposition, the solution is to choose an airquenchable steel that presents a lower crack
probability.
The existence of holes raises a problem
mainly in high-carbon steels and/or alloying
elements. The abrupt section variation and other
specific aspects of the holes (Fig. 2) must be
considered. The accumulation of quenching
liquid in the interior of blind holes leads to an
improper heat loss of the internal walls, thus
lowering the hardness. In open-ended holes, the
heat removal may not be as effective as in the
rest of the component, which is more exposed to
the quenching liquid (Ref 5). Therefore, when
the chosen steel is quenched in a liquid medium,
the components containing holes must be

quenched in specially designed devices so that
they receive strong gushes of liquid in the
interior (of the holes), or they must be arranged
so that all of the set is subject to a strong stirring.
In cases where the holes do not need to be
hardened, they can be made of low-alloy lowcarbon steel components already inserted during
molding for the casting, for example, in tool
steel components. Another possibility during
heat treatment is to fill the holes from casting
with any material that can totally inhibit contact
with the quenching fluid that would result in
hardening of this region. Regarding stress concentration, it is preferable to have the existence
of a completely quenched hole than the presence
of a mixed structure (hardened and soft). Both
methods, particularly the first, make the
achievement and finishing of holes and threads
easier.
Cotter holes, especially the rectangular section ones, are places with high stress concentration (Ref 2). Therefore, whenever
possible, they should be eliminated or substituted by channel sections, whose locking
efficiency is equivalent, but the stressconcentration factor is three to four times lower.
Another geometry that should be avoided during
project design of the cast tooling component is
the cross type, such as the furnace grid and
heating equipment shown in Fig. 3(a). It causes
serious crack problems in the cross area during
the solidification process while still inside the
sand mold or during heat treatment (Ref 3). The
solution is to use node dislocation, as illustrated
in Fig. 3(b).
Lengthy components with very thin sections
or small diameters show serious bending problems during heat treatments, mainly quenching, even when the steel is favorable to less
severe environments. The problems start from
the moment the component is put into the furnace for austenitization. If it is secured only in
two extremities, there is the risk of deflection
Fig. 1
Fig. 2
Sharp edge elimination. (a) Sharp corners create high

strain concentration. (b) Exaggerated relief causes a
shrinkage cavity. (c) Ideal relief
Types of holes. (a) Blind hole with a parallel bottom.

(b) Blind hole with a steeple bottom. (c) Passing hole
the most economic
pg 152
18/8/2008 3:10PM Plate # 0
Failures from the Casting Process / 153
Fig. 3
Grid crossings. (a) Crossed node region of crack formation. (b) With dislocated nodes, the occurrence of cracks is less likely if
the distance between nodes, d, is larger than 2r+e, where e is the thickness, and r is the curvature radius.
due to its own weight. If it is supported on the

furnace hearth, its heating will not be homogeneous, making the component subject to
bending and/or to the appearance of soft spots
(Ref 6). The correct heating method, in this case,
consists of hanging the component by one of its
extremities and using a furnace that allows the
austenitization of the component hung in the
vertical position.
Also to be avoided is the manufacturing of
too thin, lengthy components. They should be
split into components whose length-height ratio
is more favorable (Ref 3). For example, highalloy, high-carbon, steel sheets used in the
guillotine have been replaced by shorter sheets
that, after quenching and tempering are assembled in a chassis to present a continuous edge.
An advantage of this solution is that only the
damaged component of the edge can be replaced
when there is an in-service failure of one of the
sections, leading to an easier and cheaper
operation.
Very big components with a circular section
larger than 25 cm (10 in.) in diameter, or rectangular with equivalent mass, also present
problems during quenching. When carried out in
a liquid medium, the surface reaches the starting
martensitic transformation temperature long
before the central region, with the generation of
stress that may cause internal cracks as a consequence. It is recommended that big, bulky
components be replaced by sets of smaller
Fig. 4
Crack resulting from the normalization heat treatment

of an AISI 1045 steel cast hull caused by thinning of the
wall due to deficiency of the tooling or the core alignment
pieces, despite the adjustment problems that

result from this operation.
Figures 4 to 6 illustrate failures that happened
during the heat treatment operation due to poor
design considerations. Figure 4 is a crack that
occurred during the normalization heat treatment of an AISI 1045 steel cast hull. A prior
thinning of the wall (due to deficiency of the
tooling or core alignment) promoted the cracking during normalization. Figure 5 is a crack that
happened in the normalization heat treatment,
caused by stress buildup in the sharp edge
region. In Figure 6, poor design of an edge led to
cracking after quenching.
pg 153
18/8/2008 3:10PM Plate # 0
Effects due to Porosity

During the production of casting components,
the interactions between the liquid metal and
elements from the gaseous atmosphere, furnace
refractory, foundry ladle, molds, and core
materials are important factors from a technological and metallurgical point of view and are
responsible for desirable or undesirable changes
in the chemical, physical, and mechanical
properties of the metallic materials. The dimensional precision grade of the feeding system
should also be taken into consideration.
Therefore, the quality of the casting product
is related to the physical integrity of the
component, that is, the absence or minimization

of the quantity of defects present. Among these
defects, the most important ones are those generated by the interaction of gas and metal that
promote the appearance of voids. In general,
there are two kinds of voids: those generated by
gas, and shrinkage pores.
Porosity Caused by Gas
One of the factors that must be considered
in steel casting is the behavior of the gases in
the process. Generally, there are three major
sources that may contribute to porosity formation (voids caused by gases) in steel castings.
These are:

Fig. 5
Crack in the bottom of a machine molded from AISI

1030 steel that happened in the normalization heat
treatment, caused by stress buildup in the sharp edge region
Fig. 6
treatment
Plastic injection mold casting in AISI H13 steel with

a crack from a sharp edge after a quenching heat
High initial gas content of the melt originating from the charge ingredients, melting
practice, or atmospheric humidity
Reaction of carbon and dissolved oxygen
under certain melt conditions
Mold-metal reactions between the evolved
mold and core gases at the solidifying casting surface
In addition, any combination of these three

sources may have an accumulative effect in
promoting porosity formation. However, the
gases normally held responsible for subsurface
porosity defects are nitrogen and hydrogen.
Types of Gas Porosity Defects. Pinholes
and blowholes are the two main kinds of porosity caused by the presence of gas (Ref 8).
Gas porosity (pinholes) refers to hydrogen,
oxygen, and nitrogen gases within a casting.
Molten metal has such an affinity for H2, O2, and
N2 that it will disassociate it from other molecules, such as water or atmosphere gases, and
form a solution with it. As with most solutions,
as the temperature drops, these gases become
less soluble and precipitate as gas. The greater
the amount of gas in the molten metal and the
slower it solidifies, the greater the gas voids. It
should be remembered that the H2 comes from
the mold humidity, when the H2 from the metal
is eliminated in the cleaning process performed
before pouring. These voids are generally
smooth, round, or slightly elongated and may be
somewhat localized in the areas of the casting
that solidify last. This type of porosity is generally undetectable visually, since the surface
of the casting solidifies the quickest, preventing
the gases from forming holes large enough
to be visible on the surface, except through
pg 154
18/8/2008 3:10PM Plate # 0
fluorescent-penetrant inspection or crack detection during or after the heat treatment.

Gas holes (blowholes) are generally larger
and more localized voids than gas porosity, but
they retain the smooth, round, or slightly elongated shape. They are usually caused by a
reaction in the mold medium, producing gas that
bubbles through the molten metal. The humidity
contained in the mold walls and cores is the main
source of the vapor that is necessary for defect
(gas bubble) formation in the casting component. During mold filling, the gas generated by
the metal-mold reaction is eliminated to the
environment through permeability, a hole from
the exit of gases, and/or a rising gate. The residual quantity of gas that could lead to bubbles is
almost nonexistent or negligible. The exception
would be the use of low-permeable molds, for
example, the ones whose sand contains a high
percentage of fines, making the passage of the
gases to the environment difficult. For the cores,
which may become completely surrounded by
liquid metal during mold filling, the problem can
be more serious. Gas elimination to the exterior,
including the gases generated by the binder and
collapsible materials, is extremely difficult. It
may require the use of devices such as internal
wax wicks in all the core extensions, so the gases
are sucked toward the core prints and then
eliminated. A gas bubble can also occur, even
though it is not very common, as the result of an
inadequate measurement of the descent channels, distribution, and attack, which, during
pouring, can cause turbulence in the liquid metal
flow or can cause air to be inhaled to the interior
of the mold cavity, where it mixes with the liquid
metal.
Behavior of Gases. Dissolved hydrogen and
nitrogen in the molten steel can cause a porosity
defect such as a pinhole. The extent of gas porosity depends on the amount of these gases, the
alloy, chemical kinetics, and the alloy surface
tension (Ref 912).
During solidification of most steel alloys, the
component that is still liquid becomes more
concentrated in alloy elements due to its solubility. This solubility difference is expressed by
the component ratio KS/L, which is a relation
between the quantity of solute present in the
solid and in the liquid. For most alloys, this value
is usually lower than 1, which indicates a liquid
enrichment during solidification.
The hydrogen solubility in the austenite is
nearly 7 ppm, meaning that its solubility in the
molten and solid conditions is approximately the
same, and there is low hydrogen segregation

during the solidification, which reinforces the
fact that the presence of gas bubbles caused by
hydrogen has other causes, for example, the
reaction with the moisture from the mold and/or
the cores (Ref 9).
For nitrogen, its component ratio for stable
and metastable eutectic solidification is 1.9 and
2.2, respectively (Ref 13). This shows that
nitrogen is less soluble in the liquid metal than in
the solid metal. However, it is good to highlight
that in a real situation, the solidification involves
the liquid-solid diffusion and vice versa of a
larger quantity of elements, nitrogen being just
one of them. When nitrogen behavior in iron
alloys is analyzed, the most important element
whose mutual presence must be considered is
carbon. For example, in a 3.8% alloy at 1500 C
(2730 F), the nitrogen solubility at equilibrium
in the liquid metal is 110 ppm. At the time of
eutectic solidification, this value is reduced
in the liquid to 97.5 ppm due to austenite enrichment. In this sequence, because there is
an increase in the carbon percentage in the
liquid, the nitrogen solubility is reduced to
90 ppm. There is nitrogen saturation in the
liquid, and thus, this excess will cause the evolution of the gas and may originate pinholes,
as seen in Fig. 7. When the possibility of pinhole
formation is evaluated, the component pressure
of all involved gases must be considered. When
this sum is higher than 1 atm, pinholes form.
So, in the previous example, the formation of
pinholes could happen with lower nitrogen
concentrations than those mentioned, needing
just the presence of other gases, such as hydrogen or oxygen, for example. It is also important
Fig. 7
Typical morphology of a defect called a pinhole,

caused by gases
pg 155
18/8/2008 3:10PM Plate # 0
to consider the dynamic formation of these

bubbles in pinhole formation. However, the
thermodynamics of this event are very complex
and are not covered in this chapter.
How to Treat Pinhole and Blowhole Problems. In order to minimize or eliminate the
pinhole formation problem, a relatively simple
method can be used: bubbling argon in the
desulfurization reactor or ladle. When the argon
is blown into the bath, the gases in the atomic
form combine themselves on the bubble surface,
forming molecules from the respective gases
(N2 and H2). However, it is known that bubbling
is more effective in hydrogen elimination than in
nitrogen.
One of the most common alternatives to
reduce or eliminate blowholes is to increase
the pouring temperature. There will be a higher
fluidity and time interval for the beginning of
component solidification, giving time for the
gases to escape to the atmosphere. However,
care must be exercised in the decision to increase
the pouring temperature. The concentration
will also be bigger, and there is the risk of
the riser becoming undermeasured, thus allowing the occurrence of a shrinkage cavity, in
addition to the possibility of a molding and core
system collapse, causing other defects in the
casting.
In summary, to minimize the effect of void
appearance from gases, the main measures to be
taken are:

Control of the furnace atmosphere using

vacuum or gases with low solubility values
Develop a project of feeding channels to
avoid turbulence
Use sand molds and cores with the lowest
humidity and the maximum permeability
possible
Use low-solubility gases that, when injected
in the liquid metal, carry the dissolved gases
to be eliminated to the surface
system, directional solidification, and pouring

temperature.
During the liquid-to-solid transformation,
there is a grouping of atoms that forms ordered
structures. In the majority of cases, this transformation is followed by a density increase
(Fig. 8) and thus a shrinkage, because the metal
as a liquid occupies a larger volume than in the
solid state. The defect known as shrinkage pores
can be characterized as the appearance of nonsuperficial cavities in the casting component due
to the lack of predetermined and precise compensation devices for the liquid metal shrinkage
that occurs during solidification and/or the
metallic inserts for directional freezing.
If the concentration is a little higher than the
capacity of the system to compensate for it, or if
thicker pieces of the component work as risers
for thinner components, small, irregular voids
will be formed, as is seen in Fig. 9. However, if
the concentration is much higher than the compensating mechanisms, there are large voids of
irregular shapes on the surface of the component. These are called primary shrinkage cavities or simply shrinkage cavities, as seen in
Fig. 10, and are not discussed further in this
chapter.
In order to better understand shrinkage, an
example is given of the production of a cast iron
component with two kinds of molds: a nonstiff
mold with synthetic sand (green) and a stiff mold
with phenolic no-bake sand. In the one produced
with synthetic sand, there is a factor that must be
considered: the mold walls deform, increasing
the volume of the cavity when it receives the
liquid metal, and thus, it requires more metal to
Porosity Caused by Shrinkage Pores

This type of porosity has a rough, irregular
shape. It is caused by a lack of adequate feed
metal during solidification. This defect is an
internal void known as shrinkage pores. It
usually is detected only through ultrasonic or
radiographic tests or, during heat treatment,
when it causes disruption in the components.
Shrinkage pores are not related to the high or low
presence of any kind of gases but to the feeding
Fig. 8
Density variation with temperature in metals
pg 156
18/8/2008 3:10PM Plate # 0
compensate the increase in the volume. With the

rigid phenolic resin mold, there is no volume
increase in the cavity, and the additional liquid
metal is not necessary. On the contrary, when the
Fig. 9
Example of a shrinkage pore
(a)
equivalent carbon is larger than 3.9%, there is a

graphitic expansion, which is larger than the
solidification shrinkage and could cause a metal
reflux for the mold exterior.
Each metal or metal alloy presents a characteristic concentration rate during the solidification process. Therefore, it is possible to
estimate quite precisely which feeding condition
will be necessary to avoid the occurrence of
porosity and shrinkage cavity problems. The
theoretical calculations to predict metal volume
shrinkage during the casting process are based
on a model proposed by Campbell (Ref 14),
using a sphere as an example.
There are basically three different types of
shrinkage that may occur during the solidification process, as shown in Fig. 11: liquid
shrinkage, solidification shrinkage, and solid
shrinkage. During the casting process, the first
type of shrinkage observed is the liquid one,
which happens with temperature decrease.
However, this does not represent significant
problems in the quality of a casting component
when the volume reduction occurs linearly with
temperature decrease, and the necessary volume
of liquid material to compensate this volume
reduction can be given by risers.
On the other hand, the volume shrinkage that
occurs during solidification of the liquid metal
can bring more serious problems to the casting
component, and thus, it requires more attention.
The major concern is to make the feeding process, which replaces the liquid metal necessary
to compensate the shrinkage in the system, very
precise in a way that allows for the attainment of
perfect components. This shrinkage compensation process determines the precision and perfection of the casting component and is inversely
(b)
Fig. 10
Primary shrinkage cavity forming large voids

of irregular shapes on the component surface.
(a) Schematic drawing. (b) Shrinkage cavity compensated for riser
Fig. 11
Schematic representation of the three regimens of

shrinkage: in the liquid state, during solidification,
and in the solid state. Source: Ref 14
pg 157
18/8/2008 3:10PM Plate # 0
proportional to the quantity of shrinkage cavities

or shrinkage pores present in the obtained
component.
During the shrinkage process that occurs in
the solid state, the component size starts to be
reduced. At this moment, the casting component
faces the resistance of the mold and/or core. This
kind of stress from the casting component, when
trying to contract, generates residual stresses
that may cause plastic deformation of the casting
component, hot tearing, or cracks during heat
treatment later on. Yet, this shrinkage depends
more on the volume reduction intrinsic to the
cast alloy and the project of the mold than on the
casting parameters.
Six Rules for Casting Component Feeding. In the absence of gases and if the feeding of
liquid metal is appropriate, no porosity will be
found in the casting. However, because there are
many complex casting projects, there may be
regions of the mold with feeding problems,
allowing the internal hydrostatic tension in the
liquid metal to generate the conditions for the
formation of internal pores.
In the design of a component to be cast, it is
necessary to have an effective supply of material
in order to compensate the shrinkages previously mentioned. For the additional liquid
metal supplied to the system to compensate the
volume shrinkage that occurs during cooling, a
riser must be provided in the casting of the
component. The use of these risers, also known
as feeders, exothermic sleeves, or hot tops, can
eliminate the problem of shrinkage pores. The
quantity, form, and volume of these risers vary
according to the form and complexity of the
component to be cast. However, despite the fact
that there is a vast amount of literature on the
calculation and quantity of these risers, the
correct location of them depends on the experience of the process controller.
The following criteria, however, are considered fundamental for proper feeding of the
component, and thus, the defects caused by
shrinkage pores are reduced or eliminated:

Thermal transfer criterion: The riser must

solidify at the same time or slower than the
casting.
Volume criterion: The riser must contain
enough mass to fulfill the volume shrinkage
needs of the component.
However, there are still rules that are eventually

observed, and they define additional geometric,
thermal, and pressure criteria that are absolutely

necessary for perfect solidification:

The junction between the casting and the

riser must not create a hot spot. This place
cannot have a larger solidification period
than the riser or the casting component;
otherwise, it can cause the formation of a
shrinkage porosity.
There must be a way in which the liquid
metal of the riser can reach all the required
regions.
There must be a pressure variation in order to
cause a liquid material flow in the right
direction.
There must be enough pressure in all the
regions of the mold to avoid the formation
and growth of cavities.
Internal Porosity Formed from the Surface. If there is not enough internal pressure
inside the component being cast and if the liquid
inside the mold is still connected with the liquid
in the external surface, it can be sucked to the
inside, causing the growth of porosities that are
connected with the surface (Fig. 12), because the
liquid naturally drags air with itself that stays in
the interdendritic spacings of the casting component. This preforming mechanism is much
more common than imagined. It occurs mainly
in alloys with a very long cooling range, when
the development of the dendritic lattice means
that the aspiration of liquid in the neighboring
surfaces becomes easier than feeding from a
more distant point. The point at which the liquid
may be pulled from the surface can be anywhere
for an alloy with a long enough cooling period.
Thus, in an alloy with an intermediary solidification period, the starting point is usually a hot
spot, such as an internal corner or a recess angle
of the component.
The possibility of the connection of two
opposed surfaces in the same component
through the pores is one of the main reasons that
alloys with long solidification times should not
be employed in the manufacturing of components where high working pressures are applied,
such as hydraulic valves or motor cylinder
heads, because they would cause leakages. The
prerequisite in such complex components is that
the interior should have positive pressure in all
points in order to avoid the connection of the
surfaces through internal porosities, which is
rarely achieved.
Internal Porosity from Nucleation. Alloys
with very short solidification intervals, such as
pg 158
18/8/2008 3:10PM Plate # 0
the aluminum, brass, and eutectic aluminumsilicon alloys, do not present the connection
problem between the surface and pores because
they have a perfect solid layer in the first stages
of solidification, while the liquid feeding occurs
through the feeding channels. The internal
pressure decrease due to an inefficient feeding at
the end of solidification can create a pore
through nucleation in the interior of the liquid. In
this case, there is no connection with the external
surface of the casting.
So, in this kind of alloy, the porosity is usually
nucleated and is concentrated near the center of
the component. When it occurs in plates, for
example, it is referred to as axial porosity.
Unless subsequent machining operations pass
through the pores, the casting in such alloys is
tight.
After nucleation, the subsequent solidification
provides the necessary driving force for pore
growth, which, if observed structurally, has
many similarities with the one started from the
surface.
Growth of the Shrinkage Pores. The first
stage of shrinkage pore growth is very fast.
According to Davies (Ref 15), this period should
be less than 60 ms. After this first nucleation
stage, the growth of the pore happens more
slowly, being controlled by the heat extraction
rate of the mold. For the shrinkage pores that
started from the surfacethe primary shrinkage
cavities, for examplethere is not a fast first
Fig. 12
Internal porosity formed from the surface
stage of nucleation. In fact, such a pore or

shrinkage cavity is simply formed as an answer
to the shrinkage of the solidification.
During the solidification of a casting component, the liquid flow from the reservoirs to the
areas that are being solidified make the level of
the reservoirs decrease. At the same time, there
is the advance of the solidification front. This
joint action of decreasing the liquid level and the
advance of the solidification front creates a conic
cavity, as shown in Fig. 13. This cavity is called
a primary shrinkage cavity in order to differentiate it from a secondary shrinkage cavity,
which is porosity islands observed from longitudinal cuts guided according to a line from the
thinnest region of the primary shrinkage cavity.
In fact, these islands are interconnected in the
solid volume and are thus an extension of the
primary shrinkage cavity.
Example 1: Failure Analysis of a Mill
Gear with Defect Caused by a Shrinkage
Pore. The analyzed component corresponds to
a mill gear of 860 by 1900 mm (34 by 75 in.)
external diameter and 1050 mm and 15 teeth, as
shown in Fig. 14. The mill gear was purchased in
2003 and fractured in service during the 2003
2004 harvest (only 3 months working). The aim
of this study was to verify the metallurgical
properties of the material and the causes that
eventually could have contributed to generate
the crack nucleation and the component fracture
after a short period of working.
pg 159
18/8/2008 3:11PM Plate # 0
Fig. 13
Fig. 14
Primary and secondary shrinkage cavities
Aspects of the mill gear as received for analysis
A chemical analysis of the studied component

was carried out by optical emission spectroscopy, and the results are shown in Table 1.
A tensile test was performed according to
ASTM E8M-98, an impact test according to
ASTM E23-91, and a hardness test according to

ASTM E0-96 (HBS 2.5/187.5/15). The results
are shown in Table 2.
The fracture, as can be observed, occurred
from the bottom of the tooth, propagating toward
the internal diameter. The visual aspect of the
cracked surface is an indication that the crack
occurred by nucleation and propagation of the
cracks through cyclic efforts (fatigue) of the unidirectional type. The final crack happened
after the longitudinal section had approximately
30% of its area taken over by cracks, as seen
in Fig. 15. The area of crack propagation by
fatigue indicates that there was propagation
during a relatively short period of the milling
operation. Most likely the component started
the operation already cracked, that is, with
casting defects of considerable dimensions near
the surface (subsuperficial), causing the cracks
to arise and propagate just at the beginning
of the harvest, which caused its fast milling
rupture.
pg 160
18/8/2008 3:11PM Plate # 0
pg 161
Structural analyses were performed on samples from the fracture region and on the tooth
radio adjacent to the fractured region (Fig. 16,
17). It can be noted that the colony of subsurface
shrinkage cavities/porosity connected, which
probably caused the crack nucleation.
It could be concluded that the failure cause
was related to casting defects, such as connected
shrinkage cavities and porosity colonies, associated with tensile loads applied during the mill
gear operation, which caused crack nucleation.
Figure 18 shows the crack starting point, proving
the failure cause.
Example 2: Failure Analysis of a Mill Gear
with Defect Caused by a Shrinkage Pore.
Figure 19 shows a sliced sample of the mill gear
with the tooth root used in the failure analysis.
Several mill gear presented had fractures on
several teeth roots after an intermittent loading time. The chemical composition in weight
percent and the mechanical properties of the
steel in Fig. 19 are presented in Tables 3 and 4,
respectively, according to the data provided

by the manufacturer and compared to the
experimental ones. The phosphorus and sulfur
quantities are between the maximum limit
established by ASTM A148-93B, and the
quantity of other elements agreed with the
manufacturer specification. The values obtained
for the yield strength (0.2%) and for tensile
strength agree with the expected ones.
The mill gear fractured through the mechanism of crack propagation by fatigue. The cracks
nucleated from the casting defects, located
mainly in the third component of the width of the
mill gear and near the surface of the tooth root.
The low cooling rate during the solidification
process is probably the main cause of the high
susceptibility to casting defect formation, such
as shrinkage pores. Inclusions and bubbles
represent a small component in the material
embrittlement, since they are too small compared to shrinkage pores. Their presence should
be neglected.
Table 1 Chemical analysis of the studied material

Composition, wt%
Specification
ASTM A148 Gr 105-85
Si
Mn
Cr
Ni
Mo
Cu
Al
0.270
0.477
0.859
0.023
0.017
0.946
1.774
0.242
0.058
0.073
Table 2 Properties of the studied material

Test
Tensile strength, MPa

Yield strength, MPa
Yielding at 50 mm, %
Reduction in area, %
Impact, J
Hardness, HB
Specification
ASTM A148 Gr 105-85
804
679
6.0
10.0
3035
257259
Fig. 15
Aspect of the fracture surface showing that approximately 30% of the longitudinal section had been
taken over by the cracks diffused by fatigue. Many subsuperficial
casting defects were also observed where the nucleation of the
cracks started.
Fig. 16
Structural analysis performed on samples from the

fractured region. Etched with 3% nital. Original
magnification: 100
18/8/2008 3:11PM Plate # 0
Penetrating liquid analysis was performed on

the tooth root surface, and the result is presented
in Fig. 20. Localized cracks on the root center
have extended to the sides.
Fig. 17
Structural analysis performed on samples from the

fractured region. Etched with 3% nital. Original
magnification: 200
Figure 21(a) shows the observed microstructural aspect, where it is possible to note the
presence of several cracks in the sample interior.
These cracks were nucleated in several shrinkage pores and have propagated by the fatigue
mechanism to the surface during cyclic loading.
The aspect of the fracture is mainly transgranular, which suggests that the material was not
embrittled by drawing back. Beyond the casting
defects, inclusions were observed in the sample,
some of them with sharp forms, as shown in
Fig. 21(b). It is important to note that this kind of
shape is undesirable since it is a potential cracknucleating site due to stress concentration. The
presence of sharp inclusions indicates that the
globalization process during casting was not
totally efficient.
The proposed corrective actions include:

Increase and standardize the extraction heat

rate from the casting mold in the region next
to the tooth root
Improve the degassing process and impurity
control of the casting material
Increase the thickness of the on-metal along
the region of the tooth root, with the goal of
increasing the probability of defect elimination during the machining process
Effects due to Decarburization

during Microfusion
Fig. 18
Region adjacent to the fractured region showing a

transgranular crack generated in the casting process
and masked by material deformation during the radio machining
process, with propagation directed to the internal diameter
Fig. 19
Sample that was analyzed
Among the several kinds of defects that may

occur during the casting process and that are
detected after heat treatment is the surface
pg 162
18/8/2008 3:11PM Plate # 0
pg 163
Effects due to Cold Joints
decarburization layer. It occurs in carbon steel

components cast by the lost-wax or microfusion
process. The identification of this defect can
only be made after heat treatments in controlled
atmospheres; otherwise, this identification is
impossible, since the decarburization can also
come from the heat treatment.
This decarburization results from the presence of atmospheric oxygen that remains in the
mold as a consequence of the inert feature of the
mold and its permeability in relation to the surroundings. Some analysts have measured the
thickness of the decarburized layer in carbon
steels cast by lost wax and discovered that it
increases proportionally to the temperature
increase in the mold and the volume/surface
ratio of the casting component (Ref 14).
Figure 22 shows a microfused decarburized
component that underwent heat treatment.
Cold joint is a kind of defect that happens in

cast components and normally has a significant
effect on the structural integrity of the component. This serious nonconformity happens when:
1) two portions of the metal, each coming from
different feeding/distribution canals of the mold,
meet and, instead of contributing to the formation of a smooth and homogeneous surface,
provoke an undercut discontinuity called a cold
junction; 2) the solidification process occurs too
far from the metal flow coming from the feeding/
distribution place, where the liquid temperature
is lower than the necessary temperature; 3) the
pouring, feeding, and distribution channels are
underdimensioned and strangle the flow of metal
necessary for the filling of the mold; and 4) the
molds have voids that need to be filled with such
Table 3 Chemical composition of the studied material

Composition, wt%
Source of data
Manufacturer
Chemical analysis
Si
Mn
Cr
Ni
Mo
Cu
Al
Fe
0.31
0.31
0.51
0.50
0.78
0.80
0.020
0.018
0.013
0.017
0.76
0.76
1.66
1.66
0.23
0.24
0.06
...
0.046
...
bal
bal
Table 4 Mechanical properties of the studied material

Mechanical properties
Source of data
st, MPa
se, MPa
e at 50 mm, %
Hardness, HBW
Manufacturer
Experimental
777.0
780.1+16.2
626.0
606.7+15.3
19.0
10.2+1.9
38.6
35.9+14.1
228
230+4
Fig. 20
Cracks located in the tooth root revealed by the penetrating liquid technique. Detail of the central region with higher
magnification showing the machining imprints. The arrows show the extreme limits of the cracks.
18/8/2008 3:11PM Plate # 0
a thin thickness that the liquid metal, even at the

appropriate temperature, cannot fill them completely.
The component that shows this kind of defect,
depending on the size and location of the joint,
must be discarded, since recovery with a weld is
not recommended from a metallurgical point of
view or, depending on the cost-benefit ratio,
is not justified. This defect is usually seen, but
Fig. 21
can occur and be unnoticed initially, in components with complex geometry and abrupt variation of mass, where it is used to obtain a large
number of cores that could provide details difficult to be observed by quality control. In these
cases, the defect will only be located when there
are cracks/disruption in heat treatment or leakage and fracture when the component is in
service.
(a) Micrograph showing cracks connecting shrinkage pores (indicated by arrows) in the internal component of the sample.
(b) Detail of the box in (a), where an inclusion is indicated by the arrow
pg 164
18/8/2008 3:11PM Plate # 0
Fig. 22
Surface of a microfused component showing surface

decarburization
In summary, to avoid the appearance of cold

joints in cast components, it is necessary to
control several manufacturing stages of its
design; for example, prevent the component
from having regions with very thin thickness;
appropriate fusion and pouring temperatures for
each component; appropriate mold-filling
channel system; compatible pouring speed; and
well-established necessary amount of liquid
metal for filling the mold to avoid temporary
interruption in pouring.
Inclusions
Inclusions can be defined as nonmetallic and
sometimes intermetallic phases embedded in a
metallic matrix (Ref 16). They are usually simple oxides, sulfides, or nitrides. In almost all
instances of metal casting, they are considered
to be detrimental to the performance of the cast
component. Sometimes, an intentional introduction in larger quantities can lead to unique
dispersion-strengthened materials. There are
essentially two classifications for all inclusions:

Exogenousthose derived from external

causes
Indigenousthose that are native, innate,
or inherent in the molten metal treatment
process
Slag, dross, entrapped mold materials, and

refractories are examples of inclusions that
would be classified as exogenous. In most cases,
these inclusions are macroscopic or visible to
the naked eye at the casting surface. When the
casting is sectioned, they may also appear
beneath the external casting surface if they
have had insufficient time to float out or settle
due to the density differences with respect
to the molten metal. The presence of these
macroinclusions in steel castings is avoidable,

but their presence has plagued all forms of steel
casting and is particularly problematic in both
foundry processing and in the continuous casting of sheet steels and wire.
Macroinclusions are always practice related,
and analysis of the size and chemical composition of a macroinclusion can lead to the identification of potential sources of this problem.
Once an inclusional source is developed, a clear
and effective process change can be made to
eliminate such problems in the future. Therefore, the techniques already developed by integrated steel manufacturers can be readily
applied to foundries by coupling inclusion
identification with an in-depth study of steelmaking and casting practices in the foundry.
Horwath and Goodrich (Ref 17) and Svoboda
et al. (Ref 18) have studied macroinclusions and
identified that these kind of inclusions can result
in excessive casting repairs or rejected castings.
To reduce these problems, a method was developed to ensure that there are no inclusions in
cast materials above a size that results in failure
during ultrasonic or visual inspection of the
casting. In this method, the macroinclusions
should be eliminated; that is, inclusions greater
than 100 mm must be eliminated, but more
severely, inclusions greater that 50 mm should
be eliminated also.
Sulfides, nitrides, and oxides are examples of
indigenous inclusions that result from chemical
reactions of the molten metal and the local
environment. They are usually very small and
uniformly distributed inclusions, requiring
optical microscopy to visualize them. The presence of these microinclusions in castings is
generally unavoidable (Ref 9), because they
are the natural inclusions that are formed in
liquid steels due to the reaction between
alloying elements and oxygen; however, it is
necessary to minimize these inclusions as a
grain-boundary distribution of these inclusions
can be damaging to the component mechanical
properties.
Clean Steel
Clean steel is the common name attributed to
steel that has low levels of the elements sulfur,
phosphorus, nitrogen, oxygen, and hydrogen, as
well as residual elements copper, lead, zinc,
nickel, chromium, bismuth, tin, antimony, and
magnesium and almost no oxide product defects
produced during the act of steelmaking, ladle
pg 165
18/8/2008 3:11PM Plate # 0
metallurgy, casting, and rolling. Because the

clean concept is not absolute, the cleanliness standard desired by the customer is continuously changing as a function of time and
technological improvements. The term clean
steel is therefore continually variable, depending on the application and the competition
between steel suppliers.
Thus, due to the variability of the term clean,
it is typical to refer to high-purity steels as steels
with low levels of solutes, and low-residual
steels as steels with low levels of impurities. For
example, there are high-purity, low-residual
clean steels, such as ultra-deep-drawing steel
sheets for automobiles, that require ultralow
carbon contents (530 ppm), low nitrogen contents (530 ppm), and the absence of oxide
inclusions with diameters greater than 100 mm;
and there are low-residual clean steels, such as
those used for drawn and ironed cans, that are a
standard low-carbon steel (1006) without highpurity component requirements but are ultraclean, with the requirement that oxide diameters
must be less than 20 mm. In addition, in forging
and bearing grades, there are clean steels that
require strictly controlled inclusion size distributions.
The total inclusion content related to the total
oxygen content has been correlated with bearing
life, and decreasing total oxygen contents
(below 10 ppm) improve the bearing life. In
addition to total oxygen content, the total length
of stringer inclusions after forging is also related
to the bearing life, and, at low total oxygen
levels, efforts to reduce inclusion clustering lead
to very long fatigue life for bearings.
Clean steels can be classified as steels with a
low frequency of inclusions (55 mm). The
major problems in clean steel manufacture are
incomplete separation of clustered solid inclusions (45 mm in diameter), the presence of
sporadic larger liquid inclusions due to emulsification of covering slags, and the presence of
solid materials that originate from the refractories used to contain steels. The equipment used
to produce clean steel varies greatly between
different steel plants; however, current clean
steelmaking and casting practices are based on
the following principles:

The oxygen dissolved in liquid steel at the

melting stage must be transformed into a
solid or a gas and removed before casting.
The external oxygen sources that are responsible for the reoxidation of liquid steel

must be eliminated at every step in the

process.
The physical entrapment of the liquid fluxes
used during steel refining and casting must
be eliminated.
Refractories in contact with liquid steel must
be chemically stable and resistant to corrosion and erosion.
These simple principles are based on the

importance of maintaining chemical equilibrium
between the elements dissolved in liquid steel
and the slag and refractory systems that are in
contact with the liquid steel. Additionally, it is
necessary to control the fluid flow to avoid
conditions at liquid slag-steel interfaces that
could result in the physical entrapment of the
covering slag.
Clean steel manufacture is dependent on an
understanding of the fundamental steps necessary to produce a clean steel:

Generation of the inclusion

Transport of the inclusion to an interface
Separation of the inclusion at the interface
Removal of the inclusion from the interface
The production of really clean steel depends of

the correct application of these principles.
The Formation of Macroinclusions
There are four major methods of forming
macroinclusions, and all problems occur during
foundry processing:

Reoxidation
Interaction between liquid steel and liquid
slags: vortexing, ladle or mold filling, argon
stirring, and pouring through a slag layer
Erosion/corrosion during steel pouring
Inclusion agglomeration due to clogging
during steel pouring
Reoxidation. The major cause of macroinclusion formation in casting is reoxidation

(Ref 1719). To understand reoxidation, it is
necessary to understand that liquid iron is not
thermodynamically stable in the presence of
oxygen. The spontaneous reaction that occurs
results in the formation of iron oxide. As deoxidizers are added, the steel remains unstable in
the presence of oxygen as a gas, but now the
inclusions that form include the oxides of the
deoxidants. Some deoxidants, such as aluminum, magnesium, and calcium, form very stable
oxides that are more stable than some slag and
pg 166
18/8/2008 3:11PM Plate # 0
refractory chemistries. Thus, the steel reacts

with the less stable oxides. Reoxidation can
occur by reaction with:

The ambient atmosphere (air)

The slag components less stable than the
oxide of the deoxidant
The refractories that are less stable than the
oxides of the deoxidant
Interaction between Liquid Steel and

Liquid Slag. Macroinclusion formation can
occur by emulsification of liquid slags or scums
on the surface of liquid steels. All of these types
of defects are practice related and can be solved
by practice changes. The issue in understanding
emulsification is to understand the source of the
energy that allows a buoyant droplet to become
submerged. Generally, this energy comes from
the interaction of a flowing steel stream and a
liquid slag. There are four major sources of this
energy:

Open stream pouring onto or through a

liquid slag (common during lip pouring)
Filling a ladle or mold at too high a fill rate in
the presence of slags or scums
Vortexing during steel pouring from a ladle
Steering in the ladle with gas at too high a
stir rate
Vortexing during drainage in a water model of

a ladle was studied by Sankaranarayanan and
Guthrie (Ref 20, 21). They showed that the
initial rotational velocity at the surface of the
vessel is extremely important in determining
the height at which the vortex will form, and that
increased rotational velocities caused increased
vortex initiation depth. Entrainment due to fluid
flow at the interface has been examined by
Noguchi et al. (Ref 22), who attempted to
decrease the entrainment of slag in low-carbon
titanium-aluminum-killed steels. They noted
that entrainment decreased as the casting speed
was decreased. In a study conducted by Nakamura et al. (Ref 23), it was found that defects that
contained mold slag increased in ultra-lowcarbon grades as the casting rate was increased.
They also reported using as low an argon flow
rate as possible in their submerged entry nozzles
to avoid entrainment. Manabu et al. (Ref 24)
have also documented the existence of a critical
gas flow rate for entrainment in both a silicon
oil-water and a slag-steel system. These authors
mention that the slag depth, slag properties, and
gas bubble diameter play a role. The oil depth
was found to be directly proportional to the flow

rate needed to cause entrainment. The gas bubble size was found to be inversely proportional
to the flow rate needed to cause entrainment.
Manabu et al. (Ref 24), investigated the effect
of oil kinematic viscosities on emulsification
and found that although the kinematic viscosity
was varied by a factor of 10, very little change
was seen in the fluid velocity needed to cause
entrainment. Harman and Cramb (Ref 25),
documented the effect of interfacial tension and
slag viscosity on emulsification phenomena.
Erosion-Corrosion during Steel Pouring. This kind of defect is usually associated
with the higher corrosivity of some steel grades,
because high manganese and grades that are
barely killed and have high soluble oxygen
contents attack the binder or the mold sand itself,
leading to large entrapped sand components.
Reoxidation of steel leads to FeO-based inclusions that are very reactive and wet the materials
of the mold, leading to erosion of the mold in
areas of high fluid turbulence. Of course, sand
that is not pressed, sintered, or bonded in any
way can easily be entrapped in turbulent fluid
flow. Mold binders can also decompose at temperature and release mold components that can
be entrapped. Expansion due to the high thermal
gradients associated with casting can also cause
sand to loosen.
Inclusion Agglomeration due to Clogging
during Steel Pouring. The formation of clogs
when steels containing solid inclusions are cast
can result in quite large macroinclusion defects
if the clogs are released during teaming. All
solid inclusions tend to agglomerate due to surface tension effects. Clogging of pouring nozzles can be the source of large macroinclusion
defects when steels are dirty and pouring times
are long.
The Formation of Microinclusions
Microinclusions are formed due to reactions
between alloying additions and oxygen in molten steel. Their formation is generally heterogeneous or from highly supersaturated areas
during alloy addition. Due to the nature of the
formation of these inclusions (nucleation and
growth), they are generally small (less than
5 mm), unless they agglomerate due to turbulence or grow under conditions of high oxygen
flux. In this study, microinclusions are defined as
those inclusions with diameters smaller than
20 mm. In addition, they are defined as having
pg 167
18/8/2008 3:11PM Plate # 0
pg 168
diameters greater than 1 mm. Table 5 show

typical microinclusions that are found in cast
steels.
Since microinclusions form due to a reaction,
they are driven by thermodynamics; therefore,
changing composition or temperature can lead to
their precipitation. This means they can form in
the ladle, during transport to the mold, or in the
mold during solidification.
indicated. Figure 25 shows a micrograph of the

fractured surface, near the blade bottom. Several
turning gear imprints can be observed, showing
the presence of multiple sites of crack nucleation
Case Studies of Defects Caused

by Inclusions
Failure of a Steam Turbine Rotor Blade.
Possible causes were investigated for failure of
a rotor blade of a 35 MW steam turbine. One of
the rotor blades was fractured after a certain
operation time (Fig. 23). The fracture occurred
at two different regions: at the bottom and at the
top extremity, near the metallic lashing strap.
Both regions have the highest stress concentration due to the blade geometry and loading
conditions. The blade fracture occurred during
the maximum turbine operation. The rotor was
working, with new blades mounted in between
harvests. The blades were manufactured with
steel ingots with the chemical composition presented in Table 6. The specifications for the
mechanical properties of the material at room
temperature are shown in Table 7.
Figure 24 shows the fractured blade compared
to an intact one, with the fracture regions
Fig. 23
(a) Turbine stage that had the fractured blade.

(b) Detail of the fractured bottom component of
the blade
Table 5 Typical microinclusions found in cast steels

Steel type
Microinclusion type
Aluminum killed
Manganese-silicon killed
Calcium treated, aluminum killed
Aluminum killed, with residual
magnesium
Titanium treated, aluminum killed
All steels
Comments
Alumina
Manganese silicate or manganese-alumino
silicate
Calcium aluminate
Magnesium aluminate
Formed in liquid steel after deoxidation

Formed in liquid steel after deoxidation
Alumina, titania, titanium nitride
Titania forms during reoxidation. Titanium nitride forms

during cooling, usually in the mold itself.
Forms interdendritically during solidification. Often
nucleates on oxides already present in steels
Formed by reaction with alumina, liquid inclusion

Formed by reaction with alumina, solid inclusion
Manganese sulfide
Table 6 Nominal chemical composition of FV520(B) steel

Composition, wt%
C
0.07 max
Si
Mn
Cr
Ni
Cu
Mo
Nb
0.7 max
1.0 max
13.214.7
5.06.0
1.22.0
1.22.0
0.20.5
0.06 max
0.03 max
Table 7 Mechanical properties specifications of FV520(B) steel

Yielding limit, MPa
680800
Strain limit, MPa
Elongation, %
Impact energy, J
Hardness, HV
9001050
20 min
55 min
40 min
270320
18/8/2008 3:11PM Plate # 0
Fig. 24
Fig. 25
(a) Intact blade. (b) Fractured blade
Micrograph of the blade fracture surface showing

several turning gear imprints and the oxidized area
(dotted line)
by fatigue. A darkened region is observed on the

fracture surface, indicated by the dotted line,
suggesting that this area was more exposed to
steam and high temperatures during the turbine
operation time, and it occupies a significant
component of the fracture surface.
Penetrating liquid analysis indicated the presence of secondary longitudinal cracks in the
fractured material, normal to the main crack, at
the bottom of the blade. The analysis made in the

blade body indicated the presence of a large,
longitudinal crack, probably consisting of an
extension of the cracks observed at the bottom of
the blade, as shown in Fig. 26. Optical microscopy analysis of a cross section of the blade
body revealed a different microstructure from
the martensitic steel matrix located parallel to
the longitudinal crack in the blade body.
Because of this different microstructure,
electron-dispersive x-ray (EDX) analyses were
carried out in the regions around the longitudinal
crack in the structural sample. They showed a
chemical composition different from the nominal, as much for the central region as for the
blade head region. The fracture surface of the
longitudinal crack revealed a microstructure rich
in silicon, oxygen, manganese, and calcium,
suggesting that the material contains a large
number of impurities, probably slag from the
casting process and certainly introduced during
the manufacturing process of the component.
The occurrence of these impurities impedes
surface welding during the process of forging,
creating a surface with a smashed aspect.
However, the first region where the nucleation probably occurred was the one near the
pg 169
18/8/2008 3:11PM Plate # 0
longitudinal cracks detected by the penetrating

liquid. Indeed, the fractographic analysis of this
region shows a fracture morphology different
from the vicinity, with several inclusion components protruding into the fracture surface
(Fig. 27). The EDX microanalyses of these
components show a chemical composition with
a high level of carbon, which suggests that these
components are of iron carbide. Moreover,
several longitudinal cracks similar to the one
found in the blade body were observed.
Nonfusible longitudinal cracks exist along
the affected area in the blade. The large variety
of defects and the excessive mechanical vibration of the blade are probably the main causes of
crack nucleation by fatigue in the material near
the blade bottom. They culminated in the catastrophic fracture of the component.
The recommendation includes a more efficient quality control of the manufacturing process of the blade material and avoiding the
occurrence of casting defect formation, slag
inclusions, and other impurities.
Failure in the Axle of a Reduced Section in
a Rotating Component. Possible causes were
investigated for failure in the area of an intermediate reduction. The rotating component
fractured completely after intermittent loading.
Figure 28 shows an outline of the component
and the axle region where the cracks developed.
The chemical composition (in weight percent) of
the fractured axle material is provided in
Table 8. The results show that the axle material
is a DIN-specified 17CrNiMo6 steel. The specifications of the material mechanical properties
at room temperature are given in Table 9.
The visual inspection of the fracture surface
(Fig. 29) indicated an extremely flat aspect, such
as the ones typically displayed in fatigue cracks.
Fig. 26
The flat fracture surface occupied approximately

80% of the cross section (Fig. 29), exactly in the
axis of the radius change for the concordance
section. Due to the small relative section area of
the fracture axis, approximately 20% of its cross
section, it was deduced that the stress for the inservice component was relatively low.
Ten measurements of Rockwell C hardness
were carried out, according to ASTM E18, on
the surface of the axle near the fracture region. A
mean hardness of 33.9 HRC was obtained. This
value is well below the expected one of 43 HRC.
Figure 30 shows the microstructure of steel in
the reduced section on a longitudinal cut plane in
the vicinity of fatigue crack nucleation. The
material presents a large amount of globular or
granular bainite, in agreement with the relatively
low value of hardness of the fractured axle
surface.
Figure 31(a) shows a general topview of the
fracture surface in the region where there was
fatigue crack nucleation, indicated by the arrow
at bottom. The five clustered arrows point in the
direction of fatigue crack propagation advance.
The arrow at the top shows a dark region, originated by contamination of the fracture surface
with oil or grease. Figure 31(b) shows in detail
the fatigue crack nucleating site that probably
started at an inclusion located exactly on the
circumference surface of the reduced section in a
region in the internal concordance radium. The
presence of some inclusions in the proximity
of the fracture site is pointed out by white arrows
in Fig. 31(b). Indeed, fractographic analysis
has shown the possibility of the existence of a
concentration of inclusions in the nucleation
region of the fatigue crack. Figure 32(a) confirms the high level of inclusions in the region,
indicated by white arrows, with signs of moving
Longitudinal crack in the blade body revealed by penetrating liquid. The A-A section indicates the approximate position of
the cut made for structural observation.
pg 170
18/8/2008 3:11PM Plate # 0
by second-phase components indicated by black

arrows, similar to Fig. 31(b). Figure 32(b), the
same image shown in Fig. 32(a) but with backscattered electrons instead of secondary ones,
reveals the great amount of inclusions (darker)
in the metallic matrix (lighter). The chemical
Fig. 27
(a) General view of the probable initial region of

crack nucleation by fatigue crack. (b) Magnification
of the region in the box at the left in (a). (c) Magnification of the
region in the box at the right in (a)
analysis of the inclusions shows a massive

presence of aluminum, sulfur, and calcium
elements.
It is worth noting that these inclusions act,
on a microscopic scale, as metallurgical stress
concentrators. The presence of these secondphase components especially near the external
axle surface where the maximum tensile stresses
are developed during a torsional load (and even
flexion) applied to the in-service component,
drastically reduces the lifetime in fatigue of the
rotating component. This happens through the
promotion of both mechanisms of nucleation
and fatigue crack propagation in their early
stages of growth.
It was concluded that crack initiation occurred
in the reducer axle by fatigue. A single crack
probably was nucleated on a nonmetallic inclusion placed near the finished axle surface,
exactly in the internal component of the concordance radius machined in the section change.
The combination of the effects of stress concentration generated by both discontinuities,
metallurgical (inclusion) and geometric (curvature radius), created sufficient critical conditions
for fatigue crack nucleation that grew due to
the action of repetitive efforts of torsion (and
flexion) imposed in service to the rotating
component.
Failure of a 52100 Steel Axle. The raw
material (52100 steel) used in the manufacture
of an axle catastrophically fractured during
annealing heat treatment at 350 C. Figure 33
shows the fracture surface along with the circular cross section of the component (one of the
samples received for analysis). In the figure, the
arrow at left shows the main fracture plane of
the axis (i.e., along a longitudinal plane), and
the arrow at right points to the starting point
of brittle fracture in its cross section. In Fig. 34,
this starting site is shown in detail (arrow at
bottom).
Figure 35 shows the microstructure of the
52100 steel, in the central region of the part in a
longitudinal plane, after etching with nital. The
massive presence of pearlite and the existence of
free cementite in both formsglobulized
(inside the pearlitic colonies) and veins (circling
the colonies)is observed.
Figure 36 shows the vermiform discontinuities, with an appearance similar to
manganese sulfide inclusions, that are invariably
present in mechanical construction steels. The
presence of a grayish second phase, intermediate
to the metallic matrix (lighter) and the voids
pg 171
18/8/2008 3:11PM Plate # 0
pg 172
Fig. 28
Component drawing of the intermediate I axle. Highlighted are the section change region where the fracture developed and
the crack propagation path for the total fracture of the axle.
Table 8 Chemical composition of the axle

material
Table 9 Mechanical properties at room

temperature
sE, MPa
Composition, wt%
C
Mn
Si
Ni
Cr
Mo
0.17
0.63
0.23
0.10
0.011
1.45
1.59
0.30
742
sR, MPa
AF, %
QF, %
1080
20
57
Fig. 29
Complete cross-sectional fracture surface of the

intermediate I axle. The white arrow shows the
nucleating site of the fatigue crack. The surface generated by
the fatigue crack propagation is identified by F, while the final
fracture of the remaining section is indicated by FF.
Fig. 30
Microstructure of the axle according to a longitudinal cut plane. Etched with 2% nital
18/8/2008 3:11PM Plate # 0
(darker), is observed inside these discontinuities. This material component fills the
larger discontinuities, while the smaller discontinuities are almost totally filled by the second phase.
Figure 37 shows that the most subtle discontinuities have a rather slim, cracklike aspect
and consequently present a great capacity
to concentrate high tensile stresses. In these
terms, it is possible to assume that these second
phases are potential crack nuclei, and that they
also generate a preferential path for crack

propagation. It is worth emphasizing that the
majority of these discontinuities were found
aligned in the direction of the thermomechanical
work to which the axle was submitted during its
manufacture (i.e., longitudinal direction). The
inclusions are disposed on parallel planes to the
main fracture of the component during heat
treatment. This suggests the possibility that
these inclusions played a fundamental role in the
catastrophic failure of the 52100 steel axle.
Fig. 31
Fig. 32
Fatigue crack site. (a) General view. (b) Detail. The

inclusion that originated the site was removed from
the fracture surface. SEM image with secondary electrons
Concentration of inclusions near the fatigue crack

site. (a) SEM image with secondary electrons.
(b) Backscattered electrons
pg 173
18/8/2008 3:11PM Plate # 0
Figure 38 shows elongated microvoids,

obtained by SEM with secondary electrons, in
the vicinity of the tip of one of the cracks that
propagated in the fractured component. The
alignment of the discontinuities generate a

favorable path for material cracking. The voids
that are already interconnected by material
cracking are shown by the arrows in Fig. 38.
In Fig. 39, the 52100 steel microstructure in
the central axle region, cut in the longitudinal
Fig. 33
Cross section of a catastrophically fractured axle.

The arrow at left shows the main fracture plane
(longitudinal), and the arrow at right shows the starting point of
the fracture in the circular cross section.
Fig. 36
Fig. 34 Detail of the starting point of brittle fracture in the

circular cross section of the component (arrow at
bottom). The clustered arrows show the brittle crack tip front.
Fig. 35
Fractured axle microstructure at the center of the

component thickness. Etched with nital. Original
magnification: 400
Inclusion-like microdefects detected in the vicinity

of a crack in the fractured axle, located at the center
of the component
Fig. 37
Slim, cracklike inclusions in the 52100 steel. The

inclusions are oriented in the longitudinal direction
of the component. No etch
pg 174
18/8/2008 3:11PM Plate # 0
Fig. 38
Alignment of the elongated inclusions (oriented in

the longitudinal direction of the part) act as an easy
propagation path in the 52100 steel axle. The main fracture
direction, that is, longitudinal, corresponds exactly to the elongation and inclusion alignments. The arrows point to the existence of cracks among the microcavities that compose the
inclusions. SEM Original magnification: 100 ; 20 kV
plane, is shown in greater detail. An essentially

pearlitic matrix developed with cementite precipitates (Fe3C) in the globular form (solid
arrows). Free cementite exists in the pearlitic
colonies contour in the form of veins or platelets
(white arrows) that offer an easy path for brittle
crack propagation in the material.
The absorption spectra obtained by EDX of
the 52100 steel confirm that the plate precipitates, shown in Fig. 39, are made of iron
carbide or cementite in the free form (Fe3C).
However, the absorption spectra obtained in
microanalyses of the grayish material inside the
elongated microcavities, shown in Fig. 36 to 38,
indicated it is made essentially of iron oxide. At
first, the hypothesis that this contaminant comes
from, for example, the atmospheric oxidation
after the fracture event of the component, was
discarded, since the inclusions measured by
microprobe were completely isolated inside the
metallic matrix, without any possibility of
reaction with the environment.
It was concluded that the raw material used to
manufacture the fractured axle was probably
contaminated with iron oxide. The contaminant
was in the form of elongated inclusions, aligned
in the longitudinal part direction, making an
easy path for main crack propagation (longitudinal). The elongated format provided the
inclusions the capability to concentrate high
tensile stresses and then transform them into
potential crack nucleation sites.
Microanalysis also confirmed the existence
of free cementite in the pearlitic grain contours
Fig. 39
52100 steel microstructure in the center of the

component thickness. Etched with nital. Solid arrows
point to free cementite in the globular form, and white arrows
point to Fe3C in the form of platelets in the pearlite contour.
Original magnifications: (a) 3000 . (b) 10,000. (c) 18,000
that form the 52100 steel. The presence of this

fragile phase may have contributed, to a certain
extent, to the intergranular secondary brittle
crack propagation in the catastrophic fracture of the component during annealing heat
treatment.
pg 175
18/8/2008 3:11PM Plate # 0
ACKNOWLEDGMENTS
Thanks to the Department of Materials, Aeronautics and Automotive Engineering of the School of
Engineering of Sao Carlos, University of Sao Paulo,
on behalf of Professor Dr. Dirceu Spinelli, for the
collaboration on failure analysis case studies.
REFERENCES
1. Defects and Distortion in Heat-Treated

Components, Heat Treating, Vol 4, ASM
Handbook, ASM International, 1991,
p 13201325
2. R.E. Reed-Hill, Physical Metallurgy Principles, 2nd ed., Van Nostrand, 1982
3. Casting Design, Casting, Vol 15, ASM
p 13011322
4. Dimensional Tolerances and Allowances,
Casting, Vol 15, ASM Handbook, ASM
5. Quenching and Control of Distortion, ASM
International, 1988
6. Residual Stress, Heat Treating, Vol 4, ASM
p 13251343
7. M.T. Milan, O. Maluf, D. Spinelli, and
W.W. Bose Filho, MetaisUma Visao
Objetiva (Metals A Vision Object)
Suprema, 2004, p 148149, 161162
8. Gases in Metals, Casting, Vol 15, ASM
p 175189
9. Inclusion-Forming Reactions, Casting, Vol
15, ASM Handbook, ASM International,
1988, p 190211
10. R.D. Pelke and J. Elliott, Trans. TMSAIME, Vol 227, 1963, p 894
11. P.C. Glaws and R.J. Fruehan, Metall. Trans.
B, Vol 17, 1986, p 317
12. R.J. Fruehan, B. Lally, and P.C. Glaws,
Proceedings of the Fifth International Iron
and Steel Congress (Washington, D.C.),
Iron and Steel Society of AIME, 1986
13. A. Kagawa and T. Okamoto, Trans. Jpn.
Inst. Met., Vol 22 (No. 2), 1981, p 137
14. J. Campbell, Castings, ButterworthHeinenmann, 1993

15. J.G. Davies, Solidification and Casting,
Applied Science, 1973
16. AFS Inclusion Atlas Homepage, http://
neon.mems.cmu.edu/afs/afs2/ (Accessed on
March 2005)
17. J.A. Horwath and G.M. Goodrich, MicroInclusion Classification in Steel Casting,
AFS Trans., 1995, p 495510
18. J.M. Svoboda, R.W. Monroe, C.E. Bates,
and J. Griffin, Appearance and Composition
of Macro-Inclusions in Steel Castings, AFS
Trans., 1987, p 187202
19. C.R. Wanstall, J. Griffin, and C.E. Bates,
Clean Steel Cast Technology, Research
Report 106, Steel Founders Society of
America
20. R. Sankaranarayanan and R. Guthrie, Slag
Entrainment through a Funnel Vortex
during Ladle Teeming Operations, Proceedings of the International Symposium on
Developments in Ladle Steelmaking and
Continuous Casting, Aug 1990 (Ontario),
CIM, 1990, p 6687
21. R. Sankaranarayanan and R. Guthrie, A
Laboratory Study of Slag Entrainment
during the Emptying of Metallurgical Vessels, Steelmaking Conference Proceedings
(Ontario), 1992, p 655664
22. K. Noguchi et al., Zairyo to Purosesu (Curr.
Adv. Mater. Process.), Vol 4, 4th ed., 1991,
p 11941197
23. H. Nakamura, S. Kohira, J. Kubota, T.
Kondo, M. Suzuki, and Y. Shiratani, Technology for Production of High Quality Slab
at High Speed, Steelmaking Conference
Proceedings (Ontario), 1992
24. I. Manabu, S. Yutaka, O. Ryusuke, and M.
Zen-ichiro, Evaluation of the Critical Gas
Flow Rate Using Water Model for the
Entrapment of Slag into a Metal Bath Subject to Gas Injection, Tetsu-to-Hagane
(J. Iron Steel Inst. Jpn.), Vol 79 (No. 5), p 33
25. J.M. Harman and A.W. Cramb, A Study on
the Effect of Fluid Physical Properties on
Droplet Emulsification, Steelmaking Conference Proceedings, 1996, p 773784
pg 176
18/8/2008 3:16PM Plate # 0

DOI: 10.1361/faht2008p177

Sources of Failures in Carburized

and Carbonitrided Components
Magorzata Przyecka and Wojciech Gestwa, Poznan University
of Technology
Lauralice C.F. Canale, University of Sao Paulo
Xin Yao, Portland State University
G.E. Totten, Associacao Instituto Internacional de Ciencia and
Portland State University
MANY COMPONENTS, such as fasteners,

crankshafts, camshafts, bearings, and others, require a differentiated response of the surface and
core to external loading. This can be accomplished by surface (case) hardening methods
such as induction and flame hardening or by
surface diffusion processes such as carburizing
and carbonitriding. Raja et al. have reported that
case carburizing is one of the most common heat
treatments for steel, accounting for 50% of all
surface treatments (Ref 1). Case carburizing
involves the creation of a gradient that exhibits
high hardness, brittleness, and strength in the
surface and greater toughness and ductility in
the softer core in order to provide optimal
(Ref 2):

Wear resistance
Resistance to scoring
Bending and/or torsional fatigue strength
Rolling-contact fatigue strength
These properties are optimized by maximizing surface compressive stresses, and carburizing is one of the most effective and commonly
used methods to impart compressive stresses to
the surface of a component (Ref 3). The focus of
this chapter is on carburized and carbonitrided
materials.
Gas carburizing, which is the most widely
used carburizing process, is a surface diffusion
process where the carbon concentration in a
surface layer (case) of a steel matrix that is
predominantly iron, chromium, and nickel is
increased by heating the component at approximately 850 to 950 C with endothermic gas
(Endogas), which is a blend of carbon monoxide, hydrogen, and nitrogen (with smaller
amounts of carbon dioxide, water vapor, and

methane). Endogas is produced by reacting a
hydrocarbon gas, such as natural gas (methane),
propane, or butane, with air.
After the diffusion process is completed, the
component may be quenched from the carburizing temperature or reheated to austenitize the
steel, and then quenched. Bainite formation in
the case is strongly inhibited by the presence of
molybdenum and chromium. Since the surface
contains higher carbon content than the core, it is
harder than the softer core. Core hardness is
most strongly affected by the presence of molybdenum and manganese. Chromium exhibits a
moderate effect, and nickel exhibits a weak
effect (Ref 3). Core hardness is strongly affected
by the quenchant selection and quenching temperature.
In addition to strengthening the case, the
increased carbon content also provides desirable
increased compressive stresses that will inhibit
fatigue crack initiation. The lower carbon content in the core also will produce improved
fatigue strength.
Carbonitriding is similar to carburizing in that
it is a diffusion process that involves the simultaneous diffusion of carbon and nitrogen (from
ammonia) into the steel surface. To obtain
maximum strength, the carbonitriding process
produces a surface that is enriched in nitrogen
and carbon in the form of an epsilon
(e)-carbonitride layer and a diffusion zone
containing chromium-iron carbide, (Cr,Fe)7C3;
chromium carbide nitride, Cr62C3 5N0.3; chromium nitride, (Cr2N) or [Cr, Fe(2Ni . . . x)]; and
Fe2N phases (Ref 4, 5). Typical case thicknesses
range from 50 to 200 mm with a hardness
pg 177
18/8/2008 3:17PM Plate # 0
pg 178
between 750 and 900 HV. Like carburizing, the

case depth of carbonitrided steel is dependent
on both the carbonitriding diffusion time and
temperature, as illustrated in Fig. 1 (Ref 2, 6).
Deeper case hardnesses may be obtained by first
precarburizing prior to carbonitriding. Karamis
showed that carbonitrided AISI 5115 steel
exhibited greater surface hardness and wear resistance than carburized AISI 5115 steel (Ref 4).
Carbonitriding processes are typically conducted in either a gas (ammonia) or a salt bath
based on trade names such as Tufftride, Nitrotec,
and Nitrox. Alternatively, a plasma nitriding
process may be conducted. A brief summary
comparison of carburizing and carbonitriding
processes is provided in Table 1 (Ref 2).
Carter has reported that failures of carburized
gears are primarily due to service-related causes,
such as misalignment, poor lubrication, and
overloading, which constitute the greatest
source of all gear failures, as shown in Table 2
Fig. 1
Correlation of case depth of carbonitrided steels with

varying diffusion times and temperatures
Table 1 Comparison of carburizing and

carbonitriding processes
Process
Carburizing
Comments
Hard, highly wear-resistant surface (medium case

depths), excellent contact load potential, good
bending fatigue strength, good seizure
resistance, excellent quench cracking resistance,
low-to-medium-cost steels required, high capital
investment
Carbonitriding Hard, highly wear-resistant surface (shallow case

depths), fair contact load potential, good bending
fatigue strength, good seizure resistance, good
dimensional control, excellent quench cracking
resistance, low-cast steels usually satisfactory,
medium capital investment
(Ref 7). Heat treatment was the second most

often cited cause for failure. However, it is often
difficult to detect the root cause of a specific
failure under the conditions in which the failure
occurred, and many of the service-related failures could have been reduced with more attention to the other potential causes of failure
shown, since they are often interrelated.
Palaniradja et al. reported that 10 to 12% of
carburized parts are rejected due to various
process-related defects (Ref 8). To examine this
in more detail, they conducted a Taguchi analysis of gas carburization of AISI 8620 and 3310
steels, and their results showed that relative
contribution to surface hardness was holding
time (20%), carbon potential (20%), carburizing
temperature (0%), and quenching time (60%).
Similarly, they also studied the effects of process
variables on case depth and found: holding
time (60%), carbon potential (9%), carburizing
temperature (14%), and quenching time (10%)
(Ref 8). These results show that an adequate
understanding of failure analysis of carburizing
and, by implication, carbonitriding must be
accompanied by understanding the contribution
of process parameters on resulting potential
failures.
Some of the most common contributors to
failure of carburized gears include surface finish,
microstructure, excessive or inadequate case
depth, incorrect case and/or core hardness,
improper carbon concentration and hardness
gradients, undesirable surface carbon content,
excessive retained austenite, large amounts of
globular and network carbides, intergranular
oxidation, internal oxidation, residual stress,
extremely coarse case or core grain structure,
untransformed core with free ferrite, quenching
and grinding cracks, surface decarburization,
excessive heating during grinding, excessive
removal of the case during grinding, microcracking, and so on (Ref 9, 10).
This chapter provides an overview of various
contributors to failure of carburized and carbonitrided components, with the primary focus on
carburized components.
Table 2 Survey summary of sources of
gear failures
Cause of gear failure
Material quality and forming

Design
Service-related causes
Manufacturing
Heat treatment
0.8
6.9
74.7
1.4
16.2
18/8/2008 3:17PM Plate # 0
Sources of Failures in Carburized and Carbonitrided Components / 179
Design
Component design may contribute directly or
indirectly to component failure. Deficiencies
such as insufficient radii or sudden changes in
section size are significant contributors to failure. In addition, the presence of stress raisers,
such as those shown in Fig. 2, are among the
most common design contributors to quench
cracking and fatigue failure.
A more comprehensive insight into design
is provided by Kuehmann et al., who developed
a systems analysis flow chart to describe the
effects of case-core hardening in designing a
carburizing process/metallurgical structure/
resulting properties and performance for the
production of gears produced by three routes:
conventional forging, near-net shape casting,
and powder metal processing (Fig. 3) (Ref 11).
To properly design a component, it is necessary to estimate surface loading, distortion after
heat treatment, case depth and carbon profile,
Fig. 2
case and core hardness, and core strength. As an

estimate, for hardnesses within the range of 30 to
45 RC, the required case depth can be calculated
from (Ref 2):
Case depth to 50 HRC=(1:2 107 W)=F
where W is the force in pounds pressing the

surfaces together, and F is the length of the line
contact (inches).
Carter has recommended the following general design criteria (Ref 7):

If a component is carburized from both sides,

the case depth should not be greater than
20% of the wall thickness.
At the base of gear teeth, 30% of the core
material should remain uncarburized.
Shallow case depths usually require higher
case hardness.
Case depths should be five times the acceptable wear limit.
Effected of stress raisers on stress concentration and distribution of stress at several changes of form in components. (a) to (c)
Progressive increases in stress with decreasing fillet radii. (d) to (f) Relative magnitude and distribution of stress resulting from
uniform loading. (g) Stress caused by the presence of an integral collar of considerable width. (h) Decrease in stress concentration that
accompanies a decrease in collar width. (i) Stress flow at the junction of a bolt head and a shank. (j) Effect of a single sharp notch. (k) Effect
of a continuous thread. (l) Effect of a groove or gauge. Source: ASM Handbook, Volume 11, 2002, p 715
pg 179
18/8/2008 3:17PM Plate # 0
Although machining is proportional to the

case depth, it should be minimal.
Kern and Suess have recommended the following general guidelines for heat treatment of
gas-carburized gears (Ref 2):

For forgings, normalize or anneal (as required by the alloy being heat treated) from a
temperature at least 28 C (50 F) above the
carburizing temperature.
Assure that the gears are machined prior to
heat treatment.
Bring the gear to the carburizing temperature
with sufficient circulation of a neutral
atmosphere, and then introduce the gas used
for carburizing.
For deep cases (41.5 mm, or 0.060 in.),
adjust the carburizing atmosphere and time
to produce uniform carbon diffusion from
the surface to the core. A decrease of 0.15
to 0.20%/0.25 mm (0.010 in.) of depth is
Fig. 3

nearly ideal. The gradient may be steeper for

shallower case depths.
The recommended surface carbon is 0.90 to
1.10% for 4300, 4600, 8600, 8800, and 9400
carburizing steels. Although the same case
depth is generally acceptable for grades such
as 4800, they are preferably reheated for
hardening. The recommended surface carbon is 0.65 to 0.85% for high-nickel steels
such as the 4800 series, which is usually
direct quenched.
To minimize cost and distortion, use direct
quenching whenever possible.
To assure optimal dimensional control,
properly maintain quenching dies and
plugs.
Quench as rapidly and uniformly as practical, and use spray impingement fixtures on
large, solid pinions that are four pitch and
coarser.
Use hot oil quenching on fine pitch gears.
Kuehmann et al. flow chart to summarize design elements of a carburizing process/metallurgical structure/resulting properties
and performance comparison of gas carburizing gears produced by conventional forging, near-net shape casting, and powder
metal processing
pg 180
18/8/2008 3:17PM Plate # 0

Case hardness of the finished gear should be

60 HRC or greater.
If possible, test each gear for partial decarburization and/or upper transformation
products.
To minimize distortion and to permit quieter
operation, the surface carbon content should
be uniform throughout the production cycle.
Steel Selection and Hardenability

Steels typically used for case hardening contain carbon contents of less than approximately
0.25%. The carbon content of the case is usually
controlled to between 0.8 and 1% C. The actual
surface carbon content is generally limited to
0.9%, because excessively high carbon content
may lead to the presence of unacceptably high
retained austenite and brittle martensite. Some
of the most commonly used AISI grades of steel
used for carburizing are shown in Table 3
(Ref 12).
Plain carbon steels may be carburized; however, relatively poor hardenability due to the
lack of alloying elements reduces the carburizing response of the case. Because of the stabilizing effect of the nitrogen relative to austenite,
carbonitriding provides greater hardenability
than attainable with carburizing. Therefore,
plain carbon steels respond well to carbonitriding.
Proper steel selection is a critically important
process to provide the desired case depth and
microstructure and the required core properties.
Typically, the case structure should be fully
martensitic, with the exception of allowing for
required application design limits on retained
austenite content. For example, the steel must
Table 3 Common carburizing grades of steel

and their relative processing features
AISI
steel
grade
4620
8620
4320
4820
9310
Note
Lower-cost, chrome/nickel/molybdenum steel where only

nominal hardenability and core response is required
Most commonly specified grade. Excellent carburizing
response, with good hardenability for most section sizes
Higher hardenability for improved core response in heavier
sections
Increased nickel content for improved core toughness; slower
response results in longer process times
Maximum nickel content for maximum core toughness;
slower response results in longer process times
possess sufficient hardenability to provide the

desired hardness and microstructure in both the
case and the core. After carburizing, the component must possess sufficient toughness without exhibiting brittle failure.
Most steels that are carburized are deoxidized
by the addition of aluminum (commonly designated as killed steels). Deoxidation will provide
finer grain sizes to temperatures of approximately 1040 C. Coarser grained steels may be
carburized if grain refinement by double quenching is possible. Double quenching typically involves direct quenching followed by reheating
to a lower-temperature quenching a second time
(Ref 13).
Selection of proper hardenability of steels for
both carburizing and carbonitriding is critically
important, both of the core and the case, since
improper hardenability design can lead to undesirable nonmartensitic transformation products in the case, leading to a potential reduction
in static and dynamic fatigue strength of up to
30% and a reduction of impact fatigue of up to a
factor of 2.5 times (Ref 14). The hardenability
gradient of the case and the core is dependent
on a number of factors, including cooling rate
during quenching, variability of the chemical
composition (alloy content, carbon and nitrogen) of the case, and the carburizing or carbonitriding method being used.
Core hardenability is being used increasingly
to specify alloy steels used for case hardening
where the hardenability of both case and core
must be considered. Details for the traditional
approach for the experimental determination of
hardenability of carburizing steels are provided
in Ref 15. Jominy curves for a number of carburizing steel alloys with varying hardenability
are shown in Fig. 4 (Ref 16).
Procedures have also been described for
determining ideal diameter (DI) values and
hardenability of carburizing steels from Jominy
data using regression equations for composition
and grain size (Ref 17, 18). The ideal diameter is
defined as the diameter of a cylindrical steel bar
that will form 50% martensite at the center when
subjected to an ideal quench. Hardenability
differences may be substantially greater for
some case-hardening steel grades relative to
others due to the difference in carbon content in
the case and core. This is more critical for heavysectioned components that are reheated and
quenched.
The hardness gradient through the case is due
to the relationship between the thermal gradient
pg 181
18/8/2008 3:17PM Plate # 0
and the carbon gradient during quenching.

Therefore, an increase in case hardenability
required to produce greater amounts of martensite for a given carbon content will result in an
increased case depth. In such cases, a reduced
(shallower) carbon profile and shorter carburizing times will be necessary to obtain the desired
hardness profile in the carburized component.
SAE J1975 standard Case Hardenability of
Carburized Steels summarizes characteristics
of carburized steels and factors involved
in controlling hardness, microstructure, and
residual stress. Methods of determining case

hardenability are also provided.
Parrish reported the following scheme that
was developed to classify the case hardenability
of steels (Ref 19):

Level 1: Surface carbon contents 40.8% C

are martensitic.
Level 2: All carbon contents from the surface
to 50% C are martensitic.
Level 3: All carbon contents from the surface
to 0.27% C are martensitic.
Level 4: A martensitic case occurs at all
carbon levels, including the core material
just beneath the case.
Figure 5 illustrates the core hardenabilities for

a number of carburizing steels (Ref 19). This
figure is used by estimating the equivalent diameter for the critically stressed section of the
component of interest, and then the expected
level of case hardenability of that steel is determined. Figure 5 indicates that level 4 is attainable only for small section sizes of more alloyed
steels, and level 3, depending on the section size,
is more readily attainable for most of the steels
shown. Level 2 is more typical of the more
common case-hardened parts and should represent a minimum target to be attained.
Case hardenability may vary widely even for
steels with equivalent core hardenabilities. Kern
Fig. 4
Jominy hardenability data for a number of carburizing

steels
Fig. 5
Case hardenabilities of a number of carburizing steels with oil quenching
pg 182
18/8/2008 3:17PM Plate # 0
and Suess provided the following guidelines

(Ref 2):
Steel grades in which the case hardenability

is due to carbide-forming metals such as
chromium (8600 series) are sensitive to
microcracking, especially when direct quenched from the carburizing temperature. This
can be controlled by restricting the carbon
content in the case to 0.9%.
Steel grades with relatively high nickel
content, for example, 4800 and 9300 series,
may form excessive (430%) amounts of
retained austenite when direct quenched
unless the carbon content of the case is
maintained at 50.75%.
Carburizing round section sizes greater than
76 mm (3.0 in.) may lead to difficulty in
achieving the desired case and core microstructures when quenching in oil. In such
situations, consider induction hardening or
nitriding or using a highly alloyed steel
grade such as AISI 9310.
Some standard grades of steel exhibit narrower core hardenability bands than other
grades. For example, 8620H exhibits a
hardenability band spread of 14 HRC at J 4,
and 9310 exhibits only 8 HRC spread at the
same J-value. This provides a greater
amount of distortion control in addition to
some possible application-dependent property advantages as well.
One problem that can arise during the steelmaking process or that may be observed as a millto-mill variant is the presence of segregation
effects through the section of the steel billet
during a continuous casting process, which results in the presence of a white band (Ref 20).
White band is a type of negative segregation
often observed in electromagnetically stirred
continuous castings. The white banding produces a significant hardness gradient across the
billet. After subsequent rolling and forging or
machining to produce a component, the resulting
grainflow can produce nonuniform hardenability
and/or soft spots that can significantly affect
distortion.
In addition to proper hardenability selection,
to achieve maximum core toughness, proper
austenitization and quenching to martensite is
necessary. These topics are discussed subsequently.
Case Depth. The case of a carburized (or
carbonitrided) steel alloy is that portion
extending inward from the surface, where the

hardness is greater than that of the core. The total
case depth is the distance or thickness of the
carbon-enriched surface layer. The effective
case is the point where 0.4 to 0.5% C (percent is
called points in the industry) is present if the part
is hardened to 50 HRC (510 HIV). The depth of
the case is a function of carburizing time and
carbon (carbon potential) at the surface. Genel
and Demirkol have reported that the following
equation model can be used to predict effective
case depth (Ref 21):
Effective case depth (mm)=
0:41 Carburizing time (h)1=2
The carbon potential of a furnace atmosphere

at a specified temperature is defined as the carbon
content of pure iron that is in thermodynamic
equilibrium with the atmosphere. The carbon
potential of the furnace atmosphere must be
greater than the carbon potential of the surface of
the workpieces for carburizing to occur. The
carbon potential is a measure of the ability of
a gas to react with the steel surface. It is this
difference (carbon content in the gas and at
the steel surface) that provides the driving force
for carbon transfer to the parts being carburized.
The composition of a gas that will produce a
given surface carbon content is dependent on
equilibrium data for the gas. The amount of
carbon transferred will depend on factors that
include temperature, time, and steel composition.
The alloy composition of the steel will affect
the effective carbon potential at the surface. The
presence of elements such as chromium, manganese, and molybdenum that form stable carbides of iron will increase the effective carbon
potential. Elements such as silicon and nickel
form less stable carbides and reduce the effective carbon potential. Alloying elements that
stabilize austenite or ferrite also reduce the
effective carbon potential. The effect of alloying
elements on carbon potential can be calculated
from (Ref 22):
Log (correction factors)=0:005 (%Si)
0:013 (%Mn) 0:040 (%Cr)+0:014 (%Ni)
0:013 (%Mo)
It is important to note that, except for long

carburizing times (3 to 410 h), the surface
carbon content is typically not equal to the carbon potential, because the surface of the part
pg 183
18/8/2008 3:17PM Plate # 0
being carburized does not reach equilibrium with

the carburizing atmosphere. This is illustrated in
Fig. 6. A different, but similar, correlation exists
for various steel alloys and carburizing conditions. From a correlation such as Fig. 6 and
given case depth requirement and carburizing
time, it is possible to correct for the carbon
potential under nonequilibrium conditions to
achieve a specific surface carbon content, as indicated by the dashed lines in the figure (Ref 22).
For a specific temperature, the case depth (d)
will vary with the square root of the carburizing
time (t):
d=Qt1=2
Values of the depth factor (Q) as a function of

temperature are shown in Fig. 7. This equation is
reported to be valid for low-carbon steel and
some alloy steels (Ref 22). For applications such
as automotive gears, typical case depths are 0.8
to 1.4 mm (Ref 23).
An equation that is often used to calculate the
case depth (in.) as a function of both time and
temperature is the Harris equation (Ref 24):
Case depth (in:)=31:6 t0:5 =10(6700=T)
where t is the time at the carburizing temperature in hours (carburizing time), and T is the
absolute temperature [Kelvin (K) = C+273,

Rankine (R) = F+460]. In metric units, the
case depth (mm) is (Ref 24):
Case depth (mm)=660 e8287=T t0:5
At the operating temperature, the carburizing

process may be conducted in two parts. Carburizing occurs during the first part of the process in a high-carbon-potential period when the
enriching gas is added to the furnace atmosphere
to increased the carbon content of austenite (the
carburize-boost period) and the carbon potential
is greater than the desired carbon potential. This
part of the process is typically conducted at a
carbon potential close to the solubility limit of
carbon in austenite, typically between 1.0 and
1.2% C, which is dependent on the temperature
and alloy content of the steel. The time for this
part of the process to occur is called the carburizing time. This part of the process is followed by a boost-diffuse period, where the
process is operated at the equilibrium carbon
potential, which is reduced to a level that will
maintain surface content, typically 0.8 to 0.9%
C, during which time the carbon will diffuse
deeper into the case and provide a gradual case/
core transition. Together, this is called the boostdiffuse cycle. The time for this part of the process is called the diffusion time. When the
required case depth is achieved, if the component is direct quenched, the temperature is
Fig. 6
Carbon dioxide content of the atmosphere required to

produce certain surface carbon levels at different carburizing times under a given set of carburizing conditions. The
dashed lines illustrate alternative times and carbon dioxide
contents to produce a single surface carbon content.
Fig. 7
Variation of the depth factor, Q, with carburizing

temperature for low-carbon and certain alloy steels
pg 184
18/8/2008 3:17PM Plate # 0
lowered to 850 C to reduce the distortion and

then quenched (Ref 25).
Carburizing boost-diffuse cycles conducted
over 2 h are advantageous for case depths
40.50 mm (0.020 in). They are also useful
when relatively deep cases free of carbides or
retained austenite are required. This is important
when the carbon content is greater than the
eutectoid composition, where there is an increased tendency to form carbides and retained
austenite upon quenching. These effects increase with alloy content (Ref 24).
Harris also developed equations to compute
the carburizing and diffusion times to achieve a
specific case depth and surface carbon content
(Ref 24):
Carburizing time (h)=(C Ci )2 =(CO Ci )
Diffusion time (h)=Total time Carburizing time
where C is the final desired surface carbon

content, CO is the surface carbon content at the
end of the carburizing cycle, and Ci is the carbon
content at the core.
The effect of the steel alloy composition on
the carbon gradient is illustrated for AISI 1020
plain carbon steel and AISI 8620 after carburizing at three temperatures in Fig. 8 (Ref 24).
The alloy content will influence the diffusion
rate, but its greatest effect is on the case carbon
content. Normal carbon gradients, such as those
shown in Fig. 8, can be achieved by maintaining
a saturated austenite condition at the surface
during the entire boost-diffuse carburizing cycle
(Ref 24).
It is important to control the ratio between the
boost and diffuse times and to carefully control
the carbon potential to avoid obtaining a carbon
profile such as that shown in Fig. 9 (Ref 24).
Although the desired surface hardness was
obtained, the lower carbon content at the surface
can lead to a transformation that proceeds simultaneously outward from the case-core interface
and at the surface and proceeds inward such that
the last portion of the case to transform is just
below the surface. This will result in an undesirable condition where the surface is in tension
relative to the core as well as a corresponding
decrease in fatigue strength in addition to an
increased potential for cracking (Ref 24).
Boyer reports that a maximum tolerable carbon potential for carburizing cycles of up to 10 h
at 925 C is 1.3% to avoid excessive soot formation.
If excessively high carburizing temperatures

are used, the following situations may occur:

Rapid increase in grain growth and loss of

properties
Increased energy consumption
Increased deterioration of the furnace fixtures and baskets
When high carbon potentials and long carburizing times are used to produce high surfacecarbon content and deep case depths, excessive
retained austenite and/or free carbides may be
obtained as a result. These microstructural products exhibit adverse effects on residual-stress
distribution (which is discussed subsequently).
Therefore, although high carbon potentials may
be used for short carburizing times, substantial
deleterious effects may result if used over prolonged carburizing times.
Excessive carbon potentials, gaseous atmosphere composition control problems because of
carbon probe malfunctions or air ingression, and
inadequate furnace purging can lead to excess
free carbon and sooting, which may be sufficiently severe as to leave carbon deposits on the
parts (Ref 24, 26). This can lead to corresponding problems in controlling the carbon potential,
resulting in nonuniform carburizing and dimensional control problems.
Quenching. During carburizing, the steel
microstructure consists of polycrystalline austenite. Grain coarsening may occur if the carburizing times are relatively long. The austenitic
grain size will determine the size and distribution of martensite that will form as a result of
quenching. In addition, phosphorus segregation
into the grain boundaries may occur during the
carburizing cycle, which has been found to be
directly dependent on phosphorus and carbon
content. Hyde et al. found that fatigue and
fracture toughness decreased slightly when the
phosphorus content increased from 0.005 to
0.017%, and when 0.017 to 0.031% P was present, the endurance limit and fracture toughness
decreased substantially (Ref 27).
Phosphorus also affects the degree of carbon
segregation in the form of cementite at the austenite grain boundaries. During quenching,
small amounts of cementite form at the austenite grain boundaries in the high-carbon case
(Ref 25, 27). This leads to increased sensitivity
to intergranular fracture, which is a major cause
of fatigue crack initiation in carburized steels
(Ref 25).
pg 185
18/8/2008 3:17PM Plate # 0
Fig. 8
Carbon gradients for gas-carburized 1020 and 8620 steels. The 1020 steel was carburized in a batch furnace, and the 8620 was
carburized in a pit furnace.
pg 186
18/8/2008 3:17PM Plate # 0
After carburizing, the parts are then either

quenched directly or air cooled and then reheated and quenched. Quenching is performed to
harden the components. It is most desirable to
develop a martensitic or bainitic case while
controlling the formation of retained austenite to
an acceptable level and simultaneously minimizing proeutectoid and pearlitic structures. The
challenge is to quench sufficiently fast to produce the desired core structure but not so fast
that the higher-carbon and more brittle case
cracks. In addition, the desired hardness gradient
between the surface and the core is critically
Fig. 9
Carbon profile of an incorrectly carburized steel
Fig. 10
important to achieve the desired wear and fatigue properties. If the carburized gear, such as a
spiral bevel gear, is not quenched to achieve the
necessary surface hardness and hardness gradient, failures accompanied by micropitting and,
ultimately, fracture may occur (Ref 28).
The morphology of martensite is carbon
dependent, as shown in Fig. 10 (Ref 29). At
lower carbon content, a lath martensitic structure forms, while plate martensite forms at
higher levels of carbon. The two different morphologies are illustrated in Fig. 11 (Ref 30). Lath
martensite exhibits better toughness than the
higher-carbon plate martensite. Plate martensite,
as the name indicates, forms as lenticular (lensshaped) crystals and is sometimes referred to as
acicular (meaning needlelike) martensite or
high-alloy martensite. A characteristic of plate
martensite is the zigzag pattern of smaller plates,
which formed later in the transformation,
bounded by adjacent larger plates that formed in
the beginning of the transformation (Ref 30).
Typically, quenching is performed either directly from the carburizing process after furnace
cooling to approximately the Accm temperature
and then quenched, or the parts are air cooled
and then reheated and quenched. Less commonly, double reheat quenching may be performed to provide high-durability components
(Ref 19). Some quenching cycles recommended
Dependence of the martensitic structure on carbon content
pg 187
18/8/2008 3:17PM Plate # 0
both the case and core grain size. The refined

core will be soft and machineable with
maximum toughness and resistance to
impact. The refined case will be hardened for
wear and resistance. In addition to grain
refinement, double reheating and quenching
is reported to improve fatigue properties by
reducing the size and density of microcracks
in the structure (Ref 21).
by Crucible Steel for carburized 8620 steel

include (Ref 31):
Direct quench from carburizing: Quench in

oil directly from the carburizing temperature
of 925 C (1700 F). The core is hardened
but unrefined. The case is hardened to the
extent that it will be fileproof if the carbon
content is sufficiently high.
Cool, reheat and quench (1): After cooling
from the carburizing temperature (925 C,
or 1700 F), the carburized (but not yet
hardened) steel is reheated to a temperature
above the upper critical temperature, Accm,
of the core (835 C, or 1535 F) and then
quenched in oil. The core will be refined and
exhibit maximum strength and hardness.
The case will be hardened and somewhat
coarsened.
Cool, reheat, and quench (2): After cooling
from the carburizing temperature (925 C,
or 1700 F), the carburized (but not yet
hardened) steel is reheated to a temperature
above the lower critical temperature, Ac1, of
the case (730 C, or 1350 F) and then
quenched in oil to harden and refine the case.
The core will be unrefined, soft, and machineable, and the case will be hardened.
Double reheat and quench: The steel is
cooled in the furnace from the carburizing
temperature of 925 C (1700 F). The steel
is then reheated to above 730 C (1350 F)
and oil quenched to refine the core. The steel
is again reheated to 730 C (1350 F) and oil
quenched to refine the case. This double
heating and quenching procedure refines
Fig. 11
Of these methods, the most common is direct

quenching. However, there are a number of reasons why reheat quenching is favored for higheralloy, case-hardening steels, including (Ref 19):

To assure grain size and retained austenite

control
When intermediate subcritical heat treatment is required to condition the carbide
structure within the case or to facilitate
additional machining
When the parts are to be plug or die quenched for distortion control
When it is not possible to direct quench, such
as in pit carburizing
Ingham and Clarke compared the results

obtained for carburized 8620 steel with the
carbon gradient shown in Fig. 12, which was oil
quenched from a direct quench following carburizing and by a reheat and oil quench cycle
(Ref 32). The results obtained showed that the
direct quench process yielded a higher hardness
than the reheat and quench process, which exhibited a relatively lower as-quenched hardness
due to the presence of bainite in the hardened
case structure.
Martensite morphology. (a) Lath martensite. (b) Plate martensite. Source: Ref 30
pg 188
18/8/2008 3:17PM Plate # 0
Quenchants must be selected to provide

cooling rates capable of producing an acceptable
microstructure and hardness gradient through
the case and the core. However, it is not desirable to use quenchants with excessively highheat-removal rates, since the propensity to cause
increased distortion or cracking increases with
quench severity. Although a reduction of quench
severity leads to reduced distortion, it may also
be accompanied by undesirable microstructures.
Therefore, it is essential to select optimal quenchant and agitation conditions for the required
microstructure, hardness, and strength in critical
sections of the parts for each steel alloy, section
size, and required microstructural and mechanical properties. Actual cooling rates or heat
fluxes provided by a specific quenching medium
are typically unavailable. However, some illustrative comparative data are provided in Table 4
(Ref 33). Figure 13 illustrates the comparative
cooling properties of various oil-quenched steel
bars assuming a surface heat-transfer coefficient
of 0.019 cal s 1 C 1 cm2 (Ref 32).
Quench nonuniformity is a significant contributor to quench cracking. Quench nonuniformity can arise from nonuniform flow
fields around the part surface during the quench
or nonuniform wetting of the surface. Both lead
to nonuniform heat transfer during quenching,
Fig. 12
creating large thermal gradients between the

core and the surface of the part. Poor agitation
design is a major source of quench nonuniformity, since the purpose of the agitation system is not only to take hot fluid away from the
surface and to the heat exchanger but also to
provide uniform heat removal over the entire
cooling surface of all of the parts throughout the
load being quenched.
A wide range of quench media can potentially
be used when quenching carburized parts. Some
comments on quench media selection, provided
by Boyer, include (Ref 34):
For carbon steels, the most common

quenchants are water and brine. When water
is used as the quenchant, bath temperatures
of 20 to 30 C with agitation are the most
common.
In the industry, oil quenchants are the most
common, particularly when integral-quench
(sealed-quench) furnaces are used at temperatures of 25 to 70 C. The quench oils
may be classified as fast, intermediate, or
slow depending on the cooling rate, enhancing additive, and quench oil base stock
being used. When distortion control is critical, a hot oil that can be used at temperatures as high as 175 C may be used.
Comparison of direct quenching and reheat and quenching of 10 cm (4 in.) diameter AISI 8620 steel after oil quenching
pg 189
18/8/2008 3:17PM Plate # 0
It is possible to use aqueous polymer

quenchants, even in integral-quench furnaces, if appropriate structural conditions
are met. The user is advised to consult his
furnace manufacturer prior to use. A wide
range of quench severities is possible by
varying the polymer concentration, bath
temperature, and agitation. In one study, it
was shown that an aqueous polymer quenchant produced substantial improvements
in fatigue properties relative to a conventional quench oil, which was attributed to
an improvement in quench uniformity
(Ref 35).
Table 4 Comparison of typical heat-transfer

rates
Maximum surface
heat-transfer rate,
W m 2 K1
Quench medium
Still air
Nitrogen (1 bar)
Salt bath or fluidized bed
Nitrogen (10 bar)
Helium (10 bar)
Helium (20 bar)
Still oil
Hydrogen (20 bar)
Circulated oil
Hydrogen (40 bar)
Circulated water
5080
100150
350500
400500
550600
9001000
10001500
12501350
18002200
21002300
30003500
When distortion control is especially critical, salt bath quenching may be required.
However, parts should never be transferred
directly from a carburizing bath containing
45% cyanide to a nitrate-nitrite quench
bath, because this will result in a violent
reaction and possibly an explosion (Ref 34).
One often-encountered quenching problem

that may lead to increased dimensional control
problems is contamination. For example, heterogeneous quench media caused by water
contamination of oil or oil contamination of
water or aqueous polymer solutions can potentially cause cracking problems. Similarly, salt
contamination, either from salt baths or hardmetal ion contamination, can lead to problems of
cooling rate control. Solid contamination, such
as sludge or soot contamination in oil or aqueous
media, also may lead to distortion and cracking.
Finally, excessive foaming and air entrainment
of the quench media will lead to nonuniform
cooling, soft spots, increased residual stresses,
and cracking. Therefore, it is essential that the
quench bath be well maintained to assure optimal distortion control and minimize the potential for cracking.
To develop the optimal residual-stress gradient, it is important to use the proper quenching
Soaking temperature
Bar size, in.
Fig. 13
Centerline cooling curves for oil-quenched steel bars of varying section sizes, assuming a surface heat-transfer coefficient of
0.019 cal s1 C 1 cm2
pg 190
18/8/2008 3:18PM Plate # 0
conditions for the steel grade of interest.

Figure 14 shows the development of the temperature distribution through 12.5 mm diameter
bars carburized to a depth of 0.9 mm for oilquenched alloy steel and a water-quenched mild
steel (Ref 36). This figure shows that the Ms and
Mf temperatures are reduced with increasing
carbon content in the case. The isochronal lines
in the figure illustrate the cooling profile through
the case to the core at specified time intervals.
The martensitic transformation occurs at the
core/case interface first. The case transforms last
along with the corresponding expansion of
martensite in the case. Since the core is already
transformed, this process restrains further expansion during continued cooling, leading to the
development of surface compressive stresses
and placing the core in relative tension (Ref 36).
As this figure shows, since the Mf is less than
ambient temperature, retained austenite will
accompany this process.
Because carbonitriding is similar to carburizing, except that nitrogen and carbon diffusion
into the case is involved, both processes exhibit
similar microstructural transformation and
evolution of residual stresses upon quenching.
Process temperatures during carbonitriding are
typically lower (800 to 850 C), as are process
times (30 to 60 min), which provides a relatively

shallow case, usually 50.5 mm (Ref 36).
These results show any factor that affects the
cooling profile, such as bath temperature, agitation, or the quenchant selection, will exhibit a
corresponding effect on the thermal distribution
through the carburized case upon hardening
and therefore on the development, type, and
magnitude of residual stresses. To assure optimal distortion control, the following variables
should be carefully monitored and controlled:

Adequate quality-control procedures of the

quench media should be in place. For
examples, follow ASTM D6710, Standard
Guide for Evaluation of Hydrocarbon-Based
Quench Oils, for oil quenchants and ASTM
D6666, Standard Guide for Evaluation of
Aqueous Polymer Quenchants, for aqueous
polymer quenchants.
Carefully control the water content if polymer quenchants are being used.
Replace the quench mediate at regular uselevel intervals.
Carefully control the quench bath temperature.
Monitor fluid flow variation at critical locations in the quench tank. Nonuniform quenching has been reported to lead to quench
cracking of a carburized 17CrNiMo6 axle in
a reduction gearbox.
Monitor hardness and dimensional changes
of the parts being processed to look for
unexpected variance.
Retained Austenite (Ref 1, 7). The quenchant temperature is a critically important variable in controlling the amount of retained
austenite in the carburized steel. This is important because incomplete quenching and the presence of retained austenite will often seriously
affect wear resistance and pitting fatigue strength
(Ref 9). With carburized steels, the martensite
start (Ms) temperature will decrease with
increasing carbon content. To determine the
impact of the carbon content on the Ms temperature from the steel composition, the Steven
and Haynes equation may be used (Ref 37, 38):
Ms ( C)= 561 474C 33Mn 17Ni 17Cr
21Mo
Fig. 14
Temperature distribution and martensitic transformation during quenching of carburized 12.5 mm

diameter steel bar. The curves (isochronal lines) in the figure
indicate time in seconds after immersion of the carburized
(0.9 mm case) bar into the quenchant indicated.
where C, Mn, Ni, Cr, and Mo are the percent of

the element contained in the steel. This equation
is only accurate for steels containing up to 0.5%
pg 191
18/8/2008 3:18PM Plate # 0
pg 192
C. For steels with higher carbon content, Fig. 15

should be used to determine a more accurate
value for Ms (Ref 37).
Although the degree of transformation
between the Ms and martensite finish (Mf)
temperatures is not linear, the difference is essentially constant (Mf is approximately 215 C
lower than the Ms temperature) (Ref 37).
Nevertheless, if Ms Mf is approximately constant, incomplete transformation can be expected if some part of the transformation occurs at
a temperature lower than the quench bath temperature. Therefore, the volume of untransformed austenite (Vc) is related to both the Ms
and the quenchant temperature (Tq). This relationship is quantitatively defined by the wellknown Koistenen and Marburger equation:
72
Vc =e71:10 10
(Ms 7T q )
Using these equations, Parrish demonstrated

the effect of quenchant temperature on retained
austenite on a hypothetical steel. The results of
these calculations are summarized in Table 5,
which show that the amount of retained austenite
is expected to decrease with decreasing bath
temperature.
Trusova studied the formation of retained
austenite after quenching and showed that the
retained austenite content was dependent on the

carbon content and alloying elements and
the content of the carburized case (Ref 39). The
higher the steel temperature prior to quenching,
the greater the decomposition of austenite but
the greater the potential for cracking. To prevent
cracking, steel may be quenched in hot oil, but
the amount of retained austenite increases.
Additional dilatometer examination of an isothermal high-temperature tempering process
with subsequent cooling showed there was
either a volume shrinkage due to decomposition
of tetragonal martensite or an expansion caused
by the decomposition of retained austenite
(Fig. 16) (Ref 39). These transformations occurred during heating, isothermal holding, and
subsequent cooling. As a result of this work,
Trusova showed that reducing the carburized
steel temperature to 800 C prior to quenching
would reduce the potential for cracking and a
double tempering at 580 to 600 C for carburized case structures containing i1.2% C.
Because of the alloy and high carbon content
in many case-hardened steels, the Ms temperatures in the carburized case are typically
between 100 and 200 C or lower. These values
will vary with the carbon content in the case.
Therefore, the Mf temperature, which is approximately 215 C below the Ms, is also below the
ambient temperature. Under these conditions,
to reach the Mf temperature and therefore
minimize retained austenite content, a subzero
treatment is required. (This treatment is also
known as refrigeration or deep cooling.) To
minimize the possibility for the formation of
subsurface microcracking, tempering at 150 to
175 C prior to cold treatment is commonly
performed to stabilize the retained austenite.
Gulyaev reported that the use of cold treatment
to reduce retained austenite was most effective if
conducted immediately after quenching, as
shown in Table 6 (Ref 40).
Equipment for achieving temperatures as
low as 75 C may be relatively simple, such as
dry ice mixed with kerosene or alcohol in a
bucket. Temperatures down to 100 C can be
Table 5 Effect of quenchant temperature on
retained austenite
Martensite
start (Ms)
temperature, C
Fig. 15
Correlation curves for correcting the Steven and

Haynes Ms temperature equation. When the carbon
content is less than 0.9%, an 830 C soak of over 2 h is required
to produce a fully austenitic structure.
150
150
150
150
Quenchant
temperature
(Tq), C
Ms Tq,
C
Approximate
retained
austenite, %
Estimated
hardness,
HRC
80
60
40
20
70
90
110
130
45
35
29
25
52
56
57
58
18/8/2008 3:18PM Plate # 0
pg 193
achieved by mechanical refrigeration. For lower

temperatures, down to 195 C, liquid nitrogen
can be used (Ref 41).
After cold treatment, the presence of retained

austenite can be assessed by comparing the
hardness before and after refrigeration. An
increase in hardness is expected if retained
austenite was present and was transformed to
martensite as a result of the cold treatment. Since
cold treatment increases martensite as a result of
the loss of retained austenite, the carburized
steel must be tempered a second time (150 to
200 C for 1 to 2 h) to reduce the potential for
cracking. Table 7 shows the effect of subzero
treatment after quenching to reduce the presence
of retained austenite (Ref 42).
Although subzero treatment of carburized
parts does provide a reduction of retained austenite, the degree of transformation at a given
temperature is variable, depending on the
amount of retained austenite at the beginning of
the subzero treatment, the elapsed time between
quenching and subzero treatment, intermediate
thermal treatments such as tempering, the level
of compressive stress, any cold working of the
material, and part design (Ref 41). However, it
has been reported that fatigue resistance is
decreased due to localized residual stresses
imparted by the subzero treatment (Ref 36). The
case ductility also seems to be negatively
Table 6 Effect of time delay between
quenching and cold treatment on retained
austenite reduction
Holding time at room
temperature before subzero
treatment at 183 C
Fig. 16
Trusova dilatomer curves for tempering of carburized steels quenched from 950 C
Retained austenite, %
Steel Kh-12
Steel SKh-12
24
46
...
48
36.5
54.5
55
...
23 min.
24 h
45 days
60 days
Table 7 Effect of subzero cooling after quenching

Heat treatment
after carburizing
Oil quenched
from 800 C,
low-temperature
tempered
Tempered at 650 C,
oil quenched
from 800 C,
low-temperature
tempered
Air-cooled from
900750 C, oil
quenched
low-temperature
tempered
Bending
strength
Impact
strength
Condition
Retained
austenite, %
Hardness, HRC
MPa
kg/mm2
MPa
kg/mm2
As-quenched
Subzero treated
62
20
54
62
1530
1442
156
147
25.5
19.5
2.6
2.0
As-quenched
Subzero treated
34
10
60
62
1697
1608
173
164
40
...
4.1
...
As-quenched
Subzero treated
90
20
47
60
1618
1353
165
138
59
19.5
6.0
2.0
Note: Steel is 18Kh2N4VA. Subzero treatment conducted at 120 C.
18/8/2008 3:18PM Plate # 0
affected (Ref 43). In addition, subzero treatment

increases the stability and induces microresidual
stresses in any remaining retained austenite (Ref
43). Although subzero treatment does decrease
retained austenite, in view of these complications, Parrish has suggested that optimization of
the following process variables be considered
before employing this process: quenchant temperature, surface carbon content, steel composition, and the use of reheat quenching (Ref 43).
Increasing amounts of retained austenite will
produce corresponding decreases in tensile
strength, as shown by Fig. 17, although increasing strains can lead to the transformation of
retained austenite to martensite (Ref 37). However, conflicting test results make it difficult to
predict if the martensite formed by such straininduced transformations is beneficial or not,
since one study reported by Parrish stated that
the strain-induced martensite was more ductile,
and another stated that the untempered martensite was more brittle.
Koistinen showed that the distribution and
magnitude of residual stresses in carburized
steels was governed by the amount of retained
austenite (Ref 44). Figure 18 illustrates the
magnitude of residual stress as a function of
the amount of retained austenite and position
in the case for carburized SAE 8620 and 5140
and carbonitrided SAE 1118. The maximum
compressive residual stress occurs at the position where the ratio of martensite/retained austenite is maximum (Ref 37).
Fig. 17
Dependence of stress for first detectable plastic strain

(approx. 0.0001) on retained austenite content.
AQ, air quenched; OQ, oil quenched; T, tempered
It is assumed, as Fig. 19 shows, that increasing

the retained austenite content will reduce the
low-stress, high-cycle fatigue limit of carburized
steel (Ref 37, 45, 46).
In a study using a flexural four-point bending
fatigue test and carburized SAE 8620 steel test
specimens, it was shown that increasing the
Fig. 18
Residual-stress distribution and retained austenite

content in case-hardened steels
pg 194
18/8/2008 3:18PM Plate # 0
retained austenite in the case resulted in longer

fatigue life and that fatigue life is directly proportional to the square root of the grain size
(d 0.5) (Ref 47). Jeddi et al. also showed that
fatigue strength of carburized 14NiCr11 steel
improved with increasing retained austenite content in the hardened case (Ref 48). In this work, it
was shown that the level of improvement was
related to the microstructure and residual-stress
distribution within j200 mm of the surface. In
addition, although the improvements in fatigue
Fig. 19
Fatigue limits of plasma- and gas-carburized test

specimens as a function of retained austenite content
are attributable to uniform retained austenite

content throughout the microstructure, only a
maximum of approximately 40% of the retained
austenite transformed to martensite at any depth
due to the cyclic loading.
The impact fatigue resistance is also dependent on the amount of retained austenite and the
level of applied stress (Ref 37). Figure 20 shows
that impact fatigue resistance actually increases
with increasing retained austenite at the highest
stress loading, while lowest stress loading
exhibited the opposite effect.
Increasing retained austenite content resulted
in corresponding improvements of carburized
SAE 8620 steel using an abrasive wear test
utilizing a pin-on-disk tribometer, as shown in
Fig. 21 (Ref 47).
Sliding wear tests were conducted on carburized SNCM21, which corresponds to SAE
8620 steel, and carburized SCM4, which corresponds to AISI 4140. Although SCM4 is not
typically carburized for the study reported, it
was used to represent an example of high core
strength and case depth. The results are shown in
Fig. 22 (Ref 49). According to these results,
sliding wear resistance increases with increasing
retained austenite at a 40 kg applied load up to a
critical retained austenite level, which, for this
work, was approximately 30%, at which point
the wear resistance decreased.
The presence of retained austenite can also
exhibit a dramatic effect on the scoring resistance of carburized steels. Kozlovskii studied
the scoring resistance of carburized 20Kh2N4A
(0.21% C, 0.62% Mn, 0.20% Si, 3.50% Ni,
Fig. 21
Fig. 20
Effect of retained austenite on impact fatigue resistance of a carburized 1.45C-11.5Cr steel
Effect of retained austenite (RA) on abrasive wear.

Sample A, HRC = 59.7+1.8, RA = 37; sample B,
HRC = 62.7+1.2, RA = 6%; and sample C, HRC = 61.4+1.5,
RA = 23%
pg 195
18/8/2008 3:18PM Plate # 0
pg 196
and propagation. Conversely, tensile stresses

reduce desirable mechanical properties such as
fatigue, fracture, and wear.
Residual stresses are classified as the macrostress or residual stress of the first kind, which
acts over a few grains, residual stresses of the
second kind, and residual stresses of the third
kind. Residual stress of the second kind is the
difference between the average residual stress
with a grain and the residual stress of the first
kind. Residual stress of the third kind refers to
stress variations within a grain. Residual stresses
of the second and third kind are microstresses
(Ref 51). In most engineering materials, such as
steel, residual-stress variation between microstructural phases is typically more important
than microstresses. The primary focus of this
chapter is on macrostresses, residual stresses of
the first kind.
Carburizing and carbonitriding introduce
surface and subsurface compressive stresses as a
result of the formation of a carbon-enriched
case. The increased carbon content in the case
relative to the core significantly reduces the Ms
temperature in the case, as illustrated in Fig. 23
(Ref 25). Because of the depressed Ms temperature, austenite-to-martensite transformation
begins in the core before the surface, even
though the surface temperature is lower. The
volumetric expansion of the martensite in the
core can be accommodated by the relatively hot
untransformed austenite nearer the surface.
Upon further cooling, the temperature at the
surface is less than the Ms temperature of the
carbon-enriched case, and it begins to transform
to martensite. The martensite that formed first in
the core is cooler and stronger than the austenite
that is now transforming, and it resists the
expansion of the higher-carbon-containing surface martensite now forming at the surface,
which puts the surface in compression relative
to the core. To maintain balance, the core is
now in tension. This is illustrated by Fig. 23
(Ref 25).
1.42% Cr) using a roller machine test and found

that even a relatively small amount of retained
austenite could exhibit a large decrease in scoring resistance, even if the hardness was affected
only minimally, as shown in Table 8 (Ref 50).
Residual Stress
Residual stress is defined as a tensile or
compressive force within a material such as steel
without application of a thermal gradient or an
external force. Residual stresses are produced by
phase transformation, plastic deformation, or
thermal effects such as contraction upon cooling. Newtons laws require that compressive
residual stresses at the surface of a material are
balanced by tensile stresses within the material.
Typically, compressive residual stress exhibits
favorable effects such as improved fatigue life
and stress corrosion by inhibiting crack initiation
Fig. 22
Sliding wear rate (at 200 rpm) as a function of

retained austenite content. A, carburized SNCM21,
40 kg load, sliding distance of 864 m; B, carburized SCM4, 40 kg
load, sliding distance of 864 m; C, carburized SNCM21, 20 kg
load, sliding distance of 1728 m; D, carburized SCM4,
20 kg load, sliding distance of 1728 m
Table 8 Effect of retained austenite on scoring resistance of carburized and carbonitrided case
Case carbon
content, %
Case depth,
mm
Surface
hardness, HRC
On
surface
At depth
of 1.5 mm
Scoring
load, kg/cm
Compressive
stress at
scoring
load, kg/cm2
Carburize at 930 C for 8 h, slow cool to 750 C,

quench in oil, temper at 180 C for 1.5 h
0.81
1.1
5759
48
60
930
12,020
Carburize at 930 C for 8 h, slow cool to 20 C,

temper at 600 C for 2 h, quench from 830 C
in oil, temper at 180 C for 1.5 h
0.84
1.1
6062
16
355
7424
Retained sustenite
Treatment
18/8/2008 3:18PM Plate # 0
Fig. 23
Schematic illustration of carbon content, retained

austenite, and residual stresses in the case of carburized steels
Fig. 24
An evaluation was made of 70 steels that were

gas carburized, oil quenched, and tempered
between 150 and 180 C. Case depths were
1 mm or less, and the core carbon contents
varied between 0.15 and 0.20%. The range of
residual-stress profiles obtained is shown in
Fig. 24, where it can be observed that the compressive residual stresses acted throughout most
of the case (Ref 25).
It is important to consider the history of residual stresses due to dimensional changes that
may occur during any stage of the manufacturing process and therefore may contribute to the
final residual-stress condition of the heat treated
component. These contributions to the stress
history are illustrated in Fig. 25, which is a
simulation of the production of a carburized
8620 steel cylinder that was subsequently
quenched in unagitated water (Ref 52).
The development of residual stress was
achieved in the case of a 9.5 mm diameter
chromium-molybdenum SCM420 (0.23% C,
0.72% Mn, 1.12% Cr, 0.21% Mo) steel test
specimen that was gas carburized at 930 C until
the case depth shown in Fig. 26 was achieved, oil
quenched, and then tempered for 1 h at 200 C
(Ref 53). These data show that the compressive
residual stress increases with increasing case
depth. Generally, the magnitude of the surface
compressive stresses will be dependent on
the ratio of the case and core thickness. When
the core is thicker than the case, the surface
Range of residual stresses obtained for 70 carburized steels
pg 197
18/8/2008 3:18PM Plate # 0
compressive stresses will be high. When the case

is thicker than the core, the surface compressive
stresses in the case will be lower, and the tensile
stresses in the core will be higher (Ref 36). For
automotive gears, the inversion point occurs
where the surface compressive stresses become
tensile, which was shown to occur at the point
where the hardness becomes equivalent to the
core value (Ref 54). Furthermore, the formation
of surface compressive stresses was shown to
be fundamental to the prevention of fatigue
cracking.
One of the primary reasons for conducting the
carburizing process is to improve the fatigue
performance of a heat treated component.
A surface compressive stress will inhibit the
formation and growth of surface cracks, which
is important since fatigue failures are typically
initiated at the surface (Ref 55). Kanetake, for
example, showed an approximately 40% improvement in fatigue strength of the carburized
SCM420 steel (Ref 53). Shot peening is also
used to create residual surface compressive
stresses that will increase fatigue properties of
steel components (Ref 56). Therefore, it has
been of continuing interest to examine the
potential of fatigue property improvements not
achievable by either treatment alone (Ref 55,
5759). For example, Shaw et al. have reported
up to a 75% increase in fatigue strength by
Fig. 25
combining carburizing and shot peening of gears

prepared from gas-carburized 20MnCr5 steel
(Ref 57). Before continuing this discussion, a
brief overview of the shot peening process is in
order.
Shot peening (and shot blasting) involves impinging the surface to be treated with spherical
media called shot. (Shot peening should be differentiated from shot blasting. Shot blasting a
process in which an abrasive material is accelerated through a pressurized nozzle or centrifugal wheel and directed at the surface of
a part to clean or otherwise prepare the part
surface for further treatment, (Ref 60)). Shot
peening is a cold working process where each
individual spherical ball impinging the surface
acts as a miniature hammer that plastically
deforms and work hardens the surface by
creating a small indentation upon impingement,
as illustrated in Fig. 27 (Ref 61). The indentation
process causes the surface to yield in tension. To
balance the tensile forces involved in indentation, the subsurface is in a highly stressed
compression state. As the process continues, the
indentations overlap, and a uniform layer of
metal is in residual compressive stress. The
compressive stress is the result of superposition
of residual stress formed by surrounding shots
(Ref 62). The magnitude of the compressive
stresses that are formed is material dependent
Illustration of the tangential stress history over the first 20, of a water-quenched 1.27 cm diameter 8620 carburized steel
cylinder. The carbon gradient and retained austenite content are shown in Table 9.
pg 198
18/8/2008 3:19PM Plate # 0
and is generally at least 1/2 the yield strength of

the material (Ref 56). Some of the variables of
the shot peening process include the shape and
type of shot, size of shot, and impingement
velocity.
The effect of shot peening on contact fatigue
under rolling-sliding conditions was studied
using a carburized and hardened 20CrMnTi steel
(0.21% C, 1.09% Cr, 1.0% Mn, 0.025% Ti,
0.02% P, 0.008% S), and, as a result of that work,
it was shown that failure may produce three
kinds of cracks, which were classified as: surface, which initiate from 0.015 to 0.05 mm;
shallow sub surface; and deep subsurface, which
initiate from 0.3 to 0.5 mm (Ref 63). Depending on the specific contact stresses on the
Table 9 Carbon gradient and retained austenite

in the 8620 carburized steel related to Fig. 25
Depth below the surface
mm
in.
Carbon, %
Retained
austenite, %
0.0
0.5
1.0
1.5
2.0
0
0.020
0.040
0.060
0.080
1.20
1.10
0.80
0.40
0.25
32.0
28.0
15.0
4.0
0
Fig. 26
200
C
Residual-stress profiles of SCM420 steel that was gas

carburized at 930 C, oil quenched, and tempered at
shot-peened surfaces, cracking, pitting,

shallow spalling, and deep spalling may occur.
In addition to providing substantial improvements in fatigue strength, carburizing and shot
peening offer other benefits, such as reducing the
deleterious effects of internal oxidation. However, in the absence of surface oxidation and
oxide inclusions, MnS inclusions will then act as
fatigue initiation sites (Ref 57).
If a case-hardened surface is shot peened with
sufficient intensity, a stress-induced transformation of retained austenite to martensite may
be observed, and the surrounding volume constraint may result in a deepening of the surface
compressive residual stress (Ref 55, 58). Nakonieczny et al. (Ref 58) have shown that plastic
strain induced by shot peening reduces retained
austenite in tempered martensite and produces
a new e-phase (Fe2C and Fe3C). Peyrac, in
another study, also showed significant retained
austenite transformation and increases in
residual stress as a result of shot peening of gascarburized 18NCD6 steel (Ref 64). Selected
results from this study are summarized in
Table 10.
It is known that surface structure anomalies,
also known as soft skin layers, which include
internal oxides and nonmartensitic structures,
including retained austenite (Ref 65) near the
surface, will decrease fatigue properties of
Fig. 27
Illustration of the plastic deformation of the surface

and resultant stress distribution after shot peening.
Source: Ref 61
pg 199
18/8/2008 3:19PM Plate # 0
pg 200
gas-carburized steel (Ref 10). Structural surface

anomalies act as preferential zones for fatigue
crack initiation under low stress amplitudes.
Shear-type crack growth can occur in ductile
retained austenite in the near-surface region of
the case for gas-carburized steels. Kikuchi et al.
have shown that shot peening is very effective in
improving the fatigue properties of carburized
steels with these surface structure anomalies
(Ref 59, 65). In fact, in their study, the fatigue
properties were essentially comparable for a
20 mm diameter chromium-molybdenum steel
(0.16% C, 0.26% Si, 0.74% Mn, 0.012% P,
0.013% S, 1.01% Cr, 0.18% Mo) bar that was
carburized at 950 C for 1 h with and without
surface structure anomalies (to a depth of approximately 30 mm) after shot peening. As a
result of this work, it was concluded that since
internal oxides near the surface can also act as
preferential crack initiator sites, it is desirable to
avoid the presence of such defects during gas
carburizing (Ref 65).
Although components such as gears are carburized and shot peened to introduce the desired
level of surface compressive stresses, there are
other processes that can be used to introduce
compressive stresses. One such process is presetting. Presetting involves the introduction of
an overload that causes yielding in the area of
maximum stress concentration, such as in the
root area of a gear tooth. When the load is
released, a residual stress is introduced in that
area. The induced stress is compressive on the
side being loaded and tensile on the other side
(Ref 66). Woods et al. evaluated presetting to
improve the bending fatigue of carburized AISI
4120 steel spur gear teeth and found that presetting introduces compressive stresses in
the area of a gear tooth where fatigue cracks
originate. The results of this work showed
that presetting provided substantially longer
fatigue life.
Dimensional Stability
Dimensional stability has two components:
size (distortion) and shape (warpage). Distortion
is defined as an irreversible change in the
component during heat treatment (Ref 67).
While changes in shape such as straightness
(warpage) can be corrected by application of
stress or by tempering in the elastic range
(reversible), size changes are irreversible and
cannot be changed in this way. Metallurgically,
distortion may be thermally or transformationally derived. Size distortion typically refers to
dimensional variation due to growth or shrinkage that is due to volumetric changes attributable
to microstructural phase transformations (Ref
19, 68). Figure 28 shows the effect of the temperature dependence of the specific phase
volume of different steel transformation phases
(Ref 69).
Variables that affect distortion include (Ref
19, 68, 70):

Chemical composition and hardenability

chemical and phase composition as well as
hardenability (distortion increases as hardenability increases), as shown in Fig. 29
(Ref 67, 71, 72)
Steelmakinggrain size and hardenability
Hot workinghot reduction, length and
direction of fiber
Prior heat treatmentresidual stresses,
grain size and uniformity of microstructure
Geometrycheese blanks, shafts, rims
Heat treatment aspectsheat rate, cooling
rate (quench severity), quenching temperature, jigs and fixtures, plug quenching. Plug
quenching is used to minimize dimensional
change of inside diameters such as roundness and taper distortion of ring-shaped
components (Ref 67, 68) (generally, salt
bath quenching yields minimum distortion
Table 10 Effect of shot peening on retained austenite transformation and residual stress of
gas-carburized 18NCD6 steel
Heat
treatment(b)
Depth modified by
shot peening treatment, mm
Retained austenite
on surface, %
Retained austenite
converted to martensite, %
None
T1
T2
...
...
24.7
36.8
...
...
300
300
G1
T1
T2
100
100
5.7
14.9
19.0
21.9
1450
1350
G2
T1
T2
200
200
12.1
22.1
12.6
14.6
980
930
Shot peening(a)
smax, MPa
(a) G1, steel shot, BA 300, F 2530A, overlap rate 150%; G2, steel shot, BA 800, F 5580A, overlap rate 150%. (b) T1, carburize for 3 h. at 920 C, plateau at 850 C,
oil quenching at 60 C, tempering for 2 h. at 150 C; T2, carburize for 3 h at 960 C, oil quenching at 60 C, tempering for 2 h at 150 C
18/8/2008 3:19PM Plate # 0
Fig. 28
Variation of the specific phase volume of different steel transformation phases as a function of temperature. Source: Ref 69
Fig. 29
Effect of steel hardenability on shape distortion
relative to hot oil) (Ref 67). Furnace temperature uniformity, case depth uniformity,
prequench temperature, fluid flow during
the quench (Ref 73), number of times a
part is quenched, carburizing temperature
(Ref 70)

RackingVashchuk et al. reported that

high-temperature deformation of large gears
could occur due to their own weight (Ref
73). They recommended the larger diameter
of a gear wheel should be on the floor of the
furnace (bottom) to support the end face and
that the smaller diameter should be the free
surface
MachiningParts should be machined as
near final dimensions so that outer case
will not require grinding after carburizing
(Ref 24, 68). Residual stresses due to prior
machining exhibit a large effect on distortion, and as shaved thickness increases,
potential distortion increases (Ref 67)
Method of green part manufacture, for
example, parts machined from bar or tubing
or forged from bar or tubing and then
machined (Ref 70)
Growth of surfaces during carburizing
Murzin et al. reported that the cooling rate
from the carburizing temperature to the
prequench temperature exhibited one of the
pg 201
18/8/2008 3:19PM Plate # 0
strongest effects on gear wheel distortion

(Ref 74).
Variables that affect warpage include (Ref 19,
68, 75):

Stresses resulting from phase transformations

Nonuniform residual stresses in the original
blank, such as those due to prior heating,
including stress relief
Nonuniform heating or coolingfurnace
shape, part shape, heat control
Insufficient furnace timeundersoaking
Creepmethod of stacking and fixturing of
parts during heating and quenching (hanging
versus standing)
Internal stresses due to machining
Bulgakov reported that the most effective

method of controlling stresses to reduce or
stabilize warping was by controlling the hardenability of the steel (Ref 75). Hardenability
control will permit control of the phase transformation and reduction of volumetric changes
leading to structural stresses and warping. In particular, it was shown that slow cooling of carburized steels in the range of 1100 to 900 C
Fig. 30
after forging was extremely undesirable since

this facilitates the formation of hard-to-dissolve
particles and austenite grain growth during heat
treatment. Optimal warpage control was achieved with grain sizes not greater than grade 7.
A comparison was made of the effect of
transformation behavior on dimensional stability of machine parts, such as gears, constructed
from two different carburized steels: 20KhGR
(0.18% C, 0.82% Mn, 0.24% Si, 1.07% Cr,
0.20% Ni, 0.0031% B) and 12KhN3A (0.14% C,
0.45% Mn, 0.21% Si, 0.78% Cr, 0.85% Ni).
Since only very low cooling rates occur during
carburizing, thermal stresses would be expected
to be minimal, and any stresses that result would
be due to austenitic transformation, which would
be dependent on the steel chemical composition
(Ref 76).
The isothermal transformation diagrams for
both steels before and after gas carburizing are
shown in Fig. 30 (Ref 76). These diagrams
show that austenite is less stable in the 20KhGR
alloy and that relative stability of austenite
compared to the 12KhN3A alloy remains after
carburizing.
Analysis of the kinetics of austenite transformation of alloys, before and after carburization
Isothermal transformation diagrams. (a) 20KhGR and (b) 12KhN3A alloys before carburizing (c) 20KhGR and (d) 12KhN3A
alloys after carburizing. Source: Ref 76
pg 202
18/8/2008 3:19PM Plate # 0
(Fig. 31) over the range of cooling rates used

(0.79 to 175 C/min), showed that the noncarburized 20KhGR yielded mostly pearlite.
The 12KhN3A alloy yielded predominantly
bainite at the higher cooling rates, and at lower
Fig. 31
cooling rates, the pearlitic transformation

predominated for both the carburized and
uncarburized alloy.
When air cooled, the case of the carburized
20KhGR consisted of pearlite and carbides, and
Dilatometric curves for the transformation of austenite in 20KhGR and 12KhN3A steels. Curves (a) and (b) were after
carburizing, and curves (c) and (d) were before carburizing. (a,c) 20KhGR. (b,d) 12KhN3A. The cooling rates are: 1, 0.79; 2,
1.46; 3, 4.6; 4, 5.0; 5, 70; 6, 175 C/min. Source: Ref 76
pg 203
18/8/2008 3:19PM Plate # 0
the core consisted of pearlite and ferrite. Upon

rapid cooling, the carburized case had a small
amount of martensite with a correspondingly
lower amount of pearlite, and in the core, a lower
amount of ferrite with the remainder being pearlite was obtained. However, for the 12KhN3A
alloy, air cooling produced predominantly martensite and retained austenite in the case and
correspondingly less pearlite. The core contained bainite (Ref 76).
In practice, it was shown that gears manufactured from 20KhGB steel exhibited an average reduction in diameter (0.7 to 1.4 mm) from
463+0.3 after carburizing. This was attributed
to reduced stability of austenite in the pearlite
region. Conversely, for the 12KhN3A steel, the
gears increased in size (0.2 to 0.5 mm) after
carburizing because austenite was more stable in
the pearlite region. These data suggest that distortion can be controlled by alloy and cooling
rate selection, or improper selection of cooling
rate or alloy can result in unacceptable component distortion (Ref 76).
Fig. 32
The effect of phase transformation behavior

in the case and the core due to thermal gradients
such as those occurring during heating and
cooling will affect shape and size distortion of
small parts, as illustrated by Fig. 32 (Ref 19). If
the thermal and transformational stresses exceed
the yield strength of the steel, corresponding
distortion will occur.
Quenching and Grinding Cracks (Ref 77)

Quenching cracks occur when tensile
stresses of the first kind are greater than the
material strength. Quenching cracks typically
occur during, or in some cases after, quenching at
temperatures less than the Ms temperature. Susceptibility to cracking increases with the carbon
content of the steel, increasing austenitizing
temperature, and cooling rate, especially in the
Ms Mf transformation temperature range.
The probability of cracking during quenching
increases with the presence of stress raisers
Resulting distortion after heat treatment of different steels after quenching in oil and water
pg 204
18/8/2008 3:19PM Plate # 0
such as notches, abrupt changes in section size,

keyways, and holes. Quenching cracks are irreversible but can be minimized by appropriate
design modifications, reduction in hardening
temperature, and use of lower quench severity.
The reasons for the occurrence of quenching
defects include:

Poor surface cleanliness, such as residual

forging and metalworking lubricants, and so
on, which leads to nonuniform quenching,
increased thermal gradients, and soft spots
Incorrect loading and arrangement of parts
in the furnace, which leads to nonuniform
heating and related distortion
Excessive heating rates, which may lead to
warping and cracking
Lack of a protective atmosphere to eliminate
oxidation and decarburization of the steel
surface, which will lead to reduction in
mechanical properties after hardening and
a decrease of hardness of the superficial
layer
Excessive cooling rates and incorrect immersion into the quenching bath, causing
cracking, warping, and twisting
Insufficient cooling rates or undersized
quench tanks, which will inhibit the desired
martensitic transformation
Quenching cracks, which are characterized by

relatively large depth and short length, rarely
occur in the case of carburized or carbonitrided
components. This can be explained by more
Fig. 33
beneficial compressive surface stresses than

those typically formed in higher-carboncontaining through-hardened parts. However,
internal cracks may form in the core below the
carburized surface or in the transition zone, that
is, in the places with the largest tensile stresses in
carburized parts. Cracks may also form on the
surfaces and corners of carburized parts, which
is related to triaxial tensile stresses in these
locations. Therefore, to prevent cracking, the
case must be sufficiently deep so that stresses
developed at any point below the surface are less
than the fatigue limit of the material at that point
(Ref 1).
When steel contains greater than 0.5% C in a
martensite matrix, such as in the carburized case,
intergranular fracture along prior-austenite grain
boundaries may occur. In this situation, the
intergranular fracture is due to the presence of
both phosphorus and cementite formation on the
austenite grain boundaries during austenitizing
or cooling from austenitizing temperatures.
Krauss has referred to this fracture mechanism
as quench embrittlement and has suggested
that the mechanism for this to occur is analogous
to quench cracking in through-hardened steels,
which is due to the formation of tensile surface stresses during quenching, as described
previously (Fig. 33) (Ref 78). However, since
relatively high surface compressive stresses are
present in properly carburized steels, quench
cracking should not exist if conditions for potential intergranular cracking are present.
Scanning electron micrographs of overload case fracture surfaces in carburized SAE 8620 steel. (a) Quenched directly after
carburizing at 927 C (1700 F). (b) Reheated to 788 C (1450 F). Both specimens were tempered at 145 C (300 F).
pg 205
18/8/2008 3:19PM Plate # 0
Krauss has reported that carburized steels will

fail by intergranular cracking if sufficent bending or tensile stresses are applied to offset the
compressive stresses in carburized cases. The
fracture map shown in Fig. 34 illustrates three
conditions where susceptibility to intergranular
cracking can be minimized: carburizing, intercritical austenitizing, and applications where
loading is Hertzian or compressive (Ref 79).
In some cases, microcracking can occur with
higher-carbon lath martensite matrices, which
are present in the hardened carburized case.
Microcracking is due to contact of martensitic
plates with each other or with the austenitic grain
boundaries. The potential for microcracking
increases with the austenization temperature.
Prior-austenite grain size also affects microcrack density, which decreases with decreasing
prior-austenite grain size. However, microcrack
density is not affected by quench severity
(Ref 78, 80). Figure 35 illustrates examples of
microcrack formation in the carburized case of
SAE 8620 steel (Ref 78).
The presence of microcracks can further lead
to a surface defect called flaking, which refers to
Fig. 34
flakelike fracture and subsequent peeloff. (Note:

Flaking is initiated at microcracks that may also
be caused by surface damage due to lubricant
contamination by chips, shavings, burrs, or
abrasive powder ingression into the lubricating
system.)
Krauss reported that fatigue resistance
decreased with increasing microcrack formation. Crack initiation occurred at the site of the
microcrack that acts as a stress concentrator (Ref
78). Although microcracks can be removed by
surface grinding and polishing, fatigue failure
may still be initiated at prior-austenite grain
boundaries, intergranular surface oxides, or
surface defects such as scratches, machining
marks, and surface asperities due to roughness.
Aksenova et al. showed that contributing
factors to cracking of case-hardened gear wheels
included residual tensile stresses in the case-core
interface, grain growth and overheating, supersaturation of the case with carbon, excessively
high cooling rates during quenching, too low
a quenching temperature, and insufficient residence time in the quench tank after immersion
(Ref 81).
Krauss fracture map illustrating conditions where susceptibility to intergranular cracking can be minimized: carburizing,
intercritical austenitizing, and Hertzian or compressive contact loading
pg 206
18/8/2008 3:19PM Plate # 0
McEvily et al., reported the use of fractographic analysis to explain cracking of a

carburized AISI 9310 gear at the case-core
transition zone. One of the factors was the differential in the Poisson ratio (n) between the case
and the core that develops during a monotonic
bending test when the core deforms plastically
while the case deforms elastically. McEvily
used as an example to illustrate this point a
9 mm diameter round bar with a 1 mm case.
When the bar is first loaded initially and the
deformation is elastic, the Poisson ratio for the
case and the core is equal. Upon further bending,
the core will deform plastically while the case is
deforming elastically. However, because of the
difference in the Poisson ratio between the case
(n = 0.3) and the core (n = 0.5) at this point,
radial tensile stresses will develop. For a tensile
strain of 1%, the radial strain in the case (relative
to the centerline) would be 0.003, and the radial
strain in the core would be 0.005. To remain
compatible, the difference in the strain and
corresponding displacements must be accommodated by the formation of a tensile stress.
However, this radial stress will result in a state of
triaxial stress that will promote brittle behavior
by inhibiting plastic deformation. The total
tensile stress may then be developed sufficient
to result in rupture at the case-core interface
(Ref 82).
Grinding Cracks. Grinding may be used for
postprocessing of components to remove growth
and distortion that may have resulted from carburizing and carbonitriding. Grinding is also
performed to remove such metallurgical features
as carbide films, internal oxidation, and hightemperature transformation products that may
impart deleterious performance properties. In
addition, grinding processes are commonly used
Fig. 35
Microcracks in the martensitic case of a coarsegrained SAE 8620 steel
to create the desired surface finish to improve

bending and contact fatigue and lubrication
properties.
Surface cracks in the carburized case may
occur during the grinding process, which can be
attributed to microstructural transformations
and thermal stresses producing tensile forces. If
the tensile forces in the case exceed the material
strength, then surface cracks will result. This is
due to the difference in the specific volume of
the transformational phases present in the case
structure, primarily martensite and austenite.
Structural defects, such as those caused by
inclusions, will also influence the susceptibility
of the steel to cracking.
Generally, crack creation during grinding is
influenced by thermoelastic tensile stresses that
are created in the surface cooling zone during
grinding. They are dependent on thermal and
mechanical properties of the material, maximum
contact temperature, grinding feed depth, and
cooling rate.
One study conducted on the generation of
grinding cracks showed that microstructural
heterogeneity, such as the presence of carbide
inclusions (particularly those with a mean diameter of 6 to 10 mm), which were shown to be
associated with large internal residual stresses,
were a predominant cause of grinding cracks
(Ref 83).
Grinding burns arise when excessive heat is
generated during the grinding process, and this is
characterized by surface discoloration. The term
grinding burn refers to localized surface temperature increases at least sufficient to cause
tempering of the martensitic surface, resulting in
localized soft spots. Furthermore, since carbide
precipitation volume contraction accompanies
tempering, the burnt areas are in tension and, if
the resulting tensile stresses are sufficient, subject to tranverse cracking. However, in other
cases, the increase may be in excess of the Ac3
temperature, producing an austenitic surface
that, upon rapid cooling, may produce a hard,
light-etching, martensitic thin layer at the surface. This induced defect is known as a rehardening burn, which is characteristically
surrounded by a layer of tempered steel (Fig. 36)
(Ref 84). In this case, the rehardened zone is in
compression due to the martensitic volume
expansion, and the surrounding areas of tempered martensite are in relative tension. Cracking may occur in the area surrounding the
rehardened material or in the interface between
the two (Ref 84).
pg 207
18/8/2008 3:19PM Plate # 0
There are two possibilities to prevent the

occurrence of thermal defects in the surface
resulting from the grinding process:
Reduction of heating due to the rotation of a

grinding wheel by reducing the speed of the
grinding wheel will reduce heating and
friction. Conversely, increasing the speed of
the grinding wheel will increase heat production and the potential for the formation of
grinding burns due to overheating.
Increasing heat abstraction due to the
grinding process and reducing the contact
time between the grinding wheel and part
will reduce the tendency for grinding burns.
Grinding cracks may exhibit characteristic

short, parallel cracks (Fig. 37), or they may
exhibit a chicken-wire pattern and are typically between 0.076 and 0.13 mm (0.003 and
0.005 in.) deep. The parallel cracks are typically
deeper than the chicken-wire pattern. Grinding
cracks form perpendicular to the grinding
direction. As indicated previously, the potential
for grinding cracks is affected by improper heat
treatment or a metallurgical structure that is
prone to cracking. For example, if the surface
temperature exceeds the Ac3 temperature, the
steel in this region may transform to austenite,
then upon rapid cooling, a hard martensite layer
may form. This effect is called a rehardening
burn. Grinding cracks may be detected by a
magnetic particle test.
Some carburizing steels such as chromium
and chromium-manganese steels, which include
SAE 5120 and 20MnCr5, may undergo overcarburizing with subsequent cracking of the case
Fig. 36
Microstructure of a section through a rehardening

burn. Original magnification: 500
upon cooling, which then renders the part more

susceptible to the formation of grinding cracks
(Ref 1). Overcarburizing leads to the formation
of a complex carbide network that is excessively
brittle, which causes greater susceptibility to
grinding cracks.
Severe grinding may lead to the development
of residual tensile stress, which can be the
initiation point for crack formation. Since cracks
will not propagate into layers of compressed
stress, it therefore may be advantageous to shot
peen the part prior to grinding to prevent the
formation of grinding cracks (Ref 86).
Parrish reported that the potential for formation of grinding cracks may be minimized by
(Ref 84):

The thermal conductivity of the steel is an

important design variable, and free carbides
and retained austenite have an adverse effect
on thermal conductivity.
The surface carbon concentration should be
between 0.7 and 0.9%.
Parts should be tempered immediately after
quenching.
The tempering temperature should be as
high as possible while still achieving the
necessary surface hardness.
Improper Case Depth (Ref 77). In a recent

study conducted by Bahnsen et al. on carburized
SAE 5120 test specimens to rate the relative
influence of surface carbon content, case depth,
and carburizing temperature on distortion, it was
Fig. 37
Example of grinding cracks on the flank of a worm

gear. Source: Ref 85
pg 208
18/8/2008 3:19PM Plate # 0
reported that of these variables, the most dominant effect was observed for case depth (Ref 87).
Case depth of carburized steel is determined by
the carburizing time and the available carbon
potential at the surface. One of the most common defects of carburized and carbonitrided
materials is an insufficient or excessive case
depth. For example, when prolonged carburizing times are used to produce a deep case depth,
a high carbon potential will produce a high
carbon content on the surface and the possible
corresponding formation of excessive retained
austenite or free carbides, which may lead to an
improper residual-stress distribution in the casehardened part. Therefore, a high carbon potential may be suitable for short carburizing times
and shallow case depths but not for prolonged
carburizing times and deep case depths. Furthermore, the fatigue limit of carburized SAE
8620 steel was related to case depth and also
microstructure, distribution of retained austenite, depth of internal oxidation, and nearsurface compressive residual stresses (Ref 10,
21). Bending-fatigue strength decreased with
the increasing case depths due to the presence of
increasing internal oxidation and nonmartensitic
transformation products at the surface. Wear
behavior of carburized 8620 steel is also related
to case depth (Ref 88).
To assure optimal quality during the production of case-carburized parts, the following are
essential:

The processing temperature should be

accurately controlled.
Maintain temperature uniformity throughout
the load.
Rack the parts to assure uniform gas flow
throughout the load.
Use uniform circulation of atmosphere
throughout the load in the furnace
To properly control case depth, either use
shim stock or sample the parts periodically
during the carburizing cycle.
Conduct the process at the lowest acceptable
temperature and time.
The components constituting a load should
possess uniform size and surface area with
respect to each other, using empirically
established carburizing conditions.
Avoid carburizing and carbonitriding in the
same furnace. Use a separate furnace for
each process, if possible.
The optimal case depth for a specific component and steel alloy is based on the design and
service conditions of the component. Typically,

the case depth is designed to provide the
necessary residual-stress distribution for the
wear requirements for the part (Ref 89). Typically, the greater the case depth, the greater the
fatigue strength (Ref 23). For carburized steel
gears used in the automotive industry, for
example, SAE 8620, hardened case depths are
generally 0.8 to 1.4 mm. Improper case depth
may be caused by establishing an unnecessarily
restrictive case thickness specification that is not
appropriate for the process or the particular
furnace in use, which leads to decreased and
nonuniform hardness and unacceptable material
properties.
Insufficient Case Hardness and

Improper Core Hardness (Ref 77)
One reason for insufficient case hardness is
the presence of incorrect microstructure, such as
bainite. The appearance of bainite in a carburized case in even small amounts will significantly decrease fatigue strength, especially
contact fatigue strength. It is an important
microstructural defect to be avoided. The presence of bainite in the carburized case is
particularly problematic because it cannot be
detected by hardness measurements and by the
severity of the quenchant used to harden the steel
after carburizing. Figure 38(a) illustrates a bainitic case microstructure of carburized SAE
8620. Figure 38(b) shows the core structure. In
the case of carbonitrided components, this is less
important because nitrogen increases hardenability of steel more than carbon.
Insufficient as-quenched case hardness is
caused by:

Insufficient carbon content in the entire case

or in the superficial zones
Increased retained austenite content
Insufficient case hardenability
Insufficient case depth
On the other hand, core hardness is dependent

on carbon content, steel alloy hardenability, and
section size.
Case carbon content less than 0.4 to 0.5% C is
easily detected by conducting a spark test.
Insufficient case carbon content resulting from
the gas carburizing process occurs when the
process is conducted with a low carbon potential, inadequate furnace pressure, or cooling the
pg 209
18/8/2008 3:19PM Plate # 0
load from the diffusion temperature to the

quenching temperature without a protective
atmosphere or improper atmosphere composition. These problems may be rectified by additional soaking in a carburizing atmosphere with
the proper carbon potential.
Hardness reduction due to an increase in
retained austenite content may occur with a
direct quenching after the carburizing process.
This is even more critical with steels containing
increased chromium or nickel, which may produce retained austenite levels as high as 80 to
100% in the surface after quenching. As the case
thickness increases, the zone containing retained
austenite may be sufficient to cause a significant
reduction in hardness. To reduce the retained
austenite content from a direct quenching process, it is important to select the proper steel
alloy for carburizing and to use the appropriate
quenchant and quenchant conditions, including
the use of a subzero treatment, if necessary.
Reduced hardenability usually does not occur
throughout the entire case but only in the 0.001
to 0.01 mm depth from the surface. Common
problems leading to reduced hardenability include internal oxidation and overcarburization
of surface zones. To improve case hardenability,
internal oxidation and overcarburizing should be
prevented.
When the case depth is too shallow, the
observed hardness is dependent on the load
applied. For example, for thicknesses of 0.3
to 0.4 mm, if the surface hardness is determined using a 60 kg load, the value will be
approximately 12 HRC units higher than if the

hardness is determined using a 150 kg load. In
practice, it is possible for the hardness to differ
by 1 to 2 HRC units from published values for
the material. Hardnesses that are i2 HRC units
less than published values may be due to the use
of an incorrect steel grade or an insufficient
austenitizing temperature for the steel hardenability and section size in use. Hardness values
iHRC units higher than published values may
be due to the use of an incorrect steel grade,
excessive case depths, or puncturing a copper
layer or paste on surfaces that are protected from
carburization.
Core Microstructure. The design material
properties of case-hardened steels are not only
dependent on a martensitic case but also on the
microstructural composition of the core. An
important design criterion is the ultimate tensile
strength, which is dependent on the microstructure of the core. For example, soft cores
(5770 N/mm2, or 50 ton/in.2) are suggestive
of a core with high ferrite content, as shown in
Fig. 39(c) (Ref 19), and a hard core (41240 N/
mm2, or 80 ton/in.2) would be expected for a
predominantly martensitic structure, as shown in
Fig. 39(a) (Ref 19). Intermediate structures
would be bainitic structures, such as those illustrated in Fig. 39(b) (Ref 19). The effect of core
microstructure on ultimate tensile strength is
illustrated by Fig. 40 (Ref 90). The approximate
relationship between the core microstructure
and hardness for a Ni-Cr-Mo steel is illustrated
in Fig. 41 (Ref 19).
The case and core microstructure of carburized SAE 8620 test specimens (0.95% C potential); carburized at 955 C
(1750 F), quenched into a 50:50 mixture of sodium nitrate and potassium nitrate at 250 C (480 F), held 120 min, then air
cooled and tempered at 250 C (480 F) for 240 min. The case is lower bainite (56 HRC), and the core (42 HRC) is lath martensite. These
images were made by etching with 10% Na2S2O5 (sodium metabisulfite). The magnification bar is 20 mm. Courtesy of G. Vander Voort,
Buehler Ltd., Lake Bluff, IL
Fig. 38
pg 210
18/8/2008 3:19PM Plate # 0
Influence of the Transition Zone. There

is a transition zone between the case and the
core, and the thickness of a transition zone is
dependent on carburizing time (transition zone
thickness increases with carburizing time),
carburizing medium, and carburizing temperature (excessive carburizing temperature will
lead to increased pearlite content in the core).
After subsequent hardening, if there is too great
a transition between the case and the core, there
is increased potential for peeling and chipping,
resulting from the presence of a martensitic
carburized case and an uncarburized core containing sorbite.
Fatigue cracks occur most often in the transition zone, which subsequently propagate into
the core and into the carburized case. When the
gradient between the case and the core changes
too rapidly, operations such as grinding or
when a component is subjected to bending due
to heavy loadings, could lead to peeling and
chipping failures. For example, deep grinding
could not be performed. Increasing the depth
of the transition zone will increase the strength
of adhesion of the case to the core. Then, if the
transition zone is sufficiently large, deep grinding operations may be performed.
Figure 42 provides the microstructure of
two transition zones. Figure 42(a) illustrates a
gradual transition between the case-core
microstructure. A more rapid transition between
the case-core microstructure is illustrated
in Fig. 42(b). Typically, a transition zone
such as that illustrated in Fig. 42(a) is desired,
since it will exhibit a lesser tendency for chipping.
Influence of Surface Carbon Content
Fig. 39
Microstructures obtained by cooling a 0.16%C3%Ni-Cr steel from 920 C. (a) Fast cool (920
200 C in 30 s), giving low-carbon martensitic structure of
1590 MPa ultimate tensile strength (UTS). Original magnification: 800. (b) Intermediate cooling (920250 C in 200 s),
giving bainitic structure of 1360 MPa UTS. Original magnification: 800. (c) Slow cool (920250 C in 104 s), giving a ferrite/
pearlite structure of 740 MPa UTS. Original magnification:
800
Overcarburizing or Overcarbonitriding.
Important microstructural defects related to
carburized or carbonitrided case structure
include overcarburization or overcarbonitriding
of the case and coarse grain structure. Excessive
carbon content (carburizing) or carbon and
nitrogen (carbonitriding) is typified by the presence of carbides or carbonitrides in the case,
which creates an almost continuous nonetching
area. Case-hardened steels with these microstructures characteristically exhibit increased
brittleness, a propensity for chipping during
grinding and use, and decreased fatigue strength
and pitting resistance.
pg 211
18/8/2008 3:20PM Plate # 0
Fig. 40
Fig. 41
Approximate effect of microstructure on the ultimate tensile strength of low-carbon, low-alloy steels
Approximate relationship between core microstructure and hardness of a Ni-Cr-Mo carburizing steel (approximately 4%
alloy content) with approximately 0.16% C. The alloy content/carbon content extension (upper right corner of the figure)
permits phase percentage plots to be adjusted in relation to the fixed hardness scale to approximate core strength for other steels. Below
250 HV represents slow-cooled (normalized) and annealed steels, and bainite can be read as bainite, pearlite, or spheroidized carbides.
Above 250 HV refers to quenched steels. For the 180 C tempered condition, there will be zero change at 360 HV and below, but there
will be a 20 HV loss at 100% martensite.
pg 212
18/8/2008 3:20PM Plate # 0
Figure 43 illustrates the effect of carbon

content on the hardness of martensite in carbon
and alloy steels. Increasing carbon content to
0.5% increases hardness from 20 to 65 HRC.
However, increases in carbon content to approximately 1% do not produce a corresponding
increase in hardness above 65 HRC.
Hardness and mechanical properties are
related not only to carbon content but also the
composition of carburized steels, which is illustrated in Fig. 44. This figure shows the corresponding carbon gradients of the cases of three
carburized steels: chromium-molybdenum, carbon, and nickel (Ref 25). The data in Fig. 44
show that the presence of chrome and molybdenum increases the case carbon content, and
nickel decreases the case carbon content. The
case carbon content is increased due to the presence of carbide-forming elements; their structures in the carburized case influence the
mechanical properties of the steel. Overcarburization may also lead to quench cracking
(Ref 91).
The influence of steel composition on the
microstructure of a carburized and quenched
20H steel is shown in Fig. 45. Figure 45(a) shows
a martensite-retained austenite microstructure
with some carbides dispersed throughout the
Fig. 42
structure. Fig. 45(b) shows a martensiticretained austenite structure with network carbides. This structure is due to the chrome content
in the steel and leads to varying carbon content
during carburizing. Additional examples of the
case microstructure of SAE 8620 steel and core
microstructures of SAE 1524 and 8115 steel are
shown in Fig. 46 to 48 respectively.
Decarburization is the opposite of carburizing. While carburization is performed to
increase carbon content in the surface of steel,
decarburization is the process by which carbon
is lost from the surface of steel. Decarburization
can lead to catastrophic failures of components
(Ref 92) and must be minimized because of
fatigue failure such as bending and contact
fatigue (Ref 93). Figure 49 illustrates decarburization of a poorly carburized SAE 8620 steel,
and Fig. 50 shows the microstructure at higher
magnification.
Decarburization occurs at temperatures in
excess of 700 C in the presence of gases that act
as decarburization agents, which include carbon
dioxide (CO2), water vapor (H2O), hydrogen
(H2), and oxygen (O2). The decarburization
process involves the following chemical
reactions of molecules with carbon in the steel
surface (CFe) until there is an equilibrium
Micrographs illustrating transition zones between the carburized case and the uncarburized core. (a) Illustrates a gradual
transition. (b) Illustrates a rapid transition between the case and the core microstructures. Original magnification: 500
pg 213
18/8/2008 3:20PM Plate # 0
Fig. 43
Fig. 44
Martensitic hardness as a function of carbon content in carbon and alloy steel. Source: Ref 80
Dependence of the carbon gradient as a function

of case depth for three carburized steels that were
carburized under the same conditions: 925 C and 10 h. 1,
chromium-molybdenum steel (0.56% Cr, 0.16% Mo); 2, carbon
steel; 3, nickel steel (3.5% Ni)
Fig. 45
Microstructures of the carburized case structure

of two different samples of 20H steel that were
carburized in the same load. Carburizing temperature:

930 C for 7 h; hardening temperature: 860 C for 0.5 h. (a)
Surface. (b) Core. Etchant: 3% HNO3. Original magnification:
500
pg 214
18/8/2008 3:21PM Plate # 0
Fig. 46
Microstructure of SAE 8620 case of a mold taken just

below the surface. Etchant: alkaline sodium picrate
boiling (60 s), area just below the surface. Original magnification: 500. Courtesy of G. Vander Voort, Buehler Ltd., Lake
Bluff, IL
Fig. 47
Illustration of lath martensite in the core of carburized SAE 1524 steel; water quenched from 925 C
(1700 F). Etchant: nital. Scale = 10 mm. Courtesy of G. Vander
Voort, Buehler Ltd., Lake Bluff, IL
Fig. 49
Micrograph of poorly carburized SAE 8620 mold

showing decarburization at the surface (note patches
of ferrite and pearlite). Below this zone is where the grainboundary carbides are seen. Original magnification: 500 .
Courtesy of G. Vander Voort, Buehler Ltd., Lake Bluff, IL
Fig. 50
This is a higher magnification of the decarburized

microstructure shown in Fig. 49 of the surface of a
poorly carburized SAE 8620 mold (note patches of ferrite and
pearlite). Original magnification: 1000 . Courtesy of G. Vander
established between the gaseous atmosphere and

the steel surface:
CFe
CFe +H2 O
CFe +2H2
Fig. 48
Illustration of lath martensite in the core of carburized SAE 8115 steel; water quenched from 925 C
(1700
Etchant: nital. Scale = 10 mm. Courtesy of G. Vander
F).
2CO
CO+H2
CH4
When the reactions proceed from left to right,

decarburization will occur. These are the reverse
of the carburization process.
Some of the more commonly reported causes
of decarburization include a malfunctioning
endogas generator, such as soot accumulation
hindering catalyst activity; excess moisture in
pg 215
18/8/2008 3:21PM Plate # 0
the furnace atmosphere; air contamination and

leakage (Fig. 51); heating in aged (deoxidized)
salt baths; and improper carbon potential selection (Ref 94).
If a carburized steel is at a temperature above
the Ac3 (approximately 900 C) in a decarburizing atmosphere, the carbon potential will be
low and the surface carbon content will also be
low, since the carbon in the steel surface and the
related gaseous reactions shown previously will
be driven to this equilibrium condition. This will
result in a decarburized layer being produced,
and the depth of the decarburized layer will
depend on residence time in the furnace under
these conditions.
If the temperature of the steel is below Ac3
and above Ac1 (800 to 840 C), there is a different decarburization condition. In this case,
the carbon content rapidly decreases from A
to B, as shown in Fig. 52 (Ref 94). Further
decreases in carbon content will result in a
material of carbon content C in equilibrium
with material of carbon content B. Therefore,
further loss of carbon by decarburization must
result in the formation of ferrite containing
carbon content C.
If the atmosphere carbon content is controlled
to carbon potential D, then ferrite cannot form.
Instead, a gradient is formed between carbon
contents A and D.
Fig. 51
Effect of air ingression into the carburization atmosphere (N2/4% natural gas) on the decarburization of
SAE 8620 after 2 h at 850 C. Source: Ref 94
Decarburization is typically classified as total

or partial. Figure 53(a) illustrates a case of total
decarburization of 1018 steel and is characterized by a ferritic surface layer (Ref 94).
Usually, there is a gradient from total to partial
decarburization with increasing depth from the
surface. Partial decarburization is illustrated in
Fig. 53(b) and is often characterized by grainboundary ferrite at the surface (Ref 94). In this
case, nital etching will reveal a structure more
gray in color than would be achieved with a
higher-carbon martensite. The quenched surface
of partially decarburized steel is typically bainitic or martensitic. The final microstructure
produced by a specific level of decarburization is
dependent on the steel alloy and cooling rate.
Decarburization is accompanied by surface
hardness reduction. However, partial or shallow
decarburization may not necessarily be detected
by macrohardness determination. If decarburization is suspected, microhardness determinations, in view of their sensitivity to the presence
of such microstructures, should be used.
Decarburization can exhibit dramatic effects
on the residual stress of a component, as illustrated by Fig. 54 (Ref 94). In this example,
the surface carbon content of a carburized
3.5Ni-1.5Cr steel would be approximately 1%,
and the surface residual stresses would be
compressive at 4392 MPa. After decarburization reduces the carbon content to 0.64%,
the surface residual stresses were found to be
nearly 0. Finally, when decarburization reduced the surface carbon content to 0.35%,
Fig. 52
Iron-carbon equilibrium diagram to explain decarburization
pg 216
18/8/2008 3:21PM Plate # 0
the surface residual stresses were tensile at

226 MPa.
Various reports have shown that decarburization can result in large decreases in bending
and contact fatigue strength (Ref 94). Since
wear-resistance properties are typically dependent on achieving optimal hardness, reduction of
carbon content and thus surface hardness will
have a correspondingly adverse effect.
Fig. 53
Micrographs illustrating total and partial decarburization. (a) Total decarburization of 1018 steel
caused by a furnace air leak. Etchant: 1% nital. Original magnification: 500. (b) Illustration of partial decarburization. Original
magnification: 190
Influence of Grain Size

Grain size is one of the most important and
characteristic features of steel. Grain size influences mechanical and plastic properties,
especially impact resistance and also steel hardenability. Grain size is characterized by the size
of the austenite grain, and it is dependent on
various factors, such as degree of cooling and the
deoxidation process during steelmaking (Ref
95). Generally, an ASTM grain size of 6 to 8 is
specified. Figure 55 illustrates grain sizes of
ASTM No. 6 to 9.
Grain growth is affected by temperature.
Typically, grain growth increases with temperature and time at temperature. Aluminum
may be added to steel to provide resistance to
grain growth (grain refined). Alloying elements
such as nickel and molybdenum also provide
greater resistance to grain coarsening at typical
carburizing temperatures than plain carbon
steels. Although grain coarsening is usually not a
problem for carburizing temperatures up to
925 C, carburizing at temperatures greater than
1000 C is typically accompanied by some grain
coarsening, yielding a mixed-grain structure.
Reheating at 820 to 860 C can be performed to
refine the mixed-grain structure.
Fine-grained steels are less hardenable than
coarse-grained steels with the same composition, and this generalization is true for case
structure also. The slower the cooling of the steel
during steelmaking, the larger the grain, since
there are fewer nucleation sites formed. Steels
Fig. 54
Effect of decarburization on the residual stresses of

carburized and hardened 3.5Ni-1.5Cr steel. The
carbon content at 0.002 mm was approximately 1% for curve 1,
0.64% for curve 2, and 0.35% for curve 3.
pg 217
18/8/2008 3:21PM Plate # 0
are typically deoxidized by silicon and manganese prior to aluminum addition. This will
inhibit the undesirable formation of AlN or
Al2O3 particles, which provide nucleation sites
for coarse-grained carbides that lead to the
Fig. 55
Comparison of nominal ASTM No. 6 to 9 grain

sizes. Etchant: nital. Original magnification: 100 .
Source: Ref 96
formation of coarse ferrite/carbide (pearlitic)

grain structures on cooling.
Grain size is possible by control of the steel
composition during the steelmaking process.
Subsequent to this, control is by proper heat
treatment. Heating the steel to the upper critical
temperature, Ac1, will typically produce an
average minimum grain size. Heating to higher
temperatures will increase the grain size. Also,
quenching from the Ac1 temperature will produce fine grain size, and quenching from a
higher temperature would yield a coarser grain
size.
The austenite grain size at the onset of transformation during the quenching process will
influence the martensite platelet size and thus
will affect microcracking potential, the amount
of retained austenite formed, and the frequency
and depth of internal oxidation (Ref 95).
Coarse grain structure is observed on etched
microsections, most often in the form of coarse
grain structure of martensite relative to finer
retained austenite structure. Of the various factors affecting grain size, the primary factor is
furnace treatment. Although coarse-grained
steels exhibit better machinability, they generally possess lower toughness and ductility and
exhibit a greater tendency for distortion and
cracking than fine-grained steels. Coarsegrained steel also exhibits a more limited range
of thermal treatment temperatures, and they
possess better hardenability with higher asquenched hardness. Additionally, coarsegrained steels typically possess lower impact
resistance and a lower yield point.
Carburized steel case structures with a
coarse-grained structure, along with significant
amounts of overcarburizing, are characterized
by decreased mechanical properties. In this case,
the grain-boundary structure contains a continuous network carbide structure that is difficult
to remove, and within the grain structure there
are typically acicular carbides. Such microstructures are fatigue sensitive, with cracking
throughout the brittle network carbide structure.
The presence of these defective microstructures
can be prevented by controlling the carburizing
temperature and carbon potential or by using a
steel with characteristically fine-grained structure.
An example of grain size formation during
austenitization of steel 40 is shown in Fig. 56.
The irregular austenitic grain boundaries result
from the short heating time and also are related
to a prior normalizing and annealing process.
pg 218
18/8/2008 3:21PM Plate # 0
This thermal history produces the variable grain

size shown.
Isogai et al. reported that grain sizes of
approximately 8 were required for carburized
transmission gear steel to achieve the required
high fatigue strength, pitting strength, and
impact strength in severe-use environments (Ref
97). To assure a fine grain size, after the
carburizing process, the chromium-containing
steel, such as SCr420, was reheated to 820 to
870 C for 20 to 60 min in an atmosphere containing the carburizing gas (carbon potential of
0.75%), quenched in oil until the steel was
120 C, and then tempered at 120 to 200 C.
Treatment of the carburized steel in this way
permits substantial reductions in grain size and
corresponding improvements in fatigue strength
and impact strength. Similar improvements were
achievable with carbonitrided steel (Ref 97).
temperature, as well as the steel alloy chemistry.

The total depth of internal oxidation can be
calculated from (Ref 100):
Xi 2 =(2D0 C0 )=(n CM ) t=kp t
Xi =kp t0:5
The depth of internal oxidation that originates

the carburizing atmosphere may vary from 1 to
30 mm (Ref 93, 98, 99). Internal oxidation consists of a continuous oxide layer on the surface,
on the order of 0.01 mm, due to oxygen reaction
with the carburized steel surface. In this region,
the oxygen content can be 10 to 20 times that of
the core. During the gas carburizing process,
oxides will not only form on the surface but also
penetrate into the steel surface. Since internal
oxide formation is a diffusion process, the depth
and extent of oxide penetration is dependent on
the square root of the total carburizing time and
where Xi is the depth of oxygen penetration, D0

is the diffusion coefficient of oxygen in the alloy,
C0 is the oxygen concentration at the steel alloy
surface, CM is the concentration of the base
metal in the alloy (e.g., silicon, chromium,
manganese, titanium, vanadium), and n is the
stoichiometric factor.
Internal oxidation appears on the polished
metallographic specimen in the form of very
small inclusions concentrated in an austenite
grain or within the grain boundary. The probability of oxidation within the grain increases as
the grain size decreases (Ref 101). Figure 57
shows grain-boundary oxidation of carburized
20MnCr5 steel. Figure 58 also shows grainboundary oxidation, but it is accompanied by
nonmartensitic transformation products. Internal oxidation occurs in two zones: an inner zone
and an outer zone. Oxides of chromium-manganese are typically formed in the outer zone,
both within the grain and in the grain boundaries.
In the inner zone, silicon-rich oxides are typically formed exclusively within the grain
boundaries (Ref 102). Figure 59 also illustrates
intergranular oxidation of a gas-carburized steel
(Ref 79).
Typically, the greater the case depth, the
greater the degree of oxide formation at the
Fig. 56
Fig. 57
Internal Oxidation (Ref 77)
Prior-austenite grains formed in hardened steel 40,

which were due to abnormal growth during the
austenitizing process, Etched: S. Bechet and L. Beaujurda. Original magnification: 500
Illustration of grain-boundary oxidation of carburized 20MnCr5 to a depth of 30 mm. Unetched. Original magnification: 200. Courtesy of Fluidtherm Technology P.
Ltd., Ambattur, India
pg 219
18/8/2008 3:21PM Plate # 0
surface and the greater the depth of internal

oxidation. The thickness of the internal oxidation zone typically is approximately 5% of the
carburized layer thickness and, on occasion,
may be as high as 10%. The critical cooling rate
is greater in the internal oxidation zone. As a
result, an otherwise normal hardening produces
a greater amount of bainite, which will lead to
Fig. 58
Illustration of grain-boundary oxidation with nonmartensitic transformation products to a depth of

approximately 30 mm. Etchant: nital. Original magnification:
200. Courtesy of Fluidtherm Technology P. Ltd., Ambattur,
India
Fig. 59
lower surface hardness and poorer abrasive wear

resistance.
Typically, the oxides in the outer surface
region are globular in form, while intergranular
oxides were formed further from the surface. In
one analyis of a carburized chromium-manganese steel, larger globular oxides were formed in
the region closer to the surface (1.9 mm depth)
and intergranular oxides in the region farther
from the surface (2.49 mm) when the steel was
heated for 16.6 h (diffusion at 2 h at 800 C,
followed by 3 h at 930 C and a boost cycle at
930 C for additional heating times, in this case
11.6 h) (Ref 103). For this work, glow discharge
optical emission spectroscopy, in which a sputter erosion process using ionized argon gas with
a voltage of 600 V and 25 mA was used to
quantify the degree of oxidation and elemental
distribution, was performed using energy-dispersive x-ray analysis. When the carburized
samples were subjected to shorter heating times,
intergranular oxides formed relatively farther
from the surface.
Transmission electron microscopy was used
by An et al. (Ref 103) to identify oxide type
and morphology. Chromium and manganese
globular oxides formed nearer the surface after a
total heat carburizing cycle of 5.8 h (5+0.8 h).
An agglomerated internal oxide of chromium
Intergranular oxidation of the surface along prior grain boundaries in a carburized steel. Original magnification: 1000 .
Source: Ref 78
pg 220
18/8/2008 3:21PM Plate # 0
and manganese oxides in the core surrounded by

silicon oxides formed after heating for 16.6 h
(5+11.6 h) (Ref 103).
Internal oxide formation within grain boundaries provides sites for crack initiation (Ref 10).
This was shown in a study by Laue et al.
(Ref 104), who evaluated the fatigue behavior of
case-hardened SAE 5115 steel with internal
oxidation of the case structure. Fatigue studies
were conducted on this steel, and it was shown
that fatigue crack initiation occurred along the
oxidized grain boundaries of the steel (Ref 104).
Internal oxides form as a result of oxygen
diffusion into the surface, with subsequent formation of metal oxides at carburizing temperatures. The formation of these oxides is enhanced
by the presence of metals, chromium and manganese and Cr-Mn-Ti, which possess a greater
affinity for oxygen than iron. The susceptibility
for internal oxidation increases with increasing
concentration of these oxide-forming elements.
Lohrmann et al., referring to earlier work by
Kozlovskii and co-workers, reported that the
depth of internal oxidation was dependent on the
total oxidation potential (TOP) of the steel alloy
(Ref 105, 106):
TOP=4:87 Si+3:7 Mn+1:47 Cr 3:24 Ni
1:82 Mo
Fig. 60
where the elemental compositions are given in

weight percent. Figure 60 provides a correlation
of the TOP and depth of internal oxidation (Ref
106).
In addition, Kozlovskii et al. reported that
most steels, when subjected to the gas carburizing process, will undergo internal oxidation
with a corresponding surface formation of
troostite to a depth of 0.01 to 0.03 mm. If
troostite is formed at a depth greater than
0.014 mm, there is a substantial decrease in
fatigue strength. However, the potential for
internal oxide formation can be reduced by the
addition of 5 to 10% of ammonia to the furnace
for 10 min before the carburizing process is
completed or by using steels containing 0.5%
Mo and not more than 0.5% Cr (Ref 106).
Lohrmann showed that the form and type of
internal oxide obtained was dependent on the
alloy composition of the steel. For example,
depending on the steel alloy, spot-, liner-, or
lattice-type internal oxides could be formed
(Ref 105).
Kehr and Seese examined the effect of internal
oxidation during carburizing of investment-cast
ingot-iron test specimen steels containing various amounts of chromium (0.20, 0.45, 0.91,
1.85, and 4.74% Cr) (Ref 107). It was shown that
steels containing approximately 0.50% Cr, for
Correlation of calculated total oxidation potential (TOP) and average depth of internal oxidation
pg 221
18/8/2008 3:21PM Plate # 0
example, SAE 8620, were more susceptible to

internal oxidation than either plain carbon steel
or steels with greater amounts of chromium, such
as AISI 5120 (0.70 to 0.90% Cr). Figure 61
shows that the depth of oxide formation increases
as the chromium content increases up to 0.45%
(Ref 107).
The presence of internal oxidation on carburized steels greatly reduces bending and
contact fatigue strength and wear resistance
(Ref 93). However, surface oxides that are
formed may be removed by grinding or shot
peening. The potential for internal oxidation can
be reduced by heating steel to the carburizing
temperature under a nitrogen/hydrogen gas
mixture. During carburizing, these gases are
replaced by a carburizing atmosphere whose
oxygen activity is less than that required for the
formation of manganese II oxide or chromium
III oxide, and, in some cases, in the presence of
ammonia (Ref 100).
Finally, internal oxidation (and decarburization) is known to lead to variations in the surface
compressive stresses in a carburized component
(Ref 108). There is one report of the presence of
internal oxidation leading to undesirable surface
tensile stresses that then led to subsequent
cracking of a carburized idler gear when used in
a diesel engine gearbox.
Carbides and Carbide Structure

Carbides formed during carburization are
treated as undesirable products to be avoided.
Fig. 61
Effect of chromium content of steel on the depth of

oxidation
There are three types of carbides to be discussed

here: globular (or massive) carbides, network
carbides, and surface-film or flake carbides.
Carbides in steel are hard and brittle ceramiclike interstitials with a high compressive
strength but low tensile strength (approximately
35 MPa, or 5000 psi).
Carbides in steel basically form when carbon
levels exceed the solubility limits of carbon in
the iron crystal structure. The allotropic nature
of iron also has different phase structures (i.e.,
crystal) with different solubility limits for carbon. For example, the maximum solubility of
carbon in the body-centered cubic (bcc) structure of ferrite is approximately 0.025 wt% at
723 C on the iron-carbon phase diagram (see
Appendix 9). For the face-centered cubic phase
of austenite (c), the maximum solubility limit of
carbon in c is approximately 2.06 wt% C at
1147 C. At still higher temperatures, another
type of bcc solid phase is d-ferrite. The maximum solubility of carbon in d-ferrite is
0.09 wt% C at 1493 C. A peritectic also occurs
at 0.16% C at 1493 C.
The iron-carbon system has eutectic transformation at 1147 C during soldification, with
steel carbon levels of 2.06 to 6.67 wt% C. The
eutectic carbon concentration is 4.3%. Solidstate transformations in steel include the wellknown eutectoid transformation at 733 C, with
a carbon concentration of 0.83%. At 733 C,
austenite transforms to pearlite. Pearlite is a
eutectoid mixture containing 0.83% C and is
characterized by a fine ferrite-cementite structure that forms upon austenite decomposition
during slow cooling.
The upper critical temperature (A3) is the
temperature below which ferrite starts to form
due to ejection from austenite in the hypoeutectoid alloys. The lower critical temperature
(A1) is the temperature where the austenite-topearlite eutectoid transformation occurs. Austenite does not exist below this temperature. A2
is the temperature below which ferrite is ferromagnetic.
Cementite (Fe3C) is the principal carbide of
iron and carbon, with an orthrhombic crystal
structure. Cementite is harder and more brittle
than ferrite. Steel hardness increases with increasing cementite content. Increasing carbon
content increases the amount of cementite but
does not affect the amount of ferrite present,
since ferrite is saturated with 0.22% C. If there is
less than 0.83% C, iron and carbon will combine
to form Fe3C until no carbon remains. The
pg 222
18/8/2008 3:21PM Plate # 0
cementite formed in this manner will combine

with the required amount of ferrite to form
pearlite, and any remaining ferrite will be in the
structure as free ferrite (proeutectoid ferrite).
Pearlite will form if the carbon content in the
austenite is 40.83%, and excess carbon will
form cementite. Excess cementite (proeutectoid
cementite) will deposit in the grain boundaries.
Because cementite (Fe3C) contains a specific
amount of carbon and iron, pearlite also contains
a specific amount of cementite and ferrite.
As a phase in steel, the chemical composition
of cementite will contain carbides of other carbide-forming elements, such as chromium and
manganese Alloying elements of chromium,
manganese, nickel, and other elements are, of
course, commonly used in alloy steels for
property improvement. They also impact the
properties of ferrite and cementite, because they
partition differently in the phases. For example,
chromium and manganese partition in cementite
instead of ferrite. However, nickel and silicon
tend to favor partitioning in ferrite. Chromium,
manganese, molybdenum, and titanium are thus
cementite stabilizers in steels, while nickel and
silicon are ferrite stabilizers. Interestingly, while
chromium, manganese, molybdenum, and vanadium show no negative effect on cementite
formation, titanium, nickel, and silicon exhibit a
negative effect on cementite formation (Ref
109).
Cementite develops different morphologies
and distributions depending on the process of
cementite formation. Figure 62 illustrates three
microstructural forms of cementite: lamellar,
mixed, and granular (Ref 110). Cementite may
also be classified as reticular, acicular, or granular (Ref 110). Reticular cementite, also known
as shell-type cementite, possesses a cracksensitive network or platelet structure. Acicular
cementite, or needlelike structure, refers to a
Fig. 62
lamellar structure of cementite in ferrite, shown

in Fig. 62(b). Finally, cementite exhibits a
granular or grainy appearance, as shown in
Fig. 62(c).
Cementite forms during soldification from a
liquid or during solid-state transformations.
When cementite originates by crystallization
from a liquid melt, it is referred to as primary cementite (Fe3CI). Secondary cementite
(Fe3CII) is formed from austenite by hypereutectoid alloys (carbon40.8%). Tertiary cementite (Fe3CIII) is formed at temperatures below
723 C by precipitation in the grain boundaries
(which become richer in carbon with the
decreasing carbon content in a-iron).
Globular Carbides. Slowly heating a steel
to the carburizing temperature in the presence of
a carburizing atmosphere through the ferrite-toaustenite temperature transformation region will
lead to unconnected globular carbide formation
either within the ferrite grains or at the former
ferrite grain boundaries, as shown in Fig. 63.
This process is favored by high carbon potentials
and also by reduction of normal heating rates
typically involved during carburizing, by
excessive furnace loading, or by a furnace malfunction. Globular carbide formation may also
be enhanced by austenitic nuclei or by localized
concentrations of carbide-forming elements.
The problem of globular carbides may also
coexist with other problems, such as retaining
austenite or quench cracking.
When carburizing steels (0.15 to 0.25% C),
which are commonly ferritic with localized
areas of spheroidal carbides due to prior normalizing and subcritical annealing, are heated
through the Ac1 temperature, the high-carbon
regions begin to transform to austenite, resulting
in the formation of localized regions of carbon
and carbide-forming elements in addition to
undissolved carbides in the presence of the
Cementite structures of CT60 steel with (a) lamellar, (b) mixed, and (c) granular cementite. Original magnification: 500
pg 223
18/8/2008 3:21PM Plate # 0
carbon-rich carburizing atmosphere. As the

temperature increases further to Ac3, additional
ferrite will transform to austenite by combining
with carbon from the furnace atmosphere, preferentially over carbon transfer from the spheroidal carbides already present until the austenitic
transformation is complete, at which point there
will be an equilibrium between the austenite and
carbides (Ref 112).
Those steels containing alloying elements
such as chromium and manganese that will
reduce the eutectoid carbon content and Ac1
temperature are more likely to develop globular
carbides. These elements will increase the
solubility limit of carbon (Acm), thus shifting the
equilibrium diagram to the left, as shown in
Fig. 64 (Ref 112).
It should be noted that although the terms
spheroidal carbides and globular carbides are
often used interchangeably, these carbides may,
in fact, possess a round, angular, or even needlelike appearance. Also, although the structures
are typically designated as M3C, their actual
ratio of the element/carbon composing the carbide being observed is not only dependent on the
time and temperature of heat treatment but primarily on the elemental availability in the steel
during formation.
There is a three-step process of high-density
carburizing that is conducted to increase case
hardness by aggressive precipitation of cementite (Fe3C) to improve surface fatigue strength
(Ref 113). Generally, grain-boundary carbide
precipitation will form network carbide structures that are susceptible to quench cracking and
Fig. 63
Bad globular carbide formation in the case of a carburized 9310 steel. Etchant: boil in alkaline sodium
picrate solution (45 s). Scale = 10 mm. Courtesy of G. Vander
result in reduced fatigue strength. Interestingly,

it is possible to obtain desirable surface fatigue
strength by vacuum or plasma carburizing a
chromium-containing steel such as SCr420H
or a chromium-molybdenum steel such as
SCM420H (Ref 114). This process involves the
formation of a M23C6-type globular carbide that
is approximately 1 mm in diameter. These
microstructures are reported to exhibit excellent
surface fatigue strength and rolling fatigue
strength under high bearing loads (43 GPa) at
relatively high temperature (100 to 300 C),
which are unachievable in the presence of larger
(i3 mm) M3C carbides.
Reheating a carburized steel to a temperature
below Acm causes spheroidized carbide particles
to form. (Both austenite and cementite are stable
at this temperature.) Since these carbides bind
some of the carbon in the case, there will be an
increase in the Ms temperature. In addition,
grain-boundary migration is reduced (Ref 115).
Carbides are typically very hard. For example, the microhardness of globular carbides in a
carburized case of a plain carbon steel has been
measured to be 41000 HV, and carbides in a 2%
Ni-Cr steel have been measured to be approximately 800 HV (Ref 112). Therefore, the presence of free globular carbides is often assumed
to improve component wear, abrasion, and
scuffing resistance. However, in a study reported
by Parrish on the effect of globular carbides on
contact fatigue with carburized 2% Cr-Mn steel,
it was shown that the presence of massive
Fig. 64
The saturation surface carbon content (Acm) of various carburizing steels as a function of carburizing
temperature as related to the iron-carbon equilibrium diagram
pg 224
18/8/2008 3:21PM Plate # 0
carbides was detrimental, while a beneficial

effect was obtained if the carbides were finer and
better distributed, as shown in Fig. 65 (Ref 112).
The presence of globular carbides may also
lead to grinding cracks, which are also related to
surface residual stresses.
Network Carbides. Globular carbides are
most typically formed when the carbon content
is less than the Acm. However, under some
conditions, it is possible for globular carbides to
still be formed when the carbon content is
greater than the solubility limit (in excess of the
eutectoid composition) and austenite is supersaturated with respect to carbon. Most typically,
these are conditions where carbon will precipitate in the grain boundaries as cementite
(Fe3C) during slow cooling from the carburizing
temperature, leading to the formation of network
carbides. Figure 66 illustrates network carbide
structure observed in a broken carburized AISI
P5 steel tool.
The equilibrium diagram shown in Fig. 67
illustrates the conditions for the formation of
network carbides when excess carbon is precipitated from austenite as Fe3C (Ref 112).
When the carbon steel supersaturated with carbon content of C1 is cooled from t0 to t1, Fe3C
will begin to precipitate. As cooling continues
to t2, additional carbon will have precipitated
as Fe3C until the carbon content is C2. At the
Fig. 65
eutectoid temperature t3, carbon precipitation as

Fe3C will have stopped, and the austenite carbon
content C3 will transform to a pearlitic eutectoid
microstructure. The relative proportion of Fe3C
to austenite can be determined from Fig. 67
using the lever rule (CxC1/C1B).
Although austenite is supersaturated with
respect to carbon during the carburizing process,
and carbide precipitation at austenitic grain
boundaries will occur during cooling, if the steel
is quenched from the carburizing temperature,
the excess carbon can be retained by the resulting as-quenched martensitic/retained-austenite
microstructure. Typically, during carburizing,
the load is cooled in the furnace from the carburizing temperature to the temperature from
which the steel will be hardened. Since the Fe3C
migrates to the grain boundaries during slow
cooling, faster cooling by quenching will allow
retention of the carbon in the martensitic/austenitic structure. Thus, cooling rate can be used to
control the amount of network carbide formation. Alternatively, carburized steel can be
cooled to ambient temperature and reheated to
820 to 860 C and quenched. However, it has
been shown that traces of the network carbides
remained in the microstructure even after heating to 900 C (Ref 112). Parrish also has
reported that if steel containing excess carbon is
subcritically annealed prior to quenching, a fine
Contact fatigue strength of carburized 25Kh2GHTA steel (tempered at 180 to 200 C)
pg 225
18/8/2008 3:21PM Plate # 0
Fig. 66
Micrographs of a broken carburized P5 tool steel die.

Note the layer of cementite along the surface and the
heavy grain-boundary network in (a). The case was 59.5 HRC,
and the core was 22 GRC. (a) The case is shown at an original
magnification of 100 (100 mm bar). (b) The case at an original
magnification of 1000 (10 mm bar). (c) The core at an original magnification of 400 (25 mm bar) in Nomarski differential
interference contrast (note the old and new ferrite, as in dualphase steel). The austenitizing temperature for the case is
approximately 1475 F, which is in the two-phase field for the
core. New ferrite formed due to limited hardenability in the
quench. Courtesy of G. Vander Voort, Buehler Ltd., Lake Bluff, IL
dispersion of relatively unharmful spheroidized

carbides will be obtained instead of network
carbides (Ref 112).
Karpov studied the effect of cooling from the
carburizing temperature on the nature of network carbides formed when quenching gascarburized 11 by 11 by 56 mm 07Kh16N6 steel
test specimens. The test specimens were carburized, cooled, then reheated to austenitize to
1020 C, and then cooled to the quenching
temperature at 0.036 C/s, which is less than the
critical quenching rate that leads to network
carbide segregation. Quenching was conducted
in water at room temperature. From Table 11, it
is evident that network carbide formation begins
at 900 to 850 C for this alloy and is completed
at 600 C. From the phase diagram for this alloy,
the Acm temperature is 860 C (Ref 116). During the course of this work, Karpov found that a
nondestructive electromagnetic flaw detector
could be used to rank the network carbide size
(Ref 116).
Network carbides have been reported to
reduce surface fatigue (pitting) resistance of
carburized steel used for bearing applications
(Ref 114). Like globular carbides, complex
carburized networks in an overcarburized case
will reduce the potential for carbide redissolution during reheating, which will lead to increased brittleness and grinding cracks (Ref 1).
The presence of network carbides in a Kaplan
turbine blade constructed from improperly carburized 17CrNiMo6 steel was reported to be a
major contributor to failure by an intergranular
microcracking mechanism (Ref 117). The
cracks seemed to follow the path of the network
carbide structure. Parrish has summarized various studies and concluded that continuous network carbides do reduce fatigue properties,
leading to premature cracking failures by a
stiffening mechanism (Fig. 68) However, other
studies with partial nertwork carbides showed
no deleterious effects (Ref 112). The presence of
network carbides also is not expected to produce
an adverse effect with respect to wear under
heavily loaded conditions or scuffing (Ref 112).
The three most common failure modes of
carburized steels are ductile fracture, cleavage,
and intergranular fracture (Ref 23). Ductile
fracture is caused by nucleation growth and
coalescence of voids that are initiated at inclusion sites and second-phase particles. Cleavage
fracture occurs by separation at crystallographic
planes by a transgranular pathway. Intergranular
fracture, such as that involving cementite
pg 226
18/8/2008 3:21PM Plate # 0
pg 227
Fig. 67
The use of the iron-carbon equilibrium diagram to illustrate network carbide formation
Table 11 Effect of initial cool-down period on

network carbide formation of carburized
07Kh16N6 steel
Cool-down temperature(a), C
1020 (initial temperature)

950
900
850
800
700
600
Network carbide
severity rating(b)
1
1
12
13
13
34
55
(a) The cooldown occurs from 1020 to the temperature shown at a rate of
0.036 C/s. (b) 1, very fine carbide network; 34, failure rating; 5, largest
deposition at grain boundaries, involves cracking on grain boundaries and is due to (Ref 23):

Precipitation of a brittle phase (such as network carbides) on the grain boundary

Hydrogen embrittlement
Environmental-assisted cracking
Intergranular corrosion
Grain-boundary cavitation
temperature cracking
and
high-
Film or Flake Carbides. Surface-film or

flake carbides are composed of a continuous or
discontinuous carbide film with typically little or
no penetration into the case structure, which is
caused by cooling of the carburized steel in the
furnace with high carbon potential. Parrish
summarized various previously published reports that stated such carbide films contain
approximately 19% Fe3C, 16% austenite, with
the balance being martensite. These films cover
a nonmartensitic layer of approximately 30 mm
thickness, with carbides penetrating into the
grain boundaries (Ref 112). Koistinen showed
that such films have high tensile surface
(0.025 mm depth) residual stress, as shown in
Fig. 69 (Ref 44). Until now, no further information relating to the presence of these carbide
types on carburized steel properties has been
reported.
18/8/2008 3:21PM Plate # 0
Noncarbide Inclusions
Two of the primary causes of fatigue failure
are inclusions and surface defects. Inclusions
may be metallic impurities or metallic oxides
(Ref 118). Metallic element inclusions (impurities), although typically in trace quantities,
may be traced to the scrap used in the steelmaking process. These elements cause intergranular segregation, which may lead to crack
Fig. 68
Comparison of bending fatigue of carburized 12Khn3 gears showing adverse effect of network carbides
Fig. 69
Residual-stress distribution of carburized SAE 1018 steel with a film-carbide layer formed due to a high carburizing potential.
The surface layer consisted of 16% Fe3C, 16% retained austenite, and the balance was as-quenched martensite.
pg 228
18/8/2008 3:21PM Plate # 0
formation, detrimental precipitate formation,

and are often observed as slivers in the final
product.
Metallic oxide inclusions vary in morphology
and composition. Some sources of these oxides
include:
Deoxidation products, such as alumina

inclusions, that are formed by the reaction of
dissolved oxygen and the added deoxidant,
such as aluminum. Alumina inclusions are
dendritic when formed in the presence of
high oxygen concentration.
Reoxidation products are generated when
the aluminum remaining in the liquid steel is
oxidized by FeO, MnO, or SiO2 and other
oxides in the slag or refractory materials or
by exposure to the atmosphere.
Slag entrapment occurs in metallurgical
fluxes entrained in the steel. This occurs
during transfer between steelmaking vessels.
These inclusions are typically spherical.
Exogenous inclusions from other sources
include dirt, broken refractory brickwork,
and ceramic lining materials. They act as
sites for heterogeneous nucleation of alumina and may include a central particle.
Chemical reactions may produce oxides
from inclusion modification when calcium
treatment is imperfectly performed. Inclusions containing CaO may also originate
from entrained slag.
All steels contain various noncarbide inclusions. It is well known that inclusions, in
addition to surface defects and inhomogenities
such as retained austenite, nonmetallic inclusions, and inhomogeneities, can reduce fatigue
Fig. 70
strength and negatively influence ductility and

toughness (Ref 119). Although harder inclusions
such as metallic inclusions are more harmful
than softer inclusions, nonmetallic inclusions
are still generally harmful. In addition, the deleterious effects of inclusions increase with size.
In one study conducted by Bomas and
Schleicher on the effect of inclusions on the
fatigue strength of carburized 16MnCrS5 (SAE
5115) steel, it was found that subsurface fatigue
crack initiation was initiated by nonmetallic
inclusions up to depths of 1.4 mm (Ref 120).
Similar results were obtained for a study of
MnS-induced bending fatigue failure of carburized EN39B steel (Ref 121).
Even in clean steels, oxide and sulfide inclusions exist. From a study of the effect of defects
such as inclusions on the fatigue bending
strength of carburized SCM20 steel, a model of
fatigue crack initiation during fatigue bending,
such as would occur in a gear tooth, was developed (Ref 122). This model is illustrated schematically in Fig. 70. Figure 70(a) illustrates the
chromium and manganese oxides and grainboundary segregation of silicon oxides present
in a decarburized surface layer. When the surface is loaded, the grain boundary or oxides act
as a stress raiser for microcrack formation, as
shown in Fig. 70(b). Although most of the cracks
remain in the decarburized layer, the most critical crack penetrates deeper into the steel, as
shown in Fig. 70(c). The crack that initiates fatigue failure was thought to be due to a combination of the microcracks shown in Fig. 70(b). In
this case, the decarburized layer in combination
with the presence of inclusions was projected to
be the source of fatigue failure (Ref 122).
Model of fatigue crack initiation due to the presence of inclusions in a nonmartensitic (decarburized) steel layer. Source:
Ref 122
pg 229
18/8/2008 3:22PM Plate # 0
Micropitting
Micropitting fatigue usually occurs on heavily-loaded surface-hardened components and
is characterized by a frosted or gray-stained
appearance under thin-film lubrication conditions (Ref 87, 123). Numerous small cracks in
the surface may exceed the depth of the micropits (Ref 124). Under magnification, small pits
approximately 10 mm deep will be observed.
The surface will appear etched with a pattern
that sometimes follows the slightly higher ridges
left by cutter marks or other surface irregularities on the finished component. Micropitting is
influenced by high surface loads, frictional heat
generation due to poor lubrication caused by
insufficient film thicknesses in the partial elastohydrodynamic lubrication regime, excessive
retained austenite, tangential speed, and lubricant additives and oxidation (Ref 123).
Micropitting is strongly influenced by the
relative lubricating oil film thickness and can
be quantitatively related to both the surface
condition (roughness) and the thickness of
the lubricating film by the lambda (L) value
(Ref 124, 125):
h
L=h=s= q
2
(s1 +s22 )
where s1 and s2 are the root mean square surface

roughness of the two opposing wear surfaces,
and h is the lubricating oil film thickness. When
Li3, there is full film lubrication with no
asperity contact. When 0.85L53, there is
partial elastohydrodynamic lubrication. When
L50.8, there is a boundary lubrication condition. When L51, micropitting will occur, and
once micropitting occurs, pitting fatigue (macropitting) will be accelerated (Ref 124). However, if macropitting does occur, it is often
characterized by an arrowhead or fan shape (Ref
123). There have been reports of substantial
improvements in fatigue lives, such as with
carburized 9310 steel, with corresponding reductions in surface roughness (Ref 125).
For gears, there is a critical temperature where
pitting fatigue and scuffing are likely to occur.
This is called the critical scuffing temperature
(Tc), which is calculated from:
Tc =Tb +Tf
where Tb is the equilibrium temperature of

the gears before meshing, and Tf is the flash
temperature, which is the instantaneous temperature rise due to localized friction heat at the
point where the gear teeth mesh. The value of Tb
is controlled by gear geometry design, and the
value of Tf is controlled by the lubricant viscosity and surface roughness.
To minimize micropitting:

Use higher operational speeds and smooth

material surfaces.
Use the recommended amount of clean, dry
lubricant with the highest viscosity permissible.
Reduce the lubricating oil temperature and
surface loading.
Use the optimal amount of case carbon
content in carburized gear materials.
Contact Fatigue Pitting (Macropitting)

Pitting failures occur when fatigue cracks are
initiated on the tooth surface or just below the
surface. Usually, fatigue pits are the result of
surface cracks caused by metal-to-metal contact
of asperities or defects due to insufficient lubricant film thickness. They are dependent on the
Hertzian contact surface stress and the number
of stress cycles (Ref 124). Surface asperities of
the harder material of a wear contact will lead to
damage of the softer surface, sometimes by a
work-hardening mechanism, leading to the
creation of microcracks that then become fatigue
pits as the wear process continues (Ref 124).
Pitting damage is commonly encountered with
rolling element bearings, gears, and machine
components subject to cyclic rolling-sliding
motion under a load. Initially, fatigue pits may
occur in localized areas and may range in size
from 0.38 to 0.76 mm (0.015 to 0.030 in.) in
diameter (Ref 85).
Vinokur et al. examined the effect of case
carbon content of carburized 18KhGNMFL steel
(1.3% Mn, 1.3% Cr, 0.8% Ni, 0.25% Mo, 0.1%
V) on the contact endurance of fatigue tests
conducted with a wear test with an applied stress
of 3500 MPa using a lubricating oil. Contact
fatigue was the average of 10 tests to determine
the number of cycles until pitting was observed.
The case depth was approximately 1.8 mm. The
carburized steel was hardened from the intercritical range and tempered at 170 C. The
results of this study are summarized in Fig. 71
(Ref 126). The contact endurance increases with
carbon content up to approximately 0.9% C and
pg 230
18/8/2008 3:22PM Plate # 0
then decreases. The optimal hardening temperature was at approximately 850 C, which is
in the intercritical temperature range, just below
the upper critical temperature for the steel.
Although fatigue pits are usually initiated at
the surface, subsurface initiation is relatively
common in case-hardened rolling element
bearings with serious inclusion problems. In
these situations, failures usually do not follow
the case-core interface (Ref 125, 127). The
potential for micropitting, pitting, and spalling
phenomena may be assessed from the lambda
(L) value (see the section Micropitting in this
chapter); however, even relatively smooth surfaces and lubricating film thickness, such as
those used for high-speed gears, may exhibit
pitting failures due to the presence of subsurface
cracks. The subsurface cracks may be caused by
the presence of inclusions that act as stress
concentrators, causing the crack to propagate
parallel to the surface and subsequently break
through the surface.
Spalling failure occurs when the wear process
causes several pits to join together. These larger
craters are usually caused by more severe overloading conditions. As the number of stress
cycles increases, the pitting process will continue in an effort to relieve stresses. The rule is
that spalling cracks initiate where the ratio of
shear stress to Vickers hardness is maximum
(Ref 124). However, this relationship is not
correct when there are excessive amounts of
retained austenite. Generally, it is assumed that
spalling will occur when L41.
Fig. 71
Effect of case carbon content and hardening temperature on the contact endurance limit of carburized 18KhGNMFL steel (1.3% Mn, 1.3% Cr, 0.8% Ni, 0.25% Mo,
0.1% V). Source: Ref 126
To prevent pitting fatigue, either the surface

loading must be decreased to a level below the
endurance limit of the material or the hardness
must be increased to increase the endurance
limit (Ref 28). Pitting may also be reduced by
instituting a break-in period at reduced loads and
speeds to improve gear tooth contact (Ref 85,
124].
Other potential causes of fatigue pitting
include hydrogen embrittlement due to water
contamination of the lubricant and particle
contamination of the lubricant, which act as
surface stress-concentration points that lead to
pitting failure.
Case Crushing
If the case depth is too deep, case-core
separation may occur due to the tips of the gear
teeth becoming too brittle and possibly breaking, if the case depth is too thin, the strength
of the gear teeth will be reduced, causing premature pitting, or it may lead to a condition
called case crushing (Ref 131). Case crushing
occurs in heavily loaded case-hardened components such as gears. Case crushing occurs by a
subsurface fatigue process where the high-cycle
contact stress exceeds the endurance limit. This
will occur when subsurface stresses exceed the
strength of the core. Case crushing failures may
have a similar appearance to pitting, although it
often occurs as longitudinal cracks on the surface of only one or two gear teeth, where sections of the tooth surface may subsequently
break away. However, the case material may
appear to have chipped away from the core in
large flakes (Ref 85). The observed cracks will
move toward the case-to-core boundary and then
to the gear surface (Ref 129).
Adequate case support is provided by proper
core structure to not only prevent case crushing
but also to transmit torque, support bending
loads, and provide adequate toughness to prevent
brittle fracture. The presence of any ferrite will
contribute to reducing the toughness of the core.
Case crushing may be prevented by increasing
the case depth and possibly the core hardness.
For general applications where core hardnesses
of 30 to 45 HRC are specified, the required case
depth can be estimated from (Ref 128):
Case depth to 50 HRC=1:2 107 (W)=F
where W is the force in pounds pressing the

surfaces together, and F is the length of the line
pg 231
18/8/2008 3:22PM Plate # 0
contact in inches. The strength of the core can be

determined from hardness. Since case crushing is
promoted by shear, the shear strength of the core
must be determined and can be estimated from
Fig. 72 (Ref 128). Generally, the subsurface
stress/strength ratio should not be greater than
0.55.
Eryu et al. have studied case-crushing fracture
mechanisms of carburized 20CrMnTi by
scanning electron microscopy (Ref 130). They
showed that the surface features of the primary
cracks exhibited scaly features and that there
were two features of the fracture surfaces of the
branching cracks. The fatigue steps and dimples
were analogous to materials of higher strength,
and the fragmentation pattern was analogous to
brittle material. In addition, spherical particles
were observed that were composed of a-iron,
which were proposed to be caused by the
movement of the faces of the primary crack due
to shearing and compressive stresses.
Pitting Corrosion
Pitting corrosion is a localized penetrating
corrosion attack of typically corrosion-resistant
steel resulting in a mass loss of the steel (Ref
131). Pitting corrosion is related to localized
discontinuities of a passive layer caused by
mechanical imperfections, inclusions, surfacelocalized chemical attack of the passive layer by
salts such as chlorides, or by overaggressive
lubricant additives. After the corrosion pit is
Fig. 72
Shear strength of carburizing steels as a function of

hardness
created, the localized chemical surrounding is

much more aggressive than the surrounding area
of the uncorroded material.
Initiation of the pitting process is dependent on
temperature and on the steel surface, including
the presence of sediment. In some applications
where pitting corrosion is more prevalent, such
as steel in concrete structures, the pitting corrosion process is characterized by the temperature
or narrow range of temperatures above which
pitting will nucleate. The creation of corrosion
sediments will lead to a temperature decrease.
Pitting corrosion will only occur above this critical temperature. Therefore, to increase the
lifetime of steel used in reinforced concrete by
reducing the rate of pitting corrosion, frequent
sediment removal (cleaning) is recommended.
Most often, pitting corrosion is initiated by the
presence of chloride salts, and the rate of corrosive attack is steel alloy dependent. The critical concentration of chlorides for different
steel alloys cannot be defined, because corrosivity is dependent on other chemicals that may
be present, which will affect the rate of corrosion
attack. However, since pitting corrosion is
typically relatively fast, it should be prevented.
Resistance of steel to pitting corrosion is
dependent on the alloy composition (chromium,
molybdenum, tungsten, nitrogen). Relative corrosion resistance of steel alloys may be empirically quantified by:
Relative corrosion resistance=%Cr
+3:3 %Mo+0:5 (%W)+16 (%N)
Chromium and molybdenum are also useful

alloy additions to minimize the potential for
stress-corrosion cracking.
Corrosion pitting may also be caused by
chemical attack of the steel surface by lubricant
additives such as extreme-pressure additives,
particularly in the presence of acid, water, or
contaminants. Also, during use, the oil itself will
oxidatively degrade, producing acidic by-products that may lead to corrosion pitting. In
addition to pitting, corrosive attack may occur at
the grain boundaries of the carburized case. It is
particularly important that components exposed
to saltwater, liquid chemicals, or other foreign
materials during use should be sealed from their
operating environment (Ref 85).
Pitting corrosion may also occur during heat
treatment, particularly those processes involving
salt baths (Ref 132). Heating of steel with scale
on a surface not only accelerates decarburizing,
pg 232
18/8/2008 3:22PM Plate # 0
pg 233
but pitting corrosion is also accelerated. The

corrosive attack is increased with temperature
and holding time. If the scale is not uniform, then
pitting corrosion is localized to those areas
where scale is present. If residual salt from the
bath crystallizes on the surface of the steel,
violent boiling may occur during subsequent
quenching in oil, which may result in blister
formation on the surface. After cleaning, localized pitting corrosion will then occur. To avoid
pitting corrosion during furnace heating, parts
should be thoroughly cleaned.
In some cases, stainless steel machine parts
are carburized to reduce wear. However, carburizing a stainless steel (1Kh16N2AM) reduces
corrosion resistance where machine parts are
used in humid environments (Ref 133). The
potential for pitting corrosion increases as the
amount of d-ferrite increases. Corrosion resistance is reduced as the quenching temperature
prior to carburizing is increased. Although
increasing the quenching temperature after carburizing to 950 to 1100 C does not affect corrosion resistance, decreasing the quenching
temperature to 800 to 850 C reduces corrosion
resistance. Increasing the tempering temperature decreases corrosion resistance, as shown in
Fig. 73 (Ref 133). The corrosion resistance of
carburized steel is greatest after stress relieving
at 250 C.
Partial Melting
Partial melting occurs when there is nonuniform heating of the surface of the steel, such that
some areas are heated to the liquation temperature (the partial melting temperature of an alloy)
(Ref 77, 131). Corners and edges are particularly
susceptible to partial melting. Microscopically,
the presence of partial melting is typically
observed as black spots containing retained
austenite in a large cluster of carbides. Macroscopically, partial melting is accompanied by
the formation of tiny surface cracks.
Partial melting occurs when the carburized
steel is heated to an excessively high temperature, resulting in incomplete or selective carburizing of the surface. For example, partial
melting may occur during stray current flow into
the load from electrodes used to heat salt pot
furnaces or if a load is placed too close to the
furnace hearth, so that some areas of the load are
heated to an abnormally high temperature. To
avoid this defect, heating in salt baths with
appropriate composition and at appropriate austenitizing temperature should be conducted by
keeping the load at a recommended distance
from the heating electrodes. Similarly, when
heating in a conventional furnace, the load
should be properly placed to facilitate uniform
heating.
It is also important to be aware of the liquation
temperature (beginning of melting) of the alloy
being heated. Some typical examples of the
approximate liquation temperatures for different
steels are provided in Table 12, where the
soaking temperature is 2 to 3 min. Precise definition of the partial melting temperature range is
typically a difficult task, because of the relatively large data scatter due to potential compositional variation within the alloy, variation of
carbon content, and, in some cases, relatively
large carbide segregation. These structural variations favor the potential for partial melting to
Table 12 Approximate liquation temperature
for various steel alloys
Fig. 73
Effect of tempering temperature on corrosion resistance of carburized stainless steel 1Kh16N2AM. The
corrosion test was conducted in a humidity cabinet. Source:
Ref 133
Steel alloy
Approximate liquation
temperature, C
SW14
SW18
SW7Mo
SK5
SKC
SK5V
SW12C
SK10V
1320
1330
1280
1350
1280
1270
1260
1250
18/8/2008 3:22PM Plate # 0
occur or the creation of ledeburite networks.

Improvement in temperature control will reduce
the potential for partial melting defects.
REFERENCES
1. B.V.R. Raja, N. Pal, P.L. Talwar, and N.P.

Jayaswal, Engineering Steels for Automotive Applications, Steel World (No. 3),
March 2005, p 13
2. R.F. Kern and M.E. Suess, Chapter 10
Selection of Steels for Carburized Gears,
Steel Selection: A Guide for Improving
Performance and Profits, John Wiley &
Sons, New York, NY, 1979, p 181205
3. D.H. Herring, Gear Heat Treatment: The
Influence of Materials and Geometry,
Gear Technol., March/April 2004, p 3540
4. M.B. Karamis and R. Ipek, An Evaluation
of the Using Possibilities of the Carbonitrided Simple Steels Instead of Carburized
Low Alloy Steels (Wear Properties),
Appli. Surf. Sci., Vol 119, 1997, p 2533
5. R. Selcuk, R. Ipek, and M.B. Karamis,
A Study on Friction and Wear Behaviour
of Carburized, Carbonitrided and Borided
AISI 1020 and 5115 Steels, J. Mater.
Process. Technol., Vol 141, 2003, p 189
196.
6. R.F. Kern, Selecting Steels and Designing
Parts for Heat Treatment, American
7. A. Carter, Failures in Carburized Materials, FWP J., Aug 1983, p 2728
8. K. Palaniradja, N. Alagumurthi, and V.
Soundararajan, Optimization of Process
Variables in Gas Carburizing Process: A
Taguchi Study with Experimental Investigation on SAE 8620 and AISI 3310
Steels, Turkish J. Eng. Env. Sci., Vol 29,
2005, p 279284
9. J.N. Dutta, Gear Failures for Heat Treatment and Metallographic Reasons, Tech.
J. Tata Iron Steel Co. (TISCO), Vol 24
(No. 4) 1977, p 143149
10. O. Asi, A.C. Can, J. Pineault, and M.
Belassel, The Relationship Between Case
Depth and Bending Fatigue Strength of
Gas Carburized 8620 Steel, Surf. Coat.
Technol., Vol 201, 2007, p 59795987
11. C.J. Kuehmann, G.B. Olson, J.P. Wise,
and C. Campbell, Advanced Case Carburizing Secondary Hardening Steels, U.S.
Patent 6,176,946, Jan 23, 2001
12. Carburizing, http://www.treatallmetals.

com/gas.htm, Treat All Metals, Inc., Milwaukee, WI (accessed Feb 2008)
13. Carburizing, http://www.key-to-steel.
com/Articles/Art114.htm, Key to Metals
Group, Zurich, Switzerland (accessed Feb
2008)
14. V.D. Kalner, V.A. Teldekov, and S.A.
Yurasov, Microstructural Hardenability
Diagrams of the Impregnated Case and
Possibilities of Their Practical Use in
Carburizing and Carbonitriding, Met. Sci.
Heat Treat. (USSR), (No. 4) April 1984,
p 3135 (translated from Russian)
15. R.F. Kern and M.E. Suess, Appendix
VIIMethods of Calculating the Hardenability of Heat Treating Grades of Steel,
Steel Selection: A Guide for Improving
Performance and Profits, John Wiley &
Sons, New York, NY, 1979, p 405437
(see p 424434)
16. J. Grosch, Chapter 14Fundamentals of
Carburizing and Toughness of Carburised
Components, Quenching and Carburizing,
Proceedings of the Third Seminar of the
International Federation for Heat Treatment and Surface Engineering, The Institute of Materials, London, U.K., 1993,
p 227249
17. R.E. Gliner, Computer Design of Rational
Alloying of Carburized Steels, Met. Sci.
Heat Treat. (USSR), Vol 42 (No. 78),
JulyAug 2000, p 289291 (translated
from Russian)
18. J.M. Tartaglia and G.T. Eldis, Core Hardenability Calculations for Carburizing
Steels, Metall. Trans. A, Vol 15, June
1984, p 11731183
19. G. Parrish, Chapter 6Core Properties
and Case Depth, Carburizing: Microstructure and Properties, ASM International, 1999, p 135170
20. M. Cristinacce, Heat Treatment of Metals:
Distortion in Case Carburized ComponentsThe Steelmakers View, Heat
Treat. Met., (No. 1), 1999, p 1821
21. K. Genel and M. Demirkol, Effect of Case
Depth on Fatigue Performance of AISI
8620 Carburized Steel, Int. J. Fatigue,
Vol 21, 1999, p 207212
22. H.E. Boyer, Ed. Chapter 1Carburizing
Principles, Case Hardening of Steel, ASM
23. S. Farfan, C. Rubio-Gonzalez, T. Cervantes-Hernandez, and G. Mesmacque,
pg 234
18/8/2008 3:22PM Plate # 0
24.
25.
26.
27.
28.
29.
30.
31.
32.
High Cycle Fatigue, Low Cycle Fatigue

and Failure Modes of Carburized
Steel, Int. J. Fatigue, Vol 26, 2004, p 673
678
H.E. Boyer, Ed., Chapter 2Gas Carburizing Practices, Case Hardening of Steel,
ASM International, 1987, p 1369
G. Krauss, Chapter 13Microstructure,
Residual Stresses and Fatigue of Carburized Steels, Quenching and Carburizing,
Proceedings of the Third Seminar of the
p 205225
J. Titus, The Top Ten Ways to Keep Your
New or Used Carburizing Equipment
Operating at Maximum Performance and
Efficiency, Proceedings of the 22nd Heat
Treating Society Conference and the Second International Surface Engineering
Congress, N.B. Dahotre, R.J. Gaster, R.A.
Hill, and O.O. Popoola, Ed., Sept 1517,
2003 (Indianapolis, IN), ASM International, p 5262
R.S. Hyde, G. Krauss, and D.K. Matlock,
Phosphorus and Carbon Segregation:
Effects on Fatigue and Fracture of GasCarburized Modified 4320 Steel, Metall.
Mater. Trans. A, Vol 26 (No. 6), June
1994, p 12291240
T. Sekercioglu and V. Kovan, Pitting
Failure of a Truck Spiral Bevel Gear, Eng.
Fail. Anal., Vol 14, 2007, p 614619
C. Razim and H. Streng, Properties of
Case-Hardened
ComponentsBasic
Considerations Concerning the Definition
and Evaluation of Toughness, CaseHardened Steels: Microstructural and
Residual Stress Effects, D.E. Diesburg,
Ed., The Metallurgical Society of AIME,
1984, p 115
G. F. Vander Voort, Ed., Ferrous Martensite, Metallography and Microstructures,
Vol 9, ASM Handbook, ASM International, 2004.
AISI 8620 Thermal Treatments, http://
www.crucibleservice.com/eselector/
prodbyapp/carbon/8620.html, Crucible
Service Centers, Camillus, NY (accessed
Feb 2008)
D.W. Ingham and P.C. Clarke, Prediction
of Microstructure and Hardness in Carburized Gears, Case-Hardened Steels:
Microstructural and Residual Stress
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
Effects, D.E. Diesburg, Ed., The Metallurgical Society of AIME, 1984, p 89111
G.E. Totten, M. Narazaki, and R.R.
Blackwood, Failures Related to Heat
Treating Operations, Failure Analysis and
H.E. Boyer, Ed., Chapter 4Liquid CaseHardening Methods, Case Hardening of
Steel, ASM International, 1987, p 103140
G.E. Totten, M. Przylecka, and W. Gestwa,
Use of Aqueous Polymer Quenchants for
Hardening Carbonitrided Parts, Nitriding
Technology Theory and Practice, A.
Nakonieczny, Ed., Proceed. of Ninth
International Seminar International Federation for Heat Treatment and Surface
Engineering, Institute of Precision Mechanics, Warsaw Poland, 2003, p 405414
T. Reti, Residual Stresses in Carburized,
Carbonitrided, and Case-Hardened Components, Handbook of Residual Stress and
Deformation of Steel, G. Totten, M. Howes
and T. Inoue, Ed., ASM International,
2002, p. 189208
G. Parrish, Chapter 4Retained Austenite, Carburizing: Microstructure and
Properties, ASM International, 1999,
p 7797
C.Y. Kung and J.J. Rayment, An Examination of the Validity of Existing Emperical Formulae for the Calculation of Ms
Temperature, Metall. Mater. Trans. A,
Vol 13, Feb 1994, p 328331
I.L. Trusova, Problem of Retained Austenite in Deeply Cemented Steel
20Kh2N4A, Met. Sci. Heat Treat. (USSR),
Vol 40 (No. 1112), Nov-Dec 1998,
p 445448 (translated from Russian)
A.P. Gulyaev, Cold Treatment of Steel,
Met. Sci. Heat Treat. (USSR), Vol 40
(No. 1112), Nov-Dec 1998, p 449455
(translated from Russian)
H.E. Boyer, Ed. Chapter 6Heat Treatment Considerations, Case Hardening of
Steel, ASM International, 1987, p 151194
G. Parrish, Chapter 7Post Hardening
TreatmentsThermal, The Influence of
Microstructure on the Properties of CaseHardened Components, American Society
for Metals, 1980, p 167193
G. Parrish, Chapter 7Post Carburizing
Thermal Treatments, Carburizing: Microstructure and Properties, ASM International, 1999, p 171199
pg 235
18/8/2008 3:22PM Plate # 0
44. D.P. Koistinen, The Distribution of Residual Stresses in Carburized Cases and
Their Origin, Trans. ASM, Vol 50, 1958,
p 227241
45. J.L. Pacheco and G. Krauss, Microstructure and High Bending Fatigue
Strength of Carburized Steel, J. Heat
Treat., Vol 7 (No. 2), 1989, p 7786
46. O. Asi, An Investigation of Retained
Austenite Contents in Carburized SAE
8620 Steel, G. U. J. Sci., Vol 17 (No. 4),
2004, p 103113
47. V.F. da Silva, L.F. Canale, D. Spinelli,
W.W. Bose-Filho, and O.R. Crnkovic,
Influence of Retained Austenite on Short
Fatigue Crack Growth and Wear Resistance of Case Carburized Steel,
J. Mater. Eng. Perform., Vol 8, 1999,
p 543548
48. D. Jeddi, H. Sidhom, D. Ghiglione, and
H.-P. Lierade, Role of Cyclic Stability of
Retained Austenite in Fatigue Performance of Carburized 14NiCr11 Steel,
J. Mater. Eng. Perform., Vol 14, 2005,
p 3749
49. H.-J. Kim and Y.-G. Kweon, The Effects
of Retained Austenite on Dry Sliding Wear
of Carburized Steels, Wear, Vol 193, 1996,
p 815
50. I.S. Kozlovskii, S.E. Manevskii, and I.I.
Sokolov, Effect of Retained Austenite on
the Resistance to Scoring of Steel
20Kh2N4A, Met. Sci. Heat Treat. (USSR),
Vol 20 (No. 1), Jan 1978, p 7880 (translated from Russian)
51. C. Kim, Fracture of Case-Hardened Steels
and Residual Stress Effects, Case-Hardened Steels: Microstructural and Residual Stress Effects, D.E. Diesburg, Ed.,
The Metallurgical Society of AIME, 1984,
p 5987
52. J.A. Burnett, Prediction of Stresses Generated during the Heat Treating of Case
Carburized Parts, Residual Stress for
Designers and Metallurgists, L.J. Vande
Walle, Ed., American Society for Metals,
1981, p 5169
53. N. Kanetake, Chapter 17Effect of Carburising on Mechanical Properties of
Steels, Quenching and Carburizing, Proceedings of the Third Seminar of the
p 281291
54. G. Albertini, G. Bruno, F. Fiori, M. Marcantoni, E. Quadrini, and F. Turquier,

Residual Stresses in Case Hardened
Automotive Steel Gears, Mater. Res.
Forum, Vol 404407, 2002, p 191196
55. M. Benedetti, V. Fontanari, B.-R. Hohn, P.
Oster, and T. Tobie, Influence of Shot
Peening on Bending Tooth Fatigue Limit
of Case Hardened Gears, Int. J. Fatigue,
Vol 24, 2002, p 11271136
56. J. Champaigne, Shot Peening Overview,
http://www.shotpeener.com, Electronics
Inc., Mishawaka, IN, 2001 (accessed Feb
2008)
57. B.A. Shaw, C. Aylott, P. OHara, and K.
Brimble, The Role of Residual Stress on
Fatigue Strength of High Performance
Gearing, Int. J. Fatigue, Vol 25, 2003,
p 12791283
58. Nakonieczny, W. Szyrie, and J. Suwalski
Residual Stresses, Microstructure and
Fatigue Behaviour of Carburized Layers
Before and After Shot Peening, Paper
1996073, Sixth International Conference
on Shot Peening, Conference Proceedings
of ICSP-6, J. Champaigne, Ed., Sept 25,
1996 (San Francisco, CA)
59. M. Kikuchi, H. Ueda, K. Hanai, and T.
Naito, The Improvement of Fatigue Durability of Carburized Steels with Surface
Structure Anomolies by Shot Peening,
Paper 1984067, Second International
Conference on Shot Peening, Conference
Proceedings of ICSP-2, 194, p 208214
60. D.H. Herring, A Tutorial on Shot Blasting
and Shot/Laser Peening, Ind. Heat., Vol 73
(No. 8), Aug 2006, p 2022
61. Shot Peening and Shot Blasting, OSKKiefer GmbH, Oberflachen and Strahltechnik, Ettlingen, Germany
62. M. Kobayashi, T. Matsui, and Y. Murakami, Mechanism of Creation of Compressive Residual Stress by Shot Peening,
Int. J. Fatigue, Vol 20 (No. 5), 1998, p 351
357
63. C. Jia-Xiang and A. Bing-qiu, The Effect
of Shot Peening on Contact Fatigue Life of
Carburized Steel, Proceedings of First
International Conference on Shot Peening, Sept 1417, 1981 (Paris), Pergamon
Press, p 333339
64. C. Peyrac, J.F. Flavenot, and F. Convert,
Combining Case Hardening and Shot
Peening for Gear Steels: Influence on
Residual Stress, Proceedings for the
pg 236
18/8/2008 3:22PM Plate # 0
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
Seventh International Conference on Shot

Peening, A. Nakonieczny, Ed., Sept 28
Oct 1, 1999 (Warsaw, Poland), p 362370
M. Kikuchi, H. Ueda, and J. Naito, Fatigue
Behavior of Carburized Steel with Internal
Oxides Near the Surface, Metall. Mater.
Trans. A, Vol 18 (No. 1), Jan 1987, p 156
158
J.L. Woods, S.R. Daniewicz, and R. Nellums, Increasing the Bending Fatigue
Strength of Carburized Spur Gear Teeth
by Presetting, Int. J. Fatigue, Vol 21, 1999,
p 549556
K. Funatani, Residual Stresses during Gear
Manufacture, Handbook of Residual Stress
and Deformation of Steel, G. Totten, M.
Howes, and T. Inoue, Ed., ASM International, 2002, p 437457
Chapter 14Dimensional Control, Gas
Carburizing, ASM Society for Metals,
1964, p 125136
J. Koutsky and V. Pechman, HeatTreatment Procedure and Volume Change,
Heat Treatment of Engineering Components: Proceedings of The Conference
Organized by the Heat Treatment Joint
Committee of the Iron and Steel Institute,
Dec 1718, 1969 (Royal Garden Hotel,
London), The Institute of Metals and The
Institute of Metallurgists, p 129133
G.A. Fuller, Atmosphere Carburizing
at 927 C (1700 F) for Distortion Control, J. Heat Treat., Vol 8 (No. 1) 1990,
p 2126
C. Prinz, B. Clausen, F. Hoffmann, R.
Kohlmann, and H.-W. Zoch, Metallurgical
Influence on Distortion of the CaseHardening Steel 20MnCr5, Proceedings of
First International Conference on Distortion Engineering, Sept 1416, 2005
(Bremen, Germany), p 7582
D.T. Llewellyn and W.T. Cook, HeatTreatment Distortion in Case-Carburizing
Steels, Met. Technol., May 1977, p 265
278
L.N. Vashchuk, O.A. Konko, V.T. Alekseenko, and V.K. Adamanova, Volume
Changes of Large Gear Wheels in Carburization and Heat Treatment, Met. Sci.
Heat Treat. (USSR), Vol 28 (No. 12),
NovDec 1986, p 858860 (translated
from Russian)
V.A. Murzin, V.P. Ponomarev, N.N. Tolmachevakii, G.D. Glezer, A.V. Poropov,
and A.P. Kul, Effect of Separate Heat
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
Treatment Parameters on the Accuracy of

Gear Wheels, Met. Sci. Heat Treat.
(USSR), Vol 12 (No. 12), Dec 1970,
p 10071010 (tranlated from Russian)
V.A. Bulgakov, Effect of Original Structure on Deformation and Warping of
Machine Parts after Final Heat Treatment,
(No. 9), Sept 1977, p 784786 (translated
from Russian)
N.M. Grinberg and M.P. Konovalova,
Reasons for the Changes in the Dimensions of Steel Parts During Carburizing,
Met. Sci. Heat Treat. (USSR), Vol 10 (No.
6), June 1968, p 483485 (translated from
Russian)
J.Wyszykowski Chapter 12The Defects
and Quality Inspection of Heat-Treatment,
The Engineers HandbookThe HeatTreatment of Ferrous Metals, W. Luty,
Ed., Publishing House of ScientificallyTechnical, Warsaw, 1977, p 600627
G. Krauss, Chapter 10Surface Hardening, Principles of Heat Treatment of
Steel, American Society for Metals, 1980,
p 247273
G. Krauss, Steels: Processing, Structure,
and Performance, ASM International,
2005
C.R. Brooks, Chapter 4Quenching of
Steels, Principles of the Heat Treatment of
Plain Carbon and Low Alloy Steels, ASM
S.I. Aksenova, M.A. Balter, I.S. Dukarevich, and M.M. Gurarii, Causes of Cleavage on Case-Hardened Gear Wheels,
(No. 3), March 1988, p 191194 (translated
from Russian)
A.J. McEvily, K. Pohl, and P. Mayr,
Comparison of the Fractographic Features
of a Carburized Steel Fractured Under
Monotonic or Cyclic Loading, Mater.
Charact., Vol 36, 1996, p 153157
Kh.G. Tkhagapsoev and G.K. Surzhinskii,
Cracking in the Carburized Case of Gears
During Their Grinding, Met. Sci. Heat
Treat. (USSR), Vol 11 (No. 34), March
April 1969, p 224226 (translated from
Russian)
G. Parrish, Chapter 8Post Hardening
TreatmentsMechanical, The Influence
of Microstructure on the Properties of
Case-Hardened Components, American
pg 237
18/8/2008 3:22PM Plate # 0
85. G. Parrish, Chapter 8Postcarburizing

Mechanical Treatments, Carburizing:
Microstructures and Properties, ASM
86. M. Lawerenz and I. Ekis, Creating an InHouse Shot Peening Specification for
GearsPart 2, Paper 1993118, Shot
Peener, Vol 7 (No. 1), 1993, p 35
87. C. Bahnsen, B. Clausen, F. Hoffmann, P.
Mayr, and H.-W. Zoch, Influence of
Carburising on Distortion Behavior, Proceedings of First International Conference on Distortion Engineering, Sept
1416, 2005 (Bremen, Germany), p 235
242
88. M. Izciler and M. Tabur, Abrasive Wear
Behavior of Different Case Depth Gas
Carburized AISI 8620 Gear Steel, Wear,
Vol 260 (No. 12), 2006, p 9098
89. M. Boniardi, F. DErrico, and C. Tagliabue, Influence of Carburizing and Nitriding on Failure of GearsA Case Study,
Eng. Fail. Anal., Vol 13, 2006, p 312339
90. G. Parrish, Chapter 6Core Properties
and Case Depth, The Influence of Microstructure on the Properties of CaseHardened Components, American Society
for Metals, 1980, p 131165
91. L. Faqian, L. Shuanlu, l. Helin, Q.
Changyi, Z. Jie, T. Jiping, and C. Jun,
Analysis of Causes of Casing Elevator
Fracture, Eng. Fail. Anal., Vol 14, 2007,
p 606613
92. Z. Yu, X. Xu, and H. Ding, Failure Analysis of a Diesel Engine Piston-Pin, Eng.
Fail. Anal., Vol 14, 2007, p 110117
93. Z.Ya. Nostislavskii, Yu.N. Griboedov, and
S.I. Markov, Effect of the Atmosphere
During Heating of Carburized Cr-Ni Steel
to Quenching Temperature on Decomposition of Austenite, Met. Sci. Heat Treat.
(USSR), Vol 18 (No. 56), MayJune 1976,
p 443445
94. G. Parrish, Chapter 2Decarburization,
Carburizing: Microstructure and Properties, ASM International, 1999, p 3749
95. G. Parrish, Chapter 5Influential
Microstructural Features, Carburizing:
Microstructure and Properties, ASM
96. Atlas of Microstructures of Industrial
Alloys, Vol 7, Metals Handbook, 8th ed.,
American Society for Metals, 1972, p 4
97. T. Isogai, R. Uchino, M. Sato, and Y.
Kono, Method for Heat Treatment of
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
Steel, U.S. Patent 6,149,734, Nov 21,

2000
E. Gianotti, I. Calliari, M. Zanesco, and
E. Ramous, Grain Boundary Oxidation
in Endothermic Gas Carburizing Process,
Proceedings of the Seventh International
Tooling Conference, Tooling Materials
and Their Applications from Research to
Market, Vol 2, May 25, 2006, Politecnico
Di Torino, Torino, Italy, p 481490
D.K. Matlock, K.A. Alogab, M.D.
Richards, and J.G. Speer, Surface Processing to Improve the Fatigue Resistance of
Advanced Bar Steels for Automotive
Applications, Mater. Res., Vol 8 (No. 4),
2005, p 453459
H.-P. Schmidt, Process for Avoiding Surface Oxidation in the Carburization of
Steels, U.S. Patent 5,498,299, March 12,
1996
V.D. Kalner and S.A. Yurasov, Internal
Oxidation During Carburizing, Met. Sci.
Heat Treat. (USSR), Vol 12, (No. 6), June
1970, p 451454 (translated from Russian)
J.D. Verhoeven, N. Chen, and A.J. Bevolo,
Scanning Auger Microscopy Characterization of the Internal Oxidation Produced
on Carburizing, J. Heat Treat., Vol 9
(No. 2), 1992, p 7380
X. An, J. Cawley, W.M. Rainforth, and L.
Chen, A Study of Internal Oxidation in
Carburized Steels by Glow Discharge Optical Emission Spectroscopy and Scanning
Electron Microscopy, Spectrochim. Acta
B, At. Spectrosc., Vol 58, 2003, p 689698
S. Laue, H. Bomas, and F. Hoffman,
Influence of Surface Condition on the
Fatigue Behaviour of Specimens Made of
SAE 5115 Case-Hardened Steel, Fatigue
Fract. Eng. Mater. Struct., Vol 29, 2006,
p 229241
M. Lohrmann and W. Lerche, Investigation of Process and Steel Influences on the
Extent of Internal Oxidation during Gas
Carburizing, Heat Treat. Met., Vol 27
(No. 4), 2000, p 8590
I.S. Kozlovskii, A.T. Kalinin, A. Va. Novikova, E.A. Lebedeva, and A.I. Feofanova,
Internal Oxidation during Case-Hardening
of Steels in Endothermic Atmospheres,
Met. Sci. Heat Treat (USSR), Vol 9 (No. 3),
March 1967, p 157161
W.D. Kehr and H.E. Seese, Surface Attack
in Chromium-Iron Alloys, IBM J., Sept
1966, p 420421
pg 238
18/8/2008 3:22PM Plate # 0
108. Z. Yu and X. Xu, Failure Analysis of an

Idler Gear of Diesel Engine Gearbox, Eng.
Fail. Anal., Vol 13, 2006, p 10921100
109. M. Umemoto, Z.G. Liu, K. Masuyama,
and K. Tsuchiya, Influence of Alloy Additions on Production and Properties of
Bulk Cementite, Scr. Mater., Vol 45, 2001,
p 391397
110. S.N. Saltykov and N.V. Tarasova, Anodic
Dissolution of Ferrite Phases of IronCarbon Ferrite-Cementite Alloys with
Different Forms of Cementite, Prot. Met.
(Russia), Vol 42 (No. 5), 2006, p 501506
111. V.Ya. Zubov, Patenting of Steel Wire,
Met. Sci. Heat Treat., (USSR), Vol 14
(No. 9), Sept 1972, p 793800
112. G. Parrish, Chapter 3Carbides, The
Influence of Microstructure on the Properties of Case-Hardened Components, American Society for Metals, 1980, p 3960
113. T. Ichihara and H. Fujiwara, Pullu Unit
Having One-Way Clutch, U.S. Patent
Application US2003/0085091 A1, May 8,
2003
114. T. Nakamura, S. Nakamura, T. Kimura,
T. Yamaguchi, K. Otani, N. Uchiyama,
N. Kino, and T. Nanbu, High Bearing
Pressure-Resistant Member, U.S. Patent
6,569,267 B1, May 27, 2003
115. G. Krauss, Heat Treatment, Microstructures, and Residual Stresses in Carburized Steels, Conf. Proceed. Quenching
and Distortion Control, G.E. Totten, Ed.,
116. L.P. Karpov, Effect of Quenching Conditions on the Carbide Network in Steel
07Kh16N6, Met. Sci. Heat Treat. (USSR),
Vol 35 (No. 1), Jan 1993, p 2023 (translated from Russian)
117. G. Diego, M. Serrano, and A.M. Lancha,
Failure Analysis of a Multiplier from a
Kaplan Turbine, Eng. Fail. Anal., Vol 7,
2000, p 2734
118. L. Zhang and B.G. Thomas, State of the
Art in Evaluation and Control of Steel
Cleanliness, ISIJ Int., Vol 43 (No. 3),
2003, p 271291
119. Y. Murakami and M. Endo, Effects of
Defects, Inclusions, and Inhomogenities
on Fatigue Strength, Fatigue, Vol 16,
April 1994, p 163182
120. H. Bomas and M. Schleicher, Application
of the Weakest-Link Concept to the
Endurance Limit of Notched and Multiaxially Loaded Specimens of Carburized
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
Steel 16MnCrS5, Fatigue Fract. Eng.

Mater. Struct., Vol 28, 2005, p 983995
T. Wilks, G.P. Cavallaro, C. Subramanian,
K.N. Strafford, P. French, and J.E. Allison,
Conditions Prevailing in the Carburizing
Process and Their Effect on the Fatigue
Properties of Carburized Gears, J Mater.
Process. Technol., Vol 40, 1994, p 111
125
T. Masuyama, K. Inoue, M. Yamanaka, K.
Kitamura, and T. Saito, Evaluation of
Bending Strength of Carburized Gears
Based on Inferential Identification of
Principal Surface Layer Defects, JSME
Int. J., Series C, Vol 45 (No. 3), 2002,
p 794801
P. Brechot, A.B. Cardis, W.R. Murphy,
and J. Thiessen, Micropitting Resistant
Industrial Gear Oils with Balanced Performance, Ind. Lubr. Tribol., Vol 52
(No. 3), p 125136
S. Matsumoto, On the Design of Surface
Durability of Rolling-Sliding Contact
Surface, Technical Paper n77-DET-115,
American Society of Mechanical Engineers, Sept 2630, 1977
T.L. Krantz, M.P. Alanou, H.P. Evans, and
R.W. Snidle, Surface Fatigue Lives of
Case-Hardened Gears with an Improved
Surface Finish, J. Tribology, Vol 23, Oct,
p 709716
B.B. Vinokur, S.E. Kindratyuk, R.A.
Khrunik, K.B. Katsov and I.V. Kokotailo,
Influence of Carbon on the Contact Endurance of Carburized and High-Carbon
Steels, Sov. Mater. Sci., Vol 22 (No. 2),
March 1986, p 211213
R.C. Graham, A. Olver, and P.B. Macpherson, Investigation into the Mechanisms of Pitting in High-Hardness
Carburized Steels, Paper 80-C2/DET-118,
ASME, 1980
R.F. Kern and M.E. Suess, Chapter 9
General Principles for the Selection of
Steels for Case Hardening, Steel Selection:
A Guide for Improving Performance and
Profits, John Wiley & Sons, New York,
NY, 1979, p 159180
A. Bensely, S.S. Jayakumar, D. Mohan
Lal, G. Nagarajan and A. Rajadurai, Failure Investigation of Crown Wheel and
Pinion, Eng. Fail. Anal., Vol 13 (No. 8),
Dec 2006, p 12851292
S. Eryu, C. Qing, and L. Xiaogang, Fracture of Case Crushing of Carburized
pg 239
18/8/2008 3:22PM Plate # 0
Steel, Wear, Vol 122 (No. 1), 1988, p 57

62
131. A. Moszczynski, The Gas Carburized of
Steel, Publishing House of ScientificallyTechnical, Warsaw, 1983, p 259262
132. V.I. Muravev and V.P. Kurbatov, Pitting
Corrosion in the Process of Heat Treatment, Met. Sci. Heat Treat. (USSR), Vol 12
(No. 5), May 1970, p 393395 (translated

from Russian)
133. V.I. Belyakova, M.F. Alekseenko, L.Ya.
Gurvich and V.L. Erofeeva, The Corrosion
Resistance of Carburized Stainless Steels,
(No. 2), Feb 1973, p 143146 (tanslated
from Russian)
pg 240
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_241-253.pdf/F5113_07/

DOI: 10.1361/faht2008p241
18/8/2008 3:26PM Plate # 0

Fatigue Fracture of Nitrided Layers

Aleksander Nakonieczny, Institute of Precision Mechanics
THE DURABILITY of products depends

strongly on surface conditions on the order of
1 mm to several millimeters, depending on the
type of technological process applied. The condition of the surface layer is critical to wear
resistance during the process of friction and to
corrosion resistance. In the case of mechanical
loading (especially fatigue resistance) and corrosion, a critical role is played by the substrate, its
condition and properties, as well as the atomic
relationship between the surface layer and the
substrate. For this very reason, one should take
into account the substrate-surface layer system
when considering the life expectancy of machine
components and assemblies. Attributes of such a
system are the thickness of the surface layer, the
ratio of this thickness to the entire cross section,
the ratio of surface hardness to core hardness,
and the state of residual stresses, usually compressive, situated within the surface layer relative to the state of stresses in the substrate, which
are usually tensile. An incorrectly applied surface layer may cause the formation of a structural
flaw in the transition zone of the layer and may
be the location of crack initiation, especially by a
fatigue mechanism (Ref 1).
Surface engineering encompasses various
process technologies and also the service properties of products, surface-layer investigation
methodologies, and design aspects of the
substrate-layer system. In terms of service
properties of products, the functionality of
surface treatment may be assessed by defining
the fatigue limit, wear resistance, or corrosion
resistance. Such evaluations are usually performed on specimens in laboratory conditions.
However, the most valuable information is to be
gained from actual service trials, which also
unfortunately may be costly.
Fatigue Resistance
Fatigue resistance of machine components
is a function of their design, material and
technological parameters, as well as the type of

loading in service conditions. When discussing
the problem of fatigue resistance, one should
consider in detail the effect of these parameters
and, in the case of loads, define fatigue characteristics (e.g., plots for different types of loading,
such as bending, tensile, and torque). These
problems have been sufficiently dealt with in the
technical literature.
In the process of searching for methods to
increase fatigue resistance, there are some constant elements that have a favorable effect,
including enhancing treatments such as thermal,
thermochemical, as well as surface work hardening.
In order to increase fatigue resistance, it is
not sufficient to apply a chosen enhancement
treatment. Rather, it is important to select
the appropriate initial volume heat treatment
prior to successive surface, thermal, and workhardening treatments. The problem of enhancing
fatigue resistance of machine components by
technological methods involves the application
of not one chosen treatment but a cycle of successive treatments. The appropriate selection
of these treatments affects the structural flaw
formed in the process of enhancement, which has
a decisive influence on fatigue resistance (Ref 1).
A structural flaw occurs in all locations where,
as the result of heat treatment (e.g., induction
hardening), thermochemical treatments (carburizing, nitriding, etc.), or work-hardening
treatments (burnishing, shot peening) of machine components, the layer formed in these
processes has different physical-chemical properties than that of the core due to a large gradient
of property changes. The value of the structural
flaw coefficient, bs, depends on the type of
material and the parameters of the technological
processes that cause this structural flaw to form.
In other words, it depends on heat treatment and
surface hardening.
Thermal and surface work-hardening treatments used industrially cause enhancement of
fatigue resistance. Based on research carried out
pg 241
18/8/2008 3:26PM Plate # 0
pg 242
by the Institute of Precision Mechanics, it can

be accepted that the fatigue limit (s 1) rises 15
to 30% on average as a result of implementation
of such treatments. Enhancement of fatigue
resistance is obtained by structural changes,
strengthening, and favorable distribution of
residual stresses, which are formed as a result
of thermal and surface treatments.
Due to physical processes taking place within
the material during the application of surface,
thermal, and work-hardening treatments, changes in microstructural and mechanical properties
arise between the surface layer and the core
of the material. The gradient of changes of
physical-chemical properties depends on the
selected technological process and its parameters. Numerous examples have been noted
where the fatigue crack origins were traced to
the transition zone between the hardened surface
layer and the substrate. Figure 1 shows a fatigue
fracture with the origin located under the hardened layer at the point where stresses mounted.
The mounting of stresses occurs as the result of
residual stresses created during heat treatment
combined with external stresses.
Fatigue Resistance of Steel after Nitriding
and Related Nitriding Treatments. The significance and detailed assessment of the effect of
a structural flaw are explained, using investigations of the effect of variable core conditions on
fatigue resistance as an example.
Reference 2 defines the effect of tempering
temperature and time of nitriding on the
rotational-bending fatigue resistance of 40HMgrade steel (AISI 4140). This structural steel is
commonly used for the manufacture of various
types of machine components (e.g., gears,
crankshafts). Specimens used for the study were
quenched and tempered to the following hardness levels: 30 to 32 HRC, 33 to 34 HRC, and 35
to 36 HRC, applying tempering temperatures
within the range of 550 to 620 C.
Nitriding was carried out in a controlled
process at a temperature of 530 C for 4 to
16 h. Investigations encompassed metallurgical
characterization of nitrided layers as well as
determination of fatigue resistance (s 1). Moreover, the yield strength (R0.2) was determined in
conditions of shear bending (Rg0.2).
For the metallurgical investigations pertaining to surface hardnesses, hardness traverses,
layer thickness, and microstructure, the Neophot
30 metallograph and the Zwick microhardness
tester were used.
Investigation of fatigue resistance was carried
out with the aid of the PUNZ machine, manufactured by Schenk. The loading frequency was
100 Hz, and the investigations covered 107
cycles. The value of the fatigue limit was calculated by the Dixon-Mood method. Results of
fatigue tests of nitrided specimens were compared with results obtained for the same steel
(40HM grade) quenched and tempered to a
hardness of 30 to 32 HRC.
Fractographic investigations of fatigue fractures and determination of the chemical composition of visible inclusions on these fractures
were carried out by scanning electron microscopy. Yield strength values were determined
in cases of static bending with the aid of the
Instron TT-DM machine. The results of metallurgical and strength investigations are shown in
Tables 1 and 2.
Table 1 Metallurgical characteristics and
corresponding values of fatigue limits for the
investigated versions of technological processing
Case depth
Fig. 1
Fatigue source located in the transition zone between

case and core
Surface hardness
Temper
Total
Effective at core
temper- Core
ature
hardness, Nitriding at 500 +50 HV
C
mm
HRC
time, h HV, mm
HV1
HV10
s 1
MPa
...
4
16
4
8
16
4
16
...
686
743
752
777
786
772
778
...
642
657
695
707
701
698
699
550
735
745
720
725
777
820
840
620
620
3032
3032
590
3334
550
3536
...
0.13
0.26
0.14
0.19
0.27
0.16
0.30
...
0.24
0.46
0.22
0.29
0.43
0.2
0.42
18/8/2008 3:26PM Plate # 0
Fatigue Fracture of Nitrided Layers / 243
Investigations of fatigue resistance (Table 1)

showed that nitriding, independently of process
time (case depth and core hardness), caused an
increase in the fatigue limit in comparison with
quenching and tempering. The smallest increase
in fatigue limit value (by 30%) was obtained on
specimens tempered at 590 C and nitrided for
4 h. The greatest increase in the fatigue limit (49
to 53%) was obtained on specimens tempered at
550 C and nitrided for 4 and 16 h. For the
remaining versions of technological processing,
the increase in fatigue limit was almost identical
and amounted to 31 to 35%.
Investigations of quenched and tempered and
subsequently nitrided specimens showed a variation in the value of the fatigue limit (s 1),
depending on the tempering temperature.
Nitriding time did not affect the value of the
fatigue limit, both after tempering at 550 C as
well as at 620 C.
Some variation was observed in the case of
versions in which the tempering temperature
was 590 C. For this tempering temperature, the
highest value of the fatigue limit (777 MPa) was
obtained on specimens nitrided for 16 h. At the
same time, it was established that increasing the
time of nitriding from 4 to 8 h did not affect the
fatigue limit. The difference of approximately
1% is within experimental error. Extending the
time of nitriding from 4 and 8 h to 16 h caused
an increase in fatigue limit value by approximately 50 MPa, that is approximately 6.5%.
A comparison of specimens tempered at 550
and 620 C shows that lowering the tempering
temperature causes an increase in the fatigue
limit by approximately 12%. As has already
been mentioned, decreasing the nitriding time
fourfold, from 16 to 4 h, for a given tempering
temperature does not cause any significant
changes of the s 1 value.
Further comparison with results obtained on
specimens tempered at 590 C indicates that
temperatures of 550 and 620 C cause a clear
variation in the effect of lowering the tempering
temperature, from the point of view of the

fatigue limit. A tempering temperature of
590 C is intermediate between the two aforementioned temperatures, at which this effect is
manifest only after the application of a longer
nitriding time. It should therefore be emphasized
that by the appropriate selection of the tempering temperature, it is possible to achieve an
increase in the fatigue limit with shorter nitriding times. On the other hand, selection of inappropriate tempering temperatures may cause the
inability to achieve an increase in the s 1 value
when the nitriding time is too short.
An analysis of angular coefficients of simple
regressions of the Wohler plots (Table 2) indicates that, for almost all versions, the angle of
inclination of the regression plots is similar. By
the same token, the sensitivity of the material to
a change in loading within limited fatigue
strength is not connected with the time of
nitriding. Only specimens tempered at 550 C
and nitrided for 16 h were characterized by a
slightly higher sensitivity to a change in the level
of loading. By the same token, an increase in the
level of loading within limited fatigue strength
causes a more significant decrease in the number
of loading cycles to failure, compared with other
technological versions. Thus, shortening the
nitriding time coupled with a lower tempering
temperature is favorable even when the loading
level exceeds the fatigue limit.
Fractography of specimen fatigue fractures
showed they are of a fine-grained character, with
the exception of the middle zone, which has a
differing, coarse-grained structure (catastrophic
failure zone). The clearly observed fatigue
sources occur in the form of fisheyes (Fig. 2).
Table 2 Equations of regression for the

investigated technological versions
Tempering
temperature, C
Nitriding
time, h
620
4
16
4
16
4
16
590
550
Regression equations
s 1 = 141.2 lgN+1567.9
s 1 = 62.4 lgN+1146.6
s 1 = 58.6 lgN+1092.7
s 1 = 67.7 lgN+1201.5
s 1 = 137.2 lgN+1658.3
s 1 = 27.6 lgN+1025.8
Fig. 2
Appearance of fatigue source. Original magnification:

100
pg 243
18/8/2008 3:26PM Plate # 0
Initiation of fatigue fractures occurs at the

border between the layer and the substrate, in a
zone characterized by hardness higher than that
of the core by 50 HV. The features initiating
fatigue fractures are nonmetallic inclusions of
calcium sulfides (Fig. 3), originating from the
metallurgical process.
Results of investigations of yield strength in
bending indicate that variation of both the tempering temperature as well as the nitriding time
affects its value insignificantly, not exceeding
7%. The values of Rg are all within the range of
1850 to 1980 MPa.
Results of metallurgical evaluations of nitrided layers (Table 1) show that extension of the
nitriding time from 4 to 16 h causes an almost
doubled growth of the effective minimum tendency to increase the case depth with a lowering
of the tempering temperature. A similar but clear
trend is observed in the surface hardness of
layers, where lowering of the tempering temperature from 620 to 550 C with a 4 h nitriding
time cycle causes a rise in HV1 surface hardness
by approximately 100 units and in HV10 hardness by 50 units.
Figure 4 shows the changes taking place in the
character of the microhardness profile as a
function of process time, with tempering temperature at 590 C. As can be seen, extension of
the nitriding time causes a drop in the angle of
inclination of the microhardness profile. Similar
changes in the microhardness profile versus
process time occur for the remaining tempering
temperatures.
Fig. 3
Precipitation on the fracture surface of a specimen

which served as the source of the fatigue fracture.
Fatigue Evaluation of Nitrided Steels

An evaluation of fatigue properties requires
an understanding of the mechanism. This
mechanism also depends on the condition and
mutual relationship between the substrate and
the surface layer. This requires determination of
the strength condition of the system: substratesurface layer, as a function of external loading.
Among the most important parameters
describing the condition of the surface layer are
microstructure, degree of strengthening, state of
stresses, and roughness. Other important parameters include texture, surface energy, and
chemical composition (Ref 2, 3).
In engineering practice, usable properties
such as tensile strength, Rm and fatigue limit,
may be determined as functions of mechanical
parameters, that is, hardness, H, or tensile
strength, Rm, or surface roughness (Ref 4). Such
correlations maintain their validity for a material
homogeneous throughout its cross section. For
heterogeneous materials (e.g., ones that have
been surface treated), such correlations cannot
be applied directly. The character of distribution
Fig. 4
Microhardness traverses for different nitriding process

times
Fig. 5
Fatigue characteristics for 1, a homogeneous material,

and 2, a heterogeneous material
pg 244
18/8/2008 3:26PM Plate # 0
pg 245
of basic mechanical properties (hardness,

residual stresses, as well as their significant
effect on fatigue properties) vary in different
ways for homogeneous or heterogeneous materials (Fig. 5). Homogeneous materials (curve 1,
Fig. 5) have constant properties through the
cross section. Heterogeneous materials (curve 2)
have variable properties.
The fatigue characteristic, which is shown
by the distribution of the fatigue limit, s, is a
function of hardness, H, and residual stresses, sr:
s=f H,sr
(Eq 1)
The method of designing a usable characteristic

for the fatigue limit distribution has been
described in publications (Ref 35).
The distribution of stresses from extraneous
loading constitutes a significant characteristic,
because it enables the determination of the
strength condition for the surface layer. Loading
characteristics are typical distributions of
stresses across the section of a component or
specimen for the investigated types of loading:
tensile-compressive, bending, or torque. For
smooth specimens, their determination does not
present any problems. Some difficulties may
arise when determining the distribution of
stresses in a notched specimen.
Distribution of stresses from extraneous
loading for notched specimens in conditions of
bending is (Ref 3):

2 x 3a2
smax x=sn a 17
d
(Eq 2)
where smax(x) is the value of local stress at a

distance of (x) from the surface, sn is the
nominal stress, a is the coefficient of stress
concentration, and d is the cross-sectional
dimension.
Knowledge of the fatigue characteristic, as

well as the loading characteristic, allows the
determination of the fatigue strength condition
for any location on the component cross section:
s1 is0z
i
(Eq 3)
where s l is the fatigue limit at any location on

the specimen cross section, and si0z is the value
of stresses from extraneous loading at the given
location i.
The investigations were conducted on structural steels 40HM (4140) and 38HMJ (Nitralloy
135M). The steels were hardened and tempered
prior to nitriding at two temperatures: 550 and
620 C (Ref 1). Nitriding was carried out using
two types of atmosphere, that is, NH3-NH3(diss)
and NH3-N2. In the nitriding processes, the
atmosphere gas composition was varied, as were
the time of nitriding (4 and 16 h) and the
nitriding potential, KN (from 1.65 to 4.8).
Fatigue resistance tests were carried out on
the PUNZ machine (manufactured by Schenck),
applying rotational bending stresses with a
notch (a = 1.02). The specimen diameter was
W = 5.88+0.02 mm. The results of the fatigue
resistance tests, metallurgical evaluation and
process parameters are shown in Table 3.
Figure 6 shows microhardness traverses in the
nitrided case for 40HM-grade steel, while Fig. 7
shows the same for the 38HMJ grade.
Test results show that the tempering temperature has an effect on the properties of the
nitrided case. The effect of the tempering temperature on the basic properties of the nitrided
case depends on the steel grade. A higher
increase in hardness (by approximately 50%) as
well as in case depth is observed for the lowalloy chromium steel 40HM.
As can be seen in Fig. 7 for the 38HMJ steel,
the effects of tempering temperature on the
Table 3 Technological parameters and test results

HV0.5 hardness
Tempering
temperature,
C
40HM (4140)
38HMJ
(Nitralloy 135M)
Steel grade
Nitriding
time, h
Core
hardness
HV0.5
Max
on cross
section
On
surface
550
550
620
690
4
16
4
16
402
396
343
343
677
642
715
343
550
550
620
620
4
16
4
16
356
343
318
296
1030
1030
1030
1030
Residual stresses
Fatigue limit
(s 1), MPa
At
surface,
MPa
Depth at which
stress changes
sign, mm
757
757
826
642
820
840
735
745
600
650
600
900
0.32
0.52
0.37
0.55
1373
1227
1273
1304
805
785
766
810
900
600
450
800
0.25
0.48
0.30
0.45
18/8/2008 3:26PM Plate # 0
changes in hardness and case depth are smaller

than for the 4140 steel. It was found that lowering the tempering temperature raises the fatigue resistance of the core-case system. Shorter
nitriding time, following a low tempering temperature, does not affect the volume fatigue
resistance (Ref 6).
The effect of heat treatment of the core on
fatigue resistance is shown in Fig. 8. From the
illustration, it is seen that raising the core hardness moves the fatigue resistance characteristic,
that is, the distribution of the speed limit value,
across the section in the direction of higher stress
values.
Data to determine the characteristics in Fig. 8
are shown in Table 4. To calculate the fatigue
limit, the following formula was used:
s1 =1:98H70:0011HVi 2
(Eq 4)
where HV is Vickers surface hardness, and HVi

is Vickers hardness at i location on the cross
section. The relationship is valid for a hardness
range of 340jHVj900.
A significant increase in the fatigue limit
value (to 820 MPa) with a tempering temperature of 550 C and up to 735 MPa with a tempering temperature of 620 C, relative to prior
values of 618 and 550 MPa, determined at the
location of fatigue crack initiation (Fig. 9) (on an
average 0.5 mm from the surface) should be
interpreted as the favorable effect of compressive stresses in the nitrided case (Ref 7).
It follows from Fig. 8 that fatigue crack
initiation of nitrided cases (compare with Fig. 7)
takes place under the surface because stresses
Fig. 6
Microhardness traverses across a nitrided case on

40HM (4140)-grade steel. 1, tempering temperature
550
h; 2, tempering temperature 550 C, time 16 h; 3,
tempering temperature 620 C, time 4 h; 4, tempering temperature 620 C, time 16 h
C, time 4
from extraneous loading locally exceed the

value of the fatigue limit, and, in accordance
with curve 3 in Fig. 8, material decohesion must
occur.
Fatigue Property Characteristics after

Carbonitriding
In most modern methods of manufacturing, it
is recommended that the design stage consider
the different manufacturing technologies. In
connection with that, there is an urgent need to
determine material characteristics, especially of
materials after the application of modern technological property-enhancing treatments. There
also exists the need to develop calculation
methods of fatigue resistance after thermal and
thermochemical treatment. This, however, is the
next stage of activity and is possible to carry out
only when the basic fatigue properties of the
steel following heat treatment are known.
The carbonitriding treatment is used for
components exposed to lighter loads and subjected to wear as well as bending (Ref 810). For
those components, which are subjected to contact fatigue during service, the case depths are
designed deeper. For the present series of tests, a
Fig. 7
Microhardness traverses across a nitrided case on

38HMJ (Nitralloy 135M)-grade steel. 1, tempering
temperature 550 C, time 4 h; 2, tempering temperature 550 C,
time, 16 h; 3, tempering temperature 620 C, time 4 h; 4, tempering temperature 620 C, time 16 h
pg 246
18/8/2008 3:26PM Plate # 0
Fig. 8
Distributions of fatigue limit, curves 1 and 2; residual stresses, curve 3 (550 C) and curve 4 (620 C); extraneous loading,
curves 5 and 6, 40HM (4140)-grade steel
Table 4 Fatigue limit values across the specimen

section
Experimental
results
Fatigue limit (s 1),
MPa
Core
At surface
At location of initiation
of fracture
Theoretical
results
Tempering temperature
550 C
620 C
550 C
620 C
613
...
820
550
...
735
618
852
618
550
819
550
case depth of 0.7 mm was selected. The optimal

microstructure of carbonitrided components is
fine acicular martensite with a small amount of
retained austenite and containing no coarse
carbide precipitations (Ref 810).
Specimens prepared to meet the aforementioned conditions were subjected to rotationalbending and one-point bending fatigue tests.
Such types of loading were selected based on the
premise that bending is the most common
method of loading during service, as well as the
fact that it is during bending that one can observe
Fig. 9.
Location of fatigue crack initiation on nitrided 40HM

(4140)-grade steel. Original magnification: 100
the most favorable and strongest effect of surface strengthening.

Simplified Smith curves were plotted to
determine the fatigue resistance for at least three
pg 247
18/8/2008 3:26PM Plate # 0
pg 248
methods of bending (the Wohler curve), as well

as static strength and yield strength for a given
type of loading and materials. The fatigue tests
were carried out with the following coefficients
of cycle asymmetry: R = 1, R = 0.1, and R =
0.3. The coefficients of 0.1 and 0.3 were selected
to ensure the possibility of running the tests only
within the range of one-sided bending stresses.
The fatigue characteristic was developed for a
material in the quenched and tempered condition
as a reference and, for materials with a diffusion
case, heat treated to the same condition as the
heat-treated-only version. Once these values
were known, the surface coefficient of strengthening was determined from the equation:
m=
sww
1
s1
(Eq 5)
ww
where s 1
is the fatigue limit of the enhanced
specimen, and s 1 is the fatigue limit of the
reference specimen.
Rotational-bending fatigue tests were carried
out on the Schenck fatigue machine with a
constant distribution of the bending movement
along the length of the specimen. The frequency
was 100 Hz.
One-point bending tests were carried out on
the Amsler machine. In order to obtain the
bending effect on this machine, a prototype
addition was designed that, through a lever,
allows loading of the tested section of the specimen under a constant bending moment (Fig. 10).
The frequency was 150 Hz. Tests were carried
out to NG = 107 cycles.
The material used in these tests was the
18HGT grade, normalized, with a fine-grained
Fig. 10
Schematic of equipment for fatigue testing in the

rotational-bending mode
ferritic-pearlitic microstructure. The chemical

composition is given in Table 5. Tests were also
carried out on carbonitrided specimens. Carbonitriding of specimens from 18HGT-grade steel
was carried out at a temperature of 860 C in
an endothermic atmosphere enriched with ammonia and natural gas.
Metallurgical evaluations were carried out on
18HGT steel in the quenched and tempered only
and carbonitrided condition. In the normalizedonly condition, the specimens showed a ferriticpearlitic microstructure with very fine-grained
pearlite (Fig. 11). The microstructure of specimens with diffusion cases was determined based
on micrographs and microhardness measurements.
Specimens made from 18HGT steel, after
carbonitriding and quenching and tempering,
exhibit a microstructure of tempered martensite
in the subsurface zone (Fig. 12) and a bainiticmartensitic microstructure in the core (Fig. 13).
The microstructure of the core was 550 HV0.1.
To determine the fatigue resistance, a static
bending test was carried out. For the heat treated
Table 5 Chemical composition of specimens
prepared from steel grade 18HGT
Specimen
No./diameter,
mm
1 12
2 12
3 12
1 14
2 14
3 14
Fig. 11
Chemical composition, %
C
Mn
Cr
Si
Ni
Cu
Ti
0.187
0.185
0.189
0.216
0.225
0.215
0.014
0.015
0.014
0.009
0.009
0.009
0.89
0.87
0.89
1.00
1.02
0.98
1.07
1.12
1.12
1.00
1.02
1.02
0.26
0.28
0.32
0.35
0.34
0.33
0.06
0.06
0.06
0.13
0.13
0.12
0.13
0.12
0.12
0.13
0.12
0.14
0.05
0.05
0.05
0.07
0.08
0.08
Microstructure of heat-treated-only 18HGT-grade

steel. Etched with nital. Original magnification: 500
18/8/2008 3:26PM Plate # 0
(normalized)-only steel, it was not possible to

obtain a fatigue resistance value because of the
ductility of the material. Only the yield strength
was determined, and it amounted to 822.8 MPa
on specimens.
The static bending test carried out on carbonitrided specimens is shown in Fig. 14. In this
case, it was not possible to determine the yield
strength, and only the relative bending strength
was established as 2289.8 MPa.
In order to obtain a full characteristic of the

properties of surface-enhanced materials and of
the reference version, a static tensile test was
performed. The average value of tensile strength
Rm for specimens without a diffusion case was
599.8 MPa, and yield strength was determined
as Re = 430 MPa. Similarly, as in the case of the
technological bending test, it was not possible to
determine the yield strength for the carbonitrided material (Fig. 15).
An analysis of static test results delivers new
data. The tensile plot for the carbonitrided
material is characteristic of brittle materials.
There is no necking and no elongation of the
specimen. Similar behavior was noted when the
bending strength test was performed. In neither
case was it possible to determine the yield
strength.
Based on fatigue tests for rotational bending,
which were performed on specimens made from
carbonitrided and heat treated (normalized)only material, it was possible to determine the
coefficient of surface strengthening, that is, the
ratio of m = s 1 (with diffusion case) to s 1
(with no case). For 18HGT steel after carbonitriding, this coefficient was 2.48.
With the aid of results obtained in static and
fatigue tests for the case of two- and one-side
Fig. 12
Microstructure of carbonitrided case on specimen

made of carbonitrided 18HGT-grade steel. Etched
with nital. Original magnification: 500
Fig. 13
Microstructure of core of specimen made of carbonitrided 18HGT-grade steel. Etched by nital. Original magnification: 500
Fig. 14
Plot of static bending test of carbonitrided 18HGTgrade specimen
Fig. 15
Plot of static tensile test of carbonitrided 18HGTgrade specimen
pg 249
18/8/2008 3:26PM Plate # 0
pg 250
bending, simplified Smith curves were plotted

for the heat-treated-only material (Fig. 16) and
for the carbonitrided material (Fig. 17). To plot
the chart, values of unlimited fatigue resistance
were used from the Wohler curves. The upper
limit of the chart for the heat-treated-only
material is the yield strength obtained from the
bending test.
The Smith curve for materials after thermochemical treatment differs from that for the
reference heat-treated-only material, because its
upper limit is determined by the bending yield
strength, Rg (Fig. 17). The designer of the
component, basing his design on the presently
available tables containing data of the ultimate
properties of the steel after hardening and fatigue properties for alternating stresses, creates a
design that consumes large amounts of material.
As the result of using values of s 1 taken from
catalogues, the strength of the assembly is
compromised. The values for fatigue resistance,
sgR, are much higher, which can be seen from
the Smith plot (Ref 11).
Based on the results of fatigue resistance tests
for carbonitrided 18HGT steel, shown for comparison in Table 6, it can be concluded that
carbonitriding ensures good strength properties.
Summarizing the results of the tests presented
in the form of Smith plots, the designer, using
the characteristics of the steel achieved after

thermochemical treatment, will be able to
reduce the amount of material used and to
choose an optimal technological version, thus
ensuring high parameters and longer life of the
designed component.
Summary
Testing of properties of structural materials
is essential for development and manufacturing
Fig. 17
Simplified Smith plot for 18HGT-grade steel after

carbonitriding
Table 6 Comparison of fatigue properties of

18HGT-grade steel after different types of thermal
and thermochemical treatment
Heat Treatment
Strength parameter
Fig. 16
Simplified Smith plot for 18HGT-grade steel after

normalizing
Yield strength
(Re), MPa
Tensile strength
(Rm), MPa
Relative elongation
(A5), %
Necking (Z), %
Bending yield strength
(Rg0.2), MPa
Bending strength (Rg),
MPa
Bending fatigue limit
(s 1), MPa
Surface-strengthening
coefficient, m
Normalizing
Carburizing
Carbonitriding
...
...
1302.3
1303.2
30.45
...
...
70.06
822.8
...
...
...
...
...
2604.1
2889.8
358.0
745.0
887.1
430
599.8
2.08
2.48
18/8/2008 3:26PM Plate # 0
as well as for the practice of design and

construction of machines. A large number of
assessment criteria for materials properties, as
well as methods and testing equipment, stem
from the multitude of different mechanisms for
failure of structural materials from which components, assemblies, and whole machines are
made. Among the most important of these are
failures due to the action of static loading. There
is, however, an entire spectrum of different types
of dynamic loading, as well as volume fatigue,
impact loading, contact stresses, destruction by
wear of mating surfaces in friction, and others
that affect the performance of engineering
components.
Processes connected with the calculation and
design of machine components and assemblies
call for a database of materials properties.
Modern computational methods and their constant development force the necessity of determining an ever growing number of parameters
that describe the properties of materials. A good
example of this is the attempt to describe the
mechanism of material failure through the
coefficient of stress intensity (KIC), both when
the failure takes place under variable loads
(fatigue) as well as when it occurs due to wear in
processes involving friction. Computation of the
life of machine components in conditions of
variable loading calls for information not only
about the value of the fatigue limit but also about
the angle of inclination of the straight line in the
range of limited fatigue strength, and also about
the parameters of the bend point of the fatigue
curve.
Modern structural material does not need to
be homogeneous throughout its section. A great
number of steels, plastics, and other metallic
materials call for enhancement of the surface,
due to the constant quest for decreasing material
and energy consumption as well as increasing
properties. Surface layers are, in the majority of
cases, superficially hardened layers formed by
thermal and thermochemical treatment or other
enhancement technologies, such as surface work
hardening and anticorrosion coatings.
Testing of materials properties after heat
treatment shows that the achievement of desired
service properties is connected with the appropriate selection of parameters not only in the
final thermochemical treatment but also in the
prior volume heat treatment. In the case of
nitriding of machine components, this technology is usually preceded by quenching and tempering at a minimum temperature 20 C above
the subsequent nitriding temperature. Initial

hardening by quenching and tempering is critical to core hardness and to the properties of the
nitrided case and affects the fatigue resistance of
the material after nitriding.
An analysis of fractured surfaces of nitrided
specimens, exposed to service in conditions of
rotational bending, revealed that the weakest
location on the specimen cross section is the
zone of transition of the nitrided case to core.
The method of designing surface cases enables
an explanation of the root cause of fatigue crack
initiation under the nitrided case. The fatigue
limit in the cross section of the specimen was
described as a function of microhardness and
residual stresses. The initiation of fatigue cracks
takes place in the location where stresses from
extraneous sources exceed the value of the
fatigue limit, which is obviously in agreement
with conditions of strength. The favorable effect
of core hardness on fatigue resistance was
observed.
The investigations showed good fatigue
properties of cases obtained by carbonitriding,
as well as a lowering of ductility of these cases.
Results of investigations, presented in the form
of Smith plots, confirm the necessity of further
pursuing investigations in this field. In designing
practice, the application of obtained results is the
least expensive method of lowering material
consumption and enhancing the life of machine
components.
It was established that fatigue resistance is
significantly affected not only by compressive
stresses but also by tensile stresses.
Among the parameters describing the state of
residual stresses, the distribution of stresses was
of more significance, followed by the value of
the residual stresses.
Models of surface layers described in literature are difficult to implement in industrial
practice. There appears to be a need for the
creation of such a model of the surface layer.
This could be described by parameters that can
be used in strength calculations and that would
allow its application in instances of different
types of extraneous loading, depending on the
type of service of the component. This model
could become the basis for predicting the state of
the surface layer, based on required usable
properties of machine components. Work on
such a model is carried out in two directions:
Based on experimental description of the

state of the surface layer through hardness
pg 251
18/8/2008 3:26PM Plate # 0
traverses, residual-stress distributions, and

distributions of element concentrations, for
example, carbon and nitrogen
Based on a description of the material
through the theory of elasticity and solving
constitutive equations by numerical methods
Surface layer design for the criterion of fatigue failure is based on a comparison of the local
fatigue resistance with local stresses occurring at
critical locations in the investigated component.
Contemporary machines and designs should
be characterized by required life and reliability,
featuring a sufficient life between overhauls,
depending on the type of service, while at the
same time fulfilling the requirements for ecology and ergonomics. Such parameters should
be attained concurrently with a reduction of
material and energy consumption during manufacture and service. This task may be achieved
only when modern computational methods are
implemented along with modern technology and
proper service conditions at each stage of the
product life, that is, study phase, design, manufacture, service, and recycling.
The implemented computational methods
enable the design of products according to
strength criteria, somewhat less often according
to tribological criteria, and least often pertaining
to corrosion.
Contemporary machine components and
assemblies are subjected in service to the joint
interaction of strength, tribological, and corrosion hazards. On the other hand, the implemented
computational methods enable the design of
products with one selected mode of failure.
In the construction of machine components,
there are many parts (crankshafts, threaded
joints, springs) that are concurrently exposed to
different types of failure hazards during service:
mechanical, tribological, or corrosive. Similar
elements of construction (bridges, masts, cables,
earth-moving and mining machines) are exposed to concurrent hazards of fatigue-type
stresses and corrosion.
Classical strength or tribological calculations
do not take into account the factor of time.
During service, due to the processes of fatigue,
tribological, or corrosive deterioration, there
occurs a change in the properties of the system
being evaluated. Tribological and corrosive
processes cause a change in the geometry and
surface condition of the component. This, in turn,
causes a change in the state of stresses in working
systems, affecting their life and reliability.
Therefore, the development of failure criteria,

taking into account the joint effect of an accumulation of damage due to the working of
alternating loads, wear by friction, and the action
of corrosion, is a very important task, because
the determination of the criteria for failure will
enable proper selection of surface layers for the
given service condition.
REFERENCES
1. A. Nakonieczny, Podwyzszenie wytrzymaosci zmeczeniowej czesci maszyn

przez obrobke cieplna i powierzchniowa
obrobke plastyczna (Enhancement of Fatigue Strength through Heat Treatment and
Surface Work Hardening), Proc. XXIV
Seminar IMP, XI on Metallurgy and Heat
Treatment, Oct 2324, 1984 (Warsaw), IMP
(translated from Polish)
2. T. Babul, A. Nakonieczny, and J. Tacikowski, Wpyw umocnienia podoza na
wytrzymaosc zmeczeniowa azotowanej
stali 40HM (The Effect of Core Strengthening on the Fatigue Resistance of Nitrided
40HM Grade Steel), Proc. III Polish Scientific Conference on Surface Treatment,
Czestochowa-Kule, Politechnika Lodz,
1996 (translated from Polish).
3. A. Nakonieczny and J. Tacikowski, Analiza
pekania zmeczeniowego stali azotowanych
(An Analysis of Fatigue Fracturing of
Nitrided Steels), Proc. First Polish
Scientific Conference on Modern Technology in Surface Engineering, Sept
1994, Politechnika Lodz (translated from
Polish)
4. A. Nakonieczny, The Effect of Residual
Stresses and Hardness on Fatigue Behavior
of Surface Treated Materials, Proc. MATTEC 91, Technology Transfer Series,
A. Niku-Lari, Ed., 1991
5. B. Winderlich, Das Konzept der lokalen
Deuerfestigkeit und seine Anwendung auf
martensitische Randschichten, in bensondere Laserhartungsschichten, Mater. wiss.
Werkst. tech., Vol 21, 1990, p 378389 (in
German)
6. V.P. Kogaev, N.A. Machutov, and A.P.
Gusenkov, Rascety detalej masin i konstrukcij na procnost i dolgovecnost, Masinostroenie, Moskva, 1985, p 150182
(in Polish)
pg 252
18/8/2008 3:26PM Plate # 0
7. J. Tacikowski and A. Nakonieczny, Report

114,01,0163, IMP, Warsaw, 1992
8. A. Nakonieczny, Dissertation, Russian
Academy of Science, Moscow, 1991
9. W. Olszanski, More Important Problems
Pertaining to the Austenitic Carbonitriding
Process, XVI Seminar, VII on Metallurgy
and Heat Treatment Book 2, IMP, Warsaw,
1977
10. J. Wyszkowski, Nowoczesne tendencje w

zakresie naweglania i wegloazotowania
gazowego, (Modern Trends in the Field
of Gas Carburizing and Carbonitriding),
IMP, Warsaw, 1974 (translated from Polish)
11. W. Olszanski, I. Sukowski, J. Tacikowski,
and J. Zysk, Obrobka cieplno-chemiczna,
(Thermochemical Treatment) Book 5, IMP,
Warsaw, 1979
pg 253
18/8/2008 3:28PM Plate # 0

DOI: 10.1361/faht2008p255

Steel Heat Treatment Failures due to

Quenching
G.E. Totten, Associacao Instituto Internacional de Ciencia and
Portland State University
QUENCHING is one of the more important

heat treating processes, because it is so closely
related to dimensional control requirements and
control of residual stresses. Quenching is often
attributed to many distortion and cracking problems, whether the quenching process is the
actual root cause or not. Approximately 20% of
the problems in heat treating relate to heating
processes, while as much as 80% of the problems relate to cooling processes. This chapter
provides an overview of the fundamental
material- and process-related parameters of
quenching on residual stress, distortion control,
and cracking. This overview is followed by
various selected case histories of failures
attributed to the quenching process.
Phase Transformations During Heating

and Quenching
Properties such as hardness, strength, ductility, and toughness are dependent on the
microstructural products that are present in steel.
Typically, the first step in the transformation
process is to heat the steel to its austenitizing
temperature. The austenitized steel is then
cooled rapidly to avoid the formation of pearlite,
which is a relatively soft transformation product,
and to maximize formation of martensite, a
relatively hard transformation product, and to
achieve the desired as-quenched hardness.
The most common transformation products
that may be formed in quench-hardenable steels
from austenite are, in order of formation with
decreasing cooling rate, martensite, bainite,
pearlite, ferrite, and cementite. The formation of
these products and the proportions of each are
dependent on the time and temperature cooling
history of the particular alloy and the elemental
composition of that alloy. The transformation

products formed are typically illustrated with the
use of transformation diagrams that show the
temperature-time dependence of the microstructure formation process for the alloy being
studied. Two of the most commonly used transformation diagrams are the time-temperature
transformation and continuous cooling transformation diagrams.
Time-temperature transformation (TTT)
diagrams, also called isothermal transformation diagrams, are developed by heating small
samples of steel to the austenite transformation
temperature, followed by rapid cooling to a temperature intermediate between the austenitizing
and the martensite start (Ms) temperature, and
then holding for a fixed period of time until the
transformation is complete, at which point the
transformation products are determined. This is
done repeatedly until a TTT diagram is constructed, such as that shown for an unalloyed
steel (AISI 1045) in Fig. 1 (Ref 1). The TTT
diagrams can only be read along the isotherms.
Continuous Cooling Transformation Diagrams. Alternatively, a given steel may be
continuously cooled from the austenitizing temperature at different specified rates. The proportion of transformation products formed after
cooling to various temperatures intermediate
between the austenitizing temperature and the
Ms temperature is used to construct a continuous
cooling transformation (CCT) diagram, such as
the one shown for an unalloyed carbon steel
(AISI 1045) in Fig. 2 (Ref 1). The CCT curves
provide data on the temperatures for each phase
transformation, the amount of transformation
product obtained for a given cooling rate with
time, and the cooling rate necessary to obtain
martensite. The critical cooling rate is dictated
by the time required to avoid formation of
pg 255
18/8/2008 3:28PM Plate # 0
Martensite
Fig. 1
Time-temperature transformation diagram of an unalloyed steel containing 0.45% C. Austenitizing temperature: 880 C.
Source: Ref 1
Martensite
Fig. 2
Continuous cooling transformation diagram of an unalloyed steel containing 0.45% C. Austenitizing temperature: 880 C.
Source: Ref 1
pg 256
18/8/2008 3:28PM Plate # 0
Steel Heat Treatment Failures due to Quenching / 257
Fig. 3
Crystal structures. (a) Austenite, face-centered cubic. (b) Ferrite, body-centered cubic. (c) Martensite, body-centered tetragonal. Source: Ref 1
pearlite for the particular steel being quenched.

As a general rule, a quenchant must produce a
cooling rate equivalent to or faster than that
indicated by the nose of the pearlite transformation curve to maximize the martensite transformation product (Ref 1). The CCT diagrams
can only be read along the curves of different
cooling rates, and a continuous cooling curve
can only be superimposed on a CCT but not on a
TTT diagram.
Metallurgical Crystal Structure. When
steel is slowly cooled, it undergoes a crystal
structure (size) change as it transforms from a
less densely packed (face-centered cubic) austenite to a more densely packed body-centered
cubic structure of ferrite. At faster cooling rates,
the formation of ferrite is suppressed, and martensite, which is an even less densely packed
body-centered tetragonal structure than austenite, is formed. Illustrations of these crystal
structures are provided in Fig. 3 (Ref 1). This results in a volumetric expansion at the Ms temperature as shown in Fig. 4 (Ref 1).
Figure 5 shows that the crystal lattice of austenite expands with increasing carbon content
(Ref 2). It has been reported that typically when
a carbide-ferrite mixture is converted to martensite, the resulting expansion due to increasing
carbon content is approximately 0.05 mm/mm
(0.002 in./in.) at 0.25% C and 0.18 mm/mm
(0.007 in./in.) at 1.2% C (Ref 2). The fractional
increase in size when austenite is converted
to martensite is approximately 0.36 mm/mm
(0.014 in./in.) for eutectoid compositions. This
illustrates the effect of carbon structure and steel
transformation on residual stresses and distortion leading to dimensional changes.
Estimation of Volumetric Change due to
Steel Transformation upon Quenching.
Various microstructures are possible upon
Fig. 4
Steel expansion and contraction upon heating and

cooling. Source: Ref 1
quenching of steel, and the potential microstructural transformations that are possible for a
given steel are illustrated by their CCT or TTT
diagrams. Furthermore, dimensional changes depend on carbon content and the microstructural
transformation product formed. Table 1 summarizes the atomic volumes of various microstructural components as a function of carbon
content (Ref 3). Table 2 provides an estimate of
volumetric changes as a function of carbon
pg 257
18/8/2008 3:28PM Plate # 0
pg 258
content for various metallurgical transformations (Ref 4, 5).

Thelning reported that volumetric expansion
occurring as a result of quenching could be
estimated from (Ref 6):
DV=V 100=(100 Vc Va ) 1:68C
+Va ( 4:64+2:21C)
(Eq 1)
where (DV/V) 100 equals the percentage

change in volume, Vc equals the percentage by
volume of undissolved cementite, (100 Vc
Va) equals the percentage by volume of martensite, Va equals the percentage by volume of
austenite, and C equals the percentage by weight
of carbon dissolved in austenite and martensite.
Berns reported that if the value of (DV/V) is
known or can be computed, internal stresses
that are developed in a part due to temperature differences (DT) arising from either
one-dimensional heating or cooling could be

estimated from (Ref 7):
s=E e=E 1/3 (DV=V)=E a DT
(Eq 2)
where E (modulus of elasticity) = 2 105

N/mm2 and a (coefficient of thermal expansion) = 1.2 10 5. Relative volume changes
due to phase transformation are illustrated in
Fig. 6 (Ref 7).
Kunitake and Susigawa (Ref 8) reported that
the tendency for cracking decreases as the start
of the martensite transformation temperature
(Ms) increases. The Ms temperature was approximated from:
Ms ( C)=521 353C 225Si 24:3Mn
27:4Ni 17:7Cr 25:8Mo
(Eq 3)
The correlation between the occurrence of

quench cracks and Ms temperature is shown
in Fig. 7. A similar study produced a poor
Table 2 Volumetric changes with various steel
transformations
Steel transformation
Pearlite?austenite
Austenite?martensite
Austenite?acicular lower bainite
Austenite?feathered upper bainite
Volumetric change
4.64+2.21 C(a)
4.64 0.53 C(a)
4.64 1.43 C(a)
4.64 2.21 C(a)
(a) Percent carbon. Source: Ref 4, 5
Fig. 5
Carbon content versus lattice parameters of (retained)

austenite and martensite at room temperature. a at
the top of the graph is the lattice parameter of face-centered cubic
austenite. a and c in the lower half of the graph are the two lattice
parameters of tetragonal martensite. The ratio of c/a for martensite
as a function of carbon content is also given. Source: Ref 2
Table 1 Atomic volume of various

microstructural constituents of ferrous alloys
Phase
Ferrite
Cementite
Ferrite+carbides
Pearlite
Austenite
Martensite
(a) Percent carbon. Source: Ref 3
3
Apparent atomic volume, A
11.789
12.769
11.786+0.163 C(a)
11.916
11.401+0.329 C(a)
11.789+0.370 C(a)
Fig. 6
Specific volume (DV/V) of carbon steels relative to

room temperature. Source: Ref 7
18/8/2008 3:28PM Plate # 0
pg 259
correlation between grain size and quench

cracking, as shown in Fig. 8 (Ref 8).
Kunitake and Sisigawa (Ref 8) developed a
relationship to interrelate the combined effect
of both carbon content and elemental composition on cracking propensity. This was designated
as the carbon equivalent (Ceq), and it is calculated by:
Ceq =C+Mn=5+Mo=5+Cr=10+Ni=10
(Eq 4)
Figure 9 shows a good correlation between

the carbon equivalent and steel cracking. In
general, steels are classified as crack sensitive
if the Ceq value is greater than 0.52 to 0.55%
(Ref 8).
Another measure of cracking tendency is the
difference in the start and finish temperatures
of martensite formation (Ms Mf) (Ref 9).
Fig. 7
A summary of the Ms and Mf values for some

common steels is provided in Table 3.
The correlation between cracking sensitivity
and the transformation temperature range is due
in part to the low Mf caused by high-carbon steels
(which expand more) and to the fact that wide
transformation ranges may result in cracking of
the brittle untempered martensite formed at
higher temperatures in the transformation range.
Fujio et al. (Ref 10) showed that the volumetric expansion caused by martensite formation can be estimated from the maximum
cooling rate in a particular type of steel, as
shown in Fig. 10. Similar correlations were
evaluated for both cooling time and cooling rate
at the Ms temperature. However, these correlations were dependent on the cross-sectional
size and thus could not be used for gears or
other parts with complex shapes. Volumetric
expansion can be estimated for various crosssectional sizes by a correlation between the
volume fraction of martensite versus the cooling
Relationship between quench cracking frequency and

martensite start (Ms) temperature. Source: Ref 8
Fig. 9
Relationship between carbon equivalent (Ceq) and

quench cracking frequency. Source: Ref 8
Table 3 Martensite start (Ms) and martensite

finish (Mf) values for selected steels
AISI No.
Fig. 8
Relationship between quench cracking frequency and

austenitic grain size. Source: Ref 8
1065
1090
1335
3140
4130
4140
4340
4640
5140
8630
8695
9442
Austenitizing
temperature, C
Ms, C
Mf , C
815
885
845
845
870
845
845
845
845
870
845
860
275
215
340
330
375
340
290
340
330
365
135
325
150
80
230
225
290
220
165
255
240
280
...
15
18/8/2008 3:28PM Plate # 0
rate at the Ms temperature for the steel (Fig. 11).

It is well known that retained austenite
can substantially affect distortion. Geller and
Brimene (Ref 11) published a nomogram that can
Fig. 10
Fig. 11
be used to predict dimensional changes caused

by the total carbon concentration in the martensitic transformation product and the amount of
retained austenite. Steel chemical compositions
Relationship between maximum cooling rate and volumetric fraction of martensite. Source: Ref 10
Relationship between maximum cooling rate and the martensite start (Ms) temperature and volumetric fraction of martensite.
Points in the same curve are related to different positions in the bar and therefore with the degree of martensitic transformation. Source: Ref 10
pg 260
18/8/2008 3:29PM Plate # 0
pg 261
Table 4 Steel chemical compositions listed in Fig. 12

Russian
steel
designation(a)
Composition, wt%
Mo
Ti
Co
Cu
U8
0.750.84 0.170.33 0.170.33 0.15 max
...
...
...
...
...
...
...
KhVG
0.901.05 0.100.40
...
...
...
...
1.201.60
...
ShKh15
0.951.05 0.170.37 0.200.40 1.301.65
...
...
...
...
...
7KhG2VM
0.680.76 0.200.40 1.802.30 1.501.80
0.30
max
...
0.500.80 0.100.25
...
0.50.9
...
7KhG3V
Kh12M
0.680.76 0.200.40 3.03.5 1.501.80

1.451.65 1.101.40 0.150.45 11.012.5
...
...
...
0.100.25
0.400.60 0.150.30
...
...
0.50.9
...
...
...
0.30
max
0.25
max
0.30
max
...
...
4Kh5V2FS
0.350.45 0.801.20 0.150.40 4.505.50
...
...
0.600.90
...
1.602.20
...
3Kh2V8F
0.300.40 0.150.40 0.150.40 2.202.70
...
...
...
...
7.508.50
...
4Ch5W2FS
0.350.40 0.801.20 0.150.40 4.505.50
0.30 max
0.600.90
...
0.300.40 0.150.40 0.150.40 2.202.70
0.50 max
0.200.50
7.808.50
...
7.0
5.0
...
...
0.03
max
0.03
max
...
...
1.602.20
3Ch2W8F
0.35
max
0.35
max
...
...
14.0
...
25.0
9.0
0.03
max
0.03
max
0.02
max
0.03
max
...
0.03
max
0.03
max
0.03
max
0.03
max
0.03
max
...
...
0.03
max
0.03
max
0.03
max
0.03
max
...
0.03
max
0.03
max
0.03
max
0.03
max
0.03
max
...
...
R14M7K25
N18K9M5
1.0
1.0
Si
...
...
Mn
Cr
0.81.10 0.901.20
...
...
...
...
Ni
0.30
max
0.30
max
0.30
max
0.03
max
...
...
(a) These compositional data were provided by Dr. Dmitry Wainstein, Surface Phenomena Research Group, Physical Metallurgy Institute, CNIICHERMET, Moscow,
Russia.
are found in Table 4. The following comments

will assist in interpreting the nomogram shown in
Fig. 12:

The shaded line represents zero distortion.

Steels with martensitic carbon contents and
retained austenite levels falling on the line
will exhibit essentially no distortion.
Martensitic steels with carbon contents and
retained austenite levels that fall bellow the
shaded line will exhibit shrinkage upon
quenching.
Martensitic steels with carbon contents and
retained austenite levels that fall above the
shaded line will exhibit expansion.
This nomogram was developed for various

construction and tool steels. Therefore, it should
be used with caution for other steel grades (e.g.,
high-speed tool steels).
Basic Distortion Mechanism. Shape and
volume changes occurring during heating and
cooling can be attributed to three fundamental
causes (Ref 12):

Residual stresses will cause shape change

when they exceed the yield strength of the
material.
Stresses caused by differential expansion
due to thermal gradients will increase with
the thermal gradient and cause plastic
deformation as the yield strength is exceeded.
Volume changes due to transformational

phase changes will be contained as residualstress systems until the yield strength is
exceeded.
When parts are heated during heat treatment,

a thermal gradient exists across the cross section
of the component. If a section is heated so that a
portion of the component becomes hotter than
the surrounding material, the hotter material
expands and occupies a greater volume than the
adjacent material and will thus be exposed to
applied stresses that will cause a shape change
when they exceed material strength. These
movements can be related to heating rate and
section thickness of the component.
Volume Changes During Phase Transformations. When a steel part is heated, it transforms to austenite with an accompanying
reduction in volume. When it is quenched, the
structure transforms from austenite to martensite, and its volume increases. If these volume
changes cause stresses to be set up that are constrained within the strength of the material, a
residual-stress system is created. If the stresses
cannot be contained, material movement will
occur, which will cause cracking under extreme
conditions. The expansion is related to the composition of the steel. Figure 13 shows the relative
volume increase of two steels as a function of
austenitizing temperature and specimen dimensions (Ref 13).
18/8/2008 3:29PM Plate # 0
Fig. 12
Changes in linear dimensions during quenching relative to carbon concentrations in martensite and retained austenite.
Source: Ref 11
Fig. 13
Volume increase of 90MnV8 and 15CrV6 steels as a function of austenitizing temperature and specimen dimensions. Source:
Ref 13
pg 262
18/8/2008 3:29PM Plate # 0
While each of these phenomena is a wellknown physical change, the situation is made
more complex when all three events occur
simultaneously. In addition, other events, such as
heating rate, quenching, and inconsistent material composition, further complicate the process.
Relief of Residual Stresses. If a part has
locked-in residual stresses, these stresses can be
relieved by heating the part until the lockedin stresses exceed the strength of the material.
A typical stress-strain curve obtained from a
tension test is shown in Fig. 14 (Ref 12). Initial
changes in shape are elastic, but under increased
stress, they occur in the plastic zone and are
permanent. Upon heating, the stresses are gradually relieved by changes in the shape of the
part due to plastic flow. This is a continuous
process, and as the temperature of the part is
increased, the material yield stress decreases, as
shown in Fig. 15 (Ref 14). It is a function not
only of temperature but also of time, since the
material will creep under lower applied stresses.
It is apparent that the stresses can never be
reduced to zero, because the material will always
possess some level of yield strength below which
residual stresses cannot be reduced.
Effect of Materials and Quench Process

Design on Distortion
Quenchant selection and quenching conditions are critically important parameters in
quench system design. For example, one study
compared the distortion obtained with quenching of a 0.4% medium-carbon plain steel bar of
200 mm diameter by 500 mm long in water or
oil from 680 C (Ref 15, 16). The results, shown
in Fig. 16(a and c), show essentially equivalent
variation in diameter and length with both
cooling processes, which was due to thermal
strains within the steel. Interestingly, the wellknown diameter variations at the end of the bar,
known as the end effect, were observed, which
is attributable to heat extraction from both the
sides and ends of the bar (Ref 1).
If the same steel bars of the same dimensions
are heated to 850 C to austenitize the steel and
then are quenched in water or oil, the results
shown in Fig. 16(b and d), respectively, are
obtained (Ref 15, 16). Considerably greater
dimensional variation and lengthening of the bar
(for the oil quench) was obtained due to both
thermal and transformational strains within the
steel.
Thuvander and Melander modeled the
dimensional changes of a 70 mm steel (0.15%
C, 1% Mn, 0.75% Cr, 0.85% Ni) cube after
austenitizing and then quenching in water and
oil (Ref 15, 17). The results of this work are
shown in Fig. 17. They show that the edges and
faces shrink (becoming concave) and the effect
is greater when quenched in water than when
quenched in oil (Ref 1).
Various factors may affect distortion and
growth of steel during heat treating. These
include component design, steel grade and
condition, machining, component support
and loading, surface condition, heating and
atmosphere control, retained austenite, and the
quenching process (Ref 18).
Component Design
One of the overwhelming causes of steel
cracking and unacceptable distortion control is
Fig. 14
Various features of a typical stress-strain curve

obtained from a tension test. Source: Ref 12
Fig. 15
Variation of yield strength with temperature for three

generic classes of steel. Source: Ref 14
pg 263
18/8/2008 3:29PM Plate # 0
component design. Poor component design

promotes distortion and cracking by accentuating nonuniform and nonsymmetrical heat
transfer during quenching. Component design
characteristics that are common to distortion and
cracking problems include (Ref 19, 20):

Parts that are long (L) with thin (d) cross

sections. Long and thin parts are defined as
greater than L = 5d for water quenching,
L = 8d for oil quenching, and L = 10d for
austempering, where L is the length of the
parts, and d is the thickness or diameter.
Parts that possess large cross-sectional areas
(A) and are thin (t), which are defined as
A = 50t
Parts that exceed these dimensions must often

be straightened or press quenched to maintain
dimensional stability (Ref 20). If possible,
Fig. 16
materials with sufficient hardenability should be

oil or salt quenched.
Design symmetry is also an important variable to minimize distortion. For example, the
unsymmetrical gear design shown in Fig. 18(a)
may typically undergo distortion, as shown in
Fig. 18(b) (Ref 19). (The load on a gear tooth
increases by the 4.3 power of the taper, Ref 19).
The solution to the gear design problem shown
in Fig. 18 is to provide greater symmetry, as
shown in Fig. 19. If this is not possible, press
quenching or tooth-by-tooth induction hardening may be the only solutions (Ref 19, 20).
Another common design problem is parts with
holes, deep keyways, and grooves. One illustration of this problem is hardening of a shaft
with a lubrication cross hole, as illustrated in
Fig. 20 (Ref 19). Preferred alternative designs
are also shown in Fig. 20. If a radial cross hole is
Dimensional variation of a medium-carbon (0.4%) steel bar (200 mm diam by 500 mm) after the indicated heat treatments.
These bars were quenched vertically with one end down (marked 0 in the figure). (a) and (c) show no transformation, only
thermal strain after water quenching from 680 C. (b) and (d) show thermal and transformation strains after quenching from 850 C.
Source: Ref 1
pg 264
18/8/2008 3:29PM Plate # 0
mandatory, the use of a carburized steel with oil

quenching would be preferred.
Kern and Suess have reported that the size of
the tapped holes can be maintained by the
insertion of SAE grade 8 set screws or bolts
(Ref 21). Prequenching can be used to control
the taper of plain holes during heat treatment.
Some hole distortion problems may require oil
quenching (conventional or hot oil) or austempering.
The distortion encountered when quenching a
notched part, such as a shaft with a milled slot, is
illustrated in Fig. 21 (Ref 20). In this case,
nonuniform heat transfer results. The metal
within the notch is affected by the shrinkage of
the metal around it due to slower cooling within
the slot, caused by vaporization of the quenchant. Therefore, upon cooling, the metal on the
side with the shaft is too short, pulling the shaft
out of alignment. A general rule for solving such
quench distortion problems is that the short side
is the hot side, which means that the inside of the
bowed metal was quenched more slowly than
the opposite side (Ref 20).
Flat plates are also susceptible to distortion
upon quenching. If the material is flat and stress
free, round or nearly square, and free of
decarburization, Kern and Suess have reported a
guide (Table 5) to maintain a flat surface (within
0.025 mm, or 0.001/in., of size) if parts are
racked and quenched edgewise (Ref 21). Parts
exceeding these limits may require press
quenching.
Fig. 19
Design solutions to the distortion problem shown in

Fig. 18. Source: Ref 19
Fig. 17
Dimensional changes in a 70 mm steel (0.15% C,

1% Mn, 0.75% Cr, 0.85% Ni) bar after austenitizing
and then quenching in water or oil. Source: Ref 1
Fig. 18
Schematic of a gear that is difficult to harden without

the distortion shown. Source: Ref 19
Fig. 20
Design solutions to the quench-cracking problem

often encountered in shaft hardening over a cross
hole. Source: Ref 19
pg 265
18/8/2008 3:29PM Plate # 0
pg 266
Steel Grade and Condition. Although

quench cracking is most often due to nonuniform heating and cooling, material problems
may also be encountered. Some typical material
considerations include (Ref 19):

The compositional tolerances should be

checked to assure that the alloy is within
specification.
Some alloys are particularly problematic.
For example, some steel grades must be
water quenched when the alloy composition
is on the low side of the specification limit.
Conversely, if the alloy composition is on
the high side, cracking is more common.
Steel grades that exhibit this problem
include 1040, 1045, 1536, 1541, 1137, 1141,
and 1144. As a rule, steels with carbon
contents and hardenability greater than 1037
are difficult to water quench (Ref 19).
Some steel grades with high manganese are
prone to microsegregation of manganese and
gross segregation of chromium and are
prone to cracking. These include 1340, 1345,
1536, 1541, 4140, and 4150. If possible, it is
often a good choice to replace the 4100
Fig. 21
Distortion often encountered when quenching a

notch. Source: Ref 20
series with the 8600 series of steels (Ref 20).

Dirty steels, those containing greater than
0.05% S, such as 1141 and 1144, are more
prone to cracking. The reasons include:
greater alloy segregation in dirty steels leads
to alloy-rich and alloy-lean regions; there are
typically more surface seams that act as
stress raisers with dirty steels; and steels
with higher sulfur levels are often manufactured to coarse-grain practice for improved machinability, which also imparts
greater brittleness and propensity for cracking.
Decarburization of up to 0.064 mm/
1.59 mm (0.0025 in./1/16 in.) diameter may
be present.
It is well known that cracking propensity

increases with carbon content. Therefore, the
carbon content of the steel is one of the determining factors for quenchant selection. Table 6
summarizes some steel mean carbon content
concentration limits for water, brine, or caustic
quenching (Ref 22).
Regions containing high concentrations of
coarse carbide microstructure as a result of
improper forging may become the initiation
point for subsequent quench cracking, particularly with parts of complex shape (Ref 23). It is
important to provide a sufficient forging reduction ratio to allow the carbide formation to
become fine and uniform (Ref 24).
Since part manufacture, such as gear production, often requires machining, the condition
of the steel that is going to be machined is
critically important. Some workers have recommended that normalized and subcriticalannealed steel is the ideal condition (Ref 18).
Subcritical annealing is performed to relieve
stresses incurred during normalization without
softening or homogenizing the steel. The subcritical annealing process reduces the carbon
content and alloy carbide content in the austenite, allowing the production of more lath
Table 6 Suggested carbon content limits for
water, brine, and caustic quenching
Hardening method/shapes
Table 5 Guide to maintain a flat surface

Quenchant
Ratio (max) of
perimeter/thickness
Water
Oil
Austemper
Gas
30
80
125
150
Carbon, max%
Furnace hardening
General use
Simple shapes
Very simple shapes, e.g., bars
0.30
0.35
0.40
Induction hardening
Simple shapes
Complex shapes
0.50
0.33
Source: Ref 22
18/8/2008 3:29PM Plate # 0
martensite in the microstructure, which provides

higher fracture toughness and higher impact
toughness (Ref 23).
Machining. Material
removal
during
machining can result in high residual-stress
levels and ultimately unacceptable distortion
(Ref 18). When excessive machining stresses are
imparted, the process may require modification
to include a rough machining, then stress
relieving, followed by fine machining.
Component Support and Loading. Many
parts may sag and creep under their own weight
when heat treated, which is an important cause
of distortion. An example of a component that is
susceptible to such distortion is a ring gear. Dimension limits by which ring gears are classified
are provided in Fig. 22 (Ref 18). (A general
Fig. 22
Dimensions of a ring gear shape. Shape limitation:

length/wall thickness, j1.5; inside diameter/outside
diameter (ID/OD), 40.4. Minimum wall thickness
(WT) is
p
5
defined by: WT i2.25 module+0.4 mod L OD3 .
Source: Ref 18
Fig. 23
Classification of shapes. Source: Ref 18
dimensional classification of various distortionsensitive shapes is provided in Fig. 23, Ref 18).
Proper support when heating is required to
minimize out-of-flatness and ovality problems,
which may result in long grinding times,
excessive stock removal, high scrap losses, and
loss of case depth (Ref 18). To achieve adequate
distortion control, custom supports or press
quenching may be required.
Pinion shafts, as defined in Fig. 23, are susceptible to banding along their length if they are
improperly loaded into the furnace, as shown in
Fig. 24 (Ref 18). When this occurs, the pinion
shafts must then be straightened, which will add
to the production cost.
Surface Condition. Quench cracking may
be due to various steel-related problems that are
only observable after the quench, but the root
cause is not the quenching process itself (Ref
25). Many of these problems have been
reviewed earlier and include prior steel structure, stress raisers from prior machining, laps
and seams, alloy inclusion defects, grinding
cracks, chemical segregation (bonding), and
alloy depletion. In this section, three surface
condition-related problems that may contribute
to poor distortion control and cracking are discussed: tight scale formation, decarburization,
and the formation of surface seams or nonmetallic stringers.
Tight scale problems are encountered with
forgings hardened from direct-fired gas furnaces
pg 267
18/8/2008 3:30PM Plate # 0
with high-pressure burners (Ref 20, 22). The

effect of tight scale on the quenching properties
of two steels, 1095 carbon steel and 18-8
stainless steel, is illustrated in Fig. 25 (Ref 23).
These cooling curves were obtained by still
quenching into fast oil. A scale of not more than
0.08 mm (0.003 in.) increases the rate of cooling of 1095 steel as compared to the rate
obtained on a specimen without scale. However,
a heavy scale (0.13 mm or 0.005 in. deep)

retards the cooling rate. A very light scale
(0.013 mm or 0.0005 in. deep) also increased
the cooling rate of the 18-8 steel over that
obtained with the specimen without scale.
In practice, the formation of tight scale will
vary in depth over the surface of the part,
resulting in thermal gradients due to differences
in cooling rates. This problem may yield soft
spots and uncontrolled distortion and is particularly a problem with nickel-containing steels.
Surface oxide formation can be minimized by
the use of an appropriate protective atmosphere.
The second surface-related condition is decarburization, which may lead to increased distortion or cracking (Ref 24). At a given depth
within the decarburized layer, the part does not
harden as completely as it would at the same
point below the surface if there were no decarburization. This leads to nonuniform hardness, which may contribute to increased
distortion and cracking because (Ref 20):

Fig. 24
Fig. 25
Typical pinion shaft distortion due to furnace loading. Source: Ref 18
The decarburized surface transforms at a

higher temperature than the core (the Ms
temperature decreases with carbon content).
This will lead to high residual tensile stresses
at the decarburized surface or a condition of
unbalanced stresses and distortion.
Since the surface is decarburized, it will
exhibit lower hardenability than the core.
This will cause the upper transformation
products to form early, nucleating additional
undesirable products in the core. The
Centerline cooling curves showing the effect of scale on the cooling curves of steels quenched in fast oil without agitation.
(a) 1095 steel. Oil temperature: 50 C (125 F). (b) 18-8 stainless steel. Oil temperature: 25 C (75 F). Test specimens were
13 mm diam by 64 mm long (0.5 by 2.5 in.). Source: Ref 23
pg 268
18/8/2008 3:30PM Plate # 0
decarburized side will be softer than the side

that did not undergo decarburizing, which is
harder. The greater amount of martensite
leads to distortion.The solution to this problem is to restore carbon into the furnace
atmosphere or machine off the decarburized
layer.
The third surface-related condition that may
lead to cracking or material weakening is the
formation of surface seams or nonmetallic
inclusions, which may occur in hot rolled or cold
finished material. The presence of these defects
prevent the hot steel from welding to itself during the forging process, for example, creating a
stress raiser. To prevent this problem with hot
rolled bars, stock should be removed before heat
treatment. Recommendations made earlier by
Kern are provided in Table 7 (Ref 22).
Although not a published standard, Kern has
reported that a seam or nonmetallic depth of
0.025 mm/3.3 mm (0.001 in./0.13 in.) diameter
maximum is usually acceptable for cold finished
bars (Ref 22). If the seam depth is excessive, it is
recommended that the bars be magniflux inspected prior to heat treatment.
Heating and Atmosphere Control. An
important source of steel distortion and cracking
is nonuniform heating and not using the appropriate protective atmosphere. For example, if
steel is heated in a direct-gas-fired furnace with
high moisture content, the load being heated
may adsorb hydrogen, leading to hydrogen
embrittlement and subsequent cracking that
would not normally occur with a dry atmosphere
(Ref 19, 26).
One source of distortion is when a part is in
contact with the furnace hearth during heating,
which may produce sufficient nonuniform temperature distribution within the part. This will
occur because the portion of the part in contact
with the furnace hearth will be heated conductively much faster than the remainder of the
Table 7 Minimum recommended material

removal from hot rolled steel products to prevent
surface seam and nonmetallic stringer problems
during heat treatment
Minimum material removal per side(a)
Condition
Turned on centers
Centerless turned or ground
Nonresulfurized
3% of diameter
2.6%
Resulfurized
3.8% of diameter
3.4%
(a) Based on bars purchased to special straightness, i.e., 3.3 mm in 0.04 m

(0.13 in. in 5 ft) maximum. Source: Ref 22
part surface, which is heated primarily by

radiation. Thus, as the hotter surface tries to expand, it will be restrained by the cooler steel,
leading to a hot upsetting condition and possibly
significant distortion even if quenched uniformly (Ref 21). A similar condition exists if the
tray of gears is placed near radiant tube heaters
or electric heating elements in the furnace wall
and the remainder of the gear surface is heated
by radiation from the roof of the furnace.
Localized overheating is particularly a
potential problem for inductively heated parts
(Ref 4, 26). Subsequent quenching of the part
leads to quench cracks at sharp corners and areas
with sudden changes in cross-sectional area
(stress raisers). Cracking is due to increases of
residual stresses at the stress raisers during the
quenching process. The solution to the problem
is to increase the heating speed by increasing the
power density of the inductor. The temperature
difference across the heated zone is decreased by
continuous heating or scanning of several pistons together on a single bar (Ref 26).
For heat treating problems related to furnace
design and operation, it is usually suggested that
(Ref 19):

The vestibules of atmosphere-hardening

furnaces should be loaded and unloaded with
purging. Load transfer for belt and shaker
hearth furnaces should only occur with
thorough purging to minimize atmosphere
contamination.
If the load being heated in the furnace is
excessively large, either nonuniform heating
over the specified heating cycle or nonuniform cooling when quenched will result. In
such cases, either the production rate can be
increased or smaller loads should be processed.
Retained Austenite. Dimensional changes

may occur slowly or quickly and are due to the
volume composition of the transformation products formed upon quenching. One of the most
important, with respect to residual-stress variation, distortion, and cracking, is the formation
and transformation of retained austenite. For
example, the data in Table 8 illustrate the slow
conversion of retained austenite to martensite,
which was still occurring days after the original
quenching process for the two steels shown (Ref
15, 16). This is particularly a problem when
dimensional control and stability is one of the
primary goals of heat treatment. Therefore,
pg 269
18/8/2008 3:30PM Plate # 0
pg 270
microstructural determination is an essential

component of any distortion control process.
additional general comments regarding quenchant selection include (Ref 4, 20):

Quenching Process
Other than component design, the quenching
process itself is one of the most frequently
encountered problems in heat treating. When
designing a quenching process, it is important to
consider quenchant selection, quench severity,
and quench uniformity.
Quenchant Selection and Severity.
Quench severity is defined as the ability of a
quenching medium to extract heat from a hot
steel workpiece; expressed in terms of the
Grossmann number (H) (Ref 27). A typical
range of Grossmann H-values (numbers) for
commonly used quench media is provided in
Table 9. Figure 26 provides a correlation between the H-value and the ability to harden steel,
as indicated by the Jominy distance (J-distance)
(Ref 20). Although Table 9 is useful to obtain a
relative measure of the quench severity offered
by different quench media, it is difficult to apply
in practice, because the actual flow rates for
moderate, good, strong, and violent
agitation are unknown. Alternatively, the measurement of actual cooling rates or heat fluxes
provided by a specific quenching medium does
provide a quantitative meaning to the quench
severity provided. Some illustrative values are
provided in Table 10 (Ref 28).
Typically, the greater the quench severity, the
greater the propensity of a given quenching
medium to cause increased distortion or cracking. This usually is the result of increased
thermal stress, not transformational stresses.
Specific recommendations for quench media
selection used with various steel alloys is provided by standards such as AMS 2759. Some
Most machined parts made from alloy steels

are oil quenched to minimize distortion.
Most small parts or finish-ground larger
parts are free quenched. Larger gears, typically those over 20 cm (8 in.), are fixture
(die) quenched to control distortion. Smaller
gears and parts, such as bushings, are usually
plug quenched on a splined plug typically
constructed from carburized 8620 steel.
Although a reduction of quench severity
leads to reduced distortion, it may also be accompanied by undesirable microstructures,
such as the formation of upper bainite
(quenched pearlite) with carburized parts.
Quench speed may be reduced by quenching
in hot (150 to 205 C, or 300 to 400 F) oil.
When hot oil quenching is used for carburized steels, lower bainite, which exhibits
properties similar to martensite, is formed.
Table 9 Typical quenching conditions and

Grossmann H-values
Quenching medium
Poor (slow) oil quenchno agitation

Good oil quenchmoderate agitation
Very good oil quenchgood agitation
Strong oil quenchviolent agitation
Poor water quenchno agitation
Very good water quenchstrong agitation
Brine quenchno agitation
Brine quenchviolent agitation
Ideal quench
Grossmann H-value
0.20
0.35
0.50
0.70
1.00
1.50
2.00
5.00
...
Note: It is possible with high-pressure impingement to achieve H-values greater

than 5.00.
Table 8 Dimensional variation in hardened

high-carbon steel with time at ambient
temperature
Steel type
Change in length after time,

Tempering Hard% 103
temperature, ness,
C
HRC 7 days 30 days 90 days 365 days
1.1% C
tool steel,
790 C
quench
None
120
205
260
66
65
63
61.5
9.0
0.2
0.0
0.0
18.0
0.6
0.2
0.2
27.0
1.1
0.3
0.3
40.0
1.9
0.7
0.3
1% C/Cr,
840 C
quench
None
120
205
260
64
65
62
60
1.0
0.3
0.0
0.0
4.2
0.5
0.1
0.1
8.2
0.7
0.1
0.1
11.0
0.6
0.1
0.1
Fig. 26
Quench severity in terms of Grossmann (H) values.

J, Jominy distance. Source: Ref 20
18/8/2008 3:30PM Plate # 0

Excellent distortion is typically obtained

with austempering, quenching into a medium just above the Ms temperature. The
formation of retained austenite is a significant problem with austempering processes.
Retained austenite is most pronounced
where manganese and nickel are major
components. The best steels for austempering are plain carbon and chromium and
molybdenum alloy steels (Ref 20).
Aqueous polymer quenchants may often be
used to replace quench oils, but quench
severity is still of primary importance.
Gas or air quenching will provide the least
distortion and may be used if the steel has
sufficient hardenability to provide the
desired properties.
Low-hardenability steels are quenched in
brine or vigorously agitated oil. However,
even with a severe quench, undesirable
microstructures, such as ferrite, pearlite, or
bainite, can form.
Kern and Suess have provided guidelines for

hardening steels to achieve optimal microstructural control (Ref 21). To minimize the
potential for cracking:

In carbon or alloy steels containing50.3% C

maximum, use a water quench.
Steels with 0.3 to 0.38% C can be water
quenched if they are in the form of simple
shapes such as round bars.
If the carbon content is 40.38%, an oil
quench should be used. (More current
references, such as AMS 2759, would permit
polymer quenching of some alloys if
appropriate quench bath maintenance procedures are used.)
Exceptions are carbon steels with low alloy
content (maximum of 1% Mn). Carbon steels
containing40.95% C and 0.30 to 0.50% Mn
Table 10 Comparison of typical heat-transfer

rates
Quench medium
Still air
Nitrogen (1 bar)
Salt bath or fluidized bed
Nitrogen (10 bar)
Helium (10 bar)
Helium (20 bar)
Still oil
Hydrogen (20 bar)
Circulated oil
Hydrogen (40 bar)
Circulated water
Heat-transfer rate, W m
5080
100150
350500
400500
550600
9001000
10001500
12501350
18002200
21002300
30003500
can be water quenched if they are in the form

of simple shapes and have no drilled or
punched holes.
Other guidelines of Kern and Suess for hardening steels include (Ref 21):

If the part has widely varying section sizes

(ratio of 3 to 1), or if it has holes, keyways, or
grooves, water quenching may produce
cracking regardless of the carbon content.
Designing with generous fillets in these
regions may resolve the problem.
If the distortion must be as low as possible,
oil or salt quenching should be used with
appropriate qualification. More recent work
has shown that polymer quenching may be
used in some cases.
If 100% bainite is required, austempering in
molten salt should be performed. To assure
that no retained austenite remains, a final
temper slightly below the austempering
temperature is recommended.
Quenchant Uniformity. Quench nonuniformity is one of the greatest contributors to

quench cracking. Quench nonuniformity can
arise from nonuniform flow fields around the
part surface during the quench or nonuniform
wetting of the surface (Ref 20, 2932). Both lead
to nonuniform heat transfer during quenching.
Nonuniform quenching creates large thermal
gradients between the core and the surface of the
part.
When there is nonuniform cooling within the
part between the Ms and Mf, there will be a
stretching or elongation in areas where the
cooling is slow, which will act as a push stress,
leading to push cracking (Ref 33, 34). Another
form of cracking is pull cracking, which occurs
with nonuniform surface cooling between the
austenitizing temperature and Ms. Push cracking
and pull cracking are the opposite of each other,
although the cracking event takes place for both
between the Ms and Mf. Figure 27 provides illustrations of both push and pull cracking (Ref 32).
Poor agitation design is a major source
of quench nonuniformity. The purpose of the
agitation system is not only to take hot fluid
away from the surface and to the heat exchanger,
but it also provides uniform heat removal over
the entire cooling surface of all of the parts
throughout the load being quenched. The
batch quench system in Fig. 28 illustrates a
system where axial (vertical) quenchant flow
occurs throughout a load of round bars lying
pg 271
18/8/2008 3:30PM Plate # 0
horizontally in a basket (Ref 19). In this case, the

bottom surfaces of the bars experience greater
agitation than the top surfaces. Cracks form on
the upper surfaces because of the nonuniform
heat loss. Agitation produces greater heat loss at
the bottom, creating a large thermal gradient
between the top and the bottom surfaces.
If a submerged spray manifold is used to
facilitate more uniform heat removal, the following design guidelines are recommended:

The total surface of the part should experience uniform quenchant impingement.
The largest holes possible (2.3 mm or
0.09 in. minimum) should be used.
The manifold face should be at least 13 mm
(0.5 in.) from the surface of the parts being
quenched.
Repeated removal of hot quenchant and
vapor should be possible.
Excessive distortion was also obtained with an

agitation system illustrated in Fig. 29 when the
quenchant flow was either in the same direction
relative to the direction of part immersion or in
the opposite direction (Ref 31). The solution

to this problem was to minimize the quenchant
flow to that required for adequate heat transfer
during the quench and to provide agitation by
mechanically moving the part up and down in the
quenchant. Identifying sources of nonuniform
fluid flow during quenching continues to be an
important tool for optimizing distortion control
and minimizing quench cracking.
Nonuniform thermal gradients during
quenching are also related to interfacial wetting
kinematics, which are of particular interest with
vaporizable liquid quenchants, including water,
oil, and aqueous polymer solutions (Ref 32).
Most liquid vaporizable quenchants exhibit
boiling temperatures between 100 and 300 C at
atmospheric pressure. When parts are quenched
in these fluids, surface wetting is usually timedependent, which influences the cooling process
and the achievable hardness.
Another major source of nonuniform
quenching is foaming and contamination. Contaminants include sludge, carbon, and other
insolubles. It includes water in oil, oil in water,
and aqueous polymer quenchants. Foaming and
contamination lead to soft spotting, increased
distortion, and possibly cracking.
Stress Raisers and Their Role in Quench

Cracking
Fig. 27
Two forms of quench cracking. Source: Ref 32
Fig. 28
Harmful effects of impeded vertical quenchant flow

through the load of a batch quench system. Source:
Ref 19
Not all quench failures occur immediately

following the quench; some failures that occur
during subsequent use may be due to unacceptably high and/or nonuniform stresses that
are imparted during the quenching process and
may even be unpredictable.
As already discussed, quench cracking occurs
due to thermal contraction stresses coupled with
Fig. 29
Effect of quenchant flow direction on distortion.

Source: Ref 31
pg 272
18/8/2008 3:30PM Plate # 0
the volumetric expansion that accompanies the

martensitic transformation. It is directly proportional to carbon content and microstructural
factors. These cracks can be instantaneous upon
quenching, or they may be delayed. Also, some
components may be crack-free, whereas seemingly identical components may have cracked.
Delayed quench cracks can be the result of
additional transformation of retained austenite in
steel. This occurs when heavily stressed retained
austenite continues to transform to martensite
prior to tempering or even after tempering, if
there is sufficient retained austenite. However,
as mentioned previously, surface damage and
inadequate microstructure (decarburizing, banding, inclusions, coarse grain size), among others,
may also cause the part to fail (Ref 3335).
Quench cracking typically initiates at the
surface, particularly at positions where geometrical changes occur, such as at corners, defects,
and inclusions. Quench cracks always begin at
the part surface and have characteristics that are
easily recognized. First, the fracture generally
runs from the surface toward the center of the
mass in a relatively straight line, with either a
longitudinal or radial orientation unless located
by a change in section size. The crack is also
likely to open or spread and may exhibit a shear
lip. Shear lips are ledges on the side of the specimen that make a 45 angle to the plane of fracture and may be present on the edges of some
predominantly brittle fractures to form a picture
frame around the surface (Ref 35). The fracture
surfaces of quench cracks are always intergranular. It is common to find secondary cracking, which forms from and after the main crack,
indicating that the component was under high
stress.
Because the quenching process involves high
levels of thermal and transformation stresses, the
presence of imperfections in the microstructure
can increase the risk of cracking the part.
Imperfections such as very small cracks, inclusions, some second-phase particles and defects
from prior machining, and laps and seams work
as stress raisers. At positions far removed from
those defects, the stress is just nominal stress,
that is, the load divided by the cross-sectional
area. This does not pose a problem if the applied
stress is below the elastic limit. However, in the
vicinity of small defects or cracks, the situation
changes, and the stress is amplified. Because of
this, they are called stress raisers and are very
important during quenching as well as during
service.
When cracked parts are subsequently tempered, the intergranular morphology may form a
thick oxide scale from the tempering process.
The microstructure adjacent to the crack will not
be decarburized unless a specimen with an
undetected quench crack is rehardened. In
quenched and tempered steels, proof of quench
cracking is often obtained by opening the crack
and looking (visually) for temper color typical
for the temperature used.
The following are some case studies showing
sources of cracking that are often attributed to
the quench but whose root cause is not the
quench itself. The quench only exacerbates the
problem. There is only one example (case
study 1) where the cracking root cause was the
quench severity.
Case Studies in Quench Cracking

Case Study 1: As-Quenched 4340 Steel
(Ref 25, 36). A component (Fig. 30) from AISI
4340 steel cracked during heat treatment. Chemical analysis of the component confirmed a
composition compatible with AISI 4340 steel.
As shown in Fig. 30, the crack passes straight
from the surface to the core. Quench cracks
always begin at the part surface and have characteristics that are easily recognized. First, the
fracture generally runs from the surface toward
Fig. 30
Macrograph of AISI 4340 quenched and tempered

steel illustrating macroetched pure quench crack.
Source: Ref 25, 36
pg 273
18/8/2008 3:30PM Plate # 0
pg 274
the center of the mass in a relatively straight line,

with either a longitudinal or radial orientation
unless located by a change in section size
(Ref 35).
The component steel is considered a highhardenability steel because of alloying elements
such as chromium, nickel, and molybdenum. It
is recommended that when using high-hardenability steels, quenchants exhibiting lower
quench severity and time should be used. This is
important because cracking can occur during
quenching due to thermal contraction stresses
coupled with the volumetric expansion that
accompanies the martensitic transformation. As
already described, excessive cooling rates (high
quench severity) will produce greater thermal
stresses in addition to greater transformation
stresses. If the total residual stresses in the part
exceed the yield strength of the steel, distortion
will occur. If the ultimate strength is exceeded,
cracking will occur (Ref 4, 5).
With respect to the case being discussed here,

the cause of cracking was identified as due to
excessively high quench severity.
Case Study 2: Cracking of 4140 Block
Forging after Quenching and Tempering.
Cracking was observed to occur with an AISI
4140 block forging subsequent to quenching and
tempering. Chemical analysis was performed on
the steel block and compared to the specification
range for this alloy (Table 11), which confirmed
that the steel was nominally 4140.
To verify the presence or absence of inclusions (quantity, morphology, and distribution), a
metallographic examination in the unetched
condition was performed. In this condition,
although the microstructure is not revealed, it is
easier to identify inclusions. The steel was
examined near the crack, and the results are
shown in Fig. 31(a). No evidence of nonmetallic
inclusions was found that could be attributed to
the observed crack formation.
Table 11 Chemical analysis

Chemical composition, wt%
Material
AISI 4140
Block
Fig. 31
Mn
Si
Cr
Ni
Mo
0.380.43
0.39
0.751.00
0.88
0.150.35
0.15
0.035
0.013
0.040
0.028
0.801.10
0.86
0.25
0.06
0.150.25
0.16
Representative view of the surface and crack profile from the block sample. (a) Unetched condition. (b) Etched with 2% nital.
18/8/2008 3:30PM Plate # 0
pg 275

Material
AISI 4140
AISI 4130
Component
Mn
Si
Cr
Ni
Mo
Cu
0.380.43
0.280.33
0.40
0.751.00
0.400.60
0.88
0.150.35
0.150.35
0.25
0.035
0.035
0.025
0.040
0.040
0.031
0.801.10
0.801.10
1.03
0.25
0.25
0.09
0.150.25
0.150.25
0.22
...
0.35
0.25
Fig. 32
Representative view of the crack propagating from

porosity or voids within the brazed joint. Unetched.
The surface was also examined, and no evidence was observed of detrimental surface
conditions, such as small cracks or defects due to
machining that could contribute to the cracking.
This is important because quenching involves
high levels of thermal and transformation
stresses, and the presence of imperfections in the
microstructure can increase the risk of cracking
the part. Imperfections act like very small
cracks. Inclusions and other surface defects
function as stress raisers.
In Fig. 31, it is possible to see evidence of
secondary cracking. It is common to find secondary cracking that forms around the main
crack, indicating that the component was under
high stress due to thermal contraction stresses
coupled with the volumetric expansion that
accompanies the martensitic transformation.
Metallographic examination in the etched
condition is necessary to verify other microstructural characteristics. For steels, the most
common etchant is 2% nital (2 mL HNO3 +
98 mL ethanol, 95%). Figure 31(b) shows the
same region of the Fig. 31(a) but in the etched
condition.
Examination of the crack profiles revealed no
evidence of decarburization in the crack,
although tempering oxide was observed. The
fracture surfaces of quench cracks are always

intergranular, since it is a brittle crack (Ref 37).
When cracked parts are subsequently tempered,
intergranular morphology is usually observed in
the quench crack, which is due to a thick oxide
scale from the tempering process. Its presence
means that the fracture surface was present
before the tempering process.
After these analyses, it was possible to conclude that quenching stresses were the main
cause of the failure, and it was recommended
that a less severe quenching condition be used.
Case Study 3: Use of Improper Steel Alloy
and Presence of Voids in a Steel Brazed Joint
(Ref 25). A reamer fabricated from an AISI
4130 steel shaft was brazed to an AISI 4130 steel
body. After quenching and tempering, cracking
was observed at the brazed joint, which propagated into the reamer body.
The nominal range of chemical compositions
for AISI 4130 and 4140 steel are provided in
Table 12. Chemical analysis of the component,
which was thought to be AISI 4130, is also
shown, which confirms that the wrong steel alloy
was used. The chemical analysis of the component is consistent with that for AISI 4140.
The higher carbon content of AISI 4140
relative to AISI 4130 means greater hardenability and therefore greater probability for
cracking and increased distortion to occur. In
such cases, quenching should be less severe.
From these data, it would appear that the AISI
4130 steel reamer body was most likely exposed
to an excessively high cooling rate for this steel
alloy if heat treatment parameters were set for
4130 steel.
To verify the presence or absence of inclusions (quantity, morphology, and distribution), a
metallographic examination in the unetched
condition was made. In this case, the examination was also made near the crack, and the results
are shown in Fig. 32.
Figure 32 did not reveal evidence of nonmetallic inclusions, although it is possible to
observe the presence of voids in the brazed joint.
These are undesirable and should be avoided,
since voids are stress raisers by amplifying the
18/8/2008 3:30PM Plate # 0
stress during quenching and are nucleation sites

for crack formation in the brazed joints. This
occurs during fast cooling, leading to increased
thermal stresses in the component. Figure 32
shows that cracking initiated from porosity or
voids within the brazed joint and appears to have
propagated into the reamer body from quenching stresses. Secondary cracks are also observed.
Examination of the brazed joint in the etched
condition, using 2% nital (2 mL HNO3 +98 mL
ethanol, 95%), revealed the predominance of
tempered martensite and a uniform microstructure (Fig. 33).
From these analyses, it can be concluded that
the presence of voids (stress raisers) within the
brazed joint has nucleated cracks that propagated under quenching stresses. The incorrect
steel grade increased the potential for cracking.
Fig. 33
Representative view of the brazed joint between the

reamer body (bottom) and reamer shaft (top).
Microstructure is tempered martensite. Etched with 2% nital.
Case Study 4: Presence of a Seam Defect

(Ref 25, 36). An AISI 4140 steel bar cracked
after austenitization and quenching in a 20 to
21% aqueous polymer quenchant solution.
Chemical analysis confirmed that the steel used
was compatible with AISI 4140. Metallographic
examination in the unetched condition was
performed on the steel near the crack, which is
shown in Fig. 34. No evidence of nonmetallic
inclusions was observed. However, examination
of the defect profile revealed the presence of
seam defects. A seam defect is an unbounded
fold or lap on the surface of the metal that
appears as a crack and is usually the result of a
seam that was formed but not closed during the
working process (such as rolling, forging, etc.)
of the material. These defects typically exhibit
the presence of scale and high-temperature
oxidation adjacent to the crack, as shown in
Fig. 34.
Examination of the etched condition (2%
nital), shown in Fig. 35, reveled a uniform
cross-sectional martensitic microstructure.
Examination of the crack profile revealed a seam
defect. Cracking appears to have initiated from
this defect and propagated from quenching
stresses. Evidence of decarburization (lighter
regions) and high-temperature oxidation can also
be observed within the defect profile.
Thus, cracking was caused by seam defects
that nucleated crack formation, which then
propagated due to quenching stresses.
Case Study 5: Presence of Slag Inclusions
and a Lap Defect (Ref 25). Longitudinal
cracks after quenching and tempering were
obtained with an unthreaded AISI 4140 stud bolt
(25.4 mm diameter).
Fig. 35
Fig. 34
View of the identified seam defects in a bar sample of

AISI 4140. Unetched. Original magnification: 100
Aspect of the defect in the etched condition. In the

cross section, it is possible to see a uniform microstructure compounded by martensite. Decarburizing and hightemperature oxidation can be observed. Etched with 2% nital.
pg 276
18/8/2008 3:30PM Plate # 0
pg 277

Material
AISI 4140
Sample
Mn
Si
Cr
Ni
Mo
Cu
0.380.43
0.39
0.751.00
0.81
0.150.35
0.22
0.035
0.014
0.040
0.027
0.801.10
0.92
0.25
0.15
0.150.25
0.18
0.35
0.26
Fig. 36
(a) Representative view of large slag-type inclusions observed throughout the sample cross sections. Unetched. Original
magnification: 100 . (b) View of the crack profile and slag-type inclusions observed adjacent to the cracking. Unetched.
Fig. 37
Representative view of the crack surface profile.

Unetched. Original magnification: 100
Chemical analysis of this steel is shown in

Table 13, together with the nominal composition
range of AISI 4140 steel. These data confirm that
the steel is consistent with AISI 4140.
Metallography results for the steel in the
unetched condition are shown in Fig. 36(a
and b). Large slag-type inclusions were

observed throughout the cross section, as shown
in Fig. 36(a). Figure 36(b) shows these slag-type
inclusions adjacent to the crack.
The steel test specimen used for Fig. 37 is also
unetched and shows a surface profile of the
crack. A surface seam or lap-type defect is evident, and the crack appears to propagate through
or from a surface seam. A lap is a surface defect
that appears as a seam and is caused by folding
over of hot metal, fins, or sharp corners and then
rolling or forging them into the surface, although
they are not welded close by the hot surfaces
involved (Ref 38). Secondary cracks are also
observed.
Those observations are important, since
quenching involves relatively high levels of
thermal and transformation stresses, and the
presence of imperfections in the microstructure
can increase the risk of cracking. Imperfections
such as inclusions and other surface defects act
as stress raisers.
Figure 38 shows the steel in the etched condition (2% nital), which reveals a uniform tempered martensite microstructure with a slag
inclusion stress raiser.
From those observations, it can be concluded
that cracking was caused by quenching stresses
acting upon stress-concentration sites of large
18/8/2008 3:30PM Plate # 0
pg 278
cracking appeared to be intergranular, showing

tempering oxide within the profile. No evidence
of decarburization was observed. As noted
before, when cracked parts are subsequently
tempered, the intergranular morphology may
form a thick oxide scale from the tempering
process.
Quenching stresses associated with the nonuniform microstructure, caused, in part, from a
slack quench condition or inherent chemical
segregation, have contributed to the observed
cracking. Slack quenching is related to hardening of steel from the austenitizing temperature
at a rate slower than the critical cooling rate,
resulting in incomplete transformation and the
formation of one or more transformation products in addition to or instead of martensite.
Case Study 7: Network Carbides and
Coarse Grain Size. A low-alloy 17CrNiMo6
(0.18% C, 0.25% Si, 0.50% Mn, 1.65% Cr,
0.80% Mo, 1.55% Ni) carburized steel gear
produced cracks after carburizing, quenching,
and tempering. The carburized case was 1.8 to
2.0 mm, and the measured surface hardness was
57 to 61 HRC. After carburizing, the gear was
quenched from 840 C using an aqueous polymer quenchant and subsequently double tempered for 5 h at 240 C and 3 h at 260 C.
Figure 41(a) illustrates the sectioned gear.
Examination of the cracking zone in the
etched condition (2% nital) showed that
intergranular cracking occurred in the boundary
of the coarse-grained structure, as illustrated in
Fig. 41(b). Quench cracking typically initiates
at the surface, particularly at positions where
geometrical changes occur, such as at corners,
defects, and inclusions. It always begins at the
part surface and has characteristics that are
easily recognized. The fracture surfaces of
quench cracks almost always occur intergranularly (Ref 37). Quench cracking is considered a complex mechanism of intergranular
fracture and can be aggravated by the various
mechanisms of grain-boundary weakening (such
as segregation of embrittlement elements to the
grain boundary) and grain size. However, it also
is heavily influenced by volumetric expansion
during transformation hardening and the
slag inclusions adjacent to the crack and a small

seam or lap defect at the surface. Cracks were
also observed propagating from and/or through
an apparent surface seam or lap-type defect.
Case Study 6: Presence of Chemical Segregation (Ref 25, 36). A press-formed steel
flange made from AISI 1035 steel produced
cracks after quenching in an aqueous polymer
quenchant solution. Chemical composition for
this steel grade is provided in Table 14. Comparison with the actual composition of the steel
from the flange confirmed that it was the correct
AISI 1035 steel grade.
Cracking was observed at the press-formed
ring location. Examinations (unetched) of the
press-formed ring location were performed.
Figure 39, shows that in the unetched condition,
there is no evidence of surface imperfections or
inclusions that could be attributed to the cracking of the steel.
However, examinations in the etched condition revealed a microstructure that exhibited
chemical segregation in the form of banding, as
shown in Fig. 40(a). The microstructure is nonuniform and consists of bainite and tempered
martensite, as illustrated in Fig. 40(b).
Examination of the outer radius of the pressformed ring revealed evidence of cracking. The
Fig. 38
Etched condition showing tempered martensite

microstructure and slag inclusion. Etched with 2%
nital. Original magnification: 100

Material
AISI 1035
Flange
Mn
Si
Cr
Ni
Mo
Cu
0.320.38
0.33
0.600.90
0.73
...
0.18
0.04
0.012
0.050
0.003
...
0.04
...
0.02
...
50.01
...
50.01
18/8/2008 3:30PM Plate # 0
temperature extremes of quenching. Causes of

intergranular brittle fracture include brittle
second-phase particles and/or films in grain
boundaries (Ref 37).
Fig. 39
Representative view of the cracking associated with

the radius of the press-formed ring. Unetched. Original magnification: 100
The microstructure of the carburized case was

carefully examined, which also showed the
presence of brittle network carbides in the prioraustenitic grain boundaries. The formation of
network carbides is an indication that the carbon
potential employed was too high for the steel
concerned (Ref 39).
If, during carburizing, the austenite is supersaturated with carbon, that is, it contains carbon
in excess of the eutectoid composition (0.8% C),
the carbide will precipitate at the grain boundaries during slow cooling from the carburizing
temperature. Under equilibrium cooling conditions, an austenitized steel, having a carbon
content above the eutectoid carbon content, will
reject the excess carbon as carbide (Fe3C).
However, if the same austenite were to be cooled
quickly, most of the excess carbon would be
retained by the resultant martensite-austenite
structure (Ref 39).
Fig. 40
(a) Representative view of the chemical segregation (banding). Etched with 2% nital. Original magnification: 50 . (b)
Higher magnification of the microstructure showing tempered martensite and bainite. Etched with 2% nital. Original
magnification: 400
Fig. 41
(a) Carburized steel gear (17CrNiMo6). (b) Representative view of the cracking zone. Presence of coarse grains and intergranular cracking. Etched with 2% nital
pg 279
18/8/2008 3:31PM Plate # 0
Carbides, being ceramic compounds, are

brittle. The more accelerated the diffusion of
carbon, the coarser the carbides become. Oversized carbides do not resist thermal fluctuations,
because they are ceramic compounds; therefore,
cracking initiates due to thermal shock or fast
ramp-up and ramp-down of a furnace. Because
carbides in the network configuration possess a
brittle nature, once a carbide segment of a network starts cracking, crack propagation is very
fast (Ref 40).
During carburizing, it is necessary to carefully
control the carbon content of the surface case. If
this layer becomes hypereutectoid, cementite
will be present in the boundaries of the grain,
forming a network, as demonstrated in Fig. 42(a
and b).
In addition to network carbides, nonuniform
grain size (ASTM 1 and 2 ) is also observed in
Fig. 42(a and b). Nonuniform grain size is a
problem, since the hardenability of a carbon
steel may increase as much as 50% with an
increase in austenite grain size from ASTM 8
(6 to 10) to ASTM 3 (1 to 4). This phenomenon
causes a nonuniform martensitic transformation,
contributing to increased stress during quenching. The effect becomes more pronounced if
the carbon content is increased at the same time
(Ref 39). These factors contributed to the
cracking observed during quenching.
Excessive carbon content (more than 0.8%)
in the carburized layer was related to the
high carburizing temperature that was used
(950 C). Higher temperatures result in greater
solubility of the carbon in the austenite
phase. Grain growth can also be related to high
process temperature and steel chemistry.
Fig. 42
Alloying elements such as aluminum, niobium,

vanadium, or titanium function as graingrowth inhibitors, and the steel grade used
in this case did not contain these elements
(Ref 38, 39).
Therefore, it was concluded that the presence
of network carbides and nonuniform grain size,
coupled with quenching stresses, was responsible for the observed cracking.
Case Study 8: Presence of Stringer Inclusions and Chemical Segregation (Ref 25).
Pins of AISI 1144 steel (resulfurized steel grade)
were through hardened prior to induction hardening of the pin tip, and cracking and soft spots
were obtained. The nominal chemical composition of AISI 1144 steel is provided in Table 15
along with the chemical analysis of the component. These results confirm that the steel
used for the component is consistent with AISI
1144.
Examinations in the unetched condition
revealed evidence of many long stringer inclusions, which were oriented in streaks or bands
that were parallel (or longitudinal) with respect
to the length of the pins. A stringer inclusion
occurs when an impurity, either metallic or
nonmetallic, is trapped in the ingot and is elongated in the direction of hot working. It appears
as a narrow streak that is parallel to the direction
of hot working (Ref 38). Stringers were also
observed extending to the pin tip surface within
the induction-hardened case (Fig. 43). Stringers
can, like other inclusions, act as stress raisers.
At positions that may be far removed from those
defects, the applied stress (applied load/crosssectional area) may be normal and will not pose
any problem. However, if the applied stress is
(a) Microstructure of the tooth top showing boundary carbides and coarse grains. (b) Detail of the brittle carbide network
showing prior-austenitic grain size and tempered martensite
pg 280
18/8/2008 3:31PM Plate # 0
pg 281

Material
AISI 1144
Pin
Fig. 43
Mn
Si
Cr
Ni
Mo
Cu
0.400.48
0.44
1.351.65
1.50
...
0.23
0.04
0.008
0.240.33
0.29
...
0.05
...
0.02
...
0.02
...
50.01
Aspect of the stringer inclusions observed within

the pin tip location. Unetched. Original magnifica-
tion: 100
less than the elastic limit in the vicinity of small

defects or cracks, the stress is amplified. Therefore, defects such as stringer inclusions are
called stress raisers and are very important during quenching as well as during use.
Examination in the etched condition showed
that the microstructures consisted of many areas
of chemical segregation. Unavoidable chemical
segregation of alloying elements occurs during
the solidification of an ingot in the steel production process. If this occurs on a grain-sized
scale, it is called microsegregation. If chemical
segregation occurs on a much larger ingot-sized
scale, it is referred to as macrosegregation, and
the inhomogeneous steel structure will possess
nonuniform properties throughout, particularly
in the direction that is transverse to the hot
working direction (Ref 38). In the AISI 1144
steel sample being analyzed, the chemical segregation appeared as ferrite bands and was
associated with the stringer inclusions, which
can be associated with the soft spots that also can
contribute to cracking (Fig. 44a and b). It should
be noted that stringer inclusions are not uncommon for resulfurized steel grades, because
the distribution and shape of these inclusions are
often difficult variables to control.
The fracture pattern of the pin tips subsequent to induction hardening appears to have
propagated from the stringer inclusions and
associated chemical segregation in the presence

of quenching stresses.
Case Study 9: Decarburization and Oxidized Grain Boundary (Ref 25, 36). A heavy
wall tube section of AISI 4140 tube stock produced cracks after quenching and tempering.
The component was austenitized at 843 C for
2 h and then quenched into an aqueous polymer
solution (25%) and tempered at 565 C for 2 h,
then air cooled. Chemical analysis is shown in
Table 16 together with the nominal composition
of AISI 4140. These data confirm that the proper
steel alloy was used.
Although it is not shown here, the steel in the
unetched condition revealed no evidence of a
large number of nonmetallic inclusions. Examination in the etched condition of the crosssectional microstructure shows that it consists of
uniform tempered martensite (Fig. 45).
However, examination of the surface in the
etched condition revealed high-temperature
grain-boundary oxidation (Fig. 46). Hightemperature grain-boundary oxidation occurs
when grain boundaries starting at the surface of
the part are oxidized. Normally, when parts are
being heat treated, such as during carburization
or austenitization (hardening), if the furnace
contains free oxygen from air leakage (ingression) into the furnace or excessive vapor or
steam is present, oxygen will diffuse into the surface of the material, resulting in oxidation of the
grain boundaries and degradation of the engineering properties at the surface (Ref 40). Hightemperature grain-boundary oxidation also acts
as a stress-concentration site for crack initiation.
The surface profile in the etched condition
(Fig. 47) revealed evidence of partial decarburization and tempering oxide within the cracks.
When cracked parts are subsequently tempered,
the intergranular morphology may form a thick
oxide scale from the tempering process.
Decarburization appears when steels are
processed by forming, forging, heat treating, or
any other thermal treatments where the material
temperature may exceed 760 C for some time
with no atmospheric protection. If this occurs,
the steel may start losing carbon from the heated
surfaces, leading to a decarburized surface. This
18/8/2008 3:31PM Plate # 0
pg 282
Fig. 44
View of the induction-hardened pin tip location. Ferrite bands and inclusions can be observed. Etched with 3% nital.
(a) Original magnification: 100 . (b) Original magnification: 200

Material
AISI 4140
Sample
Fig. 45
Mn
Si
Cr
Ni
Mo
Cu
0.380.43
0.44
0.751.00
0.97
0.150.35
0.24
0.035
0.012
0.040
0.016
0.801.10
1.06
0.25
0.12
0.150.25
0.19
0.35
0.18
Cross-sectional microstructure showing uniform

tempered martensite. Etched with 2% nital. Original
magnification: 400
Fig. 46
Surface profile adjacent to the cracking. Evidence of

high-temperature grain-boundary oxidation. Etched
with 2% nital. Original magnification: 400
18/8/2008 3:31PM Plate # 0
Fig. 47
View of the crack profile showing tempering oxide

within the crack and decarburization at the surface.
Etched with 2% nital. Original magnification: 50
decarburized layer can be partial or full and can

degrade the engineering properties of the surface
relative to the matrix of the material.
Based on this evidence, the identified cracking was attributed to quenching stresses acting
on the oxidized surface grain boundaries.
REFERENCES
1. Y. Toshioka, Heat Treatment Deformation

of Steel Products, Mater. Sci. Technol., Vol
1, 1985, p 883892
2. S. Mocarski, Carburizing and Its Control
I. Basic Considerations, Ind. Heat., Vol 41
(No. 5), 1974, p 5870
3. A. Bavaro, Heat Treatments and Deformation, Trait. Therm., Vol 240, 1990, p 3741
4. F. Legat, Why Does Steel Crack During
Quenching, Kovine Zlitine Technol., Vol 32
(No. 34), 1998, p 273276
5. R.W. Bohl, Difficulties and Imperfections
Associated with Heat Treated Steel,
MEI Course 10, Lesson 13, ASM International
6. K.E. Thelning, Steel and Its Heat Treatment, Butterworths, London, U.K., 1985
7. H. Berns, Distortion and Crack Formation
by Heat Treatment of Tools, Radex
Rundsch., Vol 1, 1989, p 4057
8. T. Kunitake and S. Susigawa, Sumitomo

Search, May 1971, p 1625
9. R.R. Blackwood and L.M. Jarvis, Ind.
Heat., March 1991, p 2831
10. H. Fujio, T. Aida, and Y. Masumoto, Bull.
Jpn. Soc. Mech. Eng., Vol 20, 1977, p 1655
1662
11. Yu.A. Geller and V.P. Brimene, Steel
USSR, July 1971
12. G.E. Totten and M.A.H. Howes, Chapter
5Distortion of Heat Treated Components,
Steel Heat Treatment Handbook, G.E.
Totten and M.A.H. Howes, Ed., Marcel
Dekker Inc., New York, NY, 1997, p 292
13. C.E. Bates, G.E. Totten, and R.L. Brennan,
Quenching of Steel, Heat Treating, Vol 4,
ASM Handbook, ASM International, 1991,
p 67120
14. D.A. Canonico, Stress-Relief in Heat
Treating of Steel, Heat Treating, Vol 4,
p 3334
15. W.T. Cook, A Review of Selected SteelRelated Factors Controlling Distortion in
Heat-Treatable Steels, Heat Treat. Met.,
Vol 26 (No. 2), p 2736
16. Y. Toshioka, Heat Treatment Deformation
of Steel Products, Mater. Sci. Technol.,
Vol 1, Oct 1985, p 18831892
17. A. Thuvander and A. Melander, Calculation
of Distortion During Quenching of a Low
Carbon Steel, First ASM Heat Treatment
and Surface Engineering Conference in
Europe (Proc.), Part 2, Trans. Tech. Publications, 1992, p 767782
18. P.C. Clarke, Close-Tolerance Heat Treatment of Gears, Heat Treat. Met., Vol 25
(No. 3), 1998, p 6164
19. R.F. Kern, Thinking Through to Successful
Heat Treatment, Met. Eng. Q., Vol 11
(No. 1), 1971, p 14
20. R. Kern, Distortion and Cracking, II: Distortion from Quenching, Heat Treat., March
1985, p 4145
21. R. F. Kern and M. E. Suess, Steel Selection:
A Guide for Improving Performance and
Profits, Wiley-Interscience, New York,
1970, p 1550
22. R. Kern, Distortion and Cracking, III: How
to Control Cracking, Heat Treat., April
1985, p 3842
23. H.-J. Chen and Z.W. Jiang, Microstructure
Improvement and Low-Temperature Quenching of Dies Made of Cr12-Type Steel,
Jinshu Rechuli, No. 8, 1992, p 3941
pg 283
18/8/2008 3:31PM Plate # 0
24. H.-H. Shao, Analysis of the Causes of

Cracking of a 12% Cr Steel Cold Die During
Heat Treatment, Jinshu Rechuli, No. 11,
1995, p 43
25. R.R. Blackwood, L.M. Jarvis, D.G. Hoffman, and G.E. Totten, Conditions Leading
to Quench Cracking Other Than Severity of
Quench, Heat Treating Including the Liu
Dai Memorial Symposium, Proc. of the
18th Conf., R.A. Wallis and H.W. Walton,
Ed., ASM International, 1998, p 575585
26. X. Cheng and S. He, Analysis of Quenching
Cracks in Machine-Tool Pistons Under
Supersonic Frequency Induction Hardening, Heat Treat. Met. (China), No. 4, 1991,
p 5152
27. J.R. Davis, ASM Materials Engineering
Dictionary, ASM International, 1992, p 407
28. P.F. Stratton, N. Saxena, and R. Jain, Requirements for Gas Quenching Systems,
Heat Treat. Met., Vol 24 (No. 3), 1997,
p 6063
29. H.M. Tensi, G.E. Totten, and G.M.
Webster, Proposal to Monitor Agitation
of Production Quench Tanks, Heat
Treating: Including the 1997 International
Induction Heat Treating SymposiumProc.
of the 17th Conf., D.L. Milam, D.A. Poteet,
G.D. Pfaffmann, V. Rudnev, A. Muehlbauer, and W.B. Albert, Ed., ASM International, 1997, p 423441
30. S. Owaku, Quench Distortion of Steel Parts,
Netsu Shori (J. Jpn. Soc. Heat Treat.), Vol
32 (No. 4), 1992, p 198202
31. R.T. Von Bergen, The Effects of Quenchant
Media Selection on the Distortion of
Engineered Steel Parts, Quenching and
Distortion Control, G.E. Totten, Ed., ASM

32. H.M. Tensi, A. Stich, and G.E. Totten,
Fundamentals of Quenching, Met. Heat
Treat., Mar/April 1995, p 2028
33. V.D. Kalner and S.A. Yurasov, Internal
Oxidation During Carburizing, Met. Sci.
Heat Treat. (USSR), Vol 12 (No. 6), June
1970, p 451454
34. A.A. Polyakov, Quenching Properties of
Parts Having Stress Concentrators, Met.
Sci. Heat Treat., Vol 37 (No. 78), 1995,
p 324325
35. W. Becker, Fatigue Failure, Failure
Analysis and Prevention., Vol 11 ASM
p 700727
36. L.C.F. Canale, G.E. Totten, R.R. Blackwood, L.M. Jarvis, and D.G. Hoffman, An
Overview of Non-Quench Related Problems Often Attributed to the Quenchant
and Quenching Process, 59th Congresso
Anual da ABM-Internacional, July 1922,
2004 (Sao Paulo, SP, Brazil)
37. S. Lampman, Intergranular Fracture, Failure Analysis and Prevention, Vol 11, ASM
p 641649
38. Manufacturing Techniques, http://www.
mslab.boun.edu.tr/Heat_treatment.doc (Accessed April 24, 2007)
39 G. Parrish, The Influence of Microstructure
and the Properties of Case-Carburized
Components, ASM International, 1980,
p 236
40. A. Abada, Why Do Heat-Resistant Alloys
Fail?, Ind. Heat., Oct 2002, p 5559
pg 284
18/8/2008 3:32PM Plate # 0

DOI: 10.1361/faht2008p285

Steel Failures due to Tempering and

Isothermal Heat Treatment
Jan Vatavuk, Universidade Presbiteriana Mackenzie
L.C.F. Canale, Universidade de Sao Paulo USP
FAILURE of swords made by early metalsmiths was a complex phenomenon for bladesmiths. The repeated working, heating, and
cooling could cause embrittlement, with sword
failure occurring in the most critical moments of
a battle. Likewise, some of the earliest cannons
would break apart after the first shots following
manufacture. These problems occurred in the
ferrous alloy application until the benefit of
tempering became recognized. In the middle of
the 18th century, the tempering process (and/or
stress relief) received attention as a fundamentally important process in the heat treatment of
the ferrous components of tools. Some ironsmith
tools were treated by the so-called process of
water annealing, whereby steel was tempered in
the range of 300 to 600 C. The slow cooling
was substituted by water cooling.
At the beginning of the 20th century, Krupp
developed a great number of patents based on
water cooling after tempering of chromiumnickel steel. This phenomenon received attention after the start of WWI, when large amounts
of steel were used by the armament industry. In
1917, the term tempering embrittlement was
introduced by Dickenson, having been published in papers by Brarley, Hatfield, Philpot,
and Grenet. Some investigators, such as Greves
and his collaborators, began a set of experimental methods using notched bars to determine
the susceptibility of tempering embrittlement.
A relationship between the energy absorbed
after water cooling and annealing was termed
the steel susceptibility ratio. At that time, all
the experiments were performed at room temperature, because no one anticipated that temperature may also have an effect on the results.
The effect of test temperature received attention in the beginning of 1944, when Jolivet
and Vidal introduced experiments at different
temperatures, resulting in a revision of all former data.

A very important technological mark was the
development of the beneficial molybdenum
effect on the embrittlement reduction phenomenon, through work by Greaves and Jones
(Ref 1). For some time, embrittlement due to the
tempering process has been shown to be an
important failure related to heat treatment. In
this chapter, the causes and cases associated with
problems originated by tempering are reviewed.
However, to provide background on this phenomenon, a brief description of the martensite
reactions and the steel heat treatment of tempering is given to review the different stages of
microstructural transformation.
Martensite
Before describing the solid-state reactions
resulting from the tempering process in the ferrous matrix, it is important to define the martensitic structure as a function of the alloying
elements, especially for the carbon effect. Ferrous martensite is composed of a body-centered
tetragonal crystallographic structure, with lattice parameters (c and a) related to the carbon
contents of its chemical composition, as shown
in the expression (Ref 2):
c=a=1+0:0467 (wt% C)
The lattice ratio for the tetragonal structure is

approximately 1.0467, with 1 wt% C in solid
solution. As shown in Fig. 1, hardness varies
with carbon content, and that effect is strongly
related to the distortions caused by the carbon
atom in the body-centered tetragonal structure.
Martensite is extremely hard (maximum of 800
to 900 HV) and brittle.
pg 285
18/8/2008 3:32PM Plate # 0
The transformation of martensite from austenite is a nonequilibrium (athermal) process

that occurs during rapid cooling from the austenite phase. Unlike the phase transformation
from atomic diffusion at equilibrium temperatures, the martensitic transformation occurs
when many atoms together undergo a shear
displacement. This rapid shear displacement of
atoms results in a rapid change in crystal structure during cooling. Thus, the martensitic
transformation is referred to as a diffusionless
process, because the transformation involves a
nonequilibrium (athermal) crystal change from a
shear transformation during rapid cooling from
austenite.
Martensite is a nonequilibrium structure and

thus does not appear on the iron-carbon equilibrium phase diagram. The face-centered cubic
lattice shearing of austenite (Bains deformation) in a martensitic transformation is illustrated in Fig. 2. The deformation is large and
rapid over many atoms in the lattice, and the
change in the polycrystalline system is accommodated by lattice deformation. This can occur
by slide, mechanical twinning, or even a mixture
of both mechanisms according to the steel chemical composition. The crystal change results in
an expansion of the polycrystalline system.
The effect of temperature on martensite formation is directly related to the transformation
temperatures of martensite start (Ms) and martensite finish (Mf). Carbon is the alloy element
that has a higher influence on Ms temperature,
which is mainly responsible for the martensite
morphology of steels. There are several
empirical formulas to calculate Ms temperature.
Some examples are reported as follows (Ref
24):
Ms =539 432 (%C) 30:4 (%Mn) 17:7 (%Ni)
12:1 (%Cr) 7:5 (%Mo)
For medium-carbon alloy steels (Ref 4):

Ms =520 320 (%C) 50 (%Mn) 30 (%Cr)
20 %(Ni+Mo) 5 %(Cu+Si)
In Ref 3:
Ms =561 474 (%C) 33 (%Mn) 17 (%Ni)
17 (%Cr) 21 (%Mo) (Ref 3)
The higher the transformation temperature,

the higher the probability of the plastic deformation mechanism occuring by dislocation
slide, although the low temperatures provoke a
Fig. 1
Hardness of martensite as a function of carbon content.

Source: Ref 3
Fig. 2
(a) Body-centered tetragonal cell in austenite. (b) Body-centered tetragonal cell before (left) and after (right) the lattice
deformation from austenite to martensite. Source: Ref 2, 3
pg 286
18/8/2008 3:33PM Plate # 0
Steel Failures due to Tempering and Isothermal Heat Treatment / 287
plastic deformation by mechanical twinning.

This way, it is possible to establish a martensite
morphology with respect to the alloy content,
giving special attention to carbon.
Figures 3 to 5 show the morphological aspects
of martensite as a function of carbon content for
steels through optic microscopy. In the case of
lath martensite, the deformation mechanism of
the lattice is dislocation slip. This kind of martensite is also known as slipped martensite.
Morphologically, this martensite presents lath
packages, which are separated by low-angle
boundaries (Ref 5). In the past, it was thought
that the transformation units happened as lath
packages, although recently it became clear that
each lath is independently formed, and the evidence shows that an austenite film exists, which
can be seen in Fig. 6.
Measurements of dislocation density found

in martensite are on the order of 0.3 to
0.9 1013 cm/cm3 of the crystal. This dislocation density is higher than the maximum that can
be obtained by elevating the percentage of cold
plastic deformation. Some observations, made
by transmission electron microscope, show a
very small cellular structure (approximately 0.2
to 0.3 mm) inside the lath (Ref 5).
Figure 4 shows the martensite with a high
carbon content, observed with an optical
microscope. With a high carbon content, the
microstructure has twinned martensite or plate
martensite. With higher carbon contents, higher
volume of retained austenite (Fig. 5) occurs,
because a higher carbon content lowers the Ms
Fig. 3
Fig. 5
Lath (low-carbon) martensite in SAE 8620 alloy steel

(Fe, 0.2% C, 0.8% Mn, 0.55% Ni, 0.5% Cr, 0.2% Mo)
after heat treatment (954 C, or 1750 F, for 1 h, water quench)
Microstructure of quenched 1.3% C steel. Dark needles of plate martensite and white areas of retained
austenite (white arrow)
Fig. 6
Fig. 4
High-carbon martensite (0.8% C). Etched with nital
Ref 2
Illustration of the austenite film surrounding martensite

laths in a Fe-10Cr-0.2C steel. Source: Adapted from
pg 287
18/8/2008 3:33PM Plate # 0
temperature. Twinning density can be seen by

transmission electron microscopy, because
density is high even though twins are very nar ).
row (on the order of 10 A
The percentage of slip martensite and twinned
martensite in carbon steel and tool steel can be
experimentally determined, as shown in Fig. 7.
As the carbon content increases, the amount of
lath martensite decreases. The untransformed
austenite increases by the Ms temperature
(martensite start temperature) reduction. The
increase in the retained austenite volume fraction can reduce the as-quenched hardness
mainly in the higher content range.
Martensitic transformation causes an increase
in volume and size variations, which contributes
to the residual tension stresses that develop in
the surface after the heat treatment of quenching,
when transformation takes place in all of the
sample cross sections, and transformation between surfaces and nucleus occurs anachronically. The volume variation measured during the
transformation from austenite to martensite in a
1% C steel is approximately 4% (the transformation to pearlite results in a 2.4% expansion)
(Ref 2), decreasing as far as the carbon is added
in the matrix. This occurs because of the different carbon effect in the austenite related to the
Fig. 7
martensite. In the first, the deformation has a

volumetric character, while in the second, it is
more directional (Ref 7). This behavior can be
seen in Fig. 8.
It can be seen in Fig. 8 that the difference
in the specific volume between austenite and
martensite is approximately 15% from very low
carbon content to very high carbon content
(2% C). It is also interesting to observe that for
low carbon, the change of volume from an asannealed condition to an as-hardened condition
is practically nil. On the other hand, increasing
carbon content raised that difference.
These observations are important during the
component process design. As mentioned earlier, the greater the carbon content, the greater
the embrittlement of the martensite plates
(Ref 8). However, since retained austenite also
increases with carbon content, this fraction of
retained austenite will bring some toughness to
the matrix as well as reduce the volume variation as shown in the curve A-FC in Fig. 8.
This fact results in a lower load to the cold surface because of the incomplete transformation
of the nucleus. It is important to remember
that the nucleus presents a low yield strength
when retained austenite has a low mechanical resistance, decreasing the possibility of
Effect of carbon content on the lath martensite volume, retained austenite volume fraction, and Ms temperature. Source:
Adapted from Ref 6
pg 288
18/8/2008 3:33PM Plate # 0
pg 289
developing tension stress in the surface of the

component.
It can be seen in Fig. 8 that for 2% C, theoretically it could be possible to obtain a complete
transformation to martensite, but, this is not
the case. However, it is interesting to observe
that the retained austenite line matches the austenite line for very high carbon content. This is
the case for Hadfield steels (Ref 9), which have a
high manganese content (approximately 12%)
that guarantees an austenitic microstructure,
even though the carbon amount is approximately
1.2%. In this situation, it is possible to quench
large components with complex geometry
without the risk of developing cracks, even
while increasing quenchant severity.
The ability to form martensite is described
in terms of hardenability, which is related to
the presence of other alloy elements besides
carbon. For example, molybdenum and manganese increase hardenability, while cobalt lowers
the hardenability of steel. A higher hardenability
allows martensite formation with a slower
cooling rate. This is beneficial for reducing
the tensile residual stresses in the component
surface.
properties can be changed when the component

is held isothermally at a temperature where
austenite cannot form.
It is important to emphasize that tempered
martensite usually does not contain martensite.
Instead, it is a structure of fine carbide particles
in ferrite, which has formed from martensite
during the tempering. This structure has a lower
hardness than the martensite, but by proper
choice of temperature and time used, the structure developed will be fine to give the desired
hardness. Table 1 lists the colors associated with
the tempering heats, and Table 2 illustrates the
times required to reach furnace temperature
during tempering (Ref 10).
Effect on Mechanical Properties
As noted, martensitic structures are too brittle
for most practical applications. However, it
is possible to enhance the structure toughness through tempering. The toughness usually
comes at the expense of a decrease in yield
Table 1 Colors of tempering heats

Temperature(a)
Tempering
Tempering is historically associated with the
heat treatment of martensite in steels. The
resultant microstructure is called tempered
martensite. The main purpose for tempering is to
develop a usable combination of hardness and
toughness. The microstructure and mechanical
Fig. 8
Temperature(b)
Color of oxides
188
199
210
221
232
254
265
370
390
410
430
450
490
510
Faint yellow
Light straw
Dark straw
Brown
Purple
Dark blue
Light blue
238
265
293
321
337
349
376
460
510
560
610
640
660
710
(a) Temperature held for 1 h. (b) Temperature held for 8 min
Specific volume (DV/V) of carbon steels relative to room temperature. Source: Adapted from Ref 7
18/8/2008 3:33PM Plate # 0
pg 290
strength and hardness, as illustrated in Fig. 9

and 10. Figure 11 shows other modifications
of mechanical properties that occur when an oilquenched AISI 4340 steel is tempered at temperatures above 200 C.
The interrelationship between tempering
temperature, steel chemistry, and hardness can
be estimated by the equation (Ref 13):
HB=2:84Hh +75(%C) 0:78(%Si)
+14:24(%Mn)+14:77(%Cr)+128:22(%Mo)
54:0(%V) 0:55T t +435:66
where HB is the Brinell hardness after hardening

and tempering, Hh is the Rockwel (HRC) hardness after hardening, and Tt is the tempering
temperature in C. This equation was developed
for the following conditions:

p
Tt =647 [S(60 C+20)=Ht 70:9]1=4 73:45 SHt
+(5377561S)(%C)+505S(%V)+219S(%Mo)
+75S(%Cr)+66S(%Si)751 [C ]
where Ht is the hardness after hardening and

tempering (HRC), S is the degree of hardening,
Sj1.0, and the alloying elements are given in
weight percent. This formula is valid for a
tempering time of 2 h.
Tempering Reactions
Tempering is a process in which the microstructure approaches equilibrium under the
influence of thermal activation. It follows that
the tendency to temper depends on how far
the starting microstructure deviates from
Hh = 20 to 65 HRC and Tt = 500 to 600 C

C = 0.20 to 0.54%, Si = 0.17 to 1.40%,
Mn = 0.50 to 1.90%, and Cr = 0.03 to
1.20%
An average relation between the hardness

after hardening (Hh) and the hardness after
hardening and tempering (Ht) can be found
through:
Hh =(T t =167 1:2)Ht 17 HRC
where Ht is the hardness after hardening and

tempering (HRC), and Tt is the tempering temperature ( C). This equation is valid for
490 C5Tt 5610 C and for a tempering time
of 1 h.
The tempering temperature for a specified
hardness after hardening and tempering is also
possible to calculate when chemical composition and the degree of hardening are known
(Ref 13):
Fig. 9
Effect of tempering on the true stress in a carbon steel.

Source: Adapted from Ref 11
Table 2 Approximate heating times for tempering

Per inch of diameter or thickness, with furnace maintained steadily at Tmax, and steel having dark or scaled surface
Temperature
C
121
149
177
204
260
316
371
427
482
Heating time, min

F
Cubes or
spheres(a)
Squares or
cylinders(a)
Average
flats(a)
Cubes or
spheres(b)
Squares or
cylinders(b)
Average
flats(b)
250
300
350
400
500
600
700
800
900
30
30
30
25
25
25
20
20
20
55
50
50
45
40
40
35
30
30
80
75
70
65
60
55
50
45
40
15
15
15
15
15
15
15
15
15
20
20
20
20
20
20
20
20
20
30
30
30
30
30
30
30
30
30
(a) In hot air oven, without circulation. (b) In circulation air furnace or oil bath (can be used only in lower temperatures)
18/8/2008 3:33PM Plate # 0
equilibrium. Martensite microstructure is the

farthest, followed by bainite, ferrite, and
cementite (Ref 14).
When the martensitic structure is metastable,
there is a natural tendency to transform it to a
structure with more stability, and those modifications are accelerated by increasing the
temperature during the tempering.
The modification that occurs during tempering is complex, and the transformations that take
place during the treatment conditions necessary
to produce the best mechanical properties combination are a result of accumulated knowledge,
not just from the academic point of view but also
the practical aspect of observation. Most of the
time, the structures developed during isothermal
heat treatments are influenced by the low content
of other elements besides iron and carbon.
Tempering stages
Solid-state reactions follow a sequence of
precipitation that is related to variables such as:

Diffusivity of the involved element

Surface energy of interfaces produced by the
reactions
Fig. 10

Crystallographic adjustment (coherence

stresses) between the precipitated phases
and the ferrous matrix
Thermodynamic stability of reactions
During tempering, the martensitic structure is

submitted to a sequence of reactions, often
superimposed and defined as temper stages (Ref
2, 3, 4, 5, 12).
Stage 1. In high-carbon steels, the precipitation of excess carbon begins with the formation of a transition carbide, such as e (Fe2.4C).
The e-carbide can grow at temperatures as low
as 50 C. Martensite is said to be supersaturated
with carbon when the concentration exceeds its
equilibrium solubility with respect to another
phase. However, the equilibrium solubility
depends on the phase. The solubility will be
larger when the martensite is in equilibrium with
a metastable phase such as e-carbide. Approximately 0.25 wt% C is said to remain in solution
after the precipitation of e-carbide is completed.
Although most textbooks will begin a discussion
of tempering with this first stage of tempering,
involving the redistribution of carbon and precipitation of transition carbides, cementite can
Effect of carbon content on the hardness of tempered plain steels. Source: Ref 10
pg 291
18/8/2008 3:33PM Plate # 0
precipitate directly (Ref 14). This is the case for

the lath martensite structure, where the dislocation density can be as high as 1012 to 1013/cm2
(Ref 15). Trapped carbon atoms will not precipitate as transition carbides, but cementite is
more stable than trapped carbon. This stage
begins at room temperature and extends to
250 C. A fine adjustment between the ecarbides and the ferrous structure is attributed to
the precipitation-hardening effect of martensite
in high-carbon steels tempered between 50 and
100 C.
Stage 2. Tempering at higher temperatures,
in the range of 200 to 300 C, for 1 h induces the
retained austenite to decompose into a mixture
of cementite and ferrite. When the austenite is
present as a film, the cementite also precipitates
as a continuous array of particles that have the
appearance of a film (Ref 3, 5, 12, 14). The
Fig. 11
martensite of the steels with less than 0.5% C

content has a retained austenite amount lower
than 2%, reaching 6% for 0.8% C. There are
some indications that austenite decomposes,
turning into ferrite and cementite, but presently a
consensus does not exist about whether this
structure can be correlated to lower bainite,
typically from the isothermal decomposition of
austenite, in the temperature range of 230 to
300 C.
Stage 3. Tempering at even higher temperatures leads to a coarsening of the cementite
particles, with those located at the plate boundaries growing at the expense of the intraplate
particles. This precipitation is responsible for
the embrittlement phenomenon observed at the
temperature of 250 to 400 C. It can be avoided
by adding silicon, which is an insoluble element
in cementite. This allows cementite formation at
Changes in the mechanical properties of AISI 4340 steel with tempering temperature. Source: Ref 12
pg 292
18/8/2008 3:34PM Plate # 0
temperatures where silicon diffusion occurs,

thus slowing the process.
Silicon is a substitutional alloy element that
accumulates in the ferrous matrix adjacent to the
Fe3C carbide in the growth process. It increases
the carbon activity in this region, decreasing the
carbon flow to the growing particle and consequently reducing its growth velocity (Ref 5, 16,
17). This silicon effect resulted in the development of alloy 300M, which substitutes for 4340
steel in those applications where it is necessary
to use tempering temperatures that cause
embrittlement of tempered martensite, which is
soon defined. This reaction begins to occur at
temperatures on the order of 100 C. Cementite
can also be observed during quenching when the
Ms temperature is elevated, as is the case of steel
with a low carbon content, mainly in the martensite formed just below the Ms temperature.
This phenomenon is known as self-tempering
(Ref 5).
Stage 4. In carbon steels, stage 3 marks the
end of the tempering process. Spheroidization of
Fe3C occurs as cementite coalesces. This phenomenon is sometimes called stage 4 of tempering (Ref 2). The lath boundary maintains
stability up to approximately 600 C. Intense
rearrangement occurs between the lath and its
low-angle boundaries above 600 C. This
recovery process is replaced by recrystallization
and coarsening (Fig. 12) at temperatures between 600 and 700 C (Fig. 13).
Effect of Temperature and Alloying. The

effect of the tempering temperature on steels
with increasing carbon contents can be inferred
from Fig. 13. During tempering, the continuous
decomposition of martensite to ferrite and carbides changes the state of stress because of
continuous dimensional changes. At low temperatures (first stage), a volume contraction
takes place as a consequence of e-carbide precipitation. In the second stage, with the transformation of retained austenite (approximately
300 C), the volume is increased. In stage 3, the
progressive decomposition of martensite leads
to a volume decrease.
It is important to observe that the austenitization temperature, which determines the
amount of carbon dissolved and the amount of
retained austenite, has a strong influence on the
expected volume changes (Ref 2, 4, 12, 13).
Table 3 shows the changes in length for various
steels as a function of tempering temperature.
Alloyed steels can also have another stage
with the precipitation of alloy carbides, including M2C (molybdenum), M7C3, M6C, M23C6
(chromium rich), V4C3, TiC, and so on, where
the M refers to a combination of metal atoms.
However, all of these carbides require longrange diffusion of substitutional atoms. They
can only precipitate when the combination of
time and temperature is sufficient to allow this
diffusion. The alloy carbides grow at the
expense of the less stable cementite. If the concentration of strong carbide-forming elements,
such as molybdenum, chromium, titanium,
vanadium, and niobium, is large, then all of the
carbon can be accommodated in the alloy carbide, thereby completely eliminating the
cementite.
Figure 14 illustrates the effect of alloying
elements on hardness as a function of tempering
temperature in carbon steels (Ref 5). Increases in
hardness with additions of titanium, vanadium,
molybdenum, and chromium are related to the
alloy carbide precipitation. This phenomenon is
common for tool steels and can affect their
toughness, as illustrated in Fig. 15.
Embrittlement
Fe-0.17C alloy quenched in water from 900 C and

tempered at 650 C for 5 h. Microstructure shows
ferrite grains and spheroidized Fe3C
Fig. 12
Hardness decreases with increasing tempering temperature (Fig. 10, 11). Consequently,
yield strength and tensile strength decrease as
well. On the other hand, elongation and ductility
increase. In this general context, a failure related
pg 293
18/8/2008 3:34PM Plate # 0
pg 294
to tempering may be attributed to an incorrect

choice of temperature (and/or time of tempering), resulting in an incorrect hardness or low
toughness.
However, most failures are related to
embrittlement phenomena. Quenched and tempered steels are susceptible to a number of different types of embrittlement. Some of them
are due to structural modifications during tempering, as previously described. However,
there are some due to the interaction of the
environment with the quenched and tempered
microstructures, such as hydrogen embrittlement and liquid metal embrittlement. Examples
of the first type of embrittlement are tempered
martensite embrittlement and temper embrittlement, which are described as follows.
Tempered Martensite Embrittlement. It is
well known that tempered martensite embrittlement (TME) is related to tempered martensite
of specimens tempered between 250 and
370 C, as shown in Fig. 16. The impact
toughness after tempering at this temperature
range is lower than that obtained on tempering
at temperatures below 250 C. This type of
Fig. 13
brittleness is inherent to some extent in all steels,

including carbon grades. For that reason, medium-temperature tempering is, as a rule, not
employed in practice, although it can ensure a
high yield limit. According to Krauss (Ref 12),
TME may or may not be associated with
impurity atom segregation to prior-austenitic
grain boundaries, but the most common factor,
at least for medium-carbon steels, is the phenomenon that takes place due to decomposition
of retained austenite to cementite in the interlath
Table 3 Length variations related to
metallurgical reactions as a function of
tempering temperature ranges
Stage
Temperature
range, C
0200
200300
230350
350700
Metallurgical reactions
Precipitation of e-carbide;
loss of tetragonality
Decomposition of retained
austenite
e-carbides decompose to
cementite
Precipitation of alloy
carbides; grain
coarsening
Source: Ref 18
Hardness as a function of tempering temperature for plain carbon steels. Source: Ref 6
Expansion (E) or
contraction (C)
C
E
C
E
18/8/2008 3:34PM Plate # 0
region of martensite plates along the prioraustenite grain boundaries (Ref 19).
There are basically three different modes of
fracture through the tempered martensite of
specimens tempered in the range of 260 to
370 C. First is brittle transgranular fracture,
which results from the decomposition of
retained austenite in the second stage of tempering, as mentioned previously. Films of
retained austenite between laths of martensite in
quenched medium-carbon steels transform into
thin plates of cementite on tempering. The second mode of fracture associated with TME is
intergranular. This kind of fracture is quite
common and has been related to phosphorus
segregation to the austenite grain boundary
Fig. 14
Fig. 15
Hardness and toughness of a tool steel as a function

of tempering temperature. Charpy V-test performed
at room temperature in the short-transverse direction
Effects of titanium, vanadium, chromium, and molybdenum on tempering hardness behavior. Source: Ref 5
pg 295
18/8/2008 3:34PM Plate # 0
during austenitization. However, data on the

effect of phosphorus on the fracture surfaces in
the untempered condition indicate only that the
presence of phosphorus at the prior-austenite
grain boundary is not sufficient for the development of TME. So, the interaction between
phosphorus and cementite is necessary for the
intergranular mode of TME (Ref 20, 21).
It is important to understand that the fracture
is occurring along a prior-austenite grain
boundary, which is now a high-angle ferrite
grain boundary (Ref 22). Krauss (Ref 12)
describes another type of transgranular fracture
mode associated with TME, which is observed
in 4340-type steels. It is interlath cleavage
induced by cracking parallel to the cementite
formed from the retained austenite transformation. Transgranular fractures in TME may be
related to the interlath carbide thickness; thinner
carbides cause interlath fracture, and thicker
carbides promote translath cleavage. In some
low-carbon steels, embrittlement is associated
with peculiar carbide morphology that provides
numerous sites for microcrack initiation, growing by microvoid coalescence and then fracture,
with little gross plastic deformation.
Silicon additions to carbon steels raise the
temperature range in which TME occurs (as
mentioned earlier), because silicon delays both
the conversion of the transition carbide (e) to
cementite within the martensite laths as well
as cementite coarsening at boundaries at
higher tempering temperatures (Ref 16, 17). The
effect of silicon content on the impact properties of 0.6%C-0.47%Mn-0.52%Cr-1.77%Ni0.19%V-1.0%-2.5%Si steels can be seen in
Fig. 17 (Ref 16).
Fig. 16
Investigations using AISI 4140 steels, have

shown that austenitizing temperature has an
influence on the TME phenomenon. High austenitizing temperature was found to be associated with more pronounced TME, favoring
brittle failure modes, even in specimens showing
virtual absence of phosphorus segregation.
These investigations found that high austenitizing temperatures increase carbide dissolution in
austenite, apparently due to more intensive carbide precipitation and growth during tempering
(Ref 23).
When TME appears, there is no heat treatment to reverse the effect, other than to reaustenitize and quench the steel, then temper
in a temperature range where TME does not
occur. TME is also called one-step embrittlement.
Temper Embrittlement. The second type
of embrittlement from tempering at high temperatures is known as temper embrittlement
(TE). Temper embrittlement occurs when tempering in the high-temperature range of 450 to
600 C. It is not a major problem because it may
be avoided simply by quenching from the tempering temperature. This embrittlement can be
reversible under high-temperature tempering.
When steel that has undergone TE is heated
to a temperature above 600 C and then cooled
very quickly, its impact is restored. Therefore,
such brittleness is termed reversible. Temper
embrittlement is also called two-step embrittlement, because two tempering treatments or a
heating step and a cooling step are sometimes
required to induce embrittlement. Figure 18
shows different situations where embrittlement
may or may not appear.
Illustration of toughness loss after tempering in the embrittlement range. Source: Ref 17
pg 296
18/8/2008 3:35PM Plate # 0
Temper embrittlement of this category of

steel takes place due to impurity segregation at
the grain boundaries and, finally, decohesion of
the grain boundary. This leads to intergranular
fracture morphology. It is signaled by a material
toughness loss. It is pointed out that grainboundary segregation depends on the alloying
elements of the steel.
Carbon steels with less than 0.5% Mn are not
prone to reversible TE. The phenomenon can
only appear in alloy steels. Alloying elements
may have a different effect on steel after tempering at the steel propensity to TE. Unfortunately, the most widely used alloying elements,
such as chromium, nickel, and manganese,
promote TE. When taken separately, they produce a weaker effect than in the case of
combined alloying. The highest embrittlement
effect is observed in chromium-nickel and
chromium-manganese steels.
A fundamental fact is that alloy steels of very
high purity are utterly unsusceptible to TE,
which is caused by the presence of various
impurities, such as phosphorus, antimony, and
arsenic, in commercial steels. Relatively small
amounts of these elements, on the order
of 0.01% or less, have been related to TE (Ref
2427).
Steels made of pure elements do not become
brittle after tempering as can be seen in Fig. 19.
Fig. 17
Molybdenum is one of the main alloying

elements in many low-alloy steels and is an
effective method of alleviating TE. Small additions of molybdenum (0.2 to 0.3%) can diminish
TE, while greater additions enhance the effect.
The mechanism of actuation is related to
molybdenum segregation (equilibrium and nonequilibrium) during quenching, and the quenched-in vacancies play a role in the temper
process (Ref 28, 29).
As mentioned earlier, in chromiummolybdenum steel, phosphorus is the major
embrittling element. However, the kinetics
Fig. 18
Schematic diagram showing thermal cycles leading

to the reversal of temper embrittlement and the deembrittlement heat treatment
Charpy tests at room temperature as a function of tempering temperature for 0.6% C steel with different silicon contents.
pg 297
18/8/2008 3:35PM Plate # 0
pg 298
of phosphorus segregation in this steel are different from that in steels due to molybdenum.
Molybdenum has a strong effect in lowering the
solubility of phosphorus in iron. This is interpreted as due to a molybdenum-phosphorus
scavenging reaction; so, if molybdenum is free
in solution, it appears to either prevent phosphorus from segregating to the grain boundaries
or to reduce the brittleness potency of phosphorus at the grain boundary. Although molybdenum is an effective element to reduce the
susceptibility for TE, the precipitation of
molybdenum as carbide must be taken care of.
To avoid that, it is observed that vanadium is
added in this grade of steel. Vanadium is a strong
carbide former compared to molybdenum and
chromium. Vanadium initially forms MC-type
carbides; this changes the molybdenum-tocarbon and chromium-to-carbon ratios. The
increase in molybdenum-to-carbon ratio is
favorable for Mo2C-type carbides, and that of
the chromium-to-carbon ratio is favorable for
Cr7C3 carbides in this grade of steel. These
changes in carbide formation sequences basically slow down the precipitation of molybdenum as carbides. When the molybdenum in solid
solution in the ferrite matrix is fully removed,
phosphorus is free to segregate, and the material

thus becomes embrittled (Ref 28, 29).
Failure analysis of high-temperature studs
concluded that the failure related to TE could be
delayed due to the presence of molybdenum and
vanadium. However, reductions of phosphorus
contents in the steel and a grain size of
approximately 10 mm could reduce the tendency
for brittle fracture (Ref 30).
Table 4 shows elements that can segregate to
the former austenite boundaries that are now
ferrite boundaries (Ref 2).
This segregation was shown in a conclusive
form through Auger electron microscopy performed on intergranular fracture surfaces. This
technique allowed the exact determination of
atomic specie concentration segregated in con of
fined boundaries with approximately 10 A
depth. This fraction varies from 0.3 to 2.0 for
steels when these matrix elements are lower than
0.1%.
Although there are similarities in the effects
of the two types of embrittlement from a practical standpoint, TME and TE are separable into
two different phenomena because they occur in
two different ranges of temperature and also
because TME is a much more rapid process than
TE. The former develops during the first hour
of the normal tempering period, and it is independent of section size and/or cooling rate after
tempering. Second, TE needs many hours to
develop, and it is an important concern, mainly
for heavy sections that are tempered at higher
temperatures (out of embrittlement range) and
cooled very slowly over a period of many hours
through the critical range of embrittlement.
Temper embrittlement develops in relatively
soft structures because of the high tempering
temperature required to develop it (Ref 12).
Temper embrittlement is clearly associated
with the high-angle ferrite grain boundaries
(former austenite grain boundaries). Thus, it is
expected that austenite, which has the necessary
concentration of segregants to allow the development of TE when aging tempered martensite,
Table 4 Chemical elements that can segregate
to the grain boundary
Group
IV B
Fig. 19
Influence of phosphorus and antimony on roomtemperature impact energy as a function of tempering temperature in a Ni-Cr-Mo steel. Arrow shows the laboratory
alloy. Source: Ref 19
C
Si
Ge
Sn
...
VB
VI B
N
P
As
Sb
Bi
O
S
Se
Te
...
18/8/2008 3:35PM Plate # 0
will develop TE upon aging a bainitic microstructure (Ref 22).

In fact, bainitic microstructures, present in
many kinds of low- and medium-alloy structural
carbon steels, can also develop embrittlement
when tempered (TBE). Tempering at low
temperature has only a small effect on the
cementite size and morphology. Consequently,
the low-temperature embrittlement phenomena
are not found in conventional bainitic microstructure (Ref 31).
However, for higher temperatures, embrittlement can be developed. The tempering temperature range for TBE depends on the chemical
composition and microstructure of the steel
(Ref 32).
Mechanisms that provoke TBE are similar
to TME, that is, precipitation of cementite
formed by the decomposition of retained austenite film at boundaries of bainitic ferrite laths or
even around the martensite-austenite islands
during soaking at the tempering temperature
(Ref 33, 34).
Secondary quenching (formation of martensite during cooling from the tempering temperature) and transformation of retained
austenite into martensite because of plastic
deformation after tempering are other possible
reasons for the occurrence of TBE. The influence of impurity segregation as well as the
as-quenched microstructure on the TE phenomenon in low-alloy steel was verified. The
susceptibility of lower bainite or martensite to
Fig. 20
embrittlement was investigated. It was shown

that martensitic microstructures are more susceptible to intergranular fracture than bainitic
microstructures (Ref 35).
Mechanical Tests for TE Determination.
As previously mentioned, TE is characterized
by decreasing the impact resistance by heating
and maintenance in the critical temperature
where the phenomenon occurs (450 to 600 C)
or by slow cooling through this temperature
range. The impact resistance can be recovered
by heating up the embrittlement temperature
range (4600 C), followed by a quick cooling.
The TE apparently does not have an influence on
hardness, yield strength, and elongation measured in conventional tension tests. The same
can be said about fatigue resistance, although
this phenomenon produces a drastic reduction in
toughness fracture and also increases the transition temperature, mainly for steels with nickel,
chromium, and magnesium additions.
Experimental methods that are more adequate
for TE studies consider the kind of load, the test
temperature, and also the velocity of the
imposed deformation. Figure 20 shows the test
results of AISI 1340 steel, performed in different
tempering temperatures.
Figure 21 indicates that the embrittlement
phenomenon of tempered martensite is just
sensitive to determinate test conditions. The
room-temperature tensile test, using unnotched
proof tests, does not present changes in terms of
reduction of section and tension rupture in the
AISI 1340 steel quenched and tempered (at different temperatures). A, unnotched sample at room-temperature test.
B, unnotched sample at very low-temperature test. C, notched sample at room temperature. D, Charpy impact test at room
temperature. Source: Ref 19
pg 299
18/8/2008 3:35PM Plate # 0
pg 300
embrittlement temperature range. It is possible

to make an analogy with the embrittlement
phenomenon that occurs in ferritic stainless steel
having chromium above 15% when exposed
to temperatures at approximately 475 C. In
some cases, this embrittlement is not detected
in tension tests, and almost no modifications
occur in elongation during the test, but a strong
variation in impact resistance occurs, decreasing
the value compared to the same unembrittled
alloy steel, as seen in Table 5 (Ref 36).
In examining Fig. 21, it is possible to once
again realize the importance of the kind of test
used to verify the embrittlement phenomenon.
As seen in Fig. 20, a common method of
detecting loss of toughness during tempering is
the impact test, which measures the energy
absorbed in fracturing a sample in a specified
impact loading for a specified specimen size and
geometry. Results can be reported as impact
energy (for fracture) at a specific temperature or
impact energy as a function of the test temperature. It is possible also to give the impact
transition temperature, which is the temperature
where the impact energy-temperature curve
changes from a relatively high value to a relatively low value. This temperature is often taken
as the temperature at the inflection point into the
impact energy-temperature curve, as shown in
Fig. 21 (Ref 22).
In Fe-Ni-C steel, the TME associated with
the formation of grain-boundary cementite was
observed only when the impact test was performed below the critical test temperature
Fig. 21
(approximately 40 C). Impact toughness is

controlled by the intergranular fracture below
this critical temperature, while it is controlled
by transgranular fracture (i.e., the matrix
toughness) above this critical temperature.
Temperature is an important test parameter to
determine embrittlement during tempering
(Ref 37).
Temper embrittlement is not detected by
simple plots of impact energy versus temperature. However, it is detected by more extensive
impact testing that measures variation of the
impact transition temperature with the tempering temperature. In order to determine the
maximum embrittlement temperature and the
boundaries of the brittleness to reversible TE,
the secondary tempering method is recommended (Ref 38).
It is often possible to tell if a steel has failed
because of one of these embrittlement problems
by examining the fracture surface. A grainboundary fracture mode is characteristic of both
types of embrittlement. However, in steels with
Table 5 Mechanical properties of ferritic
stainless steel at room temperature under two
different heat treatment conditions
Mechanical
properties
Annealed
at 800 C
Annealed at 800 C
and tempered at
450 C for 4 h
sR, Kgf/mm2
sE, Kgf/mm2
A, %
Impact,
Kgm/cm2 (J)
58.4
33.0
22.8
12 (94.5)
59.7
40.0
23.6
1.4 (11.0)
Transition of the fracture behavior of two hypothetical steels in two situations. A, tough. B, embrittled. Source: Adapted from
Ref 22
18/8/2008 3:35PM Plate # 0
extremely low levels of phosphorus and sulfur,

TME can display a cleavage surface.
Although traditionally, TME is usually
detected by Charpy tests, a series of experiments
were carried out on three commercial steels
to explore the possibility of characterizing TME
by macro- and microhardness tests. Results
indicate distinct hardness peaks in two steels and
an inflexion in the other at approximately the
TME temperature. These experiments are based
on the fact that TME is associated with impurity segregation, and microhardness measurements have elucidated such segregation effects
(Ref 39).
Interaction of the TE phenomenon with
Hydrogen Embrittlement. Hydrogen embrittlement (HEM) occurs when high-strength steels
absorb an excessive amount of hydrogen in a
variety of environments. Hydrogen can be
introduced into the material in service or during
materials processing, for example, in the presence of hydrocarbons or hydrogen sulfide or
during pickling in acids, plating, welding, and
heat treatment (Ref 40).
When tensile stresses are applied to a hydrogen-embrittled component, it may fail prematurely. Hydrogen embrittlement failures are
frequently unexpected and sometimes catastrophic. An externally applied load is not
required, because the tensile stresses may be due
to residual stresses in the material. The threshold
stresses that cause cracking are commonly
below the yield stress of the material.
High-strength steel, such as quenched and
tempered steels or precipitation-hardened steels,
are particularly susceptible to HEM.
Tensile stresses, susceptible material, and the
presence of hydrogen are necessary to cause
HEM. Residual stresses or externally applied
loads resulting in stresses significantly below
yield stresses can cause cracking. Thus, catastrophic failure can occur without significant
deformation or obvious deterioration of the
component. Very small amounts of hydrogen
can cause HEM in high-strength steels.
It is believed that hydrogen is likely to be
trapped by structural defects due to its extremely
low solubility in the iron lattice. The most
commonly recognized structural defects that
have a strong hydrogen trapping effect are grain
boundaries, dislocations, carbides, and microvoids. In the case of carbides, their size is related
to the hydrogen trapping effect. For the boronbearing steel tempered to 1050 and 1300 MPa,
fine cementites having an effect on hydrogen
trapping were found. Lower susceptibility to

HEM was found when this steel was tempered
to a lower strength level, due to delayed onset of
brittle intergranular fracture (Ref 41).
The fracture mechanism produced by this
embrittlement is not simple, because it is related
to the imposed load as well as the hydrogen
amount in the sample. Tests to determine this
embrittlement process are frequently static tests,
taking many hours or even many days to obtain
the results. This kind of test is sometimes called
static fatigue. Figure 22 shows the relation
between the crack propagation mode for different load levels.
Hydrogen embrittlement susceptibility is
related to the hardness of the ferrous matrix,
and a very small amount of hydrogen can be
deleterious in the case of high hardness of the
matrix. Under the influence of hydrogen, the
resistance to fracture or crack growth of steels
is greatly reduced in gaseous hydrogen or
hydrogen-containing environments. Thus, some
experiments with 4140 steel under different heat
Fig. 22
Illustration of a crack growing at different load levels.

pg 301
18/8/2008 3:35PM Plate # 0
treatments were carried out. Experiment results

indicated that the acceleration of the crack
growth in hydrogen was more pronounced for
quenched and tempered conditions when the
tempered temperature range coincided with the
TME range of the 4140 steel (230 to 370 C).
At higher tempered temperature (approximately
550 C), the influence of hydrogen became
insignificant. Specimens in the austempered
condition have the best performance on the tests,
as seen in Fig. 23 (Ref 42).
The effect of the microstructure on HEM was
investigated for a low-carbon (Mn-Si-Cr) steel.
Microstructure formed by bainite and martensite
has better behavior than tempered martensite in
intermediate-temperature tempering. At the
same strength level, the impact energy of the
mixed microstructure is 17% higher than only
tempered martensite. It is attributed to the fact
that bainite-martensite delays the TME onset
(Ref 43).
Similar work was performed by Lantsman
et al. Experiments with cadmium-plated
65S2VA steel springs were carried out. In this
process, the steel absorbs hydrogen, which leads
to HEM. Results show that the susceptibility of
HEM has a strong dependence on the preliminary heat treatment and structure. With an
identical hydrogen content, the austempered
steel, with lower internal stresses, will fracture
under a higher load than the steel subjected to

standard quenching and tempering (Ref 44).
Interaction of the TE Phenomenon with
Liquid-Metal Embrittlement. The presence of
select metallic specimens on the surface of
various alloys can provide for a very detrimental
reaction under load, known as liquid metal
embrittlement (LME), causing brittle fracture by
intergranular cracking. LME is the reduction in
elongation to failure that can occur when normally ductile metals are stressed while in contact
with liquid metals. Failure of components related to LME is less common than failures caused
by other processes, such as fatigue, HEM and
stress-corrosion cracking, but a significant
number of industrial failures related to LME do
occur (Ref 4547).
LME should depend on time of contact with
liquid metal while the solid is stressed. When
this occurs, the solid metal fails instantly,
because the flow of liquid metal into the crack tip
during crack growth plays a significant role (Ref
48). The liquid metal affects the fracture behavior at the tip of the crack, reducing the critical
stress intensity for fracture and altering the
micromechanism of fracture at the crack tip
(Ref 49).
Sources of the aggressive elements vary,
including unintentional or accidental exposures,
for example, during fabrication or service when
there is intimate contact between the structural
and embrittling metals and also when there are
tensile stresses above a threshold value. Eventually, small amounts (0.1 g) of embrittling
metals can produce extensive cracking by LME
(Ref 45).
Plain carbon and low-alloy steels may be
embrittled by exposure to liquid lead, cadmium,
brass, aluminum bronze, copper, zinc, lead-tin
solders, and lithium (Ref 50, 51). Other metal
alloys also have susceptibility to LME. Table 6
lists embrittling environments for some common structural materials (Ref 45).
Table 6 Examples of embrittling metal
environments for some structural materials
Structural materials
Fig. 23
Ref 42
Impact values of 4140 specimens after various

types of tempering treatments. Source: Adapted from
High-strength martensitic steel

Austenitic stainless steel
Titanium alloys
Aluminum alloys
Copper alloys
Zirconium alloys
Nickel alloys
Magnesium alloys
Source: Ref 45
Embrittling environments
Hg, In, Sn, Pb, Cd, Zn, Li, Cu

Zn, Cu, Li
Hg, Cd, Ag, Au
Hg, Ga, In, Sn, Pb, Cd, Zn, Na
Hg, Ga, Bi, Zn, Li, Sn, Pb, In
Hg, Cd, Cd-Cs, Zn
Hg, In, Li, Zn, Ag
Na, K, Rb, Cs, Zn
pg 302
18/8/2008 3:35PM Plate # 0
Industrial environments provide diverse

sources of embrittling metals, for example,
during application of coatings or during service
if temperatures and stress are sufficiently high
and there is close contact with the substrate, as
mentioned previously.
During soldering, brazing, or welding, LME
can be developed if the residual or assembly
stresses are sufficiently high. Metal lubricants
and overheated bearings are other examples that
can provoke LME (Ref 45). Literature presents
various cases of failures associated with LME,
such as failures in gas turbines, presented by
D.W. Cameron (Ref 50). In this case, high
temperatures inherent in the gas turbine aggravated the phenomenon.
Research has shown that the martensitic steel
91 (9% Cr, 1% Mo) is prone to LME by liquid
lead when some conditions are fulfilled (Ref 52
54). Quenched and tempered steels are susceptible to lead embrittlement, and the conditions to
develop this phenomenon are:

Presence of either external or internal lead in

the steel
Tensile loading
Temperature between 200 and 480 C
If any one of these conditions is not observed,

LME is avoided (Ref 12).
There are a great number of possible
mechanisms for this embrittlement. In a general
way, it occurs in metals when they are in contact
with low-melting metals and can happen even
Fig. 24
Worn surface with excessive generation of heat (arrow)
when the stresses occur at lower temperatures

than the melting point. LME is not typical, but it
is an important failure mechanism, and industrial awareness of potential problems is still
limited.
Case Studies
Case Study I: Grinding Cracks. The origin
of grinding cracks may be related to low efficiency of the cooling system, microstructure and
material cleanliness, and also excessive rate of
material removal per pass. The heat generated
due to these parameters can produce visual
characteristics on the worn surface, as seen in the
cam shaft surface shown in Fig. 24. This component (AISI 5160) was induction hardened and
tempered.
The surface cam shaft heat effect can be
studied by metallography and microhardness
determinations (100 g load) in a cross section
containing the worn surface. Figure 25 presents
this microstructure with a low-depth white layer
(approximately 50 mm) containing the lower
indenter marks, which means fresh martensite
(indentations 1 to 3). The fresh martensite has
different chemical reactivity than tempered
martensite or even other austenite decomposition products. Because of this the overheated
surface can be identified by special chemical
etching methods, for example, cooper sulfide.
Table 7 presents the hardness values.
pg 303
18/8/2008 3:35PM Plate # 0
Fig. 25
Microstructure close to the worn surface. Etched with 2% nital
Table 7 Microhardness values of the worn cam

shaft cross section
Identation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Depth, mm
Hardness, HV0.1
Average
40
35
35
95
92
90
155
150
160
220
220
225
405
405
405
610
605
625
1000
1000
1000
Base material
Base material
Base material
782
803
782
433
433
455
520
493
563
592
606
642
642
782
690
858
690
762
724
772
803
256
230
251
...
789
...
...
440.3
...
...
525.3
...
...
613.3
...
...
704.7
...
...
770
...
...
766.3
...
...
245.3
...
Going deeper in the case (right side of Fig. 25)

close to the white layer, a darker etched zone
(indentations 4 to 9) has lower hardness (bigger
indentations), which means that the temperature
was sufficient to cause tempering and reduce the
local hardness. Indentations 13 to 21 (400 to
1000 mm) show higher hardness (Table 7) related to the process tempering operation, where
hardness values are close to the martensite asquenched condition.
This condition is deleterious because worked
surfaces are more prone to grinding cracks.
One example can be seen in Fig. 26. The
Fig. 26
Grinding cracks on AISI 5160 steel cam shaft after

induction hardening and low tempering (high-hardness tempering). Original magnification: 200
observed grinding cracks are small and difficult

to detect by nondestructive tests as well as
optical techniques classified as macrographic
techniques. In this case, they could be seen only
through microscopy techniques applied directly
on the worn surface or using replica methods.
These defects must be avoided, because they
may increase the precore failure probability.
Machining operations must be carefully performed, but surface compressive residual stresses may contribute to avoiding the grinding
cracks.
Case Study 2: Transgranular and Intergranular Crack Path. Intergranular cracks are
frequently related to the TE phenomenon.
However, during low-temperature tempering,
very close to the embrittlement range, a predominantly intergranular crack path was found
in an AISI 5160 commercial steel after a tempering operation at 2000 C (Ref 55).
pg 304
18/8/2008 3:35PM Plate # 0
In a recent failure analysis of a compact

vehicle suspension spring (chemical composition shown in Table 8) with a specified hardness
range of 53 to 55 HRC, a fatigue crack nucleated
from a surface defect that had grown quickly
(radial marks), as seen in Fig. 27. A scanning
electron microscopy image of the radial fracture
surface (Fig. 28) shows transgranular and
intergranular fracture paths.
The spring hardness was on the order of 54
HRC, which means that the tempering process
was carried out very close to the beginning of the
TE range. However, a significant intergranular
crack path was found in the fracture surface.
This case shows that the fracture can assume an
intergranular path even outside of the embrittlement region. The high hardness means high
yield stress, which prompts grain-boundary
decohesion.
Similar behavior was found when analyzing a
carbonitrided sample quenched from 880 C
and tempered at 180 C. Figures 29 and 30 show
a carbonitrided layer and core microstructure,
respectively.
The hardness of the carbonitrided layer is in
the range of 60 to 62 HRC, while the core
hardness values are in the range of 30 to 32 HRC.
The carbon-enriched surface (approximately
0.9%) shows an intergranular overload fracture

path, and the nucleus low-carbon structure
(approximately 0.2%) presents a dimpled
transgranular crack pattern (Fig. 31a, b).
The higher carbon content increases hardness
values, which can promote stresses higher than
the cohesive grain-boundary strength or even
tempering outside of the embrittling temperature
range, as mentioned previously.
ACKNOWLEDGMENT
The authors would like to acknowledge J.C. Vendramim from ISOFLAMA Ind. Com Equip. Ltd.,
Brazil, for helpful suggestions in writing this chapter.
Table 8 Vehicle suspension spring analysis

C
0.49
Si
Mn
Cr
Ni
Cu
1.12
0.65
0.59
0.008
0.21
0.017
0.22
Fig. 28
Secondary electron image showing intergranular

(horizontal arrows) and transgranular (vertical
arrows) crack path
Fig. 27
Surface fracture spring. The dotted arrow shows the

fatigue crack nucleus, and the dashed arrow shows
the fatigue-to-brittle fracture transition. The solid arrow shows the
surface analyzed by scanning electron microscopy. Original
magnification: 6
Fig. 29
Tempered martensite with some retained austenite

(approximately 20%). Etched with 2% nital
pg 305
18/8/2008 3:35PM Plate # 0
Fig. 30
Base material. Low-carbon martensite (horizontal arrow) and upper bainite (vertical arrow). Etched with 2% nital
Fig. 31
Secondary electron image. (a) Intergranular crack path on the carbonitrided case. (b) Transgranular dimpled fracture in the
base material
pg 306
18/8/2008 3:35PM Plate # 0
REFERENCES
1. M. Szczepanski, The Brittleness of Steels,

John Wiley & Sons, Inc., New YorkLondon, New York, 1963
2. R.W.K. Honeycombe, Steel Microstructure
and Properties, Edward Arnold Ltd., 1981,
p 78, 205
3. E.C. Bain and H.W. Paxton, Alloying Elements in Steel, American Society for
Metals, 1966
4. A.V. Sverdlin and A.R. Ness, Chapter 2
The Effects of Alloying Elements on the
Heat Treatment of Steel, Steel Heat Treatment Handbook, G.E. Totten and M.A.H.
Howes, Ed., Marcel Dekker Inc., New
York, NY, 1997, p 4591
5. G.R. Speich and W.C. Leslie, Tempering of
Steel, Metall. Trans., Vol 3 (No. 5), May
1972, p 10431054
6. G.R. Speich, Trans. Metall. Soc. Am. Inst.
Mech. Eng., Vol 245 (No. 12), 1969,
p 25532564
7. M. Narazaki, G.E. Totten, and G.M. Webster, Hardening by Reheating and Quenching, Handbook of Residual Stress and
Deformation, G.E. Totten, M. Howes, and
T. Inoue, Ed., ASM International, 2002,
p 499
8. A. Kelly and R.B. Nicholson, Strengthening Methods in Crystals, Materials Science
Series, Applied Science Publishers Ltd.,
London, 1971, Printed in Great Britain
by Galliard Limited, Great Yarmouth,
England
9. J. Vatavuk and H. Goldenstein, Acos Hadfield Contendo Carbonetos de Niobio e
Titanio, Anais do II Encontro de Tecnologia
e Utilizacao dos Acos Nacionais, May 18
22, 1987 (Rio de Janeiro), ABM, COPPE
(UFRJ), p. 151169
10. Heat Treating Data Book, 8th ed., Seco/
Warwick Corporation, 2000, p 96
11. T. Swarr and G. Krauss, Effect of
Structure on the Deformation of AsQuenched and Tempered Martensite in a
Fe-0.2%C Alloy, Metall. Trans. A, Vol 7,
1976, p 4148
12. G. Krauss, Chapter 8Tempering of Steel,
Principles of Heat Treatment of Steel,
American Society for Metals, 1980, p 187
227
13. B. Liscic, Chapter 8Steel Heat Treatment, Steel Heat Treatment Handbook, G.E.
Totten and M.A.H. Howes, Ed., Marcel
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Dekker Inc., New York, NY, 1997, p 527

662
H.K.D.H. Bhadeshia, Tempered Martensite, University of Cambridge, http://www.
msm.cam.ac.uk/phasetrans/2004/Tempered.
Martensite/tempered.martensite.html (accessed March 2008)
G. Krauss, Deformation and Fracture in
Martensitic Carbon Steels Tempered at Low
Temperatures, Metall. Mater. Trans. B,
Vol 32, 2001, p 205221
W.J. Nam and H.C. Choi, Effect of Silicon,
Nickel and Vanadium on Impact Toughness
in Spring Steels, Mater. Sci. Technol.,
Vol 13, 1997, p 568574
W.J. Nam and H.C. Choi, Effect of Si on
Mechanical Properties of Low Alloy Steels,
Mater. Sci. Technol., Vol 15, 1999, p 527
530
C. Jensen, Dimensional Changes during
Heat Treatment of Hot Work Steel, Ind.
Heat., March 1993, p 3338
R.W. Hertzberg, Chapter 10: Microstructural Aspects of Fracture Toughness,
Deformation and Fracture Mechanics of
Engineering Materials, 4th Ed., John Wiley
& Sons, Inc., 1996, p 460463
H. Ohtani and C.J. McMahon, Jr. Modes of
Fracture in Temper Embrittlement Steels,
Acta metall., Vol 23, p 377386
S.K. Banerj, C.J. McMahon, Jr., and H.C.
Feng, Intergranular Fracture in 4340Type Steels, Effect of Impurities and
Hydrogen, Metall. Trans. A, Vol 9, 1978,
p 237247
C.R. Brooks, Chapter 5Tempering,
Principles of the Heat Treatment of Plain
Carbon and Low Alloy Steels, ASM International, 1996, p 127204
F.A. Darwish, L.C. Pereira, C. Gattis, and
M.L. Graca, On the Tempered Martensite
Embrittlement in AISI 4140 Low Alloy
Steel, Mater. Sci. Eng., Vol A132, 1991,
p L5L9
I. Olefjord, Temper Embrittlement, Review
231, Int. Met. Rev., Vol 23, 1978, p 149163
G. Krauss, Tempered Martensite Embrittlement in AISI 4340 Steel, Metall. Trans.
A, Vol 10, 1979, p 16431649
M.A. Grossman and E.C. Bain, Chapter 5
Tempering After Quench Hardening, Principles of Heat Treatment, American Society
for Metals, 1964, p 129175
G.E. Totten, C.E. Bates, and N.A. Clinton,
Chapter 1Introduction to the Heat
pg 307
18/8/2008 3:35PM Plate # 0
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Treating of Steel, Handbook of Quenchant

and Quenching Technology, ASM International, 1993, p 133
V.V. Zabilskii, Temper Embrittlement
of Structural Alloy Steels Review, Met.
Sci. Heat Treat., Vol 29 (No. 1), 1987,
p 3242
S.-H. Song, R.G. Faulkner, and P.E.J.
Flewitt, Quenching and Tempering
Induced Molybdenum Segregation to
Grain Boundaries in a 2.25%Cr-1Mo
Steel, Mater. Sci. Eng. A, Vol 281, 2000,
p 2327
S.G. Chowdhury, P. Kumar, S.K. Das, D.K.
Bhattacharya, and N. Parida, Failure Analysis of High Temperature Studs, Eng. Fail.
Anal., Vol 8, 2001, p 521528
Y. Ohmori, H. Ohtani, and T. Kunitake,
Met. Sci., Vol 8, 1974, p 357366
G.L. Hu, Z.T. Liu, P. Wang, W.J. Hwa, and
M.K. Kang, Mechanism of Tempered
Bainite Embrittlement in Some Structural
Steels, Mater. Sci. Eng. A, Vol 141, 1991,
p 221227
H.K.D.H. Bhadeshia and D.V. Edmonds,
Met. Sci., Vol 17, 1983, p 411419
K.D.H. Bhadeshia and D.V. Edmonds, Met.
Sci., Vol 17, 1983, p 420425
S. Raoul, B. Marini, and A. Pineau, Effect
of Microstructure on the Susceptibility of a
533 Steel to Temper Embrittlement, J. Nucl.
Mater., Vol 257, 1998, p 199205
J. Vatavuk, Estudo Sobre os Efeitos do
Niobio na Resposta ao Tratamento Termico
de Ligas Fe-17Cr, Fe-17Cr-Ni, e do Cr e do
Ni na Solubilidade do Nb na Ferrita, Dissertacao de Mestrado apresentada a` Escola
Politecnica da Universidade de Sao Paulo,
1986
H. Kwon and C.H. Kim, Tempered Martensite Embrittlement Fe-Ni-C Steel,
J. Mater. Sci., Vol 18 (No. 12), 1983,
p 36713678
A.V. Taran, Determining Susceptibility to
Reversible Temper Embrittleness, Met. Sci.
Heat Treat., Vol 15 (No. 10), 1973, p 865
868
A. Ghosh, K. Bhaduri, and K. Ray,
Characterization of Tempered Martensite
Embrittlement Using Hardness Measurements, Mater. Charact., Vol 33, 1994,
p 321327
N.A. Tiner, Brittle Fracture in Equipment
Failure, Mech. Eng.-CIME, Vol 112 (No. 6),
1990, p 6064
41. M. Wang, E. Akiyama, and K. Tsuzaki,

Hydrogen Degradation of a Boron-Bearing
Steel with 1050 and 1300 MPa Strength
Levels, Scr. Mater., Vol 52, 2005, p 403
408
42. J.H. Chuang, L.W. Tsay, and C. Chen,
Crack Growth Behaviour of Heat-Treated
4140 Steel in Air and Gaseous Hydrogen,
Int. J. Fatigue, Vol 20 (No. 7), 1998, p 531
536
43. D. Liu, B. Bai, H. Fang, W. Zhang, J. Gu,
and K. Chang, Effect of Tempering Temperature and Carbide Free Bainite on
the Mechanical Characteristics of a High
Strength Low Alloy Steel, Mater. Sci.
Eng. A, Vol 371, 2004, p 4044
44. P.Sh. Lantsman, G.G. Vernovskaya, and
V.N. Kudryavtsev, Use of Austempering to
Reduce Hydrogen Embrittlement of Steel
65S2VA, Metalloved. Term. Obrab. Met.,
Vol 5, May 1973, p 7374 (translated from
Russian)
45. S.P. Lynch, Metal Induced Embrittlement
of Materials, Mater. Charact., Vol 28, 1992,
p 279289
46. P.L.J. Fernandez, R.E. Clegg, and D.R.H.
Jones, Failure by Liquid Metal Induced
Embrittlement, Eng. Fail. Anal., Vol 1
(No. 1), 1994, p 5163
47. R.E. Clegg and D.R.H. Jones, Liquid Metal
Embrittlement in Failure Analysis, Mater.
Sci., Vol 27 (No. 5), 1992, p 453459
48. K. Ina and H. Koizumi, Penetration of
Liquid Metals into Solid Metals and Liquid
Metal Embrittlement, Mater. Sci. Eng. A,
Vol 387389, 2004, p 390394
49. R.E. Clegg, A Fluid Flow Based Model to
Predict Liquid Metal Induced Embrittlement Crack Propagation Rate, Eng. Fract.
Mech., Vol 68, 2001, p 17771790
50. D.W. Cameron, Failures in Large Gas
Turbines due to Liquid-Metal Embrittlement, Mater. Charact., Vol 33, 1994, p 37
43
51. Liquid Metal Embrittlement, Failure Analysis and Prevention, Vol 10, Metals
Handbook, 8th ed., American Society for
Metals, 1975, p 228229
52. G. Nicaise, A. Legris, J.B. Vogt, and J.
Foct, Embrittlement of the Martensitic
Steel 91 Tested in Liquid Lead, J. Nucl.
Mater., Vol 296 (No. 1), July 2001, p 256
264
53. A. Legris, G. Nicase, J.B. Vogt, and
J. Foct, Liquid Metal Embrittlement of the
pg 308
18/8/2008 3:35PM Plate # 0
Martensitic Steel 91: Influence of the Chemical Composition of the Liquid-Metal

Experiments and Electronic Structure
Calculations, J. Nucl. Mater., Vol 301
(No. 1), 2002, p 7076
54. A. Legris, J.B. Vogt, A. Verleene, and I.
Serre, Wetting and Mechanical Properties,
A Case Study: Liquid Metal Embrittlement
of a Martensitic Steel by Liquid Lead and

Other Liquid Metals, J. Mater. Sci., Vol 40
(No. 910), 2005, p 24592463
55. A Reguly, T.R. Strohaeker, G. Krauss, and
D.K. Matlock, Quench Embrittlement of
Hardened 5160 Steel as a Function of
Austenitizing Temperature, Metall. Mater.
Trans. A, Vol 35, Jan 2004, p 153162
pg 309
18/8/2008 3:37PM Plate # 0

DOI: 10.1361/faht2008p311
pg 311

Failure Analysis in Tool Steels

Rafael Agnelli Mesquita and Celso Antonio Barbosa,
Villares Metals
TOOL STEELS are an important class of

steels due to their distinct applications and,
especially, their specific heat treating issues.
Tool steels are used in various industrial applications that require some kind of mold, die, or
mechanical device. Tool steels are also a complex class of steels with compositions close to
carbon steel or more highly alloyed grades. Tool
steels are generally classified according to three
main definitions (Ref 1, 2):
selection of heat treatment or tool design.

A helpful way to think of this synergistic relation
is to think of the factors as simple multiplying
factors. If one of them is zero, regardless which
one, the whole result would be zero. Of course,
the factors are not so distinctly separate.
Due to the large number of grades and applications of tool steels, several possible classifications have arisen. This division is important
for the discussion of heat treating. One of the
most well-known classifications is made by the
American Iron and Steel Institute (AISI), dividing tool steels into several classes according to
application, composition, or heat treatment. This
classification is shown in Table 1.
Another possible classification for tool steels
is their division into four groups according to the
final application: hot work, cold work, plastic
mold, and high-speed tool steels. The advantage
of this division is to deal with fewer groups and
to group the diverse grades within common
aspects of each application, such as sizes, hardness, operating conditions (chocks, wear, or
plastic deformation), and surface-finishing requirements. This chapter follows such a division, but the grade nomenclatures used here are
primarily from AISI.

They are used in some forming process or

forming operation for metal, ceramic, or
plastic shaping.
Tool steel properties are only attained after
heat treating, normally performed by hardening (quenching) followed by tempering.
This class of material is produced according
to rigorous melting and processing controls.
Therefore, even for grades with chemical
compositions very close to low-alloy carbon
or engineering steels, the production practices of tool steels lead to substantial property improvements.
According to this definition, heat treating is a

key issue for tool steels. In fact, heat treating
does have a strong effect on tool life, as described in this chapter. In some applications,
three major points are equally related to tool performance (supposing that operation is constant):

Design and manufacture finishing

Steel composition and its quality
Heat treatment applied
Even though this approach shows the importance of heat treating, it may be too simple for
determining the variables related to tool performances. That is, all these aspects interact with
each other and may influence the final result.
For example, if a good design and proper heat
treatment are applied to a tool produced with an
imporperly selected steel grade, the result can be
awful. The same may happen for an incorrect
Classification of Tool Steels
Table 1 AISI classification for tool steels (Ref 1)

Group
Water-hardened tool steels

Shock-resistant tool steels
Oil-hardening tool steels
Air-hardening tool steels
High-carbon and high-chromium die steels
Tool steel for application in plastic molds
Cr-, Mo-, or W-alloyed hot work tool steels
Tungsten-alloyed high-speed steels
Molybdenum-alloyed high-speed steels
Adopted from Ref 1
Symbol
W
S
O
A
D
P
H
T
M
18/8/2008 3:37PM Plate # 0
Cold work tool steels are employed for

tooling operations in temperatures below
200 C, typically at room temperature, and for
steel forming (Fig. 1 shows some examples).
The most common tools of this class are punches
and dies for cold drawing and stamping, knives,
thread-rolling dies, and coining or cold forging
tools. In these situations, mechanical strength
and wear resistance are the main performancerelated properties, which are only obtained
through high hardness after heat treating, normally at the 60 HRC level, and the dispersion of
coarse carbides in the tool steel microstructure.
Grades of AISI class D are the most important
examples for such tools, but A and O class
grades are also employed.
Cold work tooling is also applied to shock

operation applications, such as knives for shear
cutting thick plates (normally thicker than
13 mm, or 1/2 in.), chisels, and some powder
pressing molds or cold forging dies. In these
situations, high toughness is also important,
even if the wear resistance is reduced. Grades
of AISI class S are the most important example
in these applications. Toughness is also important in the previously discussed application,
preventing several catastrophic failures or, more
commonly, adhesive wear failures, in which
microchipping is a very important issue. In these
situations, the advent of new materials with
lower carbon and chromium contents has shown
interesting results.
Hot work tool steels are used for applications in which process temperature is an important aspect for the working property of a
tooling material. A common limit for hot work
working temperatures is 600 C, although
lower-temperature applications may also be
classified as hot (or warm) work. In hot work
tooling, the operating temperature interacts with
the steel structure, which is very important for
materials selection.
Hot work tooling may be divided into three
major applications: hot forging, especially
for steel forgings; extrusion; and die casting
(Fig. 2). The two last groups are mainly employed for aluminum alloys and for producing
construction or automotive parts, respectively.
All applications normally employ AISI grades H
Fig. 1
Examples of cold work tools. Courtesy of Villares

Metals
Fig. 2
Examples of hot work dies for (a) press forging and (b) die casting. Courtesy of Villares Metals
pg 312
18/8/2008 3:37PM Plate # 0
Failure Analysis in Tool Steels / 313
steel as core tools. However, for large forging

dies and tool holders, low-alloyed materials,
such as AISI grade 6F3 (DIN 1.2714), may be
employed.
The most important properties for hot work
tooling are hot strength and toughness. Hot
strength is normally related to the operating
conditions, specifically, how long and at what
temperature the tools are exposed. This gives
rise to an important concept called tempering
resistance, which describes the steel resistance
to hardness reduction after exposure to high
temperatures. Toughness, which is related to
crack end-life situations, is also very important
in hot work tooling. Toughness requirements are
so important that hot work tool steels normally
have carbon contents close to 0.40% and hardness below 52 HRC, both values much lower
than that of cold work grades.
Plastic mold steels have become a very
important division of tools due to the increase in
plastic material applications (Fig. 3). Mechanical requirements for strength and toughness
are less important than in the previous applications. In general-application molds, the moldmanufacturing-related properties are the most
important, because the end life of these molds is
rarely attained. Therefore, steel machinability,
polishability, heat treating response, and weldability are of special interest. Although less
common, there are high-demand applications,
such as the processing of corrosive polymers
or reinforced plastics, where the reason for steel
Fig. 3
selection is different, and corrosion or wear

resistance should be emphasized more than
manufacturing-related properties.
AISI P20 or P20-modified grades (mainly
DIN 1.2738) are the most employed steels in
the mold-making industry, delivered in the prehardened condition with 32 HRC. Nevertheless,
more specialized applications normally require
higher-alloyed steels, such as H13, modified
martensitic stainless steel (e.g., DIN 1.2083), or
even highly alloyed powder metallurgy grades.
Currently, the advance of manufacturing technologies, especially those related to high-speed
machining technologies, has increased the application of mold steels prehardened to high hardness, such as 40 HRC or higher. This new level
improves mold quality, especially the polishing
characteristics.
High-speed steels are also tool steels, but
they have important differences from other tool
steels. They are usually used in cutting tools,
whereas the other classes are mainly employed in
forming tools. Figure 4 shows some examples of
high-speed tools. Regarding the manufacturing
characteristics, high-speed steel tools are mainly
produced by single companies and in smaller
sizes but larger quantities than molds or dies.
From the metallurgical point of view, highspeed tool steels are definitely part of the tool
steel groups. The AISI classes for such grades
include the AISI M and T series, which correspond to the main alloy element present in their
composition: molybdenum (M) or tungsten (T).
(a) Plastic molds for a drilling machine body, showing a polishing operation, (b) Baby bath plastic injection mold, after final
polishing. Courtesy of Villares Metals.
pg 313
18/8/2008 3:38PM Plate # 0
pg 314
High-speed tool steel properties are normally

aimed at high wear resistance, which is the main
demand in cutting operations. Wear conditions
are complex in high-speed steel tools, with
important aspects being high hardness, normally
close to 65 HRC, and the homogeneous distribution of undissolved carbides. Also important is the material tempering resistance, in
order to retard hardness reduction at the high
temperatures developed in cutting processes.
Toughness should be of an acceptable level and
reduce chipping and cracking, but this property
is low when compared to other tool steels due
to extremely high hardness and the amount of
carbides in high-speed steel microstructures.
In the following sections, several failures and
their relation to heat treatment are described.
For better understanding, the aforementioned
division of tool steels is kept in two main groups,
cold work and hot work, which are the main
classes where failures are observed. High-speed
steels and mold steels are also discussed in some
specific cases. However, many aspects of a
given class (or for a specific failure) are also
applied to other classes; therefore, this chapter
uses references to previous or future parts of the
text and figures. Such a situation is not very
convenient for the reader but is important for
comparing diverse aspects in a given situation,
which is the final task of any failure analysis.
The first high-speed steels were the T type, with

T1 being the pioneer grade. During WW II, the
supply of tungsten, mainly from Germany or
eastern Europe, was interrupted, leading to the
development of molybdenum-rich grades. The
first developed grade was M1, which is still
employed in the United States, but M2 has
become much more important. M2 has a combination of molybdenum and tungsten, in almost
equal parts, which leads to very useful properties.
This grade is used for several applications.
Today (2008), it is the most important highspeed composition. Recently, the increase in
consumption of molybdenum, tungsten, and
vanadium has increased the cost of such alloys,
providing another driving force to the development of new grades. In China, due to the large
amounts of tungsten ferroalloys in that country,
the use of tungsten-rich compositions has gained
new attention. Another important element is
niobium (formerly known as columbium). Its
large reserves in Brazil motivated the development of niobium-modified grades in that
country.
Heat Treating Failures of Cold

Work Tools
Fig. 4
Chemical Composition and Main Characteristics of Cold Work Tool Steels. Typical
chemical compositions for the most common
cold work tool steels are presented in Table 2.
AISI nomenclature is used in most cases. However, some newly developed grades, not yet
Examples of high-speed tools. Courtesy of Villares

Metals
Table 2 Typical chemical compositions of cold work tool steels

Composition, wt%
AISI
D2
D3
D6
8%Cr-0.8%C;
e.g., VF800AT(a)
A2
O1
S1
M2 reg. C
PM M3 : 2(b)
DIN
1.2379
1.2080
1.2436
1.2363
1.2510
1.2542
1.3343
~1.3344(b,c)
UNS
Si
Mn
Cr
Mo
Fe
Others
T30402
T30403
...
...
1.50
2.25
2.10
0.85
0.3
0.3
0.3
1.0
0.3
0.3
0.3
0.3
12.0
12.0
11.5
8.5
1.0
...
...
2.1
...
...
0.7
...
0.9
...
0.2
0.5
bal
bal
bal
bal
...
...
...
Nb = 0.15
T30102
T31501
T41901
T11302
~T11323(b,c)
1.00
0.95
0.45
0.89
1.28
0.3
0.3
1.0
0.4
0.4
0.3
1.3
0.3
0.3
0.3
5.0
0.5
1.4
4.2
4.2
1.0
...
0.2
5.0
5.0
...
0.5
2.0
6.1
6.3
...
0.1
0.2
1.9
3.0
bal
bal
bal
bal
bal
...
...
...
...
...
(a)Trademark of Villares Metals Company, Brazil. VF800AT is not standardized; therefore, the brand name is given. (b) Obtained by powder metallurgy. (c)~, similar to
but not exactly the same
18/8/2008 3:38PM Plate # 0
standardized, are referred to in the commercial

grade. The high carbon level of almost all these
grades is evident (except for S1 due to its use in
shock-resistant applications). These tool steels
have high as-quenched hardness and, when
associated with high alloy content, high hardenability.
The combination of high carbon content and
alloy elements, especially chromium, leads to
the formation of large (also called blocky) carbides within the microstructures of cold work
tool steels. Along with a high matrix hardness,
normally 60 HRC, these carbides are the main
reason for the wear resistance of cold work tool
steels. A simple mechanism for abrasive wear
is shown in Fig. 5, with an abrasive particle
grooving away portions (chips) of tool steel
material. Hardness increase reduces the particle
penetration and the wear. However, steels cannot be much harder than 65 HRC (approximately
900 HV), which is much less than a typical
abrasive particle (between 1500 and 2500 HV).
Therefore, another mechanism takes place, this
being performed by the presence of carbides.
They can be as hard as or even harder than
abrasive particles, conferring to the material a
higher resistance to wear losses, which means
high wear resistance. In typical cold tooling, the
wear mechanism is more complicated, being
a combination of abrasive and adhesive wear.
The aforementioned mechanism is still valid,
but toughness is also shown to be important for
situations where microchipping or microcracking is present.
High-speed steels are similar to cold work
steels in this sense, with high hardness and dispersion of nondissolved carbides. The difference
is that high-speed steels present higher hot
resistance promoted by higher alloy content,
which leads to intense secondary hardening at
Fig. 5
high temperatures, as discussed in later sections

for hot work steels.
Although important or even essential for
some applications, high wear resistance is normally accompanied by low toughness. High
hardness naturally causes a reduction in toughness due to its effect on material fracture toughness. Carbides normally act as crack initiation
sites; therefore, the higher the volume fraction
of large carbides, the lower the tool steel
toughness.
The previous discussion can be illustrated
by materials properties and microstructures.
Typical microstructures of tool steels used in
cold work tooling are shown in Fig. 6. Coarse
carbides, larger than 10 mm, are present within
the microstructures. Such large carbides are
typically the M7C3 type, with M representing
the metal element that is primarily chromium.
They are typical for AISI D grades and are present in lower amounts in 8%Cr-0.8%C new
materials, as shown for VF800AT steel (the
commercial brand name). This is the main
advantage of the higher toughness of such
grades. A combination of toughness and abrasive
wear resistance leads to important improvements
in die life for 8% Cr steels, especially in metalforming operations where adhesive wear is
the main operating mechanism. Lower-alloyed
grades, such as O1, have almost no carbides
in the microstructure. This is caused by the
lack of chromium content to combine with
carbon. Therefore, they are normally used in
less demanding applications in terms of wear
resistance.
Besides the microstructural aspects, a discussion about tempering curves is also important
in order to understand cold work steel properties.
Figure 7 shows the tempering curves for some
of these grades, selected because they enable
Schematic showing wear caused by the movement of a hard particle through microgrooving. The base material has three
different conditions. (a) Tool steel with 50 HRC. (b) After hardness increase to 60 HRC. (c) Combination of high hardness and
presence of carbides within the microstructure
pg 315
18/8/2008 3:38PM Plate # 0
a discussion of all materials. For almost all

applications, the high hardness of cold working
tool steels is normally achieved from low tempering treatments. AISI D and O steels normally
have low secondary hardening. In fact, the cold
work application is dictated by this situation.
Because the materials show a low potential
for hardness retention at hot conditions, their
application is limited to low temperatures,
typically room temperature.
In practical terms, the aforementioned curves
show that tempering must normally be performed at temperatures near 200 C, if a hardness of approximately 60 HRC is desired. AISI S
grades, herein represented by S1 steel, are also
tempered in low temperatures, approximately
300 C. The chemical composition of these
alloys is specially designed for this condition,
with a high silicon level, because this element
Fig. 6
is known to dislocate temper embrittlement to

higher temperatures (Ref 4). A distinct behavior
is presented by the newcomer 8% Cr steels.
Besides lower carbon and chromium, this class
also has higher molybdenum contents than AISI
D grades. This enhances secondary hardening,
which enables tempering at higher than 500 C
and obtains hardness as high as 62 HRC. A
substantial improvement in toughness and surface treatment behavior is obtained through this
alternative tempering, which is discussed subsequently.
Some heat treating failures are observed
immediately after heat treatment and appear as
small or (usually) large catastrophic cracks.
However, other failure types related to heat
treatment are only observed during tool use,
when one notices premature failure or a lowerthan-normal performance.
Microstructures of cold work tool steels. (a) AISI D6, which is similar to D3. (b) AISI D2. (c) An 8% Cr tool steel with brand name
VF800AT. (d) AISI O1. Regions are typical for midradius of a 63 mm (21/2 in.) bar after hardening and tempering to 60 HRC.
(ad) Etched with 4% nital for the same amount of time. Original magnification: 100 . Source: Ref 3
pg 316
18/8/2008 3:38PM Plate # 0
Fig. 7
Tempering curves for most common tool steels used in cold working. Tempering curves are obtained after hardening
small (25 mm or 1 in.) specimens of all materials with the usual hardening temperature: 920 C for S1, 800 C for O1, 940 C
for D6 (similar to D3), 1010 C for D2, and 1030 C for the 8% Cr steel called VF800AT.
For simplification issues, heat-treatingrelated failures are divided by topics that represent the main cause. However, in many cases,
this division is not possible. That is because
several causes can and do act in synergy,
amplifying their effects and thus leading to the
observed failure. Nevertheless, the division is
kept. It is up to the reader to combine the presented information, keeping in mind the possibility for interaction when solving or analyzing a
specific troubleshooting case.
Design-Related Failures. The previous discussion of cold work tool steel metallurgy and
characteristics explains why this class of materials is so prone to fracture and cracking. Except
for AISI S grades, all other materials are very
brittle. This is due to their intrinsic naturethe
combination of high hardness and primary
carbidesand also because cold work steels
are used predominantly at room temperature,
where fracture toughness of steels is naturally
reduced (Ref 5). This fact is illustrated in Fig. 8,
where the lower toughness of A, D, and O grades
in comparison to H or S steels is obvious.
Cold work tool steels are thus prone to failure
under stress concentrators, also called stress
raisers, that are imposed by tool design or
machining. Today (2008), modern software is
able to calculate stresses and tool working conditions and can help to reduce stresses and
especially localized stresses under some regions
Fig. 8
Comparison of longitudinal Charpy V-notched impact

toughness for various tool steel specimens taken from
89 mm square stock and tested at working hardness. Source:
Ref 6
of tools. However, several tools are still

designed based only on previous experience.
Design faults may cause failures in heat
treatment but also during tool use, leading to
short service life. Failures just after heat treatment normally occur in the presence of some
of the following features: heavy sections adjacent to light sections, sharp corners, stamp
marks, blind holes, and improperly spaced holes
(Ref 7). Several of these faults are illustrated in
Fig. 9. Large section-size variation caused the
failures shown in Fig. 9(a) and 9(b), while the
presence of sharp corners or closely spaced
holes is shown in Fig. 9(c).
Other stress-concentration effects can also
cause or even facilitate cracking. One example
pg 317
18/8/2008 3:39PM Plate # 0
is presented in Fig. 9(d), showing cracking

produced by stamp marks. It is not evident
in Fig. 9(d), but this failure also had a contribution of poor machine finishing, because the
deep tool mark also acts as an important stress
concentrator. Figure 9(e) shows a typical sharpcorner crack. In this case, the corner was filled,
but there was a nick in the corner where the
cracking began. The shape of this fixture is also
poor for steel that must be oil quenched. As in
the case of Fig. 9(a), thinner outer regions cool
more rapidly, forming martensite first, while the
more massive central regions cool more slowly.
In some cases, it is not possible to eliminate
all the stress-concentration effects from a tool
Fig. 9
design. However, they can be minimized if heat

treatment and service failure issues are considered prior to tool design. In other words,
the designer should foresee possible heat treatment or service problems at the beginning of
tool design. As a result, several failures can
be avoided, and service life may be enhanced.
In this context, some basic advice is given in
Fig. 10.
Another possibility for solving heat treatment
or service cracking is tool steel selection. Instead
of using water-hardening grades, oil-hardening
ones are preferred in situations sensitive to
quench cracking. In some circumstances, it is
possible to apply air-hardening grades, such as
Examples of heat treatment cracking caused by design faults in hot work tool steels. (a) Cold work punch, made of a high-speed
steel, that cracked because of the large difference in section. Source Ref 1. (b) The same for a D2 die, also assisted by poor
machine finishing. Source: Ref 8. (c) O-type steel die cracked through the sharp corners Source: Ref 8. (d) Failure of die caused by stressconcentration effect of deep stamp marks. Source: Ref 1. (e) Fixture made from AISI O1 tool steel that cracked during oil quenching. A nick
in the fillet region helped to initiate cracking. Original magnification: 0.75 . Source Ref 9
pg 318
18/8/2008 3:39PM Plate # 0
series A or the new grade type 8% Cr. For

example, this could be a solution for the failure
shown in Fig. 9(e). Even D grades can be air
quenched, depending on die section. However,
the continuous cooling transformation diagram
of the cooling material should be analyzed for
the hardening condition and also the possibility
for carbide precipitation on the grain boundaries.
Before finishing this subject, one further point
should be considered. Stress raisers increase
failures in cold work tool steel, mainly due to
the intrinsically low toughness of such grades.
However, almost any tool displays some localized stress. Depending on working conditions,
one or more cracks can be initiated and propagate throughout, fracturing or spalling the tool.
Therefore, cold work steels are sensitive to overload failure. In many situations, no problems
exist in the steel, the design, or even in the heat
treatment; the only cause may be excessive
stressing of the tool due to its incorrect use.
Surface Damage by Grinding or Electrical
Discharge Machining. In the previous section,
the intrinsic brittleness of cold work tool steels
was discussed, as well as the correlation of tool
failures to poorly designed tools (regarding stress
raisers). This section discusses surface defects
introduced in tool manufacturing by grinding or
electrical discharge machining (EDM). Nevertheless, these processes can introduce more than
the macroscopic stress raiser effect due to two

major factors, described as follows.
First, both grinding and EDM cause local
heating in the tool surface that, depending on
operational conditions, induces local tempering
or, far worse, reaustenitizing, quenching, and
hardening. The high carbon of these grades,
normally more than 1%, promotes high hardness, more so than in other lower-carbon tool
steels (Fig. 11). As a result of heating and martensite transformation, small cracks, normally
hard to see with the unaided eye, may also be
formed, acting as stress raisers during tooling
and thus enabling premature cracking. Secondly, after such a metallurgical transformation
on the tool surface, the microstructure will be
predominantly untempered martensite (also
known as fresh martensite). This microstructure
is very brittle, especially in high-carbon steels
such as the cold work grades. The pre-existing
cracks or other cracks formed during tool
operation are much more prone to propagate,
thus accelerating tool failure.
Grinding and EDM cause heating and actually
act as a heat treatment applied to tool steel. In the
next sections, grinding and EDM are treated
separately, with some advice for avoiding
problems.
Incorrect Grinding. Hardenable steels are
more prone to grinding cracks than low-carbon,
low-alloy steels. Cold work tool steels and highspeed steels are the most sensitive tool steel
grades to such problems due to the high hardness
Fig. 11
Fig. 10
Simple possibilities for avoiding (a) sharp corners

and (b) large variation in section. Source: Ref 10
Effect of carbon content on the hardness of different microstructures. Martensite hardness increases
rapidly with carbon content. Reaustenitizing and quenching,
which can occur in the surface of ground or electrical discharge
machined tools, can cause high hardness and brittleness in highcarbon grades such as cold work steels, leading to tool failure.
Source: Ref 1
pg 319
18/8/2008 3:39PM Plate # 0
of as-quenched martensite and its low toughness.

Examples of grinding cracks are shown in
Fig. 12.
The formation of grinding cracks can be
explained as follows (Ref 11). Almost all the
energy used in grinding is converted into heat,
partly through pure friction and partly as a result
of deformation of the material. If a correct
grinding wheel has been chosen, most of the heat
will be removed in the chips, with only a smaller
part heating up the workpiece. Incorrect grinding
of a hardened tool steel can result in such a high
temperature at the ground surface that the tempering temperature of the material is exceeded.
This results in a reduction in the hardness of the
surface, causing low performance when the tool
is used in field applications. However, in addition
if the temperature is allowed to rise further, the
hardening temperature of the material can be
reached, resulting in rehardening.
Rehardening during the grinding operation
produces a mixture of nontempered and
tempered martensite in the surface layer, together with retained austenite, as shown in
Fig. 13(a). The affected layer normally shows
white under optical microscope examination
(after metallographic preparation and acid
etching); this denotes the presence of untempered martensite, which is more corrosion
resistant than tempered martensite. The diagram
in Fig. 13(b) shows the hardness profile through
the surface of a cold work tool steel, incorrectly
ground in such a way as to produce rehardening.
The surface exhibits a high hardness due to
the untempered martensite. An overtempered
zone occurs just below the surface, where the
hardness is lower than the basic hardness of the
workpiece.
The following hints may provide a solution
to grinding problems. Incorrect grinding, resulting in a modified surface layer, often reveals
itself through burn marksdiscoloration of the
ground surface (as indicatedted in Fig. 12b). In
order to avoid burning and grinding cracks, it
is necessary to keep down the temperature of
the ground part, for example, by means of good
cooling, and to employ properly dressed grinding wheels that cut the material with sharp cutting edges instead of simply generating heat
through friction (Ref 11).
The majority of grinding operations leave
residual stresses in the ground surface, usually
being at a maximum close to the surface. The
first and most common effect of such stresses
is the occurrence of cracks. Stresses can cause
permanent deformation of the ground part when

grinding thin materials. This may be accompanied by retained austenite formation, which
Fig. 12
Examples of grinding cracks. (a) Two views of an S1

tool cutter die cracked and spalled after grinding. Asreceived (left) and after magnetic particle testing (right), accentuating the cracks Source: Ref 9. (b) A D2 die that cracked due to
incorrect grinding (arrow indicates grinding marks) Failure was
also assisted by closely spaced holes and electrical discharge
machining procedures. Generally, grinding cracks are not as easy
to see as this. It is usually necessary to examine the part under a
microscope or with magnetic powder inspection in order to see
the cracks. Source: Ref 8
pg 320
18/8/2008 3:39PM Plate # 0
enhances deformation and the possibility for

cracking.
Three avenues are available to reduce grinding
assisted failures. First is to control the heat
emerging from the grinding operation by use of
a proper cooling process. Secondly, the effect
of grinding stresses and problems can be reduced
by stress-relief tempering after grinding. This
also involves the tempering of some regions of
untempered martensite, if present. The treatment
temperature should be approximately 30 C
below the previous tempering temperature to
avoid any risk of reducing the hardness of the
workpiece. Third, another way of reducing
grinding stresses is to tumble or blast the ground
parts (Ref 11). Obviously, if the heat damage is
too high, that is, cracks, stress relief may not help.
Incorrect EDM. Electrical discharge machining is often used in the production of cold
work tools for various reasons. Cold work tools

have an intrinsic high wear resistance and are
normally difficult-to-cut materials under regular
machining processes, such as milling. Finishing
the die making with EDM may be an interesting
solution, especially for complex-shaped tools.
However, new developments in high-speed
machining that feature low stock removal and
high frequency have been used on dies as hard as
60 HRC.
The use of EDM on hardened steels, however,
can produce a shallow, rehardened layer of
rapidly quenched as-cast structure and untempered martensite at the surface, beneath which
is a layer of tempered martensite (Fig. 14). The
EDM surface layer is known as the white layer
because of its lighter appearance under optical
microscope observation of etched samples (this
is caused by the higher corrosion resistance of
untempered martensite). Normally, the white
layer contains microcracks that can grow into
serious cracks when the tool is loaded in service
(Ref 7).
When used on hardened steel, EDM also adds
surface stresses to the already established residual stresses; the origins of such stresses are
the thermal and phase transformation dimensional variations that occur in EDM surfaces.
The temperatures developed in such regions are
so high that local melting and resolidification
occur, as shown in the upper-left microstructure
in Fig. 15(c).
In summary, the EDM white layer has four
major problems that can enable or accelerate
die failures: high hardness, residual stresses,
Fig. 14
Fig. 13
(a) White layer on a tool surface rehardened by an

incorrect grinding procedure. (b) Typical hardness
profile in regions close to cracks. Source: Ref 11
Electrical discharge machining (EDM) white layer

found on a die surface made of AISI D6 (similar to
D3) tool steel. Note the white aspect of untempered martensite
caused by the EDM process and the presence of small cracks in
this layer. Original magnification: 500 . Courtesy of Villares
Metals
pg 321
18/8/2008 3:39PM Plate # 0
Fig. 15
(a) A2 tool steel blanking die, 63 mm (21/2 in.) in diameter by 13 mm (1/2 in.) thick, that cracked in service because of a brittle
zone that had formed during electrical discharge machining (EDM) of the cavity at center. Arrows point to cracks emanating
from the cavity. Source: Ref 7. (b) Tool failure due to the same reason, where the 3.2 mm (1/8 in.) holes were produced by wire-EDM.
(c) The effect of EDM on surface microstructures and approximate hardness of the tool shown in (b) are presented. Etched with 3% nital.
Central image in lower magnification; all other images in the same magnification. Source for (b) and (c): Ref 9
pg 322
18/8/2008 3:39PM Plate # 0
coarse (as-cast) microstructure, and frequent

precracked regions.
Some typical examples of die failures assisted
by incorrect EDM are shown in Fig. 15. In
Fig. 15(a), a 63 mm (21/2 in.) diameter by
13 mm (0.5 in). long blanking die for a small
part cracked from the corners of the EDM cavity
after the die had produced 20,000 pieces. The
die was made of heat treated A2 tool steel. In
Fig. 15(b), the surface of an AISI A4 cup plate is
shown, with spalling at one of the holes, which
were made by EDM. A laboratory investigation
of this failure led to the typical appearance of
EDM-assisted failures: a coarse, white surface
layer that is very high and brittle, due to the
presence of an untempered martensitic matrix
and net carbides. Below, unquenched martensite
is observed, followed by a region of overtempered martensite, after which the normal (core)
microstructure is observed.
In many situations, the affected regions are not
exposed as clearly as in Fig. 15(c). However, the
typical white layer is always present on the tool
surface, as shown in Fig. 14. The coarser this
layer, the higher the probability for tool failure,
due to its brittleness and the fact that the EDM
white layer likely possesses cracks.
Three major practices are recommended for
avoiding premature cracking caused by EDM:

Reduce the stock removal when finishing the

EDM process (if low-frequency EDM was
used for roughing, high-frequency should be
used for finishing). This is helpful for minimizing the depth of the rehardened white
layer.
The white layer should be eliminated or
minimized by light grinding or lapping. This
procedure is time-consuming but, in many
situations, can lead to an impressive extension of tool life, especially when white layers
are thick and the tool is crack-sensitive.
For relieving stresses in EDM-processed
dies and increasing the toughness of the remaining white layer, a new tempering treatment should be performed. Its temperature
should be 30 to 50 C below the maximum
tempering temperature used in the heat
treatment, to avoid hardness loss. Normally,
this procedure is easy to apply and therefore
is highly recommended.
Although both grinding and EDM can damage

the tool surface, EDM problems are much
more common in industrial tool failures. This
occurs in particular for cold work tool steels and
high-speed steels, where the surface white layer

has high brittleness and the base material (i.e.,
the tool steel) has low resistance to crack propagation. However, EDM-assisted failures are
also observed in hot work dies.
Also important to mention here is the effect of
incorrect EDM in plastic molds. Although the
mechanical stressing is normally low, surface
finishing (by polishing or texturing) is crucial in
this application because plastic injected parts are
able to reproduce any problems on the mold
surface. Thus, EDM defects may cause serious
quality problems to injected parts, impairing the
mold application.
Failures due to the Heat Treating Procedure of Cold Work Steels. The heat treating
procedure can itself deeply change the microstructure and properties of all tool steels, not
only the cold work grades. This may occur even
if the specified hardness is obtained. This section
deals with failures caused by improper heat
treating procedure and is divided into the three
most common causes in cold work tool steels:
the use of incorrect temperatures, the use of
excessively short tempering times (or even no
tempering at all), and the formation of excessive
amounts of retained austenite, caused either by
improper hardening or incorrect tempering.
Incorrect Hardening or Tempering Temperatures. As for other tool steels, the same
class of cold work tool steels may present important differences in the indicated heat treating temperatures. If the temperature is higher
or lower than that indicated for a certain grade,
mechanical properties may be altered, especially
for toughness. Thus, the die performance is also
strongly influenced.
This section describes this effect in a specific
gradethe 8% Cr tool steel, which has been
highly employed in tools that traditionally use
grades from the D or O series. As discussed
previously, the 8%Cr-0.8%C steels have a distinct combination of toughness and wear resistance that makes these grades very suitable for
cold work tooling. However, their heat treating
temperatures are considerably different from
that used in the usual grades.
To illustrate this effect, an 8% Cr steel was
chosen (commercial name VF800), and various
temperatures were used for its heat treatment.
The composition of this grade is shown in
Table 2. Such conditions were analyzed in
the laboratory in terms of microstructure and
mechanical properties (measured by a bend
test, Ref 12). Four conditions were applied, as
pg 323
18/8/2008 3:39PM Plate # 0
follows, and summarized in Fig. 16. For all, the

hardness was maintained at 60 HRC:

Condition 1: standard-condition VF800AT

grade, with hardening temperature approximately of 1030 C and tempering at high
temperatures, 540 C (twice, for 2 h)
Condition 2: low hardening and low tempering temperatures, 970 and 200 C,
respectively. This condition is typical for
high-chromium and high-carbon D grades,
such as D3 and D6.
Condition 3: typical hardening but lower
tempering temperature. Although not indicated for 8% Cr steels, this condition is
typical for D2 steel, a well-known grade for
heat treaters. This is used in D2 for attaining
the 60 HRC level, due to the weak secondary
hardness of this grade. However, the 8%
Cr grades normally have higher alloy content in terms of molybdenum or vanadium,
allowing 60 HRC to be obtained after
high-temperature tempering. In the case of
VF800AT, up to 63 HRC is possible, depending on the hardening condition (Ref 13).
Condition 4: hardening temperature higher
than normal, using a condition typical
for high-speed steels (when treated to
60 HRC)hardening at 1150 C and
tempering at 570 C
The results of mechanical properties and

microstructures for all conditions are shown
in Fig. 16. A substantial reduction is observed
for conditions 2 to 4 compared to the material
treated under normal conditions (1).
This difference in mechanical properties may
be understood based on the relative microstructure for each condition. The first condition
has a relatively dark martensitic matrix and
dispersion of primary carbides, undissolved
during the hardening treatment. This is typical
for this material. The dark matrix indicates hightemperature tempering, where stress relief of
martensite transformation is well performed; the
dispersion of primary carbides is important for
wear resistance.
In the other conditions, the microstructures
show a different aspect. In conditions 2 and 3,
tempering at low temperatures is denoted by less
intense etching, converting to a lighter matrix.
In these cases, hardness is produced by a highly
unstable and stressed martensitic structure
instead of the secondary hardening of hightemperature tempering (adequate condition).
This reduces the toughness to the observed

levels. The last condition, 4, produced the lowest
toughness values. This is due to the intense
grain growth produced by the high hardening
temperatures, which are typical for high-speed
steels but not applied for this grade. In this
microstructure (Fig. 16e), the coarse martensite
plates reflect coarse austenite grains. The 8% Cr
cold work steels, as with other cold work steels,
have much lower alloy content than high-speed
steels. This reduces the pinning effect of carbide
precipitation on grain boundaries, thus causing
rapid grain growth when high-speed steel hardening temperatures are used.
Excessive high hardening temperatures are
also common problems in heat treating highspeed steels. Hardening temperatures for these
steels are close to the solidus temperature (less
than 50 C, 90 F), above which liquid formation starts within the microstructure (in carbide
rich areas), leading to expressive embrittlement.
Hardening temperature control is thus very
important. For example, M2 high-speed steel
hardening temperature is about 1200 C
(2192 F), but exceeding 1220 C (2228 F)
may cause loss of toughness without benefits
to hardness, and above 1240 C (2264 F),
liquation is likely to occur (Ref 38).
Figure 17 shows two cases of an incorrect heat
treatment procedure applied to an 8% Cr tool
steel, for two punches and a cutting blade, that
cracked prematurely. The microstructure observed was close to condition 3 of Fig. 16, but
the microstructure was very difficult to observe
after regular (nital) etching. A stronger etching
condition was applied. Besides tempering at low
temperature, there were evidence of hardening
overheating (coarse austenite grain sizes). This
is therefore a combination of two incorrect
situationsconditions 3 and 4.
Excessive Retained Austenite Content.
Hardening of tool steels involves the transformation of an initial phase, austenite, formed
during heating (austenitizing treatment). The
following transformations are directly dependent on austenite composition. Martensite formation is of particular interest, because this is
the expected phase after quenching. Martensitic
transformation is distinct from the usual solidphase transformations, because it does not
occur by the diffusion process. The formation
of martensite depends on the temperature
attained; therefore, two important temperatures
are defined: the start and finish of martensite
transformation, determined by the Ms and Mf
pg 324
18/8/2008 3:39PM Plate # 0
Fig. 16
(a) Bend strength and fracture energy (energy necessary to fracture the specimen) obtained in a static bend test. Four-point
bend test with specimens of 5 mm (thickness) per 7 mm (width) cross section. Tested material is an 8% Cr cold work steel
(brand name VF800AT, Ref 13), heat treated to 60 HRC under four different conditions, 1 to 4 (see text). The legend indicates the
hardening (hard.) and tempering (temp.) temperatures, all for 30 min and twice for 2 h, respectively. (b) to (e) Respective microstructures
for conditions 1 to 4 after etching with 4% nital for 10 s. All regions refer to the midradius of a 60 mm bar. Source: Ref 12
pg 325
18/8/2008 3:40PM Plate # 0
pg 326
temperatures, respectively. If no transformation

takes place during quenching, martensite is
formed gradually after the Ms temperature is
reached, and total transformation of austenite
to martensite takes place when the part reaches
the Mf temperature.
In tool steels, most compositions have Mf
values below room temperature; as a result, part
of the austenite is not transformed to martensite.
This austenite fraction is called retained austenite. It is distributed within the material microstructure; after etching, it is observed as a light
matrix crossed by plates or laths of martensite
(Fig. 18). As is discussed later, higher hardening temperatures (overheating) lead to larger
amounts of retained austenite. In the microstructure, it becomes more evident due to the
presence of large martensite plates (resulting
from larger austenite grain sizes) crossing the
austenite phase within the matrix (Fig. 18b).
Fig. 17
Retained austenite content is directly dependent on the chemical composition, because it

determines the Ms and Mf values. The following
equation (Ref 14) quantifies this dependence
for Ms, since it is also a similar rate for Mf. It is
important to observe that all the alloy elements
(with the exception of cobalt) reduce the Ms
values, especially carbon, thus increasing the
amount of retained austenite:
Ms ( C)=539 423%C 30:4%Mn 12:1%Cr
17:7%Ni 7:5% Mo
(all elements in weight percent)
(Eq 1)
Consider again the phase transformation taking place, keeping in mind the effect of alloy
composition on Ms values. As explained, austenite will be transformed to martensite because
of the rapid cooling of the quenching process.
Examples of failures in an 8%Cr-0.8%C tool steel, caused by an incorrect heat treating procedure. (a) Punches and (b) their
microstructures. (c) Cutting blade and (d) the microstructure observed in its failure analysis. The primary cause of failure is
low-temperature tempering and high-temperature hardening, a combination of incorrect conditions 3 and 4 shown in Fig. 16. Microstructures etched with Villellas reagent. Courtesy of Villares Metals
18/8/2008 3:40PM Plate # 0
Therefore, the composition that affects the Ms

value is the austenite composition at the moment of martensite transformation, not the alloy
Fig. 18
Retained austenite in two cold work tool steels after

hardening and tempering. (a) D2, with 60 HRC. (b)
O1, with 54 HRC. For both, the retained austenite content is
higher than expected (due to overheating in the hardening
treatment). (c) O1 punch from which the microstructure in (b) was
obtained. Cracked in service after short life. Retained austenite
and carbides are lighter in the microstructure after etching, but
they can be differentiated; carbides are either faceted or round,
but retained austenite has no delineated area, being within the
matrix. In both images, martensite plates that cross retained
austenite regions are clearly observed. In (a), the retained austenite content is high enough to reduce the desired hardness from
60 to 54 HRC. Compare to the usual microstructure of O1 and
D2 in Fig. 6.
composition. For monophase steels, such as

low-alloy, low-carbon grades, the austenite
composition is practically the alloy composition. However, in tool steels, especially cold
working tool steels and high-speed steels, high
amounts of carbon and alloy elements are
trapped in the undissolved carbides. The release
of these elements is only possible through carbide dissolution, which depends on time and
particularly on the austenitizing temperatures
involved. The higher the austenitizing temperature (hardening temperature), the larger the
amount of carbon and alloy elements that go
into solid solution, lowering the Ms temperature
(according to Eq 1) and thus increasing the
amount of retained austenite. Examples are
shown in Fig. 19 for a chromium high-carbon
tool steel and for D2. As the hardening temperature increases, the Ms value decreases
(Fig. 19a), and the amount of retained austenite
increases (Fig. 19a, b).
Although the basic mechanism of retained
austenite in tool steels has been described, its
effect on materials properties has not been
explained. In general, retained austenite is undesirable (Ref 17). First, it is softer than martensite, due to the crystallographic nature of
austenite (face-centered cubic lattice). However, the effect of the amount of retained austenite on the as-quenched hardness is normally
only a few HRCs, unless a strong deviation in
heat treating procedure has occurred. Secondly,
the retained austenite may cause lower toughness. Due to its unstable nature, austenite at
room temperature tends to transform to martensite if adequate thermodynamic conditions
are established; one possibility, besides the
temperature reduction, is stress. Such conditions
may be imposed during tool work that involves
elevated localized stress. In this condition,
austenite transforms to martensite (untempered,
of course), leading to brittleness. Because
retained austenite is distributed along all regions
of the steel microstructure, normally close to
carbides (microsegregated areas), its transformation to martensite may deeply embrittle
working regions, causing cracks or fracture
failures. Third, this transformation of retained
austenite also causes a volume increase, due
to the lattice difference between austenite
(more compact structure) and martensite. This
impairs dimensional stability, which is critical
for several tools, especially those that work
with small clearances, such as fine blanking
dies.
pg 327
18/8/2008 3:40PM Plate # 0
(a)
(b)
Fig. 19
(a) Influence of austenitizing temperature on martensite transformation of a tool steel containing 1.1% C and 2.8% Cr. Higher
austenitizing temperatures lower Ms temperatures and increase the amount of austenite retained at room temperature.
Source: Ref 15. (b) Amounts of microconstituents in D2 tool steel as a function of austenitizing temperature. Specimens austenitized for
30 min at temperature and oil quenched. Composition: 1.60% C, 11.95% Cr, 0.33% Mn, 0.32% Si, 0.79% Mo, 0.25% V, 0.18% P and
0.010% S. Source: Adopted from Ref 16
Fig. 20
Hardness and retained austenite as a function of

tempering in A2 tool steel. Source: Ref 18
Besides these three undesirable effects of

retained austenite, the cause of excess retained
austenite should also be considered when analyzing a failure. As discussed and shown in
Fig. 19, the increase in retained austenite content
in cold working tools is normally caused by
exceeding the recommended hardening temperature. However, the retained austenite content may also be reduced when it is converted to
martensite (or bainite) after the first tempering.
Transformation of retained austenite depends on

the tempering temperatures, as Fig. 20 shows for
A2 tool steel. For the purpose of converting
retained austenite, tempering must be carried out
immediately after quenching to avoid stabilization of retained austenite (Ref 15) and a second
tempering must always be applied. At the second
tempering temperature, this untempered martensite (also known as fresh martensite)
or bainite is tempered, being the final microstructure free from hard brittle phases.
For some cold work steels, such as AISI A or
8% Cr steels, tempering can be conducted at
temperatures (4500 C) that practically eliminate retained austenite. However, in AISI O or
D steels, tempering to 60 HRC or more is normally done at lower temperatures (5350 C),
where some amount of retained austenite will
exist after tempering; in these cases, the control
of retained austenite content should be done
by not exceeding the hardening temperatures.
Nevertheless, in some cases, it is possible to lose
some hardness by using higher tempering temperatures to reduce the amount of retained
austenite. This can be done for D2 steel if tempered at approximately 520 C (Fig. 7); hardness would be up to 58 HRC, but the amount of
retained austenite decreases from approximately
15% to less than 5% (Ref 18), considerably
improving toughness (Ref 19).
Application of cryogenic treatments is also
a possibility to reduce retained austenite
content, especially in cold work tool steels that
pg 328
18/8/2008 3:40PM Plate # 0
are tempered at low temperatures. In cryogenic

treatments, tools are led to very low temperatures, approaching or crossing the final Mf
temperature. They are normally performed after
a stress relief (at about 150 C) to avoid tool
cracking. After the cryogenic treatment, tempering is also necessary to avoid brittleness from
the just formed fresh martensite.
In summary, the presence of retained austenite in unusually large amounts is an indication
that either the hardening or tempering treatment
has been inadequately conducted. Other sources
of brittleness can emerge in such situations,
besides the retained austenite itself. For
example, higher hardening temperatures promote coarse grain sizes and also may increase
the potential for carbide precipitation on austenite grain boundaries; both effects cause intense
embrittlement (condition 4, Fig. 16). Incorrect
tempering, with shorter times (see next section)
or incorrect temperatures (previous section),
changes the stress relief of the martensite
structure and the whole strengthening mechanism, thus affecting material toughness as well
(condition 2, Fig. 16).
Therefore, several examples are observed
of industry failures assisted by retained
austenite that emerged from incorrect procedures. Figure 21 presents two examples. The
first, (Fig. 21a, b) shows an AISI O6 tool that
cracked after limited service. Retained austenite
is clearly observed in its microstructure. The
second tool in Fig. 21 is an AISI S7 die. This
grade has lower carbon (~0.50%) and much
lower undissolved carbides in comparison to the
O- or D-series steels. It should be less prone to
retained austenite formation and to the effect of
incorrect heat treating conditions. However, this
tool was carburized, and a surface pickup of
carbon took place, leading to a reduction of Ms
and Mf and thus causing the high amount of
retained austenite. In both cases, overaustenitizing conditions were employed, enabling the
existence of such high retained austenite content
and leading to embrittlement.
Excessively Short or Absent Tempering.
Tempering, the final heat treatment step
applied to tool steels, is defined as the heating
of a martensitic or hardened steel to some temperature below A1 temperature (initial temperature of austenite formation); this step produces
the final structure and mechanical properties of
a hardened steel. This citation, from Ref 1,
briefly explains the importance of tempering
treatment. However, in practical situations, this
is not so obvious. After austenitizing and

quenching, the steel is hardened to a very high
hardnessin many cases, the highest hardness
possible to attain for a given steel. After tempering, no significant differences can be observed in hardness measurements, especially for
cold work steels, which have a work hardness
very close to the as-quenched hardness. This can
lead to several problems regarding the embrittlement of a tool caused by poor tempering practice.
During tempering, several solid-state phenomena occur simultaneously in the steel microstructure. Depending on the alloy content of a
tool steel, the tempering curve presents a different aspect, as shown in Fig. 22 by a division in
classes. Class 1 is typical for high-carbon, lowalloy tool steels, class 2 for high-chromium cold
work steels, class 3 for high-speed steels with
strong secondary hardening, and class 4 for hot
work tool steels. Typical cold work steels, from
AISI D or O series, will have curves close to
classes 1 or 2.
As shown and discussed in Fig. 22 and
quantitatively in Fig. 7, both AISI O and D
grades must be tempered at lower temperatures
to attain the highest hardness levels60 HRC
or higher. Tempering is typically conducted
at temperatures of approximately 200 C. In
practical situations, this is not that simple. The
heat flux at such low temperatures is also low,
and thus, tempering of dies may take several
hours. As a consequence, it is common to find
examples of tools that were insufficiently tempered. Two cases are shown in Fig. 23.
Avoiding these kind of failures is not technically difficult, but it is time (and money) consuming. A basic rule should be observed: Any
operation should not be conducted with the steel
in the as-quenched condition. If the tempering
time is unknown, a rule of thumb to observe
is that a steel is hardly ever overtempered (if
the time is exceeded) but can be easily undertempered (if the time is too short). This happens
because the tempering parameter, as described
in Fig. 22, has the time in log scale. However, the
most desirable situation is to have good control
of the tool temperature during heat treating.
One example is monitoring tool heating and
treatment time with thermocouples attached
to the part; ideally, all tool regions should be
maintained at temperature for approximately
2 h.
This discussion of tempering times is important, due to the intense microstructural
pg 329
18/8/2008 3:40PM Plate # 0
transformations that occur in the initial tempering stages. Carbon rearranges into martensite
crystals, and transition carbides are precipitated
that are only 2 to 4 nm in size. These transformations considerably enhance material toughness, but hardness decreases only slightly from
the initial as-quenched hardness. In conclusion,
low-temperature tempering must not be suppressed based on hardness but kept due to the
necessary changes in steel microstructure that
it promotes.
Fig. 21
Heat Treating Failures of Hot Work Tools

Chemical Composition and Main Characteristics of Hot Work Tool Steels. Heat
treating of hot work tools is usually more critical
than for cold work tools. Hot work tools are
normally larger and have higher machining
costs, besides being applied to high-demand
applications. Although cold work tools may also
be critical, such as for drawing dies and industrial cutting blades, they usually do not exceed
(a) AISI O6 graphitic tool steel punch machined from centerless-ground bar stock that cracked prematurely. (b) Microstructural examination revealed an overaustenitized structure consisting of appreciable retained austenite and coarse plate
martensite. (c) Failed AISI S7 jewelry striking die showing cracks (arrows) that formed shortly after the die was placed in service; (d) Its
microstructural examination revealed that the surface was slightly carburized and the die had been overaustenitized. Note coarse plate
martensite and unstable retained austenite. Source: Ref 9
pg 330
18/8/2008 3:40PM Plate # 0
more than 500 kg. As discussed in the previous

section, it is useful to first describe the heat
treating conditions of the common grades and
then to discuss failure analysis. For this reason,
Fig. 24 shows tempering curves for the most
common grades, whose chemical compositions
are shown in Table 3.
By analyzing chemical compositions, one can
anticipate several characteristics of hot work
tool steels and their differences from cold work
steels. First, hot work steels normally have lower
carbon, leading to lower as-quenched hardness;
also, undissolved carbides will be much lower or
even nonexistent in these grades. Thus, abrasive
wear resistance, which depends on hardness
and coarse primary carbides, is much lower in
hot work steels. Final hardness in hot work
tool steels is normally determined by desired
toughness instead of wear resistance. In most
situations, hardness is between 40 and 50 HRC.
Other properties, in addition to mechanical
strength at room temperature and wear resistance, are important for hot work tool steels. The
main metallurgical properties are toughness
and strength at elevated temperatures; this last
property is dependent on the hardness stability
at high temperatures.
Fig. 22
Schematic diagram of hardness versus tempering

temperature (assuming constant time at each temperature) or versus a time-temperature tempering parameter for
four major types of tempering response in tool steels. Classes 1 to
4 are typical, respectively, for low-alloy tool steels, cold work
chromium steels, high-speed steels, and hot work tool steels.
Source: Ref 1
Toughness in hot work tools is an essential

property for avoiding cracks and fractures that
can be common in normal working conditions.
Unexpected and very unstable in cold work
tools, cracks may be encountered after a given
operation time for hot work dies. The most
important example is fine crack networks, large
in number but small in length, observed on a tool
surface after the tooling operation. This kind of
crack is known as heat checking and is typically
found in die-casting dies (Fig. 25) as well as in
forging dies or other hot work tools. The cause
is thermal fatigue at the tool surface, caused by
repeated temperature fluctuations (heating and
cooling) during tool operation. A full explanation of this behavior is found in the literature,
for example, Ref 20 to 22. Depending on the
intensity, heat checking can lead to tool failure,
which is typically the case in die-casting cavities
(Fig. 25). In such situations, the casting metal,
normally aluminum, enters into the heat checking cracks, making part extractions difficult or
impairing part surface quality. In cases of a lack
of a toughness, gross cracking may also occur,
leading to complete loss of the working tool.
High mechanical strength at high temperature
is also fundamental in hot working steels. High
temperature means temperatures that can affect
steel microstructures, normally above 500 C.
Depending on time and stress conditions, such
high temperatures lead to a decrease in hardness
that accelerates several die failure mechanisms,
especially hot wear, plastic deformation at
working temperatures, and heat checking. The
ability of a given grade to maintain hardness at
prolonged time at high temperatures is referred
to as tempering resistance. A simple (basic)
comparison of tempering resistance may be
done by analyzing tempering curves. Higher
tempering resistance is related to curves dislocated to the right (Fig. 24).
As shown schematically in Fig. 26, such hightemperature hardness is only possible by a very
important straightening mechanismsecondary
hardening, promoted by the precipitation of
fine alloy carbides. The stronger the secondary
hardening (meaning more intense carbide precipitation), the higher the tempering resistance
of hot work tool steels. Such precipitation
intensity depends on the amount of alloy elements in solid solution, which is related to the
alloy composition and heat treating practice.
Figure 27 explains this effect by comparing the
addition of different alloy contents to carbon
steels; a stronger precipitation hardening is
pg 331
18/8/2008 3:40PM Plate # 0
obtained by molybdenum, vanadium, or tungsten alloying.

Therefore, the desired mechanical properties
of hot work tool steels are only attained if a
proper heat treatment is applied. For example,
adequate secondary hardening will only be
present if alloy elements are in solid solution in
the as-quenched structure (usually martensite).
This is only possible through an adequate austenitizing treatment for hardening, capable of
dissolving the alloy elements present in the form
of carbides in the initial (annealed) state. In
addition, a proper hardening procedure should
avoid excessive grain growth and grain
boundary embrittlement by carbide precipitation. Tempering, on the other hand, should

eliminate retained austenite and promote
adequate precipitation of the alloy carbide. In
summary, tool steel properties and the expected
performance are only possible after a quality
heat treatment. Otherwise, failures may occur,
reducing die life and increasing tooling costs.
The following sections describe some typical
failures of tool steels after heat treatment and
the main mechanism that caused failure. The
main mechanism is described and divided systematically, but in practical situations, several
mechanisms as well as the tool use should be
(a)
(b)
Fig. 23
(a) Tool called a triturating wheel made from AISI D6 (similar to D3) that had poor performance due to a premature crack.
Tool diameter of approximately 300 mm (~11.5 in.). The identified cause was the absence of tempering. The hardness
measurement was 65 HRC, the usual value for the as-quenched condition of this grade. Courtesy of Villares Metals. (b) D2 dies cracked
during finish grinding operation. Cracks emerged due to the as-quenched condition of the microstructure (not tempered). Source: Ref 9
pg 332
18/8/2008 3:40PM Plate # 0
pg 333
tool steel. In Fig. 24, this can be observed for

high tempering temperatures, but the variation
is even higher if the whole tempering curve is
observed (see the tempering curve for H13,
Fig. 28, for example). Normally, it is possible to
attain values between 30 and 58 HRC in the most
common hot work steels. However, the steels are
not used within this hardness range. Although
some variation may exist, hardness levels higher
than 50 HRC or lower than 40 HRC are not
typical.
To improve wear resistance, one can propose
an increase in hardness. However, two aspects
should be considered before such a decision.
First and most obvious is the toughness necessary for a given application. Reduction of
investigated to discover the root cause of a

failure.
In specialized literature, it is less common to
discuss failures of hot work tools than of cold
work. This is because of the intrinsic lower
brittleness of hot work tool steels. Nevertheless,
the subject is very important for these materials,
due to the high value of a tooling set for hot
working and also because of the large production performance expected from it. Even if
no catastrophic failure occurs, the reduction of
hot work tool performance can cause serious
damage in terms of cost and time.
Premature Cracking Caused by Excessive
Hardness. A wide range of hardness levels can
result from hardening and tempering a hot work
Fig. 24
Tempering curves for the most common hot work tool steels. Tempering curves are obtained after hardening small (25 mm,
or 1 in.) specimens of all materials with the usual hardening temperature: 1020 C for H13, TENAX300 (brand name of lowsilicon H11), and VHSUPER (brand name of high-molybdenum, low-silicon modified H11); and 1100 C for H21 (higher temperatures
may be used, increasing tempered hardness but reducing toughness).
Table 3 Typical chemical composition of some hot work tool steels

Composition, wt%
AISI(a)
~H10
H11
H13
Low-Si H11;
TENAX300(b)
~6F3 or ~L6
H21
DIN
UNS
Si
Mn
Cr
Fe
Others
1.2365
1.2343
1.2344
T20810
T20811
T20813
...
0.32
0.37
0.40
0.36
0.30
1.0
1.0
0.3
0.3
0.3
0.3
0.3
2.9
5.0
5.0
5.0
Mo
2.8
1.3
1.3
1.4
...
...
...
...
0.50
0.90
0.45
0.45
bal
bal
bal
bal
...
...
...
...
1.2714
1.2581
...
T20821
0.56
0.32
0.3
0.3
0.7
0.3
1.1
3.5
0.5
...
...
9.0
0.15
0.50
bal
bal
Ni = 1.7
...
(a)~, similar to but not exactly the same. (b) Trademark of Villares Metal Company, Brazil. TENAX 300 is not standardized; therefore, the brand name is given.
18/8/2008 3:40PM Plate # 0
Fig. 25
Examples of heat checking cracks on aluminum die-casting dies. Cracks are white because they are filled with aluminum.
Courtesy of Villares Metals
Fig. 26
Schematic of hardness after tempering and the effect

of secondary hardening in high-alloy steels. Observe
that high-temperature hardness is only possible through precipitation hardening caused by alloy carbides (secondary hardening). At low temperatures, hardness is less than martensite due
to the presence of retained austenite, which is eliminated after
high-temperature tempering.
toughness may accelerate several mechanisms

of tool damage, such as heat checking, gross
cracking, and wear by chipping, in addition to
the risk of catastrophic cracking.
The relationship between hardness, toughness, and tempering temperature can be quantitatively evaluated in Fig. 28 for AISI H13. When
the tempering temperature is far from the hardness peak, toughness rapidly increases with
decreasing hardness. Temperatures of approximately 500 C can be considered to cause temper embrittlement, with toughness being rather
low (Ref 2426). Such temper embrittlement
has been studied in relation to silicon content
(Ref 2730). It has been shown, for example,
that reducing the silicon content from 1% (usual
in H-series grades) to approximately 0.3%
causes a strong toughness increase, close to peak
hardness but also for higher tempering temperatures.
Secondly, the type of wear should be considered. In cold work tooling, wear is directly
related to hardness, but in hot work tooling, the
situation is normally quite different. Wear can

occur by a combination of abrasion and adhesion
at high temperatures. The main reason to avoid
wear is the ability to keep high hardness at
working conditions (tempering resistance) as
well as adequate toughness to avoid chipping
after adhesion. One example of such an effect is
presented in Fig. 29 for a precision hot forging
punch. The traditional material for such an
application was H13 steel with high hardness
(55 HRC). However, accelerated wear was
observed. Analysis of the tool after end-life
showed a strong hardness decrease in the working regions, indicating that a higher tempering
resistance was necessary instead of higher
hardness. Such an alteration was done, substituting H13 with a higher-molybdenum
grade, the commercial brand VHSUPER (not
standardized), which led to a 50% increase in the
tool performance.
It is commonly found that excessive hardness assists failures in hot work applications.
Figure 30 shows further examples. For all, low
tempering temperature, short tempering time,
or even specification changes led to an increase
in hardness and premature failure due to the
resulting low toughness.
While excessively low hardness can lead to
problems, hardness also should not be too low.
The adequate value, as mentioned previously,
is normally between 40 and 50 HRC in hot work
tools; in some cases for warm forging, it
is possible to use tools up to 58 HRC. An interesting effect of hardness and toughness in final
performance can be discussed based on Fig. 31.
This graph shows the effect of these two properties on heat checking, focused on die-casting
applications. The higher the hardness and
toughness, the lower the heat checking damage,
measured by an appropriate chart. However, as
discussed and shown in Fig. 28, hardness and
toughness are also associated. Establishing ideal
pg 334
18/8/2008 3:40PM Plate # 0
hardness thus depends on several conditions of

tool design and use and is also based on previous
experience. As a starting value, 45 HRC would
be recommended, with increases or decreases
depending on the results experienced.
Inadequate Heat Treating Procedures in
Hot Work Tool Steels. Heat treating of hot
Fig. 27
work steels has several important parameters,

but incorrect practice does not always produce
failures observable just after heat treating. Some
examples were shown in Fig. 29 and 30; several
tools were improperly heat treated (to higher
hardness), but the problem was only observed
during tool use.
Effect of vanadium molybdenum, tungsten, and chromium additions on secondary (high-temperature) hardness of mediumcarbon steels. Source: Ref 23
pg 335
18/8/2008 3:40PM Plate # 0
This is common for hot work tools due to the

low as-quenched hardness and the high hardenability of this class of tool steels. For example,
the use of improper temperatures may lead to the
specified hardness, which can even be homogeneous, but the mechanical properties, such
as toughness or tempering resistance, can be
deeply affected. A poor practice during quenching can lead to high brittleness with no observed
changes in hardness or strength. These two aspects are described in this section.
Before continuing, a useful recommendation
should be made. Due to the high value of diecasting tools and their high productivity, a strong
effort has been made for improvements in this
field. An important recommendation for H13
tool steel was written by the North American
Die Casting Association (NADCA) (Ref 32).
This recommendation provides important
Fig. 29
Fig. 28
Impact toughness and hardness as a function of

tempering temperature. Retained austenite content is
also shown. Notice the hash-marked area, indicated as a temper
embrittlement region, where very low toughness is observed; this
region coincides with the peak hardness. Source: Ref 24
Example of the importance of tempering resistance

instead of initial working hardness. (a) Hot forging
punch showing wear and cracks as the normal failure condition.
For maximum wear resistance, initial hardness was established at
56 HRC for H13 tool steel. However, the end-life mechanism was
related to hardness reduction in the working (heated) areas, as
shown in (b). This grade was substituted by a higher-molybdenum
grade, brand name VHSUPER, with higher tempering resistance,
as shown in (c) by the longer times necessary for hardness
decrease. The substitution lead to 50% longer tool life. Courtesy
of Villares Metals
pg 336
18/8/2008 3:41PM Plate # 0
information for quality assurance of tool steel

and also for applied heat treating. It is very
helpful for die-casting applications as well as for
quality analysis of AISI H-series steels applied
to other processes.
Incorrect Hardening and/or Tempering
Temperatures. To analyze the effect of incorrect temperatures on hot work steel properties,
several studies were conducted in the laboratory (Ref 12). Toughness was evaluated by
unnotched impact testing, according to the
NADCA procedure (Ref 32). Three conditions
were simulated. To show the effect of low
hardening temperature on tempering resistance,
conditions 1 and 2 were evaluated:

Condition 1: standard condition for H13

grade, with hardening temperature at
1020 C and double tempering at 610 C
(for 2 h at temperature), leading to 45 HRC
Condition 2: Low hardening temperature (at
890 C), simulating a condition for furnaces
Fig. 30
that reach up to 900 C. For this condition,

tempering should be reduced to 250 C.
Condition 3: higher hardening temperature,
increasing from 1020 C to 1150 C. Such
temperature is currently used for high-speed
steels heat treated to lower hardness. For 45
HRC, tempering was slightly increased, to
640 C, also twice for 2 h.
Figure 32 evaluates the reduction in toughness

promoted by inadequate heat treating conditions,
as well as the respective microstructures. The
low-tempering situation, condition 2, causes a
substantial reduction in toughness (40%) as well
as a loss in tempering resistance (the hardness
decrease was six times higher than expected).
Toughness reduction in this condition, is caused
by incomplete austenite transformation, causing
a heterogeneous microstructure (Fig. 32c), as
well as by the low tempering temperature, which
does not promote adequate martensite stress
relief. On the other hand, the decrease in
Examples of failures caused by excessive hardness. (a) Tool made of DIN 1.2714 tool steel (similar composition to AISI 6F3
and L6) that fractured after a short life. For this tool, the hardness was expected to be approximately 40 HRC, but a sample was
analyzed and found to be 50 HRC. Arrows indicate cracking location and cracking initiation site. (b) Microstructure showing light areas,
indicating excess retained austenite and untempered martensite, another indication that low tempering temperature was employed and/
or only one tempering treatment. This led to high hardness as well as a brittle microstructure. (c) H13 forging die that cracked prematurely
(arrows). Hardness was measured at 52 HRC but expected values were approximately 44 HRC. The excessive hardness was caused by
short tempering times and low temperatures. (d) Typical H13 microstructure tempered at low temperatures. Etched with 4% nital. It is
lighter than usual (compare to Fig. 32b) due to low-temperature tempering, which causes poor precipitation of alloy carbides and thus
enhances corrosion resistance during etching. Courtesy of Villares Metals
pg 337
18/8/2008 3:41PM Plate # 0
tempering resistance is caused by the insufficient

dissolution and reprecipitation of alloy carbides
(poor secondary hardening), which is the main
strengthening mechanism in hot work tool steels
at high temperatures.
The most intense embrittlement was produced by condition 3too high hardening
temperature. In this situation, very coarse austenite grains are produced (Fig. 32d), leading to
increased grain-boundary embrittlement. When
high hardening temperatures are used, precipitation of proeutectoid carbides on grain boundaries intensifies, causing a great reduction in
toughness, as observed (90% lower).
Some sources studied the use of higher hardening temperature as one way to improve thermal fatigue (Ref 22, 33). In fact, the increase
in hardening temperature leads to more dissolution of alloy carbides, rich in vanadium and
molybdenum, which increases the content of
alloy elements in solid solution and enhances
secondary hardening. A simple observation of
this is shown in condition 3, where tempering
should be increased 30 C to attain the same
hardness as the usual hardening condition.
In terms of tempering resistance, 30 C (54 F)
Fig. 31
the dislocation in tempering curve indicates a

substantial increase in tempering resistance,
because temperature effect is exponential to
time effect in tempering conditions. This phenomenon also explains some advantages found
in specimens austenitized at 1100 C compared
to 1020 C (Ref 22). However, modifications in
hardening temperatures are rarely possible in
practical (industrial) conditions. Increasing
the hardening temperature deeply affects the
precipitation behavior on grain boundaries during quenching, causing intense embrittlement.
Figure 33 shows this effect (note the dashed
lines), but a full explanation is given in the following section (especially regarding Fig. 36).
Before continuing, it is interesting to show a
case of failure caused by excessive hardening
temperature. Figure 34 presents such a casea
tool that cracked after low performance. It is
easy to see the grain-boundary crack propagation, caused by the coarse grain size as well as by
precipitation of carbides on grain boundaries.
This is denoted by the preferential and strong
etching of austenite grain boundaries.
Slow Cooling during Quenching. Reaching the final hardness in tool steels is quite a
Heat checking resistance (lower readings indicate higher resistance) as a function of unnotched impact toughness and
hardness of H13 steel. Heat checking is evaluated by the photographs on the left; the rating is calculated by adding the
column representing the largest cracks (leading) and the column representing the severity of the cracks (network). See text for discussion of
these results. Source: Ref 31
pg 338
18/8/2008 3:41PM Plate # 0
simple task for quenching. Although there are

some exceptions, tool steels normally have
high hardenability, with the as-quenched hardness attained even if improper procedures are
applied. However, hardness, although necessary,
is not sufficient for the high required performance
of tool steels, as already shown. In this context,
quenching should be considered as a process for
promoting the required mechanical properties,
not only for attaining a specified hardness.
For this reason, consider Fig. 33 again. For
both hardening temperatures, hardness higher
than 500 HV (~49 HRC) is obtained within a
wide range of the continuous cooling tranformation (CCT) curve. Considering the hardness
reduction after tempering, it is therefore quite
simple to obtain the final hardness with different
(b) condition 1
Fig. 32
(c) condition 2
quenching practices, even using air quenching.

However, two further important aspects should
be considered. First, the dashed lines in the CCT
diagram (Fig. 33) indicate formation of proeutectic carbides. As mentioned in the previous
section, if austenitizing temperature increases,
more alloy elements go into solid solution by
carbide dissolution. In cooling, the process is
reversed, and such carbides tend to form again.
This happens by precipitation in high-energy
areas, the most important being the grain
boundaries. The result is a film of carbides
between grains, which weakens the interface
and promotes failure (Ref 35). Such a phenomenon is marked by two characteristics: a strong
etching at austenite grain boundaries (because
the interfaces of carbides and steel are regions
(d) condition 3
Laboratory simulation of adequate and inadequate heat treating conditions for AISI H13. The first situation (condition 1) is the
recommended heat treatment: hardening at 1020 C, followed by two tempering treatments at high temperature. In this case,
45 HRC was desired, and thus tempering was performed at 610 C. Condition 2 involves a very low hardening temperature, where
austenitizing was done at 890 C. To reach 45 HRC, specimens were heat tempered at 250 C. In addition to toughness reduction, the
heat treating condition caused reduction of tempering resistance. Condition 3 describes a situation with excessively high hardening
temperature (1150 C), with tempering done at 640 C to attain 45 HRC. In (a), the impact toughness is presented, and in (b) to (d), the
microstructure relative to each condition is shown (same magnification; etched with 4% nital). Source: Ref 12
pg 339
18/8/2008 3:41PM Plate # 0
more prone to corrosion) and, in stronger cases,

intergranular failure. As shown in Fig. 33, an
increase in austenitizing temperature causes
dislocation of the dashed lines to the left,
indicating stronger precipitation, even when
high cooling rates are applied. Secondly, a slow
cooling also affects the previous microstructure,

forming bainite instead of martensite. Although
the precipitation necessary for secondary hardening is practically not affected (Ref 33),
modification of the initial microstructure, from
martensite to bainite, also reduces toughness
Continuous cooling transformation diagrams for H13 tool steel austenitized at 1030 C (1885 F) (top) and 1100 C
(2010 F) (bottom). Note the dislocation of the dashed line, indicating more pronounced proeutectic carbide precipitation on
grain boundaries for the high austenitizing temperature condition. Source: Ref 34
Fig. 33
pg 340
18/8/2008 3:41PM Plate # 0
(ref 29, 33 and 36). The effect of decreasing the

cooling rate is thus embrittlement for both
mechanisms.
Fig. 34
Die failure caused by excessive hardening temperature. Two tools were analyzed: one made of
VHSUPER steel (commercial brand name) and the other of AISI
H13. (a) One of the tools cracked in the position denoted by the
arrow, where a sample was cut for analysis. (b) Typical microstructure from H13 tool and (c) from VHSUPER tool with 54 HRC
(45 HRC was expected). Note the coarse grain size, approximately ASTM 3 to 4. For these grades, grain size is expected to be
approximately ASTM 7 to 10. Courtesy of Villares Metals
Therefore, both bainite and carbide precipitation on grain boundaries must be avoided
by preventing slow cooling in quenching. This is
important advice for tools sensitive to failures
caused or assisted by cracks. One example
is presented in Fig. 35 for hot forging dies
that failed after short service time. The
microstructural analysis showed coarse grain
sizes and strong precipitation on grain boundaries (Fig. 35b, c), illustrating the interaction
between the two effects. As a final result, strong
embrittlement occurs (Fig. 35d) as well as a
clearly intergranular fracture (Fig. 35e).
Typically, hot work tool steels were oil
quenched, but today (2008), vacuum heat treating with pressurized nitrogen quenching has
become very popular. In this treatment, cooling
rate control is rather critical, since it is related
to nitrogen pressure and gas circulation as well.
If too strong and heterogeneously applied,
cooling may lead to strong distortion or even
quenching cracks. On the other hand, grain
boundary embrittlement occurs easily if the
cooling rate is too slow. A guide for evaluating
tool heat treating quality is described by the
NADCA (Ref 32), including the use of coupons
for destructive testing after heat treating as well
as advice for vacuum hardening.
Another important issue is the step in which
heat treatment should be applied. With the
advance in machining technology, the feasibility
for machining in higher hardness has increased;
machining hot work dies up to 50 HRC is rather
common by means of high-speed machining
technologies (high cutting speeds with low
feed). Consequently, it is common, mainly in
forging dies, to machine from prehardened
blocks. However, the probability of embrittlement increases as the section size of the tool
increases. Figure 36 shows the effect of section
size and austenitizing temperature on the toughness of H13 steel. The tendency for toughness
loss is evident when larger sizes or higher
austenitizing temperatures are used, because
they are directly related to the grain-boundary
embrittlement effect and are also affected by
bainite formation. Even if the quenching process
uses a strong cooling medium, large tools are
unavoidably sensitive to embrittlement in core
regions. Therefore, heavy-section tools with
deep engravings should be heat treated only after
rough machining to avoid embrittlement of tool
working regions. The most important example
in this field is die-casting dies. They are usually
heat treated only after machining to improve
pg 341
18/8/2008 3:41PM Plate # 0
(a)
Fig. 35
Example of die failure in a hot forging die caused by coarse grain size and strong precipitation of proeutectoid carbides on
austenite grain boundaries. (a) Aspect of the tool. (b) and (c) Microstructure showing the coarse grain size (approximately
ASTM 4; expecte d ASTM 8 to 10), marked by preferred etching on carbides present at grain boundaries and the coarse martensite laths.
Samples were taken from the tool midradius and analyzed regarding (d) impact toughness in the as-received condition and after new heat
treating to the same hardness and (e) fracture of impact-tested specimens (for the initial conditionas-received) by scanning electron
microscopy. Note the strong increase in toughness after new heat treating, indicating the deleterious effect of carbide precipitation on
grain boundaries, producing intergranular failure in impact specimens. Courtesy of Villares Metals
pg 342
18/8/2008 3:41PM Plate # 0
toughness, because die resistance to heat checking is directly related to this property (Fig. 33).
Even if the quench is applied after machining,
it is still important to control the Quench cooling
rate (avoiding too slow conditions), due to the
possibility of embrittlement of surface regions,
which are the working areas of die-casting dies
(Ref 32).
Failures of Nitrided Tools. Nitriding is
commonly used for several tools and dies. For
example, hot extrusion dies are typically nitrided
for all uses. They normally work for production
of parts in aluminum or other nonferrous alloys,
mainly for construction applications. Different
from other hot working processes, extrusion is
continuous and involves constant flow between
the conformed alloy, and the tool steel. This
enhances the wear condition, which is the usual
end-life mechanism. Nitriding considerably
improves surface hardness and, consequently,
wear resistance, this being the reason for its
application in virtually all extrusion dies. The
same approach is applied in hot forging tools.
Nitriding is currently used in several tools and

dies to improve wear resistance. However, as
hardness increases, surface toughness decreases,
and for deep tools, especially those prone to
cracks, nitriding is prevented. In die casting,
nitriding is used in some cases, claiming that the
increase in surface hardness tends to increase the
initiation of heat checking cracks; however, as
cracks cross the nitriding layer, it has no effect at
all. Depending on the nitriding layer condition, it
can be harmful to crack initiation.
Normally, the nitride layer in tool steels is 0.1
to 0.3 mm thick. Hardness is higher than in
carbon steels, due to the formation of alloy
nitrides. In hot work steels from the H series,
chromium is very important for this effect, and
the maximum surface hardness approaches 1100
HV. Typical nitrided microstructures present a
diffusion layer and a fine white layer on the tool
surface (Fig. 37a). In H-series tool steels, the
Fig. 36
Charpy V-notch (CVN) results for different heat

treating conditions of H13 tool steel, carried out at
room temperature and at 425 C. Specimens cooled at various
rates, simulating the core of 150 and 300 mm round bars, as well
as an air-cooled 25 mm specimen. Results were tested for different austenitizing temperatures. Toughness reduction is evident
at higher austenitizing temperatures and larger sizes, both related
to lower cooling rates during quenching. Source: Ref 33
Fig. 37
Tool steel surface after nitriding. (a) White and diffusion layers (b) Coarse nitrides precipitated on grain
boundaries. See text for discussion. Courtesy of Villares Metals
pg 343
18/8/2008 3:42PM Plate # 0
nitriding layer is darker after etching, due to the

depletion of chromium content from the steel
matrix, thus causing a reduction in corrosion
resistance. In carbon or low-alloy steels, the
nitride layer appears lighter in the microstructure and can increase corrosion resistance.
In hot work tool steels, it is rather common to
observe coarse nitrides precipitated on grain
boundaries inside the diffusion layer, as shown
in Fig. 37(b). Depending on the intensity, this
embrittles the tool and forms an exceptional
route for crack propagation (Fig. 42). Precipitation of coarse nitrides on grain boundaries
can also be accompanied by a thick white layer
on the surface. This layer is extremely hard
and brittle, because it is uniquely composed of
nitrides. It is therefore a common region for
crack initiation. The combination of a coarse
white layer and grain boundary precipitation of
Fig. 38
coarse nitrides commonly leads to premature

crack initiation, with damage to several tools,
especially forging tools prone to cracking or
for die-casting dies. Some examples of failures
assisted by this phenomenon are shown in
Fig. 38. It is thus recommended, for these
applications and in general, that the nitriding
process be controlled in order to avoid both a
thick white layer and grain-boundary nitrides.
Today (2008), this control is usually performed
in computer-controlled gas nitriding and plasma
nitriding processes.
Excessive Heating Causing Tool Failure. As discussed previously, hardening of hot
work tool steel mainly results from precipitation
of alloy carbides during the tempering treatment.
This phenomenon is also known as the fourth
stage of tempering and occurs after the modification of martensite and the formation
Examples of undesirable microstructures encountered on the surface of nitrided tools. For both cases, the core microstructures are correct, indicating proper hardening and tempering procedures. (a) Surface and (b) core microstructure of
a nitrided forging tool, showing (in a) the problems of a coarse white layer and nitrides on grain boundaries. (c) and (d) Extracted from a
die-casting die failure analysis, also for surface and core respectively. Note the strong precipitation on grain boundaries in (c), whereas
core regions are quite well heat treated (to approximately 44 HRC), leading to 300 J of unnotched (NADCA) impact strength. Nevertheless, an unexpected failure occurred (after less than 100 shots), caused primarily by improper tool use but also assisted by the nitriding
layer condition.
pg 344
18/8/2008 3:42PM Plate # 0
of cementite. At low temperatures, typically

below 500 C, the thermodynamic driving force
(free energy) is insufficient for such carbide
precipitation (mainly molybdenum or vanadium
carbides), and no hardness peak is observed
(Fig. 27). As the temperature increases, carbide
formation starts and very small carbides (a few
nanometers) precipitate within martensite laths.
Because they are fine and large in number, these
carbides are proper for obstructing dislocation
movement, causing substantial improvement in
strength. This occurs only at high temperatures,
with precipitation of alloy carbides pertaining
to materials used in hot working, in this case,
hot work tool steels. For higher temperatures,
carbides tend to transform to more stable types or
to coarsen, which increases the size and thus reduces the total number of precipitated carbides.
This results in loss of the dislocation blocking
effect, leading to strength reduction. The total
effect is schematically shown in Fig. 26.
Temperature and time affect the precipitation
behavior. From a thermodynamic point of view,
the lowest energy is obtained after alloy carbide
precipitation (reducing the free energy of elements in solid solution) and if their sizes are
as large as possible (reducing the surface free
energy). Consequently, the higher the time
or temperature, the easier it is for carbides to
become larger, and strength tends to decrease.
In tempering curves, such as those shown in
Fig. 24 and 28, this can be observed for a fixed
time (twice for 2 h) according to temperature.
However, a complete view is given by using
the tempering parameter, as shown in Fig. 39.
This combines the effects of both time and
temperature in only one variable, the parameter,
proportional to time and to a logarithm of the
temperature. Through this mathematical relation, it can be seen that temperature actually
has the highest effect; however, if excessively
long times are used at appreciably high temperatures, the same effect may occur. For example,
4 h (twice 2 h tempering) at 600 C (1110 F,
parameter = 32,700) leads to ~46 HRC for H11,
which is equivalent to 15 h at 577 C (1070 F),
or 3 days at 550 C (1020 F), or approximately
1 month at 500 C (930 F). That is why
below 500 C, heating has practically no effect.
However, a second situation should also be
considered. Low times at excessively high
temperatures can produce rather important
effects. This is the situation for several tool
applications, where tools are in contact at very
high temperatures but for very short times. For
example, consider a forging tool that reaches

750 C (~1380 F) for a half-second during each
stroke. After producing 1000 parts, this tool will
be exposed to 750 C for a total time of 8.3 min.
Such heating is equivalent to a tempering parameter of 35,260 or 2 h at 691 C (1275 F). In
H11, such heating is capable of reducing the
hardness from 46 to less than 30 HRC.
It is in this context that the failures described
in this section should be considered. In some
situations, tools are correctly heat treated, but,
during hot working operation, heating is so
high that strength is reduced and failure is
accelerated. One example, shown in Fig. 40, is
a hot forging punch that usually exhibits low
life if compared to the whole set of tools. It is
made of DIN 1.2885, a highly temper-resistant
hot work steel that has high amounts of
molybdenum and cobalt for improvement of
this property. Even though the analysis showed
a continuous reduction of hardness approaching
the punch tip, which contacted the hot inforging part. Besides the hardness values, it is
interesting to note the microstructural behavior.
Regions far from the hot working areas are
typically tempered martensite, becoming darker closer to the punch tip. As previously
explained, secondary carbides are not visible
Fig. 39
Tempering curve as a function of the time-temperature parameter for H11 steel containing 0.40% C,
0.92% Si, 5.09% Cr, 1.34% Mo and 0.52% V. For this curve,
t = time in hours, and T = temperature in F+460. Source:
Ref 37
pg 345
18/8/2008 3:42PM Plate # 0
under optical microscope; however, their effect

on corrosion resistance (and thus on etching
behavior) is evident. High tempering temperatures cause an increase in precipitation and
carbides, leading to reduction of corrosion
resistance due to the depletion of alloy elements in the steel matrix and to the increased
sites (interfaces) for corrosion. This explains
why the microstructure becomes darker where
heating is more intense and hardness is lower
(this will always happen if the hardening
practice is not changed). In the punch in
Fig. 40, heating is so high that secondary
carbides have become large enough to be
observable by optical microscope. This is
typically the annealed state for this alloy, not
obtained through transformation annealing but
by temper annealing.
The effect shown in Fig. 40 occurs in several
hot working tools. A further example is shown in
Fig. 41 and 42 of a forging die for automotive
valves. After some time in production, wear of
the forging die produced grooves on the surface
of forged products (Fig. 41a), leading to die
substitution. The analysis revealed that wear
Fig. 40
was deeply influenced by thermal fatigue cracking (Fig. 41c, d). By comparing heated and
not heated areas, two further problems were
observed. First, some regions on the die surface
had been heated to high temperatures, leading to
reaustenitizing, rehardening, and the formation
of untempered martensite (Fig. 42a). Second, the
tool surface had serious indications of nitriding
problems, with coarse nitride formed on grain
boundaries and crack propagations through
these regions (Fig. 42c). The combination of
intense surface heating and nitriding led to a
crazy pattern of hardness, differing substantially from that encountered in non-heat-affected
areas (Fig. 42b). This was caused by the combination of rehardening and nitrogen diffusion to
the core regions. The end result, of course, was a
deep embrittlement of the die, as observed by its
intense cracking.
In cases similar to those in Fig. 40 to 42, two
possibilities are possible to solve or reduce the
problem. The first is to employ one steel grade
with a higher tempering resistance, normally a
higher-alloyed grade; however, this approach
has some limitations. If the situation is only an
Microstructures of hot forging punch made of DIN 1.2885. Etched with 4% nital for the same time in all microstructures
pg 346
18/8/2008 3:42PM Plate # 0
increase in tool resistance to softening, changing

the steel grade may be a solution. For example,
this was the situation described in Fig. 29.
However, it should be noted that highly alloyed
grades can also be more brittle, leading to
problems if gross cracking or heat checking
are important issues for tool failure. Also,
the increased alloy is directly related to steel cost,
with an analysis of the cost and benefits of steel
substitution being important. The best way is to
consider the total tool cost for each produced
part, which includes the increase in steel cost
but also its higher performance as well as the
reduction of setup times for tool substitution.
Fig. 41
However, in some situations, heating is so

intense during tool application that it is impossible to solve the problem only by modifying
the tool steel employed; thus, there is a second
possibility concerning cooling applied during
hot work tooling. Coolants and lubricants are
both important in hot work tooling, in many cases
being the same product. In cases where appreciable regions of tools are reaching more than
700 C or even reaustenitizing, the best practice
is to work on the lubricating/cooling involved in
the process, before considering changes to the
tool steel used. It would be helpful to contact the
lubricant manufacturer and obtain updated
(a) Forging tools for production of automotive engine valves; the analyzed die was painted in blue. (b) Forged valve and, in
detail, the grooves produced when using worn dies. (c) Three analyzed regions, marked A, B and C. A is the main
wear region, but B is where a large amount of thermal fatigue cracking is observed. C is a region not affected by process heat and used
as a reference. (d) Thermal cracks of region B, under scanning electron microscopy. Courtesy of Villares Metals
pg 347
18/8/2008 3:42PM Plate # 0
(b)
Fig. 42
(a) Microstructure of working regions of valve forging die in Fig. 41 (region A). Note the intense cracking and thick nitrided
layer (double the expected). Also note the surface white layer, which corresponds to brittle untempered martensite, obtained
by rehardening of the tool surface. (b) Hardness profile of this region (A) and another region not affected by heat (C) in Fig. 41(c). (c) Two
pictures of region C, not affected by heat, that have a mechanical crack within the nitriding layer. Note the presence of a coarse white
layer and nitrides on the grain boundaries. The micrographs in (c) refer to the same region, but the one on the right is slightly underfocus for
better observation of nitrides on the grain boundaries. Courtesy of Villares Metals
pg 348
18/8/2008 3:42PM Plate # 0
information on the possibilities for a specific

application. Besides the type of lubricant, a very
important issue is the way the lubricant is
applied. In modern processes, robots may be
used for homogeneous application of lubricants.
It is also important to evaluate the possible
changes in tool design to avoid heat during hot
working. The answer is complex and should
consider tool life, process productivity, and the
necessary investments, always aiming for the
minimum cost per produced part.
4.
5.
6.
Conclusion
All theory and data discussed in this chapter
may be summarized by the following points:

Heat treating of tools and dies may be considered crucial for their performance. Several examples show that success cases as
well as premature failures are often related to
heat treating quality.
Hardness measurements alone are usually
not the best indication of proper quality in
too steel heat treating. Other properties such
as strength, toughness, and wear resistance,
as well as the microstructural features, are
better indicators as to whether a heat treatment was done correctly or not.
Assessing these properties is usually difficult
in tools and dies because it would be a
destructive (and expensive) test. Therefore,
guaranteeing the correct procedure is the
best way to assure that the heat treatment of a
tool was performed correctly and thus
enables adequate tool performance.
In order to apply such correct conditions, the
best approach is to follow the specifications
for each grade which is normally provided
by the tool manufacturer. For several specific points such as design, tool use, and
surface treatments, the examples given with
this chapter may be helpful.
7.
8.
9.
10.
11.
12.
13.
14.
15.
REFERENCES
1. G. Roberts, G. Krauss, and R. Kennedy, Tool

Steels., 5th ed. ASM International, 1998
2. Tool Steels, Steel Products Manual, Iron and
Steel Society, April. 1988
3. R.A. Mesquita and C.A. Barbosa, Avaliacao
das Propriedades de Desgaste e Tenacidade
em Acos para trabalho a Frio (Evaluation
of Wear and Toughness Properties of Cold
Working Tool Steels). Tecnol. Metal.
16.
17.
Mater., Vol 3 (No. 2), Oct/Dec 2005, p 1218

(Abstract in English; Article in Portuguese)
C.J. Altstetter, M. Cohen, and L. Averbach,
Effect of Silicon on the Tempering of AISI
43XX Steels, Trans. ASM, Vol 55, 1962,
p 287
G. Dieter, Mechanical Metallurgy, 3rd ed.,
McGraw-Hill Book Co., 1986
R.M. Hemphill and D.E. Wert, Impact and
Fracture Toughness Testing of Common
Grades of Tool Steels, Tool Materials for
Molds and Dies, G. Krauss and H. Nordberg, Ed., Colorado School of Mines Press,
1987, p 6669
Metals Handbook, 8th ed., American
Met. Eng. Q., Feb 1973, p 3141
J.R. Davis, Ed., Tool Materials, ASM Specialty Handbook, ASM International, 1995
Heat Treatment of Tool Steel, Technical
brochure, Uddeholm Tooling, http://www.
uddeholm.com/files/heattreatment-english.pdf (accessed Dec 2006)
Grinding of Tool Steel, Technical brochure, Uddeholm Tooling, http://www.
uddeholm.com/files/grinding-english.pdf
(accessed Dec 2006)
R.A. Mesquita, D.R. Leiva and C.A. Barbosa, Estudos de Tratamento Termico nos
Acos Ferramenta VH13ISO E VF800AT
(Heat Treating Studies of VH13 and
VF800AT Tool Steels), Proceedings of
Third Encontro da Cadeia de Ferramentas,
Moldes e Matrizes, ABM, 2005, p 3040,
(Abstract in English; article in Portuguese)
Cold Work Tool Steel, VF800-AT Datasheet, Villares Metals Company, http://www.
villaresmetals.com.br/ (accessed Dec 2006)
K.W. Andrews, Empirical Formulae for
the Calculation of Some Transformation
Temperatures, J. Iron Steel Inst., Vol 203,
1965, p 721727
W.J. Harris, Jr. and M. Cohen, Stabilization
of the Austenitic-Martensite Transformation. Trans AIME, Vol 80, 1949, pp. 447
470.
B.L. Averbach, S.A. Kulin, and M. Cohen,
The Effect of Plastic Deformation on Solid
Reactions, Part II: The Effect of Applied
Stress on the Martensite Reaction, Cold
Working of Metals, American Society for
Metals, 1949
Heat Treating, Vol 4, ASM Handbook,
ASM International, 1991
pg 349
18/8/2008 3:42PM Plate # 0
18. Cold Work Tool Steels, Datasheets,

Uddeholm Tooling, Hagfors, Sweden
19. A. Mendanha, H. Goldenstein and C.E.
Pinedo, The Role of Microstructure on the
Toughness Behaviour of AISI D2 Cold
Work Tool Steel, Proceedings of the Seventh
International Tooling ConferenceTooling
Materials and Their Applications from
Research to MarketTool 06, Vol 2,
(Politecnico de Torino, Torino, Italy), 2006,
p 813819
20. M.A.H. Howes, Heat Checking in Die
Casting Dies, Die Cast. Eng., March-April
1969, p 1216
21. T.C. Benedyk, D.J. Moracz, and J.F. Wallace, Thermal Fatigue Behaviour of Die
Materials for Aluminum Die Casting, Paper
111, Proceedings of Sixth SDCE International Die Casting Congress, 1970
(Cleveland, OH), The Society of Die Casting Engineers, Inc., 1970, p 120
22. J. Sjostrom and J. Bergstrom, Thermal
Fatigue Testing of Chromium Martensitic
Hot-Work Tool Steel after Different
Austenitizing Treatments. J. Mater. Process. Technol, Vol 153154, 2004, p 1089
1096
23. W. Crafts and J.L. Lamont, Secondary
Hardening of Tempered Martensitic Alloy
Steel, Vol 180, TMS-AIME, 1949, p 741
24. Hot Work Tool Steels, Datasheets,
Uddeholm Tooling, Hagfors, Sweden
25. J.R.T. Branco and G. Krauss, Toughness of
H11/H13 Hot Work Tool Steel, New
Materials Processes Experiences for Tooling. H. Berns, M. Hofmann, L.-A. Norstrom; K. Rasche, and A.-M Schiner, Ed.,
Materials Search (Interlaken, Switzerland),
1992, p 121134
26. H. Berns, Strength and Toughness of Hot
Working Tool Steels, G. Krauss and
H. Nordberg, Ed. Tool Materials for Molds
and Dies: Application and Performance,
(Ilinois), The Colorado School of Mines
Press, 1987, p 4565
27. W.M. Garrison, Jr., Influence of Silicon
on Strength and Toughness of 5wt-%Cr
Secondary Hardening Steel, Mater. Sci.
Technol., Vol 3, April 1987, p 256259
28. D. Delagnes, P. Lamesle, M.H. Mathon, N.
Mebarki and C. Levaillant, Influence of
Silicon Content on the Precipitation of
Secondary Carbides and Fatigue Properties
of a 5% Cr Tempered Martensitic Steel,
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Mater. Sci. Eng. A, Vol 394, 2005, p 435

444
M. Umino, T. Sera, K. Kondo, Y. Okada and
H. Tubakino, Effect of Silicon Content on
Tempered Hardness, High Temperature
Strength and Toughness of Hot Working
Tool Steels, Tetsu-to-Hagane (J. Iron Steel
Inst. Jpn.), Vol 89, (No. 6), June 2003,
p 673679
R.A. Mesquita and C.A. Barbosa, Effect of
Silicon and Phosphorous on the Toughness
of G11 Hot Work Tool Steel, Proceedings
of 61st Brazilian Metallurgical and Materials Science Congress, ABM, 2006 (in
Portuguese)
L. Eliasson and O. Sandberg, Effect of
Different Parameter on Heat-Checking
Properties of Hot-Work Tool Steels, New
Materials and Processes for Tooling, H.
Berns, H. Nordberg and H.-J. Fleischer, Ed.
(Bochum, Germany), Verlag Schurmann
and Klagges KG, 1989, p 17
Premium Quality H13 Steel Acceptance
Criteria for Pressure Die Casting, 207/
2003, North American Die Casting Association (NADCA)
M.L. Schmidt, Effect of Austenitizing
Temperature on Laboratory Treated and
Large Section Sizes of H-13 Tool Steel,
Tool Materials for Molds and Dies., G.
Krauss and H. Nordberg, Ed. (Illinois)
Colorado School of Mines Press Center,
1987, p 118164
K.-E. Thelming, Steel and Its Heat Treating, 2nd ed., Butterworths, London, 1984
C.L. Briant and S.K. Banerji, Intergranular
Failure in Steel: The Role of Grain Boundary Composition, Int. Met. Rev., (No. 4),
1978, p 164199
T. Okuno, Effect of Microstructure on the
Toughness of Hot Work Tool Steels, AISI
H13, H10 and H19, ISIJ, Vol 27, (No. 1),
1987
J.C. Hamaker, Jr., Die Steel Useful for Ultra
High-Strength Structural Requirements,
Met. Prog., Dec 1956, p 93
R.A. Mesquita, C.S. Goncalves, and C.A.
Barbosa, Effect of Hardening Conditions on
the Mechanical Properties of High Speed
Steels, Proceedings of the European Conference on Heat Treatment 2008 Innovation in Heat Treatment for Industrial
Competitiveness, Verona, Italy, 2008, CDROM
pg 350
18/8/2008 3:44PM Plate # 0

DOI: 10.1361/faht2008p351

Case Studies of Steel Component

Failures in Aerospace Applications
STEEL has limited use in modern aircraft, but

it is used in landing gear systems, arrestment
systems for naval aircraft, and for support
equipment. Because of the high strength levels
required (1790 to 2070 MPa, or 260 to 300 ksi,
yield) to maintain justification for the use of
steel, these very high-strength steels are prone to
environmental effects (stress-corrosion cracking
and hydrogen embrittlement), flaws created
during manufacturing (laps, seams, machining
gouges, and grinding), and heat treatment (distortion, decarburization, and quench cracking).
Care is needed to protect and inspect these highvalue and safety-critical steel components. The
following case histories illustrate typical failures
experienced by these high-strength steels.
The case histories in this chapter illustrate
a variety of failure mechanisms. The causative
reasons vary from manufacturing to operational
to environmental. Because of the many different
types of root causes, the failure engineer must be
aware of the manufacturing process, the assembly process, as well as the environment to which
component is exposed to effectively determine
the primary cause of failure.
Fig. 1
Failure Analysis of a Catapult

Holdback Bar
This investigation analyzed the failure of a
repeatable-release holdback bar. This bar consisted of a failed strain bar and a failed T-head
(rod end connector). These components failed
during characterization fatigue testing. The
strain bar and the T-head (rod end connector) had
been subjected to spectrum loading. This fatigue
spectrum consisted of two 365 MPa (53 ksi)
tensile loads followed by a single 469 MPa
(68 ksi) tensile load application. The strain bar
failed at 4875 cycles. The T-head failed at 3235
cycles; the desired lifetime for the strain bar and
the T-head was 4500 cycles.
The failed AISI 4340 steel strain bar and the
failed cadmium-plated AISI 4330V steel T-head
were submitted for analysis.
Figure 1 shows the as-received failed strain
bar. The fracture occurred at the aft radius of
the circumferential retainer ring groove located
near a slot in the strain bar. Figure 2 shows
the as-received failed T-head and the location
As-received strain bar from fatigue testing of a hold-back bar assembly
pg 351
18/8/2008 3:45PM Plate # 0
of the fracture. Fracture occurred in the run-out

area of internal threads at a distance of approximately 50 mm (2 in.) from the forward end.
All of the parts were subjected to a magnetic
particle inspection. In the strain bar, two cracks
were found approximately 180 apart in the retainer ring groove location shown in Fig. 3. The
Fig. 2
Fig. 3
crack lengths were 22 and 25 mm (0.85 and

0.99 in.). No cracks were detected in the T-head.
The radius at the primary fracture of the strain
bar was determined to be 1.3 mm (0.050 in.) At
the secondary cracks, the radius was 1.14 mm
(0.045 in.). At present, there is no dimensional
requirement for these radii.
As-received T-head connector from fatigue testing of hold-back bar assembly
Magnetic particle inspection indications found in the retainer ring groove on the strain bar. (a) Crack indication 1 showing a
crack length of 25 mm (0.99 in.) between arrows. (b) Crack indication 2 showing a crack length of 22 mm (0.85 in.) between
arrows. Original magnification: 1.4
pg 352
18/8/2008 3:45PM Plate # 0
Case Studies of Steel Component Failures in Aerospace Applications / 353
The fracture surfaces were examined visually

and at up to 50 magnification using a stereomicroscope. Figure 4 shows macrographs of the
fracture surfaces of both the strain bar and the
T-head. The strain bar primary fracture surface
exhibited multiple fatigue origins around the
outside periphery of the fracture. The T-head
fracture exhibited fatigue origins primarily near
internal thread roots.
A scanning electron microscope (SEM) was
used to examine the fracture surfaces and to
document the mode of failure. Figure 5 shows
the fracture topographies of the primary fracture
in the strain bar. The secondary fractures in the
strain bar are shown in Fig. 6. Fatigue striations
emanated from fatigue origins located on the
outside periphery of the fractures. Fracture ridges also gave evidence of fatigue origin locations on the outside periphery of the fracture.
Figure 7 shows topographic features of the
fracture surface of the T-head failures. Fatigue
striations emanated from multiple origins located on the inside peripheries of the fractures.
Hardness measurements were made on the
strain bar and T-heads to verify heat treat conditions. The hardness of the strain bar was
determined to be 41.6 HRC, and this met the
hardness requirement of 40 to 43 HRC. The
hardness of the T-head was determined to be
48.8 HRC. This hardness value exceeded the
requirement of 46 to 48 HRC.
Metallographic sections were prepared
through the strain bar and a typical T-head at
the locations shown in Fig. 4. The specimens
were prepared using standard metallographic
Fig. 4
techniques. Figure 8 shows the microstructures

of the strain bar and a typical T-head.
The inclusion contents of the metallographic
specimens appeared to be relatively high, as
is observed in air-melted steels. An energydispersive spectroscopy analysis identified manganese sulfide inclusions as being present.
Manganese sulfide inclusions are typically present in air-melted steels and can severely affect
fracture toughness (Ref 1). Inclusions can act as
initiation sites for fatigue failures; however, the
failure of these parts could not be associated
with the presence of inclusions.
Banding was also present in both the strain
bar and T-head metallographic specimens.
Microhardness measurements made on the light
and dark bands in the strain bar and T-head
yielded hardness readings of 44.1 HRC (light),
42.6 (light), 49.8 (dark), and 49.1 (dark), respectively. This microstructure is typical of an AISI
4340 component that has been improperly
normalized (Ref 2). The light etching areas are
generally untempered or lightly tempered martensite, while the darker etching regions
are martensite that has been tempered more
thoroughly. The lighter etching regions contain
higher hardenability than the darker etching
regions. This is typically due to segregation of
chromium, generally as the result of inadequate
normalizing prior to heat treatment. This type of
nonuniform structure is not optimal. The harder
regions tend to have lower fracture toughness
and have higher notch sensitivity than the darker, more tempered regions. Because of this, it
was thought that the premature fracture occurred
Macrographs of primary fracture surfaces. (a) Strain bar, with maximum depth of fatigue at arrow (4.24 mm, or 0.167 in.),
looking forward. Original magnification: 2.4. (b) T-head, with maximum depth of fatigue at arrow (2.90 mm, or 0.114 in.),
looking aft. Original magnification: 2.2
pg 353
18/8/2008 3:45PM Plate # 0
because of improper heat treatment. The presence of a higher level of inclusions could also
contribute to notch sensitivity reduction.
Cracking in a Main Landing Gear

Attach Pin
Multiple pins exhibited cracks in the flanges
during magnetic particle inspection as part of
the manufacturing process. Some parts were
returned to the vendor for rework, while others
were reworked in-house.
The pins were machined from 300M steel.
The pins were then heat treated to a strength
range of 1930 to 2070 MPa (280 to 300 ksi).
The pin was ground and chromium plated.
Figure 9 shows the main landing gear (MLG)
lever attach pin as received for examination. The
magnetic particle inspection indications in the
flange are shown in Fig. 10.
Fig. 5
After the chromium plating had been

removed, the flange area was subjected to a
temper-etch operation of 1% nitric acid in
methanol. Numerous areas on the flange etched
dark, indicating that these regions had been
overtempered (Ref 3). A micrograph of the
temper-etched inspected flange is shown in
Fig. 11.
A typical crack was opened, and the fracture
surface was examined visually and with a stereomicroscope. The fracture surface was discolored
(Fig. 12). The discoloration appeared to be similar to a temper color in steel (straw or tan). After
the chrome plating was ground to final dimensions, the attach pin had received a stress-relief
operation that could have produced this discoloration.
An SEM was used to examine the fracture
surface. Figure 13 shows the SEM fractographs
of the surface. The fracture topology was primarily intergranular.
SEM fractographs documenting typical topographic features of the strain bar primary fracture surface. (a) Overall view of strain
bar origins (micron bar is 1.5 mm long). (b) Typical origins (arrows) on outside edge of strain bar fracture (micron bar is
1500 mm long). (c) Typical fatigue striations on strain bar fracture (micron bar is 10 mm). (d) Typical worn area on strain bar fracture
(micron bar is 10 mm)
pg 354
18/8/2008 3:45PM Plate # 0
A hydrogen analysis was conducted on the

MLG lever attach pin before and after baking for
24 h at 190 C (375 F). The results of the
analysis indicated that before baking, the lever
attach pin contained 1.6 ppm of hydrogen, and
after baking, the attach pin contained 1.4 ppm
of hydrogen. These amounts of hydrogen are
considered to be low and will not produce a
hydrogen embrittlement failure (Ref 4). Also,
fractures caused by hydrogen embrittlement are
generally not discolored.
Metallographic sections were prepared
through the magnetic particle indications using
standard metallographic techniques. Shown in
Fig. 14 are areas where the chromium plating
had been reduced by possible grinding.
Figure 15 shows micrographs of a typical crack.
This crack extended through the chromium
plating and into the base metal to a depth of approximately 0.25 mm (0.010 in.). There was
no visual indication of untempered or overtempered martensite.
Fig. 6
Based on the results of this investigation,

it was concluded that the MLG lever attach pin
contained numerous probable grinding cracks in
the flange.
MLG Linear Actuating Rod and Cylinder

The part was removed from the aircraft and
was reported to have accumulated 50 flight
hours. The failed part was from an early production lot manufactured with electroless nickel
on the inner diameter surface rather than the
drawing requirement of electrolytic nickel. The
manufacturing sequences for the MLG linear
actuating piston rod cylinder were:

A 300M steel bar was machined to the required outside diameter and was then bored,
honed, and nickel plated.
The bar was swaged in an approximate
76 mm (3 in.) length on the end opposite
SEM fractographs of the fracture surfaces of cracks 1 and 2 in the strain bar. (a) Fracture surface of crack 1, showing typical
fatigue zones at arrows (750 mm). (b) Typical fatigue striations in crack 1 fatigue zones (30 mm). (c) Fracture surface at crack 2 of
the strain bar, showing typical fatigue zones at arrows (830 mm). (d) Typical fatigue striations in crack 2 fatigue zones (1.5 mm)
pg 355
18/8/2008 3:45PM Plate # 0
Fig. 7
SEM fractographs documenting typical topographic features of the T-head. (a) Typical portion of fracture surface, showing
typical origins at arrows (500 mm). (b) Typical fatigue striations on fracture surface of T-head (1 mm)
Fig. 8
Micrographs documenting the microstructure of the failed strain bar. (a) Section 1-1 from Fig. 4 showing a banded structure
(670 mm). (b) Location A, retainer ring groove (300 mm). (c) Section 2-2 from Fig. 4 showing banded structure (670 mm).
(d) Location B at origin showing banded structure (300 mm)
from the mono-ball end. The bar was supposedly swaged at 650 C (1200 F) and
then stress relieved.
The bar was then finish machined and heat

treated to a tensile strength range of 1930 to
2070 MPa (280 to 300 ksi). The part was
pg 356
18/8/2008 3:45PM Plate # 0
then magnetic particle inspected. Electrodeposited nickel was specified for plating on
the inside diameter surface.
This plating is relatively soft, ductile, and
has a relatively high melting point. However,
electroless nickel, which has high phosphorus
content, was substituted without authorization.
This plating is hard and brittle and starts to melt
at the eutectic temperature of 880 C (1616 F).
Figures 16 and 17 show the MLG linear actuating piston rod cylinder components as received for examination. Two circumferential
fractures were located approximately 165 and
215 mm (61/2 and 81/2 in.) from the mono-ball
end of the piston rod cylinder. These two fractures were joined by a longitudinal crack
approximately 50 mm (2 in.) long.
The longitudinal fracture was opened, and the
fracture surfaces were examined visually and at
up to 50 magnification using a stereomicroscope. Figure 18 shows macrographs of the
fracture surfaces. The circumferential fractures
originated and terminated on the longitudinal
crack. The longitudinal crack exhibited a single,
Fig. 9
Macrograph documenting the appearance of the

as-received main landing gear lever attach pin
intergranular-appearing origin on the inside

diameter surface of the piston rod cylinder. The
origin was located at approximately midlength
of the longitudinal crack, 200 mm (8 in.) from
the mono-ball end. An arrest mark was present
on the fracture surface at approximately 85% of
the wall thickness.
Fig. 10
Magnetic particle inspection indications on the

flange of the main landing gear lever attach pin
Fig. 11
Overtempered (darkened areas) in the flange

revealed by temper etching (2.5 mm)
pg 357
18/8/2008 3:45PM Plate # 0

surface and to document the mode of failure.
Figure 19 shows typical SEM fractographs made
of the fracture surface. The fracture topography
at the origin was intergranular. The dimensions
of the origin were 0.5 mm wide by 0.19 mm
deep (0.02 in. by 0.0075 in.). Away from the
origin was a static rupture area, which was
characterized by dimples. Approximately 85%
across the thickness, an arrest mark was present.
As shown in Fig. 19, the fracture topography in
the arrest was intergranular. Past this zone to the
edge of the specimen, the fracture topography
consisted of dimples (static rupture).
piston rod cylinder. The hardness was 54 HRC,
and this met the hardness requirement of 53 to 55
HRC.
Fig. 12
Discoloration of the fracture surface (330 mm)
Fig. 13
Topographic features of the fracture surface as observed by the SEM. (a) Overall view of the fracture surface (111 mm).
(b) Intergranular fracture observed on the fracture surface (10 mm)
Fig. 14
Metallographic examination of cracks evident in the flange of the main landing gear lever attach pin, showing loss of
chromium at cracks (20 mm)
pg 358
18/8/2008 3:46PM Plate # 0
Fig. 15
Fig. 16
Micrographs showing the morphology of the cracks. (a) Overall view (50 mm). (b) Closeup of crack (20 mm)
As-received main landing gear linear actuating

piston rod cylinder
A hydrogen analysis was conducted on the

piston rod cylinder. The hydrogen content was
determined to be 2.3 ppm. This amount of hydrogen is considered sufficient to cause hydrogen embrittlement (Ref 5).
A metallographic specimen, section 1-1, was
removed at the origin location shown in Fig. 19.
The specimen was prepared using standard
metallographic techniques. The microstructure
at the origin is shown in Fig. 20. The microstructure of the steel, which consisted of tempered martensite, was normal.
The plating at the origin consisted of three
distinct layers. The layers were subjected to an
energy-dispersive x-ray analysis. The outermost
two layers consisted essentially of nickel and
phosphorus, whereas the inner layer consisted of
nickel and iron. The outermost two layers contained cracks, and the inner layer did not. Also
evident in the micrograph is plating that had
covered a small part of the fracture surface. This
gave evidence that a crack was present prior to
plating or prior to the heating of the plating,
which occurred during swaging.
A single intergranular origin was associated
with the fracture surface, and this origin was
relatively clean. These factors suggest a delayed
mode of failure, as occurs in a hydrogen
embrittlement failure. The bulk hydrogen was

analyzed to be 2.3 ppm; this concentration
would most likely be higher at the crack tip.
The intergranular origin was separated from
the intergranular arrest zone by a ductile rupture
region, indicating that the crack had propagated
in a ductile manner before arresting. Hydrogen
most likely diffused to and concentrated at the
crack tip, which caused the intergranular arrest
region to occur. The crack then propagated to
failure by ductile rupture.
Based on the results of this investigation, it is
concluded that the MLG linear actuating piston
rod cylinder failure was most likely due to
Failure Analysis of AISI 420 Stainless

Steel Roll Pin
Several failures of AISI stainless steel roll
pins were reported. This pin is a standard part
and is used to hold pin components together by
expanding after compression. The pin is manufactured from AISI 420 stainless steel and is heat
treated to 46 to 55 HRC.
Figure 21 shows the as-received failed roll
pin. The fracture extended the length of the pin.
surface of the service failure and a laboratorycreated overload failure from the same lot of
material. Figure 22 shows the results of the SEM
examination. In this figure, the topology of
the fractures was intergranular, with some dimples present. This failure mechanism indicated
that the failure was due to overload (static
rupture).
pg 359
18/8/2008 3:46PM Plate # 0
Fig. 17
Appearance of longitudinal crack in the main landing gear linear actuating piston rod cylinder
Fig. 18
Fracture surfaces. (a) Circumferential crack 1 (4.3 mm). (b) Circumferential crack 2 (4.3 mm). (c) Longitudinal crack (5 mm)
pg 360
18/8/2008 3:46PM Plate # 0

roll pin to verify the heat treated condition.
The hardness of the roll pin was measured to
be 49 HRC, which met the hardness requirement
of 46 to 55 HRC.
A metallographic specimen was prepared
through the fracture surface. The specimen was
prepared using standard metallographic techniques. The microstructure (Fig. 23) consisted
Fig. 19
of tempered martensite, with carbides outlining the prior-austenitic grain boundaries.

This is not a typical microstructure for quenched and tempered AISI 420 stainless steel
(Ref 2).
Generally, because of their high hardenability, martensitic stainless steels can be quenched in either oil or air. Oil quenching ensures
maximum ductility and corrosion resistance. Air
SEM fractographs documenting the appearance of the fracture surface. (a) Origin location (670 mm). (b) View of box A
showing fracture origin. (c) Location A (5 mm). (d) Location B (2 mm). (e) Location C (5 mm). (f) Location D (2 mm)
pg 361
18/8/2008 3:46PM Plate # 0
quenching may cause decreases in ductility and

corrosion resistance. If slow cooled or an
inadequate quench through the critical range
of 870 to 540 C (1600 to 1000 F), this steel
will precipitate carbides at the grain boundaries
(Ref 2). Because of the presence of the intergranular carbides, it is likely that this part
experienced an inadequate quench.
Based on the results of this investigation, it
was thought that the failure mechanism was
overload, with limited ductility caused by
improper quenching during heat treatment.
Failure Analysis of a Main Landing

Gear Lever
Fig. 20
Fig. 21
Microstructure at fracture origin (12.5 mm)
As-received 420 roll pin. (a) Visual view of

as-received roll pin (2.5 mm). (b) SEM view of
as-received roll pin (200 mm)
The MLG lever was removed from service

after a hard carrier landing. Multiple cracks
developed during removal of the ion vapordeposited (IVD) aluminum with sodium hydroxide to analyze for residual stresses by x-ray
diffraction.
The lever was machined from 300M steel
forging into a hollow configuration, machined,
then heat treated to a tensile strength level of
1930 to 2070 MPa (280 to 300 ksi). The lever
was then IVD coated on the outside.
Figure 24 shows the failed MLG lever as
received for examination. The location of a
primary crack and a series of secondary cracks,
which were between the up-latch and oleo lugs,
is also shown in Fig. 24. The primary crack was
approximately 23 mm (0.9 in.) long and is
shown in Fig. 24.
The primary crack was opened, and the fracture surfaces were examined visually and at up
to 50 magnification using a stereomicroscope.
Figure 25 shows a macrograph of the fracture
surface. Also shown in Fig. 25 are the appearances of origins (some of which were discolored), which were located on the outside
surface of the lever. The origin areas had a
faceted appearance, which indicate a delayed
mode of failure, that is, stress-corrosion cracking or hydrogen embrittlement (Ref 6).
surface. Figure 26 shows SEM fractographs
documenting the topographic features of the
fracture origins. The fracture topography was
intergranular, which is indicative of a delayed
mode of failure.
The wall thickness at the origin location was
measured to be 5.8 mm (0.229 in.). The drawing
pg 362
18/8/2008 3:46PM Plate # 0
Fig. 22
SEM examination of the failed roll pin and laboratory-produced overload fractures. (a) Location A of the service failure
(20 mm). (b) Location A of the service failure showing intergranular fracture with some dimples (5 mm). (c) Laboratoryproduced overload failure showing intergranular fracture (20 mm). (d) Laboratory-produced failure showing intergranular fracture with
dimples (5 mm). Compare to (b)
Fig. 23
Microstructure of failed roll pin. Microstructure

consists of tempered martensite with carbides decorating the prior-austenite grain boundaries (10 mm)
requirements for the maximum and minimum

wall thicknesses were 5.33 and 4.45 mm (0.210
and 0.175 in.), respectively. Therefore, the
drawing requirement was exceeded by 0.48 mm
(0.019 in.).
lever to verify heat treatment. The hardness of
the lever was 54.2 HRC, and this met the hardness requirement of 53 to 55 HRC.
Hydrogen analyses conducted on the lever
yielded values of 3.0, 4.6, 3.4, and 3.5 ppm. The
average of the four values was 3.63 ppm, which
is considered high enough to produce a hydrogen embrittlement failure.
through the fracture origin area. The specimen
was prepared using standard metallographic
techniques. Figure 27 shows the microstructure
of the specimen. The microstructure of the lever
was tempered martensite, and this is normal for
a quenched and tempered high-strength, lowalloy steel.
Based on the results, it is thought that the
cracks formed in the MLG lever as a result of
pg 363
18/8/2008 3:46PM Plate # 0
Fig. 24
Appearance of the main landing gear lever showing the location of the primary and secondary cracks. (a) Overall view
(33 mm). (b) Location of primary and secondary cracks at site of ion vapor deposit (IVD) removal (10 mm)
the application of sodium hydroxide to remove

the IVD coating in addition to the residual
stresses present as a result of the hard carrier
landing. It is likely that accelerated stresscorrosion cracking occurred because of the high
residual stresses and electrolyte.
Failure Analysis of an Inboard Flap

Hinge Bolt
An inboard flap hinge bolt was found to be
failed after 286.5 flight hours. The bolt had been
machined from 4330V-modified steel bar and
pg 364
18/8/2008 3:46PM Plate # 0
Fig. 25
Appearance of primary crack removed from part.

(a) Region of primary crack where ion vapordeposited coating had been removed (13 mm). (b) Primary crack
showing branching (2 mm). (c) Fracture surface of crack after
opening (3 mm)
heat treated to the 1515 to 1655 MPa (220 to

240 ksi) tensile strength range. A nut with the
bolt was machined from 4140 steel bar heat
treated to the 1240 to 1380 MPa (180 to 200 ksi
tensile strength range.
Figure 28 shows the inboard flap hinge bolt
and nut as received for examination. The bolt
was approximately 102 mm (4 in.) long. The
bolt consisted essentially of the shank, which

was approximately 23 mm (0.9 in.) in diameter
and 64 mm (2.5 in.) long, and a threaded portion, which was 23 mm (0.9 in.) in length with
thread type 0.5000-20 UNJF-3A THD in accordance with MIL-S-8879. The unthreaded portion of the bolt was chromium plated, while the
threaded portion was aluminum IVD coated.
The fracture had occurred at the transition of the
shank to the threaded portion of the bolt. The
fracture surface was examined visually and at up
surface. Fracture ridges emanated from an origin
that was located on the outer diameter surface
of the bolt at the location shown in Fig. 29.
The origin exhibited a reflective, intergranular
appearance. Apparent corrosion products were
also observed on the fracture surface.
Residual stresses were measured on the surface of the bolt at locations 1 and 2, shown in
Fig. 28. The residual stresses were measured
using an x-ray residual-stress analyzer. Prior to
measuring the residual stresses, the chromium
plating (location 1) and the aluminum IVD
coating (location 2) were removed. The residual
stress measured at location 1 was 790+
100 MPa (115+15 ksi), which indicated that
this area was properly shot peened as required
according to the engineering drawing. The residual stress measured at location 2 was 140+
100 MPa ( 20+15 ksi). This indicated that
this area was not shot peened. The engineering
drawing does not require this location to be shot
peened.
surface and to document the topographic features. Figure 30 shows SEM fractographs taken
of the fracture surface. The fracture topography
at the origin was intergranular. The intergranular
topography extended for approximately 3/4 of
the way across the fracture surface, at which
point dimples indicative of ductile rupture were
present.
A laboratory-produced failure yielded a structure that consisted only of dimples.
The corrosion products on the service failure
fracture surface were subjected to an energydispersive x-ray analysis. Calcium, potassium,
magnesium, and chlorine were identified as
being present.
Rockwell hardness measurements were made
on the inboard flap hinge bolt to verify heat
treatment condition. The hardness of the bolt
averaged 46 HRC. This met the hardness
pg 365
18/8/2008 3:46PM Plate # 0
Fig. 26
Typical fracture features. (a) Overall view (500 mm). (b) Intergranular fracture apparent at fracture origin (7 mm)
Fig. 27
Metallographic examination of primary crack. Fracture path followed prior-austenite grain boundaries.
No precipitates in grain boundaries (50 mm) were observed.
requirement of 46 to 48 HRC for 4330Vmodified steel heat treated to the 1515 to

1655 MPa (220 to 240 ksi) tensile strength
range. The hardness of the nut was 45 HRC,
which was above the hardness requirement of
40 to 43 HRC. The high hardness value did
not appear to be detrimental.
A hydrogen analysis conducted on the hinge

bolt yielded a value of 1.5 ppm, which is considered low to cause hydrogen embrittlement.
The diameter of the bolt at the fracture was
measured to be 10.6 mm (0.419 in.), meeting
the drawing requirement of 10.7+0.25 mm
(0.422+0.010 in.). The radius at the transition
where the failure occurred was 1.5 mm
(0.060 in.), meeting the drawing requirement of
1.6+0.25 mm (0.063+0.010 in.).
A spectrographic analysis verified that the
bolt was 4330V-modified steel.
A tensile test was conducted on a production
bolt. The bolt failed in the threaded area. The
failing load was determined to be 17,600 kg
(38,820 lb). No requirement was available.
A metallographic section was removed
through the origin location shown in Fig. 29.
The specimen was prepared using standard
metallographic techniques. Figure 31 shows
micrographs taken of the fracture origin. The
tempered martensitic microstructure was normal. No untempered martensite was present.
Corrosion products were observed on the fracture surface at the origin (Fig. 32). However,
there was no evidence of pitting or other corrosion processes that would have produced the
corrosion products. No aluminum IVD was
present on the surface at the origin, because the
IVD had been removed prior to measuring
residual stresses.
Based on the results of this investigation, it is
concluded that the inboard flap hinge bolt failed
in a delayed mode of failure, which was most
likely stress-corrosion cracking. No material
anomalies were observed that would account for
the failure of the inboard flap hinge bolt.
pg 366
18/8/2008 3:46PM Plate # 0
Fig. 28
Fig. 29
As-received flap hinge bolt and nut
Fracture surface of the failed flap hinge bolt
Failure Analysis of a Nose Landing

Gear Piston Axle
This investigation determined the mode and
initiating cause of failure of a nose landing gear
(NLG) piston axle. Failure of the piston axle
occurred on the aircraft during taxiing after the
aircraft had accumulated approximately 60 h in
service. When failure occurred, the nose wheel
completely separated from the piston strut. The
NLG piston assembly was fabricated from 300M
steel and heat treated to a 1930 to 2070 MPa (280
to 300 ksi) ultimate strength range. The axle
was then shot peened and low-embrittlement
cadmium plated. White paint was applied to the
inside and outside surfaces of the axle, with
exception of the two land surfaces for the wheel

bearing cans. The failed NLG piston assembly
was removed from the wheel and submitted for
examination. Figure 33 shows the as-received
piston assembly and the failed axle.
A visual examination of the failed axle
revealed that the fracture surface followed
a circumferential path and contained a large
discolored region (Fig. 34). Except for the
discolored region, the fracture was typically
gray colored and contained chevron markings,
as shown in Fig. 35, indicating the failure originated at the discolored region. It was determined that the discolored region, from which the
failure appeared to originate, was located at the
bottom of the axle, at approximately the 6:30
oclock position. The sketch in Fig. 36 shows the
approximate failure location. Failure of the axle
initiated approximately 145 to 149 mm (5.70 to
5.85 in.) from the threaded outboard end.
An optical examination with a stereomicroscope at up to 30 magnification confirmed that
the failure of the axle originated at the discolored
region. Chevron marks were evident that indicated the crack propagated away from the ends
of the discolored region, terminating at an area
diametrically across from it, as shown in
Fig. 35(b). The discolored region extended
along the circumference of the axle for a cord
distance of 22.15 mm (0.872 in.) and a maximum depth from the inside diameter surface
of 3.38 mm (0.133 in.). Actual wall thickness
at that location was 3.81 mm (0.150 in.). The
plane of the discolored region was approximately 25 from being perpendicular to the
longitudinal axis of the axle, along machining
marks. These dimensions were shown in Fig. 33
pg 367
18/8/2008 3:46PM Plate # 0
Fig. 30
SEM examination of the flap hinge bolt. (a) Overall view of fracture surface (1.25 mm). (b) Intergranular fracture evident at
origin (10 mm)
Fig. 31
Micrographs of the fracture origin. (a) Section 1-1, Fig. 29. Fracture is intergranular at prior-austenite grain boundaries
(50 mm). (b) Location A showing evidence of corrosion product at fracture facets (25 mm)
and 36. When examined, the discolored region

showed a color variance ranging from a gold
color to reddish brown. Also, small patches of
blue were evident. This observed color variation
is a characteristic typically observed on steel
surfaces that have been exposed to elevated
temperature, such as during heat treatment
(Ref 1). The machining marks indicate a surface
much rougher than is normal.
The hardness measurements were made at
several locations around the circumference of
the axle adjacent to the fracture surface. The
hardness measured 53 to 55 HRC, which conformed to the drawing requirement of 53 to 55
HRC for 300M steel heat treated to a 1930 to
2070 MPa (280 to 300 ksi) condition.
A section containing the discolored region
was removed from the axle and examined with
the SEM. The locations where SEM examinations were performed are shown in Fig. 35(a).
Prior to SEM examination, the discolored region
was cleaned with acetone and replication in an
attempt to remove the scale from the fracture
surface. After several attempts, only a small
amount of the scale could be eliminated, indicating the scale was firmly attached to the fracture surface. The SEM fractographs adjacent to
the inside diameter surface (area 1) were primarily intergranular and appeared to be covered
with scale. At areas beyond the inside diameter
surface, the SEM fractographs revealed a mixed
intergranular with transgranular features and
patches of scale. These are shown in Fig. 37.
Along the periphery of the discolored region,
fatigue striations could be observed at a higher
magnification, as shown in Fig. 38. The depth of
pg 368
18/8/2008 3:46PM Plate # 0
the fatigue growth region was approximately

0.05 mm (0.002 in.). Rapid fracture extended
beyond the fatigue region. The SEM fractograph
in Fig. 39 shows dimpled features that characterize a ductile mode of rapid fracture. When
examining the discolored region with SEM at
low magnification, there was no evidence of
shear lip along the inner diameter surface, which
Fig. 32
Micrograph showing the microstructure at the fracture origin. Microstructure consists of quenched and
tempered martensite (25 mm). IVD, ion vapor deposited
Fig. 34
Discolored region of the fracture
indicates that the fracture did not initiate subsurface, as would be the case for delayed failure
resulting from hydrogen embrittlement.
In an attempt to determine the composition
of the observed scale or to identify any contaminant that may be associated with the fracture
surface, an energy-dispersive x-ray (EDX) analysis was performed on the discolored region.
The EDX analysis revealed no other element
except the ones common to the base metal
composition.
Fig. 33
As-received nose landing gear piston assembly and

the failed axle
pg 369
18/8/2008 3:47PM Plate # 0
Fig. 35
Fracture surface of the failed axle. (a) Black arrows show locations for SEM examination. (b) White arrows show fracture
direction and location of metallographic sections.
Fig. 36
Axle wall thickness, discolored region dimensions, and fracture location
To examine the microstructure of the discolored region, a section was removed (section
A-A, Fig. 35b) for metallographic examination.
The microstructure was tempered martensite,
which is typical for 300M steel heat treated to
the 1655 to 2070 MPa (280 to 300 ksi) ultimate
strength range (Fig. 40). A secondary crack was
observed starting at the inner diameter surface
(Fig. 40b), at a location showing large surface
irregularities. In an attempt to investigate this
secondary crack, the crack was opened and the

fracture surface examined with a light microscope and the SEM. The fracture depth was
relatively small (0.25 mm, or 0.01 in.), which
made it difficult to determine with the light
microscope if a high-temperature scale was
present. The SEM examination revealed intergranular and ductile transgranular features
(Fig. 41). These features suggest a secondary
heat treat crack.
pg 370
18/8/2008 3:47PM Plate # 0
Fig. 37
SEM fractographs showing brittle intergranular structure in discolored region of the fracture surface. (a) Intergranular fracture
partially covered with scale at the area adjacent to the inner diameter surface (10 mm). (b) Fracture surface away from the
inner diameter surface showing intergranular and transgranular features (10 mm)
Fig. 38
SEM fractographs showing fatigue growth at regions close to the outer diameter surface. (a) At boundary of discolored region
(5 mm). (b) Outside the boundary of the discolored region (5 mm)
In order to verify the material composition,

one section of the piston axle was chemically
analyzed, using the atomic absorption spectroscopy method. The results of the analysis are
shown in Table 1. These values met the requirements of AMS 6419 for 300M steel.
In summation, the results of this investigation
indicated that the failure of the NLG piston axle
was introduced from a pre-existing defect. This
defect was present on the axle prior to the final
assembly of the part to the aircraft. The defect
had a brittle intergranular fracture surface
feature and a discoloration characteristic of a
thermal or quench crack. These phenomena
demonstrated that the crack occurred on the
piston axle during or prior to the heat treatment
Fig. 39
Rapid fracture at location outside the discolored

region (5 mm)
pg 371
18/8/2008 3:47PM Plate # 0
process. Because of the machining marks and

the initiation of cracking at these asperities, it
is possible that cracking occurred during the
machining operation, in a fashion similar to the
formation of grinding cracks.
Based on the results of this examination, it is
concluded that the failure of the NLG piston
axle was due to the pre-existing crack on the
axle. This crack was created prior to or during
heat treatment of the part.
Multiple-Leg Aircraft-Handling Sling

A new multiple-leg aircraft-handling sling
that had just received proof testing was found to
have a cracked clevis. The failed adapter was
machined from 4330V-modified steel that had

been heat treated to a 1240 to 1380 MPa (180 to
200 ksi) tensile strength range.
Figure 42 shows the adapter part of the multiple-leg aircraft-handling sling as received for
examination. Also shown in Fig. 42 is the location and appearance of the crack that was at a
clevis. The crack penetrated completely through
the wall of the clevis.
The fracture surface was opened, and the
surface. Fracture ridges indicated that origins
were present on the outside surface of the clevis.
The fracture surface appeared faceted, which is
characteristic of an intergranular, delayed mode
of failure.
Fig. 40
Microstructure and secondary cracking at the discolored region. (a) Normal tempered martensite typical of 300M (25 mm).
(b) Secondary cracking apparent on inner diameter surface (100 mm)
Fig. 41
SEM fractographs of opened secondary crack. (a) Origin of secondary crack (33 mm). (b) Intergranular fracture apparent
(10 mm)
pg 372
18/8/2008 3:47PM Plate # 0

surface and the side of the clevis at the origin.
Figure 44 shows SEM photographs documenting topographic features of the fracture surface
and corrosion pitting on the side of the clevis
at the fracture origin. The pits were coated with
primer, which indicated they were present
prior to painting. Also documented in Fig. 44
are intergranular topography and corrosion products at the fracture origins. These features are
all characteristic of a stress-corrosion failure
(Ref 4).
A laboratory overload fracture produced in
the clevis was examined in the SEM. Figure 45
shows an SEM fractograph documenting dimples that are characteristic of an overload mode
of failure.
adapter to verify the heat treat condition. The
hardness of the adapter was 41.4 HRC, and this
met the hardness requirement of 40 to 43 HRC.
The hydrogen content of the adapter was
determined to be 1 ppm, which is considered
Table 1 Chemical analysis of nose landing

gear piston axle
Chemical composition, %
Si
Mn
Ni
Cr
Mo
Fe
1.42
0.72
1.79
0.68
0.49
0.008
bal
Fig. 42
As-received failed multiple-leg aircraft-handling

sling
low in relation to producing a hydrogen embrittlement failure.

through a fracture origin. The specimen was
prepared using standard metallographic techniques. Figure 46 shows the microstructure, which
was normal (tempered martensite).
Also shown in Fig. 46 is a corrosion pit that
was coated with paint, indicating the pit was
present prior to painting.
Based on the results of this investigation,
it was concluded that the adapter failed due to
stress-corrosion cracking. Corrosion pits at the
fracture origin were present prior to painting.
Failure Analysis of an Aircraft Hoist

Sling during Static Test
An aircraft hoist sling was successfully tested
to an ultimate load of 136,000 kg (300,000 lb).
However, shortly after relieving the load,
the weld on the aft right-hand fitting failed.
The fitting was fabricated from welded 17-4PH
stainless steel plates.
Figure 47 shows the as-received failed portion
of the test fixture. The failure occurred in a
consumable electrode weld that traversed
around the length of the fracture, which was
approximately 69 cm (27 in.). A dark, discolored area was present in the fractured weld
over a length of approximately 122 mm (4.8 in.)
(location C).
Lack of fusion was a general condition
of the weld. Figure 48 shows a typical area
where lack of fusion was present in the weld.
Fig. 43
Appearance of the fracture surface of the multipleleg aircraft-handling sling
pg 373
18/8/2008 3:47PM Plate # 0
Fig. 44
SEM fractographs of the service fracture. (a) Overall view. (b) Intergranular topography at origin. Original magnification:
1500 . (c) Pits on side of clevis at origin. Original magnification: 400. (d) Pit and corrosion products at origin. Original
magnification: 1000
Figure 49 documents the direction of fracture,

as evidenced by the convergence of river patterns, at a typical location. The side of the
fracture containing location C had been placed
in tension due to bending. Portions of the fracture at locations A, B, and C shown in Fig. 47
were excised and examined on an SEM. Figure 50 shows SEM fractographs documenting
topographic features of these portions of the
fracture surface. The areas at A and B showed
evidence of overload, that is, dimples. At location C in the dark, discolored area, the fracture
topography was different than at A and B. The
topography at location C was indicative of a
heated, oxidized surface.
An energy-dispersive spectrographic (EDS)
analysis conducted on the fractures at A and B
basically detected iron, chromium, nickel, and
copper. Nickel was not detected at location C.
The EDS spectra are shown in Fig. 51. The base
metal also showed the presence of iron, chromium, nickel, and copper, which are present in
17-4PH stainless steel.
Fig. 45
Dimpled rupture indicating overload failure in a

laboratory-produced failure. Original magnification:
5000
In conclusion, a different consumable welding electrode was used at location C (no nickel
present) than at locations A and B, where nickel
was present. It is recommended that 17-4PH
filler metal be used when welding 17-4PH steel.
pg 374
18/8/2008 3:47PM Plate # 0
This was apparently done at locations A and B

but not at C, where no nickel was present. Nickel
gives ductility to the weld and helps to prevent
cracking during cooling from the welding temperature (Ref 7). It appears that at location C,
where no nickel was present, a crack was
formed, and the fracture surfaces at C became
discolored and pebbled in appearance at an
elevated temperature. Lack of fusion due to insufficient heating was a general condition in the
weld, and this also contributed to the failure. In
summary, the weld, when placed under tension,
failed due to a crack and lack of fusion.
Figure 52 shows the internal gear as received

for examination. Also shown in Fig. 52 is the
appearance of typical cracks, which were located on the inside of the internal gear on areas
next to the gear spline. The cracks were located
completely around the circumference on both
sides of the gear. A crack was opened, and the
Figure 53 shows macrographs of the fracture
surface. The fracture surface was discolored
Failure Analysis of an Internal Spur Gear

This investigation analyzed cracks that were
present in an internal spur output gear. The
internal gear is part of the planetary gear system
for a canopy. The gear was made from 4340 steel
heat treated to the 1790 to 1930 MPa (260 to
280 ksi) tensile strength range. The part was
rejected after magnetic particle inspection due to
multiple crack indications along the inside surfaces next to the gear spline. The typical manufacturing sequence is forge, machine, heat treat,
grind, and plate.
Fig. 46
Microstructure consisting of tempered martensite at

fracture origin. Original magnification: 200 . Corrosion pit coated with paint is evident in the micrograph.
Fig. 47
As-received portion of the failed aft right-hand fitting

for the hoist sling
pg 375
18/8/2008 3:47PM Plate # 0
with shades of black and brown, as can be produced by heating the steel. The color brown
corresponds to the temper color, which is produced by heating steel to approximately 238 C
(460 F) in air and cooling to room temperature.
Chromium plating was observed on the outer
Fig. 48
diameter surface of the gear. There were no

shear lips evident around the periphery of the
fracture surface. Fracture ridges emanated from
origins located on the inner diameter edge, and
the fracture surface exhibited a faceted appearance indicative of a brittle intergranular failure.
Typical lack of fusion area
Fig. 50
Fig. 49
Direction of fracture propagation
SEM fractographs documenting the topographic

features of the failed weldment. (a) Location A
showing overload features. (b) Location B showing overload
features. (c) Location C showing weld defect features. Original
magnification: 1200
pg 376
18/8/2008 3:47PM Plate # 0

Figure 54 shows SEM fractographs made at a
typical fracture origin. The fracture topography
was intergranular, which is indicative of a brittle
mode of failure (Ref 3). No defects were observed that could be associated with the cause
of the fracture.
at the location shown in Fig. 52. The specimen
was prepared using standard metallographic
techniques. Figure 55 shows the microstructure
(tempered martensite) and typical cracks that
were present. The cracks were intergranular,
and an oxide was present in the cracks. The
Fig. 51
Energy-dispersive spectrographs of failed weld and

base material. (a) Location A weld metal. (b) Location B weld metal. (c) Location C weld metal
Fig. 52
As-received internal spur gear showing location of

cracks
Fig. 53
surface
Macrographs of (a) magnetic particle indications of

cracks and (b) opened crack revealing the fracture
pg 377
18/8/2008 3:47PM Plate # 0
Main Landing Gear Axle
Fig. 54
SEM fractographs showing the topographic features

of a typical fracture origin. (a) Fracture surface. Original magnification: 20. (b) Location A showing intergranular
fracture. Original magnification: 3000
morphology of the cracks is typical of grinding

cracks (Ref 8).
A hydrogen analysis conducted on the internal gear yielded a value of 1.85 ppm, which is
not considered to be high enough to cause
internal gear to verify the heat treat condition.
The hardness of the gear near the observed
cracking was 44 HRC. This hardness value was
below the drawing hardness requirement of 50
to 53 HRC. The bulk hardness of the gear was
51 HRC.
The lower-than-expected hardness at the
crack initiation sites indicates that there was
some event that caused excessive local heating
of the microstructure. Because of the nature of
the part and typical manufacturing sequence, it
is likely that the grinding cracks occurred as a
result of aggressive grinding. Based on the
results of this investigation, it is concluded that
the internal spur gear failed due to aggressive
grinding.
During taxi, the MLG axle separated. The

axle had been installed for only 90 days, with a
total flight time of 62 h. The axle was fabricated
from 300M high-strength, low-alloy steel.
Shown in Fig. 56 is the as-received failed
MLG axle. Also shown in Fig. 56 is the location
of the fracture, which occurred adjacent to a
chromium- and cadmium-plated area. Also
shown in Fig. 56 is a worn (discolored) area of
chromium plating.
The fracture surface was examined visually
and at up to 50 magnification using a stereomicroscope. Figure 57 is a macrograph of the
fracture surface. Shown in Fig. 58 is an SEM
photograph of the wear that had removed the
protective plating adjacent to the fracture surface. Multiple origins were located around the
periphery of the axle. Faceted areas characteristic of an intergranular, delayed mode of failure
had progressed through approximately 80% of
the thickness. Fatigue propagated from the ends
of the intergranular areas, with several areas of
the wall being totally penetrated by the fatigue.
The relatively large intergranular areas and the
irregular shape of the intergranular area are
indicative of stress corrosion.
Shown in Fig. 59 are SEM fractographs documenting topographic features, that is, intergranular topography and fatigue striations, on
the fracture surface. From the ends of the intergranular area, fatigue propagated through the
wall thickness. Also documented in Fig. 59 are
the lack of plating at a fracture origin and corrosion products, which were present at the
intergranular fracture origins. The lack of plating and the presence of corrosion products are
indicative of a stress-corrosion failure. Also
shown in Fig. 59 is an area where the chromium
was intact.
The hardness of the axle was determined to be
53.9 HRC. This met the hardness requirement of
53 to 55 HRC.
A chemical analysis by atomic absorption
verified that the MLG axle was fabricated from
300M steel.
through a typical fracture origin on the axle. The
specimen was prepared using standard metallographic techniques. As determined from the
metallographic specimen, the thickness of the
chromium plating was approximately 0.06 mm
pg 378
18/8/2008 3:47PM Plate # 0
Fig. 55
Micrographs showing the appearance of the cracks and microstructure. (a) Surface cracking evident. Original magnification:
50. (b) Surface cracking and MnS inclusions. Original magnification: 100 . (c) Extended surface crack. Original magnification: 100. (d) Intergranular crack along prior-austenite grain boundaries. Original magnification: 200
Fig. 56
As-received failed axle
(0.0024 in.). Shown in Fig. 60 is the microstructure of the axle at a typical fracture origin.
The chromium plating ended approximately
1.6 mm (0.062 in.) from the origin. The hardness of the light, discolored area at the
chromium runout was 58.4 HRC, which exceeds

the hardness requirement of 53 to 55 HRC. Past
the chromium runout, the part was supposed to
be cadmium plated. However, past the chromium runout to the origin, no plating was present. The wear documented in Fig. 60 had
removed the protective cadmium plating from
this location. This allowed corrosion to occur.
Shown in Fig. 61 is a secondary branch crack
characteristic of stress corrosion. This micrograph also shows the chromium plating that had
been penetrated. Documented in Fig. 62 is the
appearance of the crack (unetched) after further
polishing. Note the worn (missing) chromium.
A hydrogen analysis conducted on the axle
yielded hydrogen content of 4.2 ppm, which is
relatively high and considered high enough
to cause hydrogen embrittlement (Ref 4). This
hydrogen is in the form of atomic hydrogen and
can be generated in corrosion reactions and then
absorbed by steel.
pg 379
18/8/2008 3:47PM Plate # 0
Based on the results observed, it was determined that the MLG axle failed due to stresscorrosion cracking. The protective cadmium
plating had been worn away, allowing corrosion
to occur at locations where the cadmium was no
longer present.
Nondestructive Testing and Failure

Analysis of Fin Attach Bolts after
Full-Scale Fatigue Testing
During the inspection, after the first lifetime
of fatigue testing, one aft fin attach bolt was
found to have broken. Five additional bolts were
found to have crack indications. The location of
the bolts and the nomenclature used for identification of these bolts are shown in Fig. 63. The
failed bolt (bolt A) and the additional five bolts,
identified as B, C, G, H, and I, were submitted to
the laboratory to determine if ultrasonic testing
would be a suitable field inspection technique
for this application and to examine the failed bolt
and cracked bolts to determine the failure
modes.
The specification requirements for these bolts
are for a cadmium-plated, forged hex head
metric bolt that has a close tolerance shank.
The material can be any of four material
specifications: BS S147, S148, S149, or S158.
The compositions of these steels are shown
in Table 2. These steels are heat treated to
Fig. 57
Macrograph of fracture surface of the failed axle.

Typical fracture origins are shown at arrows.
a minimum tensile strength of 1100 MPa

(160 ksi) minimum. The threads and the headto-shank fillet radius are rolled after heat treatment. Machining of the fillet radius is not
permitted. The specified dimensions are shown
in Fig. 64 and Table 3.
The failed bolt and the additional bolts were
examined using nondestructive testing, visual
examination, scanning electron microscopy,
metallography, and analytical chemistry. The
hardness of the parts was also measured.
Nondestructive testing was performed on
the submitted bolts to determine if ultrasonic
inspection is a suitable field inspection method.
Magnetic particle inspection was used as a
confirmation of the ultrasonic inspection
method.
Ultrasonic reference standards were fabricated from AN-8 and AN-6 pan stock bolts.
These bolts were machined so that the head
thickness and lengths were similar to the test
articles. Notches were cut into the bolts at the
head radius and at the midpoint length of the grip
Fig. 58
Overall view at origin. (a) Worn cadmium plating at

origin. (b) SEM photograph of fracture surface
pg 380
18/8/2008 3:48PM Plate # 0
Fig. 59
SEM fractographs documenting the fracture features found at the origin. (a) Fatigue striations emanating from the origin
(200 mm). (b) Intergranular fracture at origin (location A, 50 mm). (c) Intergranular fracture and corrosion products found at
the origin (13 mm). (d) Intergranular fracture found at interface between chromium plating and steel (50 mm)
Fig. 60
Lack of plating at fracture origin. (a) Overall view of microstructure through fracture origin (500 mm). (b) Lack of plating at
fracture origin (50 mm)
on the shank. The dimensions of the reference

standards and notches are noted in Fig. 65 and
Table 4. While every attempt was made to make
the manufactured notches at the head-to-shank

fillet radii and the shank identical, it was not
always possible because of manufacturing
pg 381
18/8/2008 3:48PM Plate # 0
Fig. 61
Secondary branch cracking observed. (a) Overall

etched microstructure. (b) Location E, showing
cracking on prior-austentite grain boundaries. (c) Location F,
showing voids in chromium plating and associated cracking
Fig. 62
tolerances. The ultrasonic inspection was conducted with an ultrasonic transducer (20 MHz;
3.18 mm, or 0.125 in., in diameter). The delay
line was removed for this application. The
instrument settings are noted in Table 5. The
sensitivity of the instrument was adjusted so that
an 80% signal response was obtained from the
notch under the bolt head of the reference standard (4.17 mm, or 0.164 in., long; 0.686 mm,
or 0.027 in., deep) and a notch on the shank of
the machined reference standard (4.57 mm, or
0.180 in., long; 4.57 mm, or 0.18 in., deep). Signal responses for the reference standards are
shown in Fig. 66. Each of the submitted bolts
was inspected with an additional 6 dB of gain to
increase the sensitivity of the inspection. Signal
responses for each of the bolts are shown in
Fig. 67. The ultrasonic inspection showed that a
crack indication was present at the shank-tohead fillet radius of bolt C. No crack indications
were found in the other bolts.
The submitted bolts were inspected using
standard magnetic particle inspection techniques. The results indicate that bolts C, G, and I
had circumferential crack indications at the
shank-to-head fillet radius. Bolt I had a circumferential crack indication around the shank
of the bolt. The magnetic particle indications are
shown in Fig. 68.
After nondestructive testing, bolts C, G, and B
were inspected visually. The as-received bolts
are shown in Fig. 69. After inspection, the cracks
Unetched micrograph showing branch cracking and void in chromium plating. (a) Overall view of secondary crack
(500 mm). (b) Location G showing void in chromium plating and associated cracking (50 mm). (c) Location H showing
cracking along prior-austenite grain boundaries (50 mm)
pg 382
18/8/2008 3:48PM Plate # 0
indicated by magnetic particle inspection were

exposed. This was accomplished by centering
the threaded portion of the bolts in the chuck of a
lathe and drilling a small pilot hole. A larger drill
bit was then used to drill the center of the bolt,
leaving a wall thickness of approximately
0.76 mm (0.030 in.). Using a hacksaw, a small
saw cut was made opposite to the crack. The
head of the bolt was held in a vise, and the shank
Fig. 63
Schematic of bolt location and nomenclature used
of the drill bit was inserted into the hole of the

bolt to support the bolt shank. Using the drill bit,
the bolt was bent away from the crack indication, keeping the crack faces in tension. The
exposed crack was then analyzed.
Bolt B showed no crack indications, using
ultrasonic inspection or magnetic particle
inspection. This bolt showed significant circumferential scoring around the periphery of the
pg 383
18/8/2008 3:48PM Plate # 0
pg 384
Bolt C showed a highly polished shank surface, with a circumferential scoring line in the
same approximate location as bolt B. Some
rounding of the head corners was evident. No
identifying mark or part numbers were stamped
on the head of the bolt. After exposing the crack,
shank, at approximately the midpoint of the

shank. Minor deformation at the corners of the
bolt head was detected. Because no crack indication was seen by nondestructive testing
(NDT), no additional work was performed on
this bolt.
Table 2 Allowable chemistries of the submitted bolts

Allowable composition range of British Specifications, wt%
BS S147
BS S148
BS S149
BS S158
Element
Max
Min
Max
Min
Max
Min
Max
Min
C
Si
Mn
P
S
Cr
Mo
Ni
0.38
0.20
0.75
...
...
0.40
0.20
0.60
0.43
0.35
1.00
0.025
0.020
0.60
0.30
0.70
0.36
0.15
0.60
...
...
0.50
...
1.10
0.41
0.35
0.90
0.025
0.025
0.80
...
1.50
0.38
0.20
0.65
...
...
0.70
0.20
1.65
0.43
0.35
0.85
0.025
0.020
0.90
0.30
2.0
0.22
0.15
0.50
...
...
0.90
0.15
...
0.29
0.35
0.80
0.020
0.015
1.20
0.25
0.30
Fig. 64
Schematic of bolt dimensions
Table 3 Allowable bolt dimensions

Thread
Size code
03
04
05
06
08
10
12
14
16
Diameter
Pitch
A tolerance
B min
Max
Min
Max
Min
F
Min
H
+0.0 0.3
R
+0.0 0.2
M3
M4
M5
M6
M8
M10
M12
M14
M16
0.5
0.7
0.8
1.0
1.25
1.5
1.5
1.5
1.5
5.5
7.0
8.0
10.0
13.0
17.0
19.0
22.0
24.0
4.88
6.38
7.38
9.28
12.28
16.08
18.17
21.17
23.17
0.4
0.5
0.5
0.5
0.5
0.6
0.6
0.6
0.6
0.2
0.2
0.2
0.2
0.2
0.3
0.3
0.3
0.3
2.990
3.990
4.990
5.990
7.987
9.987
11.984
13.984
15.984
2.965
3.965
4.965
5.965
7.962
9.962
11.959
13.959
15.959
6.08
7.74
8.87
10.95
14.26
18.90
21.10
24.49
26.75
6.0
7.6
8.5
9.5
12.4
14.8
15.5
17.5
19.5
2.0
2.5
3.0
3.5
4.5
5.0
6.0
7.0
8.0
0.4
0.4
0.5
0.7
0.7
0.8
0.9
1.0
1.1
...
1.0
1.2
1.6
2.0
2.0
2.0
2.0
2.0
Note: All dimensions in millimeters
18/8/2008 3:48PM Plate # 0
pg 385
the resulting fracture surface was examined

(Fig. 70). The fracture surface of crack C had the
characteristic ratchet marks or shear ridges that
indicate multiple initiation sites along the outer
edge of the bolt. The shape and appearance of
the fracture surface was indicative of fatigue.
Bolt G showed circumferential scoring
(Fig. 69), and some rounding of the head corners
was evident. After the crack was exposed,
the resulting fracture surface was examined
Table 4 Dimensions of reference standards

used for ultrasonic testing
Specimen
identification
HL1
HL2
HL3
HL4
HS1
HS2
HS3
HS4
HS5
SL1
SL2
SL3
SL4
SS1
SS2
SS3
SS4
SS5
mm
in.
mm
in.
12.4
10.1
7.70
5.99
9.47
7.82
6.50
4.16
2.46
12.3
11.2
8.13
5.82
8.97
7.65
5.16
4.57
2.82
0.487
0.399
0.303
0.236
0.373
0.308
0.256
0.164
0.097
0.485
0.440
0.320
0.229
0.353
0.301
0.203
0.180
0.111
5.66
2.39
1.37
0.81
3.96
1.98
1.17
0.69
0.15
5.87
3.40
1.52
0.81
3.18
2.24
0.76
0.46
0.18
0.223
0.094
0.054
0.032
0.156
0.078
0.046
0.027
0.006
0.231
0.134
0.060
0.032
0.125
0.088
0.030
0.018
0.007
Table 5 Ultrasonic inspection techniques

equipment settings
Tune
Reject
Gain
Frequency
Video filter
Damping
Delay
Range
Velocity
Repetition rate
Fig. 65
Sketch of reference standards used during ultrasonic

inspection
Fig. 66
Ultrasonic signal responses from the manufactured reference standards
0
Off
53 dB
25 MHz
FW
Minimum
0
8.3
1
3
18/8/2008 3:48PM Plate # 0
Fig. 67
Ultrasonic signal responses for each bolt inspected
(Fig. 71). The fracture surface of crack G had the

edge of the bolt. The shape and appearance of
the exposed crack was indicative of fatigue.
Bolt H (Fig. 72) showed no crack indications,
using ultrasonic inspection or magnetic particle
inspection. This bolt showed significant circumferential scoring around the periphery of
the shank, at approximately the midpoint of the
shank. More general wear was seen on the shank
than bolt B. Because no crack indication was
seen by NDT, no additional work was performed
on this bolt.
pg 386
18/8/2008 3:48PM Plate # 0
Fig. 68
Ultraviolet light photographs of magnetic particle indications for each bolt inspected. NDT, nondestructive testing
pg 387
18/8/2008 3:49PM Plate # 0
Bolt I (Fig. 72) showed scoring along the

circumference of the shank. In this bolt, crack
indications were found by NDT in both the
shank and the head. The exposed shank crack is
shown in Fig. 73. The shank fracture surface of
crack I had the characteristic ratchet marks or
shear ridges that indicates multiple initiation
sites along the outer edge of the bolt. The shape
and appearance of the exposed crack was indicative of fatigue. The crack indication in the
head-to-shank fillet radius of bolt I was opened
in a similar manner as the other bolts, and the
fracture surface is shown in Fig. 74. The headto-shank fillet radius fracture surface had the
Fig. 69
As-received condition of bolts B, C, and G

edge of the shank fillet radius (Ref 9). The shape
and appearance of the exposed crack was indicative of fatigue.
Fig. 70
Fracture surface of bolt C with fracture schematic
Fig. 71
Fracture surface of bolt G with fracture schematic
pg 388
18/8/2008 3:49PM Plate # 0
The as-received condition of bolt A is shown

in Fig. 75. A photograph of the fracture surface
of the failed bolt is shown in Fig. 76. The
Fig. 72
Fig. 73
Fig. 74
Fracture surface of the exposed crack in the head

fillet radius of bolt I
Fig. 75
As-received condition of the failed bolt
Fig. 76
Fracture surface of the failed bolt
As-received condition of bolts H and I
Fracture surface of the exposed crack in the shank of

bolt I
pg 389
18/8/2008 3:49PM Plate # 0
fracture surface shows ratchet marks that are

characteristic of fatigue origins. Multiple fatigue
origins, located to one side of the bolt, were
observed. A large fatigue zone was apparent.

The final fracture zone was badly damaged. The
bolt showed evidence of machining marks on the
sides of the shank at the site of crack initiation
(Fig. 77). Opposite the initiation site, on the side
of the shank, the surface appeared to be polished,
Fig. 79
SEM micrograph of the initiation site of the failed

bolt, showing cracking beginning at the localized
surface wear (1 mm)
Fig. 77
Fig. 78
Machining marks and polishing evident on the shank

of the failed bolt
SEM micrograph of representative multiple fatigue

origins associated with machining marks on exposed
cracks (100 mm)
Fig. 80
SEM micrograph showing initiation of cracking on

the shank of bolt I at localized surface wear (400 mm)
Fig. 81
SEM micrograph of representative fatigue striations

found on the bolt fracture surfaces (2 mm)
pg 390
18/8/2008 3:49PM Plate # 0
with more evidence of machining marks

(Fig. 77).
The fracture surfaces on bolts A, C, G, and I
were examined using the SEM. The fracture
surfaces of the exposed cracks in the bolts were
characterized by multiple fatigue origins, initiating along machining marks (Fig. 78). In the
fractured bolt, initiation of cracking began at a
region of localized wear and scoring on the
shank of the bolt (Fig. 79). This was also evident
Fig. 82
in the initiation of cracking in the shank on bolt I

(Fig. 80). Striations, characteristic of fatigue,
were seen on the fracture surface of bolt A
(Fig. 81). Machining marks found in the headto-shank fillet radius were evidence that the
head-to-shank fillet radius was machined
(Fig. 82). This is unacceptable according to the
design specification.
using standard techniques. This specimen was
used to examine the microstructure and microhardness of bolt A. A photographic montage of
the metallographic specimen is shown in Fig. 83.
The tempered martensite microstructure of
bolt A was typical for quenched and tempered
low-carbon steel (Fig. 84). Bolt A failed in
the shank; in addition, a crack approximately
0.15 mm (0.006 in.) long was found in the headto-fillet radius (Fig. 85). Neither magnetic particle nor ultrasonic NDT detected this crack.
The failed bolt was submitted to a microhardness survey to determine the extent of decarburization. The microhardness survey was
taken with a Knoop indenter, with a 500 g load.
Indentations were taken at 0.0318 mm
SEM micrograph of machining marks at the headto-shank fillet radius of bolt I (1 mm)
Fig. 84
Micrograph of representative quenched and tempered martensite found in the submitted bolts.
Etched with 2% nital (25 mm)
Fig. 83
Microstructure at bolt A fracture through section 1-1.

The overall microstructure is shown as well as the
location of additional micrographs. There is no evidence of a
cold-formed head. Etched with 2% nital (667 mm)
Fig. 85
(25 mm)
Micrograph of the crack found in the head-to-shank

fillet radius of the failed bolt. Etched with 2% nital
pg 391
18/8/2008 3:49PM Plate # 0
(0.00125 in.) intervals until the hardness readings began to stabilize. The point at which the
hardness stabilized was called the core hardness.
For the failed bolt, the core hardness was 404
KHN. In accordance with the specification, the
point at which the hardness is within 20 KHN of
the core hardness is defined as the depth of
decarburization. For the failed bolt, this was at
0.064 mm (0.0025 in.) deep, meeting the specification requirements. The hardness indentions
are shown in Fig. 86. The results of the microhardness survey indicated that the surface
hardness (Knoop), converted to Rockwell C, is
26 HRC. The Knoop core hardness, converted to
Rockwell C, is 40 HRC. While the partial decarburization of the bolt satisfies process specification requirements, the steep gradient of the
hardness profile and the surface softness measured indicated that there would be degradation
in the potential fatigue life of the bolts.
Determination of the composition of the

failed bolt was performed by arc-spark spectrometry. The carbon content of the steel was
determined by Leco carbon analyzer. Results are
shown in Table 6. The chemical results indicate
that this bolt does not meet the composition
requirements of Table 2 for any of the specified
bolt alloys.
The hardness of each bolt was measured using
standard procedures. The calibration of the
hardness tester was verified, using two calibration test blocks (54.1+1.0 HRC and 34.6+1.0
HRC). The measured hardnesses on the test
blocks were found to be within the range of the
hardness test block calibration. A section of
the bolt was taken by cutting perpendicular to
the long axis of the bolt, in the threaded region.
The hardness of each bolt was determined by
taking a hardness measurement on the end of the
cut section. Results are shown in Table 7.
Based on this investigation, it was concluded that:

Fig. 86
Micrograph of Knoop hardness (500 g load) survey of

failed bolt (100 mm)
Table 6 Measured chemistry of the failed bolt A

Element
C
Si
Mn
P
S
Cr
Mo
Ni
0.42
0.32
0.82
0.010
0.014
0.49
0.15
0.45
Table 7 Measured hardness of submitted bolts

Bolt
C
B
H
G
I
Failed bolt A
Percent,
wt%
Hardness,
Rockwell C
43.4
36.4
41.6
40.4
39.2
36.9
Ultrasonic inspection does not have the

sensitivity to detect cracks of the size and
geometry seen in the bolts.
The fractured bolt failed by fatigue, initiating at localized wear on the bolt shank. The
soft surface hardness from partial decarburization aggravated the fatigue failure.
The exposed cracks in bolts C, G, and I
initiated at machining marks at the head-toshank fillet radius, and cracking propagated
by fatigue. Evidence of machining at this
radius was found. Bolt I had a second
initiation site on the shank.
The composition of the failed bolt did not
meet the chemical requirements of S105
bolts. However, the chemistry of the bolts is
not thought to have contributed to the failure
of bolt A.
The bolts were machined instead of cold
formed. This results in lower fatigue strength
and proof load.
REFERENCES
1. G.E. Dieter, Mechanical Metallurgy, McGrawHill, Inc., 1986

2. Metallography and Microstructures, Vol 9,
for Metals, 1985
3. D.A. Ryder et al., General Practice in Failure
Analysis, Failure Analysis and Prevention,
pg 392
18/8/2008 3:49PM Plate # 0
Vol 11, Metals Handbook, 9th ed., American

4. B.E. Wilde, Stress-Corrosion Cracking,
for Metals, 1986
5. W.J. Jensen, Failures of Mechanical Fasteners, Failure Analysis and Prevention, Vol
11, Metals Handbook, 9th ed., American
6. Threaded Steel Fasteners, Failure Analysis and Prevention, Vol 11, Metals

Metals, 1986
7. A.G. Glover et al., Failures of Weldments,
for Metals, 1986
8. E. Alban, Failures of Gears, Failure Analysis
and Prevention, Vol 11, Metals Handbook,
9th ed., American Society for Metals, 1986
9. R.J. Parrington, Fractography of Metals and
Plastics, Pract. Fail. Anal. Vol 2 (No. 5),
2002, p16
pg 393
18/8/2008 3:51PM Plate # 0

DOI: 10.1361/faht2008p395

Failure Analysis of Powder Metal

Steel Components
S. Ashok and Sundar Sriram, Sundram Fasteners Ltd.
POWDER METALLURGY (PM) technology provides a cost-effective method of producing near-net shape products, especially when
a large number of the same or similar products
are required. While the initial powder used
is expensive compared to wrought steels, the
manufacturing process can avoid machining
steps and provide more uniform composition
control. Total cost of producing PM parts can be
less expensive than conventional metalshaping
(casting, forging) in many cases. It was first
adopted by the auto industry. Since this initial
inception, other industries, such as the refrigerator, lawnmower, and hand tools industries,
have also extensively adopted PM parts (Ref 1).
Manufacture of powder metal parts involves
pressing of metal powder in a die, sintering
the green compact so that metallurgical bonding
takes place, sizing or coining for densification
and dimensional correction, followed by finishing operations. The finishing operations include
machining, case and through hardening, steam
treatment, and so on (Ref 1).
This chapter reviews failure aspects of structural ferrous PM parts, which form the bulk of
the PM industry. Focus is on conventional PM
technology of parts in the density range of 6 to
7.2 g/cc. This constitutes 90% of the parts produced by PM technology.
The PM part undergoes multiple heat treatments, some of which are unique to PM while
others are similar to that used by conventional
manufacturing processes, such as forging and
casting. These include:

Powder annealing
Sintering
Case hardening
Carbonitriding
Case carburizing
Through hardening
Induction hardening
Steam treatment
This chapter briefly introduces the processing

steps involved in PM (Fig. 1). These heat treatments and the PM process steps are essential
to understanding failure analysis of PM parts.
The methods used for analyzing the failures are
then discussed. Methods of failure analysis itself
have recently become more systematic in terms
of prevention and ensuring more robust process
and high-quality products. Some case studies
are given that illustrate different failures and the
methods of prevention of these failures.
Powder Metallurgy Process

Powder Production. Solid-state reduction
of iron ore and atomization of molten iron are
the most common methods of producing iron
powders for structural parts. The powder thus
produced should be favorable for compaction
such that:

It fills the die cavity repeatably, which is

measured in terms of the apparent density
and flow rate of the powder.
It can be compressed to the desired density
with the lowest pressure, which is measured
in terms of compressibility.
It has sufficient strength at that density so
that the green part can be handled, which is
measured in terms of its green strength.
The filling characteristic of a powder is

measured by its apparent density and its
flow rate. The compressibility of a powder is a
function of the particle morphology, processing
history, and purity. Water-atomized iron powders can be compressed to higher densities
compared to sponge iron powder. The higher
the impurity content (typically oxides), the
poorer the compressibility. As a process, all iron
powders are generally annealed in a reducing atmosphere. Special high-compressibility
pg 395
18/8/2008 3:52PM Plate # 0
Fig. 1
The powder metallurgy process
powders are produced by double-annealing

operations.
Blending. Alloying elements are typically
mixed with the iron powder by blending in
a double-cone blender. Elemental addition of
alloying elements, that is, graphite, copper, or
nickel, has its drawback, because these elements
tend to segregate in the product as the alloying
element content increases. A recent development in this regard is the introduction of diffusion-bonded iron powder and binder-treated
premixes. These processes are proprietary. For
example, the Distaloy process is diffusion
bonding of nickel, molybdenum, and copper to
iron powders developed by Hoganas AB, Sweden. The Ancorbond and Starmix methods are
binder treated premixs to reduce graphitelubricant segregation. In the case of diffusionbonded iron powders, the alloying elements,
typically copper, nickel, and molybdenum, are

mixed with iron, and the mix is annealed at
a temperature of approximately 850 to 920 C.
As a result, bonding of the alloying elements
to the iron particle takes place, and this prevents segregation of alloying elements in the
products. The same principle is used to prevent
carbon-lubricant segregation. In this case, the
binder, which is added during powder mixing,
is sprayed on the iron-carbon mix. This binder
ensures that each iron particle has a coating
of graphite and thus minimizes segregation.
Case studies in this chapter illustrate this
mechanism.
Compaction. The most widely used method
of compaction is axial pressing of loose powder.
Initially, the die cavity is filled with the loose
powder from a hopper by gravity; the density is
typically the apparent density of the iron powder, which is 2.4 to 3 g/cc. When compaction
commences, that is, the punches move toward
each other, densification takes place by particle
rearrangement and plastic deformation. Thin
sections of individual particles bend or break,
cold welding and interlocking of neighboring
grains takes place, and the voids are filled by the
material, which becomes squeezed into them.
The density of 5.5 g/cc is the lowest density
of iron powder parts fit for practical handling in
the green stage. On attaining this density, a rapid
increase in pressure is observed when the part is
compacted to densities between 6 and 7.2 g/cc.
Compacting in the higher-density range gradually increases the rate of plastic flow of the
metal. This results in the following:

Increased work hardening of powder particles

Increased friction against the die walls
Structural ferrous powder metal parts typically have sections of varying thickness. With
respect to heat treatment, it is very important that
all sections in the parts have uniform density
throughout. Low sectional densities result in
through hardening and wide variation in the case
depth within the PM part. Various methods of
compaction have been developed to produce
parts with minimal difference in density. These
include:

Double-action compaction
Spring floating die compaction
Compaction by withdrawal process
Multilevel compaction using multiple moving punches in special multiplaten presses
pg 396
18/8/2008 3:52PM Plate # 0
Failure Analysis of Powder Metal Steel Components / 397
Sintering. During sintering of ferrous PM

parts, the following reactions occur:

Metallurgical bonds are developed from the

mechanical interlocks between the powder
metal particles in the compact.
Metal oxides in the powder compact are
reduced by reaction with the carbon from the
blended graphite or with the combustible
constituents of the atmosphere.
Desired final level of combined carbon in the
microstructure is obtained by diffusion of
carbon from graphite in the powder mix.
Densification of PM compacts can be

achieved in sintering. The degree of densification depends on the sintering parameters and the
alloys involved. Generally, higher sintering
temperatures and longer sintering times promote
densification of ferrous sintered parts.
For conventional sintering, which generally is
conducted at 1120 C, mesh belt conveyor furnaces are widely used. These furnaces can be
used up to temperatures of 1150 C, primarily
due to the temperature limitation of the belt and
metallic muffle material.
For stainless steels, proper sintering is critical
for proper corrosion resistance (Ref 2).
Case Hardening
Powder metallurgy steel parts can be case
hardened by several processes, although various
available processes are not equally suited to
every application. A clear case/core relationship
can be obtained only with parts having a density
of at least 7.2 g/cc.
Carburizing. Powder metallurgy parts with
relatively low combined carbon contents of up to
0.20 wt% can be carburized by conventional
pack or gas methods. Liquid carburizing is
not recommended because of the difficulty of
washing the parts free of salt.
Gas carburizing is more practical for PM parts
than pack carburizing. For this process to be
successful, however, density as well as the precise composition of the parts should be known.
Low-density parts should not be subjected to
gas carburizing, because the carburizing gases
penetrate the voids. Consequently, a distinct
case is not achieved compared with the case
developed on wrought parts under the same
conditions. Instead, the carbon penetration in
PM parts is generally deeper and relatively
nonuniform. The extent of this condition varies
with density. In parts that have been repressed

and resintered, this condition may be tolerable or
even negligible, but for parts of lesser density,
the depth of carbon penetration may be so great
that these sections of quenched parts will be
brittle.
Another reason that conventional gas carburizing enjoys only limited use is because it does
not increase hardenability. Thus, plain carbon
grades usually must be quenched in an aqueous
medium. This may result in cracking, especially
if carbon penetration is excessive. As with
wrought parts, the depth of carburized case of the
PM parts depends on time and temperature.
Carbonitriding is a modified form of carburizing. The principal process modification
consists of introducing ammonia into the gas
carburizing atmosphere, which results in the
addition of nitrogen to the carburized case as
it is produced. Nascent nitrogen forms at
the workpiece surface by dissociation of the
ammonia in the furnace atmosphere. Nitrogen
diffuses into the steel surfaces simultaneously
with the carbon, where the austenite stability is
greatly enhanced by nitrogen in solution. This
way, the necessary quenching rate to form
martensite is reduced, and a martensitic microstructure is obtained without expensive alloying
elements. Typically, carbonitriding of PM parts
is carried out at 790 to 880 C for a duration of
30 to 60 min.
Carbonitriding is widely used for case hardening of PM parts made of ferrous powders.
Densities of the sintered compacts vary from
approximately 6.8 to 7.9 g/cc. Parts may be
infiltrated with copper prior to carbonitriding.
Carbonitriding is extremely effective for case
hardening high-density (7.2 g/cm3) parts made
from sintered iron compacts (Fig. 2). Additionally, it is reasonably effective for case hardening
parts of lower density.
Equipment and Techniques. Procedures for
carbonitriding PM parts are essentially the same
as those used for similar wrought parts. Control
of temperature and time is generally more critical than for wrought parts because of porosity.
Lower temperatures are avoided to minimize the
potential danger of explosion, and higher temperatures are avoided because case depth control
is more difficult.
The processing cycle, including composition
of the atmosphere, is critical. The ammonia content (usually 1 to 5% of carrier gas by volume)
increases hardenability and affects dimensional
stability. Because dimensional changes in heat
pg 397
18/8/2008 3:52PM Plate # 0
Fig. 2
Effect of density on the case depth as measured through

a hardness traverse from the surface. Courtesy of
P. Beiss, University of Aachen, Germany
treating are often crucial to the economic justification of producing parts by PM, gas composition, temperature, and quenching medium
must be closely controlled.
When processing to a new specification,
establishment of processing parameters is
usually made on small production quantities,
thereby requiring sacrifice of only a few parts to
arrive at optimal conditions. Such adjustments
are then recorded so they can be used when
processing the next lot of similar parts. Tempering of carbonitrided parts requires special
consideration, largely because the quenching oil
they contain will partially evaporate and pollute
the environment. Toughness of PM parts in
the hardened condition (either by oil quenching
or through sinter hardening) is significantly
improved on tempering.
Induction Hardening. Induction heating is
a method of heating electrically conductive
materials by the application of a varying magnetic field whose lines of force are intersected
by the workpiece. In this process, the varying
magnetic field induces an electric potential,
which in turn results in generation of electric
current depending on the geometry, the frequency, and the electrical characteristics of
the workpiece. The induced current, termed
eddy current, generates heat that makes it
amenable for use in many different heating
applications, of which the induction hardening
of steels and cast irons is one of the most predominant.
Induction hardening of PM parts has several
differences compared to hardening wrought
steels and cast irons. Chemical composition,
microstructural heterogeneity, and low density
due to the porosity are the causes of the different

induction-hardening responses of PM parts. The
results of hardening are more sensitive to chemical composition and the prior microstructure
when compared with alternative processes. The
electrical resistivity, thermal conductivity, and
magnetic permeability strongly depend on the
porosity of the PM part. Low density negatively
affects the hardenability of powder metal parts
(Table 1). It is recommended that the part
selected for induction hardening has a density
of at least 7 g/cc. Carbon, copper, nickel, and
molybdenum are the most commonly used
alloying elements in PM parts. The stresses due
to high carbon content or large pores aggravate
cracking. Segregation of alloying elements,
foreign inclusions, or large pores can serve as
stress raisers, making the powder metal part
susceptible to cracking.
It is quite common for PM parts to absorb
oil. Thus, intensive ventilation is required, and
one must ensure that the reused quenchant provides the required quench severity. Water-based
polymer fluids are the most common quenchants
used for induction hardening of PM parts.
Steam Treatment. Sintered parts are subjected to steam treatment, wherein controlled
oxidaton of the ferrous part is carried out in
an atmosphere of superheated steam. As a result,
a layer of Fe3O4 forms on the surface as well
as in the pores. This oxide is hard and has
excellent wear resistance (hardness of 450
HV0.05 kg), increases density, and improves
corrosion resistance. The iron oxide also serves
to seal the pores, which results in parts that
are impermeable to gases. This characteristic of
Table 1 Effect of low density on properties and

induction hardening process parameters
Property
Thermal
conductivity
Electrical
resistivity
Magnetic
permeability
Structural
homogeneity
Change
Effect on induction
hardening parameters
Decreases with Inefficient conduction

decrease in
heat transfer
density
Larger temperature
gradients(a)
Increases with Larger current
decrease in
penetration depth
density
Decreases with Larger current
decrease
penetration depth
in density
Lower coil electrical
efficiency
Decreases with Wide scatter in
decrease in
apparent hardness
density
Wide scatter in case depth
(a) The penetration of the quenchant into open porosity overcompensates the
effect of lower conductivity. As a result, low-density parts cool faster.
pg 398
18/8/2008 3:52PM Plate # 0
steam-treated PM products has enabled its

widespread use in pistons and valve plates for
compressors. The sealing of pores also helps the
PM part to be amenable to corrosion-preventive
coatings (galvanic as well as barrier coatings)
and, in some cases, to gas and salt bath nitriding.
The formation of Fe3O4 is exothermic, and thus,
a further increase in product temperature by
10 to 20 C is observed. Steam treatment is
typically carried out at 500 to 550 C for a time
of 30 to 300 min. The upper limit of 550 C
is chosen so that the products do not exceed
570 C, which in turn would result in FeO formation (Fig. 3).
Failure Analysis Techniques

Failure analysis techniques can be broadly
classified as preventive methods or corrective
methods. Preventive methods of failure analysis
are typically carried out during product or process development. These refer to the standard
techniques for failure mode and effects analysis
(FMEA), which are specified in the various
certifications, such as TS16949 and ISO
9000:2000. The engineer anticipates the failures

of the product on the basis of known failure
modes and designs the process and product
parameters accordingly.
However, in practice, failures are observed in
spite of carrying out the measures that are
recommended by an FMEA. This can be due to
the fact that assumptions which go into carrying
out the FMEA may change over a period of time,
or the basis of the assumptions is incorrect, but
more commonly, the failures occur due to the
breakdown in one or more so-called 4M parameters (man, machine, method, material). Thus,
there is a need for corrective methods of failure
analysis, which form the bulk of this chapter.
However, the techniques used in the FMEA are
useful for analyzing failures, and a few are very
relevant to heat treated products. In the case
of heat treated powder metal parts, the failure
modes and their causes are limited in number
and can be quantified easily. The common failure modes of heat treated powder metal parts
are:

Wear
Fracture
Dimensional instability because of plastic
deformation
Dimensional instability because of phase
transformation
Corrosion
These can be related to deficiencies in mechanical or metallurgical properties. Properties that

play a role in failure of heat treated PM steel
include:

Fig. 3
Fe-O-H2O diagram indicating the zones in which

various oxides of iron are stable. Courtesy of P. Beiss,
University of Aachen, Germany
Metallurgical properties: microstructure;

case depth; coating adhesion, hardness, and
thickness for wear-resistant coatings; salt
spray life in the case of coatings for corrosion prevention; density; and chemical
composition
Mechanical properties: tensile strength,
elongation, impact strength, fatigue strength,
fracture toughness, and hardness
Deficiencies in properties that can result in a

failure are caused by a combination of one
or more of the 4M parameters. For example,
low apparent hardness after carbonitriding may
be associated with carbonitriding or due to
low density in compaction. Thus, in analyzing
the failure, one needs to consider the fact that
the cause of the failure may be associated with
more than one stage of component production.
pg 399
18/8/2008 3:52PM Plate # 0
Failure Analysis Tools. The objective of the

failure analysis is to relate the failure to a product property attributed to one or more 4M
parameters of processing. The steps that are
typically carried out in the failure analysis of
heat treated powder metal parts are shown in
Fig. 4. Examination of the product involves one
or more of the following techniques:

Visual or any other method of nondestructive examination

Fractography or examination under a lowmagnification stereomicroscope
Hardness and related mechanical property
measurements
Metallography under an optical or scanning
electron microscope
Chemical analysis
The failure can occur because of a nonconforming property in the product even though
the product meets the specifications. The cause
of the former is related to the manufacturing
process, and the latter is related to the design.
However, both failures involve determining the
4M conditions that cause the failure and correcting the same. In the case of failure of a heat
treated powder metal part, the failure mode as
well as the causes of failure are limited. As
mentioned previously, a principal activity is an
operation-wide 4M analysis to determine the
Fig. 4
Steps in a failure analysis
cause of the failure. Two tools are invaluable

aids in this regard. These are process maps and a
cause-and-effect matrix.
Process Maps. In charting a process map,
the entire manufacturing process is considered.
The implicit assumption is that the cause of
failure can be related to not only the final
operation but also to prior operations. Every
operation in the manufacturing process is
regarded as a process with inputs and outputs.
The inputs consist of the 4M conditions and the
output of the product requirements. All possible
4M conditions that can result in a nonconforming product are listed as inputs, with the nonconforming properties of the products as output.
Table 2 is an example of a detailed process
map for the PM process. This exercise is carried
out for all stages in the manufacturing process.
Once the process map is charted, the potential
causes at different manufacturing steps are
identified.
Cause-and-Effect Matrix. The process map
is an effective tool that eliminates the possibility
of ignoring an operation that can result in failure
of the product. However, information presented
by the process map requires sorting for easy and
meaningful interpretation. The cause-and-effect
(CE) matrix ensures this. In the CE matrix, the
4M conditions that can cause the failure are
listed in the y-axis and the defective properties
pg 400
18/8/2008 3:52PM Plate # 0
pg 401
Table 2 Detailed powder metallurgy process map

Inputs
Compaction
Single platen press
No pressway
No double action
High apparent density
Low apparent density
Segregation of alloys
Segregation of carbon and
lubricant
Sintering
Low cooling rate
High cooling rate
Low temperature
High temperature
Low soaking time
High soaking time
Oxidizing atmosphere
Decarburizing atmosphere
Carburizing atmosphere
High rapid burnoff
temperature
Low rapid burnoff
temperature
Low density
Carbon-lubricant
segregation
Alloy element
segregation
Low sectional density
Poor sinter hardenability
Carburizing/carbonitriding
Low carbonitriding
temperature
High carbonitriding
temperature
Low soaking time
Outputs
...
Low density
Density difference within
compact exceeding 0.2 g/cc
Carbon and lube segregation
Alloy segregation
Dimensions not conforming
to specification
...
...
Low hardness
Decarburization
Oxidation
Excess growth
Shrinkage
Carburized microstructure
Low mechanical properties
Blistered product
Pinholes
Low sintered density
...
...
...
...
...
...
...
High surface hardness
on the x-axis. These are usually transferred from

the process map. A coordinate in this matrix
links the potential cause, that is, the 4M conditions, to the failure, that is, the defective product characteristic. Once all the coordinates are
filled, all the potential causes of a failure are
listed in a user-friendly format. An example of
a process map and CE matrix is illustrated in
Fig. 5.
Inputs
Induction hardening
Too high frequency
Too low frequency
High heating rate
Low heating rate
High heating time
Delayed quench
Severe quench
High temperature
Low temperature
High carbon
Low density
Steam treatment
Continous mesh belt
furnace
Batch furnace
Low temperature
Insufficient time
Insufficient steam
Oil in pores
Case Study 1: Wear after Sinter Hardening

Sinter-hardened bushes were developed for
an application that required high wear and
impact fatigue resistance. The bushes were sinter hardened to a martensitic-bainitic microstructure using a prealloyed Fe-Cr-Mo powder.
Accelerated rig testing for 24 h yielded no wear,
...
...
Cracked product
High case depth
Low case depth
Low surface hardness
Localized zones of melting
...
...
...
...
...
Low hardness
Leaky product (pores not well sealed)
Low oxide layer
Poor surface appearance
Loose rust
Red rust
but the bushes were observed to fail in the field.

The parameters that can cause the failure
on listed in the CE matrix for wear of sinterhardened bushes, as follows:
Operation
Powder
chemistry
Blending
Compaction
Sintering
Case Studies of PM Steel Failures
Outputs
Carburizing/carbonitriding (continued)
High time of cabon
...
potential attainment
High soaking time
Low surface hardness
Decarburization
Oxidation
Excess growth
High nitrogen potential
Shrinkage
Masking of products
...
Delayed quench
Carbide network in case
Hot oil quench
Nonmartensitic transformation
product in case
High-viscosity quench oil
...
Delayed tempering
Low mechanical properties
Poor hardenability
High retained austenite
Low density
Low case depth
Carbon segregation
High case depth
Alloy segregation
High core hardness
Low core hardness
4M parameter
Observations
Low carbon
Yes

alloy elements
Low density
Low sintering temperature
Low soaking time
High rapid burnoff temperature
Bainite in microstructure
Overloading of parts
No
No
No
No
No
No
No
Yes
No
Figures 6(a) and (b) show the failed bush.

Figure 7 shows the microstructure of the failed
bush, which consists of martensite and upper and
18/8/2008 3:52PM Plate # 0
pg 402
Blending
X
X
X
X
X
X
X
Poor mechanical
properties
Nonconforming
dimensions
Blistered product
Pinholes
Poor surface apearance
Loose rust
Red rust
Thick oxide layer
Thin oxide layer
Melt zones
Cracked product
Decarburization
Oxidation
Free ferrite in core

Nonmartensitic
transformation product
Low core hardness

High core hardness
High case depth
Low case depth
Process inputs
Excess time
Insufficient time
Elemental alloy additions
High carbon
High lubricant
High apparent hardness
Process step
Low apparent hardness
X
X
X
X
Compaction
Single platen press
No pressway
No double action
High apparent density
Low apparent density
Segregation of alloys
lubricant
Sintering
Low cooling rate

High cooling rate
Low temperature
High temperature
Low soaking time
High soaking t ime
High rapid burnoff
temperature
Low rapid burnoff
temperature
Low d ensity
Carbon-lubricant
segregation
Alloy element segregation
Low sectional d ensity
Poor s inter hardenability
Carburizing/
carbonitriding
Fig. 5
X
X
X
X
X
X
X X
X
X
X
X
X
X
X
X
X X
X X
X
X
X
X
X
X
X
X
X
X
X
X
Low carbonitriding
X
temperature
High carbonitriding
temperature
X
Low soaking time
High time of carbon
potential attainment
High soaking t ime
X
Decarburizing atmosphere X
X
High nitrogen potential
X
Masking of products
X
Delayed quench
X
Hot oil quench
X
High-viscosity quench oil
Delayed t empering
X
X X
X
X X
X X
X
X
X
X
XX
X
X
X
X
Cause-and-effect matrix derived from powder metallurgy process map
X
X
18/8/2008 3:52PM Plate # 0
pg 403
Carburizing/
carbonitriding
Induction
hardening
Steam
t reatment
Too high frequency

Too low frequency
High heating r ate
Low heating rate
High heating time
Delayed quench
Severe quench
High t emperature
Low t emperature
High carbon
Low density
Continous mesh belt
furnace
Batch furnace
High t emperature
Low temperature
Insufficient time
Insufficient steam
Oil in pores
X
X
Poor mechanical
properties
Nonconforming
dimensions
Blistered product
Pinholes
Poor surface apearance
Loose rust
Red rust
Thick oxide layer
Thin oxide layer
Melt zones
Cracked product
Decarburization
Oxidation
Free ferrite in core

Nonmartensitic
transformation product
Low core hardness

High core hardness
High case depth
Low case depth
Process inputs
Poor hardenability
Low density
Carbon segregation
Alloy segregation
High apparent hardness
P ro c e s s s te p
Low apparent hardness
X X
X
X
X X
X
X
X
X X
X
X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
Fig. 5 (continued)
Fig. 6
Sintered bush. (a) Outside diameter wear. (b) Outside diameter crack
lower bainite. The failure was due to the softer

bainite being worn out, resulting in eventual
accelerated fatigue failure of the bush.
Bainite is an essential component for
toughness. A fully martensitic microstructure
obtained by increasing the cooling rate and the
carbon content has poor toughness. Thus, the
decision was made to use a different material
that can give the required hardness and toughness.
Corrective Measures. An improved chemistry was derived, where the base iron powder
was prealloyed with nickel and molybdenum,
and the carbon content was increased to 0.9%.
This resulted in a predominantly martensitic
microstructure. The presence of nickel ensured
the toughness requirements, and the increased
carbon ensured the wear requirement. Figure 8
shows the new microstructure, wherein martensite with some retained austenite is observed.
18/8/2008 3:52PM Plate # 0
Fig. 7
Micrographs of the failed bush. (a) Martensite and lower bainite. (b) Upper bainite
Results. The improved bushes were assembled in chains and tested in a rig as well as in the
field. Figure 9 shows the comparative evaluation
of the PM and wrought chains after the change in
material of the PM bush. It is observed that in
addition to withstanding the tests, the chains
with PM bushes have less elongation compared
to chains with wrought bushes.
Case Study 2: Pinholes after Sintering
Valve seats have a chemistry that is highly
alloyed and a carbon content greater than 1%.
Because most of the elements are admixed, segregations of these elements are not uncommon.
The parameters that can cause pinholes in valve
seats one:
Operation
Blending
Compaction
Sintering
Machining
4M parameter
Segregation of carbon and lubricant

Copper agglomeration
No inspection for agglomerates
in powder
Not applicable
Low soaking time
Pullout of nonmetallic inclusions
in machining
Fig. 8
Martensite and some retained austenite in the modified

chemistry
Observations
Yes
No
Yes
...
No
No
No
In all of these above cases, the pinholes that

are observed after sintering or machining are in
the regions where the agglomerates were present, prior to sintering or machining. The copper
melts and diffuses into the iron matrix, carbon
diffuses in the matrix, and the lubricant burns
off, leaving the pinhole in each case. In the case
of nonmetallic inclusions, they are pulled out
during machining, thus leaving a pinhole.
Scanning electron microscopy/electron dispersive x-ray analysis was carried out to determine the cause. Figures 10 to 13 reveal
high carbon content near the pinhole. Thus, the
likely cause of the pinholes was concluded to
be carbon-lubricant agglomeration during
blending.
Corrective Measures. Binder-treated premixes that minimize carbon-lubricant segregation were recommended for the products.
An inspection procedure was evolved to check
the premix for the presence of agglomerates and
coarse particles, wherein the premix was sieved
and the +150 mm fraction was inspected for
agglomerates under a stereomicroscope. This
ensured that a premix with agglomerates is not
issued for compacting the products.
pg 404
18/8/2008 3:52PM Plate # 0
pg 405
Fig. 9
Comparison of sintered versus regular bush in 12B model chain
Fig. 10
Fig. 11
Pinhole in the valve seat at a magnification of 40 : 1
Results. The rejections due to pinholes

decreased significantly as a result of using binder-treated premix.
SEM micrograph of the pinhole of Fig.10 indicating a

suspected graphite particle
The parameters that can cause blistered sintered

products are:
Operation
Powder chemistry
Blending
Case Study 3: Blistered Sintered Products

Synchronizer keys used in an automobile gear
box were produced with Fe-3%Ni-0.5%C. Elemental nickel and carbon were mixed with the
iron powder. Blistering of the products after
sintering was observed in ~5% of the products.
Compaction
Sintering
4M parameter
Observations
High nickel content

Segregation of nickel powder
Elemental nickel additions
Not applicable
Low rapid burnoff temperature
H2 in hot zone
Yes
No
Yes
...
No
Yes
It was observed that delubing the keys was not

completed in rapid burnoff, and the lubricant
18/8/2008 3:53PM Plate # 0
Fig. 12
Region of suspected graphite agglomeration in the pinhole (Fig. 11) at higher magnification
Fig. 13
High carbon in electron-dispersive x-ray analysis, confirming the likely cause of the pinhole (Fig. 1012) as carbon-lubricant
segregation in blending
was carried into the hot zone, where the atmosphere consists of N2-10%H2. In the presence
of H2, nickel acts as a catalyst favoring the
cracking of the lubricant. This results in the
lubricant exploding, and consequently, blistered
products.
Corrective Measures. Two possible solutions to the problem were considered:

Ensure complete removal of the lubricant

prior to contact with H2
Reduce the catalytic activity of nickel
Due to the limitations of sintering in a continous mesh belt furnace, the former could not be
completely ensured. Thus, it was resolved that

reduction in catalytic activity of nickel could be
an effective solution. Sulfur effectively poisons
nickel and limits its catalytic properties, so
0.03% S was added to the blend.
Results. The addition of sulfur to the blend
effectively reduced the rejections from 5%
to nil.
Case Study 4: Dimensional Instability
during Shrink Fitting
Exhaust valve seats with tool steel powder
as a major constituent were shrink-fitted in a
pg 406
18/8/2008 3:53PM Plate # 0
pg 407
cylinder head of a sport utility vehicle. It was

observed that the valve seats were expanding
during shrink-fitting instead of contracting.
This resulted in interference between the
valve seat and the cylinder head, causing rejection of the cylinder heads. The parameters that
can cause dimensional instability during shrinkfitting are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Tempering
4M parameter
Observations
Not applicable
Not applicable
Not applicable
Not applicable
...
...
...
...
Yes
A significant amount of retained austenite

was observed even after tempering. This
retained austenite was transforming to martensite during shrink-fitting, where the products
were cooled to 60 C, which resulted in a
volume expansion.
Corrective Measures. A complete transformation of retained austenite can be effected
by:

Multiple tempering
Cooling to below the martensite finish temperature (in this case, a deep cryogenic
treatment)
Combination of tempering and deep cryogenic treatment
Trials revealed that a combination of subzero

treatment and tempering gave the best results
with respect to dimensional stability as well as
metallurgical properties.
Results. The valve seats subjected to a
combination of subzero treatment and tempering did not expand during shrink-fitting.
Further dimensional measurements taken
at 60 C were in line with the theoretically
calculated shrinkage in contrast to the earlier
expansion.
parameters that can cause wear of a free-graphite

bush are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Oil impregnation
4M parameter
Observations
Not applicable
Not applicable
High density
High sintering temperature
High sintering time
Cementite network
Insufficient free graphite
Low oil content
...
...
No
Yes
No
Yes
Yes
No
Metallography of the worn sample revealed a

cementite network in the product (Fig. 14). This
is a clear indication that excess carbon had gone
into solution, which resulted in the solid lubricant being depleted.
Corrective Measures. Sintering was carried
out at 1080 C for 20 min, which resulted
in excess combined carbon. The temperature
was reduced so that the combined carbon
was 50.8%. The sintering temperature for the
product was fixed at between 1040 and 1060 C
Results. Bushes sintered at 1060 C
for 20 min resulted in a pearlitic-steaditic
microstructure with no cementite network
(Fig. 15).
Case Study 6: Fracture of Steam-Treated

Part
Valve plates for compressors were subjected
to grinding after steam treatment. Fracture of
a steam-treated valve plate was observed during
grinding after steam treatment. The parameters
Case Study 5: Wear after Sintering

Bushes were produced with Fe-0.45%P
2%Cu-2.5%C and supplied in the oil-impregnated condition. These bushes were deliberately
undersintered so that some free graphite was
retained, which acted as a solid lubricant that
in turn minimized wear in application. The
Fig. 14
Pearlite-steadite microstructure with cementite network
18/8/2008 3:53PM Plate # 0
pg 408
Fig. 15
Pearlite-steadite microstructure free from cementite

network
that can cause the fracture of a steam-treated

part are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Steam treatment
4M parameter
Observations
Not applicable
Not applicable
Low density
Low soaking time
High dewaxing temperature
High temperature
Insufficient steam
Low soaking time
Masking of parts
...
...
No
No
No
No
No
No
Yes
No
No
No
No
No
were subjected to standard metallographic

analysis. The product with low hardness
revealed a predominantly ferritic structure with
free copper (Fig. 16), even though the blend
graphite addition was 0.8% (Fig. 17). Thus,
undersintering was suspected to be the possible
cause. However, the sintering control charts
revealed that the temperature as well as the
belt speed were meeting the specifications.
The loading pattern was then checked. It was
observed that the loading pattern followed
that which was recommended for a 46 cm
(18 in.) belt, whereas the belt width was 30 cm
(12 in.). This resulted in overloading of
parts, which consequently resulted in undersintering.
Corrective Measures. Suitable one-point
lessons were imparted to the sintering operators
to ensure that the right process plans were followed.
Case Study 7: Oxidation after Sintering
Exhaust valve seats for internal combustion
engines are produced with a chromium-rich
hard-phase alloy. The hard-phase alloy is
essential for wear resistance at elevated temperatures, to which the exhaust valve seat
is subjected. Oxidation of the hard phase was
observed after sintering. The parameters that can
cause oxidation after sintering are:
Operation
Hardness measurements revealed very low

hardness. Product with hardness conforming to
specification and product with low hardness
Fig. 16
Low-hardness sample with almost no combined

carbon. Only ferrite and free copper are observed.
Powder chemistry
Blending
Compaction
Sintering
Fig. 17
4M parameter
Observations
Chromium-rich alloy
Not applicable
Rusted compacts
High cooling zone dewpoint
Yes
...
No
Yes
Yes
Typical microstructure with ~0.7% combined

carbon
18/8/2008 3:53PM Plate # 0
pg 409
Chromium-rich particles are susceptible to

oxidation when the product is heated and cooled
in the range of 650 to 1050 C. Oxidation of the
hard-phase particles is favored by:


High cooling zone dewpoint
Figure 18 shows the microstructure with

oxidized hard-phase particles. The oxide film
around the hard phase does not allow the hard
phase to interact with the matrix. As a result, the
matrix will not have chromium carbides, which
helps the valve seats to retain wear resistance at
high temperatures. The rapid burnoff temperature was observed to be 650 C, and the cooling
zone dewpoint was between 20 to 30 C.
These are potential causes for oxidation.
Corrective Measures. The furnace was
allowed to stabilize for a period of 4 to 16 h. The
gas ratio was maintained at 70% N2 and 30% H2
during this period, and the gas volume was
increased by ~25%. The rapid burnoff temperature was restricted below 600 C. The valve
seats were sintered when the operating dewpoint
in the hot and cooling zones is 35 C or lower.
Results. The modification in sintering parameters resulted in the production of oxidationfree valve seats (Fig. 19).
density of 7 g/cc. Subsequently, the gear was

sintered, and a sizing operation was then carried
out for dimensional correction. Induction hardening of the gear was carried out after sizing, and
cracks were observed after induction hardening.
The parameters that can cause cracks after
induction hardening are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Sizing
Induction hardening
4M parameter
Observations
High carbon
Not applicable
Low density
Low sintering time
Not applicable
High heating time
Severe quench
Yes
...
No
No
No
...
No
No
Figure 20 shows the crack. This crack originates in the middle of the root of the gear and
Case Study 8: Cracks after Induction

Hardening
A PM transmission gear was subjected to
induction hardening for an application that
demanded high wear resistance as well as high
contact and bending strength. The gear had an
Fe-Cu-C chemistry and was compacted to a
Fig. 18
Exhaust valve seat with oxidized hard-phase

particles
Fig. 19
No oxide layer observed in the hard-phase particles

after implementation of corrective measures
Fig. 20
Crack in the middle of the root of the primary driven

gear
18/8/2008 3:53PM Plate # 0
pg 410
does not originate at any sharp corners. Hardness

profiles from the root indicated that the root has
been hardened to a very high hardness (Fig. 21),
compared to the earlier sample in which no crack
was observed. Microstructure near the root
revealed untempered martensite, which gradually became a mixture of martensite and bainite
and subsequently ferrite and pearlite. This is in
contrast to the earlier sample where the root was
not hardened and had a bainitic structure. Secondly, the combined carbon estimated from the
pearlite content in the sintered microstructure
was 0.7% compared to 0.6% in the product
that was successfully induction hardened. The
strain caused by the martensitic transformation
increases in proportion to the combined carbon
in the microstructure. Thus, the possible causes
for the crack were:

Too high a soaking time, which resulted in a

severe quench
High combined carbon in the sintered
microstructure
Corrective Measures. The formation of

martensite in the root was a result of quenching
from very high temperatures. The time was
reduced, and the coil design and quench were
modified. However, the martensitic microstructure could not be avoided in the root, and
cracking was still observed. It was concluded
that the product was unable to withstand the
stresses associated with the transformation. The
Fig. 21
Hardness traverse from root
carbon content was reduced in the blend to 0.6%,

which resulted in a lesser volumetric strain due
to martensitic transformation in the product.
Induction hardening with this chemistry yielded
no cracks, and the hardness traverse matched
that obtained in the rapid prototype samples.
Results. Products with a lower combined
carbon content did not crack when induction
hardened.
Case Study 9: Cracks after Quenching
A flyweight for a governor assembly in a
diesel engine was carbonitrided to meet the
product requirements for strength. This is a
complex part wherein multiple punches are used
to form the sections of different thicknesses.
The product was observed to crack after carbonitriding. The parameters that can cause cracks
after quenching are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Carbonitiriding
4M parameter
Observations
Not applicable
Not applicable
High density difference
Underfilling of powder
Low sintering time
High hardening temperature
Severe quench
...
...
No
Yes
No
No
No
No
The parts were subjected to magnetic particle

inspection prior to heat treatment. No cracks
were observed. Metallography was then carried
18/8/2008 3:53PM Plate # 0
out at the location where the crack occurred in

heat treatment. This is shown in Fig. 22. Insufficient powder filling was observed (Fig. 23).
The underfill results in a localized region of low
strength, which gives way during subsequent
quenching and results in crack formation.
Corrective Measures. The localized underfilling was due to the presence of a sharp corner
in the tool. A trial was conducted wherein the
sharp corner was smoothed by filing the parts.
No crack was observed after carbonitriding. It
was resolved to increase the radius in the tool so
that a sharp corner and consequently an underfill
are avoided. Products thus produced were heat
treated, and no cracks were observed.
Results Products produced with an improved
radius in the tool resulted in elimination of
product rejections due to quench cracks.
Case Study 10: Dimensional Change
in Carbonitriding
An angle lever for a diesel injection pump was
produced with FD-02-00 chemistry and carbonitrided to a case depth of 0.2 to 1.2 mm. When
the carbonitriding was established at a new heat
treatment source, it was observed that the bore
did not shrink sufficiently during carbonitriding.
The parameters that could cause a dimensional
change in carbonitriding are:
Operation
4M parameter
Powder chemistry
Blending
Compaction
Sintering
Carbonitiriding
Not applicable
Not applicable
Bore undersized due to tool wear
High sintering temperature
High sintering time
Low soaking time
Insufficient carbon and nitrogen
enrichment
High quench oil temperature
Insufficient transformation
Observations
...
...
No
No
No
Yes
No
Carbonitriding was carried out at 840 C for

45 min with a carbon potential of 0.8% in the
atmosphere. In the case of the vendor, better
furnace gas sealing ensured that this case depth
was achieved in 20 min. The hardness traverse
indicated that the longer soaking time had a
higher carbon martensite in the case as well as
a higher core hardness, which in turn resulted in
a case depth at the lower end of the specification
for a lower soaking time. This is shown in
Fig. 24.
Corrective Measures. Higher shrinkage is
ensured by a higher amount of martensitic
transformation. This was ensured by increasing
the soaking time from 20 min to 45 min. The
increased soaking time resulted in better exposure of the product to carbon and nitrogen
enrichment, which in turn resulted in higher
martensite content and the desired level of
dimensional shrinkage.
Results. The increased martensite content
resulted in increased volume expansion. Thus,
the shrinkage in the bore is higher with a
soaking time of 45 min. Products soaked at
840 C for 45 min were found to shrink sufficiently, and dimensions conformed to specifications.
Case Study 11: Low Surface Hardness

after Carbonitriding
Low surface hardness was observed in
an iron-copper part after carbonitriding. The
No
Yes
Fig. 23
Fig. 22
Quench crack in flyweight
Metallography of the region where the crack occurred in the sintered sample, showing a pre-existing
defect caused by poor powder filling
pg 411
18/8/2008 3:53PM Plate # 0
Fig. 24
Hardness traverse in bell crank lever
parameters that can cause low surface hardness

after carbonitriding are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Carbonitiriding
4M parameter
Not applicable
Not applicable
Low density
Low sintering time
Low soaking time
Insufficient carbon and
nitrogen enrichment
Excess carbon and nitrogen
enrichment
Delayed quench
Pearlite/bainite in case
Observations
...
...
No
No
No
No
No
Yes
No
No
Yes
After ensuring that the density conformed

to specification, the low-hardness products
were subjected to metallographic analysis. The
case microstructure revealed a high amount
of retained austenite, which is the cause of low
hardness (Fig. 25). High retained austenite
can result if the carbon potential as well as
the nitrogen potential are high. The oxyprobe,
which senses the carbon potential, was examined and found to be covered with soot. This
resulted in improper sensing of carbon potential,
leading to an excess of active carbon in the
furnace atmosphere, which in turn resulted in
high retained austenite.
Corrective Measures. The temporary corrective measure adopted to salvage the parts
was to carry out a deep freeze, because reprocessing may result in scrapping of the product
due to dimensional distortion. On a permanent

basis, a periodic preventive maintainence
schedule for the oxyprobe was drawn up and
adhered to.
Results. Correcting the oxyprobe yielded
products with a martensitic/510% retained
austenite structure that met the requirements for
surface hardness.
Case Study 12: Variation in Bore Diameter
after Heat Treatment
A valve retainer for a shock absorber was
produced from an FC-02-08 blend that was
hardened for wear resistance. High ovality was
observed after heat treatment. This resulted in
~10% of the products being out of specification.
Fig. 25
High retained austenite in the case
pg 412
18/8/2008 3:53PM Plate # 0
The parameters that can cause a wide variation

in bore diameter after heat treatment are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Carbonitiriding
4M parameter
Observations
Excess copper content

Elemental additions
Low density
Low sintering time
Volume expansion due to
martensitic transformation
Thermal contraction during
quenching
No
No
No
No
No
Yes
Yes
The ovality was checked after sintering and

after heat treatment. The data revealed that
hardening contributes significantly to distortion,
and the contribution of sintering is negligible.
Thermal contraction due to quenching and the
accompanying martensitic phase transformation
contributed significantly to the distortion.
Corrective Measures. To minimize the
distortion, it was resolved to eliminate the
hardening operation and achieve the required
hardness in sintering. Prealloyed powder containing nickel and molybdenum, blended specially to produce products with negligible
distortion after sintering, was adopted to achieve
both the hardness and dimensions after sintering.
Results. Changing the process from conventional hardening to sinter hardening reduced
the dimensional spread to 50% of the variation
observed in products with a separate heat treatment operation.
Case Study 13: Low Breaking Load after
Carbonitriding
A component for a gearbox application produced with a chemistry of FD-02-00 and having
an overall density of 7.2 g/cc was subjected
to carbonitriding to meet the product requirements for wear and mechanical strength. During
assembly, the product failed by fracture. The
parameters that can cause low breaking load
after carbonitriding are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Carbonitiriding
4M parameter
Not applicable
Not applicable
Low density
Low sintering time
High carbonitriding temperature
High time of carbon potential
attainment
Variation in batch quantity
Martensite in core
No case depth inspection
in thin section
Low breaking load indicated poor toughness

of the product. Poor toughness is related to both
microstructure and density. Lower densities
and a through-hardened microstructure with
martensite in the core can cause premature
fracture. Metallography was carried out in different sections of the failed product. Metallography of the sections indicated that the thinner
sections in the product were through hardened.
Figure 26 shows the through-hardened microstructure. Sectional density measurements indicated that the thicker sections had a density of
47.2 g/cc, and the thinner sections had a density
of 7.1 g/cc.
However, this through hardening of thinner
sections was not observed in the initial samples
that were tested. To determine the cause of this
variation across different heat treated batches,
the carbon potential attainment was studied for
various batches.
During establishment of the carbonitriding
process, the process parameters were fixed for
~500 parts. The same was subsequently fixed
for the full batch quantity of 3500. In this particular case, the carbon potential pickup in the
furnace was completely different for the trial
and bulk. Figure 27 shows the difference in
carbon potential pickup. There is a significant
difference in the time for attainment of a carbon
potential of 0.8%. Longer times between 0.6 and
0.8% in the bulk lots result in additional carbon
pickup, which resulted in through hardening of
the thinner sections.
Corrective Measures. The carbonitriding
temperature and time were reduced. The batch
quantity was fixed. The density in the component was increased to 7.2 g/cc min in all
Observations
...
...
No
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Fig. 26
Martensite in the core of thin sections in the product
pg 413
18/8/2008 3:53PM Plate # 0
Fig. 27
Difference in carbon potential attainment for change in batch quantity
sections, which ensured closure of interconnected pores. Inspection standards for sectional
density and case depth measurements in the
thinner section were introduced so that any
nonconformance would be detected and not
reach the customer.
Results. The corrective measures resulted in
a pearlitic-bainitic microstructure in the core of
the thin section, and the toughness and, consequently, the breaking load were achieved consistently.
Case Study 14: Dimensional Change in
Steam Treatment
Pistons having a density of 7 g/cc were steam
treated for wear resistance and corrosion resistance. During steam treatment, it was important
that the bore dimensions be maintained within
specification. It was observed that after steam
treatment, the bore was undersized. The parameters that can cause dimensional change in
steam treatment are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Machining
Steam treatment
4M parameter
Observations
Not applicable
Not applicable
Not applicable
Not applicable
Bore undersized
High temperature
Thick oxide layer
...
...
...
...
No
Yes
Yes
After ensuring that the machined dimensions

of the bore were in accordance with specification
prior to steam treatment, metallography was

carried out on the samples with the undersized
bore. An oxide layer of up to 15 mm was
observed (Fig. 28). Typically, Fe3O4 layers are
3 to 5 mm, because the lattice of Fe3O4 does not
allow further diffusion of oxygen molecules
once this layer is formed. Such a thick layer is
due to the formation of FeO. The steam treatment set temperature was 560 C for this product. Because the Fe3O4 reaction is exothermic,
the temperature of the products exceeded
570 C, resulting in the formation of FeO.
Corrective Measures. The steam treatment temperature was subsequently reduced to
530 C, where an oxide layer of 4 to 6 mm was
obtained (Fig. 29). This resulted in the bore
dimensions meeting the specification.
Fig. 28
560 C
Thick oxide layer indicative of FeO formation that

was obtained at a steam treatment temperature of
pg 414
18/8/2008 3:53PM Plate # 0
pg 415
Fig. 29
530 C
Oxide layer of 6 mm, indicative of Fe3O4 obtained on

reduction of the steam treatment temperature to
Case Study 15: Low Core Hardness after

Steam Treatment
Crankshaft sprockets with an Fe-Cu-C
chemistry and a density of 6.6 to 6.8 g/cc were
supplied in the steam-treated condition. The
customer demanded that the hardness in the
surface and core meet the specification of
140 HV10 min. The parameters that can cause
low core hardness after steam treatment are:
Operation
Powder chemistry
Blending
Compaction
Sintering
Sizing
Steam treatment
4M parameter
Observations
Low carbon
Not applicable
High density
Low sintering time
High ferrite content
Closed surface pores
Oil in pores
Insufficient time
Low temperature
High temperature (4570 C)
Masking of products
Nonuniform steam circulation
Yes
...
No
No
No
Yes
No
Yes
Yes
Yes
No
Yes
No
It was observed that within a batch, some

products had a core hardness greater than 140
HV10. Low and high core hardness samples were
subjected to metallography. It was observed that
higher-hardness samples had better steam
penetration compared to low-hardness samples.
Corrective Measures. To ensure higher
core hardness, the products required greater
exposure to steam. This was achieved by changing the steam treatment temperature and time
from 530 C/30 min to 550 C/2 h. Masking of
products can also potentially prevent uniform
exposure to steam. To eliminate masking, the
loading pattern was changed from a double layer

to a single layer. However, even after increasing
the steam exposure time and improving the
loading pattern, variation of hardness was
observed within the same batch. A trial was
conducted wherein oil-free sintered samples
were steam treated with sized and oiled parts. It
was observed that the sintered samples have
uniform steam penetration, and the thickness of
the steam oxide layer is uniform throughout the
part. This is in contrast to the sized products with
low core hardness, wherein the oxide layer is
minimal. The presence of oil or a related organic
residue was suspected of interfering with the
steam penetration. An oil that has a higher
volatile content and a lower organic residue
content was used for rust prevention, and the
same was used as a lubricant in sizing. The
change of oil resulted in a core hardness conforming to specification. However, the hardness
obtained ranged from 145 to 190 HV. The presence of products with hardness marginally
above 140 HV10 indicated that the hardness
must exceed 160 HV to avoid the chances
of core hardness going below specification.
This additional improvement in hardness
was achieved by having a completely pearlitic
microstructure instead of a ferrite+pearlite
microstructure, which was effected by a blend
addition of 0.8% C.
Results. Implementation of the corrective
measures resulted in the core hardness being
achieved consisently in accordance with specification, as follows:
Properties
Specification
Surface
hardness
Core hardness
140
HV10 min
140
HV10 min
Sintered
...
microstructure
Steam-treated
...
microstructure
Before
After
152208
185235
110173
170207
Ferrite+
100% pearlite
pearlite
Complete steam
Incomplete
penetration with
steam
oxide layer of
penetration
with low-oxide 46 mm in core
layer52 mm
in core
REFERENCES
1. Asm Handbook, Volume 7, Powder Metal

Technologies and Applications, 1998
2. E. Klar and P. Samel, Powder Metallurgy
Stainless Steels: Processing, MicroStructures, and Properties, ASM International, 2007
18/8/2008 3:58PM Plate # 0

DOI: 10.1361/faht2008p417

Induction Hardening
Janez Grum, University of Ljubljana
INDUCTION HEATING, in most applications, is used to selectively heat only a portion of

the workpiece that requires treatment. This
usually means that the process can be accomplished in a relatively short time and with high
efficiency, because energy is applied to the
workpiece only where it is needed.
Induction surface hardening is applicable to
axisymmetric or near-axisymmetric machine
parts in steel or cast iron that are produced in
substantial volumes. There are two basic techniques for induction hardening machine parts:
single-shot and scanning. The former employs
selective heating and quenching to harden a
specific area or areas of the machine part in one
operation. The latter is usually applied to harden
progressively long, continuous sections, such as
shafts and spindles. In this instance, the scanning
inductor traverses the length of the section,
heating only a relatively small area at any given
time, and is followed closely by the quench
arrangement, which is often an integral part of
the inductor.
These advantages make it possible for induction hardening to be fully automated and are
especially suitable for a large series of workpieces.
The induction-hardening procedure enables
an engineer, by simply adapting the shape of the
induction coil, to ensure the desired shape of the
hardened profile of the surface layer. Likewise,
the engineer can surface harden only that part of
the surface (local hardening) on which a certain
increased level of hardness and wear resistance
are desired. One of the main advantages of induction hardening is the ability to harden a surface layer only in certain places at a defined
penetration depth and shape.
For dynamically loaded machine parts, it is
very important to ensure the total compressive
stresses in the thin, most-loaded surface layer.
The total stresses are a sum of residual stresses in
a machine part and of load stresses produced by
the action of external forces and moments. To

ensure a long life of the machine part, knowledge of the residual stresses in the machine part
and how to adjust the size and distribution of the
residual stresses by means of the selection of an
appropriate production technology are very
important.
In surface hardening, compressive residual
stresses always occur in the thin surface layer
due to martensite transformation. The size and
variation of the residual stresses depend primarily on carbon content and less on the type
and content of alloying elements in heat treatment and surface-hardening steels. The variation
of residual stresses in the surface layer can be
modified by varying the induction-heating conditions and by a quenching method.
Induction surface hardening creates very
desirable residual stresses in the hardened surface layer. Residual stresses are always of a compressive nature and are usually present to the
depth of the induction-hardened layer. Residual
internal stresses, that is, the so-called residual
stresses, are the stresses present in a material
or a workpiece when there is no external force
and/or external moment acting on it. The residual stresses in metallic machine parts have
attracted the attention of technicians and engineers only after manufacturing processes improved to the level at which the accuracy of the
manufacture exceeded the size of deformation,
that is, distortion, of a workpiece/product.
Thus, it was almost 150 years ago that the
effect of internal stresses on plasticizing, that is,
destruction, distortion, and plastic deformation,
of workpieces was already known. It was then
that experts introduced measurement of individual dimensions of products. For a given type of
machining process, they connected the influence
of the selected machining conditions with the
size of dimensional deviations. This was also the
beginning of an expert approach to the selection
of the most suitable machining and/or heat
pg 417
18/8/2008 3:58PM Plate # 0
treating conditions based on the criterion of

minimum dimensional deviations, that is, minimum workpiece distortion. Currently, measurement of individual workpiece dimension is a
very practical, uncomplicated, and reliable
method of product quality assessment. Through
the appropriate selection of the grinding wheel
and grinding conditions and taking into account
the physical and mechanical properties of the
workpiece material, very favorable compressive
residual stresses in the hardened surface layer
can be retained.
How is it possible to assure a desirable surface
and surface layer quality after induction hardening and fine grinding? Finding an answer to
this question requires a very good knowledge of
the grinding process on the microlevel as well as
knowledge of mechanical and heat effects acting
on the layer of the workpiece, including the type
of grinding wheel and the grinding conditions.
An all-inclusive consideration of the numerous influences of the tool type and condition on
the changes on the surface and in the surface
layer of the workpiece in the given machining
conditions is described by the term surface
integrity. This is a scientific discipline providing
an integral assessment of the surface and subsurface layer. It was defined at the beginning of
the 1960s. For high-quality machine parts and
parts subjected to heavy thermomechanical
loads, different levels of description of the surface integrity were defined. A basic level of the
surface-integrity description includes measurement of roughness and analysis of the microstructure and microhardness in the thin surface
layer resulting from the machining process
under given machining conditions. The second
level of the surface-integrity description includes studies of residual stresses in the surface
layer and of mechanical properties of the given
material. The third level of the surface-integrity
description includes tests making clear the behavior of the given part under the operating
conditions.
As essential advantage of induction surface
hardening is that it is possible to achieve a sufficient repeatability of the hardened layer thickness on the workpiece as well as a desirable or
even prescribed hardened layer profile, ensuring
sufficient hardness and favorable distribution
of residual stresses in the hardened layer. A
variety of steels and a whole range of inductionhardening methods provide the possibilities
for very accurate planning of residual stress size
and distribution. This is of growing importance,
since manufacturers are frequently required to

produce machine parts that, among other surface
properties, must have quite specific residualstress distribution along the depth of the hardened layer. It has become a proven fact that high
compressive stresses ensure high fatigue
strength of machine components and reduce the
danger of crack occurrence and growth on the
surface of components.
In thermal hardening, the surfaces of suitable
materials, usually plain carbon, low-alloy steels,
or cast irons, are austenitized and then quenched
to produce a hard martensitic case that is usually
tempered in a subsequent operation. Case depths
are normally in the range of 0.5 to 5 mm. Case
hardness is typically approximately 700 HV on
hardening and 600 to 650 HV after tempering at
200 C (Ref 1).
Heating processes include electrical induction
and resistance (Ref 24), and direct impingement methods using flames, (Ref 5), lasers
(Ref 611), and electron beams (Ref 12). Of
these, induction heating is the most widely used.
Laser and electron beam heating have recently
become established in a number of applications,
mainly where distortion was a problem with the
induction method.
Thermal treatments are mainly employed
when only local areas on machine parts require
hardening and wear resistance. Consequently,
such methods are energy efficient compared
with thermochemical methods where bulk heating of batches of components is undertaken.
Normally, the surface-hardening process also
introduces compressive stresses into the surface
layers, leading to an improvement in fatigue
properties. For example, the drive shafts of
heavy lorries and buses are induction surface
hardened to improve their fatigue properties
(Ref 10, 11).
The skills and experience required for manual
flame hardening have been largely superseded
by automated flame techniques or by the induction method. However, for one-off machine
parts, the simplicity of manual flame hardening
and the extremely low capital investment ensure
that the method is still used on a regular basis
(Ref 5, 13).
From a heat treatment point of view, laser can
be considered as a versatile and flexible highintensity heat source that can operate in air. It is
capable of undertaking a range of processes,
essentially simultaneously, since the laser beam
can be directed through air by metal mirrors and
switched and shared among a number of work
pg 418
18/8/2008 3:58PM Plate # 0
pg 419
Induction Hardening / 419
stations. Manipulative techniques using mirrors

allow the beam to be directed to areas not
accessible by other techniques, for example, the
bores of tubes. Set against these considerations
are the high capital cost and low energy efficiency of the technique (Ref 1, 6, 10).
Steels for Surface Hardening

Heat treatable steels contain between 0.2 and
0.6% C and can be carbon steels or low-alloyed
steels (Table 1) (Ref 14, 15).
In practice, engineers choose these steels
because, in the soft condition (normalized or soft
annealed), they are more suitable for machining,

while their strength properties and hardness can
be subsequently refined by additional heat
treatment. Soft steels have good machinability,
while additional heat treatment improves their
mechanical properties to a desirable level.
Figure 1 presents a schematic of the heat treatment methods used to achieve the desired
mechanical properties (Ref 15, 16). Besides the
mentioned surface hardening, other thermochemical methods can be used to create a desirable wear resistance. The steels for surface
hardening are from the group of heat treatable
steels, except that they usually contain between
0.35 and 0.5% C. Heat treatable steels can be
Table 1 Suitable steels for induction surface hardening

Chemical composition %
Type of steel
1
2
3
4
5
6
7
8
9
10
11
Si
Mn
P, max
S, max
Cr
Mo
Ni
0.330.39
0.380.44
0.430.49
0.480.55
0.500.57
0.420.48
0.340.40
0.380.44
0.380.44
0.380.44
0.370.43
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.150.40
0.500.80
0.500.80
0.500.80
0.600.90
0.400.70
0.500.80
0.600.90
0.600.90
0.500.80
0.701.00
0.500.80
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.035
...
...
...
...
...
0.400.60
0.901.20
0.901.20
0.901.20
0.400.60
0.600.90
...
...
...
...
...
...
...
...
0.150.30
0.150.30
0.150.30
...
...
...
...
...
...
...
...
...
0.400.70
0.701.00
Source: Ref 14, 15
Fig. 1
Heat treatment methods for carbon and low-alloyed steels. Source: Ref 15
18/8/2008 3:58PM Plate # 0
supplied in soft or refined conditions. Carbon

steels with less than 0.35% C are in the normalized condition, and the products made from
them can be in the soft or refined condition. For
improvement of the wear resistance of these
steels, additional surface hardening can be
applied (Ref 14, 17, 18).
Below 0.35% C, the treated surface is insufficiently hard, and above 0.6% C, there is a
danger of surface cracking after quenching. To
obtain a satisfactory hardening response in
induction surface hardening, it is necessary for
the carbides to dissolve when the steel is heated.
Thus, steels that have been quenched and tempered and have small carbide particles respond
better than steels with large spheroidized carbides. Similarly, the rate of solution of carbides
in steels other than plain carbon and low-alloy
steel is too slow for them to be satisfactorily
induction hardened. Steels must also have sufficient hardenability to achieve the required
hardness at the specified depth (Ref 17).
The procedures are similar for fabrication of
machine parts made of carbon steels with more
than 0.35% C and alloyed steels, expect that
because of the higher machinability required,
they must be in the soft-annealed condition.
For induction surface hardening, it is recommended to use steels with an appropriate carbon
content (0.35 to 0.45%) and a careful selection
of alloyed elements. Usually, these steels have
an increased silicon (0.15 to 0.40%) and manganese (0.50 to 1.00%) content and other
alloying elements, such as chromium, molybdenum, and nickel. More highly alloyed tool
steels (O1, D2, D3, A1, and S1) and some martensitic stainless steels (AISI 416, 420, and
440C) are also sometimes induction hardened.
Among alloyed steels, chromium, chromiummolybdenum, and chromium-molybdenumnickel steels prevail. With given combinations of
alloying elements in induction-hardened machine parts, the internal stresses in heating as well
as in cooling can be controlled. The steels are
normally quenched in water. In certain cases, the
alloy steels can be cooled by means of an oil
emulsion. Steels can also be oil quenched. Having selected the right shape of the product and the
right choice of technology, it is possible to
expect only minimal distortions of the workpiece
after completion of the heat treatment (Ref 14).
If the starting points are the hardness and
residual-stress profiles, then the induction-heating conditions should be adapted to the selected
steel. Induction heating of a machine-part
surface to the austenitic zone above TA3 is of

major importance, since not only the case depth
but also the right through-depth residual-stress
and microhardness profiles of the machine part
are to be provided. The through-depth microhardness profile of the surface layer depends,
with the given steel, on the induction-hardening
conditions. Induction heating may be controlled
by infrared thermometers. An infrared thermometer is placed close to the induction coil, which
means that only the maximum surface temperature may be measured. This technique is quite
simple and practical. It allows a user to determine
the required, that is, optimal, power density for
heating. When a multiturn coil is used, for
example, with longer workpieces, several infrared optical-fiber thermometers may be employed. Thus, with sufficient spacing between
two adjoining turns, the temperature may be
measured at different locations on the machine
part. A difficulty may arise due to nonuniform
heating along the coil movement. By collecting
and processing the gathered measurement data,
the process quality may be assessed. In scientific
research, temperature measurement with
thermocouples may be applied too. In induction
surface heating, temperature measurement using
thermocouples is quite exacting, since they are
embedded in the machine part. The hardened
state of steel is to be obtained at an appropriate
depth. The thermocouples embedded at a particular location in the workpiece permit temperature measurements at this location only. In
progressive hardening, however, the momentary
temperature is measured. Thus, the so-called
temperature cycles in heating, as well as in
cooling, are obtained. The temperature cycles
make it possible to predict the efficiency of surface hardening. The choice of energy input, that
is, power density, in an appropriate nomogram,
that is, the choice of power and heating time with
reference to the shape of the induction coil and
case depth, applies only to stationary heating and
hardening of machine parts. This means that in
progressive hardening the data obtained on the
energy input are of informative character only.
Main Features of Induction Heating

Induction heat treatment is a segment of the
much larger technical field of induction heating,
which combines many other industrial processes
using the phenomenon of heating by induction
(Ref 1, 16, 1822).
pg 420
18/8/2008 3:58PM Plate # 0
Induction heating is often one of the most

effective heat treatment processes available for a
variety of applications, including:

Surface hardening
Through hardening
Tempering and stress relief
Annealing and normalizing
Grain refinement
Precipitation hardening or aging
Sintering of powdered metals
In most of these applications, induction

heating is used to selectively heat only the portion of the workpiece that requires treatment
(Ref 2). This usually means that the process can
be accomplished in a relatively short time and
with high efficiency, because energy is applied
to the workpiece only where it is needed.
One of the main features of induction heating
compared with conventional heating procedures
is that heat is generated in the workpiece itself.
In conventional heating procedures, the heat
input achieved is only 5 to 200 kJ/m2s energy,
whereas in induction heating this energy input
is 300 MJ/m2s. In induction heating, heat
penetrates into the workpiece by the aid of
Fig. 2
high-frequency alternating current, the choice of

frequency depending on heating requirements.
Induction heating power supplies are frequency
changers that convert the available utility line
frequency power to the desired single-phase
power at the frequency required by the induction
heating process. They are often referred to as
converters, inverters, or oscillators, but they are
generally a combination of these. The converter
portion of the power supply converts the line
frequency alternating current input to direct
current, and the inverter or oscillator portion
changes the direct current to single-phase alternating current of the required heating frequency.
Many different power supply types and
models are available to meet the heating requirements of a nearly endless variety of induction
heating applications (Ref 21, 22). The specific
application will dictate the frequency, power
level (Fig. 2), and other inductor parameters
such as coil voltage, current, and power factor
(cos Q) or Q factor (Ref 15, 19).
Advantages in Surface Hardening of
Machine Components. Induction hardening
is most often used for surface hardening of
machine components and has the following
Typical power-frequency regions of induction heat treatment applications. Source: Ref 15, 19
pg 421
18/8/2008 3:58PM Plate # 0
advantages over other procedures (Ref 2,

1922):

Relatively short heating times

Heating procedure is not strictly governed
by hardening temperature. All that matters is
that the heating process does not end at too
low a temperature, because sometimes it is
necessary for the transformation into austenite. The heating temperature is limited by
the solidus-line temperature, since the process is carried out while the material is in the
solid condition. Due to a short heating time,
there is no danger that the austenite grains
would grow at higher austenitization temperatures, which also means there is no
danger of coarse and brittle martensite formation.
The quenching procedure is easy to perform,
contributing to short surface-hardening
times. In progressive induction hardening, a
spray coil is located directly below the inductor and quenches the heated surface. In
single-shot hardening, the inductor is designed to perform the function of heating as
well as quenching. The coil around the workpiece functions as an inductor in the heating
phase. After the austenitization temperature
has been reached, the current is interrupted,
and the coil functions as a spray for
quenching.
Induction hardening is a short procedure that
does not require any additional protection
against oxidation. Thus, compared with
other similar procedures, such as cementation, it does not require much subsequent
machining.
Due to the nature of the procedure, the
workpieces, especially if symmetrically
shaped, are less susceptible to undesirable
deformations after induction hardening. The
volume changes in the workpieces after hardening the surface layer can be very accurately
predicted or estimated. The volume changes
after induction hardening of thin layers are so
small that quite often the function of the
machine component is not affected.
Especially in induction hardening of thin
layers and workpieces with low mass, it is
possible to achieve the desired critical rate of
cooling by self-cooling in air alone, that is,
by heat conduction from the heated surface
layer into the remaining cold part of the
workpiece. With thicker layers and workpieces of greater mass, it is necessary to use
quenching agents that move the actual

cooling rates close to the critical cooling
rate. These requirements can be met with the
right selection of quenching oils or polymer
water solutions. Practical experience has
shown that polymer water solutions are very
suitable for quenching of induction-heated
surfaces, since optimal quenching can be
ensured with the right choice of concentration of the polymer water solution.
The induction-hardening procedure enables
the engineer, by simply adapting the shape
of the induction coil, to ensure the desired
shape of the hardened profile of the surface
layer. Likewise, the engineer can surface
harden only that part of the surface (local
hardening) on which a certain increased level
of hardness and wear resistance is desired.
One of the main advantages of induction
hardening is the possibility of hardening a
surface layer only in certain places, at a
defined penetration depth and shape.
These advantages make it possible for
induction hardening to be fully automated
and are especially suitable for a large series
of workpieces.
Induction hardening always leaves compressive residual stresses in the surface layer,
which makes machine components more
resistant to dynamic loads. Compressive
residual stresses in the surface layer after
induction hardening prevent the occurrence
of cracks in dynamically loaded components
and prevent the growth of existing cracks
on the workpiece surface, if these are present
due to hardening or grinding.
Induction hardening is appropriate for smallsized workpieces, since, by a well-chosen
technology of heating and cooling or quenching, a hardened surface layer and a refined
core can be ensured. Thus, the required wear
resistance of the machine component at a
certain location as well as the required loadbearing capacity of the component can be
created, experiencing only a slight loss in
toughness of the core.
Induction Hardening of Machine Parts

Working Coil and Procedure during
Induction Hardening. Methods of induction
hardening of thin surface layers are always
adapted to the product size and shape and the
pg 422
18/8/2008 3:58PM Plate # 0
requirements of location, size, and shape of the

hardened layers profile.
Designing an induction coil, however, is very
specific to a given application and can be considered the most important aspect of the engineering of the system. There are several major
functions that an induction coil must perform to
make a job successful:

Induce current in the load so that the proper

heating pattern is obtained
Accomplish the proper heating pattern with
as great an efficiency as possible
Provide an impedance match to the generator so that adequate power can be transferred to the load
Have a geometry that will accomplish the
aforementioned three major functions and
permit easy loading and unloading of the
part being heated
First, the outer and inner surface on machine

components must be distinguished. The size and
shape of the induction coil must be adapted to
the size and shape of the product, and during the
heat treatment process, the distance should be
adapted to the workpiece. It is important that the
design of the machine component be planned for
induction-hardening specifics. A uniform gap
size between the induction coil and the workpiece surface is very important, since the energy
penetration depends on it. This is why singleshot hardening, where the workpiece and the coil
are at a standstill, is rarely used. To ensure a
uniform gap, single-shot hardening is almost
Fig. 3
always combined with a rotating motion of the

workpiece, or, on long workpieces, progressive
hardening is used.
Figure 3 shows different types of coils for
induction hardening designed for external surfaces (Fig. 3a), internal surfaces (Fig. 3b), and
front surfaces (Fig. 3c) of the workpiece. In all
three cases, the shape of the induction coil is
adapted to the size and shape of the workpiece surface that is to be hardened (Ref 2, 15, 16,
1922).
Workpiece sizes depend on the manufacturing
possibilities of high-frequency coils and their
types. This restricts the use of induction hardening primarily to small-sized machine components. Induction hardening of internal surfaces
requires a great deal of knowledge and experience in the design and manufacturing of induction coils of small diameters. The smallest
internal diameter of a machine component that is
to be induction hardened depends on the manufacturing possibilities of making small-diameter
coils, the capacity of the high-frequency generator (power, frequency), and the positioning
accuracy of the induction coil inside the
workpiece.
The induction coils are made of materials with
the highest possible electrical conductivity.
Consequently, copper or silver is usually chosen
to produce the induction coils. The induction
coils are adapted to the shape of the machine part
at the location where heating, that is, surface
hardening, is performed. The choice of the size
and shape of the induction coil depends on the
Significant types of coils for induction heating. Source: Ref 15, 16
pg 423
18/8/2008 3:58PM Plate # 0
Fig. 4
Influence of shape and arrangement of induction coil around machine part on heating, that is, hardened-pattern profile
method of heating, that is, hardening. With the

single-shot heating method, the induction coil
may be adapted by choosing an adequate pipe
cross section and by adapting the number of
turns used so as to provide simultaneous heating
of the entire area at the machine part. Figure 4
shows the influence of the induction coil shape
on the heating profile (Fig. 4a) and that of the
distance between the individual turns of the
induction coil and the gap on the heating profile
(Fig. 4b).
The concentration of the heat generated in the
machine part heated by the induction coil is
symmetrically and uniformly distributed across
the cross section, so that no distortion of the part
may occur after quenching. The symmetry of the
hardened-pattern profile thus reduces the risk of
distortion of the machine part. The induction
coils are made of pure copper, which can be
nicely shaped as required to produce coils. Because the electrical conductivity of the material
is reduced due to plastic deformation of the
material, it is necessary to adequately anneal the
coils to soften the material. Thus, the highest
possible electrical conductivity is ensured. It is
also necessary to prevent the formation of copper oxide at the coil surface to preserve highest
possible electrical conductivity of the material.
This will provide a high efficiency of induction
heating.
In single-shot induction heating, that is,
induction surface hardening, it is important to
provide the same gap size all around the cylindrical machine part. Because such an arrangement can be very exacting, the heating process is
usually performed with the rotation of the
machine part. Figure 5(a) shows single-shot
hardening performed with a different number of
Fig. 5
(a) Influence of multiturn coil on hardened-pattern

profile. (b) Influence of shape and gap size of induction
coil turns on hardened-pattern profile. (c) Influence of number of
induction coil turns on hardened-pattern depth
inductor turns along the specified hardened

pattern. This means different divisions among
the individual coil turns. Under the same heating
conditions, considerable difference in the
hardened-pattern depth will be obtained with
the same heating time.
Figure 5(b) shows different examples of
shaping a multiple-turn induction coil around
the cylindrical machine part. The figure indicates that the gap size along the cylindrical part
changes, which produces different hardenedpattern profiles. Figure 5(b) shows the influence
pg 424
18/8/2008 3:58PM Plate # 0
of a multiple-turn coil with too-large divisions of

the individual coil turns; therefore, the same heating depth, that is, hardened-pattern depth, along
the coil length is not achieved. Thus, individual
coil areas showing a smaller hardened-pattern
depth are obtained at the machine part.
Figure 5(c) shows the influence of the cross
section of a single-turn coil on the hardenedpattern profile.
For hardening of axles, shafts, and similar
lengthy machine parts, the scan-hardening process with a single-turn coil of an adapted cross
section is chosen. Such a hardening process is
necessary, otherwise, high-energy densities and
very long multiple-turn induction coils would be
needed. Figure 6 shows a diagram of heating, that
is, through hardening, of a cylindrical rod of
different diameters (D, in centimeters) with different inductor powers (P, in kilowatts), and a
given frequency ( f, kilohertz) and inductor
movement (v, in cm/s). The diagram shows a
linear dependence between the connected power
at the coil and the depth of hardening, which
means that the choice of a higher power produces
hardening at a greater depth, the movement
velocity being the same. It can also be established
which diameters of the axles or shafts can be
through hardened if the same power is available
and different movement velocities are chosen.
It is possible to distinguish between:

Single-shot hardening
Scanning or progressive hardening
Tooth-by-tooth hardening or gap-by-gap
hardening
Fig. 6
Quenching techniques are an important

design feature of induction-hardening equipment. The important questions to be answered
when determining quenching systems include
(Ref 2, 17, 23, 24):

Workpiece size and geometry

Hardenability of steel
Type of austenitizing operation (surface or
through hardening)
Type of heating method (single-shot or
scanning)
Type of quenchant
The two most common types of systems

consist of spray quench rings and immersion
techniques. When quench rings are used for
round bars, their shape, like the coil, is generally
round (Ref 2, 23, 25). Figure 7 shows different
ways to single-shot harden the surface layer.
Common to all of them is that heating is
performed along the length of the layer, so that
the induction coil embraces the particular part of
the workpiece where heat treatment is to be
performed (Ref 15, 16). The ring may be located
concentric with the coil (Fig. 7c) or directly
underneath or alongside it (Fig. 7b), as in the
single-shot induction hardening setup.
In any of these ways, it is important to ensure a
constant gap size between the induction coil and
the workpiece surface. This is usually achieved
by a rotating motion of the workpiece. In the first
case, shown in Fig. 7(a), the cooled induction
coil embraces the object at a certain height. After
heating to the austenitization temperature is
Selection of heating conditions in scan hardening to provide through hardening of a cylindrical rod
pg 425
18/8/2008 3:58PM Plate # 0
completed, the workpiece is quenched in a

quenching bath, or quenching is done by moving
the workpiece into the quenching spray located
beneath the induction coil (Fig. 7b). Figure 7(c)
shows a method of heating and cooling in the
induction coil, which is made so that it enables
heating of the workpiece surface for a certain
period, and then, on switching the current off, the
same coil is activated as a spray for quenching.
The induction coil is usually made of copper
tubing with additions of alloying elements
that do not affect the electrical conductivity of
copper but contribute to good machinability
of the material. Good machinability is required
because it is necessary to make the openings
for the quenching spray. Thus, during the entire
Fig. 7
heat treatment, the workpiece is rotating, and in

this way, uniform heating and quenching are
ensured. The heating and quenching conditions
created in this way produce a uniform hardened
surface layer and ensure repeatability of thickness and shape of the hardened profile.
In induction scanning or progressive hardening, workpieces move through the quench ring
and coil, with quenching occurring immediately
after heating (Fig. 8). For nonsymmetric workpieces, the quenching system, like the coil, is
generally the same shape as the workpiece (Ref
15, 16).
The second method of induction hardening
involves a group of procedures of progressive
or progressive-rotating hardening. This type of
Single-shot induction surface hardening of a cylindrical workpiece. Source: Ref 15, 16
pg 426
18/8/2008 3:58PM Plate # 0
hardening is employed when the required length

of the hardened surface is longer than the length
(size) of the coil. In this case, the correct heating
mechanism can be ensured by the optimal induction coil design (number of coil turns, length
of the coil) and suitable energy inputs for a given
size and profile of the hardened layer (Ref 23,
25, 26). To ensure the uniformity and repeatability of the hardened layer, progressive motion
of the induction coil must be provided, or the
workpiece is fed progressively through the coil
while rotation of the workpiece takes place at the
same time. This method of induction hardening
is presented in Fig. 8. It is possible to design the
induction coil separately from the spray (Fig. 8a)
or combined so that the induction coil has
downward-directed spray openings for quenching in its bottom part (Fig. 8b).
Gear wheels belong to very demanding
machine elements, from the point of view of
mechanical machining as well as heat treatment.
In hardening smaller gears, that is, gears with
small modules or small diameters, the method of
single-shot hardening can be used to ensure a
varying thickness of the hardened layer (Ref 15,
16). The thickness is maximum on the tooth tip
and then decreases toward the tooth root.
Fig. 8
Figure 9(a) shows single-shot hardening of tooth

tips, where the coil encircles the whole gear.
This method is very simple and suitable for highfrequency heating of gears with a module up
to 3 mm and for smaller gear diameter. In this
case, it must be ensured that the tooth flanks are
hardened at least along the length that is otherwise activated in the mesh, since the thickness of
the hardened layer varies, being highest at the
tooth tip and gradually decreasing toward the
tooth root. This method of hardening is suitable
only for gears subjected to low loads, just to
increase the wear resistance of the gear (Ref 16,
23, 26, 27).
Figure 9(b) shows induction surface hardening, where the whole tooth and a certain area
below the tooth root are hardened. This method
of induction surface hardening is appropriate for
gears with a module up to 5 mm. Heating is
achieved by two systems: first, with a current of
medium frequency, and then in the final phase
for a relatively short period by heating with a
high-frequency current. By quenching, a desirable hardness of the tooth flank surface and
increased strength of the gear tooth can be
achieved. In the root of the tooth, due to a refined
microstructure, an increased material fatigue
Progressive induction hardening of a cylindrical workpiece. Source: Ref 16
pg 427
18/8/2008 3:59PM Plate # 0
strength can be noted. Gears heat treated in this

way are suitable for highly loaded machine parts
(Ref 2830).
Figure 9(c) shows hardening of individual
gear teeth, where the coil encircles the tooth and
the individual teeth are heated in turn and
quenched directly after heating. In the case of
large-module gears, simultaneous heating of the
left and right flank of the same gear tooth can be
performed (Fig. 9d).
Induction hardening of all gear teeth (Fig. 10)
can be done only for gears with a module smaller
than 3 mm and a gear diameter up to 100 mm.
This kind of hardening requires high-power
frequency generators, for example, 200 kW.
For gears exposed to higher loads, where
refinement of the whole tooth is desirable, a
power of 30 to 40 kW and longer heating times
can be selected.
To ensure a desired penetration depth and
profile of the induction-hardened layer, the following heat treatment parameters can be varied
in a single-shot or progressive induction hardening by choosing a suitable power density and
feed rate:

The size and shape of the induction coil

adapted to the workpiece
Kind of steel and its thermal properties
Size and mass of the workpiece on the
location where induction hardening is to be
performed
Quenching agent and method of quenching
Fig. 9
A number of graphs and nomographs are

available for this purpose, offering a selection of
heat treatment conditions for heating and cooling. The most important data are the starting
point data on energy input and frequency of the
current, and the temperature to achieve in
induction hardening. From these data, the time
necessary for heating, in the case of single-shot
hardening, or the feed rate of the workpiece or
the rate at which the coil should move along the
workpiece can be defined (Ref 2, 1820, 23, 27,
31, 32).
Figure 11(a) illustrates induction surface
heating with a medium-and high-frequency
current. The procedure is known as doublefrequency heating (Ref 27). Here, the gear is first
placed into the coil fed by the mediumfrequency current. Then, the gear is moved into
the high-frequency coil, where only the surface
layer of the gear tooth is reheated with a highfrequency current. When both phases are completed, the gear is dropped or moved into a
quenching bath. In this method of heating with a
double frequency, the progressive motion of the
workpiece can be combined with additional
rotation.
Additional rotating motion ensures a uniform
reheating of the surface layer and results in
uniform microstructural changes on the left and
right side of the tooth. In this way, undesirable
and nonuniform dimensional deviations between the left and right tooth flanks are prevented. The same method of induction surface
Different methods of induction surface hardening of gear wheels. Source: Ref 15, 16
pg 428
18/8/2008 3:59PM Plate # 0
heating is shown in Fig. 11(b), where the same

coil is used for heating with a medium- and highfrequency current (Ref 27).
The second method is hardening of individual
gear wheels or individual gear gaps. Figure 12
shows induction heating of a gear wheel toothby-tooth and Fig. 13 shows induction heating
with the gap-by-gap method. In both methods,
the tooth surface is first heated with the induction coil and then quenched with a specially
adapted spray system.
Figure 14 shows the shape and position of the
massive inductor placed around the gear tooth
(Ref 27). The coil is shaped so that the gap
between the coil and the tooth surface varies.
Only in this way is it possible to ensure a uniform thickness of the layer in the middle and on
the edge of the tooth (Ref 3337).
Difficulties with this method of gear wheel
hardening occur when the tooth gaps are too
small and the coil heats up the adjacent flank.
This method of induction surface heating of gear
teeth is often not appropriate, since after heating
the adjacent tooth, the next step does not ensure
Fig. 10
the desirable hardness of the adjacent tooth. To

avoid this, special protection made of thin copper sheets is used to prevent heating of the
adjacent tooth flanks. The conditions are presented in Fig. 15(ac) (Ref 27). In addition to
protecting adjacent tooth flanks, the shape of
the coil around the tooth has other effects. By
changing the shape and position of the coil
around a particular tooth, it is possible to achieve
equal hardened layer profiles, as in the case of
single-shot hardening. The thickness of the
hardened layer is at a maximum at the tooth
tip and then gradually decreases toward the tooth
root. In Fig. 15(c), the coil is placed slightly
lower, heating only the tooth flank along
the entire height from the root to the tip.
Figure 15(b) shows the lowest position of the
coil while still reaching below the tip of the
tooth. The coil positioned in this way does
not heat the tip of the tooth but only the surface
of the tooth flank from the tip to the root and
yields maximum thickness of the hardened layer
in the middle of the tooth. In this case, too, small
gears are heated with a high-frequency current,
Single-shot surface induction hardening of gear wheel tooth tips
pg 429
18/8/2008 3:59PM Plate # 0
Fig. 11
Fig. 12
Induction heating system for gear wheel heating by double frequency (medium/high). Source: Ref 27
Tooth-by-tooth induction surface heating process of

installing the induction coil around the gear wheel
tooth for heating, followed by shifting spray quenching
Fig. 13
Gap-by-gap gear wheel induction surface heating by

moving the induction coil for heating and moving or
installing spray quenching
pg 430
18/8/2008 3:59PM Plate # 0
whereas large-sized gears are first heated with a

medium-frequency current, and then heating is
performed with a high-frequency current (Ref 2,
16, 18, 19, 27, 32).
The third method of induction surface hardening is appropriate for large gear modules and
is known as tooth-gap hardening, which belongs
to the progressive hardening methods. In this
case, the coil is placed so that it ensures a uniform gap between the coil and the flanks of
two adjacent teeth. The tooth-gap hardening
method is very demanding and requires much
Fig. 14
Relative position of the massive induction coil to the

gear wheel tooth at induction heating. Source: Ref 27
Fig. 15
Influence of induction coil height on profile of induction heating surface layer in an individual tooth. Source: Ref 27
pg 431
18/8/2008 3:59PM Plate # 0
experience and knowledge to achieve the

desirable properties of the gear.
This method is also known as contour hardening. It is an ideal method for heat treatment of
gears, because it increases the hardness on the
tooth surface only slightly while decreasing the
load-bearing capacity in the root of the tooth.
Gears heat treated in this way exhibit very good
behavior in operation, because compressive residual stresses are present in the root of the tooth.
Gears with induction-hardened flanks, given that
the dimensioning is carefully carried out, can
achieve the highest fatigue strength. To verify
the results of induction surface hardening, it is
necessary to take certain measures for controlling the quality of the hardened layer. For this
purpose, hardness and microhardness measurements, supported by microstructural analysis,
are commonly used. A disadvantage of this procedure is that, due to the method of heating and
quenching (nonuniformly overheated left and
right tooth flank), slightly higher dimensional
deviations may be obtained than in the case of
simultaneous hardening of both flanks of the
same tooth (Fig. 15d) (Ref 27).
Metallurgical Aspect of Induction Surface
Heating. Prior to transformation hardening,
an operator should calculate the processing
parameters for the given power system. The procedure is as follows. Some of the processing
parameters are chosen, some calculated. The
choice is usually left to the operator and his
experience. Optimization is then based only on
the selection of power density and scan speed.
The correctly set parameters of transformation
hardening ensure the right heating rate, heating
to the right austenitizing temperature, TA3 , and a
sufficient austenitizing time, tA. Consequently,
with regard to the specified depth of the hardened layer, a temperature a little higher than the
transition temperature, TA3 , should be ensured.
Because of a very high heating rate, the equilibrium diagram of, for example, steel, is not
suitable; therefore, it is necessary to correct the
existing quench temperature with reference to
the heating rate. Thus, with higher heating rates,
a higher austenite transformation temperature
should be ensured in accordance with a timetemperature-austenitizing (TTA) diagram.
The diagram in Fig. 16(a) is such a TTA
diagram for 1053 steel in the quenched and
tempered state, whereas Fig. 16(b) is for the
same steel in the normalized state (Ref 38).
Because the steel shows a pearlitic-ferritic
microstructure, a sufficiently long time should
be ensured to permit austenitizing. In fast heating, austenitizing can be accomplished only

by heating the surface and subsurface to an
elevated temperature. For example, with a
heating time, t, of 1 s, for total homogenizing,
a maximum surface temperature, Ts, of 880 C
should be ensured in the first example and a
much higher surface temperature, 1050 C, in
the second example. This indicates that approximately 170 C higher surface temperature,
DTs, should be ensured in the second example
(normalized state) than in the first example
(quenched and tempered state).
Figure 17 shows a space TTA diagram
including numerous carbon steels with different
carbon contents. The TTA diagram gives particular emphasis to the characteristic steels, that
is, 1015, 1035, 1045, and 1070 steels, and their
variations of the transition temperature, TA3 ,
with reference to the given heating rate and the
corresponding heating time (Ref 38). Such a
temperature difference ensures, with regard to
the heating and cooling conditions of the specimen, the time required for austenite homogenizing, tA, in the given depth.
Figure 18 shows a shift of the transformation
temperature, which ensures the formation of
inhomogeneous and homogeneous austenite
within the selected interaction times (Ref 39).
A shorter interaction time will result in a
slightly higher transformation temperature, TA1 ,
and also a higher transformation temperature,
Fig. 16
Time-temperature-austenitizing diagram for steel

1053 in various states. Source: Ref 38
pg 432
18/8/2008 3:59PM Plate # 0
TA3 . To ensure the formation of homogeneous

austenite with shorter interaction times, considerably higher temperatures are required.
Fig. 17
Ref 38
Fig. 18
Influence of heating rate and carbon content

on austenitic transformation temperature. Source:
Figure 18(a) shows a temperature-time diagram

for austenitizing of Ck 45 steel.
The isohardnesses obtained at different interaction times in heating to the maximum
temperature ensure that partial or complete
homogenizing of austenite is plotted.
Figure 18(b) shows the same temperature-time
diagram for austenitizing of 100Cr6 hypereutectoid alloyed steel.
The diagram indicates that with short interaction times, which in laser hardening vary
between 0.1 and 1.0 s, homogeneous austenite
cannot be obtained; therefore, the microstructure consists of austenite and undissolved
carbides of alloying elements that produce a
relatively high hardness, even up to 920 HV0.2.
After common quenching of this alloyed steel
at a temperature of homogeneous austenite, a
considerably lower hardness, only 750 HV0.2,
but a relatively high content of retained austenite
were obtained. Retained austenite is unwanted,
since it will produce unfavorable residual
stresses and reduce wear resistance of such a
material.
Temperature-time-austenitizing diagrams with lines of resulting hardness for various steels. Source: Ref 39
pg 433
18/8/2008 3:59PM Plate # 0
The distribution of residual stresses in heat

treatment procedures where only the surface of
the workpiece is heated (induction hardening,
flame hardening) differs greatly from the procedures where heating is performed throughout
the entire volume (nitriding, cementation). In
nitriding and cementation, the aforementioned
second layer in the subsurface does not appear at
all, because the direction of the heat flow is
opposite to the direction of the heat flow in
induction and flame hardening. The resultant
operating tensile stresses on the surface or in the
surface layer can thus be considerably smaller.
Due to the surface hardness, induction and flame
hardening lowers the fatigue strength of machine components; therefore, care should be taken
to diminish all detrimental effects in the surface
layer.
A typical example of induction surface hardening is surface hardening of gears that are
heated with a low heating rate and relatively low
current frequency. The outer hardened zone includes almost the entire height of the gear teeth,
whereas the second zone is in the tooth root area.
A gear heat treated in this way will meet the
wear-resistance requirements expected of the
gear tooth, while the strength of the other part of
the tooth is of minor importance. The fatigue
strength in this case will be relatively low due to
high tensile residual stresses in the tooth root,
that is, in the second zone where the operating or
load tensions and the tensile residual stresses are
summed up. Figure 19 shows another example
of an induction-hardened gear where the frequency of the current was so high that the gear
tooth is heated along the flank surface and tip, as
is the case in cementation. The energy input in
heating a gear tooth or the whole gear was such
Fig. 19
that the second zone has not shown up. A similar

heat treatment can be applied to the spline inside
the gear. A gear heat treated in this way is more
resistant to wear and corrosion and should have
high resistance to fatigue in bending because of
a smaller thickness of the layer in the second
zone.
Many induction-hardened gears are treated in
the tooth gap, that is, in simultaneous heating
and subsequent quenching of two adjacent
flanks of the left and right tooth of the gear with a
hardened root area. In view of the variety of
methods to induction harden gears and the other
possible ways of gear hardening, it is unwise to
make a hasty decision in selecting the procedure.
It is necessary to make a thorough analysis,
including answers about the expected quality of
the hardened layer and analysis of the operation
loads of the machine components.
Highly loaded gears can be successfully
induction hardened if a high-frequency current
and high-input power are used. High frequency
is necessary in order to obtain a sufficient thickness of the hardened layer on the tooth flanks and
a fine hardened layer in the tooth root area. A
high-input power is necessary to increase the
heat gradient, which makes the size of the second zone smaller, resulting in a thinner second
layer with tensile residual stresses.
On gears with a small diameter, the inductionhardening equipment should be able to harden
the gear throughout its volume. Induction hardening of entire gears has some advantages, such
as rapid heating, no danger of decarbonization
and oxidation, high productivity, and repeatability in the gear quality. Thus, the hardened
layer is only defined by the first zone to the depth
that is greater than the height of the gear tooth.
Typical examples of induction surface hardening of (a) carbon steel and (b) alloyed steel gears produced from carbon steel (a)
and alloyed steel (b)
pg 434
18/8/2008 3:59PM Plate # 0
In induction hardening and quenching, lowering

or even the disappearance of compressive residual stresses is achieved in the tooth root area,
causing a considerable decrease in fatigue
strength and a higher propensity for fatigue.
In flame heating, lower temperature gradients
are reached than in induction heating. This
results in increased thickness of the layer in the
second zone. Possible harmful effects due to the
disappearance of the second zone of the hardening layer can be avoided by heating the entire
gear, which is possible to do in commonly used
furnaces.
When the induction coil has stopped heating,
an austenitic microstructure in the surface layer
should be obtained. Then, the cooling process
for the austenitic layer begins. To accomplish
martensite transformation, it is necessary to
ensure a critical cooling rate that depends on
the material composition. Figure 20 shows a
continuous cooling transformation diagram
for EN19B steel, including the cooling curves
(Ref 38).
Because carbon steels have different carbon
contents, their microstructures also show different contents of pearlite and ferrite. An increased
carbon content in steel decreases the temperature of the beginning of martensite transformation, TMS as well as of its finish, TMF
Fig. 20
Figure 21 shows the dependence between

carbon content and the two martensite transformations (Ref 38).
Consequently, an increase in carbon content
in steel results in the selection of a lower critical
cooling rate. In general, the microstructures
formed in the surface layer after transformation
hardening can be divided into three zones:

A zone with completely martensitic microstructure

A semi-martensitic zone or transition
microstructure
A quenched and tempered or annealed zone
with reference to the initial state of steel
Transformation hardening of steel starts

from its initial microstructure, which is ferriticpearlitic, pearlitic-ferritic. or pearlitic. In steel
heating, transformation into a homogeneous
austenitic microstructure should be ensured.
Figure 22 shows the dependence of the maximum surface temperature obtained in induction
heating with a machine part made of steel with
0.8% C and a pearlitic microstructure (Ref
40, 41). How the through-depth heating of the
machine part will proceed depends on the maximum surface temperature obtained and the
power density, which means that the throughdepth hardness profiles will differ.
Continuous cooling transformation diagram of EN19B steel. Source: Ref 38
pg 435
18/8/2008 4:00PM Plate # 0
With the same power density, this means that

a higher maximum surface temperature will be
accomplished with a longer heating time. Since
the heating times are usually short, austenitic
grains have little time left to grow. It can be

assumed that the grains remain fine and do not affect through-hardenability, that is, the throughdepth hardness profile. The diagram shows the
heating conditions that provide, in all cases, a
completely homogeneous austenitic microstructure to a certain depth and, consequently, constant hardness. Then follows a transition zone
consisting of homogeneous austenite and some
inhomogeneous austenite and pearlite. Consequently, hardness in the transition zone will
gradually decrease to that of the parent metal. It
will be approximately 240 HK. It is important
that the transition zone consists of a mixture
showing different ratios of the microstructures
concerned. The different microstructure ratios in
the transition zone, however, define the hardness
profile in this zone. Thus, at the maximum surface temperature, a maximum hardness of approximately 700 HK, without a constant part
with homogeneous martensite, is obtained, and
then it decreases immediately to the hardness of
the parent metal.
With a maximum temperature of 800 C,
there will be constant hardness to a depth of
0.6 mm; from that point to a depth of 0.85 mm,
the hardness will slowly decrease to that of the
parent metal, that is, 240 HK. The highest
hardness, 850 HK, is obtained with the maximum austenitizing temperature, 850 C, and is
found to a depth of 1.10 mm. From that point
to a depth of 2.70 mm, the hardness slowly
decreases to that of the parent metal. With higher
maximum surface temperatures, the lowest
Fig. 21
Influence of carbon content in steel according to start

and finish temperature of martensitic transformation
Fig. 22
Hardness profiles for an induction-hardened 0.8% C steel for various maximum temperatures. The initial microstructure of
the steel was pearlite. Source: Ref 40, 41
pg 436
18/8/2008 4:00PM Plate # 0
hardness is obtained at the surface, and then,

with a greater depth, it slowly decreases. With
steel having a pearlite-ferrite or ferrite-pearlite
microstructure, a microstructure consisting not
only of inhomogeneous martensite but also of
pearlite-ferrite grains will occur in the transition
zone. Such a microstructure results in a stronger
decrease in hardness than with the steel having a
pearlite microstructure. It is essential for efficient induction surface hardening that constant
hardness is obtained to a sufficient depth and that
the hardness profile of the transition zone is
adequate. It is only in this way that notch defects
in the hardened layer may be prevented and
better operation of the machine part under
dynamic load may be ensured.
Magnetic Flux Concentrators

In induction surface heating, unwanted areas
on the workpiece are often heated or even hardened. This type of problem appears when the
shape of the product surface is very complex,
and it is therefore difficult to adjust the coil for
local heating and quenching. For more demanding shapes, shields from materials with
good heat conductivity, for example, copper, are
often used on the product to prevent heating of
the workpiece. Heating the copper shield prevents undesirable heating of certain areas as well
as loss of heat. Such an example can be seen in
Fig. 15, which shows the protection of two adjacent heat flanks while heating the tooth in the
middle with an induction coil (Ref 15, 1921,
28, 42).
Heat losses occur due to local heating of the
workpiece surface that is not to be hardened;
however, this is done because it is not possible
Fig. 23
to adjust the shape of the induction coil.

The heating temperature is higher than the
hardening temperature, so that after quenching,
unwanted increased surface hardening is obtained. In cases of treatment of two local surface
areas on the workpiece that are in direct proximity, if heat is applied two times in sequence,
the heating of the second area may result in
tempering of the previously treated area. In these
cases, the area hardened first may have a tempered microhardness with a lower hardness.
Therefore, research has been done on how to
form and adapt the coil to offer a more concentrated magnetic flux. By adjusting the concentration of the magnetic flux, it is possible to
achieve localized heating of only those areas on
the workpieces that are to be hardened. In the
last decade, the development of induction coils
has been directed toward achieving localized
concentration of magnetic flux (Ref 19, 21). The
purpose of magnetic flux concentration is to
improve the efficiency of surface heating and
reduce heat losses. The use of a magnetic flux
concentrator enables selective local heating on
workpiece/product areas with complex geometry. Figure 23 shows a straight conductor with
current density distribution in points A and
B for three cases (Ref 19, 42):

Current density distribution in a straight

conductor (Fig. 23a)
Current density distribution in a straight
conductor when the conductive material
(workpiece) is approached (Fig. 23b). The
current density is greater in point A, closer
to the workpiece material, than in point B.
As a result, the workpiece is locally heated
over a longer lengthlonger than the width
of the conductor. This is referred to as current
redistribution due to the proximity effect.
Current distribution in an inductor without/with a magnetic flux concentrator and its effect on the heating profile of the
workpiece. Source: Ref 19, 42
pg 437
18/8/2008 4:00PM Plate # 0
By placing the conductor into the magnetic

flux concentrator made of powder metals
based on iron, nickel, cobalt, and other
powder materials, it is possible to achieve a
higher density of the magnetic field
(Fig. 23c).
The concentrated magnetic field results in

localized heating on only those areas that are
meant to be hardened. The magnetic flux concentrator surrounds the conductor, so that the
current density on the conductor surface is
redistributed, as is the heating of the workpiece
material.
The effects of the magnetic flux concentrator
depend on:

Workpiece material and shape of the local

area to which heat is applied
Workpiece material and concentrator shape
The current density in this case is the highest

in point A of the conductor and is considerably
higher than in point B. This results in effective
local heating, where the length of the heated area
is only slightly greater than the width of the conductor, but, due to the high density of the magnetic flux, a considerably greater depth of the
heated area is achieved.
As for concentrator material, different applications require the use of different materials.
The material for the concentrator must be chosen
after considering several factors:

Relative magnetic permeability

Magnetic reluctance
Flux density in saturation
Losses under magnetization
Resistance to high temperatures
Cooling abilities
Resistance to cooling effects of fire extinguishers
Good machinability
Adjustment to different shapes and sizes of
coils
Ease of assembly and disassembly
Manufacturing costs, depending on the kind
of material, induction heating parameters,
and geometry features between the concentrator and the workpiece
Figure 24 shows a local surface area with a

complex shape that is induction heated with a
single coil. Current density is in the inductor,
which results from the position of the inductor
with respect to the workpiece. This results in
differences in the current density, which are
reflected in the different heating/workpiece

depth profiles created by the differences in the
power density distribution (Ref 19, 42).
Figure 25 shows the same area and shape, but
heating was performed with a magnetic flux
concentrator (Ref 19, 42). This is placed on the
left and right sides of the surrounding coil, which
contributes to higher concentration of the magnetic flux to prevent heating of the sides of the
workpiece at this place. A redistribution of
power density took place, and a desirable local
heating profile of the workpiece was achieved.
Advantages offered by the magnetic flux
concentrator are:

Smaller consumption of power

Improved efficiency of heating
Better exploitation of equipment due to
shorter heating times
More selective heating of the workpiece
areas
Fig. 24
Heating profile on rotational workpiece with

induction coil. Source: Ref 42
Fig. 25
Heating profile on rotational workpiece with flux

concentrator. Source: Ref 42
pg 438
18/8/2008 4:00PM Plate # 0

Achievement of desirable heating profiles of

hardened layer profiles
High repeatability of the procedure in terms
of hardened profile size and hardened layer
microstructure
Efficient protection from unwanted heating
of adjacent areas and successful prevention
of reheating and softening the already hardened layers
Elimination of detrimental effects on the
operators health due to exposure to the
magnetic field
Improved operation lifetime of the heating
equipment and higher productivity
Less workpiece distortion and lower costs in
straightening and/or final grinding subsequent to hardening
Reduction of the quantity of rejected parts in
terms of required size and shape of the hardened layer and workpiece distortion and
cracking
Figure 26 shows two hardening procedures

for tooth profiles of gear wheels using appropriate induction coils with a ferritic core concentrating the magnetic field (Ref 19).
The first procedure is known as gap-by-gap
hardening (Fig. 26a), because the induction coil
with the ferritic core is moved in the tooth space
so that a suitable gap between a tooth profile and
the inductor is provided. The advantage of using
magnetic flux concentrators rather than the
conventional coil is that, with the same energy
input, heating time is shortened. Unfortunately,
the use of concentrators also shows some deficiencies, since the transition zone between the
hardened and nonhardened microstructures,
which is very important for hardening, is lost due
to rapid heating. This results in less favorable
residual-stress and microhardness profiles of the
hardened layer. Thus, a greater risk of failure
Fig. 26
Gap-by-gap and tooth-by-tooth induction hardening

of gear wheels. Source: Ref 19
is incurred at the point where teeth or other

machine parts are most strongly loaded, which is
at the transition from the hardened subsurface to
the nonhardened core, where an increased stress
concentration will occur. Such heating circumstances may be avoided by choosing lowerenergy inputs or power densities. In this case,
however, the case depth is more difficult to control. If a hardened-and-tempered microstructure
is to be provided at the tooth inside as well,
heating with two frequencies is applied. The
hardened-and-tempered microstructure may be
accomplished with only one frequency in heat
treatment of smaller gear wheels when the induction coil encompasses the entire gear wheel.
The second procedure is tooth-by-tooth
heating (Fig. 26b). In this case, the induction coil
is moved against a tooth so that an appropriate
gap is provided between the concentrator and the
surface of both flank profiles of the same tooth.
The difference between the two techniques is
that in gap-by-gap heating, two adjacent tooth
profiles and the root section of a gear wheel are
heated, whereas in tooth-by-tooth heating the
entire gear tooth is heated. Thus, in the first
example, only hardened gear wheel tooth profiles are obtained, whereas in the second example, hardened tooth profiles of the gear wheel
and quenched and tempered tooth inside are
obtained.
Figures 27(a to d) show various automobile
parts that were induction surface hardened using
the single-shot or scan-hardening technique (Ref
19). They were prepared for macroscopic and
microscopic examinations of the hardened layer.
The segments of the individual specimens of
various characteristic and exacting automobile
parts after grinding, polishing, and macroetching permit identification of the profile of the
hardened pattern, for example, in the crosssectional and longitudinal direction of the part,
respectively, as well as an analysis of the microstructure and microhardness. Each figure also
indicates which technique of heating, that is,
single-shot hardening or scan hardening, and
which heating conditions (P, f ), using singleturn or multiple-turn induction coils, were
employed.
The quality of the surface-hardening process
can often be efficiently assessed by measuring
the hardened-layer depth at different locations
on the individual parts. If the achieved depth is
very uniform, regardless of the location of measurement, then the part was not subjected to
distortion. A sufficiently high initial hardening
pg 439
18/8/2008 4:00PM Plate # 0
because the part must be cut in the longitudinal

and transverse directions respectively, so that
the information required on the hardened pattern
may be obtained. Such methods are appropriate
only for periodical control, particularly statistical control with periodical sampling. Based on
the results obtained for the characteristics of the
selected hardened pattern, the quality of an
article is confirmed or accepted and uninterrupted production provided. These types of nondestructive testing are long-standing; therefore,
nondestructive methods of testing the material
condition after induction hardening are gradually being introduced. Of all the methods, magnetic and magnetic-induction methods have
established themselves because they are very
fast, reliable, and provide reproducible results.
The nondestructive methods of testing these
parts allow on-line supervision of part quality,
since, due to the speed of testing, all parts may be
tested, which is a general tendency in mass
production today.
Conditions in Induction Heating and

Quenching of Machine Parts
Fig. 27
(a) Single-shot inductors used for both track (lobes)

and shaft of this automotive component. The part is
sectioned and acid etched to show the hardness pattern. All tracks
are hardened at the same time using 250 kW/30 kHz; the stem is
hardened using 135 kW/10 kHz. (b) Automotive right and left
cam shafts that have been selectively induction surface hardened.
The cam shaft was forged, heat treated, then final ground. No
premachining was necessary. The equipment used a dual-spindle
transfer mechanism; the coil was a four-turn coil that heated four
lobes per spindle at a time. Power was applied for each set of four
lobes: 150 kW, 10 kHz. The result is a 4.2 mm deep case depth
in the base circle of the cam. (c) Hardness patterns on carbon steel
crankshaft journals resulting from the stationary inductionhardening process. (d) (Left) An unacceptable nonuniform hardness pattern due to the nonuniformity of the workpiece, scanned
with a single-turn inductor. (Right) An acceptable hardness
profile achieved with a single-shot inductor. Source: Ref 19.
Courtesy of Inductoheat, Inc.
temperature provides an adequate through-depth

microhardness of the hardened profile. The
analysis is considered a destructive method,
Heating of workpieces is done so that a

magnetic field is created in the inductor, which is
connected to a high-frequency generator. When
a ferromagnetic material or workpiece is introduced into a magnetic field, eddy currents are
induced. The distribution of eddy currents in the
workpiece is specific, their density being highest
on the surface and decreasing toward the inside.
This phenomenon is known as the surface effect
or skin effect. Due to resistance offered by the
workpiece material, heating takes place mainly
in the thin surface layer, whereas the inner
core remains cold or is only slightly heated
(low-mass workpieces).
The depth of current penetration depends on
workpiece permeability, resistivity, and the
alternating-current frequency. Because the first
two factors vary comparatively little, the greatest variable is frequency. Depth of current
penetration decreases as frequency increases.
High-frequency current generally is used when
shallow heating is desired; intermediate and low
frequencies are used in applications requiring
deeper heating.
Most induction surface-hardening applications require comparatively high power densities and short heating cycles to restrict heating
pg 440
18/8/2008 4:00PM Plate # 0
to the surface area. The principal metallurgical

advantages that may be obtained by surface
hardening with induction are the same as for
flame hardening. The drop in magnetic permeability of steels depends on the temperature line,
TA2 , where steel transforms from magnetic into
nonmagnetic ferrite. The larger the effect of the
magnetic permeability change on the temperature line, TA2 , the smaller the carbon content in
the steel (the larger the proportion of ferrite in
the steel) and vice versa. Due to rapid heating,
phase transformation moves upward toward
higher temperatures. The temperature-time
curves of heating along the depth of a cylindrical
component depend on the kinetics of the magnetic transformation, TA2 , and the effects of other
phase transformations during induction heating.
A thickness of 1.0 to 1.5 mm is reached with a
medium-frequency current. The temperaturetime variation over the cross section of the steel
workpiece is a function of the following factors:

Penetration depth of eddy currents

Heat conduction of the material
Heating rate of the surface
Initial temperature of the surface
Size and shape of the workpiece
The depth of penetration of the heat is governed mainly by the power and frequency employed. The normal power density is 0.1 to 2 kW/
cm2 of the heated surface. The relationship
between depth of penetration and frequency can
be calculated approximately by using simplified
expressions, which are valid for the temperature
rise in steel up to the hardening temperature
(Ref 16):
20
dCS = p
f
cold state (20 C)
500
dHS = p
f
hot state (800 C)
where dCS is the depth of penetration in the cold

state, measured in millimeters; dHS is the depth
of penetration in the hot state, measured in
millimeters; and f is the frequency, measured in
hertz.
Due to heat conduction in the material during
heating, the overall depth of penetration is larger.
It is possible to calculate the additional penetration due to heat conduction from the expression:
p
dHC =0:2 t
where t is time, measured in seconds; and dHC is

the depth of penetration for heat conduction,
measured in millimeters.
The total depth of penetration is obviously
dT = dCS +dHC. It should be stressed that these
expressions give only a rough estimate of the
depth of penetration, and they have been included here only to show the fundamental effects of
frequency and time.
In flame heating, the temperature achieved on
the surface at equal energy input is considerably
higher than in induction heating, the overheating
and the hardened layer thickness being dependent on the heat conduction of the workpiece
material. Figure 28(a) shows the temperaturetime variation over the cross section of the
workpiece in flame heating (Ref 15). Characteristic of this variation is that the temperature
rapidly changes with time, and therefore, the
conditions for the formation of a homogeneous
austenitic microstructure are not fulfilled.
Figure 28(b) shows the temperaturetime
variation over the cross section of the workpiece
in induction heating. The temperature variation
is very similar to that in flame heating up to
magnetic transformation, that is, to line A2. At
temperatures higher than line A2, eddy currents
grow characteristically, and the rate of heating
decreases sharply. This slows down the heating
above temperature line A2. A reduced rate of
heating on the surface provides the conditions
for faster heating into the depth of the workpiece. This figure shows that a relatively thin
layer is heated up, but the layer has a rather
homogeneous austenitic microstructure. The
temperature-time variation on the workpiece
cross section, or the temperature field, depends
on the workpiece size and shape. Thus, in heavymass workpieces, faster heat abduction into the
remaining cold part of the workpiece is achieved,
and that is why the actual variation of temperature over the cross section is steeper. This means
that in heavy-mass workpieces, a higher surface
temperature than in low-mass workpieces must
be ensured to grant the same penetration depth.
The microstructural changes in induction
hardening depend to a large extent on the rate of
heating and subsequent cooling. The rates of
heating range from one to a few seconds, which
means that the diffusion processes may become
jeopardized. In steel, transformation of pearlite
into austenite takes place in induction heating at
almost the same temperature as in conventional
heating. In subeutectic steels suitable for surface
hardening, it is important that the induction
pg 441
18/8/2008 4:00PM Plate # 0
Fig. 28
Temperature profile across the workpiece diameter in (a) flame surface heating and (b) induction surface heating. Source:
Ref 15
heating procedure ensures enough time for the

diffusion of carbon for transformation of ferrite
into homogeneous austenite (Ref 2, 16, 27, 29,
30).
Figure 29 shows the temperature line of
through hardening of subeutectic steels with
different carbon contents versus different heating rates (Ref 15, 27). The graph shows that
austenitization is clearly influenced by the
heating rate, especially when carbon concentrations are low to medium. Likewise, in the
case of rapid heating for through hardening, a
considerably higher temperature is needed than
for normal hardening. Thus, in surface hardening carbon steels, it is difficult to ensure
enough homogeneous martensite, whereas in the
transition temperature range (TA1 to TA3 ), a
higher homogenization of martensite is achieved
but with a presence of ferrite. The proportion of
ferrite in the transition temperature range is thus
higher with faster surface heating and smaller
carbon content in the steel.
In surface hardening alloyed steels, a better
homogenization of the austenite is achieved in
the heating phase, and when quenching is completed, a very homogeneous martensite with a
uniform microhardness along the depth of the
hardened layer is derived. Unfortunately, there
is a transition temperature range with a smaller
ferrite content, which causes a sharp drop in
Fig. 29
Influence of induction surface heating rate on hardening temperature for subeutectic steels. Source:
Ref 15, 27
hardness in the transition area with a hardened

and nonhardened microstructure.
A very important heating requirement in the
hardening procedure, besides the austenitization
temperature, is the time necessary for austenitization, since both of these heating parameters
pg 442
18/8/2008 4:00PM Plate # 0
affect the size of austenitic grains. Austenitic

grain size is, on the other hand, dependent on the
martensite formed subsequent to quenching and
is reflected in the toughness of the surface layer.
That is why heat treatment conditions are sought
that would ensure the finest and most homogeneous austenitic microstructure in the heating
and overheating phases, leading to the formation
of very fine martensite with the highest possible
toughness of the hardened layer after quenching
(Ref 27, 29). This microstructural condition and
the achieved mechanical properties ensure good
wear resistance of machine components (gears)
and their good response under dynamic loads.
Figure 30 shows the growth of austenitic
grains for different rates of induction heating for
a steel with 0.4% C that can be used for induction hardening (Ref 15, 27). With increasing
rates of heating, the austenitic transformation
moves toward higher temperatures. A higher
heating temperature creates a higher formation
rate of austenitic crystal and therefore fine grains
of austenite. Subsequent to quenching, these fine
grains of austenite ensure a very fine martensitic
microstructure.
Figure 31 shows the relationship between
the hardened layer hardness and the heating
rate and temperature for a steel with 0.45% C
(Ref 15, 27).
For each steel, there exists a certain temperature range after hardening that yields the
best microstructural condition and thus the best
properties. With higher rates of heating, this
Fig. 30
Influence of surface heating rate on austenitic grain

size. Source: Ref 15, 27
range moves to higher temperatures. This means

that quenching from lower temperatures leads to
imperfect hardening, while higher temperatures
yield medium or rough needles of martensite.
The heating rate of 50 C/s is sufficient to reach
a hardening temperature range between 850 and
925 C. In the case of a higher-energy input that
heats up the thin surface layer with a rate higher
than 140 C/s, the required temperature range
becomes 870 to 970 C. In both cases, a surface
hardness of 60 HRC is reached subsequent to
quenching.
Figure 32(ae) shows the entire process of
induction hardening a cylindrical component
with a small diameter or cross section, which,
subsequent to hardening, leads to self-tempering
(Ref 15, 27).
Figure 32(a and b) show induction heating of
a thin surface layer to the austenitic temperature
range, ensuring, a sufficient thickness of the
austenite layer d2 subsequent to quenching. The
process of quenching or self-tempering is shown
in Fig. 32(c and d), where, due to heat conduction into the remaining cold part of the workpiece, the temperature on the workpiece surface
layer increases, and thus, the thickness of the
austenite layer increased to d3 or d4, respectively.
Due to a small workpiece cross section or low
workpiece mass causing heat conduction, the
temperature in the middle of the workpiece rose
to a point corresponding to the tempering temperature of the given material. Since the process
of tempering takes place in the workpiece with
the available heat needed for austenitization,
self-tempering of the workpiece is indicated
(Ref 27, 31, 32, 43).
Specific properties of the hardened layer can
be described by analyzing the microstructure
Fig. 31
Hardness reached after induction surface hardening

at various heating rates in steel with 0.45% C.
Source: Ref 15, 27
pg 443
18/8/2008 4:00PM Plate # 0
Fig. 32
Individual phases in induction heating and spray quenching in the workpiece surface layer and corresponding temperaturediameter diagrams. Source: Ref 15, 27
with hardness measurements on the surface or

microhardness measurements in the cross section of the hardened layer, and by measuring
residual stresses (Ref 27, 44, 45).
Figure 33 shows the hardness profile in the
hardened layer subsequent to induction surface
hardening (line 1) as a function of different
carbon content in the steel after conventional
hardening (line 2) (Ref 15, 27). Induction
surface-hardened layers normally have on the
average 3 HRC higher hardness on the surface
than that achieved in the same kind of steel after
conventional hardening. This is primarily due to
a finer martensite and compressive residual
stresses present in the induction surfacehardened layer (Ref 2, 27, 46).
Time-Temperature Dependence in
Induction Heating
The time variation of temperature in induction
heating of a thin surface layer depends on the
Fig. 33
Influence of carbon content on steel hardness after

various heat treatments. Source: Ref 15, 27
type and shape of the induction coil used, the

alternating-current frequency, f (Hz), the and
power density Q (W/cm2). Power density is
defined by the selected power of the highfrequency generator and the surface layer of
the workpiece. Surface heating depends on the
coupling between the induction coil and the
workpiece. Figure 34 shows the influence of
pg 444
18/8/2008 4:01PM Plate # 0
the selected power density and frequency on the

reference depth of the skin effect in a ferromagnetic material (Ref 2, 20).
A higher power density results in a greater
reference depth of the skin effect and a greater
depth of the heated layer with the same maximum temperature obtained at the workpiece
surface. Figure 35 shows the interdependence
between the heating parameters, that is, power
density and generator frequency, as a function of
the specified depth of the hardened layer and the
heating time required for single-shot techniques
of surface induction hardening (Ref 2, 20).
The data supplied by the diagram can have a
character of information only, but they make the
selection of an optimal surface induction condition easier. With the scanning technique of
surface induction hardening, however, the speed
of the workpiece movement, rather than time,
assures the depth of the required hardened layer
and should be defined. Generally, longer heating
times are required with smaller power densities
and vice versa. For the same depth of hardened
layer, longer heating times with lower current
frequencies are also required. With regard to the
depth of hardened layer selected between 0.5
and 10.0 mm, generator frequencies of 450, 10,
and 3 kHz can be selected in the single-shot

surface-hardening technique, in which case
appropriate power densities between 2 and
50 MW/m2 are obtained. With lower high frequencies, such as 10 kHz, the same depth of
surface-hardened layer, that is, 2.0 mm, can be
ensured only when the power density is changed
to 50 MW/m2. The lowest generator frequency,
3 MW/m2, shown in Fig. 35 cannot ensure the
depth of a hardened layer smaller than 2.5 mm.
Immediately after tempering, an intensive inverse heat flow is expected as well.
The power density and frequency used in
induction hardening are usually based on the
shape and size of the machine part to be surface
hardened; the case depth is specified also. In
addition to the depth of hardening, the case
pattern along the entire length of the machine
part is important.
Regardless of the complexity of a workpiece
shape, case depth and transition that are as uniform as possible as well as regular-shaped hardened ends should be provided. Inadequate
pattern transitions may produce high stress
concentrations related to the given loads in its
vicinity, so the material cannot resist dynamically loaded parts.
Fig. 34
Fig. 35
Reference depth of skin effect as a function of power

density and selected generator frequency for ferromagnetic steel. Source: Ref 2, 20
Influence of high-frequency generator on selection

of power density and heating time with given thickness of surface induction-hardened layer. Source: Ref 2, 20
pg 445
18/8/2008 4:01PM Plate # 0
Various heating conditions that are defined by

the power density and frequency provide different pattern depths. The third influencing
parameter is heating time. Thus, different
hardness and residual-stress profiles may be
achieved. The steel grade to be hardened, the
loop shape, and the gap size between the
inductor and the workpiece should be considered. The data found in the diagram apply
only to stationary hardening; therefore, with
scanning surface hardening, these values
should be suitably corrected. For progressive
hardening, the hardening conditions should
be slightly corrected to allow for the loop
movement.
With the surface heat treatment processes,
studies are often conducted on the influence of
the selected heating and quenching conditions
on the depth of the hardened layer and the size of
the transition zone between the hardened and
nonhardened microstructures. One simple and
practical procedure to control surface heat
treatment is to measure the time variation of
temperature from the beginning to the end of the
heating process and also from the beginning to
the end of the quenching process. The heat
process may be changed by changing the power
density and the generator frequency, whereas the
quenching process may be changed by the
selection of different quenching agents and
quenching processes.
Figure 36 shows the measured and calculated
temperature cycles for the surface, the core, and
in a radius, r, of 7 mm in a depth of 1 mm in a
cylindrical specimen (Ref 47).
A comparison of the temperature cycles
shows that in surface induction hardening, a
thermal flow of 3 MW/mm2 in heating and
5.8 MW/mm2 in quenching was selected. Under
Fig. 36
Comparison of variations of calculated and measured temperature cycles for a cylindrical specimen
16 mm in diameter. Source: Ref 47
such heating conditions, a maximum temperature of nearly 1000 C was attained, while
heating above a temperature of 800 C was
somewhat slowed down. The data in the diagram
show that the time required for heating the specimen from the ambient temperature to 800 C is
equal to the time required for heating from the
latter to the maximum temperature obtained at
the surface, that is, 1.6 s. A temperature cycle at
the surface takes 3.2 to 3.3 s. The temperature
differences between the surface and the core in a
given moment are the greatest during the heating
process, that is, DTmax 600 C. During the
quenching process, however, they can reach up
to 360 C. Temperature gradient changes are
much stronger in heating than in quenching. In
material heating, there is also a great difference
in yield stress of the material, which can produce
plastic deformation. Another very important
finding (Ref 47) is that the theoretical model
is appropriate, since the results obtained
were confirmed by the standard experimental
methods such as temperature measurement with
thermocouples, diamond pyramid hardness test,
and measurement of residual stresses with x-ray
diffraction. The difference between the measured and the calculated temperature cycles is
very small. It occurs mainly in heating and
reaches up to 60 C at maximum, not taking into
account the losses due to eddy currents.
Figure 37 shows the variation of temperature
from the surface toward the core with various
volume power densities, Q, that is, 0.4
109, 1.2 109, and 2.4 109 W/m3 (Ref 48).
Fig. 37
Calculated variation of temperature through specimen cross section in induction heating up to

hardening temperature with different volume power densities.
Source: Ref 48
pg 446
18/8/2008 4:01PM Plate # 0
With smaller volume power densities, the maximum temperature differences between the surface and the core also become smaller and the
austenitizing times longer. Because of higher
temperatures attained in the core, the inverse heat
flow may even be so high as to produce a change
in microstructure and, consequently, a reduction
of hardness in the surface layer.
The induction heating industry has standardized power supply frequencies, and probably
99% of the power used will be at the frequencies
listed in Table 2. Also included in Table 2 is the
type of equipment used to a change 50 Hz (60 in
some countries) to a higher frequency and the
conversion efficiency (Ref 23).
Melander (Ref 49, 50) first treated single-shot
surface induction hardening of low-alloy steel
with 0.4% C, 0.7% Mn, and 1.1% Cr for tempering and hardening as well as surface hardening. For an analysis, a representative-sized
machine part that is most often used, that is,
a cylindrical specimen 40 mm in diameter,
was chosen. Induction-heating conditions were
Table 2 Power sources, frequencies, efficiency,
and power for induction heating equipment
Type
Vacuum tube
oscillators
Motor generators
Frequency multipliers
Frequency inverters
Source: Ref 23
Fig. 38
Power (P),
kW
Efficiency (g),
%
Frequency
( f ), kHz
5600
5060
200450
7.5500
1001000
501500
7580
9095
8595
1, 3, 10
180 and 540
0.5, 1, 3, 10
selected so that the temperature of the diameter,

TA1 , was exceeded to a depth of 5.0 mm. This
means that a change of microstructure and hardness was expected even to the depth of 5.0 mm,
where only partial austenitization was obtained.
Induction-heating conditions were chosen that
subjected the surface layer to heating for up to
35 s. Time variations of temperature at the surface of the cylindrical specimen and in its subsurface at depths of 2.0, 4.0, and 10.0 mm are
shown in Fig. 38 (Ref 49, 50). Time-temperature
diagrams differ from the previous ones, since a
distinctive deviation occurs in heating the specimen material when the temperature of magnetic
domain, TA2 , is reached and exceeded.
The course of surface heating indicates that
the transformation temperature TA2 was obtained in 10 s. In spite of the same power density, further heating of the surface up to a
temperature of 850 C was very slow due to the
nonmagnetic character of the steel surface layer;
it took another 28 s. It is difficult to assess which
models of induction heating are more suitable
than others. However, with the heating process
suggested, it is possible, after quenching, to
ensure a homogeneous, fine austenitic microstructure exhibiting the finest martensite with
the highest possible hardness of the given steel.
Due to the presence of alloying elements and
high cooling rates of the surface layer, in addition to fine martensite, up to 3% residual austenite also appears. Based on the time-temperature
variation of heating, the depth of the hardened
layer ranges between 2.0 and 4.0 mm. From the
Time-temperature cycle during single-shot surface induction heating and quenching. Source: Ref 49, 50
pg 447
18/8/2008 4:01PM Plate # 0
time-temperature variation at a depth of 10 mm,

it can be assumed that the specimen was heated
through the entire volume, that is, to the very
core of the specimen. Because of a different time
variation of temperature in the fourth temperature cycle, it may be concluded that the maximum temperature obtained in the depth of
10 mm is lower than the magnetic transformation temperature of the given steel. Because of
strong overheating of the cylindrical specimen
toward its center, lower temperature gradients
occur, which result in a reduction of thermal
stresses during the heating process. According to
a comparatively high temperature in the core,
temperature gradients between the surface and
the core are generally lower, which produce a
decrease in axial internal stresses generated
during quenching and also a decrease in axial
residual stresses.
Figure 39 shows the time variations of
through-thickness temperature in single-shot
induction heating with high-frequency generator
powers of 40, 60, 100, and 180 kW and a
medium-frequency current (Ref 51). Characteristic temperature transformations, TA1 and
Fig. 39
TA3 , for equilibrium heating are plotted (Ref 51).

Induction heating is a very fast process; therefore, the temperature transformations shift to
higher temperatures. To obtain homogeneous
austenite in the surface layer, it is necessary to
heat the surface layer to the hatch-marked temperature range. The results shown in the four
diagrams make it possible to draw the following
conclusions:

The heating curves differ strongly.

Only with the power of 40 kW does the temperature not reach the hatch-marked temperature range in 30 s.
With all other powers higher than 60 kW,
the hatch-marked temperature range is
reached in a shorter heating time.
With all powers higher than 60 kW, more
intensive heating of the surface layer occurs,
which is shown in a steeper hardness curve in
the hatch-marked temperature range.
The first question raised by an engineer would

be how to evaluate the through-thickness variations of hardness and residual stresses with
reference to the time variation of temperature.
Time-temperature variations in single-shot heating at various powers. Source: Ref 51
pg 448
18/8/2008 4:01PM Plate # 0
The second question would be how to select

induction-heating parameters to obtain optimal
properties of the surface-hardened layer with
a minimal energy input. In answer, the initial
steel microstructure must be taken into account
and an appropriate quenchant selected to ensure a cooling rate in the surface layer equal to
or a bit higher than the critical cooling rate
required.
Figure 40 shows the time and through-thickness temperature variations of the specimen on
heating with different powers and different
velocities of workpiece movement, that is,
v1 = 140 mm/min (Fig. 40a), v2 = 220 mm/
min (Fig. 40b), v3 = 370 mm/min (Fig. 40c),
and v4 = 680 mm/min (Fig. 40d), in scan hardening (Ref 51). In the first case, with v1, the
hardened layer is obtained with the powers of 74
and 56 kW. In the second case, with a higher
velocity of movement, v2, the hatch-marked
temperature range is reached only in heating
with the powers of 78 and 59 kW. In the third
case, with v3, the hatch-marked temperature
range is reached only in heating with the power
of 112 kW, whereas in the fourth case, with
the highest velocity, v4, this temperature range is
Fig. 40
not reached in spite of a very high power,

151 kW.
Quenching Systems for Induction

Hardening
Control of the quenching process of a gear
wheel from high hardening temperatures
ensures martensite microstructure. The cooling
rate should be high enough to prevent formation
of unwanted softer microstructures, such as
pearlitic or bainitic ones; therefore, it is very
important that in the development of new systems of induction hardening, quenching systems
are designed properly.
Process parameters should be accurately
controlled to ensure permanent and reproducible
results after hardening of machine parts. It is also
indispensable to define individual parameters
and determine their permissible deviation in
operation, presuming that the deviations have a
negligible influence on the results of heat treatment. The quenching system for induction hardening is defined by eight parameters (Ref 17,
Time-temperature variations in scan hardening at various scan speeds. Source: Ref 51
pg 449
18/8/2008 4:01PM Plate # 0
pg 450
52): heat time/scan rate, power level, power

frequency, part position/rotation, quenchant
concentration, quenchant flow, quenchant temperature, and quenching time.
Important simultaneous changes of one or
more of these parameters can produce unwanted
effects on the workpiece, which, in extreme
cases, result in an unsuitable microstructure, deviations in the depth of the hardened profile,
unsuitable hardness variation (too low hardness,
soft spots), and exceeding the distortion of the
machine element.
In practice, there are machine parts of different shapes and sizes, requiring different depth of
the hardened layer. In these cases, the type of
material chosen and its through hardenability
should be considered.
Thus, with alloyed steels having good through
hardenability and/or machine parts with a comparatively thin hardened surface layer, the martensite microstructure can be obtained without
the application of a quenchant. In such cases,
heat sinks from the surface into the cold workpiece core, so that the critical self-cooling rate
obtained at the surface is higher than the critical
cooling rate (Ref 24, 53).
Figure 41(a) shows the time variation of the
temperature measured at the individual measuring points of the gear-wheel tooth during
quenching with a pressurized water jet (Ref 52).
Data on the temperature variation show that
the cooling temperatures were very similar at
measuring points 1 and 2, somewhat lower at
measuring point 3, and the lowest at measuring
point 4, where, after 40 s of quenching, it still
equaled approximately 350 C.
Figure 41(b) shows the calculated time variation of temperature at the same measuring
points of the gear-wheel tooth during quenching,
taking into account selected heat-transfer coefficients, a0, of 5000 and 10,000 kcal/m2 Ch
(Ref 52). The calculated time variations of
temperature differ from the measured ones only
at those measuring points with higher cooling
rates. The time variation of temperature refers
only to the selected measuring points; therefore,
assessment of the stress state during the quenching process and the distortion of gear-wheel
teeth or a tooth is very difficult.
The authors focused on an analysis of the conditions during the quenching process by means
of a calculated distribution of temperature,
shown as isotherms (a0 = 5000 kcal/m2 Ch)
at the half-cut of the gear-wheel tooth. Figure 42
shows the variation of isotherms during the
quenching process after 0.1, 1.0, and finally 10 s

of cooling (Ref 52).
In quenching, it is very important that the
temperature gradients are high enough to prevent plastification, that is, distortion, of the gear
wheel. Such quenching conditions, including
high-temperature gradients, result in a decrease
Fig. 41
Temperature profiles of tooth gear at given measuring points during cooling process. (a) Measured. (b)
Calculated. Source: Ref 52
Fig. 42
Temperature distributions
process. Source: Ref 52
during
the
cooling
18/8/2008 4:01PM Plate # 0
in favorable compressive residual stresses after

hardening. It follows that a sufficiently high
cooling rate must be ensured at the tooth surface
as well as at the required depth of hardening. That
means that temperature gradients in the gearwheel tooth must be attained so that gear-wheel
teeth are hardened with the minimum required
cooling rate, ensuring the required martensite
transformation. A sequence of images showing
isotherms in very short time intervals from the
beginning of quenching allows assessment of the
size of distortion of the tooth and residual stresses
to be expected after quenching.
Figure 43(a) shows the measured volumefraction distribution of martensite and Fig. 43(b)
the calculated volume-fraction distribution of
martensite when different values of the heattransfer coefficient, a, that is, 5000, 7500, and
10,000 kcal/m2 Ch, are taken into account (Ref
52). The main difficulty encountered in the calculation is how to determine the heat-transfer
coefficient to obtain a description of real quenching conditions. Quenching is very intensive,
since it is carried out under a pressurized waterjet; the gear wheel has been induction heated,
whereas the core is cold.
Two common quenching methods that use a
quenchant are spray quenching and immersion
quenching. Particularly popular spray quenching
techniques that offer different possibilities are:

Spray with progressive scan heating (scan

hardening)
Spray after heating in position (single-shot
hardening)
Fig. 43
Spray quenching out of location after

heating
Spray quenching is carried out immediately

after heating with a short pause. It is used with
machine parts made of steels with good hardenability and with a comparatively thin hardened layer. Immersion quenching is carried out
after the inductor has been disconnected from
the high-frequency generator. Gradual hardening is very efficient and prevents the influence
of inverse heat from the core toward the surface,
so that no reverse heat flow and no heating of the
already cooled surface occur. This problem is
characteristic of large parts, where the workpiece surfaces should be hardened selectively.
In the past, straight oils and water-soluble oils
were used for quenching after surface induction
heating. Straight oils and water-soluble oils
produce mild quenching effects, which reduce
difficulties due to distortion and/or crack initiation.
Straight oils require immersion quenching
to minimize the risk of oil ignition, whereas
water-soluble oils are more suitable for spray
quenching. Other quenchants used are water,
polymeric water solutions of different concentration, water-soluble oils, saltwater, and
so on.
Polymeric water solutions are inflammable
quenchants. They are prepared in various concentrations to be suitably adapted to various
cooling rates. For spray quenching, various flow
rates of quenchants can be selected. They
depend on the size of inlet openings and the
Martensite distribution of the hardened tooth gear. Source: Ref 52
pg 451
18/8/2008 4:01PM Plate # 0
number of outlet nozzle boreholes in one, two, or

more rows at the quenching ring. The fluid flow
rate can also be regulated by pressure. Technical
literature provides data on shaping of quenching
nozzles as well as on the selection of the quenchant flow rate required. It is very important that
a sufficient flow rate of the quenchant is chosen
to ensure sufficient heat removal from the workpiece surface layer during quenching.
The flow rate is also controlled by the number
of boreholes for the flow and spraying of the
quenchant, respectively. An appropriate arrangement of the boreholes in the ring permits a
uniform heat removal after spray and proper
and uniform cooling. The borehole cross sections should be only 15% of the total available
ring surface and of the inductor in the singleshot-type application, respectively. Rotation
velocities of the workpieces in surface induction
heating are relatively high and can be selected
between 800 and 1000 rotations per minute,
whereas the rotation velocity of the workpieces
during quenching is considerably lower, that is,
40 to 60 rotations per minute. Very high rotation
velocities of the workpieces are required when a
uniform and reproducible depth of hardened
layer is to be ensured, and particularly with
relatively short heating times.
Figure 44 shows the dependence of cooling
rate changes on the momentary temperature at
the workpiece surface for four different concentrations of polymeric water solutions (Ref
24). The selection of an appropriate concentration of the polymeric water solution for the
selected steel and the required depth of hardened
layer should ensure minimum distortion of the
workpiece, which is the final purpose of any heat
treatment. Thus, the required heat removal from
the heated surface layer of the workpiece, considering the thickness of the heated layer, should
be ensured by an appropriate quenching system.
control. An appropriate machining technology

and appropriate heat treatment processes should
be selected, and internal stresses lower than the
yield stress should be ensured at any moment
and any location of a machine part during heating and/or cooling. Analytical methods provide
an insight into heat treatment conditions if the
time variation of internal stresses is monitored
and the dependence between the cooling time
and the specimen temperature at each point is
known. With regard to the specimen temperature determined in this way, the specimen yield
stress at each point can be determined as well.
Consequently, with different heat treatment conditions, different values of physical quantities
can be selected. They are reflected in the changed conditions in the material, which make it
possible to study distortion of the machine part
during cooling and to determine the magnitude
of the residual stresses.
Figure 45 shows the time variation of axial
stresses at the surface and in individual depths,
that is, from 2.0 to 4.0 mm under the surface
and in the middle of a cylindrical specimen
(Ref 50).
In surface heating, tensile stresses occurred
at the surface due to thermal extension of the
Time Variation of Stresses and

Residual Stresses
With a certain heat treatment performed, the
required microstructural changes and an appropriate magnitude and variation of hardness are
obtained. It is also required that distortion of a
machine part be as small as possible so that the
final size of a machine part can be obtained with
minimum precision machining. Consequently, it
is very important that distortion be kept under
Fig. 44
Polymer additive ratio effects on the cooling rate.

Source: Ref 24
pg 452
18/8/2008 4:01PM Plate # 0
Fig. 45
Axial stress distribution at various depths below the surface during single-shot induction surface hardening. Source: Ref 50
surface layer, but when the transformation

temperature of the pearlitic- to austenitic microstructure was exceeded, additional compressive
stresses occurred at the surface.
The compressive axial stresses transform
into tensile stresses in the zone of undercooled
austenite. In the transformation of undercooled
austenite into martensite, the compressive
stresses increase with the increase in martensite
fraction. In the core, the opposite sign of the
stress was obtained, that is, the tensile stress. At
the end of quenching, the compressive residual
stresses obtained in the surface-hardened layer
were approximately 1600 N/mm2, and the
tensile stresses in the core were approximately
+870 N/mm2.
Internal stresses are induced in heat treatment
by temperature and microstructural changes.
Residual stresses in the induction surfacehardened layer are always of a compressive
nature, are relatively high, and have a good
effect on dynamically loaded components. The
existence of residual stresses in the radial
direction, that is, into the depth of the hardened
layer, is very important, as is the absolute value
of residual stress on the surface and the stress
profile in the transition from compressive into
tensile stresses (Fig. 46) (Ref 15).
In the case of induction hardening, a maximum compressive residual stress in the surface
layer is achieved, which is very desirable for
dynamically loaded components. The transition
from compressive into tensile stresses should be
Fig. 46
Residual stress profile below the surface after

induction surface hardening. Source: Ref 15
as gentle as possible, lessening the effect of

stress concentration in loaded components. This
contributes to the fact that a machine component
is less susceptible to overloads in operation. It
has been shown that residual stresses are closely
linked to hardness variation and microstructure
in the transition zone of the hardened layer, that
is, in the narrow range between the hardened and
nonhardened microstructure.
Some examples of the dependence between
microhardness and residual stresses are shown in
pg 453
18/8/2008 4:01PM Plate # 0
Fig. 47 (Ref 15). Figure 47(a) shows a very steep

microhardness profile in the transition zone and
the highest compressive stress in the surface,
and, related to this, a very steep transition of
residual stresses into the tensile range. The
change from compressive into residual stresses
happens at the transition between the hardened
and nonhardened areas. Figures 47(b and c)
show graphs of broader transition zones, which
produce a change in the microhardness and also
the residual stresses (Ref 15). Thus, Fig. 47(c)
shows a very slight drop in hardness in the
transition zone for the same microhardness on
the surface.
This is reflected in lower compressive stresses
in the surface, accompanied by a slight change
of stresses in the transition zone into lower
tensile residual stresses. In induction surface
hardening, the engineer should choose the kind
of heat treatment conditions that will result in
the microhardness and residual-stress profiles
shown in the last two examples in Fig. 47.
Investigations of residual stresses after induction surface hardening have confirmed that when
the hardened layer is 2 mm thick, the change of
compressive into tensile stresses happens in
compliance with the transition zone, that is,
the achieved depth of the hardened layer. When
the thickness of the hardened layer is greater
than 2 mm, the transition from compressive
into tensile stresses happens in the hardened
zone, that is, the martensite microstructure. This
means that induction surface hardening is much
Fig. 47
more difficult if the thickness of the layer is

above 2 mm. Due to intensive cooling, the
internal stresses in the surface layer may become
so high that they cause failure of the component.
This failure is very typical, because the surface layer separates from the core due to high
radial stresses. Important features on the curve
are:

Maximum value of compression residual

stress in the hardened surface layer
Maximum value of tension residual stress in
the hardened surface layer
Transition width of compression to tension
of residual stress in the hardened surface
layer
Transition steepness of compression to tension of residual stress profile
Layer depth with transition microstructure
The examples shown in Fig. 46 and 47 were

based on a presumption that the maximum
hardness and the maximum value of the residual
stresses obtained were independent from the
heating power density and the heating time
(Ref 15). In the case of induction surface hardening, a maximum compressive residual stress
in the surface layer is achieved. The transition
from compressive into tensile residual stresses
should be as gentle as possible, lessening the
effect of stress concentration in loaded components. This contributes to the fact that a machined component is less susceptible to overloads
in operation. It has been shown that residual
Various residual-stress and hardness profiles below the surface. Source: Ref 15
pg 454
18/8/2008 4:01PM Plate # 0
stresses are closely linked to hardness variation

and microstructure in the transition zone of the
hardened layer and base material, that is, in the
narrow range between the hardened and nonhardened microstructure.
On the basis of the heating temperature cycle,
microstructural changes in the surface layer
after quenching can be predicted. Consequently,
the temperature cycles at the surface, that is, a
through-thickness heated layer, have a decisive
influence on the variation of residual stresses
after quenching. The variation of the temperature cycles in heating can be adjusted by the
selection of adequate energy input with reference to the steel microstructure and its thermal
properties.
Figure 48 shows three examples of induction
surface hardening, each with different energy
input during heating, whereas the quenching
process was the same in all cases (Ref 15, 54).
The variation of residual stresses after induction
surface hardening is affected by the energy input
through the workpiece surface and thermal
conductivity of the workpiece material. The
energy input in the workpiece depends on the
power chosen, the duration of high-frequency
current, the shape of an induction loop, the size
of a gap between the induction loop and the
workpiece, and the area exposed to the energy
input.
In progressive hardening, the velocity of
progression of the high-frequency loop along the
heated area must be considered.
The transition from tensile to compressive
residual stresses presents a serious danger for
catastrophic failure. From the aforementioned,
it can be understood that the workpiece must
be slowly heated and also quenched with the
correct cooling rate. The actual cooling rate is
very important throughout the martensitic transformation and must be as close as possible to
the critical cooling rate. The quenching process
must be carried out very carefully in the martensitic transformation temperature range to
ensure internal stresses lower than the yield
point. The only barely-seen difference is in the
transition area, when the compressive residual
stress starts to fall rapidly. In the transition area,
the residual-stress curve is steeper and less
desirable as far as notch effects under dynamic
loads are concerned. In general, efforts should
be made to achieve low or moderate residual
stresses in the transition area after induction
surface hardening. Manufacturing engineers are
often confronted with the question of how to
obtain this kind of residual-stress profile. It

should be emphasized that because of their
lower heat conductivity, alloyed steels for surface hardening are very difficult to treat, and it is
difficult to ensure desirable residual stresses in
the transition area. This is why special alloyed
steels are available that display a more favorable
behavior in heating and quenching. Because of
these difficulties, delamination of the surface to
the very core may occur. Figure 48(a) shows that
that the energy input was too weak, since only a
partial transformation of the steel matrix into
austenite occurred (Ref 15, 54). After quenching, this will produce only a small portion of
martensitic microstructure, which is followed by
a measured variation of residual stresses. The
second example, shown in Fig. 48(b), indicates
that heating with a high-power density was very
intensive. The result of a too-high power density
is a distinct influence of heat conduction from
Fig. 48
Residual-stress profiles after induction surface hardening at various input energies. Source: Ref 15, 54
pg 455
18/8/2008 4:01PM Plate # 0
the surface to the subsurface, which, in turn,

resulted in a higher temperature measured in the
subsurface than at the surface after the delayed
interruption of heating at the beginning of the
quenching process. Because of such a heating
regime and energy input at the lower limit of
the required energy, considerably lower compressive residual stresses were achieved at the
surface. In this case, the highest value of compressive residual stress was obtained at a depth
of 1.0 mm. The variation of the stress was not
the most adequate, since induction surface hardening can provide considerably higher compressive residual stresses in the surface layer.
The anticipated variation of the residual stresses,
that is, very high compressive residual stresses
ranging between 800 and 1000 N/mm2, is
shown in Fig. 48(c) (Ref 15, 54).
A very favorable variation of residual stresses
can be found next to the transition zone between
the quenched and unquenched surface layers,
which ensures favorable behavior of a machine
part under dynamic loads.
In practice, the progress of heating and
quenching can be monitored by physical modelling, in which case the temperature cycle at the
workpiece surface as well as the temperature
cycles in the individual depths are obtained. The
variations of the temperature cycles in heating
and the known temperatures, that is, temperature
ranges, make it possible to determine changes
of a pearlite-ferrite microstructure into an
austenitic one. The latter makes it possible to
predict the microstructure after surface hardening.
Fig. 49
The progress of induction surface heating can

be supervised by measuring the temperature at
the workpiece surface using a pyrometer. The
temperature measurement is a determination of
temperature cycles in individual depths during
induction heating and subsequent cooling. There
are three basic and important metallurgical
conditions that will affect a successful induction
surface-hardening operation:
Lower critical temperature, where metallurgical phase transformation begins
Upper critical temperature, where the formation of austenitic grain microstructure is
complete
Surface temperature, which affects the
resultant grain size of the heat treated microstructure
For a typical AISI 140 material, the steel
reference books will list TA1 =750 C,
TA3 =795 C, and Tsmax =900 C. However,
because of the shorter heating times with
induction heating, the actual values required for
a satisfactory microstructure and process may be
30 to 60 C higher than for conventional furnace
heat treatment processes.
Figure 49 shows the temperature cycles in the
surface and in the individual depths during induction heating and quenching (Ref 15, 54).
The temperature range of the pearlite-ferrite
transformation into austenite, which is a condition for the formation of the martensitic microstructure after quenching, is plotted also.
On the basis of the heating temperature cycle,
microstructural changes in the surface layer after
Temperature cycles during single-shot induction surface hardening, at various depths. Source: Ref 15, 54
pg 456
18/8/2008 4:01PM Plate # 0
quenching can be predicted. Consequently, the

temperature cycles at the surface, that is, a
through-thickness heated layer, have a decisive
influence on the variation of residual stresses
after quenching. The variation of the temperature cycles in heating can be adjusted by the
selection of adequate power density with reference to the steel microstructure and its thermal
properties.
Figure 50 shows graphic representations of
temperature cycles at the surface at various
power densities in single-shot induction hardening (Ref 15, 54). Each temperature cycle
includes induction heating, which is followed by
quenching a short while after the interruption
of heating. Four temperature cycles of induction
heating and quenching measured at different
heating power densities using high-frequency
current were plotted. With the same inductor
having the same gap width between the inductor
and the workpiece surface, but with different
power densities, different temperature cycles of
heating were obtained. With a higher power
density, the heating curve was steeper and the
heating period shorter (Ref 54).
In all cases of heating, however, the same
depth of the hardened layer was ensured. With
the given power density, however, it is more
difficult to ensure the same depth of the hardened layer, since the heating times are considerably reduced. With very short heating
Fig. 50
times, it is much more difficult to ensure the

same depth of the hardened layer, which indicates a worse repeatability of the results. The
shorter heating times with the same cooling
times contribute to a shorter hardening cycle and
result in higher productivity. This higher productivity, however, entails a high consumption
of electric energy for heating and less favorable
through-thickness microhardness and residualstress variations in the hardened layer. Figure 51
shows microhardness and residual-stress variations with the highest power density (Q1) and
different heating times, tH1 . . . tH6 (Ref 15, 54).
The influence of heating time affects the
microhardness and residual-stress variations.
With a correctly chosen condition of induction
heating of the thin surface layer, adequate values
of the maximum compressive residual stresses
are obtained at the surface. The variation of
residual stresses across the hardened surface
layer, that is, in the transition zone between
the hardened and the unhardened material, is
more important. The variation should not be
steep, which means that the gradient of residual
stresses in the transition zone between the hardened and unhardened material should be very
gently sloping. The microhardness variation in
the transition zone, with a gradient of residual
stresses as gently sloping as possible, can be
accomplished by selecting the given power
density and an appropriate heating time, since
Temperature cycles at the surface in induction surface heating and quenching at various power densities. Source: Ref 15, 54
pg 457
18/8/2008 4:01PM Plate # 0
full use of compressive residual stresses in dynamically loaded parts can be made in this way.
Figure 52 shows a cylindrical workpiece or
sample with initial diameter DI and initial height
HI. Its purpose is to show the process of residual
stresses after induction surface treating in the
central part of the workpiece (Ref 15, 27). With
energy input to the thin surface layer, the initial
diameter increases to the value of DA. Change of
diameter is associated with thermal expansion of
the material and is due to ferrite and/or pearlite
transformed into austenite.
After quenching, a thin surface layer with a
martensite microstructure is obtained that has a
different specific volume than the initial microstructure. The hardened surface layer is ready to
receive a greater diameter (DH), which is resisted by the initial microstructure. Of interest are
the residual stresses in the radial and axial
directions. Residual stresses are compressive in
the hardened layer, then they change into tensile.
Residual stresses in the axial direction are
compressive in the middle part of the hardened
layer and turn into tensile below the hardened
layer. In this way, very high tensile residual
stresses are achieved at the bottom of the
Fig. 51
hardened layer. Compressive residual stresses

prevail in the remaining part of the nonhardened
surface layer. A residual-stress profile is unfavorable in the transition zone between the
hardened layer and the rest of the nonhardened
part of the workpiece.
The transition from tensile to compressive
residual stresses presents a serious danger for
catastrophic failure. From the aforementioned
it can be understood that the workpiece must
be slowly heated and also quenched with the
correct cooling rate. The actual cooling rate is
very important throughout the martensitic transformation and must be as close as possible to the
critical cooling rate. The quenching process
must be carried out very carefully in the martensitic transformation temperature range to
ensure internal stresses lower than the yield
point.
Figure 53 presents a crankshaft formed by hot
forging, and Fig. 54 shows the manufacturing
procedure from blank to crankshaft (Ref 15, 36).
The procedure of forming should be carefully
prescribed, including the initial and final temperatures of forging and the uniform plastic
deformation rate for the entire volume.
Microhardness and residual-stress profiles at various heating times, tH1tH4. Source: Ref 15, 54
pg 458
18/8/2008 4:01PM Plate # 0
Fig. 52
Radial and axial residual-stress profiles after induction hardening the surface layer in the central part of a cylinder steel rod.
Source: Ref 15, 27
This will ensure a fine-grained, banded

microstructure contributing to a favorable relationship between strength and toughness of the
product. The forging procedure is followed by
annealing to remove the residual stresses incurred by the refinement procedure applied to
forgings. Deformation of the crankshaft occurs
after heat treatment, as early as in the phase of
heating to the austenitization temperature. Due
to stress relief as well as the effects of nonuniform cooling of the product, it is necessary for
the forgings to be subjected to straightening
prior to mechanical treatment. If necessary additional annealing can be prescribed to remove
internal stresses induced by straightening. This
is followed by turning and rough grinding to
approach the final dimensions of the product.
The technology of manufacturing the crankshaft
involves careful selection of the conditions of
turning and subsequent rough grinding to avoid
the occurrence of internal stresses that would
remain in the material even after induction surface hardening and grinding, thus reducing the
fatigue strength of the material. Induction hardening may be preceded by stress annealing if the
depth of the surface hardening is smaller than the
depth of the damaged layer, since, in this way, it
is possible to change the unfavorable stress state
in the surface layer induced by machining. In
this case, the depth of the induction surface
Fig. 53
Schematic presentation of a crankshaft with marked

main bearing locations. Source: Ref 15, 36
hardening was greater than the depth of the

damaged surface layer; therefore, machining
could immediately be followed by induction
surface hardening. After induction surface
hardening, the size and distribution of residual
stresses contribute to toughness and fatigue
strength of the material. In these tests, induction
surface hardening was followed by finish grinding and nondestructive magnetic inspection of
the surface to reveal the possible existence
of cracks on the product surface.
Crankshafts were taken from production after
induction surface hardening with the heat treatment and machining conditions as specified in
the technology sheet. The residual stresses on
the main crankshaft bearings were measured on
the bearing location in the middle (sample A in
Fig. 53), on the extreme left side (sample C) and
on the extreme right side (sample G).
pg 459
18/8/2008 4:02PM Plate # 0
Fig. 54
Machining and heat treatment procedure from blank to crankshaft. Source: Ref 15, 36
Figure 55 shows residual-stress distribution

after induction surface hardening in the central
bearing location (sample A) and on the extreme
left side (sample C) (Ref 15, 20, 44). For both
locations, residual stresses were measured on
two samples. The distribution of residual stresses on location A is very similar on both samples, as expected, with the highest compressive
stress ranging between 1020 and 1060 N/mm2 at
a depth of approximately 250 mm and then
slowly dropping to a depth of 3.5 mm.
The residual-stress distribution after induction surface hardening on bearing location C is
very similar to that in the central bearing location A, except that its absolute values are slightly
lower, and a distinct decrease in the residual
stresses can be noted as early as a depth of approximately 3 mm, reaching its minimum value
at a depth of approximately 5.0 mm.
The residual-stress distribution is just as
favorable as in the central location, except that
its absolute values are slightly lower. The difference in the residual-stress distribution can
be related to the period of overheating in the
austenitization temperature, which resulted in a
thinner layer in austenitization and also a thinner
hardened surface layer.
Figure 56 shows a zone of the measured
values of residual stresses at four bearing locations of the knee shaft (Ref 54). The arithmetic
mean value for characteristic depths of the hardened layer was determined, and the highest and
lowest values of residual stresses at the given
depth were established, respectively.
The upper and lower values of the residual
stress with the individual depths of the hardened
layer were defined by the range of scatter of
residual-stress values. The upper and lower confidence limits were defined using a statistical
Fig. 55
Residual-stress profile after induction surface hardening on sample A of the mean bearing location in
the middle of the crankshaft and on sample C on the extreme left
side. Source: Ref 15, 20, 44
data analysis of the measured values of the residual stresses through the hardened-layer depth.
Figure 57 shows a zone of the scattered values
of residual stresses, the calculated variation of
mean values in the individual depths, and the
zone determined by the upper and lower
confidence limits for the hardened-layer depth,
to a depth of 5.5 mm (Ref 54).
The zone between the upper and lower confidence limits is important and requires a very
definite gradient of residual stresses. Where the
limit, that is, the depth, should be set and where
the gradient should be controlled depends on the
chosen induction-heating conditions. The heating conditions applied in the study indicated that
the specified gradient of residual stresses should
be ensured at a depth ranging between 3.0 and
5.0 mm.
A practical application showed that the stress
gradient between the surface and a depth of
pg 460
18/8/2008 4:02PM Plate # 0
3.0 mm was not problematic, since the values

obtained lay in this range in all cases. Difficulties may be encountered in induction heating
and induction hardening, respectively, if the
residual-stress value obtained at the surface lies
at the lower limit of the scattered results. This
means that the measured compressive residual
stresses are at the lower limit as well, which
may result in a steeper gradient to the depth of
3.0 mm. Consequently, the required residualstress variation cannot be achieved through the
entire hardened layer.
The cylindrical specimen was surface induction heated to a temperature of 980 C and
then quenched in saltwater. Figure 58 shows
the calculated distribution of individual
microstructural phases from the surface to the

center of the cylindrical specimen at the end of
cooling (Ref 20, 47).
The initial microstructure was preserved to a
radius, r, of 5.5 mm. In the radii between 5.5 and
6.6 mm, the pearlitic-ferritic microstructure and
low-carbon martensite appeared. With the radii
exceeding 6.6 mm, a fine martensitic microstructure with approximately 6.0% residual
austenite was obtained due to intensive cooling.
With the selected conditions of induction heating and fast cooling of the specimen, no homogeneous martensite was formed in the surfacehardened layer. Figure 59 shows the calculated
and measured variations of individual components of residual stresses (Ref 20, 47). The
Fig. 56
Residual-stress profiles for six measurements on four bearing locations after induction hardening. Source: Ref 54
Fig. 57
Determination of upper and lower confidence limit and arithmetic mean values of residual stresses through the hardened
depth. Source: Ref 54
pg 461
18/8/2008 4:02PM Plate # 0
calculated variations of residual stresses represent high compressive stresses at the surface,
that is, the axial component of residual stresses,
sz, equals 803 N/mm2, and the tangential one,
sT, is 588 N/mm2. On the contrary, the tensile
residual stresses were calculated after hardening
in the core, with the preserved pearlitic-ferritic
microstructure. Thus, the axial component of
residual, stresses, sz, calculated for the core
equalled to +370 N/mm2 and the tangential
one, sT, was +62 N/mm2. The diagram in
Fig. 59 indicates that the maximum tensile
stresses were attained in the transition zone
between the hardened layer and the unhardened
one (Ref 47). At a greater depth, very low stress
gradients occurred, and in the opposite direction,
that is, in the thin surface layer to a depth of
2.5 mm, very high gradients of residual stresses
occurred. The variations of residual stresses
were determined experimentally by the x-ray
diffraction method. The large gradient changes
of the measured residual stresses in the
thin surface layer can also be confirmed by
measurement. The results of the measured and
calculated variation of residual stresses in the
surface-hardened layer sufficiently agree with
small local deviations. To determine the local
deviations of the variation and of residual
stresses, numerous calculations were made with
varying physical parameters of the material as
well as different process parameters.
Fig. 58
Ref 20, 47
Calculated distribution of microstructures along

the cylinder radius at the end of cooling. Source:
A particular problem with steels having

ferritic-pearlitic and pearlitic-ferritic microstructures, respectively, is that heating of short
duration does not ensure complete homogenization of austenite.
Figure 60 shows the influence of inhomogeneity of austenite on the level of residual
stresses, which is particularly noticeable in the
martensite zone (Ref 20, 47). The initial inhomogeneous austenitic microstructure resulted
in the appearance of a martensitic transformation with a small fraction of residual austenite. With regard to the volume fraction of
martensite and residual austenite, plastification
of the material occurred, which produced internal stresses and the variation of residual
stresses, particularly in the thin surfacehardened layer.
Figure 61 shows the calculated distribution of
residual stresses due to heating of the specimen
with heating rates of 200 and 800 C/s to a
temperature of 1050 C, followed by cooling
with a cooling rate of 1500 C/s (Ref 20, 47).
With the high heating rate, 800 C/s, a very steep
transition of residual stresses from the compressive to the tensile zone was obtained, which
resulted in a decrease in fatigue strength. With
the considerably lower heating rate, 200 C/s,
the variation of residual stresses in the thin
surface layer was essentially more favorable than
that with the high heating rate. The variation of
Fig. 59
Calculated and measured residual-stress profiles

after induction surface hardening. Source: Ref 20, 47
pg 462
18/8/2008 4:02PM Plate # 0
the tangential and axial components of residual

stresses permits the following observations:

With the lower heating rate, the residual

stresses at the surface are lower by 100 to
200 N/mm2.
The stress gradient for the tangential and
axial components, st and sz, is very small in
the subsurface, from a depth of 0.7 to
4.5 mm.
The transition from compressive to tensile
residual stresses does not occur before a
depth of 2.6 mm due to a small stress

gradient.
The radial component of residual stresses is
0 at the surface. In the subsurface, it is of the
tensile character. Thus, it equals approximately
50 N/mm2 with the higher heating rate, 800 C/
s, and is 400 N/mm2 with the lower rate,
200 C/s.
Figure 62 shows simulation of the variation of
residual stresses with a favorable rate of surface
induction heating, 200 C/s, which gives the
Fig. 60
Residual-stress profiles after induction surface hardening for heterogeneous and homogeneous austenite at austenitizing
temperature. Source: Ref 20, 47
Fig. 61
Simulated residual-stress profiles at maximum surface temperature (Tmax = 1050 C) with various heating rates
(VH1 = 200 C/s and VH2 = 800 C/s) and at a given cooling rate, VC, of 1500 C/s. Source: Ref 20, 47
pg 463
18/8/2008 4:02PM Plate # 0
maximum temperature obtained at the surface,

that is, 1050 C (Ref 20, 47). This was followed
by quenching with two cooling rates, 1500 and
300 C/s.
The variation of the tangential and axial
components of residual stresses permits the
following observations:
With the higher cooling rate, compressive

residual stresses are obtained at the surface,
that is, the axial component of 800 N/mm2
and the tangential component of 1050
N/mm2.
With a low gradient, the axial and tangential
components of the stresses vary; they change
their sign to the tensile zone only at a depth
between 2.6 and 2.8 mm.
With the lower cooling rate, considerably
lower compressive residual stresses are
obtained at the surface, that is, an axial
component of 275 N/mm2 and a tangential component of 390 N/mm2.
A comparison of the axial and tangential
components of residual stresses indicates
that because of the considerably reduced
cooling rate, the latter is lower in the thin
surface layer by a factor of 3.
The gradients of residual stresses with a
higher or lower cooling rate are favorable,
since a slow decrease of compressive to
tensile residual stresses results in a minor
susceptibility of a machine part to fatigue
under dynamic loads.
Fig. 62
Residual stresses due to surface induction

hardening of the gear wheels were calculated
along the tooth surface during the heating process as well the quenching process.
The series of images in Fig. 63(a) show distributions of internal stresses during the heating
process after 20, 60, 100, and 187 s (Ref 52).
The profile variation of internal stresses indicates that:
In the initial heating phase, the internal

stresses in the tooth root are of the compressive character and reach up to 700 N/
mm2; with further heating, they change into
tensile internal stresses ranging from +200
to +300 N/mm2.
They are considerably lower in the zone
reaching from the root to the pitch circle.
Up to a heating time of 60 s, they are of
the tensile character and reach up to 700
N/mm2. With further heating, they gradually
change to compressive stresses attaining
100 N/mm2.
They are obviously very low and insignificant in the upper part of the gear-wheel
tooth, from the pitch-circle diameter to the
tip of the tooth, therefore, they are not plotted. For this reason, deformations are the
greatest at the tip of the gear-wheel tooth.
The series of images in Fig. 63 show changes

of internal stresses during the heating process (Fig. 63a) and the cooling or quenching
Simulated residual-stress profiles at maximum surface temperature (Tmax = 1050 C) with heating rate, VH, of 200 C/s and at
a various cooling rates (VC1 = 1500 C/s and VC2 = 300 C/s). Source: Ref 20, 47
pg 464
18/8/2008 4:02PM Plate # 0
process (Fig. 63b) (Ref 20, 52). It is characteristic of the quenching process that considerably higher internal stresses occur and are
more important in the tooth root. The figures
show the magnitude and variation of internal
stresses after the cooling times of 0.1, 1, 10, and
300 s, when the gear wheel is finally cool. The
series of graphic representations of the internal
stresses at the tooth surface indicate that:
In the initial quenching phase, at 0.1 and 1 s,

the internal stresses are of the tensile character along the entire tooth height. They are
the highest in the tooth root and gradually
decrease toward the tip of the tooth.
Between the quenching times of 1 and 10 s,
the sign of the internal stresses changes,
becoming gradually of the compressive
character and ranging between 100 and
150 N/mm2 in the upper part of the tooth and
up to 600 N/mm2 in the tooth root.
At the end of quenching, after the cooling
time of 300 s, the internal stresses change
considerably only in the tooth root and attain up to 1500 N/mm2.
20 s
60 s
0.1 s
1s
100 s
187 s
10 s
300 s
The magnitude of residual stresses was also

measured with the x-ray diffraction method and
strain gages. Figure 64 shows the results of stress
measurements in the pitch circle, that is, at the
middle of the tooth surface for both teeth and in
the tooth root where critical residual stresses
occur (Ref 20, 52). By means of the x-ray diffraction method, it was found that:
The highest residual stresses occur in the

root. At the surface they equal approximately
540 N/mm2, then increase to 750 N/
mm2 at a depth of 30 mm to reach 870 N/
mm2 at a depth of 60 mm.
The residual stresses are a bit lower in the
middle of the tooth surface, that is, at the
pitch circle. At the surface, they equal approximately 90 N/mm2, then gradually increase with a greater depth to reach 400 N/
mm2 at a depth of 60 mm.
A general conclusion can be drawn that the

experimental as well as theoretical results agree
very well and provide useful information,
especially to a technologist in the manufacture
of gear wheels.
Figure 65 shows a part of a gear in cross
section, with residual stress distribution in the
surface-hardened layer of the gear tooth (Ref 15,
20). Also in this case, a residual-stress distribution typical of induction hardening is obtained, that is, compressive residual stresses in
the hardened layer with a martensite microstructure, followed by tensile residual stresses,
and, at greater depth, compressive residual
stresses once again. The size and distribution of
residual stresses can be influenced by power
density.
(a)
Stress
0
(b)
daN/mm2
50
Fig. 63
Ref 20, 52
Principal stress during (a) induction surface

heating process and (b) quenching process. Source:
Fig. 64
Residual-stress measurements on the root below the

tooth surface. Source: Ref 20, 52
pg 465
18/8/2008 4:02PM Plate # 0
A number of authors have done research into

the effects of power density or energy input used
in heating, with the purpose of determining optimal induction-hardening conditions that lead
to good mechanical properties of gears. Lower
power density at the same current frequency
requires longer heating times at the same depth
of the hardened layer and results in higher compressive residual stresses with a moderate transition of residual stresses from the compressive
into the tensile region.
In induction hardening of gears, it is necessary
to ensure the most uniform depth of the hardened
layer to achieve a symmetric distribution of
residual stresses in the gear tooth cross section.
Workpiece Distortion in Induction

Surface Hardening
Dimensional changes of the workpiece are
closely related to the internal stresses occurring
in heating and cooling. During heat treatment,
when internal stresses in the workpiece are
higher than the yield point, distortion of the
workpiece takes place. When the workpiece has
an axisymmetric shape, it is very important to
Fig. 65
Residual-stress distribution in the induction surfacehardened layer of the gear tooth. Source: Ref 15, 20
place the induction coil so that symmetric

heating and quenching is achieved, resulting in
uniform thickness of the hardened layer with a
martensitic microstructure.
Figure 66 shows induction surface hardening
of a cylindrical rod and tube. When the gap
between the coil and the workpiece varied from
the maximum gap, r1, to the minimum gap, r2,
that is, r14r2, the hardened layer thickness
varied also (Ref 15, 27). This means that the side
of the workpiece with a thicker hardened layer
will suffer greater changes in volume than the
side with the thin layer. A changing thickness of
the hardened layer will result in distortion or
curvature of the cylindrical rod in the direction
of the thinner induction-hardened layer.
Figure 66(a) shows a cylindrical tube with an
induction-hardened inner surface. Asymmetrical placement of the coil inside the tube (hole,
bore) causes differences in the gap between the
coil and the bore surface inside the workpiece.
The conditions of heating are very similar to
those mentioned earlier, and the cylindrical rod
bends in the direction of the thicker hardened
layer.
Similar distortion happens on prismatic bodies where the specimen dimensions are a h l.
Distortion of workpieces subsequent to induction surface hardening on rectangular crosssectional rods, strips, plates, and similar forms is
defined or estimated with respect to the size of
the object and thickness of the hardened surface
layer (Ref 27).
In unilateral surface hardening of prismatic
bodies, the distortion will be greater with a
smaller height (h) and greater length (l). Distortion also grows with the depth of the hardened
surface layer. Figure 67 shows distortion or,
more precisely, bending of a steel rod with a
rectangular cross section, where the rod has the
same cross section a h and the same length, l
(Ref 15). Figure 67(a) presents specimens of
equal size and shape that have been induction
hardened to different depths. For an ideal comparison, choose heat treatment conditions that
will compare easily, that is, equal power (P) and
frequency (f ), and change the time of heating.
Changing the induction heating time allows the
identification of different depths of the hardened
layer and very similar microhardness or microstructure in the transition zone. Figure 67(b)
shows the conditions in induction surface heating when the temperature transformation is not
exceeded, TA1 . This means that on the completion of quenching, the hardened layer has not
pg 466
18/8/2008 4:02PM Plate # 0
Fig. 66
Influence of nonuniform thickness of surface-hardened layer on distortion for cylindrical steel rod and tube. Source:
Ref 15, 27
Fig. 67
Bending of a steel rod of rectangular cross section as a function of the thickness and location of the hardened layer. Source:
Ref 15
pg 467
18/8/2008 4:02PM Plate # 0
been reached (d0B = 0), and this is reflected on

only a partial depth of the overheated layer, d00B.
Despite surface quenching, volume changes in
the surface layer are not affected, which means
that there is no distortion of the workpiece, so the
size of bending is f = 0. Bending distortion
may be noted quite rarely, and only if very intensive cooling is applied with temperature differences between the surface and subsurface
defined by the depth at which the temperature
in the specimen is equal to the temperature of
the environment. However, such cases can be
neglected, since the goal is to achieve the austenitization and homogenization temperature,
which is much higher than the temperature of the
ferrite/pearlite-to-austenite transformation, TA3 .
In practice, the cooling rate is adapted to the
quenching medium, which ensures the formation of a hardened microstructure with minimum
internal stresses.
Figure 67(c) illustrates conditions after
induction surface hardening achieves a very
small depth of the hardened layer, d0C , and a
correspondingly small depth of the heat-affected
zone. The achieved depth of penetration (conduction) of heat after induction hardening is
equal to d0C +d00C =dC . The volume changes in
the specimen are influenced only by the hardened layer, that is, the volume of the hardened
layer versus the entire volume of the specimen.
If the volume changes are small, then the
bending of the specimen is negligibly small. For
a given hardened layer depth, it is possible at
each point during the cooling process to ensure a
higher yield point of the material (Rp, 0:2 ) than
the internal stress in the axial direction of the
specimen. When the depth of the hardened layer
is equal to or smaller than one-twentieth of the
0 41/ h, there is no bendspecimen height, dC
20
ing deformation of the specimen, only a slight
increase in residual stresses in the axial direction, despite a minimum volume change of the
hardened surface layer.
Figure 67(d) illustrates the conditions after
induction surface hardening of a specimen with a
slightly greater depth of the hardened layer, dC0 ,
than on the specimen in Fig. 67(b), dC4dB. The
achieved depth of penetration (heat conduction)
after induction heating is also slightly greater
and is dC or dC4dB. If the achieved hardenedlayer thickness is dC41/10 h, then hardening is
accompanied by deformation of the specimen,
described as bending deformation in the middle
of the specimen ( f40). Surface hardening
is followed first by annealing to remove the
stresses and then by mechanical straightening

and finally grinding. The procedures of subsequent treatment of machine components are
chosen mainly with respect to the size of the
bending deformation.
Figure 67(e) shows the conditions after
induction surface hardening of the specimen
on both sides. Several different procedures can
be chosen to ensure two-sided hardening. The
first procedure involves hardening of the top
surface to a certain depth, d0E , and then of
the bottom surface to the same depth. In the
second procedure, both top and bottom surface
are hardened at the same time. The third procedure involves hardening of the top surface
from the left to the right side and the bottom
surface from the right to the left. The most suitable procedure is the second one, where the
top and bottom surfaces are hardened simultaneously to the same depth. The achieved depth
00
of heating is equal to dE , whereas the depth of
the hardened layer is equal to d0E . When equal
heat treatment conditions are ensured on both
sides of the specimen, the depth of the hardened
layer on the top surface, dEU, is equal to the
depth of the hardened surface on the bottom
surface, dED.
This induction surface-hardening procedure
ensures that the bending deformation in the
middle of the specimen is equal to zero ( f = 0).
Figure 68 shows two crankshafts that differ in
the position of the crankpin with respect to the
main journal. The first example presents a
crankshaft where the distance, a, between the
axes of the crankpin and the main journal is
greater than the sum of the radii of both the
crankpin and the journal:
a4
dGL dKL
+
2
2
In the second example, the crankshaft is

slightly different, having the distance between
the two axes equal to the sum of the radii of the
crankpin and journal:
a=
dGL dKL
+
2
2
This must be considered in the analysis of the

deformation of the crankshaft after hardening
cylindrical parts that alternate in following one
another (crankpin and main journal).
For the conditions after induction surface
hardening of the crankshaft described in the first
pg 468
18/8/2008 4:02PM Plate # 0
example, Fig. 68(a) shows induction surface

hardening of the crankpin, which causes the
deformation of the main journal (Ref 15). Then
follows induction surface hardening of both
main journals on the left and right side. Due to
tensile residual stresses in the web between both
main journals, the web sections come closer to
one another, and the length of this crankshaft
section shortens by 2Dl. Considering the number
of crankpins or the number of main journals, the
entire length of the crankshaft decreases by Dl.
This value depends on the number of the main
journals, N, and the size of contraction per one
embedment (Dl). The entire contraction of the
crankshaft is thus DLC = N Dl.
In the second example (Fig. 68b), the crankshaft is designed differently, with the axial distance between the crankpin and the main journal
being equal to the sum of both radii (Ref 15). As
in the previous case, induction surface hardening
of the crankpin induces tensile stresses in the
web, causing distortion of the main journal. Both
adjacent main journals come closer to one
another, but the distortion is smaller than in the
first case. After induction surface hardening of
both adjacent main journals, any temporary
distortion that may have occurred vanishes, and
the crankshaft has equal length, just as prior to
heat treatment. Research has confirmed that the
stress state in the workpiece is more favorable
Fig. 68
and distortions are smaller in progressive hardening than in single-shot hardening. The only
problem is whether the size and shape of the
workpiece allow the application of progressive
hardening.
Fujio et al. (Ref 52), in their third report on
induction hardening, focused on studies of distortion of gear-wheel teeth and residual stresses
in gear wheels. The authors measured the outer
diameter and the root diameter of the gear wheel
across two opposite teeth and the root parts of
the teeth with a micrometer before and after
quenching. The same dimensions were also
calculated theoretically, taking into account the
volume changes due to phase transformations,
and serve as a basis for determination for
deviations.
Figure 69 shows a change of the outer diameter of the gear-wheel tooth for gear wheels 2,
3, and 4 (Ref 20, 52). The gear wheel has 26
teeth, which means that the deviation was measured between the first and twenty-sixth tooth,
the third and sixteenth, and so on, so that thirteen
measurements were made in total. The results of
the measurements are shown as points in the
diagram. The dotted line represents the theoretically calculated increase in diameter due to
heat treatment.
Figure 70 shows a change of tooth height for
gear wheels 2, 3, and 4, as in the previous
Crankshaft distortion after surface induction hardening of individual journal locations. Source: Ref 15
pg 469
18/8/2008 4:02PM Plate # 0
example (Ref 20, 52). The dotted lines in the

individual diagrams represent the theoretical
deviations of height of the individual teeth by
0.18 mm and the deviations found by measurement, which are mainly greater than the theoretical values. The greatest deviation can be
found with gear wheel 4. It exceeds 0.2 mm.
Figure 71 shows changes of the root diameter.
Calculations show that after heat treatment,
the root diameter decreases by approximately
0.02 mm (Ref 20, 52). The measured values
of the same diameter, however, remained
Fig. 69
Change of tip gear-wheel diameter after quenching.

Source: Ref 20, 52
Fig. 70
Change of whole depth of gear wheel after quenching. Source: Ref 20, 52
unchanged for gear wheels 2 and 3 and decreased by 0.1 mm maximum for gear wheel 4.
A comparison of the outer tooth diameter and
the tooth root diameter, also taking into account
the tooth height, shows that distortion of individual gear wheels is considerably more complicated and cannot be described by the selected
measurement methods; therefore, changes of
tooth profiles along with tooth height were
measured. Figure 72 shows tooth profile error
curves after induction surface heating and after
quenching (Ref 20, 52). Prior to quenching, the
left tooth surface was marked with the letter a
and the right tooth surface with the letter b.
Measurements of the height profile are shown
for two teeth, those marked 1 and 14, of gear
wheels 2 and 3. Considering that the tooth height
is 10 mm and the deviations are plotted in millimeters, their absolute value can be evaluated. It
ranges between 9 and 100 mm at each tooth
surface concerned.
One method of induction surface hardening
appropriate for large gear modules is known as
gap-by-gap hardening. It belongs to the progressive hardening methods. In this case, the
coil is placed so that it ensures a uniform gap
between the coil and the flanks of two adjacent
teeth. The tooth gap-hardening method is very
demanding and requires much experience and
knowledge to achieve the desirable properties of
the gear. This method is also known as contour
Fig. 71
Change of root gear-wheel meter after quenching.

Source: Ref 20, 52
pg 470
18/8/2008 4:03PM Plate # 0
Fig. 72
Tooth profile error curves (a) after induction surface heating and (b) after quenching. Source: Ref 20, 52
hardening. It is an ideal method for heat treatment of gears because it increases the hardness
on the tooth surface only slightly and decreases
the load-bearing capacity in the tooth root. Gears
heat treated in this way exhibit very good
behavior in operation, because compressive
residual stresses are present in the root of the
tooth. Gears with induction-hardened flanks,
given that the dimensioning is carefully carried
out, can achieve the highest fatigue strength. To
verify the results of induction surface hardening,
it is necessary to discuss certain measures for
controlling the quality of the hardened layer. For
this purpose, hardness and microhardness measurements supported by microstructural analysis
are commonly used. A disadvantage of this
procedure is that, due to the method of heating
and quenching (nonuniformly overheated left
and right tooth flanks), slightly higher dimensional deviations may be obtained than in
simultaneous hardening of both flanks of the
same tooth (Ref 27).
Figure 73 shows the deviation in the dimensions of the tooth after induction surface hardening by heating with a coil that encircles the
gear tooth (Ref 15, 20, 55).
Measurements of gear teeth and gear gaps
after induction surface hardening show an
increased volume in the tooth root and thus a
smaller gap and increased volume at the tip of
the tooth (increased gear diameter). These
Fig. 73
Distortion of individual tooth shape after induction

hardening caused by volume changes. Source:
Ref 15, 20, 55
volume changes result in a slightly smaller width

of the tooth above the pitch circle and slightly
increased tooth width below the pitch circle
of the gear. These dimensional deviations are
relatively small and negligible in the case of
gears with small diameter or small module, but
may become more important in gears with larger
modules and greater tooth width (Ref 55, 56).
Surface hardening of gears is one of the most
frequent applications of induction surface hardening. In the first part of this chapter, different
methods of induction heating of gears was discussed, depending on gear size but also on the
properties expected of gears after heat treatment.
pg 471
18/8/2008 4:03PM Plate # 0
pg 472
Gears are axisymmetric machine elements;

therefore, volume changes due to the transformation into a martensitic microstructure induced by heat treatment are to be expected.
Teeth themselves are symmetrical, which means
that careful application of uniform heating of
particular gear teeth and uniform quenching,
regardless of the method applied, should not
cause distortion after heat treatment. In induction surface hardening, compressive residual
stresses are created in the surface, which, in
addition to hardness and wear resistance of
the surface, create a considerably high fatigue
strength and resistance to bending loads.
Figure 74 shows shape distortion after singlefrequency and dual-frequency induction surface
hardening (Ref 20, 57). Figure 74(a) shows a
simulation of distortion by showing the initial
shape of the tooth half-cut prior to hardening
(dotted line) and the modified shape after dualfrequency induction surface hardening (continuous line). Figure 74(b) shows a similar
simulation of distortion but for conventional
single-frequency induction surface hardening.
In Fig. 74, the calculated actual size of distortion is 50 mm, with an equivalent geometrical
scale of 2.0 mm. A comparison of both cases of
hardening from the viewpoint of tooth distortion
clearly shows that distortion in single-frequency
surface induction hardening is twice that
obtained after dual-frequency hardening.
This difference in tooth distortion can be
attributed to the fact that in single-frequency
surface induction hardening, the martensitic

transformation occurs in the whole tooth. Consequently, the volume change of the tooth
is greater and so is the distortion. In dualfrequency surface induction hardening, only
the tooth contour becomes hardened. Consequently, the volume change of the tooth is
smaller and so is the distortion, that is, it equals
only half of that obtained after single-frequency
hardening.
Residual Stresses after Induction Surface

Hardening and Finish Grinding
The last phase in the manufacturing of
crankshafts is fine grinding, where achieving the
desirable condition of the surface and the surface
layer requires:

Suitable dimensions of the particular bearing

locations with respect to the allowable
deviations
Suitable surface roughness
Grinding stresses are compressive or the
lowest tensile to maintain the favorable
stress profile obtained by induction surface
hardening of the surface layer
Smallest changes possible in the microstructure and also smallest changes in the
hardness and microhardness profiles in the
heat-affected zone after grinding
How is it possible to assure a desirable surface

and surface layer quality after induction surface
hardening and fine grinding? Finding an answer
to this question requires a very good knowledge
of the process of grinding on the microlevel as
well as all mechanical and heat effects acting on
the layer of the workpiece, including the type
and condition of the grinding wheel. An allinclusive consideration of the numerous influences of the kind and condition of the tool on the
changes on the surface and in the surface layer of
the workpiece in the given machining conditions
can be based on the descriptions of surface
integrity (Ref 5863).
For the grinding process, the following conditions have been selected:

Fig. 74
Ref 20, 57
Distortion after induction surface hardening with

(a) dual frequency and (b) single frequency. Source:
Different kinds of grinding

Different grinding conditions
normal, abusive)
(gentle,
Because of thermomechanical loads in the

thin surface layer during the grinding process,
18/8/2008 4:03PM Plate # 0
very complex physical and chemical processes

take place on the microlevel. For an accurate
description of the conditions in the contact zone
for machining a given material, it is necessary to
consider the grinding method as well as the kind
of material from which the grinding wheel is
made, its structure, and the grinding conditions.
The heat conditions in grinding are a result of the
contact conditions between the individual grinding grains and the workpiece material as well as
deformation work in chip formation in the shear
zone. The amount of generated heat is strongly
dependent on the chosen machining conditions
and is abducted mainly through the chip, while a
smaller part of the generated heat is transferred
through heat transfer into the thin surface layer
of the workpiece. Heating of the chip does not
cause any particular difficulties, but heating of
the thin surface layer of the workpiece creates
the conditions for different mechanical and
thermokinetic processes, which cause microchemical changes. Heating up the thin surface
layer of the workpiece can leave certain undesirable effects that change the properties of the
part surface layer and thus harm its operational
abilities. The generated friction heat is transferred through the heat-transfer phenomenon
from the contact between the grinding grain and
the workpiece into the grinding grain.
The increased amount of heat on the grinding
grain or grinding wheel intensifies the wear
processes and damage of particular grinding
grains, which follow very different and complex
mechanisms and, in the final phase, affect the
serviceability and operational life of the tool.
The grinding tool consists of grinding grains,
each representing a process. Grinding grains are
interconnected with an appropriate binder,
which is defined by different degrees of porosity
of the grinding wheel structure. Therefore, it is
necessary to also know the force needed to break
the grinding grain from a given structure, expressed by the grinding wheel hardness. Simultaneous changes in volume proportions of the
grinding grains and binder can create different
structures on the grinding wheel that do not behave in the same way. This means that by
changing the kind of material for the grinding
wheel and the binder, it is possible to achieve
equal effects by changing the structure of the
grinding wheel. It follows that by a suitable
combination of influential parameters, it is possible to achieve a longer life and wear resistance
of the wheel in equal kinematic conditions of the
wheel and the workpiece. The wear of the
grinding grains is a result of mechanical and

thermal effects, which are reflected in reduced
cuttability. Figure 75 shows the basic forms of
wear and damage on the grinding wheel grains,
expressed by characteristic changes on the
grinding grains (Ref 15, 64). Due to mechanical
loads on particular grinding grains, short-lived
but intensive heat effects are created in the contact or on the friction surfaces with the workpiece. In these cases, blunting of the grinding
grains (Fig. 75a), breaking off of the grinding grains (Fig. 75b), or splitting of the grinding
grains (Fig. 75c) may happen. Requirements are
often set that a worn-out grain should fall from
the grinding wheel at a certain moment. A wornout grain typically has an increased contact
surface, which causes the forces on the grinding
grain during the cutting process to be higher than
the binding forces between the grains, and therefore, the grain falls out (Fig. 75d). It is expected
that the manufacturing engineer will carefully
choose all the parameters of the process so that
the worn-down grains fall out. This ensures
more efficient machining and reduced thermomechanical effects in the thin surface layer of the
workpiece material. To create a suitable quality
of the workpiece surface and surface layer, it
is necessary to ensure that grinding grains will
fall out, and that the activation of new, sharp
grinding grains is made possible.
It is very important to choose a cutting condition that would make the grains fall out only
when their cuttability is reduced. Figure 75(e)
illustrates characteristic blunting of the grinding
grain due to chemical reactions at high temperatures. Chemical reactions at high temperatures are frequently followed by filling the pores
with overheated, highly plastic chips of the
workpiece material (Fig. 75f). When the pores
Fig. 75
Basic forms of wear and damage on the grinding

wheel grain. Source: Ref 15, 64
pg 473
18/8/2008 4:03PM Plate # 0
on the wheel surface become filled, the temperature in the contact zone rises and changes
the progress of the temperature cycle into the
depth of the workpiece. The maximum temperature of the temperature cycle on the surface
rises, as does the temperature in the particular
depths of the workpiece material. This results in
greater depth of the heat-affected zone in the
material, which may have fatal consequences in
terms of surface layer properties of the workpiece material.
Generally, blunting of the grinding grains
makes the contact surfaces between the grains
and the workpiece larger, which creates the
conditions for increased mechanical effects in
the contact zone and higher heat input, accompanied by stronger heat effects, in the surface
layer of the workpiece. A plastically deformed
layer is created because of interaction between
the grinding tool and the workpiece on the place
where the chip separates from the base.
A result is hardening of the thin surface layer
of the workpiece material and occurrence of
internal stresses that may lead to failure of the
thin surface layer and/or deformation of the
workpiece, with the presence of residual stresses
at the end of machining.
Macro- and microanalysis with optical and/or
electronic microscopes shows microcracks and/
or other damage on the surface caused by an
inadequate grinding method or procedure.
Microscopic assessment of the surface state and
damage on the surface quite often points to
inadequate grinding conditions. The most frequent surface damage includes hollows, mars,
torn-off areas, built-up edges of the workpiece or
the tool, and so on. It is often necessary to consider the generated heat effects that cause microstructural and/or chemical changes accompanied
by dimensional changes. The damage on the
workpiece surface should be taken very seriously, since this may give rise to very detrimental friction conditions during operation with
another element in the mating pair.
In the analysis of microstructural changes in a
thin surface layer of the material after finish
grinding, it is possible to evaluate the size of the
heat-affected layer. Figure 76 shows the time
variation of the maximum temperature on the
surface at the particular depths with respect to
the workpiece speed, VW (work speed, VW) (Ref
15, 65, 66). By knowing the melting temperature
and the austenitization temperature of the discussed steel, the depths of the remelted layer and
the heat-affected layer can be defined. Under
different grinding conditions, different temperature cycles were obtained on the surface and
in the depth of the heat-affected layer, which has
effected microstructural changes and changes
in the microhardness and residual stresses
(Ref 6466).
Figure 77 presents the temperature cycles on
the specimen surface as a function of depth in
the hardened steel at given grinding conditions
(Ref 15, 65). The temperature cycles can be
treated separately as a heating phase and then a
cooling phase.
The maximum temperatures achieved at the
surface and in the surface layer, respectively, are
also very important (Ref 64). A distinction can
be made between three characteristic cases of
temperature cycles:
The maximum temperatures at the surface

and in the surface layer, respectively, are
higher than the melting temperature of the
specimen material, depending on the temperature cycles (Fig. 78a) (Ref 65). Such
conditions may occur due to very sharp
grinding conditions or the selection of an
inappropriate grinding wheel with regard to
the specimen material. The depth of the remelted layer is only a few micrometers and
makes a very fine ledeburite microstructure
containing fine cementite spread in residual
austenite. The newly formed microstructure
has a slightly lower hardness than martensite. The residual stresses in the thin surface
layer will be tensile, due to plastic deformation of the surface layer in grinding
caused by tensile forces in the contact zone
of the specimen material. To this should be
Fig. 76
Maximum temperature drop as a function of depth in

the hardened steel during grinding with various work
speeds, Vw. Source: Ref 15, 65, 66
pg 474
18/8/2008 4:03PM Plate # 0
added the tensile stresses induced by the

occurrence of residual austenite.
The maximum temperature in the contact
zone is lower than the temperature required
for the beginning of melting of the given
material and higher than the austenitizing
temperature. The lower temperature, that is,
the austenitizing temperature, will shift
because of a high rate of heating of the
specimen toward higher temperatures, as
known from transformation diagrams
(Fig. 78b) (Ref 65). Provided the previous
microstructure of the surface layer was
martensite-cementite-carbide, a finer martensitic microstructure with a higher carbon
content, obtained at the expense of cementite-carbide phases and with the possibility
of a lower content of residual austenite, may
be expected in the thin surface layer after
grinding. The modified content of the
cementite-carbide phase depends on the
heating conditions, whereas the content of
the residual austenite depends on the cooling
conditions.
Fig. 77
Grinding temperature cycles in different depths in

the hardened steel at given grinding conditions.
Source: Ref 15, 65
Fig. 78
The maximum temperature in the contact

zone is lower than the temperature required
for the beginning of austenitization and
higher than the lower temperature that is
limited by the temperature of steel tempering. It is equal to approximately 200 C
(Fig. 78c) (Ref 65). The grinding conditions
are very mild, so that with the selection of
the correct type of grinding wheel, no
important changes are expected in the surface layer. Only martensitic tempering may
occur in the surface layer if it was not performed during the induction surface hardening of the specimen.
The engineer should be aware that surface

integrity depends on the tribological conditions
in the operation of a component/assembly.
Therefore, adequate knowledge for the assessment of tribological conditions of a component
in operation is very important for the prescription of machining that would result in the desired
surface condition and subsurface layer. In general, there are two tribological systems. The first
is the one present during machining, and the
second is acting in operation. In both tribological systems, however, it is the properties of the
workpiece material and its condition prior to and
after machining that play a key role. Under different machining conditions of grinding, different temperature cycles were obtained on the
surface and in the depth of the heat-affected
zone, which has effected microstructural changes and changes in the microhardness and residual stresses. Thus, on the surface, a maximum
temperature higher than the temperature of
melting of the workpiece material was obtained.
The depth of the remelted layer is only a few
micrometers and makes a very fine ledeburite
microstructure containing fine cementite spread
in residual austenite. The newly formed
Maximum temperature drop as a function of depth in the induction surface-hardened steel at various speed, Vw. Source:
Ref 65
pg 475
18/8/2008 4:03PM Plate # 0
microstructure has a slightly lower hardness than

martensite. The residual stresses in the thin surface layer will be tensile, due to plastic deformation of the surface layer in grinding caused
by tensile forces in the contact zone of the
workpiece material as well as the tensile stresses
induced by the occurrence of residual austenite.
The relative grinding stress is obtained by
measuring the residual stress after induction
surface hardening, followed by measuring the
same spot after induction hardening and grinding, then calculating their difference. Figure 79
shows the measured absolute residual-stress
profile after induction surface hardening and the
measured residual-stress profile after induction
surface hardening and grinding (Ref 15).
A relative grinding stress represents the difference between measured residual stress on
specimens that were induction surface hardened
and induction surface hardened and grinded.
A relative grinding tensile stress equals
+425 N/mm2, and then the sign changes in the
depth at approximately 175 mm, as shown in
Fig. 80.
Figure 80 shows the absolute residual-stress
profile after induction surface hardening and
grinding on bearing location A as well as the
average residual stress after induction surface
hardening. From this, the relative grinding stress
can be calculated (Ref 15).
Figure 81 shows a zone of relative grinding
stresses occurring under gentle grinding conditions (Ref 54).
With gentle grinding conditions, compressive
relative grinding stresses will prevail at the
Fig. 79
Subsurface residual-stress profile after induction

surface hardening and grinding (absolute stress) on
bearing location A. Source: Ref 15
surface. It is only in the subsurface that the

tensile influence can be felt.
The relative grinding stresses are low and
practically negligible, although they exert a
positive influence on the original residual
stresses after induction hardening. The zone was
defined by the upper limiting values characteristic of greater depths under normal grinding
conditions and by the lower limiting values
characteristic of gentle grinding conditions.
Generally, care should be taken when choosing
grinders and grinding conditions. The chosen
grinder should ensure regeneration of the abrasive grains, so that the worn-out grains may fall
off and the new ones grind the surface. The
selected grinding conditions should ensure that
a suitable force be applied to the worn-out
abrasive grains so that they may fall off and new,
sharp abrasive grains become active. Figure 82
Fig. 80
Subsurface profile of relative grinding stress on

bearing location A. Source: Ref 15
Fig. 81
Relative grinding stress region through thin surface

layer after gentle grinding conditions. Source: Ref 54
pg 476
18/8/2008 4:03PM Plate # 0
shows a zone of relative grinding stresses

occurring under abusive or normal grinding conditions (Ref 54). The zone of grinding stresses
was determined from the profile of residual
stresses after induction hardening and also after
induction hardening and grinding (Ref 64).
By deducting the two residual stresses, the socalled relative grinding stresses were determined. The abusive and normal grinding conditions produce remelting of the thin surface
layer, that is, its rehardening. With the abusive
and normal grinding conditions, the tensile
relative grinding stresses are always predominant at the surface. At greater depths, there is
still the influence of remelting under the abusive
grinding conditions, whereas rehardening occurs under normal grinding conditions.
For the sake of clearness, the zone was defined
by the upper limit of residual-stress values,
which is characteristic of the abusive grinding
conditions, and the lower limit, which is characteristic of the normal grinding conditions
(Ref 64).
The results confirm as predominant the residual stresses, stresses induced by the plastic
deformation of the material, and as a lesser
influence, tensile stresses caused by the formation of residual austenite. On the basis of the
measurements of residual stresses after induction surface hardening or induction surface
hardening and grinding, it can be concluded:
For residual stresses after induction surface

hardening and grinding, the conditions of
abusive grinding are a more favorable
choice. They lower to a lesser extent the
desirable compressive residual stresses after
induction surface hardening.
Grinding conditions can be chosen so that

the melting temperature of the workpiece
material (gentle grinding conditions) is not
exceeded. Then, the favorable compressive
stresses after induction surface hardening
are lowered due to plastic deformation of the
workpiece material during the process, and
thus, relatively low tensile residual stresses
are obtained. However, this will significantly lower the productivity.
Special attention should be paid to the
selection of the type of grinding wheel in
terms of grinding wheel material, binding
agent, hardness, and pore density, since a
correct selection can contribute to higher
cutting efficiency concerning the plastic deformation of the workpiece material. In this
way, the grinding tensile stresses are kept as
low as possible and make the compressive
residual stresses induced by induction surface hardening of the prevailing variety.
Induction surface hardening creates a very

desirable residual stress state. Residual stresses
are always of a compressive nature and are
usually present to the depth of the induction
surface-hardened layer.
However, a major difficulty in induction surface hardening is ensuring a very slight/slow
variation in microhardness and the existence of
compressive residual stresses in transition areas
to the microhardness of the core material. By
gently grinding and varying the hardness and
existence of compressive stresses in the transition area, it is possible to diminish the notch
effect induced by stress concentration. Additional grinding of an induction surface-hardened
surface deteriorates the stress state in the surface
layer, since grinding has always induced tensile
stresses. By the correct selection of machining
conditions and grinding wheel properties, the
engineer will contribute to less tensile residual
stresses and will avoid deteriorating the favorable residual-stress state after induction surface
hardening (Ref 44, 59).
Hardness Profiles in the Induction

Surface-Hardened Layer
Fig. 82
Relative grinding stress region through thin surface

layer after abusive or normal grinding conditions.
Source: Ref 54
The induction surface-hardened layer was

analyzed by measuring the hardness and
microhardness and their relation to the microstructure (Ref 44). To establish some relationships between microstructural changes versus
pg 477
18/8/2008 4:03PM Plate # 0
hardness across the entire hardened layer and the

microhardness in the very thin surface layer, the
latter were measured by the Vickers method.
Thus, Fig. 83 and 84 show the hardness profiles
versus the induction-hardened layer depth
according to Vickers at a load of 20 N or measurements of microhardness according to Vickers in a very thin surface layer to a depth of
150 mm at a load of 2 N (Ref 44).
The hardness of the surface layer after
induction surface hardening is very uniform in
all the investigated main bearing locations and
equals approximately 520 to 550 HV2.0. The
hardness profile highly conforms to the residualstress profile, which is confirmed by a decrease
in hardness in the transition zone. The hardness
profile in the transition area is likewise very
steep and points to high stress concentrations in
this location when the crankshaft is in the loaded
state.
Figure 83 presents the hardness profile in

an induction surface-hardened layer to a depth
of 5.0 mm and the microhardness profile to
a depth of 150 mm on bearing location A
(Ref 44).
The hardness measurements show that the
quenched and tempered steel has a hardness of
approximately 220 to 260 HV2.0, and the surface-hardened layer has a hardness of approximately 540 HV2.0. In the surface-hardened
layer, a slight increase in hardness as a function
of depth is evident, which is conditioned by
microstructural differences due to varied cooling rates of the surface layer. The surface is
cooled under the effects imposed by the cooling
medium, yet at a greater depth, the effects of the
medium are accompanied by a more expressed
effect of the cold mass of the core, resulting in
the formation of very fine martensite and greater
hardness at greater depth (Ref 64).
Fig. 83
Hardness profile in the induction surface-hardened layer and microhardness profile in a very thin surface layer for bearing
location A. Source: Ref 44
Fig. 84
Hardness profile in the induction surface-hardened layer and microhardness profile in a very thin surface layer for bearing
location C. Source: Ref 44
pg 478
18/8/2008 4:03PM Plate # 0
Kosel and Kosec (Ref 67) conducted an

investigation on the cracks formed at the
surface-hardened and ground bearing location
of a crankshaft. Crankshafts are made of highquality chromium-molybdenum heat treatable
steel. The shafts were forged from square shafts
with a cross-sectional area of 120 by 140 mm. In
the course of mechanical and heat treatments,
the workpieces were tested with various nondestructive methods, for example, visual
inspection, and, if required, magnetic particle
and/or penetrant inspections. A peculiarity of
shaft production is that during forging, the
material flows from the middle of a slab ingot to
the outer section of a forging due to plastic
deformation. The steel in the middle is usually of
lower quality than the steel at the surface of a
cast iron-works blank and that of a formed steelworks semiproduct. The character of forming
the steel-works semiproduct tool, steel-works
blank, and crankshaft was such that a reverse
material flow occurred from the middle to the
surface. The authors reported the presence of
weblike cracks at the surface of the crankshaft
bearing location (Fig. 85) and gave two reasons
for the occurrence of cracks immediately after
forging (Ref 67):

Inhomogeneity, that is, nonuniform crosssectional chemical composition of steel

Inadequate conditions of bearing-location
grinding
After macrostructural and microstructural

examinations as well as a microchemical analysis were performed, it was found that, from the
metallurgical point of view, the forging showed
quality, and the defect may be attributed to the
grinding process alone. Figure 86 shows a
macrosection of the forging at the bearing, and
Fig. 85
Weblike surface cracks at bearing location of

crankshaft. Source: Ref 67
Fig. 87 shows a forging made of low-quality slab

ingot (Ref 67).
The main purpose of the investigation was to
find the cause of the weblike cracks after
induction surface hardening and the final
grinding on the bearing location on the crankshaft. Grinding was studied in the same way as at
the cut-out bearing locations. From the steelworks slab ingots prepared for forging, specimens of a suitable length were cut out and turned
to size.
Fig. 86
Macrograph of etched favorable billet cross section.

Source: Ref 67
Fig. 87
Macrograph of etched unfavorable billet cross

section. Source: Ref 67
pg 479
18/8/2008 4:03PM Plate # 0
The specimens prepared in this way were

hardened and tempered at 560 C and then surface hardened, with a maximum surface temperature ranging between 820 and 900 C. The
case depth ranged between 3 and 4 mm.
The macrograph in Fig. 88 clearly identifies
the bearing location and shows a considerable
difference in case depth. Generally, it was found
that:

There were insignificant differences in chemical compositions.

There were segregations in the central part of
the crankshaft.
There were shrinkage cavities mainly on the
inside and also on the outside part of the
forging.
There were microsegregations.
Some of the hardened bearing locations
showed dimensional variations with reference to the anticipated thickness.
The microhardness obtained at the surface
was within the expectations and somewhat
above a value of 700 HV2.0.
The depth of the hardened bearing locations
ranged between 2.0 and 2.5 mm.
No differences in the microstructures could
be found using a common optical microscope analysis or a scanning electron microscope.
The microhardness at the cracked cylindrical
specimens was somewhat lower, approximately 600 HV2.0.
Figure 91 shows two temperature cracks, with

a pore at the crack location being visible as well.
An analytical model was elaborated to clearly
demonstrate the causes of crack formation at the
hardened surface after grinding (Ref 67).
The stress condition, sR, occurring at the
hardened-layer surface after heat treatment and
grinding was known. In several cases, the surface-hardened layers took the shape of an
eccentric ring, which means a nonuniform thickness of the hardened layer. The cracks occurred
at the location of the smallest layer thickness. As
expected, residual stress occurs at the specimen
surface where cracks initiate when stress occurs due to phase transformation to martensite,
thermal stresses during hardening, stresses due
to martensitic tempering during grinding, and
thermal stress after grinding (Ref 67).
In the phase transformation from austenite to
martensite, steel volume will increase. The volume strain can be calculated using the equation:
eV =
VM 7VA
VA
where VM is the volume of an elementary cell of

martensite, which can be calculated for the given
steel:
VM =(2:861 0:013% C)2 (2:861+0:116% C)
=23:7262 103 nm3
Figure 89 shows the microhardness profile

of the cross section of an induction-hardened
layer for a favorable bearing location, and
Fig. 90 shows the profile for an unfavorable
bearing location without cracks (Ref 67).
Fig. 88
Macroscopic examination of case depth at cross

section of crankshaft bearing. Source: Ref 67
Fig. 89
Hardness characteristic at cross section of weblike

cracked neck as a function of depth, z. Source: Ref 67
pg 480
18/8/2008 4:03PM Plate # 0
pg 481
and VA is the volume of an elementary cell of

austenite of the same steel:
VA =0:5(3:548+0:044% C)3=22:68242 103 nm3
then eV = 0.046.
A stress condition with such a volume strain
would exceed the material strength. Table 3
indicates that the smallest hardened-layer thickness required is obtained under the condition
that the resulting stress shall be higher than the
breaking stress, according to Hooks model, and
no cracks occur at the workpiece surface after
grinding (zmin = 2.873 mm) (Ref 67).
Using the proposed analytical model, the

researchers (Ref 67) evaluated the magnitude of
stresses in steel after surface hardening and
grinding. They found a mutual dependence of
the hardened-layer profile and the resistance
against material cracking during grinding.
Cracks will appear at the surface of those
bearings in which short overheating of steel
occurred and the hardness in a narrow zone
decreased strongly during grinding. In such
cases, through hardenability of a material
reaches a depth of 150 mm, where tempered
martensite will form. The thickness of the
surface-hardened layer with all the cracked
bearings was approximately 2.0 mm, which
indicated that an appropriate model was chosen.
Fatigue Strength of Materials
Fig. 90
Hardness characteristic of neck cross section without

cracks as a function of depth, z. Source: Ref 67
A heat treatable AISI 4140 steel was used

for manufacturing crankshafts. This steel is
very appropriate for statically and dynamically
loaded parts of car engines and machines
because of its high hardness achieved after hardening (57 HRC). The steel is characterized by
good hardenability and is thus suitable for manufacturing machine parts with large cross sections
in which a very high strength can be obtained
after refinement. After tempering, the steel does
not show a tendency to brittleness, and therefore,
no special heat treatment procedures are required. This steel is also suitable for surface
hardening (flame surface hardening, induction
surface hardening) and displays a very good
resistance to wear.
However, special attention must be paid to the
part design phase, and great care should be given
to the design of radius and transition areas
to prevent notch effects under dynamic loads.
The steel is adapted for use in a wide range of
Table 3 Minimum case depth required for no
cracks at workpiece surface
Fig. 91
Thermal cracks in surface-hardened neck section.

Source: Ref 67
Excentricity
of hardened
layer
(e), mm
Depth
(z), mm
Temperature
prior to
quenching
used in
Hooks model
(TK), C
0.594
0.709
0.877
1.146
3.156
3.0412
2.873
2.6045
343
336
325
307
Source: Ref 67
N/mm2
Resultant
stress after
hardening
and grinding
(sR(T0)),
N/mm2
1123
1144
1177
1232
1306
1327
1360
1414
Yield
stress of
martensite

,
RK
0:2, (T K )
18/8/2008 4:03PM Plate # 0
temperatures, and it preserves high toughness

even at low temperatures (Ref 44, 68, 69).
Impact tests provide useful quantitative comparative data with relatively simple test specimens and equipment. However, these tests do
not provide property data for design purposes for
material sections containing cracks or defects.
Data of this type are obtained from fracture
mechanics, in which theoretical and/or experimental ways determine the fracture of structural
materials containing pre-existing cracks and
defects.
The number of cycles to failure (N ) obtained
by the S-N curve under load or a stress-control
condition is related to the total life of the part up
to failure. Fatigue cycles throughout the life of
the machine part consist of crack initiation and
its propagation. Crack propagation refers to
stable crack growth up to the crack instability.
The S-N curve approach does not separate the
crack initiation phase from propagation. Industrial experts assume that the crack has already
initiated in the machine part, and only the total
number of cycles associated with the propagation are of interest. Existing or initiated cracks
assumed in the machine part are the result of
dynamic load cycles induced during manufacturing, prior to its use. The size of the preexisting crack can be assumed based on the
capability of suitable inspection. The expert may
assume an initial surface crack after testing and
makes a decision about the application of the
part.
Using the available initial defects in the
material, the total life of the part can be assessed
by an appropriate fracture mechanics method.
Basic factors affecting the shape of the S-N
curve are:

Materials selection and heat treating or cold

working conditions
Various types of loading on the specimen,
such as tension, compression, torsion, or a
combination
Loading conditions described by medium
stress, amplitude stress, and frequency
Influences on the environment carried by
temperature, corrosion, and other factors
Some major factors that affect the strength of

a metal include:
Stress concentration: Fatigue strength is

greatly reduced by the presence of stress
raisers, such as notches, holes, keyways, or
sharp changes in cross sections.

Surface roughness: In general, the smoother

the surface finish on the metal sample, the
higher the fatigue strength.
Surface condition: Since most fatigue failures originate at the metal surface, any major
change in the surface condition will affect
the fatigue strength of the metal.
Environment: If a corrosive environment is
present during the cyclic stress of a metal,
the chemical attack greatly accelerates the
rate at which fatigue cracks propagate.
During machining processes, various defects

occur on the surface, such as small scratches and
grooves, and are introduced into the workpiece
surface. Typical failures are moving machine
parts, such as shafts, connecting rods, and gears.
It is estimated that failures of machine parts in
machines contribute approximately 80% of
fatigue failures. These surface detects can limit
the fatigue life. Improving the surface finish by
polishing will increase fatigue life significantly.
One of the most effective methods of increasing fatigue life is the existence of residual
compressive stresses in a thin surface layer.
Thus, applied surface tensile stress will be partially reduced in magnitude by the residual
compressive stress. The net effect is the probability of crack formation and a consequent reduction in fatigue failure.
According to the AISI standard, the heat
treatable structural steel 4140 contains between
0.38 and 0.45% C, 0.90 and 1.2% Cr, and 0.15
and 0.30% Mo. It has very high hardenability,
contributing to high strength values in products
with high mass. Molybdenum yields a desirable
fine microstructure after hot working as well as
heat treatment, contributing to a good strengthto-toughness ratio. Due to its fine-grained
microstructure, it also reaches a relatively high
toughness in the heat treated condition.
The strength of the steel as well as its surface hardness and wear resistance may be increased by heat treatment and thermochemical
treatment.
Mechanical properties of steel having a diameter of up to 40 mm and between 40 to
100 mm are given in Table 4.
Tensile strength of the steel varies between
880 and 1080 N/mm2, and a minimum toughness value, r3, equals approximately 41 J. The
steel is very sensitive to notch and transition
on machine parts subjected to fatigue loading.
Fatigue strength of the material is lowest under
torsional load, sT, and varies for the diameters
pg 482
18/8/2008 4:03PM Plate # 0
mentioned, that is, 16 to 40 mm, so that

sT = 285 N/mm2, and for diameters from 40 to
100 mm, sT = 255 N/mm2. Fatigue strength
under torsional load is three or four times lower
than the tensile strength of steel under static
load, according to the data in the table.
In professional literature, data on fatigue
strength of materials are usually presented for a
prescribed specimen shape and size that has
been adjusted to the testing device. Specimens
for fatigue strength are usually of cylindrical
shape, with a smaller diameter in the middle part
and a rounded transition into the larger-diameter
part. The latter is then usually clamped for
testing. Modes of loading the specimens vary but
are usually either torsion, bending, or tension/
compression. The highest fatigue strength is
displayed by a specimen subjected to bending
loads (Ref 27, 36). For other modes of loading,
the relation with bending fatigue strength, swb,
is expressed empirically, that is, fatigue strength
in torsion, tw, is 0.58 swb, or fatigue strength in
tension/compression, swz, is 0.70 swb.
Table 4 Mechanical properties of heat treated
structural steel 4140
Diameter
(D), mm
16540
405100
Source: Ref 15
Fig. 92
Tensile strength
(Rm), N/mm2
Yield point
(Rp0.2), N/mm2
9801180
8801080
769
635
Extension Toughness
(r3), J
(A5), %
11
12
41
41
Figure 92 presents curves for different modes

of dynamically loaded specimens made from
various steels that had been heat treated in
assorted ways (Ref 15). From among six curves,
four represent specimens made from heat treatable steel and two for cementation steel specimens. The heat treatable steel is Cr-Mo-Ni steel
with 0.37% Ni, and the specimens were heat
treated in two different ways:

Surface hardening applied to specimens with

a smooth shape (curve 1) and specimens
with a slot (curve 2)
Quenching or tempering applied to smooth
specimens (curve 4) and specimens with a
slot (curve 6)
The cementation steel is a chromium-nickel

steel with 0.15% C, where the specimens were
smoothly shaped (curve 3) and slotted (curve 5).
Axle shafts used in cars, trucks, and farm
vehicles are, with few exceptions, surface hardened by induction. Although a portion of the
hardened surface is used as a bearing in some
axles, the primary purpose of induction hardening is to put the surface under a state of compressive residual stress (Ref 45). By this means,
the bending and torsional fatigue life of an axle
may be increased by as much as 200% over that
for parts conventionally heat treated (Fig. 93)
(Ref 45).
Graphs of the fatigue strength of surface-hardened and carburized specimens. Source: Ref 15
pg 483
18/8/2008 4:03PM Plate # 0
Induction-hardened axles consist of a hard

surface, high-strength core, and tough outer case
with good torsional strength and a tough, ductile
core. Many axles also have a region in which the
case depth is kept very shallow, so that the part
can be readily straightened following heat
treatment. In addition to substantially improving
strength, induction hardening is also very costeffective. This is because most shafts are made
of inexpensive, unalloyed medium-carbon steel
that is surface hardened to case depths of 2.5 to
8 mm, depending on the cross-sectional size. As
with crankshafts, typical hardness (after tempering) is approximately 50 HRC. Such hard,
deep cases improve yield strength considerably
as well (Ref 45).
Modern transmission shafts, particularly
those for cars with automatic transmissions, are
required to have excellent bending and torsional
strength, as well as surface hardness for wear
resistance. Under well-controlled conditions,
induction-hardening processes are able to satisfy
these needs, as shown by the data in Fig. 94,
which compares the fatigue resistance of
through-hardened axles (Ref 45). The inductionhardening methods employed are varied and
Fig. 93
include both single-shot and scanning techniques.

The hardness achieved on the surface of the
hardened specimens was 56 to 59 HRC and on
the cemented specimens, 58 to 59 HRC. To test
the effects of the slot on fatigue strength, a slot
of equal size and shape was made on all the
specimens, whether they were quenched or
tempered, hardened or cemented. It was made
in the middle of the cylindrical specimen to an
equal depth of 0.4 mm. The depth on the surface-hardened and cementation specimens was
1.5 mm. The results of testing showed that there
are significant differences in terms of heat
treatment methods and that the highest fatigue
strength was found in surface-hardened specimens.
A comparison of the fatigue testing results
showed that:
In surface-hardened specimens with a

smooth cylindrical shape and with a slot, the
difference in the achieved fatigue strength is
minimal. This can be attributed to a very desirable distribution and size of compressive
residual stresses throughout the hardened
Bending fatigue response of furnace-hardened and induction-hardened medium-carbon steel tractor axles. Shaft diameter:
70 mm. Fillet radius: 1.6 mm. Source: Ref 45
pg 484
18/8/2008 4:03PM Plate # 0
layer. Since the depth of the slot reaches only

one-quarter of the hardened surface layer,
the size of compressive residual stresses at
the slot is still very high, so that the weakening due to the slot and stress concentration
along the slot does not cause any essential
drop in fatigue strength.
In quenched and tempered specimens with
no earlier prestressing of the surface layer,
the fatigue strength was considerably lower
than in surface-hardened specimens. The
fatigue strength of quenched and tempered
specimens with a slot was also remarkably
lower. The results show that the difference in
fatigue strength in slotted induction surfacehardened specimens and slotted quenched
and tempered specimens is 5 to 1.
Smooth cemented specimens displayed 25%
lower fatigue strength than the surfacehardened specimens of the same shape,
whereas the cemented specimens with a slot
displayed 50% lower fatigue strength than
the same surface-hardened specimens.
The question arises about what the fatigue
strength is in those specimens where the slot
reaches deeper than the hardened layer. A
regular problem in these cases is crack
occurrence and propagation of cracks starting from the slot. Due to the shape of the
specimen and the slot, stresses start concentrating at these places, depending on the
type and size of external loads. It should not
be forgotten that there are no compressive
residual stresses along the slot, and the size
of tensile stresses along the slot plays a
decisive role in crack occurrence.
Figure 95 shows bending stress in a tooth root

subjected to dynamic load versus the number of
Fig. 94
oscillations (Ref 36). Figure 95(a) shows the

bending stress for induction surface hardening
of adjacent flanks of two teeth with a coil
reaching into the tooth gap. In the process, the
tooth flank as well as the tooth root are hardened
(Ref 36). This kind of heat treatment of gears
from steels for induction surface hardening
provides a fatigue bending strength in the range
of 320 to 490 N/mm2. Figure 95(b) shows the
same relationship when the induction coil
encircles an individual gear tooth. In this process, the tooth flank is hardened, and the microstructure and hardness in the tooth root are
preserved (Ref 36). A result of this method of
hardening is that the fatigue strength is drastically lowered to values ranging from 200 to
300 N/mm2 for the entire range of steels suitable
for induction surface hardening. This is a considerable drop in fatigue strength for the material
in the tooth root (Ref 45, 70).
Stress Profiles in Machine Parts in the

Loaded State
Heat treatment engineers must be very careful
in choosing the conditions of induction surface
hardening in order to benefit from the distribution of residual stresses achieved in dynamically
loaded parts. In industrial practice, induction
surface hardening should satisfy the requirement
of fatigue resistance of machine components.
The main reason for this worsening of the
properties of the machine part is attributed to
tensile residual stresses in the hardened layer
and undesirable hardness distribution in the
transition zone from the hardened into the
unhardened part of the subsurface. These effects
Comparison of fatigue life of induction surface-hardened transmission shafts with that of through-hardened and carburized
shafts. Arrow in lower bar (induction-hardened shafts) indicates that one shaft had not failed after testing for the maximum
number of cycles shown. Source: Ref 45
pg 485
18/8/2008 4:03PM Plate # 0
are quite natural, and in the first phase are a

result of very rapid local heating of the thin
surface layer, while in the second phase this is
accompanied by forced quenching, which
ensures a critical cooling rate and the occurrence
of a martensitic microstructure. Both phases in
induction surface hardening can increase the risk
Fig. 95
of fatigue, especially if the latter is assessed only

from the point of view of surface hardness. To
successfully estimate the quality of the hardened
layer, one must select the optimal synergetic
effects between the input electric energy and the
interdependence between the induction coil and
the workpiece surface, connected with the
Bending fatigue strength of gear teeth at (a) tooth gap hardening and (b) flank hardening for various steels. Broken lines denote
confidence limit according to DIN 3990. Source: Ref 36
pg 486
18/8/2008 4:03PM Plate # 0
occurrence of eddy currents in the workpiece

surface layer that leads to heating. Due to complex synergetic effects in induction heating
or hardening, it is necessary to carefully study
each influence on the properties of the hardened
surface layer (Ref 27, 45). Fatigue strength
in machine components that have been induction
surface hardened is increased if the total sum
of the load tensions and residual stresses in
the surface layer is of a compressive nature.
To ensure the highest fatigue strength of a component, it is necessary to provide the following:
be treated from the point of view of heating,

overheating, and cooling/quenching as well as
the internal stresses created at a certain point
during the treatment. During heat treatment,
internal stresses are created by the temperature
differences and phase transformations between
the core and the surface, which are a result of the
volume differences between the core and the
surface. The created volume differences
between the core and the surface then give rise to
internal stresses. During the process of heating
and cooling, internal stresses may produce the
following effects:
In dynamically loaded components, the

surface is prone to fatigue occurrence, so the
surface must have the highest compressive
stresses.
If the total sum of stresses, that is, load tension plus residual stresses on the surface, is
always of a compressive nature, then there is
no chance for the occurrence of cracks and
crack growth.
To ensure good behavior of the surface and
the hardened surface layer in the loaded
condition, it is necessary to induce a suitable
prestressing in the surface layer. This can be
achieved by a carefully selected heat treatment method that would create the highest
compressive residual stresses on the surface
and a desirable profile of the latter in the
hardened subsurface layer.
Induction surface hardening offers opportunities to ensure a considerable amount of
compressive stresses in the machine component surface and to ensure restrained
transformation of compressive surface
stresses into tensile residual stresses in the
subsurface layer.
The endurance of machine components subjected to bending and torsion loads can be successfully increased by ensuring sufficiently high
compressive residual stresses. A manufacturing
goal is to create a sufficient amount of compressive residual stresses with a favorable distribution, since this is the only way to increase
the reliability of components in operation. An
early failure of a component in operation may
cause catastrophic damage on a machine and
thus a loss in profit. A decisive role in the occurrence of residual stresses is played by the
synergetic effects between the heat treatment
method, the type of material, and the shape of the
workpiece. For this reason, heat treatment must
When internal stresses are lower than that of

the yield point, higher residual stresses are
induced by heat treatment in the workpiece,
but these would not cause distortions,
cracks, or failure.
During a certain moment in heat treatment,
internal stresses exceed the yield point,
which leads to distortions and lower residual
stresses in the workpiece.
During very detrimental conditions in heat
treatment, internal stresses are higher than
the tensile strength of the material, causing
the workpiece to crack and creating larger
distortions and high residual stresses.
Numerous changes that take place in the

hardened surface layer of the workpiece are
always a result of the heating and quenching
conditions. Therefore, it is necessary to study the
events taking place in the workpiece directly
after the hardening temperature is reached.
Three zones are distinguishable in a workpiece
heated to the hardening temperature (Fig. 96a):
the first zone, where the outer layer is heated to
the hardening temperature; the second zone,
which is heated below the hardening temperature between the temperatures TA1 and TA3 for
rapid heating; and the third zone, where the
temperature is lower than TA1 (Ref 15).
Heating to the hardening temperature at a
certain depth is followed by quenching. Quenching results in the occurrence of compressive
residual stresses (Fig. 96b), when the familiar
transformations in the hardened layer take place.
The second layer does not suffer the same distortions as the surface layer, although the heating there has been sufficient enough to improve
the properties of the material. In the second
layer, hardening is incomplete, which, in comparison with the first layer, results in lower
hardness and strength of the material.
pg 487
18/8/2008 4:03PM Plate # 0
When a machine component that has been

surface induction hardened in this way is subjected to an external load (Fig. 96c), additional
tensile residual stresses can be noted in the first
and second layers (Ref 15).
Fig. 96
Since fatigue of a machine component is a

very delicate problem connected with the total
sum of tensile stresses in the second zone, there
is a danger that the effects of fatigue are transferred to the surface, due to the size, shape, and
Stress profile in a round bar in the loaded state, where residual stresses after induction surface hardening and loading stresses
add up. Source: Ref 15
pg 488
18/8/2008 4:04PM Plate # 0
location of the hardened trace. When a loaded

layer is exposed to tensile external stress up to
the surface, the surface becomes very sensitive
to the occurrence of cracks (Fig. 96d) (Ref 15).
This propensity to cracking is further increased
if defects are present in the surface of the
workpiece material.
A critical state in the second zone of the hardened surface layer occurs on locally hardened
workpieces, where the surface was overheated
or nonuniformly heated. The conditions are
quite the opposite in surface heating of workpieces, such as in nitriding and cementation
or intensive but shallow quenching of steel
(Fig. 97) (Ref 15). In these surface heat treatment procedures, heating is carried out throughout the whole machine component, followed by
slow cooling, for example, nitriding or quenching such as in cementation. Residual stresses
in nitrided and cemented surface layers are
compressive, while tensile residual stresses
occur in the central part with the refined pearlite
microstructure.
The distribution of residual stresses in heat
treatment procedures where only the surface of
Fig. 97
Stress profile in a round bar in the loaded state, where

residual stresses after carburizing or nitriding and
loading stresses add up. Source: Ref 15
the workpiece is heated (induction hardening,

flame hardening) differs greatly from the procedures where heating is performed throughout
the entire volume (nitriding, cementation). In
nitriding and cementation, the aforementioned
second layer in the subsurface does not appear at
all, because the direction of the heat flow is
opposite to the direction of the heat flow in
induction and flame hardening. The resultant
operating tensile stresses on the surface or in the
surface layer can thus be considerably smaller.
Due to the hardness of the surface, induction and
flame hardening lowers the fatigue strength of
machine components. Therefore, care should be
taken to diminish all detrimental effects in the
surface layer.
A typical example of induction surface hardening is surface hardening of gears that are
heated with a low heating rate and relatively low
current frequency. The outer hardened zone
includes almost the entire height of the gear
teeth, whereas the second zone is in the tooth
root area. A gear heat treated in this way will
meet the wear resistance requirements expected
of the gear tooth, while the strength of the other
part of the tooth is of minor importance. The
fatigue strength will be relatively low due to
high tensile residual stresses in the tooth root,
that is, in the second zone where the operating
or load tensions and the tensile residual stresses
are summed up. The energy input in heating a
gear tooth or the whole gear was such that
the second zone has not appeared. A similar
heat treatment can be applied to the spline inside
the gear. A gear heat treated in this way is
more resistant to wear and corrosion and should
have high resistance to fatigue in bending
because of a smaller thickness of the layer in the
second zone.
Residual Stresses in Carburized Machine
Parts. Carburizing is a process in which an
austenitized ferrous alloy is brought into contact
with an environment of sufficient carbon
potential to cause absorption of carbon at the
surface and, by diffusion, to create a carbonconcentration gradient in the thin surface layer.
As this definition clearly indicates, two factors
may control carburizing. Either the carbonabsorption reaction at the surface or the diffusion of carbon in the steel will determine the
rate of carburizing. Carburizing is done at elevated temperature, generally in the range of
850 to 950 C, although occasionally at temperatures as low as 790 C and as high as
1095 C.
pg 489
18/8/2008 4:04PM Plate # 0
With the hardened steel, that is, hardened

machine part, the following was studied:

Effective case depth

Depth at which martensitic microstructure is
present
Hardness variation in the transition zone
between martensite and matrix
Microstructural composition in the transition zone with bainite-pearlite-ferrite
The characteristics obtained are essentially

affected by:

The grade, chemical composition, and

microstructure of steel in the soft state
The size, that is, mass, of the machine part
Considering that certain properties described

by the previously mentioned characteristics are
specified for the machine part, adequate conditions for quenching from the austenitizing temperature should be specified too. The quenching
conditions are described with reference to the
mass of the machine part.
Thus, in steel quenching, the following are
studied:

Volume, that is, dimensional changes due to

the differences in microstructure between
the initial material condition and the hardened condition of the same material
Distortion of the machine part due to incorrect shaping of the machine part and/or an
improper quenching procedure
Residual stresses due to volume changes
occurring between the hardened zone and
unhardened zone of the machine part
Residual stresses may result from a deformation of the machine part during the quenching
process. Internal forces due to temperature
stresses during quenching may exceed the yield
stress of the material, which results in plastic
deformation during the quenching process and
residual stresses after cooling. The magnitude of
residual stresses is related to the yield stress of
the material at the temperature at which the
deformation occurred.
Thus, in the case of through hardening in
which a complete and homogeneous austenitic
microstructure turns into a martensitic one, a 4%
volume change of the machine part or a 1.3%
linear increase in the machine-part size is
obtained:
DV
100=4%
V
In incomplete hardening, the volume changes at

the surface are greater than in the core, which
produces transformation strains. The residualstress variation thus depends on:

Quenching/cooling conditions
Temperature difference between the surface
and the core during the quenching process
Temperature interval between the beginning
and end of martensitic transformation and
the cooling rate in this zone
An image of the magnitude of internal stresses

during the quenching process can be provided by
a test of quenching cylindrical specimens from a
temperature lower than the transformation temperature. During the quenching process, temperatures are measured at the specimens, that is,
at their surface and in their core. Maximum
thermal stresses can thus be calculated (Ref 1,
20, 71, 72). The maximum thermal stresses
found in the cylindrical specimens with the
smallest diameter, 25 mm, and further specimen
diameters in a geometrical ratio with a factor of 2
are given in Table 5.
As an approximate orientation to the cooling
rates obtained, a calculation was made for a
temperature of 500 C. Comparative data on
maximum thermal stresses refer to air cooling
and oil quenching. With air cooling, the stresses
obtained in the specimen with the smallest diameter, 25 mm, equal only 7 MPa, whereas for
the specimen with the largest diameter, 800 mm,
the stresses are as much as 200 MPa. The maximum thermal stresses, however, are considerably higher with oil quenching; they range
between 230 MPa with the 25 mm diameter
specimen and 620 MPa with the 800 mm diameter specimen. The latter represents an extremely high internal thermal stress that may result
in material plastification. The data exclude
stresses due to phase changes and stresses due to
inhomogeneity, which additionally increase
their value.
Results of the measured residual stresses with
surface-hardened steels are given in Table 6.
The stresses were measured by the x-ray diffraction technique just below the surface, that is,
0.05 mm. The results refer to steel grades
832M13, 805A20, 805A17, 897M39, 905M39,
and cold rolled steel after induction hardening,
with the longitudinal residual stresses being
measured. The values of longitudinal stresses
after carburizing to 1.0 to 1.5 mm case with
0.8% surface carbon and direct quenching
without tempering range between 190 and
pg 490
18/8/2008 4:04PM Plate # 0
400 MPa, which depends on the steel grade

selected. With steel 832M13, testing of specimens after quenching and by undercooling to
80 and 90 C with tempering and without
tempering was performed. Tempering contributes to obtaining tempered martensite and
possible transformation of retained austenite,
which produces a reduction, that is, a compensation, of the existing residual-stress profile.
Thus, the lowest residual stress is obtained in
steel quenching by undercooling and subsequent
tempering. In nitriding of steel 897M39 to a
depth of 0.5 mm, however, a residual stress
ranging between 400 and 600 MPa is obtained.
In steel 905M39, however, a considerably
higher residual stress, ranging between 800 and
Table 5 Cooling rate thermal stresses in simple
rounds with no transformation
Air cooled
Diameter
of round
(D), mm
25
50
100
200
300
400
800
Oil quenched
Cooling rate
at 500 C
(v500), C/s
Maximum
thermal
TH stress
smax , MPa
Cooling
rate at 500 C
(v500), C/s
Maximum
thermal
TH stress
smax , MPa
0.662
0.312
0.146
0.070
0.0445
0.0326
0.0158
7
15
28
54
73
100
200
20.0
6.12
1.88
0.59
0.29
...
...
230
290
370
450
510
(540)
(620)
Source: Ref 1, 71
Table 6 Residual stresses measured in surface

heat treated steels
Steel
832M13
805A20
805A20
805A17
805A17
897M39
905M39
Cold
rolled
steel
Heat treatment
Carburized at 970 C to 1.0 mm

case with 0.8% surface carbon
Direct quenched, 80 C subzero
treatment, no temper
Direct quenched, 90 C subzero
treatment, tempered
Carburized and quenched
Carburized to 1.11.5 mm case
at 920 C, direct oil quench,
no temper
Carburized to 1.11.5 mm case
at 920 C, direct oil quench,
tempered at 150 C
Nitrided to case depth of
approximately 0.5 mm
Induction hardened, untempered
Induction hardened, tempered
at 200 C
Induction hardened, tempered at
300 C
Induction hardened, tempered at
400 C
(a) Immediately subsurface, that is, 0.05 mm. Source: Ref 71
Residual stress
(longitudinal),
MPa
1000 MPa, is obtained in nitriding to the same

depth. The highest residual stresses were measured in cold-rolled steel subjected to additional
induction surface hardening without tempering.
A maximum value obtained was 1000 MPa.
With additional tempering at 200 C, martensite
will become tempered and the residual stresses
reduced to 650 MPa. By additional tempering at
an even higher temperature, at which martensite
disintegrates, the maximum measured stresses
decrease. For example, at a tempering temperature of 400 C, they decrease to 170 MPa or
a factor of 6 lower than those obtained immediately after steel quenching.
Figure 98 shows the results of carbonitriding
SAE 1118 steel with reference to the contents of
carbon and nitrogen at the surface and the gradient of the two in the subsurface (Ref 71). The
austenite content, that is, the austenite gradient
of the two in the subsurface, and the way a
complementary residual-stress variation proceeds in the subsurface are very important.
With a gradual reduction of austenite content
to a depth of 0.5 mm, the residual stresses
decrease as well, and the maximum pressure
achieved is 200 MPa. To a depth of 45 mm,
the stresses gradually increase to tensile values,
and at a depth of 1.5 mm, they are as much as
+100 MPa.
Input and Output Control of Steel for

Induction Surface Hardening of Gears
Production automation necessitates an everincreasing demand for greater uniformity of the
280
340
200
240340(a)
190230
400
150200
400600
8001000
1000
650
350
170
Fig. 98
Residual stress, carbon, nitrogen, and retained austenite through a carbonitrided case on SAE 1118
steel. Source: Ref 71
pg 491
18/8/2008 4:04PM Plate # 0
properties of heat treated steel parts (Ref 73).

Therefore, a general market demand is for a
certain prescribed quality that is usually the
result of various properties (Ref 41, 7482). One
of the most important properties is hardenability, which should be a uniform and constant property of steel regardless of the treatment
conditions in the ironworks. In the experimental
part of the study, the starting point was the
analysis of steel hardenability according to the
Jominy test (Ref 40, 7476). On this basis, the
conditions for induction hardening were prescribed for a worm gear tooth. The success of the
induction hardening was verified by the microhardness measurement along the tooth profile
centerline (Ref 83, 84).
Figure 99 shows Jominy curves for the hardness profiles of the specimen front of AISI 1050,
4150, and 4340 steels (Ref 40, 41). It generally
applies that Jominy curves for common testspecimen heating differ strongly from those for
the specimens induction heated and quenched in
the same way. Hardness differences in the
Jominy test specimen are minimal at the surface
and increase through the depth. It is important
that the hardness profiles of the furnace-heated
test specimens or short-term induction-heated
specimens differ less if carbon steels have a
smaller carbon content (AISI 1050) and differ
more strongly with alloyed steel, as was the case.
Figure 100 shows Jominy curves for furnace
heating and induction heating. Both figures
indicate that the through-height hardness profile
of a Jominy specimen depends on the maximum
temperature obtained at the surface (Fig. 100b)
and on heating time (Fig. 100a), the power
density being given (Ref 40, 41). Induction
heating was performed under the conditions
given and with different heating times. Furnace
heating of Jominy specimens at 870 C provides
the highest hardness values for the hardened
front.
In induction heating at the same temperature,
870 C, a quite different hardness profile of the
Jominy specimen hardened front is obtained
(Fig. 100a). With an increasing surface austenitizing temperature, the through-height hardness profile of the specimen increases so that at a
temperature of 1040 C, there is negligible difference in the hardness profile of the Jominy
specimen, regardless whether it was furnace
heated or induction heated.
The prescription for the verification of the
effects of induction hardening was carried out
for the upper- and lower-limit microhardness
from the tooth top and to a depth of 9/6 of the

tooth height. The experiments on heat treatable
carbon steel C35 with a different history have
Fig. 99
Jominy curves for end-quenched bars of (a) AISI

1050, (b) 4150, and (c) 4340 steels, austenitized
conventionally and by short-time induction heating. Source: Ref
40, 41
pg 492
18/8/2008 4:04PM Plate # 0
shown that with the existing heat treatment conditions, it is not possible to achieve the microhardness profile required by the user (Ref 79).
First, it was necessary to find a Slovenian
substitute for the French steel CC35. This was
done with the help of a cooperative industry.
Steel CC35 is a heat treatable steel that has been
mechanically machined and induction hardened
to ensure a uniform quality.
Steel C35, chosen as a substitute, is classified
among the high-grade, unalloyed, heat treatable
carbon steels with a maximum phosphorus and
sulfur content lower than 0.035%. It is intended
for parts smaller than 100 mm, subjected to
lower loads, and with high requirements on
homogeneity of material after heat treatment
and mechanical machining.
Effect of (a) time at an 870 C austenitizing temperature and (b) maximum surface temperature
on the Jominy curves for induction-hardened AISI 4150 steel.
The curve for conventional furnace-heated 4150 is also shown
in (b). Source: Ref 40, 41
Fig. 100
For the selected heat treatable carbon steel

C35, a suitable acceptance control procedure
had to be prescribed, with special emphasis on
hardenability.
Based on the corresponding criteria, the
acceptance conditions should enable the classification of steels into quality classes in accordance with the different heat treatment
prescriptions. The chemical composition of the
steel is presented according to International
Organization for Standardization (ISO) standards and that of the French according to the
Association Francaise de Normalisation
(AFNOR). The chemical composition is given
by quoting the upper and lower content limit for
each particular part. The limits of the Slovenian
steel are much more constrained than that of the
French. In the French steel, there is a higher
content of carbon, manganese, and traces of
chromium, which leads to somewhat better
hardenability compared to the Slovenian steel
C35.
Figure 101 presents the hardness curves of the
Jominy specimen versus the distance from the
face (Ref 83). The starting point for the analysis
was the upper and lower confidence limit for
Slovenian steel C35 as prescribed by the producer, Ravne Ironworks.
The given confidence limits are presented by
the hatched area on the extreme right of the
figure. Since testing enables a successful selection of steel as well as optimal heat treatment
conditions, it was decided to test the upper and
lower confidence limits of a small-sized, hardened worm gear (Fig. 102a) (Ref 83). This gear
must have hard, wear-resistant surfaces and an
increased strength due to loading conditions or
expected load-carrying capacity of the teeth.
Both of these properties are achieved by induction hardening. The success of the heat treatment
was verified by Vickers microhardness measurements at a loading of 0.3 daN acting along
the centerline of the tooth profile. The verification of the effects of surface hardening and
simultaneous heat treatment of the teeth was
carried out in ten measurements: the first on the
tooth top and then each subsequent measurement at a distance of 1/6 h or 0.367 mm. Thus,
in ten measurements the microhardness was
measured down to the depth of 9/6 of the tooth
height, as shown in Fig. 102(b) (Ref 83). When
ordering the gear, the customer specified
the prescribed allowable hardness for forming
the upper and lower confidence limits along the
height of the gear tooth.
pg 493
18/8/2008 4:04PM Plate # 0
The microhardness limit values were:

On the tooth top at h = 0, the microhardness

is 680 to 800 HV0.3.
In the depth of 4/6 h or 1.45 mm, the microhardness is 520 HV0.3 maximum.
In the depth of 8/6 h or 2.90 mm, the microhardness is 520 HV0.3 maximum.
In the depth of 9/6 h or 3.30 mm, the microhardness is 220 to 275 HV0.3.
On the basis of hardenability testing, the following conclusions can be drawn:

Due to a higher content of carbon, manganese, and some traces of chromium, the
French steel CC35 displays a very favorable

hardenability according to the Jominy tests.
Even to a depth of 7 mm, the hardenability
curve shows 40 HRC, then hardness begins to
decrease rapidly and, at a depth of 10 mm, is
only 22 HRC.
The Slovenian steel C35 produced by the
Ravne Ironworks has a slightly lower content of carbon and a considerably lower
content of manganese. For example, for steel
of charge B, the carbon content is lower by
0.06% and the manganese content by 0.35%,
according to the French steel. On the basis of
the data in the literature, it was determined
that this considerably lower manganese
Fig. 101
Hardenability of the analyzed steels and determination of the upper and lower confidence limit. Source: Ref 83
Fig. 102
Worm characteristics after induction surface hardening. Source: Ref 83
pg 494
18/8/2008 4:04PM Plate # 0
content reduces the effect of hardenability

by a factor of 1.5. The relative effect of
silicon on hardenability is practically negligible, although the content may vary within
0.1 to 0.4%.
The results of the hardenability testing of the
discussed steels are presented in Fig. 101 (Ref
83). Based on the mean values and their deviations, it was determined that the curve of hardenability or hardness shows solid progress into
the depth in the French steel CC35 and steel C35
of charge C and D. On the other side, the
hardenability of Slovenian steel C35 of charge
C and D is much lower, so that at a depth of
4 to 5 mm, the lower confidence limit is exceeded. The input control of steel should be focused
mainly on the lower confidence limit of hardenability, because the progress of hardness has a
decisive influence on the strength properties of
the gear teeth. It was determined that the upper
confidence limit corresponds to the maximum
hardness achieved, and that it cannot be exceeded in any way. It was therefore decided that
the place of the corrected lower confidence limit
would be defined using Jominy hardenability
tests, determining the actual progress of hardness. As a criterion for the correction of the
lower confidence limit, the microhardness along
the tooth height (9/6 h), limited by the maximum
and minimum microhardness, was used. This
helped to establish the primary criteria conditioning the microhardness distribution on the
Fig. 103
product, which was also limited by the upper and

lower confidence limit. The answer to which
charges of the steel are more suitable for use can
be obtained by associating the corresponding
microhardness confidence limits on the product
and also on the Jominy specimen. The desired
product characteristics were defined on the basis
of the following conditions of induction hardening:

Generator power22.5, 24.0 and 25.5 kW

Heating time3.9, 4.4, and 4.9 s
Time/pause between the end of heating and
the beginning of quenching0.1 s
Quenching time4.0 s
The results of induction hardening on the gear

are presented in Fig. 103 for the French steel
CC35 (Ref 83), and in Fig. 104 for the Slovenian
steel C35 at the given power values of a highfrequency generator and heating times (Ref 83).
The diagrams show the upper and the lower
confidence limits for microhardness along the
tooth height. The results of microhardness distribution along the tooth height must fall within
the mentioned limits. However, it can be seen
that it is possible to reach only a lower microhardness on the tooth that results from a lower
hardenability of the steel or from unsuitable heat
treatment conditions. The lower confidence
limit is defined only by the microhardness of the
basic material at a depth of 4/6 to 9/6 h and
represents only a theoretical limit. The data on
Hardness distribution along the tooth symmetry line after heat treatment of the French steel CC35. Source: Ref 83
pg 495
18/8/2008 4:04PM Plate # 0
the steel properties and heat treatment conditions provide various actual measured microhardness distributions, where the microhardness
of the basic material is reached in the depth from
h to 9/6 h in the French steel CC35 and from 11/12
h to h in the Slovenian steel C35. From all these
data, it can be stated that it is possible, with
adequate heating and quenching conditions and
adequate hardenability, to reach the desired
microhardness distribution in the tooth. From
Fig. 103, it can be noted that in the case of the
French steel CC35 only three out of nine heat
treatment conditions fall out of the confidence
limit at the minimum power P = 22.5 kW (Ref
83). Quite the contrary can be found from
Fig. 104 presenting Slovenian steel C35. Here,
a desirable distribution of microhardness
along the tooth height was achieved only in
the conditions of maximum power (Ref 83). This
means that for an equal efficiency of inductionhardened gears from the French steel CC35, a
significantly shorter heating time is necessary
than with steel C35. This time is estimated to be
for even a second shorter and represents a 25%
shorter heat treatment cycle, contributing to
lower costs of manufacturing.
Induction surface hardening of machine
components and especially gears is a very
complex process involving a whole range of
possible heat treatment methods, which are all
reflected in either good or bad serviceability of
machine components. The heat treatment engineer must be aware of the different effects of
particular design shapes of induction coils, be
Fig. 104
familiar with electromagnetic phenomena and

eddy currents, and have some experience in the
right choice of energy inputs necessary for
heating. The energy needed for heating can be
provided by changing the generator power as
well as the frequency of the current. In progressive hardening, to achieve a suitable energy
input, the workpiece feed rate or the rate at
which the coil is moved must be adjusted,
whereas in single-shot hardening, suitable
energy input is achieved by adjusting the heating
time with a high-frequency current. An essential
advantage of induction surface hardening is that
it is possible to achieve a sufficient repeatability
of the hardened layer thickness on the workpiece
as well as a desirable or even prescribed hardened-layer profile, ensuring sufficient hardness
and favorable distribution of residual stresses in
the hardened layer. A variety of steels and a
whole range of induction-hardening methods
provide the possibilities for very accurate planning of the size and distribution of residual
stresses. This is of growing importance, since
manufacturers are frequently required to produce machine components that, among other
surface properties, must possess quite specific
residual-stress distribution along the depth of the
hardened layer. It has become a proven fact that
high compressive stresses ensure high fatigue
strength of machine components and reduce the
danger of the occurrence and growth of cracks
on the surface of components. As far as induction surface hardening is concerned, it is also
quite important to choose the right quenching
Hardness distribution along the tooth symmetry line after heat treatment of the Slovenian steel C35. Source: Ref 83
pg 496
18/8/2008 4:04PM Plate # 0
medium and method of quenching. For this

reason, engineers must direct their attention not
only to the method of heating and possible
overheating of the surface layer but also to the
methods of quenching and the right choice of the
medium for quenching.
With increasing surface heating power in
the final phase, an increase in microhardness
along the tooth symmetry line is also reached,
especially in the depth from 0.5 to 2.5 mm,
whereas by making the heating time longer, a
more pronounced effect directed into the depth
can be achieved. In this way, an adequate heat
treatment can ensure the required surface hardness (52 HRC) as well as an increased strength of
the gear in the tooth-root part, which is of
extreme importance for certain operating conditions.
Tests have shown that it is necessary to move
the lower hardenability limit at least on the level
indicated by the scattered microhardness values
of the French steel CC35. The new lower confidence limit is named the corrected limit,
ensuring the prescribed microhardness distribution in the subsurface at a power P =
24.0 kW and P = 25.5 kW.
On the basis of the tests, it can be concluded
that the criterion of hardenability can be very
successfully applied in the input control of
steels. Hardness decreases with the distance
from the face of the Jominy specimen and must
fall within the confidence limits. In some cases,
that is, with Slovenian steel C35 of charge A
and B, when the hardenability exceeds the
lower confidence limit, an adequate selection of
induction-hardening conditions (power in kilowatts and heating time in seconds) can ensure
the desired microhardness along the tooth symmetry line. The results confirm that heat treatment conditions can be successfully determined
by relatively simple experiments that also make
the procedure more economical.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
REFERENCES
1. H.C. Child, Surface Hardening of Steel,

Engineering Design Guides 37, Published
for the Design Council, The British Standards Institution and the Council of Engineering Institutions, Oxford University
Press, Oxford, U.K., 1980
2. P.A. Hassell and N.V. Ross, Induction Heat
Treating of Steel, Heat Treating, Vol 4,
14.
15.

p 164202
R.E. Haimbaugh, Practical Induction Heat
Treating, ASM International, 2001
S. Lampman, Introduction to Surface
Hardening of Steels, Heat Treating, Vol 4,
p 264265
T. Ruglic, Flame Hardening of Steels, Heat
Treating, Vol 4, ASM Handbook, ASM
K. Sridhar and A.S. Khanna, Laser Surface
Heat Treatment, Chapter 3, Lasers in
Surface Engineering, N.B. Dahotre, Ed.,
Surface Engineering Series, Vol 1, ASM
J. Grum and R. Sturm, Characteristics of
Laser Surface Melt-Hardening and of
Optimizing the Process, Second Int. Conf.
on Quenching and the Control of Distortion
(Cleveland, OH), 1996, p 193200
J. Grum and R. Sturm, Properties of Laser
Surface Melt Hardening on Cast Iron,
Proceedings of the Sixth European Conference on Laser Treatment on Materials,
ECLAT96 (Stuttgart, Germany), 1996,
p 501508
J. Grum and P. Zerovnik, Residual Stresses
in Laser Heat Treatment of Plane Surface,
Proc. of the First Int. Conf. on Quenching
and Control of Distortion (Chicago, IL),
1992, p 333341
O.A. Sandven, Laser Surface Hardening,
D. Bialod, Ed., Electromagnetic Induction
and Electric Conduction in Industry, Centre
Francais de LElectricite, Diffusion: CFE
and TEC and DOC, Paris, 1997
S. Schiller, S. Panzer, and B. Furchheim,
Electron Beam Surface Hardening of Steels,
K.E. Thelning, Chapter 6.7: Flame Hardening, Steel and Its Heat Treatment, Bofors
Handbook, Butterworth, London and
Boston, 1975, p 451465
K.E. Thelning, Chapter 6.6.2: Steel Grades
for Induction Hardening, Steel and Its Heat
Treating, Bofors Handbook, Butterworth,
London and Boston, 1975, p 434435
J. Grum, Induction Hardening, Materials
Science and Technology Series, Vol 1,
Faculty of Mechanical Engineering,
Ljubljana, 2001
pg 497
18/8/2008 4:04PM Plate # 0
16. K.E. Thelning, Chapter 6.6: Induction

Hardening, Steel and Its Heat Treating,
Bofors Handbook, Butterworth, London
and Boston, 1975, p 432451
17. K.E. Thelning, Chapter 6.2: Quenching and
Tempering of Constructional Steels, Steel
and Its Heat Treating, Bofors Handbook,
Butterworth, London and Boston, 1975,
p 319325
18. P.G. Simpson, Induction HeatingCoil and
System Design, McGraw-Hill, New York,
Toronto, London, 1960
19. V. Rudnev, D. Loveless, R. Cook, and
M. Black, Handbook of Induction Heating,
Marcel Dekker, Inc., New York, Basel,
2003
20. J. Grum, Induction Hardening, Handbook of
Residual Stress and Deformation of Steel,
G.E. Totten, M.A.H. Howes, and T. Inoue,
21. V.I. Rudnev, R.L. Cook, D.L. Loveless, and
M.R. Black, Induction Heat Treatment,
Basic Principles, Computation, Coil Construction and Design Considerations,
Chapter 11A, Steel Heat Treatment Handbook, G.E. Totten and M.A.H. Howes,
Ed., Marcel Dekker, Inc., New York, 1997,
p 765767
22. V.I. Rudnev, R.L. Cook, D.L. Loveless, and
M.R. Black, Introduction Heat Treatment,
Modern Power Supplies, Load Matching,
Process Control, and Monitoring, Chapter
11B, Steel Heat Treatment Handbook,
G.E. Totten and M.A.H. Howes, Ed.,
Marcel Dekker, Inc., New York, 1997,
p 873874
23. G. Pfaffmann, Introduction to Induction
Heating, Fundamentals of Induction Heating, Society of Manufacturing Engineers,
Anaheim, CA, 1998
24. D.J. Williams, Quench Systems for Induction Hardening, Met. Heat Treat., 1995
25. K.E. Thelning, Chapter 6.6.3, Equipment
for Induction Hardening, Steel and Its
Heat Treatment, Bofors Handbook, Butterworth, London and Boston, 1975, p 435
437
26. V.K. Rudnev, D.L. Loveless, M.R. Black,
and P.J. Miller, Progress in Study of
Induction Surface Hardening of Carbon
Steels, Gray Irons and Ductile (Nodular)
Irons, Ind. Heat., March 1996
27. M.G. Lozinskii, Industrial Applications
of Induction Heating, Pergamon Press,
Oxford, 1969
28. R.S. Ruffini and V. Nemkov, Induction

Heating Systems Improvement by Application of Magnetic Flux Controllers, Proc.
of the Int. Induction Heating Seminar,
Padna, 1998, p 133139
29. C. Durban, D. Durand, and P. Chevre,
Determination of Austenitic Transformation During Fast Heat Treatment, 17th ASM
Heat Treating Society Conference Proceedings, Including the First International
Induction Heat Treating Symposium, D.L.
Milan, D.A. Poteet, G.D. Pfaffmann, V.
Rudnev, A. Muehlbauer, and W.A. Albert,
30. K.Z. Shepelyakovski and F. Bezmenov,
Advanced Steels for Modern Induction
Through and Surface Hardening, 17th ASM
Induction Heat Treating Symposium, D.L.
Milan, D.A. Poteet, G.D. Pfaffmann, V.
Rudnev, A. Muehlbauer, and W.A. Albert,
31. E. Hohne, Inductionsharten, Werkstattbucher fur Betriebsfachlente, Konstrukteure und Studierende, Heft 116,
Herausgeber, H. Haake, Ed., SpringerVerlag, Berlin, Gottingen, Heidelberg, 1955
(in German)
32. G. Benkowsky, Induktionserwarmung,
VEB Verlag Technik, Berlin, 1985 (in
German)
33. G.D. Pfaffmann, Selective Surface Treatment of Gears by Induction Profile Hardening, Mater. Sci. Forum, Vol 102104,
1992, p 345364
34. B. Crique, F. Rukcksuhl, O. Longeot,
C. Delalean, S. Plano, B. Ottosson, G.
Bonzano, and C. Pichard, Multi Frequency
Induction Hardening of Gears to Replace
CarburizingMethodology of Development, 17th ASM Heat Treating Society
Conference Proceedings, Including the
First International Induction Heat Treating
Symposium, D.L. Milan, D.A. Poteet, G.D.
Pfaffmann, V. Rudnev, A. Muehlbauer, and
W.A. Albert, Ed., ASM International, 1998,
p 887894
35. AGME Standard, American National Standard, Gear Materials and Heat Treatment
Manual, American Gear Manufacturers
Association, Virginia, 1989
36. G. Parrish and D.W. Ingham, The Submerged Induction Hardening of Gears, Heat
Treat. Met., (No. 2), 1998, p 4350
pg 498
18/8/2008 4:04PM Plate # 0
37. H. Sandinger, The Submerged Induction

Hardening of Gears, Heat Treat. Met.,
Vol 21 (No. 2), 1996, p 111123
38. W. Amende, Transformation Hardening of
Steel and Cast Iron with High-Power
Lasers, Chapter 3, Industrial Applications
of Lasers, H. Koebner, Ed., John Wiley &
Sons, Chicheste, 1984, p 7999
39. J. Meijer, R.B. Kuilboer, P.K. Kirner, and
M. Rund, Laser Beam Hardening: Transferability of Machining Parameters, Proceedings of the 26th International CIRP
Seminar on Manufacturing Systems
LANE94, Laser Assisted Net Shape Engineering, M. Geiger and F. Vollertsen, Ed.,
Erlangen Meisenbach-Verlag, Bamberg,
p 243252
40. J.F. Libsch, W.P. Chuang, and W.I.
Murphy, The Effect of Alloying Elements
on the Transformation Characteristics of
Induction-Heated Steels, Trans. ASM, Vol
42, 1950, p 121
41. C.R. Brooks, The Metallurgy of Induction
Surface Hardening, Adv. Mater. Process.,
2000, p H19H23
42. V. Rudnev and R. Cook, Magnetic Flux
Concentrators: Myths, Realities, and Profits, Met. Heat Treat., March/April 1995
43. W. Brunst, Die Inductive Warmebehandlung, Springer-Verlag, Berlin, 1957
44. J. Grum and D. Ferlan, Residual Internal
Stresses after Induction Hardening and
Grinding, 17th ASM Heat Treating Society
Conference Proceedings, Including the
First International Induction Heat Treating
p 629639
45. S.L. Semiatin and D.E. Stutz, Induction
Heat Treatment of Steel, 2nd printing,
American Society for Metals, 1987
46. N. Stevens, Induction Hardening and Tempering, Heat Treating, Vol 4, Metals
Metals, 1981, p 451483
47. S. Denis, M. Zandona, A. Mey, and S.A.
Boufoussi, Calculation of Internal Stresses
during Surface Heat Treatment of Steels,
Proceedings of European Conference on
Residual Stresses, V. Hauk, H.P. Hougardy,
E. Macherauch, and H.D. Tietz, Ed.,
(Frankfurt, Germany), 1992, DEM Informationsgesselschaft mbH, Oberursel, 1993,
p 10111020
48. U. Bruckner, W. Schuler, and H. Walter,

Untersuchungen zum Eigenspannungszubeim inductiven Randschnitharten, Eigenspannungen: Entstehung-MessungBewertung, Band 1, E. Macherauch and V. Hank,
Ed., Deutsche Gesselschaft fur Metallkunde
E.V., Oberursel, 1983, p 293308 (in
German)
49. M. Melander, Computer Calculations of Residual Stresses due to Induction Hardening,
Eigenspannungen
Entstehung-MessungBewertung, Band 1, E. Macherauch and V.
Hank Ed., Deutsche Gesselschaft fur Metallkunde E.V., Oberursel, 1983, p 309328
50. A.J. Fletcher, Thermal Stress and Strain
Generation in Heat Treatment, Chapter 9.2,
Induction Hardening, Elsevier Applied
Science, London and New York, 1989,
p 182187
51. G. Seulen and H. Voss, Oberflachenhartung
mil Induktionserhitzung bei mittleren Frequenzen, Stahl Eisen, Vol 63 (No. 51),
1943, p 929935 (in German)
52. H. Fujio, T. Aida, Y. Masumoto, and T.
Tsuruki, Paper 169-14, Distortion and Residual Stresses of Gears Caused by Hardening, Third Report, Induction Hardening of
Gears Made of 535 C Carbon Steel, Bull.
JSME, Vol 22 (No. 169), 1979, p 10011008
53. G.H. Ledl, Why Quenching Is Important
in Induction Heating, Met. Prog., 1967,
p 200204
54. J. Grum, How to Select Induction Surface
Hardening and Finished Grinding Conditions in Order to Ensure High Compressive
Residual Stresses on Machine Parts Surface,
Mater. Sci. Forum, Vol 426432, 2003,
p 25992604
55. G. Parrish, D.W. Ingham, and M. Chaney,
The Submerged Induction Hardening of
Gears, Part 1: Processing and Application
Aspects, Heat Treat. Met., Vol 25 (No. 1),
1998, p 18
56. G. Parrish, D.W. Ingham, and M. Chaney,
The Submerged Induction Hardening of
Gears, Part 2: Properties Imparted and
Problems Solved, Heat Treat. Met., Vol 25
(No. 2), 1998, p 4350
57. T. Inoue, H. Inoue, F. Ikuda, and T. Horino,
Simulation of Dual Frequency Induction
Hardening Process of a Gear Wheel, Proc.
of Third International Conf. on Quenching
and Control of Distortion, G.E. Totten,
B. Lisicic, and H.M. Tensi, Ed., ASM
pg 499
18/8/2008 4:04PM Plate # 0
58. M. Field, J.F. Kahles, and J.T. Cammet,

Review of Measuring Method for Surface
Integrity, Ann. CIRP, Vol 21 (No. 2), 1971,
p 219237
59. E. Brinksmeier, A Model for the Development of Residual Stresses in Grinding,
Advanced In-Surface Treatments, TechnologyApplicationsEffects, Vol 5, A.
Niku-Lari, Ed., Published in Cooperation
with the Institute for Industrial Technology
Transfer International, Pergamon Press,
Oxford, 1987, p 173189
60. R. Snoeys, M. Maris, and J. Peters, Thermally Induced Damage in Grinding, Key
Note Papers, Ann. CIRP, Vol 27 (No. 2),
1978, p 571581
61. P.G. Althaus, Residual Stresses in Internal
Grinding, Ind. Diamond Rev., (No. 3), 1985,
p 124127
62. M. Moris and R. Snoeys, Heat Affected
Zone in Grinding Operations, Proceedings of the 14th International Machine
Tool Design and Research Conference,
F. Koenigsberger and S.A. Tobias, Ed.,
Manchester, 1973, p 569669
63. J. Triemel, Grinding with Cubic Boron
Nitride, Proceedings of the 14th International Machine Tool Design and
Research Conference, F. Koenigsberger
and S.A. Tobias, Ed., Manchester, 1973,
p 671676
64. J. Grum, A Review of the Influence of
Grinding Conditions on Resulting Residual
Stresses after Induction Surface Hardening
and Grinding, J. Mater. Process. Technol.,
Vol 114 (No. 3), 2001, p 212226
65. J. Grum, How to Select Induction Surface
Hardening and Finished Grinding Conditions in Order to Ensure High Compressive
Residual Stresses on Machine Parts Surface,
Mater. Sci. Forum, 2002, p 404, 407, 623,
628
66. J. Grum, Measuring and Analysis of Residual Stresses after Induction Hardening and
Grinding, Mater. Sci. Forum, 2002, p 347,
349, 453458
67. F. Kosel and L. Kosec, Internal Stresses in a
Surface Hardened and Ground Steel, Mech.
Eng. J., Vol 31 (No. 910), 1985, p 225230
68. H.
Staudinger,
Induktive
Oberflachenharung und Eigenspannungen,
Untersuchungen uber Stahlauswahl und
geeignete Warmevorbehandlung, Hart.Tech. Mitt., Vol 2 (No. 21), 1966, p 111123
(in German)
69. V. Nemkov, Role of Computer Simulation

in Induction Heating Technique, Proc. of
the Int. Induction Heating Seminar (Padua),
1998, p 301308
70. P.K. Braisch, The Influence of Tempering
and Surface Conditions on the Fatigue
Behaviour of Surface Induction Hardened
Parts, Mater. Sci. Forum, Vol 102104,
1992, p 319334
71. H.C. Child, Residual Stress in Heat-Treated
Components, Heat Treat. Met., Vol 4, 1981,
p 8994
72. H.C. Child, The Heat Treatment of Tools
and DiesA Review of Present Status and
Future Trends, Tools and Dies for Industry
Book, The Metals Society, p 185, 291321
73. Electromagnetic Induction and Electric
Conduction in Industry, Chapter 9, Heat
Treatments by Induction, Centre Francais de
IElectricite, 1997
74. R. Jonsson, Toleranzen bei der Warmebehandlung von Stahlen in Harterein, TZ Met.
bearb., Vol 80 (No. 5), 1986 (in German)
75. K.E. Thelning, Chapter 4: Hardenability,
Steel and Its Heat Treatment, Bofors
Handbook, Butterworths, London and
Boston, 1975, p 127181
76. H. Arend and W. Neuhaus, Die Hartbarkeit
des Stahles, Verlag W. Girardet, Essen,
1955 (in German)
77. D.H. Breen and G.H. Walter, ComputerBased System Selects Optimum Cost
SteelI, Source Book on Materials Selection, Vol I, American Society for Metals,
1977, p 182185
78. D.H. Breen, G.H. Walter, and C.J. Keith,
Computer-Based System Selected Optimum Cost SteelsII, Source Book on
Materials Selection, Vol II, American
79. D.H. Breen, G.H. Walter, and C.J. Keith,
Computer-Based System Selected Optimum Cost SteelsIII, Source Book on
Materials Selection, Vol III, American
80. D.H. Breen, G.H. Walter, and J.T. Sponzilli,
Computer-Based System Selected Optimum Cost SteelsIV, Source Book on
Materials Selection, Vol IV, American
81. D.H. Breen, G.H. Walter, and J.T. Sponzilli,
Computer-Based System Selected Optimum Cost SteelsV, Source Book on
Materials Selection, Vol V, American
pg 500
18/8/2008 4:04PM Plate # 0
82. H. Matthes, Novel Process of Quality Control during Inductive Hardening Process,
17th ASM Heat Treating Society Conference Proceedings, Including the First
International Induction Heat Treating
p 727733
83. J. Grum, Input and Output Control of Steel
Intended for Surface Hardening, 17th ASM

Induction Heat Treating Symposium,
D.L. Milan, D.A. Poteet, G.D. Pfaffmann,
V. Rudnev, A. Muehlbauer, and W.A.
Albert, Ed., ASM International, 1989,
p 763768
84. P. Birk, Hartenscnutzmassen bei der Einstazhartung, Hart.-Tech. Mitt., Herausgegeben von Reibensahm, Vol 10 (No. 3),
1957, p 919 (in German)
pg 501
18/8/2008 4:05PM Plate # 0

DOI: 10.1361/faht2008p503
pg 503

Failure Analysis of Steel Welds

J.H. Devletian, Portland State University
D. Van Dyke, MEI-Charlton, Inc.
FAILURE ANALYSIS OF STEEL WELDS

may be divided into three categories. They include failures due to: design deficiencies, weldrelated defects usually found during inspection,
and failures in field service. Failures due to
design deficiencies usually result in a ductile
overload due to overstressing the component or
exposing the component to service conditions
well beyond expected performance requirements. Weld-related defects, which fall outside
the quality-acceptance criteria specified by the
applicable welding code, contribute to premature failures upon testing as well as in service.
Failures in the field may be insidious, because an
unanticipated fracture mechanism has been
causing problems without the designers knowledge. In this chapter, the failures due to various
discontinuities in the steel weldment are emphasized. These include poor workmanship, a
variety of hydrogen-assisted cracking (HAC)
failures, stress-corrosion cracking, fatigue, and
solidification cracking in steel welds.
butt welds are listed in Table 1 and illustrated in

Fig. 1. In this figure, discontinuities 1 through 7
are caused by poor workmanship. These include
(1) porosity, (2) inclusions, (3) lack of fusion,
(4) lack of penetration, (5) undercut, (6) underfill, and (7) overlap. Discontinuities 8, 9, and 10
are associated with the rolling operation in the
steel mill. These include (8) laminations, (9)
delaminations, and (10) seams and laps in the
steel plate.
The most insidious discontinuities are those
that cause brittle cracking due to metallurgical
origins, such as HAC, solidification cracking,
or stress-corrosion cracking. Such brittle failures are characterized by fracture surfaces
exhibiting intergranular, interdendritic, or cleavage modes of fracture, which occur at stress
levels well below the yield stress. Additionally, even failures by quasi-cleavage and/or
Table 1 Common types of weld discontinuities

illustrated in Fig. 1
No.
Discontinuities in Steel Welds
By far, the most common steel welding problems are associated with poor workmanship,
such as lack of fusion, lack of penetration, porosity, undercut, arc strikes, and others. The number and size of such discontinuities may be cause
for rejection by most codes and can act as stressconcentration sites to reduce weld joint strength
and promote cracking, such as fatigue failures. A
description of the wide variety of discontinuities
that can occur in steel weld metal and base metal
is provided by many welding codes, one of
which is the American Association of State
Highway Transportation Officials/American
Welding Society (AASHTO/AWS) D1.5 Bridge
Welding Code (Ref 1). From this code, rejectable discontinuities that can occur in common
2
3
4
5
6
7
8
9
10
11
12
Discontinuity
Porosity(b)
Cluster(1b in Fig. 1)
Piping(1d in Fig. 1)
Inclusions
Non-metallic slag(2b)
Lack of fusion(b)
Lack of joint penetration(b)
Undercut(b)
Underfill(b)
Overlap(b)
Laminations(b)
Delaminations(b)
Seams and laps(b)
Lamellar tears
Cracks(b)
a. Longitudinal(12a)
b. Transverse(12b)
c. Crater(12c)
d. Throat(12d)
e. Toe(12e)
f. Root(12f)
g. Underbead and HAZ(12g)
Location(a)
W
W
W
BM
W
W
BM
BM
BM
BM
W, HAZ, BM
W, HAZ, BM
W
W
HAZ
(a) W, weld; HAZ, heat-affected zone; BM, base metal. (b) See Fig. 1. Source:
Ref 1
18/8/2008 4:06PM Plate # 0
microvoid coalescence can occur at reduced

levels of ductility. Such cracking is illustrated in
discontinuities 11 and 12a through 12g in Fig. 1.
These types of brittle failures are especially
insidious, because the fracture occurs at stress
levels and ductility levels well below the design
parameters.
Two examples of brittle failure in steel welds
are hydrogen-assisted cracking and solidification cracking. Among the discontinuities illustrated in Fig. 1, HAC can commonly take the
form of lamellar tears, longitudinal cracks in the
heat-affected zone (HAZ) (12a), transverse
cracks (12b), toe cracks in the HAZ (12e), and
underbead cracks in the HAZ (12g). Solidification cracking can commonly take the form
of longitudinal cracks along the weld centerline
(12a), crater cracks (12c), and fissures and
microfissures in the weld metal.
Example 1: Porosity in Weld Metal. Twopass submerged arc welding was used to butt
weld two 19 mm (3/4 in.) thick A709-grade 250
plates. The joint design was a simple square
groove butt joint with 1.6 mm (1/16 in.) root
opening. It was designed to be welded in two
passes, with a single pass on each side of the
plate. After the first side was welded, the weld
bead exhibited an unusually large amount of
reinforcement. Subsequent magnetic particle
Fig. 1
inspection indicated extensive discontinuities in

the weld bead. The first pass of the weld was
sectioned, polished, and etched in 5% nital.
The macrograph of the welded plate is shown
in Fig. 2. Extensive rejectable porosity (in accordance with AWS D1.1 Structural Welding
Code) developed within the weld metal but was
not readily visible on the surface.
The welding system was inspected for possible sources of porosity. Inspection revealed that
the mating flanges were covered with thick mill
scale and were neither ground clean nor preheated prior to welding. The welding electrode
and flux combination was F7A2-EM12K. Normally, this flux/wire combination is designed to
be tolerant of light rust. However, in this case,
the porous mill scale was so thick that it acted as
a sink for moisture, oil, and other volatile contaminants.
Excessive porosity was caused by welding on
heavily rusted steel. Simply rough grinding the
faying surfaces to bare metal prior to welding
completely eliminated the porosity problem.
If rough grinding is not possible, the welding
engineer can choose a more active flux that is
designed for welding rusted plate. Active fluxes
contain strong deoxidizers, such as titanium
and aluminum, but should only be used in a
single pass.
Schematic illustration of common types of discontinuities in welds. Reproduced with permission of the American Welding
Society. See also Table 1. Source: Ref 1
pg 504
18/8/2008 4:06PM Plate # 0
Failure Analysis of Steel Welds / 505
Example 2: Lack of Penetration. A36

structural steel was butt welded by automatic
flux-cored arc welding. A double V-groove joint
preparation was used for a two-pass welding
operation on 6.4 mm (1/4 in.) thick plate. By
visual examination, the weld appeared satisfactory. Full penetration of the butt joint was
required. Transverse-to-weld tensile testing
revealed inadequate yield strength. The tensile
test failure occurred in the weld metal. Normally,
previous tensile tests always failed in the unaffected base metal.
Radiographic examination of the weld joint
and the mandatory transverse-to-weld tensile
test showed the weld to have inadequate penetration. Inspection of the fracture surface of the
tensile specimen revealed unfused metal at the
plate midthickness. The fractured steel exhibited
a ductile slant fracture. By measuring the approximate area of fused cross section and area of
unfused cross section, it was determined that
approximately 80% of the joint thickness carried
the load.
Even though the weld fracture was ductile, the
reduced cross-sectional area of fused metal (due
to lack of penetration) did not provide enough
sound metal to meet the required joint strength.
Welding parameters were changed to ensure full
penetration.
Example 3: Lack of Side-Wall Fusion. DH36 structural steel for shipbuilding was butt
welded by robotic gas metal arc welding. A
double V-groove joint preparation was used for a
two-pass welding operation. By visual examination, the weld appeared satisfactory. Dyepenetrant testing revealed a discontinuity along
one edge of the weld. Face bend testing revealed
Fig. 2
Wormhole or piping porosity in weld metal deposited

by submerged arc welding. Plate is 19 mm thick.
a rejectable discontinuity along the same edge of

the weld.
Visual inspection of the discontinuity in the
bend test showed that the base metal and weld
metal were not fused at the toe. Subsequent
metallographic examination showed that the
lack of side-wall fusion existed throughout the
length of the weld.
Clearly, the robot and fixturing were slightly
misaligned. The robot welder directed the
welding heat too far on one side of the joint and
not enough on the other side. The robot was reprogrammed to provide adequate weave to melt
both faying surfaces sufficiently to prevent a
recurrence of the problem.
Fatigue of Welded Joints

Fatigue cracking is a result of repetitive
fluctuating stress causing fracture well below the
yield strength of the steel. Factors required for
fatigue fracture include a sufficiently high tensile stress, a sufficiently large variation in applied stress, and a sufficient number of cycles of
applied stress. Cracks resulting from fatigue
account for more than half of all failures. Fatigue
fractures are insidious because there is no visible
plastic deformation preceding fracture. Fatigue
fractures are promoted by stress concentrators
such as notches, sharp fillets, corners, holes,
threads, splines, keyways, dents, gouges, laps,
folds, flakes, and delaminations in plates, sheets,
and forgings. Tensile residual stresses arise
from punched holes, heat treatment, welding,
and so on.
Visible beach marks indicate the location of
the fracture initiation site. Beach marks are
produced when oxidation at the crack tip
propagating under an alternating stress is interrupted from time to time, for example, constantspeed machinery that is turned on and off.
Also, visible beach marks are generated when
non-steady-state loading occurs or with rapid
changes in loading, for example, auto/truck
components and road signs.
Usually, fatigue cracks have multiple initiation sites, particularly in rotating shafts subject
to bending loads with stress concentrations.
Multiple initiation sites are easily recognized because of the presence of ratchet marks. A ratchet
mark ridge or ligament occurs at the merger of
two adjacent fatigue cracks that propagate on
different planes. When these two fatigue cracks
meet, they are connected by a ridge of deformed
pg 505
18/8/2008 4:06PM Plate # 0
metal called a ratchet mark. For example, three

adjacent propagating fatigue cracks will have
two ratchet marks connecting them.
When viewing a fatigue fracture surface at
high magnification of approximately 500 to
5000 under the scanning electron microscope
(SEM), striations are usually visible. In ductile
structural materials, crack propagation is by
microvoid coalescence, and the crack path becomes perpendicular to the tensile axis. The
correlation between striation spacing and tensile
load has been well established, where one striation represents one fatigue load cycle. Striation
spacing increases with increasing maximum
tensile loading. Secondary cracks sometimes
occur alongside striations, which better resolves
striation spacing.
For brittle materials, crack propagation occurs
in the cleavage or intergranular mode. Since
brittle cracks can grow unstably, the crack path
provides less clear correlation between striation
spacing and loading. Although striations are less
well defined, secondary cracks are helpful in
resolving some striations. An absence of observable striations is possible when metals with
complex metallurgical microstructures obscure
striations, or when the fatigue crack closes and
plastically deforms the striations. The latter case
usually produces a very shiny fracture surface.
Example 4: Fatigue Cracking of Welded
Pipe Flange. A short drain pipe from a heatrecovery steam generator failed in service by
developing leaks around a circumferential weld
near a flange. The pipe was a 25 mm (1 in.)
schedule 80 pipe (25 mm, or 1 in., internal diameter by 4.8 mm, or 3/16 in., wall thickness)
section 127 mm (5 in.) long with flanges welded
onto both ends. The pipe had been in service for
5 years with an operating pressure of 8.3 MPa
(1200 psi). Upon failure, the crack completely
separated the flange from the pipe.
The 127 mm (5 in.) section of pipe with
flanges was cut from the boiler and submitted for
examination, as shown in Fig. 3. The fracture
surface of the pipe is shown in Fig. 4. There are
clearly defined ratchet marks in both the top and
bottom edges of the fracture surface. Beach
marks are present on both the top and bottom
of the fracture surface that run toward the center, where there is a brightly colored band.
Beach marks are not present in the brightly
colored band.
Metallographic sectioning of the weld profile
showed that the crack initiated at the toe of the
weld and proceeded in between the weld filler
metal and HAZ. There was no evidence of weld

defects such as undercut or porosity present in
the section.
Scanning electron microscopy showed fine
striations within the beach mark areas of the
pipe. The SEM showed a ductile dimpling structure within the brightly colored band of the fracture surface. Energy-dispersive spectroscopy of
several areas of the fracture surface revealed
only iron, manganese, and oxygen.
The pipe fractured due to fatigue cracks that
initiated at several locations at the toe of the
weld. The fatigue cracks propagated until there
was insufficient cross section to support the
service loads. The morphology of the fracture
suggests the cracking propagated under reversed
bending stresses.
Hydrogen-Assisted Cracking Theory

In addition to discontinuities resulting from
poor workmanship, a very common cause of
Fig. 3
Overview of pipe section. Cracking is visible on right

end of the pipe at the toe of the weld. Courtesy of MEICharlton, Inc.
Fig. 4
Overview of pipe end. Ratchet marks and beach marks

are clearly visible. Courtesy of MEI-Charlton, Inc.
pg 506
18/8/2008 4:06PM Plate # 0
failure in welded steel plates, castings, and

structural components is HAC. Although HAC
of steel weldments has been studied for over
a half-century, the mechanism of cracking is
still uncertain. Fortunately, judicious hydrogen
management has permitted the crack-free
welding of high-strength steels primarily by
preheating the weld to a temperature that must
be increased with increasing:

Diffusible hydrogen content

Yield strength
Hardness
Carbon equivalent
Thickness
Restraint
Susceptible microstructure
Cracking can take place in either the weld

metal or the HAZ, depending on which is more
susceptible. Generally, the peak hardness,
strength, and carbon equivalent values of the
HAZ are greater than those of the weld metal.
So, cracking usually takes place in the HAZ.
This is the most common form of HAC and is
called underbead cracking. Because HAC is
dependent on time for hydrogen atoms to diffuse
at room temperature to crack nucleation sites, it
is also called delayed cracking. Such cracking
can take place anytime, from minutes to several
days after the weld has cooled. Several other
forms of hydrogen cracking are also possible
and are discussed.
After over a half-century of research on the
deleterious effects of hydrogen in steel, the
fundamental mechanism of HAC is still uncertain. Preheating, however, is the most reliable
method to prevent weld cracking. Unfortunately, preheating is labor-intensive and costly.
More recent studies have shown that preheating
temperatures can be either reduced or eliminated
if the composition of the steel base metal and
filler metal is low in crack-promoting elements.
(Ref 219). For example, an empirical carbon
equivalent formula (CEN), developed by Yurioka et al. (Ref 2), clearly reflects the fact that
alloying elements such as carbon and boron
produce high CEN values and have a strong
detrimental effect on hydrogen crack resistance:
CEN= C+ A( C) fSi=24+ Mn=6+ Cu=15
+ Ni=20+( Cr+ Mo+ Nb+ V)=5+ 5Bg
where A(C) = 0.75+0.25 tanh{20(C0.12)},

and all elements are in weight percent.
Thus, modern structural steel plate and filler

metals contain reduced carbon for increased
resistance to hydrogen cracking and, coincidently, for increased fracture toughness. Many
low-carbon steels, such as the U.S. Navys
HSLA-65, HSLA-80, and HSLA-100, are
designed to be welded without preheating (Ref
2022).
The presence of hydrogen in welds can cause
failure at levels of strength and plastic strain well
below design yield strength and ductility. The
current models of HAC involve pre-existing
defect sites in the weld metal. These initiation
sites include inclusions, minor phase particles,
microscopic cracks, and other discontinuities.
Many theories attempt to explain the brittleness
caused by the presence of as little as a few parts
per million (ppm) of hydrogen in steel. The
decohesion theory of Oriani (Ref 2329) states
that dissolved hydrogen migrates to regions of
triaxial tensile stress, and that the cohesive forces between atoms in the iron lattice are reduced
in proportion to the interstitial hydrogen concentration. This theory provides for the observed
increase in hydrogen solubility at the tip of a
crack (Ref 30). Diffusible hydrogen has the
additional effect of supplying internal mechanical stresses that produce an apparent softening
of the steel. Hydrogen also causes growth of
microvoids in undeformed specimens and the
initiation of microcracks in specimens deformed
to the critical strain (Ref 29). During cooling and
shrinkage of a weld, these defect sites become
localized regions of triaxial tensile stress concentration. Atomic hydrogen diffuses interstitially to these regions of the expanded lattice.
As the concentration of hydrogen increases, the
cohesive strength between iron atoms decreases
below the local intensified stress level, at which
point cracking occurs. A definitive reason why
the cohesive strength between iron atoms
decreases in the presence of hydrogen atoms
has not been fully explained. In 1960, Troiano
(Ref 31) suggested that certain sites act to concentrate stress to promote cracking, and that
hydrogen reduces the cohesive strength between
iron atoms.
In a classic 1972 paper (Ref 32), C.D. Beachem presented a model suggesting that the
presence of sufficiently concentrated hydrogen
dissolved in the lattice just ahead of the crack
tip aids whatever deformation processes the
microstructures will allow. Intergranular, quasicleavage, or microvoid coalescence fracture
modes are dependent on the microstructure, the
pg 507
18/8/2008 4:06PM Plate # 0
pg 508
crack-tip stress intensity, and the concentration

of hydrogen. That is, hydrogen reduces the stress
intensity needed for cracking regardless of the
mode of fracture, as shown in Fig. 5. Even the
stress intensity needed for HAC by microvoid
coalescence is reduced in the presence of hydrogen. Beachems model unifies several theories
but shows how the stress-sorption and lattice embrittlement models are unnecessary. The model
shows that the planar pressure effects are necessary at low stress intensities and are necessary
only to augment the driving force from the applied loads. The basic hydrogen-steel interaction
appears to be an easing of dislocation motion or
dislocation generation, or both.
In 2003, Albert et al. (Ref 33) found that
hydrogen traps increased with alloying elements. Measurements of diffusible hydrogen in
weld metal decreased in steel welds containing
increasing amounts of chromium and molybdenum alloying elements. The HAC in steel
is dependent on hydrogen trapping, which is
the interaction between diffusible hydrogen
and lattice defects. Several investigators have
examined hydrogen trapping in steel (Ref 34
40). These traps include point defects (vacancies, solute atoms), dislocations, interfaces and
surfaces (grain boundaries, particle-matrix interfaces, cracks, external surfaces), and volume
defects (voids, second-phase particles).
Gibala and DeMiglio (Ref 34) experimentally
measured binding energies (Table 2) and saturabilities of hydrogen traps in AISI 4340 steel,
which led them to predict probable susceptibilities to HAC. Both primary and secondary
crack paths were explained in terms of the types
of microstructural traps that predominate in a
given material. Management of hydrogen traps

can have a profound effect on the susceptibility
to HAC. In theory, decreasing the amount of
diffusible hydrogen by introducing permanent
or irreversible traps has the beneficial possibility
of reducing the susceptibility to HAC.
Olson et al. (Ref 3740) have shown that
effective hydrogen traps can include microscopic particles of Ce2O3, TiC, Y2O3 VC, and
NbC. Although these traps would decrease
the amount of diffusible hydrogen in the weld,
large quantities of such particles would have a
deleterious effect on ductility and fracture toughness. Steel can be embrittled by hydrogen if it is
stressed at temperatures high enough to allow
hydrogen to diffuse to potential embrittlement
sites, but also low enough that the hydrogen
is not depleted from embrittlement-producing
traps. With increasing temperature, steel is
expected to exhibit mechanical behavior determined largely by the nature of hydrogen
transport among the populations of various
hydrogen traps (Ref 34).
Increasing the tensile strain or residual tensile
stress also increases the ability of the steel to
retain hydrogen in body-centered cubic (bcc)
iron. Normally, the solubility (s) of hydrogen in
bcc iron is given by:
s=47:66 p1=2 exp (72:72 107=RT)
where s is in ppm by mass, R is the gas constant,

T is the absolute temperature, and p is the pressure in atmospheres. For example, at room
temperature and pressure, the solid solubility of
hydrogen in bcc iron is only 6.7 10 4 ppm.
Solubility is increased by the presence of traps.
For example, by increasing tensile strain, the
solid solubility of iron (Ref 30) increased by a
Table 2 Hydrogen trap interactions in steel

Hydrogen trap type
Fig. 5
Effect of hydrogen content on hydrogen-assisted

cracking (HAC) for microvoid coalescence (MVC),
quasi-cleavage (QC), and intergranular (IG) fracture modes.
Adapted from Beachem. Source: Ref 32
Interstitial solutes
Ti atom
Vacancy
Screw dislocation core
Mixed dislocation core
Hydrogen vapor/void
Grain boundary
Free surface
AlN interface
Fe3C interface
TiC interface
Source: Ref 34
Binding energy
(Ed), kJ/mol
315
26
46
2030
59
29
59
7095
65
84
96
18/8/2008 4:06PM Plate # 0
factor given by:

c=cn =exp(Asys 1=2 Bsys )
A={2(1+n)V=3RT}{2E=p}1=2
B={2(1+n)V=RT}{2=p}
where c is the hydrogen concentration at the

crack tip, cn is the hydrogen concentration in the
bulk steel, n is Poissons ratio, V is the partial
molar volume of hydrogen in iron, E is the
elastic modulus, and sys is the yield strength.
For a high-yield-strength steel having sys =
1000 MPa, the solubility of hydrogen in steel
increases by a c/cn factor of 6.5. The tip of a
crack is an excellent site for hydrogen accumulation, because the strain is greater than that
associated with sys.
Although the mechanism for HAC is still
debated, equations to prevent HAC have been
developed empirically. For example, the preheat
parameter (Pw) and the hydrogen accumulation
parameter (PHA), were developed from experimental work in Japan by Yurioka et al. (Ref 15,
41) as practical tools to empirically predict a
preheating temperature to prevent HAC:
Pw =Pcm +HD =60+R=40,000
where Pcm is the Ito-Bessyo (Ref 42) carbon

equivalent given as:
P cm =C+Si=30+(Mn+Cu+Cr)=20+Ni=60
+Mo=15+V=10+5B
where HD is the diffusible hydrogen content

in mL/100 g, elements are in weight percent,
and R is the restraint intensity in MPa. The HAC
can be avoided if the value of Pw50.3. In the
Pw equation, the effects of alloy composition,
hydrogen content, and restraint are taken into
account empirically.
Consumables manufacturers have recognized
that filler metals containing less carbon than the
steel plate would produce not only enhanced
weld metal toughness at required strength levels
but also greater resistance to HAC. In fact, virtually all of the older algorithms for determining preheating temperatures were based on the
composition of the base metal and not the weld
metal. Cracking of the HAZ by HAC was so
common in structural steels that it was assumed
that the coarse-grainced HAZ would always be

the location of maximum susceptibility to
cracking. Using the advantages of modern lowcarbon steels and consumables, Nippon Steel
(Ref 4) designed a series of commercial lowcarbon steels for line pipes that could be welded
while maintaining high strength and toughness
with equally low-carbon filler metals. These
pipeline steels included the X-65, X-70, and
X-80 grades, which contain very low carbon
(50.03%), high manganese for strength and to
control the bainite transformation, and 0.001%
boron to suppress the proeutectoid ferrite nucleation at austenite grain boundaries. It is desirable
to have a large amount of acicular ferrite in the
weld metal for optimal strength and toughness as
well as good resistance to HAC (Ref 43, 44).
Microconstituents detrimental to weld metal
toughness and possibly increased susceptibility
to HAC include grain-boundary ferrite, martensite, and side-plate ferrite, because these
structures provide a continuous path for cleavage crack propagation.
Types of Hydrogen-Assisted Cracking

Hydrogen-assisted cracking can appear in
four common forms:

Underbead or delayed cracking

Weld metal fisheyes
Ferrite vein cracking
Hydrogen-assisted reduced ductility
As mentioned earlier, the mechanism of HAC

is not clear, but management of hydrogen and
the prevention of HAC are well established.
Preheating the weld area prior to and during
welding provides the most reliable resistance to
HAC. There are many empirically-derived
methods to calculate preheat temperatures to
prevent HAC. All of the various types of HAC
can be avoided by good welding practice. The
forms of HAC are discussed in the following
sections.
Underbead or Delayed Cracking
By far, the most common form of HAC is
underbead or delayed cracking, schematically
illustrated as discontinuity 12g in Fig. 1 and
described in Table 1. Typically, this form of
cracking occurs in the coarse-grained HAZ up to
72 h after the weld has cooled. This is because
the HAZ typically has higher carbon content and
pg 509
18/8/2008 4:06PM Plate # 0
generally a higher carbon equivalent level than

the weld metal. Even though the source of
hydrogen is virtually always from the welding
consumable, atomic hydrogen rapidly diffuses
to crack nucleation sites in the HAZ. The diffusion of hydrogen explains the time-dependent
nature of HAC, thus the name delayed cracking.
The coarse-grained HAZ is the zone adjacent to
the weld and represents a region that has been
heated to nearly the melting temperature, followed by rapid cooling. Because of its higher
carbon content and large austenite grain size, the
coarse-grained HAZ develops a significantly
higher hardness than the weld metal. The HAZ
will transform to martensite upon cooling if the
carbon equivalent is high enough. Since the
harder HAZ is more susceptible to HAC than
either the weld metal or the unaffected base
metal, the cracking in a butt weld is typically
confined to a narrow strip of metal immediately
adjacent to the weld bead. Cracking in fillet
welds occurs at the toe of the weld because that
is the location of highest stress concentration.
Toe cracking is schematically illustrated as
discontinuity 12e in Fig. 1 and described in
Table 1.
Example 5: Underbead Cracking. Crosscountry line pipe is welded continuously for
long distances. At regular intervals, a flange
needs to be welded onto the pipe for coupling to
a valve or other device. Recently, a section of
pipe was removed from service because of
cracking that had occurred in the toes of the fillet
welds joining the flange to the pipe. The pipe
was 203 mm (8 in.) outside diameter by 6.4 mm
(0.25 in.) wall thickness, and the flange was
205 mm (8.1 in.) inside diameter by 305 mm
(12 in.) outside diameter by 18 mm (0.71 in.)
thick. Since the pipe was only 6.4 mm thick, the
weld was not preheated. Cracks measuring
approximately 10 cm long developed at the toes
of the fillet welds on the flange side, as shown in
Fig. 6.
From the illustration in Fig. 6, the cracking
occurred only in the toes of the two fillet welds
on the flange side. No cracking was observed at
the toes of the two fillet welds on the pipe side. A
metallographic section of the crack, shown in
Fig. 7, clearly reveals that the crack was confined to the brittle martensitic HAZ on the flange
side. Scanning electron microscopy of the
cracked area clearly shows an intergranular
mode of fracture (Fig. 8). Chemical analysis of
the pipe and flange in Table 3 revealed that an
incorrect steel was used for the flange. The
Fig. 6
Underbead cracking at the toe of the fillet weld on the

flange
Fig. 7
Toe cracking on the flange side of the flange-to-pipe

fillet weld, showing the weld metal, heat-affected
zone, and unaffected base metal. Cracking occurred in the martensitic (white) heat-affected zone of the flange.
Fig. 8
Fracture surface of flange failure in the as-received

condition. Intergranular fracture is shown as well as
debris retained from the field.
pg 510
18/8/2008 4:06PM Plate # 0
correct flange steel was supposed to be a lowcarbon steel. Instead, the flange was a 0.8%
(high) carbon steel, which was very susceptible
to HAC in the HAZ when welded without preheating.
Cracking in the HAZ on the flange side of the
fillet weld was due to the mistaken use of a high
0.8% C steel instead of the specified low-carbon
steel. Quality control measures need to be followed to prevent mixed steels from being used.
Weld Metal HAC and Fisheyes
Welding of modern low-carbon steels often
results in HAZs with greater resistance to HAC.
Thus, the weld metal composition is now as
susceptible as the HAZ. If the weld metal contains sufficient diffusible hydrogen content, has
high yield strength, and is in a highly stressed
condition, the susceptibility of such weld metal
to HAC is very possible. For example, in
the line-pipe industry, new thermomechanicalcontrolled processing steels achieve high yield
strength through thermal processing in the rolling mill, so that the carbon content and carbon
equivalent levels for a given yield strength have
dropped substantially. With this reduction in
carbon equivalent, the susceptibility to HAZ
cracking has also declined significantly. Since
the as-deposited weld metal achieves strength
primarily through alloying, the weld metal is
now very susceptible to HAC. Often, field
welding of X-65 and X-70 line pipe is performed
with high-hydrogen E8010G cellulosic electrodes. In this case, the weld metal yield strength
is greater than both the HAZ and unaffected base
metal. Thus, the weld metal has become the
weak link and is most susceptible to HAC.
The strong influence of hydrogen on weld
metal cracking can be observed in tensile testing
and bend testing as well as in failures of welds
subject to slowly applied tensile stress. Fisheyes
occur typically on the fracture surface of steel
all-weld-metal tensile specimens that fail due to

HAC. In tensile testing, fisheyes reduce the weld
metal ductility measurements, such as percent
elongation and percent reduction of area. Fisheyes are local areas within the weld that are more
hardenable due to solute banding, cellular, or
dendritic segregation of alloying elements (Ref
45). These initiation areas may also be richer in
localized hydrogen due to their proximity to
hydrogen traps such as inclusions. Since these
alloy-rich segregated areas are more susceptible
to brittle HAC, small, localized brittle-fracture
zones appear visually on a tensile test fracture
surface as bright round spots surrounded by gray
ductile fracture. The bright round spot may
consist of a local region of typically intergranular or possibly cleavage fracture surrounded by ductile dimpled failure. Both
intergranular and cleavage failures are brittle
fracture modes that appear much more brightly
than the surrounding material, which is ductile
dimpled and gray-appearing.
Example 6: Fisheyes on Fracture Surface. A high-strength steel, HSLA-100, was
butt-welded with a matching-strength filler
metal using gas metal arc welding (GMAW) and
argon-5%CO2 shielding gas at a heat input of
1.1 kJ/mm (28 kJ/in.) without preheating. The
filler metal contained Fe-0.03%C-1.4%Mn3%Ni-0.7%Mo. Because of the very low carbon
content, the weld metal hardness did not exceed
24 HRC. Tensile test results showed inadequate
ductility of only 8% elongation.
Upon examining the fracture surface of the
tensile specimen, multiple fisheyes were
observed, as shown in Fig. 9. An SEM image of
Table 3 Chemical analysis of the flange and pipe

Chemical
element
Flange
Pipe
C
Mn
Si
Ni
Cr
S
P
Al
Nb
0.80
0.67
0.25
0.01
0.23
0.13
0.018
...
...
0.07
1.10
0.24
0.01
0.01
0.005
0.017
0.038
0.02
Fig. 9
Brittle fisheyes appear as bright spots in a gray ductile

matrix.
pg 511
18/8/2008 4:06PM Plate # 0
the center of a bright fisheye exhibited brittle

intergranular fracture, as shown in Fig. 10. The
SEM images outside of the fisheyes showed
ductile dimpled microvoid coalescence. Welds
were repeated to determine the amount of diffusible hydrogen in similar welds in accordance
with AWS A4.3 (a standard welding test for
diffusible hydrogen). Despite the fact that
GMAW is known as a very low-hydrogen process, the value obtained for diffusible hydrogen
was 9 mL/100 g, which was far greater than
expected. It was found that the filler metal
manufacturer used excess hydrocarbon-base
lubricant during the wire-drawing operation
because of the high strength of the wire.
Fisheyes were caused by excessive amounts
of diffusible hydrogen in the weld metal due to
the lubricant residue on the filler metal.
Ferrite Vein Cracking
A very unexpected form of HAC is ferrite
vein cracking, which can occur in slowly cooled
electroslag welds. In recent studies of electroslag welding of 50 and 75 mm thick low-carbon
steel at Portland State University (Ref 46),
ferrite vein cracking of A709-grade 245 steel
occurred only in welds that were made with flux
and/or filler metal known to have high moisture
content. Although the mechanism is not certain,
diffusible hydrogen causes the ferrite at prioraustenite grain boundaries to crack under the
residual tensile stress produced by contraction
during weld cooling. This is very unusual,
because typical HAC is associated with hard
martensitic microstructures. Ferrite was always
thought to be immune to HAC because of its low
Fig. 10
SEM image of center of fisheye showing intergranular

fracture
strength and low hardness. It was also found that

nickel alloying additions tended to promote
HAC in the form of ferrite vein cracking, while
an equivalent amount of molybdenum resisted
cracking. Both nickel and molybdenum are
essential alloying elements for enhancing fracture toughness in both the weld metal and base
metal. The mechanism by which nickel and
molybdenum appear to have virtually opposite
effects on susceptibility to HAC is not known.
Example 7: Ferrite Vein Cracking in
High-Heat-Input Welds. Single-pass fullpenetration electroslag welds were deposited on
50 mm (2 in.) thick ASTM A588 steel using a
heat input of 42 kJ/mm (1070 kJ/in.) for bridge
applications. The ASHTO/AWS D1.5 Bridge
Welding Code required both radiographic and
ultrasonic testing (UT) of the completed welds.
The UT revealed possible indications of cracking around the weld center. The weld metal was
sectioned for metallographic examination, and
ferrite vein cracking was found, as shown in
Fig. 11.
Clearly, cracking was confined to the grainboundary ferrite, which was nucleated at the
prior-austenitic grain boundaries.
Fig. 11
grade 50W
Ferrite vein crack occurring in the prior-austenite

grain boundaries of weld metal deposited on A709-
pg 512
18/8/2008 4:06PM Plate # 0
Review of the welding procedure revealed

that the flux and tubular wire used in this process
were not baked prior to welding. The unbaked
flux was a source of moisture. Despite its extremely high weld heat input and the use of mild
steel electrodes, electroslag weld metal has been
shown to be susceptible to HAC. The fabricated
metal-cored tubular wire was also unbaked.
Under the welding heat, moisture in the flux
and filler metal produces atomic oxygen and
hydrogen. Since electroslag welds are large
single-pass deposits, the weld center is under
substantial tensile stress. The combination of
ample diffusible hydrogen and high-shrinkage
tensile residual stress at the weld center provides
the necessary ingredients for ferrite vein cracking. Subsequent welds were made with flux that
was baked to 204 C (400 F) and a new metalcored wire that was baked at an elevated temperature prior to shipment. The resulting welds
have since been free of ferrite vein cracking.
Although the mechanism of HAC of grainboundary ferrite is unknown, elimination of
cracking was achieved by reducing the sources
of moisture or hydrogen.
exhibited an intergranular mode of fracture.

Scanning electron microscopy at 50 revealed
that the crack propagated intergranularly along
the prior-austenite grain boundaries, as shown in
Fig. 13. However, at approximately 4000 , the
intergranular fracture surface exhibited shallow
dimples, as shown in Fig. 14. This dimpled
intergranular mode of fracture is due to the
weakness of the ferrite envelopes surrounding
each prior-austenite grain, as shown in Fig. 15.
The presence of diffusible hydrogen caused a
reduction of ductility of grain-boundary ferrite
sufficient to fail the bend test. The Beachem
diagram in Fig. 5 shows that microvoid coalescence can also be adversely affected by diffusible hydrogen. Using low-hydrogen practices,
such as baking the flux prior to use, eliminated
the cracking problem during bend testing.
Stress-Corrosion Cracking of Steel

Stress-corrosion cracking of steels is possible
when the steel is subject to both adequate tensile
Hydrogen-Assisted Reduced Ductility

This form of HAC occurs when the damage
due to diffusible hydrogen is not sufficient to
cause cracking in the weldment but is sufficient
to cause reduced ductility in subsequent tensile
and bend tests. This is a clear illustration of the
principles reported by Beachem (Ref 32) and
shown schematically in Fig. 5. All fracture
modes become more severe with increasing
diffusible hydrogen. Even the ductility associated with ductile microvoid coalescence is
substantially reduced in the presence of diffusible hydrogen.
Example 8: Failure to Pass Bend Tests due
to Hydrogen. Multipass submerged arc welds
deposited on 50 mm thick A588 steel were
subject to inspection in accordance with the
AWS D1.1 Structural Welding Code. The following tests were performed for the procedure
qualification welds: tensile testing, bend testing,
Charpy V-notch impact toughness testing, as
well as both ultrasonic and radiographic testing.
All of these tests were passed successfully except the guided bend test. As shown in Fig. 12,
the side-bend specimen cracked well before the
prescribed bend radius could be achieved.
Visual inspection and low-magnification optical microscopy of the cracked bend specimen
Fig. 12
Side-bend test failure of weld
Fig. 13
Scanning electron micrograph of fracture surface of

bend failure
pg 513
18/8/2008 4:06PM Plate # 0
stress (but below the yield strength) and an

aggressive environment. The absence of one of
these elements will eliminate stress-corrosion
cracking. Welds are particularly good sites for
stress-corrosion cracking, because substantial
tensile residual stresses are always present. The
shrinkage stresses associated with the solidification and cooling of welds produce near-yield
tensile residual stresses in and around the weld.
Relatively mild chemical environments can
activate the stress-corrosion cracking process.
Environments known to cause stress-corrosion
cracking of plain carbon and alloy steels include
liquefied ammonia, hydrogen sulfide, molybdenum disulfide, sodium hydroxide sour gas, high
pH values, nitrate solutions, and many other
corrosive environments. Stress-corrosion cracking of carbon steel can even take place in pure
water under high temperature and pressure.
Fig. 14
Higher-magnification image of fracture surface in

Fig. 13 showing dimpled intergranular fracture
Fig. 15
Optical microscopy of grain structure of electroslag

weld metal. Original magnification: 50
Although many theories for stress-corrosion

cracking have been suggested, only two appear
to be the basis for such cracking. These are the
stress-sorption theory and the electrochemical
theory.
The stress-sorption theory states that damaging substances in the environment are chemically absorbed onto the surface of the steel,
causing a reduction in the cohesive bonding
force between iron atoms. Only an atom-thick
surface layer is needed to seriously affect the
bonding forces between surface atoms. There
is a threshold stress necessary to initiate stresscorrosion cracking. In some ways, this mechanism resembles HAC in steels (discussed earlier),
where only a few ppm of hydrogen in concert
with tensile stress are needed to reduce the
cohesive bonds between iron atoms and cause
cracking. In stress-corrosion studies, acoustic
emission sensors have been attached to the
cracking sample to monitor the propagation of
the crack. Strong acoustic emissions were emitted and recorded each time the advancing crack
jumped or burst. Acoustic emission sensors used
to monitor HAC of steel displayed a similar
jump or burst behavior.
The electrochemical theory involves the
setting up of galvanic cells within the microstructure of the steel. Anodic dissolution paths
are produced along concentration gradients in
the metal or in grain boundaries. When the grain
boundaries are anodic to the bulk of the metal,
tensile stresses (although below yield) are
necessary to continue the cracking process in
order to open up dissolved pathways for further
penetration by the corrosive environment. As
evidence of the electrochemical nature of this
cracking process, stress-corrosion cracking can
be stopped by applying cathodic protection. As
soon as cathodic protection is removed, stresscorrosion cracking continues.
Example 9: Stress-Corrosion Cracking of a
Weld. After 30 years in service, a low-pressure
steam supply line developed a reoccurring
cracking problem in a circumferential butt weld.
The weld was on a 25 cm (10 in.) supply line
that carried 0.4 MPa (55 psig), 205 C (400 F)
steam to a paper machine. This line was 30 m
(100 ft) downstream from a spray attemperator
that cooled higher-temperature steam by spraying boiler feedwater into the line.
When cracks were initially discovered, they
were ground out and rewelded. The repair welds
reportedly cracked after only a few days in
service.
pg 514
18/8/2008 4:06PM Plate # 0
A visual examination revealed that neither the

original welds nor the repair welds penetrated
through to the inside of the pipe. The cracking
consisted primarily of circumferential cracks
down the center of the weld. In addition, there
were a few locations where short longitudinal
cracks emanated from the primary circumferential crack. Macroscopically, the cracks were
generally irregular and branching.
Sections of the cracked region were prepared
for metallographic examination of the weld
geometry and crack morphology, as shown in
Fig. 16. A microhardness scan of the base metal
revealed a hardness of 79 HRB, with a hardness
of 87 to 94 HRB in the original weld. The repair
weld was significantly harder, with a hardness of
22 to 34 HRC.
Metallographic investigations revealed an
intergranular branching morphology, as shown
in Fig. 17. Scanning electron microscopy also
showed the presence of several secondary
branching cracks emanating from the primary
crack. These cracks also proceeded in an intergranular fashion. Energy-dispersive spectroscopy did not reveal the presence of foreign
materials at the fracture surface in a measurable
quantity.
The pipe fractured due to stress-corrosion
cracking (SCC) precipitated by the presence of
geometric stress concentrations and high residual tensile stresses in the weld. The SCC is an
environmentally induced cracking mechanism
that can occur in a susceptible material in the
presence of tensile stress and an aggressive
chemical agent. Although the particular agent involved could not be identified, this is not unusual,
because small amounts of a caustic agent can
often cause cracking in the proper conditions.
Due to the location of the cracking, it is likely
that the agent entered the line in the spray
Fig. 16.
Cross section of weld at butt joint. Etchant: 2% nital.

Courtesy of MEI-Charlton, Inc.
attemperator process. It is likely that an upset in

the chemistry of the boiler feedwater at some
time in the past contaminated the downstream
lines and led to the SCC in this instance.
The consequences of additional steam line
failures need to be evaluated. Given the nature of
the cracking, inspection methods capable of
detecting the cracks in early stages are limited. It
may be most cost-effective to replace the steam
lines downstream of the attemperator and reevaluate the methods used for ensuring the
proper chemistry of the boiler feedwater.
Future repairs should ensure complete crack
removal and full penetration welds using proper
preheat and interpass temperatures to minimize
hardness gradients.
Solidification Cracking of Steel

Solidification cracking is one of several forms
of hot cracking. Solidification cracking in steel
and steel alloys occurs near the end of the solidification process and is caused by two dominant
factors: tensile stress acting on the weld during
solidification, and a large temperature range
between the solidus and liquidus temperatures or
the presence of low-melting impurities such as
sulfer and phosphorus. The tensile stress acting
on the weld can arise from either shrinkage
tensile stresses produced during solidification
and cooldown, or externally applied tensile
stress or tensile restraint stress. The effect of the
liquidus-to-solidus temperature range has been
Fig. 17
Micrograph of the crack near the weld root. Original

magnification: 100 . Etchant: 2% nital. Courtesy of
MEI-Charlton, Inc.
pg 515
18/8/2008 4:06PM Plate # 0
dealt with generally by several empirical solidification cracking equations. For example,
Matsuda (Ref 47) developed a very popular
parameter for solidification cracking of steels
called Lt, where increasing Lt increased susceptibility to cracking:
Lt =70(C Si=12 Mn=9+3P+4S+Ni=23
+Cr=35+Mo=70)
where all elements are in weight percent.

Clearly the effects of carbon and alloying
elements such as manganese, molybdenum,
chromium, and nickel were assumed to be linear,
as shown in the Lt equation. As the Lt equation
suggests, decreasing carbon content has always
been assumed to decrease solidification cracking
susceptibility in steel weld metal. Prediction
formulas for solidification cracking show
the effect of carbon on cracking to be linear.
Fig. 18
Maximum crack length (MCL) as a function of carbon

content in iron weld metal obtained in transvarestraint tests at 4% augmented strain. Source: Ref 53, 54
Fig. 19
Solidification cracking in weld metal
However, recent research has shown that the

effect of carbon on solidification cracking of
low-carbon steels is far more complex and
nonlinear than predicted by the Lt formula.
For example, Masumoto (Ref 48) showed that
solidification cracking was enhanced for carbon
contents 40.1%. Conversely, Ohshita et al. (Ref
41) reported that cracking was enhanced for
carbon 50.1% and that nickel additions were
beneficial in reducing the cracking effect of
carbon, Karjalainen et al. (Ref 49) surveyed the
technical literature and reported that there was a
least-susceptible range of carbon contents
between 0.1 and 0.17%. Within this range, the
cracking susceptibility was minimized. Ichikawa et al. (Ref 50) reported peak solidification
cracking susceptibility at 0.035% C, followed by
enhanced cracking when the carbon content
exceeded 0.1%. Most recently, Kim et al. (Ref
51) and Won et al. (Ref 52) showed a peak
in solidification cracking susceptibility at
approximately 0.10% C. It was apparent from
the literature that the effect of carbon on solidification cracking susceptibility of steel weld
metal required further study.
In the most recent work by Shankar and
Devletian (Ref 53, 54), the effect of carbon on
solidification cracking was nonlinear, with a
peak in cracking susceptibility at 0.1% C, as
shown in Fig. 18. In this study, testing was
performed on high-purity iron-carbon alloy
castings using the varestraint and transvarestraint tests. Maximum crack distance and
maximum crack length were measured at a 4%
augmented strain.
pg 516
18/8/2008 4:06PM Plate # 0
The data (schematically plotted in Fig. 18)

clarify some of the conflicting issues in the
literature. Solidification cracking susceptibility
is controlled by the brittle temperature range
(BTR), the d?c transformation stress, and the
iron-carbon peritectic reaction. According to
Shankar and Devletian (Ref 53, 54), there are
four distinct %C ranges that produced characteristic solidification behavior. These include:

Region 1: less than ~0.09% C (maximum

solid solubility of carbon in d-iron)
Region 2: ~0.09 to 0.11% C (maximum
solidification cracking)
Region 3: ~0.11 to 0.16% C
Region 4: greater than 0.16% C (iron-carbon
peritectic point)
In region 1, there was no cracking below

0.01% C in transvarestraint tests. This is because
the solidus/liquidus temperature range was
negligible. However, as the carbon content increased, the cracking susceptibility increased
rapidly up to approximately 0.06% C. The
cracking dropped slightly at 0.075% and then
continued to increase with carbon content up to
0.09%, due to the increasing solidus/liquidus
temperature range. In region 2, a large peak in
solidification cracking was observed. This critical cracking peak, centered at approximately
0.1% C, was found to be due to the simultaneous
action of three factors: the maximum solidus/
liquidus temperature range, the d?c transformation stresses, and the occurrence of the
BTR. At 0.1% C, cracking occurred with a
minimum critical strain and low fracture stress.
In region 3, solidification cracking decreased
with increasing carbon content because of the

decreasing solidus/liquidus temperature range.
In region 4, the solidification cracking susceptibility increased due to the increasing solidus/
liquidus temperature range.
Example 10: Solidification Cracking of
Steel Weld. Welds were deposited by fluxcored arc welding on 12 mm thick AISI/SAE
1020 steel plate at high travel speeds for maximum cost-effectiveness. Weld joints were
highly restrained during welding to prevent
distortion. Visible longitudinal centerline cracks
were observed, as shown in Fig. 19. The portions
of the weld seam that were not visibly cracked
failed the root bend test.
The cracked specimen was broken open and
observed under the SEM. Rejected bend-test
specimens were also broken open for examination by SEM. The crack surface clearly showed
that the mode of fracture was solidification
cracking, as shown in Fig. 20. Spectrographic
analysis of the weld metal admixture revealed a
carbon content of 0.1%. Subsequent welds were
deposited with reduced welding travel speed in
order to reduce the length of the teardrop shape
of the weld puddle. Subsequent welds deposited
at the reduced welding speed were crack-free.
Centerline cracking failure was caused by
solidification cracking. Decreasing welding
speed reduced susceptibility to solidification
cracking in the weld metal. Reducing restraint
during welding and reducing the weld metal
carbon content (or Matsudas Lt factor, mentioned previously) would have also decreased
the occurrence of solidification cracking.
REFERENCES
Fig. 20
Scanning electron micrograph of fracture surface

showing solidification cracking
1. AASHTO/AWS D1.5 Bridge Welding

Code, American Welding Society
2. N. Yurioka, M. Okumura, T. Kasuya, and
S. Ohshita, Welding Note, 3rd ed., Nippon
Steel, Japan, 1985
3. N. Yurioka, Weldability of Modern High
Strength Steels, Nippon Steel Corporation,
Japan, 1990
4. K. Shinada, Y. Horrii, and N. Yurioka,
Development of Weld Metal with High
Toughness and Low Hardenability, Nippon
Steel Corporation, 1989
5. J.M. Sawhill, J.C. Baker, and P. Howe,
Hydrogen-Assisted Cracking in High
Strength Pipeline Steels, Weld. J., Vol 65
(No. 7), 1986, p 175s183s
pg 517
18/8/2008 4:06PM Plate # 0
6. N. Yurioka and T. Kasuya, A Chart Method

to Determine Necessary Preheat in Welding,
Weld. World, Vol 35 (No. 5), 1995, p 3138
7. N. Yurioka and T. Kasuya, A Chart
Method to Determine Necessary Preheat in
Welding, IIW Doc. II-1230-94 and IIW
Doc. IX-1740-94, International Institute of
Welding, 1994, p 12
8. T. Kasuya and N. Yurioka, Determination
of Necessary Preheat Temperature to Avoid
Cold Cracking under Varying Ambient
Temperature, ISIJ Int., Vol 35 (No. 10),
1995, p 11831189
9. M. McParlan and B.A. Graville, Hydrogen
Cracking in Weld Metals, Weld. J., Vol 55
(No. 4), 1976, p 95s102s
10. J. Brozda, A Comparison Between the
Levels of Preheat Necessary to Prevent
Cold Cracking during Welding of Low
Alloy High Strength Steels, Weld. Int.,
Vol 11 (No. 2), 1997, p 2532
11. B.A. Graville, Interpretive Report on
Weldability Tests for Hydrogen Cracking of
Higher Strength Steels and their Potential
for Standardization, Bulletin 400, Welding
Research Council, New York, 1995
12. T. Terasaki, G.T. Hall, and P.L. Harrison,
Predictive Equation for Cooling Time of
CTS Test Welds, Trans. Jpn. Weld. Soc.,
Vol 21 (No. 2), Oct 1990, p 5155
13. N. Yurioka, H. Suzuki, H. Okumura, S.
Ohshita, and S. Saito, Carbon Equivalents
to Assess Cold Cracking Sensitivity and
Hardness of Steel Welds, Nippon Steel
Technical Report 20, Dec 1982, p 6173
14. F.R. Coe, Welding Steels without Hydrogen
Cracking, The Welding Institute, Abington,
U.K., 1973
15. N. Yurioka, H. Suzuki, S. Ohshita and
S. Saito, Determination of Necessary Preheat Temperature in Steel Welding, Weld.
J., Vol 62 (No. 6), 1983, p 147s153s
16. C.L.M. Cottrell, An Improved Predictive
Method for Avoiding Hydrogen Cracking,
Weld. Met. Fabr., April 1990, p 178183
17. N. Yurioka, Studies on Delayed Cracking
in Steel Welds (Report 1); Prevention
of Root Cracking in Root-Pass Welds,
Nippon Steel Corporation, Dec 1981
in Steel Welds (Report 2); Prevention of
Cracking in Restraint Multi-Run Welds,
Nippon Steel Corporation, Dec 1981
in Steel Welds (Report 3); Prevention of
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Micro- and Transverse Cracking in Weld

Metals, Nippon Steel Corporation, Oct
1983
R.J. Wong and R.D. Hayes, Arc Welding
Consumables for HSLA Steels with Yield
Strengths of 80 ksi and Above, The Metallurgy of Welding and Qualification of
Microalloyed Steel Weldments, J.T. Hickey
et al., Ed., American Welding Society,
1990, p 450489
P.W. Holsberg and R.J. Wong, Welding of
HSLA-100 Steel for Naval Applications,
Weldability of Materials, R.A. Patterson
and K.W. Mohin, Ed., ASM International,
1990, p 219239
J.J. DeLoach, C. Null, S. Fiore, and
P. Konkol, The Right Welding Wire Could
Help the U.S. Navy Save Millions, Weld. J.,
Vol 78 (No. 6), June 1999, p 5558
R.A. Oriani and P.H. Josephic, Equilibrium
Aspects of Hydrogen-Induced Cracking of
Steels, Acta Metall., Vol 22 (No. 9), 1974,
p 10651074
R.A. Oriani, Effects of Hydrogen on the
Plastic Response of Steels, Hydrogen
Effects in Metals, I.M. Bernstein and A.W.
Thompson, Ed., Metallurgical Society of
AIME, 1981, p 235245
R.A. Oriani, Hardening and Softening
Induced by Hydrogen in Carbon Steels,
NATO Conf. Series, Mater. Sci., Vol 5,
1983, p 795798
J.K. Lin and R.A. Oriani, Effect of Hydrogen on the Initiation of Shear Localization
in Plain Carbon Steels, Perspectives in
Hydrogen in Metals, Pergamon Press, 1986,
p 615621
R.A. Oriani, HydrogenThe Versatile
Embrittler, Corrosion, Vol 43 (No. 7), July
1987, p 390397
R.A. Oriani, Effects of Hydrogen on the
Plastic Properties of Medium Carbon Steels,
Metall. Trans. A, Vol 11 (No. 11), 1980,
p 18091820
J.K. Lin and R.A. Oriani, Effect of Hydrogen on the Initiation of Shear Localization
in Plain Carbon Steels, Acta Metall., Vol 31
(No. 7), 1983, p 10711077
H.A. Wriedt and R.A. Oriani, Effect of
Tensile and Compressive Elastic Stress
on Equilibrium Hydrogen Solubility in a
Solid, Acta Metall., Vol 18 (No. 7), 1970,
p 753760
A.R. Troiano, The Role of Hydrogen
and Other Interstitials in the Mechanical
pg 518
18/8/2008 4:06PM Plate # 0
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
Behavior of Metals, Trans. ASM, Vol 52,

1960, p 5480
C.C. Beachem, New Model for HydrogenAssisted Cracking (Hydrogen Embrittlement), Metall. Trans., Vol 3 (No. 2), Feb
1972, p 437451
S.K. Albert, V. Ramasubbu, N. Parvathavarthini, and T.P.S. Gill, Influence of
Alloying on Hydrogen-Assisted Cracking
and Diffusible Hydrogen Content in Cr-Mo
Steel Welds, Sadhana, Vol 28 (No. 3, 4),
June/Aug 2003, p 383393
R. Gibala and D.S. DeMiglio, Hydrogen in
Steels: Interactions, Traps and Crack Paths,
Hydrogen Effects in Metals, I.M. Bernstein
and A.W. Thompson, Ed., Metallurgical
Society of AIME, 1981, p 113122
G.M. Pressouyre, Hydrogen Traps, Repellers, and Obstacles in Steel: Consequence
on Hydrogen Diffusion, Solubility, and
Embrittlement, Metall. Trans. A, Vol 14
(No. 10), 1983, p 21892193
G.M. Pressouyre and I.M. Bernstein,
Example of the Effect of Hydrogen Trapping on Hydrogen Embrittlement, Metall.
Trans. A, Vol 12 (No. 5), 1981, p 835844
C.A. Lensing, Y.D. Park, I.S. Maroef, and
D.L. Olson, Yttrium Hydrogen Trapping to
Mangage Hydrogen in HSLA Welds, Weld.
J., Vol 83 (No. 9), 2004, p 254s266s
I.S. Maroef and D.L. Olson, Evaluation of
Hydrogen Trapping for Hydrogen Management in Steel Welding, Proc. of Sixth
International Trends in Welding Research,
D.L. Olson, I. Maroef, C. Lensing, D.
Smith, T. Wildeman, and M. Eberhart,
Hydrogen Management in Steel Weldments,
J.L. Davidson and D.L. Olson, Ed., Organizing Committee of the Joint Seminar on
behalf of Defense Science and Technology
Assoc. and Welding Technology Institute of
Australia, 1996, p 119
I. Maroef, D.L. Olson, M. Eberhart, and
G.R. Edwards, Hydrogen Trapping in Ferritic Steel Weld Metal, Int. Mater. Rev.,
Vol 47 (No. 4), 2002, p 191223
S. Ohshita, N. Yurioka, N. Mori, and
T. Kimura, Prevention of Solidification
Cracking in Very Low Carbon Steel, Weld.
J., Vol 62 (No. 5), p 129s136s
Y. Ito and K. Bessyo, Determination of
Preheating Conditions to Avoid Weld
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Cracking in Steel Construction: Report 1,

J. Jpn. Weld. Soc., Vol 40 (No. 11), 1971,
p 11171127
O. Grong and D.K. Matlock, Microstructural Development in Mild and Low
Alloy Steel Weld Metals, Int. Met. Rev.,
Vol 31, 1986, p 2748
P.L. Harrison and R.A. Farrar, Acicular
Ferrite in C-Mn Weld Metal, J. Mater. Sci.,
Vol 22 (No. 11), 1987, p 38123820
G.L.F. Powell, I.H. Brown, V.M. Linton,
and J.L. Davidson, Cold Cracking and
Segregation in Multi-Pass Welds of Quenched and Tempered Steel, Trends in Welding Research, S.A. David et al., Ed., ASM
W. Wood, J.H. Devletian, and R. Turpin,
NGI-ESW Consumables Development,
Final Report, FHWA Contract DTFH61-970030, Aug 2003
F. Matsuda, Hot Cracking Susceptibility of
Weld Metal, First U.S.-Japan Symposium
on Weldability of Materials, AWS, JWS,
and JWES, 1990, p 1935
I. Masumoto, Solidification, Blow Hole Formation and Cracking of Steel Weld Metal,
Weld. World, Vol 31 (No. 3), p 211221
L.P. Karjalainen, V.P. Kujanpaa, and N.
Suutala, Hot Cracking in Iron Base Alloys
Effect of Solidification Mode, Advances in Welding Science and Technology, American Society for Metals, 1986,
p 145149
K. Ichikawa, H.K.D.H. Bhadeshia, and
D.J.C. MacKay, Model for Solidification
Cracking in Low Alloy Steel Weld Metal,
Sci. Technol. Weld. Joining, Vol 1 (No. 1),
p 4350
J.G. Kim, K.T. Rie, and J. Ruge, Contribution to Hot Cracking Susceptibility of Submerged Arc Steel Weldments, Steel Res,
Vol 56 (No. 12), p 639644
Y.M. Won, H.N. Han, T.J. Yeo, and K.H.
Oh, Analysis of Solidification Cracking
Using the Specific Crack Susceptibility,
ISIJ Int., Vol 40 (No. 2), p 129136
J.H. Devletian, Final Report, Award
0096290, National Science Foundation,
2004
V. Shankar and J.H. Devletian, Solidification Cracking in Low Alloy Steel Welds,
Sci. Technol. Weld. Joining, Vol 10 (No. 2),
2005, p 236243
pg 519
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_521-523.pdf/Appendix_1/
18/8/2008 4:24PM Plate # 0

DOI: 10.1361/faht2008p521
pg 521

APPENDIX 1
Metric Conversion Guide

THIS APPENDIX is intended as a guide for
expressing weights and measures in the Systeme
International dUnites (SI). The purpose of SI
units, developed and maintained by the General
Conference of Weights and Measures, is to
provide a basis for worldwide standardization of
units and measure (Table A1.1). For more information on metric conversions (Table A1.2), the
reader should consult the following references:
Standard for Metric Practice, E380,

Annual Book of ASTM Standards, American
Society for Testing and Materials, 1916
Race Street, Philadelphia, PA 19103
Metric Practice, ANSI/IEEE 268
1982, American National Standards
Institute, 1430 Broadway, New York, NY

10018
The International System of Units, SP 330,
1986, National Institute of Standards and
Technology. Order from Superintendent of
Documents, U.S. Government Printing
Office, Washington, DC 20402-9325
Metric Editorial Guide, 4th ed. (revised),
1985, American National Metric Council,
1010 Vermont Avenue NW, Suite 1000,
Washington, DC 20054960
ASME Orientation and Guide for Use of SI
(Metric) Units, ASME Guide SI 1, 9th ed.,
1982, The American Society of Mechanical
Engineers, 345 East 47th Street, New York,
NY 10017
Table A1.1 Base, supplementary, and derived SI units

Measure
Unit
Symbol
Base units
Amount of substance
Electric current
Length
Luminous intensity
Mass
Thermodynamic
temperature
Time
mole
ampere
meter
candela
kilogram
kelvin
mol
A
m
cd
kg
K
second
radian
steradian
rad
sr
gray
meter per second squared
becquerel
Gy
m/s2
Bq
radian per second squared

radian per second
square meter
farad
mole per cubic meter
rad/s2
rad/s
m2
F
mol/m3
siemens
ampere per square meter
kilogram per cubic meter
S
A/m2
kg/m3
Supplementary units
Plane angle
Solid angle
Derived units
Absorbed dose
Acceleration
Activity (of
radionuclides)
Angular acceleration
Angular velocity
Area
Capacitance
Concentration (of amount
of substance)
Conductance
Current density
Density, mass
Measure
Electric charge density

Electric field strength
Electric flux density
Electric potential,
potential difference,
electromotive force
Electric resistance
Energy, work, quantity
of heat
Energy density
Entropy
Force
Frequency
Heat capacity
Heat flux density
llluminance
Inductance
Irradiance
Luminance
Luminous flux
Magnetic field strength
Magnetic flux
Magnetic flux density
Molar energy
Molar entropy
Molar heat capacity
Moment of force
Permeability
Permittivity
(continued)
Unit
Symbol
coulomb per cubic meter

volt per meter
coulomb per square meter
volt
C/m3
V/m
C/m2
V
ohm
joule
V
J
joule per cubic meter

joule per kelvin
newton
hertz
joule per kelvin
watt per square meter
lux
henry
candela per square meter
lumen
ampere per meter
weber
tesla
joul per mole
joule per mole kelvin
joule per mole kelvin
newton meter
henry per meter
farad per meter
J/m3
J/K
N
Hz
J/K
W/m2
lx
H
W/m2
cd/m2
lm
A/m
Wb
T
J/mol
J/mol K
J/mol K
Nm
H/m
F/m
18/8/2008 4:24PM Plate # 0
pg 522
Table A1.1 (continued)

Measure
Power, radiant flux

Pressure, stress
Quantity of electricity,
electric charge
Radiance
Radiant intensity
Specific heat capacity
Specific energy
Unit
Symbol
watt
pascal
coulomb
W
Pa
C

steradian
watt per steradian
joule per kilogram kelvin
joule per kilogram
W/m2 sr
W/sr
J/kg K
J/kg
Measure
Specific entropy
Specific volume
Surface tension
Thermal conductivity
Velocity
Viscosity, dynamic
Viscosity, kinematic
Volume
Wavenumber
Unit
Symbol
joule per kilogram kelvin

cubic meter per kilogram
newton per meter
watt per meter kelvin
meter per second
pascal second
square meter per second
cubic meter
1 per meter
J/kg K
m3/kg
N/m
W/m K
m/s
Pa s
m2/s
m3
l/m
Table A1.2 Conversion factors

To convert from
to
multiply by
Angle
degree
rad
1.745 329 E02
mm2
cm2
m2
m2
6.451
6.451
6.451
9.290
600 E+02
600 E+00
600 E04
304 E02
Nm
Nm
Nm
Nm
1.129
1.355
9.806
7.061
848 E01
818 E+00
650 E+00
552 E03
Bending moment or torque per unit length

lbf in./in.
lbf ft/in.
N m/m
N m/m
4.448.222 E+00
5.337 866 E+01
A/cm2
A/mm2
A/m2
1.550 003 E01

1.550 003 E03
1.076 400 E01
Current density
A/in.2
A/in.2
A/ft2
T
mWb
S
A/m
Vm
mV m
1.000
1.000
1.000
7.957
1.000
1.662
000 E04
000 E02
000 E+00
700 E+01
000 E02
426 E 03
Energy (impact, other)

ft lbf
Btu (thermochemical)
cal (thermochemical)
kW h
Wh
J
J
J
J
J
1.355
1.054
4.184
3.600
3.600
818 E+00
350 E+03
000 E+00
000 E+06
000 E+03
L/min
L/min
L/min
L/min
4.719
2.831
6.309
3.785
475 E01
000E+01
020 E+02
412 E+00
N
N
kN
N
4.448
4.448
8.896
9.806
222 E+00
222 E+03
443 E+00
650 E+00
Flow rate
ft3/h
ft3/min
gal/h
gal/min
N/m
N/m
Fracture toughness
p
ksi in:
MPa
1.459 390 E+01

1.751 268 E+02
p
m
1.098 800 E+00
Heat content
kJ/kg
kJ/kg
2.326 000 E+00

4.186 800 E+00
J/in.
kJ/in.
J/m
kJ/m
3.937 008 E+01

3.937 008 E+01
Length
A
min.
mil
in.
in.
ft
yd
mile
nm
mm
mm
mm
cm
m
m
km
1.000
2.540
2.540
2.540
2.540
3.048
9.144
1.609
oz
lb
ton (short, 2000 lb)
ton (short, 2000 lb)
ton (long, 2240 lb)
kg
kg
kg
kg 103(a)
kg
2.834 952 E02

4.535.924 E01
9.071 847 E+02
9.071 847 E01
1.016 047 E+03
Mass per unit area

oz/in.2
oz/ft2
oz/yd2
lb/ft2
kg/m2
kg/m2
kg/m2
kg/m2
4.395
3.051
4.882
4.882
kg/m
kg/m
1.488 164 E+00

1.785 797 E+01
kg/s
kg/s
kg/s
1.259 979 E04

7.559 873 E03
4.535 924 E01
Heat input
000 E01
000 E02
000 E+01
000 E+01
000 E+00
000 E01
000 E01
300 E+00
000 E+01
517 E01
428 E+00
428 E+00
Mass per unit length

lb/ft
lb/in.
Mass per unit time
Force
lbf
kip (100 lbf )
tonf
kgf
multiply by
Mass
Electricity and magnetism

gauss
maxwell
mho
Oersted
V cm
V circular-mil/ft
lbf/ft
lbf/in.
Btu/lb
cal/g
Bending moment or torque

lbf in.
lbf ft
kgf m
ozf in.
to
Force per unit length
Area
in.2
in.2
in.2
ft2
To convert from
lb/h
lb/min
lb/s
Mass per unit volume (includes density)

g/cm3
lb/ft3
lb/ft3
lb/in.3
lb/in.3
(continued)
(a) kg 103 = 1 metric ton or 1 megagram (Mg)
kg/m3
g/cm3
kg/m3
g/cm3
kg/m3
1.00 000 E+03

1.601 846 E02
2.767 846 E+01
2.767 846 E+04
2.767 990 E+04
18/8/2008 4:25PM Plate # 0
pg 523
Appendix 1: Metric Conversion Guide / 523

To convert from
to
multiply by
Btu/s
Btu/min
Btu/h
erg/s
ft lbf/s
ft lbf/min
ft lbf/h
hp (550 ft lbf/s)
hp (electric)
kW
kW
W
W
W
W
W
kW
kW
1.055
1.758
2.928
1.000
1.355
2.259
3.766
7.456
7.460
056 E+00
426 E02
751 E01
000 E07
818 E+00
697 E02
161 E04
999 E01
000 E01
to
multiply by
W/in.
W/m
1.550 003 E+03
Press capacity
See Force
Pressure (fluid)
atm (standard)
bar
in. Hg (32 F)
in. Hg (60 F)
lb/in.2 (psi)
torr (mm Hg, 0 C)
Pa
Pa
Pa
Pa
Pa
Pa
1.013
1.000
3.386
3.376
6.894
1.333
250 E+05
000 E+05
380 E+03
850 E+03
757 E+03
220 E+02
J/kg K
J/kg K
4.186 800 E+03

4.186 800 E+03
Specific heat
Btu/lb F
cal/g C
tonf/in. (tsi)
kgf/mm2
ksi
lbf/in.2 (psi)
MN/m2
5.192
1.730
1.442
4.184
m/m k
m/m k
1.000 000 E+00

1.800 000 E+00
m/s
m/s
m/s
m/s
m/s
mm/h
8.466
5.080
3.048
2.540
2.777
1.609
rad/s
rad/s
1.047 164 E01

6.283 185 E+00
pa s
m2/s
m2/s
mm2/s
1.000
1.000
9.290
6.451
000 E01
000 E04
304 E02
600 E+02
m3
m3
m3
m3
1.638
2.831
2.957
3.785
706 E05
685 E02
353 E05
412 E03
204 E+02
735 E+00
279 E01
000 E+02
Thermal expansion
in./in. C
in./in. F
ft/h
ft/min
ft/s
in./s
km/h
mph
667 E05
000 E03
000 E01
000 E02
778 E01
344 E+00
Velocity of rotation
rev/min (rpm)
rev/s
Viscosity
poise
stokes
ft2/s
in.2/s
951 E+01
650 E+00
757 E+00
757 E03
000 E+00
in.3
ft3
fluid oz
gal (U.S. liquid)
MPa
MPa
MPa
MPa
MPa
1.378
9.806
6.894
6.894
1.000
C
K
5/9 ( F32)
5/9
ft3/min
ft3/s
in.3/min
m3/s
m3/s
m3/s
4.719 474 E04

2.831 685 E02
2.731 177 E07
C
5/9
Wavelength
nm
1.000 000 E01
Temperature
F
R
W/m K
W/m K
W/m K
W/m K
Volume
Stress (force per unit area)

2
Btu in./s ft2 F

Btu/ft h F
Btu in./h ft2 F
cal/cm s C
Velocity
Power density
Temperature interval
F
To convert from
Power
(a) kg 103 = 1 metric ton or 1 megagram (Mg)
Volume per unit time
18/8/2008 4:26PM Plate # 0

DOI: 10.1361/faht2008p525
pg 525

APPENDIX 2
Temperature Conversion Table

Temperature Conversions
The general argument of this conversion table was devised by Sauveur and Boylston. The middle columns of numbers (in boldface type)
contain the temperature readings ( F or C) to be converted. When converting from degrees Fahrenheit to degrees Celsius, read the
Celsius equivalent in the column headed C. When converting from Celsius to Fahrenheit, read the Fahrenheit equivalent in the column
headed F.
F
...
...
...
...
...
458
456
454
452
450
272.22
271.11
270.00
268.89
267.78
...
...
...
...
...
368
366
364
362
360
222.22
221.11
220.00
218.89
217.73
...
...
...
457.6
454.0
278
276
274
272
270
172.22
171.11
170.00
168.89
167.78
306.4
302.8
299.2
295.6
292.0
188
186
184
182
180
122.22
121.11
120.00
118.89
117.78
...
...
...
...
...
448
446
444
442
440
266.67
265.56
264.44
263.33
262.22
...
...
...
...
...
358
356
354
352
350
216.67
215.56
214.44
213.33
212.22
450.4
446.8
443.2
439.6
436.0
268
266
264
262
260
166.67
165.56
164.44
163.33
162.22
288.4
284.8
281.2
277.6
274.0
178
176
174
172
170
116.67
115.56
114.44
113.33
112.22
...
...
...
...
...
433
436
434
432
430
261.11
260.00
258.89
257.78
256.67
...
...
...
...
...
348
346
344
342
340
211.11
210.00
208.89
207.78
206.67
432.4
428.8
425.2
421.6
418.0
258
256
254
252
250
161.11
160.00
158.89
157.78
156.67
270.4
266.8
263.2
259.6
256.0
168
166
164
162
160
111.11
110.00
108.89
107.78
106.67
...
...
...
...
...
428
426
424
422
420
255.56
254.44
253.33
252.22
251.11
...
...
...
...
...
338
336
334
332
330
205.56
204.44
203.33
202.22
201.11
414.4
410.8
407.2
403.6
400.0
248
246
244
242
240
155.56
154.44
153.33
152.22
151.11
252.4
248.8
245.2
241.6
238.0
158
156
154
152
150
105.56
104.44
103.33
102.22
101.11
...
...
...
...
...
418
416
414
412
410
250.00
248.89
247.78
246.67
245.56
...
...
...
...
...
328
326
324
322
320
200.00
198.89
197.78
196.67
195.56
396.4
392.8
389.2
385.6
382.0
238
236
234
232
230
150.00
148.89
147.78
146.67
145.56
234.4
230.8
227.2
223.6
220.0
148
146
144
142
140
100.00
98.89
97.78
96.67
95.56
...
...
...
...
...
408
406
404
402
400
244.44
243.33
242.22
241.11
240.00
...
...
...
...
...
318
316
314
312
310
194.44
193.33
192.22
191.11
190.00
378.4
374.8
371.2
367.6
364.0
228
226
224
222
220
144.44
143.33
142.22
141.11
140.00
216.4
212.8
209.2
205.6
202.0
138
136
134
132
130
94.44
93.33
92.22
91.11
90.00
...
...
...
...
...
398
396
394
392
390
238.89
237.78
236.67
235.56
234.44
...
...
...
...
...
308
306
304
302
300
188.89
187.78
186.67
185.56
184.44
360.4
356.8
353.2
349.6
346.0
218
216
214
212
210
138.89
137.78
136.67
135.56
134.44
198.4
194.8
191.2
187.6
184.0
128
126
124
122
120
88.89
87.78
86.67
85.56
84.44
...
...
...
...
...
388
386
384
382
380
233.33
232.22
231.11
230.00
228.89
...
...
...
...
...
298
296
294
292
290
183.33
182.22
181.11
180.00
178.89
342.4
338.8
335.2
331.6
328.0
208
206
204
202
200
133.33
132.22
131.11
130.00
128.89
180.4
176.8
173.2
169.6
166.0
113
116
114
112
110
83.33
82.22
81.11
80.00
78.89
...
...
...
...
...
378
376
374
372
370
227.78
226.67
225.56
224.44
223.33
...
...
...
...
...
288
286
284
282
280
177.78
176.67
175.56
174.44
173.33
324.4
320.8
317.2
313.6
310.0
198
196
194
192
190
127.78
126.67
125.56
124.44
123.33
162.4
158.8
155.2
151.6
148.0
108
106
104
102
100
77.78
76.67
75.56
74.44
73.33
(continued)
18/8/2008 4:27PM Plate # 0
pg 526
144.4
140.8
137.2
133.6
130.0
98
96
94
92
90
72.22
71.11
70.00
68.89
67.78
+89.6
+93.2
+96.8
+100.4
+104.0
+32
+34
+36
+38
+40
+0.00
+1.11
+2.22
+3.33
+4.44
323.6
327.2
330.8
334.4
338.0
162
164
166
168
170
72.22
73.33
74.44
75.66
76.67
557.6
561.2
564.8
568.4
572.0
292
294
296
298
300
144.44
145.56
146.67
147.78
148.89
126.4
122.8
119.2
115.6
112.0
88
86
84
82
80
66.67
65.56
64.44
63.33
62.22
+107.6
+111.2
+114.8
+118.4
+122.0
+42
+44
+46
+48
+50
+5.56
+6.67
+7.78
+8.89
+10.00
341.6
345.2
348.8
352.4
356.0
172
174
176
178
180
77.78
78.89
80.00
81.11
82.22
575.6
579.2
582.8
586.4
590.0
302
304
306
308
310
150.00
151.11
152.22
153.33
154.44
108.4
104.8
101.2
97.6
94.0
78
76
74
72
70
61.11
60.00
58.89
57.78
56.67
+125.6
+129.2
+132.8
+136.4
+140.0
+52
+54
+56
+58
+60
+11.11
+12.22
+13.33
+14.44
+15.56
359.6
363.2
366.8
370.4
374.0
182
184
186
188
190
83.33
84.44
85.56
86.67
87.78
593.6
597.2
600.8
604.4
608.0
312
314
316
318
320
155.56
156.67
157.78
158.89
160.00
90.4
86.8
83.2
79.6
76.0
68
66
64
62
60
55.56
54.44
53.33
52.22
51.11
143.6
147.2
150.8
154.4
158.0
62
64
66
68
70
16.67
17.78
18.89
20.00
21.11
377.6
381.2
384.8
388.4
392.0
192
194
196
198
200
88.89
90.00
91.11
92.22
93.33
611.6
615.2
618.8
622.4
626.0
322
324
326
328
330
161.11
162.22
163.33
164.44
165.56
72.4
68.8
65.2
61.6
58.0
58
56
54
52
50
50.00
48.89
47.78
46.67
45.56
161.6
165.2
168.8
172.4
176.0
72
74
76
78
80
22.22
23.33
24.44
25.56
26.67
395.6
399.2
402.8
406.4
410.0
202
204
206
208
210
94.44
95.56
96.67
97.73
98.89
629.6
633.2
636.8
640.4
644.0
332
334
336
338
340
166.67
167.78
168.89
170.00
171.11
54.4
50.8
47.2
43.6
40.0
48
46
44
42
40
44.44
43.33
42.22
41.11
40.00
179.6
183.2
186.8
190.4
194.0
82
84
86
88
90
27.78
28.89
30.00
31.11
32.22
413.6
417.2
420.8
424.4
428.0
212
214
216
218
220
100.00
101.11
102.22
103.33
104.44
647.6
651.2
654.8
658.4
662.0
342
344
346
348
350
172.22
173.33
174.44
175.56
176.67
36.4
32.8
29.2
25.6
22.0
38
36
34
32
30
38.89
37.78
36.67
35.56
34.44
197.6
201.2
204.8
208.4
212.0
92
94
96
98
100
33.33
34.44
35.56
36.67
37.78
431.6
435.2
438.8
442.4
446.0
222
224
226
228
230
105.56
106.67
107.78
108.89
110.00
665.6
669.2
672.8
676.4
680.0
352
354
356
358
360
177.78
178.89
180.00
181.11
182.22
18.4
14.8
11.2
7.6
4.0
28
26
24
22
20
33.33
32.22
31.11
30.00
28.89
215.6
219.2
222.8
226.4
230.0
102
104
106
108
110
38.89
40.00
41.11
42.22
43.33
449.6
453.2
456.8
460.4
464.0
232
234
236
238
240
111.11
112.22
113.33
114.44
115.56
683.6
687.2
690.8
694.4
698.0
362
364
366
368
370
183.33
184.44
185.56
186.67
187.78
0.4
+3.2
+6.8
+10.4
+14.0
18
16
14
12
10
27.78
26.67
25.56
24.44
23.33
233.6
237.2
240.8
244.4
248.0
112
114
116
118
120
44.44
45.56
46.67
47.78
48.89
467.6
471.2
474.8
478.4
482.0
242
244
246
248
250
116.67
117.78
118.89
120.00
121.11
701.6
705.2
708.8
712.4
716.0
372
374
376
378
380
188.89
190.00
191.11
192.22
193.33
+17.6
+21.2
+24.8
+28.4
+32.0
8
6
4
2
+0
22.22
21.11
20.00
18.89
17.78
251.6
255.2
258.8
262.4
266.0
122
124
126
128
130
50.00
51.11
52.22
53.33
54.44
485.6
489.2
492.8
496.4
500.0
252
254
256
258
260
122.22
123.33
124.44
125.56
126.67
719.6
723.2
726.8
730.4
734.0
382
384
386
388
390
194.44
195.56
196.67
197.78
198.89
+35.6
+39.2
+42.8
+46.4
+50.0
+2
+4
+6
+8
+10
16.67
15.56
14.44
13.33
12.22
269.6
273.2
276.8
280.4
284.0
132
134
136
138
140
55.56
56.67
57.78
58.89
60.00
503.6
507.2
510.8
514.4
518.0
262
264
266
268
270
127.78
128.89
130.00
131.11
132.22
737.6
741.2
744.8
748.4
752.0
392
394
396
398
400
200.00
201.11
202.22
203.33
204.44
+53.6
+57.2
+60.8
+64.4
+68.0
+12
+14
+16
+18
+20
11.11
10.00
8.89
7.78
6.67
287.6
291.2
294.8
298.4
302.0
142
144
146
148
150
61.11
62.22
63.33
64.44
65.56
521.6
525.2
528.8
532.4
536.0
272
274
276
278
280
133.33
134.44
135.56
136.67
137.78
755.6
759.2
762.8
766.4
770.0
402
404
406
408
410
205.56
206.67
207.78
208.89
210.00
+71.6
+75.2
+78.8
+82.4
+86.0
+22
+24
+26
+28
+30
5.56
4.44
3.33
2.22
1.11
305.6
309.2
312.8
316.4
320.0
152
154
156
158
160
66.67
67.73
68.83
70.00
71.11
539.6
543.2
546.8
550.4
554.0
282
284
286
288
290
138.89
140.00
141.11
142.22
143.33
773.6
777.2
780.8
784.4
788.0
412
414
416
418
420
211.11
212.22
213.33
214.44
215.56
(continued)
18/8/2008 4:27PM Plate # 0
Appendix 2:
pg 527
Temperature Conversion Table / 527
791.6
795.2
798.8
802.4
806.0
422
424
426
428
430
216.67
217.78
218.89
220.00
221.11
1040.0
1058.0
1076.0
1094.0
1112.0
560
570
580
590
600
293.33
298.89
304.44
310.00
315.56
2210.0
2228.0
2246.0
2264.0
2282.0
1210
1220
1230
1240
1250
654.44
660.00
665.56
671.11
676.67
3380.0
3398.0
3416.0
3434.0
3452.0
1860
1870
1880
1890
1900
1015.6
1021.1
1026.7
1032.2
1037.8
809.6
813.2
816.8
820.4
824.0
432
434
436
438
440
222.22
223.33
224.44
225.56
226.67
1130.0
1148.0
1166.0
1184.0
1202.0
610
620
630
640
650
321.11
326.67
332.22
337.78
343.33
2300.0
2318.0
2336.0
2354.0
2372.0
1260
1270
1280
1290
1300
682.22
687.78
693.33
698.89
704.44
3470.0
3488.0
3506.0
3524.0
3542.0
1910
1920
1930
1940
1950
1043.3
1048.9
1054.4
1060.0
1065.6
827.6
831.2
834.8
838.4
842.0
442
444
446
448
450
227.78
228.89
230.00
231.11
232.22
1220.0
1238.0
1256.0
1274.0
1292.0
660
670
680
690
700
348.89
354.44
360.00
365.56
371.11
2390.0
2408.0
2426.0
2444.0
2462.0
1310
1320
1330
1340
1350
710.00
715.56
721.11
726.67
732.22
3560.0
3578.0
3596.0
3614.0
3632.0
1960
1970
1980
1990
2000
1071.1
1076.7
1082.2
1087.8
1093.3
845.6
849.2
852.8
856.4
860.0
452
454
456
458
460
233.33
234.44
235.56
236.67
237.78
1310.0
1328.0
1346.0
1364.0
1382.0
710
720
730
740
750
376.67
382.22
387.78
393.33
398.89
2480.0
2498.0
2516.0
2534.0
2552.0
1360
1370
1380
1390
1400
737.78
743.33
748.89
754.44
760.00
3650.0
3668.0
3686.0
3704.0
3722.0
2010
2020
2030
2040
2050
1098.9
1104.4
1110.0
1115.6
1121.1
863.6
867.2
870.8
874.4
878.0
462
464
466
468
470
238.89
240.00
241.11
242.22
243.33
1400.0
1418.0
1436.0
1454.0
1472.0
760
770
780
790
800
404.44
410.00
415.56
421.11
426.67
2570.0
2588.0
2606.0
2624.0
2642.0
1410
1420
1430
1440
1450
765.56
771.11
776.67
782.22
787.78
3740.0
3758.0
3776.0
3794.0
3812.0
2060
2070
2080
2090
2100
1126.7
1132.2
1137.8
1143.3
1148.9
881.6
885.2
888.8
892.4
896.0
472
474
476
478
480
244.44
245.56
246.67
247.78
248.89
1490.0
1508.0
1526.0
1544.0
1562.0
810
820
830
840
850
432.22
437.78
443.33
448.89
454.44
2660.0
2678.0
2696.0
2714.0
2732.0
1460
1470
1480
1490
1500
793.33
798.89
804.44
810.00
815.56
3830.0
3848.0
3866.0
3884.0
3902.0
2110
2120
2130
2140
2150
1154.4
1160.0
1165.6
1171.1
1176.7
899.6
903.2
906.8
910.4
914.0
482
484
486
488
490
250.00
251.11
252.22
253.33
254.44
1580.0
1598.0
1616.0
1634.0
1652.0
860
870
880
890
900
460.00
465.56
471.11
476.67
482.22
2750.0
2768.0
2786.0
2804.0
2822.0
1510
1520
1530
1540
1550
821.11
826.67
832.22
837.78
843.33
3920.0
3938.0
3956.0
3974.0
3992.0
2160
2170
2180
2190
2200
1182.2
1187.8
1193.3
1198.9
1204.4
917.6
921.2
924.8
928.4
932.0
492
494
496
498
500
255.56
256.67
257.78
258.89
260.00
1670.0
1688.0
1706.0
1724.0
1742.0
910
920
930
940
950
487.78
493.33
498.89
504.44
510.00
2840.0
2858.0
2876.0
2894.0
2912.0
1560
1570
1580
1590
1600
848.89
854.44
860.00
865.56
871.11
4010.0
4028.0
4046.0
4064.0
4082.0
2210
2220
2230
2240
2250
1210.0
1215.6
1221.1
1226.7
1232.2
935.6
939.2
942.8
946.4
950.0
502
504
506
508
510
261.11
262.22
263.33
264.44
265.56
1760.0
1778.0
1796.0
1814.0
1832.0
960
970
980
990
1000
515.56
521.11
526.67
532.22
537.78
2930.0
2948.0
2966.0
2984.0
3002.0
1610
1620
1630
1640
1650
876.67
882.22
887.78
893.33
898.89
4100.0
4118.0
4136.0
4154.0
4172.0
2260
2270
2280
2290
2300
1237.8
1243.3
1248.9
1254.4
1260.0
953.6
957.2
960.8
964.4
968.0
512
514
516
518
520
266.67
267.78
268.89
270.00
271.11
1850.0
1868.0
1886.0
1904.0
1922.0
1010
1020
1030
1040
1050
543.33
548.89
554.44
560.00
565.56
3020.0
3038.0
3056.0
3074.0
3092.0
1660
1670
1680
1690
1700
904.44
910.00
915.56
921.11
926.67
4190.0
4208.0
4226.0
4244.0
4262.0
2310
2320
2330
2340
2350
1265.6
1271.1
1276.7
1282.2
1287.8
971.6
975.2
978.8
982.4
986.0
522
524
526
528
530
272.22
273.33
274.44
275.56
276.67
1940.0
1958.0
1976.0
1994.0
2012.0
1060
1070
1080
1090
1100
571.11
576.67
582.22
587.78
593.33
3110.0
3128.0
3146.0
3164.0
3182.0
1710
1720
1730
1740
1750
932.22
937.78
943.33
948.89
954.44
4280.0
4298.0
4316.0
4334.0
4352.0
2360
2370
2380
2390
2400
1293.3
1298.9
1304.4
1310.1
1315.6
989.6
993.2
996.8
1000.4
1004.0
532
534
536
538
540
277.78
278.89
280.00
281.11
282.22
2030.0
2048.0
2066.0
2084.0
2102.0
1100
1120
1130
1140
1150
598.89
604.44
610.00
615.56
621.11
3200.0
3218.0
3236.0
3254.0
3272.0
1760
1770
1780
1790
1800
960.00
965.56
971.11
976.67
982.22
4370.0
4388.0
4406.6
4424.0
4442.0
2410
2420
2430
2440
2450
1321.1
1326.7
1332.2
1337.8
1343.3
1007.6
1011.2
1014.8
1018.4
1022.0
542
544
546
548
550
283.22
284.44
285.56
286.67
287.78
2120.0
2138.0
2156.0
2174.0
2192.0
1160
1170
1180
1190
1200
626.67
632.22
637.78
643.33
648.89
3290.0
3308.0
3326.0
3344.0
3362.0
1810
1820
1830
1840
1850
987.78
993.33
998.89
1004.4
1010.0
4460.0
4478.0
4496.0
4514.0
4532.0
2460
2470
2480
2490
2500
1348.9
1354.4
1360.0
1365.6
1371.1
(continued)
18/8/2008 4:27PM Plate # 0
pg 528
4550.0
4568.0
4586.0
4604.0
4622.0
2510
2520
2530
2540
2550
1376.7
1382.2
1387.8
1393.3
1398.9
5090.0
5108.0
5126.0
5144.0
5162.0
2810
2820
2830
2840
2850
1543.3
1548.9
1554.4
1560.0
1565.6
5702.0
5792.0
5882.0
5972.0
6062.0
3150
3200
3250
3300
3350
1732.2
1760.0
1787.7
1815.5
1843.3
8402.0
8492.0
8582.0
8672.0
8762.0
4650
4700
4750
4800
4850
2565.5
2593.3
2621.1
2648.8
2676.6
4640.0
4658.0
4676.0
4694.0
4712.0
2560
2570
2580
2590
2600
1404.4
1410.0
1415.6
1421.1
1426.7
5180.0
5198.0
5216.0
5234.0
5252.0
2860
2870
2880
2890
2900
1571.1
1576.7
1582.2
1587.8
1593.3
6152.0
6242.0
6332.0
6422.0
6512.0
3400
3450
3500
3550
3600
1871.1
1898.8
1926.6
1954.4
1982.2
8852.0
8942.0
9032.0
9122.0
9212.0
4900
4950
5000
5050
5100
2704.4
2732.2
2760.0
2787.7
2815.5
4730.0
4748.0
4766.0
4784.0
4802.0
2610
2620
2630
2640
2650
1432.2
1437.8
1443.3
1448.9
1454.4
5270.0
5288.0
5306.0
5324.0
5342.0
2910
2920
2930
2940
2950
1598.9
1604.4
1610.0
1615.6
1621.1
6602.0
6692.0
6782.0
6872.0
6962.0
3650
3700
3750
3800
3850
2010.0
2037.7
2065.5
2093.3
2121.1
9302.0
9392.0
9482.0
9572.0
9662.0
5150
5200
5250
5300
5350
2843.3
2871.1
2898.8
2926.6
2954.4
4820.0
4838.0
4856.0
4874.0
4892.0
2660
2670
2680
2690
2700
1460.0
1465.6
1471.1
1476.7
1482.2
5360.0
5378.0
5396.0
5414.0
5432.0
2960
2970
2980
2990
3000
1626.7
1632.2
1637.8
1643.3
1648.9
7052.0
7142.0
7232.0
7322.0
7412.0
3900
3950
4000
4050
4100
2148.8
2176.6
2204.4
2232.2
2260.0
9752.0
9842.0
9932.0
10022.0
10112.0
5400
5450
5500
5550
5600
2982.2
3010.0
3037.7
3065.5
3093.3
4910.0
4928.0
4946.0
4964.0
4982.0
2710
2720
2730
2740
2750
1487.8
1493.3
1498.9
1504.4
1510.0
5450.0
5468.0
5486.0
5504.0
5522.0
3010
3020
3030
3040
3050
1654.4
1660.0
1665.5
1671.1
1676.7
7502.0
7592.0
7682.0
7772.0
7862.0
4150
4200
4250
4300
4350
2287.7
2315.5
2343.3
2371.1
2398.8
5000.0
5018.0
5036.0
5054.0
5072.0
2760
2770
2780
2790
2800
1515.6
1521.1
1526.7
1532.2
1537.8
5540.0
5558.0
5576.0
5594.0
5612.0
3060
3070
3080
3090
3100
1682.2
1687.8
1693.3
1698.9
1704.4
7952.0
8042.0
8132.0
8222.0
8312.7
4400
4450
4500
4550
4600
2426.6
2454.4
2482.2
2510.0
2537.7
18/8/2008 4:28PM Plate # 0

DOI: 10.1361/faht2008p529

APPENDIX 3
Steel Hardness Conversions

FROM A PRACTICAL STANDPOINT, it is
important to be able to convert the results of
one type of hardness test into those of a different
test. Because a hardness test does not measure a
well-defined property of a material and because
all the tests in common use are not based on the
same type of measurements, it is not surprising
that universal hardness conversion relationships
have not been developed. Hardness conversions
instead are empirical relationships that are defined by conversion tables limited to specific
categories of materials. That is, different conversion tables are required for materials with
greatly different elastic moduli or with different
strain-hardening capacity.
The most reliable hardness-conversion data
exist for steel that is harder than 240 HB. The
indentation hardness of soft metals depends on
the strain-hardening behavior of the material
during the test, which in turn depends on the
previous degree of strain hardening of the
material before the test. The modulus of elasticity also has been shown to influence conversions at high hardness levels. At low hardness
levels, conversions between hardness scales
measuring depth and those measuring diameter
are likewise influenced by differences in the
modulus of elasticity.
Hardness conversions are covered in standards such as SAE J417, Hardness Tests and
Hardness Conversions; ISO 4964, Hardness
ConversionsSteel; and ASTM E140, Standard Hardness Conversion Tables for Metals.
Conversion tables for nickel and high-nickel
alloys, cartridge brass, austenitic stainless steel
plate and sheet, and copper can be found in
ASTM E140. Recently, ASTM committee E-28
on indentation hardness has developed mathematical conversion formulas based on the
conversion-table values fround in ASTM E140.
Over 60 conversion formulas are listed in the
appendix of ASTM E140, and these formulas
can be used in place of the tables. A computer is
helpful in performing the calculations quickly.
Other hardness conversion formulas for various materials have also been published, and a
list of some other conversion formulas is given
in Table A3.1. The standard procedure for
Table A3.1 Examples of published hardness

conversion equations
Steels
HB=
7300
1307HRB
(40100 HRB)
HB=
3710
1307HRE
(30100 HRE)
HB=
1,520,00074500 HRC
(1007HRC)2
(540 HRC)
25,000710(577HRC)2
1007HRC
6700
HRB=1347
HB

1=2
2:43 106
HRC=119:07
HV

1=2
6:85 105
HRA=112:37
HV

1=2
5:53 105
HR15N=117:947
HV

1=2
1:88 106
HR30N=129:527
HV

1=2
3:132 106
HR45N=133:517
HV
HB=
HB = 0.951 HV
HB = 0.941 HV
(4070 HRC)
(+7 HRB, 95% CL)
(2401040 HV)
(2401040 HV)
(2401040 HV)
(2401040 HV)
(2401040 HV)
(steel ball, 200400 HV)
(tungsten-carbide ball,
200700 HV)
Cemented carbides

1=2
2:43 106
HRC=117:357
HV

1=2
2:437106
2117
HV
HRA=
1:885
(9001800 HV)
(9001800 HV)
Rockwell from Knoop for steels
HRC = 64.934 log HK

HRC = 67.353 log HK
HRC = 71.983 log HK
HRC = 76.572 log HK
HRC = 79.758 log HK
HRC = 82.283 log HK
HRC = 83.58 log HK
HRC = 85.848 log HK
140.38
144.32
154.28
163.89
170.92
176.92
179.30
184.55
(15 gf)
(25 gf)
(50 gf)
(100 gf)
(200 gf)
(300 gf)
(500 gf)
(1000 gf)
White cast irons
HB = 0.363 (HRC)2 22.515

(HRC)+717.8
HV = 0.343 (HRC)2 18.132
(HRC)+595.3
HV = 1.136 (HB)2 26.0
1
=0:0001304(1307HRB)
HB
Stable alpha-beta titanium alloys
HRC = 0.078 HV+8.1
(6090 HRB, 110192 HB)
pg 529
18/8/2008 4:28PM Plate # 0
pg 530
from ASTM E140, are for conversion among

Rockwell, Brinell, and Vickers hardness for heat
treated carbon and alloy steels, almost all constructional alloy steels, and tool steels in the asforged, annealed, normalized, and quenched and
tempered conditions. The tables are also summarized in graphical form in Fig. A3.1.
reporting
converted
hardness
numbers
indicates the measured hardness and test scale in
parenthesesfor example, 451 HB (48 HRC).
The method of conversion (table, formula, or
other method) should also be defined.
When making hardness correlations, it is best
to consult ASTM E140. Tables A3.2 to A3.5,
Table A3.2 Approximate Rockwell B hardness conversion numbers for nonaustenitic steels
Rockwell
B, 100 kgf,
1/16 in. ball
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
A, 60 kgf,
diamond
61.5
60.9
60.2
59.5
58.9
58.3
57.6
57.0
56.4
55.8
55.2
54.6
54.0
53.4
52.8
52.3
51.7
51.1
50.6
50.0
49.5
48.9
48.4
47.9
47.3
46.8
46.3
45.8
45.3
44.8
44.3
43.8
43.3
42.8
42.3
41.8
41.4
40.9
40.4
40.0
39.5
39.0
38.6
38.1
37.7
37.2
36.8
36.3
35.9
35.5
35.0
34.6
34.1
33.7
33.3
32.9
32.4
Superficial Rockwell
E, 100 kgf, 15T, 15 kgf,
1/8 in. ball 1/16 in. ball
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
100
99.5
99.0
98.0
97.5
97.0
96.0
95.5
95.0
94.5
93.5
93.0
92.5
92.0
91.0
90.5
90.0
89.5
89.0
88.0
87.5
87.0
86.5
85.5
85.0
84.5
84.0
83.5
93.1
92.8
92.5
92.1
91.8
91.5
91.2
90.8
90.5
90.2
89.9
89.5
89.2
88.9
88.6
88.2
87.9
87.6
87.3
86.9
86.6
86.3
86.0
85.6
85.3
85.0
84.7
84.3
84.0
83.7
83.4
83.0
82.7
82.4
82.1
81.8
81.4
81.1
80.8
80.5
80.1
79.8
79.5
79.2
78.8
78.5
78.2
77.9
77.5
77.2
76.9
76.6
76.2
75.9
75.6
75.3
74.9
30T, 30 kgf,
1/16 in. ball
83.1
82.5
81.8
81.1
80.4
79.8
79.1
78.4
77.8
77.1
76.4
75.8
75.1
74.4
73.8
73.1
72.4
71.8
71.1
70.4
69.7
69.1
68.4
67.7
67.1
66.4
65.7
65.1
64.4
63.7
63.1
62.4
61.7
61.0
60.4
59.7
59.0
58.4
57.7
57.0
56.4
55.7
55.0
54.4
53.7
53.0
52.4
51.7
51.0
50.3
49.7
49.0
48.3
47.7
47.0
46.3
45.7
45T, 45 kgf,
1/16 in. ball Vickers
72.9
71.9
70.9
69.9
68.9
67.9
66.9
65.9
64.8
63.8
62.8
61.8
60.8
59.8
58.8
57.8
56.8
55.8
54.8
53.8
52.8
51.8
50.8
49.8
48.8
47.8
46.8
45.8
44.8
43.8
42.8
41.8
40.8
39.8
38.7
37.7
36.7
35.7
34.7
33.7
32.7
31.7
30.7
29.7
28.7
27.7
26.7
25.7
24.7
23.7
22.7
21.7
20.7
19.7
18.7
17.7
16.7
240
234
228
222
216
210
205
200
195
190
185
180
176
172
169
165
162
159
156
153
150
147
144
141
139
137
135
132
130
127
125
123
121
119
117
116
114
112
110
108
107
106
104
103
101
100
...
...
...
...
...
...
...
...
...
...
...
Knoop,
500 gf
and over
Brinell,
3000 kgf,
10 mm ball
Tensile
strength
MPa (ksi)
Brinell,
500 kgf,
10 mm ball
251
246
241
236
231
226
221
216
211
206
201
196
192
188
184
180
176
173
170
167
164
161
158
155
152
150
147
145
143
141
139
137
135
133
131
129
127
125
124
122
120
118
117
115
114
112
111
110
109
108
107
106
105
104
103
102
101
240
234
228
222
216
210
205
200
195
190
185
180
176
172
169
165
162
159
156
153
150
147
144
141
139
137
135
132
130
127
125
123
121
119
117
116
114
112
110
108
107
106
104
103
101
100
...
...
...
...
...
...
...
...
...
...
...
800 (116)
787 (114)
752 (109)
724 (105)
704 (102)
690 (100)
676 (98)
648 (94)
634 (92)
620 (90)
614 (89)
607 (88)
593 (86)
579 (84)
572 (83)
565 (82)
558 (81)
552 (80)
524 (76)
503 (73)
496 (72)
482 (70)
475 (69)
469 (68)
462 (67)
455 (66)
448 (65)
441 (64)
434 (63)
427 (62)
421 (61)
414 (60)
407 (59)
400 (58)
393 (57)
386 (56)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
201
195
189
184
179
175
171
167
163
160
157
154
151
148
145
142
140
137
135
133
130
128
126
124
122
120
118
116
114
112
110
109
108
106
104
102
100
99
98
96
95
94
92
91
90
89
87
86
85
84
83
82
81
80
80
79
78
Data are only approximate conversions for carbon and low-alloy steels in the annealed, normalized, and quenched-and-tempered conditions; less accurate for cold-worked
condition and for austenitic steels. Source: ASTM E140, except for values for E scale and tensile strength, which are not from standards
18/8/2008 4:28PM Plate # 0
Appendix 3:
pg 531
Steel Hardness Conversions / 531
Table A3.3 Approximate Rockwell C hardness conversion numbers for nonaustenitic steels, according
to ASTM E140
C, 150 kgf,
diamond
68
67
66
65
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
A, 60 kgf,
diamond
D, 100 kgf,
diamond
15 N, 15 kgf,
diamond
30 N, 30 kgf,
diamond
45 N, 45 kgf,
diamond
Vickers
Knoop, 500 gf
and over
Brinell, 3000 kgf,

10 mm ball
Tensile strength,
MPa (ksi)
85.6
85.0
84.5
83.9
83.4
82.8
82.3
81.8
81.2
80.7
80.1
79.6
79.0
78.5
78.0
77.4
76.8
76.3
75.9
75.2
74.7
74.1
73.6
73.1
72.5
72.0
71.5
70.9
70.4
69.9
69.4
68.9
68.4
67.9
67.4
66.8
66.3
65.8
65.3
64.8
64.3
63.8
63.3
62.8
62.4
62.0
61.5
61.0
60.5
76.9
76.1
75.4
74.5
73.8
73.0
72.2
71.5
70.7
69.9
69.2
68.5
67.7
66.9
66.1
65.4
64.6
63.8
63.1
62.1
61.4
60.8
60.0
59.2
58.5
57.7
56.9
56.2
55.4
54.6
53.8
53.1
52.3
51.5
50.8
50.0
49.2
48.4
47.7
47.0
46.1
45.2
44.6
43.8
43.1
42.1
41.6
40.9
40.1
93.2
92.9
92.5
92.2
91.8
91.4
91.1
90.7
90.2
89.8
89.3
88.9
88.3
87.9
87.4
86.9
86.4
85.9
85.5
85.0
84.5
83.9
83.5
83.0
82.5
82.0
81.5
80.9
80.4
79.9
79.4
78.8
78.3
77.7
77.2
76.6
76.1
75.6
75.0
74.5
73.9
73.3
72.8
72.2
71.6
71.0
70.5
69.9
69.4
84.4
83.6
82.8
81.9
81.1
80.1
79.3
78.4
77.5
76.6
75.7
74.8
73.9
73.0
72.0
71.2
70.2
69.4
68.5
67.6
66.7
65.8
64.8
64.0
63.1
62.2
61.3
60.4
59.5
58.6
57.7
56.8
55.9
55.0
54.2
53.3
52.1
51.3
50.4
49.5
48.6
47.7
46.8
45.9
45.0
44.0
43.2
42.3
41.5
75.4
74.2
73.3
72.0
71.0
69.9
68.8
67.7
66.6
65.5
64.3
63.2
62.0
60.9
59.8
58.6
57.4
56.1
55.0
53.8
52.5
51.4
50.3
49.0
47.8
46.7
45.5
44.3
43.1
41.9
40.8
39.6
38.4
37.2
36.1
34.9
33.7
32.5
31.3
30.1
28.9
27.8
26.7
25.5
24.3
23.1
22.0
20.7
19.6
940
900
865
832
800
772
746
720
697
674
653
633
613
595
577
560
544
528
513
498
484
471
458
446
434
423
412
402
392
382
372
363
354
345
336
327
318
310
302
294
286
279
272
266
260
254
248
243
238
920
895
870
846
822
799
776
754
732
710
690
670
650
630
612
594
576
558
542
526
510
495
480
466
452
438
426
414
402
391
380
370
360
351
342
334
326
318
311
304
297
290
284
278
272
266
261
256
251
...
...
...
739(a)
722(a)
705(a)
688(a)
670(a)
654(a)
634(a)
615
595
577
560
543
525
512
496
481
469
455
443
432
421
409
400
390
381
371
362
353
344
336
327
319
311
301
294
286
279
271
264
258
253
247
243
237
231
226
...
...
...
...
...
...
...
...
...
2420 (351)
2330 (338)
2240 (325)
2158 (313)
2075 (301)
2013 (292)
1951 (283)
1882 (273)
1820 (264)
1758 (255)
1696 (246)
1634 (237)
1579 (229)
1524 (221)
1482 (215)
1434 (208)
1386 (201)
1344 (195)
1296 (188)
1254 (182)
1220 (177)
1179 (171)
1137 (166)
1110 (161)
1075 (156)
1048 (152)
1027 (149)
1006 (146)
972 (141)
951 (138)
930 (135)
903 (131)
882 (128)
861 (125)
848 (123)
820 (119)
806 (117)
792 (115)
772 (112)
758 (110)
Data are only approximate conversions for carbon and low-alloy steels in the annealed, normalized, and quenched-and-tempered conditions; less accurate for cold-worked
condition and for austenitic steels. (a) Hardness values outside the recommended range for Brinell testing per ASTM E10. Source: ASTM E140, except for values for
tensile strength, which are not from standards
18/8/2008 4:28PM Plate # 0
pg 532
Table A3.4 Approximate equivalent hardness numbers for Brinell hardness numbers for steel
Brinell hardness number(a)
3000 kgf load, 10 mm
ball(a)
Brinell
indentation
diam, mm
2.25
2.30
2.35
2.40
2.45
2.50
2.55
2.60
2.65
2.70
2.75
2.80
2.85
2.90
2.95
3.00
3.05
3.10
3.15
3.20
3.25
3.30
3.35
3.40
3.45
3.50
3.55
3.60
3.65
3.70
3.75
3.80
3.85
3.90
3.95
4.00
4.05
4.10
4.15
4.20
4.25
4.30
4.35
4.40
4.45
4.50
4.55
4.60
4.65
4.70
4.75
4.80
Standard
ball
Tungstencarbide
ball
...
...
...
...
...
...
...
...
...
...
(495)
...
(477)
...
(461)
...
444
...
429
415
401
388
375
363
352
341
331
321
311
302
293
285
277
269
262
255
248
241
235
229
223
217
212
207
201
197
192
187
183
179
174
170
167
163
159
156
(745)
(712)
(682)
(653)
627
601
578
555
534
514
...
495
...
477
...
461
...
444
429
415
401
388
375
363
352
341
331
321
311
302
293
285
277
269
262
255
248
241
235
229
223
217
212
207
201
197
192
187
183
179
174
170
167
163
159
156
Rockwell hardness No.

C scale, D scale,
A scale,
B scale,
150 kgf
100 kgf
60 kgf
100 kgf
Vickers
load,
load,
load,
load,
1
/16 in.
hardness diamond
diamond diamond
No.
indenter diam ball indenter indenter
840
783
737
697
667
640
615
591
569
547
539
528
516
508
495
491
474
472
455
440
425
410
396
383
372
360
350
339
328
319
309
301
292
284
276
269
261
253
247
241
234
228
222
218
212
207
202
196
192
188
182
178
175
171
167
163
84.1
83.1
82.2
81.2
80.5
79.8
79.1
78.4
77.8
76.9
76.7
76.3
75.9
75.6
75.1
74.9
74.3
74.2
73.4
72.8
72.0
71.4
70.6
70.0
69.3
68.7
68.1
67.5
66.9
66.3
65.7
65.3
64.6
64.1
63.6
63.0
62.5
61.8
61.4
60.8
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
(110.0)
(109.0)
(108.5)
(108.0)
(107.5)
(107.0)
(106.0)
(105.5)
(104.5)
(104.0)
(103.0)
(102.0)
(101.0)
100.0
99.0
98.2
97.3
96.4
95.5
94.6
93.7
92.8
91.9
90.9
90.0
89.0
88.0
87.0
86.0
85.0
83.9
82.9
65.3
63.4
61.7
60.0
58.7
57.3
56.0
54.7
53.5
52.1
51.6
51.0
50.3
49.6
48.8
48.5
47.2
47.1
45.7
44.5
43.1
41.8
40.4
39.1
37.9
36.6
35.5
34.3
33.1
32.1
30.9
29.9
28.8
27.6
26.6
25.4
24.2
22.8
21.7
20.5
(19.0)
(17.7)
(16.4)
(15.2)
(13.8)
(12.7)
(11.5)
(10.2)
(9.0)
(8.0)
(6.7)
(5.4)
(4.4)
(3.3)
(2.0)
(0.9)
74.8
73.4
72.0
70.7
69.7
68.7
67.7
66.7
65.8
64.7
64.3
63.8
63.2
62.7
61.9
61.7
61.0
60.8
59.7
58.8
57.8
56.8
55.7
54.6
53.8
52.8
51.9
51.0
50.0
49.3
48.3
47.6
46.7
45.9
45.0
44.2
43.2
42.0
41.4
40.5
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
Rockwell superficial
hardness No.,
diamond indenter
Knoop
15 N
30 N
45 N
scale, scale, scale, hardness No.,
Scleroscope
15 kgf 30 kgf 45 kgf 500 gf load
and greater hardness No.
load
load
load
92.3
91.6
91.0
90.2
89.6
89.0
88.4
87.8
87.2
86.5
86.3
85.9
85.6
85.3
84.9
84.7
84.1
84.0
83.4
82.8
82.0
81.4
80.6
80.0
79.3
78.6
78.0
77.3
76.7
76.1
75.5
75.0
74.4
73.7
73.1
72.5
71.7
70.9
70.3
69.7
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
82.2
80.5
79.0
77.5
76.3
75.1
73.9
72.7
71.6
70.3
69.9
69.4
68.7
68.2
67.4
67.2
66.0
65.8
64.6
63.5
62.3
61.1
59.9
58.7
57.6
56.4
55.4
54.3
53.3
52.2
51.2
50.3
49.3
48.3
47.3
46.2
45.1
43.9
42.9
41.9
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
72.2
70.4
68.5
66.5
65.1
63.5
62.1
60.6
59.2
57.6
56.9
56.1
55.2
54.5
53.5
53.2
51.7
51.5
49.9
48.4
46.9
45.3
43.6
42.0
40.5
39.1
37.8
36.4
34.4
33.8
32.4
31.2
29.9
28.5
27.3
26.0
24.5
22.8
21.5
20.1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
852
808
768
732
703
677
652
626
604
579
571
558
545
537
523
518
499
496
476
459
441
423
407
392
379
367
356
345
336
327
318
310
302
294
286
279
272
265
259
253
247
242
237
232
227
222
217
212
207
202
198
194
190
186
182
178
91
...
84
81
79
77
75
73
71
70
...
68
...
66
...
65
...
63
61
59
58
56
54
52
51
50
48
47
46
45
43
42
41
40
39
38
37
36
35
34
...
33
32
31
...
30
29
...
28
27
...
26
...
25
...
24
(continued)
Note: Values in parentheses are beyond normal range and are given for information only. Data are for carbon and alloy steels in the annealed, normalized, and quenchedand-tempered conditions; less accurate for cold-worked condition and for austenitic steels (a) Brinell numbers are based on the diameter of impressed indentation. If the
ball distorts (flattens) during test, Brinell numbers will vary in accordance with the degree of such distortion when related to hardnesses determined with a Vickers diamond
pyramid. Rockwell diamond indenter, or other indenter that does not sensibly distort. At high hardnesses, therefore, the relationship between Brinell and Vickers or
Rockwell scales is affected by the type of ball used. Standard steel balls tend to flatten slightly moe than tungsten-carbide balls, resulting in a larger indentation and a lower
Brinell number than shown by a tungsten carbide ball. Thus, on a specimen of about 539547 HV, a standard ball will leave a 2.75 mm indentation (495 HB), and a
tungsten carbide ball a 2.70 mm indentation (514 HB). Conversely, identical indentation diameters for both types of ball will correspond to different Vickers and
Rockwell values. Thus, if indentation in two different specimens both are 2.75 mm diameter (495 HB), the specimen tested with a standard ball has a Vickers hardness of
539, whereas the specimen tested with a tungsten-carbide ball has a Vickers hardness of 528. Source: ASTM E140
18/8/2008 4:28PM Plate # 0
Appendix 3:
pg 533

Brinell hardness number(a)
3000 kgf load, 10 mm
ball(a)
Brinell
indentation
diam, mm
4.85
4.90
4.95
5.00
5.10
5.20
5.30
5.40
5.50
5.60
Standard
ball
Tungstencarbide
ball
152
149
146
143
137
131
126
121
116
111
152
149
146
143
137
131
126
121
116
111

C scale, D scale,
A scale,
B scale,
150 kgf
100 kgf
60 kgf
100 kgf
Vickers
load,
load,
load,
load,
1
/16 in.
hardness diamond
diamond diamond
No.
indenter diam ball indenter indenter
...
...
...
...
...
...
...
...
...
...
159
156
153
150
143
137
132
127
122
117
81.9
80.8
79.7
78.6
76.4
74.2
72.0
69.8
67.6
65.4
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
hardness No.,
diamond indenter
Knoop
15 N
30 N
45 N
scale, scale, scale, hardness No.,
Scleroscope
15 kgf 30 kgf 45 kgf 500 gf load
and greater hardness No.
load
load
load
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
23
...
22
21
...
20
19
18
17
174
170
166
163
157
151
145
140
135
131
Note: Values in parentheses are beyond normal range and are given for information only. Data are for carbon and alloy steels in the annealed, normalized, and quenchedand-tempered conditions; less accurate for cold-worked condition and for austenitic steels (a) Brinell numbers are based on the diameter of impressed indentation. If the
ball distorts (flattens) during test, Brinell numbers will vary in accordance with the degree of such distortion when related to hardnesses determined with a Vickers diamond
pyramid. Rockwell diamond indenter, or other indenter that does not sensibly distort. At high hardnesses, therefore, the relationship between Brinell and Vickers or
Rockwell scales is affected by the type of ball used. Standard steel balls tend to flatten slightly moe than tungsten-carbide balls, resulting in a larger indentation and a lower
Brinell number than shown by a tungsten carbide ball. Thus, on a specimen of about 539547 HV, a standard ball will leave a 2.75 mm indentation (495 HB), and a
tungsten carbide ball a 2.70 mm indentation (514 HB). Conversely, identical indentation diameters for both types of ball will correspond to different Vickers and
Rockwell values. Thus, if indentation in two different specimens both are 2.75 mm diameter (495 HB), the specimen tested with a standard ball has a Vickers hardness of
539, whereas the specimen tested with a tungsten-carbide ball has a Vickers hardness of 528. Source: ASTM E140
Table A3.5 Approximate equivalent hardness numbers for Vickers (diamond pyramid) hardness
numbers for steel
Vickers
hardness
No.
940
920
900
880
860
840
820
800
780
760
740
720
700
690
680
670
660
650
640
630
620
610
600
590
580
570
Standard
ball
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
A scale,
60 kgf
Tungstenload,
carbide diamond
ball
indenter
...
...
...
(767)
(757)
(745)
(733)
(722)
(710)
(698)
(684)
(670)
(656)
(647)
(638)
(630)
620
611
601
591
582
573
564
554
545
535
(diamond pyramid)
hardness No.,
diamond indenter
Brinell hardness No.,

3000 kg load,
10 mm ball
85.6
85.3
85.0
84.7
84.4
84.1
83.8
83.4
83.0
82.6
82.2
81.8
81.3
81.1
80.8
80.6
80.3
80.0
79.8
79.5
79.2
78.9
78.6
78.4
78.0
77.8
B scale,
100 kgf
load,
1/16 in.
diam ball
C scale,
150 kgf
load,
diamond
indenter
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
68.0
67.5
67.0
66.4
65.9
65.3
64.7
64.0
63.3
62.5
61.8
61.0
60.1
59.7
59.2
58.8
58.3
57.8
57.3
56.8
56.3
55.7
55.2
54.7
54.1
53.6
D scale,
100 kgf
load,
15 N scale,
diamond
15 kgf
indenter
load
76.9
76.5
76.1
75.7
75.3
74.8
74.3
73.8
73.3
72.6
72.1
71.5
70.8
70.5
70.1
69.8
69.4
69.0
68.7
68.3
67.9
67.5
67.0
66.7
66.2
65.8
93.2
93.0
92.9
92.7
92.5
92.3
92.1
91.8
91.5
91.2
91.0
90.7
90.3
90.1
89.8
89.7
89.5
89.2
89.0
88.8
88.5
88.2
88.0
87.8
87.5
87.2
30 N scale,
30 kgf
load
84.4
84.0
83.6
83.1
82.7
82.2
81.7
81.1
80.4
79.7
79.1
78.4
77.6
77.2
76.8
76.4
75.9
75.5
75.1
74.6
74.2
73.6
73.2
72.7
72.1
71.7
Knoop
45 N scale, hardness No.,
500 gf load
45 kgf
and greater
load
75.4
74.8
74.2
73.6
73.1
72.2
71.8
71.0
70.2
69.4
68.6
67.7
66.7
66.2
65.7
65.3
64.7
64.1
63.5
63.0
62.4
61.7
61.2
60.5
59.9
59.3
920
908
895
882
867
852
837
822
806
788
772
754
735
725
716
706
697
687
677
667
657
646
636
625
615
604
Scleroscope
hardness
No.
97
96
95
93
92
91
90
88
87
86
84
83
81
...
80
...
79
78
77
76
75
...
74
73
72
...
(continued)
Note: Values in parentheses are beyond normal range and are given for information only. Data are for carbon and alloy steels in the annealed, normalized, and quenchedand-tempered conditions; less accurate for cold-worked condition and for austenitic steels. Source: ASTM E140
18/8/2008 4:28PM Plate # 0
pg 534
Table A3.5
(continued)
Vickers
hardness
No.
560
550
540
530
520
510
500
490
480
470
460
450
440
430
420
410
400
390
380
370
360
350
340
330
320
310
300
295
290
285
280
275
270
265
260
255
250
245
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
95
90
85
Standard
ball
...
(505)
(496)
(488)
(480)
(473)
(465)
(456)
(448)
441
433
425
415
405
397
388
379
369
360
350
341
331
322
313
303
294
284
280
275
270
265
261
256
252
247
243
238
233
228
219
209
200
190
181
171
162
152
143
133
124
114
105
95
90
86
81
A scale,
60 kgf
Tungstenload,
carbide diamond
ball
indenter
525
517
507
497
488
479
471
460
452
442
433
425
415
405
397
388
379
369
360
350
341
331
322
313
303
294
284
280
275
270
265
261
256
252
247
243
238
233
228
219
209
200
190
181
171
162
152
143
133
124
114
105
95
90
86
81
(diamond pyramid)
hardness No.,
diamond indenter
Brinell hardness No.,

3000 kg load,
10 mm ball
77.4
77.0
76.7
76.4
76.1
75.7
75.3
74.9
74.5
74.1
73.6
73.3
72.8
72.3
71.8
71.4
70.8
70.3
69.8
69.2
68.7
68.1
67.6
67.0
66.4
65.8
65.2
64.8
64.5
64.2
63.8
63.5
63.1
62.7
62.4
62.0
61.6
61.2
60.7
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
B scale,
100 kgf
load,
1/16 in.
diam ball
C scale,
150 kgf
load,
diamond
indenter
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
(110.0)
...
(109.0)
...
(108.0)
...
(107.0)
...
(105.5)
...
(104.5)
...
(103.5)
...
(102.0)
...
(101.0)
...
99.5
...
98.1
96.7
95.0
93.4
91.5
89.5
87.1
85.0
81.7
78.7
75.0
71.2
66.7
62.3
56.2
52.0
48.0
41.0
53.0
52.3
51.7
51.1
50.5
49.8
49.1
48.4
47.7
46.9
46.1
45.3
44.5
43.6
42.7
41.8
40.8
39.8
38.8
37.7
36.6
35.5
34.4
33.3
32.2
31.0
29.8
29.2
28.5
27.8
27.1
26.4
25.6
24.8
24.0
23.1
22.2
21.3
20.3
(18.0)
(15.7)
(13.4)
(11.0)
( 8.5)
(6.0)
(3.0)
(0.0)
...
...
...
...
...
...
...
...
...
D scale,
100 kgf
load,
15 N scale,
diamond
15 kgf
indenter
load
65.4
64.8
64.4
63.9
63.5
62.9
62.2
61.6
61.3
60.7
60.1
59.4
58.8
58.2
57.5
56.8
56.0
55.2
54.4
53.6
52.8
51.9
51.1
50.2
49.4
48.4
47.5
47.1
46.5
46.0
45.3
44.9
44.3
43.7
43.1
42.2
41.7
41.1
40.3
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
86.9
86.6
86.3
86.0
85.7
85.4
85.0
84.7
84.3
83.9
83.6
83.2
82.8
82.3
81.8
81.4
80.8
80.3
79.8
79.2
78.6
78.0
77.4
76.8
76.2
75.6
74.9
74.6
74.2
73.8
73.4
73.0
72.6
72.1
71.6
71.1
70.6
70.1
69.6
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
30 N scale,
30 kgf
load
71.2
70.5
70.0
69.5
69.0
68.3
67.7
67.1
66.4
65.7
64.9
64.3
63.5
62.7
61.9
61.1
60.2
59.3
58.4
57.4
56.4
55.4
54.4
53.6
52.3
51.3
50.2
49.7
49.0
48.4
47.8
47.2
46.4
45.7
45.0
44.2
43.4
42.5
41.7
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
Knoop
45 N scale, hardness No.,
45 kgf
500 gf load
load
and greater
58.6
57.8
57.0
56.2
55.6
54.7
53.9
53.1
52.2
51.3
50.4
49.4
48.4
47.4
46.4
45.3
44.1
42.9
41.7
40.4
39.1
37.8
36.5
35.2
33.9
32.5
31.1
30.4
29.5
28.7
27.9
27.1
26.2
25.2
24.3
23.2
22.2
21.1
19.9
...
...
. . ..
...
...
...
...
...
...
...
...
...
...
...
...
...
...
594
583
572
561
550
539
528
517
505
494
482
471
459
447
435
423
412
400
389
378
367
356
346
337
328
318
309
305
300
296
291
286
282
277
272
267
262
258
253
243
234
226
216
206
196
185
175
164
154
143
133
123
112
107
102
97
Scleroscope
hardness
No.
71
70
69
68
67
...
66
65
64
...
62
...
59
58
57
56
55
...
52
51
50
48
47
46
45
...
42
...
41
...
40
39
38
...
37
...
36
35
34
33
32
30
29
28
26
25
23
22
21
20
18
...
...
...
...
...
Note: Values in parentheses are beyond normal range and are given for information only. Data are for carbon and alloy steels in the annealed, normalized, and quenchedand-tempered conditions; less accurate for cold-worked condition and for austenitic steels. Source: ASTM E140
18/8/2008 4:28PM Plate # 0
1000
1000
900
900
800
800
Vickers hardness
Vickers hardness
Appendix 3:
700
600
500
45 N
600
500
400
300
300
200
0
10
20
30
40
50
60
70
80
90
100
10
700
700
Brinell hardness number
800
600
500
400
300
40
50
60
70
80
90
100
45 N
30 N
500
15 N
400
300
10
20
30
40
50
60
70
80
90
200
100
10
1000
70
800
60
600
400
200
100 200 300 400 500 600 700 800 900 1000
20
30
40
50
60
70
80
90
100
Rockwell superficial hardness number
Rockwell C hardness
Vickers hardness
30
600
Rockwell C hardness
Fig. A3.1
20
800
15 N
30 N
Rockwell C hardness
200
700
400
200
pg 535
45 N
50
30 N
40
15 N
30
20
0
10
20
30
40
50
60
70
80
90
100
Approximate equivalent hardness numbers for steel. Points represent data from the hardness conversion tables.
18/8/2008 4:29PM Plate # 0

DOI: 10.1361/faht2008p537
pg 537

APPENDIX 4
Austenitizing Temperatures for Steels

TEMPERATURES RECOMMENDED for
austenitizing carbon and low-alloy steels prior
to hardening are given in Table A4.1 (for
direct-hardening grades) and Table A4.2 (for
carburized steels). Table A4.2 is applicable to
carburized steels that have been cooled slowly
from the carburizing temperature and are to be
furnace hardened in a subsequent operation.
For most applications, the rate of heating to
the austenitizing temperature is less important
than other factors in the hardening process, such
as maximum temperature attained throughout
the section, temperature uniformity, time at
temperature, and rate of cooling. The thermal
conductivity of the steel, the nature of the
furnace atmosphere (scaling or nonscaling),
thickness of section, method of loading (spaced
or stacked), and the degree of circulation of
the furnace atmosphere all influence the rate of

heating of the steel part to the required temperature selected from Tables A4.1 and A4.2.
The difference in temperature rise within
thick and thin sections of articles of varying
cross section is a major problem in practical
heat-treating operations. When temperature
uniformity is the ultimate objective of the heating cycle, this is more safely attained by slowly
heating than by rapidly heating. Furthermore,
the maximum temperature in the austenite range
should not exceed that required to achieve
the necessary extent of solution of carbide. The
temperatures listed in Tables A4.1 and A4.2
conform with this requirement. When heating
with significant cross-section variations, provisions should be made for slower heating to
minimize thermal stresses and distortions.
Table A4.1 Austenitizing temperatures for direct-hardening carbon and alloy steels (SAE)
Temperature
Steel
Temperature
F
Carbon steels
1025
1030
1035
1037
1038(a)
1039(a)
1040(a)
1042
1043(a)
1045(a)
1046(a)
1050(a)
1055
1060
1065
1070
1074
1078
1080
1084
1085
1086
1090
1095
Steel
Free-cutting carbon steels

855900
845870
830855
830855
830855
830855
830855
800845
800845
800845
800845
800845
800845
800845
800845
800845
800845
790815
790815
790815
790815
790815
790815
790815(a)
15751650
15501600
15251575
15251575
15251575
15251575
15251575
14751550
14751550
14751550
14751550
14751550
14751550
14751550
14751550
14751550
14751550
14501500
14501500
14501500
14501500
14501500
14501500
14501500(b)
1137
1138
1140
1141
1144
1145
1146
1151
1536
1541
1548
1552
1566
830855
815845
815845
800845
800845
800845
800845
800845
815845
815845
815845
815845
855885
15251575
15001550
15001550
14751550
14751550
14751550
14751550
14751550
15001550
15001550
15001550
15001550
15751625
830855
815845
815845
815845
815845
830855
830855
815855
15251575
15001550
15001550
15001550
15001550
15251575
15251575
15001575
Alloy steels
1330
1335
1340
1345
3140
4037
4042
4047
(continued)
(a) Commonly used on parts where induction hardening is employed. All steels from SAE 1030 up may have induction hardening applications. (b) This temperature range
may be employed for 1095 steel that is to be quenched in water, brine, or oil. For oil quenching, 1095 steel may alternatively be austenitized in the range 815 to 870 C
(1500 to 1600 F). (c) This range is recommended for steel that is to be water quenched. For oil quenching, steel should be austenitized in the range 815 to 870 C (1500 to
1600 F).
18/8/2008 4:30PM Plate # 0
pg 538
Table A4.1
(continued)
Temperature
Temperature
C
Steel
4063
4130
4135
4137
4140
4142
4145
4147
4150
4161
4337
4340
50B40
50B44
5046
50B46
50B50
50B60
5130
5132
5135
5140
5145
5147
5150
800845
815870
845870
845870
845870
845870
815845
815845
815845
815845
815845
815845
815845
815845
815845
815845
800845
800845
830855
830855
815845
815845
815845
800845
800845
800845
800845
800845
775800(c)
775800(c)
775800(c)
845885
815855
830870
830855
830855
815855
815855
815855
815855
800845
800845
830855
830855
815900
815900
815900
845885
845885
830855
14751550
14751550
14751550
14251475(c)
14251475(c)
14251475(c)
15501625
15001575
15251600
15251575
15251575
15001575
15001575
15001575
15001575
14751550
14751550
15251575
15251575
15001650
15001650
15001650
15501625
15501625
15251575
Steel
14751550
15001600
15501600
15501600
15501600
15501600
15001550
15001550
15001550
15001550
15001550
15001550
15001550
15001550
15001550
15001550
14751550
14751550
15251575
15251575
15001550
15001550
15001550
14751550
14751550
5155
5160
51B60
50100
51100
52100
6150
81B45
8630
8637
8640
8642
8645
86B45
8650
8655
8660
8740
8742
9254
9255
9260
94B30
94B40
9840
(a) Commonly used on parts where induction hardening is employed. All steels from SAE 1030 up may have induction hardening applications. (b) This temperature range
may be employed for 1095 steel that is to be quenched in water, brine, or oil. For oil quenching, 1095 steel may alternatively be austenitized in the range 815 to 870 C
(1500 to 1600 F). (c) This range is recommended for steel that is to be water quenched. For oil quenching, steel should be austenitized in the range 815 to 870 C (1500 to
1600 F).
Table A4.2 Reheating (austenitizing) temperatures for hardening of carburized carbon and alloy
steels (SAE)
Carburizing is commonly carried out at 900 to 925 C (1650 to 1700 F); slow cooled and reheated to given austenizing temperature.
Temperature
Steel
Temperature
F
Carbon steels
1010
1012
1015
1016
1017
1018
1019
1020
1022
1513
1518
1522
1524
1525
1526
1527
790830
830845
815845
815845
815845
815845
815845
815845
815845
800830
800830
800830
845870
845870
845870
845870
845870
845870
845870
845870
845870
790830
14501525
15251550
15001550
15001550
15001550
15001550
15001550
15001550
15001550
14751525
14751525
14751525
15501600
15501600
15501600
15501600
15501600
15501600
15501600
15501600
15501600
14501525
Alloy steels
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
760790
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
14001450
760790
760790
760790
760790
14001450
14001450
14001450
14001450
Free-cutting carbon steels

1109
1115
1117
1118
Steel
3310
4320
4615
4617
4620
4621
4626
4718
4720
4815
4817
4820
8115
8615
8617
8620
8622
8625
8627
8720
8822
9310
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_539-540.pdf/Appendix-5/
18/8/2008 4:30PM Plate # 0

DOI: 10.1361/faht2008p539
pg 539

APPENDIX 5
Temper Colors for Steels

IT IS COMMON and long-time knowledge
that steel heated in contact with air at temperatures in the tempering range takes on various
temper colors due to the formation of a thin oxide
film. As an example, the 1948 ASM Metals
Handbook gives the following table:
Heating temperature, F
400
440
475
520
540
590
640
Copper/Steel/Stainless Steel Test.
Place a piece of bright copper strip, a bright

carbon steel strip and a bright stainless steel
strip side by side on the belt and send them
through the furnace.
Use the following chart to determine the
source of oxidation and the color chart to
determine its location:
Color
Faint straw
Straw
Deep straw
Bronze
Peacock
Full blue
Light blue
Color Chart (back cover). Another example of temper colors is shown on the back cover
of this book. This example was developed to
illustrate the effect of oxidation with furnace
leaks. Oxidation of steel in the front of a continuous furnace results in a frosted or matt surface finish with a loss of surface brightness.
Oxidation in the high heat section is evidenced
by a black oxide that is often scaled and flakes
from the surface of the steel. If oxidation is
occurring in the cooling section of a continuous
furnace, use the following test to determine the
cause:

Change the atmosphere flow in the furnace

to 100% Nitrogen
Make absolutely sure that every zone of the
furnace is below 1900 F to prevent copper
from melting
Sample surface appearance

Copper
Oxidized
Bright
Bright
Carbon steel
Stainless steel
Conclusions
Oxidized
Oxidized
Bright
Oxidized
Oxidized
Oxidized
Air leak
Water leak
Very small air
Or water leak
Time-Temperature Effect. A study at the

Illinois Institute of Technology obtained detailed information on temper colors for plain
carbon steel, especially on the effect of time and
temperature. This was obtained using 16 mm
(5/8 in.) diam hot-rolled bars of a standard SAE
1035 steel. Samples that were 50 mm (2 in.)
long were cut from the bars and carefully
machined and cleaned to give smooth, bright
surfaces. They were then heated for various
times at several temperatures in air-circulating
furnaces controlled to within +3 C (+5 F)
of the desired temperatures. The results of this
study are shown in Fig. A5.1 for 1035 steel.
18/8/2008 4:30PM Plate # 0
Fig. A5.1
Temper colors after heating 1035 steel in circulating air (atmospheric pressure)
pg 540
18/8/2008 4:32PM Plate # 0

DOI: 10.1361/faht2008p541
pg 541

APPENDIX 6
Physical Properties of Carbon

and Low-Alloy Steels
Thermal Expansion
Table A6.1 Coefficients of linear thermal expansion for carbon and low-alloy steels
Average coefficient of expansion, mm/m K, at C (F)(a)
AISI-SAE
grade
1008
1008
1010
1010
1010
1015
1015
1016
1017
1018
1019
1020
1020
1021
1022
1023
1025
1026
1029
1030
1035
1037
1039
1040
1043
1044
1045
1045
1046
1050
1052
1053
1055
1060
Treatment or condition
Annealed
Annealed
Annealed
Unknown
Unknown
Rolled
Annealed
Annealed
Unknown
Annealed
Unknown
Annealed
Unknown
Unknown
Annealed
Unknown
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Unknown
Annealed
Annealed
Unknown
Annealed
Annealed
20100
(68212)
12.6(b)
11.6
12.2(b)
11.9(c)
...
11.9(b)
12.2(b)
12.0(b)
12.2(b)
12.0(b)
12.2(b)
11.7
12.2(b)
12.0(b)
12.2(b)
12.2(b)
12.0(b)
12.0(b)
12.0(b)
11.7(b)
11.1
11.1(b)
11.1(b)
11.3
11.3(b)
11.1(b)
11.6(b)
11.2(e)
11.1(b)
11.1(b)
11.3(e)
11.1(b)
11.0
11.1(e)
20200
(68392)
20300
(68572)
20400
(68752)
20500
(68932)
20600
(681112)
20700
(681292)
13.1(b)
12.5
13.0(b)
12.6
...
12.5(b)
...
...
...
...
...
12.1
...
...
12.7(b)
...
...
...
...
...
11.9
...
...
12.0
...
12.0(b)
12.3(b)
11.9(e)
...
12.0(b)
11.8(e)
...
11.8
11.9(e)
13.5(b)
13.0
13.5(b)
13.3
...
13.0(b)
...
...
...
...
...
12.8
...
...
13.1(b)
...
...
...
...
...
12.7
...
...
12.5
...
...
13.1(b)
12.7(e)
...
...
12.7(e)
...
12.6
12.9(e)
13.8(b)
13.6
13.9(b)
13.8
15.1(d)
13.6(b)
13.4(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.4
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
13.4
13.5(b)
13.5(b)
13.3
13.5(b)
13.3(b)
13.7(b)
13.5(e)
13.5(b)
13.5(b)
13.7(e)
13.5(b)
13.4
13.5(e)
14.2(b)
14.2
14.3(b)
14.3
...
14.2(b)
...
...
...
...
...
13.9
...
...
13.9(b)
...
...
...
...
...
14.0
...
...
13.9
...
...
14.2(b)
14.1(e)
...
...
14.5(e)
...
14.0
14.1(e)
14.6(b)
14.6
14.7(b)
14.7
...
...
14.2(b)
14.4(b)
14.5(b)
14.4(b)
14.7(b)
14.4
14.2(b)
14.3(b)
14.4(b)
14.4(b)
14.4(b)
14.4(b)
14.4(b)
14.4(b)
14.4
14.6(b)
14.6(b)
14.4
14.6(b)
...
14.7(b)
14.5(e)
...
...
14.7(e)
...
14.5
14.6(e)
15.0(b)
15.0
15.0(b)
14.9
...
...
...
...
...
...
...
14.8
...
...
14.9(b)
...
...
...
...
...
14.8
...
...
14.8
...
...
15.1(b)
14.8(e)
...
...
15.0(e)
...
14.8
14.9(e)
(continued)
(a) To obtain coefficients in min./in. F, multiply stated values by 0.556. (b) Stated value represents average coefficient between 0 C (32 F) and indicated temperature.
(c) 10.3 mm/m K from 100 to 20 C ( 148 to 68 F); 9.8 mm/m K from 150 to 20 C ( 238 to 68 F). (d) Stated value represents average coefficient between 20
and 650 C (68 and 1200 F). (e) Stated value represents average coefficient between 25 C (75 F) and indicated temperature. (f) Stated value represents average
coefficient between 20 and 95 C (68 and 200 F). (g) Stated value represents average coefficient between 20 and 370 C (68 and 700 F). (h) 8.6 mm/m K from 195 to
20 C ( 320 to 68 F); 10.0 mm/m K from 130 to 20 C ( 200 to 68 F). (i) Stated value represents average coefficient between 20 and 260 C (68 and 500 F).
(j) Stated value represents average coefficient between 20 and 540 C (68 and 1000 F). (k) Stated value represents average coefficient between 18 and 95 C (0 and
200 F). (l) Stated value represents average coefficient between 18 and 650 C (0 and 1200 F). (m) 11.2 mm/m K from 100 to 20 C ( 148 to 68 F) 10.4 mm/m K
from 150 to 20 C ( 238 to 68 F). (n) Stated value represents average coefficient between 20 and 205 C (68 and 400 F). (o) Stated value represents average
coefficient between 20 and 315 C (68 and 600 F). ( p) Stated value represents average coefficient between 25 and 270 C (77 and 518 F). (q) Stated value represents
average coefficient between 20 and 275 C (68 and 525 F). (r) Stated value represents average coefficient between 18 and 260 C (0 and 500 F). (s) Stated value
represents average coefficient between 18 and 540 C (0 and 1000 F).
18/8/2008 4:32PM Plate # 0
pg 542
Table A6.1
(continued)
AISI-SAE
grade
1064
1065
1070
1078
1080
1080
1085
1086
1095
1095
1095
1117
1118
1132
1137
1139
1140
1141
1144
1145
1145
1146
1151
1330
1335
1345
1522
1524
1524
1526
1541
1548
1551
1552
1561
1566
2330
2515
3120
3130
3140
3150
4023
4027
4028
4032
4042
4047
4130
4135
4137
4140
4142
4145
4147
4161
4320
4337
20100
(68212)
20200
(68392)
20300
(68572)
20400
(68752)
20500
(68932)
20600
(681112)
20700
(681292)
Unknown
Unknown
Unknown
Unknown
Annealed
Unknown
Annealed
Unknown
Unknown
Annealed
Hardened
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Annealed
Annealed
Unknown
Unknown
Unknown
Unknown
Unknown
Annealed
Unknown
Annealed
Annealed
Annealed
Unknown
Annealed
Unknown
Annealed
Annealed
Annealed
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Oil hardened, tempered
Unknown
Unknown
Unknown
Unknown
Unknown
11.1(b)
11.1(b)
11.5(b)
11.3(b)
11.0
11.7(b)
11.1(b)
11.1(b)
...
11.4
13.0(b)
12.2(f)
12.2(f)
12.6(f)
12.8
12.6(f)
12.6(f)
...
13.3
11.2(b)
11.6(b)
12.8
...
12.0
12.2
12.0
12.0(b)
11.9
12.0(b)
12.0(b)
12.0(b)
11.9(b)
11.7(b)
11.1(b)
11.1(b)
11.5(b)
10.9(e)
10.9(f)(h)
11.3(k)
11.3(k)
11.3(k)
11.3(k)
11.7(k)
11.7(k)
11.9
11.9
11.9
11.9
12.2
11.7
11.7
12.3
11.7
11.7
11.7
11.5
11.3(k)
11.3(k)
...
...
...
...
11.6
12.2(b)
11.7(b)
...
...
...
...
...
...
...
...
...
...
12.6(b)
...
12.1(b)
12.3(b)
...
12.6(b)
12.8
12.8
12.6
...
12.7
...
...
...
...
...
...
...
...
11.2(e)
...
...
...
...
...
...
...
12.4
12.4
12.4
12.4
...
12.2
12.2
12.7
12.2
12.2
12.2
12.2
...
...
...
...
...
...
12.4
...
12.5(b)
...
...
...
...
...
...
...
...
...
...
...
...
13.0(b)
13.1(b)
...
...
13.3
13.3
13.3
...
...
...
...
...
...
...
...
...
...
12.1(e)
12.6(i)
...
...
...
...
...
...
12.9
12.9
12.9
12.9
...
12.8
12.8
...
12.8
12.8
12.8
12.9
...
...
13.5(b)
13.5(b)
13.3(b)
13.3
13.2
...
13.2(b)
13.1(b)
...
...
...
13.1(g)
13.3(g)
...
...
...
...
...
...
13.6(b)
13.7(b)
...
...
...
...
...
13.5(b)
13.9
13.5(b)
13.5(b)
13.5(b)
13.3(b)
13.5(b)
13.5(b)
13.5(b)
13.5(b)
12.9(e)
...
...
...
...
...
...
...
...
...
...
...
13.7
...
...
13.7
...
...
...
...
...
...
...
...
...
...
13.8
...
13.6(b)
...
...
...
...
...
...
...
...
...
...
...
...
14.0(b)
14.2(b)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
13.4(e)
13.5(j)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
14.2
...
14.2(b)
...
...
...
...
...
...
...
...
...
...
...
...
14.6(b)
14.7(b)
...
...
...
...
...
14.4(b)
14.7
14.4(b)
14.4(b)
14.4(b)
14.6(b)
14.6(b)
...
14.6(b)
14.7(b)
13.8(e)
...
...
...
...
...
...
...
...
...
...
...
14.6
...
...
14.5
...
...
...
...
...
...
...
...
...
...
14.7
...
14.7
...
14.6(d)
...
...
...
...
...
...
...
...
...
...
14.8(b)
15.1(b)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
14.6(l)
14.6(l)
14.6(l)
14.6(l)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
14.6(l)
14.6(l)
(continued)
20 C ( 320 to 68 F); 10.0 mm/m K from 130 to 20 C ( 200 to 68 F). (i) Stated value represents average coefficient between 20 and 260 C (68 and 500 F).
Appendix 6:
18/8/2008 4:32PM Plate # 0
pg 543
Physical Properties of Carbon and Low-Alloy Steels / 543

AISI-SAE
grade
4340
4340
4422
4427
4615
4617
4626
4718
4815
4820
5046
50B60
5117
5120
5130
5132
5135
5140
5145
5150
5155
52100
52100
6150
6150
615 0
8115
81B45
8617
8622
8625
8627
8630
8637
8645
8650
8655
8720
8822

600 C(1110 F)
630 C (1170 F)
Unknown
Unknown
Unknown
Carburized and hardened
Normalized and tempared
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Annealed
Unknown
Unknown
Unknown
Annealed
Hardened
Annealed
Hardened, tempered
205 C (400 F)
Annealed
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
20100
(68212)
12.3
. . .(m)
20200
(68392)
20300
(68572)
20400
(68752)
20500
(68932)
20600
(681112)
20700
(681292)
12.7
...
13.7
...
14.5
...
12.4
...
13.6
...
14.3
...
11.7(k)
12.6
11.5
12.5
11.70(f)
11.3
11.5(f)
11.3(f)
11.9
11.9
12.0
12.0
12.2
12.2
12.0
...
12.2
12.8
12.2
11.9(b)
12.6(b)
12.2
12.0
...
...
12.1
13.1
...
12.2
12.2(n)
12.2(n)
12.4
12.4
12.8
12.8
12.9
12.9
12.8
12.6
...
...
...
...
...
12.7
12.5
...
13.8
12.7
...
12.6(i)
13.1
13.1(o)
12.9(o)
12.9
12.9
13.5
13.5
13.5
13.5
13.5
13.4
13.1( p)
13.7( p)
13.1(q)
...
...
12.3
12.9
...
...
13.2
...
...
...
...
...
...
...
...
...
...
...
...
13.9
...
...
...
...
...
13.7
13.0
...
...
13.7
...
...
...
...
...
...
...
...
...
...
...
...
14.3
...
...
...
...
...
14.1
13.3
...
15.1
14.1
...
13.8(j)
...
...
...
...
...
...
...
...
...
...
14.6
...
...
...
...
...
14.4
13.7
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
15.0
...
...
...
...
...
...
...
12.4(e)
11.9
11.9
11.9(k)
11.1
11.1
11.3
11.3(k)
11.3
11.7
11.7
11.7
11.3(k)
11.3
12.6(e)
12.6
12.6
...
12.2
12.2
12.2
...
12.2
12.2
12.2
12.2
...
12.2
13.3(e)
13.3
13.3
12.8(r)
12.9
12.9
12.9
...
12.8
12.8
12.8
12.8
...
12.9
13.8(e)
...
...
...
...
...
...
...
...
...
...
...
...
...
14.2(e)
...
...
...
...
...
...
...
...
...
...
...
...
...
14.5(e)
...
...
14.0(s)
...
...
...
...
...
...
...
...
...
...
14.7(e)
...
...
...
...
...
...
14.6(l)
...
...
...
...
14.6(l)
...
20 C ( 320 to 68 F);10.0 mm/m K from 130 to 20 C ( 200 to 68 F). (i) Stated value represents average coefficient between 20 and 260 C (68 and 500 F).
18/8/2008 4:32PM Plate # 0
Table A6.2 Summary of thermal expansion

Coefficient of linear thermal expansion (CTE), approximate ranges at room temperature to 100 C (212 F), from lowest to highest
CTE value
CTE
CTE
10 6/K
10 6/F
Material
10 6/K
10 6/F
2.63.3
2.26.1
4.54.6
0.68.7
4.85.1
5.6
6.0
6.1
5.77.0
6.36.6
6.26.7
6.5
4.98.2
6.8
2.012
7.1
7.27.3
5.19.6
4.511
7.19.7
8.38.5
8.38.4
5.511
8.48.6
8.68.7
7.69.9
7.710
4.014
8.89.1
7.611
9.39.6
9.39.9
9.110
8.411
8.611
9.9
9.810
1011
6.814
9.312
7.614
11
8.912
9.512
9.912
11
1012
1012
9.312
9.813
1012
1.41.8
1.23.4
2.52.6
0.34.8
2.72.8
3.1
3.3
3.4
3.23.9
3.53.7
3.43.7
3.6
2.74.6
3.8
1.16.7
3.9
4.04.1
2.85.3
2.56.2
3.95.4
4.64.7
4.64.7
3.16.3
4.74.8
4.84.8
4.25.5
4.35.7
2.27.8
4.95.1
4.25.9
5.25.3
5.25.5
5.15.6
4.76.3
4.86.3
5.5
5.45.8
5.65.9
3.87.8
5.26.5
4.27.5
5.9
4.96.9
5.36.6
5.56.5
6.1
5.66.6
5.66.5
5.26.9
5.46.9
5.86.7
1112
12
1113
6.16.9
6.5
5.97.1
1113
1014
12
8.815
1114
9.415
1213
6.27.0
5.67.6
6.6
4.98.4
5.97.5
5.28.2
6.57.0
12
1114
1114
7.617
6.8
5.97.6
6.07.5
4.29.4
11
8.514
11
7.116
9.313
1112
11
11
1013
7.615
1112
6.317
1013
6.2
4.77.8
6.3
3.98.7
5.27.2
6.16.6
6.3
6.4
5.77.0
4.28.5
6.16.8
3.59.4
5.77.3
Pure Silicon (Si)

Pure Osmium (Os)
Pure Tungsten (W)
Iron-cobalt-nickel alloys
Pure Molybdenum (Mo)
Pure Arsenic (As)
Pure Germanium (Ge)
Pure Hafnium (Hf)
Pure Zirconium (Zr)
Pure Cerium (Ce)
Pure Rhenium (Re)
Pure Tantalum (Ta)
Pure Chromium (Cr)
Pure Iridium (Ir)
Magnetically soft iron alloys
Pure Technetium (Tc)
Pure Niobium (Nb)
Pure Ruthenium (Ru)
Pure Praseodymium (Pr)
Beta and near beta titanium
Pure Rhodium (Rh)
Purr Vanadium (V)
Zirconium alloys
Pure Titanium (Ti)
Mischmetal
Unalloyed or low-alloy titanium
Alpha beta titanium
Molybdenum alloys
Pure Platinum (Pt)
Alpha and neat alpha titanium
High-chromium gray cast iron
Ductile high-chromium cast iron
Pure Gadolinium (Gd)
Pure Antimony (Sb)
Maraging steel
Protactinium (Pa)
Water-hardening tool steel
Molybdenum high-speed too steel
Niobium alloys
Ferritic stainless steel
Pure Neodymium (Nd)
Cast ferritic stainless steel
Hot work tool steel
Martensitic stainless steel
Cast martensitic stainless steel
Cermet
Ductile silicon-molybdenum cast iron
Iron carbon alloys
Pure Terbium (Tb)
Cobalt chromium nickel tungsten
High-carbon high-chromium cold work
tool steel
Tungsten high-speed tool steel
Commercially pure or low-alloy nickel
Low-alloy special purpose tool steel
Pure Dysprosium (Dy)
Nickel molybdenum alloy steel
Pure Palladium (Pd)
Pure Thorium (Th)
Wrought iron
Oil-hardening cold work tool steel
Pure Scandium (Sc)
Pure Beryllium (Be)
Carbide
Nickel chromium molybdenum alloy steel
1114
1213
4.820
1015
9.913
9.016
1213
1114
1015
6.020
1115
9.017
13
7.020
1116
13
14
1214
1017
1315
8.119
14
7.020
14
1019
7.921
1316
14
1415
1218
1020
9.719
15
1219
8.822
1418
1319
4.527
1618
17
1519
1718
1618
17
9.825
1619
1619
18
6.27.5
6.47.4
2.711
5.68.3
5.57.3
5.08.9
6.57.4
5.98.0
5.68.6
3.311
6.08.5
5.09.6
7.4
3.911
6.18.6
7.4
7.5
6.87.7
5.69.6
7.08.2
4.511
7.8
3.911
7.8
5.311
4.412
7.09.0
7.8
7.78.4
6.710
5.611
5.411
8.5
6.710
4.912
7.59.8
7.010
2.515
8.810
9.4
8.311
9.29.8
9.110
9.6
5.414
8.911
8.911
10
(continued)
Source: Thermal Properties of Metals and Alloys, ASM, 2002
Material
Shock-resisting tool steel

Structural steel
Air hardening medium-alloy cold work
tool steel
High manganese carbon steel
Malleable cast iron
Mold tool steel
Nonresulfurized carbon steel
Chromium molybdenum alloy steel
Chromium alloy steel
Molybdenum/molybdenum sulfide alloy
steel
Chromium Vanadium alloy steel
Cold work tool steel
Ductile medium-silicon cast iron
Nickel with chromium and/or iron,
molybdenum
Resulfurized carbon steel
High strength low-alloy steel (HSLA)
Pure Lutetium (Lu)
Duplex stainless steel
High strength structural steel
Pure Promethium (Pm)
Pure Iron (Fe)
Metal matrix composite aluminum
Cobalt alloys (including Stellite)
Pure Yttrium (Y)
Gray cast iron
Precipitation hardening stainless steel
Pure Bismuth (Bi)
Pure Holmium (Ho)
Nickel copper
Pure Nickel (Ni)
Palladium alloys
Pure Cobalt (Co)
Cast austenitic stainless steel
Gold alloys
High-nickel gray cast iron
Bismuth tin alloys
Pure Uranium (U)
Pure Gold (Au)
Pure Samarium (Sm)
Pure Erbium (Er)
Nickel chromium silicon gray cast iron
Tungsten alloys
Beryllium alloys
Manganese alloy steel
Iron alloys
Proprietary alloy steel
White cast iron
Austenitic cast iron with graphite
Pure Thulium (Tm)
Wrought copper nickel
Ductile high-nickel cast iron
Pure Lanthanum (La)
Wrought high copper alloys
Cast high copper alloys
Wrought bronze
Cast copper
Wrought copper
Cast copper nickel silver
Cast bronze
Wrought copper nickel silver
Pure Barium (Ba)
pg 544
Appendix 6:
18/8/2008 4:32PM Plate # 0

CTE
CTE
10 6/K
10 6/F
Material
10 6/K
10 6/F
Material
18
18
1820
19
1721
1623
10
10
9.911
11
9.412
8.913
1624
1624
1624
8.913
8.913
9.113
20
2021
1824
11
1112
1013
2526
2526
2526
2526
26
1736
2527
26
2033
2528
2729
28
2324
1415
1415
1415
1415
15
9.220
1415
15
1118
1416
1516
16
1319
2022
1923
1232
22
1222
22
2124
2224
23
23
2225
1112
1113
6.418
12
6.912
12
1213
1213
13
13
1214
2325
2425
2129
2228
2327
1314
1314
1216
1215
1315
Cast copper nickel

Pure Tellurium (Te)
Silver alloys
Pure Silver (Ag)
Wrought brass
3xx.x series cast aluminum
silicon+copper or magnesium
2xxx series wrought aluminum copper
Zinc copper titanium alloys
6xxx series wrought aluminum magnesium
silicon
Pure Strontium (Sr)
Cast brass
1xx.x series commercially pure cast
aluminum
4xx.x series cast aluminum silicon
2xx.x series cast aluminum copper
Pure Gallium (Ga)
Manganese (Mn)
4xxx series wrought aluminum silicon
Pure Calcium (Ca)
7xxx series wrought aluminum zinc
3xxx series wrought aluminum manganese
8xx.x series cast aluminum tin
Unalloyed aluminum ingot
1xxx series commercially pure wrought
aluminum
5xx.x series cast aluminum magnesium
7xx.x series cast aluminum zinc
Tin lead
Zinc aluminum
Zinc copper aluminum
29
2830
2632
2040
2240
3032
3335
35
3749
56
64
6971
83
90
14203
125
97291
16
1617
1418
1112
1222
1718
1819
19
2127
31
36
3839
46
50
7.8113
70
54162
Cast magnesium aluminum zinc

Pure Magnesium (Mg)
Wrought magnesium aluminum zinc
Cast magnesium aluminum manganese
Cast magnesium rare earth
Commercially pure tin
Commercially pure magnesium
Pure Ytterbium (Yb)
Pure Indium (In)
Lead tin solder
Commercially pure or low-alloyed lead
Tin silver
9xx.x series cast aluminum plus other
elements
Pure Lead (Pb)
Pure Thallium (Tl)
Magnesium alloys
Unalloyed or low-alloy zinc
5xxx series wrought aluminum magnesium
Pure Cadmium (Cd)
Zinc copper
Pure Europium (Eu)
Pure Selenium (Se)
Pure Lithium (Li)
Pure Sulfur (S)
Pure Sodium (Na)
Pure potassium (K)
Pure Rubidium (Rb)
Pure Plutonium (Pu)
Pure Phosphorus (P)
Pure Cesium (Cs)
Source: Thermal Properties of Metals and Alloys, ASM, 2002
Table A6.3 Summary of thermal conductivity

Thermal conductivity, approximate ranges at room temperature to 100 C (212 F), from lowest to highest value
W/m K
23
0.25
3
5
58
8
8
611
811
612
810
614
417
11
11
11
11
1113
915
13
13
13
13
13
1116
1017
Btu/
(h ft F)
12
0.13
2
3
35
5
5
37
47
37
56
48
210
6
6
6
7
67
59
7
7
8
8
8
610
610
Material
Pure Tellurium (Te)

Pure Selenium (Se)
Pure Sulfur (S)
Pure Technetium (Tc)
Pure Plutonium (Pu)
Permanent magnet iron alloy
Pure Manganese (Mn)
Beta and near beta titanium
Pure Bismuth (Bi)
Alpha beta titanium
Pure Mercury (Hg)
Cobalt chromium nickel tungsten alloys
Alpha and near alpha titanium
Pure Terbium (Tb)
Pure Cerium (Ce)
Cermet
Nickel molybdenum alloys
Mischmetal
Pure Lanthanum (La)
Pure Samarium (Sm)
Pure Yttrium (Y)
W/m K
1315
14
1416
25
15
15
1121
16
16
16
16
133
17
17
1224
18
1522
1522
33
1128
835
1330
22
1925
2025
1431
Btu/
(h ft F)
Material
79
8
89
15
8
9
612
9
9
9
10
0.619
10
10
714
11
913
913
19
616
520
717
13
1115
1115
818
Cobalt chromium tungsten alloys

Pure Europium (Eu)
Pure Erbium (Er)
Pure Scandium (Sc)
Pure Gallium (Ga)
Pure Holmium (Ho)
Pure Lutetium (Lu)
Cobalt alloys
Pure Neodymium (Nd)
Pure Thulium (Tm)
Cast austenitic stainless steel
Pure Barium (Ba)
Pure Titanium (Ti)
Low-alloy titanium
Pure Rubidium (Rb)
Uranium alloys
Nickel with chromium, iron, molybdenum
Pure Rhenium (Re)
Maraging steel
pg 545
18/8/2008 4:32PM Plate # 0
Table A6.3
(continued)
W/m K
Btu/
(h ft F)
Material
W/m K
Btu/
(h ft F)
1332
2224
1927
24
840
1742
2229
1734
2230
2032
2429
1836
1342
2530
2036
2434
1050
1645
2345
34
1355
3435
2941
2350
1163
38
39
3643
40
2754
3557
2558
3748
3154
3946
3454
44
883
1281
4657
47
4550
4650
4752
2970
2080
5051
2972
51
1886
5354
3871
5257
5457
58
4275
4475
4871
5665
5274
67
6076
21116
6973
7176
6880
819
1314
1116
14
523
1024
1317
1020
1317
1219
1417
1121
824
1517
1221
1420
622
926
1326
20
732
1920
1724
1328
637
22
22
2125
23
1631
2033
1433
2128
1831
2327
2029
26
548
747
2733
27
2629
2729
2730
1740
1246
2930
1742
30
1150
3031
2241
3033
3133
34
2443
2643
2841
3238
3043
39
3544
1267
4042
4144
3947

Pure Hafnium (Hf)
Pure Zirconium (Zr)
Pure Indium (In)
Zirconium alloys
Hot work tool steel
Cast nickel-silver copper alloys
Pure Antimony (Sb)
White cast iron
Shock-resisting tool steel
Pure Cesium (Cs)
Ductile cast iron
Pure Uranium (U)
High-speed tool steel
Lead antimony
Nickel steel
Nickel copper alloys
Cast copper-nickel
Low-alloy lead
Magnetically soft iron alloy
Pure Lead (Pb)
Pure Gallium (Ga)
Specialty stainless steel
Pure Thorium (Th)
Pure Ytterbium (Yb)
Tin antimony
Tantalum alloys
Lead tin
Ultrahigh strength steel
Molybdenum alloy steel
Pure Vanadium (V)
Pure Thallium (Tl)
Nickel chromium molybdenum alloy steel
Nickel specialty alloys
Cast bronze
Tin lead alloys
Pure Protactinium (Pa)
Niobium alloys
High-manganese carbon steel
Niobium alloys
Gray cast iron
High-strength low-alloy steel
Mold steel
Malleable cast iron
Low-alloy nickel
Pure Niobium (Nb)
Pure Tantalum (Ta)
Tin silver alloys
Pure Rubidium (Rb)
Cast magnesium aluminum manganese
Pure Rhenium (Re)
Pure Germanium (Ge)
Pure Tin (Sn)
Pure Chromium (Cr)
Pure Palladium (Pd)
Wrought nickel-silver copper alloys
Pure Platinum (Pt)
Pure Lithium (Li)
Pure Iron (Fe)
77
35121
47110
80
51113
7292
6999
87
88
88
8791
8497
56142
65138
96108
100108
25189
21195
108113
105125
105125
116
88146
125126
26230
100159
92168
44
2070
2764
47
3065
4253
4057
50
51
50
5053
4956
3282
3880
5697
5863
14109
12115
6265
6172
6172
67
5185
7273
15133
5892
5397
117146
131138
133138
138155
121163
141147
134155
138156
145151
130178
112205
6885
7680
7780
8090
7094
8285
7790
8090
8487
75103
65118
153167
109212
100234
154193
8897
63123
58135
89112
121227
167184
146210
112251
85315
167244
70131
97106
85121
65145
49182
97141
142284
82164
209222
121128
221247
80410
117431
194391
292316
346391
360413
417428
415493
633716
128143
46237
21249
112226
169182
200226
208239
241247
240285
366414
Material
Pure Thorium (Th)

Carbide based material
Pure Indium (In)
Cast rare earth magnesium alloys
Pure Nickel (Ni)
Pure Cobalt (Co)
Pure Osmium (Os)
Pure Rhodium (Rh)
Pure Gallium (Ga)
Pure Chromium (Ch)
Pure Cadmium (Cd)
Magnesium alloys
Tungsten alloys
(2xx.x series) Cast aluminum copper
Potassium (K)
Cast brass
Wrought copper nickel
Low-alloy zinc
Pure Zinc (Zn)
Zinc aluminum
Pure Ruhenium (Ru)
(5xx.x series) Cast aluminum magnesium
Pure Calcium (Ca)
Wrought bronze
(7xx.x series) Cast aluminum zinc
(3xx.x series) Cast aluminum silicon plus
copper or magnesium
Wrought magnesium zinc
Pure Sodium (Na)
Cast magnesium zinc
(4xxx series) Wrought aluminum silicon
(4xx.x series) Cast aluminum silicon
Pure Iridium (Ir)
Molybdenum (Mo)
Pure Silicon (Si)
Pure Rhodium (Rh)
Pure Tungsten (W)
(5xxx series) Wrought aluminum
magnesium
Pure Magnesium (Mg)
Beryllium alloys
Wrought brass
manganese
(7xxx series) Wrought aluminum zinc
(8xx.x series) Cast aluminum tin
Pure Beryllium (Be)
(2xxx series) Wrought aluminum copper
(1xxx series) Commercially pure wrought
aluminum
magnesium silicon
(1xx.x series) Commercially pure cast
aluminum
Pure Aluminum (Al)
High copper alloys
Molybdenum alloys
Wrought coppers
Pure Gold (Au)
Cast copper
Silver alloys
Pure Silver (Ag)
Silver tungsten
Silver tungsten carbide
pg 546
Appendix 6:
18/8/2008 4:32PM Plate # 0
pg 547
Table A6.4 Thermal conductivities of carbon and low-alloy steels

Conductivity, W/m K, atC (F) (a)
AISI-SAE
grade
1008
1008
1010
1015
1016
1018
1020
1022
1025
1026
1029
1030
1035
1037
1039
1040
1042
1043
1044
1045
1046
1050
1055
1060
1064
1070
1078
1078
1080
1086
1095
1117
1118
1141
1151
1522
1524
1526
1541
1548
1551
1561
1566
2515
4037
4130
4140
4145
4161
4427
4626
5132
5140
8617
8622
8627
8637
8822
Treatment or
condition
0
(32)
Unknown
Annealed
Unknown
Annealed
Annealed
Annealed
Unknown
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Unknown
Annealed
Unknown
Unknown
Unknown
Unknown
Annealed
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Annealed
Annealed
Annealed
Annealed
Unknown
Annealed
Annealed
Annealed
Unknown
Hardened and
tempered
Hardened and
tempered
Hardened and
tempered
Hardened and
tempered
Hardened and
tempered
Unknown
Unknown
Unknown
Hardened and
tempered
Unknown
Unknown
Unknown
Unknown
Unknown
59.5
65.3(b)
...
51.9
51.9
51.9
51.9
51.9
51.9
51.9
51.9
...
...
...
...
...
51.9
...
...
...
51.2
51.2
51.2
50.5
51.2
49.9
47.8
49.6
50.5
49.9
...
51.9(b)
51.5(b)
...
...
51.9
51.9
51.9
51.9
50.5
50.7
51.2
51.2
34.3(b)
...
100
(212)
200
(392)
300
(572)
400
(752)
500
(932)
600
(1112)
700
(1292)
800
(1472)
1000
(1832)
1200
(2192)
57.8
60.3
46.7
51.0
50.2
50.8
51.0
50.8
51.1
50.1
50.1
51.0
50.8
51.0
50.7
50.7
50.7
50.8
50.8
50.8
49.7
49.7
49.7
...
49.7
48.4
48.2
48.1
...
48.4
46.7
...
...
50.5
50.5
50.1
50.1
50.1
50.1
49.0
49.3
49.7
49.7
...
48.2
53.2
54.9
...
48.9
47.6
48.9
48.9
48.8
49.0
48.4
48.4
...
...
...
...
...
48.2
...
...
...
...
46.8
...
46.8
...
...
45.2
...
46.8
...
...
...
...
47.6
47.6
48.4
48.4
48.4
48.4
48.3
48.4
...
...
...
45.6
49.4
...
...
...
...
...
...
...
46.1
...
...
...
...
...
...
...
45.6
...
...
...
...
...
...
...
...
...
41.4
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
38.2(c)
...
45.6
45.2
...
...
...
...
...
...
42.7
...
...
...
...
...
...
...
41.9
...
...
...
...
...
...
...
...
...
38.1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
39.4
41.0
...
...
...
...
...
...
...
39.4
...
...
...
...
...
...
...
38.1
...
...
...
...
...
...
...
...
...
35.2
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
36.8
36.4
...
...
...
...
...
...
35.6
...
...
...
...
...
...
...
33.9
...
...
...
...
...
...
...
...
...
32.7
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
33.9
33.1
...
...
...
...
...
...
...
31.8
...
...
...
...
...
...
...
30.1
...
...
...
...
...
...
...
...
...
30.1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
28.5
28.5
...
...
...
...
...
...
26.0
...
...
...
...
...
...
...
24.7
...
...
...
...
...
...
...
...
...
24.3
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
27.6
27.6
...
...
...
...
...
...
27.2
...
...
...
...
...
...
...
26.8
...
...
...
...
...
...
...
...
...
26.8
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
29.7
...
...
...
...
...
...
...
29.7
...
...
...
...
...
...
...
29.7
...
...
...
...
...
...
...
...
...
30.1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
42.7
...
...
37.3
...
31.0
...
28.1
30.1
...
42.7
42.3
...
37.7
...
33.1
...
...
...
...
40.6
41.8(b)
...
...
...
...
...
...
...
...
...
...
42.7(b)
...
...
...
...
...
...
...
...
...
...
36.8(b)
...
48.6
...
...
44.1
46.5
44.8
...
...
...
...
...
43.3
37.5(d)
37.5(d)
37.5(d)
37.5(d)
...
...
44.4
43.5
...
...
42.3
...
...
...
38.5
37.7
...
...
35.6
...
...
...
31.8
31.4
...
...
28.9
...
...
...
26.0
...
...
...
28.1
...
...
...
30.1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
(a) To obtain conductivities in Btu/ft h F, multiply values in table by 0.5777893; to obtain conductivities in Btu in./ft2 h F, multiply values by 6,933472; to obtain
conductivities in cal/cm s C, multiply values in table by 0.0023884. (b) Thermal conductivity at 20 C (68 F). (c) Thermal conductivity at 260 C (500 F).
(d) Thermal conductivity at 50 C (120 F)
18/8/2008 4:32PM Plate # 0
Table A6.5 Summary of heat capacity

Specific heat capacity, approximate ranges at room temperature to 100 C (212 F), from lowest to highest value
J/kg K
113117
92142
116117
104140
121129
126130
112133
130132
130131
129134
128138
136
134142
137140
111167
130150
134151
130159
131159
144148
147
150
154
160
162
165
168
171
182
182
188
190
193195
150239
195
195
196
201
204213
219
213239
205238
205247
230
202267
236
201272
235239
238
201280
236246
234251
244248
229264
233264
254
246276
268
285
281289
298300
308
310322
310328
325331
272418
330380
Btu/lb F)
0.0270.028
0.0220.034
0.028
0.0250.033
0.0290.031
0.0300.031
0.0270.032
0.0310.032
0.031
0.0310.032
0.0310.033
0.032
0.0320.034
0.033
00270.040
0.0310.036
0.0320.036
0.0310.038
0.0310.038
0.0340.035
0.035
0.036
0.037
0.038
0.039
0.039
0.040
0.041
0.043
0.043
0.045
0.045
0.0460.047
0.0360.057
0.047
0.047
0.047
0.048
0.0490.051
0.052
0.0510.057
0.0490.057
0.0490.059
0.055
0.0480.064
0.056
0.0480.065
0.0560.057
0.057
0.0480.067
0.0560.059
0.0560.060
0.0580.059
0.0550.063
0.0560.063
0.061
0.0590.066
0.064
0.068
0.0670.069
0.0710.072
0.074
0.0740.077
0.0740.078
0.0780.079
0.0650.100
0.0790.091

Material
Pure Thorium (Th)

Tantalum alloys
Pure Uranium (U)
Pure Mercury (Hg)
Pure Bismuth (Bi)
Pure Gold (Au)
Pure Lead (Pb)
Pure Iridium (Ir)
Pure Osmium (Os)
Low alloyed lead
Commercial gold
Pure Platinum (Pt)
Pure Tungsten (W)
Pure Rhenium (Re)
Uranium alloys
Pure Thallium (Tl)
Tungsten alloys
Lead antimony alloys
Pure Plutonium (Pu)
Pure Hafnium (Hf)
Pure Tantalum (Ta)
Pure Lutetium (Lu)
Pure Ytterbium (Yb)
Pure Thulium (Tm)
Pure Cerium (Ce)
Pure Holmium (Ho)
Pure Erbium (Er)
Pure Europium (Eu)
Pure Terbium (Tb)
Pure Neodymium (Nd)
Mischmetal
Tin lead alloys
Pure Lanthanum (La)
Pure Samarium (Sm)
Pure Tellurium (Te)
Pure Antimony (Sb)
Tin silver alloys
Lead-tin alloys
Pure Tin (Sn)
Commercially pure tin
Cadmium
Pure Cesium (Cs)
Niobium alloys
Pure Silver (Ag)
Pure Ruthenium (Ru)
Molybdenum alloys
Pure Palladium (Pd)
Commercial silver
Commercial palladium
Pure Indium (In)
Indium Semi-conductor Grade
Pure Rhodium (Rh)
Pure Molybdenum (Mo)
Pure Niobium (Nb)
Pure Barium (Ba)
Pure Zirconium (Zr)
Pure Yttrium (Y)
Pure Strontium (Sr)
Pure Germanium (Ge)
Pure Arsenic (As)
Gallium compounds
Zinc alloys
Pure Rubidium (Rb)
J/kg K
Btu/lb F)
337376
335390
377
377.0
375398
374425
0.0800.090
0.0800.093
0.090
0.090
0.0900.095
0.0890.102
360420
387393
317462
360420
372409
382400
280502
371418
398
377419
377420
383418
377439
389420
335481
376440
352473
285545
0.0860.100
0.0920.094
0.0760.110
0.0860.100
0.0890.098
0.0910.096
0.0670.120
0.0890.100
0.095
0.0900.100
0.0900.100
0.0910.100
0.0900.105
0.0930.100
0.0800.115
0.0900.105
0.0840.113
0.0680.130
335502
418421
377565
414452
368502
439
420460
0.0800.120
0.1000.101
0.0900.135
0.0990.108
0.0880.120
0.105
0.1000.110
448
420481
431477
456
393525
444473
460
460
460
0.107
0.1000.115
0.1030.114
0.109
0.0940.125
0.1060.113
0.110
0.110
0.110
460
460
460
460
420502
460462
453473
442487
440494
470
440502
460486
461486
461486
448502
418540
460502
430532
460502
486
481502
475508
460527
0.110
0.110
0.110
0.110
0.1000.120
0.110
0.1080.113
0.1060.116
0.1050.118
0.112
0.1050.120
0.1100.116
0.1100.116
0.1100.116
0.1070.120
0.1000.129
0.1100.120
0.1030.127
0.1100.120
0.116
0.1150.120
0.1130.121
0.1100.126
(continued)
Material
Cast beryllium copper nickel alloys

Cast brass
Cast copper-nickels
Cast nickel-silvers
Wrought copper-nickels
Cobalt chromium nickel tungsten
alloys
Wrought high-copper alloys
Pure Copper (Cu)
Pure Selenium (Se)
Wrought bronze
Gallium (Ga)
Pure Zinc (Zn)
Zirconium alloys
Wrought brass
Beryllium copper
Wrought nickel silvers
Wrought coppers
Unalloyed or low-alloy zinc
Nickel molybdenum alloys
Cast bronze
Cobalt alloys
Nickel with chromium and/or iron,
molybdenum
Cobalt chromium tungsten alloys
Pure Cobalt (Co)
Nickel steel
Chrome-nickel-iron superalloy
Air-hardening medium-alloy cold
work tool steel
Ultrahigh-strength steel
Maraging high strength steel
Molybdenum alloy steel
Zinc aluminum alloys
Pure Iron (Fe)
Mold steel
High-carbon high-chromium cold
work tool steel
Molybdenum high-speed tool steel
Gray cast iron
Water-hardening tool steel
Hot work tool steel
Pure Nickel (Ni)
Pure Chromium (Cr)
Permanent magnet iron alloys
High strength low alloy steel (HSLA)
High-manganese carbon steel
Nickel copper alloys
Austenitic cast stainless steel
Commercially pure or low-alloy nickel
Ductile cast iron
High strength structural steel
Pure Manganese (Mn)
Malleable cast iron
pg 548
Appendix 6:
18/8/2008 4:32PM Plate # 0
pg 549


Btu/lb F)
Material
500
498506
420586
460572
370674
460586
510553
490580
513599
515597
448670
0.119
0.1190.121
0.1000.140
0.1100.137
0.0880.161
0.1100.140
0.1220.132
0.1170.139
0.1230.143
0.1230.143
0.1070.160
561567
623655
691
712728
733
741
770
864
0.1340.135
0.1490.156
0.165
0.1700.174
0.175
0.177
0.184
0.206
816963
0.1950.230
893920
0.2130.220
895933
875963
920
0.2140.223
0.2090.230
0.220
879963
0.2100.230

Pure Vanadium (V)
Alpha and near alpha titanium alloys
Alpha beta titanium alloys
Cast stainless steels
Beta and near beta titanium alloys
Pure Titanium (Ti)
Unalloyed or low-alloy titanium
Nickel chromium molybdenum alloy
steel
Pure Scandium (Sc)
Pure Calcium (Ca)
Carbon
Pure Silicon (Si)
Pure Sulfur (S)
Pure Phosphorus (P)
Pure Potassium (K)
silicon
copper
manganese
Pure Aluminum (Al)
(2xx.x series) Cast aluminum copper
(8xxx series) Wrought
aluminum+other elements
magnesium
J/kg K
Btu/lb F)
Material
887963
0.2120.230
8001050
7951050
900963
0.1910.251
0.1900.251
0.2150.230
960963
0.2290.230
963
963
0.230
0.230
963
963
9661050
10101050
9621109
0.230
0.230
0.2310.251
0.2410.251
0.2300.265
10001080
0.2390.258
10251064
1050
1060
12221234
12851620
15001630
18862070
33003515
0.2450.254
0.251
0.253
0.2920.295
0.3070.387
0.3580.389
0.4500.494
0.7880.840

magnesium silicon
(7xxx series) Wrought aluminum zinc
(1xxx series) Commercially pure
wrought aluminum, 99.00% or
greater
(3xx.x series) Cast aluminum
silicon+copper or magnesium
(4xx.x series) Cast aluminum silicon
(5xx.x series) Cast aluminum
magnesium
(7xx.x series) Cast aluminum zinc
(8xx.x series) Cast aluminum tin
Cast magnesium rare earth alloys
Commercially pure magnesium
alloys
Cast magnesium aluminum
manganese
Wrought magnesium zinc alloys
Cast magnesium zinc
Pure Magnesium (Mg)
Pure Sodium (Na)
Pure Boron (B)
Beryllium alloys
Pure Beryllium (Be)
Pure Lithium (Li)
J/kg K
Table A6.6 Specific heats of carbon and low-alloy steels

Mean apparent specific heat, J/Kg K, at C (F)
AISISAE
grade
1008
1010
1015
1016
1017
1018
1020
1025
1030
1035
1040
1042
1045
1050
1060
1070
1078
1086
1095
1117
1140
1151
1522
1524
Treatment
or
condition
Annealed
Unknown
Annealed
Annealed
Unknown
Annealed
Unknown
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Annealed
Unknown
Unknown
Annealed
Unknown
Unknown
Unknown
Unknown
Unknown
Annealed
Annealed
50100
150200
200250
250300
300350
350400
450500
(122212) (302392) (392482) (482572) (572662) (662752) (842932)
481
450
486
481
481(a)
486
486
486
486
486
486
486
486
486
502
490
490
500
461(b)
481
461(b)
502(b)
486
477
519
500
519
515
...
519
519
519
519
519
519
515
519
519
544
532
532
532
...
...
...
...
519
511
536
520
...
...
...
...
...
532
...
...
...
528
...
...
...
...
548
...
...
...
...
...
...
528
553
535
...
...
...
...
...
557
...
...
...
548
...
...
...
...
565
...
...
...
...
...
...
544
574
565
...
...
...
...
...
574
...
...
...
569
...
...
...
...
586
...
...
...
...
...
...
565
595
590
599
595
...
599
599
599
599
586
586
586
586
590
...
...
607
...
...
...
...
...
599
590
662
650
...
...
...
...
...
662
. . ..
...
...
649
...
...
...
...
670
...
...
...
...
...
...
649
550600 650700 700750 750800 850900

(1022
(1202
(1292
(1382
(1562
1112)
1292)
1382)
1472)
1652)
754
730
...
...
...
...
...
749
...
...
...
708
...
...
...
...
712
...
...
...
...
...
...
741
867
825
...
...
...
...
...
846
...
...
...
770
...
...
...
...
770
...
...
...
...
...
...
837
1105
...
...
...
...
...
...
1432
...
...
...
1583
...
...
...
...
2081
...
...
...
...
...
...
1449
875
...
...
...
...
...
...
950
...
...
...
624
. . ..
...
...
...
615
...
...
...
...
...
...
821
846
...
...
...
...
...
...
...
...
...
...
548
...
...
...
...
...
...
...
...
...
...
...
536
(continued)
C (75200 F). (b) Specific heat at 20100 C (68212 F). (c) Specific heat at 20200 C (68392 F). (d) Value presented is mean value of
(a) Specific heat at 2595

temperatures between 20 C (68 F) and the higher of the cited temperatures. (e) Specific heat at 1025 C (50 80 F). (f) Average specific heat from 25 540 C
(80 1000 F)
18/8/2008 4:32PM Plate # 0
pg 550
Table A6.6
(continued)
Mean apparent specific heat, J/Kg K, at C (F)
AISISAE
grade
Treatment
or
condition
1561
4032
4130
Annealed
Unknown
Hardened
and
tempered
Hardened
and
tempered
Unknown
Normalized
and
tempered
Unknown
Unknown
Hardened
and
tempered
Unknown
Unknown
Unknown
4140
4142
4626
4815
5132
5140
8115
8617
8637
50100
150200
200250
250300
300350
350400
450500
(122212) (302392) (392482) (482572) (572662) (662752) (842932)
550600 650700 700750 750800 850900

(1022
(1202
(1292
(1382
(1562
1112)
1292)
1382)
1472)
1652)
486
...
477
519
461(c)
515
...
...
...
...
...
544
...
...
...
...
...
595
...
...
657
...
...
737
...
...
825
...
...
...
...
...
833
...
...
...
...
473(d)
...
...
...
519(d)
...
561(d)
...
...
...
...
...
335(e)
502(c)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
615(f)
...
...
...
...
...
...
481(b)
494
452(d)
...
523
473(d)
...
536
...
...
553
...
...
574
...
...
595
519(d)
...
657
...
...
741
561(d)
...
1499
...
...
934
...
...
574
...
461(b)
481(a)
...
...
...
502(c)
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
837
...
...
...
...
(a) Specific heat at 2595 C (75200 F). (b) Specific heat at 20100 C (68212 F). (c) Specific heat at 20200 C (68392 F). (d) Value presented is mean value of
temperatures between 20 C (68 F) and the higher of the cited temperatures. (e) Specific heat at 1025 C (50 80 F). (f) Average specific heat from 25 540 C
(80 1000 F)

DOI: 10.1361/faht2008p551
21/8/2008 4:56PM Plate # 0
pg 551

APPENDIX 7
AISI to Non-AISI Steel Cross Reference

Table A7.1 Cross-references of standard SAE carbon and low-alloy steels to selected chemically
similar steels
Specifications established by standards organizations from the Federal Republic of Germany, Japan, the United Kingdom, France, Italy,
and Sweden.
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)
Carbon steels
1005
1.0288, D5-2
1.0303, QSt32-2
1.0312, D5-1
1.0314, D6-2
1.0393, ED3
1.0394, ED4
1.1012, RFe120
...
970 015A03
1006
1.0311, D7-1
1.0313, D8-2
1.0317, RSD4
1.0321, St23
1.0334, StW23
1.0335, StW24
1.0354, St14Cu3
1.0391, EK2
1.0392, EK4
1.1009, Ck7
...
970 030A04
970 040A04
970 050A04
1008
1.0010, D9
1.0318, St28
1.0320, St22
1.0322, USD8
1.0326, RSt28
1.0330, St2, St12
1.0333, St3, St13
1.0331, RoSt2
1.0332, StW22
1.0336, Ust4, Ust14
1.0337, RoSt14
1.0344, St12Cu3
1.0347, RRSt13
1.0357, USt28
1.0359, RRSt23
1.0375, Feinstblech
T57, T61, T65, T70
1.0385, Weissblech
T57, T61, T65, T70
1.0744, 6P10
1.0746, 6P20
1.1116, USD6
G34445
STKM11A
(11A)
1449 3CR
1449 3CS
1449 3HR
1449 3HS
1717 ERW101
3606 261
1010
1.0204, UQSt36
1.0301, C10
1.0328, USD10
1.0349, RSD9
1.1121, Ck10
1.1122, Cq10
G4051 S10C
G4051 S9Ck
1449 40F30, 43F35,

A33-101 AF34
46F40, 50F45,
CC10
60F55, 68F62,
C10
75F70 (available in HR,
HS, CS conditions)
1449 4HR, 4HS, 4CR, 4CS
970 040A10 (En2A,
En2A/1, En2B)
970 045A10, 045M10
(En32A)
(continued)
...
5598 3CD5
1160
A35-564 XC6FF
5598 3CD6
5771 C8
1147
1225
A35-551 XC10
XC6
XC6FF
5598 3CD8
1142
1146
5331 C10
6403 C10
7065 C10
7846 C10
5598 ICD10
5598 3CD12
5771 C12
1232
1265
1311
7356 CB10FF,
CB10FU
21/8/2008 4:57PM Plate # 0
pg 552
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)
Carbon steels (continued)

970 050A10
970 060A10
980 CEW1
1012
1.0439, RSD13
1013
1.0036, USt37-2
1.0037, St37-2
1.0038, RSt37-2
1.0055, Ust34-1
1.0057, RSt34-1
1.0116, St37-3
1.0218, RSt41-2
1.0219, St41-3
1.0307, StE210-7
1.0309, St35.4
1.0315, St37.8
1.0319, RRStE210.7
1.0356, TTSt35
1.0417
1.0457, StE240.7
1015
1.0401, C15
1.1132, CQ15
1.1135, Ck16A1
1.1140, Cm15
1.1141, Ck15
1.1144
1.1148, Ck15AI
1016
1.0419, RSt44.2
1.0467, 15Mn3
1.0468, 15Mn3A1
1.1142, GS-Ck16
1017
...
1018
1.0453, C16.8
G4051 S12C
...
1449 12HS, 12CS

1501 141-360
970 040A12 (En2A,
En2A/1, En2B)
970 050A12
970 060A12
A33-101 AF37
A-35 551 XC12
C12
3059 360
3061 360
3603 360
A35-551 XC12
CC12
5869 Fe360-1KG,
Fe360-2KW
6403 Fe35-2
7070 Fe34CFN
7091 Fe34
1233
1234
1330
G4051 F15Ck
G4051 S15C
970 040A15
970 050A15
970 060A15
970 080A15, 080M15
970 173H16
XC15
5331 C16
7065 C16
7356 CB15
7846 C15
1370
...
3059 440
3606 440
970 080A15, 080M15
970 170H15
970 17H16
...
G4051 S17C
...
1019
...
...
1020
1.0402, C22
1.0414, D20-2
G4051 S20C
G4051 S20CK
1.0427, C22.3
1.0460, C22.8
1.1149, Cm22
1.1151, Ck22
...
1332
1431
...
1370
2101
1449 17HS, 17CS

970 040A17
970 050A17
970 060A17
A35-551 XC18
A35-552 XC18
A35-566 XC18
A35-553 XC18S
A35-554 XC18S
...
1312
970 080A17
A33-101 AF42
C20
...
...
...
...
...
970 040A20
970 050A20
(En2C, En2D)
970 060A20
A35-551 XC18
A35-552 XC18
5598 1CD20
5598 3CD20
A35-566 XC18
A35-553 C20
A35-553 XC18S
A35-554 XC18S
CC20
6922 C21
7356 CB20FF
...
1450
1021
...
...
970 070M20
970 080A20
A35-551 21B3
A35-552 21B3
A35-553 21B3
A35-557 21B3
A35-566 21B3
5332 C20
7065 C20
...
1022
1.0432, C21
1.0469, 21Mn4
1.0482, 19Mn5
1.1133, 20Mn5,
GS-20Mn5
...
3111 Type 9
970 120M19
970 170H20
A35-551 20MB5
A35-552 20MB5
A35-553 20MB5
A35-556 20MB5
5771 20Mn4
...
(continued)
Appendix 7:
21/8/2008 4:57PM Plate # 0
pg 553
AISI to Non-AISI Steel Cross Reference / 553

United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1.1134, Ck19
A35-557 20MB5
A35-566 20MB5
A35-566 20M5
1023
1.1150, Ck22.8
1.1152, Cq22
G4051 S22C
1025
1.0406, C25
1.0415, D25-2, D26-2
1.1158, Ck25
G4051 S25C
1026
1.1155, GS-Ck25
1.1156, GS-Ck24
1029
1.0562, 28Mn4
...
...
1449 2HS, 22CS

970 040A22 (En2C, En2D)
970 050A22
970 060A22
970 080A22
...
A35-552 XC25
A35-566 XC25
...
970 070M26
970 080A25
970 080A27
G3445
STKM15A
(15A),
STKM15C
(15C)
G4051 S28C
970 060A27
970 080A27 (En5A)
A33-101 AF50
CC28
5332 C20
7065 C20
...
5598 ICD25
5598 3CD25
...
7845 C25
7847 C25
...
...
...
C30
1030
1.0528, C30
1.0530, D30-2
1.1178, Ck30
1.1179, Cm30
1.1811, G-31Mn4
G4051 S30C
1449 30HS, 30CS

970 060A30
970 080A30 (En5B)
970 080M30 (En5)
A35-552 XC32
A35-553 XC32
5332 C30
6403 C30
7065 C30
7845 C30
7874 C30
5598 3CD30
6783 Fe50-3
7065 C31
...
1035
1.0501, C35
1.0516, D35-2
1.1172, Cq35
1.1173, Ck34
1.1180, Cm35
1.1181, Ck35
G4051 S35C
1717 CDS105/106
970 060A35
970 080A32 (En5C)
970 080A35 (En8A)
980 CFS6
A33-101 AF55
A35-553 C35
A35-553 XC38
A35-554 XC38
XC35
XC38TS
C35
5333 C33
5598 1CD35
5598 3CD35
7065 C35
7065 C36
7847 C36
7356 CB35
1550
1572
1037
1.0520, 31Mn4
1.0561, 36Mn4
G4051 S35C
3111 type 10
970 080M36
970 170H36
1038
No international
equivalents
1039
1.1190, Ck42A1
1040
1.0511, C40
1.0541, D40-2
1.1186, Ck40
1.1189, Cm40
1042
1043
...
...
970 060A40
970 080A40 (En8C)
970 080M40 (En8)
970 170H41
40M5
A35-552 XC38H2
A35-553 38MB5
A35-556 38MB5
A35-557 38MB5
A35-557 XC38H2
XC42, XC42TS
G4051, S40C
1287
1449 40HS, 40CS
3146 Class 1 Grade C
3146 Class 8
970 060A40
970 080A40 (En8C)
970 080M40 (En8)
A33-101 AF60
C40
1.0517, D45-2
G4051 S43C
970 060A42
970 080A42 (En8D)
A35-552 XC42H1
A35-553 C40
CC45
XC42, XC42TS
1.0558, GS-60.3
G4051 S43C
970 060A42
970 080A42 (En8D)
970 080M46
A35-552 XC42H2
(continued)
...
...
...
...
5598 ICD40
5598 3CD40
6783 Fe60-3
6923 C40
7065 C40
7065 C41
...
7847 C43
...
...
...
21/8/2008 4:57PM Plate # 0
pg 554
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1044
1.0517, D45-2
1045
1.0503, C45
1.1184, Ck46
1.1191, Ck45,
GS-Ck45
1.1192, Cq45
1.1194, Cq45
1.1201, Cm45
1.1193, Cf45
1046
1049
1.0503, C45
1.0519, 45MnAl
1.1159, GS-46Mn4
...
...
G4051 S45C
G5111 SCC5
...
...
970 060A47
970 080A47
970 080M46
...
...
A33-101 AF65
A35-552 XC48H1
A35-553 XC45
3545 C45
5332 C45
7065 C45
A35-554 XC48
XC48TS
C45
7845 C45
7874 C45
5598 1CD45
5598 3CD45
7065 C46
7847 C46
...
3100 AW2
970 080M46
45M4TS
A35-552 XC48H1
A35-552 XC48H2
XC48TS
G3445 STKM17A
(17A)
G3445 STKM17C
(17C)
970 060A47
A35-552 XC48H1
6403 C48
970 080A47
A35-554 XC48
7847 C48
...
1672
...
...
XC48TS
1050
1.0540, C50
1.1202, D53-3
1.1206, Ck50
1.1210, Ck53
1.1213, Cf53
1.1219, Cf54
1.1241, Cm50
G4051 S50C
G4051 S53C
1549 50HS
1549 50CS
970 060A52
970 080A52 (En43C)
970 080M50 (En43A)
A35-553 XC50
5332 C50
7065 C50
7065 C51
7845 C50
7874 C50
5598 ICD50
5598 3CD50
6783 Fe70-3
7847 C53
1674
1053
1.1210 Ck53
1.1213 Cf53
1.1219 Cf54
G4051 S53C
970 080A52 (En43C)
52M4TS
A35-553 XC54
7847 C53
1674
1055
1.0518, D55-2
1.0535, C35
1.1202, D53-3
1.1203, Ck55
1.1209, Cm55
1.1210, Ck53
1.1213, Cf53
1.1219, Cf54
1.1220, D55-3
1.1820, C55W
G4051 S53C
G4051 S55C
3100 AW3
970 060A57
970 070M55
970 080A52 (En43C)
970 080A57
A33-101 AF70
A35-552 XC55H1
A35-552 XC55H2
A35-553 XC54
XC55
C55
5598 3CD55
7065 C55
7845 C55
7874 C55
7065 C56
7847 C53
1059
1.0609, D58-2
1.0610, D60-2
1.0611, D63-2
1.1212, D58-3
1.1222, D63-3
1.1228, D60-3
970 060A62
A35-553 XC60
1060
1.0601, C60
1.0642, 60Mn3
1.1221, Ck60
1.1223, Cm60
1.1740, C60W
1449 60HS
1449 60CS
970 060A57
970 080A57
A35-553 XC60
3545 C60
7064 C60
7065 C60
7845 C60
7874 C60
5598 3CD60
7065 C61
1678
1064
1.0611, D63-2
1.0612, D65-2
1.0613, D68-2
1.1222, D63-3
1.1236, D65-3
...
970 060A62
970 080A62 (En43D)
...
5598 3CD65
...
1065
1.0627, C68
1.0640, 64Mn3
1.1230, Federstahldraht
FD
...
970 060A67
970 080A67 (En43E)
...
G4051 S58C
(continued)
XC65
...
...
...
...
...
Appendix 7:
21/8/2008 4:57PM Plate # 0
pg 555

United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1.1233
1.1240, 65Mn4
1.1250, Federstahldraht
VD
1.1260, 66Mn4
1069
1.0615, D70-2
1.0617, D73-2
1.0627, C68
1.1232, D68-3
1.1237
1.1249, Cf70
1.1251, D70-3
1.1520, C70W1
1.1620, C70W2
...
1070
1.0603, C67
1.0643, 70Mn3
1.1231, Ck67
...
1449 70HS, 70CS

970 060A72
970 070A72 (En42)
970 080A72
XC70
3545 C70
1770
1074
1.0605, C75
1.0645, 76Mn3
1.0655, C74
1.122, D73-3
...
970 070A72 (En42)

970 080A72
A35-553 XC75
XC70
3545 C75
7064 C75
1774
1075
1.0614, D75-2
1.0617, D73-2
1.0620, D78-2
1.1242, D73-3
1.1252, D78-3
1.1253, D75-3
...
A35-553 XC75
XC70
3545 C75
7064 C75
5598 3CD70
5598 3CD75
...
1078
1.0620, D78-2
1.0622, D80-2
1.0626, D83-2
1.1252, D78-3
1.1253, D75-3
1.1255, D80-3
1.1262, D83-3
1.1525, C80W1
970 060A78
XC80
5598 3CD80
...
1080
1.1259 80Mn4
1.1265 D85-2
...
1449 80HS, 80CS

970 060A78
970 060A83
970 070A78
970 080A78
970 080A83
XC80
5598 3CD80
5598 3CD85
...
1084
1.1830, C85W
...
970 060A86
970 080A86
XC85
G4801 SUP3
...
...
A35-553 XC68
XC70
...
...
...
...
...
...
...
1085
1.0647, 85Mn3
1.1273, 90Mn4
1.1819, 90Mn4
...
970 080A83
1086
1.0616, C85, D85-2

1.0626, D83-2
1.0628, D88-2
1.1262, D83-3
1.1265, D85-3
1.1269, Ck85
1.1272, D88-3
...
970 050A86
A35-553 XC90
5598 3CD85
5598 3CD90
...
1090
1.1273, 90Mn4
1.1819, 90Mn4
1.1282, D95S3
...
1449 95HS
1449 95CS
970 060A96
...
3545 C90
7064 C90
5598 3CD95
...
1095
1.0618, D95-2
1.1274, Ck101
1.1275, Ck100
1.1282, D95S3
1.1291, MK97
1.1545, C105W1
1.1645, C105W2
G4801 SUP4
1449 95HS
1449 95CS
970 060A99
(continued)
A35-553 XC100
3545 C100
7064 C100
1870
21/8/2008 4:57PM Plate # 0
pg 556
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)
Carbon-manganese steels
1513
1.0424, Schiffbaustahl
CS:DS
1.0479, 13Mn6
1.0496, 12Mn6
A32
B32
E32
1.0549
1.0579
A36
D36
E36
1.0599
1.8941, QStE260N
1.8945, QStE340N
1.8950, QStE380N
...
12M5
1449 40/30 HS
1449 40/30 CS
1453 A2
A33-101 AF50-S
A35-501 E35-4
A35-501 E36-2
2772 150M12
A35-501 E36-3
970 130M15
970 130M15 (En201)
1.0471, 21MnS15
1.0529, StE350-Z2
1.1120, GS-20Mn5
1.1138, GS-21Mn5
1.1169, 20Mn6
1.8970, StE385.7
1.8972, StE415.7
1.8978
1.8979
G4106 SMn21
1503 221-460
1503 223-409
1503 224-490
3146 CLA2
980 CFS7
1524
1.0499, 21Mn6A1
1.1133, 20Mn5,
GS-20Mn5
1.1160, 22Mn6
G4106 SMn21
G5111 SCMn1
1456 Grade A
970 150M19 (En14A,
En14B)
970 175H23
980 CDS9, CDS10
...
...
970 120M28
A35-551 20MB5
A35-552 20M5
A35-556 20M5
A35-552 20MB5
A35-553 20MB5
A35-556 20MB5
A35-557 20MB5
A35-566 20MB5
...
...
1.0412, 27MnSi5
1.1161, 26Mn5
1.1165, 30Mn5
1.1165, GS-30Mn5
1.1170, 28Mn6
G5111 SCMn2
1453 A3
1456 Grade B1, Grade B2
3100 A5
3100 A6
970 150M28 (En14A,
En14B)
1536
1.0561, 36Mn4
1.1165, 30Mn5
1.1165, GS-30Mn5
1.1166, 34Mn5
1.1167, 36Mn5,
GS-36Mn5
1.1813, G-35Mn5
G4052 SMn1H
G4052 SMn433H
G4106 SMn1
G4106 SMn433
G5111 SCMn2
1045
3100 A5, A6
970 120M36 (En15B)
970 150M36 (En15)
A35-552 32M5
A35-552 38MB5
A35-553 38MB5
A35-556 38MB5
A35-557 38MB5
1541
1.0563, E
G4106 SMn2,
SMn438
G4052 SMn2H,
SMn438H
G4106 SMn3,
SMn443
G4052 SMn3H,
SMn443H
G5111 SCMn5
970 135M44
40M5
970 150M40
45M5
1.1127, 36Mn6
1.1168, GS-40Mn5
1548
1.1128, 46Mn5
1.1159, GS-46Mn4
4010 FeG52
6930 20Mn6
7660 Fe510
...
A35-566 25MS5
1527
1.0564, N-80
...
970 125A15
1522
1526
...
1449 40/30 HR
2165
2168
...
2130
4010 FeG60
7874 C28Mn
...
4010 FeG60
...
G5111 SCMn3
...
2120
2128
A35-552 40M6
...
...
...
...
...
1551
1.0542, StSch80
...
...
24M4TS
...
...
1552
1.0624, StSch90B
1.1226, 52Mn5
...
...
55M5
...
...
(continued)
Appendix 7:
21/8/2008 4:57PM Plate # 0
pg 557

United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)
Carbon-manganese steels (continued)

1561
1.0908, 60SiMn5
...
...
...
...
...
1566
1.1233
1.1240, 65Mn4
1.1260, 66Mn7
...
...
...
...
...
...
A35-562 10F1
...
...
Resulfurized carbon steels

1108
1.0700, U7S10
1.0702, U10S10
G4804 SUM12
1110
1.0703, R10S10
G4804 SUM11
1117
...
G4804 SUM31
970 210A15
970 210M17 (En32M)
970 214A15
970 214M15 (En202)
1118
...
...
970 214M15 (En201)
1137
...
G4804 SUM41
970 212M36 (En8M)

970 216M36 (En15AM)
970 225M36
35MF4
A35-562 35MF6
1139
1.0726, 35S20
...
970 212A37 (En8BM)

970 212M36 (En8M)
970 216M36 (En15AM)
970 225M36
35MF4
A35-562 35MF6
...
1957
1140
No international
equivalents
1141
...
G4804 SUM42
970 212A42 (En8DM)

970 216A42
A35-562 45MF4
...
...
1144
1.0727, 45S20
G4804 SUM43
970 212A42 (En8DM)

970 212M44
970 216M44
970 225M44
970 226M44
A35-562 45MF6
970 212M44
45MF4
...
1146
1.0727, 45S20
...
1151
1.0728, 60S20
1.0729, 70S20
...
...
...
...
...
...
...
...
...
...
...
...
4838 CF35SMn10
4838 CF44SMn28
...
...
1973
...
...
...
1973
Resulfurized/rephosphorized carbon steels

1211
No international
equivalents
1212
1.0711, 9S20
1.0721, 10S20
1.1011, RFe160K
G4804 SUM21
1213
1.0715, 9SMn28
1.0736, 9SMn36
1.0740, 9SMn40
G4804 SUM22
1215
1.0736, 9SMn36
G4804 SUM23
12L14
No international
equivalents
...
10F2
12MF4
S200
4838 10S20
4838 10S22
4838 CF9S22
970 220M07 (En1A)

970 230M07
970 240M07 (En1A)
A35-561 S250
S250
4838 CF9SMn28
4838 CF9SMn32
1912
970 240M07 (En1B)
A35-561 S300
4838 CF9SMn32
4838 CF9SMn36
...
Alloy steels
1330
1335
1340
1345
4023
4024
No international
equivalents
1.5069, 36Mn7
1.5223, 42MnV7
1.0625, StSch90C
1.0912, 46Mn7
1.0913, 50Mn7
1.0915, 50MnV7
1.5085, 51Mn7
1.5225, 51MnV7
1.5416, 20Mo3
1.5416, 20Mo3
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
(continued)
21/8/2008 4:57PM Plate # 0
pg 558
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)
Alloy steels (continued)

4027
1.5419, 22Mo4
...
4028
...
...
...
...
...
...
...
...
970 605A32
970 605H32
970 605M30
970 605M36 (En16)
...
...
...
4032
1.5411
4037
1.2382, 43MnSiMo4
1.5412, GS-40MnMo4 3
1.5432, 42MnMo7
...
3111 Type 2/1

3111 Type 2/2
970 605A37
970 605H37
...
...
...
4042
1.2382, 43MnSiMo4
1.5432, 42MnMo7
...
...
...
...
...
4047
No international
equivalents
4118
1.721, 23CrMoB4
1.7264, 20CrMo5
...
7846 18CrMo4
...
4130
...
G5111 SCMnM3
...
970 605M30
G4052 SCM15H
G4105 SCM21H
G4052 SCM418H
G4105 SCM418H
970 708H20
970 708M20
G4105 SCM1
G4105 SCM432
G4105 SCM2
G4105 SCM430
G4106 SCM2
1717 CDS110
970 708A30
A35-552 30CD4
A35-556 30CD4
A35-557 30CD4
30CrMo4
6929 35CrMo4F
7356 34CrMo4KB
7845 30CrMo4
7874 30CrMo4
2233
4135
1.2330, 35CrMo4
1.7220, 34CrMo4
1.7220, GS-34CrMo4
1.7226, 34CrMoS4
1.7231, 33CrMo4
G4054 SCM3H
G4054 SCM435H
G4105 SCM1
G4105 SCM432
G4105 SCM3
G4105 SCM435
970 708H37
970 708H37
35CD4
A35-552 35CD4
A35-553 35CD4
A35-556 35CD4
A35-557 34CD4
5332 35CrMo4
6929 35CrMo4F
7356 34CrMo4KB
7845 35CrMo4
7874 35CrMo4
2234
4137
1.7225, GS-42CrMo4
G4052 SCM4H
G4052 SCM440H
G4105 SCM4
3100 type 5
970 708A37
970 708H37
970 709A37
40CD4
42CD4
A35-552 38CD4
A35-557 38CD4
5532 40CrMo4
5333 38CrMo4
7356 38CrMo4KB
...
4140
1.3563, 43CrMo4
1.7223, 41CrMo4
G4052 SCM4H
G4052 SCM440H
3100 Type 5
4670 711M40
3160 G40CrMo4
5332 40CrMo4
1.7225, 42CrMo4
G4103 SNCM4
970 708A40
1.7225, GS-42CrMo4
G4105 SCM4
970 708A42 (En19C)
1.7227, 42CrMoS4
G4105 SCM440
970 708H42
40CD4
A35-552 42CD4,
42CDTS
A35-553 42CD4,
42CDTS
A35-556 42CD4,
42CDTS
A35-557 42CD4,
42CDTS
40CD4
A35-552 42CD4,
42CDTS
A35-553 42CD4,
42CDTS
A35-556 42CD4,
42CDTS
A35-557 42CD4,
42CDTS
7845 42CrMo4
7874 42CrMo4
2244
7845 42CrMo4
7847 41CrMo4
7874 42CrMo4
970 708M40
970 709A40
970 709M40
4142
1.3563, 43CrMo4
1.7223, 41CrMo4
...
970 708A42 (En19C)

970 708H42
970 709A42
2244
4145
1.2332, 47CrMo4
G4052 SCM5H
G4052 SCM445H
G4105 SCM5,
SCM445
970 708H45
A35-552 45SCD6
A35-553 45SCD6
...
...
4147
1.2332, 47CrMo4
1.3565, 48CrMo4
G4052 SCM5H
G4052 SCM445H
970 708A47
A35-552 45SCD6
A35-553 45SCD6
...
...
(continued)
Appendix 7:
21/8/2008 4:57PM Plate # 0
pg 559

United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1.7228, 50CrMo4
G4105 SCM5,
SCM445
A35-571 50SCD6
1.7228, GS-50CrMo4
1.7230, 50CrMoPb4
1.7238, 49CrMo4
...
4150
1.3565,48CrMo4
1.7228, 50CrMo4
1.7228, GS-50CrMo4
1.7230, 50CrMoPb4
1.7238, 49CrMo4
4161
1.7229, 61CrMo4
G4801 SUP13
1.7266, GS-58CrMnMo4
43
4320
4340
...
1.6565, 40NiCrMo6
G4103 SNCM23
G4103 SNCM420
G4103 SNCM420H
...
A35-571 50SCD6
...
3100 BW4
3146 CLA12 Grade C
...
20NCD7
A35-565 18NCD4
A35-565 20NCD7
...
G4103 SNCM8
4670 818M40
G4103 SNCM439 970 2S.119
G4108 SNB23-1-5
G4108 SNB24-1-5
E4340
1.6562, 40NiCrMo7 3
...
4422
1.5419, 22Mo
...
4427
No international
equivalent
...
23D5
...
15ND8
4615
...
...
4617
...
...
970 665A17
970 665H17
970 665M17 (En34)
4620
...
...
970 665A19
970 665H20
970 665M20
4626
...
...
970 665A24 (E35B)
4718
No international
equivalent
...
...
...
18NCD4
...
...
...
A35-552 38C2
A35-556 38C2
A35-557 38C2
A35-552 42C2
A35-556 42C2
A35-557 42C2
7356 41Cr2KB
...
45C2
7847 45Cr2
...
50B40
1.7003, 38Cr2
1.7023, 28CrS2
50B44
...
...
...
...
...
...
1.2101, 62SiMnCr4
...
5115
1.7131, 16MnCr5,
GS-16MnCr5
1.7139, 16MnCrS5
...
3608 G20Mo5
...
No international
equivalent
1.7138, 52MnCrB3
...
...
4820
5060
5332 40NiCrMo7
6926 40NiCrMo7
7845 40NiCrMo7
7874 40NiCrMo7
7356
40NiCrMo7KB
...
...
No international
equivalent
50B50
2523
2523-02
...
4817
1.3561, 44Cr2
3097 20NiCrMo7
5331 18NiCrMo7
7846 18NiCrMo7
...
...
No international
equivalent
...
...
No international
equivalent
5046
...
...
4815
50B46
...
...
4720
G4052 SMnC3H
G4052 SMnC443H
G4106 SMnC3
G4106 SMnC443
G5111 SCMnCr4
...
970 2S.119
...
...
...
2ND8
...
...
...
55C2
...
...
970 526M60 (En11)
61SC7
A35-552 60SC7
...
...
G4052 SCr21H
970 527A17
16MC5
G4052 SCr415H
970 527H17
A35-551 16MCS
(continued)
7846 16MnCr5
2127
21/8/2008 4:57PM Plate # 0
pg 560
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1.7142, 16MnCrPb5
1.7160, 16MnCrB5
G4104 SCr21
G4104 SCr415
970 527M17
5117
1.3521, 17MnCr5
1.7016, 17Cr3
1.7131, 16MnCr5,
GS-16MnCr5
1.7139, 16MnCrS5
1.7142, 16MnCrPb5
1.7168, 18MnCrB5
...
5120
1.2162, 21MnCr5
1.3523, 19MnCr5
1.7027, 20Cr4
1.7028, 20Cr5 4
1.7121, 20CrMnS3 3
1.7146, 20MnCrPb5
1.7147, GS-20MnCr5
1.7149, 20MnCrS5
G4052 SCr22H
G4052 SCr420H
G4052 SMn21H
G4052 SMn421H
G4104 SCr22
G4104 SCr420
5130
1.8401, 30MnCrTi4
G4052 SCr2H
G4052 SCr430H
G4104 SCr2
G4104 SCr430
970 530A30 (En18A)

970 530H30
28C4
5132
1.7033, 34Cr4
1.7037, 34CrS4
G4104 SCr3
G4104 SCr435
970 530A32 (En18B)

970 530A36 (En18C)
970 530H32
5135
1.7034, 37Cr4
1.7038, 37CrS4
1.7043, 38Cr4
G4052 SCr3H
G4052 SCr435H
5140
1.7035, 41Cr4
1.7039, 41CrS4
1.7045, 42Cr4
G4052 SCr4H
G4052 SCr440H
G4104 SCr4
G4104 SCr440
5147
1.7145, GS-50CrMn4 4
5150
1.7145, GS-50CrMn4 4
1.8404, 60MnCrTi4
5155
1.7176, 55Cr3
G4801 SUP11
G4801 SUP9
5160
1.2125, 65MnCr4
G4801 SUP9A
51B60
No international
equivalent
E50100
1.2018, 95Cr1
1.3501, 100Cr2
...
E51100
1.2057, 105Cr4
1.2109, 125CrSi5
1.2127, 105MnCr4
1.3503, 105Cr4
...
E52100
1.2059, 120Cr5
1.2060, 105Cr5
1.2067, 100Cr6
1.3505, 100Cr6
1.3503, 105Cr4
...
...
18Cr4
A35-551 16MC5
...
...
...
A35-551 20MC5
A35-552 20MC5
7846 20MnCr5
...
...
...
A35-552 32C4
A35-553 32C4
A35-556 32C4
A35-557 32C4
7356 34Cr4KB
7874 34Cr4
...
3111 Type 3
970 530A36 (En18C)
970 530H36
38C4
A35-552-38Cr4
A35-553 38Cr4
A35-556 38Cr4
A35-557 38Cr4
5332 35CrMn5
6403 35CrMn5
5333 36CrMn4
7847 36CrMn4
7356 38Cr4KB
7845 36CrMn5
7874 36CrMn5
7847 38Cr4
...
3111 Type 3
970 2S.117
970 530A40 (En18D)
970 530H40
970 530M40
A35-552 42C4
A35-557 42C4
A35-556 42C4
5332 40Cr4
7356 41Cr4KB
7845 41Cr4
7874 41Cr4
2245
...
3100 BW2, BW3

3146 CLA 12 Grade A
3146 CLA 12 Grade B
50C4
...
3100 BW2
3100 BW3
3146 CLA 12 Grade A
3146 CLA 12 Grade B
...
...
...
...
2230
A35-571 55C3
...
...
...
...
...
...
A35-565 100C2
...
...
...
...
3160 G90Cr4
...
...
100C6
3097 100Cr6
2258
...
970 527A60 (En48)
970 527H60
970 534A99 (En31)

970 535A99 (En31)
(continued)
Appendix 7:
21/8/2008 4:57PM Plate # 0
pg 561

United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

1.3514, 101Cr6
1.3520, 100CrMn6
6118
No international
equivalent
6150
1.8159, GS-50CrV4
8115
No international
equivalents
81B45
No international
equivalents
G4801 SUP10
8615
...
...
8617
...
...
8620
1.6522, 20NiCrMo2
1.6523, 21NiCrMo2
1.6526, 21NiCrMoS2
1.6543, 21NiCrMo2 2
8622
1.6541,
23MnNiCrMo5 2
8625
...
8627
No international
equivalents
8630
1.6545, 30NiCrMo2 2
8637
8640
...
1.6546, 40NiCrMo2 2
G4052 SNCM21H
G4052 SNCM220H
G4103 SNCM21
G4103 SNCM220
...
...
...
970 735A50 (En47)

970 S.204
...
A35-552 50CV4
A35-553 50CV4
A35-571 50CV4
3545 50CrV4
7065 50CrV4
7845 50CrV4
7874 50CrV4
15NCD2
15NCD4
3097 16NiCrMo2
5331 16NiCrMo2
7846 16NiCrMo2
18NCD4
18NCD6
2772 806M20
970 805A20
970 805H20
970 805M20 (En362)
18NCD4
5331 20NiCrMo2
20NCD2
6403 20NiCrMo2
A35-551 19NCDB2 7846 20NiCrMo2
A35-552 19NCDB2
A35-551 20NCD2
A35-553 20NCD2
A35-565 20NCD2
A35-566 20NCD2
2772 806M22
23NCDB4
970 805A22
970 805H22
970 805M22
A35-556 23MNCD5
A35-556 23NCDB2
A35-566 22NCD2
970 805H25
970 805M25
25NCD4
A35-556 25MNCD6
A35-566 25MNDC6
...
No international
equivalents
8645
No international
equivalents
86B45
No international
equivalents
8650
No international
equivalents
8655
No international
equivalents
8660
...
...
...
2506-03
2506-08
...
...
...
...
30NCD2
7356
30NiCrMo2KB
...
...
970 945M38 (En100)
40NCD3
5332 38NiCrMo4
7356
38NiCrMo4KB
7845 39NiCrMo3
7874 39NiCrMo3
...
...
...
3111 Type 7, 2S.147

970 945A40 (En 100C)
40NCD2
40NCD2TS
5333 40NiCrMo4
7356
40NiCrMo2KB
7845 40NiCrMo2
7874 40NiCrMo2
7847 40NiCrMo3
...
40NCD3TS
40NCD3
8642
...
970 805A17
970 805H17
970 805M17 (En 361)
2230
...
970 805A60
970 805H60
(continued)
...
...
...
21/8/2008 4:57PM Plate # 0
pg 562
Table A7.1
(continued)
United states
(SAE)
Fed. R. of Germany
(DIN)
Japan
(JIS)
United Kingdom
(BS)
France
(AFNOR NF)
Italy
(UNI)
Sweden
(SS14)

8720
No international
equivalents
8740
1.6546, 40NiCrMo2 2
...
3111 Type 7, 2S.147
40NCD2
40NCD2TS
40NCD3TS
8822
No international
equivalents
9254
No international
equivalents
9260
...
E9310
1.6657, 14NiCrMo13 4
94B15
No international
equivalents
94B17
No international
equivalents
94B30
No international
equivalents
G4801 SUP7
...
970 250A58 (En45A)

970 250A61 (En45A)
60S7
61S7
970 832H13
970 832M13 (En36C)
S.157
16NCD13
7356
40NiCrMo2KB
7845 40NiCrMo2
7874 40NiCrMo2
...
6932 15NiCrMo13
9335 10NiCrMo13
...
...
...

DOI: 10.1361/faht2008p563
21/8/2008 5:03PM Plate # 0
pg 563

APPENDIX 8
Non-AISI to AISI Steel Cross Reference

Table A8.1 Cross reference to steels by country
Country
FRANCE
Designation
AISI
Country
AFNOR
20 MC 5
20 MC 5
20 NCD 2
20 NCD 2
20 NCD 2
20 NCD 2
22 NCD 2
22 NCD 2
22 NCD 2
22 NCD 2
25 CD 4 (S)
25 CD 4 (S)
32 C 4
32 C 4
32 DCV 28
35 CD 4
35 CD 4
35 CD 4 TS
35 CD 4 TS
35 M 5
38 C 4
38 C 4
40 CD 4
40 CD 4
40 CD 4
40 CD 4
40 M 5
40 M 5
42 C 2
42 C 2
42 C 4
42 C 4
42 CD 4
42 CD 4
42 CD 4
42 CD 4
45 C 2
45 C 2
50 CV 4
50 CV 4
55 C 3
55 C 3
55 S 7
55 WC 20
60 S 7
60 S 7
61 SC 7
61 SC 7
90 MV 8
100 C 6
CC 20
CC 35
Designation
AISI
FRANCE (continued) AFNOR

5120
5120H
8617
8617H
8620
8620H
8617
8617H
8620
8620H
4130
4130H
5130H
5132
H10
4135
4135H
4135
4135H
1039
5132H
5135
4137
4137H
4140
4140H
1335
1335H
5140H
5150
5135H
5140
4137
4137H
4140
4140H
5140H
5150
6150
6150H
5155
5155H
9255
S1
9260
9260H
9260
9260H
O2
E52100
1020
1035
(continued)
Adapted from Engineering Properties of Steel, ASM International, 1982, p 509523
CC 55
XC 10
XC 15
XC 15
XC 18
XC 18
XC 18 S
XC 25
XC 32
XC 35
XC 38
XC 38 TS
XC 38 TS
XC 42
XC 42
XC 42 TS
XC 42 TS
XC 45
XC 45
XC 48
XC 48
XC 60
XC 65
XC 68
XC 90
1060
1010
1015
1017
1015
1017
1023
1023
1034
1034
1034
1038
1038H
1045
1045H
1045
1045H
1045
1045H
1045
1045H
1064
1064
1070
1086
Z 2 CND 17.12
Z 2 CND 19.15
Z 6 CA 13
Z 6 CN 18.09
Z 6 CND 17.11
Z 6 CNN6 18.10
Z 6 CNT 18.10
Z 6 CNU 17.04
Z 8 C 17
Z 8 CD 17.01
Z 10 C 13
Z 10 C 14
Z 10 CF 17
Z 10 CNF 18.09
Z 12 C 13
Z 12 C 13 M
Z 12 CN 17.08
Z 12 CNS 25.20
Z 12 CNS 25.20
Z 15 CN 16.02
Z 15 CN 24.13
Z 18 N 5
Z 20 C 13
Z 30 WCV 9
Z 38 CDV 5
Z 40 COV 5
316L
317L
405
304
316
347
321
431
430
434
410
410
430F
303
410
403
301
310
314
431
309S
A2515
420
H21
H11
H13
21/8/2008 5:04PM Plate # 0
pg 564
Table A8.1
(continued)
Country
Designation
Country
AISI
GERMANY
(continued)
FRANCE (continued) AFNOR

Z 80 WCV
18-04-01
Z 80 WKCV
18-05-04-01
Z 85 DCWV
08-04-02-01
Z 85 DCWV
08-04-02-01
Z 85 WDCV
06-05-04-02
Z 90 WDCV
06-05-04-02
Z 100 CDV 5
Z 110 WKCDV
07-05-04-04-02
Z 110 WKCDV
07-05-04-04-02
Z 120 WDCV
06-05-04-03
Z 130 WDCV
06-05-04-04
Z 200 C 12
GERMANY
T1
T4
H41
M1
M2
M3 (Class 1)
A2
M41
M42
M3 (Class 2)
M3 (Class 2)
D3
DIN
1.0204
1.0402
1.0419
1.0501
1.0601
1.0700
1.0702
1.0711
1.0715
1.0718
1.0718
1.0904
1.0909
1.0909
1.0912
1.0912
1.1121
1.1133
1.1141
1.1141
1.1151
1.1157
1.1158
1.1165
1.1165
1.1167
1.1167
1.1172
1.1176
1.1176
1.1181
1.1186
1.1191
1.1191
1.1209
1.1210
1.1221
1.1226
1.1230
1.1231
1.1269
1.1273
1008
1020
1016
1035
1060
1108
1109
1212
1213
12L13
12L14
9255
9260
9260H
1345
1345H
1010
1022
1015
1017
1023
1039
1025
1330
1330H
1335
1335H
1030
1038
1038H
1034
1040
1045
1045H
1055
1050
1064
1548
1065
1070
1086
1090
(continued)
Designation
AISI
DIN
1.1274
1.2080
1.2330
1.2341
1.2343
1.2344
1.2363
1.2365
1.2379
1.2510
1.2550
1.2581
1.2606
1.2625
1.2735
1.2842
1.3202
1.3246
1.3246
1.3249
1.3249
1.3255
1.3265
1.3342
1.3343
1.3344
1.3346
1.3346
1.3348
1.3355
1.3501
1.3503
1.3505
1.4001
1.4002
1.4005
1.4006
1.4016
1.4021
1.4024
1.4057
1.4104
1.4112
1.4113
1.4125
1.4301
1.4303
1.4303
1.4305
1.4306
1.4310
1.4401
1.4404
1.4438
1.4449
1.4510
1.4512
1.4532
1.4541
1.4546
1.4550
1.4568
1.4828
1.4833
1.4841
1.4841
1095
D3
P20
P4
H11
H13
A2
H10
D2
O1
S1
H21
H12
H23
P6
O2
T15
M41
M42
M33
M34
T4
T5
M3 (Class 1)
M2
M3 (Class 2)
H41
M1
M7
T1
E50100
E51100
E52100
410S
405
416
410
430
420
403
431
430F
440B
434
440C
304
305
308
303
304L
301
316
316L
317L
317
430Ti
409
632
321
348
347
631
309
309S
310
314
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 565
Non-AISI to AISI Steel Cross Reference / 565

Country
GERMANY
(continued)
Designation
DIN
1.4935
1.4971
1.4980
1.5069
1.5419
1.5419
1.5419
1.5680
1.5711
1.6511
1.6523
1.6523
1.6523
1.6523
1.6543
1.6543
1.6543
1.6543
1.6543
1.6543
1.6545
1.6545
1.6546
1.6546
1.6546
1.6546
1.6562
1.6562
1.6565
1.6565
1.6755
1.6755
1.7006
1.7006
1.7007
1.7007
1.7030
1.7033
1.7033
1.7034
1.7034
1.7035
1.7035
1.7138
1.7138
1.7147
1.7147
1.7176
1.7176
1.7218
1.7218
1.7220
1.7220
1.7223
1.7225
1.7225
1.7225
1.7225
1.7228
1.7228
1.7228
1.7228
1.7362
1.7511
1.7511
1.8159
1.8159
Country
AISI
ITALY
422
661
660
1340H
4419
4419H
4422
A2515
3140
9840
8617
8617H
8620
8620H
8622
8622H
8720
8720H
8822
8822H
8630
8630H
8640
8640H
8740
8740H
E4340
E4340H
4340
4340H
4718
4718H
5140H
5150
50B40
50B40H
5130
5130H
5132
5132H
5135
5135H
5140
50B50
50B50H
5120
5120H
5155
5155H
4130
4130H
4135
4135H
4142H
4137
4137H
4140
4140H
4147
4147H
4150
4150H
501
6118
6118H
6150
6150H
Designation
9 SMn 23
9 SMnPb 23
9 SMnPb 23
10 S 20
20 NiCrMo
20 NiCrMo 2
20 NiCrMo 2
20 NiCrMo 2
25 CrMo 4
25 CrMo 4
25 CrMo 4 KB
25 CrMo 4 KB
30 NiCrMo 2 KB
30 NiCrMo 2 KB
34 Cr 4 KB
34 Cr 4 KB
34 CrMo 4 KB
34 CrMo 4 KB
35 CrMo 4
35 CrMo 4
35 CrMo 4 F
35 CrMo 4 F
38 Cr 4 KB
38 Cr 4 KB
38 CrB 1 KB
38 CrB 1 KB
38 CrMo 4
38 CrMo 4 KB
38 CrMo 4 KB
38 CrMo 4 KB
38 CrMo 4 KB
38 NiCrMo 4
40 Cr 4
40 Cr 4
40 CrMo 4
40 CrMo 4
40 CrMo 4
40 CrMo 4
40 NiCrMo 2 KB
40 NiCrMo 2 KB
40 NiCrMo 2 KB
40 NiCrMo 2 KB
40 NiCrMo 7
40 NiCrMo 7
40 NiCrMo 7 KB
40 NiCrMo 7 KB
41 Cr 4 KB
41 Cr 4 KB
50 CrV 4
50 CrV 4
55 Si 8
58 WCr 9 KU
88 MnV 8 KU
100 Cr 6
C 20
C 35
C 60
CB 10 FU
CB 35
G 22 Mn 3
G 22 Mo 5
G 22 Mo 5
G 22 Mo 5
G 40 CrMo 4
G 40 CrMo 4
G 40 CrMo 4
(continued)
AISI
UNI
1213
12L13
12L14
1212
8617H
8617
8620
8620H
4130
4130H
4130
4130H
8630
8630H
5130H
5132
4135
4135H
4135
4135H
4135
4135H
5132H
5135
50B40
50B40H
4142H
4137
4137H
4140
4140H
9840
5135H
5140
4137
4137H
4140
4140H
8640
8640H
8740
8740H
E4340
E4340H
E4340
E4340H
5135H
5140
6150
6150H
9255
S1
O2
E52100
1020
1035
1060
1008
1030
1022
4419
4419H
4422
4137
4137H
4140
21/8/2008 5:04PM Plate # 0
pg 566
Table A8.1
(continued)
Country
ITALY (continued)
Designation
G 40 CrMo 4
ICL 472 T
X 2 CrNi 18 11
X 2 CrNi 18 11
KG
X 2 CrNi 18 11
KW
X 2 CrNiMo 17 12
X 3 Cr Ni 18 11
X 5 CrNi 18 10
X 5 CrNiMo 17 12
X 5 CrNiMo 18 15
X 6 CrAl 13
X 6 CrNi 23 14
X 6 CrNiTi 18 11
X 6 CrNiTi 18 11
KG
X 6 CrNiTi 18 11
KT
X 6 CrNiTi 18 11
KW
X 8 Cr 17
X 8 CrMo 17
X 8 CrNi 19 10
8 CrNi 19 10
X 8 CrNiNb 18 11
X 10 CrNiS 18 09
X 10 CrS 17
X 12 Cr 13
X 12 CrNi 17 07
X 12 CrS 13
X 16 CrNi 16
X 16 CrNi 23 14
X 16 CrNiSi 25 20
X 16 CrNiSi 25 20
X 20 Cr 13
X 22 CrNi 25 20
X 22 CrNi 25 20
X 28 W 09 KU
X 35 CrMo 05
KU
X 35 CrMoV 05
KU
X35 CrMoW 05
KU
X 75 W 18 KU
X 78 WCo 1805
KU
X 80 WCo 1810
KU
X 82 MoW 09 KU
X 82 MoW 09 KU
X 82 WMo 0605
KU
X 150 CrMo 12
KU
X 150 WCoV
130505 KU
X 210 Cr 13 KU
JAPAN
Country
AISI
JAPAN (continued)
UNI
S 15 CK
S 15 CK
S 17 C
S 17 C
S 20 C
S 20 CK
S 22 C
S 25 C
S 28 C
S 38 C
S 40 C
S 45 C
S 45 C
S 48 C
S 48 C
S 53 C
S 55 C
SCCrM 1
SCCrM 1
SCCrM 3
SCCrM 3
SCM 1
SCM 1
SCM 2
SCM 2
SCM 4
SCM 4
SCM 4
SCM 4
SCM 4 H
SCM 4 H
SCM 4 H
SCM 4 H
SCM 5
SCM 5
SCM 5
SCM 5
SCM 5 H
SCM 5 H
SCM 5 H
SCM 5 H
SCMn 2
SCMn 2
SCMn 3
SCMn 3
SCPH 11
SCPH 11
SCPH 11
SCr 2
SCr 2
SCr 2H
SCr 2H
SCr 3 H
SCr 3 H
SCr 4 H
SCr 4 H
SCS 19
SKD 1
SKD 5
SKD 6
SKD 12
SKD 61
SKD 62
SKH 2
SKH 3
SKH 4A
4140H
321
304L
304L
304L
316L
304L
304
316
317
405
309S
321
321
321
321
430
434
305
308
347
303
430F
410
301
416
431
309
310
314
420
310
314
H21
H11
H13
H12
T1
T4
T5
H41
M1
M2
D2
T15
D3
JIS
S 9 CK
S 10 C
S 12 C
S 15 C
S 15 C
Designation
1010
1010
1010
1015
1017
(continued)
AISI
JIS
1015
1017
1015
1017
1023
1023
1023
1025
1025
1034
1040
1045
1045H
1045
1045H
1050
1050
4130
4130H
4315
4135H
4135
4135H
4130
4130H
4137
4137H
4140
4140H
4137
4137H
4140
4140H
4147
4147H
4150
4150H
4147
4147H
4150
4150H
1330
1330H
1335
1335H
4419
4419H
4422
5130H
5132
5130H
5132
5132H
5135
5135H
5140
304L
D3
H21
H11
A2
H13
H12
T1
T4
T5
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 567

Country
JAPAN (continued)
Designation
Country
AISI
SWEDEN
JIS
SKH 9
SKH 52
SKH 53
SMn 1 H
SMn 1 H
SMn 2
SMn 2
SMn 2 H
SMn 2 H
SMnC 21
SNCM 8
SNCM 8
SNCM 21
SNCM 21
SNCM 21
SNCM 21
SNCM 21 H
SNCM 21 H
SNCM 21 H
SNCM 21 H
SUH 309
SUH 310
SUH 409
SUH 616
SUM 11
SUM 12
SUM 21
SUM 22
SUM 22 L
SUM 22 L
SUM 23 L
SUM 24 L
SUM 24 L
SUP 4
SUP 10
SUP 10
SUP 11
SUP 11
SUS 301
SUS 303
SUS 304
SUS 304 L
SUS 305
SUS 305
SUS 305 J1
SUS 305 J1
SUS 316
SUS 316 L
SUS 317
SUS 321
SUS 347
SUS 403
SUS 405
SUS 410
SUS 410S
SUS 416
SUS 420 J1
SUS 430
SUS 430
SUS 431
SUS 434
SUS 440 B
SUS 440 C
SUS Y 310
SUS Y 310
SUS Y 316
M2
M3 (Class 2)
M3 (Class 2)
1330
1330H
1335
1335H
1335
1335H
1022
4340
4340H
8617
8617H
8620
8620H
8617
8617H
8620
8620H
316
316L
409
422
1109
1109
1212
1213
12L13
12L14
12L13
12L13
12L14
1095
6150
6150H
50B50
50B50H
301
303
304
304L
305
308
305
308
316
316L
317
321
347
403
405
410
410S
403
420
430
430F
431
434
440B
440C
310
314
316
1370
1370
1450
1550
1665
1672
1672
1678
1770
1778
1870
1914
1914
2090
2120
2120
2225
2225
2230
2230
2234
2234
2242
2244
2244
2244
2244
2260
2302
2303
2320
2325
2332
2337
2338
2346
2347
2348
2352
2367
2383
2722
AISI
1015
1017
1020
1035
1064
1045
1045H
1064
1070
1070
1095
12L13
12L14
9255
1335
1335H
4130
4130H
6150
6150H
4135
4135H
H13
4137
4137H
4140
4140H
A2
410
420
430
434
304
321
347
303
316
316L
304L
317L
430F
M2
UNITED KINGDOM B.S.
(continued)
Designation
SS14
040 A 20
060 A 35
060 A 62
060 A 96
070 M 20
080 A 32
080 A 35
080 A 37
080 A 40
080 M 36
2 S 93
2 S 117
2 S 117
2 S 119
2 S 119
2 S 130
2 S 516
2 S 516
2 S 517
2 S 517
3 S 95
3 S 95
1020
1035
1060
1095
1020
1035
1035
1035
1040
1035
1040
5135H
5140
4340
4340H
348
1345
1345H
1345
1345H
4340
4340H
21/8/2008 5:04PM Plate # 0
pg 568
Table A8.1
Country
(continued)
Designation
UNITED KINGDOM B.S.

(continued)
5 S 80
120 M 36
150 M 36
220 MO7
250 A 53
250 A 58
250 A 58
302 S 17
303 S 21
304 S 12
304 S 14
304 S 15
304 S 16
304 S 18
304 S 22
304 S 25
304 S 40
310 S 24
310 S 24
316 S 12
316 S 14
316 S 16
316 S 18
316 S 22
316 S 24
316 S 25
316 S 26
316 S 29
316 S 30
316 S 30
316 S 31
316 S 37
316S 40
316 S 41
316 S 82
317 S 12
321 S 12
321 S 18
321 S 22
321 S 27
321 S 40
321 S 49
321 S 50
321 S 59
321 S 87
347 S 17
347 S 17
347 S 18
347 S 40
403 S 17
405 S 17
409 S 17
410 S 21
416 S 21
420 S 29
420 S 37
430 S 15
431 S 29
434 S 19
530 A 30
530 A 32
530 A 32
530 A 36
530 A 36
530 A 40
530 A 40
Country
AISI
Designation
UNITED KINGDOM B.S.

(continued)
530 H 30
530 H 32
530 H 32
530 H 36
530 H 36
530 H 40
530 H 40
530 M 40
530 M 40
534 A 99
535 A 99
640 M 40
708 A 37
708 A 37
708 A 42
708 A 42
708 A 42
708 A 42
708 M 40
708 M 40
708 M 40
708 M 40
708 A 40
709 M 40
709 M 40
709 M 40
735 A 50
735 A 50
805 A 20
805 A 20
805 A 20
805 A 20
805 A 20
805 A 20
805 H 20
805 H 20
805 H 20
805 H 20
805 M 20
805 M 20
805 M 20
805 M 20
816 M 40
817 M 40
817 M 40
3111 Type 6
3111 Type 6
ANC 1 Grade A
ANC 3 Grade B
BA 2
BD 2
BD 3
BH 11
BH 12
BH 13
BH 21
BM 1
BM 1
BM 2
BM 34
BM 34
BO 1
BO 2
BT 1
BT 4
BT 5
431
1039
1039
1213
9255
9260
9260H
304
303
304L
304L
304
304
304
304L
304
304
310
314
316L
316L
316
316
316L
316L
316
316
316L
316
316L
316L
316L
316
316
316L
317L
321
321
321
321
321
321
321
321
321
347
348
348
348
410S
405
409
410
416
403
420
430
431
434
5130
5130H
5132
5132H
5135
5135H
5140
(continued)
AISI
5130
5130H
5132
5132H
5135
5135H
5140
5135H
5140
E52100
E52100
3140
4135
4135H
4137
4137H
4140
4140H
4137
4137H
4140
4140H
4137H
4137
4140
4140H
6150
6150H
8622
8622H
8720
8720H
8822
8822H
8617
8617H
8620
8620H
8617
8617H
8620
8620H
9840
4340
4340H
4340
4340H
410
347
A2
D2
D3
H11
H12
H13
H21
H41
M1
M2
M33
M34
O1
O2
T1
T4
T5
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 569

Country
Designation
UNITED KINGDOM B.S.

(continued)
BT 15
CDS-18
CDS-20
CDS 105/106
CDS 110
CDS 110
En. 44 B
En. 47
En. 47
En. 56 A
En. 56 B
En. 58 B
En. 58 C
En. 58 E
En. 58 F
En. 58 G
En. 58 H
S. 139
S. 139
S. 525
S. 527
S. 536
S. 537
Type 3
Type 3
Type 7
Type 7
Type 7
Type 7
Type 8
Type 8
UNITED STATES
Country
AISI
UNITED STATES
(continued)
T15
420
321
1039
4130
4130H
1095
6150
6150H
410
403
321
321
304
347
347
316
E4340
E4340H
348
348
304L
316L
5132H
5135
8640
8640H
8740
8740H
E4340
E4340H
AMS
5010 D
5024 C
5032
5040
5042
5044
5045
5047
5053
5060
5069
5070
5075
5077
5080
5082
5085
5110
5112
5112 E
5115
5120 D
5121
5122
5132
5331
5333
5342
5343
5344
5354
5355
5369
1212
1137
1020
1010
1010
1010
1020
1010
1010
1015
1018
1022
1025
1025
1035
1035
1050
1080
1090
1086
1070
1074
1095
1095
1095
4340
8615
630
630
630
615
630
651
(continued)
Designation
AISI
AMS
5376
5501
5502
5503
5504
5505
5506
5507
5508
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5531
5532
5546
5547
5548
5549
5554
5556
5557
5558
5559
5560
5561
5565
5566
5567
5568
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5585
5591
5592
5594
5602
5604
5610
5611
5612
5613
661
304
501
430
410
410
420
316L
615
321
304L
347
304
305
302
302
301
301
301
632
310S
314
309S
316
660
651
651
631
631
661
661
633
634
633
634
633
347
321
347
321
304
304
304
304
631
321
347
310S
316
309S
347
321
310S
651
661
410
330
634
501
630
416Se
403
403
410
21/8/2008 5:04PM Plate # 0
pg 570
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
AMS
5615
5615
5618
5620
5620
5621
5622
5626
5627
5628
5630
5631
5632 (Type 1)
5632 (Type 2)
5636
5637
5639
5640 (Type 1)
5640 (Type 2)
5641
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5657
5673
5674
5678
5680
5681
5685
5686
5688
5689
5690
5691
5694
5695
5697
5716
5720
5721
5722
5731
5732
5734
5735
5736
5737
5738
5742
5743
5744
5745
5768
5769
5774
414
615
440C
420F
420F(Se)
420
630
T1
430
431
440C
440A
440F
440F(Se)
302
302
304
303
303Se
303Se
630
631
321
347
304L
316
316F
309S
310S
314
316L
347
422
632
631
347
631
347
347
305
305
302
321
316
316
310
310
304
330
651
651
651
660
660
660
660
660
660
303Se
634
634
634
633
661
661
633
(continued)
Designation
AISI
AMS
5775
5776
5780
5781
5794
5795
5804
5805
5812
5813
5817
5821
5824
5825
6260 F
6265 B
6272
6274
6275
6275 A
6276
6277
6280
6281
6290
6294
6322
6323
6325
6327
6342 C
6350
6355
6356
6358
6359
6360
6361
6362
6365 C
6370
6371
6372 C
6373
6381
6382
6390
6395
6414
6415
6437
6440
6441
6442 B
6443
6444
6444
6446
6447
6448
6449
6450
6455
6466
6467
6485
633
410
634
634
661
661
660
660
632
632
615
410
631
630
E9310
E9310
8617
8620
94B17
94B15
8620
8620
8630
8630
4615
4620
8740
8740
8740
8740
9840
4130
8630
4130
8740
4340
4130
4130
4130
4135
4130
4130
4135
4130
4140
4140
4140
4140
4340
4340
H11
E52100
E52100
E50100
E51100
E52100
E52100
E51100
E52100
6150
E51100
6150
6150
502
502
H11
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 571

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
AMS
6487
6488
6530
6550
7240
7301
7304
7470
H11
H11
8630
8630
1060
6150
1095
615
ASME
5041
SA182
SA182
SA182
SA182
SA182
SA182
SA182
SA182
SA182
SA182
SA182
SA193
SA193
SA193
SA193
SA194
SA194
SA194
SA194
SA194
SA194
SA194
SA194
SA194 (Type 3)
SA194 (Type 6)
SA194 (Type 8)
SA213
SA213
SA213
SA213
SA213
SA213
SA213
SA213
SA213
SA213
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
SA240
1006
304
304L
304N
310
316
316L
316N
321
347
348
430
305
316
321
347
303
303Se
305
316
321
347
416
416Se
501
410
304
304
304L
304N
310
316
316L
316N
321
347
348
302
304
304L
304N
305
309S
310S
316
316L
316N
317
317L
321
347
348
405
410
410S
430
(continued)
Designation
AISI
ASME
SA240 (XM-8)
SA240 (XM-21)
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249
SA249 (XM-19)
SA249 (XM-29)
SA268
SA268
SA268
SA268
SA268
SA268
SA268 (XM-8)
SA268 (XM-27)
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312
SA312 (XM-19)
SA312 (XM-29)
SA320
SA320
SA320
SA320
SA320 (B8)
SA320 (B8C)
SA358
SA358
SA358
SA358
SA358
SA358
SA358
SA358
SA358
SA376
SA376
SA376
SA376
SA376
SA376
SA376
SA387 (Type 5)
SA387 (Type 5)
SA403
SA403
SA403
SA403
S43035
304HN
304
304L
304N
309
310
316
316L
316N
317
321
347
348
329
405
409
410
430
446
S43035
304
304L
304N
309
310
316
316L
316N
317
321
347
348
303
303Se
316
321
304
347
304
304N
309
310
316
316N
321
347
348
304
304N
316
316N
321
347
348
501
502
304
304L
309
310
21/8/2008 5:04PM Plate # 0
pg 572
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASME
SA403
SA403
SA403
SA403
SA403
SA403
SA403
SA403 (XM-19)
SA409
SA409
SA409
SA409
SA409
SA409
SA409
SA409
SA412
SA412
SA412 (XM-19)
SA430
SA430
SA430
SA430
SA430
SA430
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479
SA479 (XM-8)
SA479 (XM-19)
SA479 (XM-27)
SA564
SA564 (XM-25)
SA638
SA688
SA688
SA688
SA688
SA688 (XM-29)
SA705
SA705
SA705 (XM-12)
SA705 (XM-13)
SA705 (XM-25)
SA737 (XM-27)
316
316L
316N
317
321
347
348
304
309
310
316
317
321
347
348
201
304
304N
316
316N
321
347
302
304
304L
304N
310S
316
316L
316N
321
347
348
405
410
430
S43035
630
660
304
304L
316
316L
630
631
ASTM
A26
A29
A29
A29
A29
A29
A29
A29
1064
1005
1006
1008
1010
1012
1015
1016
(continued)
Designation
AISI
ASTM
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A29
A36
A57
A59
A107
A107
A107
A108
A108
A108
A108
A108
A108
A108
A108
A108
A108
1017
1018
1019
1020
1021
1022
1023
1025
1026
1030
1034
1035
1038
1038H
1039
1040
1044
1045
1045H
1046
1050
1055
1059
1060
1064
1065
1070
1074
1080
1086
1090
1095
1108
1109
1116
1119
1132
1137
1141
1144
1211
1212
1213
12L13
12L14
1215
1547
1548
15B48H
9260
1064
9260
1117
1118
1141
1008
1010
1015
1016
1017
1018
1020
1030
1035
1040
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 573

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A108
A108
A108
A108
A108
A108
A108
A108
A108
A108
A131
A131 (A)
A131 (AH32, DH32, EH32)
A131 (AH36, DH36, EH36)
A131 (B)
A131 (CS, DS)
A131 (D)
A131 (E)
A135
A139
A139 (B)
A139 (C)
A139 (D)
A139 (E)
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A167
A176
A176
A176
A176
A176
A176
A176
A176
A177
A181
A182
A182
A182
A182
A182
A182
A182
A182
A182
A182
A182
A193
A193
A193
1117
1118
1137
1141
1144
1211
1212
1213
12L14
1215
301
302
302B
304
304L
305
308
309
309S
310
310S
316
316L
317
317L
321
347
348
403
405
409
410
410S
430
442
446
301
1034
304
304L
304N
310
316
316L
316N
321
347
348
430
304
316
321
(continued)
Designation
AISI
ASTM
A193
A193
A193
A194
A194
A194
A194
A194
A194
A194
A194
A194
A194
A194 (grade 8F)
A213
A213
A213
A213
A213
A213
A213
A213
A213
A213
A228
A229
A230
A231
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240
A240 (XM-8)
A240 (XM-21)
A249
A249
A249
A249
A249
A249
A249
A249
A249
A249
A249
A249
A249
A268
A268
A268
A268
A268
347
410
501
303
303Se
304
316
321
347
410
416
416Se
501
303Se
304
304L
304N
310
316
316L
316N
321
347
348
1086
1065
1064
6150
302
304
304L
304N
305
309S
310S
316
316L
316N
317
317L
321
348
405
410
410S
430
S43035
304HN
304
304L
304N
305
309
310
316
316L
316N
317
321
347
348
329
405
409
430
430Ti
21/8/2008 5:04PM Plate # 0
pg 574
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A268
A268
A268 (XM-8)
A269
A269
A269
A269
A269
A269
A269
A270
A271
A271
A271
A273
A274
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276
A276 (XM-21)
A276 (XM-27)
A284 (C)
A295
A295
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
443
446
S43035
304
316
316L
317
321
347
348
304
304
321
347
1026
9840
302
302B
304
304L
304N
305
308
309
309S
310
310S
314
316
316L
316N
317
321
347
348
403
405
410
414
420
430
431
440A
440B
440C
446
304HN
304HN
304HN
E50100
E51100
1330H
1335H
1340H
1345H
15B48H
4027H
4028H
4037H
4042H
4047H
4118H
4130H
4135H
4137H
4140H
(continued)
Designation
AISI
ASTM
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A304
A311
A311
A311
A312
A312
A312
A312
4142H
4145H
4147H
4150H
4161H
4320H
4340H
4419H
4620H
4626H
4718H
4720H
4815H
4817
4817H
4820H
50B40H
50B44H
5046H
50B46H
50B50H
50B60H
5120H
5130H
5132H
5135H
5140H
5150H
5155H
5160H
51B60H
6118
6150H
81B45H
8617H
8620H
8622H
8625H
8627H
8630H
86B30H
8637H
8640H
8642H
8645H
86B45H
8650H
8655H
8660H
8720H
8740H
8822H
9260H
94B15H
94B17H
94B30H
E4340
E4340H
E9310H
1137
1141
1144
304
304L
304N
309
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 575

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A312
A312
A312
A312
A312
A312
A312
A312
A313
A313
A313
A313
A313
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A314
A320
A320
A320
A320
A320
A320
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
310
316
316L
316N
317
321
347
348
302
304
305
316
631
202
302
302B
303
303Se
304
304L
305
308
309
309S
310
310S
314
316
316L
317
321
347
348
403
405
410
414
416
416Se
420
430
430F
430F(Se)
431
440A
440B
440C
446
501
502
303
303Se
304
316
321
347
1330
1335
1340
1345
3140
4023
4024
4027
4028
4037
(continued)
Designation
AISI
ASTM
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A322
A331
4042
4047
4118
4130
4137
4140
4142
4145
4147
4150
4161
4320
4340
4419
4615
4620
4626
4718
4720
4815
4817
4820
50B40
50B44
50B46
50B50
50B60
5120
5130
5132
5135
5140
5150
5155
5160
51B60
6118
6150
81B45
8615
8617
8620
8622
8625
8627
8630
8637
8640
8642
8645
8650
8655
8660
8720
8740
8822
9255
9260
94B17
94B30
94B40
9840
E9310
E51100
E52100
1330
21/8/2008 5:04PM Plate # 0
pg 576
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A331
A332
A331
A331
A331
A331
A331
A331
A331
A331
A355
A358
A358
A358
1335
1340
1345
3140
4023
4024
4027
4028
4037
4042
4047
4118
4130
4137
4140
4142
4145
4147
4150
4161
4320
4340
4419
4615
4620
4626
4718
4720
4815
4817
4820
50B60
5120
5130
5132
5135
5140
5150
5155
5160
51B60
6150
8617
8620
8622
8625
8627
8630
8637
8640
8642
8645
8655
8660
8720
8740
8822
9260
94B17
94B30
E4340
E52100
4135
304N
309
310
(continued)
Designation
AISI
ASTM
A358
A358
A358
A358
A358
A368
A368
A368
A368
A376
A376
A376
A376
A376
A376
A376
A387 (5)
A387 (5)
A403
A403
A403
A403
A403
A403
A403
A403
A403
A403
A403
A409
A409
A409
A409
A409
A409
A409
A412
A412
A429
A429
A429 (XM-19)
A430
A430
A430
A430
A430
A430
A453
A453
A457
A458
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
316
316N
321
347
348
302
304
305
316
304
304N
316
316N
321
347
348
501
502
304L
304N
309
310
316
316L
316N
317
321
347
348
304
309
310
316
317
321
347
201
202
201
202
304
304N
316
316N
321
347
651
660
651
651
202
302
302B
303
303Se
304
304L
305
308
309
309S
310
310S
314
316
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 577

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A473
A477
A478
A478
A478
A478
A478
A478
A478
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479
A479 (XM-8)
A492
A492
A492
A493
A493
A493
A493
A493
A493
A493
A493
A493
A493
A505
A505
A505
A505
A505
A505
A505
316L
317
321
347
403
405
410
410S
414
416
416Se
420
430
430F
430F(Se)
431
440A
440B
440C
446
501
502
651
302
304
304L
305
316
316L
317
302
304
304L
304N
310S
316
316L
316N
321
347
348
403
405
410
430
S43035
302
304
305
302
304
305
321
347
384
410
430
431
440C
4118
4130
4137
4140
4142
4145
4147
(continued)
Designation
AISI
ASTM
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A505
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A510
A511
A511
A511
4150
4320
4340
4615
4620
4718
4815
4820
5130
5132
5140
5150
5160
6150
8615
8617
8620
8630
8640
8642
8645
8650
8655
8660
8720
8740
9260
E4340
E51100
E52100
1005
1006
1008
1010
1012
1015
1016
1017
1018
1019
1020
1021
1022
1023
1025
1026
1030
1035
1038
1039
1040
1044
1045
1046
1050
1055
1060
1070
1080
1090
1095
1547
1548
302
304
304L
21/8/2008 5:04PM Plate # 0
pg 578
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A511
A512
A512
A513
A513
A513
A513
A513
A513
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
305
309
309S
310
310S
316L
317
321
329
347
403
405
410
414
416Se
430
440A
443
446
1025
1030
1016
1017
1018
1019
4130
8620
1008
1010
1012
1015
1017
1018
1019
1020
1021
1022
1025
1026
1030
1035
1040
1045
1050
1330
1335
1340
1345
3140
4023
4024
4027
4028
4037
4042
4047
4118
4130
4135
4137
4140
4142
4145
4147
4150
4320
(continued)
Designation
AISI
ASTM
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A519
A534
A535
A535
A535
A535
A535
A544
A544
A544
A544
A544
A544
A544
A545
A545
A545
A545
A545
A545
A545
A545
A545
4340
4422
4427
4720
4817
4820
50B40
50B44
5046
50B46
50B50
50B60
5120
5130
5132
5135
5140
5150
5155
5160
51B60
81B45
8630
8637
8640
8642
8645
86B45
8650
8660
8720
8740
8822
9260
94B15
94B17
94B30
94B40
9840
E4340
E9310
E50100
E51100
E52100
4023
4320
4620
4720
4820
E52100
1017
1018
1020
1022
1030
1035
1038
1006
1008
1010
1012
1015
1016
1018
1019
1021
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 579

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A545
A545
A545
A545
A545
A546
A546
A546
A546
A546
A547
A547
A547
A547
A547
A547
A547
A548
A548
A548
A548
A548
A549
A549
A549
A549
A549
A549
A549
A554
A554
A554
A554
A554
A554
A554
A554
A554
A554
A554
A554
A554
A554
A564
A564
A564
A564
A565
A565
A567
A575
A575
A575
A575
A575
A575
A575
A575
A575
A576
A576
A576
A576
A576
A576
A576
1022
1026
1030
1035
1038
1030
1035
1038
1039
1040
1335
1340
4037
4137
4140
4142
4340
1016
1018
1019
1021
1022
1008
1010
1012
1015
1016
1017
1018
301
302
304
304L
305
309
309S
309S(Cb)
310S
316L
317
347
430
430Ti
630
631
632
634
422
615
661
1008
1010
1012
1015
1017
1020
1023
1025
1044
1008
1010
1012
1015
1016
1017
1018
(continued)
Designation
AISI
ASTM
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A576
A579
A579
A579 (grade 61)
A579 (grade 62)
A579 (grade 63)
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A580
A581
A581
A581
A581
A581
A581
A581 (XM-2)
A582
A582
A582
A582
A582
A582
1019
1020
1021
1022
1023
1025
1026
1030
1035
1038
1039
1040
1044
1045
1046
1050
1055
1060
1070
1080
1090
1095
1547
1548
632
634
633
631
431
302B
304
304L
305
308
309
309S
310S
314
316L
317
347
348
403
405
410
414
420
430
431
440A
440B
440C
446
303
303Se
416
416Se
430F
430F(Se)
303MA
303
303Se
416
416Se
420F(Se)
430F
21/8/2008 5:04PM Plate # 0
pg 580
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
ASTM
A582
A582 (XM-2)
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A600
A632
A632
A632
A632
A632
A632
A633
A638
A639
A646
A646
A646
A646
A651
A651
A651
A651
A651
A651 (XM-8)
A659
A659
A659
A659
A659
A659
A659
A666
A666
A666
A666
A666
A681
A681
A681
A681
A681
A681
A681
A681
430F(Se)
303MA
M1
M2
M3 (Class 1)
M3 (Class 2)
M4
M6
M7
M10
M30
M33
M34
M36
M41
M42
M43
M44
M46
M47
T1
T2
T4
T5
T6
T8
T15
304
304L
310
316L
317
348
347
660
661
4130
4140
4340
E52100
304
409
430
430Ti
434
S43035
1015
1016
1017
1018
1020
1021
1023
201
202
301
302
304
A2
A3
A4
A5
A6
A7
A8
A9
UNITED STATES
(continued)
(continued)
Designation
AISI
ASTM
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A681
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A682
A686
A686
A688
A688
A688
A693
A693
A693
A693
A699
A705
A705
A705
A10
D2
D3
D4
D5
D7
H10
H11
H12
H13
H14
H19
H21
H22
H23
H24
H25
H26
H41
H42
H43
O1
O2
O6
O7
P2
P3
P4
P5
P6
P20
P21
S1
S2
S4
S5
S6
S7
1030
1035
1040
1045
1050
1055
1060
1064
1065
1070
1074
1080
1086
1095
W1
W5
304
304L
316L
630
631
632
633
634
630
631
632
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 581

Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
ASTM
A705
A711
A711
A711
A711
A711
A711
A711
A711
B511
B512
B535
B536
B546
634
3140
4135
4720
8660
E9310
E50100
E51100
E52100
330
330
330
330
330
FED
QQ-S-633
(C12L13)
QQ-S-635 (C1030)
QQ-S-635 (C1035)
QQ-S-635 (C1045)
QQ-S-635 (C1050)
QQ-S-637
QQ-S-637
QQ-S-637 (C1008)
QQ-S-637 (C1109)
QQ-S-637 (C1116)
QQ-S-637 (C1117)
QQ-S-637 (C1118)
QQ-S-637 (C1119)
QQ-S-637 (C1132)
QQ-S-637 (C1137)
QQ-S-637 (C1144)
QQ-S-637 (C1211)
QQ-S-637 (C1212)
QQ-S-637 (C1913)
QQ-S-698 (C1008)
QQ-S-698 (C1015)
QQ-S-700 (C1025)
QQ-S-700 (C1030)
QQ-S-700 (C1035)
QQ-S-700 (C1045)
QQ-S-700 (C1050)
QQ-S-700 (C1055)
QQ-S-700 (C1065)
QQ-S-700 (C1074)
QQ-S-700 (C1080)
QQ-S-700 (C1086)
QQ-S-700 (C1095)
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
12L13
1030
1035
1045
1050
1141
1215
1008
1109
1116
1117
1118
1119
1132
1137
1144
1211
1212
1213
1008
1015
1025
1030
1035
1045
1050
1055
1065
1074
1080
1086
1095
202
302
304
304L
305
309
310
316
316L
317
321
347
403
405
410
414
420
(continued)
Designation
AISI
FED
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-763
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-S-766
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-570
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
430
440A
440B
440C
446
201
202
301
302
304
304L
309
310
316
316L
321
347
348
420
430
446
A2
A3
A4
A5
A6
A7
A8
A9
A10
D2
D3
D4
D5
D7
H10
H11
H12
H13
H14
H19
H21
H22
H23
H24
H25
H26
H41
H42
H43
O1
O2
O6
O7
S1
S2
S4
S5
S6
M1
M2
M3 (Class 1)
M3 (Class 2)
M4
M6
M7
21/8/2008 5:04PM Plate # 0
pg 582
Table A8.1
(continued)
Country
UNITED STATES
(continued)
Designation
Country
AISI
UNITED STATES
(continued)
FED
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-T-590
QQ-W-412
QQ-W-412 (II)
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-423
QQ-W-461
STD-66
STD-66
STD-66
STD-66
M10
M30
M33
M34
M36
M41
M42
M43
M44
M46
T1
T2
T4
T5
T6
T8
T15
302
304
305
310
316
321
347
410
416
420
430
1006
202
304
310
430
MIL SPEC
MIL-C-24111
MIL-F-20138
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
MIL-S-862
630
304
302
303
303Se
304
304L
309
310
316
316L
317
321
347
403
405
410
416
416Se
420
430
430F
430F(Se)
431
440A
440B
440C
440F
440F(Se)
(continued)
Designation
AISI
MIL SPEC
MIL-S-862
MIL-S-866
MIL-S-866
MIL-S-980
MIL-S-5000
MIL-S-5059
MIL-S-5059
MIL-S-5059
MIL-S-6049
MIL-S-7420
MIL-S-7493
(A4615)
MIL-S-7493
(A4620)
MIL-S-8503
MIL-S-11310
(CS1005)
MIL-S-11310
(CS1006)
MIL-S-11310
(CS1008)
MIL-S-11310
(CS1010)
MIL-S-11310
(CS1012)
MIL-S-11310
(CS1017)
MIS-L-11310
(CS1018)
MIL-S-11310
(CS1020)
MIL-S-11310
(CS1022)
MIL-S-11310
(CS1025)
MIL-S-11310
(CS1030)
MIL-S-11310
(CS1040)
MIL-S-11310
(ORD4150)
MIL-S-11713 (2)
MIL-S-16788 (C10)
MIL-S-16974
MIL-S-16974
MIL-S-16974
(Gr. 1060)
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-16974
MIL-S-18411
MIL-S-18411
MIL-S-18733
MIL-S-20166
446
1016
8615
E52100
E4340
301
304
316
8740
E52100
4615
4620
6150
1005
1006
1008
1010
1012
1017
1018
1020
1022
1025
1030
1040
4150
1070
1095
1015
1050
1060
1080
1330
1335
1340
3140
4130
4135
4140
4145
4340
8620
8625
8630
8640
8645
1117
12L13
4135
Appendix 8:
21/8/2008 5:04PM Plate # 0
pg 583

Country
UNITED STATES
(continued)
Designation
AISI
MIL SPEC
(CS1116)
MIL-S-22141
MIL-S-22698 (B)
MIL-S-23195
MIL-S-23195
MIL-S-23195
MIL-S-23195
MIL-S-23196
MIL-S-23196
MIL-S-23196
MIL-S-23196
1116
E52100
304
304L
347
348
304
304L
347
348
Country
UNITED STATES
(continued)
Designation
AISI
MIL SPEC
MIL-S-25043
MIL-S-46042
MIL-S-46049
MIL-S-46049
MIL-S-46409
MIL-S-81506
MIL-S-81591
MIL-T-6845
MIL-T-8504
MIL-T-8506
MIL-W-46078
631
651
1065
1074
1065
630
630
304
304
304
631
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_585.pdf/Appendix_9/

DOI: 10.1361/faht2008p585
18/8/2008 4:38PM Plate # 0

APPENDIX 9
Iron-Carbon Equilibrium Diagram
Fig. A9.1
Iron-carbon equilibrium diagram from Metal Progress Materials and Process Engineering Databook, American Society for
Metals, 1968, p39
pg 585
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_587-599.pdf/F5113_Appendix-10/

DOI: 10.1361/faht2008p587
18/8/2008 4:39PM Plate # 0

APPENDIX 10
Isothermal Diagrams of Selected Steels

A selection of isothermal diagrams for

Carbon steels (1019, 1030, 1050, 1060, 1080)

Cr-Mo steels (4130, 4140)
Ni-Cr-Mo steels (4340, 8620)
Ni-Mo Steel (4640)
Cr steel (5160, 52100)
REFERENCE
1. From Atlas of Isothermal Transformation

and Cooling Transformation Diagrams,
pg 587
Fig. A10.1
Carbon steels, 1019. Source: Ref 1
18/8/2008 4:40PM Plate # 0
pg 588
18/8/2008 4:40PM Plate # 0
Appendix 10: Isothermal Diagrams of Selected Steels / 589
Fig. A10.2
pg 589
Fig. A10.3
18/8/2008 4:40PM Plate # 0
pg 590
18/8/2008 4:40PM Plate # 0
Fig. A10.4
pg 591
Fig. A10.5
18/8/2008 4:40PM Plate # 0
pg 592
18/8/2008 4:41PM Plate # 0
Fig. A10.6
Chromium-molybdenum steels, 4130. Source: Ref 1
pg 593
Fig. A10.7
Chromium-molybdenum steels, 4140. Source: Ref 1
18/8/2008 4:41PM Plate # 0
pg 594
18/8/2008 4:41PM Plate # 0
Fig. A10.8
Ni-Cr-Mo steels, 4340. Source: Ref 1
pg 595
Fig. A10.9
Nickel-molybdenum steels, 4640. Source: Ref 1
18/8/2008 4:41PM Plate # 0
pg 596
18/8/2008 4:41PM Plate # 0
Fig. A10.10
Chromium steels, 5160. Source: Ref 1
pg 597
Fig. A10.11
Nickel-chromium-molybdenum steels, 8620. Source: Ref 1
18/8/2008 4:42PM Plate # 0
pg 598
18/8/2008 4:42PM Plate # 0
Fig. A10.12
Chromium steels, 52100. Source: Ref 1
pg 599
18/8/2008 4:47PM Plate # 0

DOI: 10.1361/faht2008p601

APPENDIX 11
Continuous Cooling Diagrams of

Selected Steels
Selected continuous cooling transformation
diagrams for:

Carbon steels with nominal carbon contents

of 0.8, 0.44, 0.86, 0.96 wt% C
Mn steels (13/4 M, 11/2 Mn)
Mn-Mo, Mn-Ce, Mn-Ni-Mo, and Mn-Ni-CrMo steels
Silicon steels (Si-Mn, Si-Cr, Si-Cr-Mo)
1/4 Mo steel
Nickel steels (31/2 Ni, 11/2 Ni-Mn, 13/4 Ni-Mo,
11/4 NiCr)

Ni-Cr-Mo steels (1/2 Ni-Cr-Mo, 11/2 Ni-CrMo, 31/2 Ni-Cr-Mo)

Chromium steels (1/2 Cr, 1Cr, 11/4 Cr-Mo,
1Cr-V, 11/2 Cr-Al-Mo)
REFERENCE
1. From Atlas of Continuous Cooling Transformation Diagrams for Engineering Steels,

pg 601
18/8/2008 4:48PM Plate # 0
pg 602
Si
0.18 0.20
Mn
Cr
Mo
Ni
Al
Nb
0.45 0.020 0.020
900
800
Start
Ac3
10%
50%
700
Ac1
90%
Finish
P
600
C
B
500
400
300
200
100
1000
500
200
100
50
Cooling rate at 800 C
0
mm 0.1
Bar
diameter
0.2
5
0.5
1
10
10
20
20
10
20
50
100
100 150 200 300
50
50
20
100
10
C per min
150 200
300
200
500
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
50
HRC
40
300
200
30
As cooled
100
Fig. A11.1
0.18 C (10171022), analysis wt%, austenitized at 900 C, previous treatment rolled
20
10
18/8/2008 4:48PM Plate # 0
pg 603
Appendix 11: Continuous Cooling Diagrams of Selected Steels / 603
Si
0.44 0.28
Mn
Cr
Mo
0.81 0.035 0.037 0.14 0.04
Ni
Al
Nb
Cu
0.15
Sn
0.12 0.016
900
800
Ac3
700
Start
10%
Ac1
50%
90%
F
Finish
P
600
C
500
400
300
Ms
M
200
100
1000
500
200
100
50
0
mm 0.1
0.2
Bar
5
diameter
10
800
0.5
5
50
20
10
20
50
Hardenability band
BS 970 080H46
10
20
100
100
20
10
50
150 200
150 200
100
300
300
200
500
500
HV
400
1000 2000 mm
Air
500
mm Oil
500
mm Water
Hardness after
transformation
700
600
C per min
As cooled
T 500 C 1 h
T 550 C 1 h
T 600 C 1 h
T 650 C 1 h
T 700 C 1 h
60
50
HRC
40
300
30
200
20
10
100
Fig. A11.2
18/8/2008 4:48PM Plate # 0
pg 604
Si
0.86 0.20
Mn
Cr
Mo
Ni
Al
Nb
0.60 0.020 0.020
900
800
Ac1a
Ac1b
A + C
Start
700
P
Finish
600
C
500
400
10% 50% 90%
300
Ms
200 10%
50%
M
100 90%
1000
500
200
100
50
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
10
50
50
20
20
50
100
100 150 200 300

100
10
C per min
150
200 300
200
500
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
As cooled
50
HRC
40
300
30
200
20
10
100
Fig. A11.3
18/8/2008 4:49PM Plate # 0
pg 605
Si
Mn
0.96 0.20
Cr
Mo
Ni
AI
Nb
0.60 0.020 0.020
900
800
700
Ac1a
Ac1b
Start
P
Finish
600
C
500
10% 50% 90%
400
300
200 Ms
10%
50%
100
M
1000
90%
500
200
100
50
20
10
C per min
100
500
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
10
20
50
50
100
100
50
150 200 300

150 200 300
200
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
As cooled
500
HV
T 400 C 1 h
T 500 C 1 h
400
HRC
40
300
30
20
10
200
100
Fig. A11.4
50
18/8/2008 4:49PM Plate # 0
pg 606
Si
0.46 0.25
Mn
Cr
Mo
Ni
AI
Nb
1.80 0.020 0.015
900
800
Ac3
700
Ac1
A
Start
600
10%
50%
90%
F
Finish
500
400
300 Ms
200
M
100
1000
500
200
100
50
20
10
C per min
0
mm 0.1
Bar
diameter
0.2
10
10
800
0.5
20
20
50
50
10
100
20
50
150 200
100 150 200
100
300
300
200
1000
2000 mm
Air
500
500
mm Oil
500
Hardenability band
SAE 1345H
mm Water
Hardness after
transformation
700
60
600
50
500
HV
HRC
400
300
40
As cooled
30
20
10
200
100
Fig. A11.5
13/4 Mn (1547) (1345), analysis wt%, austenitized at 850 C, previous treatment rolled
18/8/2008 4:49PM Plate # 0
pg 607
Si
0.44 0.20
Mn
Cr
Mo
Ni
AI
Nb
1.50 0.020 0.250
900
800
Ac3
700
Ac1
Start
10%
50%
90%
600
Finish
C
500
400
300 Ms
200
M
100
1000
500
200
100
50
20
10
C per min
100
500
0
0.2
mm 0.1
Bar
diameter
0.5
10
10
20
20
50
50
10
20
50
100 150 200

100
800
150 200
300
300
200
500
1000
2000 mm
Air
mm Oil
500
mm Water
Hardness after
transformation
700
60
600
50
500
As cooled
HV
400
40
300
30
20
10
200
100
Fig. A11.6
HRC
11/2 Mn+S (1139), analysis wt%, austenitized at 850 C, previous treatment rolled
18/8/2008 4:49PM Plate # 0
pg 608
Si
0.38 0.25
Mn
Cr
Mo
1.50 0.020 0.020
Ni
AI
Nb
0.45
900
800 Ac3
700
Ac1
Start
10%
600
50%
C
500
90%
400
Finish
B
Ms
300
200
M
100
1000
500
200
100
50
20
10
C per min
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
10
50
50
100
100
20
150 200
150 200
50
100
200
300
300
500
500
500
1000
2000 mm
Air
mm Oil
mm Water
800
700
600
500
Hardenability band
BS 970 608H37
As cooled
T 550 C 1 h
T 600 C 1 h
T 650 C 1 h
T 700 C 1 h
HV
Hardness after
transformation
60
50
HRC
400
40
300
30
20
10
200
100
Fig. A11.7
11/2 Mn Mo, analysis wt%, austenitized at 845 C, previous treatment rolled
18/8/2008 4:50PM Plate # 0
pg 609
Si
Mn
0.20 0.25
Cr
1.25 0.025 0.015 1.15
Mo
Ni
0.02
0.15
AI
Nb
900
Ac3
800
Start
Ac1
10%
700
50%
90%
F
Finish
600
P
C
500
B
400 M
s
300
M
200
100
1000
500
200
100
50
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
50
50
10
20
100
100
20
10
C per min
50
150 200
150 200
100
300
300
200
500
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
600
500
HV
60
Hardenability band
DIN 17210 20MnCr5
50
HRC
400
40
300
As cooled
200
100
Fig. A11.8
11/4 Mn Cr, analysis wt%, austenitized at 870 C, previous treatment rolled
30
20
10
18/8/2008 4:50PM Plate # 0
pg 610
Si
Mn
Cr
0.19 0.20 1.60 0.020 0.020
Mo
Ni
0.25
0.55
AI
Nb
900
Ac3
800
A
700
Start
10%
Ac1
50%
600
C
500
90%
B
Ms
400
Finish
300
M
200
100
1000
500
200 100
50
0
mm 0.1
Bar
diameter
5
10
0.2
0.5
10
20
20
50
50
10
100
100
20
50
150 200
150 200
20
10
C per min
100
500
300
300
200
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
50
HRC
500
HV
400
300
40
As cooled
200
100
Fig. A11.9
11/2 Mn Ni Mo, analysis wt%, austenitized at 870 C, previous treatment rolled
30
20
10
18/8/2008 4:50PM Plate # 0
pg 611
Si
Mn
0.43 0.24
Cr
Mo
Ni
1.35 0.025 0.025 0.45 0.20
AI
Nb
0.75
900
800
700
Ac3
Ac1
A
Start
10%
600
C
500
50%
400
90%
300 Ms
Finish
200
M
100
1000
500
200 100
50
20
10
C per min
0
0.2
mm 0.1
Bar
diameter
5
10
0.5
10
20
20
50
50
10
100
100
20
50
150 200
150 200
100
300
300
800
700
200
500
500
500
1000
2000 mm
Air
mm Oil
mm Water
Hardness after
transformation
60
As cooled
600
T 550 C 1 h
500
T 620 C 1 h
T 650 C 1 h
40
300
30
20
10
200
100
Fig. A11.10
50
HRC
HV
400
11/2 Mn Ni Cr Mo, analysis wt%, austenitized at 850 C, previous treatment rolled and softened 650 C 1 h
18/8/2008 4:51PM Plate # 0
pg 612
Si
0.40 1.75
Mn
Cr
Mo
Ni
Al
Nb
0.85 0.030 0.030
900
Ac3
800
Start
Ac1
10%
50%
90%
700
Finish
F
P
600
C
500
B
400
Ms
300
200
100
1000
500
200
100
50
0
mm 0.1
0.2
0.5
10
20
10
C per min
20
50
100
200
500
150 200
300
500
1000
2000 mm
Air
Bar
diameter
10
10
20
20
50
50
100
100
150 200
300
800
700
500
mm Oil
mm Water
Hardness after
transformation
Hardenability band
NF A-35 41S7
60
600
500
HV
400
50
As cooled
40
300
30
20
10
200
100
Fig. A11.11
HRC
13/4 Si Mn, analysis wt%, austenitized at 910 C, previous treatment rolled
18/8/2008 4:51PM Plate # 0
pg 613
Si
Mn
0.45 3.40
900 Ac3
Cr
Mo
Ni
Al
Nb
0.60 0.015 0.010 8.50
Start
800
A + C
10%
50%
Finish
700
600
C
500
400
300
B
200
Ms
100
1000
500
200
100
50
20
10
C per min
50
100
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
50
50
10
100
20
150 200 300
100 150
200 300
200
500
500
500
1000
2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
60
700
600
As cooled
50
HRC
500
HV
400
40
300
30
200
20
10
100
Fig. A11.12
31/2 Si Cr, analysis wt%, austenitized at 1050 C, previous treatment rolled and softened 650 C 1 h
18/8/2008 4:51PM Plate # 0
pg 614
Si
0.60 1.90
Mn
0.85
Cr
0.025 0.025 0.30
Mo
Ni
Al
Nb
0.25
900
800
Ac3
Start
Ac1
10%
50%
90%
A
700
Finish
P
600
C
500
400
B
300
Ms
200
100
1000 500
200
100
50
0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
50
50
10
20
100
100
50
150 200
150 200
20
10
C per min
100
300
300
200
500
500
500
1000 2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
50
HRC
500
HV
400
As cooled
40
300
30
200
20
10
100
Fig. A11.13
2 Si Cr Mo, analysis wt%, austenitized at 910 C, previous treatment rolled
18/8/2008 4:52PM Plate # 0
pg 615
Si
Mn
0.40 0.30
Cr
Mo
0.80 0.025 0.020
Ni
AI
Nb
0.26
900
800
Ac3
A
700
Start
Ac1
10%
50%
90%
F
600
Finish
C
B
500
400
Ms
300
M
200
100
1000
500
200
50
100
20
10
C per min
0
0.2
mm 0.1
Bar
diameter
0.5
10
10
20
20
10
50
50
20
100
100
50
100
200
150 200
300
500
150
200
300
500
1000
2000 mm
Air
mm Oil
500
mm Water
800
700
Hardenability band
SAE 4042H
Hardness after
transformation
60
600
500
HV
400
As cooled
50
HRC
40
300
30
200
20
10
100
Fig. A11.14
1/4
Mo (40374042), analysis wt%, austenitized at 810 C, previous treatment rolled
18/8/2008 4:52PM Plate # 0
pg 616
Si
Mn
0.40 0.26
Cr
0.62 0.007 0.005 0.23
Mo
Ni
0.10
3.45
AI
Nb
900
800
Ac3
700
Ac1
Start
600
10%
50%
500
B
90%
400
Finish
300
Ms
200
M
100
1000 500
200
100
50
20
10
C per min
1000
2000 mm
Air
0
0.2
mm 0.1
Bar
diameter
0.5
10
10
20
20
50
50
10
100
100
20
50
150 200
150 200
100
300
300
200
500
500
500
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
As cooled
500
HV
400
T 575 C 1 h
T 625 C 1 h
50
HRC
40
300
30
200
20
10
100
Fig. A11.15
31/2 Ni, analysis wt%, austenitized at 860 C, previous treatment rolled
18/8/2008 4:52PM Plate # 0
pg 617
Si
0.16 0.25
Mn
Cr
1.40 0.020 0.015 0.20
Mo
Ni
0.05
1.50
AI
Nb
900
800 Ac3
A
700
Start
Ac1
10%
50%
600
C
500
90%
400
Ms
Finis
300
200
100
1000
500
200
100
50
20
10
C per min
100
500
0
0.2
mm 0.1
Bar
diameter
0.5
10
10
20
20
50
50
10
20
100
100
150
150
50
200
200
300
300
200
500
500
1000
2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
50
As cooled
HRC
40
300
30
200
20
10
100
Fig. A11.16
11/2 Ni Mn, analysis wt%, austenitized at 840 C, previous treatment rolled
18/8/2008 4:53PM Plate # 0
pg 618
Si
0.40 0.15
Mn
Cr
0.48 0.016 0.040 0.15
Mo
Ni
0.25
1.75
AI
Nb
900
800
Ac3
700
Ac1
Start
10%
50%
90%
F
600
Finish
C
500
B
400
300
Ms
200
M
100
1000
500
200 100
50
20
10
C per min
50
100
0
0.2
mm 0.1
Bar
5
diameter
0.5
10
10
20
20
50
50
10
20
100
100
150
200 300
150 200
300
200
500
500
500
1000
2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
600
500
HV
400
60
As cooled
T 500 C 1 h
T 600 C 1 h
50
HRC
40
300
30
200
20
10
100
Fig. A11.17
13/4 Ni Mo, analysis wt%, austenitized at 845 C, previous treatment rolled
18/8/2008 4:53PM Plate # 0
pg 619
Si
Mn
0.40 0.23
Cr
Mo
Ni
0.75 0.020 0.020 0.65
Al
Nb
1.30
900
800
Ac3
700
A
Start
Ac1
10%
50%
90%
F
Finish
600
P
C
500
400
300
Ms
200
M
100
1000
500
200 100
50
20
10
C per min
50
100
0
mm 0.1
0.2
Bar
5
diameter
0.5
10
10
20
20
50
50
10
100
100
20
150
150
200
200
300
300
200
500
500
500
1000
2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
60
700
As cooled
600
T 550 C 1 h
T 600 C 1 h
500
T 650 C 1 h
50
HRC
HV
400
40
300
30
200
20
10
100
Fig. A11.18
11/4 Ni Cr, analysis wt%, austenitized at 850 C, previous treatment rolled
18/8/2008 4:53PM Plate # 0
pg 620
Si
0.24 0.20
Mn
Cr
0.80 0.020 0.020 0.50
Mo
Ni
0.20
0.55
AI
Nb
900
800
Ac3
Start
10%
Ac1
50%
700
90%
F
600
P
Finish
C
500
B
400 Ms
300
M
200
100
1000 500
200
100
50
20
10
500
1000
C per min
0
mm 0.1
Bar
diameter
0.2
5
0.5
10
10
10
20
20
50
50
20
100
100
50
100
200
150 200 300
500
150 200 300
500
2000 mm
Air
mm Oil
mm Water
800
Hardness after
transformation
700
600
60
Hardenability band
BS 970 805H25
500
HV
50
HRC
400
40
300
30
As cooled
200
100
Fig. A11.19
1/2
Ni Cr Mo (86228627) (8720) (8822), analysis wt%, austenitized at 830 C, previous treatment rolled
20
10
18/8/2008 4:54PM Plate # 0
pg 621
Si
Mn
0.40 0.25
Cr
0.60 0.020 0.020 1.20
Mo
Ni
0.15
1.50
AI
Nb
900
800
Ac3
700
Ac1
F
P
600
C
Start
500
10%
400 Ms
90%
50%
B
300
Finish
M
200
100
1000 500
200
100
50

0
mm 0.1
Bar
5
diameter
10
0.2
0.5
10
20
20
5
50
50
10
100
100
20
150 200
150
20
10
C per min
50
300
200 300
100
500
500
200
500
1000 2000 mm
Air
mm Oil
mm Water
800
700
600
500
HV
400
Hardness after
transformation
As cooled
T 550 C 1 h
T 600 C 1 h
T 650 C 1 h
60
50
HRC
40
300
30
200
20
10
100
Fig. A11.20
11/2 Ni Cr Mo, analysis wt%, austenitized at 850 C, previous treatment rolled, softened 650 C 1 h
18/8/2008 4:54PM Plate # 0
pg 622
Si
0.13 0.20
Mn
Cr
0.50 0.020 0.020 0.85
Mo
Ni
0.18
3.40
Al
Nb
900
800 Ac3
700
Ac1
Start
10%
50%
600
F
C
500
400
90%
Ms
300
Finish
M
200
100
1000
500
200
100
50
20
10
C per min
0
mm 0.1
0.2
Bar
5
diameter
10
0.5
10
20
20
50
50
10
20
100
100
50
150 200
150 200
100
300
300
200
500
1000 2000 mm
Air
500
500
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
Hardenability band
BS 970 832H13
As cooled
50
HRC
40
300
30
200
20
10
100
Fig. A11.21
31/2 Ni Cr Mo (9310), analysis wt%, austenitized at 820 C, previous treatment blank carburized 900 C 4 h A.C.
18/8/2008 4:54PM Plate # 0
pg 623
Si
Mn
0.59 0.25
Cr
Mo
Ni
0.60 0.025 0.025 0.65
Al
Nb
0.20
900
800
Ac3
Ac1
A
Start
700
Finish
P
600
C
10%
500
50%
90%
400
B
300
Ms
200
100
1000
500
200
100
50

0
mm 0.1
Bar
5
diameter
0.2
0.5
10
10
20
20
5
50
50
10
100
100
20
10
C per min
20
150
150
50
200
200
100
300
300
200
500
1000
2000 mm
Air
500
500
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
Hardenability band
BS 970 527H60
50
HRC
400
300
40
As cooled
200
100
Fig. A11.22
21/2 Cr (5060) (51555160), analysis wt%, austenitized at 830 C, previous treatment rolled
30
20
10
18/8/2008 4:55PM Plate # 0
pg 624
Si
Mn
0.39 0.20
Cr
Mo
Ni
Al
Nb
0.70 0.020 0.020 1.05
900
800 Ac3
A
Ac1
Start
700
Finish
F
600
P
C
500
10%
B
50%
90%
400
Ms
300
200
M
100
1000
500
200
100
50

0
mm 0.1
0.2
Bar
5
diameter
10
0.5
10
1
20
20
2
50
50
10
20
100
100
50
150 200
150 200
20
10
C per min
100
500
300
300
200
1000 2000 mm
Air
500
500
mm Oil
mm Water
800
700
600
500
HV
400
Hardenability band
ISO R 683 VII 3
As cooled
T 500 C 1 h
T 550 C 1 h
T 600 C 1 h
T 650 C 1 h
T 700 C 1 h
60
50
HRC
40
300
30
20
10
200
100
Fig. A11.23
Hardness after
transformation
1 Cr (5140), analysis wt%, austenitized at 870 C, previous treatment rolled
18/8/2008 4:55PM Plate # 0
pg 625
Si
0.42 0.25
Mn
Cr
Mo
0.85 0.020 0.020 1.15
Ni
Al
Nb
0.20
900
800
700
Ac3
Start
Ac1
10%
50%
90%
Finish
F
600
P
C
500
400
B
300
Ms
200
M
100
1000 500
200
100
50
20
10
C per min
0
0.2
mm 0.1
Bar
5
diameter
0.5
10
10
20
20
5
50
50
10
100
20
50
150 200
100 150
200
100
300
300
200
500
1000 2000 mm
Air
500
500
mm Oil
mm Water
800
Hardenability band
SAE 4142H
700
600
500
HV
400
As cooled
T 450 C 1 h
T 500 C 1 h
T 550 C 1 h
T 600 C 1 h
T 650 C 1 h
T 700 C 1 h
Hardness after
transformation
300
50
HRC
40
30
20
10
200
100
Fig. A11.24
60
11/4 Cr Mo (41404142), analysis wt%, austenitized at 860 C, previous treatment rolled
18/8/2008 4:55PM Plate # 0
pg 626
Si
Mn
0.50 0.25
Cr
0.75 0.025 0.025 0.95
Mo
Ni
0.05
0.15
Al
Nb
V
0.20
900
800 Ac3
Ac1
Start
700
10%
90%
50%
P
Finish
600
C
500
B
400
300
Ms
M
200
100
1000 500
200
100
50
10
20
C per min
0
0.2
mm 0.1
Bar
5
diameter
0.5
10
10
20
20
10
50
50
20
100
100
50
150 200
150
200
100
300
300
200
500
1000 2000 mm
Air
500
500
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
Hardenability band
ISO R 683 XIV 13
40
300
As cooled
200
100
Fig. A11.25
50
HRC
1 Cr V, (6150), analysis wt%, austenitized at 875 C, previous treatment rolled
30
20
10
18/8/2008 4:56PM Plate # 0
pg 627
Si
Mn
0.42 0.30
Cr
0.65 0.020 0.020 1.65
Mo
Ni
Al
0.33
Nb
1.00
900
Start
800 Ac1
10%
50%
90%
A
700
Finish
600
C
500
400
300
B
Ms
200
M
100
1000 500
200
100
50
10
20
C per min

0
mm 0.1
Bar
diameter
0.2
0.5
10
10
20
20
50
50
10
100
100
20
150
150
50
200
200
100
200
500
1000
2000 mm
Air
300
300
mm Oil
mm Water
800
Hardness after
transformation
700
60
600
500
HV
400
As cooled
Hardenability band
SIS 14 29 40
40
300
30
20
10
200
100
Fig. A11.26
50
HRC
11/2 Cr Al Mo, analysis wt%, austenitized at 900 C, previous treatment softened 650 C 1 h
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_629-640.pdf/F5113_Index/
20/8/2008 5:06PM Plate # 0

Index
A
abrasive erosion, 129
abrasive wear, 129, 130(F), 137, 195(T), 315(F)
acute-angled keyway, 11, 12(F)
adhesion colloids, 138
adhesive wear, 129(F), 137, 315
aerospace applications (case studies)
cracking in a main landing gear attach pin, 354355,
357(F), 358(F), 359(F)
failure analysis of a catapult holdback bar, 351354(F),
355(F), 356(F)
failure analysis of a main landing gear lever, 362364(F),
365(F), 366(F)
failure analysis of a nose landing gear (NLG) piston axle,
367372(F), 373(F)
failure analysis of an aircraft hoist sling during static test,
373375(F), 376(F), 377(F)
failure analysis of an inboard flap hinge bolt, 364366(F),
367(F), 368(F), 369(F)
failure analysis of an internal spur gear, 375378(F),
379(F)
fracture analysis of ASI 420 stainless steel roll pin,
359362(F), 363(F)
main landing gear (MLG) axle, 378380(F), 381(F),
382(F)
MLG linear actuating rod and cylinder, 355359(F),
360(F), 361(F), 362(F)
multiple-leg aircraft-handling sling, 372373(F), 374(F),
375(F)
nondestructive testing and failure analysis of fin
attach bolts after full-scale fatigue testing,
380392(F&T)
aging tendency, 3536
Agricultural Ammonia Institute, 75
AISI to non-AISI steel cross reference, 551562(T)
Aluminizing, 33
American Association of State Highway Transportation
Officials/American Welding Society (AASHTO/
AWS), 503
American Iron and Steel Institute (AISI), 311(T)
annealing
austenitic stainless steels, 39
definition, 3(T)
diffusion (homogenizing annealed), 3(T)
double-annealing, 395396
full, 3(T)
incorrect, 107(F)
intercritical ( partial), 3(T)
pg 629
isothermal, 3(T)
powder, 395
recrystallization, 3(T)
soft, 3(T)
solution, 38, 3940
spheroidized, 107(F)
stress, 459
subcritical, 3(T), 16, 266267
temper, 346
time range (normal), 40
transformation, 346
water, 285
anode, 131132
anodic stress, 74(F)
Association Francaise de Normalisation (AFNOR), 493
asymptotic, definition, 116
austenitic stainless steels, 4, 5(T), 36, 39
austenitizing temperatures for steel, 537538(T)
austenitizing temperatures for direct-hardening carbon
and alloy steels (SAE), 537538(T)
reheating (austenitizing) temperatures for hardening
of carburized carbon and alloy, 538(T)
autofrettage, 66
B
backscattered Z contrast, 113
bainite, 2, 209, 210(F), 279(F), 341, 403404. See also lower
bainite; upper bainite
banding, 103(F), 278, 279(F), 353
beach marks, 77, 78(F), 97, 113, 118, 505, 506(F)
Biot modulus, 6
blind holes, 14, 33(T), 152(F)
blowholes, 155, 156
boost-diffuse cycle, 184185
boriding, 33
brittle fracture, 30, 3334, 5359(F), 120122(F)
brittle temperature range (BTR), 517
brittle transgranular fracture, 295
buckling, 5, 30, 56, 88(F), 89, 128(F)
bulk composition evaluation, 114
C
carbides
cementite, 222223(F)
film or flake, 227, 228(F)

20/8/2008 5:06PM Plate # 0
630 / Index
carbides (continued)
globular, 223225(F)
network, 225227(F&T), 228(F)
overview, 222
spheroidal, 224
carbon steels
coefficients of linear thermal expansion, 541543(T)
heat capacity, summary of, 548549(T)
physical properties of, 541550(T)
specific heats of, 549550(T)
thermal conductivities of, 547(T)
thermal conductivity, summary of, 545546(T)
thermal expansion, summary of, 544545(T)
carbonitrided components
carburizing and carbonitriding, comparison of, 178(T)
case, definition, 183
case depth, 183, 209
case depth, improper, 208
case hardness, 209
core hardness, 209
core microstructure, 210211, 212(F)
dimensional stability, 200204(F)
overcarbonitriding, 211, 213, 214(F), 215(F)
overview, 177178(F&T)
quenching, 191
quenching cracks, 205
residual stresses, 196
carbonitriding
case hardening, 397398(F)
case studies, 411412(F), 413414(F)
dimensional change in, 411412(F)
fatigue property characteristics, 246250(F&T)
process, 177178(F&T)
SAE 1118 steel, 491(F)
steel selection, 181
stress, 196
carburized components
carbides, 222227(F&T)
carburizing and carbonitriding, comparison of,
178(T)
carburizing process, 184185
case, definition, 183
case crushing, 231232(F)
case depth, 183185(F), 186(F)
case depth, improper, 207208(F)
case hardness, 209, 210(F)
decarburization, 213, 215217(F)
design, 179181(F)
dimensional stability, 200204(F)
grain size, 217219(F)
internal oxidation, 219222(F)
macropitting, 230231(F)
micropitting, 230
noncarbide inclusions, 228229(F)
overcarburizing, 211, 213, 214(F), 215(F)
overview, 177178(F&T)
partial melting, 233234(T)
pitting corrosion, 232233(F)
quenching, 185, 187191
pg 630
quenching cracks, 204207(F)

residual stresses, 196200(F&T)
retained austenite, 191196(F&T)
steel selection and hardenability, 181196(F&T)
survey summary of sources of gear failures, 178(T)
transition zone, 211, 213(F)
carburized steels
carbon potential, 216
case depth, 209
case structure, 218
corrosion resistance, 233
decarburization, 95(F)
film or flake carbides, 227
gas-carburized, 199200
incorrectly, 187(F)
intergranular cracking, 96, 118
intergranular oxidation, 219, 220(F)
network carbides, 225, 226
pitting, 226, 233
puzzle piece, 108(F)
spheroidized carbide particles, 224
case crushing, 231232(F)
casting process, failures
clean steel, 165166
cold joints, 163165
component lifetime factors, 151
decarburization during microfusion, 162163, 165(F)
improper cast design, 151153(F), 154(F)
inclusions, 165175(F&T)
macroinclusions, 166167
microinclusions, 167168(T)
porosity, effects due to, 154162(F), 163(F)
cathode, 131132
cause-and-effect (CE) matrix, 400401, 402403(F)
cavitation creep, 128
cementite, 222223(F)
definition, 1
characteristics vs. defect, 88
Charpy impact test, 49, 55, 58, 295(F), 297, 297(F)
Charpy V-notch (CVN) impact toughness testing, 317(F),
343(F), 513
chevron marks, 48, 5354(F), 119(T), 367, 370(F)
chicken-wire pattern, 208
chromizing, 33
cladding, 39
clean steel, 165166, 229
clogging, 167
coefficient of thermal expansion, 45
cold cracking, 36
cold forming, 44
cold joints, 163165
cold junction, 163
cold work tools, heat treating failures of, 314330(F&T)
characteristics, 315317(F)
chemical composition, 314317(F)
design-related failures, 317319(F)
hardening temperatures, incorrect, 323324, 325(F),
326(F)

20/8/2008 5:06PM Plate # 0
pg 631
Index / 631
heat treating processing, 323

incorrect EDM, 321323(F)
incorrect grinding, 319321(F)
large (blocky) carbides, 315
retained austenite content, excessive, 324, 326329(F),
330(F)
short or absent tempering, 329330, 331(F), 332(F)
surface damage, 319323(F)
tempering temperatures, incorrect, 323324, 325(F),
326(F)
component design
design, overview, 1
design aspects
general thermal stress, 7, 32
overview, 2, 4
peak stress, 7, 32, 38
primary stress, 7, 8, 32
residual stress, 78
secondary stress, 7, 32
distortion, techniques for controlling, 1618
hardening, 2
heat treat processing, introduction to, 12(T)
heat treated components, 814(F)
heat treatment design, 2931
heat treatment failures, examples, 1829(F)
heat treatment, modeling of, 3133(F&T)
heat-transfer theory, 47(F&T)
quenching medium, 17
quenching techniques, 1718
risk-based approach to, 4041(F)
steel grade/condition, 1416
tempering, 2
welded components, 3336
contact fatigue pitting. See macropitting
contact loading, 95, 98(F), 206(F)
continuous cooling diagrams, 601627(F)
1/4 Mo steel, 615(F)
carbon steels, 602605(F)
chromium steels, 623627(F)
Mn steels, 606607(F)
Mn-Cr steels, 609(F)
Mn-Mo steels, 608(F)
Mn-Ni-Cr-Mo steels, 611(F)
Mn-Ni-Mo steels, 610(F)
nickel steels, 616619(F)
Ni-Cr-Mo steels, 620622(F)
silicon steels, 612614(F)
continuous cooling transformation (CCT) diagrams, 32,
255257(F), 339340(F)
contour hardening, 432, 470471
corrosion
crevice, 131
definition, 131
galvanic, 131132
pitting, 131
stress-corrosion cracking (SCC), 131
types of, 131
corrosion fatigue, 100101(F), 118, 120
corrosive wear, 129, 138
cotter holes, 152
crack
blunts, 116
brittle, 92, 121(F), 275, 506
brittle fracture, 30
carbides, 315
circumferential, 360(F), 382
coloration, 60(F), 90
crack path, 96
crack-free welding, 507
ductile, 92, 120(F)
ductile fracture, 33
extension, 56
faces, 44(F), 50, 63(F), 64(F)
fatigue, 101, 171(F), 172(F), 173(F), 229(F), 242(F).
See also fatigue crack initiation
fatigue failures, 30
forging, 142
formation, 228
formulation, 99, 103, 143(F), 153(F)
fracture, 9097(F), 98(F), 118
grinding, 207208(F)
growth mechanism, 51
growth rates, 49, 50, 80
heat checking, 334(F)
intergranular path, 306(F)
lateral oxide formation, 221
longitudinal, 169, 170(F), 357, 360(F)
macrobrittle, 92
macroductile, 92
mechanical, 348(F)
metallography, 5051
microscale crack path, 97
networks, 33
normalization, 153(F), 154(F)
nucleation, 128, 130, 159, 160, 168(F), 170, 171(F)
optical micrograph, 27(F)
order, 48
origin, 21, 25(F), 90, 162(F)
oxide filled, 104(F)
pearlitic matrix, 53
pre-existing, 116
profile, 101(F)
propagation, 301, 482, 506. See also crack propagation
quench, 21, 24(F), 25(F), 27(F), 102(F), 154(F), 411(F).
See also quench cracking
resistance, 107
resistance to propagation, measuring, 34
secondary, 27(F), 50, 101(F), 274(F), 370, 372(F), 382(F)
shank, 388, 389(F)
sharp corner, 318(F)
stage I, 78
stage II, 78
stage III, 78
surface, 482
temperature, 480, 481(F)
through-the-wall, 76
tip, 78, 80, 112, 116, 126(F), 302
transgranular, 162(F)
transgranular path, 304305(F&T)
vertical, 64(F)

20/8/2008 5:06PM Plate # 0
632 / Index
crack extension, 38, 56, 118

crack initiation
carbides, 315
classifying, 118
electroplating, 45
fatigue crack initiation, 122124(F)
fatigue cycles, 482
fatigue failures, structural features of, 78
high-temperature grain-boundary oxidation, 281,
283(F)
hot work tool steels, 344
inhibiting, 196
internal oxidation formation, 221
microcracks, 206
nitriding, 343
retard, 123
S-N curves, 482
subsurface, 93, 94(F), 229(F)
crack propagation, 124127(F)
beach marks, 77, 78(F)
brittle fracture, 121, 175
carbides, 280
cleavage crack propagation, 509
cleavage fracture, 54(F)
definition, 482
ductile fracture, 120(F)
ductile structural materials, 506
fatigue crack propagation, 162, 170, 172(F),
173(F)
grain-boundary, 338, 341(F)
hydrogen embrittlement, 301(F)
intergranular brittle fracture, 121122(F)
nitrided tools, 344, 348(F)
rate of, 80
resistance, 34
SCC fractures, 73
stress analyses, 116
cracking
cold, 36
delayed. See underbead cracking
fatigue, 506(F)
ferrite vein, 512513(F)
flat, 82
HAC, 509513(F&T)
hydrogen-induced, 36
intergranular, 59, 63(F), 96(F), 100(F), 205206(F),
279(F)
microcracking, 183, 192, 206(F), 218
premature, 333335, 336(F), 337(F), 338(F)
quench, 5965(F), 273283(F&T)
SCC, 28, 36, 70, 7276(F&T), 77(F)
solidification, 515517(F)
toe, 503(t), 504(t)
underbead, 507, 510511(F)
creep, 128129(F)
creep rupture, 7677(F), 78(F)
critical flaw size, 49, 116
critical scuffing temperature, 230
cyclic fracture propagation, 80, 82(F)
cyclic loading, 77, 118, 122(F)
pg 632
D
damage, definition, 87
damage mechanism, definition, 113
damage mode, definition, 113
decarburization, 215217(F)
on carburized steel, 95(F)
deformation, 8890(F)
buckling, 89
elastic, 89, 98(F)
gradual onset, 8889(F)
plastic, 89
sudden, 88(F), 89
die-casting dies, 331, 334(F), 341, 343
dimensional stability
distortion, 200202(F)
isothermal-transformation diagrams, 202(F)
warpage, 200, 202
discontinuities, 133, 503505(F&T)
distortion
bending, 468
buckling, 128(F)
creep, 128129(F)
reasons for, 127129(F)
residual stresses, 129
techniques for controlling, 1618
yielding, 127128
double-frequency heating, 428
ductile crack path. See microvoid coalescence (MVC)
ductile dimples. See microvoid coalescence (MVC)
ductile fracture, 28(F), 5153(F), 91(F), 9293(F), 96,
120(F), 226227
E
elastic collapse. See buckling
electrical discharge machining (EDM), 2021
electrochemical theory, 514
electron-dispersive x-ray (EDX), 169
electroplating, 2829, 45
embrittlement, 293303(F&T)
definition, 38
temper (two-step embrittlement), 296
tempered martensite embrittlement, (TME), 294296(F)
endogas, 177
energy-dispersive x-ray analysis, 113
energy-dispersive x-ray spectroscopy (EDS), 26
environmentally assisted failure. See corrosion
etching, 11(F), 64(F), 104(F), 210(F)
acid, 320
cooper sulfide, 303
corrosion resistance during, 337(F)
nital, 171, 174(F), 216, 324, 325(F)
retained austenite, 326, 327(F)
stain, 113
temper, 45(F), 62(F), 357(F)
Euler buckling, 30, 89
exogenous inclusions, 165, 229
exothermic sleeves, 158

20/8/2008 5:06PM Plate # 0
pg 633
Index / 633
F
failure, definition, 40, 87
failure analysis, guidelines of
background data, collection of, 111112
chemical analysis, 114115
conclusions, formulation of, 117
evidence, analysis of, 116117
failed part, preliminary examination of, 112
failure mechanism, determination of, 113114
fracture mechanics analysis, 90, 116
macroscopic examination/analysis, 112
mechanical testing, 112
metallographic specimens, 113
microchemical analysis, 114
microscopic examination/analysis, 112113
nondestructive testing, 112
photographic examination, 112
the report, 117118
specimens, 112
stress analysis, 114115
testing, 116
failure analysis, stages of
background information, collection of, 47
macroscopic examination, 50
mechanical testing, 49
metallography, 5051(F)
microscopic examination, 50(F)
preliminary visual examination, 4748
specimens, collection/preservation of, 4950
specimens, sectioning of, 50
failure mechanism, definition, 113
failure mode and effects analysis (FMEA), 40, 399
failure mode, definition, 113
failures, mechanisms/causes of, 4386(F&T), 87109(F)
corrosion and environmental damage, 99101(F)
design deficiencies, 43
failure analysis, 4751(F)
fracture mechanism, 5183(F)
fractures, 9097(F)
heat treating process, 106108(F)
heat treatment, poor response from, 101103(F),
104(F)
manufacture, 4446(F)
material issues, 4344(F)
processing, 4446(F)
service conditions, 4647(F)
wear, 9799(F)
fatigue, 7783
carburized steels, 185
inhibit, 177
structural surface anomalies, 199200
fatigue fracture, 122127
fatigue crack propagation, 124127(F)
nitrided layers, 241253(F&T)
fatigue intrusions and extrusions, 80(F), 122(F)
fatigue resistance, 241244(F&T)

after nitriding/nitriding treatments, 242244(F&T)
fatigue striations
aluminum alloy, 125(F)
formation of, 80, 83(F)
interstitial-free steel, 125(F)
representative, 381(F), 390(F)
stage II crack growth, 78(F), 81(F)
typical, 354(F), 355(F), 356(F)
feeders, 158
ferrite vein cracking, 512513(F)
fine grinding, 472477(F)
finite element analysis (FEA), 116
fisheyes, 115, 243(F), 511512(F)
flat cracking, 82, 83(F)
forging, failure in
case studies
avoidance of flow through, lap, and crack, 145148(F)
crankcase underfill, 138139(F)
spade bit, 140142(F)
summarized, 133(T)
trim tear, 142143(F)
tube bending, 139140(F), 141(F)
upset forging, 143145(F), 146(F)
discontinuities, 133
factors in analysis of cold forging failures, 134(T)
factors in analysis of hot forging failures, 135(T)
forging process design, 134138(F&T)
forging tolerances, 135137
lubricant performance, 138
lubrication, 137, 138
wear types, 137138
Fouriers first law, 5
Fouriers second law, 4
fracture mechanics analysis, 115116
fracture mechanism, 5183(F)
brittle fracture, 5359(F)
ductile fracture, 5153(F)
fatigue, 7783(F)
intergranular brittle fracture, 5977(F&T)
fracture modes
brittle, 28(F)
ductile, 28(F)
grain-boundary, 300
identification chart, 119(T)
intergranular, 26, 28(F), 80, 82(F)
microvoid coalescence (MVC), 51, 52(F), 507508(F)
TME, 65
transgranular, 80, 82(F), 296
fracture surface matching, 118
fracture toughness, 49, 55, 185, 522523(T)
fracture toughness testing, 49
fractures
brittle, 96. See also brittle fracture
corrosion-fatigue cracks, 118, 120
cyclic loading, 118
examining, 9093(F)
fracture modes, 19(T), 118
hydrogen embrittlement, 96. See also hydrogen
embrittlement (HEM)

20/8/2008 5:06PM Plate # 0
634 / Index
fractures (continued)
intergranular fracture, 96. See also intergranular brittle
fracture
macroscale brittle fracture, 91(F), 92(F)
macroscale ductile fracture, 91(F), 93(F)
macroscale features, 9193(F)
microscale features, 9597(F)
origin of, 118
process of, 118
river line features, 97(F)
slant, 56. See also brittle fracture
stress versus strength, 9395(F)
stress-corrosion cracks, 118, 120
striations, 118
fretting, 97, 98(F), 130131(F)
G
galling or seizing, 129
galvanic corrosion, 131132
gap-by-gap hardening, 429, 430(F), 439(F), 470471
gas bubble, 155, 167
gas holes (blowholes), 154, 155, 156
gas metal arc welding (GMAW), 511
gas porosity, 154156(F)
glow discharge optical emission spectroscopy, 220
gouging, 119(T)
grain size, influence of, 217219(F)
grain-boundary sliding, 128(F)
graphite products, 138
gray cast iron, 121(F)
Great Boston Molasses Disaster, 5657(F)
grid crossings, 153(F)
grinder burn, 98
grinding
fatigue strength, 80
incorrect, 319321(F)
manufacture and processing, 44, 45(F)
porosity, eliminating, 504
quench cracking, 62(F)
residual stresses after, 472477(F)
stock removal, 16
surface damage, 319
surface oxidation, removing, 222
thermal defects, preventing, 208
grinding burn, definition, 207
grinding burns, 207, 208
grinding cracks, 207208(F), 303304(F)
grinding grains, 473474(F)
Grossman H-values (numbers), 270(T)
H
halos (or fisheyes), 115. See also fisheyes
heat checking, 331, 334(F)
heat checking cracks, 334(F)
heat treat processing, introduction to, 12
pg 634
heat treated steel parts

component characteristics, 104106(F)
corrosion damage, 99101(F)
environmental damage, 99101(F)
fracture, 9097(F)
heat treatment, poor response to, 101108(F)
wear, 9799(F)
heat treater, 89, 91, 92, 104, 106, 107108
heat treating process
atmosphere, improper, 107108(F)
design engineers and, 106
errors, 106107
heat treaters and, 106
heating rate, inadequate, 107108(F)
temperature, inadequate, 107108(F)
time at temperature, inadequate, 107108(F)
heat treatment
component design process, 2930
design, 2931
failures, examples of, 1829(F)
heat treating errors, 18
heating errors, 18
material behavior, 3031
modeling of, 3133(F&T)
phase 2 design review to avoid failures, 33(T)
poor response to
component characteristics, 104106(F)
raw material characteristics contributing to,
101103(F), 104(F)
solid-phase transformation model, 3132
temperature errors, 18
thermomechanical modeling, 3233
heat-affected zone (HAZ), 35
heat-transfer theory, 47(F&T)
Hertzian stresses, 97, 98(F), 206(F), 230
high-speed steels, 313314(F)
high-velocity oxy-fuel (HVOF) coating, 62
holes, types of, 152(F). See also individual types
hot tops, 158
hot work tools, heat treating failures of, 330349(F&T)
characteristics of, 330333(T), 334(F), 335(F)
chemical composition, 330333(T), 334(F), 335(F)
cooling (slow) during quenching, 338343(F)
heat treating procedures, inadequate, 335337(F)
heating, excessive, 344349(F)
incorrect hardening temperatures, 337338, 339(F),
340(F), 341(F), 343(F)
incorrect tempering temperatures, 337338, 339(F),
340(F), 341(F), 343(F)
nitrided tools, 343344(F), 348(F)
premature cracking, 333335, 336(F), 337(F),
338(F)
hydrogen embrittlement (HEM)
electroplating, 2829, 45
intergranular brittle fracture, 59(F), 6870(F), 71(F),
72(F), 73(F)
TE, interaction with, 301302(F)
hydrogen traps, 508(T)
20/8/2008 5:06PM Plate # 0

pg 635
Index / 635
hydrogen-assisted cracking (HAC)

examples
failure to pass bend tests due to hydrogen, 513(F)
ferrite vein cracking in high-heat output welds,
512513(F)
fisheyes on fracture surface, 511512(F)
underbead cracking, 510511(F)
ferrite vein cracking, 512513(F)
fisheyes, 511512(F)
hydrogen-assisted reduced ductility, 508(F), 513(F),
514(F)
underbead cracking, 503(T), 504(F), 509511
weld metal HAC, 511
hydrogen-assisted cracking theory, 506509(F&T)
hydrogen-assisted reduced ductility, 508(F), 513(F), 514(F)
hydrogen-induced cracking, 36
I
impact testpieces, 121(F)
impact tests, usefulness of, 482
impact toughness testing, 49
inclusions
case studies
failure in the axle of a reduced section in a rotating
component, 170171, 172(F&T), 173
failure of a 52100 steel axle, 171, 173175(F)
failure of a steam turbine rotor blade, 168170(F&T),
171(F)
categories of, 53
classifications, 165
definition, 165
stringer, 166, 280281(F)
indigenous inclusions, 165
induction coils
materials for, 423424(F)
multiple-turn induction coil, 424425(F)
types of, 423
induction hardening
fatigue strength, 481485(F&T)
fine grinding, 472477(F)
induction heating, 420422(F), 440444(F),
444449(F&T)
induction surface heating. See induction surface heating
induction surface-hardened layer, 477481(F&T)
machine parts, 422432(F), 485491(F&T)
magnetic flux concentrators, 437440(F)
overview, 417419
quenching systems for, 449452(F)
steels, 419420(F&T)
stresses/residual stresses, time variation of, 452466(F)
surface hardening, 421422, 466472(F)
induction heating
coils for, 423(F)
definition, 398
features of, 420422(F)
gear wheels, 429, 430(F), 431(F)
Jominy curves, 4150 steel, 492, 493(F)
machine parts, 440444(F)

power supplies, 421
rotation velocities, 452
supervising, 456(F)
temperature cycles, 456457(F), 458(F)
time-temperature dependence in, 444449(F&T)
use of, 417
induction scanning, 426, 427(F)
induction surface hardening
advantages of, 421422
gears, 491497(F)
residual stresses, 472477(F)
workpiece distortion in, 466472(F)
induction surface heating, 430(F), 432437(F), 442(F), 456,
457(F), 465(F), 471(F)
induction surface-hardened layer, 444, 477481(F&T)
inhomogeneity, 462, 463(F), 479
Instron TT-DM machine, 242
intergranular brittle fracture
causes of, 59, 279. See also individual causes
creep process and, 128(F)
creep rupture, 7677(F), 78(F)
hydrogen embrittlement, 59(F), 6870(F), 71(F), 72(F),
73(F)
liquid metal embrittlement, 6667(F&T)
SAE 3161, 122(F)
solid metal embrittlement, 6768(F&T), 69(F)
stress-corrosion cracking (SCC), 70, 7276(F&T), 77(F)
temper embrittlement, 6566(F)
TME and, 65(F), 295296
intergranular cracking, 59, 63(F), 96(F), 100(F), 205206(F),
279(F)
interlath cleavage, 296
internal oxidation, 219222(F)
internal porosity, 158159(F)
International Organization for Standardization (ISO), 493
ion vapor-deposited (IVD), 362, 364(F), 365, 367(F), 369(F)
iron-carbon equilibrium diagram, 585(F)
isothermal diagrams
carbon steels, 588592(F)
chromium steels, 597(F), 599(F)
chromium-molybdenum steels, 593594(F)
nickel-chromium-molybdenum steels, 598(F)
nickel-molybdenum steels, 596(F)
Ni-Cr-Mo steels, 595(F)
isothermal-transformation diagrams. See time-temperature
transformation (TTT) diagrams
J
jig-bore grinding, 2(F), 20
job shop heat treater, 87
Jominy distance (J-distance), 270
L
lap, definition, 145146, 277
lath martensite, 187, 188(F)
20/8/2008 5:06PM Plate # 0

636 / Index
Liberty ships, 54
linear thermal expansion, 5(T), 541(T), 544(T)
liquid metal embrittlement (LME), 6667(F&T), 302303(T)
liquid shrinkage, 157(F)
liquidation temperature, definition, 233(T)
low-alloy steels
coefficients of linear thermal expansion, 541543(T)
cold forging, strain-limiting criteria for, 142(F)
embrittlement, 302(T)
HAZ maximum hardness, 3536
heat treatment procedures, 3637
linear thermal expansion, 5(T)
quench age embrittlement, 59
SCC and, 36, 73, 74(T)
specific heats, 549550(T)
thermal conductivities, 547(T)
thermal expansion, summary of, 544545(T)
thermal hardening, 418
lower bainite, 2, 210(F), 270, 292, 299, 403(F)
lubricant performance, 138
lubrication
graphite products, 138
inadequate, 4647(F)
micropitting, 230
wear and, 137138(F)
M
machine parts, induction hardening of
carburized parts, residual stresses in, 489491(F&T)
contour hardening, 432
double-frequency heating, 428
gap-by-gap, 429, 430(F), 439(F), 470471
gear wheels, 427432(F), 439(F)
induction scanning, 426, 427(F)
progressive hardening, 426427(F)
quenching, 442444(F)
quenching techniques, 425426(F)
scan-hardening process, 425(F)
single-shot induction hardening, 422, 423, 424(F), 427,
428(F), 440(F), 496
single-shot induction heating, 424(F), 448(F)
stress profiles, 485491(F&T)
tooth-by-tooth, 430(F), 439(F)
tooth-gap hardening, 431432(F)
macroinclusions, formation of, 166167
macropitting, 230231(F)
macroscale brittle fracture, 91(F), 92(F)
magnetic flux concentrators, 437440(F)
martempering, 1718, 60
martensite, 285289(F)
definition, 2
diffusionless process, 286
morphology of, 187(F)
slipped martensite, 287
tempered martensite, 289
pg 636
mechanical cracks, 348(F)

mechanism and mode, difference between, 113114
mechanism of failure, definition, 87
metric conversion guide
base SI units, 521(T)
conversion factors, 522523(T)
derived SI units, 521522(T)
supplementary SI units, 521(T)
microchemical analysis, 114
microcracking, 183, 192, 206(F), 218
microinclusions, formation of, 167168(T)
micropitting, 230
microscale fracture features, 9597(F)
microscopic examination, 50(F), 112113
microvoid coalescence (MVC), 96
modulus of elasticity, 89, 529
N
National Safety Transportation Board, 75
nil-ductility temperature (NDT), 5556
niobium (formerly columbium), 39, 314
nitrided layers, fatigue fracture of
carbonitriding, 246250(F&T)
fatigue evaluation, 244246(F&T), 247(F&T)
fatigue resistance, 241244(F&T)
nitrided steels, 244246(F&T), 247(F&T)
nitrocarburizing, 33
node dislocation, 152, 153(F)
non-AISI to AISI steel cross reference, 563583(T)
dye-penetrant method, 48
eddy-current methods, 48
magnetic particle inspection, 48
ultrasonic testing (UT), 4849
North American Die Casting Association (NADCA),
336337
notch, 12(F), 55, 56, 70(F), 179(F), 266(F)
O
Occupational Safety and Health Administration (OSHA),
4041
P
pancake (flattened disc), forging, 144(F)
Paris regime, 126127(F)
part-way downs, 146147
passing hole, 152(F)
pearlite
crack/void formation, 53
decarburization on carburized steel, 95(F)
definition, 1

20/8/2008 5:06PM Plate # 0
pg 637
Index / 637
phase transformations, 255263(F&T)

CCT diagrams, 255257(F)
distortion mechanism, 261
metallurgical crystal structure, 257(F), 258(F)
residual stresses, relief of, 263(F)
TTT diagrams, 255, 256(F)
volume changes during, 261263(F)
volumetric change upon quenching, 257261(F&T),
262(F)
phosphorus, 55, 65, 115, 295296, 297298
phosphorus segregation, 185, 295296, 297298
pinholes, 154, 155156(F), 404405(F), 406(F)
pitting, definition, 113114
pitting corrosion, 232233(F)
plastic blunting process, 78, 80
plastic deformation
avoiding, 8, 30
brittle fracture and, 34, 53, 54(F), 120
buckling and, 128
creep and, 128
ductile fracture and, 51, 96, 120
formation of, 7, 52, 286287, 331, 446
fracture surface matching, 118
shot peening and, 199(F)
plastic mold steels, 313(F)
plate martensite, 187, 188(F)
polishing, 51, 113, 313(F), 390(F), 482
porosity
blowholes, treating, 156
casting component feeding, 159
effects due to, 154162(F), 163(F)
gas, 154156(F)
internal, from nucleation, 158159
internal, from the surface, 158, 159(F)
pinholes, treating, 156
shrinkage pores, 156158(F)
postweld heat treatment (PWHT), 37
powder metal steel components
case hardening
carbonitriding, 397398(F)
carburizing, 397
induction hardening, 398(T)
steam treatment, 398399(F)
definition, 395, 396(F)
failure analysis techniques, 399401(F&T)
CE matrix, 400401, 402403(F)
failure mode and effects analysis (FMEA), 399
process maps, 400, 401(T)
steps, 400(F)
tools, 400(F)
powder metallurgy process, 395396, 397(F)
powder metal steel failures, case studies
blistered sintered products, 405406
cracks after induction hardening, 409410(F)
cracks after quenching, 410411(F)
dimensional change in carbonitriding, 411, 412(F)
dimensional change in steam treatment, 414(F),
415(F)
dimensional instability during shrink fitting, 406407
fracture of steam-treated part, 407408(F)
low breaking load after carbonitriding, 413414(F)

low core hardness after steam treatment, 415
low surface hardness after carbonitriding, 411412(F)
oxidation after sintering, 408409(F)
pinholes after sintering, 404405(F), 406(F)
variation in bore diameter after heat treatment,
412413
wear after sinter hardening, 401, 403404(F)
wear after sintering, 407(F), 408(F)
powder metallurgy process
blending, 396
compaction, 396
powder production, 395396
sintering, 397
progressive hardening, 420, 426, 427(F), 446, 455, 469,
496. See also gap-by-gap hardening
puzzle piece carbide, 108(F)
Q
definition, 5960
intergranular fracture, 96
localized overheating, 62
martempering, 60
mitigation of, 60
stress raisers role in, 272273
quench cracking, case studies
as-quenched 4340 steel, 273274(F)
cracking of 4140 block forging after quenching and
tempering, 274275(F&T)
decarburization and oxidized grain boundary,
281283(F)
network carbides and coarse grain size, 278280(F)
presence of a seam defect, 276(F)
presence of chemical segregation, 278, 279(F)
presence of slag inclusions and a lap
defect, 276278(F&T)
presence of stringer inclusions and chemical
segregation, 280281(F&T)
use of improper steel alloy and presence of voids in a steel
brazed joint, 275276(F&T)
quench nonuniformity, 189
quenching
carbonitrided components, 191
carburized components, 185, 187191
component design, 263265(F), 266(F&T)
contamination, 190
dry die quenching, 18
failures due to, 255284(F&T)
immersion quenching, 451
induction hardening, 449452(F)
martempering, 1718
phase transformations, 255263(F&T)
press quenching, 18
quench nonuniformity, 189
quenchant nonuniformity, 271272(F)
quenchant selection, 270271(T)

20/8/2008 5:06PM Plate # 0
638 / Index
quenching (continued)
slack quenching, 278
spray quenching, 430(F), 444(F), 451452
steel grade/condition, 266270(F&T)
atmosphere control, 269
component support/loading, 267(F), 268(F)
heating control, 269
machining, 267
retained austenite, 269270(T)
surface condition, 267269(F&T)
quenching cracks, 204207(F), 341
R
ratchet mark, definition, 505506
ratchet marks, 126(F), 388(F), 389(F), 506(F)
recognized and generally accepted good engineering practice
(RAGAGEP), 40
rehardening burn, 207
reoxidation, 166167
residual stresses, 196200(F&T), 472477(F)
retained austenite, 191196(F&T), 269270(T)
risers, 158
river patterns, 54, 55(F), 97(F), 376(F)
rolling, 130
hot rolling, 89
surface rolling, 122, 123(F), 124
S
scanning electron microscope (SEM), 21, 50(F)
shank crack, 388, 389(F)
sharp asperity, 130(F)
shear lip, 50, 51, 118, 119(T), 273
Sherlock Holmes rule, 114
shot blasting, 198
shot peening, 198199(F), 200(T)
shrinkage pores, 159, 160(F)
examples: failure analysis of a mill gear, 159161(F&T),
161162(F), 163(F), 164(F&T)
silicon, 56, 183, 223, 293, 296, 297(F), 334
siliconizing, 33
single-shot induction hardening, 422, 423, 424(F), 427,
428(F), 440(F), 496
single-shot induction heating, 424(F), 448(F)
sintering
ferrous PM parts, 397
oxidation after, 408409(F)
pinholes after, 404405, 406(F)
PM process map, 401(T), 402(F)
wear after, 407, 408(F)
slack quenching, 278
slant fracture, 56, 505
sliding, 56, 128(F), 129, 130(F), 195, 196(F), 199
slip band extrusions, 78, 80(F)
slip band intrusions, 78, 80(F), 122
Smith curve, 247248, 250(F)
pg 638
soft skin layers, 199200

softening replica tape, 4950
solid metal embrittlement, 6768(F&T), 69(F)
solid shrinkage, 158
solidification cracking, 515517(F)
solidification shrinkage, 157158(F)
solid-phase transformation model, 3132
spalling failure, 20(F), 23(F), 138, 231, 322(F), 323
spheroidization, 107(F)
spray quenching, 430(F), 444(F), 451452
stainless steels, PWHT of, 3740
austenitic chromium-nickel, 3839
chromium steels, 37
duplex austenitic-ferritic chromium-nickel, 3940
fully austenitic, 39
martensitic, 38
soft martensitic, 38
stabilized, 39
unstabilized austenitic, 39
static fatigue test, 301
static tests, 301
steady-state creep rate, 77, 139
steel hardness conversions
approximate equivalent hardness numbers for Brinell
hardness numbers for steel, 532533(T)
approximate equivalent hardness numbers for
steel, 535(F)
approximate equivalent hardness numbers for Vickers
(diamond pyramid) hardness numbers for steel,
533534(T)
approximate Rockwell B hardness conversion numbers for
nonaustenitic steels, 530(T)
approximate Rockwell C hardness conversion numbers for
nonaustenitic steels, 531(T)
examples of published hardness conversion equations,
529(T)
steel susceptibility ratio, 285
steel welds, failure analysis of
discontinuities, 503505(F&T)
examples
fatigue cracking of welded pipe flange, 506(F)
lack of penetration, 505
lack of side-wall fusion, 505
porosity in weld metal, 504, 505(F)
SCC of a weld, 514515(F)
solidification cracking of steel weld, 516(F), 517(F)
fatigue, 505506(F)
hydrogen-assisted cracking (HAC), 509513(F&T)
hydrogen-assisted cracking theory, 506509(F&T)
solidification cracking, 515517(F)
stress-corrosion cracking (SCC), 513515(F)
steels
alloyed, tempering, 293, 294(F&T), 295(F)
carbon. See carbon steels
carbon, isothermal diagrams
1019, 588(F)
1030, 589(F)
1050, 590(F)
1060, 591(F)
1080, 592(F)

20/8/2008 5:06PM Plate # 0
pg 639
Index / 639
carbon steels with nominal carbon contents

0.18 C (10171022), 602(F)
0.44 C (10391046), 603(F)
0.86 C (10801090), 604(F)
0.96 C (10901095), 605(F)
carburized. See carburized steels
chromium, 3739
1 Cr (5140), 624(F)
1 Cr V, (6150), 626(F)
11/4 Cr Mo (41404142), 625(F)
11/2 Cr Al Mo, 627(F)
21/2 Cr (5060) (51555160), 623(F)
5160, 597(F)
52100, 599(F)
chromium-molybdenum
4130, 593(F)
4140, 594(F)
clean, 165166, 229
cold work tool, 312(F)
continuous cooling diagrams, 601627(F)
high-speed, 313314
hot work tool, 312313(F)
induction hardening, 419420(F&T)
interstitial-free, 124, 125(F)
isothermal diagrams, 587599(F)
low-alloy steels. See low-alloy steels
Mn
11/2 Mn+S (1139), 607(F)
13/4 Mn (1547) (1345), 606(F)
Mn-Cr, 11/4 Mn Cr, 609(F)
Mn-Mo, 11/2 Mn Mo, 608(F)
Mn-Ni-Cr-Mo, 11/2 Mn Ni Cr Mo, 611(F)
Mn-Ni-Mo, 11/2 Mn Ni Mo, 610(F)
Mo, 1/4 Mo (40374042), 615(F)
nickel
1/2 Ni Cr Mo (86228627) (8720) (8822), 620(F)
11/4 Ni Cr, 619(F)
11/2 Ni Cr Mo, 621(F)
11/2 Ni Mn, 617(F)
13/4 Ni Mo, 618(F)
31/2 Ni, 616(F)
31/2 Ni Cr Mo (9310), 622(F)
nickel-chromium-molybdenum, 8620, 598(F)
nickel-molybdenum, 4640, 596(F)
Ni-Cr-Mo, 4340, 595(F)
nitrided, 244246(F&T), 247(F&T)
plastic mold, 313(F)
as quenched 4340, 273274(F)
silicon
13/4 Si Mn, 612(F)
2 Si Cr Mo, 614
31/2 Si Cr, 613(F)
tool, 311314(F&T), 314330(F&T)
stress
analysis, 115116
concentrations, 3233, 45, 46(F), 98(F)
machine parts, 485491(F&T)
versus strength, 9395(F)
types of, 78
stress raisers, 179(F), 272273
stress-corrosion cracking (SCC), 28

intergranular brittle fracture, 70, 7276(F&T), 77(F)
stress-sorption theory, 514
striations, 78, 80, 97, 124, 506. See also fatigue striations
stringer inclusion, 166, 280281(F)
stringers, 25(F), 103(F), 280, 281(F)
structural flaw, 241
subcritical annealing, 16, 266267
sudden onset deformation, 89
sulfidizing, 33
superposition principle, 127
surface (or skin) effect, 440
surface carbon content
decarburization, 213, 215217(F)
surface fatigue wear, 138
surface integrity, 418, 475
surface rolling, 122, 123(F), 124
surface structure anomalies, 199200
Systeme International dUnites (SI), 521
T
TBE, 299
temper colors for steels
color chart (back cover)
time-temperature effect, 539, 540(F)
temper embrittlement (TE), 296299(F)
HEM, interaction with, 301302(F)
LME, interaction with, 302303(T)
mechanical tests for, 299301(F&T)
mechanism of, 6566(F)
molybdenum, effect on, 297298
phosphorus, effect on, 297298
vanadium, effect on, 298
temperature conversion table, 525528
tempered martensite embrittlement, (TME), 65(F),
294296(F)
tempering
alloyed steels, 293, 294(F&T), 295(F)
case study: grinding cracks, 303304(F)
case study: transgranular and intergranular crack path,
304305(F&T), 306(F)
colors of tempering heats, 289(T)
embrittlement, 293303(F&T)
heating times, 290(T)
martensite, 285289(F)
mechanical properties, effect on, 289290(F), 291(F),
292(F)
reactions, 290291
stages, 291293(F)
tempering embrittlement, 285
tempering resistance, 36(F), 313, 331, 334, 338, 339
tensile, definition, 92
thermal expansion
austenite versus ferrite, 5
20/8/2008 5:06PM Plate # 0

640 / Index
thermal expansion (continued)

carbon and low-alloy steels, specific heats of, 549550(T)
coefficients of, 45, 6(F), 9(F), 32
coefficients of linear thermal expansion for carbon and
low-alloy steels, 541543(T)
ferrous materials, 5(T)
induced strain, 7
residual stress, influence on, 5, 32
summary of, 544545(T)
thermal conductivities of carbon and low-alloy steels,
547(T)
thermal shock, 56(T), 32
thermochemical modeling, 33
thermomechanical modeling, 3233
through hardening, 395, 396, 413, 421, 425(F),
442(F), 490
through-hardened, 413(F), 484, 485(F)
time-temperature transformation (TTT) diagrams, 2, 32,
202(F), 255, 256(F)
time-temperature-austenitizing (TTA) diagram, 432(F),
433(F)
toe cracking, 503(T), 504(F)
tool steels, failure analysis in
classification of, 311314(F&T)
cold work, 312(F)
cold work tools, 314330(F&T)
high-speed steels, 313314(F)
hot work, 312313(F)
hot work tools, 330349(F&T)
plastic mold steels, 313(F)
tooth-by-tooth induction hardening, 430(F), 439(F)
tooth-gap hardening, 431432(F)
torsional fatigue, 177, 483
total oxidation potential (TOP), 221(F)
total-life fatigue analysis, 177
transition temperature, 35, 55, 66, 67(F)
transmission electron microscopy, 65, 220, 288
tribological systems, 475
twinning density, 288
two-step embrittlement, 296
pg 640
U
ultrasonic testing (UT), 512
underbead cracking, 507, 510511(F)
undercut, 20(F), 46(F), 503(T)
undercutting, 4546(F)
unnotched impact testing, 337
upper bainite, 2, 270, 306(F), 403(F)
upset forging, 143145(F&T), 146(F)
V
volume wear, 138
W
wear, 9799(F)
wear, definition, 129
wear-assisted failure, 129131(F)
welded components, failure aspects of
aging tendency, 35
austenitic stainless steels, SCC of, 36
brittle fracture, 3334
cold cracking, 36
hardening tendency, 3536
segregation tendency, 36
temper embrittlement, 3435
welded components, heat treatment procedures
austenitic-ferritic dissimilar joints, 40
carbon steels, 3637
low-alloy steels, 3637
stainless steels, PWHT of, 3740
Y
yield point, 483(T)
yield strength, 144(T), 161(T), 250(T), 507
yielding, 127128
ASM International is the society for materials
engineers and scientists, a worldwide network

dedicated to advancing industry, technology, and
applications of metals and materials.
ASM International, Materials Park, Ohio, USA
This publication is copyright ASM International. All rights reserved.
Publication title
Product code
05113G
To order products from ASM International:

Online Visit www.asminternational.org/bookstore
Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US)
Fax 1-440-338-4634
Mail
Customer Service, ASM International

9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
Email CustomerService@asminternational.org
American Technical Publishers Ltd.

27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX,
In Europe United Kingdom
Telephone: 01462 437933 (account holders), 01462 431525 (credit card)
www.ameritech.co.uk
Neutrino Inc.
In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan
Telephone: 81 (0) 424 84 5550
Terms of Use. This publication is being made available in PDF format as a benefit to members and
customers of ASM International. You may download and print a copy of this publication for your
personal use only. Other use and distribution is prohibited without the express written permission of
ASM International.
No warranties, express or implied, including, without limitation, warranties of merchantability or
fitness for a particular purpose, are given in connection with this publication. Although this
information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be
obtained from the use of this publication alone. This publication is intended for use by persons having
technical skill, at their sole discretion and risk. Since the conditions of product or material use are
outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this
information. As with any material, evaluation of the material under end-use conditions prior to
specification is essential. Therefore, specific testing under actual conditions is recommended.
Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale,
use, or reproduction, in connection with any method, process, apparatus, product, composition, or
system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this
publication shall be construed as a defense against any alleged infringement of letters patent,
copyright, or trademark, or as a defense against liability for such infringement.

Failure Analysis of Heat Treated Steel Components (2008) PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Failure Analysis of Heat Treated Steel Components (2008) PDF

Transféré par

Droits d'auteur :

Formats disponibles

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_i.

12/8/2008 8:00PM Plate # 0

2008 ASM International. All Rights Reserved.

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_ii.pdf/Front Matter_Copyright/

12/8/2008 5:09PM Plate # 0

2008 ASM International. All Rights Reserved.

Name ///sr-nova/Dclabs_wip/Failure_analysis/5113_iii.pdf/Front Matter_Dedication Page/

19/8/2008 2:16PM Plate # 0

2008 ASM International. All Rights Reserved.

This book is dedicated to our families, without whose continued

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/

2008 ASM International. All Rights Reserved.

20/8/2008 12:41PM Plate # 0

Component Design ............................................................................................................ 1

Overview of the Mechanisms of Failure in Heat Treated Steel Components .................... 43

Mechanisms and Causes of Failures in Heat Treated Steel Parts ....................................... 87

General Aspects of Failure Analysis ............................................................................... 111

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/

2008 ASM International. All Rights Reserved.

20/8/2008 12:41PM Plate # 0

Failure in Steel Forging .................................................................................................. 133

Failures from the Casting Process .................................................................................. 151

Sources of Failures in Carburized and Carbonitrided Components ................................ 177

Fatigue Fracture of Nitrided Layers ............................................................................... 241

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/

2008 ASM International. All Rights Reserved.

20/8/2008 12:41PM Plate # 0

Steel Heat Treatment Failures due to Quenching ........................................................... 255

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/

2008 ASM International. All Rights Reserved.

20/8/2008 12:41PM Plate # 0

Main Features of Induction Heating .................................................................................

Failure Analysis of Steel Welds ...................................................................................... 503

Index ............................................................................................................................. 629

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_ix.pdf/Front Matter_Preface/

2008 ASM International. All Rights Reserved.

19/8/2008 3:39PM Plate # 0

ASM International is the society for materials

engineers and scientists, a worldwide network

Failure Analysis of Heat Treated Steel Components

To order products from ASM International:

Customer Service, ASM International

American Technical Publishers Ltd.

Failure Analysis of Heat Treated Steel Components

18/8/2008 2:47PM Plate # 0

Copyright 2008 ASM International

DESIGN involves different creative aspects:

Control of microstructure formation

Introduction to Heat Treat Processing

involve a complex interrelationship of variables

18/8/2008 2:49PM Plate # 0

2 / Failure Analysis of Heat Treated Steel Components

Important Design Aspects

18/8/2008 2:49PM Plate # 0

Table 1 Heat treating process for carbon steels

The optimal austenitization temperature is 3050 C (5590 F) above Ac3 for

The primary purpose of annealing is to soften

The hardenability depends on the concentration of dissolved carbon in the austenitic

18/8/2008 2:49PM Plate # 0

4 / Failure Analysis of Heat Treated Steel Components

promote heat treatment failures before the

where Q is the quantity of heat flowing through

where T is the temperature, t is the time, and a is