Failure Analysis of Heat Treated Steel Components PDF

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_i.
pdf/Front Matter_Title-page/ 12/8/2008 8:00PM Plate # 0 pg i
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
www.asminternational.org
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_ii.pdf/Front Matter_Copyright/ 12/8/2008 5:09PM Plate # 0 pg ii

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
Name ///sr-nova/Dclabs_wip/Failure_analysis/5113_iii.pdf/Front Matter_Dedication Page/ 19/8/2008 2:16PM Plate # 0 pg iii

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.
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
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/ 20/8/2008 12:41PM Plate # 0 pg vi

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
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/ 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
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_v-viii.pdf/Front Matter_Contents/ 20/8/2008 12:41PM Plate # 0 pg viii

Main Features of Induction Heating ................................................................................. 420

Induction Hardening of Machine Parts ............................................................................. 422
Magnetic Flux Concentrators ............................................................................................ 437
Conditions in Induction Heating and Quenching of Machine Parts ................................. 440
Time-Temperature Dependence in Induction Heating ..................................................... 444
Quenching Systems for Induction Hardening ................................................................... 449
Time Variation of Stresses and Residual Stresses ............................................................ 452
Workpiece Distortion in Induction Surface Hardening .................................................... 466
Residual Stresses after Induction Surface Hardening and Finish Grinding ..................... 472
Hardness Profiles in the Induction Surface-Hardened Layer ............................................ 477
Fatigue Strength of Materials ............................................................................................ 481
Stress Profiles in Machine Parts in the Loaded State ........................................................ 485
Input and Output Control of Steel for Induction Surface Hardening of Gears ................ 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 .......................................................................... 521

Appendix 2: Temperature Conversion Table .................................................................. 525
Appendix 3: Steel Hardness Conversions ....................................................................... 529
Appendix 4: Austenitizing Temperatures for Steels ........................................................ 537
Appendix 5: Temper Colors for Steels ............................................................................ 539
Appendix 6: Physical Properties of Carbon and Low-Alloy Steels ................................... 541
Appendix 7: AISI to Non-AISI Steel Cross Reference ..................................................... 551
Appendix 8: Non-AISI to AISI Steel Cross Reference ..................................................... 563
Appendix 9: Iron-Carbon Equilibrium Diagram ............................................................. 585
Appendix 10: Isothermal Diagrams of Selected Steels ................................................... 587
Appendix 11: Continuous Cooling Diagrams of Selected Steels ..................................... 601
Index ............................................................................................................................. 629
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 pro-
cessing, 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
Customer Service, ASM International
Mail
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/ 18/8/2008 2:47PM Plate # 0 pg 1
Failure Analysis of Heat Treated Steel Components Copyright 2008 ASM International
L.C.F. Canale, R.A. Mesquita, and G.E. Totten, editors, p 1-42 All rights reserved.
DOI: 10.1361/faht2008p001 www.asminternational.org
Component Design
Mario Solari, CTI Consultores de Tecnologa e Ingeniera SRL
Pablo Bilmes, Universidad Nacional de La Plata
DESIGN involves different creative aspects: involve a complex interrelationship of variables

planning, development, procedures, availability, associated with the material, manufacturing
and fitness concerning the materials and pro- processes, and service conditions (Ref 2).
cesses used to manufacture the component. The ability of ferrous materials to develop
Design is an iterative process, often based on required properties through heat treatment is
experience, to provide an assessment of the a broad concept that refers both to the ease
performance of a component for a certain period with which a material may be heat treated
of time of expected or intended service life. The and the resulting in-service fitness of the com-
design process culminates in a technical speci- ponent.
fication for the part or system and suitable The iron allotropic transformation between
manufacturing processes. Another obvious aim more densely packed face-centered cubic iron,
of design is to prevent failures throughout the nonmagnetic gamma (c) phase designated as
component lifetime cycle and avoid situations austenite, and the less densely packed body-
resulting in severe failure. centered cubic iron, alpha (a) phase designated
Heat treating achieves the desired changes in as ferrite, is the basis for heat treatment of steels.
structure and properties, and various types of Austenite can dissolve up to approximately
heat treatments may be employed to meet design 2.0 wt% C and in most steels is not stable at
requirements for mechanical strength, corro- low temperature. On the other hand, the inter-
sion, wear, and so on. Heat treatments include stitial sites in ferrite are much smaller than in
stress relieving, austenitizing, normalizing, austenite; therefore, ferrite can only dissolve
annealing, quenching, and tempering (Ref 1). very small concentrations of carbon (0.025 wt%
Heat treating may also involve chemical or maximum) and is relatively soft and stable at
additional physical processes. A systematic room temperature.
procedure for minimizing risks involved in heat The iron-carbon phase diagram shows the
treated steel components requires a combination compositional limits of the different transfor-
of metallurgical failure analysis and fitness for mational phases formed by a steel alloy that
service with respect to safety and reliability exist during heating or cooling as a function
based on risk analysis. The effects of steel heat of temperature. In hypoeutectoid steels (those
treatment may include (Ref 1): with 50.80 wt% C), upon cooling two different
phases can exist, ferrite and austenite, each con-
Control of microstructure formation
taining different amounts of carbon. Upon fur-
Increase of strength, toughness, or perhaps
ther cooling, the microstructure of these steels
creep resistance
exhibits ferrite grains in a pearlite island. Pearlite
Relief of residual stresses and prevention of
is a metastable microstructure formed during
cracking
Control of hardness (and softness) austenite decomposition. The pearlite structure
is an aggregate consisting of alternating lamellae
Improvement of machinability
of ferrite and cementite that is formed on slow
Improvement of corrosion resistance or wear
cooling during the eutectoid reaction. Cementite
resistance
is a very hard and brittle compound of iron
Introduction to Heat Treat Processing and carbon (Fe3C). Depending on the thermal
history, cementite will appear as lamellae (with
Material behavior related to heat treatment ferrite), spheroids, or globules in a ferritic
can be analyzed by developing models that matrix.
2 / Failure Analysis of Heat Treated Steel Components
Microstructures that are formed upon cool- these conditions, the surface of the part is initi-
ing and the proportions of each are dependent on ally cooled much more quickly than the core.
austenitization time, temperature, cooling his- Therefore, the specific volume in the core is
tory of the particular alloy, and the composition greater than at the surface, and the reduction in
of the alloy. The transformation products from volume at the surface is resisted by the greater
austenite decomposition change from a mixture volume in the core, resulting in the surface being
of ferrite and pearlite to bainite or martensite in tension and the core in compression. After the
with increasing cooling rate. cooling processes have been completed, the
Bainite is a two-phase mixture of ferrite residual-stress distribution between the surface
and cementite consisting of fine lines of iron and core is obtained. If the surface stresses
carbide in acicular ferrite. Upper bainite has a exceed the hot yield strength of the material, it
feathery appearance and forms just below the plastically deforms, resulting in thermally
temperature where fine pearlite is formed. induced dimensional changes (Ref 3). When
Lower bainite exhibits an acicular micro- steels that undergo transformational changes are
structure that is formed just above martensite, quenched, the possibility of the formation of
which is produced at approximately 350 C both thermal and transformational stresses must
(660 F). be considered.
Martensite is a supersaturated solid solu- Steel parts are often tempered by reheating
tion of carbon in alpha iron (ferrite) that is after quench hardening to obtain specific
less densely packed than the c body-centered mechanical properties. The tempering process
tetragonal lattice and is a magnetic platelike involves heating hardened steel to some tem-
structure formed by a diffusionless shear type perature below the eutectoid temperature for the
of transformation of austenite below the mar- purpose of decreasing hardness and increas-
tensite start (Ms) temperature. The amount of ing ductility and toughness while relieving
transformation depends on the martensitic quench stresses and ensuring dimensional stab-
temperature range (Ms to Mf). (Mf is the mar- ility. Tempering processes include temper-
tensite finish temperature.) The three forms of ing of martensite, transformation of retained
martensite are lath, plate, and tempered mar- austenite to martensite, tempering of the de-
tensite. composition products of martensite, and de-
Transformation from austenite to martensite composition of retained austenite to martensite.
results in a volumetric expansion at the Ms In addition, tempering may also lead to dimen-
temperature. Dimensional changes are possible, sional variation due to relaxation of residual
depending on the carbon content and micro- stress and plastic deformation, which is due
structural transformation product formed. The to the temperature dependence of yield
volume change (%) is [4.640.53 (%C)] for the strength. Tempering may lead to an increase in
reaction from austenite to martensite. hardness if secondary hardening occurs, which
The two most commonly used transforma- is due to precipitation of a compound or to
tion diagrams are time-temperature transfor- the formation of martensite or bainite from
mation for isothermal transformation, and retained austenite, decomposition during tem-
continuous cooling transformation diagrams. pering, or destabilization during this process
These diagrams can be used to predict steel and then transformation during subsequent
microstructures and hardness after heat treat- cooling.
ment, or they may be used to design a heat Quenchant selection and quenching con-
treatment process. Heat treating processes ditions are critically important parameters in
include hardening, austenitization, annealing quench system design. For example, the dim-
(full annealing, intercritical annealing and sub- ensional changes after austenitizing and then
critical annealing, recrystallization annealing, quenching in water are greater than quenching in
isothermal annealing, soft annealing, diffusion oil (Ref 3).
annealing), normalizing, stress relieving,
quenching and tempering, and austempering,
and are summarized in Table 1 (Ref 1).
Hardening and tempering are common Important Design Aspects
heat treatment processes. If steel is cooled suf-
ficiently fast, without microstructural transfor- The importance of good design cannot be
mation, thermal stresses can develop. Under overemphasized. Poor design can cause or
Component Design / 3
Table 1 Heat treating process for carbon steels

Process Characteristics
Austenitization: The optimal austenitization temperature is 3050 C (5590 F) above Ac3 for
Complete transformation to austenite by hypoeutectoid steels and 3050 C (5590 F) above Ac1 for hypereutectoid steels.
heating the steel above the critical Ac3 is the temperature at which the transformation of ferrite to austenite is completed
temperature for austenitic formation 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: Full annealing: Heat 3050 C (5590 F) above Ac3 for hypoeutectoid steels, then
Heat treatment consisting of heating and furnace cool through the critical temperature range at a specified cooling rate. The aim
soaking at suitable temperature followed by is to break the continuous carbide network of high-carbon steels. It improves
cooling under conditions such that, after machinability.
return to ambient temperature, the metal will Partial (intercritical) annealing: Heating within the critical temperature range (Ac1Ac3),
be in a structural state closer to that of equi- followed by slow furnace cooling. It improves machinability.
librium Subcritical annealing: Heating 1020 C (2035 F) below Ac1 followed by cooling in
The primary purpose of annealing is to soften still air. It can be used to temper bainitic or martensitic structures to produce softened
the steel to enhance its workability and microstructures containing spheroidal carbides in ferrite. Improves the cold working
machinability. Also, it relieves internal properties of low carbon steels (525% C) or softens high-carbon and alloy steel
stresses, restores ductility and toughness, Recrystallization annealing: Heat the steel for 30 min1 h at temperature above the
refines grains, reduces gaseous content in the recrystallization temperature (TR=0.4 Tm), then the steel is cooled. The treatment
steel, and improves homogenization of temperature depends on prior deformation, grain size, and holding time. The
alloying elements. 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.
Normalizing: Heat the steel to 4050 C (8090 F) above Ac3 for hypoeutectoid steels and 4050 C
The aim is to provide a uniform (8090 F) above Acm for hypereutectoid steels. The holding time depends on the size,
microstructure of ferrite plus pearlite (small and then the steel is cooled in still air. It produces grain refinement and improved
grains and finer lamellae than in annealing). homogenization.
Stress relieving: Heat to a temperature below Ac1for the required time to achieve the desired reduction in
It is typically used to remove residual residual stresses, and then the steel is cooled at a rate sufficiently slow to avoid the
stresses that have accumulated from prior formation of excessive thermal stresses. Below 300 C (570 F), faster cooling rates
manufacturing processes. Stress relieving can be used. No microstructural changes occur during stress-relief processing. The
results in a significant reduction of yield recommended heating temperature range is 550700 C (10201290 F), depending on
strength in addition to reducing the residual the type of steel. These temperatures are above the recrystallization temperature. Little
stresses to some safe value. 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: The hardenability depends on the concentration of dissolved carbon in the austenitic
Ability to develop hardness to a given phase, alloying elements, austenitizing temperature, austenitic grain size at the moment
depth after having been austenitized and of quenching, size and shape of the cross section, and quenching conditions.
quenched
Quenching: Specific recommendations for quench media selection for use with various steel alloys are
Quench severity is the ability of a provided by standards such as SAE AMS 2759. Quench media include water, brine,
quenching medium to extract heat from a aqueous polymer, gas or air quenching, and caustic quenching.
hot steel workpiece.
Tempering: The tempering process involves heating steel to any temperature below the Ac1
Tempering is the thermal treatment of temperature. During tempering, as-quenched martensite is transformed into tempered
hardened and normalized steels to obtain martensite, which is composed of highly dispersed spheroids of cementite (carbides)
the desired mechanical properties, which dispersed in a soft matrix of ferrite, resulting in reduced hardness and increased
include improved toughness and ductility, toughness. The objective is to allow the hardness to decrease to the desired level and
lower hardness, and improved dimensional then stop the carbide decomposition by cooling. The extent of the tempering effect
stability. is determined by the temperature and time of the process.
Source: Ref 1
promote heat treatment failures before the properties l, r, and Cp vary with temperature.
component is put into service, or it may reduce The differential equation of heat conduction
service life, sometimes dramatically. The establishes the relation between the time and
objective of proper design for heat treatment is space variation of temperature at any point
to provide the minimum engineering require- of the body in which conduction takes place.
ments, the desired material properties at the The factor of proportionality thermal diffusivity,
lowest total cost, and, in particular, to minimize a, represents a physical property of the material,
the expense of scrap due to rework of parts that is essential for transient processes of heat flow,
may have undergone excessive distortion or and defines the rate of change of temperature.
cracked. If the thermal conductivity, l, is the ability of
The Heat-Transfer Theory Applied to Heat a solid to conduct heat, thermal diffusivity is
Treatments. The laws that govern heat trans- the measure of a material thermal inertia. The
mission are very important to the engineer in quantity rCp is the volumetric specific heat;
heat treatment design. There are three different this product is approximately constant for solid
types of heat transfer: conduction, convection, metals. So, in the case of austenitic stainless
and radiation (Ref 4). They have in common that steels, low thermal conductivities correspond
temperature difference (thermal gradient) must to low thermal diffusivities. In other words,
exist and that the heat is always transferred in the equalization of temperature at all points of
direction of decreasing temperature. When the space will proceed at a lower rate in austenitic
temperature profile does not change with time, stainless steels, with respect to ferritic steel, due
the fundamental relation for the unidirectional to its lower thermal diffusivity, and there are
steady flow of heat through a solid by conduc- difficulties in homogenizing temperature during
tion, Fouriers first law, can be expressed by: heat treatments.
Heat transfer by convection occurs between
T
Q=7l (Eq 1) the surface of the body and surrounding fluids;
x for this type of heat transmission, the following
where Q is the quantity of heat flowing through Newtons equation is in general use:
the unit area of a wall per unit time in the
Qh =hDT (Eq 4)
direction of the x-axis and is directly propor-
tional to the thermal conductivity, l, and the It simply states that an invariable temperature
thermal gradient in x-direction. Thermal con- difference, DT, between a surface and a fluid in
ductivity has a nearly linear dependence on contact with it causes a steady heat flow of Qh.
temperature. The factor of proportionality, h, is called the
During heat treatments, the temperature var- coefficient of heat transfer.
ies in time as well as in space; these processes The third type of heat-transfer mode is
are called unsteady, nonstationary, or transient. radiation. The heat flow by radiation is com-
As the body heats, the temperature at each point monly written:
asymptotically approaches the temperature of

the medium. The temperature of points near the Q=es T14 7T04 (Eq 5)
surface of the body changes most rapidly. The
differential equation for one-dimensional tran- where e is the emissivity (1 for a black body),
sient heat conduction, Fouriers second law, in T is the temperature, and s is the Stefan
the absence of inner heat sources is: constant (Ref 5). The three different types of
heat transfer, conduction, convection, and
T 2 T radiation, are present during the heat treatment
=a 2 (Eq 2)
t x processes.
Perhaps the most important physical property
where T is the temperature, t is the time, and a is of steel to be considered in design is its coeffi-
the thermal diffusivity of the metal and is: cient of thermal expansion. Most heat treating
l
problems could be solved if this coefficient
a= (Eq 3) could be controlled. Because it cannot, it is
rCp
necessary to learn to design with it (Ref 6).
where l is the thermal conductivity, r is the Almost all solids expand on heating. As the
density, and Cp is the specific heat at constant temperature is raised, the thermal vibration
pressure of the material. The thermophysical pushes the atoms apart, increasing their mean
spacing. The effect is measured by the linear Thermal expansion of austenite is larger than
expansion coefficient: that of ferrite. These thermal expansion values
are representative of solution-annealed material.
1 dl Subsequent precipitation hardening treatments
a= (Eq 6)
L dT may affect thermal expansion (Ref 7).
where L is a linear dimension of the body In high-temperature components design,
(Ref 5). incompatibility of thermal expansion becomes a
The relationship between thermal con- major problem. Choice of materials and designs
ductivity and thermal expansion is important in should take this into account. Bolts used to hold
designing against thermal distortion. Thermal high-temperature casings together must be
gradients can cause a change of shape, which is a selected to have sufficient elevated-temperature
distortion of the component. The strain is related strength and make a good thermal expansion
to temperature by: match with the casing material. When rotor and
casing are made of ferritic steel, modified 12%
e=aT0 7T (Eq 7) Cr bolts work well up to 565 C (1050 F), but
where T0 is ambient temperature, and a is ther- nickel-base superalloys are needed at 595 C
mal expansion (Ref 5). (1100 F) or higher.
The ability of a material to resist thermal
The distortion is proportional to the gradient
of the strain, so it is proportional to the thermal shock, due to a sudden immersion in a cold
gradient. By Fouriers first law, the heat flow is medium, without cracking depends on its ther-
mal expansion coefficient, a; tensile strength,
proportional to the thermal gradient through
the thermal conductivity, l. For a given geo- st, for metals; Youngs modulus, E; thermal
metry and heat flow, the distortion is mini- conductivity, l; and heat-transfer coefficient, h.
mized by selecting materials with large values A temperature change of DT applied to a
constrained body or a sudden change DT of the
of l/a (Ref 5). For example, austenitic stain-
less steels have low thermal conductivity and surface temperature of the unconstrained com-
high thermal expansion, related to ferritic steel, ponent induces a stress:
so distortion during welding becomes a pro- EaDT
blem. s= (Eq 8)
C
Thermal expansion has a strong influence on
the development of residual stress. Whenever where C is equal to 1 for axial constraint, (1n)
the thermal expansion or contraction of a body is for biaxial constraint, (12n) for triaxial con-
prevented, thermal stresses appear; if large straint, and n is Poissons ratio. If this thermally
enough, they cause yielding, fracture, or elastic
collapse (buckling). For axial constraint, the Table 2 Linear thermal expansion for ferrous
stress, Ds, produced by a temperature change of materials
1 C or the stress per C caused by a sudden Linear thermal expansion
in temperature range
change of surface temperature in one that is not from room temperature
constrained is equal to aE, where a is the to 704 C (1300 F)
expansion coefficient, and E is the elastic mod- Steel mm/m in./ft
ulus of the material. For biaxial and triaxial Carbon and low-alloy steels:
constraint, the stresses shall be multiplied by C; C-Mn; C-Si; C-Mn-Si; C-1/2 Mo to 10.3 0.124
11/4 Cr-1/2 Mo; Mn-1/2 Mo-1/2 Ni
(1n) and (12n) respectively, where n is 5Cr-1Mo and 29Cr-7Ni-2Mo-N steels 9.5 0.114
Poissons ratio. These stresses are large and can 9Cr-1Mo steel 8.8 0.106
cause a material to yield, crack, spall, or buckle 12Cr; 12Cr-1Al; 13Cr; and 13Cr-4Ni 8.5 0.102
steels
(Ref 5). 15Cr and 17Cr steels 8 0.096
Linear thermal expansion (Table 2) in going 27Cr steels 7.3 0.088
from room temperature to 700 C (1300 F) is Austenitic stainless steels:
approximately: 16Cr-12Ni-2Mo; 16Cr-12Ni-2Mo-N;
16Cr-12Ni-2Mo-Ti; 18Cr-8Ni 13.1 0.157
10.3 mm/m (0.124 in./ft) for low-alloy steel
Austenitic stainless steels:
8.5 mm/m (0.102 in./ft) for martensitic
29Ni-20Cr-3Cu-2Mo; 20Cr-18Ni-6Mo;
stainless steel (type 13Cr) 22Cr-13Ni-5Mn 12.25 0.147
13.1 mm/m (0.157 in./ft) for austenitic Source: Ref 7
stainless steel (type 18Cr-8Ni).
induced stress exceeds the local tensile strength other hand, for fast air flow (h=102 W/m2K),
of the material, yielding (permanent plastic the factor A results are equal to 3 10 2 (for
deformation) or cracking results (Ref 5). This section s=10 mm), and the thermal shock
plastic flow causes permanent shape change resistance DT is larger by the factor 1/A (Ref 5).
(distortion) and impacts the magnitude and dis- As an example of the use of the aforemen-
tribution of residual stresses. Water quenching tioned equations, Fig. 1 shows schematically the
gives a high h, and then the values of DT cal- effect of the thermal expansion coefficient (a)
culated from the previous equation give an and the heat-transfer coefficient (h) in thermal
approximate ranking of thermal shock resistance shock resistance (DT) for a hypothetical steel
(Ref 5). with 800 MPa (~120 ksi) tensile strength, bi-
However, when heat transfer at the surface is axial constraint, and thermophysical properties
poor and the thermal conductivity of the solid is constant with temperature. Three cases were
high, the thermal stress is less than that given by analyzed:
the previous equations. A measure of the thermal
shock resistance that takes into account the finite a) In the first case, a ferritic steel with low
rate of heat transfer at the surface, a heat-transfer thermal expansion and a very high-heat
coefficient that is never infinite, is given by: transfer coefficient (fast water quench,
st h=104 W/m2K) was considered. The max-
BDT= (Eq 9) imum temperature change (thermal shock
aE
resistance) that induces stresses below the
where st, a, and E were defined previously, tensile strength, avoiding yielding or crack-
B=C/A, where C also was defined previously; ing, is 240 C (470 F).
and A is: b) The second example similar to case (a) but
sh=l with high thermal expansion. The thermal
A= (Eq 10)
1+sh=l shock resistance results in temperatures
above 135 C (270 F) increasing the failure
where s is a typical dimension of the sample in
risk.
the direction of heat flow, h is the heat-transfer
c) The third example is similar to case (a) but
coefficient, and l is the thermal conductivity
with a lower heat-transfer coefficient (air
(Ref 5). The quantity Bi=sh/l is usually called
flow, h=50 W/m2K). The thermal shock
the Biot modulus. If Bi41, heat flow is limited
resistance results in 280 C (535 F)
by conduction. For fast water quench of metals,
decreasing the risk of failure.
the heat-transfer coefficient, h, is high (h=
104 W/m2K), and the thermal conductivity is These factors (residual stresses and dimen-
also high, so the factor A approaches 1. On the sional changes) have the greatest influence on
900
Tensile strength
800
(120)
700
Induced thermal stress for very
600 high heat-transfer coefficient, h,
Stress, MPa (ksi)
(90) and low thermal expansion

500 Induced thermal stress for very
400 high heat-transfer coefficient, h,
(60) and high thermal expansion
300 Induced stress for low heat-
transfer coefficient and low
200
(30) Thermal shock thermal expansion
100 resistance
0
0 50 100 150 200 250 300
(90) (180) (270) (360) (450) (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
the design process of a component. In addition to lysis or by using approximate formulas that
thermal strains, many materials systems undergo account for the most common cases. Design
phase transformations as a function of tem- should be verified to confirm whether there are
perature. Often, the new phase(s) that forms has stress-concentration points that may activate
a different volume and different coefficient of failure mechanisms due to brittle fracture, cor-
expansion as well as different mechanical rosion, or fatigue.
behavior(s) than the parent phase(s). For exam- Examples of peak stresses are thermal stresses
ple, phase transformation from austenite to in the austenitic steel cladding of a carbon steel
martensite results in a volumetric expansion at vessel, thermal stresses in the wall of a vessel or
the martensite start (Ms) temperature. These pipe caused by rapid change in temperature of
differences increase the complexity of under- the contained fluid, and the stress at a local
standing the effect of thermal gradients on the structural discontinuity.
strains produced and the resulting plastic Residual stresses (Ref 9) can be defined as
deformation (Ref 8). Thermal and transforma- those stresses that remain in a material or body
tion-induced strains can result in substantial after being manufactured and processed in
plastic deformation and residual stresses. The the absence of external forces or thermal gra-
total induced strain is the result of the sum dients. Residual stresses can be defined as either
of the strain produced by thermal expansion macro- or microstresses, and both may be pre-
(aE DT=eth) of a piece with initial length (E) sent in a component. Macroresidual stresses
and the transformation strain (etr). The total vary within the body of the component over a
induced strain must be accommodated through much larger range than the grain size. Micro-
either elastic (eel) or plastic (eep) strain, which residual stresses, which result from differences
sums to the total strain et=eth +etr= within the microstructure of a material, operate
(eel +eep). In order to determine the accom- at the grain-size level or at the atomic level.
modation strain values, Youngs modulus (E) Microresidual stresses often result from the
and the yield strength (sys) are required as a presence of different phases or constituents in a
function of phase and temperature. Most of the material.
plastic deformation occurs during the heat-up Residual stresses develop during most
and cool-down stages of the process (Ref 8). manufacturing processes involving material
Primary Stresses, Secondary Stresses, Peak deformation, heat treatment, machining, or
Stresses, and Residual Stresses. Primary processing operations that transform the shape
stress is a normal or shear stress developed by or change the properties of a material. They arise
the imposed loading that is necessary to satisfy from a number of sources and can be present in
the laws of equilibrium of external and internal the unprocessed raw material, introduced during
forces and moments. The basic characteristic manufacturing, or can arise from in-service
of a primary stress is that it is not self-limiting. loading.
Primary stress that considerably exceeds the In heat treated parts, residual stresses may be
yield strength will result in failure or at least classified as those caused by a thermal gradient
in great distortion. Secondary stress is a normal alone or a thermal gradient in combination
or shear stress developed by the constraint of with a microstructural change (phase transfor-
adjacent parts or by self-constraint of a structure. mation). When a steel part is quenched from
The basic characteristic of the secondary stress the austenitizing temperature to room tempera-
is that it is self-limiting. An example of sec- ture, a residual-stress pattern is established due
ondary stress is a general thermal stress. to a combination of a thermal gradient and a
The elastic stresses calculated previously are local transformation-induced volume expan-
nominal values, that do not take into account sion. Thermal contraction develops nonuniform
local discontinuities such as holes, notches, or thermal (or quenching) stress due to different
section changes. Even on a structure where rates of cooling experienced by the surface
stress intensity has been limited by yield criteria, and interior of the steel part. Transformational
there may exist highly localized regions where volume expansion induces transformation stress
peak stresses are several times higher than yield. arising from the transformation of austenite into
Maximum local stresses on a structure can be martensite or other transformation product
determined by considering nominal stresses (Ref 10).
multiplied by a stress-concentration factor and Residual stresses may be sufficiently large to
can be estimated through a detailed stress ana- cause local yielding and plastic deformation,
both on microscopic and macroscopic levels, stress gradients in the surface layers of the
and can severely affect component perfor- component.
mance. Both the magnitude and distribution of The criterion applied to avoid plastic defor-
the residual stress can be critical to perfor- mation states that the calculated stress intensity
mance and should be considered in the design or effective stress must be lower than the yield
of a component. In any free-standing body, and design life creep-rupture stresses of the
stress equilibrium must be maintained, which material. When effective stress is exceeded
means that the presence of a tensile residual somewhere within the component, it does not
stress in the component will be balanced by a necessarily indicate plastic collapse of the entire
compressive stress elsewhere in the body. structure. Primary stresses may locally exceed
Tensile residual stresses in the surface of a yield, within certain limits, provided that there is
component are generally undesirable, since enough ductility to allow the material to yield
they can contribute to, and are often the major without cracking. Plastic collapse occurs when
cause of, fatigue failure, quench cracking, and primary stresses are uniform on the entire
stress-corrosion cracking. Compressive resi- structure and exceed effective stress. To prevent
dual stresses in the surface layers are usually an incremental collapse or thermal stress ratchet
beneficial, since they increase both fatigue in each loading cycle, the total elastic stress-
strength and resistance to stress-corrosion intensity range, considering residual and applied
cracking and increase the bending strength of stresses, should be limited to twice the yield
brittle ceramics and glass. In general, residual stress.
stresses are beneficial when they operate in Factors Leading to Size and Shape Changes
the plane of the applied load and are opposite in Heat Treated Components. Within a typi-
to it (for example, a compressive residual stress cal component manufacturing process, there are
in a component subjected to an applied tensile seven major factors that lead to size and shape
load). changes and the development of residual stres-
The origins of residual stresses in a compo- ses in heat treated components (Ref 8):
nent may be classified as mechanical, thermal,
Variation in structure and material compo-
and chemical. Mechanically generated residual
sition throughout the component, leading to
stresses are often a result of manufacturing
anisotropy in properties and transformation
processes that produce nonuniform plastic
behavior
deformation. They may develop naturally during
Movement due to relief of residual stresses
processing or treatment or may be introduced
from prior machining and forming opera-
deliberately to develop a particular stress profile
tions
in a component. Examples of operations that
Creep of the part at elevated temperature
produce undesirable surface tensile stresses
under its own weight or as a result of fix-
or residual-stress gradients are rod or wire
turing
drawing (deep deformation), welding, machin-
Large differences in section size and asym-
ing (turning, milling), and grinding (normal or
metric distribution of material, causing dif-
harsh conditions).
ferential heating and cooling during
On a macroscopic level, thermally generated
quenching
residual stresses are often the consequence of
Volume changes caused by phase transfor-
nonuniform heating or cooling operations.
mation
These, together with the material constraints
Nonuniform heat extraction from the part
in the bulk of a large component, can lead to
during quenching
severe thermal gradients and the development
Thermal expansion
of large internal stresses. An example is the
quenching of steel (or aluminum alloys), All of these factors, except relief of prior
which leads to surface compressive stresses residual stresses (second item) and creep at
balanced by tensile stresses in the bulk of the elevated temperature (third item), can be
component. directly related to thermal and transformation-
Chemically generated stresses can develop induced strains in the component (Ref 8).
due to volume changes associated with chemical A simple example of how this thermophysical
reactions, precipitation, or phase transformation. property affects heat treating is given in Fig. 2
Chemical surface treatments and coatings can (Ref 11). As the shaft is quenched, the corner
lead to the generation of substantial residual- cools first, and as it shrinks, it mechanically
upsets the hot steel beneath it. As the quench Fig. 3 (Ref 11). Because its thin sections are
progresses, the entire shaft cools, but now, relatively rapidly heated, this part could not be
because the end is hot upset, the diameter is too made to the required tolerances. The redesign
small to accommodate the circumference. As a shown on the right was an improvement, but the
result, the end is (usually) in a high state of necessary large access holes were still trouble-
residual tensile stress, and if the steel is brittle, some.
quench cracks may develop. Several other factors at the design stage can
The coefficient of expansion is a factor that contribute to problems traceable to austeniti-
requires serious design consideration because it zation:
affects a part during austenitization. With fur-
Combinations of components with widely
nace heating, a part is heated to the austenitiza-
tion temperature mainly by radiation (80 to varying (nonuniform) section sizes
Designs requiring contact with furnace
98%) and partly by convection (2 to 20%). By
hearths or placement near walls
radiantly heating particular portions of a part,
thin sections heat fastest, especially those that Designs requiring processing that results in a
state of high residual stress before austen-
expose a large surface area, such as a spline or
itization.
gear. A typical example is the gear on the left in
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 inter-
dependent. The choice of a certain grade of steel,
such as one that must be brine or water quen-
ched, will affect all aspects of design and man-
ufacture. 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 quench-
ing. The design must also be compatible
Fig. 2 Effect of coefficient of thermal expansion in heat with the equipment available, for example,
treating a shaft. Source: Ref 6 heat treatment furnaces and surface-finishing
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
devices. Designing tools and dies is more diffi- knowledge available for guidance.
cult than designing components made from Analysis of many tool and die failures shows
structural steels because of the difficulty in that two relatively simple design problems
predicting service stresses. Despite advances cause the most failures. These design short-
made in design procedures, much of the design comings are the presence of sharp corners and
work is still empirically based. Such experience the presence of extreme changes in section mass
is primarily based on past failures; therefore, it is (Ref 12). A sharp corner concentrates and
important that the findings of the failure analyst magnifies applied stresses, stresses that arise in
be incorporated into future work. Despite the tool and die manufacturing (such as during
shortcomings of the empirical approach, there quenching), or stresses that occur during service.
is a vast body of common-sense engineering 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 liquid-
quenching 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
Fig. 4(a, b) Lathe tool bit of 1.45% C and 1.4% Cr steel elevator that failed as a result of reverse-bending
with acute-angled and rough-machined cross- fatigue. In this case, the fracture also propagated
sectional transition that fractured during hardening. (a) Fracture.
Original magnification: 1. (b) View into angle. 2. Source: from a sharp-edged cross-sectional transition.
Ref 13 To avoid further damage and prevent potential
Fig. 5 Bolt of a self-service elevator that failed as a result of reverse-bending fatigue. Source: Ref 13
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 cause many errors (Ref 12). Not every steel user
were examined metallographically or in bending is in a position to select the most suitable
tests. Eight of these proved to have incipient material for his purpose from the many varieties
fatigue cracks in the cross-sectional transitions. available. When in doubt, consult the steel
The bolts were partially normalized and partially manufacturer whose materials specialists pos-
heat treated. Their strength was determined from sess the necessary knowledge of mechanical and
Brinell hardness to be between 440 and technological properties of the required mater-
700 MPa. Cracks had occurred in the annealed ials. A close cooperation between the materials
as well as the heat treated bolts, that is, in soft as specialist of the producer and the designer and
well as hard bolts. The higher strength of the heat plant engineers of the user is the best formula for
treated bolts was made ineffective by the unfa- success.
vorable design. One of the overwhelming causes of steel
Figures 6(ae) show different parts of rock cracking and unacceptable distortion control
drills of different durability and made of a steel is part design (Fig. 911). Poor part design
with composition 0.95% C, 1.2% Cr, and 0.25% promotes distortion, cracking, and nonsym-
Mo. They had failed after a short period of ser- metrical heat transfer during heating and cool-
vice in the hexagonal shaft, while others had ing.
proved free of defects. At first glance, it could be Sometimes, designers make designs in which
seen that the broken drills had sharp edges combinations of parts intended to reduce costs
(Fig. 6a), while those free of defects were can actually increase cost due to problems
well rounded off (Fig. 6b). Fatigue fractures during austenitization. The classic example is
propagated from the sharp edges. These, in turn, the gear-and-hub combination shown in Fig. 12
led to catastrophic failures under shock loading (Ref 11). As this part is heated, the thin extrem-
(Fig. 6c). The design differences could be ities at the top of the hub heat faster than the
clearly seen in the cross section (Fig. 6d,e). sections near the gear. Accordingly, this area has
Etching showed that the failed drills also were a propensity to increase in size but is restrained
surface decarburized, which further reduced the by the colder metal nearer the gear; therefore, it
fatigue strength. upsets itself (yields in compression). Finally, the
Figures 7(a,b) show a compressor transmis- entire part comes to the prescribed tempera-
sion shaft with a fracture propagating from an ture. On cooling, however, and even without
acute-angled keyway, and Fig. 8 shows a drive quenching, the top end of the hub pinches in
shaft pinion with fatigue fractures propagating because it has upset itself. This upsetting can
from the acute-angular edge of the helical gear. result in a serious taper condition in the bore. (If
On the other hand, ignorance, carelessness, the bore is broached before heat treatment, the
and false economies in the selection of materials extremities of the hub will stretch and then close
in, thus causing additional taper.) The top austenitization, making forging easier and, in
portion of the hub of the pinion in Fig. 12(a) is most cases, reducing total cost. The bevel pinion
used only as a spacer and need not be heat shown in Fig. 13 presents a similar problem,
treated. A much shorter hub with a steel tubing although the hub extension is necessary. Here,
spacer (Fig. 12b) would solve the problem in steel or, preferably, a heat-resistant alloy cap
Fig. 7(a, b) 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. 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 Fig. 11 Distortion often encountered when quenching a
the distortion shown. Source: Ref 3 notch. Source: Ref 3
that will create mass can be put over the thin hub As in quenching, however, induction tooling
before austenitization to retard the heating rate can be designed to concentrate flux by using
during carburizing. appropriate coil configuration and laminated
Because of the extremely rapid heating core material. The different frequencies avail-
rate, austenitization with high-frequency elec- able provide not only for various depths of
tric current can be likened to reverse rapid heating but also for the sharpness of the heating
quenching. Accordingly, design is of utmost effect, because the induction-heated layer is
importance. As a general rule, steel exposed often much thinner than the hardened depth of
to the flux of the inductor will heat fastest on austenitized steel. The extent of conduction
corners, around holes (as shown in Fig. 14), and is a function of the differential between the
through thin sections. The bottoms of keyways surface temperature and that of the core. Thus,
and the roots of gear teeth and splines are aus- preheating, either in the induction coil with a
tenitized last, often mainly by conduction from suitable delay or in a furnace, can be employed
adjacent areas. 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
Finally, part design recommendations to fillets wherever possible.

avoid distortion and cracking problems (Fig. 15) Use air-hardening or high-carbon (oil- and
include: air-hardening) tool steel on unbalanced and
intricately shaped dies.
Parts that exceed the following dimensions Add extra holes, if possible, on heavy,
often must be straightened or press quenched unbalanced sections to allow for faster and
to maintain dimensional stability: long and more uniform cooling when quenched.
thin parts, L=5d for water quenching and Do not machine knife blades to a sharp cut-
L=8d for oil quenching (L is the length, ting edge before hardening.
and d is the thickness or diameter); and parts Avoid deep scratches and tool marks. The
that possess large cross-sectional area (A) insertion of identification marks on the
and are thin (t), which are defined as A=50t. hardened component is recommended, pre-
Balance the areas of mass. ferably after hardening, with tools having
Avoid sharp corners and reentrant angles. well-rounded edges and minimum defor-
Avoid sharp corners between heavy and thin mation (shallow penetration depth) and at
sections. positions far away from the high-stress-
Avoid single internal or external keys, key- concentration zones (reentrant angles,
ways, or splines. bends, and so on).
Provide adequate fillet or radius at the base On long, delicate parallels, shafts, and so on,
of gear teeth, splines, and serrations. rough out and have pieces annealed to
Do not have holes in direct line with the remove stresses before finish machining.
sharp angles of cutouts. Always use the grade or composition of steel
Avoid sharp corners at the bottom of small most suitable for the work that the part has to
openings, such as in drawing or piercing perform.
dies, because spalling or flaking is likely to Design symmetry is also an important vari-
result at these points. able to minimize distortion.
Keep hubs of gears, cutters, and so on as near A general rule for solving such quench dis-
the same thickness as possible, because tortion problems is that the short side is the
dishing is likely to occur. hot side, which means that the inside of the
Order stock large enough to allow for bowed metal was quenched more slowly
machining to remove decarburized surfaces than the opposite side.
and surface imperfections, such as laps and
seams. Steel Grade and Condition. Although steel
Do not drill screw holes closer than 6.35 mm cracking is most often due to nonuniform
(0.25 in.) from the edges of die blocks heating and cooling, material problems may
or large parts, where possible. Cracking be encountered. Some typical material prob-
may be avoided by using steel that may be lems include the compositional tolerances:
hardened by using lower quench severity, dirty steels, those containing greater than
or, if possible, pack the bolt hole to 0.05% S (such as SAE 1141 and SAE 1144),
reduce thermal stresses arising due to are more prone to cracking. It is well known
quenching. that cracking propensity increases with carbon
Avoid blind holes, if possible. content. Therefore, the carbon content of the
Design all parts with round corners and steel is one of the determining factors for
Fig. 14 Section through a hole in a part following rapid heating in an induction coil, showing distortion that leads to cracking.
Source: Ref 11
quenchant selection. As a rule of thumb, plain means that higher-cost, higher-alloy steels are
carbon steels with less than 0.35% C rarely required for adequate response to the slower
crack on hardening, even under severe quench- oil quench when carbon content exceeds
ing conditions. The carbon content of steel 0.35%.
should never be greater than necessary for However, when strength alone (and/or
the specific application of the part. As a general hardness), without the toughness of a quenched
rule, steels with more than 0.35% C will and tempered microstructure, fulfills the mini-
require oil quenching to avoid cracking. This mum engineering requirements, the use of
Fig. 15 Part design recommendations for minimal internal stresses. Source: Ref 3
cold-finished bars made with extra-heavy draft In considering the alternatives that relate to
or elevated-temperature drawing should receive minimization of distortion, it is assumed that the
consideration. In fact, why heat treat when it is grade of material is fixed, and no deviations are
unnecessary? Thus, SAE 1050, SAE 1140, or allowed in this area. There are often instances,
SAE 1144 steels should be particularly attractive especially for parts of complex design, where a
to firms with no heat treating facilities or with no change in steel composition will permit a less
commercial heat treater nearby. By adjustment drastic quench and thereby reduce distortion.
of their composition and the degree to which Such changes are usually to steels with higher
they are cold (or warm) worked, these steels can hardenability.
be made to have good machining characteristics In most instances, however, immediate
(Ref 6). changes in workpiece composition are not fea-
On the other hand, it is well known that sible. Pros and cons of the three most likely
regions containing high concentrations of coarse alternatives (listed previously) are discussed
carbide microstructure as a result of improper separately in the paragraphs that follow.
forging may become the initiation point for Consider Change to Another Process. In
subsequent quench cracking, particularly with this area, there are sometimes two or three pos-
parts of complex shape. It is important to sibilities, such as changing from a through-
provide a sufficient forging for microstructure to hardening steel to a case-hardening type or
become fine and uniform. Because part manu- changing to one that does not require rapid
facture, such as gear production, often requires cooling, such as nitriding. One of the most likely
machining, the condition of the steel that is changes that is often made in this area is to the
going to be machined is critically important. use of localized heating, such as induction. For
Some workers have recommended normalized example, a shaftlike member requires hardening
and subcritical-annealed steels as the ideal only in certain bearing areas. This can be
condition. The subcritical annealing process accomplished easily by induction and, in addi-
reduces the carbon content and alloy carbide tion to eliminating distortion, is often more
content in the austenite, allowing the production economical for other reasons. Parts such as ring
of more lath martensite in the microstructure, gears represent other examples where a change
which provides higher fracture toughness and to induction hardening resulted in keeping dis-
higher impact toughness. tortion within acceptable limits.
Steel hardenability is determined by its Increase Stock Allowance. In many in-
chemistry. The quench conditions required to stances, allowance for stock removal in the
obtain the desired properties are a function finishing operation (usually grinding) is the most
of the hardenability. Therefore, if the steel economical approach. Under these conditions,
chemistry is incorrect, the selected quench pro- some study is usually necessary to determine the
cess conditions may, if too severe, lead to magnitude of distortion caused by heat treating
cracking. Unfortunately, this problem is not and thereby how much stock allowance is
uncommon. required for cleanup. Frequently, it is necessary
to take reasonable precautions (perhaps some
special procedures) in heat treating and then
Techniques for Controlling Distortion take further steps by increasing stock removal
in finishing.
In applying one or more of the effective Mechanical straightening, either during
methods of minimizing distortion, cost is usually processing or applied to heat treated parts, offers
the major consideration. Therefore, in planning a third approach for solving distortion problems.
manufacturing operations, it behooves the pru- Straightening is sometimes used as the sole
dent processor to evaluate the costs of mini- technique for correcting distortion, but more
mizing distortion against the alternatives (Ref often it is used in conjunction with systematic
14). In almost any instance, there are at least stock removal.
three alternatives: Heat Treating Practices for Minimizing
Change to another heat treating process Distortion. Positioning in the furnace may
Make allowances for stock removal in fin- have a marked influence on total distortion,
ishing operations to correct the distortion especially for parts having a relatively large
Incorporate straightening operations as re- length-to-cross section ratio. For example, for
quired long, shaftlike parts (solid or tubular), the
poorest loading technique would be to pile them Fog or mists

horizontally and more or less indiscriminately Air
on the furnace hearth. Under these conditions,
distortion begins immediately and continues as The higher rate of heat extraction (quenching
the parts heat up and lose strength. Parts at the power) is obtained by agitated brine. Minimum
bottom of the pile will naturally distort the most, distortion would be obtained by vertical heating,
because they are subject to the greatest stress then cooling (quenching) by hanging in still air.
during heating. Of course, this technique would not usually be
In contrast to the poor technique described practical, largely because the parts would have
previously, the best technique for such parts to be made from air-hardening steels. There are
would be to hang them (preferably with spaces several factors that influence choice of quench-
between each part) in a vertical furnace for ing medium, but hardenability of the steel is
heating. As a rule, some further improvement usually the key factor. Cooling rate thus has a
can be achieved by heating in molten salt as marked effect on the amount of distortion.
opposed to a gaseous atmosphere. This is due to Consequently, the quenching speed should
the fact that some support is supplied by the never be faster than is required to attain the
buoyancy effect of molten salt. required critical cooling rate, when distortion is
One possible disadvantage (relating to dis- an important consideration.
tortion) of heating in molten salt is the heating Special quenching techniques may be
rate. Parts are heated four or five times as fast in needed. Vertical heating and quenching one part
a medium of molten salt compared to heating in at a time could be construed as a special tech-
a gaseous atmosphere. Rapid heating sometimes nique. However, commonly recognized special
increases distortion, especially when various techniques include:
section thicknesses are involved. Martempering
Position during quenching may also have a Press quenching
marked effect on the total amount of distortion. Cold-die quenching
Parts that are hung vertically in the furnace to
minimize distortion should likewise be hung All of these methods can be used effectively to
vertically in the quenching tank; that is, deep reduce distortion, but as a rule, they are all
tanks are preferred for this type of work. In relatively costly and will greatly increase the
most instances, minimum distortion of specific total manufacturing cost.
workpieces is achieved when the quenching Whenever the handling of parts individually
medium is not agitated. To obtain full hardness, is involved, the cost of heat treating increases
however, agitation usually must be used. When rapidly. For example, tubes made from 52100
minimum distortion is required, if agitation is steel, 1.2 m (4 ft) long, 63.5 mm (2.5 in.) in
used, the quenching medium should be agitated diameter with a 3.175 mm (0.12 in.) wall thick-
with agitating force at the bottom of the tank. ness were required to be hardened to a minimum
Although this specific system is used for water, of 60 HRC with a maximum total indicator
the principle applies to any quenching medium. reading of 0.75 mm (0.030 in.). No salt bath of
The quenching medium should never be agitated sufficient depth was available for quenching;
from the side in such a system when minimum thus, martempering could not be considered.
distortion is important. Several different procedures were tried, but the
Choice of quenching medium can affect one that was ultimately successful consisted of
distortion. Typical quenching media listed in heating vertically (hanging) and quenching in
approximate order of decreasing cooling power, unagitated oil, one tube at a time. This procedure
are as follows: cost more than six times the estimated cost for
bundling the tubes, then mass heating and
Water quenching. Therefore, the cost of such a practice
Brine solutions (aqueous) would normally be considered prohibitively
Caustic solutions (aqueous) expensive. Straightening during processing
Polymer solutions may have solved the problem, but it also is very
Oils expensive.
Molten salts Martempering can, under many conditions,
Molten metals be used to effectively reduce distortion and
Gases, including still or moving can still be applied to mass production. This
depends largely on workpiece shape and size. metal dusting; sulfidation; nitridation; dis-
However, for long, shaftlike or otherwise bonding of chromium-molybdenum steels
unwieldy workpieces, martempering requires overlayed by austenitic stainless steels)
handling each workpiece individually and is
thus a relatively expensive operation, particu- Influence of design, steel grade, and condition
larly when simultaneous straightening is incor- are illustrated in the following examples.
porated. Example 1, in Fig. 16 (Ref 16), shows two
Press quenching is probably the oldest special AISI W1 carbon steel concrete roughers that
quenching technique and is still used to a con- failed after a few minutes of service. Cracking
siderable extent. The greatest use of press occurred at the change in section size due
quenching is for gears that cannot be heat treated to bending stresses. Although the section
with sufficient dimensional accuracy by mass change has a smooth, filleted surface, it is still
quenching in baskets. a very effective stress concentrator. Sub-
Selection of press quenching should be done sequent design changes involved a tapered
with the full knowledge that it is very expensive. change in section at the cracked location and
First, the presses are expensive machine tools. later at the start of the wrench above the
Second, the dies are expensive, and the dies must cracked region.
be tailored to the specific workpiece. Also, press Example 2 illustrates that holes placed too
quenching is slow, tedious, and thus expensive. close to the edges of components are a common
Minimum distortion is achieved by press source of failure during heat treatment or in
quenching, but heat treating cost is high. service. Figures 17(a) and (b) (Ref 16) show an
Dry die quenching is another tedious and AISI O1 tool steel die that cracked during oil
expensive process and should be considered quenching. The die face contained numerous
only for highly specialized applications. 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
Examples of Failures due the 205 C (400 F) temper used in this case, is
to Heat Treatment apparent. This indicates the depth of the crack
produced during quenching that was open dur-
Different types of errors and failures produced ing tempering. Coarse machining marks and
in heat treatment (Ref 15) include: deep stamp marks were also present. Sharp,
unfilleted corners may also promote quench
Heating errors: heating too fast causing cracking.
stresses in outer zones; heating non- Example 3, Fig. 18 (Ref 16) shows a 76 by 87
uniformly or locally overheating; heating at by 64 mm (3 by 37/16 by 2.5 in.) AISI O1 tool
too high a temperature or for too long a time steel die that cracked during oil quenching. The
(distortion, cracking, residual stresses, de- cracking pattern (emphasized using magnetic
carburization, alloy depletion) particles) that emanates from the sharp corners is
Temperature errors: overheating (scaling, visible. A few cracks are also associated with the
burning, internal oxidation, hot shortness, holes that are rather close to the edges. Temper
grain coarsening, aging, phases precipi- color was observed on the crack surfaces, indi-
tation) cating that the cracks were present before tem-
Heat treating errors: improper thermal pering.
cycle; too high a temperature; too low a Example 4, in Fig. 19 (Ref 16), shows an-
temperature; improper heating rate; impro- other example of a quench crack initiated by a
per cooling rate; improper soaking from sharp corner. This fixture was also made of AISI
timing errors or nonuniform quenchant; O1 tool steel that was oil quenched. In this case,
improper atmosphere control, which is cri- the corner was filleted, but there was a nick in the
tical in carburizing and nitriding; delay corner where cracking began. The shape of this
between quench and temper; improper aging fixture is also poor for steel that must be oil
treatment or postweld heat treatment (im- quenched. The thinner outer regions cool more
proper, unacceptable, or mixed structures and rapidly, forming martensite first, while the more
microstructural features; temper embrittle- massive central region cools at a slower rate. An
ment; sigma-phase embrittlement; 475 C air-hardenable steel would be a better choice for
embrittlement; sensitization; carburization; this part.
Example 5 provides another case of a poor Example 6, in Fig. 21 (Ref 16), is a punch
design for liquid quenching, as shown in Fig. 20 made of AISI S7 tool steel that cracked during
(Ref 16). This 76 mm diameter by 76 mm long quenching because of rough machining marks (a
(3 in. by 3 in.) threaded part made of AISI W2 common cause of quench cracking). Because of
carbon tool steel cracked in half at an undercut at the section size, the punch was oil quenched to
the base of the threads. Figure 20 shows the two 540 C (1005 F), then air cooled. The cracking
broken halves along with a cold-etched disk pattern has been emphasized with magnetic
taken from the hollow portion of the part. The particles. Temper color was observed on the
hardened outer case can be seen in the fracture crack walls.
detail and in the cold-etched disk. Similar Example 7, Fig. 22(a,b) (Ref 16) show a
parts, without the undercut, were successfully classic example of a failure due to improper
hardened.
Fig. 17 (a) Front view of an AISI O1 tool steel die that

cracked during oil quenching. The die face contains
Fig. 16 AISI W1 (0.85% C) tool steel concrete roughers that holes that are close to the edge for safe quenching. (b) Side view of
failed after short service (2 min for S, 7 min for S11). broken die halves showing the mating fracture surfaces and
Failures occur at the change of sections. Source: Ref 16 temper color (arrow) on the crack surfaces. Source: Ref 16
electrical discharge machining (EDM) techni-

que. 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 tem-
pered 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.
Fig. 18 AISI O1 tool steel that cracked during oil quenching. An as-cast region was present at the extreme
Note the cracks emanating from the sharp corners.
The four holes, which are close to the edge, also contribute to
edge (approximately 35.5 HRC). Beneath this
cracking. Source: Ref 16 layer was a region of as-quenched martensite
(approximately 63.5 HRC). Next was a back-
tempered 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 pre-
sent. 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 as-
quenched martensite surface layer (thin, white
Fig. 19 Fixture made from AISI O1 tool that cracked during
layer), while the internal structure was grossly
oil quenching. The design is poor for liquid
quenching. Source: Ref 16
(a) (b)
Fig. 20 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
overaustenitized (note the retained austenite, (SEM) micrograph showing intergranular frac-
white, and coarse plate martensite). Both factors ture at a crack origin. Figure 26(c) is an SEM
led to cracking. If the EDM surface layer was not micrograph illustrating the brittle mode of
present, poor service life would have resulted failure associated with the fracture. Figure 26(d)
anyway due to poor microstructural condition. is a micrograph showing the typical con-
Example 8, in Fig. 24(a,b) (Ref 16), shows a centrations of nonmetallic stringers in the tube
41 mm (1.6 in.) square 1.4 kg (3 lb) AISI S5 material, and Fig. 26(e) is a micrograph show-
tool steel sledgehammer head that cracked during a quench crack with a heavy oxide. Although
ing quenching. A disk cut from the head was the steel met the compositional requirements
macroetched, revealing a heavily decarburized of SAE 4140, the sulfur level was 0.022%
surface (Fig. 24b). Such a condition promotes and would account for the formation of the sul-
quench cracking, particularly in liquid-quench- fide stringers observed. The combination of the
ing grades such as S5 (oil quenched), due to clustered, stringer-type inclusions and the
differential surface stresses. A deep stamp mark quenching conditions was too severe for this
also helped promote cracking. component geometry. The result was a high
Example 9 presents a situation with stamp incidence of quench cracks that rendered the
marks, such as that shown in Fig. 25 (Ref 16) parts useless.
that commonly promote quench cracks. This Example 11 presents the case of six wrist
was present on an air-quenched die made from pins designed especially for a high-performance
AISI S7 tool steel. In this case, the die was six-cylinder automotive engine (Ref 18) that
not tempered, another prime cause of quench failed after 4800 km (3000 mi) of normal
cracking. operation. The wrist pins were made of low-
Example 10, Fig. 26 (Ref 17) shows SAE carbon steel carburized on both the outer and
4140 grade steel seamless tubing that inner diameters. Two failed wrist pins were
failed because of quench cracks. During pro- submitted for examination. Sample 1 had frac-
duction of hydraulic cylinder housings being tured into three pieces (Fig. 27). Sample 2 had
fabricated from this steel seamless tubing, not fractured but exhibited circumferential
magnetic particle inspection indicated the pre- cracks on the surface of the central zone. Some
sence of circumferential and longitudinal cracks of the cracks had progressed for most of the 360
in a large number of cylinders. Figure 26(a) is a of the pin. Both samples showed some evidence
cross section of the tube showing extensive of scoring on the outer diameter.
cracking revealed by dye-penetrant inspection. The fractured faces of sample 1 were battered
Figure 26(b) is a scanning electron microscope 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 dimen-
sions of a pin. The machining on the inside
diameter surface (indicated by B) was rela-
tively 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 micro-
structure 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 cemen-
tite. 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
Fig. 21 Punch made of AISI S7 tool steel that cracked during
value of 50 HRC was measured. Results indi-
quenching because of rough machining marks
(a common cause of quench cracking). Source: Ref 16 cated a case depth of 0.89 to 1.0 mm (0.035
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
to 0.040 in.). Determination of case depth by diameter, 111 mm (4.375 in.) grip length, and
visual examination on a microspecimen was 134.5 mm (5.294 in.) overall length. It was heat
0.89 mm (0.035 in.). treated to a tensile strength of 1240 to 1380 MPa
The principal causes of failure were inade- (180 to 200 ksi) and a hardness of 39 to 43 HRC
quate heat treatment of the case and a design that and then cadmium plated per QQ-P-416 type II
incorporated a raised central section of the inner class 2 (23 h bake).
diameter, which acted as a stress raiser. Rough The bolt fractured through the head-to-shank
machining of the inner diameter aggravated the fillet, a type of failure usually traceable to a
situation. The case with the cementite grain- poorly controlled manufacturing process, such
boundary network had not been heated to a high as heat treating (quench cracking) or chemical
enough temperature or long enough to take the plating (hydrogen embrittlement). In this in-
cementite into solution in the austenite. It was stance, delayed cracking caused by hydrogen
suspected that after slow cooling from the car- embrittlement was initially suspected, because
burizing temperature, pins were heated slightly the bolt reportedly had passed a magnetic par-
above the Ae1 prior to quenching and then given ticle inspection.
a low temper. The case was refined, and the core The fracture surface (Fig. 32), has two distinct
was unrefined. zones. Zone 1 was covered with a thick layer of
Thus, poor heat treatment, resulting in a brittle baked-on scale. The scale was removed and the
grain-boundary network of cementite, and a area examined using an SEM. The fracture
design that formed locations of stress con- topography shown in Fig. 33 is a combination of
centration in the inner diameter were the most tearing and intergranular hairline cracks, fea-
probable causes of failure. tures often associated with both hydrogen
The pins should be carburized to a double heat embrittlement and quench cracking. However,
treatment to refine both case and core and to the heavy adherent nature of the scale suggests
eliminate the brittle grain-boundary network of that it formed on the crack surface at high tem-
cementite. The pin design should be changed to perature, that is, during heat treating but before
eliminate the central raised section of the inner conventional quenching and subsequent cad-
diameter to avoid the fillets acting as stress rais- mium plating. Zone 2 is characterized by
ers. The machining of the inner diameter should equiaxed dimples, a common feature of ductile
be improved to avoid a rough surface. The depth tension overload (final stage of fracture). The
of the carburized case should be reduced to bolt head was cut through its centerline, and
approximately 0.38 mm (0.015 in.) to increase the crack cross section was metallographically
pin toughness. examined in a zone 1 area to further explore
Example 12 features cracking of an alloy whether failure was caused by hydrogen em-
steel bolt. A heat treated, cadmium-plated AISI brittlement (step cracking, no decarburization)
8740 steel bolt broke through the head-to- or quench cracking (temper scale). The crack
shank fillet while being handled during assem- had both heavy decarburization and temper scale
bly (Ref 19). Dimensions of the alloy steel bolt (Fig. 34), ruling out hydrogen embrittlement
(MSD 21250-10070) were 15.9 mm (0.625 in.) cracking. However, because quench cracking
(a) (b) (c)
Fig. 23 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
occurs at relatively low temperatures, de-

carburization 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, dimen-
sionally inspected, and returned to the produc-
tion lot instead of being scrapped. The heavy
decarburization layer on the crack surface sup-
ports this scenario. A schematic of the quench
crack formation is shown in Fig. 35.
Example 13 involves hydrogen embrittle-
ment 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 sig-
nificant 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 sec-
tioned 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
Fig. 25 Quench crack promoted by the presence of a deep,
sharp stamp mark in a die made of AISI S7 tool steel.
mark (arrows). Actual size. Source: Ref 16 Source: Ref 16
(a) (b) (c)
(d) (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 mag-
nification: 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 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 sur-
Fig. 27 Failed wrist pin (sample 1), showing fractured faces. faces displayed similar fracture modes. Figure 37
Source: Ref 18 shows the overall fracture surface. Figure 38
shows a ductile fracture mode, which was To determine whether the cap screws were
observed over the majority of the fracture sur- plated, a fracture surface and the outside surface
faces. Figure 39 shows an intergranular fracture of the unthreaded part of the shank were cleaned
mode, which was observed around the circum- of oil and other deposits and analyzed using
ference of the screw, next to the plated surface. energy-dispersive x-ray spectroscopy (EDS) in
The part of the shank containing the fracture conjunction with SEM examination. The EDS
surfaces was metallurgically prepared in cross
section to look for secondary cracking and
pitting. No secondary cracking or pitting was
observed.
Fig. 31 Macrograph of sample 2 taken along the longitudinal

axis, showing cracks emanating from both the inner
and outer diameters. Unetched. Original magnification: 15.
Fig. 29 Central longitudinal zone of sample 2, showing Source: Ref 18
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 speci-

men 2. The dark matrix is tempered martensite; the Fig. 32 Close-up view of the bolt-shank fracture surface.
light-colored grain-boundary network is cementite. Nital etch. Note the heavy scale on the zone 1 surface. Source:
Original magnification: 200 . Source: Ref 18 Ref 19
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. 37 Fracture surface of the crack in the failed screw after

Fig. 36 As-received socket head cap screws. Arrow indicates the crack was opened in the laboratory. L indicates
a secondary crack in the screw thread root. Source: the laboratory-induced overload region. Original magnification:
Ref 20 20. Source: Ref 20
Fig. 38 SEM micrograph showing a ductile fracture mode Fig. 39 SEM micrograph showing an intergranular fracture
mode, observed around the entire circumference at
observed over the majority of both fracture surfaces. both fractures in the screw. Structure at top is the base metal;
Original magnification: 1000 . Source: Ref 20 structure at bottom is cadmium plating. Original magnification:
1000 . Source: Ref 20
80.5 HRN, which converts to approximately

LT= 100 SECS
BASE METAL OF SCREW 40 HRC.
20K
The presence of a ductile fracture mode at the
Iron
core and an intergranular fracture mode at the

15K outer surface of a plated bolt is typical of
hydrogen embrittlement but could also be stress-
COUNTS
10K
corrosion cracking (SCC). However, SCC can
be eliminated, because the metallographic
results showed no evidence of secondary cracks
Manganese
5000 or other corrosion mechanisms, such as pitting.

Iron
Many hydrogen embrittlement mechanisms

0 have been proposed, but none is universally
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000 accepted. However, the phenomenon of hydro-
(a) ENERGY keV gen embrittlement is widely known. The pre-
LT= 100 SECS sence of hydrogen in steel reduces the ductility
15K PLATING ON THE SCREWS SURFACE of the steel and causes premature failure under a
Cadmium
static load. The time for failure depends on

the stress applied to the component and the
Cadmium
10K
amount of hydrogen that has diffused into the
COUNTS
steel. A component may fail initially when put

under load or may fail several weeks after being
loaded. Because of this characteristic, hydrogen
Cadmium
5000
Cadmium
embrittlement is sometimes called hydrogen-

Cadmium
induced delayed failure.

Electroplating is a common cause of hydro-
0 gen embrittlement in bolts and screws, because
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 hydrogen is evolved (or liberated) during the
process. The screws had been plated, and be-
Fig. 40 Energy-dispersive x-ray spectroscopy spectra of (a) cause no other source of hydrogen was iden-
the base metal of the screw and (b) the plating on the tified, it is likely that the plating process was the
outside surface. Source: Ref 20
source of the hydrogen.
After most plating processes, bolts and screws
results are shown in Fig. 40. It was determined are usually baked for several hours at 190 C
that the screws were plated with cadmium. (375 F) to diffuse any hydrogen out of the steel.
Superficial Rockwell hardness measurements However, when a bolt or screw is cadmium
were taken on the metallographic section. The plated, it requires a much longer time (approxi-
average hardness for the failed cap screw was mately 24 h) for baking, because hydrogen
diffuses less readily through cadmium than other systematic component design process. The
electrodeposited metals. In the case of the cap typical phases of component design include
screws, the screws either were not baked at all or planning and requirements definition, concept
were not baked for a sufficient period of time design, detail design, and test and validation.
and/or at a high enough temperature. The component is designed to provide a specific
The cap screws failed because of hydrogen mechanical, thermal, and chemical function
embrittlement. The most probable root cause throughout its life cycle and is often limited by
was absence of baking or insufficient baking space, cost, and safety considerations. The
of the cap screws after the cadmium plating selection of materials and manufacturing pro-
process. cesses for a cost-effective component design is a
To eliminate the possibility of future hydro- complex process and often involves iterative
gen embrittlement failures, the screws should decision making. The iterative nature of design
be baked at approximately 190 C (375 F) requires a continuous analysis and redesign
for 24 h. process.
Process design employs stress-analysis tools
with stress-concentration factors, design rules
Heat Treatment Design based on experimental data, material property
databases, and mechanical properties resulting
As was analyzed in previous sections, heat from a broad range of heat treatment processes.
treatments are a series of operations in the course Computer modeling is a valuable design tool for
of which a solid ferrous product is totally or heat treated components. A computer process
partially exposed to thermal cycles to achieve simulation model allows a particular design to
the desired change in structures and properties be tested under a specific set of process condi-
(Ref 2). The chemical composition of the tions. The computer software can graphically
material may possibly be modified during these display not only the resulting residual stresses
operations (thermochemical treatment). and distortions in the component but also the
In addition to the ability of the heat treatment associated transient evolution of temperature,
to achieve desired mechanical properties, heat metallurgical phases, volume changes, and
treatment also produces dimensional changes stresses.
and residual-stress patterns that, in some cases, Regardless of the procedure used for devel-
can lead to component cracking and distortions. oping the design, at the end of the design pro-
In the following section, a procedure is ana- cess, the required quality of the product should
lyzed to improve the performance of the design be defined and then described in a technical
process. A typical component manufacturing document, definitive layout, or final design
process includes metalworking, machining, or intended to accomplish the product manu-
other forming operations, followed by heat facturing. Only the minimum quality needed
treatment. for the product to perform the function intended
Different types of heat treatment of steels should be specified. Overspecifying and in-
are usually employed by industry: hardening, cluding restrictive features in the quality des-
austenitization, annealing, normalizing, stress cription can cause delays and increase costs to
relieving, quenching and tempering, and aus- the buyer.
tempering. The design process requires developing
Heat treatment processes include component operational definitions. Operational definitions
heatup, holding at temperature for through-heat are, for example, a specific test of a piece of
solutionizing, or thermochemical treatments material or criterion for judgment. Without
such as carburizing or nitriding, quenching from operational definitions, a specification is mean-
elevated temperature, postquench tempering, or ingless. A specification for heat treatment con-
aging treatment. All the steps can influence taining the clause Avoid long, thin sections
dimensional changes, residual-stress patterns, requires operational meaning of long, thin
and cracking in heat treated components. section. This definition depends greatly on the
quenching media; however, any section length
greater than 15 times the diameter is almost
The Process of Component Design always characterized as such, and the slightest
In order to avoid failures associated with nonuniformity in quench will cause it to distort.
heat treatments, it is necessary to develop a As the quenching medium becomes more severe
(i.e., water, caustic quench), this criterion is temperatures. If the material is heat treated
reduced to as low as 5 times the diameter. For during fabrication, it will have adequate
larger length-to-diameter ratios, consideration toughness in its final condition. Design must be
should be given to fixture quenching or induc- reviewed in order to minimize the presence of
tion hardening. notches and defects that concentrate stresses.
The component design process can be When mechanical loads are an alternative,
divided into two phases: there may be a risk of fatigue failures.
In fatigue failures, a crack grows in each
Phase 1, which corresponds to the basic
loading cycle until the remaining ligament fails
definition of the product (including concept
due to ductile or brittle fracture. This phe-
and detail design)
nomenon can occur at stress levels lower than
Phase 2, which corresponds to the design
the allowable stresses for static loads. It should
review aiming to prevent failures and mini-
be emphasized that fatigue failures strongly
mize risks
depend on design and manufacturing quality,
Phase 1 deals with the basic definition of the which is accomplished by increasing fatigue
product. As a first step, the generic type of the strength and minimizing stress concentrators.
material as well as its geometric configuration Technical requirements should be complied
should be selected. The criterion applied to with at the lowest cost. A phase 1 basic design
avoid plastic deformation states that the cal- detailed analysis is beyond the scope of this
culated effective stress must be lower than the work.
yield and design life creep-rupture stresses of Phase 2 design review has the purpose of
the material. If the product to be designed is an assuring that the basic design fulfills the
element that must withstand not only tension requirements and reviews the design to avoid
loads but also bending, torsional, and axial- failures. This step verifies that the basic types of
compressive loads, then the combined effect of failure modes have been properly controlled by
the applied load type, shape, and size and the design and determines the types of damages
material properties should be analyzed. Design associated with each failure mode in order to
for those elements subjected to axial compres- implement methods for detection.
sion should be intended to avoid not only Material behavior can be analyzed by
plastic collapse but also elastic instabilities. developing models that relate materials attri-
Elastic instabilities may cause Euler buckling butes, required functions, and manufacturing
and local buckling. The occurrence of this processes. Due to the large number of aspects
failure mode depends on the geometry of the involved, the problem can be simplified by
elements and on the Youngs modulus of the considering blocks of knowledge that corre-
material. In general, the relationship between spond to specific mechanisms and functions.
the Youngs modulus (E) and the density (r), Each block of knowledge represents a simplified
E/r, should be maximized in order to increase model that relates some properties to the
stiffness. By doing so, yield occurs before required functions, through the knowledge pro-
buckling, whereas by increasing specific vided by materials science and engineering. The
strength (sf/r), strength also increases; hence, use of state-of-the-art criteria, which, in some
buckling occurs before yield. During the basic cases, are based on practical experience, can
design step, brittle fracture should be avoided. optimize the accomplishment of the analyzed
Brittle fracture is associated with very little or functions. The results of the analysis are syn-
no plastic deformation. A material may fail in a thesized in the definition of design.
catastrophicbrittlemanner under stresses Consider two groups of behavior models.
even lower than the allowable design stresses The first group relates materials attributes
used to avoid ductile failures. The material generally known as propertiesthat are well
property that controls brittle fracture strength is defined and individually determined. These
toughness. Other factors that have an effect on properties are component shape and size,
brittle fracture are material thickness; local modulus of elasticity, Poissons ratio, ultimate
stress level, including nominal stresses, resi- tensile strength, yield strength in tension, shear
dual stresses, and stress-concentration factors; strength, compressive strength, ductility, elon-
temperature; and loading rate. Carbon and low- gation, fracture toughness, hardness, thermal
alloy steels undergo a transition from ductile conductivity, thermal diffusivity, thermal
failure mode to brittle failure mode at low expansion, specific heat, density, fracture
toughness, creep strength, creep-rupture Heat Transfer Model

strength, and so on. For these properties, test Heat treatments are a series of operations in
and criteria have accurate operational defini- the course of which a solid ferrous product is
tions. exposed to thermal cycles. There are different
The second group of behavior models in- types of heat transfer: conduction, convection,
cludes attributes that involve the complex and radiation. During heat treatments, the tem-
interrelation of a number of variables associated perature varies in time as well as in space; these
with materials, manufacturing (heat treatment) processes are called unsteady, nonstationary, or
processes, and service conditions. For this transient. The factor of proportionality thermal
second group of blocks of knowledge, phy- diffusivity, a, defines the rate of change of
sical metallurgy is intensely used, together temperature, and the heat-transfer coefficient, h,
with the laws of mechanics and empiric knowl- controls the heat flow through the component
edge. surface. The thermophysical properties, l, r,
and Cp, vary with the temperature. The geo-
metric design will try to avoid mass asymmet-
Modeling of Heat Treatment ries. Large differences in section size and
distribution of material causes differential
Modeling of heat treatment processes, like heating and cooling during heat treating. The
other materials processes such as casting and thermal cycle induces phase transformation and
welding, is quite complex due to the tight thermomechanical stresses.
coupling of various metallurgical transfor-
mations and the associated changes in thermal
Solid-Phase Transformation Model
and mechanical states. A heat-transfer model,
coupling with a phase transformation model, a The thermal cycle induces solid-phase
thermomechanical model, and a thermo- transformation. Microstructures that are formed
chemical model (Fig. 41), is considered. upon cooling and the proportions of each are
Fig. 41 Behavior model to analyze heat treatment design

dependent on austenitization time, temperature, gradient through the thermal conductivity, l.

cooling history of the particular alloy, and the Thermal expansion has a strong influence on the
composition of the alloy. The transformation development of residual stress; whenever the
products from austenite decomposition change thermal expansion or contraction of a body is
from a mixture of ferrite and pearlite to bainite prevented, thermal stresses appear. If large
or martensite with increasing cooling rate. enough, they cause yielding, fracture, or elastic
Isothermal transformation (time-temperature collapse (buckling).
transformation, or TTT) and continuous cooling The total induced strain is the result of the
transformation (CCT) diagrams can be used to sum of the strain produced by thermal (aE
predict steel microstructures and hardness after DT=eth) and transformation (etr) strain due to
heat treatment. These diagrams are a set of local transformation-induced volume expan-
curves drawn in a semilogarithmic coordinate sion. The total induced strain must be accom-
system with logarithmic time/temperature modated through either elastic (eel) or plastic
coordinates that define, in the case of the TTT (eep) strain, which sums to the total strain:
diagram, for each level of temperature, the
beginning and end of the transformation of et =eth +etr = (eel +eep ) (Eq 11)
austenite under isothermal conditions, and, in
the case of the CCT diagram, define each var- Thermal Shock. The ability of a material to
iation in temperature as a function of time during resist thermal shock, due to a sudden immersion
cooling, the temperature at which the austenite in a cold ambient, without cracking depends on
begins and ends its transformation. Industrial its thermal expansion coefficient, a; the tensile
heat treatments consist of heating and soaking strength, st, for metals; the Youngs modulus, E;
at a suitable temperature, followed by cooling the thermal conductivity, l; and the heat-transfer
at an appropriate rate in order to obtain a struc- coefficient, h. A temperature change of DT
tural state closer to that equilibrium (annealing), applied to a constrained body or a sudden
reduce the internal stresses without substantially change, DT, of the surface temperature of the
modifying the structure (stress relieving), unconstrained component induces a thermal
increase hardness by more or less complete strain.
transformation of austenite to martensite and Stress Concentration. The elastic strain and
possibly bainite (quench harden), or obtain a stresses calculated previously are nominal
uniform and fine-grained structure with pearlite values that do not take into account local dis-
(normalizing). Other heat treatments are applied continuities such as holes, notches, or section
to ferrous products after quench hardening changes. Even on a structure where stress
(tempering) or solution treatment (aging) to intensity has been limited by yield criteria, there
bring the properties to the required level. The may exist highly localized regions where
process conditions that shall be taken into peak stresses are several times higher than
account usually include the furnace atmosphere yield. Maximum local stresses on a structure can
(for example, temperature and carbon potential), be determined considering nominal stresses
heating rates, and quench conditions. multiplied by a stress-concentration factor and
can be estimated through a detailed stress ana-
lysis or by using approximate formulas that
Thermomechanical Modeling account for the most common cases. Design
Primary stress is a normal or shear stress should be verified to confirm whether there are
developed by the imposed loading that is stress-concentration points that may activate
necessary to satisfy the laws of equilibrium of failure mechanisms due to brittle fracture, cor-
external and internal forces and moments. Sec- rosion, or fatigue. A sharp corner concentrates
ondary stress is a normal or shear stress devel- and magnifies applied stresses, stresses that arise
oped by the constraint of adjacent parts or by in tool and die manufacturing (such as during
self-constraint of a structure. The basic char- quenching), or stresses that occur during service.
acteristic of the secondary stress is that it is self- In addition to promoting cracking during liquid
limiting. An example of secondary stress is a quenching, sharp corners promote buildup of
general thermal stress. residual stresses that may not be fully relieved
Thermal stresses are related to temperature by tempering and can therefore reduce service
by the thermal expansion coefficient. The life. The largest possible fillet should be used
deformation is proportional to the thermal at all sharp corners. Air-quenching grades of
steel are more tolerant of sharp corners than Quench Cracking. If thermally induced
liquid-quenching grades. Fatigue fractures pro- stress exceeds the local tensile strength of the
pagate from the sharp edges. material, cracking results.
Residual stresses can be defined as those
stresses that remain in a material or body after Thermochemical Model
being manufactured and processed in the
absence of external forces or thermal gradients. Thermochemical treatments may be applied
After the cooling processes have been com- to a ferrous product in the austenite state to
pleted, the residual-stress distribution between obtain a surface enrichment in carbon (carbur-
the surface and core is obtained. If the surface izing), which is in solid solution in the austenite.
stresses exceed the hot yield strength of the Carburizing can be done in gas atmosphere,
material, it plastically deforms, resulting in in solid medium, or in a bath of molten salt.
thermally induced dimensional changes and Other thermochemical treatments can be applied
thermal and transformational stresses. In heat to produce surface enrichment in nitrogen
treated parts, residual stresses may be classified (nitriding); in nitrogen and carbon (nitro-
as those caused by a thermal gradient alone and a carburizing); in sulfur, carbon, and nitrogen
thermal gradient in combination with a micro- (sulfidizing); in silicon (siliconizing); in chro-
structural change (phase transformation). mium (chromizing); in boron (boriding); and in
Distortion is any change in the shape and aluminum (aluminizing). By error in the heat
original dimensions of a ferrous product occur- treatment, the surface can be decarburized.
ring during heat treatment. For a given geometry Table 3 shows guidelines to avoid heat treat-
and heat flow, the distortion is minimized by ment failures during the design review (phase 2
selecting materials with large values of l/a. design review); the recommendations come from
Shape and volume changes during heating and the examples presented in the previous sections.
cooling can be attributed to three fundamental
causes:
Failure Aspects of Welded Components
Residual stresses that can cause shape
change when they exceed material yield Brittle Fracture. Low-temperature/low-
strength toughness fracture is sudden failure of a struc-
Stresses caused by differential expansion tural component that is usually initiated at a
due to thermal gradients crack or defect. This is an unusual occurrence,
Volume changes due to transformational because design stresses are normally sufficiently
phase change low to prevent such an occurrence. However,
Table 3 Phase 2 design review to avoid heat treatment failures

Characteristic of the material Always use the grade or composition of steel most suitable for the work that the part has to perform.
prior to heat treatment The carbon content of the steel is one of the determining factors for quenchant selection; plain carbon steels
operation 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 Avoid asymmetric design.
dimensions 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.
some older equipment with thick walls, equip- Residual stresses and postweld heat treat-
ment that may be subjected to low temperature ment
due to an upset, or equipment that may be Thickness
modified could be susceptible to varying degrees
of embrittlement. Temper embrittlement is one of the main
Brittle fracture is associated with very little or causes of toughness degradation in ferritic steels
no plastic deformation and with a cleavage during high-temperature service. This degrada-
fracture or intergranular surface, unlike ductile tion may lead to component failure during ser-
fracture, which is associated with a fibrous sur- vice. The problem arises when some types of
face. A material may fail in an unstable and steels are exposed to temperatures between 345
catastrophic brittle manner under stresses even and 565 C (650 and 1050 F). Typically, 21/4
lower than the allowable design stresses used to Cr-1=2 Mo steels with a bainitic structure are
avoid ductile failures. This may occur with a the most susceptible to this phenomenon.
combination of material properties and applied Temper embrittlement can also occur in
stress levels. C-1=2 Mo, 1Cr-1=2 Mo, 11=4 Cr-1=2 Mo, 3Cr-1Mo, and
The material property that controls brittle 5Cr-1=2 Mo steels. Conversely, 9Cr-1Mo steels
fracture strength is toughness. Other factors that are less susceptible. Welded joints (weld metal
have an effect on brittle fracture are material and heat-affected zone) are the most susceptible
thickness; local stress level, including nominal zones.
stresses, residual stresses, and stress-concentra- In all cases, the solution to the problem lies in
tion factors; temperature; and loading rate. alloy purity. Exposures within the critical tem-
Carbon and low-alloy steels undergo a transition perature range may occur during temper or
from ductile failure mode to brittle failure postweld heat treatments or during service, and
mode at low temperatures. Resistance to crack these conditions should be avoided. However,
propagation is measured through fractomechan- many components operate within the critical
ical tests. Crack propagation will occur when the temperature range.
stress intensity at the crack tip, K, reaches a The segregation of residual elements anti-
critical value, Kc (MPa m1/2). Brittle materials mony, arsenic, phosphorus, and tin toward
are those that remain elastic until breaking austenitic grain boundaries is the main cause
(break occurs before yield). of temper embrittlement. Also, manganese and
According to API 581 (Ref 21) low- silicon play an important role in this segre-
temperature/low-toughness fracture of steel is gation, and their content should be limited.
affected by: Both residual and alloy elements can segre-
gate, but the former can be concentrated up to
Applied loads: Fracture is less likely at low 300 times their average value in the material.
applied loads. Segregation only occurs in ferrite within a 315
Materials specification: Some materials are to 540 C (600 to 1005 F) temperature range
manufactured to exhibit good fracture but never occurs during austenitization. In
properties or toughness properties. Materials addition to segregation in grain boundaries, a
are often qualified for use by performing an fine precipitation can occur within the grains,
impact test that measures the energy needed resulting in a strength increase (Mo2C pre-
to break a notched specimen. Fine-grained cipitation).
structures, such as tempered martensite, with The phenomenon associated with changes in
low impurity content are associated with grain boundaries causes intergranular brittle
a high degree of toughness. Other micro- fracture. In general, ductility and rupture
structural elements, such as precipitates, strength are not affected; nevertheless, both
second-phase particles, dislocations, and can be reduced under severe conditions.
solutes in a solid solution, contribute to Toughness is affected by up to a 100 C
increased yield strength but reduce tough- (212 F) shift toward the right of the ductile-
ness. brittle transition curve, as evidenced from
Temperature: Many materials (especially impact testing.
ferritic steels) become brittle at a tempera- When a material is exposed to a 370 to 565 C
ture called the transition temperature. Brittle (700 to 1050 F) temperature range, property
fracture is typically not a concern above degradation may become irreversible; in this
300 C (570 F). case, temper embrittlement and creep operate
simultaneously. During equipment operation r/ti3.0 and tj24.0 mm (0.94 in.)

in a hydrogen service environment, hydrogen r/ti10.0 and all thicknesses
may diffuse into the metal. During cooling
These conditions correspond to an elongation,
from the operating temperature (shutdown), the
e=1/(2r/t+1), in the cold-worked areas of 33,
material becomes oversaturated with hydrogen.
25, 20, 14, and 5%, respectively.
The combination of thermal stresses and
Hardening Tendency. The maximum hard-
hydrogen oversaturation may lead to hydrogen-
ness in the heat-affected zone (HAZ) depends
induced cracking. If the material toughness has
on the chemical composition (carbon equiva-
been reduced considerably due to temper
lent) and the cooling rate during welding (t8/5).
embrittlement, the risk of a catastrophic crack-
ing is high. Carbon most affects hardenability, and its
Material degradation due to temper embrit- effect and that of other elements have been
tlement may result not only in brittle fractures included in carbon equivalent formulas. The
with catastrophic consequences but also in a International Institute of Welding carbon equi-
reduction of the equipment useful life and in a valent formula, recommended for steel with
decrease in the equipment reliability and effi- more than 0.18% C, is:
ciency, since it may be necessary to operate at Mn Cr+Mo+V Cu+Ni
lower temperatures to avoid temper embrittle- CEIIW =C+ + +
6 5 15
ment or to depressurize to avoid stresses when
(in wt%) (Eq12)
the equipment is cold.
Temper embrittlement is reversible. Heat When carbon is 50.18%, it is generally re-
treatment for a short period of time at tempera- commended that the percentage of cementite
tures above 565 C (1050 F), followed by (Pcm) formula be used:
quick cooling, can restore the initial properties.
However, if material thickness is high and Si Mn Cu Ni Cr Mo V
cooling rates required by precipitation kinetics PCM =C+ + + + + + +
30 20 40 60 20 15 10
are not achieved, embrittlement may reoccur. +5B (in wt%) (Eq13)
Moreover, the component may crack during
heat treatment due to the effect of thermal Hardenability of steel is not necessarily an
stresses. indicator of HAZ hardness. It is important to
Material thickness over 25 mm (0.98 in.) is control the maximum hardness in the HAZ in
more susceptible to brittle fracture due to temper order to avoid two main problems: cold
embrittlement. Postweld heat treatments mini- cracking (hydrogen-assisted cold cracking)
mize the susceptibility to brittle fracture through during welding fabrication, in which the limit
this mechanism. of maximum hardness is usually 350 HV; and
Aging Tendency. The tendencies for aging in-service cracking in hydrogen environments,
after cold working increase hardness and tensile where the maximum hardness shall be less
strength, with a simultaneous reduction in duc- than 220 HV. Maximum HAZ hardness can
tility. Therefore, the risk of embrittlement due to be accurately calculated as a function of
welding in cold-worked areas increases. The chemical composition and the cooling time
killed steels show enhanced resistance to aging from 800 to 500 C (1470 to 930 F) (t8/5)
in normalized conditions. When the degree of (Ref 22).
deformation is high, the material should be The HAZ maximum hardness for low-alloy
thermally treated (normalized or stress relieved) steels can be calculated using the following
before the deformed zone is welded. The risk of expressions by Duren (Ref 22). For 100% mar-
embrittlement due to welding in cold-worked tensite (alloying elements in wt%):
areas to be welded, without special require-
ments, is allowable under the following con- HVM =802 C+305 (Eq 14)
ditions. The relationship between the inner For 100% bainite:
bending radius (r) and the plate thickness (t)
shall be: HVB =

Si Mn Cu Cr Ni Mo V
r/ti1.0 and tj4.0 mm (0.16 in.) 350 C+ + + + + + +
11 8 9 5 17 6 3
r/ti1.5 and tj8.0 mm (0.32 in.)
r/ti2.0 and tj12.0 mm (0.47 in.) +101 (Eq15)
For 05%martensite5100: internal stresses, restoring ductility and tough-

h ness, refining grains, reducing gaseous content
HVx =2019 C (170:5 log t8=5 ) in the steel, and improving homogenization of
alloying elements.
Si Mn Cu Cr Ni Mo V Stress-corrosion cracking (SCC) of auste-
+0:3 + + + + + +
11 8 9 5 17 6 3 nitic stainless steels can be caused by chlor-
+66 (170:8 log t8=5 ) (Eq16) ides and polythionic acids, by hydrosulfuric acid
on carbon and low-alloy steels, and by caustic
where the hardness of martensite is HVM, bai- corrosion on carbon steels. Stress-corrosion
nite is HVB, the variable amount of martensite is cracking may arise when a susceptible material
HVx, and t8/5 is the cooling time between 800 is simultaneously combined with certain levels
and 500 C (1470 and 930 F). of tensile stresses and a critical environment
Cold cracking can occur after the welding within a specific temperature range. Tensile
process if four factors are present: local hydro- residual stresses, resulting from manufacturing
gen concentration, susceptible weld metal or processes such as welds, contribute to cause this
HAZ, local metal hardness, and a high level of type of damage. Also, cold plastic deformation
residual stress remaining after welding, which causes hazardous residual stresses. Tensile
may cause cracking at temperatures less than stresses must be reduced by controlling manu-
100 C (212 F). facturing processes and design. Stress relieving
The causes of cold cracking are related to postweld heat treatments can be used to mini-
many factors, including initial weld metal mize susceptibility to SCC in austenitic stainless
hydrogen content, residual hydrogen content at steel, attempting to avoid sensitization by using
100 C, steel carbon equivalent, yield stress of stabilized or low-carbon grades.
steel or weld metal, heat input, preheat tem-
perature, material thickness, joint restraint
intensity, notch concentration factor, welding Heat Treatment Procedures Applied
process thermal efficiency, and others. Cold to Welded Components
cracking is a diffusion-controlled phenomenon
that requires days or weeks at room temperature
to develop cracks. Carbon and Low-Alloy Steels
Hydrogen-induced cracking in the HAZ can Postweld heating (soaking) is an option for
be parallel to the fusion boundary adjacent to a carbon-manganese and low-alloy steels. Mat-
fillet weld or in the form of toe cracks. Weld erial of thick-walled low-alloy steel pressure
metals are by no means immune when the steels vessels is susceptible to hydrogen-induced
possess high yield strength. Thus, cold cracking cracking, and there are many difficulties in
induces surface-connected cracking or subsur- detecting small cracks in the HAZ of heavy-
face cracking, which may provide initiation section steel by conventional nondestructive
points for further cracking by brittle fracture or examination methods. Therefore, it is very
fatigue. important to reduce to a minimum the risk of
Segregation Tendency. Element (phos- hydrogen-induced cold cracking of weldment of
phorus, sulfur, carbon, etc.) segregation impairs thick plate under the heavily restricted condition
weldability. Chemical heterogeneity can con- during the entire welding process. One possible
tribute to localized increases of hardenability. solution is to apply both preheating and post-
Thus, a normal chemical composition of the heat heating (Ref 23). There is a temperature range
of the steel may exhibit hardness higher than the where weld hydrogen-induced cracking may
maximum allowable hardness in certain parts of occur and a temperature above the upper limit
the HAZ, despite a normal chemical composi- where no delay cracking occurs, although the
tion. From the point of view of segregation hydrogen content or the restraint intensity is
behavior, semikilled and killed steels are better high. Therefore, if postheating is carried out
than rimmed steels. If segregation zones are above this critical temperature, the discharge of
involvedas in butt weldingcare should be hydrogen is allowed; then, the material can
taken to limit penetration and hence minimize reach room temperature with a minimum risk of
weld metal dilution. In addition, suitable filler hydrogen-induced cracking.
metal and low-hydrogen basic electrodes should The purpose of soaking is to allow hydrogen
be used. The annealing treatment relieves diffusion to avoid critical values in the
weldments. Welding must be performed, main- heating rates (particularly at temperatures below
taining preheating and interpass temperatures 300 C, or 570 F, with complex components),
that depend on the material, process, thickness, time at soak temperature, soak temperature, and
type of joints, and heat input. The postheating cooling rate to avoid such undesirable events as
must be carried out immediately after welding, flame impingement, distortion, overheating, air
not allowing the temperature to be lower than quenching, and reheat cracking. Very large-
120 C (250 F). The temperature for post- diameter pressure vessels that have transport
heating generally is 300 C (570 F) for 4 to difficulties may be erected on site and PWHTed
6 h, depending on the steel type. Then, the from inside by gas burners, using the vessel as its
weldment is cooled to room temperature. Gen- own furnace (Ref 23).
erally, the same devices recommended for pre- It is well known that the mechanical properties
heating are used. This treatment does not of material are degraded due to stress relieving in
produce either stress relief or microstructural some materials (Ref 23). Generally, in low-alloy
changes. pressure vessel steels, the yield strength, tensile
Postweld heat treatment (stress relieving) strength, and toughness diminish, and elongation
may be used for carbon-manganese and low- and reduction of area increase as the temper
alloy steels. Postweld heat treatment (PWHT) is parameter (TP) is increased:
a uniform heating of a weldment at a tempera-
ture below the critical range to relieve the major TP=T (log t+20) 103 (Eq 17)
part of the residual stresses, followed by uniform where T is temperature (K), and t is time (h). The
cooling in still air. The PWHT is carried out to PWHT is not considered to be a significant
fabricate vessels to increase fracture toughness variable for carbon and carbon-manganese steels
and minimize the levels of residual stress, which up to 50 mm (2 in.) (Ref 23).
confers resistance to brittle fractures, tempers
material structure, and removes the possibility
Postweld Heat Treatment of Stainless Steels
of SCC. The accrued benefits depend on the
material under consideration. Postweld heat treatment is used for stainless
It is usually required by codes and customers steels (Ref 2426). Welding stainless steels in
that manufacturers should heat treat all welds on thick sections, when the thickness exceeds
a thick-wall pressure vessel, including any repair approximately 20 mm (0.8 in.), is a complex
welds. The PWHT is conducted at a temperature operation. In addition to codes or engineering
and for the period specified in the applicable specifications, which may impose definitive
fabrication codes. There are two possibilities: procedures, it is important to have sufficient
one is to stress relieve the completed vessel in metallurgical background to understand what
the furnace as a whole; the other is to stress may happen during welding and the subsequent
relieve the subassemblies separately in the fur- heat treatment operation.
nace and then heat treat the final circumferential The stresses induced by welding often need
seam locally. Techniques for PWHT are some- to be eliminated if dimensional stability of the
what similar to those thermal methods in pre- construction is to be guaranteed and if resistance
heating and soaking. When local treatment is to SCC is mandatory. The properties of welded
carried out, care must be taken to ensure full joints of stainless steels in thick sections that
through-thickness heating, and similarly, the may be modified during a PWHT are principally
temperature gradient is such that the length the corrosion resistance and the mechani-
of material on either side of the weld at a cal properties. The possibility of distortion
temperature exceeding half the treatment tem- occurring during heat treatment must also be
perature is at least 2.5(r t)1/2, where r is the bore considered.
radius, and t is the material thickness. Problems For the purpose of analyzing PWHT in
can arise where butt welds attaching pipes to stainless steels, the materials considered were
nozzles positioned close to nozzle/vessel welds divided into four groups.
require separate heat treatment. In such cases, Chromium Steels. For both ferritic and
it may be necessary to apply individual stress martensitic types, some codes recommend pre-
analysis to verify the proposed conditions heating in the temperature range from 150 to
(Ref 23). 400 C (300 to 750 F) to avoid problems with
For both local and furnace treatment techni- hydrogen-induced cracking in welding these
ques, it is very important to exercise control over steels.
Martensitic stainless steels are normally deleterious to impact toughness. For enhanced
PWHTed between 600 and 800 C (1110 and corrosion, temper embrittlement, and tempering
1470 F), whereas ferritic steels are sometimes resistance, 0.5 to 2% Mo is added, depending on
heated between 730 and 800 C (1345 and the intended use. In order to develop the max-
1470 F), with rapid cooling in order to avoid imum strength and toughness, the steel must be
embrittlement. Treatments above 900 C mostly martensitic after cooling, with limited
(1650 F) in these materials are usually intended delta ferrite. The martensite must be tempered to
as homogenizing treatments in order to achieve obtain good toughness, ductility, and stress-
better properties after the tempering treatment corrosion resistance. In the as-welded condition,
that follows. The mechanical properties are the microstructure consists of low-carbon mar-
usually improved after this double heat treat- tensite, some presence of delta ferrite, and
ment. Low-temperature stress relieving should retained austenite in agreement with the nickel
not be applied to the straight chromium stainless content of the alloy. Postweld heat treatments
steels, since it may markedly affect the ductility are necessary to satisfy the service mechanical
and toughness. The phenomenon is known as property requirements (Ref 31). If high impact
475 C (885 F) embrittlement, and it is due to values are required, PWHTs such as solution
the coherent precipitation of chromium-rich annealing plus tempering or double tempering
ferrite, known as alpha prime, within the mis- are necessary (Ref 32, 33). The aim of solution
cibility gap of the iron-chromium system. This annealing is the homogenization of the micro-
precipitation leads to a slow increase in hardness structure by dissolution of the delta ferrite,
accompanied by a corresponding loss of tough- which is a nonequilibrium solidification product.
ness. The alpha-prime phase also decreases the The delta ferrite is harmful since it increases the
corrosion resistance. This type of structural ductile-brittle transition temperature. On the
change can be reversed by an annealing treat- other hand, intercritical tempering at 600 C
ment at approximately 600 C (1110 F). In (1110 F) or double tempering (hypercritical
17% Cr steels containing nickel and molybde- plus intercritical) at 670+600 C (1240+
num, the toughness is increased by tempering at 1110 F) produces tempered martensite with
630 to 650 C (1165 to 1200 F), below the finely dispersed austenite that is stable and not
temperature where austenite or ferrite is formed. transformable during cooling (Ref 33). It is
In straight 17% Cr steels, the precipitation of known that this austenite, which can be observed
sigma phase can occur between 550 and 800 C only by scanning electron microscopy, increases
(1020 and 1470 F), and it is accompanied by a toughness sharply, although it slightly reduces
loss of ductility. The sigma phase is formed only the strength. It has been argued that when
after a very long time and may be eliminated by retained austenite is present near a propagating
heat treating above 800 C (1470 F). crack, the concentrated strain at the crack tip
Soft martensitic stainless steels have re- induces transformation into martensite. This
sulted in an increasingly worldwide use in mechanically induced transformation would
petrochemical and chemical plants or industries, absorb energy and thus increase the toughness.
gas turbine engines, turbine blades, compressors The associated volumetric expansion of the
and discs, and in a variety of aircraft structural martensitic transformation would tend to close
and engine applications (Ref 27). They have the crack and relieve stresses at its tip. The
high proof strength and high toughness even in latter mechanism absorbs strain energy during
very low temperatures or thick cross sections fracture and therefore limits crack extension
(Ref 28, 29). If the 12% Cr stainless steels are (Ref 34).
used as high-strength structural steels, they must In many codes for austenitic chromium-
be weldable, formable, and have good impact nickel stainless steels with low ferrite content
toughness (Ref 30). Hence, in soft martensitic and fully austenitic alloys, no PWHT is pre-
stainless steels, the carbon content is kept below scribed. In the case of austenitic chromium-
mass 0.1% to improve weldability by promoting nickel steels, low-temperature treatment (400 to
a structure with fewer tendencies for cold 525 C, or 750 to 975 F) will help to achieve
cracking, better corrosion resistance, and better dimensional stability of the construction by
toughness. Because of the lower carbon, the reducing peak stresses, although this treatment is
addition of 4 to 6% Ni (the most powerful not frequently used. Treatment in the tempera-
austenite former after carbon and nitrogen) ture range from 550 to 1170 C (1020 to
is required to avoid delta ferrite, which is 2140 F) is a true stress-relieving treatment. The
highest part of the range (960 to 1170 C, or apart from M23C6 precipitation, particularly
1760 to 2140 F) involves solution treatment regarding mechanical properties. Due to the
and achieves maximum relief of stress; it may be highest chromium content of delta ferrite, the
performed to dissolve most of the carbides and weld metal containing delta ferrite is often more
sigma phase or delta ferrite. Treatments in the precipitation prone than the base metal of similar
range from 1040 to 1170 C (1905 to 2140 F), composition (because the solution-annealing
followed by water quenching after annealing, treatment normally applied to the base metal
are applied to prevent intercrystalline corrosion dissolves any delta ferrite).
and SCC. When PWHT is recommended, it only Stabilized Austenitic Stainless Steels. This
applies to heavy thickness, and a temperature group of steels contains grades that are alloyed
between 900 and 1000 C (1650 and 1830 F) is with titanium or niobium in order to improve
chosen, followed by water quenching or air their intergranular corrosion resistance. Tita-
cooling, depending on the thickness of the nium- and niobium-stabilized steels can be
component. Although it may seem difficult to welded using niobium-stabilized filler metals
imagine such treatments being applied to heavy with delta ferrite contents in the range of 7 to 15
components, there are some industrial examples. ferrite number. Whenever possible, heat treat-
The stress relieving of austenitic steels is not ment after welding, for example, stress-relieving
usually applied except for very thick sections. treatment, should be avoided. If it cannot be
An exception is in the case of cladding, where avoided, it is important to use special filler
the stress-relieving temperature is chosen with metals with less delta ferrite content to avoid
respect to the base material, and it is frequently sigma precipitation and detrimental effects on
in the range of 540 to 700 C (1005 to 1290 F). intergranular corrosion resistance. Stabilized
Austenitic stainless steels have a higher coeffi- steels are somewhat more susceptible to sigma-
cient of thermal expansion and lower thermal phase precipitation or to knife-line attack.
conductivity than ordinary ferritic steels, so Fully Austenitic Stainless Steels. This
greater distortion of welded components must be group of stainless steels has a stable austenitic
expected. structure that must normally be welded with
Unstabilized Austenitic Stainless Steels fully austenitic filler metals that do not produce
(UNS S30400, S31600, S30403, and any ferrite in the weld deposit. In the event of a
S31603). These grades normally possess PWHT, the intergranular corrosion attack range
excellent weldability, provided they are welded is relatively strongly influenced by the chro-
with filler metals that yield an austenitic-ferritic mium and nitrogen content. The presence of a
weld metal to avoid hot cracking during welding small amount of sigma-phase or chi-phase pre-
(5 to 15 ferrite number). It is necessary to follow cipitation is sufficient to give a marked drop
certain procedures in order to achieve sufficient in the pitting resistance. A 475 C (890 F)
corrosion resistance, cracking resistance, and embrittlement does not occur in fully austenitic
toughness. It is well known that austenitic weld metal, because this only occurs in the
stainless steels are always subjected by the steel presence of a ferritic structure.
manufacturer to a solution-annealing treatment, Duplex austenitic-ferritic chromium-
normally in the range of 1050 to 1100 C (1920 nickel stainless steels having delta content in
to 2010 F). In the course of this heat treat- the range of 30 to 60% are considered. An alloy
ment, carbide M23C6, sigma phase, and delta containing 25% Cr and 5% Ni, which, in
ferrite are completely dissolved, and the anneal- metallurgical terms, is very close to the widely
ing process produces a homogeneous, fully used duplex stainless steel 22Cr-3Mo-5Ni,
austenitic structure. With a subsequent quench- solidifies completely to delta ferrite from melt-
ing treatment, this state is maintained up to ing. During further cooling, d-c transformation
room temperature. If possible, the PWHT of starts at approximately 1200 C (2190 F) with
welded components should be avoided, with the the precipitation of predominantly nodular aus-
exception of a solution-annealing treatment. tenite at the ferrite grain boundary. During fur-
However, if heat treatment cannot be avoided, ther cooling to room temperature, there is only a
special attention must be paid to the influence of partial transformation to austenite. The structure
carbide and phase precipitations on the corro- now contains some 60% primary delta ferrite
sion-resistance and toughness properties of the and approximately 40% secondary precipitated
weld. The precipitation of sigma phase is the austenite. If such an alloy is now subjected to a
most important of all precipitation phenomena, solution-annealing treatment, this ratio can be
shifted to slightly higher austenite contents. to PWHT at temperatures above 600 C

Preheating of duplex steels is not normally (1110 F). Nickel alloy filler metals, abundant
required. With thick material, a preheating and versatile, are frequently used for piping and
between 100 and 150 C (212 and 300 F) may pressure vessel applications in refineries, che-
be advantageous. In order to obtain high ducti- mical plants, and power plants (Ref 35).
lity 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 tem- The Risk-Based Approach
perature must be set according to manufacturers and Heat Treatments
specifications. With a solution-annealing treat-
ment in the range of 1020 to 1100 C (1870 to Risk analysis is a powerful tool to rationalize
2010 F), very close to a metallurgical equili- the decision-making process and is applicable to
brium, it is possible to reverse any harmful heat treated steel components. Increasing inci-
structures in the HAZ that may have occurred dents led to regulatory action. Some regulations
during welding. It should be noted that this may require that recognized and generally accepted
lead to severe distortion. good engineering practice (RAGAGEP) must be
Heat Treatment of Austenitic-Ferritic followed. This also created the desirable objec-
Dissimilar Joints. The PWHT of austenitic- tive for industry to document what RAGAGEP
ferritic dissimilar joints or weldments should be was.
avoided whenever possible. However, heat The concepts of risk-based design, inspection,
treatments such as annealing or stress relieving and maintenance have been developed and are
may at times be unavoidable or even mandatory. being implemented (Fig. 42). Risk is the com-
They should always be adapted to suit the bination of the probability (or frequency of
requirements of the low-alloy steel section of occurrence) and consequence (or severity) of a
the joint. Often, the steel in question is a low- hazard. Its scope is limited to a specific envir-
alloy creep-resistant steel type used in boiler onment during a certain period of time. The
and pressure vessels that demand a PWHT to intent of risk-based initiatives is to use finite
suit each particular steel grade. In austenitic- resources and allocate these resources in a
ferritic dissimilar joints, such heat treatments manner that achieves the greatest overall
may lead to the occurrence of the following reduction in risk. Flaws that exceed the limits
phenomena: permitted by codes may be found. After a defect
that is not acceptable is found via a flaw eva-
Carbon enrichment in the weld metal due to
luation/fitness-for-service analysis, a repair or
the diffusion of carbon from the low-alloy
steel into the austenitic weld metal replacement is required for continued operation.
Various types of damage may occur
As a consequence thereof, carbon depletion
throughout the different heat treatment stages.
in the HAZ of the low-alloy steel
These types of damage may lead to failures
Coarse grain formation in the HAZ of the
in the product during its useful life. The term
low-alloy steel due to recrystallization pro-
failure refers to the inability of part of the
cesses
component or the entire component to perform
Embrittlement of the austenitic weld metal
the functions it was designed for, leading to an
due to precipitation of brittle phases, for
unprofitable and technically useless product.
example, sigma phase
Failure modes analysis is a procedure in
All these processes are time and temperature which each potential failure mode is analyzed to
dependent. Considering that the normal anneal- determine its effects and the criticality of these
ing time is in the range of 2 to 10 h, eventual effects on the system and to rank each potential
damage will normally occur, depending on the failure according to its severity. The criticality
material combination, at temperatures above analysis involves the use of risk analysis tech-
600 C (1110 F). Nickel-base weld metals, niques based on the assessment of the likelihood
due to their high nickel content outside the range of failure and its potential consequences.
of sigma-phase precipitation, show no signs Among the most widely used tools in the
of embrittlement during heat treatments. The analysis of design is failure mode and effects
use of nickel-base filler metals is recommended analysis (FMEA), which is accepted by the
if austenitic-ferritic dissimilar joints are subject Occupational Safety and Health Administration
Risk Based Initiatives
Economic
Factor
Risk-Based Risk-Based
Design Inspection
Risk Assessment,
HAZOP, FMEA, WI, FTA,
RCA, RCM, RBI,
Fitness-For-Service,
Life Extension
Material Science
FEA, Welding, etc.
Fitness- Failure
For-Service 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 allow the development of new tools to assist
reliability and safety. the designers and manufacturers in minimizing
By using the FMEA methodology in the failures related to heat treating operations.
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 REFERENCES
to avoid them, thus eliminating them completely
or achieving a more robust redesign that is less 1. L. Campos Franeschini Canale, G. Totten,
sensitive to failures. The FMEA has evolved and D. Pye, Heat-Treating Process Design,
from an ad hoc technique, dependent on a Handbook of Metallurgical Process Design,
designers experience, to a formal and accepted G. Totten, K. Funatani, and L. Xie, Ed.,
analysis technique. Failure modes can be Marcel Dekker Inc., New York, 2004
eliminated by removing their causes or at least 2. M. Solari, Risk Based Design, Chapter 2,
having their probabilities of failure reduced to Handbook of Mechanical Alloy Design, G.
acceptable levels. Totten, K. Funatani, and L. Xie, Ed., Marcel
There are several methodologies to assist in Dekker Inc., New York, 2003
logical thinking to resolve undesirable events. 3. G. Totten, M. Narazaki, R.R. Blackwood,
These tools include, among others, reliability- and L.M. Jarvis, Failures Related to Heat
centered maintenance, risk-based inspection, Treating Operations, Failure Analysis and
FMEA, modified FMEA, and root-cause Prevention, Vol 11, ASM Handbook, ASM
analysis. International, 2002, p 192223
Working into an integrated system with risk- 4. M. Jacob and G.A. Hawkins, Elements of
based efforts, fractomechanical-based structural Heat Transfer, John Wiley & Sons, Inc.,
integrity approaches, and failure analysis will 1957
5. M.F. Ashby, Material Selection in Mech- 21. Risk-Based Inspection, 1st ed., API Pub-
anical Design, Butterworth Heinemann, lication 581, American Petroleum Institute,
1999 Washington, D.C., 2000
6. R.F. Kern and M.E. Suess, Steel Selection, 22. C. Duren, Equations for the Prediction of
John Wiley & Sons, 1979, p 35 Cold Cracking in Field-Welding Large
7. ASME Boiler and Pressure Vessel Code, Diameter Pipes, IIW Document IX-1356-
Section 2 D, ASME International, 1999 85, The International Institute of Welding,
8. W.E. Dowling, Jr. and N. Palle, Design for Cambridge, England, 1985
Heat Treatment, Materials Selection and 23. R.W. Nichols, Ed., Developments in Pres-
Design, Vol 20, ASM Handbook, ASM sure Vessel Technology 3, Applied Science
International, 1997 Publishers LTD, London, 1980
9. F.A. Kandil, J.D. Lord, A.T. Fry, and P.V. 24. E. Folkhard, Welding Metallurgy of Stain-
Grant, A Review of Residual Stress Mea- less Steels, Springer-Verlag, New York,
surement MethodsA Guide to Technique 1994
Selection, Report MATC(A)04 Project 25. A.W. Marshall, Document IIW-IX-H-422-
CPM4.5, Measurement of Residual Stress in 98, The International Institute of Welding,
Components, Materials Centre, Middlesex, Cambridge, England, 1998
U.K., Feb 2001 26. J.C.M. Farrar, Document IIW-IX-H-423-
10. R.W.K. Honeycombe and H.K.D.H. Bha- 98, The International Institute of Welding,
deshia, Steels: Microstructure and Proper- Cambridge, England, 1998
ties, 2nd ed., Arnold, 1995 27. P. Bilmes, C. Llorente, and M. Solari, Role
11. R.F. Kern, Selecting Steels and Designing of the Retained Austenite on the Mech-
Parts for Heat Treatment, American anical Properties of 13Cr-4NiMo Weld
Society for Metals, 1969 Metals, Proc. the 20th ASM Heat Treating
12. K.E. Thelning, Steel and Heat Treatment, Society International Conference and
Butterworth, London, The Institute of Mat- Exposition, Oct 2000 (St. Louis, MO), ASM
erial, 1993, p 637 International
13. F.R. Hutchings and P.M. Unterweiser, 28. P. Brezina, Escher Wyss News, Vol 12,
Failure Analysis, American Society for 1980, p 218
Metals, 1981, p 3542 29. H. Niederau, Stahl Eisen, Vol 98 (No. 8),
14. H.E. Boyer, Quenching and Control of 1978, p 385
Distortion, ASM International, 1988, p 245 30. F.B. Pickering, Physical Metallurgy and the
15. Failure Analysis and Prevention, Vol 11, Design of Steels, Applied Science Publish-
ASM Handbook, ASM International, 2002 ers, London, 1978, p 165
16. G. Vander Voort, Failure Analysis and 31. T.G. Gooch, Weld. J., July 1995, p 213s
Prevention, Vol 11, ASM Handbook, ASM 32. R.D. Kane, Corrosion, Vol 33 (No. 7),
International, 1995, p 565 1977, p 231
17. Failure Analysis and Prevention, Vol 11, 33. P. Bilmes, C. Llorente, and M. Solari, Effect
Metals Handbook, 9th ed., American of Post Weld Heat Treatments on the
Society for Metals, 1986, p 335 Microstructure and Mechanical Behaviour
18. W.B.F. Mackay, Failure of Wrist Pins in an of 13Cr-4NiMoL and 13Cr-6NiMoL Weld
Automotive Engine, Handbook of Case Metals, Proc. the 18th ASM Heat Treating
Histories in Failure Analysis, Vol 1, ASM Society Conference and Exposition, Oct
International, 1992 1998 (Chicago, IL), ASM International
19. E. Levy, Cracking of an Alloy Steel Bolt, 34. P. Bilmes, C. Llorente, and M. Solari,
Handbook of Case Histories in Failure Characteristics and Effects of Retained
Analysis, ASM International, 1992 Austenite Resulting from Tempering of
20. G.M. Tanner, Hydrogen Embrittlement 13Cr-NiMo Weld Metals, Mater. Charact.,
Failure of Socket Head Cap Screws, Vol 46 (No. 4), April 2001, p 285296
Handbook of Case Histories in Failure 35. H. Ebert, Nickel Alloy Filler Metal Review,
Analysis, ASM International, 1992, p 332 Weld. J., July 2004, p 6062
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 unanticipated failures. Higher stresses or unan-
other material, may have various consequences, ticipated service conditions can cause unfore-
such as: seen failure because of complex or increased
stress fields. Stress concentrations may become
Making the device or component completely
more critical because of the increase in loading
inoperable
for the new application.
Preventing an operable device from func-
Insufficient design criteria can also be the
tioning satisfactorily cause of unforeseen failures. Inadequate
Making the device or component unsafe or
knowledge of the stress state in the component
unreliable, with immediate removal from
or inadequate stress calculations can contribute
service required
to failures. Much higher stress states than initi-
Many aspects may also be involved in tracing ally assumed or improper stress assumptions can
back to the possible sources of failure of a result in premature service failures. Lack of
component. Some of these sources include: consideration of severe environmental, fatigue,
or impact conditions may contribute to failure.
Design Material issues can usually be attributed to
Material issues, such as improper materials either selection of material or material imper-
selection or material imperfections (laps, fections rendering it unsuitable for service. In-
seams, inclusions, porosity, etc.) adequate material data can also result in
Fabrication and processing conditions that may contribute to failure. For
Rework example, adequate fatigue data, elevated-tem-
Assembly perature tensile data, or creep or corrosion data
Inspection may not be available, and the designer may have
Storage and shipment to extrapolate or estimate the effects or these
Service conditions properties.
Maintenance Other sources of failure can be attributed to
Unanticipated service conditions material imperfections. For wrought products,
Many times, more than one factor contributes to this could be related to segregation, inclusions,
a part failure. Rarely is it only one factor. porosity, laps, and seams. For castings, these
imperfections could be cold shuts, inclusions,
shrinkage, voids, and porosity. Forgings can
General Sources of Failure have laps, seams, segregation, and anisotropy in
properties from forging flow lines.
Design deficiencies are a common source of In one example (Fig. 1), a large roll was heat
component failure. Examples include the pre- treated, and several large cracks were observed
sence of a sharp notch in regions of high stress or after inspection. This was originally attributed to
a fillet radii that is too sharp. Using a component quench cracking. On further examination, it was
design for a new application can also lead to determined that a lap was present in the forging,
indicated by the presence of high-temperature abusive, can cause large temperature gradients
oxides in the crack along the crack faces. and localized overheating. This overheating can
Manufacture and Processing. Processing cause changes in microstructureeither loca-
can have a large influence on properties and the lized softening of the material or localized
resulting residual stresses. Typically, this is transformation to martensite and other trans-
related to wrong procedures or improperly spe- formation productsresulting in hard spots.
cified procedures. Ambiguous processes or In Fig. 2, a large gear was ground after heat
specifications can also contribute to failures due treatment. Because of abusive grinding, local
to interpretation or application. Simple things temperatures exceeded the austenitization tem-
like improper selection of processing sequences perature, and transformation to martensite
or procedures or specifications that were not occurred upon cooling. This transformation and
followed can also contribute to failure. the resulting residual stresses caused cracking of
Cold forming, such as stretching or deep the gear. Temper etch examination of the gear
drawing, can develop highly localized residual using dilute nitric acid in water in the regions of
stresses. Local changes in microstructure can cracking showed evidence of localized abusive
occur. Because of the changes in reduction, a grinding.
large anisotropy in material properties also re- Identification of parts can also cause failure to
sults. Due to the drawing operation, cracks or initiate. This is from localized impact or electro-
microcracking can occur. This could be due to etching. Localized mechanical stress concen-
improper lubrication or improper die design. trations or changes in microstructure can occur.
The localized changes in ductility can also con- This creates either a mechanical or micro-
tribute to failure. structural notch or stress concentration.
Machining and grinding can create high Heat treatment can cause a variety of different
residual stresses from either machining practice root causes for failures. Overheating, decarbur-
(feeds and speeds) or improper cutting tool ization, quenching, tempering, annealing, and
selection, material, or geometry. Grinding, if other heat treatments can cause failure to occur.
Fig. 1 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.
Overview of the Mechanisms of Failure in Heat Treated Steel Components / 45
This could also include improper austenitization There are also several embrittlement mech-
temperatures and times. Decarburization is the anisms caused by the use of improper tempering
result of a low-carbon surface from improper temperatures. Temper embrittlement and blue
atmosphere control. Typically, there is a deple- brittleness are just two of the common mech-
ted carbon layer at the surface that, when anisms that can occur from improper heat
quenched, is softer than the core material. This treatment and tempering operations.
soft layer can be completely devoid of carbon Cleaning, pickling, and electroplating op-
(complete decarburization) or only partially erations can also cause potential failures or
depleted in carbon (partial decarburization). contribute to them. Hydrogen charging of high-
This decarburized layer can contribute to pre- strength steels from the dissociation of hydro-
mature fatigue failures, because the surface gen on the surface of high-strength steel can
material is different than the designer expected, occur from cleaning operations in acids. Char-
or failure can result from high residual stresses ging of hydrogen from high current densities
created at the surface from the quenching in electroplating can cause hydrogen embrittle-
operation. The low-carbon surface area can also ment unless proper baking procedures are used
result in distortionagain, high residual tensile to allow the hydrogen to diffuse out. Electro-
stresses at the surface with low surface hardness. plating can also cause high residual tensile
Carburization is similar to the effects of de- stresses, which can contribute to crack initiation.
carburization. In this case, there is a higher sur- Welding can cause many different problems.
face carbon than expected. High residual tensile These problems can be cracks that are initiated
stresses can result as well as increased distortion. from improper welding procedures, high resi-
Quenching can also contribute to high residual stresses, porosity from inadequately dried
dual stresses or the formation of cracks or weld rods, or dirty workpieces. Microstructural
microcracking. Transformation stresses from notches or stress concentrations from the heat-
quenching cause the high residual stresses. affected zone and the transition to the base
These high residual tensile stresses can drasti- material can be the result of improper preheat
cally reduce the fatigue strength or have other and postheat. Improper weld penetration, weld
ramifications in service. geometry, and excessive weld current (under-
Overheating can cause excessive grain cutting) can also cause mechanical stress con-
growth, with resulting increases in hardenability centrations (Fig. 3).
and increased embrittlement. Underheating can The mast arm failure shown in Fig. 3 (Ref 2)
cause poor mechanical properties, because there was the result of weld bead undercutting and
was an incomplete transformation to austenite poor weldment design. Fatigue cracking initi-
and therefore an incomplete transformation to ated at the site of the weld toe undercut. This
martensite. Poor mechanical properties, such as location was a highly stressed area and the loca-
low tensile and yield stress, and poor impact tion of a large mechanical stress-concentration
properties may occur. factor because of the weld toe undercut. Typical
Fig. 2 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.
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 com-
ponent 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 tor-
que 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, tem-
perature (high or low), or a chemical environ-
ment, 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 Fig. 4 The probable cause of this accident was a loss of air-
plane pitch control resulting from the in-flight failure of
lubrication can also play a role in failure (Fig. 4). the acme nut threads on the horizontal stabilizer trim system
jackscrew assembly. The thread failure was caused by excessive
In the case of Fig. 4 (Ref 3), the lubrication wear resulting from insufficient lubrication of the jackscrew
schedule was extended to reduce aircraft assembly. Source: Ref 3
downtime. This, and other contributing factors, Nondestructive testing

resulted in the loss of 88 lives. Selection and preservation of specimens
Stresses from startup can also contribute, Mechanical testing
along with rapid temperature gradients and rapid Macroexamination
localized changes in the environment. Start-up Microexamination
procedures and maintenance are critical for Metallographic examination
intermittent operations. Shut-down procedures Determination of the fracture mechanism
and resulting stresses are just as critical as proper Chemical analysis (bulk and microanalysis)
startup. Inspection procedures to prevent failure Exemplar testing
are also important. Failure to properly inspect Analysis and writing the report
for problems or cracking can be catastrophic
(Fig. 5), (Ref 4). In this case, maintenance and These stages are described as follows, and ad-
ditional information on failure analysis proce-
inspection personnel failed to detect a fatigue
dures is given in the chapter General Aspects of
crack in the compressor stage of an aircraft
engine. Upon application of power, the com- Failure Analysis in this book.
pressor stage ruptured, with shrapnel severing
fuel lines and igniting the fuel, ultimately lead- Collection of Background Information
ing to the loss of the aircraft. During the collection of background data, the
engineer is trying to gather an understanding of
the purpose of the part. The engineer is attem-
General Practice Conducting pting to discern the design criteria, service
a Failure Analysis conditions, and failure conditions. In the back-
ground information, the operating details and
The primary objective of any failure analysis manufacturing history should be examined and
is to determine the primary root cause of failure collected. This manufacturing history should
and to establish the appropriate corrective include all the mechanical processing, thermal
action. There are several stages of an analysis, history or processing, and any chemical process
which can proceed one after the other or occur at performed on the part.
the same time. There is no set fixed-in-stone The service history should include all the
procedure, because it is highly dependent on the maintenance records of the part. It should also
part and procedures/capabilities of the specific include the expected environment and loading at
laboratory. the time of failure, as well as the normal environ-
These stages of analysis are: ment and loading. Any quality records should be
Collection of background data examined for discrepancies. Unfortunately,
Preliminary visual examination 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

Documenting the failure or fracture is extre-
mely important. There can never be too many
drawings or photographs. The cost of photo-
graphs (especially digital) is cheap compared to
analysis. A high-quality camera with macro-
capability 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
Fig. 5 The probable cause of this accident was the failure of the color of scale or oxides can often give an
maintenance and inspection personnel to perform a
proper inspection of a seventh-stage high compressor disk, thus
indication of the temperatures that the part has
allowing the detectable crack to grow to a length at which the disk experienced.
ruptured under normal operating conditions, propelling engine Sample selection is also very important. All
fragments into the fuselage. The fragments severed the right
engine main fuel line, which resulted in a fire that rapidly associated debris should be collected and iden-
engulfed the cabin area. Source: Ref 4 tified. Similar parts should also be collected for
comparison. In the case of a fastener failure, it is Magnetic particle inspection uses discon-
important that the nut and washer be collected, tinuities in the magnetic field to identify cracks
too. All mating pieces should be gathered for or discontinuities. Fluorescent dyes with small
subsequent analysis. magnetic particles are used. These magnetic
Any abnormal conditions should be observed particles gather at the discontinuities in the
and compared with new and used components. magnetic field, indicating flaws or indications. It
Any discoloration or debris should be noted and is a common, sensitive, and reliable method that
collected. Any distortion of the part should be is simple to learn and use. This method has no
noted, along with dimensions of the part. limitation in part size but is limited to magnetic
Weather conditions at the time of failure should materials. No elaborate precleaning of the sur-
be collected, as well as all bearing and lubrica- faces is necessary. Detection is limited to the
tion conditions and records. surface of the part or section examined. Care
During the initial wreckage analysis, the must be exercised to prevent local arcing.
determination of all wreckage should be identi- The dye-penetrant method is useful for
fied and located on a map or grid before any is examining surface flaws or cracks. It is used
touched or moved. Photograph each piece of primarily for nonferrous alloys but is used for
wreckage and its surroundings. Inventory the examining ferrous weldments for cracks and
parts present or missing. Determine the operat- porosity. In this method, a high-wetting liquid is
ing conditions at time of failure. This should spread on the surface of the part. Excess liquid
include the position of control surfaces, power is wiped off. A developer is applied to the part
settings, position of throttles, and any lights or surface. Any cracks, flaws, or other indications
annunciations that occurred. will appear. Limitations of this method are the
As best as possible during the initial exam- necessity of cleaning the surface prior to and
ination of the wreckage, the sequence of failure after application of the indication fluid and
should be determined. This can be accomplished developer solution. Surface features may also
by examining chevron markings and crack mask indications. It is simple to use, but an
order. The parts should then be closely examined understanding of the limitations must be
and reassembled. DO NOT allow the fracture understood prior to application to a part.
surfaces to touch each other, because this can Eddy-current methods depend on the princi-
cause potential damage to the delicate surfaces. ple that all metals conduct electricity. An alter-
This analysis can also help determine the nating current is applied, and eddy currents
sequence of events leading up to failure. Pre- occur by electromagnetic induction. Cracks or
liminary examination of the part should note any other flaws cause distortions in the electro-
paint, debris, or deposits present. Always magnetic fields, with a result of changing the
remember to do no harm. field impedance. The advantage of this method
The visual examination should be detailed. is that subsurface discontinuities can be detec-
Fracture surface crack directions should be ted. No special skill is required to use this
noted, identified, and documented. Any abuse or method, and the method can be automated.
discoloration should be identified, and a general Probe contact with the part is not needed. Lim-
assessment of the workmanship should be itations of this method are that the depth pene-
determined. Document all findings with photo- tration is limited, and the part must be capable of
graphy, with multiple photographs taken from conducting electricity. Reference standards are
different directions. The incorporation of rulers needed for specific flaw sizes and materials.
or scales is important to determine the size and Many things can influence readings, including
direction of fracture. segregation, carburized layers, and changes in
profile.
Ultrasonic testing uses high-frequency sound
waves transmitted through a conducting med-
Nondestructive Testing ium. Any discontinuous boundary can cause a
Nondestructive testing is very useful for deflection. This method is very sensitive and has
determining the extent of cracking. Magnetic high penetration. It is possible to get accurate
particle inspection is useful for ferrous alloys, measurements of flaw position and size, but
with dye-penetrant and ultrasonic inspection reference standards must be used. Shape and size
as additional methods available for initial can cause errors in interpretation. Experienced
inspection. operators are required to properly interpret the
results of testing. Effects of grain size, porosity, test. These methods also do not incorporate the
and inclusions can also hinder interpretation. effects of residual stresses.
Radiography, using x-rays, neutrons, or
gamma rays, is also often used to examine
structures. Film or sensors (charge-coupled de- Selection and Preservation of Specimens
vices) pick up the emitting radiation, with the The selection and preservation of fracture
intensity proportional to the density of the surfaces is vital to prevent the destruction of
sample. Light areas indicate a dense region, and evidence. Unprotected, the fracture surfaces or
dark areas indicate a greater exposure or less parts can become mechanically or chemically
dense region. Advantages of radiography are the damaged. This damage can obliterate evidence
detection of subsurface and internal features at and make the determination of fracture difficult
various depths and the documentation of these or impossible. Both sides of the fracture must be
features by film or other imaging techniques. The protected. This is in the event that if one surface
primary disadvantage is that reference standards is damaged, the other side can be examined.
must be used, and the area for testing must be Protection of the specimens during shipment is
enclosed to prevent radiation from leaking out. also very important, because evidence could be
destroyed. Avoid touching surfaces with the
hands, because the chemicals and acids present
Mechanical Testing can cause artifacts or destroy data. NEVER fit
Mechanical testing is useful to determine the surfaces together, because the delicate fracture
properties of the part and to verify that it meets features can be destroyed. Since both surfaces
expected properties and specifications. There are would be damaged, it could destroy the chances
many types of mechanical testing available, for determining the fracture mechanism.
including hardness, tensile testing, and impact Cleaning of specimens is to be done only
fracture testing. when absolutely necessary. For the most part,
Hardness testing is probably the most versa- it is required to prepare the sample for the
tile and widely used. It is often used to evaluate scanning electron microscope (SEM). Dry air
heat treatment and can be used as an approx- blasts or soft artist brushes are typically all that
imation for tensile strength. It can be used to is needed. Rinsing in organic solvents then
detect the presence of work hardening or soft- evaporating the solvent with dry air is useful for
ening and hardening or softening from localized preparing specimens for the SEM.
thermal events such as grinding. For the most Chemical cleaning is generally not recom-
part, it is a nondestructive test. For microhard- mended under any circumstance. Foreign sub-
ness testing, it is necessary to use a metallo- stances such as scale or debris should be
graphic specimen. preserved. Do not use rust inhibitors, because of
Tensile testing is used more to establish con- the inevitable damage to the part and fracture
formance to specification. It is not necessary to surfaces. These rust inhibitors are also extremely
show inadequate ductility because of service difficult to remove. Avoid washing the sample or
loads. Because of the size of the tensile speci- parts with water unless seawater or other che-
men, it may not be possible to excise an appro- mical is present. In this case, gently wash with
priately sized sample from the part. Anisotropy distilled water and follow that with high-quality
of properties can be expected to lower measured alcohol or acetone. Allow to dry and place in a
tensile and yield strength properties. dessicator.
Impact and fracture toughness testing is Plastic replicas are useful in preserving
typically used to determine conformance to fracture surfaces and removing debris for fur-
specifications. Charpy impact testing has a high ther analysis. Softening replica tape (available
variability in results and may be temperature at transmission electron microscope supply
related. Results must be taken with temperature houses) with small amounts of acetone forms
in mind and may not correlate with real results plastic replicas. The softened tape is pressed
because of size limitations. Fracture toughness gently onto the fracture surface. Additional
testing and the results from KIc testing can be layers of tape, softened with acetone, are applied
used in design, and the results are useful for to the fracture surface. After multiple layers
calculating critical flaw sizes. It can also be have been applied, the entire replica is allowed
used to examine estimated crack growth rates; to dry and then is placed in a dessicator. When
however, samples are difficult to prepare and the part is ready to be examined, the replica is
carefully removed using tweezers. Any debris information on the location of fracture origins,
on the surface is also preserved for further direction of cracking, configuration of the stress
analysis in the replica. Multiple plastic replicas state, and the last region to fail (shear lip). The
can be used to clean a surface of a part. This can presence of chevron marks can indicate the
be repeated as necessary. direction of rapid crack growth, and the different
textures of the fracture can differentiate between
fast final fracture and the initiating mechanism
Sectioning of Specimens of fracture. Different textures from the region of
Sectioning is very important, because it cap- fast fracture can indicate a different mechanism,
tures the portion of the fracture surface for such as fatigue, stress-corrosion cracking, or
examination or the appropriate metallographic hydrogen embrittlement.
specimen. The biggest limitation is size. It is
important that the portion to be removed is docu-
Microscopic Examination
mented by photographs and sketches, showing
the location of the specimen to be removed. The microscopic examination is usually con-
Preserve any fracture surface by plastic replicas ducted with an SEM (Fig. 6). This instrument
or other method to prevent damage or attack. is probably the most useful of all instruments
Regions adjacent to cracks are also to be pre- for determining the mechanism of failure. It is
served and protected. Cutting the specimens capable of a large depth of field, with magnifi-
should be done very carefully so as not to cause cations of 10 to 300,000 . It allows for direct
any heat damage. Coolants are not recommen- examination of specimens, and when coupled
ded, unless the material cannot be cut without with an energy-dispersive spectrometer, very
heat generation. The use of plastic replicas is small regions can be examined and analyzed for
useful for protecting surfaces and preserving any chemistry. It is very easy to use and requires very
debris present. little training to take quality images.
Opening secondary cracks is useful when the Interpretation of the images requires experi-
primary fracture surface is damaged. These sec- ence and understanding of the four basic modes
ondary cracks may provide better information, of failure: dimpled rupture, cleavage, brittle
because they are tightly closed, and the fracture intergranular, and fatigue. From these four basic
surfaces are not exposed to surface contaminants modes, the detailed mode can be examined, and
and corrosion. Care must be taken not to damage the failure mechanism is fit to the evidence.
the primary fracture surface. Bending to open A greater discussion of the mechanisms of fail-
the crack is preferable, to expose the crack face. ure is found later in this chapter and elsewhere in
Often, the use of a sawcut to the back of the part this book.
will reduce the amount of force necessary to
open the secondary crack. Another method is
Metallography
to use a tensile machine to open the crack face.
The crack opening should be measured prior to Metallography is a vital part of a failure
opening, and the crack opening displacement analysis investigation. It can examine crack
can also be measured as the crack is slowly
opened and exposed. One technique is to im-
merse 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.
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 Fig. 6 Typical scanning electron microscope used for
macroscopic examination can reveal a wealth of microscopic analysis of a fracture surface
morphology and its relationship with the treatment. It often provides the needed docu-
microstructure present. It can help determine the mentation and support for the fracture analysis
thermal history of a component or region of a and determination of the root cause of failure.
part and can show if work hardening was pre-
sent. There can never be too many photographs
and metallographic sections. Metallographic
sections should be taken away from the crack Determination of the Fracture
and near the determined origins of cracking. Mechanism
Because this method is destructive, it is under-
taken last. Typically, the crack face and edges Examination of the fracture surface and
are protected from rounding by applying sup- metallography are used to determine the cause of
port. This support can be electroless nickel plate failure. First, it is necessary to determine the
or the use of alumina beads or steel shot in the fracture mode. Unfortunately, there is no clear
metallographic specimen, adjacent to the sur- or logical classification of fracture. Generally,
face. classification is based on the crack growth mech-
Metallographic specimens are prepared using anism (see also the chapter General Aspects of
an epoxy or phenolic resin. The sample is placed Failure Analysis in this book).
into a small press, and phenolic resin is poured
over the section. The press compacts the resin
Ductile Fracture
and forms a small, round sample that is then
polished, etched, and examined under a metal- On a macroscopic scale, a ductile fracture is
lographic microscope. When the specimen has accompanied by a relatively large amount of
cooled, it is taken out of the press and ground plastic deformation before the part fails. After
through a sequence of sandpapers. Typically, the failure, the cross section is reduced or distorted.
sequence is 240, 320, 400, and 600 grit. The Shear lips are observed at the latter part of
specimen is ground very flat before polishing. the fracture and indicate the final failure of the
During polishing, the metallographic specimen part. The fracture surface is dull, with a fibrous
is polished using a flat platen and 3 mm alumina appearance. Microscopically, ductile fracture
slurries. Final polish is accomplished using is characterized by several distinct stages
0.15 mm alumina slurry. Other polishing agents (Ref 58); an example is shown in Fig. 8. In this
can be used, with diamond being a very common case, an ISO 12.9 low-alloy bolt failed by ductile
polishing agent. A finished metallographic torsional overload. The fracture was smooth,
sample used for the determination of the fracture with fracture initiating from the threads. The
mechanism in a steel weldment is shown in fracture mode was microvoid coalescence
Fig. 7. (Ref 9), which occurs by the following process:
Examination of the metallographic specimen
A free surface is created from a small parti-
reveals surface imperfections, inclusions, and
cle. This particle can be a second-phase
microstructural details. It can reveal the pre-
particle, dispersoid, or inclusion. The separ-
sence of decarburization and improper heat
ation 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 sur-
face.
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
Fig. 7 Typical metallographic specimen. This specimen was the particle, then the initial formation of voids is
used to examine microstructures in a failed weldment. the critical step. Fracture occurs shortly after
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 coal-

escence
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 sub-
stantially reduce the cross section, with the
resulting local plastic instability (Ref 11). These Fig. 10 Schematic representation of the creation of dimples
voids coalesce to form a central crack perpen- in a loaded member by (a) simple tension, (b) shear
loading, and (c) tearing
dicular to the applied tensile stress. Depending
on the applied stresses, the shape and config-
uration of the dimple shape can be changed
(Fig. 10). This fact is important in determining
the type of loading during a postfracture inves-
tigation. 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 impor-
tant 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. 11 Microvoid coalescence as seen through the SEM
important to the ductility of the steel. All other amounts of inclusions reduce the ductility of the
things being equal, the steel with the lower steel. A reduction in the yield stress, due to the
inclusion size, shape, and frequency will have a stress concentrations around the inclusions, is
greater ductility than another steel with a greater evident at low volume concentrations of inclu-
inclusion count. Modern steelmaking practices sions.
generally produce low inclusion content. Often, The presence of inclusions in the size range of
steels for aerospace applications require a fre- 1 to 30 mm reduces the energy absorbed during
quency/severity determination of inclusions in ductile fracture. Fine dispersions of ductile
accordance with AMS 2300, AMS 2301, AMS inclusions will delay the onset of cleavage-type
2303, or AMS 2304 (Ref 1316). A specific- fracture by localized relaxation of stresses. At
sized test specimen must be heat treated and the same time, the yield stress also increases.
examined using magnetic particle inspection. During deformation, forming, or forging, the
The procedures are outlined in the aforemen- ductile inclusion MnS has a marked effect on the
tioned specifications. ductility of the final product. Types 1 and 2 MnS
The inclusions found in steels have been inclusions will elongate on deformation, while
divided into five categories related to their type 3 MnS inclusions will rotate into the rolling
deformation behavior (Ref 17): plane. This will reduce toughness and ductility
in the transverse direction. Type 2 inclusions are
The inclusions Al2O3 and calcium alumi-
the most harmful to ductility and toughness, so
nates are produced during deoxidation of
some effort is being made to eliminate these
steel during the production of molten steel.
inclusions by ladle additions of other strong
They are brittle at practically all tempera-
sulfide formers, such as titanium, zirconium, and
tures.
calcium.
Spinel-type oxides are not deformable up to
Ductility is also influenced by the fact that
1200 C but may be deformed above this
MnS contracts more than the iron matrix upon
temperature.
cooling. The bond between the MnS inclusion
Silicates of calcium, manganese, iron, and
and the matrix is not strong enough to prevent
aluminum in various proportions are brittle
microvoid formation. Because MnS inclusions
inclusions at room temperature but become
tend to form as strings or stringers along the
more deformable at higher temperatures.
rolling direction, the toughness and ductility
The formability increases as the melting
are strongly influenced in the rolling direction.
temperature of the silicate decreases. There-
Transverse to the rolling direction, ductility and
fore, aluminum silicate has much less form-
toughness are much worse.
ability than the lower-melting manganese
In a similar fashion to that of inclusions, the
silicates.
distribution of carbides can also influence the
FeO and (FeMn)O are deformable at room
toughness and ductility of the steel. The strain
temperature but gradually become more
needed for void formation decreases with
brittle at temperatures above 400 C.
increasing carbide volume fraction. Spheroidal
Manganese sulfide (MnS) is the most com-
carbides will not crack at small strains and
mon inclusion found in steel, and it is
exhibit decohesion. Spheroidized steel is much
increasingly deformable as the temperature
more ductile than similar steel of the same
falls. The morphology of the MnS inclusions
hardness containing only ferrite and pearlite.
changes, depending on how they were
Pearlite has a lower critical strain for void for-
formed.
mation. In addition, when a crack or void forms
Ductile failure can occur with any of the in a pearlitic matrix, it will tend to run along
types of inclusions. This is true whether it is the length of a pearlite lamella. Examining this
the brittle alumina-type inclusions or the more type of fracture under the SEM reveals that the
ductile sulfide-type inclusions. Inclusions gen- base of the dimples contain fractured pearlite
erally initiate ductile cracking above a critical lamella.
size. Coarser inclusion sizes tend to have a larger
local stress-concentration factor, which can
Brittle Fracture
cause local decohesion and microcrack forma-
tion. Work by Maropoulos and Ridley (Ref 18) Very little plastic deformation and a shiny
has shown the effect of volume fraction of iron- fracture surface characterize brittle fractures.
alumina on the ductility of steel. Increasing Often, chevron patterns point back to the origin
of failure (Fig. 12) (Ref 19). It can occur at low heat treatment caused metallurgical or
stress and propagate with rapidity, often at microstructural notches to occur from abrupt
speeds approaching the speed of sound in the changes in microstructure or the production
failed material. of microstructures that were brittle. Features
Since the early 1940s, there has been tre- such as porosity from welding or casting also
mendous growth in the number of large welded caused brittle fracture initiation.
structures. Many of these structures have failed Material flaws: These flaws resulted from
catastrophically in service, most notably the melt practice at the mill and appeared as
Liberty ships (Ref 20) used to transport war large inclusions, internal oxidation, porosity,
material during World War II. Analysis of the or segregation.
fracture surfaces of the failures (Ref 21) indi-
cated that they initiated at a notch and propa- In brittle fractures, limited energy is absorbed
gated with no plastic deformation. These by the fracture. Energy is absorbed through
notches were of three types: regions of small plastic deformation. Individual
grains separate by cleavage along specific
Design features: Structural members were crystallographic planes. This is shown in Fig. 13.
rigidly joined at angles less than 90 and Visually, little or no plastic deformation or
then welded. distortion of the shape of the part characterizes
Fabrication details: Procedures used during brittle fractures. The fracture is usually flat and
the manufacture of the part caused the perpendicular to the stress axis. The fracture
formation of notches. Welding arc strikes, surface is shiny, with a grainy appearance.
gouges, and fitting procedures created Failure occurs rapidly, often with a loud report.
physical notches. Welding procedures and 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 schemati-
cally in Fig. 14.
There are three basic factors that contribute to
a cleavage type of fracture in steels. They are:
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
Fig. 12 Chevron markings point back to the origin of failure
triaxial stress state and low temperature. This is
in brittle steels. Source: Ref 19
Fig. 13 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
actuated by a high rate of loading. Many types of Changes in the NDT can be produced by
tests have been developed to determine the changes in microstructure and chemistry. The
susceptibility of steels to brittle behavior. These largest change can be effected by changes in the
tests include the Charpy impact test (ASTM amount of carbon and manganese. The NDT is
E23) (Ref 24) and the fracture toughness test lowered by approximately 6 C (10 F) for
(ASTM E399) (Ref 25). Others include the nil- every 0.1% increase in the manganese con-
ductility test (ASTM E208) (Ref 26) and centration. Increasing the carbon content also
dynamic tear test (ASTM E604) (Ref 27). lowers the NDT. The manganese-carbon ratio
The notch toughness of low- and medium- should be approximately 3 to 1 for good notch
strength steels is highly dependent on tempera- toughness.
ture. There is a transition from ductile fracture to Decreasing the concentration of phosphorus
brittle fracture as the temperature decreases. also decreases the NDT. Nitrogen causes the
One criterion for the transition temperature is the NDT to increase (more brittle). However,
nil-ductility temperature (NDT). The NDT is the because of the interaction with other alloying
temperature where fracture becomes 100% elements in steel, it is difficult to quantify the
cleavage, and there is essentially no plastic increase of NDT with increasing nitrogen con-
deformation. centration.
Fig. 14 Schematic of river patterns formed in brittle materials. (a) Tilt boundary. (b) Twist boundary. Source: Ref 23
Nickel is beneficial for increasing ductility. In the intermediate range, the fracture beha-
Up to 2% Ni is effective in lowering the NDT. vior is complicated. The fracture does not con-
Increasing concentrations of silicon have the sist of entirely slant-type fracture, nor does it
effect of increasing the NDT. Chromium has contain entirely a flat plane-strain-type fracture.
nearly no effect, while molybdenum is extre- Instead, the regions of flat and slant fracture are
mely effective in increasing the ductility of approximately equal. At the thin end of the
steels and drastically decreasing the NDT. thickness range, the slant ligaments on either
Oxygen strongly decreases the ductility. It can side of the testpiece carry most of the load. At
also cause an increased propensity for inter- the thick end of the range, the side ligaments
granular fracture by creating brittle oxides at the carry a much smaller percentage of the load. The
grain boundaries. Decreasing the grain size has a amount of flat fracture increases. This is shown
strong effect on increasing the ductility and schematically in Fig. 17. It has been found (Ref
notch toughness. 28) that the amount of flat fracture depends only
Section thickness can also influence ductile on the thickness of the test specimen and was
and brittle behavior (Ref 28). The results showed independent of crack length.
that there was considerable variation of tough- In the third region, the fracture consists of
ness with the thickness of the specimen (Ref 29, predominantly flat fracture. Some evidence of
30). Further, at large thickness, the toughness very small shear lips may be present at the later
appeared to reach a constant value (Fig. 15) part of fracture. Fracture is catastrophic and
(Ref 31). Within this curve, there are three rapid. No plastic deformation is evident. In this
apparent regions. First, there is the region where third region, any increase in the thickness of the
maximum toughness is obtained (thin sections). testpiece causes no further decrease in the
Second, there is the region of intermediate toughness.
toughness, and lastly, a region with relatively These fracture patterns are useful in deter-
constant toughness (thick sections). mining the state of stress within a failed com-
In the first region, the fracture appears to ponent and can help to understand the
consist entirely of a shear lip, or, in other words, mechanism of failure.
the fracture surface is inclined at an angle One famous failure involving brittle fracture
of approximately 45 to the tensile axis. In this was the Great Boston Molasses Disaster
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. Be-
cause 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.
Fig. 15 Variation of toughness with thickness Fig. 16 Mode of separation in a thin sheet
(Ref 34). In this failure, the United States he have any technical training. No engineers or
Alcohol Company fabricated a large cast iron architects were consulted to ensure that the tank
molasses tank in Boston in December 1915. This was constructed safely. On January 15, 1919, the
tank was 27 m (90 ft) wide and 17.7 m (58 ft) tank exploded with great force, and the streets of
tall, with a head of 15 m (49.5 ft) of molasses. It Boston were flooded with waves of molasses
was fabricated of cast iron plates riveted toge- from 2 to over 4 m (8 to 15 ft) tall (Fig. 18). This
ther. It held 8.7 106 L (2.3 million gal) of great wall of molasses was reported to have
molasses, ostensibly used for the fermentation of moved at speeds up to 35 miles (56 km) per
ethanol used for liquor. The man who oversaw hour and devastated a large section of Boston
construction could not read blueprints, nor did 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 build-
ings 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 transi-
tion 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
Fig. 17 Schematic of fracture in the intermediate range 4330V hook-point, used for the arrestment of
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.
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 micro-
structure 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 quasi-
cleavage, suggestive of brittle failure (Fig. 20).
Charpy impact testing showed that the impact
toughness of the as-received part was sig-
nificantly lower than a part of the same chem-
istry properly tempered to HRC 46. Finite Fig. 20 SEM fractographs showing (a) location of origin at the
inner fillet radius and (b) quasi-cleavage evident on
element analysis showed a high localized stress the fracture surface
concentration at the lug inside fillet radius.
It also showed that the stresses were highly and that the failure was aggravated by high local
triaxial. Based on the analysis, it was determined stress concentration at the fillet radius, improper
that the hook-point lug failed by quasi-cleavage, heat treatment (making the material more
brittle), and extremely high dynamic loading. It penetrate by diffusion. An example would be
was recommended that the radius be made larger galvanized steel where the zinc has diffused
to reduce the stress concentration and also to into the steel in the vicinity of 420 C
retemper the hook-points to meet specification. (787 F).
Hydrogen embrittlement: The presence of
hydrogen and static loads or a low strain rate
Intergranular Brittle Fracture can result in hydrogen embrittlement.
Another form of brittle fracture is called Stress-corrosion cracking
intergranular cracking. In this fracture mechan- Grain-boundary decohesion at elevated
ism, failure occurs by decohesion along grain temperatures (creep rupture)
boundaries and not on specific crystallographic The fracture surface appearance of inter-
planes, such as in cleavage fracture. Inter- granular cracking is generally shiny and faceted.
granular cracking can have several different It has a rock-candy appearance. Often, when
causes. Typical causes of intergranular cracking the mechanism is from corrosion, the corrosion
in steel alloys include: product is present. This can dull the appearance
Quench-age embrittlement: Cooling of car- of the facets. The appearance of intergranular
bon steels and low-alloy steels from sub- fracture is most clearly seen in the electron
critical temperatures can precipitate carbides microscope, and an example is shown in Fig. 21.
within the microstructure. The strength is Quench cracking is the limiting case of
raised, but toughness is lost. excessive residual stresses exceeding the tensile
Quench cracking: During quenching, the strength of the material. Two processes con-
transformational and residual stresses tribute to quench cracking, as well as distortion
developed during quenching of steels can and residual stresses. The first process is the
cause cracking during heat treatment. stress from the volume expansion of martensite
Tempered martensite embrittlement: Within during transformation from austenite to mar-
the range where blue-purple oxides can form tensite. The second source is from thermal stress
on steels (230 to 370 C, or 450 to 700 F), due to differential contraction due to different
precipitates can form that increase the tensile cooling rates in the steel. The transformational
strength and hardness while reducing the stress from the formation of martensite is pri-
ductility and toughness. marily responsible for cracking during quench-
Temper embrittlement: Quenched steels ing, and thermal stresses from differential
containing appreciable amounts of manga- cooling are usually from subcritical heat treat-
nese, silicon, nickel, or chromium are sus- ments such as annealing.
ceptible to temper embrittlement if they During quenching, the volume expands from
contain even trace amounts of antimony, tin, the close-packed face-centered cubic structure
or arsenic. Embrittlement of susceptible of austenite to the body-centered tetragonal
steels can occur after heating in the range of structure of martensite. This volume expansion
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 com-
mon failures in high-temperature, oxidizing
conditions.
Liquid metal embrittlement or solid metal
embrittlement: Intermetallic compounds
form at grain boundaries when low-melting-
temperature metals (cadmium, zinc, etc.) Fig. 21 Intergranular fracture from hydrogen embrittlement,
as seen through the SEM
is approximately 4% and is related to the carbon Martempering is the process of using high-
content of the steel. During quenching, the outer temperature quench oils and quench oil tem-
surface of the part cools first and transforms to peratures of 90 to approximately 200 C (200 to
martensite. There is an attendant volume 400 F). The part is quenched into the high-
expansion at the surface, and the untransformed temperature oil, and the parts are allowed to
and still hot interior surface usually has suffi- equilibrate or at least minimize the temperature
cient plasticity to accommodate the changes in gradient across the interior of the part. The part
the part volume. The outside surface is in com- is then removed from the oil and allowed to cool
pression. Upon cooling, the interior of the part in any convenient manner. This method has
also transforms to martensite but is constrained proven to be very effective in reducing quench
by the hard outside surface layer of previously cracking as well as distortion from quenching.
transformed martensite. On the transformation A long pinion gear failed in service near the
of the inner core, a volume expansion occurs in midlength of the shaft (Fig. 22). One gear tooth
the interior of the part, and the outer surface is fractured during service, resulting in the gear
placed in tension. If quenching is severe, the being removed from service and sent to the
resulting tensile residual stresses can exceed the laboratory for failure analysis. Magnetic particle
ultimate tensile stress of the surface untempered inspection, using a fluorescent dye, revealed
martensite. Cracking is intergranular and often the presence of multiple linear indications on
exhibits an oxide scale on the fracture surface. If cracking of the gear tooth faces (Fig. 23).
cracking occurred during quenching, remnants Examination of the fracture surface showed a
of quench oil can be found on the surface of the discolored region at the origin of cracking
crack, and often, elevated-temperature scale is (Fig. 24). This discolored region was attributed
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 tem-
perature. 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.
Fig. 22 As-received pinion gear that failed in service
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 higher-
hardenability alloys will also reduce the pro-
pensity for cracking, because it will allow a
reduced quench rate to achieve the same prop-
erties. Reducing the austenitizing temperature
or reducing the temperature differential between
the austenitizing temperature and the quenchant
temperature will reduce the propensity for crack-
ing. 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 Fig. 23 Magnetic particle inspection of the failed pinion gear
quench cracking. showed arc-shaped cracks on the gear tooth faces.
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.
tempering (Fig. 25, 26). If the crack was pre-

existing prior to heat treatment, it would be
Fig. 25 Closeup of the fracture region showing the dis-
darker and thicker.
colored region. The color of the oxidation indicated
that the crack occurred after quenching and during the tempering Examination of the tooth faces showed sec-
operation.
ondary cracking at regions of tearing and
smearing along the tooth face (Fig. 27), sug-
to oxidation that occurred during heat treat- gestive of abusive machining practice, including
ment. The coloration of the oxide scale sug- the use of a tool that was dull or excessive feeds
gested that the oxidation occurred during and cutting speeds.
Fig. 27 Secondary cracking evident at regions of abusive Fig. 28 Metallographic specimen of the pinion showing
machining 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 evi- caused by quench cracking, aggravated by

dence of burning or excessive temperature. The improper concentration control and induction-
root of the tooth showed little evidence of proper hardening parameters. The situation was further
hardening or case (Fig. 28). The microstructure aggravated by poor machining practice, creating
in the root consisted of ferrite and pearlite, with tearing and smearing at the surface.
lightly tempered martensite, further suggesting Often, quench cracking can result not from
inadequate heat treatment. The tooth tip showed heat treating operations but from other sources,
a fine-grained martensitic structure. No evidence such as abusive grinding (Fig. 29). In this case, a
of overheating was present. Examination of large gear was found to be cracked. As is usually
the tooth surface showed tears and smearing. the case, the heat treater was blamed. Temper
Microhardness of the hardened regions of the etching of the region of cracking showed a dar-
tooth showed a hardness of HRC 58, while the kened region, suggesting overtempering of the
root of the tooth was HRC 29, consistent with part in a localized region.
the observed microstructure. Localized overheating during service can also
Investigation of the induction heat treating result in quench cracking. A hook-point, used
conditions revealed that the concentration of the for catching the large cable on an aircraft carrier,
quenchant used was approximately 5%, while showed evidence of cracking in the cable groove
6 to 10% was specified. The concentration was (Fig. 30) after a carrier landing. The hook-point
controlled solely by refractometer. Contamina- was manufactured from AMS 6411 (AISI
tion of the quenchant was unknown. 4330V), heat treated, and a high-velocity oxy-
Based on the evidence, it was determined that fuel (HVOF) coating was applied. Imprints of
fracture and failure of the pinion gear tooth was the arresting cable were left in the cable groove.
The vertical cracks were exposed using liquid prior-austenite grains. The laboratory fracture
nitrogen and impact loading. The crack faces showed microvoid coalescence. The cracks
were discolored with a golden tint; they were showed three distinct regions: incipient melting
subsequently examined using the SEM at the surface, intergranular regions, and
(Fig. 31). The cracks were intergranular along laboratory-induced ductile fracture (Fig. 32).
A metallographic section (Fig. 33) through
the vertical cracks showed untempered marten-
site at the surface, a transition region of over-
tempered 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 heat-
affected zone and shows that a significant heat
event occurred.
Chemical analysis showed that the material
conformed to AMS 6411, with alloying ele-
ments at the top of the range increasing the
sensitivity to quench cracking. Hydrogen ana-
lysis indicated 0.8 ppm hydrogen. The levels of
Fig. 30 Arresting gear hook-point showing vertical cracking
hydrogen present and the high strain rate of
in the cable groove and evidence of localized heat-
ing (from the temper colors in the cable groove)
Fig. 31 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.
loading precludes hydrogen embrittlement as a result of transformational stresses from fric-

possible failure mechanism. tional heating during capture of the arresting
Based on the analysis, it was determined that cable by the hook-point. The mechanism is
the vertical cracking in the cable groove was the similar to quench cracking and was aggravated
Fig. 32 Exposed crack face showing two distinct regions on the crack face. A, region of incipient melting. B, intergranular
fracture
Fig. 33 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
by the higher-than-normal hardenability of the Rail steels have been documented to fail
alloy. Recommendations included the applica- because of TME (Ref 40). This was especially
tion of a different HVOF coating to better resist true of older rails manufactured in open-hearth
the frictional heating of the cable during carrier furnaces with high phosphorus content. This
arrestments. occurred because of slow cooling through the
Tempered martensite embrittlement 500 C (930 F) range or from isothermal hold-
(TME) may not be associated with impurity ing at 500 C. Figure 34 shows a representative
atoms segregating to prior-austenite grain boun- SEM fractograph of an Fe-0.26C-2.11Si-
daries. The most common factor in TME is 2.27Mn-1.59Cr wt% carbide-free bainitic rail
the formation of cementite during tempering steel that has been temper embrittled by heat
(Ref 35). When a given steel has a low impurity treatment at 500 C for 5 h (Ref 40).
content, the source of TME is the decomposition Temper embrittlement is only now becom-
of retained austenite during the second stage of ing understood with regard to its mechanism.
tempering. Thomas first proposed this mech- However, the conditions of temper embrittle-
anism (Ref 36). This was found when transmis- ment are well known (Ref 41, 42).
sion electron microscopy showed the presence Steels must be heat treated or cooled through
of thin regions of retained austenite between the range of 375 to 575 C (706 to 1070 F) in
martensite in as-quenched steels, which subse- order to become temper embrittled. Temper em-
quently transformed to cementite on tempering brittlement is typically detected by an increase
in the range of 230 to 470 C (450 to 700 F). in the ductile-to-brittle transition temperature.
The presence of phosphorus also plays a role. This is shown in Fig. 35 for AISI 3140 steel
If two steels are compared, one containing a temper embrittled by furnace cooling through
higher concentration of phosphorus, the steel the critical range and holding at 550 C
with the higher phosphorus content will have (1020 F) (Ref 35). The embrittlement reaction
poorer impact properties than an identical steel follows a typical C-curve, with the minimum in
with a lower phosphorus level. This will remain embrittling time at approximately 1 h at 550 C
true through the entire range of tempering tem- (1020 F) and several hundred hours at 375 C
perature up to approximately 500 C (932 F). (706 F) (Ref 43). By heating to approximately
The fracture mode is intergranular along prior- 575 C (1070 F), temper embrittlement is
austenite grain boundaries (Ref 37). It is likely reversible and can be eliminated after holding
that phosphorus is present at the prior-austenite for only a few minutes at temperature.
grain boundaries. It is only after cementite pre- For temper embrittlement to occur, specific
cipitates in the tempered martensite that TME is embrittling impurities must be present. These
fully present. Often, the presence of molybde- include antimony, phosphorus, tin, and arsenic.
num at concentrations up to approximately 0.5% Quantities of less than 0.01% are enough to
will reduce the effect of TME. cause temper embrittlement. For the most part,
On June 19, 1974, during a cold start after a simple plain carbon steels are not considered
long shutdown for repairing the Tennessee to be susceptible to temper embrittlement
Valley Authority Gallatin No. 2 unit, the inter-
mediate-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 Fig. 34 SEM fractograph of Fe-0.26C-2.11Si-2.27Mn-
1.59Cr wt% carbide-free bainitic rail steel that has
TME (Ref 39). been temper embrittled by heat treatment at 500 C for 5 h
as long as manganese concentrations are held must be high enough that the embrittling metal
to below 0.5%. Alloy steels containing chro- can melt. Lastly, tensile stresses must be present
mium and nickel are the most prone; however, as externally applied or internal residual stres-
additions of molybdenum at a concentration ses. Should any of these conditions not be met,
of up to 0.5% are effective in reducing the then it is unlikely that the steel will fail by liquid
susceptibility of these steels to temper embrit- metal embrittlement.
tlement. Liquid metal embrittlement has been known
Large forgings have been prone to temper to embrittle gun tubes. In 1977, during the
embrittlement because of the slow cooling that manufacture of a 105 mm M68 gun tube, lead
occurs during fabrication. These large forgings was electroplated to the tube and used as a
are also prone because of the operating lubricant during the autofrettage process. During
temperatures applied, especially in large turbine the postautofrettage thermal treatment, the lead
rotors. melted and embrittled the gun tube. A complete
Liquid Metal Embrittlement or Solid Metal transverse brittle failure occurred. The axial
Embrittlement. Exposure of steels to liquid tensile residual stresses from the autofrettage
metals has been observed to result in brittle process were adequate to completely fracture the
fracture along prior-austenite grain boundaries tube, even though the hoop stresses were much
(Ref 44). Steels may be embrittled by exposure greater (Ref 46).
to any of the low-melting metals shown in In another example, an ISO 8.8 low-alloy
Table 1 (Ref 45). Embrittlement occurs by steel bolt that was electroplated with cadmium
wetting of the prior-austenite grain boundaries was used for an extended time at an elevated
with a thin film of the molten metal. Usually, temperature of 230 C (455 F). The resulting
very low tensile stresses are required to fail parts failure showed intergranular fracture, with cad-
that are liquid metal embrittled. mium penetration along grain boundaries. This
In general, three conditions are necessary for cadmium penetration was detected by x-ray
liquid metal embrittlement. First, the embrittling diffraction (Fig. 36) (Ref 9).
metal must be present, either externally as a On April 28, 1997, United Flight 1210, a
coating or internally. Internal sources can Boeing 737222 equipped with Pratt and
include lead used to enhance machinability. Whitney JT8D-7B engines, experienced an un-
Second, temperatures that the part is exposed to contained 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 pas-
120
sengers were slightly injured during evacuation.
(160) Postincident examination of the engine revealed
100 that two-fragments (approximately 50 by
Energy absorbed, ft-lbf (J)
(130) 100 mm, or 2 by 4 in.) separated from the disk.

80
Examination of the disk revealed a 100 mm
(100) (4 in.) circumferential fracture around the dia-
60 meter, with three additional fractures emanating
(80) diagonally outward toward the rim. Also, cracks
40
(50)
Table 1 Melting temperatures of metals known
20
(30)
to embrittle high-strength steels
Melting temperature
0
Metal C F
Mercury 39 38
Gallium 29 85
Indium 156 313
Lithium 180 356
Cadmium 321 610
Tin 232 449
Lead 327 620
Fig. 35 Shift in ductile-brittle transition temperature curve to Zinc 419 787
a higher temperature for AISI 3140 steel by holding at Antimony 642 1187
500 C
and continuous cooling through the temper embrittle-
Source: Ref 45
ment critical range. Source: Ref 35
emanated radially outward from two tie-rod discrepant parts were segregated. A series of
holes, one of which bisected the fracture at the parts were then retempered to the specified
diameter. No cracks were detected on any of the hardness of HRC 46 from HRC 51. The lug
other high-pressure compressor disks in the radius was enlarged, and a new HVOF coating
engine. was applied. During field trials, multiple hook-
On further examination by the National points were identified by magnetic particle
Transportation Safety Board, the fractures were inspection as having cracks in the lug radius
found to have large intergranular areas in the (Fig. 37). No through cracks were found. The
steel compressor disk (Ref 47). Solid molten cracks were exposed, and a narrow uniform
cadmium was detected along the prior-austenite region of intergranular fracture (approximately
grain boundaries, indicative of liquid metal 200 mm) was observed (Fig. 38). Metallography
(cadmium) embrittlement. Inadequate nickel indicated that the microstructure was quenched
plating prestrike thickness was observed at the and tempered martensite, and the hardness was
surface of the disk. The failed disk was plated by within the specification of 46 to 48 HRC.
a trainee and inadequately plated with nickel. It Metallography revealed that no decarburization
was found that the nickel plating was approxi- or precipitates were found at the prior-austenite
mately 0.003 mm (0.00012 in.) in thickness, grain boundaries. Hydrogen analysis showed a
which is below the Pratt and Whitney specified concentration of 0.2 ppm of hydrogen. The low
thickness of 0.015 to 0.02 mm (0.0006 to concentration of hydrogen and the rapid rate of
0.0008 in.). This thickness was inadequate to loading eliminated hydrogen embrittlement as a
prevent migration of cadmium into the steel cause of cracking. Auger analysis of the grain
grain boundaries. boundaries within the intergranular region
This was not the first time that liquid metal showed the presence of cadmium at the grain
embrittlement occurred in a compressor disk. On boundaries. The concentration of cadmium also
July 23, 1990, the crew of a JT8D-9-equipped decreased as the grain boundaries were ion
Boeing 737100 reported that they heard a milled away. This analysis indicated that the
muffled explosion during climb, followed by fracture occurred because of liquid metal
a loss of rpm of the No. 1 (left) engine. The crew embrittlement or solid metal embrittlement.
returned to Houston, Texas, without incident. Solid metal embrittlement is similar to liquid
Engine examination revealed a failure of a metal embrittlement, except that temperatures
disk spacer due to liquid metal embrittlement are not high enough to cause melting of the
(Ref 48). cadmium. For cadmium, solid metal embrittle-
An AISI 4330V hook-point failed during field ment can occur at temperatures above 230 C
trials. This hook-point is used to grab the cable (450 F). A review of the planning showed that
on aircraft carriers and arrest the forward the work order release did not include removal
movement of the aircraft during landing. Pre- of the cadmium plating prior to retempering
vious hook-points failed because of excessive of the hook-points. Tempering to bring the hook-
hardness and high triaxial stresses during impact points to the proper hardness was above 320 C
loading. These hook-points were evaluated, and (610 F). Based on this, it was determined that
Fig. 36 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
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 manu-
facturing 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 hydro-
gen 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
Fig. 37 AISI 4330V cadmium-plated hook-point used to
the part to diffuse out. This is based on the
arrest landings of naval aircraft. Overall view of the
part, showing location of cracks observed using nondestructive study by Johnson, Morlet, and Troiano (Fig. 39)
testing
(Ref 59). This hydrogen embrittlement relief is
usually mandated whenever parts are plated,
cleaned, or exposed in some fashion to aqueous
the cracking observed in the hook-points was solutions such as coolants or acid. Alkaline
due to liquid metal embrittlement caused by solutions are not generally prone to causing
failure to remove the cadmium plating prior to hydrogen embrittlement.
tempering. Fasteners are prone to hydrogen embrittle-
Hydrogen embrittlement is a particularly ment. In this example, an ISO 10.9 low-alloy
insidious form of failure. Often, failure is de- steel bolt grade that was zinc plated failed during
layed for hours, months, and possibly years after service. Multiple fracture initiation sites were
the component has been fabricated. The results evident along the bolt head transition, with
may be catastrophic and unpredictable. The intergranular fracture morphology and heavy
failure mode is typically intergranular along secondary cracking. A hydrogen source was
prior-austenite grain boundaries (Fig. 21). suggested from manufacturing (pickling stage)
Hydrogen can come from either external or and/or cathodic hydrogen charging due to ano-
internal sources. One common source of dic zinc plating (Fig. 40) (Ref 9).
To verify that the baking process after chro- specimens. These test specimens were heat
mium plating was adequate, a plating shop tes- treated to 1515 MPa (220 ksi) and chromium
ted four chromium-plated 4340 notched tensile plated. During a sustained load test, one of the
Fig. 38 SEM examination of the hook-point showing a narrow region of intergranular fracture along prior-austenite grain boundaries
Normal Notch Strength = 300,000 psi (2070 MPa)

(2070) 300
Uncharged
(1895) 275
(1725) 250 Bake 24 hr

Applied Stress, 1000 psi (MPa)
(1550) 225 Bake 18 hr
(1380) 200
(1205) 175 Bake 12 hr
(1035) 150
Bake 17 hr
(860) 125
Bake 3 hr
(690) 100
Bake 0.5 hr
(520) 75
(345) 50
0.01 0.1 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
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
Evidence of light corrosion products was found

at the fracture origin of origin 2. SEM exam-
ination 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.
Fig. 41 As-received notched tensile specimen showing
A metallographic specimen was removed
location of fracture. Tensile specimen was fabricated
from 4340 steel, heat treated to 1515 MPa (220 ksi), and chro- from the largest origin location and examined
mium plated. (Fig. 47). The microstructure of the steel was
quenched and tempered martensite, typical of
300M heat treated to HRC 54 to 55. Chromium
specimens failed prematurely (Fig. 41). The plating was found to be intact at the fracture
fracture was located at the notch. The fracture origin.
surface (Fig. 42) was examined, and the origin Hydrogen analysis conducted on the jack pad
showed a shiny, faceted surface. At the origin, showed hydrogen concentrations of 4 and
the fracture was intergranular, while away from 6 ppm, which is considered adequate hydrogen
the origin, near the center of the fracture surface, to cause hydrogen embrittlement in 300M steel
the fracture mechanism was microvoid coales- heat treated to this hardness.
cence. Hydrogen analysis on the notched Based on this investigation, it was concluded
tensile specimen yielded an average hydrogen that the jack pad most likely failed from
concentration of 12 ppm hydrogen. This is hydrogen embrittlement.
considered very high and is sufficient to cause Stress-corrosion cracking is the attack of a
hydrogen embrittlement. Metallography of the material by the combined action of tensile stress
test specimen showed a normal quenched and on a part, either externally from an applied force
tempered microstructure, typical of a steel heat or internally from residual stresses, and a spe-
treated to this hardness. cific corrosive environment. Common features
During a routine wheel and tire change, a new are brittle fracture with little ductility, localized
jack pad for a military aircraft failed, causing an corrosive attack, and a specific environmental-
aircraft to drop prematurely onto the new wheel. alloy system. Failure by stress-corrosion crack-
No damage occurred to the aircraft. The jack pad ing (SCC) is characterized by exposure to a
was machined from 300M steel that was heat specific chemical environment and the simulta-
treated to HRC 54 to 55. The jack pad was neous application of a tensile stress. Without one
chromium plated. or the other, SCC will not occur. Fine cracks can
The as-received jack pad (Fig. 43) was penetrate deeply into the part without obvious
examined, and two fracture surfaces were iden- signs of attack. Impending failure can occur
tified (Fig. 44). These were identified as origins without warning.
1 and 2. Ridges emanated away from a distinct The applied tensile stresses can be from
origin location on each of the fracture surfaces. the service environment or from any of the
Fig. 43 As-received jack pad showing the locations of the

(a) 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 loca-

tion 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: Fig. 44 Fracture surfaces of the jack pad showing location of
1000 . (b) Microvoid coalescence at location B the origins. Original magnification: 2
numerous sources of residual stresses from the alloy, the heat treated condition, and the
manufacturing (thermal processing, machining, environment.
grinding, surface finishing, fabrication, or The site of initiation of SCC may be micro-
assembly). The tensile stress is important in scopic. This could be from local differences in
the rupture of any protective film during initia- metal composition or stress concentrations. A
tion and subsequent propagation of the crack. pre-existing mechanical flaw or discontinuity
There appears to be a threshold tensile stress may act as a stress raiser and serve as a site for
intensity, KISCC, below which SCC does not SCC initiation.
occur. This stress intensity is dependent on Stress corrosion cracking usually exhibits ex-
tensive branching and propagates in a direction
(a)
(a)
(b) (b)
(c) (c)
Fig. 45 SEM examination of origin 1. (a) Location of the Fig. 46 SEM examination of origin 2. (a) Location of the
fracture origin. Original magnification: 20 . (b) fracture origin. Original magnification: 100 .
Location A showing a region of intergranular fracture along prior- (b) Location C showing a region of intergranular fracture along
austenite grain boundaries. Original magnification: 1000 . (c) prior-austenite grain boundaries. Original magnification: 1000 .
Location B, at a distance away from origin 1, showing microvoid (c) Location D, at a distance away from origin 2, showing
coalescence. Original magnification: 2000 microvoid coalescence. Original magnification: 2000
perpendicular to the tensile stresses contributing It is often difficult to differentiate between

to propagation and initiation. However, this is SCC and hydrogen-induced damage solely from
not always the case. Structural steels exposed to the fracture surface. Fractures of both types
agricultural ammonia may exhibit nonbranched exhibit intergranular features and tend to follow
cracking. prior-austenite grain boundaries. Metallography
Stress-corrosion cracking has several special is important to determine if branched cracking
characteristics that differentiate it from other has occurred. Even so, the absence of branched
forms of cracking: cracking may not preclude SCC. In general, the
environment that the part was exposed to can be
Only certain specific environments for a
the deciding factor of whether it is SCC or
specific alloy system cause SCC. There is
hydrogen embrittlement (Fig. 48).
no general pattern regarding the corroding
Low-carbon steels generally become more
environments or alloy systems.
susceptible to SCC as the carbon concentration
Pure metals are much less susceptible to
increases. Decarburized steels and pure iron are
SCC.
resistant to SCC. Microstructure plays a greater
Cathodic protection has been successful in
role in susceptibility to SCC than does the
preventing the initiation of SCC or in stop-
alloying elements. High-alloy steels in a variety
ping the propagation of SCC.
of environments show that the heat treated
Addition of certain soluble salts effectively
strength of the alloy is more important than
can poison the environment and either re-
strictly the concentration. Steels that have been
duce or stop the propagation of SCC cracks.
heat treated to 1240 MPa (180 ksi) or higher are
Certain metallurgical features, such as grain
especially susceptible to SCC. Typical envir-
size, can influence the susceptibility of an
onments that can cause SCC in steels are shown
alloy system to SCC attack.
in Table 2.
Macroscopically, fractures produced by SCC Caustic cracking in boilers is a serious SCC
show little ductility and nearly always appear problem and has caused many failures in steam
brittle. The fracture surfaces usually contain re- boilers. These failures usually initiate in riveted
gions that are identifiable as the crack initiation and welded structures, where small leaks allow
site, slow crack propagation, and final failure. buildup of caustic soda and silica. Cracking is
The regions containing the slow propagation usually intergranular. Failures of this type have
often contain corrosion products or are occurred with concentrations of NaOH as low as
discolored. This region extends to the region of 5% in water. Failures take place when the
final fast fracture. However, this can also operating temperature is in the range of 200 to
be misleading, because the fracture could have 250 C (390 to 480 F). The concentration of
corroded before inspection, or the environment NaOH needed to cause cracking initiation
may not be conducive to straining the fracture. 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 enhan-
ces 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.
Fig. 47 Micrograph showing quenched and tempered mar-
Halide-containing environments, such as
tensite, typical of 300M heat treated to HRC 54 to 55.
Note that the chromium plating is intact. seawater, are particularly severe for alloy steels
Direction of advancing cracking into metal
H+ + e H
++
M
++ +
M M+
Anodic Stress
Corrosion Cracking
Time to Cracking Hydrogen
Embrittlement
Region of Anodic Region of Region of

Stress Corrosion Cracking Immunity Hydrogen Embrittlement
Anodic Current Cathodic Current

M M++ + 2e 2e + 2H+ 2H
Fig. 48 Schematic differentiation of anodic stress-corrosion cracking and cathodic hydrogen embrittlement
Table 2 Environments that produce in preventing SCC in steels such as 300M

stress-corrosion failures in carbon and or 4340.
low-alloy steels On August 22, 2003, an empty cargo tanker
Type of pulled upto a tank containing anhydrous
fracture
Medium (a) Comments ammonia. Approximately 1 hour after being
Aqueous chloride I,T Prevalent in high-strength steels heat filled, the front head cracked open (Fig. 49)
environments treated to 1380 MPa (200 ksi) or and started to release anhydrous ammonia.
greater Approximately 100 workers were evacuated
Caustic solutions I Well known as caustic embrittlement
Nitrates I Examples of bridge cable failures in from the building. Five people were treated
ammonium nitrate or sodium for inhalation injuries and released. The
nitrate solutions
HNO3 I ...
cost to repair the trailer was approximately
HCN I ... $25,000.
Seacoast and I High-strength steels heat treated to Examination revealed a 40 cm (16 in.) long
industrial 1380 MPa (200 ksi) or greater are
environments especially prone. through-wall crack next to the radial weld in the
Water, humid air, I High-strength steels heat treated to front head at the 1 oclock position (Fig. 50).
and gas 1380 MPa (200 ksi) or greater are Internal examination using magnetic particle
especially prone.
inspection found two additional cracks that had
(a) I, intergranular failure; T, transgranular failure
not yet propagated through the wall of the tan-
ker. SEM examination of the cracks (Ref 60)
heat treated to above 1380 MPa (200 ksi). found that the fracture was branched and inter-
The use of cadmium plating, low-hydrogen granular, with extensive surface corrosion on the
practices, and adequate baking are helpful crack faces.
In the 1950s, the Agricultural Ammonia regulation further required tankers to be

Institute determined that caustic cracking of placarded with signs indicating QT or NQT,
ammonia-containing tanks was the reason that a for quenched and tempered or not quenched and
number of carbon steel tanks had failed (Ref 61). tempered.
They further determined that the addition of The National Transportation Safety Board
0.1% water to anhydrous ammonia inhibited determined that the failure of the tank and the
SCC in carbon steel. The committee recom- subsequent release of anhydrous ammonia were
mended that at least 0.2% water be added to due to caustic cracking (SCC) of the tank from
inhibit cracking. Further cracking occurred in the transport of anhydrous ammonia containing
the 1960s in quenched and tempered ASTM less than 0.2% water.
A517 steel, because purity levels had increased A Boeing 7572008 was parked at a gate at
and water was no longer being added. In 1975, Copenhagen, Denmark, and boarding of pas-
the Department of Transportation adopted regu- sengers was nearly completed when the right-
lations (Title 49 Code of Federal Regulations, hand main landing gear truck beam failed (Ref
Parts 171 to 180) that required cargo tanks 62). As the beam failed, the right side of the
fabricated from quenched and tempered steel aircraft rested on the shock strut instead of on the
should only be used for anhydrous ammonia wheels. Figure 51 shows the failed truck beam
if the solution contained 0.2% water. The and the aircraft resting on the shock strut. A
Fig. 49 Accident cargo tank with QT designation, which indicates quenched and tempered steel
sketch of the main landing gear assembly of a are commonplace in everyday living. Examples
B-757 is shown in Fig. 52. The fracture surface, include pipes bursting in the middle of winter,
showing evidence of corrosion, is shown in the expansion of a bridge in the middle of sum-
Fig. 53. mer, and the sagging of a fireplace grate. Each of
Metallurgical analysis indicated that the these examples is an indication that properties
fracture mode was due to SCC. Examination of change with temperature. In addition, the pre-
the finish on the inner diameter showed that the vious discussion indicated that steels become
plating on the inside diameter was thin or non- more brittle as the temperature is decreased.
existent, and that it did not receive the required There are many other effects of temperature that
shot peening. Because the truck beam was have been cited (Ref 63). Even the concept of
overhauled and the original plating was retained elevated temperature is relative (Ref 64). What
and worn during service, it was likely that the is considered hot for one material may be con-
overhaul was inadequate or improper. This had sidered cold for another; for example, gallium
the result of minimal cadmium protection on the has a melting point of 30 C, while tungsten has
inner diameter surface of the truck beam. Sub- a melting point of approximately 3400 C.
sequent loss of the plating led to premature and Creep is the continuous deformation of a
severe corrosion in service and eventual fracture material as a function of time and temperature.
due to SCC. This topic is treated very thoroughly in Ref 65.
Creep Rupture The effects of temperature The creep of a material is shown in Fig. 54. It can
on mechanical properties and material behavior 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 steady-
state 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 Fig. 52 Schematic of the assembly of a Boeing 757 main
Icelandic Air, aircraft registration TF-FIJ. landing gear showing the location of fracture
exposed, the surrounding atmosphere, and the testing, the time to failure is measured at a con-
prior strain history. Andrade and Chalmers (Ref stant stress and constant temperature. This test
66) were pioneers in the study of creep and has gained acceptance for elevated-temperature
proposed that creep followed the equation: testing of turbine blade materials in jet engines.
Using a tensile machine and high-temperature
e=e0 (1+bt1=3 )ekt furnace, the strain is measured in creep testing
by special extensometers suited for elevated
where b and k are material constants that can be temperatures. In stress-rupture testing, a simple
evaluated by several different methods (Ref 67). apparatus such as a dial calipers is used, since
A better fit for the creep of materials was only the overall strain at constant time and
proposed by Garofalo (Ref 68). He indicated temperature is needed.
that:
Fatigue
de
e=e0 +et (17e7n )+ t Parts are subject to varying stresses during
dt service. These stresses are often in the form of
where de/dt is the steady-state creep rate, e0 is repeated or cyclic loading. After enough appli-
the strain on loading, n is the ratio of the tran- cations of load or stress, the components fail at
sient creep rate to the transient creep strain, and stresses significantly less than their yield
et is the transient creep strain. strength. Fatigue is a measure of the decrease in
Very early, it was recognized that fractures resistance to repeated stresses.
at elevated temperatures occurred along grain Fatigue failures appear brittle, with no gross
facets (Ref 69). In stage III creep, intergranular deformation. The fracture surface is usually
wedge cracks and cavities form. Wedge-shaped normal to the main principal tensile stress. Fati-
cracks and creep cavities usually initiate at or gue failures are recognized by the appearance of
near grain-boundary triple points and propagate a smooth, rubbed type of surface, generally in a
along grain boundaries normal to the applied semicircular pattern. The progress of the fracture
tensile stress. Creep cavities form at higher (and crack propagation) is generally suggested
temperatures and lower working stresses. These by beach marks. This is illustrated in Fig. 56
structural features are shown in Fig. 55. and 57. The initiation site of fatigue failures is
Creep testing is usually performed for 1000 generally at a stress-concentration site or stress
to 10,000 h with strains of up to 0.5%. Stress- raiser. A typical fracture appearance is shown
rupture testing, or testing to failure, uses much schematically in Fig. 58.
higher loads and temperatures, and the test is Three factors are necessary for fatigue to
usually terminated after 1000 h. In stress-rupture 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. 53 Fracture surface of Boeing 757 main landing gear

truck beam on Icelandic Air aircraft TF-FIJ Fig. 54 Schematic representation of creep
fatigue life of a component is affected by a num- Fatigue usually occurs at a free surface, with
ber of variables, including stress concentration, the initial features of stage I growth, fatigue
corrosion, temperature, microstructure, residual cracks, being initiated at slip band extrusions and
stresses, and combined stresses. intrusions (Ref 71, 72). Cottrell and Hull (Ref 73)
The structural features of fatigue failures are proposed a mechanism for the formation of these
generally divided into four distinct areas extrusions and intrusions (shown schematically
(Ref 70): in Fig. 59) that depends on the presence of slip,
with slip systems at 45 angles to each other
Crack initiation, the early development of
operating sequentially on loading and unloading.
fatigue damage
Wood (Ref 74) suggested that the formation of
Slip band crack growth, the early stages of
the intrusions and extrusions was the result of
crack propagation. This is often called stage fine slip and buildup of notches (Fig. 60). The
I crack growth.
notch created on a microscopic scale would
Stable crack growth, which is usually normal
be the initiation site of stable fatigue crack
to the applied tensile stress. This is called growth.
stage II crack growth.
In stage II, stable fatigue crack growth, stria-
Unstable crack growth, with final failure
tions (Fig. 61) often show the successive posi-
from overload. This is called stage III crack
tion of the crack front at each cycle of stress.
growth.
Fatigue striations are usually detected using
electron microscopy and are visual evidence that
fatigue occurred. However, the absence of fati-
gue 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
Fig. 55 Creep cavities and creep wedges forming at grain crack tip. When the load is released, the slip
boundaries direction reverses, and the crack tip is com-
pressed and sharpened. This provides a sharp
notch at the new crack tip where propagation
Fig. 56 Actual fatigue failure of a crankshaft showing

characteristic beach marks. Fatigue initiated at the Fig. 57 Fatigue failure of a fastener, with initiation of fatigue
radius of the journal and exhibits classic bending fatigue. occurring at the threads
Fig. 58 Schematic illustration of simple fatigue failures

can occur. This is shown schematically in fatigue striations form by the plastic blunting
Fig. 62. process.
An alternative hypothesis on striation forma- It has also been found that the thicker the
tion was presented by Forsyth and Ryder testpiece, the faster the crack propagation rate
(Ref 76). In their model, the triaxial stress state (Ref 81). It is likely that the propagation rates for
at the crack tip forms a dimple ahead of the crack thicker pieces are due to increased plane-strain
front. The material between the crack tip and the conditions, with a small plastic zone at the crack
dimple contracts and eventually ruptures, tip. Since there is a greater stress gradient for
forming a fatigue striation. This is shown sche- a small plastic zone, a faster crack propagation
matically in Fig. 63. rate may be expected. Also, in thicker
In mild steel, well-defined striations are panels there is a higher state of triaxial stress,
observed but not as well defined or as specta- which would also tend to increase crack growth
cular as in aluminum. This was first assumed rates.
to be due to the crystal lattice structure, since Since fatigue failures usually begin at the
face-centered cubic austenitic steels show well- surface, the surface condition is very important.
defined striations, and mild steels (base- Surface roughness is a primary factor influen-
structured) do not (Ref 77). Other alloys, such as cing fatigue. Highly polished specimens exhibit
titanium alloys, with a hexagonal close-packed the longest fatigue life, with increasingly
crystal structure show very defined striations rougher surfaces yielding decreased fatigue life.
(Ref 78). However, aluminum alloys (body- Rough lathe or coarse grinding reduces the
centered cubic) show strongly defined striations fatigue strength by approximately 20% below
(Ref 79). Therefore, attributing defined stria- polished specimens (Ref 82). Electropolished
tions to crystal lattice alone was discounted as specimens have lower fatigue limits than
a viable theory. mechanically polished specimens, by up to 25%
Deformation and available slip systems were (Ref 83). This reduction is due to the removal of
presumed to be more significant (Ref 80). surface compressive residual layers induced
However, this does not follow, because mild during mechanical finishing.
carbon steels are more ductile than austenitic An example of a typical fatigue failure in an
steels. It is now generally accepted that 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
Fig. 59 Schematic representation of the mechanism of fati-
gue intrusions and extrusions 2250 arrestments. The part is designed to last
Fig. 60 Mechanism of intrusions and extrusions. PSB, persistent slip bands. Source: Ref 74
two lifetimes, or 4500 arrestments, without diameter. The bore is then injection drilled to the
cracking (0.25 mm, or 0.010 in. detectible final dimensions. A follower supported the
flaw). The arresting shank was fatigue tested in a injection drill. This is not a method that is
fixture, with hydraulic cylinders providing loads commonly used for final machining operations.
at the vertical damper and hook-point cable It is heat treated in vacuum to 1930 MPa
groove. The maximum applied load was 90 mg (280 ksi) ultimate tensile strength. The part is
(200,000 lb). A schematic of the arresting hook inspected using dye-penetrant and magnetic
shank is shown in Fig. 65. particle nondestructive testing methods. The
The arresting hook shank was fabricated from bore is visually inspected using a bore scope.
an AerMet 100 rotary forging. It is rough turned This is a difficult inspection because of the long
on a lathe on the outside, then gun drilled to length and narrow bore.
create a pilot hole down the length of the for- Examination of the fracture surface showed
ging. The outer surface is turned to the final that cracking initiated at the hook-point side,
on the inner diameter, at a location approxi-
mately 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 well-
defined surface layer was observed. This layer
Fig. 61 Typical fatigue striations in 7075 aluminum 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

Fig. 62 Mechanism for fatigue striation formation ahead of a crack front
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 Schematic of the failed arresting hook shank showing location of loads
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
Fig. 66 Machining marks found on the inside of the bore, at the chips and bore surface can occur. As the
the origin of cracking cutter boring bar moves, pullout can occur. The
Fig. 68 Metallography of the arresting hook shank. (a) Typi-

cal 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

Summary
striations emanating from the fracture origin. (b)
Machining marks found on the surface of the inner bore. (c) Well-
defined layer showing fatigue emanating from the damaged In this short overview of the possible
material at the surface of the inner bore
mechanisms of failure for steels, the following
were discussed:
Techniques for examining fractures
examined flaws matched the machining feeds Ductile and brittle failures
and speeds. Intergranular failure mechanisms
The arresting hook failed by fatigue, Fatigue
initiating at flaws created during the final
machining process. The defect morphology The previous discussion has shown that it is
suggested localized solid-state welding and important to look at not only the fracture surface
pullout from chip contact with the freshly but at all the factors (manufacturing history,
machined surface. The surface roughness and service conditions, and loading). All the tools
finish of the inner bore did not meet drawing available to the metallurgist should be used
requirements. these include photography, fractography, and
metallographyto understand the sources and 15. Aircraft-Quality Steel Cleanliness: Mart-
root cause of failure. ensitic Corrosion-Resistant Steels Magnetic
Particle Inspection Procedure, AMS 2303
16. Special Aircraft-Quality Steel Cleanliness:
REFERENCES Magnetic Particle Inspection Procedure,
AMS 2304
1. Hippocrates, Epidemics, Book I, Section XI 17. R. Kiessling and N. Lange, Non-Metallic
2. D.S. MacKenzie, G. Chen, M. Barker, Inclusions in Steel, Part IIV, Iron and Steel
C. Ramsay, J. Alderson, L. Dharani, and Institute, London, 1978
J. Yu, Forensic Investigation of Failed Mast 18. S. Maropoulos and N. Ridley, Inclusions
Arms of Traffic Signal Supported Struc- and Fracture Characteristics of HSLA Steel
tures, Transport. Res. Record, Des. Struc. Forgings, Mater. Sci. Eng. A, Vol 384 (No.
2002, 2002, p 916 12), Oct 25, 2004, p 6469
3. Loss of Control and Impact with Pacific 19. Railway Investigation Report, Derailment
Ocean Alaska Airlines Flight 261 McDon- and Collision Canadian National Train
nell Douglas MD-83, N963AS about 2.7 No. U-783-21-30 and Train No. M-306-31-
Miles North of Anacapa Island, California, 30 Mile 50.84, Saint-Hyacinthe Subdivision
January 31, 2000, NTSB/AAR-02/01, Mont-Saint-Hilaire, Quebec, December 30
National Transportation Safety Board, 1999, Report R99H0010, Transportation
Adopted Dec 30, 2002, Notation 7263E Safety Board of Canada
4. Uncontained Engine Failure/Fire Valujet 20. M.L. Williams, STP 158, ASTM, 1954,
Airlines Flight 597 Douglas DC-9-32, p 1144
N908VJ, Atlanta, Georgia, June 8, 1995, 21. M.E. Shank, STP 158, ASTM, 1954, p
NTSB/AAR-96/03, National Transporta- 45110
tion Safety Board, Adopted July 30, 1996, 22. Fractography, Vol 12, Metals Handbook,
Notation 6579A 9th ed., ASM International, 1987
5. W.M. Garrison and N.R. Moody, J. Phys. 23. J. McCall and P. French, Metallography in
Chem. Solids, Vol 48, 1987 p 1035 Failure Analysis, Plenum, New York, 1978,
6. J.F. Knott, Advances in Fracture Research, p6
Proc. Seventh International Conference on 24. Standard Test Methods for Notched
Fracture (Oxford, U.K.) 1989, p 125 Bar Impact Testing of Metallic Methods,
7. J.F. Knott, Fracture 1977, Vol 1, Proc. E2388, American Society for Testing and
Fourth International Conference on Frac- Materials
ture, June 1977 (Waterloo Canada), p 61 25. Standard Test Method for Plane Strain
8. H.G. Wilsforf, Mater. Sci. Eng., Vol 59, Fracture Toughness, KIc, of Metallic Mate-
1983, p 1 rials, E39974, American Society for
9. B. Bgner, G. Rrvik, and L. Marken, Bolt Testing and Materials
FailuresCase Histories from the Nor- 26. Standard Test Method for Conducting
wegian Petroleum Industry, Microsc. Micro- Drop-Weight Test to Determine Nil-
anal., Vol 11 (Suppl. 2), 2005, p 16041605 Ductility Transition Temperature of Ferritic
10. T.L. Anderson, Fracture Mechanics, CRC Steels, E20895A, American Society for
Press, New York, 1995 Testing and Materials, 1983
11. C. Ruggieri, Numerical Investigation of 27. Standard Test Method for Dynamic Tear
Constraint Effects on Ductile Fracture in Testing of Metallic Materials, E60483,
Tensile Specimens, J. Braz. Soc. Mech. Sci. American Society for Testing and Materi-
Eng., Vol 26 (No. 2), 2004 als, 1983
12. Standard Test Methods for Determining 28. J.M. Kraft, A.M. Sullivan, and R.W. Boyle,
the Inclusion Content of Steel, E4505, Proc. Symp. Crack Propagation, Cranfield,
ASTM International U.K., 1961
13. Premium Aircraft-Quality Steel Clean- 29. W.F. Brown and J.E. Strawley, STP 381,
liness: Magnetic Particle Inspection Proce- American Society for Testing and Materi-
dure, AMS 2300 als, 1965, p 133
14. Aircraft-Quality Steel Cleanliness: 30. A.S. Tetelman and A.J. McEvily, Fracture
Magnetic Particle Inspection Procedure, of Structural Materials, Wiley, New York,
AMS 2301 1967
31. J.F. Knott, Fundamentals of Fracture Benet Laboratories, Watervliet Arsenal,

Mechanics, Wiley, New York, 1973 June 1999
32. A.H. Cottrell, Proc. R. Soc. A, Vol 276 47. NTSB CHI97IA117, National Trans-
(No. 1), 1963 portation Safety Board, adopted June 28,
33. A.H. Knott, Mater. Sci. Eng., Vol 7 (No. 1), 1998
1971 48. NTSB Safety Recommendation, in memo to
34. J.S. Colton, Class notes, Great Boston J. Garvy, Administrator A-97-83, FAA,
Molasses Disaster, Jan 15, 1919, ME6222 Aug 29, 1997
Manufacturing Processes and Systems, 49. S.F. Carter, Effect of Melting Practice on
Georgia Institute of Technology Hydrogen, J. Met. (AIME), Vol 188,
35. G. Krauss, Steels: Heat Treatment and Jan 1950, p 3040 and J. Met. (AIME), Vol
Processing Principles, ASM International, 188, Feb 1950, p 245255
1990, p 231, 237 50. F.B. Foley, Flakes and Cooling Cracks
36. G. Thomas, Retained Austenite and Tem- in Forgings, Met. Alloys, Vol 12, 1940,
pered Martensite Embrittlement, Metall. p 442445
Trans. A, Vol 10, 1979, p 16431651 51. F.J. Shortsleeve and A.R. Troiano, A Lit-
37. S. Banerji, C. McMahon, and H. erature Survey on the Mechanism of Flake
Feng, Intergranular Fracture in 4340- Formation, Interim Technical Report 1,
Type Steels: Effects of Impurities and Contract DA 33-019-ORD-897
Hydrogen, Metall. Trans. A, Vol 9, 1978, 52. C.A. Edwards, Pickling or the Action
p 237247 of Acid Solution on Mild Steel and the
38. D. Kalderon, Steam Turbine Failure at Diffusion of Hydrogen Through the Metal,
Hinkley Point A, Proc. Inst. Mech. Eng., J. Iron Steel Inst., Vol 110, 1924, p 960
Vol 186, 1972, p 341377 53. C.A. Zapffe and M.E. Haslem, Acid Com-
39. L.D. Kramer and D. Randolph, Analysis position, Concentration Temperature and
of TVA Gallatin No. 2 Rotor Burst: Part Pickling Time as Factors in the Hydrogen
1Metallurgical Considerations, Proc. Embrittlement of Mild Steel and Stainless
1976 ASME-MPC Symposium on Creep- Steel Wire, Trans. ASM, Vol 39, 1947,
Fatigue Interaction, p 124 p 213236
40. H.K.D.H. Bhadeshia, High Performance 54. J.R. Gustafson, Some of the Effects of
Bainitic Steels, Mater. Sci. Forum, Vol 500, Cadmium, Zinc, and Tin Plating on Springs,
2005, p 6374 Proc. ASTM, Vol 47, 1947, p 782
41. I. Olefjord, Temper Embrittlement, Re- 55. K.B. Valentine, Stress Cracking of
view 231, Int. Met. Rev., Vol 23, 1978, p Electroplated Lockwashers, Trans. ASM,
149163 Vol 38, 1947, p 488504
42. B. Woodfine, Temper Brittleness, A Critical 56. C.A. Zapffe and M.E. Haslem, Measure-
Review of the Literature, JISI, Vol 173, ment of Embrittlement during Chromium
1953, p 229240 and Cadmium Electroplating and the Nature
43. F. Carr, M. Golman, L. Jaffee, and D. Buf- of Recovery of Plated Articles, Trans. ASM,
fum, Isothermal Temper Embrittlement of Vol 39, 1947, p 241258
SAE 3140 Steel, Trans. TMS-AIME, Vol 57. G. Nagu and T. Namboodhiri, Effect of
197, 1953, p 998 Heat Treatments on the Hydrogen Embrit-
44. Liquid Metal Embrittlement, Failure Ana- tlement Susceptibility of API X-65 Grade
lysis and Prevention, Vol 10, Metals Line-Pipe Steel, Bull. Mater. Sci., Vol 26
Handbook, 8th Ed., American Society for (No. 4), June 2003, p 435439
Metals, 1975, p 228229 58. A. Thompson and I. Berstein, Ed., Effect
45. P. Fernandez, R. Clegg, and D. Jones, of Hydrogen on Behavior of Materials,
Failure by Liquid Metal Induced Embrit- TMS-AIME, 1976
tlement, Eng. Fail. Anal., Vol 1 (No. 1), 59. H. Johnson, J. Morlet, and A. Troiano,
1994, p 5163 Hydrogen Crack Initiation and Delayed
46. G. Vilganlte, G. Troiano, and C. Mossey, Failure in Steel, Trans. TMS-AIME, Vol
Liquid Metal Embrittlement of ASTM 212, 1958, p 528536
A723 Gun Steel by Indium and Gallium, 60. Release of Hazardous Materials from
ARCCB-TR-99011, Army Research, Cargo Tank in Middletown, Ohio, on August
Development and Development Center, 22, 2003, Accident DCA03MZ002, NTSB
Hazardous Materials Accident Brief, 73. A.H. Cottrel and D. Hull, Proc. R. Soc.
Adopted July 22, 2004 (London) A, Vol 242A, 1953, p 211
61. A.W. Loginow, Stress-Corrosion Cracking 74. W.A. Wood, Bull. Inst. Met., Vol 3, 1955,
of Steel in Liquefied Ammonia Service, p5
Bull., Natl. Board Boiler Pressure Vessel 75. C. Laird, Fatigue Crack Propagation, in
Insp., Vol 45 (No. 6), Oct 1988 STP 415, American Society for Testing and
62. Report by the Danish Accident Investiga- Materials, 1967, p 136
tion Board into the Incident to Boeing 757, 76. P.J. Forsyth and D.A. Ryder, Metallurgica,
TF-FIJ at Kastrup Airport, Denmark on Vol 63, 1961, p 117
June 28, 2001, M-04001/AIG-09, Danish 77. G. Jacoby, Current Aeronautical Fatigue
Accident Review Board Problems, J. Schijve, Ed., Pergamon, New
63. J.E. Dorn, Ed., Mechanical Behavior of York, 1965, p 78
Materials at Elevated Temperature, 78. W.R.Warke and J.M. McCall, Fracto-
McGraw-Hill, New York, 1961 graphy Using the Electron Microscope,
64. R.W. Guard, Prod. Eng., Vol 27 (No. 10), ASM Technical Report We-2-65, American
1956, p 160174 Society for Metals, 1965
65. I. Finnie and W.R. Heller, Creep of Engi- 79. G. Jacoby, Fractographic Methods, Exp.
neering Materials, McGraw-Hill, New Mech., 1965, p 65
York, 1959 80. P.J. Forsyth, A Two Stage Process
66. E.N. da C. Andrade and B. Chalmers, Proc. of Fatigue Crack Growth, Symp. Crack
R. Soc. (London) A, Vol 138, 1932, p 348 Propagation, Vol 2, Cranfield, U.K., 1961,
67. J.B. Conway, Trans. Metall. Soc. AIME, p 76
Vol 223, 1965, p 2018 81. D. Broek and J. Schijve, The Influence
68. F. Garofalo, Properties of Crystalline of Sheet Thickness in the Fatigue Crack
Solids, STP 283, ASTM, 1965 Propagation in 2024-T3 Alclad Sheet
69. W. Rosenhahn and D. Ewen, J. Inst. Met., Material, NLR Technical Report M2129,
Vol 10, 1913, p 119 Amsterdam, 1963
70. W.J. Plumbridge and D.A. Ryder, Metall. 82. N.E. Frost, K.J. Marsh, and L.D. Pook,
Rev., Vol 14, 1969, p 136 Metal Fatigue, Oxford University Press,
71. P.J. Forsyth and C.A. Stubbington, J. Inst. London, 1974
Met., Vol 83, 1955, p 395 83. T.T. Oberg and E.J. Wad, Technical Note
72. W.A. Wood, Some Basic Studies of Fatigue Report 56-289, Wright Air Development
in Metals, Wiley, New York, 1959 Department, 1956
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_87-109.pdf/Chap_02b/ 21/8/2008 10:50AM Plate # 0 pg 87
Mechanisms and Causes of Failures

in Heat Treated Steel Parts
Debbie Aliya, Aliya Analytical, Inc.
THE TERM mechanism of failure means If the part does not meet the specification for
different things to different people. One possible mechanical or physical characteristics after heat
definition refers to a particular product appli- treating, it may be the heat treaters fault.
cation, a particular type of component, and a However, there may have been something
particular type of industry where certain envir- wrong with the raw material, or prior manu-
onmental conditions are common. In many facturing processes that allowed a part that went
cases, these mechanisms may apply only to a through the normal heat treating and inspection
limited type of material or an alloy family. In process to have substandard properties, which
order to understand particular failure mechan- would not be the heat treaters fault after all.
isms, it is important to understand the causes of This chapter gives some examples of lack of
failure and categories of damage. This chapter conformance to specification that may at first
reviews various ways to classify failure cate- look like the heat treater did something wrong,
gories. Information on mechanisms of damage, but where other contributing factors made it
in particular environment-material-type pairs, is difficult or impossible for the heat treater to meet
pro-vided toward the end of this chapter. the specification.
The term damage is often preferable to the This chapter also summarizes the basic types
term failure, because damage is a technical term of damage, with particular consideration given
that is very clear and has a specific meaning, to whether their likelihood can or cannot be
generally related to the physical condition of the influenced by the heat treating process. The
component. The term failure, on the other hand, classical organization for types of damage
has many more philosophical connotations. It is (failures) is as follows: deformation, fracture,
possible to see at once, perhaps, that the part is wear, corrosion or other environmental damage,
damaged. Only at the end of the investigation and multiple or complex damage. Separately
does one have a good chance of knowing with from what the damage type is, one also should
a high degree of engineering certainty whether look at the potential causes, sources, or factors
the part itself failed, rather than the design, the promoting the damage.
design system, the employee training system, Failure analysts used to be taught to classify
the procurement system, or the user-certification failures as a result of defects or abuse. There is
system. Thus, it is important to avoid using the still a large amount of literature that presents
term failed part until one has truly determined failure causes in such binary terms. By limiting
that the part was the problem. the analysis to one of two possible causes,
Often, one of the important goals in a failure opportunities may be missed for improving
analysis of a heat treated part is to determine the product. There are actually several
whether the damage is the result of improper different ways to classify causes of failures and
heat treatment, that is, the heat treaters fault. damage.
For many people in manufacturing, especially if This chapter also describes a process that can
an independent or job shop heat treater is be used to demonstrate likelihood that a product
involved, it is easy to blame the heat treater. was abused. Sometimes, the physical evidence
After all, he was the last one to touch the part speaks clearly. If this is not the case, it may be
before assembly, and most people in the general necessary to quantitatively prove, for example,
manufacturing arena understand heat treating that the part was overloaded. This may be dif-
poorly, if at all. ficult to do, because exact service conditions are
frequently impossible to determine with any cannot be created or facilitated as a result of a

degree of certainty. problem in the heat treating process.
Besides service or maintenance abuse, four
potential sources or origins of damage and fail-
ure relate to the product life cycle, as follows: Deformation
Figure 1 shows the result of a major defor-
Design process omissions
mation event. Note the image of a cylindrical
Undesired raw material characteristics
structure. There is a dark sign-wave-shaped
Undesired component manufacturing char-
band where the material is crumpled. The weight
acteristics (includes heat treating)
of the structure above that area created the force
Improper service or maintenance conditions
that caused the cylinder to suddenly deform.
In this chapter, a modified classification is There were also some tears or fractures as well,
developed specifically for failure analysis of but the main visible damage type is deformation.
heat treated parts where the heat treating is There are different ways to categorize defor-
suspected to be the cause of the failure. The first mation. One way is to compare gradual to sud-
logical potential source of poor component den deformation. Gradual deformation can
performance is raw material characteristics. occur when something is loaded and the load is
Note the use of the word characteristics rather sustained. Due to the sustained load, the struc-
than defects. This is an important thing to keep in ture can stretch or bend. This type of deforma-
mind when performing this type of work. The tion can happen during heat treating. Imagine a
term defect has a specific legal definition that part that has a protrusion that is not supported by
does not necessarily apply to all cases of phy- an appropriate fixture. The part is heated to red
sical flaws or suboptimal material properties. heat. Depending on the particular configuration
Thus, the use of the word defect may have and the presence of residual stresses in the
undesirable cascading consequences, especially component, the protruding feature may droop
when personal injury or large financial losses due to gravity, or it may change its shape in some
were a result of the component malfunction. other way to relieve residual stresses from ear-
The second potential source of poor perfor- lier parts of the processing. It is important to
mance is undesirable component characteristics. consider this type of deformation in a compo-
For example, castings, forgings, and machined nent that has a critical dimensional envelope. If
or molded components may have discontinuities the part goes outside of the specified or required
or microstructural features present that make it envelope and is placed into service, the stresses
appear to the casual observer that there was a that are experienced by the dimensionally non-
heat treating problem. These discontinuities can conforming component may be quite different
contribute to a poor heat treating outcome, a from the original design intent.
poor service outcome, or both. The third poten- Gradual onset deformation may also happen
tial trouble spot is design characteristics, many in parts that are loaded near the elastic limit.
of which may fall under the subclassification of It is important to note that the published elas-
inadequate attention to detail. The fourth cate- tic limit for different materials is a numerical
gory of causes of failure in this scheme is true quantity that is determined by using a specific set
heat treating process problems. Note that in this of test parameters, usually of relatively short
classification, these are, so far, all things that can duration. Long times at a stress level that may be
go wrong in the engineering and manufactur-
ing of the component. Finally, service and/or
maintenance abuse can be considered.
Types of Damage and Failure
The four basic types of damage (deformation,

fracture, wear, and corrosion) are briefly re-
viewed in this section, and discussions address
Fig. 1 Example of a sudden deformation event due to
whether each of the four types of damage can or buckling
Mechanisms and Causes of Failures in Heat Treated Steel Parts / 89
considered to be below the yield strength materials, where differences can occur due to
may actually cause unacceptable permanent preferred crystal orientation. This is less likely
deformation. Since the yield strength or elastic in heat treated steel than in steel used in the as-
limit can be affected by heat treating, it is poss- worked condition. In general, in most heat
ible, although probably not a frequent cause, treated steels, the modulus of elasticity will vary
that this type of service-related gradual onset little. In attempting to determine the predomi-
deformation is also related to a heat treating nant factor or predominant cause in a sudden
problem. Gradual onset deformation could also onset deformation, one should look first at the
be a result of loading the part beyond its design geometry and the applied loading. Heat treating
intent or actual physical limitations. issues would be considered in puzzling cases
The alternative to gradual onset of deforma- that did not lend themselves to ready analysis.
tion is sudden onset. Examples of sudden onset For gradual onset deformation, the geometry
deformation are buckling instability in com- and the elastic modulus are still important, but
pression of columns or torsion of tubes. A roof the heat treating can have a much more sig-
truss that collapses under a snow load is such an nificant effect. If the yield strength is not as high
example. Roof trusses are not usually heat as it is supposed to be, unacceptable levels of
treated for strength, at least not separately from gradual onset deformation may occur. This
the thermal element of the hot rolling. However, concludes the introduction on how to think about
it is possible that annealed steel may be used in deformation failures and whether the heat treater
some structural components. Most sudden onset even needs to be involved.
damage is primarily related to the basic geo- Figure 1 shows what was reported to be a
metry and modulus of elasticity, which is not a sudden onset event. This is actually an example
strong function of any heat treating process. of a complex failure or damage mode, because
Thus, heat treating problems are generally a this structure was standing for quite a few
minor or insignificant factor in most sudden years, and then one day, the wind blew and
onset damage events. the damage happened. Investigators found evi-
Another way to classify deformation is by dence of long-term corrosion on the inside,
level or degree: elastic (in other words, if the associated with significant wall thinning in some
load is removed, deformation is relieved) or areas.
plastic (which is permanent deformation). A As another example, if a hollow tube for a
possible example of an elastic deformation truck drive shaft is not heat treated properly,
failure is a spring that does not have the correct could it buckle more readily? If the damage is
spring constant. Imagine a coil spring that is buckling deformation, the primary factor would
supposed to stretch out 0.1 mm when subjected be the wall thickness and the modulus of elas-
to 12 N of force. What if the spring stretches out ticity. However, there may also be some strength
0.05 or 0.2 mm? Can that type of failure be due issues. When people talk about buckling due to
to heat treating? What are the causes or the instability (or Eulers buckling, or what is
factors that allow these two types and two levels referred to as sudden onset), the main factors are
of deformation to happen? the geometry and the modulus of elasticity.
As with sudden onset, the two main factors However, with geometrical configurations that
that control elastic deformation are geometry are not exactly like the extreme examples that
and modulus of elasticity. These are not factors Euler used to develop his theory, one can
that are greatly influenced by the heat treater. appreciate that the equations become very com-
Thus, the heat treater is usually not at fault in the plicated and actually do have factors based on
case of sudden onset buckling of a column, strength values. However, the predominant
sudden onset buckling of a tube in torsion, or for factors, in general, are the geometry and the
elastic deformation failures. In general, sudden modulus. In failures of heat treated structural
onset and elastic failures are a result of the steel, the modulus will be 30 million psi, with
combination of the design or actual geometry some variations. For stainless steel, the modulus
and the elastic modulus being insufficient to may be a little different. If the component
sustain the loading conditions. Sudden onset and material is a heavily cold-worked steel with
elastic deformation damage is only secondarily oriented, heavily textured microstructure, then
related to the yield strength. The modulus of the modulus of elasticity may be different
elasticity is, in general, a constant. However, in the different directions. However, most of
modulus is not totally a constant in wrought this orientation may be eliminated by any
subsequent heat treating that included an auste- The other question that is frequently asked is,
nitization phase. Note again that this discussion Why did this one break at this specific time?
is strictly speaking about deformation in the This is a different question, and to answer it,
absence of a primary fracture event. fracture mechanics type of explanations and
Before leaving the subject of deformation, theories must be explored. In such a case, one
consider a small spring clip that was designed to assumes that every structure has some small
be used at the top of the linear portion of the discontinuity and a related crack growth rate,
stress-strain curve. Why would anyone design a which is a function of the stress intensity (mea-
spring to be used at a stress almost at the nominal sured in terms of the mathematical product of the
yield strength? No experienced metallurgical nominal component stress and the crack size)
engineer would expect each one of a quarter of a and the fracture toughness. These are functions
million such spring clips to sustain multiple load of service condition and material parameters.
cycles to the theoretical elastic limit and not Particularly for parts that appear to have broken
have any permanent shape change. These clips in fatigue, (consider a two-year service life when
could not even sustain the minimum 15 load- no similar component had previously cracked in
cycle requirement without excessive permanent under five years), it may be informative to look
deformation. This was a case of a nominally at the microstructure and how that may have
elastic spring application where deformation impacted the fracture toughness. Microstructure
failure was caused by the heat treatment, speci- and fracture toughness could definitely be rela-
fically by normal variations of the heat treating ted to heat treating issues. Other chapters in
process. It would be difficult to blame the heat this book give more information on fracture
treater in the absence of very strict quality- mechanics.
control specifications. In this case, the major For a more conceptual, lower-math-content
cause of the failure was the design engineers methodology to understand why something
unrealistic expectations of uniformity of heat cracked or why it broke, start by reviewing the
treated components or the design process in the stress and strength variations that are at work in
company in which the designer was working. 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
Fracture lower than the local stress, then it is possible to
How can fractures of heat treated steel parts initiate a crack. Once a crack exists, it may or
be examined to determine the existence of any may not progress to complete component frag-
factors related to the steel itself or its heat mentation. Sometimes, cracks do not propagate.
treating? Basically, there are two types of However, cracks generally do not heal them-
questions to ask. One of the things that people selves, and they often do propagate. It is actually
like to ask is, Why did this one break? That is a reasonably straightforward to learn to look at
useful question in the case of a part that has been a component, if one has an understanding of
used successfully for many years. Maybe there is the loading geometry and the heat treating,
one part that failed out of a half-million parts that and determine if there was something wrong
are in service. Proper examination and evalu- with the component in question. Someone
ation of the physical evidence can reveal much inspecting a component with a complex shape
to answer this question. Before evaluating the or complicated loading history may benefit
effects of the heat treatment itself, one must first from a good finite element analysis. However,
examine the physical condition of the damaged many components do lend themselves to ready
property. It must be understood how the loads evaluation regarding the presence or absence of
interacted with the component to create new an extraordinary factor promoting premature
surface area where none used to exist. The visual fracture.
appearance in three dimensions can reveal a To use the stress/strength distribution concept
large amount of information on issues related to to analyze fractures, one should be familiar with
how the part was really loaded. The colorations the six basic loading geometries, including
and surface texture of the newly formed (unde- tension, compression, bending, torsion, contact
sired) crack surfaces can indicate how the crack stress, and shear. A review of some of the
happened and how long it took for the crack loading geometries is presented in Ref 1; how-
to grow. ever, the three-dimensional characteristics of the
fragments are not systematically presented. cracks. In fact, much of the published literature
Armed with information on these basic fun- is unclear on this issue. To understand this more
damental loading geometries, one can learn to clearly, see Fig. 2. The image is of a broken
predict macroscale fracture appearance in duc- tensile bar. Near the fracture, the material is
tile and brittle materials given different loading necked down. Because of the visible shape
conditions, along with the more well-known change, this is an obvious ductile fracture. The
fracture surface features, such as beach marks, image in Fig. 3 is of a threaded fastener. Based
ratchet marks, and chevrons. on the ridge patterns in evidence, the crack
This is a much more powerful method of started at the root of one of the threads and went
visual examination than simply interpreting back into the page. There is very little indication
surface texture features. With practice, one can of any deformation visible at the macroscale.
look at a broken component and obtain a realistic This is correctly called a macroscale brittle
idea of what loading conditions actually caused fracture.
the fracture. Sometimes, it can be shown that the Figure 4 has more challenging fragments for
real loading was very different from the design fracture analysis. The image shows a chain link,
loading. which did not originally have an open shape. The
Before leaving this subject, a review is needed two protruding ends were touching; there was no
on details that are often poorly understood gap. The lower portion of the link did not have a
relating to how to distinguish ductile and brittle curved arc shape as depicted. A significant
cracks on the macroscale. This is clearly very amount of deformation was clearly associated
important for a heat treater to know to defend with this fracture event. Does that make this a
against incorrect accusations of embrittlement.
There are some types of macroscale ductile
cracks that can easily be misinterpreted as brittle
Fig. 2 Example of a macroscale ductile fracture in tensile Fig. 3 Example of a macroscale brittle fracture in tensile
loading loading
ductile fracture, or could this be a brittle frac- path. In this case, a fatigue crack grew due to a
ture? Imagine that only this one photograph is tensile load in the horizontal portions of the link
available, and it must be determined whether (as shown).
this was a ductile or brittle crack. To add a few more details to this case study,
Clearly, the steel that was used to make the the chain was in service at a plant that processes
link has the capability of being deformed. In meat, and strong acids were used to clean the
other words, the steel is ductile at the conditions conveyor systems. This crack actually initiated
present at the time of bending. The analyst must at a corrosion pit on the inside surface. The
not be satisfied with that answer, though. The cleaner reached the inner surfaces, but the em-
proper analysis includes a determination of ployees may not have rinsed the chain very well.
when the deformation happened. Postfracture This allowed a corrosion pit to form, which then
deformation does not make the crack event itself allowed a fatigue crack to grow. Again, it is
ductile. The macroscale deformation must occur important to understand that the material itself is
during and as an inseparable part of the fracture ductile; there is nothing wrong with the material.
process for the crack itself to be a ductile crack People involved with failure analysis need to
event. keep in mind that material behavior is or at
For those unfamiliar with this methodology, least may be different from material capability.
imagining that the link broke suddenly while it People doing failure analysis work need to be
was under load can be helpful. If the crack were able to distinguish inherent capability and actual
ductile, one would likely see some necking at the behavior.
crack location, since chain links are generally To underscore the importance of separating
loaded in tension. No such localized deforma- the behavior from the capability, imagine the
tion is observed in this case. On closer exam- potential corrections that may be considered if
ination, one can see a tiny shear lip at the top someone found this to be a ductile overload
edge of the fracture surface. That makes this fracture. The cure may be to make it harder. In
crack, for the most part, a brittle crack at the the case of the acid cleaning, harder steels are
macroscale, despite the presence of available often more susceptible to stress corrosion than
ductility in the material. The visible deformation softer steels. If the harder-is-the-answer the-
near the center of the lower (originally straight) ory were put into practice, an undesirably short
portion of the link happened after the crack was life may become a horribly short life. It is
completely formed. important to be sure that a crack that is diag-
In doing fracture analysis, it is important nosed as ductile is really ductile and one that is
to distinguish the capability of ductility in the diagnosed as brittle is really brittle. Finally,
material from the behavior at the time of returning to the heat treating issues, the fact that
the crack event. Despite the material ductility, someone misdiagnosed this crack as a ductile
the crack happened in a brittle way. Closer fracture may lead to the heat treater being
examination of other views not shown provide blamed for overtempering or inadequate hard-
clear evidence that this was a fatigue crack. ening. In fact, until now, the potential blame or
Beach marks were visible. Fatigue cracks grew innocence of the heat treater has not been
below the yield strength of the component, investigated in any thorough manner. It is poss-
creating macroscale brittle features. To clarify ible that poor heat treating or poor material
one other potential source of confusion, it is manufacture contributed to the ease of corrosion
important to remember that tensile refers to a attack, and the meat processing plant employees
loading geometry. Fatigue is a type of crack 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 reduc-
tion 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
Fig. 4 Example of a macroscale brittle fracture in tensile the transverse orientation to the length of the
loading shaft. To best understand macroscale ductile and
brittle fracture, one must be familiar with normal possible loading geometries that could have
and shear stresses. On a simple level, normal created the fracture must be reviewed to make a
stresses cause macrobrittle cracks due to crack proper determination of whether or not the crack
opening forces created at the crack tip. Shear is macroscale ductile or brittle. That determina-
stresses allow slip and are the basis of the tion cannot be made without assessing what the
deformation that creates the ductile crack event. loading geometries may have been. If this frag-
So, even though there is no necking, this is a ment had been totally covered with red rust, it
ductile crack. would have been even more difficult to deter-
Often, it is relatively easy to see some evi- mine the basic ductile or brittle behavior of the
dence of twisting on the side of a ground shaft material without knowledge of the loading
that has failed from ductile fracture by torsional geometries and expected fragment shape.
forces, which would lend more credibility to this Stress versus Strength. Almost all real
diagnosis of ductile fracture. However, this shaft loading geometries cause the stress to be highest
was extremely smooth, and it required a long somewhere along the part surface. If the strength
etching time in a heated acid solution to reveal is uniform, for example, if there is a piece of hot
permanent twisting on the original cylindrical rolled 1050 steel that does not have any de-
surface. carburization or carburization and has not been
Another possible source of confirmation that shot peened, the crack initiation is expected
this is a ductile crack is the classical smear somewhere at the original part surface. In the
features on the fracture face. It has been argued presence of any type of bending or torsional
that these smeared features could be a result of stresses, the highest stress will be at the surface
postfracture damage. While this is a possibility of the part. In the presence of pure tensile
that should be considered, since the background loading, theoretically the crack could start any-
evaluation revealed that the shaft was loaded in where in the cross section. Such pure tensile
torsion, and there are no crack opening stresses loading is rare. Imagine the case of a hydraulic
operating on the transverse planes, this must be cylinder rod. Even here, there must be a section
a macroductile crack. A macrobrittle crack in change, a fillet, at some point. The loading at the
torsion is helical. fillet is not uniform; there is a stress concentra-
In closing this section on fracture, note how tion. Even a tensile test coupon that is forced to
important it is to follow the advice of the many break in an area of nominally uniform strength
authors and teachers who state that background and stress is not totally uniformly loaded. Most
research is step one of a failure analysis. The 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 sub-
surface. It is important to note such an unusual
situation, where the location of the ratio of the
Fig. 5 Example of a macroscale ductile fracture in torsional stress to the strength was highest at a subsurface
loading position.
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
some other measure of the strength of the com-

ponent. 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
Fig. 7 Stress and strength as a function of position in a
the strength drops off as the dotted line shows,
cylindrical component loaded in torsion. Fracture ini-
tiation may be at either the surface or subsurface. Subsurface the stress is higher than the strength within a
initiation depends strongly on the hardness profile from surface to subsurface band, which allows subsurface crack
center if loading is in bending or torsion.
initiation.
This figure shows a powerful technique for
specification of case depths, which has the
Imagine a simpler case of a cylindrical com- potential to complement the usual experiential
ponent (Fig. 7). The surface of the part is shown method of case depth specification. Anyone
along the left side of the graph, and the center- doing fracture analysis on a case-hardened part
line is shown at right. For such a cylindrical can also use this information to obtain an idea
component that is loaded in either bending or about the appropriateness of the hardening spe-
torsion, the stress will be highest at the cification.
outside surface, and at the centerline, or neutral It is important to realize that there is one
axis, it will be nominally zero. A carburized or other case where a subsurface initiation may
induction-hardened material is actually stronger occur that is not related to the heat treating or
at the surface layers where the stress is highest. specification quality. Imagine the presence of
The y-axis, instead of being the stress level, can a subsurface discontinuity, such as an inclusion,
be conceptually viewed as either the hardness or a void, or a tiny crack. Even if the steel has
constant strength all the way through its cross or other rolling elements and races.
section, an inclusion that is big enough to locally Understanding that most loading geometries
increase the stress above the fracture strength of create the highest stresses at the surface also
the part makes it possible to create a subsurface allows one to understand why decarburized
initiation (Fig. 8). layers can be so damaging. A decarburized layer
Another potential cause of subsurface fracture is softer and lower in strength than the material
initiation is contact loading. A well-lubricated with the desired nominal amount of carbon.
bearing without any friction has the highest Decarburization can even occur on carburized
stresses in its subsurface layers. Thus, fracture steel. Figure 9 shows a metallographic cross
initiation will be at a subsurface location where section of a piece of steel that is carburized and
contact loads are the predominant source of has quite a bit of retained austenite. Note the
stress. Inclusions can thus be very damaging in dark constituent at the surface (arrows). On
contact-stress applications, such as bearing balls carbon steels in the medium-carbon range, de-
carburization 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 con-
ditions. Fracture analysts must start with the
Fig. 8 Stress and strength as a function of position in a macroscale, or the big picture. Many people start
cylindrical component loaded in torsion with subsur- with the details, or the little picture, and move on
face discontinuities. Surface conditions may include: inadvertent
decarburization, typically thin and may not be easy to find; deep
to the big picture, but this can be a problem that
case from induction or carburization; nitrided, thin case, often facilitates mistakes on the part of the analyst.
not more than 510 mils. Part of the case may be ground off in the Microscale Fracture Features. Scanning
finishing operations. The defect could be a faceted inclusion
(nitride) in a low-ductility matrix. Nondeformable nitride causes electron microscopy can be used to reveal the
stress concentration in the matrix. microscale fracture features. Figure 10 shows a
Fig. 9 Decarburization of carburized steel. Each small scale Fig. 10 Scanning electron micrograph of microvoid coales-
division is 2 mm. cence
ductile crack path, or what is often called ductile location? The situation is obviously more com-
dimples, microvoids, or microvoid coalescence plicated than just described; obviously, the grain-
(MVC). The MVC is a characteristic fracture boundary strength is a function of temperature,
morphology that indicates a component was which is rapidly changing, but it is helpful to
subjected to stresses in excess of the nominal think about what could be causing that inter-
ultimate tensile strength of the component granular crack and why quench cracks are gen-
material. erally intergranular.
The MVC is often an indication that the heat Note also that carburized steel often will have
treated steel did not have a gross problem with intergranular cracking in fatigue. Sometimes,
the heat treating that caused or contributed to an even the best metallography cannot show any
embrittlement problem. Brittle fracture features, problem at the grain boundaries, that is, no
including cleavage and (most) intergranular grain-boundary carbides, oxides, nitrides, poro-
features, are often indicative of a heat treating sity, and so on. It is true that there are many
problem. The arrows in Fig. 10 show the non- heat treating problems that can facilitate, or be a
metallic inclusions that initiated the void for- factor in, intergranular cracking at lower stress
mation. levels than the part usually sustains. However, it
Brittle fractures are often unexpected and is important to note that just because there is an
occur suddenly without any prior warning. intergranular crack, it does not mean there is
Ductile fracture by MVC is typically accom- surely a heat treating problem.
panied by prior plastic deformation, which gives Hydrogen embrittlement is often revealed in
advanced warning of the impending fracture part by its intergranular crack habit, particularly
event. This prior warning makes MVC the pre- in steels heat treated to high strength levels.
ferred mechanism if fracture occurs. Hydrogen embrittlement is not always purely
While MVC is generally desirable, it can intergranular; sometimes, there will be tiny,
indicate that the material is too soft if a high- shallow microvoids on the grain-boundary sur-
strength material is in question. The MVC can faces. Hydrogen embrittlement could be the heat
also reveal that the heat treater made a mistake, treaters fault (if the hardness is too high), and it
such as no heat treatment, despite the often could be an issue with the plating (if records do
preferred MVC fracture path. not show proper baking); it could be an inter-
Classical microscale brittle crack paths are action between these two factors.
along the grain boundaries (intergranular), The important point is not to confuse identi-
(Fig. 11) or cause the grains themselves to split fication of the crack path with the cause. There
(transgranular or cleavage). Many heat treating is a difference between identification of the
and other processing problems can cause unde- physical shape of the crack and the physical
sirable intergranular cracking at the microscale. characteristics relating to the crack event. These
To have intergranular cracking, something may do not automatically lead to the cause.
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 gen-
erally 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 Fig. 11 Scanning electron micrograph of intergranular
that create a shear stress at the grain-boundary cracking
The last microscale crack path is cleavage, Wear

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

Fig. 12 Scanning electron micrograph of cleavage cracking of steel segment shown is 23 mm (0.9 in.).
shows a bearing race and a bearing ball. The ball portion. Could this have been caused by
is being pressed into the race. This figure is the user?
a rough schematic, conceptually showing an To answer the question, one would have to do
exaggerated view of the most highly stressed a large amount of background information col-
area due to the elastic deformation of the two lection. The important point to realize is that the
components. Without the load, there is essen- crack location is totally unexpected, and it is
tially a point. Under load, the contact area difficult to think of something that could have
becomes circular or elliptical; in other words, happened in service to create a weak spot at this
there is a contact footprint. The actual high- hefty location.
stress location is just below the surface. In the If it cannot be qualitatively demonstrated that
presence of friction, the highest stresses are somebody abused something, then it must be
moved toward the surface. An important part of quantitatively demonstrated. This is often diffi-
bearing wear failures is determining whether the cult. It is important to know what the material
crack initiated at the surface, showing the pos- strength, fracture toughness, and other material
sibility of a lubrication issue, or if it was truly a properties were at the time of the damage. It is
subsurface initiation, in which case it may be a important to determine what the load or loading
microstructure problem, an inclusion, or another geometry was that caused the damage. One must
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 surface-
initiated 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 Fig. 15 Scanning electron micrograph of fatigue crack initi-
ating on worn carbonitrided steel. Original magni-
as using a screwdriver as a pry. Understanding fication: approximately 4000
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 Race
Bearing ball
Stress
concentrations
Fig. 14 Scanning electron micrograph of a worn piece of

hardened medium-carbon plate showing details of Fig. 16 Rough schematic of stresses in contact loading of a
the wear mechanism bearing ball on a race
consider how the suspected loading geometry heat treating problem with the crack, because
differs from the actual design intent. The analyst there is no information about the stress
again must remember that just because it is levels. Stress-corrosion cracking, by definition,
broken does not mean it was abused. Even if it requires a threshold level of corrosion and
meets the specification, it does not prove that the stress. In this case, there is no evidence that the
user abused the product. crack was caused by bad heat treatment. This
seemingly subtle distinction may be very
important in the case of a catastrophic failure
Corrosion and Environmental Damage event.
Figure 18 shows an example of a 300-series Figure 19 shows an example that involves
stainless steel that was probably not heat treated another stainless steel weld. Welding is a kind
in the most ideal manner, since small pre- of heat treatment, although not as controlled as
cipitates can be seen along the grain boundaries. an intentional heat treatment. An acid sub-
Note the crack location, which seems to be stance, polythionic acid, was in contact with
seeking the grain boundaries. To create a crack, the weldment. There are some cracks on one
there must be a load or a stress. There is corro- side of the weld, while the other side is free of
sion present too, so this may be a stress- cracks. What kind of cracks are these? The
corrosion crack. Did the nonideal heat treating crack path is intergranular (Fig. 20). The
condition cause the crack? In this case, a chain micrograph in Fig. 21 was taken after an
of reasoning cannot directly link the specific ASTM International test, and it shows ditching
Fig. 18 Grain-boundary precipitates in a 300-series stainless

steel
Fig. 17 Steel bar with crack in unexpected position. Orig-

inally shock loaded in compression. Threaded
portion diameter is approximately 2 cm. Fig. 19 Stainless steel weldment
characteristics. This indicates that whatever the damage mechanism, stress-corrosion cracking.
heat sequence from the weld, it did create this When the regular 300-series stainless heated
condition that appears to have facilitated up, the chromium and carbon combined and
intergranular corrosion. However, steel would precipitated along the grain boundaries, taking
have been considered to pass the test for free- the chromium out of solution and making the
dom from sensitization. Despite this, the ther- material less corrosion resistant. Frequently,
mal experience of the material can be directly such situations lead to pit formation, which
linked to the form of the crack. Although this is then allows the crack to propagate from the
not a heat treating example, this demonstrates stress concentration at the pit. This is an
the line of reasoning that is required to deter- example of a complex damage mechanism.
mine cause. It is interesting to note that one Another commonly named damage mech-
side cracked and the other side of the weld did anism is corrosion fatigue. It must be under-
not. It is likely that the side that did not crack stood that a corroded part that broke due to
was an L-series, a low-carbon series, specifi- repeated crack extension under load did not
cally made to minimize the chance of cracking necessarily experience the mechanism called
in weld heat-affected zones in stainless corrosion fatigue. Corrosion fatigue is a dam-
steels. The cracked material was probably not age mechanism that is studied in the laboratory.
an L-series. This is an example of a classical 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. 20 Scanning electron micrograph of stainless steel

weldment with intergranular cracking
Fig. 21 ASTM International sensitization test results showing Fig. 22 Cracked shaft used in a corrosive environment.
ditching characteristics Diameter of the shaft is approximately 10 cm (4 in.).
obviously corroded. The band highlighted Factors Contributing to Poor Response

by the arrows was covered with a layer of red from Heat Treatment
rust. Figure 23 shows the crack created by the
corrosion and stress combination. Note that Raw Material Characteristics That
there are also secondary cracks. The tip of this
Can Contribute to Poor Response
crack looks like a witchs broom or the trails of
a sparkler. This is not a normal fatigue crack. A from Heat Treatment
normal fatigue crack does not branch out and What are the raw material characteristics that
have multiple tips. In this case, the corrosion can contribute to poor heat treating outcome?
definitely has some kind of significant impact One very important characteristic is composi-
on the damage mechanism. In many cases, tion. There will be a range of values for each
examination of field failures is likely to leave a type of atom that is specified for the grade in
question mark regarding the quantitative eva- question, as well as for unspecified elements.
luation of how much faster the crack is growing A heat treater may receive material of an iron
because of the corrosion. matrix that could, with the same name, have a
To demonstrate that corrosion is the cause very wide range of responses to the heat, heating
or the fault of a failure, it is not enough simply rate, heating dwell time, cooling rate, and so on.
to say the component was in a corrosive en- Lean and rich alloy content can have a strong
vironment and it cracked because of that. It must influence on whether or not quench cracks occur.
be assured that the specific conditions, the spe- Lean and rich compositions also strongly influ-
cific material, and the specific process condition ence how readily the hardness specification is
of the material were in the realm that has been attained. For a hardness specification that is
demonstrated to be a problem for the damage toward the upper limit of what can be reliably
mechanism invoked. obtained for a particular grade, it may be diffi-
For example, concentration of an aggressive cult to meet the specification for a given lot if all
substance, the threshold stress level, and the the elements are on the lean end. If the part(s)
temperature may be required to be in a restricted crack on quenching, an important task in the
range before a particular mechanism can be troubleshooting process is to determine whether
properly said to have been acting. it is due to a lean or rich (more likely) alloy
Fig. 23 Crack profile and adjacent secondary crack tip. Original magnification: 50 . Inset is of a different secondary crack tip.
Original magnification: 500
composition that is facilitating martensitic trans- Section change was more severe (for ex-
formation. This must be differentiated from the ample, the fillet radius was too sharp)
situation where the major cause of the crack is
To confirm the suspicion of a pure quench
that the heat treater quenched the part(s) more
severity/rich composition-related quench crack,
severely than usual.
it is advisable to confirm that the microscale
It is difficult to distinguish quench cracks due
crack path is intergranular (usually SEM is
to extreme quench severity from those due to
required). If a seam had been found at the
rich composition. However, it is easier to de-
quench-opened crack, it would not be correct to
termine whether the crack has an expected
blame the heat treatment. Figure 25 shows an
geometry for a quench crack. In Fig. 24, the
interesting crack. There is a very heavy oxide
crack is at the section change, which is a prime
layer revealed by the cross section. The inset
location for a quench crack, as are sharp corners.
shows a higher-magnification view of the seam
When examining a part to determine if rich
detail. There is a rounded particle that is totally
alloy/quench severity issues were at stake,
covered by a heavy, rounded oxide layer. The
examine the crack surfaces for traces of temper
crack surface does not have blue or brown
colors, including blues and browns. A dark
temper colors but was found to be a dark char-
matte gray or black surface may indicate the
coal black. This type of heavy oxide is unlikely
presence of an oxide-filled discontinuity that
to have happened between when the part was
simply opened due to thermal stresses rather
quenched (and presumably cracked) and when it
than a rich alloy/quench severity problem.
exited the temper furnace. Another unusual
If temper colors are visible at the portion of
feature of this crack is that it changed direction
the crack surface nearest the part surface, it is
multiple times (large arrows show crack growth
likely that one or more of the following factors
direction changes). The fact that these are not
was present:
predominantly intergranular cracks deals the
Quench was more severe than was appro- final blow to any theory stating that this crack
priate was due to a heat treating problem.
Alloy was richer than usual 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. Limita-
tions 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
Fig. 24 Quench crack with typical geometry to differential cooling rates, the experienced
Fig. 25 Oxide layer along a seam most likely present in the raw material. Original magnification: 50 . Inset original magnification:
200
analyst looks for some type of discontinuity. The a particular direction, and that may be an
discontinuity does not have to be very thick. This advantage for a particular application. Samurai
example was particularly heavy, but even a very and Damascus swords from antiquity had basi-
thin oxide layer is enough to be the predominant cally banded microstructures with very desirable
cause of the crack. characteristics. While banding is not necessarily
Macro- or microsegregation (otherwise bad, it can cause some variation in the response
known as banding) are other raw material char- to the heat treatment.
acteristics that can interfere with the expected If a material has intermittent coarse grains, it
outcome of a heat treating process. Figure 26 may be easier to form martensite in the large
is a longitudinal cross section from a medium- grains and pearlite, ferrite, or bainite in the
carbon piece of steel. If the low-carbon layer is surrounding fine grains. Can the heat treater
right at the surface, it may be difficult to meet a create such a grain size distribution? This may
minimum hardness specification. be possible by overheating.
Banding, or microsegregation, is not always However, it is also possible that the coarsen-
bad. It can make it easier for the crack to grow in ing came from a subcritical amount of cold work
stored in the material. Normalizing the steel
prior to hardening may eliminate the nonuni-
form 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 tor-
sional 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
Fig. 26 Longitudinal cross section showing microsegreg-
decarburized layer on either side of an inclusion
ation. Original steel segment shown is approximately
2 cm wide. 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 por-
tion of the core able to form martensite. Original magnification: Fig. 28 Scanning electron micrograph of sulfide stringers in a
100 piece of bar stock
Component Characteristics script features may make the case even more
brittle than usual. The heat treater can control the
Figure 30 shows a cast steel product that was
carbon potential and the heating cycle. How-
quite uniformly carburized, except for the white
ever, the heat treater cannot locally put fewer
script features. The pattern looks very similar to
carbon atoms into the steel at the locations that
microsegregation between the dendrites and the
already have too many. When performing fail-
matrix. It would be very difficult for the heat to
ure analysis of steel components, analysts must
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 Fig. 30 Cast steel after carburizing. Original magnification:
100
Dark etching
area possible
indication of
heat treating
problem
0 10 20 30 40
Unusual crack
shape
Oxide filled crack
Evidence of Note multiple
pre-existing cracks in one
discontinuities thread root
Fig. 31 Thread root of a steel fastener. Original magnification: 100

always be on the lookout for the whole history of Testing competence

the part that led up to the heat treating event. Hardness scale selection
Forging discontinuities are another situation Frequency of part testing
that could cause a problem. Figure 31 shows a
Figure 32 shows two micrographs at the same
cross section of a threaded fastener with a locally
magnification from the same component,
different compositional steel inclusion. Some-
25 mm (1 in.) apart from each other. One of
thing unusual happened to create this feature.
them is virtually all martensite. In the other
Other problems relating to poor heat treating
location, there are wide grains of pearlite inter-
or poor service outcome include many different
spersed between the martensite grains. There
types of design details. Other chapters in this
was as much as 12 Rockwell C points difference
book address these issues in greater detail. To
summarize, some areas of common oversight between the two microstructures. The designer
never indicated where to test the component!
include:
This is a typical design issue.
Materials selection Another common cause of disputes between
Heat treating process selection purchasers and providers of heat treating ser-
Hardness level specification and range and vices is that designers specify Rockwell tests
position on part for hardness test when there is no way to perform anything but a
Process details (batch or continuous Knoop or Vickers test. There are still many
oven, etc.) newly minted engineering prints with case
Heat treaters familiarity with the size and depth specifications that are very unclear. There
complexity of the part and the quality level are standard methods for specifying carburized,
needed induction-hardened, or carbonitrided case
Distortion control depths, and it is helpful to use an industry
Fig. 32 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.
standard when possible. The automotive manu- only the heat treater could know. Design engi-
facturers have done a good job of providing a neers specify materials and thus need to know
range of methods at a range of ease of testing. If how to specify the type of testing and evaluation
a standard method is not being used, it is often required for application. A design engineer
difficult to determine the designers intent. cannot know how fast a load of parts will be
The specifications for 400-series stainless heated in a particular companys individual
steel can be particularly difficult for the average furnace. The design engineer cannot be expected
mechanical designer to write. Many designers to know what type of fixturing may be necessary
specify 400-series stainless steel because they to maintain required distortion levels in the
want stainless, but they want to be able to heat part. These are aspects the heat treaters must
treat it to obtain higher hardness than standard know. The heat treaters must know what load
300-series annealed bar stock. For a number of size can be treated in their own furnaces and how
the 400-series grades, one must determine in the load should be distributed. The heat treaters
advance whether maximum strength or max- must understand the characteristics of the
imum corrosion resistance is desired. Two interactions between their equipment and the
totally different heat treating processes attain full range of part sizes and load sizes they are
those goals. The heat treater has no means of processing.
guessing which characteristic is required. Heat Treating Errors. Excessive heating
rate, excessive time at temperature, and exces-
sive temperature can lead to excessive distor-
What Are the Things That Can Go Wrong tion. Excessive temperature can cause problems
in the Heat Treating Process? with excessive autotempering. If a massive part
This chapter has attempted to cover all the is heated to 75, 100, or more degrees hotter than
aspects for which the heat treater should not be it needs to be to obtain the uniform austenite
blamed. What are the aspects for which the heat required before quench, then all the extra
treater may or should bear responsibility? heat must be removed, which can make it diffi-
One approach is to say that heat treaters must cult to create the desired martensite at all. Fur-
take responsibility for those aspects that are thermore, the extra heat may act to partially
specific to their equipment. These are details that temper the martensite that is present. Excessive
Fig. 33 Undissolved ferrite and martensite in improperly specified and improperly induction-hardened medium-carbon steel part.
Each small scale division is 5 mm.
heating can cause trouble in obtaining minimum the added ductility is a necessity. The annealing
specified as-quenched hardness values. These line management may be tempted to cut down
low as-quenched hardness characteristics may the process time to save money. The minimum
or may not make a difference in the component spheroidization that the material processor
performance. However, when as-quenched believes will work is what is often provided.
hardness tests are required, it is important to In the case of varying incoming microstructures,
know that excessive temperature may be a the part may not be as ductile as usual. Examine
cause. the situation where the forming process creates a
Inadequate Heating Rate, Inadequate crack that was undetected, and then the part is
Time at Temperature, and Inadequate Tem- heat treated for hardening. If the crack remains
perature. Figure 33 shows the microstructure undetected, there is now a part with a dis-
of a part that was supposed to be induction continuity due to the spheroidizing being done
hardened. There is some martensite. The light poorly. This type of situation may be difficult to
constituent is ferrite that never went into sol- figure out, especially if the failure happens some
ution. The initial microstructure was probably a years after the fact.
mixture of ferrite and pearlite. The pearlite Insufficient time or temperature could apply
transformed into martensite, but little of the to formed parts requiring stress relief. Even at
ferrite did. On quenching, the material produced relatively low hardness values, excessive resi-
islands of martensite with islands of ferrite. The dual stresses can make the part sensitive to
irregular shapes of the ferrite islands are classi- hydrogen embrittlement. A low-carbon steel
cal undissolved ferrite. This is not a typical part that has been heavily deformed and
shape of ferrite grains formed on cooling. improperly stress relieved can crack after a very
Because of the severity of the consequences, short service life or even while sitting on a stor-
most of the common problems in heat treating age shelf. Stress relief is often used on weld-
are related to hardening. Annealing can also be ments, and if it is not done properly, fatigue
done incorrectly. Figure 34 shows the microcracks can initiate more readily at weld toes.
structure of a steel that was supposed to be Machined parts that are improperly stress
spheroidized annealed. Spheroidization is a pro- relieved can distort or crack at a later time,
cess that may take 12, 15, or even 20 h at 600 to because the stresses are higher than one may
700 C. Spheroidized annealed steel is generally think.
quite expensive. Its applications are usually Cooled Too Fast. The part that is cooled too
reserved for severe forming operations where fast due to cold quenchant or excessive quen-
chant agitation may crack or suffer excessive
distortion. Undesired microstructures may also
result from excessively fast cooling. Bainite
may be desired, but the process formed mar-
tensite.
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 atmo-
sphere 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
Fig. 34 Incompletely spheroidized annealed steel. Each treaters in business who believe that it is good to
small scale division is 2 mm.
use some ammonia when carburizing, even if the
customer did not ask for carbonitriding. These a 1050 steel that is 100 mm in diameter and
heat treaters may believe that the customer is 100 mm long and water quenched, should it be
happy to get the lower price due to a faster martensite? There probably will be some mar-
process that meets a surface hardness specifi- tensite near the surface, but it may not be a very
cation. The problem is that if the stress state deep layer. The exact depth will depend on how
requires a certain level of strength at a certain hot it was heated prior to quench, the details of
depth, the faster ammonia-enhanced process the quench tank design, and many other factors
may be inadequate. that may not be readily apparent.
Retained austenite in greater amounts than It is important to have a large amount of
normally found in parts that are straight reasonably deep knowledge to be able to make a
carburized can also be a problem. Porosity can fair and correct determination of where there
also be created in the case with nitrogen atmo- may have been a problem in the entire process of
spheres. This can be a problem if the surface designing a part, procuring material, and making
hardening is desired for strength rather than just a component. This knowledge base includes
scratch resistance. failure analysis, fracture analysis, and micro-
Figure 35 shows an example of bad carbur- structure analysis and interpretation in order to
izing. There is a significant fraction of retained read the process history. Simply checking the
austenite in this case-hardened part, as well as a hardness and the composition to see if they meet
large, chunky, unusually shaped puzzle piece the specification is not failure analysis. Failure
carbide. This feature could be a problem for analysis includes a determination of the loading
some applications of carburized parts and may geometry and background information, at the
be the result of excessive carbon potential. very least, in addition to the basic certification
conformance tests.
Some people legitimately perform internal
process failure analysis on a part that never left
Concluding Comments the door of the manufacturing plant. In failure
analysis of a field return, even from a non-end-
For those who do failure analysis of heat user assembly problem, it is important to do
treated steel parts, it is important to understand more than simply look at the composition and
what microstructure is expected, given the heat the hardness. The type of damage needs to be
treating process that is specified. For example, in identified, as well as the possible sources of that
0 10 20 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.
type of damage. A review of the comprehen- unanswered questions at the end of the project. It
siveness of the design process may be in order. If is difficult to overemphasize the importance of
recurrence prevention is a goal of the failure spending enough time initially figuring out
analysis, the damage specialist may lend some exactly what the goals of the failure analysis
understanding to the design engineers to help project are and how much detail is required.
them clarify the requirements of the component
characteristics.
Failure analysis can be very routine, or it can
be extremely involved. This chapter has con- ACKNOWLEDGMENTS
sidered only a few categories of the analysis
The author thanks Mrs. W.T. Becker for permission
procedures and some of the reasoning involved
to use the copyrighted material of William T. Becker
in determining what went wrong and at what part in Fig. 6 to 8.
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, espe- REFERENCE
cially, or a very limited number of failed
parts, inadequate planning can leave inadequate 1. D. Wulpi, Understanding How Components
specimen material for testing in the case of Fail, American Society for Metals, 1985
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 5. Macroscopic examination and analysis and

collecting, examining, and interpreting damage photographic documentation
evidence. The objective is to understand the 6. Microscopic examination and analysis
possible conditions leading to a failure and 7. Selection, preparation, examination, and
perhaps prevent similar failures in the future. analysis of metallographic specimens
A failure analysis should provide a well- 8. Determination of failure mechanism
documented chain of evidence that either 9. Chemical analysis
excludes or supports possible interpretation of 10. Fracture mechanics analysis
the damage evidence. Clear-cut conclusions 11. Testing under simulated service conditions
do not always occur, and the tendency of 12. Analysis of all the evidence, formulation of
developing preconceived interpretations should conclusions, and writing the report
be avoided. These stages or steps are briefly outlined as
Various publications (e.g., Ref 16) describe follows.
the guidelines and methods of failure analysis, Collection of Background Data and
and this chapter briefly outlines some of the Selection of Samples. There are basically
basic aspects of failure analysis. The first section three fundamental principles to be carefully
describes some of the basic steps and major followed when collecting damage evidence
concerns in conducting a failure analysis. This from a fractured material (Ref 2):
is followed by a brief review of failure types
from fracture, distortion, wear, and corrosion. Locate the origin(s) of the fracture. The
Fracture is a common damage feature, because whole fracture surface should be visually
the vast majority of mechanical failures involve inspected to identify the location of the
crack propagationtypically classified as duc- fracture-initiating site(s) and to isolate the
tile, brittle, and fatigue, as briefly described areas in the region of crack initiation that
in more detail. Distortion, wear, and corrosion will be most fruitful for further micro-
also can be important damage factors in failure analysis. Where the size of the failed part
analysis. permits, visual examination should be con-
ducted with a low-magnification wide-field
stereomicroscope having an oblique source
of illumination (Ref 3).
General Guidelines of Failure Analysis
Do not put the mating pieces of a fracture
back together, except with considerable care
For a complete evaluation, the sequence of
and protection. Protection of the surfaces is
stages in the investigation and analysis of fail-
particularly important if electron micro-
ure, as detailed in Ref 5, is as follows (Ref 2):
scopic examination is to be part of the pro-
1. Collection of background data and selection cedure (Ref 2). Appropriate packaging of
of samples failed components for shipping is equally
2. Preliminary examination of the failed part important. Wrapping them directly into a
3. Nondestructive and mechanical testing plastic bag, or placing pieces directly into
4. Selection, identification, preservation, and/ a plastic bottle or container, can intro-
or cleaning of specimens duce unwanted hydrocarbon contaminants.
Fingerprints on the failed surfaces can also surface coatings must be used to inhibit oxida-
introduce contamination (Ref 4); tion and corrosion of the fracture surface. The
Do not conduct a destructive testing without primary disadvantage of using these surface
considerable thought. Alterations such as coatings is that fracture surface debris, which
cutting, drilling, and grinding can ruin often provides clues to the cause of fracture, may
an investigation if performed prematurely. be displaced during removal of the coating.
Destructive testing must be performed only However, it is still possible to recover the sur-
after all possible information has been face debris from the solvent used to remove
extracted from the part in the original con- these surface coatings by filtering the spent
dition and after all significant features have solvent and capturing the residue. In regard to
been carefully documented by photography cleaning techniques, fracture surfaces exposed
(Ref 2). to various environments generally contain un-
wanted surface debris, corrosion or oxidation
Preliminary Examination of the Failed products, and accumulated artifacts that must be
Part. In addition to locating the failure origin, removed before meaningful fractography can
visual analysis is necessary to reveal stress con- be performed. Before any cleaning procedures
centrations, material imperfections, presence of begin, the fracture surface should be surveyed
surface coatings, case-hardened regions, welds, with a low-power stereobinocular microscope,
and other structural details that contribute to and the results should be documented with ap-
cracking. A careful macroexamination is neces- propriate sketches or photographs. Low-power
sary to characterize the condition of the fracture microscope viewing will also establish the
surface so that the subsequent microexamination severity of the cleaning problem and should also
strategy can be determined. Corrodents often be used to monitor the effectiveness of each
do not penetrate the crack tip, and this region subsequent cleaning step. It is important to
remains relatively clean. The visual macro- emphasize that the debris and deposits on the
analysis will often reveal secondary cracks that fracture surface can contain information that is
have propagated only partially through a crac- vital to understanding the cause of fracture. The
ked member. These part-through cracks can be most common techniques for cleaning fracture
opened in the laboratory and are often in much surfaces, in order of increasing aggressiveness,
better condition than the main fracture (Ref 3). are (Ref 3):
Nondestructive and Mechanical Testing.
Dry air blast or soft organic-fiber brush
A wide variety of nondestructive testing is
cleaning
available, including dye penetrant, ultrasonics,
Replica stripping
x-ray, and eddy current, which can help in
Organic-solvent cleaning
the failure analysis task in order to unveil
Water-based detergent cleaning
even subtle and/or internal defects in a part.
Cathodic cleaning
Mechanical property tests are also ready to use,
Chemical-etch cleaning
ranging from a sample hardness test to elevated-
temperature tensile and impact testing. These Macroscopic Examination and Analysis
tests are often used to determine if degradation is and Photographic Documentation. More
related to fabrication or to the service environ- often than not, the investigation starts with a
ment. Sometimes, a standard test can be adapted low-magnification, if any, observation of the
to simulate manufacturing or in-service condi- failed part. This visual examination can often
tions more closely (Ref 4). quickly answer questions such as: What was the
Selection, Identification, Preservation, mode of failure? Did it crack, or was there a
and/or Cleaning of Specimens. Unless a uniform or pitting corrosion failure? Did the
fracture is evaluated immediately after it is protective oxide film break down? Were the
produced, it should be preserved as soon as welds visibly contaminated? A variable magni-
possible to prevent attack from the environment. fication stereoscope equipped with a ring light
The best way to preserve a fracture is to dry it and directional fiberoptic lighting is a powerful
with a gentle stream of dry compressed air, then tool for macroscopic visual examination.
store it in a desiccator, a vacuum storage vessel, Contemporary stereoscopes can operate over a
or a sealed plastic bag containing a desiccant. range of 2.5 to 50 (Ref 4).
However, such isolation of the fracture is often Microscopic Examination and Analysis.
not practical. Therefore, corrosion-preventive Once the area of interest is isolated, a smaller
General Aspects of Failure Analysis / 113
portion can be cut from the sample and mounted an undisturbed state prior to analysis, because
for metallographic polishing and microscopic the culprit is often found in minute surface
examination. The microstructure of specimens features or traces of impurities. Fracture surfaces
may be enhanced by a wide variety of metallo- must remain untouched so that high-magnifica-
graphic techniques that include, for example, tion images can accurately determine the failure
heat tinting, stain etching, anodizing, and illu- mode. The sample must be removed carefully.
mination by bright-field and polarizing light. Important evidence can be destroyed by over-
Optical microscopic examination generally be- heating or by allowing adjacent fracture surfaces
gins at 50 magnification and continues through to fret or rub together during sectioning. The
1000 or even 1500 . Higher levels are best ideal method would be to unbolt the component
supplemented by differential interference con- or to provide adequate support so that a slow-
trast lighting, which allows theoretical resolu- speed saw can be used to cut out the component.
tion of features as fine as one-third of a However, sawing lubricants can mask or
micrometer. Features that are important to destroy residual chemicals or elements on the
recognize include the uniformity and size of the failed surface, so precautions become extremely
grain structure, the size distribution and shape of necessary. If the component has failed in the
intermetallic particles, and inclusions. Scanning middle of a large area, more aggressive cutting/
electron microscopy (SEM) is most useful sectioning techniques may be warranted, but
where extreme depth of focus and high magni- keep a good distance from the failed region
fications are needed. Fractures generally are (Ref 4).
complex, undulating surfaces that are difficult to Determination of Failure Mechanism
image, and an optical microscope can only focus (with Adapted Text from Ref 7). A thorough
on a very narrow region because of the very investigation should ensure that all damage is
shallow depth of field. However, the SEM excels found and documented, because multiple modes
at imaging fracture surfaces, and it can be and mechanisms may be present in most real-
operated in many different modes. The most world failure analyses. It is also important to
common mode is secondary electron imaging, recognize that many unique mechanisms may be
which provides a detailed, high-depth-focus driven by more than one environmental factor,
image that is easy to interpret. Backscattered such as stress, temperature, corrosion, wear,
Z contrast is used to identify regions of radiation, or electrical factors.
impurities within a matrix. High-atomic-number The term failure mechanism, or damage
species produce a light appearance, whereas mechanism, is meant to convey the specific
low-atomic-number species create a darker series of events that describe both how the
appearance. The topographic backscattered damage was incurred and the resulting con-
mode enhances the surface topography of the sequences. Examples of damage mechanisms
sample and accentuates height or elevation dif- include high-temperature creep, hydrogen
ferences on a fracture surface. The characteristic embrittlement, stress-corrosion cracking, and
x-rays can be detected and analyzed according to sulfidation. A failure or damage mechanism
their energy. This is called energy-dispersive describes how damage came to be present.
x-ray analysis. The x-ray wavelength corre- This definition of failure mechanism also
sponds to the presence of a specific element, and should not be confused with the description of
its amplitude corresponds to the quantity of the physical characteristics of damage observed.
such element. This technique allows quantita- For example, intergranular fracture, buckling,
tive characterization of elements within a given transgranular beach marks, and pits can all be
phase. Bulk chemistry is typically analyzed thought of as damage modes. The term damage
during failure analysis to verify conformance mode or failure mode is best used to describe
with industry-accepted chemical limits. In the what damage is present.
case of reactive metals, light elements can Much confusion has occurred because of
embrittle them due to improper processing or the tendency of engineers to use the terms
service conditions (Ref 4). mechanism and mode interchangeably; in doing
Selection, Preparation, Examination, and so, it is unclear that two distinct characteristics
Analysis of Metallographic Specimens. One need to be assessed. Sometimes this occurs
of the worst things that can happen to the sample because, within a given system, the same
is inadequate handling, examination, or pack- wording is used to describe both the failure
aging. It is imperative that the sample remains in mode and mechanism. For example, pitting
describes a damage mode because the surface of environments. What properties were re-
a material is pitted. In certain systems, pitting is quired during service? How were properties
also a possible damage mechanism. In boiler expected to change from service exposure?
tubing, for example, a pitting damage mecha- How was the part inspected during service
nism describes a specific localized corrosion intervals? What information was found
mechanism where pits form through dissolution during these inspections?
of metal either from low-pH or high-oxygen What material characteristics were specified
conditions. The metal under the pit surfaces is for the part (e.g., composition, strength,
unaffected. In this system, pitting is a specific hardness, impact, and stress-rupture proper-
damage mechanism, but many other damage ties)? What specifications, industry stan-
mechanisms also result in a pitting damage dards, and contracts govern these properties?
mode in boiler tubing, including hydrogen What were the various ways the part could
damage, phosphate corrosion, and caustic fail?
gouging.
It is helpful to be as specific as possible in The last item is a key question to repeatedly
differentiating damage mechanisms in a system. ask throughout a failure investigation. The list of
For example, fatigue is often identified as both various damage mechanisms by which a part can
a damage mode and a damage mechanism. A fail can be narrowed down through two basic
fatigue damage mode is the observable damage concepts (Ref 7). Limiting conditions that refine
that occurs under fatigue loading cycles (e.g., the scope of explanations for observed damage
the presence of beach marks). Classifying fati- can be defined by using the following two rules
gue as a damage mechanism is not necessarily of thumb:
complete because it does not point to the specific When the impossible is eliminated, whatever
environment that results in a fatigue damage remains, however improbable, must be
mode. Instead, specific mechanisms that can considered (Sherlock Holmes rule).
result in a fatigue damage mode must be When two or more explanations exist for a
examined. Examples include corrosion fatigue, sequence of events, the simple explanation is
thermomechanical fatigue, creep-fatigue inter- more likely to be the correct one (Occams
action, and mechanical fatigue. razor).
Determination of damage mechanisms starts
by characterizing the component(s) being ex- Chemical Analysis. In a failure investiga-
amined. It is impossible to know what is dif- tion, routine analysis of the material is usually
ferent about a failure without first understanding recommended. There are two main categories of
what is expected from unfailed components. chemical analysis that are often used by failure
In general, the analyst should obtain as much analysts:
information as possible about a part and its Bulk composition evaluation: often per-
background during the course of an investiga- formed in order to determine whether the
tion. Some key questions worth evaluating correct alloy was used in the subject com-
include: ponent
Microchemical analysis: to find evidence of
What was the part supposed to do? How was contamination, to evaluate the composition
it supposed to work? of microphases revealed on a metallographic
How was the part made? What processes specimen, or to evaluate corrosion products
were involved in its manufacture (e.g.,
forming, joining, and heat treatment)? What Often, chemical analysis is done last, because
properties were expected at the time of an analysis usually involves destroying a certain
manufacture? amount of material. There are instances where
What were the specified dimensions and the wrong material was used, under which con-
tolerances for the as-manufactured part? ditions the material may be the major cause of
How was the part installed? failure. In many cases, however, the difficulties
To what service environment(s) was the part are caused by factors other than material com-
exposed? Typical environments to examine position.
include operating temperatures, stresses Extreme care must be used in interpretation of
(steady state or slowly rising and cyclic), chemical analysis work performed as part of a
oxidizing/corrosive environments, and wear failure investigation. Minor deviations from
specified composition must not be interpreted as absorption are found in other metals and alloys.
the sole cause of a failure, without much addi- For example, excessive levels of nitrogen in
tional supporting evidence. In most instances, superalloys can lead to brittle nitride phases that
slight deviations from specified compositions cause failures of highly stressed parts.
are not likely to be of major importance in failure Various analytical techniques can be used
analysis. However, small deviations in alumi- to determine elemental concentrations and to
num content can lead to strain aging in steel, and identify compounds in alloys, bulky deposits,
small quantities of impurities can lead to temper and samples of environmental fluids, lubricants,
embrittlement. In specific investigations, parti- and suspensions. Semiquantitative emission
cularly where corrosion and stress corrosion spectrography, spectrophotometry, and atomic-
are involved, chemical analysis of any deposit, absorption spectroscopy can be used to deter-
scale, or corrosion product, or a substance with mine dissolved metals (as in analysis of an
which the affected material has been in contact, alloy), with wet chemical methods used where
is required to assist in establishing the primary greater accuracy is needed to determine the
cause of failure. concentration of metals. Combustion methods
Where analysis shows that the content of a ordinarily are used for determining the con-
particular element is slightly greater than that centration of carbon, sulfur, nitrogen, hydrogen,
required in the specifications, it should not be and oxygen.
inferred that such deviation is responsible for the Wet chemical analysis methods may be
failure. Often, it is doubtful whether such a employed for determining the presence and
deviation has played even a contributory part in concentration of anions such as Cl , NO3 , and
the failure. For example, sulfur and phosphorus S . These methods are very sensitive.
in structural steels are limited to 0.04% in X-ray diffraction identifies crystalline com-
many specifications, but rarely can a failure in pounds either on the metal surface or as a mass
service be attributed to sulfur content slightly in of particles and can be used to analyze corrosion
excess of 0.04%. Within limits, the distribution products and other surface deposits. Minor and
of the microstructural constituents in a material trace elements capable of being dissolved can be
is of more importance than their exact pro- determined by atomic-absorption spectroscopy
portions. An analysis (except a spectrographic of the solution. X-ray fluorescence spectro-
analysis restricted to a limited region of the graphy can be used to analyze both crystalline
surface) is usually made on drillings represent- and amorphous solids, as well as liquids and
ing a considerable volume of material and gases.
therefore provides no indication of possible Stress Analysis and Fracture Mechanics
local deviation due to segregation and similar Analysis. When confronted with a cracked,
effects. fractured, or deformed component, the failure
Also, certain gaseous elements, or inter- analyst will usually seek to answer some basic
stitials, normally not reported in a chemical questions:
analysis, have profound effects on the mechan-
ical properties of metals. In steel, for example, Were the loads and stresses encountered by
the effects of oxygen, nitrogen, and hydrogen the part at the level anticipated during
are of major importance. Oxygen and nitrogen design? Or did some unexpected condi-
may give rise to strain aging and quench aging. tion(s) contribute to the failure?
Hydrogen may induce brittleness, particularly Was the material in the area of the cracking
when absorbed during welding, cathodic clean- or deformation capable of meeting the con-
ing, electroplating, or pickling. Hydrogen is ditions anticipated during design? Was there
also responsible for the characteristic halos or some deficiency or discontinuity that con-
fisheyes on the fracture surfaces of welds in tributed to the failure, or was there a local
steels, in which instance the presence of hydro- stress raiser at the critical location? Was this
gen often is due to the use of damp electrodes. taken into account by the designer?
These halos are indications of local rupture In general, there are two types of conditions that
that has taken place under the bursting micro- may lead to structural failure:
stresses induced by the molecular hydrogen,
which diffuses through the metal in the atomic Net-section instability, where the overall
state and collects under pressure in pores and structural cross section can no longer sup-
other discontinuities. Various effects due to gas port the applied load
The critical flaw size (ac) is exceeded service conditions are fully known or under-
by some preexisting discontinuity or when stood. Corrosion failures, for example, are
subcritical cracking mechanisms (for exam- difficult to reproduce in a laboratory, and some
ple, fatigue, stress-corrosion cracking, or attempts to reproduce them have given mis-
creep) reach the critical crack size leading results. Serious errors can arise when
attempts are made to reduce the time required for
Failures due to net-section instability typi- a test by artificially increasing the severity of one
cally occur when a damage process such as of the factorssuch as the corrosive medium or
corrosion or wear reduces the thickness of a the operating temperature. Similar problems are
structural section. This type of failure can be encountered in wear testing.
evaluated by traditional stress analysis or finite On the other hand, when its limitations are
element analysis (FEA), which are effective clearly understood, the simulated testing and
methods in evaluating the effects of loading and statistical experimental design analysis of the
geometric conditions on the distribution of stress effects of certain selected variables encountered
and strain in a body or structural system. in service may be helpful in planning corrective
However, stress analyses by traditional action or, at least, may extend service life. Most
methods or FEA do not easily account for of the metallurgical phenomena involved in
crack propagation from preexisting cracks or failures can be satisfactorily reproduced on a
sharp discontinuities in the material. When a laboratory scale, and the information derived
preexisting crack or discontinuity is present, from such experiments can be helpful to the
the concentration of stresses at the crack tip investigator, provided the limitations of the tests
becomes asymptotic (infinite) when using the are fully recognized.
conventional theory of elasticity. In this regard, Analysis of All the Evidence, Formulation
fracture mechanics is a useful tool, because it is a of Conclusions, and Writing the Report.
method that quantifies stresses at a crack tip Before starting this final step, some questions
in terms of a stress-intensity parameter (K). must already be answered:
The fracture mechanics of cracking from a dis- Fracture surface:
continuity or crack in a statically loaded com- a. What is the fracture mode?
ponent has two possible situations: b. Is the origin of the fracture visible?
The crack reaches a critical length with rapid c. What is the relation between the fracture
(brittle) separation. direction and the normal or expected fra-
The crack blunts, redistributing the stress cture directions?
state, with continued loading creating a tear d. How many fracture origins are there?
zone (and sharpened crack-tip radius) in e. Is there evidence of corrosion, paint, or
front of the crack. In steels, this tear zone some other foreign material on the fracture
can then cause the critical crack length to be surface?
exceeded, such that unstable cleavage frac- f. Was the stress unidirectional or was it
ture occurs or unstable microscale ductile reversed in direction?
fracture is induced. The surface of a part:
a. What is the contact pattern on the surface
Which event occurs depends on the temperature of the part?
and the loading rate, but in either event, crack b. Has the surface of the part been deformed
propagation is unstable (i.e., does not require an by loading during service or by damage
increasing load after creation of the tear zone). after fracture?
Fracture mechanics is a tool to help evaluate the c. Is there evidence of damage on the surface
implications of preexisting discontinuities or of the part by manufacturing, assembling,
cracks. repairing, or service?
Testing under Simulated Service Con- Geometry and design:
ditions. During the concluding stages of an a. Are there any stress concentrations related
investigation, it may be necessary to conduct to the fracture?
tests that simulate the conditions under which b. Is the part intended to be relatively rigid,
failure is believed to have occurred. Often, or is it intended to be flexible, like a
simulated-service testing is not practical be- spring?
cause elaborate equipment is required, and even c. Does the part have a basically flawless
where practical it is possible that not all of the design?
d. How does the partand its assembly problem. The following questions should
work? also be considered:
e. Is the part dimensionally correct? a. Is there evidence that the mechanism was
Manufacturing and processing: overspeeded or overloaded?
a. Are there internal discontinuities or b. Is there evidence that the mechanism was
stress concentrations that could cause a abused during service or used under con-
problem? ditions for which it was not intended?
b. If it is a wrought metal, does it contain c. Did the mechanism or structure receive
serious seams, inclusions, or forging pro- normal maintenance with the recom-
blems, such as end grains, laps, or other mended materials?
discontinuities, that could have an effect d. What is the general condition of the
on performance? mechanism?
c. If it is a casting, does it contain shrinkage Environmental reactions: The problems
cavities, cold shuts, gas porosity, or related to the environment can arise anywhere
other discontinuities, particularly near the in the history of the part: manufacturing,
surface of the part? shipping, storage, assembly, maintenance,
d. If a weldment was involved, was the and service. None of these stages should be
fracture through the weld itself or through overlooked in a thorough investigation that
the heat-affected zone in the parent metal asks:
adjacent to the weld? If through the weld, a. What chemical reactions could have taken
were these problems something like gas place with the part during its history?
porosity, undercutting, underbead crack- b. To what thermal conditions has the part
ing, or lack of penetration? If through the been subjected during its existence?
heat-affected zone adjacent to the weld, Report writing: Finally, the report analyzing
how were the parent metal properties the failure should be written in a clear,
affected by the heat of welding? concise, logical manner. It should be clearly
e. If the part was heat treated, was the treat- structured with sections covering the fol-
ment properly performed? lowing (Ref 6):
Material properties: a. Description of the failed item
a. Are the mechanical properties of the metal b. Conditions at the time of failure
within the specified range, if this can be c. Background history important to the
ascertained? failure
b. Are the properties of the metal suitable for d. Mechanical and metallurgical study of the
the application? failure
c. Residual and applied stress relationship. e. Evaluation of the material quality
The residual-stress system that was within f. Discussion of any anomalies
the part prior to fracture can have a pow- g. Discussion of the mechanism or possible
erful effectgood or badon the perfor- mechanisms that caused the failure
mance of a part. h. Recommendations for the prevention of
d. What was the influence of adjacent parts future failures or for action to be taken with
on the failed part? similar pieces of equipment
e. Were fasteners tight? Irrelevant data should be omitted, and,
Assembly: depending on the nature of the problem and the
a. Is there evidence of misalignment of the data, not every report will need full treatments
assembly that could have had an effect on for every one of the sections listed previously.
the fractured part? Many times, the readership may include pur-
b. Is there evidence of inaccurate machin- chasing, operating, or accounting personnel
ing, forming, or accumulation of toler- who are not technically trained. If this is the
ances? situation, the report should be written so that it
c. Did the assembly deflect excessively under is comprehensible to these persons. At least,
stress? those sections of the report that bear on their
Service conditions: It is important to deter- decision-making or information needs should be
mine if there were any unusual occurrences, written in language that is accessible to them.
such as strange noises, smells, fumes, or Frequently, a cover letter summarizing the most
other happenings, that could help explain the important findings and the suggested action is a
good vehicle for reaching top executives who thickness) at and well away from the fracture can
are not as interested in the technical specifics but serve to define macrodeformation after assur-
need key findings and recommendations as a ance that both points of measurement had the
basis for decision making. Followup on the same dimension before fracture.
recommendations is frequently a difficult task Fracture-surface matching is also used to
but should be undertaken for the more critical determine the presence or absence of plastic
failures. Cooperation between the investigator, deformation. It is very important, however, to
the designer, the manufacturer, and the user is resist the temptation to fit the matching fracture
critical in developing good, workable changes. surfaces together, because this almost always
destroys (smears) microscopic features. The
fracture surfaces should never actually touch
Fracture during fracture-surface matching.
The origin of a fracture may be indicated by a
The process of fracture, in general terms, can discoloration or by the topography of the frac-
be described in terms of the mechanisms of ture surface. A discolored area on a fracture
crack initiation and/or crack extension (growth). surface may be produced by a preexisting crack
Different mechanisms may occur for crack whose surfaces have been corroded or oxidized.
initiation and the subsequent process of crack For example, the surfaces of a quench crack can
growth. For example, crack extension may occur be oxidized during a subsequent tempering heat
by the brittle mechanism of cleavage, even treatment; the oxide film gives a bluish-black
though extensive elongation accompanied or color to the surfaces of the crack. Topographical
preceded crack initiation. The fracture may be features that often reveal the origin of a fracture
classified as either ductile or brittle, depending are either chevron or river patterns or a set
on whether the mechanism is describing crack of diverging ledges. If the fracture surface is
initiation or crack growth, respectively. Like- essentially featureless, the presence of a shear
wise, the low-energy catastrophic fracture of a lip can be used to locate, within limits, the origin
high-strength aluminum alloy by microvoid of a fracture. For example, a shear lip is not
coalescence is also difficult to classify because, formed at the origin of a stress-corrosion crack,
although the fracture energy is low and failure but when the crack begins to propagate rapidly, a
initiates by fracture or decohesion of brittle shear lip is formed wherever the crack front
particles, the growth and coalescence of the exits from the interior to the free surface. Beach
microvoids occurs by plastic deformation. marks, which are associated with fatigue-
Another difficulty is that cleavage fracture may initiated fractures, also provide a definite indi-
be initiated by dislocation interactions that, by cation of the crack origin; however, it should be
definition, involve plasticity. This is why frac- noted that fracture surfaces having an appear-
tures are sometimes difficult to logically classify ance similar to that of the beach-mark pattern
(Ref 5). Therefore, it is helpful to be clear can be produced by stress corrosion.
whether fracture mechanisms are describing the Generally, cyclic loading produces only a
process of crack initiation or extension. Crack single crack, which is usually located at a site of
extension also can be multimode over time (e.g., stress concentration or of a metallurgical defect,
fatigue crack growth followed by overload). whereas additional cracks, formed indepen-
In terms of fracture appearances (or fracture dently of the main crack and at a distance from it,
modes, defined earlier in the section Deter- may be observed on the surface of a structural
mination of Failure Mechanism in this chap- or machine component subjected to corrosion
ter), a general summary of the visual and fatigue or stress corrosion.
microscopic aspects of fracture surfaces for On the microscopic level, striations on the
metallic materials is provided in Table 1 (Ref 8). fracture surface are unique to fatigue, and the
Several analytical procedures are available for crack path, although normally transgranular, can
distinguishing among the various types of frac- be intergranular. For example, intergranular
ture. For example, the presence or absence of fatigue cracking can occur in the case of a car-
plastic macrodeformation can be determined burized steel or in a material that has a high
with the unaided eye or by use of a steel scale, a density of second-phase particles at the grain
machinists micrometer, or a machinists or boundaries.
measuring microscope. Differences in some Corrosion-fatigue and stress-corrosion cracks
dimensional attribute of parts (such as width or may propagate transgranularly, intergranularly,
Table 1 Fracture mode identification chart

Instantaneous failure mode(a) Progressive failure mode(b)
Method Ductile overload Brittle overload Fatigue Corrosion Wear Creep
Visual, 1 to Necking or Little or no Flat progressive General Gouging, Multiple brittle-
50 distortion in distortion zone with beach wastage, rough- abrasion, appearing fissures
(fracture direction Flat fracture marks ening, pitting, or polishing, External surface
surface) consistent with Bright or coarse Overload zone trenching or erosion and internal
applied loads texture, consistent with Stress-corrosion Galling or storing fissures contain
Dull, fibrous crystalline, applied loading and hydrogen in direction of reaction scale
fracture grainy direction damage may motion coatings
Shear lips Rays or Ratchet marks create multiple Roughened areas Fracture after
chevrons where origins cracks that with compacted limited
point to origin join appear brittle powdered debris dimensional
(fretting) change
Smooth gradual
transitions in
wastage
Scanning Microvoids Cleavage or Progressive Path of penetra- Wear debris and/or Multiple
electron (dimples) intergranular zone: worn tion may be abrasive can be intergranular
microscopy, elongated fracture appearance, irregular, characterized as to fissures covered
20 to in direction of Origin area may flat, may show intergranular, or morphology and with reaction scale
10,000 loading contain an striations at a selective phase composition Grain faces may
(fracture Single crack with imperfection magnification attacked Rolling-contact show porosity
surface) no branching or stress above 500 EDS may fatigue appears
Surface slip band concentrator Overload zone: help identify like wear in early
emergence may be either corrodent(c) stages
ductile or brittle
Metallographic Grain distortion Little distortion Progressive General or May show Microstructural
inspection, and flow near evident zone: usually localized surface localized change typical of
50 to 1000 fracture Intergranular or transgranular attack (pitting, distortion at overheating
(cross Irregular, transgranular with little cracking) surface consistent Multiple inter-
section) transgranular May relate to apparent Selective phase with direction of granular cracks
fracture notches at distortion attack motion Voids formed on
surface or brittle Overload zone: Thickness and Identify embedded grain boundaries
phases internally may be either morphology of particles or wedge-shaped
ductile or brittle corrosion scales cracks at grain
triple points
Reaction scales
or internal
precipitation
Some cold flow
in last stages of
failure
Contributing Load exceeded the Load exceeded Cyclic stress Attack morphol- For gouging or Mild overheating
factors strength of the part the dynamic exceeded the ogy and alloy abrasive wear: and/or mild
Check for proper strength of the endurance limit type must be check source of overstressing at
alloy and proces- part of the material evaluated abrasives elevated
sing by hardness Check for proper Check for proper Severity of Evaluate effec- temperature
check or destruc- alloy and strength, surface exposure tiveness of lubri- Unstable micro-
tive testing, processing as finish, assembly, conditions may cants structures and
chemical analysis well as proper and operation be excessive; Seals or filters may small grain size
Loading direction toughness, grain Prior damage by check: pH, have failed increase creep
may show failure size mechanical or temperature, Fretting induced rates
was secondary Loading corrosion modes flow rate, by slight looseness Ruptures occur
Short-term, direction may may have dissolved in clamped joints after long
high-temperature, show failure was initiated oxidants, elec- subject to exposure times
high-stress rupture secondary or cracking trical current, vibration Verify proper
has ductile impact induced Alignment, metal coupling, Bearing or materi- alloy
appearance Low vibration, aggressive als engineering
(see creep) temperatures balance agents design may reduce
High cycle low Check bulk or eliminate
stress: large composition and problem
fatigue zone; contaminants Water
low cycle high contamination
stress: small High velocities
fatigue zone or uneven flow
distribution,
cavitation
(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.
or by a combination of both modes. A distin- The fracture process begins in the center of
guishing feature of stress corrosion is the the testpiece with microvoid nucleation along
branching of the main crack. If corrosion pits grain boundaries or from interfaces such as those
or corrosion products are found only on the found in base metal/inclusions boundaries. As
slow-growth region of a fracture surface, the the applied stress increases, microvoids grow
environment was in all probability sufficiently and coalesce, forming a crack in the center of the
corrosive to affect the fracture mechanism. part. This process, depicted in Fig. 2, ends up in
However, if evidence of corrosion is found on rapid crack propagation by shearing of the
both the slow-growth and fast-growth areas, remaining ligament of the neck region, at an
some corrosion took place subsequent to frac- angle of 45 in relation to the loading direction.
ture, and the environment may or may not have It is important to emphasize that a cup-cone
influenced fracture. geometry will depend on the geometry and
dimensions of the part and mechanical proper-
ties of the material. Thin sheets, for instance,
Ductile Fracture present neck formation and a fracture surface
Ductile fracture takes place when a material oriented at an angle of 45 in relation to the
capable of undergoing plastic deformation is applied load, as observed in Fig. 3. Ductile
subjected to stresses that culminate in its rup- fracture takes place intergranularly, unless some
ture. Macroscopically, the ductile fracture pro- sort of mechanism weakens the grain bound-
cess presents some peculiarities that allow it to aries. The microscopic aspect of the fracture
be identified immediately. The first feature is the surface consists of several small elliptical
presence of plastic deformation that may be cavities, or microvoids, as depicted in Fig. 4.
accompanied by neck formation. In tensile
testpieces of ductile materials, besides necking, Brittle Fracture
the fracture surface presents a fibrous aspect and
a cup-cone geometry, as seen in Fig. 1. 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. 1 Ductile fracture showing the typical cup-cone geo-

metry
Microvoids
Fig. 2 Schematic representation of the cup-cone geometry Fig. 4 Microvoids on the fracture surface of AA6061-T1
formation during the ductile fracture process tensile testpiece
associated with materials of high strength and that the superposition of high hydrostatic stres-
low ductility or materials that were subjected ses on the material reduces the triaxiality levels,
to an embrittlement process. The crack, once increasing ductility. High applied loading rates
nucleated, propagates very quickly in a direction are likely to make plastic deformation more
perpendicular to the applied load. Figure 5 pre- difficult because shearing processes are time-
sents an example of a gray cast iron testpiece that dependent, resulting in brittle behavior.
presented brittle fracture. Crack propagation by brittle fracture can
Besides the mechanical properties, several occur across the grains (transgranular) or
other factors may result in a brittle behavior, along the grain boundaries (intergranular). In
such as temperature, loading rate, presence of the transgranular mode, the fracture process
stress concentrators, and dimensions. Low tem- takes place by cleavage along specific crystal-
peratures tend to reduce the ductility of metals, lographic planes. Figure 7 presents cleavage
especially those possessing a body-centered regions in a microalloyed low-carbon steel,
cubic structure, resulting in a typically brittle which can be identified by flat regions on the
fracture. Figure 6 shows that as the temperature fracture surface. Additionally, it is worth men-
drops, the brittle aspect on the fracture surface of tioning that most parts of steels will present
impact testpieces increases. The presence of alternate regions consisting of cleavage areas
stress raisers or larger dimensions introduces a and microvoids, evidencing a mixed mode of
more severe triaxial stress state within the crack propagation.
material, and thus, there is larger probability that In another situation, fracture can take place
brittle fracture will occur. However, it is known intergranularly, because the grain boundary is a
Fig. 5 Tensile testpiece of gray cast iron presenting brittle

fracture
Fig. 7 (a) Cleavage region observed in low-carbon steel.

(b) Magnification of the region delimited by the rec-
Fig. 6 Fracture surfaces of SAE 4140 impact testpieces. tangle in (a) showing an inclusion in the center of the cleavage
Tested at room temperature, right, and at 196 C, left region
weaker path for crack propagation. Normally, life of a component during both the nucleation
this fracture mode will occur when some and crack propagation phases.
embrittlement process resulted in grain bound- Fatigue Crack Initiation. Generally, fatigue
aries being more susceptible to crack propaga- cracks are initiated at free surfaces, where there
tion than the core of the grain, such as an is no constraint to material deformation; how-
unsuitable heat treating or by environmental ever, in some cases, cracks may be initiated in
factors. Figure 8 presents an example of inter- the interior of the material where interfaces are
granular brittle fracture in an austenitic stainless present, such as the interface of a carburized
steel SAE 316L, where grain boundaries can surface layer and the base metal or the interface
clearly be observed on the fracture surface. 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
Fatigue Fracture compressive residual stresses were introduced
According to the definition given by ASTM by shot peening or surface rolling.
E1823, fatigue is the process of progressive One of the classic models of fatigue crack
localized permanent structural change occurring nucleation considers that when a material is
in a material subjected to conditions that pro- under loading (monotonic or cyclic), slips occur
duce fluctuating stresses and strains at some at the high-shear-stress planes, creating steps on
point or points and that may culminate in cracks the material surface. Under cyclic loading,
or complete fracture after a sufficient number of the formation of intrusions and extrusions is
fluctuations. A material subjected to fatigue observed, as schematically represented in Fig. 9.
can fracture at applied stresses much lower than Slip band intrusions are excellent stress raisers
those necessary to fracture the same material that can be sites of crack nucleation.
under monotonic conditions. The fluctuating Besides the applied stress amplitude, DS/2,
stresses can be originated from mechanical, several other factors are likely to affect the
thermal, or vibration loading conditions, and the nucleation of a fatigue crack, such as the mean
phenomenon is responsible for more than 80% stress, Sm, or load ratio, R; geometry and surface
of mechanical failures of components. For more finishing of the part; mechanical properties; and
than 150 years, the study of metals fatigue environment. Here, the R ratio is defined as the
has involved engineers, physicists, chemists, ratio between the minimum and maximum loads
and mathematicians, and everyday this study during the fatigue cycle.
becomes more and more complex and impor- A large proportion of fatigue data found in the
tant. The theory about fatigue is extremely vast, literature refers to tests conducted at Sm = 0,
and for each question answered, another one, that is, for a load ratio R = 1. However, in
more instigating, appears, requiring a broad many engineering situations, the fluctuating
knowledge of materials science. In the following stresses are superimposed to a static stress.
topics, a brief overview is given about the main Larger mean stresses reduce the nucleation time
mechanisms and factors influencing the fatigue 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 Fig. 9 Schematic representation of an intrusion formation on
fracture the surface of a metallic material
The mechanism proposed in Fig. 9 is employed to describe the actual stress state in
adequate to explain the initiation of cracks materials containing residual stresses. There-
on polished testpieces or components without fore, the effective stress, S0 , is given by the
the presence of geometric discontinuities. sum of the applied stress, S, to the residual stress,
However, in engineering components, there are Sres:
several stress concentrators, such as scratches,
notches, machining marks, corrosion pits, and S0 =S+Sres (Eq 1)
microconstituents such as grain boundaries,
triple points, and inclusions, that individually or Similarly, the effective minimum and maximum
synergistically can reduce the initiation time. stresses are defined, respectively, as:
Since the initiation depends essentially on
plastic deformation mechanisms, high-strength
S0max =Smax +Sres (Eq 2)
materials normally present a higher resistance to
fatigue crack nucleation. In this sense, several
surface-hardening treatments are employed to S0min =Smin +Sres (Eq 3)
selectively reinforce the material, aiming to
retard crack initiation and therefore to increase Consequently, the effective stress amplitude,
fatigue life. mean stress, and load ratio are given, respec-
The chemical composition and/or the micro- tively, by:
structure of the surface can be modified by
thermochemical treatments, such as carburizing DS0 S0max 7S0min (Smax +Sres )7(Smin +Sres )
= =
or nitriding, or by cold deformation processes, 2 2 2
such as shot peening or surface rolling. Smax 7Smin DS
= = (Eq 4)
Mechanical parts that necessarily present stress 2 2
concentrators, such as crankshafts, gears, and
bolts, can be subjected to these treatments S0max +S0min (Smax +Sres )+(Smin +Sres )
S0m = =
to increase the fatigue limit of the material. 2 2
Figure 11 shows a micrograph of the transverse Smax +Smin
= +Sres =Sm +Sres (Eq 5)
section of a bolt, where the thread was cold 2
formed by surface rolling. As a consequence,
surface grains are flattened due to the mechan- S0min S +Sres
R0 = = min (Eq 6)
ical deformation imposed. In this case, besides S0max Smax +Sres
increasing hardness and mechanical strength,
the process avoids the introduction of harmful Therefore, the presence of a residual-stress
machining marks. field does not affect the stress amplitude but
Surface treatments may also increase fatigue affects the mean stress and the load ratio.
life by the introduction of compressive residual A compressive residual stress reduces the mean
stresses on the surface of the material. As long as stress and the load ratio, increasing the number
the material remains in linear elastic conditions,
the principle of stress superposition can be
Increasing
Sm
S/2
Nf
Fig. 11 Optical micrograph of the transverse section of a
thread fillet machined by surface rolling. The ma-
Fig. 10 Mean stress effect on S-N fatigue curves terial consists of duplex stainless steel
of cycles for crack nucleation and vice versa. peening and surface rolling, contribute to the
In some situations, where high surface com- increase in the number of microstructural
pressive residual stresses are found, such as in barriers per unit of length due to the flattening of
materials subjected to surface-hardening treat- the grains.
ments, a crack may initiate below the surface, When the stress-intensity factor, K, increases
where the compressive residual-stress level is as a consequence of crack growth or higher
lower. An example of subsurface crack nuclea- applied loads, slips start to occur in different
tion is observed in Fig. 12 for a surface-rolled planes close to the crack tip, initiating stage II of
ductile cast iron subjected to bending-rotating propagation. While stage I of propagation is
fatigue. orientated 45 in relation to the applied load,
Fatigue Crack Propagation. Basically, fati- propagation in stage II is perpendicular to load
gue crack propagation can be divided into three direction, as depicted in Fig. 13. An important
stages: stage I (short cracks), stage II (long characteristic of stage II propagation is the
cracks), and stage III (final fracture). presence of ripples on the fracture surface,
A fatigue crack, once initiated, propagates known as striations, which are only visible with
along high shear-stress planes (45 ), as sche- the aid of a scanning electron microscope. Not
matically represented in Fig. 13. This is known all engineering materials exhibit striations. They
as stage I or the short crack growth propagation are clearly seen in pure metals and many ductile
stage. The crack propagates until it is deceler- alloys, such as aluminum alloys. In steels, they
ated by a microstructural barrier, such as a grain are frequently observed in cold-worked alloys.
boundary, inclusions, or pearlitic zones, that Figure 14 shows examples of fatigue striations
cannot accommodate the initial crack growth in an interstitial-free steel and in aluminum
direction. Therefore, grain refinement is capable alloys. The most accepted mechanism for the
of increasing fatigue strength of the material due formation of striations on the fatigue fracture
to the insertion of a large quantity of micro- surface of ductile metals (Ref 9) consists of
structural barriers, that is, grain boundaries, that successive blunting and resharpening of the
must be overcome in stage I of propagation. crack tip, as represented in Fig. 15.
Surface mechanical treatments, such as shot 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 micro-
structure, load ratio, and stress state (plane-
stress 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 suc-
cessive arrests or decrease in the fatigue crack
growth rate due to a temporary load drop or to an
Fig. 12 Probable subsurface crack nucleation site in a sur-
overload that introduces a compressive residual-
face-rolled ductile cast iron testpiece tested under
bending-rotating conditions stress field ahead of the crack tip.
The other region corresponds to the final
fracture and presents a fibrous and irregular
Stage I Stage II aspect. In this region, the fracture can be either
brittle or ductile, depending on the mechanical
properties of the material, dimensions of the
Surface
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,
Fig. 13 Stages I and II of fatigue crack propagation as depicted in Fig. 16(a). On the other hand,
if lower loads are applied, the fatigue crack

(a)
will have to grow longer before the applied
stress-intensity factor, K, reaches the fracture
toughness value of the material, resulting in a (d)
smaller area of fast fracture (Fig. 16b). (b)
Ratcheting marks are another macroscopic
feature that can be observed in fatigue fracture (e)
surfaces. These marks originate when multiple
(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 den- Fig. 16 Fatigue fracture surface. (a) High applied load.
drites on the surface. Arrow at top right indicates fatigue striations. (b) Low applied load
cracks, nucleated at different points, join to- Near the threshold stress-intensity factor,
gether, creating steps on the fracture surface. DKth, the effects of R ratio are mainly attributed
Therefore, counting the number of ratchet marks to crack closure effects, where crack faces come
is a good indicator of the number of nucleation in contact at an applied Kcl that is higher than the
sites. Figure 17 presents in detail some ratchet minimum applied stress-intensity factor, Kmin.
marks found on the fracture surface of a large Several different mechanisms may contribute
SAE 1045 rotating shaft, fractured by fatigue. to premature crack closure. One of them consists
Similar to the initiation phase, many factors of plasticity-induced closure, represented in
can affect long fatigue crack propagation rates. Fig. 19(a). As the crack grows, the material that
Among them, special attention should be given has been previously permanently deformed
to the effects of load ratio and the presence of within the plastic zone now forms an envelope of
residual stresses. plastic zones in the wake of the crack front. This
Increasing the load ratio has a tendency to leads to displacements normal to the crack sur-
increase the long crack growth rates in all faces as the restraint is relieved. This is no pro-
regions of the fatigue crack growth rate versus blem while the crack is open; however, as the
applied stress-intensity factor range curve, or load decreases, the crack surfaces touch before
simply, da/dN versus applied DK curve. Gen- the minimum load is reached, shielding the
erally, the effect of increasing load ratio is less crack. This type of premature contact can also
significant in the Paris regime than in near- occur due to the crack-wake roughness and
threshold and near-failure regions (Fig. 18). 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 con-
sidered closure-free and Kmax-independent, and
Fig. 17 Ratcheting marks, indicated by the arrows, in an SAE
1045 shaft fractured by fatigue
Plastic deformation
Plastic zone
envelope
Crack tip
Final failure
(a)
Increasing
da/dN Paris regime R
(b)
Near threshold Premature contact points

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 Fig. 19 Crack closure mechanisms induced by (a) plasticity,
curves. (b) roughness, and (c) oxide
the crack growth rates are generally very similar 0 j0, then:
If Kmin
for tests conducted under different R ratios. Near
the final failure, the effects of R ratio are related R0 =0 (Eq 11)
to the higher monotonic fracture component as
Kmax approaches KIc. Therefore, for the same DK 0 =Kmax
0
=Kmax +Kr (Eq 12)
applied DK, Kmax values are higher for tests
conducted under higher applied R ratios, and It is important to note that these equations
consequently, da/dN values are higher. assume that the part of the fatigue cycle during
The effects of residual stress on fatigue crack which the crack is closed at its tip (i.e., K050)
growth are related to alterations in the R ratio makes no contribution to crack growth.
and in the applied DK. In other terms, the resi-
dual stresses affect the two parameters that
control the crack driving force, that is, Kmax and Distortion
DKeff. When a crack is introduced in a plate
subjected to a residual-stress field, a residual Distortion is the least serious mode of failure,
stress-intensity factor, Kr, arises that can either but it can lead a part to failure or a structure to
decrease or increase the crack driving force collapse. It is easy to recognize but very difficult
parameters. to prevent. This is due to the fact that distortion
The superposition principle can also be does not involve the part itself but its use and
applied in terms of the stress-intensity factor, design. There are four reasons for distortion:
provided that the material remains linearly yielding, buckling, creep, and residual stresses.
elastic. In this sense, Kr can be added to Kmax and Yielding. When a load is put on a part, and it
Kmin: causes the part to be permanently distorted, it is
unable to perform the intended function and
0
Kmax =Kmax +Kr (Eq 7) therefore must be considered failed. In a well-
designed part, the stresses never exceed the yield
0
Kmin =Kmin +Kr (Eq 8) point, and the part deforms only elastically; that
is, when the load is released, the part returns to
As a result, R0 and DK0 are defined as follows. If its original dimensions.
0 40, then:
Kmin In a good design, the part operates in the
elastic range, that is, below yielding point;
0
Kmin K +Kr beyond this, the part will be permanently
R0 = 0
= min (Eq 9) deformed, and greater loads will cause the part
Kmax Kmax +Kr
to actually break. This point is considered to be a
very basic point to design and applies when the
DK 0 =Kmax
0 0
7Kmin =Kmax +Kr 7Kmin +Kr
load on a part is applied in a quasi-static way,
=Kmax 7Kmin =DK such as the load on a building structure or the
(Eq 10) stress in the legs of a desk. A ductile failure is
Kmax
Kmax
Kap=Keff
K Keff K
Kap
Kmin
Kcl Kcl
Kmin
Time Time
(a) (b)
Fig. 20 Load ratio effect on DKeff in a fatigue cycle. (a) Kmin5Kcl. (b) Kmin4Kcl
one where there is a great deal of distortion of the distortion that a part can no longer perform its
failed part. Commonly, a ductile part fails when intended function. The two general types of
it distorts and can no longer carry the needed creep processes are grain-boundary sliding and
load. However, some ductile parts break into voids at grain boundaries (cavitation creep).
two pieces and can be identified because there is The creep processes are easily identified by
a great deal of distortion around the fracture the local ductility and large numbers of inter-
face, similar to what would happen if too much is granular cracks that will depend on the tem-
placed load on a low-carbon steel bolt. perature and strain rate imposed. In general, a
Buckling. The failure of an engineering high strain rate combined with high temperature
component is not always caused by materials results in ductile fracture, followed by a large
fracture. In many occasions, the component elongation and neck formation. Additionally,
distortion may be sufficient to put it out of the grains near the fracture surface tend to be
function. The distortion can be elastic or plastic. elongated. On the other side, the combination of
The elastic distortions are temporary; however, low strain rate and high temperature results in
they may be sufficient to cause interference on intergranular brittle fracture, with low elonga-
the mobile parts. The plastic distortion is per- tion or necking. Intergranular fracture in such
manent and can be a result of an overload or conditions normally initiates by grain-boundary
creep deformation. The overload causes per- sliding from triple points or at grain-boundary
manent plastic deformation when the material intersections with second-phase particles, caus-
yield limit is overcome. This may happen in the ing cavities on the material microstructure, as
presence of stress concentrators, high tempera- presented in Fig. 22.
ture, inadequate heat treatment, or incorrect Once the crack nucleates, it propagates by
materials selection for the component applica- grain boundaries, and given that some sig-
tion. Compressive overloads may lead the nificant plastic deformation may take place,
material to overcome the buckling strength the fracture surface tends to exhibit grains of
limit, such as the one shown in Fig. 21 for an equiaxial shape. Therefore, to increase creep
aluminum part. The buckling strength is essen- strength, the material is normally heat treated to
tially a design problem (not metallurgical), and increase the grain size, reducing the ratio
the load depends on the dimensions of the part between the grain surface area and volume. In
and the Youngs modulus of the material (the turbines that work at very high temperatures, the
only materials factors involved). creep mechanism must be considered. In this
Creep is a time-dependent phenomenon that case, the component may be produced from
causes a part failure if it is under both quasi- monocrystals that significantly increase the
static load and temperatures higher than 0.3 Tm creep resistance.
(absolute melting temperature). Creep strain Most creep curves show three distinct stages
may produce sufficiently large deformation or (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) (b)
Fig. 22 Intergranular crack formation at high temperature by

grain-boundary sliding at (a) triple points and
Fig. 21 Aluminum part that suffered buckling (b) inclusions
Residual stresses can play a significant role Adhesive wear has been commonly identified
in explaining or preventing failure of a compo- by the terms galling or seizing. It is caused by
nent. One example of residual stresses prevent- the material transference from one surface to
ing failure is the use of shot peening processes another during their relative movement due to a
that increase the fatigue life of a component by solid-state welding process. Figure 24 shows a
inducing surface compressive stresses. schematic representation of this process. High
Unfortunately, there are also processes or contact pressure among the surface roughness
processing errors that can induce excessive results in local plastic deformation and points
tensile residual stresses in locations that may of microwelding. The movement between the
promote failure of a component. The internal surfaces causes the rupture of the junctions,
state of stress is caused by thermal and/or resulting in a rough peak in one surface and a
mechanical processing of the parts. Common valley on the other. Eventually, the tip of a peak
examples of these are bending, rolling, or may break, and an abrasive particle is formed.
forging a part. Thermal residual stresses are Abrasive wear, or abrasion, is caused by the
primarily due to differential expansion when a displacement of material from a solid surface
metal is heated or cooled. Two control factors due to hard particles or protuberances sliding
are thermal treatment (heating or cooling) along the surface. The particles may be found
and restraint. Both the thermal treatment and free between two surfaces or attached to one of
restraint of the component must be present to them, and the wear level depends on the relative
generate residual stresses. Residual stresses hardness between the particle and the surface
can result in visible distortion of a component. (Fig. 25). The abrasion may also happen due to
However, in the case of residual stresses, the the protuberances or sharp asperities on one of
distortion can also be useful in estimating the the surfaces in contact. The process of abrasive
magnitude or direction of these stresses. erosion may be considered as abrasive wear.
Erosion, or erosive wear, is the loss of
material from a solid surface due to relative
Wear-Assisted Failure motion in contact with a fluid that contains solid
particles. In this case, the particle is found to be
Wear may be defined as damage to a solid dispersed in a fluid or gas means, and it reaches
surface caused by the removal or displacement the surface under relatively high velocity
of material by the mechanical action of a con- (Fig. 25d). Figure 26 shows the microstructure
tacting solid, liquid, or gas. It may cause sig- of the transversal section of an H11 tool steel
nificant surface damage, and the damage is that has been subject to abrasive erosion.
usually thought of as gradual deterioration. Fatigue wear can be characterized by the
While the terminology of wear is unresolved, formation of cracks superficially and/or sub-
the following categories are commonly used: superficially and the removal of posterior
adhesive wear, abrasive wear, erosive wear, material due to cyclic loading of solid surfaces.
fretting, cavitation, rolling, contact fatigue, and The sliding contact and/or rolling between solids
corrosive wear.
Adhesion
Fracture X

t = creep rate
0 Stage I Stage II Stage III Particle
Time
Fig. 24 Transference mechanism of a material from one
Fig. 23 Schematic strain-time curve at constant load and surface to another and the formation of an abrasive
temperature showing the three stages of creep particle in the process of adhesive wear
or the repetitive impact of solids and/or liquids Fretting fatigue is considered a phenomenon
in a surface are responsible for the superficial where the damage is introduced by a conjunction
fatigue. When two surfaces of this nature inter- of events consisting of adhesion, oscillatory
act due to load application, the area effectively movement of very low amplitude, oxidation, and
in contact may be very small, resulting in high abrasion. The small oscillatory movements may
compressive and shear stresses that may lead to cause points of adhesion on the surface that
crack nucleation. If only rolling is present, the eventually break, forming oxidized particles that
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 in-
service 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.
Fig. 27 Schematic representation of contact fatigue under

pure rolling between two surfaces
(a) (b)
(c) (d)
Fig. 25 Abrasive wear. (a) Free particle between two sur-

faces. (b) Particle attached to one of the surfaces.
(c) Sharp asperity. (d) Erosion
Fig. 28 Damage by contact fatigue in rolling combined with

sliding conditions in gears produced from a quen-
Fig. 26 Fractography showing an H11 tool steel that has ched and tempered AISI 8620 carburized steel. (a) Transversal
suffered abrasive erosion section. (b) Frontal view from a formed cavity
act as abrasives on the surface, since the small- metallic and nonmetallic coatings are regularly
amplitude movements avoid their dispersion used to protect metal parts from corrosion.
apart from the source point. Figure 29 presents a Corrosion may result in failure of the com-
micrograph from a plasma nitrided Cr-Mo-V ponent. Several factors should be considered
steel, where a microcrack formed in the fretting during a failure analysis to determine the effect
region. corrosion played in a failure, such as type of
More than one mechanism can be responsible corrosion, corrosion rate, the extent of the cor-
for the wear observed on a particular part. The rosion, and the interaction between corrosion
most critical function provided by lubricants is and other failure mechanisms.
to minimize friction and wear to extend equip- Uniform, pitting crevice, galvanic, and stress-
ment service life. Gear failures can be traced to corrosion cracking are the most common types
mechanical problems or lubricant failure. of corrosion. Uniform corrosion is characterized
Lubricant-related failures are usually traced to by corrosive attack proceeding evenly over the
contamination, oil film collapse, additive entire surface area or a large fraction of the total
depletion, and use of improper lubricant for the area. General thinning takes place until failure.
application. The most common failures are due On the basis of tonnage wasted, this is the most
to particle contamination of the lubricant. Dust important form of corrosion.
particles are highly abrasive and can penetrate Stress-corrosion cracking necessitates a
through the oil film, causing plowing wear or tensile stress, which may be caused by residual
ridging on metal surfaces. Water contamination stresses, and a specific environment to cause
can cause rust on working surfaces of gears and progressive fracture of a metal. Aluminum
eventually destroy metal integrity. To prevent and stainless steel are well known for stress-
premature failure, gear selection requires careful corrosion cracking problems. However, all
consideration of the following: gear tooth geo- metals are susceptible to stress-corrosion crac-
metry, tooth action, tooth pressures, construc- king in the right environment.
tion materials and surface characteristics, Pitting corrosion is a localized form of cor-
lubricant characteristics, and operating environ- rosion by which cavities or holes are produced
ment. 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
Environmentally Assisted Failure the pits. A small, narrow pit with minimal
overall metal loss can lead to the failure of
Corrosion is chemically induced damage an entire engineering system. Pitting corrosion,
to a material that results in deterioration of which, for example, is almost a common
the material and its properties. Corrosion denominator of all types of localized corrosion
can seldom be totally prevented, but it can be attack, may assume different shapes.
minimized or controlled by proper choice of Crevice corrosion is a localized form of
material, design, coatings, and occasionally corrosion usually associated with a stagnant
by changing the environment. Various types of 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 chem-
istry within the crevice.
Galvanic corrosion (also called dissimilar-
metal corrosion or, wrongly, electrolysis) refers
to corrosion damage induced when two dis-
similar materials are coupled in a corrosive
electrolyte. It occurs when two (or more) dis-
similar metals are brought into electrical contact
under water. When a galvanic couple forms, one
Fig. 29 Fretting fatigue at the surface of a Cr-Mo-V steel of the metals in the couple becomes the anode
and corrodes faster than it would all by itself, Analysis, Failure Analysis and Prevention,
while the other becomes the cathode and cor- Vol 11, Metals Handbook, 9th ed., American
rodes slower than it would alone. Society for Metals, 1986, p 1546
6. G.F. Vander Voort, Conducting the Failure
Examination, Prac. Fail. Anal., Vol 1 (No 2),
REFERENCES April 2001, p 1446 and Failure Analysis and
Prevention, Vol 11, ASM Handbook, ASM
1. D. Dennies, How to Organize a Failure International, 2002
Investigation, ASM International, 2005 7. A. Tanzer, Determination and Classification
2. D.J. Wulpi, Chapter 1: Techniques of Failure of Damage, Failure Analysis and Prevention,
Analysis, Understanding How Components Vol 11, ASM Handbook, ASM International,
Fail, 2nd ed., ASM International, 2000, 2002
p 111 8. G. Powell, Identification of Types of Failure,
3. C.R. Brooks and A. Choudhury, Chapter 1: Failure Analysis and Prevention, Vol 11,
Introduction, Metallurgical Failure Analysis, Metals Handbook, 9th ed., American Society
McGraw-Hill, 1993, p 172 for Metals, 1986, p 7581
4. R. Graham, Strategies for Failure Analysis, 9. C. Laird, The Influence of Metallurgical
Adv. Mater. Process. Aug 2004, p 4550 Structure on the Mechanisms of Fatigue
5. D.A. Ryder, T.J. Davies, I. Brough, and F.R. Crack Propagation, Fatigue Crack Propa-
Hutchings, General Practice in Failure gation, STP 415, ASTM, p 131168
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 topics that are discussed in the case studies
forgings may occur for a variety of reasons. The include:
starting material may be of insufficient quality to
Validity checks for buster and blocker
be adequately formed without cracking, or the
design
forging process may introduce various types of
Lubrication and wear
discontinuities that cause failure during services.
For example, well-known forging-related dis- Mechanical surface phenomenon
Forging process design
continuities include:
Forging tolerances
Laps
As case studies were being selected, each of
Bursts
the aforementioned supporting topics was
Flakes
Segregation reviewed for any impact that particular study
had on the case being examined. It is a well-
Cavity shrinkage
known fact that forging solutions have several
Centerline pipe
possible avenues to follow. There is no unique
Parting-line grain flow
theory in plasticity that leads to the solution.
Inclusions
Most of the work reported here was performed
Forging discontinuities are discussed in more using the minimum amount of energy to create
detail in the texts on forging (Ref 14). the particular product. Factors unrelated to the
This article describes six case studies of deformation process, such as chemistry, micro-
failures with steel forgings (summarized in structure, phase, grain size, segregation, and
Table 1). The case studies illustrate difficulties prior strain history, are not addressed here.
encountered in either cold forging or hot forg- Instead, factors directly related to the deforma-
ing in terms of preforge factors and/or dis- tion process itself are presented in this abbre-
continuities generated by the forging process. viated discussion.
Tables 2 and 3 summarize these factors for Wear, plastic deformation processes, and
cold and hot forging, respectively. Supporting laws of friction are introduced as a group of
Table 1 Failure analysis of steel forgings and components

Case study Defect Solution
Crankshaft underfill Unable to fill crankshaft flanges with existing press Introduce creep stages for last increment of
capacity displacements
Tube bending Unable to control exterior wall thinning and interior wall Introduce induction heating and cooling to limit the
thickening heated axial tube length prior to making the bend
Spade bit Unable to achieve center web thickness at programmed Adjust the die angle to create more shear stress, enabling
force and sufficient flow to wings full flow to the wings
Trim tear Forge material tore at trimline when forging was Introduce a delay time after forge and prior to trim,
trimmed immediately following finish forging allowing the forge material to cool and gain strength
Upset forging Cracking at circumferential bulge after upset Re-examine the strain and strain rate and process map
for stable flow
Flow-through laps Material foldover at tops of rib and flange intersections Replace the input piece with a newly designed preform
and avoidance and cases of material flow under previously filled piece, following the design procedures given in this
flanges work
subjects that have been considered in the case shape (not the same for aluminum-, mag-
studies. Added factors that were evaluated in the nesium-, steel-, titanium-, or nickel-base
case studies are: alloys)
Lubrication: select one that provides the
Crankshaft underfill: induction coil inside lowest coefficient of friction and other
diameter and stock diameter, equivalent acceptable properties
current depth and subsequent time for con- Forge process: total process for entire
duction to reach a uniform stock tempera- manufacturing train, including heat treat-
ture, total heating time for scale control, ment and product testing
transfer time to press, and forging force Forge checking: fixture check for critical
applied dimensions
Tube bending: precise heat input, control of Forge tolerances: component to fit the cus-
temperature, and heated axial length of tube tomers assembly
Spade bit: direction of forging relative to Simulation of process: verify that laws of
part shape and assessment of shear effect in plasticity are met
extended wings
Trim tear: trimmer tool tolerances, part
temperature, and process time Forging Process Design
Upset forge: principal strains and equivalent
plastic strain Forging process design requires the applica-
Flow through: strains going from round or tion of integrated engineering principles that
flattened piece to finish, and assessing the bring together factors such as:
need for a more generalized shape for input Relationship between the important sub-
to finish die. Preform designstreamline system 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
Failure in Steel Forging / 135
Table 3 Factors in analysis of hot forging failures

Preforge factors
1. Raw materialchemistry, microstructure, mechanical properties, size, surface finish, and cleanliness
2. Shape sequencingshape nature, temperature, strain and strain rate, upset tooling, fuller (roll), open-die tooling
3. Hot forgingtemperature, strain and strain rate, forge center cell, loading and transfer device, lubrication, parts collection, and inspection
4. 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 behavior to achieve stable deformation at

Equation
a specified rate and proper evolution of
Material
Equipment
system
microstructures and properties
Forging Tolerances
The need for verification of the nominal
ity
Co stem
bil
dimensions and application of forging toler-

Sy
ka
ntr
or
ol
ances is important for quality assurance. Toler-

W
Control ances are required on forged products to allow

system
for practical variations in die preparation, tem-
perature effects during forging, equipment, and
Fig. 1 Relationship between important subsystems of a de- distortion during and after heat treatment. Forg-
formation system. Source: Ref 5
ing tolerance review is a basic requirement to
ensure that the part meets the multitude of design
Interdependence of forging process para-
features and tolerances. A listing of the more
meters (Fig. 2)
important forging tolerances includes:
Forging process design task overview
(Fig. 3) Dimensionallength, width, center-to-
Relationship between process and machine center, and external-internal
variables (Fig. 4) Die weargenerally approximately
Characteristics of forging machines 0.102 mm (0.004 in.)/surface
(Table 4) Die closurethickness of approximately
Workability modeling (process maps show- 0.813 to 6.35 mm (0.032 to 0.250 in.) as a
ing zones of stable flow) of workpiece function of plan form area at trimline
Matchalignment of the top and bottom Flash extension

dies Straightnesstaken as a separate feature
Radiistrong influence on material dis- and then assessing its effect on the remainder
placements of other tolerances
Data on
Billet material
Flow stress/
Ram velocity Strain rate
forgeability
Billet/Forging Contact time

Geometry, under pressure
Volume and thickness
Temperature
distribution in
forging
Die temperature,
cooling
Friction
Conditions and
Interface
coefficient
lubrication
Metal flow
Forging load
Forging energy
Fig. 2 Interdependence of forging process parameters
Fig. 3 Forging process design task overview

Draft angle gages can confirm the accuracy of dimensions

Datum plane location for three-plane (x, y, x) that are critical to the function of the component
setup dimension. Large forgings are good candidates
Alternative machined tooling points for fixture checking.
Finish allowance between forging and
machined part Wear and Lubrication
Surface interactions of two materials are
Tolerance review is conducted in various influenced by small regions where contact is
ways, and in the past, numerous forgings have made at the atomic level. The real area of contact
been rejected and held for material review until is determined by elastic and plastic deformation
some decision could be reached regarding their under consideration of loading. Lubrication
disposition for rejection or alternative repair. reduces friction by introducing a viscous and
Even though this has typically been from review low-shear-strength layer at junctions. Surface
of the part drawings, another useful way to interactions can lead to wear, or the removal of
assess a completed forging is a fixture check. A material as a result of mechanical action.
uniquely designed fixture in conjunction with a Wear types include:
dimensional inspector sets the forging into fixed
tooling point locations and proceeds with the Adhesion wear: particle transfers (pulled
check go or no-go, determining whether off) from one and adheres to the other
the part will or will not serve its function in the Abrasive wear: a hard, rough surface plows
assembly. In many cases, special fixtures and grooves into the softer one
Fig. 4 Forging equipment characteristics; relationship between process and machine variables
Table 4 Characteristics of forging presses

Equipment type Deformation rate Temperature loss Consistency Production rate
Hydraulic press Slow High Very good Low
Mechanical press Slow to medium Moderate Good Moderate to high
Screw press Moderate to high Moderate Fair to good Moderate to high
Hammer High Low Fair to good Moderate
Corrosive wear: mechanical action removes GP 100low dilution ratios and spray
a protective layer from a surface and exposes application
it to corrosive attack LSoil and water
Surface fatigue wear: spalling occurs after Precoat workpiececontains graphite as a
the formation of surface or subsurface cracks lubricant pigment
Volume wear: proportional to the load and
Adhesion colloids are reliable for high
distance traveled and inversely proportional
pressure and temperature. Types include:
to the material hardness
Colloidaldispersions
During the early 1950s, the importance of
Delta forge lubricantsfor hammer, press,
proper lubrication was recognized on the shop
and upsetters
floor. If an inexperienced oiler inadequately
Deltaglazeprotective lubricants for billets
applied lubricant (in spray or paste form), then
applicable to steel
forging problems could occur even for an
acceptable preliminary workpiece (preform).
Alternatively, a questionable preform for the
Case Studies
first closed impression die would be proven
acceptable if an experienced oiler knew where
Case Study 1: Crankshaft Underfill. There
the lubricant should be applied over the die
are several large steel forging components, such
impression and also when the die impression
as ship crankshafts and airplane landing gear,
needed additional heavy lubricant in given die
being manufactured successfully in the United
locations that appeared difficult to fill.
States and throughout the world today (2008).
These anecdotes vanished quickly as more
Crankshaft forgings in the weight range of 2268
science replaced art in forging. Presently, there
to 4536 kg (5000 to 10,000 lb) are products
are numerous ways that lubricants are used in the
made by a forging process creating a pair of
forging industry. Wrapping the workpiece dur-
flanges and a pinion shaft diameter at one time.
ing heating is an approach to prevent the for-
The inboard and outboard flanges along with the
mation of scale in the case of steel or thin metal pinion diameter become integral parts of the
sheets or cloths with impregnated graphite, in
main shaft diameter.
addition to the automatic spraying of lubricants.
The forging operation creates one set of
Lubricants play an important part in forging by
flanges by means of a working stroke in line with
minimizing the load required for maximizing
the major shaft diameter, while a 90 off-set load
material flow, protecting the die surface finish
forges the pinion shaft between the flanges.
(critical for a specific lubricant), and assisting
These operations are generally performed fol-
the entire forging process.
lowing one local heating of the starting bar
Lubricant performance factors include:
diameter for forging a set of crankshaft throws,
Adequate lubricity including the two flanges and an offset pin dia-
Stability in gas-fired and electric furnaces meter. The forging process is repeated until all
Protect stock against atmospheric conta- of the flanges plus the offset pinion diameter are
minants created along the major diameter of the crank-
Provide good surface finish shaft. The nature of the ready-for-assembly fin-
Act as a release agent ish forging design for the incrementally forged
No buildup in die cavity crankshafts includes locations where material is
Ease of application and removal provided for machining along with selected
Conform to Environmental Protection as-forged surfaces.
Agency (EPA) and Occupational Safety and During the forging of the flanges, there had
Health Administration (OSHA) require- been cases of small amounts of underfill at the
ments flange extremities, as shown in Fig. 5. That
Acceptable cost extent of underfill has caused the entire com-
Compatible with die materials ponent to be rejected.
A test run was planned to measure material
Graphite products for forging lubrication
displacements while the flanges were being
are:
forged at the prior selected process variables of
GPCfor hot and warm forging strain, strain rate, temperature of workpiece and
Die lubricantsGP series dies, and forging force exerted. The conclusion
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 rela-
tively 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 relation- Fig. 6 Both crankshaft flanges filled. Pinion shaft diameter
ship between temperature (T), stress (s), and located between flanges of single throw
strain rate (e_ ) in the steady-state condition is best
expressed as: In this case study, several time elements (all
0
e_ =A(sinh as)n exp (7Q=RT) less than 60 s) were established in subsequent
trials where all process variables were moni-
where A, a, and n0 are temperature-independent tored (including a lower-than-press-capacity
constants, Q is the activation energy, R is the force), and flange fill results were measured.
universal gas constant, and T is the temperature Finally, the proper combination of the important
in Kelvin. process variables, of which temperatures played
At low stresses characteristic of creep (as5 an important part, enabled the consistent filling
0.8), this equation reduces to: of the flange extremities, as shown in Fig. 6.
0
e_ =A0 sn exp (7Q=RT) Case Study 2: Tube Bending. Bending
large-diameter tube (310 stainless steel) created
which describes the relationship among three a 90 bend in the diameter range of 635 to
variables under creep conditions. 762 mm (25 to 30 in.) and at a nominal wall
Evaluation of experimental data of the acti- thickness of 12.7 mm (0.50 in.) This offered a
vation energy, Q, indicates that some metals in major challenge to manufacture. The challenge
hot working soften the recovery process of was to create a 90 angle bend without excess
repolygonization, and others soften by dynamic thinning at the outer wall and excess thickening
recrystallization. Thus, there is a distinct corre- at the inner wall.
spondence between hot working and viscous Earlier efforts had centered on using gas
creep deformation. heaters around the circumference of the pipe
and, in some cases, heating inside the tube. During these trials, process data were gener-
However, success was limited since the axial ated to show the variables that were used, such as
length of the heated zone was excessively long machine force, bending moment, axial tube
to permit control of the incremental bending length inside the induction coil, tube axial speed,
strain and strain rate. and amount of heat supplied, all of which were
As developments continued, a European used to control the shift of the tube neutral axis
company (Cojafax) had designed and manu- and thus the thinning and thickening of the tube
factured a large bending machine to handle a walls.
762 mm (30 in.) diameter stainless steel as the Optimizing the process variables for a 90
input and to impart a 90 bend with controlled bend on the 762 mm (30 in.) stainless steel tube
thinning/thickening of the tube walls. An resulted in minimal thinning of the outer wall, as
external axially thin induction coil supplied the shown in Fig. 8. The thinning of the outer wall of
heat and enabled control of the heated zone several tubes met the initial program goal of
while a 90 bend was being made. 18%.
The approach in the manufacturing process Case Study 3: Spade Bit. The cutting end of
was to exert an axial force in line with the a proposed wood boring spade bit (AISI 1000
straight tube axis and then push the tube for a series) consists of a central web connected to
low strain rate through the induction coil until angular extensions from each side of the web, as
the heat input was sufficient to cause material shown in Fig. 9. The forging process for the
displacements through the heated zone. In some spade bit forge was designed so that the finish
cases, cooling rings were added to control the shape could be cold forged in a continuous
axial heated length, as shown in Fig. 7. This line, starting out with wire, straighten, clean,
process was used for a number of trials in an lubricate, room-temperature forge, trip flash,
attempt to achieve the goal of a 90 bend on a and followed by heat treatment. The inherent
762 mm (30 in.) diameter tube with 12.7 mm tooling design presented an opportunity to run a
(0.50 in.) initial wall thickness. series of numerical experiments using computer
Fig. 7 Closeup of tube bending assembly. (a) Induction coil and water ring. (b) Induction coil and partially heated tube and water ring
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
Bent product tube with 90 bend and minimal
strains to be encountered when forging a round
Fig. 8
thinning at outer tube wall of ~18% of starting tube bar to a flat central section with two attached
wall thickness wings off at different angles. Shear stresses were
Fig. 9 (a) Spade bit drawing. (b) Photograph of each half

determined by Mohrs three-dimensional (3-D) Excessive clearance encourages material

analysis to quantify the stresses and strains bending displacements, which lead to shearing
throughout the central web and extended wings. in the zone. Figure 11 depicts a schematic of the
Forming simulation software Antares (from effect of punch-to-die clearance on character-
UES-Software, now defunct) was used for the istics of edges of holes produced by piercing a
deformation analysis, followed by a shop trial. low-carbon steel. In any event, the flash shearing
Strains in the wings as well as in the center is not clean and adjacent to the draft wall of the
section were calculated along with the com- forging. In some cases, the shear mechanism
pressive force and stress on the center section, initiates a crack on the edge of the forging and
before/after which simulations verified the ana- flash. The case that is being reported here is one
lysis. Following a few shop trials supported by where cracks initiated and propagated into the
simulations, an optimized process was deter- forging proper. Micrographs of crack formation
mined that led to the minimum amount of energy at the flash edge are shown in Fig. 12.
to be used to forge the spade bit center and During the search for a single or multiple
wings. solutions, several conventional avenues of
Case Study 4: Trim Tear. During the forging attack were followed:
sequence of a typical mechanic tool product
(AISI 4000 series), the process was running Trimmer tool setup for proper clearance at
satisfactorily, except that tearing occurred at the the trimline, seating of the forging inside the
flash trimline and then propagated into the forg- trip plates, and proper contact of the work-
ing proper. Trimming of the flash around the piece in the trim plate nest should be assured.
perimeter of a finish forging has traditionally Raw material condition did show some evi-
been a very dependable operation by maintain- dence of banding and inclusions.
ing the proper clearance between the punch and Temperature in the workpiece may be too
the trim tool blades, so that the deformation zone high due to high speed of production and
between the punch, forging with flash, and the resulting lower strength at trim.
trimmer sheared the flash with no bending of the Measured clearance between punch and die
flash extension. is reduced.
Fig. 10 Strain-limiting criteria for cold forging low-alloy steels

The final analysis concluded that the tem- preferably in some cases, 2.5 to 1 using flat dies.
perature of the trimming operation (immediately The material displacements are primarily radial,
following forging) was excessive and triggered extending out from the billet center and forming
the initiation of a small crack. During further the outside diameter. Material radial displace-
trials in the shop, the time between forge and ments come from the decreasing axial length of
trim was increased. This made the workpiece the starting billet. Material displacements during
have a higher strength because of the lower upsetting have been well analyzed in the past by
temperature, and no cracking occurred even numerous investigators who have identified the
though there was evidence of material banding. steel alloy; axial, radial, and tangential strain;
Other changes were made, such as decreasing strain rate; temperature of the workpiece and
the clearance between the punch and die from die; along with the frictional effect at the
0.127 to 0.076 mm (0.005 to 0.003 in.), which workpiece-die interface. During an extended
made the trimming operation cleaner and more working stroke, there is a period of time where
robust. the entire reduction operation is considered as
Case Study 5: Upset Forging. One of the one of nonsteady state. During this operation,
most important operations in the forging process there is one plasticity law that states that the sum
is the upsetting of a billet of material with ratios of the principal strain at any time is equal to 0.
of axial length to diameter of less than 3 to 1 and, This has proven to be a significant benchmark to
Edge characteristic Type 1 Type 2 Type 3 Type 4 Type 5

Fracture angle 1416 811 711 611 ...
Rollover(a) 1020% t 810% t 68% t 47% t 25% t
Burnish(a) 1020% t(b) 1525% t 2540% t 3555% t(c) 5070% t(d)
Fracture 7080% t 6075% t 5060% t 3550% t(e) 2545% t(f)
Burr Large, tensile plus Normal, Normal, Medium, tensile Large, tensile
part distortion tensile only tensile only plus compressive(g) 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

evaluate the effects of the lubrication at the die- operation, cracking on the barrel of the upset
material interface. piece often occurs because of higher strain levels
Many forged circular components and also than the material can sustain and, in particular,
circular components with off-center port bosses higher effective strain levels when each of the
begin as a billet of material of a given diameter principal strainsaxial, tangential, and radial
and are then forged between flat dies to a larger are considered.
diameter and a reduced axial dimension. Sub- The fix for a problem of this nature is to cal-
sequent forging operations on the forged initial culate the strain field in advance when plans are
preform create cylindrical walls along with made to create the upset piece. A numerical or
various configurations, depending on the final computer simulation of the process can provide
forging design. a go or no-go on the planned process. In the
Thus, an initial flattening (or sometimes past, there have been cases reported of increased
referred to as pancaking) between flat dies is a upset temperature within the workpiece, and
common initial forge operation for a number of the solution to that problem is a reduction in the
steel components, for example, missile cases strain field. Most of the cases fall into the
and bowl shapes. A typical steel upset forged situation of an unanticipated strain-rate effect,
between flat dies made on a screw press is shown requiring the rate of forging to be reduced.
in Fig. 13. However, during the upsetting 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
Fig. 13 Steel upset forged between flat dies made on a screw cracks in the material. When the critical value of
press 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: Equivalent plastic strain:
Axial?ln [1.05/4.00]=1.337 1.389
Radial+tangential?2 (ln[7.8/4.0])=2 (0.668)=+1.336 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 1.894 108 Pa (27,469 psi)
divided by disc (after working stroke) plan form area (PVA) 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 1.2554 108 Pa (18,207 psi)
strength is 19,809 psi/16,215 psi=1.16, which is in the range of Z stresses 1.211 108 Pa (17,563 psi)
reported. 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 (Disc center-to-barrel-edge readings)
plotting the dimensionless parameter of coefficient of friction disc diameter/part 1.162 108 Pa (16,853 psi)
thickness when forging a flat disc (Fig. 14) 8.247 108 Pa (11,961 psi)
4.466 108 Pa (6,477 psi)
Yield strength: 1.118 108 Pa (16,215 psi)
Fig. 14 Pressure multiplication factors for forging of thin panels. Source: Ref 7
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 con-
tinue to move outward toward the flash opening
and underneath the filled flange. This combina-
tion 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
Fig. 15 MSC superforge simulation. Disc upset forged primarily 3-D. Additional material is required to
between flat dies, showing (a) start position and fill the top of the flange and rib intersection
(b) end position after 74.93 mm (2.95 in.) stroke
because of the volume required. That additional
volume is provided by a preform shape with
deduce that the material is irreversibly damaged. increased fillet radii or taper at the base of
A typical result plot is shown in Fig. 16. the web.
Case Study 6: Avoidance of Flow Through, The term lap describes a defect that forms
Lap, and Crack. During the forging of an H- whenever material folds over itself during the
shaped cylinder of a flat web with projecting ribs forging of a new shape, using a previously
and flanges, material displacements are required designed preform as the input to follow on the
to turn 90 in the direction from the web within set of dies. Laps occur when both vertical and
Fig. 16 Typical state of strain in hot upset forging of steel showing fracture criteria in MSC superforge simulation
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 dis-
placement field could also be applied to other
alloy systems, such as carbon steels, so that the
same level of forging pressure would be used
Fig. 17 Typical forging defect caused by excessive natural
with similar results. The results are shown in
flow through the forging of a rib (flange)-and-web
part. Flow through is the tendency of a metal to flow naturally past Fig. 18.
the rib (flange) opening.
In the fifth and sixth part-way down, there is
evidence that the material front is moving away
horizontal sections intersect. When this occurs, from the entrance to the flange, and die contact is
it is an indication that a preliminary shape is lost until an additional stroke is applied. Then,
required as input to the next die in the forg- the outer die design adjacent to the flash gutter
ing sequence to provide material to fill inter- creates sufficient back pressure to displace
secting elements. Also, these types of problems material into the flange. Thus, a shape is made
are analyzed by a technique referred to as that will be suitable as an input to the finish dies,
meeting the web thickness tolerances and filling three principal axes. Preforms in the series must
the flange. satisfy subsequent shapes of the finish product
The process is to reverse-integrate from the design and local volume requirements, provid-
finish design to determine a more generalized ing die materials for intersecting product fea-
shape with precise volume distribution along the tures and other geometric attributesround
Fig. 18 Creation of a streamlined preform serving as an input to a finish die rib (flange) and web design, avoiding flow through
(simple and compound), elliptical, and tapered 4. J.E. Johnson, Ed., Forging Industry Hand-
shapes must all be accommodated. book, Forging Industry Association, Cleve-
Admissible criteria and specific tasks for land, OH, 1966
preform designs include: 5. H. Gegel, G. Huang, and S. Manna, Pre-
cision ForgingQualityProductivity
Specific alloy characteristics EquipmentA Technical Article, UES
Microstructure requirements of a finished Software Inc., Dayton, OH
product in terms of percent reduction at each 6. Piercing of Low-Carbon Steel, Metalwork-
forging operation ing: Sheet Forming, Vol 14B, ASM Hand-
Match areas and volumes along principal book, ASM International, 2006, p 159
axes and location of centroids 7. M.D. Stone, The Design and Construction
Perform reverse-integration, streamlining of Large Forging and Extrusion Presses
the finish shape to more generalized fea- for Light Metals, United Engineering and
tureslower ribs and rails (flanges) coupled Foundry, Pittsburg, PA
with increased web thickness and connect-
ing radii
Calculate principal strains when comparing
finish cross sections to preform cross sec- SELECTED REFERENCES
tions
J. Burke and V. Weiss, Advances in Defor-
Generate an overall preform shape to obtain
mation Processing, Sagamore Army Mate-
uniform deformation in the finish die
rials Research, Vol 21, Army Materials and
Examine the nature of material displace-
Mechanics Research, Massachusetts and
ments over the die contour for unsupported
Syracuse University, NY
material fronts as the working stroke pro-
H. Chandler, Metallurgy for the Non
gresses
Metallurgist, ASM International, 1998
Examine preform locations in the finish (or
G.E. Dieter, Mechanical Metallurgy, 3rd
subsequent) die and initial die contacts
ed., McGraw-Hill Book Co., 1986
Examine unsupported webs at die contact to
D.D. Fuller, Theory and Practice of Lubri-
prevent buckling
cation for Engineers, John Wiley and Sons,
Make short plots of the displacement field in
Inc., Chapman and Hall, Ltd., 1956
terms of the material contacting fixed die
C.G. Johnson, Forging Practice, American
boundaries and the change of shape of the
Technical Society Publisher, Chicago, IL,
material front being generated as the defor-
1954
mation progresses
S. Kalpakjian, Manufacturing Processes of
Follow the continuous trace of the dis-
Engineering Materials, 3rd ed., Addison-
placement field in terms of the material
Wesley, 1977
contacting fixed die boundaries and the
A. Kannappan, Wear in Forging DiesA
change of the shape of the materials front
Technical Paper, Swedish Institute of
being generated as the deformation pro-
Product Engineering Research, Goteborg,
gresses
Sweden, 1969
Determine the amount of energy expended
C. Lipson, WearConsideration in
for each preform evaluated and then the
Design, Residual Stresses and Contact
entire shape sequence
StressesA Technical Paper, University of
Michigan, Ann Arbor, MI
F.A. McClintock and A.S. Argon, Ed., An
REFERENCES Introduction to the Mechanical Behavior of
Metals, School of Engineering, Massachu-
1. G.E. Dieter, H.A. Kuhn, and S.L. Semiatin, setts Institute of Technology, Cambridge,
Ed., Handbook of Workability and Process MA, 1962
Design, ASM International, 2003 Research Report: Work at IIT Research
2. Metalworking: Bulk Forming, Vol 14A, ASM Institute, Committee of Hot Rolled and
Handbook, ASM International, 2005 Cold Finished Bar Products, American Iron
3. T. Altan, G. Ngaile, and G. Shen, Cold and and Steel Institute, New York
Hot Forging: Fundamentals and Applica- A.M. Sabroff, F.W. Boulger, and H.J. Hen-
tions, ASM International, 2005 ning, Forging Materials and Practices,
Battelle Memorial Institute, Columbus, OH, Study, American Iron and Steel Institute,
Reinhold Book Company, 1968 New York, 1966
J.A. Schey, Introduction to Manufacturing T.M. Silva and T.A. Dear, Wear in Drop
Processes, McGraw-Hill Book Company, Forging DiesA Technical Paper, De-
New York, 1977 partment of Mechanical Engineering, Uni-
J.A. Schey, Ed., Metal Deformation Process, versity of Birmingham, 1969
Marcel Dekker Inc., 1970 J.W. Spretnak, Technical Notes on Forg-
J.A. Schey and P.W. Wallace, Research ing, Forging Industry Education and
Report: Metal Flow in Closed Die Forging Research Foundation, Cleveland, OH,
of Steel, Part 2: Speed and Lubrication 1976
Effects, American Iron and Steel Institute, Technical Notes: Mechanical and Physical
New York, 1966 Properties of Ferrous Forging, Committee
J.A. Schey, P.W. Wallace, and F.A. Shunk, of Hot Rolled and Cold Finished Bar Pro-
Research Report: Metal Flow in Closed Die ducts, American Iron and Steel Institute,
Forging of Steel, Part 1: Fundamental New York
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 com- E1316-2005, Standard Terminology for Non-
ponent is often the last step or near the end of destructive Testing, components have defects
a somewhat complex manufacturing process. only when they fail to meet their specification
Finished products require attention to each step requirements. If a component has a large amount
of the long operation chain from raw material to of porosity, for example, it is not a defect unless
finished product. (1) an inspection porosity is specified, (2) its
Early in-service failures of components after amount exceeds the required acceptance criter-
heat treatment may result from improper plan- ion, or (3) the component fails because of this
ning, lack of required equipment, nonqualified porosity. This chapter describes cast steel fea-
personnel, not enough time to execute the tures that may be identified or attributed to
expected operations, or even a combination of component failure during heat treatment or
some or all of these deficiencies (Ref 1). How- subsequent processing or service. As such, these
ever, most of the early failures that happen casting features are referred to as defects in this
during the heat treatment process are the result chapter.
of features generated in previous manufacturing
stages. A component lifetime basically depends
on the following factors:
Global component project
Failures due to Improper Cast Design
Materials selection
The engineers designing job may face
Material quality
inevitable weak points due to some inherent
Processing methods, such as casting and
component characteristic in use. The engineer
machining operations prior to the heat
should, however, try to overcome these incon-
treatment
veniences by looking for alternative solutions
Heat treatment
and finding the middle ground between the
Final finishing operations
component functionality and the manufacturing
Mechanical solicitation of the component
difficulties. The project aspects that should be
and the service environment
avoided at all costs in the cast component pro-
However, it is not a simple task to identify duction are (Ref 27)
which of these items is responsible for the early
Sharp edges, sharp corners, and nonround
failure of a component during heat treatment or
edges
in service. The failure analyst uses these seven
Abrupt section changes
items as a guide in a failure analysis. This
Holes, especially when located near the
chapter deals specifically with improper casting
external wall of the component
projects and those features that originated in
Sections with cross connections
the casting process itself, including porosity
Unfavorable length/width relationship
(generated by the presence of gas as well as by
shrinkage pores), decarburization, cold joint, Rounding the corners, as shown in Fig. 1,
and inclusions. These features may not be should always be performed in order to avoid the
called defects because, according to ASTM stress concentration that can originate from
cracks formed during casting solidification the components containing holes must be
in the mold, during heat treatment, especially quenched in specially designed devices so that
quenching and tempering, or even during heat- they receive strong gushes of liquid in the
ing for austenitization (Ref 2). interior (of the holes), or they must be arranged
The risks will be reduced if steel with so that all of the set is subject to a strong stirring.
increased temperability is chosen, which re- In cases where the holes do not need to be
quires a less severe quenching medium, such as hardened, they can be made of low-alloy low-
oil or air. Other strategies in the design of cast carbon steel components already inserted during
tooling are to avoid creating components with molding for the casting, for example, in tool
right angles or to machine the corners to make steel components. Another possibility during
them round. Another option is to quench and heat treatment is to fill the holes from casting
temper the component and then remove the with any material that can totally inhibit contact
exceeding material to give the component sharp with the quenching fluid that would result in
corners if they are required for its function. This hardening of this region. Regarding stress con-
last strategy requires a steel with good temper- centration, it is preferable to have the existence
ability. Otherwise, when the exceeding material of a completely quenched hole than the presence
is removed, that region in the component will of a mixed structure (hardened and soft). Both
present a surface with lower hardness and less methods, particularly the first, make the
resistance to abrasion than the previous one. achievement and finishing of holes and threads
The quenching of components with abrupt easier.
section variations in a liquid environment Cotter holes, especially the rectangular sec-
always represents a serious problem due to the tion ones, are places with high stress con-
associated stress concentrations, even if the centration (Ref 2). Therefore, whenever
transitions are made using the apparently correct possible, they should be eliminated or sub-
concordance radius resource (Ref 4, 7). In this stituted by channel sections, whose locking
case, the solution is to create the component efficiency is equivalent, but the stress-
in different parts, treat them separately, and concentration factor is three to four times lower.
assemble later on. However, if the component Another geometry that should be avoided during
must be made as one unit due to a functional project design of the cast tooling component is
imposition, the solution is to choose an air- the cross type, such as the furnace grid and
quenchable steel that presents a lower crack heating equipment shown in Fig. 3(a). It causes
probability. serious crack problems in the cross area during
The existence of holes raises a problem the solidification process while still inside the
mainly in high-carbon steels and/or alloying sand mold or during heat treatment (Ref 3). The
elements. The abrupt section variation and other solution is to use node dislocation, as illustrated
specific aspects of the holes (Fig. 2) must be in Fig. 3(b).
considered. The accumulation of quenching Lengthy components with very thin sections
liquid in the interior of blind holes leads to an or small diameters show serious bending pro-
improper heat loss of the internal walls, thus blems during heat treatments, mainly quench-
lowering the hardness. In open-ended holes, the ing, even when the steel is favorable to less
heat removal may not be as effective as in the severe environments. The problems start from
rest of the component, which is more exposed to the moment the component is put into the fur-
the quenching liquid (Ref 5). Therefore, when nace for austenitization. If it is secured only in
the chosen steel is quenched in a liquid medium, two extremities, there is the risk of deflection
Fig. 1 Sharp edge elimination. (a) Sharp corners create high Fig. 2 Types of holes. (a) Blind hole with a parallel bottom.
strain concentration. (b) Exaggerated relief causes a (b) Blind hole with a steeple bottom. (c) Passing hole
shrinkage cavity. (c) Ideal relief the most economic
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 homo-
geneous, 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, high-
alloy, high-carbon, steel sheets used in the
guillotine have been replaced by shorter sheets Fig. 4 Crack resulting from the normalization heat treatment
of an AISI 1045 steel cast hull caused by thinning of the
that, after quenching and tempering are assem- wall due to deficiency of the tooling or the core alignment
bled 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 pieces, despite the adjustment problems that
when there is an in-service failure of one of the result from this operation.
sections, leading to an easier and cheaper Figures 4 to 6 illustrate failures that happened
operation. during the heat treatment operation due to poor
Very big components with a circular section design considerations. Figure 4 is a crack that
larger than 25 cm (10 in.) in diameter, or rec- occurred during the normalization heat treat-
tangular with equivalent mass, also present ment of an AISI 1045 steel cast hull. A prior
problems during quenching. When carried out in thinning of the wall (due to deficiency of the
a liquid medium, the surface reaches the starting tooling or core alignment) promoted the crack-
martensitic transformation temperature long ing during normalization. Figure 5 is a crack that
before the central region, with the generation of happened in the normalization heat treatment,
stress that may cause internal cracks as a con- caused by stress buildup in the sharp edge
sequence. It is recommended that big, bulky region. In Figure 6, poor design of an edge led to
components be replaced by sets of smaller cracking after quenching.
Effects due to Porosity component, that is, the absence or minimization

of the quantity of defects present. Among these
During the production of casting components, defects, the most important ones are those gen-
the interactions between the liquid metal and erated by the interaction of gas and metal that
elements from the gaseous atmosphere, furnace promote the appearance of voids. In general,
refractory, foundry ladle, molds, and core there are two kinds of voids: those generated by
materials are important factors from a techno- gas, and shrinkage pores.
logical and metallurgical point of view and are
responsible for desirable or undesirable changes
in the chemical, physical, and mechanical Porosity Caused by Gas
properties of the metallic materials. The dim- One of the factors that must be considered
ensional precision grade of the feeding system in steel casting is the behavior of the gases in
should also be taken into consideration. the process. Generally, there are three major
Therefore, the quality of the casting product sources that may contribute to porosity form-
is related to the physical integrity of the ation (voids caused by gases) in steel castings.
These are:
High initial gas content of the melt orig-

inating 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 cast-
ing 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 por-
osity caused by the presence of gas (Ref 8).
Fig. 5 Crack in the bottom of a machine molded from AISI Gas porosity (pinholes) refers to hydrogen,
1030 steel that happened in the normalization heat
treatment, caused by stress buildup in the sharp edge region 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 mol-
ecules, 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 gen-
erally undetectable visually, since the surface
of the casting solidifies the quickest, preventing
Fig. 6 Plastic injection mold casting in AISI H13 steel with
the gases from forming holes large enough
a crack from a sharp edge after a quenching heat
treatment to be visible on the surface, except through
fluorescent-penetrant inspection or crack detec- same, and there is low hydrogen segregation
tion during or after the heat treatment. during the solidification, which reinforces the
Gas holes (blowholes) are generally larger fact that the presence of gas bubbles caused by
and more localized voids than gas porosity, but hydrogen has other causes, for example, the
they retain the smooth, round, or slightly elon- reaction with the moisture from the mold and/or
gated shape. They are usually caused by a the cores (Ref 9).
reaction in the mold medium, producing gas that For nitrogen, its component ratio for stable
bubbles through the molten metal. The humidity and metastable eutectic solidification is 1.9 and
contained in the mold walls and cores is the main 2.2, respectively (Ref 13). This shows that
source of the vapor that is necessary for defect nitrogen is less soluble in the liquid metal than in
(gas bubble) formation in the casting compo- the solid metal. However, it is good to highlight
nent. During mold filling, the gas generated by that in a real situation, the solidification involves
the metal-mold reaction is eliminated to the the liquid-solid diffusion and vice versa of a
environment through permeability, a hole from larger quantity of elements, nitrogen being just
the exit of gases, and/or a rising gate. The resi- one of them. When nitrogen behavior in iron
dual quantity of gas that could lead to bubbles is alloys is analyzed, the most important element
almost nonexistent or negligible. The exception whose mutual presence must be considered is
would be the use of low-permeable molds, for carbon. For example, in a 3.8% alloy at 1500 C
example, the ones whose sand contains a high (2730 F), the nitrogen solubility at equilibrium
percentage of fines, making the passage of the in the liquid metal is 110 ppm. At the time of
gases to the environment difficult. For the cores, eutectic solidification, this value is reduced
which may become completely surrounded by in the liquid to 97.5 ppm due to austenite en-
liquid metal during mold filling, the problem can richment. In this sequence, because there is
be more serious. Gas elimination to the exterior, an increase in the carbon percentage in the
including the gases generated by the binder and liquid, the nitrogen solubility is reduced to
collapsible materials, is extremely difficult. It 90 ppm. There is nitrogen saturation in the
may require the use of devices such as internal liquid, and thus, this excess will cause the evo-
wax wicks in all the core extensions, so the gases lution of the gas and may originate pinholes,
are sucked toward the core prints and then as seen in Fig. 7. When the possibility of pinhole
eliminated. A gas bubble can also occur, even formation is evaluated, the component pressure
though it is not very common, as the result of an of all involved gases must be considered. When
inadequate measurement of the descent chan- this sum is higher than 1 atm, pinholes form.
nels, distribution, and attack, which, during So, in the previous example, the formation of
pouring, can cause turbulence in the liquid metal pinholes could happen with lower nitrogen
flow or can cause air to be inhaled to the interior concentrations than those mentioned, needing
of the mold cavity, where it mixes with the liquid just the presence of other gases, such as hydro-
metal. gen or oxygen, for example. It is also important
Behavior of Gases. Dissolved hydrogen and
nitrogen in the molten steel can cause a porosity
defect such as a pinhole. The extent of gas por-
osity 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 solu-
bility. 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
Fig. 7 Typical morphology of a defect called a pinhole,
molten and solid conditions is approximately the caused by gases
to consider the dynamic formation of these system, directional solidification, and pouring
bubbles in pinhole formation. However, the temperature.
thermodynamics of this event are very complex During the liquid-to-solid transformation,
and are not covered in this chapter. there is a grouping of atoms that forms ordered
How to Treat Pinhole and Blowhole Pro- structures. In the majority of cases, this trans-
blems. In order to minimize or eliminate the formation is followed by a density increase
pinhole formation problem, a relatively simple (Fig. 8) and thus a shrinkage, because the metal
method can be used: bubbling argon in the as a liquid occupies a larger volume than in the
desulfurization reactor or ladle. When the argon solid state. The defect known as shrinkage pores
is blown into the bath, the gases in the atomic can be characterized as the appearance of non-
form combine themselves on the bubble surface, superficial cavities in the casting component due
forming molecules from the respective gases to the lack of predetermined and precise com-
(N2 and H2). However, it is known that bubbling pensation devices for the liquid metal shrinkage
is more effective in hydrogen elimination than in that occurs during solidification and/or the
nitrogen. metallic inserts for directional freezing.
One of the most common alternatives to If the concentration is a little higher than the
reduce or eliminate blowholes is to increase capacity of the system to compensate for it, or if
the pouring temperature. There will be a higher thicker pieces of the component work as risers
fluidity and time interval for the beginning of for thinner components, small, irregular voids
component solidification, giving time for the will be formed, as is seen in Fig. 9. However, if
gases to escape to the atmosphere. However, the concentration is much higher than the com-
care must be exercised in the decision to increase pensating mechanisms, there are large voids of
the pouring temperature. The concentration irregular shapes on the surface of the compo-
will also be bigger, and there is the risk of nent. These are called primary shrinkage cavi-
the riser becoming undermeasured, thus allow- ties or simply shrinkage cavities, as seen in
ing the occurrence of a shrinkage cavity, in Fig. 10, and are not discussed further in this
addition to the possibility of a molding and core chapter.
system collapse, causing other defects in the In order to better understand shrinkage, an
casting. example is given of the production of a cast iron
In summary, to minimize the effect of void component with two kinds of molds: a nonstiff
appearance from gases, the main measures to be mold with synthetic sand (green) and a stiff mold
taken are: with phenolic no-bake sand. In the one produced
with synthetic sand, there is a factor that must be
Control of the furnace atmosphere using considered: the mold walls deform, increasing
vacuum or gases with low solubility values the volume of the cavity when it receives the
Develop a project of feeding channels to liquid metal, and thus, it requires more metal 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
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
compensate the increase in the volume. With the equivalent carbon is larger than 3.9%, there is a
rigid phenolic resin mold, there is no volume graphitic expansion, which is larger than the
increase in the cavity, and the additional liquid solidification shrinkage and could cause a metal
metal is not necessary. On the contrary, when the reflux for the mold exterior.
Each metal or metal alloy presents a char-
acteristic concentration rate during the solidi-
fication 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 solidifi-
cation 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
Fig. 9 Example of a shrinkage pore 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 pro-
cess, which replaces the liquid metal necessary
to compensate the shrinkage in the system, very
precise in a way that allows for the attainment of
(a)
perfect components. This shrinkage compensa-
tion process determines the precision and per-
fection of the casting component and is inversely
(b)
Fig. 10 Primary shrinkage cavity forming large voids Fig. 11 Schematic representation of the three regimens of
of irregular shapes on the component surface. shrinkage: in the liquid state, during solidification,
(a) Schematic drawing. (b) Shrinkage cavity compensated for riser and in the solid state. Source: Ref 14
proportional to the quantity of shrinkage cavities thermal, and pressure criteria that are absolutely
or shrinkage pores present in the obtained necessary for perfect solidification:
component.
The junction between the casting and the
During the shrinkage process that occurs in
riser must not create a hot spot. This place
the solid state, the component size starts to be
cannot have a larger solidification period
reduced. At this moment, the casting component
than the riser or the casting component;
faces the resistance of the mold and/or core. This
otherwise, it can cause the formation of a
kind of stress from the casting component, when
shrinkage porosity.
trying to contract, generates residual stresses
There must be a way in which the liquid
that may cause plastic deformation of the casting
metal of the riser can reach all the required
component, hot tearing, or cracks during heat
regions.
treatment later on. Yet, this shrinkage depends
There must be a pressure variation in order to
more on the volume reduction intrinsic to the
cause a liquid material flow in the right
cast alloy and the project of the mold than on the
direction.
casting parameters.
There must be enough pressure in all the
Six Rules for Casting Component Feed-
regions of the mold to avoid the formation
ing. In the absence of gases and if the feeding of
and growth of cavities.
liquid metal is appropriate, no porosity will be
found in the casting. However, because there are Internal Porosity Formed from the Sur-
many complex casting projects, there may be face. If there is not enough internal pressure
regions of the mold with feeding problems, inside the component being cast and if the liquid
allowing the internal hydrostatic tension in the inside the mold is still connected with the liquid
liquid metal to generate the conditions for the in the external surface, it can be sucked to the
formation of internal pores. inside, causing the growth of porosities that are
In the design of a component to be cast, it is connected with the surface (Fig. 12), because the
necessary to have an effective supply of material liquid naturally drags air with itself that stays in
in order to compensate the shrinkages pre- the interdendritic spacings of the casting com-
viously mentioned. For the additional liquid ponent. This preforming mechanism is much
metal supplied to the system to compensate the more common than imagined. It occurs mainly
volume shrinkage that occurs during cooling, a in alloys with a very long cooling range, when
riser must be provided in the casting of the the development of the dendritic lattice means
component. The use of these risers, also known that the aspiration of liquid in the neighboring
as feeders, exothermic sleeves, or hot tops, can surfaces becomes easier than feeding from a
eliminate the problem of shrinkage pores. The more distant point. The point at which the liquid
quantity, form, and volume of these risers vary may be pulled from the surface can be anywhere
according to the form and complexity of the for an alloy with a long enough cooling period.
component to be cast. However, despite the fact Thus, in an alloy with an intermediary solidifi-
that there is a vast amount of literature on the cation period, the starting point is usually a hot
calculation and quantity of these risers, the spot, such as an internal corner or a recess angle
correct location of them depends on the experi- of the component.
ence of the process controller. The possibility of the connection of two
The following criteria, however, are con- opposed surfaces in the same component
sidered fundamental for proper feeding of the through the pores is one of the main reasons that
component, and thus, the defects caused by alloys with long solidification times should not
shrinkage pores are reduced or eliminated: be employed in the manufacturing of compo-
nents where high working pressures are applied,
Thermal transfer criterion: The riser must such as hydraulic valves or motor cylinder
solidify at the same time or slower than the heads, because they would cause leakages. The
casting. prerequisite in such complex components is that
Volume criterion: The riser must contain the interior should have positive pressure in all
enough mass to fulfill the volume shrinkage points in order to avoid the connection of the
needs of the component. surfaces through internal porosities, which is
rarely achieved.
However, there are still rules that are eventually Internal Porosity from Nucleation. Alloys
observed, and they define additional geometric, with very short solidification intervals, such as
the aluminum, brass, and eutectic aluminum- stage of nucleation. In fact, such a pore or
silicon alloys, do not present the connection shrinkage cavity is simply formed as an answer
problem between the surface and pores because to the shrinkage of the solidification.
they have a perfect solid layer in the first stages During the solidification of a casting compo-
of solidification, while the liquid feeding occurs nent, the liquid flow from the reservoirs to the
through the feeding channels. The internal areas that are being solidified make the level of
pressure decrease due to an inefficient feeding at the reservoirs decrease. At the same time, there
the end of solidification can create a pore is the advance of the solidification front. This
through nucleation in the interior of the liquid. In joint action of decreasing the liquid level and the
this case, there is no connection with the external advance of the solidification front creates a conic
surface of the casting. cavity, as shown in Fig. 13. This cavity is called
So, in this kind of alloy, the porosity is usually a primary shrinkage cavity in order to differ-
nucleated and is concentrated near the center of entiate it from a secondary shrinkage cavity,
the component. When it occurs in plates, for which is porosity islands observed from long-
example, it is referred to as axial porosity. itudinal cuts guided according to a line from the
Unless subsequent machining operations pass thinnest region of the primary shrinkage cavity.
through the pores, the casting in such alloys is In fact, these islands are interconnected in the
tight. solid volume and are thus an extension of the
After nucleation, the subsequent solidification primary shrinkage cavity.
provides the necessary driving force for pore Example 1: Failure Analysis of a Mill
growth, which, if observed structurally, has Gear with Defect Caused by a Shrinkage
many similarities with the one started from the Pore. The analyzed component corresponds to
surface. a mill gear of 860 by 1900 mm (34 by 75 in.)
Growth of the Shrinkage Pores. The first external diameter and 1050 mm and 15 teeth, as
stage of shrinkage pore growth is very fast. shown in Fig. 14. The mill gear was purchased in
According to Davies (Ref 15), this period should 2003 and fractured in service during the 2003
be less than 60 ms. After this first nucleation 2004 harvest (only 3 months working). The aim
stage, the growth of the pore happens more of this study was to verify the metallurgical
slowly, being controlled by the heat extraction properties of the material and the causes that
rate of the mold. For the shrinkage pores that eventually could have contributed to generate
started from the surfacethe primary shrinkage the crack nucleation and the component fracture
cavities, for examplethere is not a fast first after a short period of working.
Fig. 12 Internal porosity formed from the surface

Fig. 13 Primary and secondary shrinkage cavities
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 uni-
directional 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
Fig. 14 Aspects of the mill gear as received for analysis during a relatively short period of the milling
operation. Most likely the component started
the operation already cracked, that is, with
A chemical analysis of the studied component casting defects of considerable dimensions near
was carried out by optical emission spectro- the surface (subsuperficial), causing the cracks
scopy, and the results are shown in Table 1. to arise and propagate just at the beginning
A tensile test was performed according to of the harvest, which caused its fast milling
ASTM E8M-98, an impact test according to rupture.
Structural analyses were performed on sam- respectively, according to the data provided
ples from the fracture region and on the tooth by the manufacturer and compared to the
radio adjacent to the fractured region (Fig. 16, experimental ones. The phosphorus and sulfur
17). It can be noted that the colony of subsurface quantities are between the maximum limit
shrinkage cavities/porosity connected, which established by ASTM A148-93B, and the
probably caused the crack nucleation. quantity of other elements agreed with the
It could be concluded that the failure cause manufacturer specification. The values obtained
was related to casting defects, such as connected for the yield strength (0.2%) and for tensile
shrinkage cavities and porosity colonies, asso- strength agree with the expected ones.
ciated with tensile loads applied during the mill The mill gear fractured through the mechan-
gear operation, which caused crack nucleation. ism of crack propagation by fatigue. The cracks
Figure 18 shows the crack starting point, proving nucleated from the casting defects, located
the failure cause. mainly in the third component of the width of the
Example 2: Failure Analysis of a Mill Gear mill gear and near the surface of the tooth root.
with Defect Caused by a Shrinkage Pore. The low cooling rate during the solidification
Figure 19 shows a sliced sample of the mill gear process is probably the main cause of the high
with the tooth root used in the failure analysis. susceptibility to casting defect formation, such
Several mill gear presented had fractures on as shrinkage pores. Inclusions and bubbles
several teeth roots after an intermittent load- represent a small component in the material
ing time. The chemical composition in weight embrittlement, since they are too small com-
percent and the mechanical properties of the pared to shrinkage pores. Their presence should
steel in Fig. 19 are presented in Tables 3 and 4, be neglected.
Table 1 Chemical analysis of the studied material

Composition, wt%
Specification C Si Mn P S Cr Ni Mo Cu Al
ASTM A148 Gr 105-85 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

Specification
Test ASTM A148 Gr 105-85
Tensile strength, MPa 804
Yield strength, MPa 679
Yielding at 50 mm, % 6.0
Reduction in area, % 10.0
Impact, J 3035
Hardness, HB 257259
Fig. 15 Aspect of the fracture surface showing that approxi-

mately 30% of the longitudinal section had been
taken over by the cracks diffused by fatigue. Many subsuperficial Fig. 16 Structural analysis performed on samples from the
casting defects were also observed where the nucleation of the fractured region. Etched with 3% nital. Original
cracks started. magnification: 100
Penetrating liquid analysis was performed on Figure 21(a) shows the observed micro-
the tooth root surface, and the result is presented structural aspect, where it is possible to note the
in Fig. 20. Localized cracks on the root center presence of several cracks in the sample interior.
have extended to the sides. These cracks were nucleated in several shrink-
age pores and have propagated by the fatigue
mechanism to the surface during cyclic loading.
The aspect of the fracture is mainly transgra-
nular, 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 crack-
nucleating site due to stress concentration. The
presence of sharp inclusions indicates that the
globalization process during casting was not
totally efficient.
Fig. 17 Structural analysis performed on samples from the
The proposed corrective actions include:
fractured region. Etched with 3% nital. Original
magnification: 200
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 elim-
ination during the machining process
Effects due to Decarburization

during Microfusion
Fig. 18 Region adjacent to the fractured region showing a Among the several kinds of defects that may
transgranular crack generated in the casting process
and masked by material deformation during the radio machining occur during the casting process and that are
process, with propagation directed to the internal diameter detected after heat treatment is the surface
Fig. 19 Sample that was analyzed

decarburization layer. It occurs in carbon steel Effects due to Cold Joints

components cast by the lost-wax or microfusion
process. The identification of this defect can Cold joint is a kind of defect that happens in
only be made after heat treatments in controlled cast components and normally has a significant
atmospheres; otherwise, this identification is effect on the structural integrity of the compo-
impossible, since the decarburization can also nent. This serious nonconformity happens when:
come from the heat treatment. 1) two portions of the metal, each coming from
This decarburization results from the pre- different feeding/distribution canals of the mold,
sence of atmospheric oxygen that remains in the meet and, instead of contributing to the form-
mold as a consequence of the inert feature of the ation of a smooth and homogeneous surface,
mold and its permeability in relation to the sur- provoke an undercut discontinuity called a cold
roundings. Some analysts have measured the junction; 2) the solidification process occurs too
thickness of the decarburized layer in carbon far from the metal flow coming from the feeding/
steels cast by lost wax and discovered that it distribution place, where the liquid temperature
increases proportionally to the temperature is lower than the necessary temperature; 3) the
increase in the mold and the volume/surface pouring, feeding, and distribution channels are
ratio of the casting component (Ref 14). underdimensioned and strangle the flow of metal
Figure 22 shows a microfused decarburized necessary for the filling of the mold; and 4) the
component that underwent heat treatment. molds have voids that need to be filled with such
Table 3 Chemical composition of the studied material

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

Mechanical properties
Source of data st, MPa se, MPa e at 50 mm, % Reduction in area, % Hardness, HBW
Manufacturer 777.0 626.0 19.0 38.6 228
Experimental 780.1+16.2 606.7+15.3 10.2+1.9 35.9+14.1 230+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.
a thin thickness that the liquid metal, even at the can occur and be unnoticed initially, in compo-
appropriate temperature, cannot fill them com- nents with complex geometry and abrupt vari-
pletely. ation of mass, where it is used to obtain a large
The component that shows this kind of defect, number of cores that could provide details dif-
depending on the size and location of the joint, ficult to be observed by quality control. In these
must be discarded, since recovery with a weld is cases, the defect will only be located when there
not recommended from a metallurgical point of are cracks/disruption in heat treatment or leak-
view or, depending on the cost-benefit ratio, age and fracture when the component is in
is not justified. This defect is usually seen, but service.
Fig. 21 (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
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 cast-
ing of sheet steels and wire.
Macroinclusions are always practice related,
and analysis of the size and chemical composi-
tion of a macroinclusion can lead to the identi-
fication of potential sources of this problem.
Once an inclusional source is developed, a clear
and effective process change can be made to
Fig. 22 Surface of a microfused component showing surface eliminate such problems in the future. There-
decarburization
fore, the techniques already developed by inte-
grated steel manufacturers can be readily
In summary, to avoid the appearance of cold applied to foundries by coupling inclusion
joints in cast components, it is necessary to identification with an in-depth study of steel-
control several manufacturing stages of its making and casting practices in the foundry.
design; for example, prevent the component Horwath and Goodrich (Ref 17) and Svoboda
from having regions with very thin thickness; et al. (Ref 18) have studied macroinclusions and
appropriate fusion and pouring temperatures for identified that these kind of inclusions can result
each component; appropriate mold-filling in excessive casting repairs or rejected castings.
channel system; compatible pouring speed; and To reduce these problems, a method was dev-
well-established necessary amount of liquid eloped to ensure that there are no inclusions in
metal for filling the mold to avoid temporary cast materials above a size that results in failure
interruption in pouring. during ultrasonic or visual inspection of the
casting. In this method, the macroinclusions
should be eliminated; that is, inclusions greater
Inclusions than 100 mm must be eliminated, but more
severely, inclusions greater that 50 mm should
Inclusions can be defined as nonmetallic and be eliminated also.
sometimes intermetallic phases embedded in a Sulfides, nitrides, and oxides are examples of
metallic matrix (Ref 16). They are usually sim- indigenous inclusions that result from chemical
ple oxides, sulfides, or nitrides. In almost all reactions of the molten metal and the local
instances of metal casting, they are considered environment. They are usually very small and
to be detrimental to the performance of the cast uniformly distributed inclusions, requiring
component. Sometimes, an intentional intro- optical microscopy to visualize them. The pre-
duction in larger quantities can lead to unique sence of these microinclusions in castings is
dispersion-strengthened materials. There are generally unavoidable (Ref 9), because they
essentially two classifications for all inclusions: are the natural inclusions that are formed in
liquid steels due to the reaction between
Exogenousthose derived from external
alloying elements and oxygen; however, it is
causes
necessary to minimize these inclusions as a
Indigenousthose that are native, innate,
grain-boundary distribution of these inclusions
or inherent in the molten metal treatment
can be damaging to the component mechanical
process
properties.
Slag, dross, entrapped mold materials, and
refractories are examples of inclusions that
Clean Steel
would be classified as exogenous. In most cases,
these inclusions are macroscopic or visible to Clean steel is the common name attributed to
the naked eye at the casting surface. When the steel that has low levels of the elements sulfur,
casting is sectioned, they may also appear phosphorus, nitrogen, oxygen, and hydrogen, as
beneath the external casting surface if they well as residual elements copper, lead, zinc,
have had insufficient time to float out or settle nickel, chromium, bismuth, tin, antimony, and
due to the density differences with respect magnesium and almost no oxide product defects
to the molten metal. The presence of these produced during the act of steelmaking, ladle
metallurgy, casting, and rolling. Because the must be eliminated at every step in the
clean concept is not absolute, the clean- process.
liness standard desired by the customer is con- The physical entrapment of the liquid fluxes
tinuously changing as a function of time and used during steel refining and casting must
technological improvements. The term clean be eliminated.
steel is therefore continually variable, depend- Refractories in contact with liquid steel must
ing on the application and the competition be chemically stable and resistant to corro-
between steel suppliers. sion and erosion.
Thus, due to the variability of the term clean,
These simple principles are based on the
it is typical to refer to high-purity steels as steels
importance of maintaining chemical equilibrium
with low levels of solutes, and low-residual
between the elements dissolved in liquid steel
steels as steels with low levels of impurities. For
and the slag and refractory systems that are in
example, there are high-purity, low-residual
contact with the liquid steel. Additionally, it is
clean steels, such as ultra-deep-drawing steel
necessary to control the fluid flow to avoid
sheets for automobiles, that require ultralow
conditions at liquid slag-steel interfaces that
carbon contents (530 ppm), low nitrogen con-
could result in the physical entrapment of the
tents (530 ppm), and the absence of oxide
covering slag.
inclusions with diameters greater than 100 mm;
Clean steel manufacture is dependent on an
and there are low-residual clean steels, such as
understanding of the fundamental steps neces-
those used for drawn and ironed cans, that are a
sary to produce a clean steel:
standard low-carbon steel (1006) without high-
purity component requirements but are ultra- Generation of the inclusion
clean, with the requirement that oxide diameters Transport of the inclusion to an interface
must be less than 20 mm. In addition, in forging Separation of the inclusion at the interface
and bearing grades, there are clean steels that Removal of the inclusion from the interface
require strictly controlled inclusion size dis-
tributions. The production of really clean steel depends of
The total inclusion content related to the total the correct application of these principles.
oxygen content has been correlated with bearing
life, and decreasing total oxygen contents The Formation of Macroinclusions
(below 10 ppm) improve the bearing life. In There are four major methods of forming
addition to total oxygen content, the total length macroinclusions, and all problems occur during
of stringer inclusions after forging is also related foundry processing:
to the bearing life, and, at low total oxygen
levels, efforts to reduce inclusion clustering lead Reoxidation
to very long fatigue life for bearings. Interaction between liquid steel and liquid
Clean steels can be classified as steels with a slags: vortexing, ladle or mold filling, argon
low frequency of inclusions (55 mm). The stirring, and pouring through a slag layer
major problems in clean steel manufacture are Erosion/corrosion during steel pouring
incomplete separation of clustered solid inclu- Inclusion agglomeration due to clogging
sions (45 mm in diameter), the presence of during steel pouring
sporadic larger liquid inclusions due to emulsi-
fication of covering slags, and the presence of Reoxidation. The major cause of macro-
solid materials that originate from the refrac- inclusion formation in casting is reoxidation
tories used to contain steels. The equipment used (Ref 1719). To understand reoxidation, it is
to produce clean steel varies greatly between necessary to understand that liquid iron is not
different steel plants; however, current clean thermodynamically stable in the presence of
steelmaking and casting practices are based on oxygen. The spontaneous reaction that occurs
the following principles: results in the formation of iron oxide. As deoxi-
dizers are added, the steel remains unstable in
The oxygen dissolved in liquid steel at the the presence of oxygen as a gas, but now the
melting stage must be transformed into a inclusions that form include the oxides of the
solid or a gas and removed before casting. deoxidants. Some deoxidants, such as alumi-
The external oxygen sources that are res- num, magnesium, and calcium, form very stable
ponsible for the reoxidation of liquid steel oxides that are more stable than some slag and
refractory chemistries. Thus, the steel reacts was found to be directly proportional to the flow
with the less stable oxides. Reoxidation can rate needed to cause entrainment. The gas bub-
occur by reaction with: ble size was found to be inversely proportional
to the flow rate needed to cause entrainment.
The ambient atmosphere (air) Manabu et al. (Ref 24), investigated the effect
The slag components less stable than the of oil kinematic viscosities on emulsification
oxide of the deoxidant and found that although the kinematic viscosity
The refractories that are less stable than the was varied by a factor of 10, very little change
oxides of the deoxidant was seen in the fluid velocity needed to cause
entrainment. Harman and Cramb (Ref 25),
Interaction between Liquid Steel and
documented the effect of interfacial tension and
Liquid Slag. Macroinclusion formation can
slag viscosity on emulsification phenomena.
occur by emulsification of liquid slags or scums
Erosion-Corrosion during Steel Pour-
on the surface of liquid steels. All of these types
ing. This kind of defect is usually associated
of defects are practice related and can be solved
with the higher corrosivity of some steel grades,
by practice changes. The issue in understanding
because high manganese and grades that are
emulsification is to understand the source of the
barely killed and have high soluble oxygen
energy that allows a buoyant droplet to become
contents attack the binder or the mold sand itself,
submerged. Generally, this energy comes from
leading to large entrapped sand components.
the interaction of a flowing steel stream and a
Reoxidation of steel leads to FeO-based inclu-
liquid slag. There are four major sources of this
sions that are very reactive and wet the materials
energy:
of the mold, leading to erosion of the mold in
Open stream pouring onto or through a areas of high fluid turbulence. Of course, sand
liquid slag (common during lip pouring) that is not pressed, sintered, or bonded in any
Filling a ladle or mold at too high a fill rate in way can easily be entrapped in turbulent fluid
the presence of slags or scums flow. Mold binders can also decompose at tem-
Vortexing during steel pouring from a ladle perature and release mold components that can
Steering in the ladle with gas at too high a be entrapped. Expansion due to the high thermal
stir rate gradients associated with casting can also cause
sand to loosen.
Vortexing during drainage in a water model of Inclusion Agglomeration due to Clogging
a ladle was studied by Sankaranarayanan and during Steel Pouring. The formation of clogs
Guthrie (Ref 20, 21). They showed that the when steels containing solid inclusions are cast
initial rotational velocity at the surface of the can result in quite large macroinclusion defects
vessel is extremely important in determining if the clogs are released during teaming. All
the height at which the vortex will form, and that solid inclusions tend to agglomerate due to sur-
increased rotational velocities caused increased face tension effects. Clogging of pouring noz-
vortex initiation depth. Entrainment due to fluid zles can be the source of large macroinclusion
flow at the interface has been examined by defects when steels are dirty and pouring times
Noguchi et al. (Ref 22), who attempted to are long.
decrease the entrainment of slag in low-carbon
titanium-aluminum-killed steels. They noted
that entrainment decreased as the casting speed The Formation of Microinclusions
was decreased. In a study conducted by Naka- Microinclusions are formed due to reactions
mura et al. (Ref 23), it was found that defects that between alloying additions and oxygen in mol-
contained mold slag increased in ultra-low- ten steel. Their formation is generally hetero-
carbon grades as the casting rate was increased. geneous or from highly supersaturated areas
They also reported using as low an argon flow during alloy addition. Due to the nature of the
rate as possible in their submerged entry nozzles formation of these inclusions (nucleation and
to avoid entrainment. Manabu et al. (Ref 24) growth), they are generally small (less than
have also documented the existence of a critical 5 mm), unless they agglomerate due to turbu-
gas flow rate for entrainment in both a silicon lence or grow under conditions of high oxygen
oil-water and a slag-steel system. These authors flux. In this study, microinclusions are defined as
mention that the slag depth, slag properties, and those inclusions with diameters smaller than
gas bubble diameter play a role. The oil depth 20 mm. In addition, they are defined as having
diameters greater than 1 mm. Table 5 show indicated. Figure 25 shows a micrograph of the
typical microinclusions that are found in cast fractured surface, near the blade bottom. Several
steels. turning gear imprints can be observed, showing
Since microinclusions form due to a reaction, the presence of multiple sites of crack nucleation
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.
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 concentra-
tion 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 pre-
sented in Table 6. The specifications for the
mechanical properties of the material at room
temperature are shown in Table 7.
Fig. 23 (a) Turbine stage that had the fractured blade.
Figure 24 shows the fractured blade compared (b) Detail of the fractured bottom component of
to an intact one, with the fracture regions the blade
Table 5 Typical microinclusions found in cast steels

Steel type Microinclusion type Comments
Aluminum killed Alumina Formed in liquid steel after deoxidation
Manganese-silicon killed Manganese silicate or manganese-alumino Formed in liquid steel after deoxidation
silicate
Calcium treated, aluminum killed Calcium aluminate Formed by reaction with alumina, liquid inclusion
Aluminum killed, with residual Magnesium aluminate Formed by reaction with alumina, solid inclusion
magnesium
Titanium treated, aluminum killed Alumina, titania, titanium nitride Titania forms during reoxidation. Titanium nitride forms
during cooling, usually in the mold itself.
All steels Manganese sulfide Forms interdendritically during solidification. Often
nucleates on oxides already present in steels
Table 6 Nominal chemical composition of FV520(B) steel

Composition, wt%
C Si Mn Cr Ni Cu Mo Nb S P
0.07 max 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 Strain limit, MPa Elongation, % Reduction in area, % Impact energy, J Hardness, HV
680800 9001050 20 min 55 min 40 min 270320
Fig. 24 (a) Intact blade. (b) Fractured blade
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 micro-
scopy 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 nom-
Fig. 25 Micrograph of the blade fracture surface showing inal, as much for the central region as for the
several turning gear imprints and the oxidized area
(dotted line) blade head region. The fracture surface of the
longitudinal crack revealed a microstructure rich
in silicon, oxygen, manganese, and calcium,
by fatigue. A darkened region is observed on the suggesting that the material contains a large
fracture surface, indicated by the dotted line, number of impurities, probably slag from the
suggesting that this area was more exposed to casting process and certainly introduced during
steam and high temperatures during the turbine the manufacturing process of the component.
operation time, and it occupies a significant The occurrence of these impurities impedes
component of the fracture surface. surface welding during the process of forging,
Penetrating liquid analysis indicated the pre- creating a surface with a smashed aspect.
sence of secondary longitudinal cracks in the However, the first region where the nuclea-
fractured material, normal to the main crack, at tion probably occurred was the one near the
longitudinal cracks detected by the penetrating The flat fracture surface occupied approximately
liquid. Indeed, the fractographic analysis of this 80% of the cross section (Fig. 29), exactly in the
region shows a fracture morphology different axis of the radius change for the concordance
from the vicinity, with several inclusion com- section. Due to the small relative section area of
ponents protruding into the fracture surface the fracture axis, approximately 20% of its cross
(Fig. 27). The EDX microanalyses of these section, it was deduced that the stress for the in-
components show a chemical composition with service component was relatively low.
a high level of carbon, which suggests that these Ten measurements of Rockwell C hardness
components are of iron carbide. Moreover, were carried out, according to ASTM E18, on
several longitudinal cracks similar to the one the surface of the axle near the fracture region. A
found in the blade body were observed. mean hardness of 33.9 HRC was obtained. This
Nonfusible longitudinal cracks exist along value is well below the expected one of 43 HRC.
the affected area in the blade. The large variety Figure 30 shows the microstructure of steel in
of defects and the excessive mechanical vibra- the reduced section on a longitudinal cut plane in
tion of the blade are probably the main causes of the vicinity of fatigue crack nucleation. The
crack nucleation by fatigue in the material near material presents a large amount of globular or
the blade bottom. They culminated in the cata- granular bainite, in agreement with the relatively
strophic fracture of the component. low value of hardness of the fractured axle
The recommendation includes a more effi- surface.
cient quality control of the manufacturing pro- Figure 31(a) shows a general topview of the
cess of the blade material and avoiding the fracture surface in the region where there was
occurrence of casting defect formation, slag fatigue crack nucleation, indicated by the arrow
inclusions, and other impurities. at bottom. The five clustered arrows point in the
Failure in the Axle of a Reduced Section in direction of fatigue crack propagation advance.
a Rotating Component. Possible causes were The arrow at the top shows a dark region, orig-
investigated for failure in the area of an inter- inated by contamination of the fracture surface
mediate reduction. The rotating component with oil or grease. Figure 31(b) shows in detail
fractured completely after intermittent loading. the fatigue crack nucleating site that probably
Figure 28 shows an outline of the component started at an inclusion located exactly on the
and the axle region where the cracks developed. circumference surface of the reduced section in a
The chemical composition (in weight percent) of region in the internal concordance radium. The
the fractured axle material is provided in presence of some inclusions in the proximity
Table 8. The results show that the axle material of the fracture site is pointed out by white arrows
is a DIN-specified 17CrNiMo6 steel. The spe- in Fig. 31(b). Indeed, fractographic analysis
cifications of the material mechanical properties has shown the possibility of the existence of a
at room temperature are given in Table 9. concentration of inclusions in the nucleation
The visual inspection of the fracture surface region of the fatigue crack. Figure 32(a) con-
(Fig. 29) indicated an extremely flat aspect, such firms the high level of inclusions in the region,
as the ones typically displayed in fatigue cracks. indicated by white arrows, with signs of moving
Fig. 26 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.
by second-phase components indicated by black analysis of the inclusions shows a massive

arrows, similar to Fig. 31(b). Figure 32(b), the presence of aluminum, sulfur, and calcium
same image shown in Fig. 32(a) but with back- elements.
scattered electrons instead of secondary ones, It is worth noting that these inclusions act,
reveals the great amount of inclusions (darker) on a microscopic scale, as metallurgical stress
in the metallic matrix (lighter). The chemical concentrators. The presence of these second-
phase 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 inclu-
sion placed near the finished axle surface,
exactly in the internal component of the con-
cordance radius machined in the section change.
The combination of the effects of stress con-
centration generated by both discontinuities,
metallurgical (inclusion) and geometric (curva-
ture 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 cir-
cular 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 dis-
continuities, with an appearance similar to
manganese sulfide inclusions, that are invariably
Fig. 27 (a) General view of the probable initial region of present in mechanical construction steels. The
crack nucleation by fatigue crack. (b) Magnification
of the region in the box at the left in (a). (c) Magnification of the presence of a grayish second phase, intermediate
region in the box at the right in (a) to the metallic matrix (lighter) and the voids
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 Table 9 Mechanical properties at room

material temperature
Composition, wt% sE, MPa sR, MPa AF, % QF, %
C Mn Si P S Ni Cr Mo 742 1080 20 57
0.17 0.63 0.23 0.10 0.011 1.45 1.59 0.30
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 Fig. 30 Microstructure of the axle according to a long-
fracture of the remaining section is indicated by FF. itudinal cut plane. Etched with 2% nital
(darker), is observed inside these dis- also generate a preferential path for crack
continuities. This material component fills the propagation. It is worth emphasizing that the
larger discontinuities, while the smaller dis- majority of these discontinuities were found
continuities are almost totally filled by the sec- aligned in the direction of the thermomechanical
ond phase. work to which the axle was submitted during its
Figure 37 shows that the most subtle dis- manufacture (i.e., longitudinal direction). The
continuities have a rather slim, cracklike aspect inclusions are disposed on parallel planes to the
and consequently present a great capacity main fracture of the component during heat
to concentrate high tensile stresses. In these treatment. This suggests the possibility that
terms, it is possible to assume that these second these inclusions played a fundamental role in the
phases are potential crack nuclei, and that they catastrophic failure of the 52100 steel axle.
Fig. 31 Fatigue crack site. (a) General view. (b) Detail. The Fig. 32 Concentration of inclusions near the fatigue crack
inclusion that originated the site was removed from site. (a) SEM image with secondary electrons.
the fracture surface. SEM image with secondary electrons (b) Backscattered electrons
Figure 38 shows elongated microvoids, alignment of the discontinuities generate a

obtained by SEM with secondary electrons, in favorable path for material cracking. The voids
the vicinity of the tip of one of the cracks that that are already interconnected by material
propagated in the fractured component. The 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 Inclusion-like microdefects detected in the vicinity

of a crack in the fractured axle, located at the center
Fig. 34 Detail of the starting point of brittle fracture in the of the component
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 Fig. 37 Slim, cracklike inclusions in the 52100 steel. The
component thickness. Etched with nital. Original inclusions are oriented in the longitudinal direction
magnification: 400 of the component. No etch
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 elon-
gation and inclusion alignments. The arrows point to the exis-
tence 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 pre-
cipitates (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 pre-
cipitates, 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. Fig. 39 52100 steel microstructure in the center of the
It was concluded that the raw material used to component thickness. Etched with nital. Solid arrows
point to free cementite in the globular form, and white arrows
manufacture the fractured axle was probably point to Fe3C in the form of platelets in the pearlite contour.
contaminated with iron oxide. The contaminant Original magnifications: (a) 3000 . (b) 10,000. (c) 18,000
was in the form of elongated inclusions, aligned
in the longitudinal part direction, making an
easy path for main crack propagation (long-
itudinal). The elongated format provided the that form the 52100 steel. The presence of this
inclusions the capability to concentrate high fragile phase may have contributed, to a certain
tensile stresses and then transform them into extent, to the intergranular secondary brittle
potential crack nucleation sites. crack propagation in the catastrophic frac-
Microanalysis also confirmed the existence ture of the component during annealing heat
of free cementite in the pearlitic grain contours treatment.
ACKNOWLEDGMENTS 14. J. Campbell, Castings, Butterworth-

Heinenmann, 1993
Thanks to the Department of Materials, Aero- 15. J.G. Davies, Solidification and Casting,
nautics and Automotive Engineering of the School of
Applied Science, 1973
Engineering of Sao Carlos, University of Sao Paulo,
on behalf of Professor Dr. Dirceu Spinelli, for the 16. AFS Inclusion Atlas Homepage, http://
collaboration on failure analysis case studies. neon.mems.cmu.edu/afs/afs2/ (Accessed on
March 2005)
17. J.A. Horwath and G.M. Goodrich, Micro-
Inclusion Classification in Steel Casting,
REFERENCES AFS Trans., 1995, p 495510
18. J.M. Svoboda, R.W. Monroe, C.E. Bates,
1. Defects and Distortion in Heat-Treated and J. Griffin, Appearance and Composition
Components, Heat Treating, Vol 4, ASM of Macro-Inclusions in Steel Castings, AFS
Handbook, ASM International, 1991, Trans., 1987, p 187202
p 13201325 19. C.R. Wanstall, J. Griffin, and C.E. Bates,
2. R.E. Reed-Hill, Physical Metallurgy Prin- Clean Steel Cast Technology, Research
ciples, 2nd ed., Van Nostrand, 1982 Report 106, Steel Founders Society of
3. Casting Design, Casting, Vol 15, ASM America
Handbook, ASM International, 1988, 20. R. Sankaranarayanan and R. Guthrie, Slag
p 13011322 Entrainment through a Funnel Vortex
4. Dimensional Tolerances and Allowances, during Ladle Teeming Operations, Pro-
Casting, Vol 15, ASM Handbook, ASM ceedings of the International Symposium on
International, 1988, p 13361347 Developments in Ladle Steelmaking and
5. Quenching and Control of Distortion, ASM Continuous Casting, Aug 1990 (Ontario),
International, 1988 CIM, 1990, p 6687
6. Residual Stress, Heat Treating, Vol 4, ASM 21. R. Sankaranarayanan and R. Guthrie, A
Handbook, ASM International, 1991, Laboratory Study of Slag Entrainment
p 13251343 during the Emptying of Metallurgical Ves-
7. M.T. Milan, O. Maluf, D. Spinelli, and sels, Steelmaking Conference Proceedings
W.W. Bose Filho, MetaisUma Visao (Ontario), 1992, p 655664
Objetiva (Metals A Vision Object) 22. K. Noguchi et al., Zairyo to Purosesu (Curr.
Suprema, 2004, p 148149, 161162 Adv. Mater. Process.), Vol 4, 4th ed., 1991,
8. Gases in Metals, Casting, Vol 15, ASM p 11941197
Handbook, ASM International, 1988, 23. H. Nakamura, S. Kohira, J. Kubota, T.
p 175189 Kondo, M. Suzuki, and Y. Shiratani, Tech-
9. Inclusion-Forming Reactions, Casting, Vol nology for Production of High Quality Slab
15, ASM Handbook, ASM International, at High Speed, Steelmaking Conference
1988, p 190211 Proceedings (Ontario), 1992
10. R.D. Pelke and J. Elliott, Trans. TMS- 24. I. Manabu, S. Yutaka, O. Ryusuke, and M.
AIME, Vol 227, 1963, p 894 Zen-ichiro, Evaluation of the Critical Gas
11. P.C. Glaws and R.J. Fruehan, Metall. Trans. Flow Rate Using Water Model for the
B, Vol 17, 1986, p 317 Entrapment of Slag into a Metal Bath Sub-
12. R.J. Fruehan, B. Lally, and P.C. Glaws, ject to Gas Injection, Tetsu-to-Hagane
Proceedings of the Fifth International Iron (J. Iron Steel Inst. Jpn.), Vol 79 (No. 5), p 33
and Steel Congress (Washington, D.C.), 25. J.M. Harman and A.W. Cramb, A Study on
Iron and Steel Society of AIME, 1986 the Effect of Fluid Physical Properties on
13. A. Kagawa and T. Okamoto, Trans. Jpn. Droplet Emulsification, Steelmaking Con-
Inst. Met., Vol 22 (No. 2), 1981, p 137 ference Proceedings, 1996, p 773784
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, amounts of carbon dioxide, water vapor, and
crankshafts, camshafts, bearings, and others, re- methane). Endogas is produced by reacting a
quire a differentiated response of the surface and hydrocarbon gas, such as natural gas (methane),
core to external loading. This can be accom- propane, or butane, with air.
plished by surface (case) hardening methods After the diffusion process is completed, the
such as induction and flame hardening or by component may be quenched from the carbur-
surface diffusion processes such as carburizing izing temperature or reheated to austenitize the
and carbonitriding. Raja et al. have reported that steel, and then quenched. Bainite formation in
case carburizing is one of the most common heat the case is strongly inhibited by the presence of
treatments for steel, accounting for 50% of all molybdenum and chromium. Since the surface
surface treatments (Ref 1). Case carburizing contains higher carbon content than the core, it is
involves the creation of a gradient that exhibits harder than the softer core. Core hardness is
high hardness, brittleness, and strength in the most strongly affected by the presence of molyb-
surface and greater toughness and ductility in denum and manganese. Chromium exhibits a
the softer core in order to provide optimal moderate effect, and nickel exhibits a weak
(Ref 2): effect (Ref 3). Core hardness is strongly affected
by the quenchant selection and quenching tem-
Wear resistance
perature.
Resistance to scoring
Bending and/or torsional fatigue strength In addition to strengthening the case, the
increased carbon content also provides desirable
Rolling-contact fatigue strength
increased compressive stresses that will inhibit
These properties are optimized by maximiz- fatigue crack initiation. The lower carbon con-
ing surface compressive stresses, and carburiz- tent in the core also will produce improved
ing is one of the most effective and commonly fatigue strength.
used methods to impart compressive stresses to Carbonitriding is similar to carburizing in that
the surface of a component (Ref 3). The focus of it is a diffusion process that involves the simul-
this chapter is on carburized and carbonitrided taneous diffusion of carbon and nitrogen (from
materials. ammonia) into the steel surface. To obtain
Gas carburizing, which is the most widely maximum strength, the carbonitriding process
used carburizing process, is a surface diffusion produces a surface that is enriched in nitrogen
process where the carbon concentration in a and carbon in the form of an epsilon
surface layer (case) of a steel matrix that is (e)-carbonitride layer and a diffusion zone
predominantly iron, chromium, and nickel is containing chromium-iron carbide, (Cr,Fe)7C3;
increased by heating the component at approxi- chromium carbide nitride, Cr62C3 5N0.3; chro-
mately 850 to 950 C with endothermic gas mium nitride, (Cr2N) or [Cr, Fe(2Ni . . . x)]; and
(Endogas), which is a blend of carbon mon- Fe2N phases (Ref 4, 5). Typical case thicknesses
oxide, hydrogen, and nitrogen (with smaller range from 50 to 200 mm with a hardness
between 750 and 900 HV. Like carburizing, the (Ref 7). Heat treatment was the second most
case depth of carbonitrided steel is dependent often cited cause for failure. However, it is often
on both the carbonitriding diffusion time and difficult to detect the root cause of a specific
temperature, as illustrated in Fig. 1 (Ref 2, 6). failure under the conditions in which the failure
Deeper case hardnesses may be obtained by first occurred, and many of the service-related fail-
precarburizing prior to carbonitriding. Karamis ures could have been reduced with more atten-
showed that carbonitrided AISI 5115 steel tion to the other potential causes of failure
exhibited greater surface hardness and wear re- shown, since they are often interrelated.
sistance than carburized AISI 5115 steel (Ref 4). Palaniradja et al. reported that 10 to 12% of
Carbonitriding processes are typically con- carburized parts are rejected due to various
ducted in either a gas (ammonia) or a salt bath process-related defects (Ref 8). To examine this
based on trade names such as Tufftride, Nitrotec, in more detail, they conducted a Taguchi ana-
and Nitrox. Alternatively, a plasma nitriding lysis of gas carburization of AISI 8620 and 3310
process may be conducted. A brief summary steels, and their results showed that relative
comparison of carburizing and carbonitriding contribution to surface hardness was holding
processes is provided in Table 1 (Ref 2). time (20%), carbon potential (20%), carburizing
Carter has reported that failures of carburized temperature (0%), and quenching time (60%).
gears are primarily due to service-related causes, Similarly, they also studied the effects of process
such as misalignment, poor lubrication, and variables on case depth and found: holding
overloading, which constitute the greatest time (60%), carbon potential (9%), carburizing
source of all gear failures, as shown in Table 2 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
Fig. 1 Correlation of case depth of carbonitrided steels with and grinding cracks, surface decarburization,
varying diffusion times and temperatures excessive heating during grinding, excessive
removal of the case during grinding, micro-
cracking, and so on (Ref 9, 10).
Table 1 Comparison of carburizing and This chapter provides an overview of various
carbonitriding processes contributors to failure of carburized and carbo-
Process Comments nitrided components, with the primary focus on
Carburizing Hard, highly wear-resistant surface (medium case carburized components.
depths), excellent contact load potential, good
bending fatigue strength, good seizure
resistance, excellent quench cracking resistance, Table 2 Survey summary of sources of
low-to-medium-cost steels required, high capital gear failures
investment
Cause of gear failure %
Carbonitriding Hard, highly wear-resistant surface (shallow case
depths), fair contact load potential, good bending Material quality and forming 0.8
fatigue strength, good seizure resistance, good Design 6.9
dimensional control, excellent quench cracking Service-related causes 74.7
resistance, low-cast steels usually satisfactory, Manufacturing 1.4
medium capital investment Heat treatment 16.2
Sources of Failures in Carburized and Carbonitrided Components / 179
Design case and core hardness, and core strength. As an

estimate, for hardnesses within the range of 30 to
Component design may contribute directly or 45 RC, the required case depth can be calculated
indirectly to component failure. Deficiencies from (Ref 2):
such as insufficient radii or sudden changes in
section size are significant contributors to fail- Case depth to 50 HRC=(1:2 107 W)=F
ure. In addition, the presence of stress raisers,
such as those shown in Fig. 2, are among the where W is the force in pounds pressing the
most common design contributors to quench surfaces together, and F is the length of the line
cracking and fatigue failure. contact (inches).
A more comprehensive insight into design Carter has recommended the following gen-
is provided by Kuehmann et al., who developed eral design criteria (Ref 7):
a systems analysis flow chart to describe the
effects of case-core hardening in designing a If a component is carburized from both sides,
carburizing process/metallurgical structure/ the case depth should not be greater than
resulting properties and performance for the 20% of the wall thickness.
production of gears produced by three routes: At the base of gear teeth, 30% of the core
conventional forging, near-net shape casting, material should remain uncarburized.
and powder metal processing (Fig. 3) (Ref 11). Shallow case depths usually require higher
To properly design a component, it is neces- case hardness.
sary to estimate surface loading, distortion after Case depths should be five times the accep-
heat treatment, case depth and carbon profile, table wear limit.
Fig. 2 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
Although machining is proportional to the nearly ideal. The gradient may be steeper for
case depth, it should be minimal. shallower case depths.
The recommended surface carbon is 0.90 to
Kern and Suess have recommended the fol- 1.10% for 4300, 4600, 8600, 8800, and 9400
lowing general guidelines for heat treatment of carburizing steels. Although the same case
gas-carburized gears (Ref 2): depth is generally acceptable for grades such
For forgings, normalize or anneal (as re- as 4800, they are preferably reheated for
quired by the alloy being heat treated) from a hardening. The recommended surface car-
temperature at least 28 C (50 F) above the bon is 0.65 to 0.85% for high-nickel steels
carburizing temperature. such as the 4800 series, which is usually
Assure that the gears are machined prior to direct quenched.
heat treatment. To minimize cost and distortion, use direct
Bring the gear to the carburizing temperature quenching whenever possible.
with sufficient circulation of a neutral To assure optimal dimensional control,
atmosphere, and then introduce the gas used properly maintain quenching dies and
for carburizing. plugs.
For deep cases (41.5 mm, or 0.060 in.), Quench as rapidly and uniformly as prac-
adjust the carburizing atmosphere and time tical, and use spray impingement fixtures on
to produce uniform carbon diffusion from large, solid pinions that are four pitch and
the surface to the core. A decrease of 0.15 coarser.
to 0.20%/0.25 mm (0.010 in.) of depth is Use hot oil quenching on fine pitch gears.
Fig. 3 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
Case hardness of the finished gear should be possess sufficient hardenability to provide the
60 HRC or greater. desired hardness and microstructure in both the
If possible, test each gear for partial dec- case and the core. After carburizing, the com-
arburization and/or upper transformation ponent must possess sufficient toughness with-
products. out exhibiting brittle failure.
To minimize distortion and to permit quieter Most steels that are carburized are deoxidized
operation, the surface carbon content should by the addition of aluminum (commonly desig-
be uniform throughout the production cycle. nated as killed steels). Deoxidation will provide
finer grain sizes to temperatures of approxi-
mately 1040 C. Coarser grained steels may be
carburized if grain refinement by double quench-
Steel Selection and Hardenability ing is possible. Double quenching typically in-
volves direct quenching followed by reheating
Steels typically used for case hardening con- to a lower-temperature quenching a second time
tain carbon contents of less than approximately (Ref 13).
0.25%. The carbon content of the case is usually Selection of proper hardenability of steels for
controlled to between 0.8 and 1% C. The actual both carburizing and carbonitriding is critically
surface carbon content is generally limited to important, both of the core and the case, since
0.9%, because excessively high carbon content improper hardenability design can lead to un-
may lead to the presence of unacceptably high desirable nonmartensitic transformation pro-
retained austenite and brittle martensite. Some ducts in the case, leading to a potential reduction
of the most commonly used AISI grades of steel in static and dynamic fatigue strength of up to
used for carburizing are shown in Table 3 30% and a reduction of impact fatigue of up to a
(Ref 12). factor of 2.5 times (Ref 14). The hardenability
Plain carbon steels may be carburized; how- gradient of the case and the core is dependent
ever, relatively poor hardenability due to the on a number of factors, including cooling rate
lack of alloying elements reduces the carburiz- during quenching, variability of the chemical
ing response of the case. Because of the stabi- composition (alloy content, carbon and nitro-
lizing effect of the nitrogen relative to austenite, gen) of the case, and the carburizing or carbo-
carbonitriding provides greater hardenability nitriding method being used.
than attainable with carburizing. Therefore, Core hardenability is being used increasingly
plain carbon steels respond well to carboni- to specify alloy steels used for case hardening
triding. where the hardenability of both case and core
Proper steel selection is a critically important must be considered. Details for the traditional
process to provide the desired case depth and approach for the experimental determination of
microstructure and the required core properties. hardenability of carburizing steels are provided
Typically, the case structure should be fully in Ref 15. Jominy curves for a number of car-
martensitic, with the exception of allowing for burizing steel alloys with varying hardenability
required application design limits on retained are shown in Fig. 4 (Ref 16).
austenite content. For example, the steel must Procedures have also been described for
determining ideal diameter (DI) values and
hardenability of carburizing steels from Jominy
data using regression equations for composition
Table 3 Common carburizing grades of steel
and their relative processing features and grain size (Ref 17, 18). The ideal diameter is
defined as the diameter of a cylindrical steel bar
AISI
steel that will form 50% martensite at the center when
grade Note subjected to an ideal quench. Hardenability
4620 Lower-cost, chrome/nickel/molybdenum steel where only differences may be substantially greater for
nominal hardenability and core response is required
8620 Most commonly specified grade. Excellent carburizing some case-hardening steel grades relative to
response, with good hardenability for most section sizes others due to the difference in carbon content in
4320 Higher hardenability for improved core response in heavier the case and core. This is more critical for heavy-
sections
4820 Increased nickel content for improved core toughness; slower sectioned components that are reheated and
response results in longer process times quenched.
9310 Maximum nickel content for maximum core toughness;
slower response results in longer process times
The hardness gradient through the case is due
to the relationship between the thermal gradient
and the carbon gradient during quenching. residual stress. Methods of determining case
Therefore, an increase in case hardenability hardenability are also provided.
required to produce greater amounts of marten- Parrish reported the following scheme that
site for a given carbon content will result in an was developed to classify the case hardenability
increased case depth. In such cases, a reduced of steels (Ref 19):
(shallower) carbon profile and shorter carburiz-
Level 1: Surface carbon contents 40.8% C
ing times will be necessary to obtain the desired
are martensitic.
hardness profile in the carburized component.
Level 2: All carbon contents from the surface
SAE J1975 standard Case Hardenability of
to 50% C are martensitic.
Carburized Steels summarizes characteristics
Level 3: All carbon contents from the surface
of carburized steels and factors involved
in controlling hardness, microstructure, and 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 dia-
meter for the critically stressed section of the
component of interest, and then the expected
level of case hardenability of that steel is deter-
mined. Figure 5 indicates that level 4 is attain-
able 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 repre-
sent a minimum target to be attained.
Fig. 4 Jominy hardenability data for a number of carburizing Case hardenability may vary widely even for
steels steels with equivalent core hardenabilities. Kern
Fig. 5 Case hardenabilities of a number of carburizing steels with oil quenching

and Suess provided the following guidelines extending inward from the surface, where the
(Ref 2): hardness is greater than that of the core. The total
case depth is the distance or thickness of the
Steel grades in which the case hardenability carbon-enriched surface layer. The effective
is due to carbide-forming metals such as case is the point where 0.4 to 0.5% C (percent is
chromium (8600 series) are sensitive to called points in the industry) is present if the part
microcracking, especially when direct quen- is hardened to 50 HRC (510 HIV). The depth of
ched from the carburizing temperature. This the case is a function of carburizing time and
can be controlled by restricting the carbon carbon (carbon potential) at the surface. Genel
content in the case to 0.9%. and Demirkol have reported that the following
Steel grades with relatively high nickel equation model can be used to predict effective
content, for example, 4800 and 9300 series, case depth (Ref 21):
may form excessive (430%) amounts of
retained austenite when direct quenched Effective case depth (mm)=
unless the carbon content of the case is 0:41 Carburizing time (h)1=2
maintained at 50.75%.
Carburizing round section sizes greater than The carbon potential of a furnace atmosphere
76 mm (3.0 in.) may lead to difficulty in at a specified temperature is defined as the carbon
achieving the desired case and core micro- content of pure iron that is in thermodynamic
structures when quenching in oil. In such equilibrium with the atmosphere. The carbon
situations, consider induction hardening or potential of the furnace atmosphere must be
nitriding or using a highly alloyed steel greater than the carbon potential of the surface of
grade such as AISI 9310. the workpieces for carburizing to occur. The
Some standard grades of steel exhibit nar- carbon potential is a measure of the ability of
rower core hardenability bands than other a gas to react with the steel surface. It is this
grades. For example, 8620H exhibits a difference (carbon content in the gas and at
hardenability band spread of 14 HRC at J 4, the steel surface) that provides the driving force
and 9310 exhibits only 8 HRC spread at the for carbon transfer to the parts being carburized.
same J-value. This provides a greater The composition of a gas that will produce a
amount of distortion control in addition to given surface carbon content is dependent on
some possible application-dependent prop- equilibrium data for the gas. The amount of
erty advantages as well. carbon transferred will depend on factors that
include temperature, time, and steel composi-
One problem that can arise during the steel- tion.
making process or that may be observed as a mill- The alloy composition of the steel will affect
to-mill variant is the presence of segregation the effective carbon potential at the surface. The
effects through the section of the steel billet presence of elements such as chromium, man-
during a continuous casting process, which re- ganese, and molybdenum that form stable car-
sults in the presence of a white band (Ref 20). bides of iron will increase the effective carbon
White band is a type of negative segregation potential. Elements such as silicon and nickel
often observed in electromagnetically stirred form less stable carbides and reduce the effec-
continuous castings. The white banding pro- tive carbon potential. Alloying elements that
duces a significant hardness gradient across the stabilize austenite or ferrite also reduce the
billet. After subsequent rolling and forging or effective carbon potential. The effect of alloying
machining to produce a component, the resulting elements on carbon potential can be calculated
grainflow can produce nonuniform hardenability from (Ref 22):
and/or soft spots that can significantly affect
distortion. Log (correction factors)=0:005 (%Si)
In addition to proper hardenability selection, 0:013 (%Mn) 0:040 (%Cr)+0:014 (%Ni)
to achieve maximum core toughness, proper 0:013 (%Mo)
austenitization and quenching to martensite is
necessary. These topics are discussed subse- It is important to note that, except for long
quently. carburizing times (3 to 410 h), the surface
Case Depth. The case of a carburized (or carbon content is typically not equal to the car-
carbonitrided) steel alloy is that portion bon potential, because the surface of the part
being carburized does not reach equilibrium with absolute temperature [Kelvin (K) = C+273,
the carburizing atmosphere. This is illustrated in Rankine (R) = F+460]. In metric units, the
Fig. 6. A different, but similar, correlation exists case depth (mm) is (Ref 24):
for various steel alloys and carburizing con-
ditions. From a correlation such as Fig. 6 and Case depth (mm)=660 e8287=T t0:5
given case depth requirement and carburizing
time, it is possible to correct for the carbon At the operating temperature, the carburizing
potential under nonequilibrium conditions to process may be conducted in two parts. Car-
achieve a specific surface carbon content, as in- burizing occurs during the first part of the pro-
dicated by the dashed lines in the figure (Ref 22). cess in a high-carbon-potential period when the
For a specific temperature, the case depth (d) enriching gas is added to the furnace atmosphere
will vary with the square root of the carburizing to increased the carbon content of austenite (the
time (t): carburize-boost period) and the carbon potential
is greater than the desired carbon potential. This
d=Qt1=2
part of the process is typically conducted at a
carbon potential close to the solubility limit of
Values of the depth factor (Q) as a function of
carbon in austenite, typically between 1.0 and
temperature are shown in Fig. 7. This equation is
1.2% C, which is dependent on the temperature
reported to be valid for low-carbon steel and
and alloy content of the steel. The time for this
some alloy steels (Ref 22). For applications such
part of the process to occur is called the car-
as automotive gears, typical case depths are 0.8
burizing time. This part of the process is fol-
to 1.4 mm (Ref 23).
lowed by a boost-diffuse period, where the
An equation that is often used to calculate the
process is operated at the equilibrium carbon
case depth (in.) as a function of both time and
potential, which is reduced to a level that will
temperature is the Harris equation (Ref 24):
maintain surface content, typically 0.8 to 0.9%
C, during which time the carbon will diffuse
Case depth (in:)=31:6 t0:5 =10(6700=T) deeper into the case and provide a gradual case/
where t is the time at the carburizing tem- core transition. Together, this is called the boost-
perature in hours (carburizing time), and T is the diffuse cycle. The time for this part of the pro-
cess is called the diffusion time. When the
required case depth is achieved, if the compo-
nent is direct quenched, the temperature is
Fig. 6 Carbon dioxide content of the atmosphere required to

produce certain surface carbon levels at different car-
burizing times under a given set of carburizing conditions. The
dashed lines illustrate alternative times and carbon dioxide Fig. 7 Variation of the depth factor, Q, with carburizing
contents to produce a single surface carbon content. temperature for low-carbon and certain alloy steels
lowered to 850 C to reduce the distortion and If excessively high carburizing temperatures
then quenched (Ref 25). are used, the following situations may occur:
Carburizing boost-diffuse cycles conducted
over 2 h are advantageous for case depths Rapid increase in grain growth and loss of
40.50 mm (0.020 in). They are also useful properties
when relatively deep cases free of carbides or Increased energy consumption
retained austenite are required. This is important Increased deterioration of the furnace fix-
when the carbon content is greater than the tures and baskets
eutectoid composition, where there is an in-
creased tendency to form carbides and retained When high carbon potentials and long car-
austenite upon quenching. These effects in- burizing times are used to produce high surface-
crease with alloy content (Ref 24). carbon content and deep case depths, excessive
Harris also developed equations to compute retained austenite and/or free carbides may be
the carburizing and diffusion times to achieve a obtained as a result. These microstructural pro-
specific case depth and surface carbon content ducts exhibit adverse effects on residual-stress
(Ref 24): distribution (which is discussed subsequently).
Therefore, although high carbon potentials may
Carburizing time (h)=(C Ci )2 =(CO Ci ) be used for short carburizing times, substantial
deleterious effects may result if used over pro-
Diffusion time (h)=Total time Carburizing time longed carburizing times.
Excessive carbon potentials, gaseous atmo-
where C is the final desired surface carbon sphere composition control problems because of
content, CO is the surface carbon content at the carbon probe malfunctions or air ingression, and
end of the carburizing cycle, and Ci is the carbon inadequate furnace purging can lead to excess
content at the core. free carbon and sooting, which may be suffi-
The effect of the steel alloy composition on ciently severe as to leave carbon deposits on the
the carbon gradient is illustrated for AISI 1020 parts (Ref 24, 26). This can lead to correspond-
plain carbon steel and AISI 8620 after carbur- ing problems in controlling the carbon potential,
izing at three temperatures in Fig. 8 (Ref 24). resulting in nonuniform carburizing and dimen-
The alloy content will influence the diffusion sional control problems.
rate, but its greatest effect is on the case carbon Quenching. During carburizing, the steel
content. Normal carbon gradients, such as those microstructure consists of polycrystalline aus-
shown in Fig. 8, can be achieved by maintaining tenite. Grain coarsening may occur if the car-
a saturated austenite condition at the surface burizing times are relatively long. The austenitic
during the entire boost-diffuse carburizing cycle grain size will determine the size and distribu-
(Ref 24). tion of martensite that will form as a result of
It is important to control the ratio between the quenching. In addition, phosphorus segregation
boost and diffuse times and to carefully control into the grain boundaries may occur during the
the carbon potential to avoid obtaining a carbon carburizing cycle, which has been found to be
profile such as that shown in Fig. 9 (Ref 24). directly dependent on phosphorus and carbon
Although the desired surface hardness was content. Hyde et al. found that fatigue and
obtained, the lower carbon content at the surface fracture toughness decreased slightly when the
can lead to a transformation that proceeds simul- phosphorus content increased from 0.005 to
taneously outward from the case-core interface 0.017%, and when 0.017 to 0.031% P was pre-
and at the surface and proceeds inward such that sent, the endurance limit and fracture toughness
the last portion of the case to transform is just decreased substantially (Ref 27).
below the surface. This will result in an unde- Phosphorus also affects the degree of carbon
sirable condition where the surface is in tension segregation in the form of cementite at the aus-
relative to the core as well as a corresponding tenite grain boundaries. During quenching,
decrease in fatigue strength in addition to an small amounts of cementite form at the auste-
increased potential for cracking (Ref 24). nite grain boundaries in the high-carbon case
Boyer reports that a maximum tolerable car- (Ref 25, 27). This leads to increased sensitivity
bon potential for carburizing cycles of up to 10 h to intergranular fracture, which is a major cause
at 925 C is 1.3% to avoid excessive soot for- of fatigue crack initiation in carburized steels
mation. (Ref 25).
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.
After carburizing, the parts are then either important to achieve the desired wear and fati-
quenched directly or air cooled and then rehe- gue properties. If the carburized gear, such as a
ated and quenched. Quenching is performed to spiral bevel gear, is not quenched to achieve the
harden the components. It is most desirable to necessary surface hardness and hardness gra-
develop a martensitic or bainitic case while dient, failures accompanied by micropitting and,
controlling the formation of retained austenite to ultimately, fracture may occur (Ref 28).
an acceptable level and simultaneously mini- The morphology of martensite is carbon
mizing proeutectoid and pearlitic structures. The dependent, as shown in Fig. 10 (Ref 29). At
challenge is to quench sufficiently fast to pro- lower carbon content, a lath martensitic struc-
duce the desired core structure but not so fast ture forms, while plate martensite forms at
that the higher-carbon and more brittle case higher levels of carbon. The two different mor-
cracks. In addition, the desired hardness gradient phologies are illustrated in Fig. 11 (Ref 30). Lath
between the surface and the core is critically martensite exhibits better toughness than the
higher-carbon plate martensite. Plate martensite,
as the name indicates, forms as lenticular (lens-
shaped) 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 di-
rectly 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 com-
monly, double reheat quenching may be per-
formed to provide high-durability components
Fig. 9 Carbon profile of an incorrectly carburized steel (Ref 19). Some quenching cycles recommended
Fig. 10 Dependence of the martensitic structure on carbon content

by Crucible Steel for carburized 8620 steel both the case and core grain size. The refined
include (Ref 31): core will be soft and machineable with
maximum toughness and resistance to
Direct quench from carburizing: Quench in
impact. The refined case will be hardened for
oil directly from the carburizing temperature
wear and resistance. In addition to grain
of 925 C (1700 F). The core is hardened
refinement, double reheating and quenching
but unrefined. The case is hardened to the is reported to improve fatigue properties by
extent that it will be fileproof if the carbon reducing the size and density of microcracks
content is sufficiently high. in the structure (Ref 21).
Cool, reheat and quench (1): After cooling
from the carburizing temperature (925 C, Of these methods, the most common is direct
or 1700 F), the carburized (but not yet quenching. However, there are a number of rea-
hardened) steel is reheated to a temperature sons why reheat quenching is favored for higher-
above the upper critical temperature, Accm, alloy, case-hardening steels, including (Ref 19):
of the core (835 C, or 1535 F) and then
To assure grain size and retained austenite
quenched in oil. The core will be refined and
control
exhibit maximum strength and hardness.
The case will be hardened and somewhat When intermediate subcritical heat treat-
ment is required to condition the carbide
coarsened.
structure within the case or to facilitate
Cool, reheat, and quench (2): After cooling
from the carburizing temperature (925 C, additional machining
When the parts are to be plug or die quen-
or 1700 F), the carburized (but not yet
ched for distortion control
hardened) steel is reheated to a temperature
When it is not possible to direct quench, such
above the lower critical temperature, Ac1, of
as in pit carburizing
the case (730 C, or 1350 F) and then
quenched in oil to harden and refine the case. Ingham and Clarke compared the results
The core will be unrefined, soft, and machine- obtained for carburized 8620 steel with the
able, and the case will be hardened. carbon gradient shown in Fig. 12, which was oil
Double reheat and quench: The steel is quenched from a direct quench following car-
cooled in the furnace from the carburizing burizing and by a reheat and oil quench cycle
temperature of 925 C (1700 F). The steel (Ref 32). The results obtained showed that the
is then reheated to above 730 C (1350 F) direct quench process yielded a higher hardness
and oil quenched to refine the core. The steel than the reheat and quench process, which ex-
is again reheated to 730 C (1350 F) and oil hibited a relatively lower as-quenched hardness
quenched to refine the case. This double due to the presence of bainite in the hardened
heating and quenching procedure refines case structure.
Fig. 11 Martensite morphology. (a) Lath martensite. (b) Plate martensite. Source: Ref 30
Quenchants must be selected to provide creating large thermal gradients between the
cooling rates capable of producing an acceptable core and the surface of the part. Poor agitation
microstructure and hardness gradient through design is a major source of quench nonuni-
the case and the core. However, it is not de- formity, since the purpose of the agitation sys-
sirable to use quenchants with excessively high- tem is not only to take hot fluid away from the
heat-removal rates, since the propensity to cause surface and to the heat exchanger but also to
increased distortion or cracking increases with provide uniform heat removal over the entire
quench severity. Although a reduction of quench cooling surface of all of the parts throughout the
severity leads to reduced distortion, it may also load being quenched.
be accompanied by undesirable microstructures. A wide range of quench media can potentially
Therefore, it is essential to select optimal quen- be used when quenching carburized parts. Some
chant and agitation conditions for the required comments on quench media selection, provided
microstructure, hardness, and strength in critical by Boyer, include (Ref 34):
sections of the parts for each steel alloy, section
size, and required microstructural and mech- For carbon steels, the most common
anical properties. Actual cooling rates or heat quenchants are water and brine. When water
fluxes provided by a specific quenching medium is used as the quenchant, bath temperatures
are typically unavailable. However, some illus- of 20 to 30 C with agitation are the most
trative comparative data are provided in Table 4 common.
(Ref 33). Figure 13 illustrates the comparative In the industry, oil quenchants are the most
cooling properties of various oil-quenched steel common, particularly when integral-quench
bars assuming a surface heat-transfer coefficient (sealed-quench) furnaces are used at tem-
of 0.019 cal s 1 C 1 cm2 (Ref 32). peratures of 25 to 70 C. The quench oils
Quench nonuniformity is a significant con- may be classified as fast, intermediate, or
tributor to quench cracking. Quench non- slow depending on the cooling rate, enhan-
uniformity can arise from nonuniform flow cing additive, and quench oil base stock
fields around the part surface during the quench being used. When distortion control is cri-
or nonuniform wetting of the surface. Both lead tical, a hot oil that can be used at tempera-
to nonuniform heat transfer during quenching, tures as high as 175 C may be used.
Fig. 12 Comparison of direct quenching and reheat and quenching of 10 cm (4 in.) diameter AISI 8620 steel after oil quenching
It is possible to use aqueous polymer When distortion control is especially cri-

quenchants, even in integral-quench fur- tical, salt bath quenching may be required.
naces, if appropriate structural conditions However, parts should never be transferred
are met. The user is advised to consult his directly from a carburizing bath containing
furnace manufacturer prior to use. A wide 45% cyanide to a nitrate-nitrite quench
range of quench severities is possible by bath, because this will result in a violent
varying the polymer concentration, bath reaction and possibly an explosion (Ref 34).
temperature, and agitation. In one study, it
One often-encountered quenching problem
was shown that an aqueous polymer quen-
that may lead to increased dimensional control
chant produced substantial improvements
problems is contamination. For example, het-
in fatigue properties relative to a conven-
tional quench oil, which was attributed to erogeneous quench media caused by water
contamination of oil or oil contamination of
an improvement in quench uniformity
water or aqueous polymer solutions can poten-
(Ref 35).
tially cause cracking problems. Similarly, salt
contamination, either from salt baths or hard-
Table 4 Comparison of typical heat-transfer
metal ion contamination, can lead to problems of
rates
cooling rate control. Solid contamination, such
Maximum surface
heat-transfer rate, as sludge or soot contamination in oil or aqueous
Quench medium W m 2 K1 media, also may lead to distortion and cracking.
Still air 5080 Finally, excessive foaming and air entrainment
Nitrogen (1 bar) 100150
Salt bath or fluidized bed 350500
of the quench media will lead to nonuniform
Nitrogen (10 bar) 400500 cooling, soft spots, increased residual stresses,
Helium (10 bar) 550600 and cracking. Therefore, it is essential that the
Helium (20 bar) 9001000
Still oil 10001500 quench bath be well maintained to assure opti-
Hydrogen (20 bar) 12501350 mal distortion control and minimize the poten-
Circulated oil 18002200 tial for cracking.
Hydrogen (40 bar) 21002300
Circulated water 30003500 To develop the optimal residual-stress gra-
dient, 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
conditions for the steel grade of interest. times (30 to 60 min), which provides a relatively
Figure 14 shows the development of the tem- shallow case, usually 50.5 mm (Ref 36).
perature distribution through 12.5 mm diameter These results show any factor that affects the
bars carburized to a depth of 0.9 mm for oil- cooling profile, such as bath temperature, agi-
quenched alloy steel and a water-quenched mild tation, or the quenchant selection, will exhibit a
steel (Ref 36). This figure shows that the Ms and corresponding effect on the thermal distribution
Mf temperatures are reduced with increasing through the carburized case upon hardening
carbon content in the case. The isochronal lines and therefore on the development, type, and
in the figure illustrate the cooling profile through magnitude of residual stresses. To assure opti-
the case to the core at specified time intervals. mal distortion control, the following variables
The martensitic transformation occurs at the should be carefully monitored and controlled:
core/case interface first. The case transforms last
along with the corresponding expansion of Adequate quality-control procedures of the
martensite in the case. Since the core is already quench media should be in place. For
transformed, this process restrains further ex- examples, follow ASTM D6710, Standard
pansion during continued cooling, leading to the Guide for Evaluation of Hydrocarbon-Based
development of surface compressive stresses Quench Oils, for oil quenchants and ASTM
and placing the core in relative tension (Ref 36). D6666, Standard Guide for Evaluation of
As this figure shows, since the Mf is less than Aqueous Polymer Quenchants, for aqueous
ambient temperature, retained austenite will polymer quenchants.
accompany this process. Carefully control the water content if poly-
Because carbonitriding is similar to carbur- mer quenchants are being used.
izing, except that nitrogen and carbon diffusion Replace the quench mediate at regular use-
into the case is involved, both processes exhibit level intervals.
similar microstructural transformation and Carefully control the quench bath tempera-
evolution of residual stresses upon quenching. ture.
Process temperatures during carbonitriding are Monitor fluid flow variation at critical loca-
typically lower (800 to 850 C), as are process tions in the quench tank. Nonuniform quen-
ching 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 quen-

chant temperature is a critically important vari-
able in controlling the amount of retained
austenite in the carburized steel. This is impor-
tant because incomplete quenching and the pre-
sence 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 tem-
perature 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 transfor-
mation during quenching of carburized 12.5 mm where C, Mn, Ni, Cr, and Mo are the percent of
diameter steel bar. The curves (isochronal lines) in the figure
indicate time in seconds after immersion of the carburized the element contained in the steel. This equation
(0.9 mm case) bar into the quenchant indicated. is only accurate for steels containing up to 0.5%
C. For steels with higher carbon content, Fig. 15 retained austenite content was dependent on the
should be used to determine a more accurate carbon content and alloying elements and
value for Ms (Ref 37). the content of the carburized case (Ref 39). The
Although the degree of transformation higher the steel temperature prior to quenching,
between the Ms and martensite finish (Mf) the greater the decomposition of austenite but
temperatures is not linear, the difference is es- the greater the potential for cracking. To prevent
sentially constant (Mf is approximately 215 C cracking, steel may be quenched in hot oil, but
lower than the Ms temperature) (Ref 37). the amount of retained austenite increases.
Nevertheless, if Ms Mf is approximately con- Additional dilatometer examination of an iso-
stant, incomplete transformation can be expec- thermal high-temperature tempering process
ted if some part of the transformation occurs at with subsequent cooling showed there was
a temperature lower than the quench bath tem- either a volume shrinkage due to decomposition
perature. Therefore, the volume of untrans- of tetragonal martensite or an expansion caused
formed austenite (Vc) is related to both the Ms by the decomposition of retained austenite
and the quenchant temperature (Tq). This rela- (Fig. 16) (Ref 39). These transformations oc-
tionship is quantitatively defined by the well- curred during heating, isothermal holding, and
known Koistenen and Marburger equation: subsequent cooling. As a result of this work,
72 Trusova showed that reducing the carburized
Vc =e71:10 10 (Ms 7T q )
steel temperature to 800 C prior to quenching
Using these equations, Parrish demonstrated would reduce the potential for cracking and a
the effect of quenchant temperature on retained double tempering at 580 to 600 C for carbur-
austenite on a hypothetical steel. The results of ized case structures containing i1.2% C.
these calculations are summarized in Table 5, Because of the alloy and high carbon content
which show that the amount of retained austenite in many case-hardened steels, the Ms tempera-
is expected to decrease with decreasing bath tures in the carburized case are typically
temperature. between 100 and 200 C or lower. These values
Trusova studied the formation of retained will vary with the carbon content in the case.
austenite after quenching and showed that the Therefore, the Mf temperature, which is approxi-
mately 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 Quenchant Approximate Estimated
start (Ms) temperature Ms Tq, retained hardness,
temperature, C (Tq), C C austenite, % HRC
150 80 70 45 52
Fig. 15 Correlation curves for correcting the Steven and 150 60 90 35 56
Haynes Ms temperature equation. When the carbon 150 40 110 29 57
content is less than 0.9%, an 830 C soak of over 2 h is required 150 20 130 25 58
to produce a fully austenitic structure.
achieved by mechanical refrigeration. For lower After cold treatment, the presence of retained
temperatures, down to 195 C, liquid nitrogen austenite can be assessed by comparing the
can be used (Ref 41). 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 aus-
tenite, 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 Retained austenite, %
temperature before subzero
treatment at 183 C Steel Kh-12 Steel SKh-12
23 min. 24 36.5
24 h 46 54.5
45 days ... 55
Fig. 16 Trusova dilatomer curves for tempering of carbur-
60 days 48 ...
ized steels quenched from 950 C
Table 7 Effect of subzero cooling after quenching

Bending Impact
strength strength
Heat treatment Retained
after carburizing Condition austenite, % Hardness, HRC MPa kg/mm2 MPa kg/mm2
Oil quenched As-quenched 62 54 1530 156 25.5 2.6
from 800 C, Subzero treated 20 62 1442 147 19.5 2.0
low-temperature
tempered
Tempered at 650 C, As-quenched 34 60 1697 173 40 4.1
oil quenched Subzero treated 10 62 1608 164 ... ...
from 800 C,
low-temperature
tempered
Air-cooled from As-quenched 90 47 1618 165 59 6.0
900750 C, oil Subzero treated 20 60 1353 138 19.5 2.0
quenched
low-temperature
tempered
Note: Steel is 18Kh2N4VA. Subzero treatment conducted at 120 C.
affected (Ref 43). In addition, subzero treatment It is assumed, as Fig. 19 shows, that increasing
increases the stability and induces microresidual the retained austenite content will reduce the
stresses in any remaining retained austenite (Ref low-stress, high-cycle fatigue limit of carburized
43). Although subzero treatment does decrease steel (Ref 37, 45, 46).
retained austenite, in view of these complica- In a study using a flexural four-point bending
tions, Parrish has suggested that optimization of fatigue test and carburized SAE 8620 steel test
the following process variables be considered specimens, it was shown that increasing the
before employing this process: quenchant tem-
perature, surface carbon content, steel compo-
sition, 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 in-
creasing strains can lead to the transformation of
retained austenite to martensite (Ref 37). How-
ever, conflicting test results make it difficult to
predict if the martensite formed by such strain-
induced 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 marten-
site 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 posi-
tion where the ratio of martensite/retained aus-
tenite is maximum (Ref 37).
Fig. 17 Dependence of stress for first detectable plastic strain

(approx. 0.0001) on retained austenite content. Fig. 18 Residual-stress distribution and retained austenite
AQ, air quenched; OQ, oil quenched; T, tempered content in case-hardened steels
retained austenite in the case resulted in longer are attributable to uniform retained austenite
fatigue life and that fatigue life is directly pro- content throughout the microstructure, only a
portional to the square root of the grain size maximum of approximately 40% of the retained
(d 0.5) (Ref 47). Jeddi et al. also showed that austenite transformed to martensite at any depth
fatigue strength of carburized 14NiCr11 steel due to the cyclic loading.
improved with increasing retained austenite con- The impact fatigue resistance is also depen-
tent in the hardened case (Ref 48). In this work, it dent on the amount of retained austenite and the
was shown that the level of improvement was level of applied stress (Ref 37). Figure 20 shows
related to the microstructure and residual-stress that impact fatigue resistance actually increases
distribution within j200 mm of the surface. In with increasing retained austenite at the highest
addition, although the improvements in fatigue 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 car-
burized SNCM21, which corresponds to SAE
8620 steel, and carburized SCM4, which corre-
sponds 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 resis-
tance of carburized steels. Kozlovskii studied
the scoring resistance of carburized 20Kh2N4A
Fig. 19 Fatigue limits of plasma- and gas-carburized test (0.21% C, 0.62% Mn, 0.20% Si, 3.50% Ni,
specimens as a function of retained austenite content
Fig. 21 Effect of retained austenite (RA) on abrasive wear.

Sample A, HRC = 59.7+1.8, RA = 37; sample B,
Fig. 20 Effect of retained austenite on impact fatigue resis- HRC = 62.7+1.2, RA = 6%; and sample C, HRC = 61.4+1.5,
tance of a carburized 1.45C-11.5Cr steel RA = 23%
1.42% Cr) using a roller machine test and found and propagation. Conversely, tensile stresses
that even a relatively small amount of retained reduce desirable mechanical properties such as
austenite could exhibit a large decrease in scor- fatigue, fracture, and wear.
ing resistance, even if the hardness was affected Residual stresses are classified as the mac-
only minimally, as shown in Table 8 (Ref 50). rostress 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
Residual Stress difference between the average residual stress
with a grain and the residual stress of the first
Residual stress is defined as a tensile or
kind. Residual stress of the third kind refers to
compressive force within a material such as steel
stress variations within a grain. Residual stresses
without application of a thermal gradient or an
of the second and third kind are microstresses
external force. Residual stresses are produced by
(Ref 51). In most engineering materials, such as
phase transformation, plastic deformation, or
steel, residual-stress variation between micro-
thermal effects such as contraction upon cool-
structural phases is typically more important
ing. Newtons laws require that compressive
than microstresses. The primary focus of this
residual stresses at the surface of a material are
chapter is on macrostresses, residual stresses of
balanced by tensile stresses within the material.
the first kind.
Typically, compressive residual stress exhibits
Carburizing and carbonitriding introduce
favorable effects such as improved fatigue life
surface and subsurface compressive stresses as a
and stress corrosion by inhibiting crack initiation
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 tem-
perature, 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 sur-
Fig. 22 Sliding wear rate (at 200 rpm) as a function of face martensite now forming at the surface,
retained austenite content. A, carburized SNCM21, which puts the surface in compression relative
40 kg load, sliding distance of 864 m; B, carburized SCM4, 40 kg
load, sliding distance of 864 m; C, carburized SNCM21, 20 kg
to the core. To maintain balance, the core is
load, sliding distance of 1728 m; D, carburized SCM4, now in tension. This is illustrated by Fig. 23
20 kg load, sliding distance of 1728 m (Ref 25).
Table 8 Effect of retained austenite on scoring resistance of carburized and carbonitrided case
Compressive
Retained sustenite
stress at
Case carbon Case depth, Surface On At depth Scoring scoring
Treatment content, % mm hardness, HRC surface of 1.5 mm load, kg/cm load, kg/cm2
Carburize at 930 C for 8 h, slow cool to 750 C, 0.81 1.1 5759 48 60 930 12,020
quench in oil, temper at 180 C for 1.5 h
Carburize at 930 C for 8 h, slow cool to 20 C, 0.84 1.1 6062 8 16 355 7424
temper at 600 C for 2 h, quench from 830 C
in oil, temper at 180 C for 1.5 h
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 com-
pressive residual stresses acted throughout most
of the case (Ref 25).
It is important to consider the history of resi-
dual stresses due to dimensional changes that
may occur during any stage of the manufactur-
ing 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
Fig. 23 Schematic illustration of carbon content, retained
the ratio of the case and core thickness. When
austenite, and residual stresses in the case of car-
burized steels the core is thicker than the case, the surface
Fig. 24 Range of residual stresses obtained for 70 carburized steels

compressive stresses will be high. When the case combining carburizing and shot peening of gears
is thicker than the core, the surface compressive prepared from gas-carburized 20MnCr5 steel
stresses in the case will be lower, and the tensile (Ref 57). Before continuing this discussion, a
stresses in the core will be higher (Ref 36). For brief overview of the shot peening process is in
automotive gears, the inversion point occurs order.
where the surface compressive stresses become Shot peening (and shot blasting) involves im-
tensile, which was shown to occur at the point pinging the surface to be treated with spherical
where the hardness becomes equivalent to the media called shot. (Shot peening should be dif-
core value (Ref 54). Furthermore, the formation ferentiated from shot blasting. Shot blasting a
of surface compressive stresses was shown to process in which an abrasive material is accel-
be fundamental to the prevention of fatigue erated through a pressurized nozzle or cen-
cracking. trifugal wheel and directed at the surface of
One of the primary reasons for conducting the a part to clean or otherwise prepare the part
carburizing process is to improve the fatigue surface for further treatment, (Ref 60)). Shot
performance of a heat treated component. peening is a cold working process where each
A surface compressive stress will inhibit the individual spherical ball impinging the surface
formation and growth of surface cracks, which acts as a miniature hammer that plastically
is important since fatigue failures are typically deforms and work hardens the surface by
initiated at the surface (Ref 55). Kanetake, for creating a small indentation upon impingement,
example, showed an approximately 40% im- as illustrated in Fig. 27 (Ref 61). The indentation
provement in fatigue strength of the carburized process causes the surface to yield in tension. To
SCM420 steel (Ref 53). Shot peening is also balance the tensile forces involved in indenta-
used to create residual surface compressive tion, the subsurface is in a highly stressed
stresses that will increase fatigue properties of compression state. As the process continues, the
steel components (Ref 56). Therefore, it has indentations overlap, and a uniform layer of
been of continuing interest to examine the metal is in residual compressive stress. The
potential of fatigue property improvements not compressive stress is the result of superposition
achievable by either treatment alone (Ref 55, of residual stress formed by surrounding shots
5759). For example, Shaw et al. have reported (Ref 62). The magnitude of the compressive
up to a 75% increase in fatigue strength by stresses that are formed is material dependent
Fig. 25 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.
and is generally at least 1/2 the yield strength of shot-peened surfaces, cracking, pitting,
the material (Ref 56). Some of the variables of shallow spalling, and deep spalling may occur.
the shot peening process include the shape and In addition to providing substantial improve-
type of shot, size of shot, and impingement ments in fatigue strength, carburizing and shot
velocity. peening offer other benefits, such as reducing the
The effect of shot peening on contact fatigue deleterious effects of internal oxidation. How-
under rolling-sliding conditions was studied ever, in the absence of surface oxidation and
using a carburized and hardened 20CrMnTi steel oxide inclusions, MnS inclusions will then act as
(0.21% C, 1.09% Cr, 1.0% Mn, 0.025% Ti, fatigue initiation sites (Ref 57).
0.02% P, 0.008% S), and, as a result of that work, If a case-hardened surface is shot peened with
it was shown that failure may produce three sufficient intensity, a stress-induced transfor-
kinds of cracks, which were classified as: sur- mation of retained austenite to martensite may
face, which initiate from 0.015 to 0.05 mm; be observed, and the surrounding volume con-
shallow sub surface; and deep subsurface, which straint may result in a deepening of the surface
initiate from 0.3 to 0.5 mm (Ref 63). Depend- compressive residual stress (Ref 55, 58). Nako-
ing on the specific contact stresses on the nieczny 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
Table 9 Carbon gradient and retained austenite another study, also showed significant retained
in the 8620 carburized steel related to Fig. 25 austenite transformation and increases in
Depth below the surface
Retained
residual stress as a result of shot peening of gas-
mm in. Carbon, % austenite, % carburized 18NCD6 steel (Ref 64). Selected
0.0 0 1.20 32.0 results from this study are summarized in
0.5 0.020 1.10 28.0 Table 10.
1.0 0.040 0.80 15.0
1.5 0.060 0.40 4.0 It is known that surface structure anomalies,
2.0 0.080 0.25 0 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. 26 Residual-stress profiles of SCM420 steel that was gas Fig. 27 Illustration of the plastic deformation of the surface
carburized at 930 C, oil quenched, and tempered at and resultant stress distribution after shot peening.
200 C Source: Ref 61
gas-carburized steel (Ref 10). Structural surface Dimensional Stability

anomalies act as preferential zones for fatigue
crack initiation under low stress amplitudes. Dimensional stability has two components:
Shear-type crack growth can occur in ductile size (distortion) and shape (warpage). Distortion
retained austenite in the near-surface region of is defined as an irreversible change in the
the case for gas-carburized steels. Kikuchi et al. component during heat treatment (Ref 67).
have shown that shot peening is very effective in While changes in shape such as straightness
improving the fatigue properties of carburized (warpage) can be corrected by application of
steels with these surface structure anomalies stress or by tempering in the elastic range
(Ref 59, 65). In fact, in their study, the fatigue (reversible), size changes are irreversible and
properties were essentially comparable for a cannot be changed in this way. Metallurgically,
20 mm diameter chromium-molybdenum steel distortion may be thermally or transformation-
(0.16% C, 0.26% Si, 0.74% Mn, 0.012% P, ally derived. Size distortion typically refers to
0.013% S, 1.01% Cr, 0.18% Mo) bar that was dimensional variation due to growth or shrink-
carburized at 950 C for 1 h with and without age that is due to volumetric changes attributable
surface structure anomalies (to a depth of ap- to microstructural phase transformations (Ref
proximately 30 mm) after shot peening. As a 19, 68). Figure 28 shows the effect of the tem-
result of this work, it was concluded that since perature dependence of the specific phase
internal oxides near the surface can also act as volume of different steel transformation phases
preferential crack initiator sites, it is desirable to (Ref 69).
avoid the presence of such defects during gas Variables that affect distortion include (Ref
carburizing (Ref 65). 19, 68, 70):
Although components such as gears are car-
burized and shot peened to introduce the desired Chemical composition and hardenability
level of surface compressive stresses, there are chemical and phase composition as well as
other processes that can be used to introduce hardenability (distortion increases as hard-
compressive stresses. One such process is pre- enability increases), as shown in Fig. 29
setting. Presetting involves the introduction of (Ref 67, 71, 72)
an overload that causes yielding in the area of Steelmakinggrain size and hardenability
maximum stress concentration, such as in the Hot workinghot reduction, length and
root area of a gear tooth. When the load is direction of fiber
released, a residual stress is introduced in that Prior heat treatmentresidual stresses,
area. The induced stress is compressive on the grain size and uniformity of microstructure
side being loaded and tensile on the other side Geometrycheese blanks, shafts, rims
(Ref 66). Woods et al. evaluated presetting to Heat treatment aspectsheat rate, cooling
improve the bending fatigue of carburized AISI rate (quench severity), quenching tempera-
4120 steel spur gear teeth and found that pre- ture, jigs and fixtures, plug quenching. Plug
setting introduces compressive stresses in quenching is used to minimize dimensional
the area of a gear tooth where fatigue cracks change of inside diameters such as round-
originate. The results of this work showed ness and taper distortion of ring-shaped
that presetting provided substantially longer components (Ref 67, 68) (generally, salt
fatigue life. bath quenching yields minimum distortion
Table 10 Effect of shot peening on retained austenite transformation and residual stress of
gas-carburized 18NCD6 steel
Heat Depth modified by Retained austenite Retained austenite
Shot peening(a) treatment(b) shot peening treatment, mm on surface, % converted to martensite, % smax, MPa
None T1 ... 24.7 ... 300
T2 ... 36.8 ... 300
G1 T1 100 5.7 19.0 1450
T2 100 14.9 21.9 1350
G2 T1 200 12.1 12.6 980
T2 200 22.1 14.6 930
(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
Fig. 28 Variation of the specific phase volume of different steel transformation phases as a function of temperature. Source: Ref 69
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 distor-
tion, and as shaved thickness increases,
Fig. 29 Effect of steel hardenability on shape distortion potential distortion increases (Ref 67)
Method of green part manufacture, for
example, parts machined from bar or tubing
relative to hot oil) (Ref 67). Furnace tem- or forged from bar or tubing and then
perature uniformity, case depth uniformity, machined (Ref 70)
prequench temperature, fluid flow during Growth of surfaces during carburizing
the quench (Ref 73), number of times a Murzin et al. reported that the cooling rate
part is quenched, carburizing temperature from the carburizing temperature to the
(Ref 70) prequench temperature exhibited one of the
strongest effects on gear wheel distortion after forging was extremely undesirable since
(Ref 74). this facilitates the formation of hard-to-dissolve
particles and austenite grain growth during heat
Variables that affect warpage include (Ref 19,
treatment. Optimal warpage control was ac-
68, 75):
hieved with grain sizes not greater than grade 7.
Stresses resulting from phase transforma- A comparison was made of the effect of
tions transformation behavior on dimensional stabi-
Nonuniform residual stresses in the original lity of machine parts, such as gears, constructed
blank, such as those due to prior heating, from two different carburized steels: 20KhGR
including stress relief (0.18% C, 0.82% Mn, 0.24% Si, 1.07% Cr,
Nonuniform heating or coolingfurnace 0.20% Ni, 0.0031% B) and 12KhN3A (0.14% C,
shape, part shape, heat control 0.45% Mn, 0.21% Si, 0.78% Cr, 0.85% Ni).
Insufficient furnace timeundersoaking Since only very low cooling rates occur during
Creepmethod of stacking and fixturing of carburizing, thermal stresses would be expected
parts during heating and quenching (hanging to be minimal, and any stresses that result would
versus standing) be due to austenitic transformation, which would
Internal stresses due to machining be dependent on the steel chemical composition
(Ref 76).
Bulgakov reported that the most effective The isothermal transformation diagrams for
method of controlling stresses to reduce or both steels before and after gas carburizing are
stabilize warping was by controlling the hard- shown in Fig. 30 (Ref 76). These diagrams
enability of the steel (Ref 75). Hardenability show that austenite is less stable in the 20KhGR
control will permit control of the phase trans- alloy and that relative stability of austenite
formation and reduction of volumetric changes compared to the 12KhN3A alloy remains after
leading to structural stresses and warping. In par- carburizing.
ticular, it was shown that slow cooling of car- Analysis of the kinetics of austenite transfor-
burized steels in the range of 1100 to 900 C mation of alloys, before and after carburization
Fig. 30 Isothermal transformation diagrams. (a) 20KhGR and (b) 12KhN3A alloys before carburizing (c) 20KhGR and (d) 12KhN3A
alloys after carburizing. Source: Ref 76
(Fig. 31) over the range of cooling rates used cooling rates, the pearlitic transformation
(0.79 to 175 C/min), showed that the non- predominated for both the carburized and
carburized 20KhGR yielded mostly pearlite. uncarburized alloy.
The 12KhN3A alloy yielded predominantly When air cooled, the case of the carburized
bainite at the higher cooling rates, and at lower 20KhGR consisted of pearlite and carbides, and
Fig. 31 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
the core consisted of pearlite and ferrite. Upon The effect of phase transformation behavior
rapid cooling, the carburized case had a small in the case and the core due to thermal gradients
amount of martensite with a correspondingly such as those occurring during heating and
lower amount of pearlite, and in the core, a lower cooling will affect shape and size distortion of
amount of ferrite with the remainder being pear- small parts, as illustrated by Fig. 32 (Ref 19). If
lite was obtained. However, for the 12KhN3A the thermal and transformational stresses exceed
alloy, air cooling produced predominantly mar- the yield strength of the steel, corresponding
tensite and retained austenite in the case and distortion will occur.
correspondingly less pearlite. The core con-
tained bainite (Ref 76).
In practice, it was shown that gears manu- Quenching and Grinding Cracks (Ref 77)
factured from 20KhGB steel exhibited an aver-
age reduction in diameter (0.7 to 1.4 mm) from Quenching cracks occur when tensile
463+0.3 after carburizing. This was attributed stresses of the first kind are greater than the
to reduced stability of austenite in the pearlite material strength. Quenching cracks typically
region. Conversely, for the 12KhN3A steel, the occur during, or in some cases after, quenching at
gears increased in size (0.2 to 0.5 mm) after temperatures less than the Ms temperature. Sus-
carburizing because austenite was more stable in ceptibility to cracking increases with the carbon
the pearlite region. These data suggest that dis- content of the steel, increasing austenitizing
tortion can be controlled by alloy and cooling temperature, and cooling rate, especially in the
rate selection, or improper selection of cooling Ms Mf transformation temperature range.
rate or alloy can result in unacceptable compo- The probability of cracking during quenching
nent distortion (Ref 76). increases with the presence of stress raisers
Fig. 32 Resulting distortion after heat treatment of different steels after quenching in oil and water
such as notches, abrupt changes in section size, beneficial compressive surface stresses than
keyways, and holes. Quenching cracks are irre- those typically formed in higher-carbon-
versible but can be minimized by appropriate containing through-hardened parts. However,
design modifications, reduction in hardening internal cracks may form in the core below the
temperature, and use of lower quench severity. carburized surface or in the transition zone, that
The reasons for the occurrence of quenching is, in the places with the largest tensile stresses in
defects include: carburized parts. Cracks may also form on the
surfaces and corners of carburized parts, which
Poor surface cleanliness, such as residual
is related to triaxial tensile stresses in these
forging and metalworking lubricants, and so
locations. Therefore, to prevent cracking, the
on, which leads to nonuniform quenching,
case must be sufficiently deep so that stresses
increased thermal gradients, and soft spots
developed at any point below the surface are less
Incorrect loading and arrangement of parts
than the fatigue limit of the material at that point
in the furnace, which leads to nonuniform
(Ref 1).
heating and related distortion
When steel contains greater than 0.5% C in a
Excessive heating rates, which may lead to
martensite matrix, such as in the carburized case,
warping and cracking
intergranular fracture along prior-austenite grain
Lack of a protective atmosphere to eliminate
boundaries may occur. In this situation, the
oxidation and decarburization of the steel
intergranular fracture is due to the presence of
surface, which will lead to reduction in
both phosphorus and cementite formation on the
mechanical properties after hardening and
austenite grain boundaries during austenitizing
a decrease of hardness of the superficial
or cooling from austenitizing temperatures.
layer
Krauss has referred to this fracture mechanism
Excessive cooling rates and incorrect im-
as quench embrittlement and has suggested
mersion into the quenching bath, causing
that the mechanism for this to occur is analogous
cracking, warping, and twisting
to quench cracking in through-hardened steels,
Insufficient cooling rates or undersized
which is due to the formation of tensile sur-
quench tanks, which will inhibit the desired
face stresses during quenching, as described
martensitic transformation
previously (Fig. 33) (Ref 78). However, since
Quenching cracks, which are characterized by relatively high surface compressive stresses are
relatively large depth and short length, rarely present in properly carburized steels, quench
occur in the case of carburized or carbonitrided cracking should not exist if conditions for po-
components. This can be explained by more tential intergranular cracking are present.
Fig. 33 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).
Krauss has reported that carburized steels will flakelike fracture and subsequent peeloff. (Note:
fail by intergranular cracking if sufficent bend- Flaking is initiated at microcracks that may also
ing or tensile stresses are applied to offset the be caused by surface damage due to lubricant
compressive stresses in carburized cases. The contamination by chips, shavings, burrs, or
fracture map shown in Fig. 34 illustrates three abrasive powder ingression into the lubricating
conditions where susceptibility to intergranular system.)
cracking can be minimized: carburizing, inter- Krauss reported that fatigue resistance
critical austenitizing, and applications where decreased with increasing microcrack forma-
loading is Hertzian or compressive (Ref 79). tion. Crack initiation occurred at the site of the
In some cases, microcracking can occur with microcrack that acts as a stress concentrator (Ref
higher-carbon lath martensite matrices, which 78). Although microcracks can be removed by
are present in the hardened carburized case. surface grinding and polishing, fatigue failure
Microcracking is due to contact of martensitic may still be initiated at prior-austenite grain
plates with each other or with the austenitic grain boundaries, intergranular surface oxides, or
boundaries. The potential for microcracking surface defects such as scratches, machining
increases with the austenization temperature. marks, and surface asperities due to roughness.
Prior-austenite grain size also affects micro- Aksenova et al. showed that contributing
crack density, which decreases with decreasing factors to cracking of case-hardened gear wheels
prior-austenite grain size. However, microcrack included residual tensile stresses in the case-core
density is not affected by quench severity interface, grain growth and overheating, super-
(Ref 78, 80). Figure 35 illustrates examples of saturation of the case with carbon, excessively
microcrack formation in the carburized case of high cooling rates during quenching, too low
SAE 8620 steel (Ref 78). a quenching temperature, and insufficient resi-
The presence of microcracks can further lead dence time in the quench tank after immersion
to a surface defect called flaking, which refers to (Ref 81).
Fig. 34 Krauss fracture map illustrating conditions where susceptibility to intergranular cracking can be minimized: carburizing,
intercritical austenitizing, and Hertzian or compressive contact loading
McEvily et al., reported the use of fracto- to create the desired surface finish to improve
graphic analysis to explain cracking of a bending and contact fatigue and lubrication
carburized AISI 9310 gear at the case-core properties.
transition zone. One of the factors was the dif- Surface cracks in the carburized case may
ferential in the Poisson ratio (n) between the case occur during the grinding process, which can be
and the core that develops during a monotonic attributed to microstructural transformations
bending test when the core deforms plastically and thermal stresses producing tensile forces. If
while the case deforms elastically. McEvily the tensile forces in the case exceed the material
used as an example to illustrate this point a strength, then surface cracks will result. This is
9 mm diameter round bar with a 1 mm case. due to the difference in the specific volume of
When the bar is first loaded initially and the the transformational phases present in the case
deformation is elastic, the Poisson ratio for the structure, primarily martensite and austenite.
case and the core is equal. Upon further bending, Structural defects, such as those caused by
the core will deform plastically while the case is inclusions, will also influence the susceptibility
deforming elastically. However, because of the of the steel to cracking.
difference in the Poisson ratio between the case Generally, crack creation during grinding is
(n = 0.3) and the core (n = 0.5) at this point, influenced by thermoelastic tensile stresses that
radial tensile stresses will develop. For a tensile are created in the surface cooling zone during
strain of 1%, the radial strain in the case (relative grinding. They are dependent on thermal and
to the centerline) would be 0.003, and the radial mechanical properties of the material, maximum
strain in the core would be 0.005. To remain contact temperature, grinding feed depth, and
compatible, the difference in the strain and cooling rate.
corresponding displacements must be accom- One study conducted on the generation of
modated by the formation of a tensile stress. grinding cracks showed that microstructural
However, this radial stress will result in a state of heterogeneity, such as the presence of carbide
triaxial stress that will promote brittle behavior inclusions (particularly those with a mean dia-
by inhibiting plastic deformation. The total meter of 6 to 10 mm), which were shown to be
tensile stress may then be developed sufficient associated with large internal residual stresses,
to result in rupture at the case-core interface were a predominant cause of grinding cracks
(Ref 82). (Ref 83).
Grinding Cracks. Grinding may be used for Grinding burns arise when excessive heat is
postprocessing of components to remove growth generated during the grinding process, and this is
and distortion that may have resulted from car- characterized by surface discoloration. The term
burizing and carbonitriding. Grinding is also grinding burn refers to localized surface tem-
performed to remove such metallurgical features perature increases at least sufficient to cause
as carbide films, internal oxidation, and high- tempering of the martensitic surface, resulting in
temperature transformation products that may localized soft spots. Furthermore, since carbide
impart deleterious performance properties. In precipitation volume contraction accompanies
addition, grinding processes are commonly used tempering, the burnt areas are in tension and, if
the resulting tensile stresses are sufficient, sub-
ject 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 sur-
face. This induced defect is known as a rehar-
dening 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 tem-
pered martensite are in relative tension. Crack-
ing may occur in the area surrounding the
Fig. 35 Microcracks in the martensitic case of a coarse- rehardened material or in the interface between
grained SAE 8620 steel the two (Ref 84).
There are two possibilities to prevent the upon cooling, which then renders the part more
occurrence of thermal defects in the surface susceptible to the formation of grinding cracks
resulting from the grinding process: (Ref 1). Overcarburizing leads to the formation
of a complex carbide network that is excessively
Reduction of heating due to the rotation of a brittle, which causes greater susceptibility to
grinding wheel by reducing the speed of the grinding cracks.
grinding wheel will reduce heating and Severe grinding may lead to the development
friction. Conversely, increasing the speed of of residual tensile stress, which can be the
the grinding wheel will increase heat pro- initiation point for crack formation. Since cracks
duction and the potential for the formation of will not propagate into layers of compressed
grinding burns due to overheating. stress, it therefore may be advantageous to shot
Increasing heat abstraction due to the peen the part prior to grinding to prevent the
grinding process and reducing the contact formation of grinding cracks (Ref 86).
time between the grinding wheel and part Parrish reported that the potential for forma-
will reduce the tendency for grinding burns. tion of grinding cracks may be minimized by
Grinding cracks may exhibit characteristic (Ref 84):
short, parallel cracks (Fig. 37), or they may The thermal conductivity of the steel is an
exhibit a chicken-wire pattern and are typi- important design variable, and free carbides
cally between 0.076 and 0.13 mm (0.003 and and retained austenite have an adverse effect
0.005 in.) deep. The parallel cracks are typically on thermal conductivity.
deeper than the chicken-wire pattern. Grinding The surface carbon concentration should be
cracks form perpendicular to the grinding between 0.7 and 0.9%.
direction. As indicated previously, the potential Parts should be tempered immediately after
for grinding cracks is affected by improper heat quenching.
treatment or a metallurgical structure that is The tempering temperature should be as
prone to cracking. For example, if the surface high as possible while still achieving the
temperature exceeds the Ac3 temperature, the necessary surface hardness.
steel in this region may transform to austenite,
then upon rapid cooling, a hard martensite layer Improper Case Depth (Ref 77). In a recent
may form. This effect is called a rehardening study conducted by Bahnsen et al. on carburized
burn. Grinding cracks may be detected by a SAE 5120 test specimens to rate the relative
magnetic particle test. influence of surface carbon content, case depth,
Some carburizing steels such as chromium and carburizing temperature on distortion, it was
and chromium-manganese steels, which include
SAE 5120 and 20MnCr5, may undergo over-
carburizing with subsequent cracking of the case
Fig. 36 Microstructure of a section through a rehardening Fig. 37 Example of grinding cracks on the flank of a worm
burn. Original magnification: 500 gear. Source: Ref 85
reported that of these variables, the most domi- service conditions of the component. Typically,
nant effect was observed for case depth (Ref 87). the case depth is designed to provide the
Case depth of carburized steel is determined by necessary residual-stress distribution for the
the carburizing time and the available carbon wear requirements for the part (Ref 89). Typi-
potential at the surface. One of the most com- cally, the greater the case depth, the greater the
mon defects of carburized and carbonitrided fatigue strength (Ref 23). For carburized steel
materials is an insufficient or excessive case gears used in the automotive industry, for
depth. For example, when prolonged carburiz- example, SAE 8620, hardened case depths are
ing times are used to produce a deep case depth, generally 0.8 to 1.4 mm. Improper case depth
a high carbon potential will produce a high may be caused by establishing an unnecessarily
carbon content on the surface and the possible restrictive case thickness specification that is not
corresponding formation of excessive retained appropriate for the process or the particular
austenite or free carbides, which may lead to an furnace in use, which leads to decreased and
improper residual-stress distribution in the case- nonuniform hardness and unacceptable material
hardened part. Therefore, a high carbon poten- properties.
tial may be suitable for short carburizing times
and shallow case depths but not for prolonged
carburizing times and deep case depths. Fur-
thermore, the fatigue limit of carburized SAE Insufficient Case Hardness and
8620 steel was related to case depth and also Improper Core Hardness (Ref 77)
microstructure, distribution of retained auste-
nite, depth of internal oxidation, and near- One reason for insufficient case hardness is
surface compressive residual stresses (Ref 10, the presence of incorrect microstructure, such as
21). Bending-fatigue strength decreased with bainite. The appearance of bainite in a carbur-
the increasing case depths due to the presence of ized case in even small amounts will sig-
increasing internal oxidation and nonmartensitic nificantly decrease fatigue strength, especially
transformation products at the surface. Wear contact fatigue strength. It is an important
behavior of carburized 8620 steel is also related microstructural defect to be avoided. The pre-
to case depth (Ref 88). sence of bainite in the carburized case is
To assure optimal quality during the produc- particularly problematic because it cannot be
tion of case-carburized parts, the following are detected by hardness measurements and by the
essential: severity of the quenchant used to harden the steel
after carburizing. Figure 38(a) illustrates a bai-
The processing temperature should be
nitic case microstructure of carburized SAE
accurately controlled.
8620. Figure 38(b) shows the core structure. In
Maintain temperature uniformity throughout
the case of carbonitrided components, this is less
the load.
important because nitrogen increases harden-
Rack the parts to assure uniform gas flow
ability of steel more than carbon.
throughout the load.
Insufficient as-quenched case hardness is
Use uniform circulation of atmosphere
caused by:
throughout the load in the furnace
To properly control case depth, either use Insufficient carbon content in the entire case
shim stock or sample the parts periodically or in the superficial zones
during the carburizing cycle. Increased retained austenite content
Conduct the process at the lowest acceptable Insufficient case hardenability
temperature and time. Insufficient case depth
The components constituting a load should
possess uniform size and surface area with On the other hand, core hardness is dependent
respect to each other, using empirically on carbon content, steel alloy hardenability, and
established carburizing conditions. section size.
Avoid carburizing and carbonitriding in the Case carbon content less than 0.4 to 0.5% C is
same furnace. Use a separate furnace for easily detected by conducting a spark test.
each process, if possible. Insufficient case carbon content resulting from
the gas carburizing process occurs when the
The optimal case depth for a specific com- process is conducted with a low carbon poten-
ponent and steel alloy is based on the design and tial, inadequate furnace pressure, or cooling the
load from the diffusion temperature to the approximately 12 HRC units higher than if the
quenching temperature without a protective hardness is determined using a 150 kg load. In
atmosphere or improper atmosphere composi- practice, it is possible for the hardness to differ
tion. These problems may be rectified by addi- by 1 to 2 HRC units from published values for
tional soaking in a carburizing atmosphere with the material. Hardnesses that are i2 HRC units
the proper carbon potential. less than published values may be due to the use
Hardness reduction due to an increase in of an incorrect steel grade or an insufficient
retained austenite content may occur with a austenitizing temperature for the steel harden-
direct quenching after the carburizing process. ability and section size in use. Hardness values
This is even more critical with steels containing iHRC units higher than published values may
increased chromium or nickel, which may probe due to the use of an incorrect steel grade,
duce retained austenite levels as high as 80 to excessive case depths, or puncturing a copper
100% in the surface after quenching. As the case layer or paste on surfaces that are protected from
thickness increases, the zone containing retained carburization.
austenite may be sufficient to cause a significant Core Microstructure. The design material
reduction in hardness. To reduce the retained properties of case-hardened steels are not only
austenite content from a direct quenching pro- dependent on a martensitic case but also on the
cess, it is important to select the proper steel microstructural composition of the core. An
alloy for carburizing and to use the appropriate important design criterion is the ultimate tensile
quenchant and quenchant conditions, including strength, which is dependent on the micro-
the use of a subzero treatment, if necessary. structure of the core. For example, soft cores
Reduced hardenability usually does not occur (5770 N/mm2, or 50 ton/in.2) are suggestive
throughout the entire case but only in the 0.001 of a core with high ferrite content, as shown in
to 0.01 mm depth from the surface. Common Fig. 39(c) (Ref 19), and a hard core (41240 N/
problems leading to reduced hardenability in- mm2, or 80 ton/in.2) would be expected for a
clude internal oxidation and overcarburization predominantly martensitic structure, as shown in
of surface zones. To improve case hardenability, Fig. 39(a) (Ref 19). Intermediate structures
internal oxidation and overcarburizing should be would be bainitic structures, such as those illu-
prevented. strated in Fig. 39(b) (Ref 19). The effect of core
When the case depth is too shallow, the microstructure on ultimate tensile strength is
observed hardness is dependent on the load illustrated by Fig. 40 (Ref 90). The approximate
applied. For example, for thicknesses of 0.3 relationship between the core microstructure
to 0.4 mm, if the surface hardness is deter- and hardness for a Ni-Cr-Mo steel is illustrated
mined using a 60 kg load, the value will be in Fig. 41 (Ref 19).
Fig. 38 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
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 tempera-
ture (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 con-
taining sorbite.
Fatigue cracks occur most often in the tran-
sition 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 grind-
ing 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 chip-
ping.
Influence of Surface Carbon Content
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 pre-
Fig. 39 Microstructures obtained by cooling a 0.16%C-
sence of carbides or carbonitrides in the case,
3%Ni-Cr steel from 920 C. (a) Fast cool (920
200 C in 30 s), giving low-carbon martensitic structure of which creates an almost continuous nonetching
1590 MPa ultimate tensile strength (UTS). Original magnifica- area. Case-hardened steels with these micro-
tion: 800. (b) Intermediate cooling (920250 C in 200 s),
giving bainitic structure of 1360 MPa UTS. Original magnifica- structures characteristically exhibit increased
tion: 800. (c) Slow cool (920250 C in 104 s), giving a ferrite/ brittleness, a propensity for chipping during
pearlite structure of 740 MPa UTS. Original magnification:
800
grinding and use, and decreased fatigue strength
and pitting resistance.
Fig. 40 Approximate effect of microstructure on the ultimate tensile strength of low-carbon, low-alloy steels
Fig. 41 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.
Figure 43 illustrates the effect of carbon structure. Fig. 45(b) shows a martensitic-
content on the hardness of martensite in carbon retained austenite structure with network car-
and alloy steels. Increasing carbon content to bides. This structure is due to the chrome content
0.5% increases hardness from 20 to 65 HRC. in the steel and leads to varying carbon content
However, increases in carbon content to ap- during carburizing. Additional examples of the
proximately 1% do not produce a corresponding case microstructure of SAE 8620 steel and core
increase in hardness above 65 HRC. microstructures of SAE 1524 and 8115 steel are
Hardness and mechanical properties are shown in Fig. 46 to 48 respectively.
related not only to carbon content but also the Decarburization is the opposite of carbur-
composition of carburized steels, which is illu- izing. While carburization is performed to
strated in Fig. 44. This figure shows the corre- increase carbon content in the surface of steel,
sponding carbon gradients of the cases of three decarburization is the process by which carbon
carburized steels: chromium-molybdenum, car- is lost from the surface of steel. Decarburization
bon, and nickel (Ref 25). The data in Fig. 44 can lead to catastrophic failures of components
show that the presence of chrome and molyb- (Ref 92) and must be minimized because of
denum increases the case carbon content, and fatigue failure such as bending and contact
nickel decreases the case carbon content. The fatigue (Ref 93). Figure 49 illustrates decarbur-
case carbon content is increased due to the pre- ization of a poorly carburized SAE 8620 steel,
sence of carbide-forming elements; their struc- and Fig. 50 shows the microstructure at higher
tures in the carburized case influence the magnification.
mechanical properties of the steel. Over- Decarburization occurs at temperatures in
carburization may also lead to quench cracking excess of 700 C in the presence of gases that act
(Ref 91). as decarburization agents, which include carbon
The influence of steel composition on the dioxide (CO2), water vapor (H2O), hydrogen
microstructure of a carburized and quenched (H2), and oxygen (O2). The decarburization
20H steel is shown in Fig. 45. Figure 45(a) shows process involves the following chemical
a martensite-retained austenite microstructure reactions of molecules with carbon in the steel
with some carbides dispersed throughout the surface (CFe) until there is an equilibrium
Fig. 42 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
Fig. 43 Martensitic hardness as a function of carbon content in carbon and alloy steel. Source: Ref 80
Fig. 45 Microstructures of the carburized case structure

Fig. 44 Dependence of the carbon gradient as a function of two different samples of 20H steel that were
of case depth for three carburized steels that were carburized in the same load. Carburizing temperature:
carburized under the same conditions: 925 C and 10 h. 1, 930 C for 7 h; hardening temperature: 860 C for 0.5 h. (a)

chromium-molybdenum steel (0.56% Cr, 0.16% Mo); 2, carbon Surface. (b) Core. Etchant: 3% HNO3. Original magnification:
steel; 3, nickel steel (3.5% Ni) 500
Fig. 46 Microstructure of SAE 8620 case of a mold taken just

Fig. 49 Micrograph of poorly carburized SAE 8620 mold
below the surface. Etchant: alkaline sodium picrate
boiling (60 s), area just below the surface. Original magnifica- showing decarburization at the surface (note patches
of ferrite and pearlite). Below this zone is where the grain-
tion: 500. Courtesy of G. Vander Voort, Buehler Ltd., Lake
Bluff, IL boundary carbides are seen. Original magnification: 500 .
Courtesy of G. Vander Voort, Buehler Ltd., Lake Bluff, IL
Fig. 47 Illustration of lath martensite in the core of carbur-

ized SAE 1524 steel; water quenched from 925 C Fig. 50 This is a higher magnification of the decarburized
(1700 F). Etchant: nital. Scale = 10 mm. Courtesy of G. Vander microstructure shown in Fig. 49 of the surface of a
Voort, Buehler Ltd., Lake Bluff, IL poorly carburized SAE 8620 mold (note patches of ferrite and
pearlite). Original magnification: 1000 . Courtesy of G. Vander
Voort, Buehler Ltd., Lake Bluff, IL
established between the gaseous atmosphere and

the steel surface:
CFe 2CO
CFe +H2 O CO+H2
CFe +2H2 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
Fig. 48 Illustration of lath martensite in the core of carbur- of decarburization include a malfunctioning
ized SAE 8115 steel; water quenched from 925 C
(1700 F).
Etchant: nital. Scale = 10 mm. Courtesy of G. Vander endogas generator, such as soot accumulation
Voort, Buehler Ltd., Lake Bluff, IL hindering catalyst activity; excess moisture in
the furnace atmosphere; air contamination and Decarburization is typically classified as total
leakage (Fig. 51); heating in aged (deoxidized) or partial. Figure 53(a) illustrates a case of total
salt baths; and improper carbon potential selec- decarburization of 1018 steel and is charac-
tion (Ref 94). terized by a ferritic surface layer (Ref 94).
If a carburized steel is at a temperature above Usually, there is a gradient from total to partial
the Ac3 (approximately 900 C) in a decarbur- decarburization with increasing depth from the
izing atmosphere, the carbon potential will be surface. Partial decarburization is illustrated in
low and the surface carbon content will also be Fig. 53(b) and is often characterized by grain-
low, since the carbon in the steel surface and the boundary ferrite at the surface (Ref 94). In this
related gaseous reactions shown previously will case, nital etching will reveal a structure more
be driven to this equilibrium condition. This will gray in color than would be achieved with a
result in a decarburized layer being produced, higher-carbon martensite. The quenched surface
and the depth of the decarburized layer will of partially decarburized steel is typically bai-
depend on residence time in the furnace under nitic or martensitic. The final microstructure
these conditions. produced by a specific level of decarburization is
If the temperature of the steel is below Ac3 dependent on the steel alloy and cooling rate.
and above Ac1 (800 to 840 C), there is a dif- Decarburization is accompanied by surface
ferent decarburization condition. In this case, hardness reduction. However, partial or shallow
the carbon content rapidly decreases from A decarburization may not necessarily be detected
to B, as shown in Fig. 52 (Ref 94). Further by macrohardness determination. If decarbur-
decreases in carbon content will result in a ization is suspected, microhardness determina-
material of carbon content C in equilibrium tions, in view of their sensitivity to the presence
with material of carbon content B. Therefore, of such microstructures, should be used.
further loss of carbon by decarburization must Decarburization can exhibit dramatic effects
result in the formation of ferrite containing on the residual stress of a component, as illu-
carbon content C. strated by Fig. 54 (Ref 94). In this example,
If the atmosphere carbon content is controlled the surface carbon content of a carburized
to carbon potential D, then ferrite cannot form. 3.5Ni-1.5Cr steel would be approximately 1%,
Instead, a gradient is formed between carbon and the surface residual stresses would be
contents A and D. compressive at 4392 MPa. After decarburiza-
tion reduces the carbon content to 0.64%,
the surface residual stresses were found to be
nearly 0. Finally, when decarburization re-
duced the surface carbon content to 0.35%,
Fig. 51 Effect of air ingression into the carburization atmo-

sphere (N2/4% natural gas) on the decarburization of Fig. 52 Iron-carbon equilibrium diagram to explain dec-
SAE 8620 after 2 h at 850 C. Source: Ref 94 arburization
the surface residual stresses were tensile at Influence of Grain Size

226 MPa.
Various reports have shown that decarbur- Grain size is one of the most important and
ization can result in large decreases in bending characteristic features of steel. Grain size in-
and contact fatigue strength (Ref 94). Since fluences mechanical and plastic properties,
wear-resistance properties are typically depen- especially impact resistance and also steel har-
dent on achieving optimal hardness, reduction of denability. Grain size is characterized by the size
carbon content and thus surface hardness will of the austenite grain, and it is dependent on
have a correspondingly adverse effect. 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 tem-
perature 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 composi-
tion, 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. 53 Micrographs illustrating total and partial decarbur-

ization. (a) Total decarburization of 1018 steel Fig. 54 Effect of decarburization on the residual stresses of
caused by a furnace air leak. Etchant: 1% nital. Original magni- carburized and hardened 3.5Ni-1.5Cr steel. The
fication: 500. (b) Illustration of partial decarburization. Original carbon content at 0.002 mm was approximately 1% for curve 1,
magnification: 190 0.64% for curve 2, and 0.35% for curve 3.
are typically deoxidized by silicon and manga- formation of coarse ferrite/carbide (pearlitic)
nese prior to aluminum addition. This will grain structures on cooling.
inhibit the undesirable formation of AlN or Grain size is possible by control of the steel
Al2O3 particles, which provide nucleation sites composition during the steelmaking process.
for coarse-grained carbides that lead to the 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 pro-
duce fine grain size, and quenching from a
higher temperature would yield a coarser grain
size.
The austenite grain size at the onset of trans-
formation 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 fac-
tors affecting grain size, the primary factor is
furnace treatment. Although coarse-grained
steels exhibit better machinability, they gen-
erally possess lower toughness and ductility and
exhibit a greater tendency for distortion and
cracking than fine-grained steels. Coarse-
grained steel also exhibits a more limited range
of thermal treatment temperatures, and they
possess better hardenability with higher as-
quenched hardness. Additionally, coarse-
grained 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 con-
tinuous network carbide structure that is difficult
to remove, and within the grain structure there
are typically acicular carbides. Such micro-
structures 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 struc-
ture.
An example of grain size formation during
austenitization of steel 40 is shown in Fig. 56.
The irregular austenitic grain boundaries result
Fig. 55 Comparison of nominal ASTM No. 6 to 9 grain
from the short heating time and also are related
sizes. Etchant: nital. Original magnification: 100 .
Source: Ref 96 to a prior normalizing and annealing process.
This thermal history produces the variable grain temperature, as well as the steel alloy chemistry.
size shown. The total depth of internal oxidation can be
Isogai et al. reported that grain sizes of calculated from (Ref 100):
approximately 8 were required for carburized
transmission gear steel to achieve the required Xi 2 =(2D0 C0 )=(n CM ) t=kp t
high fatigue strength, pitting strength, and
impact strength in severe-use environments (Ref Xi =kp t0:5
97). To assure a fine grain size, after the
carburizing process, the chromium-containing where Xi is the depth of oxygen penetration, D0
steel, such as SCr420, was reheated to 820 to is the diffusion coefficient of oxygen in the alloy,
870 C for 20 to 60 min in an atmosphere con- C0 is the oxygen concentration at the steel alloy
taining the carburizing gas (carbon potential of surface, CM is the concentration of the base
0.75%), quenched in oil until the steel was metal in the alloy (e.g., silicon, chromium,
120 C, and then tempered at 120 to 200 C. manganese, titanium, vanadium), and n is the
Treatment of the carburized steel in this way stoichiometric factor.
permits substantial reductions in grain size and Internal oxidation appears on the polished
corresponding improvements in fatigue strength metallographic specimen in the form of very
and impact strength. Similar improvements were small inclusions concentrated in an austenite
achievable with carbonitrided steel (Ref 97). grain or within the grain boundary. The prob-
ability of oxidation within the grain increases as
the grain size decreases (Ref 101). Figure 57
Internal Oxidation (Ref 77) shows grain-boundary oxidation of carburized
20MnCr5 steel. Figure 58 also shows grain-
The depth of internal oxidation that originates boundary oxidation, but it is accompanied by
the carburizing atmosphere may vary from 1 to nonmartensitic transformation products. Inter-
30 mm (Ref 93, 98, 99). Internal oxidation con- nal oxidation occurs in two zones: an inner zone
sists of a continuous oxide layer on the surface, and an outer zone. Oxides of chromium-man-
on the order of 0.01 mm, due to oxygen reaction ganese are typically formed in the outer zone,
with the carburized steel surface. In this region, both within the grain and in the grain boundaries.
the oxygen content can be 10 to 20 times that of In the inner zone, silicon-rich oxides are typi-
the core. During the gas carburizing process, cally formed exclusively within the grain
oxides will not only form on the surface but also boundaries (Ref 102). Figure 59 also illustrates
penetrate into the steel surface. Since internal intergranular oxidation of a gas-carburized steel
oxide formation is a diffusion process, the depth (Ref 79).
and extent of oxide penetration is dependent on Typically, the greater the case depth, the
the square root of the total carburizing time and greater the degree of oxide formation at the
Fig. 56 Prior-austenite grains formed in hardened steel 40, Fig. 57 Illustration of grain-boundary oxidation of carbur-
which were due to abnormal growth during the ized 20MnCr5 to a depth of 30 mm. Unetched. Ori-
austenitizing process, Etched: S. Bechet and L. Beaujurda. Ori- ginal magnification: 200. Courtesy of Fluidtherm Technology P.
ginal magnification: 500 Ltd., Ambattur, India
surface and the greater the depth of internal lower surface hardness and poorer abrasive wear
oxidation. The thickness of the internal oxida- resistance.
tion zone typically is approximately 5% of the Typically, the oxides in the outer surface
carburized layer thickness and, on occasion, region are globular in form, while intergranular
may be as high as 10%. The critical cooling rate oxides were formed further from the surface. In
is greater in the internal oxidation zone. As a one analyis of a carburized chromium-manga-
result, an otherwise normal hardening produces nese steel, larger globular oxides were formed in
a greater amount of bainite, which will lead to 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 sput-
ter 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-dis-
persive 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
Fig. 58 Illustration of grain-boundary oxidation with non- and morphology. Chromium and manganese
martensitic transformation products to a depth of
approximately 30 mm. Etchant: nital. Original magnification:
globular oxides formed nearer the surface after a
200. Courtesy of Fluidtherm Technology P. Ltd., Ambattur, total heat carburizing cycle of 5.8 h (5+0.8 h).
India An agglomerated internal oxide of chromium
Fig. 59 Intergranular oxidation of the surface along prior grain boundaries in a carburized steel. Original magnification: 1000 .
Source: Ref 78
and manganese oxides in the core surrounded by where the elemental compositions are given in
silicon oxides formed after heating for 16.6 h weight percent. Figure 60 provides a correlation
(5+11.6 h) (Ref 103). of the TOP and depth of internal oxidation (Ref
Internal oxide formation within grain bound- 106).
aries provides sites for crack initiation (Ref 10). In addition, Kozlovskii et al. reported that
This was shown in a study by Laue et al. most steels, when subjected to the gas carbur-
(Ref 104), who evaluated the fatigue behavior of izing process, will undergo internal oxidation
case-hardened SAE 5115 steel with internal with a corresponding surface formation of
oxidation of the case structure. Fatigue studies troostite to a depth of 0.01 to 0.03 mm. If
were conducted on this steel, and it was shown troostite is formed at a depth greater than
that fatigue crack initiation occurred along the 0.014 mm, there is a substantial decrease in
oxidized grain boundaries of the steel (Ref 104). fatigue strength. However, the potential for
Internal oxides form as a result of oxygen internal oxide formation can be reduced by the
diffusion into the surface, with subsequent for- addition of 5 to 10% of ammonia to the furnace
mation of metal oxides at carburizing tempera- for 10 min before the carburizing process is
tures. The formation of these oxides is enhanced completed or by using steels containing 0.5%
by the presence of metals, chromium and man- Mo and not more than 0.5% Cr (Ref 106).
ganese and Cr-Mn-Ti, which possess a greater Lohrmann showed that the form and type of
affinity for oxygen than iron. The susceptibility internal oxide obtained was dependent on the
for internal oxidation increases with increasing alloy composition of the steel. For example,
concentration of these oxide-forming elements. depending on the steel alloy, spot-, liner-, or
Lohrmann et al., referring to earlier work by lattice-type internal oxides could be formed
Kozlovskii and co-workers, reported that the (Ref 105).
depth of internal oxidation was dependent on the Kehr and Seese examined the effect of internal
total oxidation potential (TOP) of the steel alloy oxidation during carburizing of investment-cast
(Ref 105, 106): ingot-iron test specimen steels containing var-
ious amounts of chromium (0.20, 0.45, 0.91,
TOP=4:87 Si+3:7 Mn+1:47 Cr 3:24 Ni 1.85, and 4.74% Cr) (Ref 107). It was shown that
1:82 Mo steels containing approximately 0.50% Cr, for
Fig. 60 Correlation of calculated total oxidation potential (TOP) and average depth of internal oxidation
example, SAE 8620, were more susceptible to There are three types of carbides to be discussed
internal oxidation than either plain carbon steel here: globular (or massive) carbides, network
or steels with greater amounts of chromium, such carbides, and surface-film or flake carbides.
as AISI 5120 (0.70 to 0.90% Cr). Figure 61 Carbides in steel are hard and brittle ceramic-
shows that the depth of oxide formation increases like interstitials with a high compressive
as the chromium content increases up to 0.45% strength but low tensile strength (approximately
(Ref 107). 35 MPa, or 5000 psi).
The presence of internal oxidation on car- Carbides in steel basically form when carbon
burized steels greatly reduces bending and levels exceed the solubility limits of carbon in
contact fatigue strength and wear resistance the iron crystal structure. The allotropic nature
(Ref 93). However, surface oxides that are of iron also has different phase structures (i.e.,
formed may be removed by grinding or shot crystal) with different solubility limits for car-
peening. The potential for internal oxidation can bon. For example, the maximum solubility of
be reduced by heating steel to the carburizing carbon in the body-centered cubic (bcc) struc-
temperature under a nitrogen/hydrogen gas ture of ferrite is approximately 0.025 wt% at
mixture. During carburizing, these gases are 723 C on the iron-carbon phase diagram (see
replaced by a carburizing atmosphere whose Appendix 9). For the face-centered cubic phase
oxygen activity is less than that required for the of austenite (c), the maximum solubility limit of
formation of manganese II oxide or chromium carbon in c is approximately 2.06 wt% C at
III oxide, and, in some cases, in the presence of 1147 C. At still higher temperatures, another
ammonia (Ref 100). type of bcc solid phase is d-ferrite. The max-
Finally, internal oxidation (and decarburiza- imum solubility of carbon in d-ferrite is
tion) is known to lead to variations in the surface 0.09 wt% C at 1493 C. A peritectic also occurs
compressive stresses in a carburized component at 0.16% C at 1493 C.
(Ref 108). There is one report of the presence of The iron-carbon system has eutectic trans-
internal oxidation leading to undesirable surface formation at 1147 C during soldification, with
tensile stresses that then led to subsequent steel carbon levels of 2.06 to 6.67 wt% C. The
cracking of a carburized idler gear when used in eutectic carbon concentration is 4.3%. Solid-
a diesel engine gearbox. state transformations in steel include the well-
known eutectoid transformation at 733 C, with
a carbon concentration of 0.83%. At 733 C,
austenite transforms to pearlite. Pearlite is a
Carbides and Carbide Structure eutectoid mixture containing 0.83% C and is
characterized by a fine ferrite-cementite struc-
Carbides formed during carburization are ture that forms upon austenite decomposition
treated as undesirable products to be avoided. during slow cooling.
The upper critical temperature (A3) is the
temperature below which ferrite starts to form
due to ejection from austenite in the hypoeu-
tectoid alloys. The lower critical temperature
(A1) is the temperature where the austenite-to-
pearlite eutectoid transformation occurs. Aus-
tenite does not exist below this temperature. A2
is the temperature below which ferrite is ferro-
magnetic.
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 in-
creasing 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
Fig. 61 Effect of chromium content of steel on the depth of less than 0.83% C, iron and carbon will combine
oxidation to form Fe3C until no carbon remains. The
cementite formed in this manner will combine lamellar structure of cementite in ferrite, shown
with the required amount of ferrite to form in Fig. 62(b). Finally, cementite exhibits a
pearlite, and any remaining ferrite will be in the granular or grainy appearance, as shown in
structure as free ferrite (proeutectoid ferrite). Fig. 62(c).
Pearlite will form if the carbon content in the Cementite forms during soldification from a
austenite is 40.83%, and excess carbon will liquid or during solid-state transformations.
form cementite. Excess cementite (proeutectoid When cementite originates by crystallization
cementite) will deposit in the grain boundaries. from a liquid melt, it is referred to as pri-
Because cementite (Fe3C) contains a specific mary cementite (Fe3CI). Secondary cementite
amount of carbon and iron, pearlite also contains (Fe3CII) is formed from austenite by hyper-
a specific amount of cementite and ferrite. eutectoid alloys (carbon40.8%). Tertiary cem-
As a phase in steel, the chemical composition entite (Fe3CIII) is formed at temperatures below
of cementite will contain carbides of other car- 723 C by precipitation in the grain boundaries
bide-forming elements, such as chromium and (which become richer in carbon with the
manganese Alloying elements of chromium, decreasing carbon content in a-iron).
manganese, nickel, and other elements are, of Globular Carbides. Slowly heating a steel
course, commonly used in alloy steels for to the carburizing temperature in the presence of
property improvement. They also impact the a carburizing atmosphere through the ferrite-to-
properties of ferrite and cementite, because they austenite temperature transformation region will
partition differently in the phases. For example, lead to unconnected globular carbide formation
chromium and manganese partition in cementite either within the ferrite grains or at the former
instead of ferrite. However, nickel and silicon ferrite grain boundaries, as shown in Fig. 63.
tend to favor partitioning in ferrite. Chromium, This process is favored by high carbon potentials
manganese, molybdenum, and titanium are thus and also by reduction of normal heating rates
cementite stabilizers in steels, while nickel and typically involved during carburizing, by
silicon are ferrite stabilizers. Interestingly, while excessive furnace loading, or by a furnace mal-
chromium, manganese, molybdenum, and van- function. Globular carbide formation may also
adium show no negative effect on cementite be enhanced by austenitic nuclei or by localized
formation, titanium, nickel, and silicon exhibit a concentrations of carbide-forming elements.
negative effect on cementite formation (Ref The problem of globular carbides may also
109). coexist with other problems, such as retaining
Cementite develops different morphologies austenite or quench cracking.
and distributions depending on the process of When carburizing steels (0.15 to 0.25% C),
cementite formation. Figure 62 illustrates three which are commonly ferritic with localized
microstructural forms of cementite: lamellar, areas of spheroidal carbides due to prior nor-
mixed, and granular (Ref 110). Cementite may malizing and subcritical annealing, are heated
also be classified as reticular, acicular, or gran- through the Ac1 temperature, the high-carbon
ular (Ref 110). Reticular cementite, also known regions begin to transform to austenite, resulting
as shell-type cementite, possesses a crack- in the formation of localized regions of carbon
sensitive network or platelet structure. Acicular and carbide-forming elements in addition to
cementite, or needlelike structure, refers to a undissolved carbides in the presence of the
Fig. 62 Cementite structures of CT60 steel with (a) lamellar, (b) mixed, and (c) granular cementite. Original magnification: 500
carbon-rich carburizing atmosphere. As the result in reduced fatigue strength. Interestingly,

temperature increases further to Ac3, additional it is possible to obtain desirable surface fatigue
ferrite will transform to austenite by combining strength by vacuum or plasma carburizing a
with carbon from the furnace atmosphere, pre- chromium-containing steel such as SCr420H
ferentially over carbon transfer from the spher- or a chromium-molybdenum steel such as
oidal carbides already present until the austenitic SCM420H (Ref 114). This process involves the
transformation is complete, at which point there formation of a M23C6-type globular carbide that
will be an equilibrium between the austenite and is approximately 1 mm in diameter. These
carbides (Ref 112). microstructures are reported to exhibit excellent
Those steels containing alloying elements surface fatigue strength and rolling fatigue
such as chromium and manganese that will strength under high bearing loads (43 GPa) at
reduce the eutectoid carbon content and Ac1 relatively high temperature (100 to 300 C),
temperature are more likely to develop globular which are unachievable in the presence of larger
carbides. These elements will increase the (i3 mm) M3C carbides.
solubility limit of carbon (Acm), thus shifting the Reheating a carburized steel to a temperature
equilibrium diagram to the left, as shown in below Acm causes spheroidized carbide particles
Fig. 64 (Ref 112). to form. (Both austenite and cementite are stable
It should be noted that although the terms at this temperature.) Since these carbides bind
spheroidal carbides and globular carbides are some of the carbon in the case, there will be an
often used interchangeably, these carbides may, increase in the Ms temperature. In addition,
in fact, possess a round, angular, or even nee- grain-boundary migration is reduced (Ref 115).
dlelike appearance. Also, although the structures Carbides are typically very hard. For exam-
are typically designated as M3C, their actual ple, the microhardness of globular carbides in a
ratio of the element/carbon composing the car- carburized case of a plain carbon steel has been
bide being observed is not only dependent on the measured to be 41000 HV, and carbides in a 2%
time and temperature of heat treatment but pri- Ni-Cr steel have been measured to be approxi-
marily on the elemental availability in the steel mately 800 HV (Ref 112). Therefore, the pre-
during formation. sence of free globular carbides is often assumed
There is a three-step process of high-density to improve component wear, abrasion, and
carburizing that is conducted to increase case scuffing resistance. However, in a study reported
hardness by aggressive precipitation of cemen- by Parrish on the effect of globular carbides on
tite (Fe3C) to improve surface fatigue strength contact fatigue with carburized 2% Cr-Mn steel,
(Ref 113). Generally, grain-boundary carbide it was shown that the presence of massive
precipitation will form network carbide struc-
tures that are susceptible to quench cracking and
Fig. 63 Bad globular carbide formation in the case of a car-

burized 9310 steel. Etchant: boil in alkaline sodium Fig. 64 The saturation surface carbon content (Acm) of var-
picrate solution (45 s). Scale = 10 mm. Courtesy of G. Vander ious carburizing steels as a function of carburizing
Voort, Buehler Ltd., Lake Bluff, IL temperature as related to the iron-carbon equilibrium diagram
carbides was detrimental, while a beneficial eutectoid temperature t3, carbon precipitation as
effect was obtained if the carbides were finer and Fe3C will have stopped, and the austenite carbon
better distributed, as shown in Fig. 65 (Ref 112). content C3 will transform to a pearlitic eutectoid
The presence of globular carbides may also microstructure. The relative proportion of Fe3C
lead to grinding cracks, which are also related to to austenite can be determined from Fig. 67
surface residual stresses. using the lever rule (CxC1/C1B).
Network Carbides. Globular carbides are Although austenite is supersaturated with
most typically formed when the carbon content respect to carbon during the carburizing process,
is less than the Acm. However, under some and carbide precipitation at austenitic grain
conditions, it is possible for globular carbides to boundaries will occur during cooling, if the steel
still be formed when the carbon content is is quenched from the carburizing temperature,
greater than the solubility limit (in excess of the the excess carbon can be retained by the result-
eutectoid composition) and austenite is super- ing as-quenched martensitic/retained-austenite
saturated with respect to carbon. Most typically, microstructure. Typically, during carburizing,
these are conditions where carbon will pre- the load is cooled in the furnace from the car-
cipitate in the grain boundaries as cementite burizing temperature to the temperature from
(Fe3C) during slow cooling from the carburizing which the steel will be hardened. Since the Fe3C
temperature, leading to the formation of network migrates to the grain boundaries during slow
carbides. Figure 66 illustrates network carbide cooling, faster cooling by quenching will allow
structure observed in a broken carburized AISI retention of the carbon in the martensitic/auste-
P5 steel tool. nitic structure. Thus, cooling rate can be used to
The equilibrium diagram shown in Fig. 67 control the amount of network carbide forma-
illustrates the conditions for the formation of tion. Alternatively, carburized steel can be
network carbides when excess carbon is pre- cooled to ambient temperature and reheated to
cipitated from austenite as Fe3C (Ref 112). 820 to 860 C and quenched. However, it has
When the carbon steel supersaturated with car- been shown that traces of the network carbides
bon content of C1 is cooled from t0 to t1, Fe3C remained in the microstructure even after heat-
will begin to precipitate. As cooling continues ing to 900 C (Ref 112). Parrish also has
to t2, additional carbon will have precipitated reported that if steel containing excess carbon is
as Fe3C until the carbon content is C2. At the subcritically annealed prior to quenching, a fine
Fig. 65 Contact fatigue strength of carburized 25Kh2GHTA steel (tempered at 180 to 200 C)
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 net-
work carbides formed when quenching gas-
carburized 11 by 11 by 56 mm 07Kh16N6 steel
test specimens. The test specimens were car-
burized, 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). Dur-
ing 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 redissolu-
tion during reheating, which will lead to in-
creased brittleness and grinding cracks (Ref 1).
The presence of network carbides in a Kaplan
turbine blade constructed from improperly car-
burized 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 var-
ious studies and concluded that continuous net-
work 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).
Fig. 66 Micrographs of a broken carburized P5 tool steel die.
The three most common failure modes of
Note the layer of cementite along the surface and the
heavy grain-boundary network in (a). The case was 59.5 HRC, carburized steels are ductile fracture, cleavage,
and the core was 22 GRC. (a) The case is shown at an original and intergranular fracture (Ref 23). Ductile
magnification of 100 (100 mm bar). (b) The case at an original
magnification of 1000 (10 mm bar). (c) The core at an origi- fracture is caused by nucleation growth and
nal magnification of 400 (25 mm bar) in Nomarski differential coalescence of voids that are initiated at inclu-
interference contrast (note the old and new ferrite, as in dual-
phase steel). The austenitizing temperature for the case is
sion sites and second-phase particles. Cleavage
approximately 1475 F, which is in the two-phase field for the fracture occurs by separation at crystallographic
core. New ferrite formed due to limited hardenability in the planes by a transgranular pathway. Intergranular
quench. Courtesy of G. Vander Voort, Buehler Ltd., Lake Bluff, IL
fracture, such as that involving cementite
Fig. 67 The use of the iron-carbon equilibrium diagram to illustrate network carbide formation
Table 11 Effect of initial cool-down period on Grain-boundary cavitation and high-

network carbide formation of carburized temperature cracking
07Kh16N6 steel
Network carbide Film or Flake Carbides. Surface-film or
Cool-down temperature(a), C severity rating(b) flake carbides are composed of a continuous or
1020 (initial temperature) 1 discontinuous carbide film with typically little or
950 1
900 12 no penetration into the case structure, which is
850 13 caused by cooling of the carburized steel in the
800 13 furnace with high carbon potential. Parrish
700 34
600 55 summarized various previously published re-
(a) The cooldown occurs from 1020 to the temperature shown at a rate of ports that stated such carbide films contain
0.036 C/s. (b) 1, very fine carbide network; 34, failure rating; 5, largest 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
deposition at grain boundaries, involves crack-
grain boundaries (Ref 112). Koistinen showed
ing on grain boundaries and is due to (Ref 23):
that such films have high tensile surface
Precipitation of a brittle phase (such as net- (0.025 mm depth) residual stress, as shown in
work carbides) on the grain boundary Fig. 69 (Ref 44). Until now, no further infor-
Hydrogen embrittlement mation relating to the presence of these carbide
Environmental-assisted cracking types on carburized steel properties has been
Intergranular corrosion reported.
Noncarbide Inclusions (Ref 118). Metallic element inclusions (impu-

rities), although typically in trace quantities,
Two of the primary causes of fatigue failure may be traced to the scrap used in the steel-
are inclusions and surface defects. Inclusions making process. These elements cause inter-
may be metallic impurities or metallic oxides granular 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.
formation, detrimental precipitate formation, strength and negatively influence ductility and
and are often observed as slivers in the final toughness (Ref 119). Although harder inclusions
product. such as metallic inclusions are more harmful
Metallic oxide inclusions vary in morphology than softer inclusions, nonmetallic inclusions
and composition. Some sources of these oxides are still generally harmful. In addition, the de-
include: leterious effects of inclusions increase with size.
In one study conducted by Bomas and
Deoxidation products, such as alumina
Schleicher on the effect of inclusions on the
inclusions, that are formed by the reaction of
dissolved oxygen and the added deoxidant, fatigue strength of carburized 16MnCrS5 (SAE
5115) steel, it was found that subsurface fatigue
such as aluminum. Alumina inclusions are
crack initiation was initiated by nonmetallic
dendritic when formed in the presence of
inclusions up to depths of 1.4 mm (Ref 120).
high oxygen concentration.
Similar results were obtained for a study of
Reoxidation products are generated when
MnS-induced bending fatigue failure of car-
the aluminum remaining in the liquid steel is
burized EN39B steel (Ref 121).
oxidized by FeO, MnO, or SiO2 and other
Even in clean steels, oxide and sulfide inclu-
oxides in the slag or refractory materials or
sions exist. From a study of the effect of defects
by exposure to the atmosphere.
such as inclusions on the fatigue bending
Slag entrapment occurs in metallurgical
strength of carburized SCM20 steel, a model of
fluxes entrained in the steel. This occurs
fatigue crack initiation during fatigue bending,
during transfer between steelmaking vessels.
such as would occur in a gear tooth, was devel-
These inclusions are typically spherical.
oped (Ref 122). This model is illustrated sche-
Exogenous inclusions from other sources
matically in Fig. 70. Figure 70(a) illustrates the
include dirt, broken refractory brickwork,
chromium and manganese oxides and grain-
and ceramic lining materials. They act as
boundary segregation of silicon oxides present
sites for heterogeneous nucleation of alu-
in a decarburized surface layer. When the sur-
mina and may include a central particle.
face is loaded, the grain boundary or oxides act
Chemical reactions may produce oxides
as a stress raiser for microcrack formation, as
from inclusion modification when calcium
shown in Fig. 70(b). Although most of the cracks
treatment is imperfectly performed. Inclu-
remain in the decarburized layer, the most cri-
sions containing CaO may also originate
tical crack penetrates deeper into the steel, as
from entrained slag.
shown in Fig. 70(c). The crack that initiates fati-
All steels contain various noncarbide in- gue failure was thought to be due to a combi-
clusions. It is well known that inclusions, in nation of the microcracks shown in Fig. 70(b). In
addition to surface defects and inhomogenities this case, the decarburized layer in combination
such as retained austenite, nonmetallic inclu- with the presence of inclusions was projected to
sions, and inhomogeneities, can reduce fatigue be the source of fatigue failure (Ref 122).
Fig. 70 Model of fatigue crack initiation due to the presence of inclusions in a nonmartensitic (decarburized) steel layer. Source:
Ref 122
Micropitting temperature, which is the instantaneous tem-

perature rise due to localized friction heat at the
Micropitting fatigue usually occurs on heav- point where the gear teeth mesh. The value of Tb
ily-loaded surface-hardened components and is controlled by gear geometry design, and the
is characterized by a frosted or gray-stained value of Tf is controlled by the lubricant visc-
appearance under thin-film lubrication condi- osity and surface roughness.
tions (Ref 87, 123). Numerous small cracks in To minimize micropitting:
the surface may exceed the depth of the micro-
pits (Ref 124). Under magnification, small pits Use higher operational speeds and smooth
approximately 10 mm deep will be observed. material surfaces.
The surface will appear etched with a pattern Use the recommended amount of clean, dry
that sometimes follows the slightly higher ridges lubricant with the highest viscosity permis-
left by cutter marks or other surface irregula- sible.
rities on the finished component. Micropitting is Reduce the lubricating oil temperature and
influenced by high surface loads, frictional heat surface loading.
generation due to poor lubrication caused by Use the optimal amount of case carbon
insufficient film thicknesses in the partial elas- content in carburized gear materials.
tohydrodynamic lubrication regime, excessive
retained austenite, tangential speed, and lubri-
cant additives and oxidation (Ref 123). Contact Fatigue Pitting (Macropitting)
Micropitting is strongly influenced by the
relative lubricating oil film thickness and can Pitting failures occur when fatigue cracks are
be quantitatively related to both the surface initiated on the tooth surface or just below the
condition (roughness) and the thickness of surface. Usually, fatigue pits are the result of
the lubricating film by the lambda (L) value surface cracks caused by metal-to-metal contact
(Ref 124, 125): of asperities or defects due to insufficient lubri-
cant film thickness. They are dependent on the
h Hertzian contact surface stress and the number
L=h=s= q
(s1 +s22 )
2 of stress cycles (Ref 124). Surface asperities of
the harder material of a wear contact will lead to
where s1 and s2 are the root mean square surface damage of the softer surface, sometimes by a
roughness of the two opposing wear surfaces, work-hardening mechanism, leading to the
and h is the lubricating oil film thickness. When creation of microcracks that then become fatigue
Li3, there is full film lubrication with no pits as the wear process continues (Ref 124).
asperity contact. When 0.85L53, there is Pitting damage is commonly encountered with
partial elastohydrodynamic lubrication. When rolling element bearings, gears, and machine
L50.8, there is a boundary lubrication condi- components subject to cyclic rolling-sliding
tion. When L51, micropitting will occur, and motion under a load. Initially, fatigue pits may
once micropitting occurs, pitting fatigue (mac- occur in localized areas and may range in size
ropitting) will be accelerated (Ref 124). How- from 0.38 to 0.76 mm (0.015 to 0.030 in.) in
ever, if macropitting does occur, it is often diameter (Ref 85).
characterized by an arrowhead or fan shape (Ref Vinokur et al. examined the effect of case
123). There have been reports of substantial carbon content of carburized 18KhGNMFL steel
improvements in fatigue lives, such as with (1.3% Mn, 1.3% Cr, 0.8% Ni, 0.25% Mo, 0.1%
carburized 9310 steel, with corresponding reduc- V) on the contact endurance of fatigue tests
tions in surface roughness (Ref 125). conducted with a wear test with an applied stress
For gears, there is a critical temperature where of 3500 MPa using a lubricating oil. Contact
pitting fatigue and scuffing are likely to occur. fatigue was the average of 10 tests to determine
This is called the critical scuffing temperature the number of cycles until pitting was observed.
(Tc), which is calculated from: The case depth was approximately 1.8 mm. The
carburized steel was hardened from the inter-
Tc =Tb +Tf critical range and tempered at 170 C. The
results of this study are summarized in Fig. 71
where Tb is the equilibrium temperature of (Ref 126). The contact endurance increases with
the gears before meshing, and Tf is the flash carbon content up to approximately 0.9% C and
then decreases. The optimal hardening tem- To prevent pitting fatigue, either the surface
perature was at approximately 850 C, which is loading must be decreased to a level below the
in the intercritical temperature range, just below endurance limit of the material or the hardness
the upper critical temperature for the steel. must be increased to increase the endurance
Although fatigue pits are usually initiated at limit (Ref 28). Pitting may also be reduced by
the surface, subsurface initiation is relatively instituting a break-in period at reduced loads and
common in case-hardened rolling element speeds to improve gear tooth contact (Ref 85,
bearings with serious inclusion problems. In 124].
these situations, failures usually do not follow Other potential causes of fatigue pitting
the case-core interface (Ref 125, 127). The include hydrogen embrittlement due to water
potential for micropitting, pitting, and spalling contamination of the lubricant and particle
phenomena may be assessed from the lambda contamination of the lubricant, which act as
(L) value (see the section Micropitting in this surface stress-concentration points that lead to
chapter); however, even relatively smooth sur- pitting failure.
faces and lubricating film thickness, such as
those used for high-speed gears, may exhibit
pitting failures due to the presence of subsurface Case Crushing
cracks. The subsurface cracks may be caused by
the presence of inclusions that act as stress If the case depth is too deep, case-core
concentrators, causing the crack to propagate separation may occur due to the tips of the gear
parallel to the surface and subsequently break teeth becoming too brittle and possibly break-
through the surface. ing, if the case depth is too thin, the strength
Spalling failure occurs when the wear process of the gear teeth will be reduced, causing pre-
causes several pits to join together. These larger mature pitting, or it may lead to a condition
craters are usually caused by more severe over- called case crushing (Ref 131). Case crushing
loading conditions. As the number of stress occurs in heavily loaded case-hardened com-
cycles increases, the pitting process will con- ponents such as gears. Case crushing occurs by a
tinue in an effort to relieve stresses. The rule is subsurface fatigue process where the high-cycle
that spalling cracks initiate where the ratio of contact stress exceeds the endurance limit. This
shear stress to Vickers hardness is maximum will occur when subsurface stresses exceed the
(Ref 124). However, this relationship is not strength of the core. Case crushing failures may
correct when there are excessive amounts of have a similar appearance to pitting, although it
retained austenite. Generally, it is assumed that often occurs as longitudinal cracks on the sur-
spalling will occur when L41. face of only one or two gear teeth, where sec-
tions 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):
Fig. 71 Effect of case carbon content and hardening tem- Case depth to 50 HRC=1:2 107 (W)=F
perature on the contact endurance limit of carbur-
ized 18KhGNMFL steel (1.3% Mn, 1.3% Cr, 0.8% Ni, 0.25% Mo, where W is the force in pounds pressing the
0.1% V). Source: Ref 126 surfaces together, and F is the length of the line
contact in inches. The strength of the core can be created, the localized chemical surrounding is
determined from hardness. Since case crushing is much more aggressive than the surrounding area
promoted by shear, the shear strength of the core of the uncorroded material.
must be determined and can be estimated from Initiation of the pitting process is dependent on
Fig. 72 (Ref 128). Generally, the subsurface temperature and on the steel surface, including
stress/strength ratio should not be greater than the presence of sediment. In some applications
0.55. where pitting corrosion is more prevalent, such
Eryu et al. have studied case-crushing fracture as steel in concrete structures, the pitting corro-
mechanisms of carburized 20CrMnTi by sion process is characterized by the temperature
scanning electron microscopy (Ref 130). They or narrow range of temperatures above which
showed that the surface features of the primary pitting will nucleate. The creation of corrosion
cracks exhibited scaly features and that there sediments will lead to a temperature decrease.
were two features of the fracture surfaces of the Pitting corrosion will only occur above this cri-
branching cracks. The fatigue steps and dimples tical temperature. Therefore, to increase the
were analogous to materials of higher strength, lifetime of steel used in reinforced concrete by
and the fragmentation pattern was analogous to reducing the rate of pitting corrosion, frequent
brittle material. In addition, spherical particles sediment removal (cleaning) is recommended.
were observed that were composed of a-iron, Most often, pitting corrosion is initiated by the
which were proposed to be caused by the presence of chloride salts, and the rate of cor-
movement of the faces of the primary crack due rosive attack is steel alloy dependent. The cri-
to shearing and compressive stresses. tical concentration of chlorides for different
steel alloys cannot be defined, because corro-
sivity is dependent on other chemicals that may
Pitting Corrosion be present, which will affect the rate of corrosion
attack. However, since pitting corrosion is
Pitting corrosion is a localized penetrating typically relatively fast, it should be prevented.
corrosion attack of typically corrosion-resistant Resistance of steel to pitting corrosion is
steel resulting in a mass loss of the steel (Ref dependent on the alloy composition (chromium,
131). Pitting corrosion is related to localized molybdenum, tungsten, nitrogen). Relative cor-
discontinuities of a passive layer caused by rosion resistance of steel alloys may be empiri-
mechanical imperfections, inclusions, surface- cally quantified by:
localized chemical attack of the passive layer by
salts such as chlorides, or by overaggressive Relative corrosion resistance=%Cr
lubricant additives. After the corrosion pit is +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-pro-
ducts 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
Fig. 72 Shear strength of carburizing steels as a function of salt baths (Ref 132). Heating of steel with scale
hardness on a surface not only accelerates decarburizing,
but pitting corrosion is also accelerated. The Partial Melting

corrosive attack is increased with temperature
and holding time. If the scale is not uniform, then Partial melting occurs when there is nonuni-
pitting corrosion is localized to those areas form heating of the surface of the steel, such that
where scale is present. If residual salt from the some areas are heated to the liquation tempera-
bath crystallizes on the surface of the steel, ture (the partial melting temperature of an alloy)
violent boiling may occur during subsequent (Ref 77, 131). Corners and edges are particularly
quenching in oil, which may result in blister susceptible to partial melting. Microscopically,
formation on the surface. After cleaning, loca- the presence of partial melting is typically
lized pitting corrosion will then occur. To avoid observed as black spots containing retained
pitting corrosion during furnace heating, parts austenite in a large cluster of carbides. Macro-
should be thoroughly cleaned. scopically, partial melting is accompanied by
In some cases, stainless steel machine parts the formation of tiny surface cracks.
are carburized to reduce wear. However, car- Partial melting occurs when the carburized
burizing a stainless steel (1Kh16N2AM) reduces steel is heated to an excessively high tempera-
corrosion resistance where machine parts are ture, resulting in incomplete or selective car-
used in humid environments (Ref 133). The burizing of the surface. For example, partial
potential for pitting corrosion increases as the melting may occur during stray current flow into
amount of d-ferrite increases. Corrosion resis- the load from electrodes used to heat salt pot
tance is reduced as the quenching temperature furnaces or if a load is placed too close to the
prior to carburizing is increased. Although furnace hearth, so that some areas of the load are
increasing the quenching temperature after car- heated to an abnormally high temperature. To
burizing to 950 to 1100 C does not affect cor- avoid this defect, heating in salt baths with
rosion resistance, decreasing the quenching appropriate composition and at appropriate aus-
temperature to 800 to 850 C reduces corrosion tenitizing temperature should be conducted by
resistance. Increasing the tempering tempera- keeping the load at a recommended distance
ture decreases corrosion resistance, as shown in from the heating electrodes. Similarly, when
Fig. 73 (Ref 133). The corrosion resistance of heating in a conventional furnace, the load
carburized steel is greatest after stress relieving should be properly placed to facilitate uniform
at 250 C. 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 defi-
nition of the partial melting temperature range is
typically a difficult task, because of the rela-
tively large data scatter due to potential com-
positional variation within the alloy, variation of
carbon content, and, in some cases, relatively
large carbide segregation. These structural var-
iations favor the potential for partial melting to
Table 12 Approximate liquation temperature

for various steel alloys
Approximate liquation
Steel alloy temperature, C
SW14 1320
SW18 1330
SW7Mo 1280
SK5 1350
SKC 1280
Fig. 73 Effect of tempering temperature on corrosion resis-
SK5V 1270
tance of carburized stainless steel 1Kh16N2AM. The
SW12C 1260
corrosion test was conducted in a humidity cabinet. Source:
SK10V 1250
Ref 133
occur or the creation of ledeburite networks. 12. Carburizing, http://www.treatallmetals.

Improvement in temperature control will reduce com/gas.htm, Treat All Metals, Inc., Mil-
the potential for partial melting defects. waukee, WI (accessed Feb 2008)
13. Carburizing, http://www.key-to-steel.
com/Articles/Art114.htm, Key to Metals
REFERENCES Group, Zurich, Switzerland (accessed Feb
2008)
1. B.V.R. Raja, N. Pal, P.L. Talwar, and N.P. 14. V.D. Kalner, V.A. Teldekov, and S.A.
Jayaswal, Engineering Steels for Auto- Yurasov, Microstructural Hardenability
motive Applications, Steel World (No. 3), Diagrams of the Impregnated Case and
March 2005, p 13 Possibilities of Their Practical Use in
2. R.F. Kern and M.E. Suess, Chapter 10 Carburizing and Carbonitriding, Met. Sci.
Selection of Steels for Carburized Gears, Heat Treat. (USSR), (No. 4) April 1984,
Steel Selection: A Guide for Improving p 3135 (translated from Russian)
Performance and Profits, John Wiley & 15. R.F. Kern and M.E. Suess, Appendix
Sons, New York, NY, 1979, p 181205 VIIMethods of Calculating the Hard-
3. D.H. Herring, Gear Heat Treatment: The enability of Heat Treating Grades of Steel,
Influence of Materials and Geometry, Steel Selection: A Guide for Improving
Gear Technol., March/April 2004, p 3540 Performance and Profits, John Wiley &
4. M.B. Karamis and R. Ipek, An Evaluation Sons, New York, NY, 1979, p 405437
of the Using Possibilities of the Carboni- (see p 424434)
trided Simple Steels Instead of Carburized 16. J. Grosch, Chapter 14Fundamentals of
Low Alloy Steels (Wear Properties), Carburizing and Toughness of Carburised
Appli. Surf. Sci., Vol 119, 1997, p 2533 Components, Quenching and Carburizing,
5. R. Selcuk, R. Ipek, and M.B. Karamis, Proceedings of the Third Seminar of the
A Study on Friction and Wear Behaviour International Federation for Heat Treat-
of Carburized, Carbonitrided and Borided ment and Surface Engineering, The Insti-
AISI 1020 and 5115 Steels, J. Mater. tute of Materials, London, U.K., 1993,
Process. Technol., Vol 141, 2003, p 189 p 227249
196. 17. R.E. Gliner, Computer Design of Rational
6. R.F. Kern, Selecting Steels and Designing Alloying of Carburized Steels, Met. Sci.
Parts for Heat Treatment, American Heat Treat. (USSR), Vol 42 (No. 78),
Society for Metals, 1968 JulyAug 2000, p 289291 (translated
7. A. Carter, Failures in Carburized Materi- from Russian)
als, FWP J., Aug 1983, p 2728 18. J.M. Tartaglia and G.T. Eldis, Core Hard-
8. K. Palaniradja, N. Alagumurthi, and V. enability Calculations for Carburizing
Soundararajan, Optimization of Process Steels, Metall. Trans. A, Vol 15, June
Variables in Gas Carburizing Process: A 1984, p 11731183
Taguchi Study with Experimental Inves- 19. G. Parrish, Chapter 6Core Properties
tigation on SAE 8620 and AISI 3310 and Case Depth, Carburizing: Micro-
Steels, Turkish J. Eng. Env. Sci., Vol 29, structure and Properties, ASM Interna-
2005, p 279284 tional, 1999, p 135170
9. J.N. Dutta, Gear Failures for Heat Treat- 20. M. Cristinacce, Heat Treatment of Metals:
ment and Metallographic Reasons, Tech. Distortion in Case Carburized Compo-
J. Tata Iron Steel Co. (TISCO), Vol 24 nentsThe Steelmakers View, Heat
(No. 4) 1977, p 143149 Treat. Met., (No. 1), 1999, p 1821
10. O. Asi, A.C. Can, J. Pineault, and M. 21. K. Genel and M. Demirkol, Effect of Case
Belassel, The Relationship Between Case Depth on Fatigue Performance of AISI
Depth and Bending Fatigue Strength of 8620 Carburized Steel, Int. J. Fatigue,
Gas Carburized 8620 Steel, Surf. Coat. Vol 21, 1999, p 207212
Technol., Vol 201, 2007, p 59795987 22. H.E. Boyer, Ed. Chapter 1Carburizing
11. C.J. Kuehmann, G.B. Olson, J.P. Wise, Principles, Case Hardening of Steel, ASM
and C. Campbell, Advanced Case Car- International, 1987, p 112
burizing Secondary Hardening Steels, U.S. 23. S. Farfan, C. Rubio-Gonzalez, T. Cer-
Patent 6,176,946, Jan 23, 2001 vantes-Hernandez, and G. Mesmacque,
High Cycle Fatigue, Low Cycle Fatigue Effects, D.E. Diesburg, Ed., The Metal-
and Failure Modes of Carburized lurgical Society of AIME, 1984, p 89111
Steel, Int. J. Fatigue, Vol 26, 2004, p 673 33. G.E. Totten, M. Narazaki, and R.R.
678 Blackwood, Failures Related to Heat
24. H.E. Boyer, Ed., Chapter 2Gas Carbur- Treating Operations, Failure Analysis and
izing Practices, Case Hardening of Steel, Prevention, Vol 11, ASM Handbook, ASM
ASM International, 1987, p 1369 International, 2002, p 192223
25. G. Krauss, Chapter 13Microstructure, 34. H.E. Boyer, Ed., Chapter 4Liquid Case-
Residual Stresses and Fatigue of Carbur- Hardening Methods, Case Hardening of
ized Steels, Quenching and Carburizing, Steel, ASM International, 1987, p 103140
Proceedings of the Third Seminar of the 35. G.E. Totten, M. Przylecka, and W. Gestwa,
International Federation for Heat Treat- Use of Aqueous Polymer Quenchants for
ment and Surface Engineering, The Insti- Hardening Carbonitrided Parts, Nitriding
tute of Materials, London, U.K., 1993, Technology Theory and Practice, A.
p 205225 Nakonieczny, Ed., Proceed. of Ninth
26. J. Titus, The Top Ten Ways to Keep Your International Seminar International Fed-
New or Used Carburizing Equipment eration for Heat Treatment and Surface
Operating at Maximum Performance and Engineering, Institute of Precision Mech-
Efficiency, Proceedings of the 22nd Heat anics, Warsaw Poland, 2003, p 405414
Treating Society Conference and the Sec- 36. T. Reti, Residual Stresses in Carburized,
ond International Surface Engineering Carbonitrided, and Case-Hardened Com-
Congress, N.B. Dahotre, R.J. Gaster, R.A. ponents, Handbook of Residual Stress and
Hill, and O.O. Popoola, Ed., Sept 1517, Deformation of Steel, G. Totten, M. Howes
2003 (Indianapolis, IN), ASM Interna- and T. Inoue, Ed., ASM International,
tional, p 5262 2002, p. 189208
27. R.S. Hyde, G. Krauss, and D.K. Matlock, 37. G. Parrish, Chapter 4Retained Auste-
Phosphorus and Carbon Segregation: nite, Carburizing: Microstructure and
Effects on Fatigue and Fracture of Gas- Properties, ASM International, 1999,
Carburized Modified 4320 Steel, Metall. p 7797
Mater. Trans. A, Vol 26 (No. 6), June 38. C.Y. Kung and J.J. Rayment, An Exam-
1994, p 12291240 ination of the Validity of Existing Empe-
28. T. Sekercioglu and V. Kovan, Pitting rical Formulae for the Calculation of Ms
Failure of a Truck Spiral Bevel Gear, Eng. Temperature, Metall. Mater. Trans. A,
Fail. Anal., Vol 14, 2007, p 614619 Vol 13, Feb 1994, p 328331
29. C. Razim and H. Streng, Properties of 39. I.L. Trusova, Problem of Retained Auste-
Case-Hardened ComponentsBasic nite in Deeply Cemented Steel
Considerations Concerning the Definition 20Kh2N4A, Met. Sci. Heat Treat. (USSR),
and Evaluation of Toughness, Case- Vol 40 (No. 1112), Nov-Dec 1998,
Hardened Steels: Microstructural and p 445448 (translated from Russian)
Residual Stress Effects, D.E. Diesburg, 40. A.P. Gulyaev, Cold Treatment of Steel,
Ed., The Metallurgical Society of AIME, Met. Sci. Heat Treat. (USSR), Vol 40
1984, p 115 (No. 1112), Nov-Dec 1998, p 449455
30. G. F. Vander Voort, Ed., Ferrous Marten- (translated from Russian)
site, Metallography and Microstructures, 41. H.E. Boyer, Ed. Chapter 6Heat Treat-
Vol 9, ASM Handbook, ASM Interna- ment Considerations, Case Hardening of
tional, 2004. Steel, ASM International, 1987, p 151194
31. AISI 8620 Thermal Treatments, http:// 42. G. Parrish, Chapter 7Post Hardening
www.crucibleservice.com/eselector/ TreatmentsThermal, The Influence of
prodbyapp/carbon/8620.html, Crucible Microstructure on the Properties of Case-
Service Centers, Camillus, NY (accessed Hardened Components, American Society
Feb 2008) for Metals, 1980, p 167193
32. D.W. Ingham and P.C. Clarke, Prediction 43. G. Parrish, Chapter 7Post Carburizing
of Microstructure and Hardness in Car- Thermal Treatments, Carburizing: Micro-
burized Gears, Case-Hardened Steels: structure and Properties, ASM Interna-
Microstructural and Residual Stress tional, 1999, p 171199
44. D.P. Koistinen, The Distribution of Resi- 54. G. Albertini, G. Bruno, F. Fiori, M. Mar-
dual Stresses in Carburized Cases and cantoni, E. Quadrini, and F. Turquier,
Their Origin, Trans. ASM, Vol 50, 1958, Residual Stresses in Case Hardened
p 227241 Automotive Steel Gears, Mater. Res.
45. J.L. Pacheco and G. Krauss, Micro- Forum, Vol 404407, 2002, p 191196
structure and High Bending Fatigue 55. M. Benedetti, V. Fontanari, B.-R. Hohn, P.
Strength of Carburized Steel, J. Heat Oster, and T. Tobie, Influence of Shot
Treat., Vol 7 (No. 2), 1989, p 7786 Peening on Bending Tooth Fatigue Limit
46. O. Asi, An Investigation of Retained of Case Hardened Gears, Int. J. Fatigue,
Austenite Contents in Carburized SAE Vol 24, 2002, p 11271136
8620 Steel, G. U. J. Sci., Vol 17 (No. 4), 56. J. Champaigne, Shot Peening Overview,
2004, p 103113 http://www.shotpeener.com, Electronics
47. V.F. da Silva, L.F. Canale, D. Spinelli, Inc., Mishawaka, IN, 2001 (accessed Feb
W.W. Bose-Filho, and O.R. Crnkovic, 2008)
Influence of Retained Austenite on Short 57. B.A. Shaw, C. Aylott, P. OHara, and K.
Fatigue Crack Growth and Wear Re- Brimble, The Role of Residual Stress on
sistance of Case Carburized Steel, Fatigue Strength of High Performance
J. Mater. Eng. Perform., Vol 8, 1999, Gearing, Int. J. Fatigue, Vol 25, 2003,
p 543548 p 12791283
48. D. Jeddi, H. Sidhom, D. Ghiglione, and 58. Nakonieczny, W. Szyrie, and J. Suwalski
H.-P. Lierade, Role of Cyclic Stability of Residual Stresses, Microstructure and
Retained Austenite in Fatigue Perfor- Fatigue Behaviour of Carburized Layers
mance of Carburized 14NiCr11 Steel, Before and After Shot Peening, Paper
J. Mater. Eng. Perform., Vol 14, 2005, 1996073, Sixth International Conference
p 3749 on Shot Peening, Conference Proceedings
49. H.-J. Kim and Y.-G. Kweon, The Effects of ICSP-6, J. Champaigne, Ed., Sept 25,
of Retained Austenite on Dry Sliding Wear 1996 (San Francisco, CA)
of Carburized Steels, Wear, Vol 193, 1996, 59. M. Kikuchi, H. Ueda, K. Hanai, and T.
p 815 Naito, The Improvement of Fatigue Dur-
50. I.S. Kozlovskii, S.E. Manevskii, and I.I. ability of Carburized Steels with Surface
Sokolov, Effect of Retained Austenite on Structure Anomolies by Shot Peening,
the Resistance to Scoring of Steel Paper 1984067, Second International
20Kh2N4A, Met. Sci. Heat Treat. (USSR), Conference on Shot Peening, Conference
Vol 20 (No. 1), Jan 1978, p 7880 (trans- Proceedings of ICSP-2, 194, p 208214
lated from Russian) 60. D.H. Herring, A Tutorial on Shot Blasting
51. C. Kim, Fracture of Case-Hardened Steels and Shot/Laser Peening, Ind. Heat., Vol 73
and Residual Stress Effects, Case-Har- (No. 8), Aug 2006, p 2022
dened Steels: Microstructural and Resi- 61. Shot Peening and Shot Blasting, OSK-
dual Stress Effects, D.E. Diesburg, Ed., Kiefer GmbH, Oberflachen and Strahl-
The Metallurgical Society of AIME, 1984, technik, Ettlingen, Germany
p 5987 62. M. Kobayashi, T. Matsui, and Y. Mur-
52. J.A. Burnett, Prediction of Stresses Gen- akami, Mechanism of Creation of Com-
erated during the Heat Treating of Case pressive Residual Stress by Shot Peening,
Carburized Parts, Residual Stress for Int. J. Fatigue, Vol 20 (No. 5), 1998, p 351
Designers and Metallurgists, L.J. Vande 357
Walle, Ed., American Society for Metals, 63. C. Jia-Xiang and A. Bing-qiu, The Effect
1981, p 5169 of Shot Peening on Contact Fatigue Life of
53. N. Kanetake, Chapter 17Effect of Car- Carburized Steel, Proceedings of First
burising on Mechanical Properties of International Conference on Shot Peen-
Steels, Quenching and Carburizing, Pro- ing, Sept 1417, 1981 (Paris), Pergamon
ceedings of the Third Seminar of the Press, p 333339
International Federation for Heat Treat- 64. C. Peyrac, J.F. Flavenot, and F. Convert,
ment and Surface Engineering, The Insti- Combining Case Hardening and Shot
tute of Materials, London, U.K., 1993, Peening for Gear Steels: Influence on
p 281291 Residual Stress, Proceedings for the
Seventh International Conference on Shot Treatment Parameters on the Accuracy of

Peening, A. Nakonieczny, Ed., Sept 28 Gear Wheels, Met. Sci. Heat Treat.
Oct 1, 1999 (Warsaw, Poland), p 362370 (USSR), Vol 12 (No. 12), Dec 1970,
65. M. Kikuchi, H. Ueda, and J. Naito, Fatigue p 10071010 (tranlated from Russian)
Behavior of Carburized Steel with Internal 75. V.A. Bulgakov, Effect of Original Struc-
Oxides Near the Surface, Metall. Mater. ture on Deformation and Warping of
Trans. A, Vol 18 (No. 1), Jan 1987, p 156 Machine Parts after Final Heat Treatment,
158 Met. Sci. Heat Treat. (USSR), Vol 19
66. J.L. Woods, S.R. Daniewicz, and R. Nel- (No. 9), Sept 1977, p 784786 (translated
lums, Increasing the Bending Fatigue from Russian)
Strength of Carburized Spur Gear Teeth 76. N.M. Grinberg and M.P. Konovalova,
by Presetting, Int. J. Fatigue, Vol 21, 1999, Reasons for the Changes in the Dimen-
p 549556 sions of Steel Parts During Carburizing,
67. K. Funatani, Residual Stresses during Gear Met. Sci. Heat Treat. (USSR), Vol 10 (No.
Manufacture, Handbook of Residual Stress 6), June 1968, p 483485 (translated from
and Deformation of Steel, G. Totten, M. Russian)
Howes, and T. Inoue, Ed., ASM Inter- 77. J.Wyszykowski Chapter 12The Defects
national, 2002, p 437457 and Quality Inspection of Heat-Treatment,
68. Chapter 14Dimensional Control, Gas The Engineers HandbookThe Heat-
Carburizing, ASM Society for Metals, Treatment of Ferrous Metals, W. Luty,
1964, p 125136 Ed., Publishing House of Scientifically-
69. J. Koutsky and V. Pechman, Heat- Technical, Warsaw, 1977, p 600627
Treatment Procedure and Volume Change, 78. G. Krauss, Chapter 10Surface Hard-
Heat Treatment of Engineering Com- ening, Principles of Heat Treatment of
ponents: Proceedings of The Conference Steel, American Society for Metals, 1980,
Organized by the Heat Treatment Joint p 247273
Committee of the Iron and Steel Institute, 79. G. Krauss, Steels: Processing, Structure,
Dec 1718, 1969 (Royal Garden Hotel, and Performance, ASM International,
London), The Institute of Metals and The 2005
Institute of Metallurgists, p 129133 80. C.R. Brooks, Chapter 4Quenching of
70. G.A. Fuller, Atmosphere Carburizing Steels, Principles of the Heat Treatment of
at 927 C (1700 F) for Distortion Con- Plain Carbon and Low Alloy Steels, ASM
trol, J. Heat Treat., Vol 8 (No. 1) 1990, International, 1996, p 87126
p 2126 81. S.I. Aksenova, M.A. Balter, I.S. Dukar-
71. C. Prinz, B. Clausen, F. Hoffmann, R. evich, and M.M. Gurarii, Causes of Clea-
Kohlmann, and H.-W. Zoch, Metallurgical vage on Case-Hardened Gear Wheels,
Influence on Distortion of the Case- Met. Sci. Heat Treat. (USSR), Vol 30
Hardening Steel 20MnCr5, Proceedings of (No. 3), March 1988, p 191194 (translated
First International Conference on Dis- from Russian)
tortion Engineering, Sept 1416, 2005 82. A.J. McEvily, K. Pohl, and P. Mayr,
(Bremen, Germany), p 7582 Comparison of the Fractographic Features
72. D.T. Llewellyn and W.T. Cook, Heat- of a Carburized Steel Fractured Under
Treatment Distortion in Case-Carburizing Monotonic or Cyclic Loading, Mater.
Steels, Met. Technol., May 1977, p 265 Charact., Vol 36, 1996, p 153157
278 83. Kh.G. Tkhagapsoev and G.K. Surzhinskii,
73. L.N. Vashchuk, O.A. Konko, V.T. Alek- Cracking in the Carburized Case of Gears
seenko, and V.K. Adamanova, Volume During Their Grinding, Met. Sci. Heat
Changes of Large Gear Wheels in Car- Treat. (USSR), Vol 11 (No. 34), March
burization and Heat Treatment, Met. Sci. April 1969, p 224226 (translated from
Heat Treat. (USSR), Vol 28 (No. 12), Russian)
NovDec 1986, p 858860 (translated 84. G. Parrish, Chapter 8Post Hardening
from Russian) TreatmentsMechanical, The Influence
74. V.A. Murzin, V.P. Ponomarev, N.N. Tol- of Microstructure on the Properties of
machevakii, G.D. Glezer, A.V. Poropov, Case-Hardened Components, American
and A.P. Kul, Effect of Separate Heat Society for Metals, 1980, p 195228
85. G. Parrish, Chapter 8Postcarburizing Steel, U.S. Patent 6,149,734, Nov 21,
Mechanical Treatments, Carburizing: 2000
Microstructures and Properties, ASM 98. E. Gianotti, I. Calliari, M. Zanesco, and
International, 1999, p 199230 E. Ramous, Grain Boundary Oxidation
86. M. Lawerenz and I. Ekis, Creating an In- in Endothermic Gas Carburizing Process,
House Shot Peening Specification for Proceedings of the Seventh International
GearsPart 2, Paper 1993118, Shot Tooling Conference, Tooling Materials
Peener, Vol 7 (No. 1), 1993, p 35 and Their Applications from Research to
87. C. Bahnsen, B. Clausen, F. Hoffmann, P. Market, Vol 2, May 25, 2006, Politecnico
Mayr, and H.-W. Zoch, Influence of Di Torino, Torino, Italy, p 481490
Carburising on Distortion Behavior, Pro- 99. D.K. Matlock, K.A. Alogab, M.D.
ceedings of First International Con- Richards, and J.G. Speer, Surface Proces-
ference on Distortion Engineering, Sept sing to Improve the Fatigue Resistance of
1416, 2005 (Bremen, Germany), p 235 Advanced Bar Steels for Automotive
242 Applications, Mater. Res., Vol 8 (No. 4),
88. M. Izciler and M. Tabur, Abrasive Wear 2005, p 453459
Behavior of Different Case Depth Gas 100. H.-P. Schmidt, Process for Avoiding Sur-
Carburized AISI 8620 Gear Steel, Wear, face Oxidation in the Carburization of
Vol 260 (No. 12), 2006, p 9098 Steels, U.S. Patent 5,498,299, March 12,
89. M. Boniardi, F. DErrico, and C. Taglia- 1996
bue, Influence of Carburizing and Nitrid- 101. V.D. Kalner and S.A. Yurasov, Internal
ing on Failure of GearsA Case Study, Oxidation During Carburizing, Met. Sci.
Eng. Fail. Anal., Vol 13, 2006, p 312339 Heat Treat. (USSR), Vol 12, (No. 6), June
90. G. Parrish, Chapter 6Core Properties 1970, p 451454 (translated from Russian)
and Case Depth, The Influence of Micro- 102. J.D. Verhoeven, N. Chen, and A.J. Bevolo,
structure on the Properties of Case- Scanning Auger Microscopy Character-
Hardened Components, American Society ization of the Internal Oxidation Produced
for Metals, 1980, p 131165 on Carburizing, J. Heat Treat., Vol 9
91. L. Faqian, L. Shuanlu, l. Helin, Q. (No. 2), 1992, p 7380
Changyi, Z. Jie, T. Jiping, and C. Jun, 103. X. An, J. Cawley, W.M. Rainforth, and L.
Analysis of Causes of Casing Elevator Chen, A Study of Internal Oxidation in
Fracture, Eng. Fail. Anal., Vol 14, 2007, Carburized Steels by Glow Discharge Op-
p 606613 tical Emission Spectroscopy and Scanning
92. Z. Yu, X. Xu, and H. Ding, Failure Ana- Electron Microscopy, Spectrochim. Acta
lysis of a Diesel Engine Piston-Pin, Eng. B, At. Spectrosc., Vol 58, 2003, p 689698
Fail. Anal., Vol 14, 2007, p 110117 104. S. Laue, H. Bomas, and F. Hoffman,
93. Z.Ya. Nostislavskii, Yu.N. Griboedov, and Influence of Surface Condition on the
S.I. Markov, Effect of the Atmosphere Fatigue Behaviour of Specimens Made of
During Heating of Carburized Cr-Ni Steel SAE 5115 Case-Hardened Steel, Fatigue
to Quenching Temperature on Decom- Fract. Eng. Mater. Struct., Vol 29, 2006,
position of Austenite, Met. Sci. Heat Treat. p 229241
(USSR), Vol 18 (No. 56), MayJune 1976, 105. M. Lohrmann and W. Lerche, Investiga-
p 443445 tion of Process and Steel Influences on the
94. G. Parrish, Chapter 2Decarburization, Extent of Internal Oxidation during Gas
Carburizing: Microstructure and Proper- Carburizing, Heat Treat. Met., Vol 27
ties, ASM International, 1999, p 3749 (No. 4), 2000, p 8590
95. G. Parrish, Chapter 5Influential 106. I.S. Kozlovskii, A.T. Kalinin, A. Va. Novi-
Microstructural Features, Carburizing: kova, E.A. Lebedeva, and A.I. Feofanova,
Microstructure and Properties, ASM Internal Oxidation during Case-Hardening
International, 1999, p 99133 of Steels in Endothermic Atmospheres,
96. Atlas of Microstructures of Industrial Met. Sci. Heat Treat (USSR), Vol 9 (No. 3),
Alloys, Vol 7, Metals Handbook, 8th ed., March 1967, p 157161
American Society for Metals, 1972, p 4 107. W.D. Kehr and H.E. Seese, Surface Attack
97. T. Isogai, R. Uchino, M. Sato, and Y. in Chromium-Iron Alloys, IBM J., Sept
Kono, Method for Heat Treatment of 1966, p 420421
108. Z. Yu and X. Xu, Failure Analysis of an Steel 16MnCrS5, Fatigue Fract. Eng.
Idler Gear of Diesel Engine Gearbox, Eng. Mater. Struct., Vol 28, 2005, p 983995
Fail. Anal., Vol 13, 2006, p 10921100 121. T. Wilks, G.P. Cavallaro, C. Subramanian,
109. M. Umemoto, Z.G. Liu, K. Masuyama, K.N. Strafford, P. French, and J.E. Allison,
and K. Tsuchiya, Influence of Alloy Ad- Conditions Prevailing in the Carburizing
ditions on Production and Properties of Process and Their Effect on the Fatigue
Bulk Cementite, Scr. Mater., Vol 45, 2001, Properties of Carburized Gears, J Mater.
p 391397 Process. Technol., Vol 40, 1994, p 111
110. S.N. Saltykov and N.V. Tarasova, Anodic 125
Dissolution of Ferrite Phases of Iron- 122. T. Masuyama, K. Inoue, M. Yamanaka, K.
Carbon Ferrite-Cementite Alloys with Kitamura, and T. Saito, Evaluation of
Different Forms of Cementite, Prot. Met. Bending Strength of Carburized Gears
(Russia), Vol 42 (No. 5), 2006, p 501506 Based on Inferential Identification of
111. V.Ya. Zubov, Patenting of Steel Wire, Principal Surface Layer Defects, JSME
Met. Sci. Heat Treat., (USSR), Vol 14 Int. J., Series C, Vol 45 (No. 3), 2002,
(No. 9), Sept 1972, p 793800 p 794801
112. G. Parrish, Chapter 3Carbides, The 123. P. Brechot, A.B. Cardis, W.R. Murphy,
Influence of Microstructure on the Proper- and J. Thiessen, Micropitting Resistant
ties of Case-Hardened Components, Am- Industrial Gear Oils with Balanced Per-
erican Society for Metals, 1980, p 3960 formance, Ind. Lubr. Tribol., Vol 52
113. T. Ichihara and H. Fujiwara, Pullu Unit (No. 3), p 125136
Having One-Way Clutch, U.S. Patent 124. S. Matsumoto, On the Design of Surface
Application US2003/0085091 A1, May 8, Durability of Rolling-Sliding Contact
2003 Surface, Technical Paper n77-DET-115,
114. T. Nakamura, S. Nakamura, T. Kimura, American Society of Mechanical Engi-
T. Yamaguchi, K. Otani, N. Uchiyama, neers, Sept 2630, 1977
N. Kino, and T. Nanbu, High Bearing 125. T.L. Krantz, M.P. Alanou, H.P. Evans, and
Pressure-Resistant Member, U.S. Patent R.W. Snidle, Surface Fatigue Lives of
6,569,267 B1, May 27, 2003 Case-Hardened Gears with an Improved
115. G. Krauss, Heat Treatment, Micro- Surface Finish, J. Tribology, Vol 23, Oct,
structures, and Residual Stresses in Car- p 709716
burized Steels, Conf. Proceed. Quenching 126. B.B. Vinokur, S.E. Kindratyuk, R.A.
and Distortion Control, G.E. Totten, Ed., Khrunik, K.B. Katsov and I.V. Kokotailo,
ASM International, 1992, p 181191 Influence of Carbon on the Contact En-
116. L.P. Karpov, Effect of Quenching Condi- durance of Carburized and High-Carbon
tions on the Carbide Network in Steel Steels, Sov. Mater. Sci., Vol 22 (No. 2),
07Kh16N6, Met. Sci. Heat Treat. (USSR), March 1986, p 211213
Vol 35 (No. 1), Jan 1993, p 2023 (trans- 127. R.C. Graham, A. Olver, and P.B. Mac-
lated from Russian) pherson, Investigation into the Mech-
117. G. Diego, M. Serrano, and A.M. Lancha, anisms of Pitting in High-Hardness
Failure Analysis of a Multiplier from a Carburized Steels, Paper 80-C2/DET-118,
Kaplan Turbine, Eng. Fail. Anal., Vol 7, ASME, 1980
2000, p 2734 128. R.F. Kern and M.E. Suess, Chapter 9
118. L. Zhang and B.G. Thomas, State of the General Principles for the Selection of
Art in Evaluation and Control of Steel Steels for Case Hardening, Steel Selection:
Cleanliness, ISIJ Int., Vol 43 (No. 3), A Guide for Improving Performance and
2003, p 271291 Profits, John Wiley & Sons, New York,
119. Y. Murakami and M. Endo, Effects of NY, 1979, p 159180
Defects, Inclusions, and Inhomogenities 129. A. Bensely, S.S. Jayakumar, D. Mohan
on Fatigue Strength, Fatigue, Vol 16, Lal, G. Nagarajan and A. Rajadurai, Fail-
April 1994, p 163182 ure Investigation of Crown Wheel and
120. H. Bomas and M. Schleicher, Application Pinion, Eng. Fail. Anal., Vol 13 (No. 8),
of the Weakest-Link Concept to the Dec 2006, p 12851292
Endurance Limit of Notched and Multi- 130. S. Eryu, C. Qing, and L. Xiaogang, Frac-
axially Loaded Specimens of Carburized ture of Case Crushing of Carburized
Steel, Wear, Vol 122 (No. 1), 1988, p 57 (No. 5), May 1970, p 393395 (translated
62 from Russian)
131. A. Moszczynski, The Gas Carburized of 133. V.I. Belyakova, M.F. Alekseenko, L.Ya.
Steel, Publishing House of Scientifically- Gurvich and V.L. Erofeeva, The Corrosion
Technical, Warsaw, 1983, p 259262 Resistance of Carburized Stainless Steels,
132. V.I. Muravev and V.P. Kurbatov, Pitting Met. Sci. Heat Treat. (USSR), Vol 15
Corrosion in the Process of Heat Treat- (No. 2), Feb 1973, p 143146 (tanslated
ment, Met. Sci. Heat Treat. (USSR), Vol 12 from Russian)
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_241-253.pdf/F5113_07/ 18/8/2008 3:26PM Plate # 0 pg 241
Fatigue Fracture of Nitrided Layers

Aleksander Nakonieczny, Institute of Precision Mechanics
THE DURABILITY of products depends technological parameters, as well as the type of

strongly on surface conditions on the order of loading in service conditions. When discussing
1 mm to several millimeters, depending on the the problem of fatigue resistance, one should
type of technological process applied. The con- consider in detail the effect of these parameters
dition of the surface layer is critical to wear and, in the case of loads, define fatigue charac-
resistance during the process of friction and to teristics (e.g., plots for different types of loading,
corrosion resistance. In the case of mechanical such as bending, tensile, and torque). These
loading (especially fatigue resistance) and cor- problems have been sufficiently dealt with in the
rosion, a critical role is played by the substrate, its technical literature.
condition and properties, as well as the atomic In the process of searching for methods to
relationship between the surface layer and the increase fatigue resistance, there are some con-
substrate. For this very reason, one should take stant elements that have a favorable effect,
into account the substrate-surface layer system including enhancing treatments such as thermal,
when considering the life expectancy of machine thermochemical, as well as surface work hard-
components and assemblies. Attributes of such a ening.
system are the thickness of the surface layer, the In order to increase fatigue resistance, it is
ratio of this thickness to the entire cross section, not sufficient to apply a chosen enhancement
the ratio of surface hardness to core hardness, treatment. Rather, it is important to select
and the state of residual stresses, usually com- the appropriate initial volume heat treatment
pressive, situated within the surface layer rela- prior to successive surface, thermal, and work-
tive to the state of stresses in the substrate, which hardening treatments. The problem of enhancing
are usually tensile. An incorrectly applied sur- fatigue resistance of machine components by
face layer may cause the formation of a structural technological methods involves the application
flaw in the transition zone of the layer and may of not one chosen treatment but a cycle of suc-
be the location of crack initiation, especially by a cessive treatments. The appropriate selection
fatigue mechanism (Ref 1). of these treatments affects the structural flaw
Surface engineering encompasses various formed in the process of enhancement, which has
process technologies and also the service prop- a decisive influence on fatigue resistance (Ref 1).
erties of products, surface-layer investigation A structural flaw occurs in all locations where,
methodologies, and design aspects of the as the result of heat treatment (e.g., induction
substrate-layer system. In terms of service hardening), thermochemical treatments (car-
properties of products, the functionality of burizing, nitriding, etc.), or work-hardening
surface treatment may be assessed by defining treatments (burnishing, shot peening) of ma-
the fatigue limit, wear resistance, or corrosion chine components, the layer formed in these
resistance. Such evaluations are usually per- processes has different physical-chemical prop-
formed on specimens in laboratory conditions. erties than that of the core due to a large gradient
However, the most valuable information is to be of property changes. The value of the structural
gained from actual service trials, which also flaw coefficient, bs, depends on the type of
unfortunately may be costly. material and the parameters of the technological
processes that cause this structural flaw to form.
In other words, it depends on heat treatment and
Fatigue Resistance surface hardening.
Thermal and surface work-hardening treat-
Fatigue resistance of machine components ments used industrially cause enhancement of
is a function of their design, material and fatigue resistance. Based on research carried out
by the Institute of Precision Mechanics, it can rotational-bending fatigue resistance of 40HM-

be accepted that the fatigue limit (s 1) rises 15 grade steel (AISI 4140). This structural steel is
to 30% on average as a result of implementation commonly used for the manufacture of various
of such treatments. Enhancement of fatigue types of machine components (e.g., gears,
resistance is obtained by structural changes, crankshafts). Specimens used for the study were
strengthening, and favorable distribution of quenched and tempered to the following hard-
residual stresses, which are formed as a result ness levels: 30 to 32 HRC, 33 to 34 HRC, and 35
of thermal and surface treatments. to 36 HRC, applying tempering temperatures
Due to physical processes taking place within within the range of 550 to 620 C.
the material during the application of surface, Nitriding was carried out in a controlled
thermal, and work-hardening treatments, chan- process at a temperature of 530 C for 4 to
ges in microstructural and mechanical properties 16 h. Investigations encompassed metallurgical
arise between the surface layer and the core characterization of nitrided layers as well as
of the material. The gradient of changes of determination of fatigue resistance (s 1). More-
physical-chemical properties depends on the over, the yield strength (R0.2) was determined in
selected technological process and its para- conditions of shear bending (Rg0.2).
meters. Numerous examples have been noted For the metallurgical investigations pertain-
where the fatigue crack origins were traced to ing to surface hardnesses, hardness traverses,
the transition zone between the hardened surface layer thickness, and microstructure, the Neophot
layer and the substrate. Figure 1 shows a fatigue 30 metallograph and the Zwick microhardness
fracture with the origin located under the hard- tester were used.
ened layer at the point where stresses mounted. Investigation of fatigue resistance was carried
The mounting of stresses occurs as the result of out with the aid of the PUNZ machine, manu-
residual stresses created during heat treatment factured by Schenk. The loading frequency was
combined with external stresses. 100 Hz, and the investigations covered 107
Fatigue Resistance of Steel after Nitriding cycles. The value of the fatigue limit was cal-
and Related Nitriding Treatments. The sig- culated by the Dixon-Mood method. Results of
nificance and detailed assessment of the effect of fatigue tests of nitrided specimens were com-
a structural flaw are explained, using investiga- pared with results obtained for the same steel
tions of the effect of variable core conditions on (40HM grade) quenched and tempered to a
fatigue resistance as an example. hardness of 30 to 32 HRC.
Reference 2 defines the effect of tempering Fractographic investigations of fatigue frac-
temperature and time of nitriding on the tures and determination of the chemical com-
position of visible inclusions on these fractures
were carried out by scanning electron micro-
scopy. Yield strength values were determined
in cases of static bending with the aid of the
Instron TT-DM machine. The results of metal-
lurgical 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 Surface hardness
Temper Total
temper- Core Effective at core
ature hardness, Nitriding at 500 +50 HV s 1
C HRC time, h HV, mm mm HV1 HV10 MPa
620 3032 ... ... ... ... ... 550
620 3032 4 0.13 0.24 686 642 735
16 0.26 0.46 743 657 745
590 3334 4 0.14 0.22 752 695 720
8 0.19 0.29 777 707 725
16 0.27 0.43 786 701 777
550 3536 4 0.16 0.2 772 698 820
Fig. 1 Fatigue source located in the transition zone between 16 0.30 0.42 778 699 840
case and core
Fatigue Fracture of Nitrided Layers / 243
Investigations of fatigue resistance (Table 1) temperature, from the point of view of the
showed that nitriding, independently of process fatigue limit. A tempering temperature of
time (case depth and core hardness), caused an 590 C is intermediate between the two afore-
increase in the fatigue limit in comparison with mentioned temperatures, at which this effect is
quenching and tempering. The smallest increase manifest only after the application of a longer
in fatigue limit value (by 30%) was obtained on nitriding time. It should therefore be emphasized
specimens tempered at 590 C and nitrided for that by the appropriate selection of the temper-
4 h. The greatest increase in the fatigue limit (49 ing temperature, it is possible to achieve an
to 53%) was obtained on specimens tempered at increase in the fatigue limit with shorter nitrid-
550 C and nitrided for 4 and 16 h. For the ing times. On the other hand, selection of inap-
remaining versions of technological processing, propriate tempering temperatures may cause the
the increase in fatigue limit was almost identical inability to achieve an increase in the s 1 value
and amounted to 31 to 35%. when the nitriding time is too short.
Investigations of quenched and tempered and An analysis of angular coefficients of simple
subsequently nitrided specimens showed a vari- regressions of the Wohler plots (Table 2) indi-
ation in the value of the fatigue limit (s 1), cates that, for almost all versions, the angle of
depending on the tempering temperature. inclination of the regression plots is similar. By
Nitriding time did not affect the value of the the same token, the sensitivity of the material to
fatigue limit, both after tempering at 550 C as a change in loading within limited fatigue
well as at 620 C. strength is not connected with the time of
Some variation was observed in the case of nitriding. Only specimens tempered at 550 C
versions in which the tempering temperature and nitrided for 16 h were characterized by a
was 590 C. For this tempering temperature, the slightly higher sensitivity to a change in the level
highest value of the fatigue limit (777 MPa) was of loading. By the same token, an increase in the
obtained on specimens nitrided for 16 h. At the level of loading within limited fatigue strength
same time, it was established that increasing the causes a more significant decrease in the number
time of nitriding from 4 to 8 h did not affect the of loading cycles to failure, compared with other
fatigue limit. The difference of approximately technological versions. Thus, shortening the
1% is within experimental error. Extending the nitriding time coupled with a lower tempering
time of nitriding from 4 and 8 h to 16 h caused temperature is favorable even when the loading
an increase in fatigue limit value by approxi- level exceeds the fatigue limit.
mately 50 MPa, that is approximately 6.5%. Fractography of specimen fatigue fractures
A comparison of specimens tempered at 550 showed they are of a fine-grained character, with
and 620 C shows that lowering the tempering the exception of the middle zone, which has a
temperature causes an increase in the fatigue differing, coarse-grained structure (catastrophic
limit by approximately 12%. As has already failure zone). The clearly observed fatigue
been mentioned, decreasing the nitriding time sources occur in the form of fisheyes (Fig. 2).
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
Table 2 Equations of regression for the

investigated technological versions
Tempering Nitriding
temperature, C time, h Regression equations
620 4 s 1 = 141.2 lgN+1567.9
16 s 1 = 62.4 lgN+1146.6
590 4 s 1 = 58.6 lgN+1092.7
16 s 1 = 67.7 lgN+1201.5
550 4 s 1 = 137.2 lgN+1658.3 Fig. 2 Appearance of fatigue source. Original magnification:
16 s 1 = 27.6 lgN+1025.8 100
Initiation of fatigue fractures occurs at the Fatigue Evaluation of Nitrided Steels

border between the layer and the substrate, in a
zone characterized by hardness higher than that An evaluation of fatigue properties requires
of the core by 50 HV. The features initiating an understanding of the mechanism. This
fatigue fractures are nonmetallic inclusions of mechanism also depends on the condition and
calcium sulfides (Fig. 3), originating from the mutual relationship between the substrate and
metallurgical process. the surface layer. This requires determination of
Results of investigations of yield strength in the strength condition of the system: substrate-
bending indicate that variation of both the tem- surface layer, as a function of external loading.
pering temperature as well as the nitriding time Among the most important parameters
affects its value insignificantly, not exceeding describing the condition of the surface layer are
7%. The values of Rg are all within the range of microstructure, degree of strengthening, state of
1850 to 1980 MPa. stresses, and roughness. Other important para-
Results of metallurgical evaluations of nitri- meters include texture, surface energy, and
ded layers (Table 1) show that extension of the chemical composition (Ref 2, 3).
nitriding time from 4 to 16 h causes an almost In engineering practice, usable properties
doubled growth of the effective minimum ten- such as tensile strength, Rm and fatigue limit,
dency to increase the case depth with a lowering may be determined as functions of mechanical
of the tempering temperature. A similar but clear parameters, that is, hardness, H, or tensile
trend is observed in the surface hardness of strength, Rm, or surface roughness (Ref 4). Such
layers, where lowering of the tempering tem- correlations maintain their validity for a material
perature from 620 to 550 C with a 4 h nitriding homogeneous throughout its cross section. For
time cycle causes a rise in HV1 surface hardness heterogeneous materials (e.g., ones that have
by approximately 100 units and in HV10 hard- been surface treated), such correlations cannot
ness by 50 units. be applied directly. The character of distribution
Figure 4 shows the changes taking place in the
character of the microhardness profile as a
function of process time, with tempering tem-
perature 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. 4 Microhardness traverses for different nitriding process

times
Fig. 3 Precipitation on the fracture surface of a specimen

which served as the source of the fatigue fracture. Fig. 5 Fatigue characteristics for 1, a homogeneous material,
Original magnification: 500 and 2, a heterogeneous material
of basic mechanical properties (hardness, Knowledge of the fatigue characteristic, as

residual stresses, as well as their significant well as the loading characteristic, allows the
effect on fatigue properties) vary in different determination of the fatigue strength condition
ways for homogeneous or heterogeneous mate- for any location on the component cross section:
rials (Fig. 5). Homogeneous materials (curve 1,
Fig. 5) have constant properties through the s1 is0z (Eq 3)
i
cross section. Heterogeneous materials (curve 2)
have variable properties. where s l is the fatigue limit at any location on
The fatigue characteristic, which is shown the specimen cross section, and si0z is the value
by the distribution of the fatigue limit, s, is a of stresses from extraneous loading at the given
function of hardness, H, and residual stresses, sr: location i.
The investigations were conducted on struc-
s=f H,sr (Eq 1) tural steels 40HM (4140) and 38HMJ (Nitralloy
135M). The steels were hardened and tempered
The method of designing a usable characteristic prior to nitriding at two temperatures: 550 and
for the fatigue limit distribution has been 620 C (Ref 1). Nitriding was carried out using
described in publications (Ref 35). two types of atmosphere, that is, NH3-NH3(diss)
The distribution of stresses from extraneous and NH3-N2. In the nitriding processes, the
loading constitutes a significant characteristic, atmosphere gas composition was varied, as were
because it enables the determination of the the time of nitriding (4 and 16 h) and the
strength condition for the surface layer. Loading nitriding potential, KN (from 1.65 to 4.8).
characteristics are typical distributions of Fatigue resistance tests were carried out on
stresses across the section of a component or the PUNZ machine (manufactured by Schenck),
specimen for the investigated types of loading: applying rotational bending stresses with a
tensile-compressive, bending, or torque. For notch (a = 1.02). The specimen diameter was
smooth specimens, their determination does not W = 5.88+0.02 mm. The results of the fatigue
present any problems. Some difficulties may resistance tests, metallurgical evaluation and
arise when determining the distribution of process parameters are shown in Table 3.
stresses in a notched specimen. Figure 6 shows microhardness traverses in the
Distribution of stresses from extraneous nitrided case for 40HM-grade steel, while Fig. 7
loading for notched specimens in conditions of shows the same for the 38HMJ grade.
bending is (Ref 3): Test results show that the tempering tem-
perature has an effect on the properties of the
nitrided case. The effect of the tempering tem-
2 x 3a2
smax x=sn a 17 (Eq 2) perature on the basic properties of the nitrided
d
case depends on the steel grade. A higher
where smax(x) is the value of local stress at a increase in hardness (by approximately 50%) as
distance of (x) from the surface, sn is the well as in case depth is observed for the low-
nominal stress, a is the coefficient of stress alloy chromium steel 40HM.
concentration, and d is the cross-sectional As can be seen in Fig. 7 for the 38HMJ steel,
dimension. the effects of tempering temperature on the
Table 3 Technological parameters and test results

HV0.5 hardness Residual stresses
Tempering Core Max At Depth at which
temperature, Nitriding hardness on cross On Fatigue limit surface, stress changes
Steel grade C time, h HV0.5 section surface (s 1), MPa MPa sign, mm
40HM (4140) 550 4 402 677 757 820 600 0.32
550 16 396 642 757 840 650 0.52
620 4 343 715 826 735 600 0.37
690 16 343 343 642 745 900 0.55
38HMJ 550 4 356 1030 1373 805 900 0.25
(Nitralloy 135M) 550 16 343 1030 1227 785 600 0.48
620 4 318 1030 1273 766 450 0.30
620 16 296 1030 1304 810 800 0.45
changes in hardness and case depth are smaller from extraneous loading locally exceed the
than for the 4140 steel. It was found that low- value of the fatigue limit, and, in accordance
ering the tempering temperature raises the fati- with curve 3 in Fig. 8, material decohesion must
gue resistance of the core-case system. Shorter occur.
nitriding time, following a low tempering tem-
perature, does not affect the volume fatigue
resistance (Ref 6). Fatigue Property Characteristics after
The effect of heat treatment of the core on Carbonitriding
fatigue resistance is shown in Fig. 8. From the
illustration, it is seen that raising the core hard- In most modern methods of manufacturing, it
ness moves the fatigue resistance characteristic, is recommended that the design stage consider
that is, the distribution of the speed limit value, the different manufacturing technologies. In
across the section in the direction of higher stress connection with that, there is an urgent need to
values. determine material characteristics, especially of
Data to determine the characteristics in Fig. 8 materials after the application of modern tech-
are shown in Table 4. To calculate the fatigue nological property-enhancing treatments. There
limit, the following formula was used: also exists the need to develop calculation
methods of fatigue resistance after thermal and
s1 =1:98H70:0011HVi 2 (Eq 4) thermochemical treatment. This, however, is the
next stage of activity and is possible to carry out
where HV is Vickers surface hardness, and HVi only when the basic fatigue properties of the
is Vickers hardness at i location on the cross steel following heat treatment are known.
section. The relationship is valid for a hardness The carbonitriding treatment is used for
range of 340jHVj900. components exposed to lighter loads and sub-
A significant increase in the fatigue limit jected to wear as well as bending (Ref 810). For
value (to 820 MPa) with a tempering tempera- those components, which are subjected to con-
ture of 550 C and up to 735 MPa with a tem- tact fatigue during service, the case depths are
pering temperature of 620 C, relative to prior designed deeper. For the present series of tests, a
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 compres-
sive 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 Fig. 7 Microhardness traverses across a nitrided case on
40HM (4140)-grade steel. 1, tempering temperature 38HMJ (Nitralloy 135M)-grade steel. 1, tempering
550 C, time 4h; 2, tempering temperature 550 C, time 16 h; 3, temperature 550 C, time 4 h; 2, tempering temperature 550 C,
tempering temperature 620 C, time 4 h; 4, tempering tem- time, 16 h; 3, tempering temperature 620 C, time 4 h; 4, tem-
perature 620 C, time 16 h pering temperature 620 C, time 16 h
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 Theoretical
results results
Tempering temperature
Fatigue limit (s 1),
MPa 550 C 620 C 550 C 620 C
Core 613 550 618 550
At surface ... ... 852 819
At location of initiation 820 735 618 550
of fracture
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). Fig. 9. Location of fatigue crack initiation on nitrided 40HM
Specimens prepared to meet the aforemen- (4140)-grade steel. Original magnification: 100
tioned conditions were subjected to rotational-
bending and one-point bending fatigue tests.
Such types of loading were selected based on the the most favorable and strongest effect of sur-
premise that bending is the most common face strengthening.
method of loading during service, as well as the Simplified Smith curves were plotted to
fact that it is during bending that one can observe determine the fatigue resistance for at least three
methods of bending (the Wohler curve), as well ferritic-pearlitic microstructure. The chemical
as static strength and yield strength for a given composition is given in Table 5. Tests were also
type of loading and materials. The fatigue tests carried out on carbonitrided specimens. Carbo-
were carried out with the following coefficients nitriding of specimens from 18HGT-grade steel
of cycle asymmetry: R = 1, R = 0.1, and R = was carried out at a temperature of 860 C in
0.3. The coefficients of 0.1 and 0.3 were selected an endothermic atmosphere enriched with am-
to ensure the possibility of running the tests only monia and natural gas.
within the range of one-sided bending stresses. Metallurgical evaluations were carried out on
The fatigue characteristic was developed for a 18HGT steel in the quenched and tempered only
material in the quenched and tempered condition and carbonitrided condition. In the normalized-
as a reference and, for materials with a diffusion only condition, the specimens showed a ferritic-
case, heat treated to the same condition as the pearlitic microstructure with very fine-grained
heat-treated-only version. Once these values pearlite (Fig. 11). The microstructure of speci-
were known, the surface coefficient of strength- mens with diffusion cases was determined based
ening was determined from the equation: on micrographs and microhardness measure-
ments.
sww
1 Specimens made from 18HGT steel, after
m= (Eq 5)
s1 carbonitriding and quenching and tempering,
exhibit a microstructure of tempered martensite
ww in the subsurface zone (Fig. 12) and a bainitic-
where s 1 is the fatigue limit of the enhanced
specimen, and s 1 is the fatigue limit of the martensitic microstructure in the core (Fig. 13).
reference specimen. The microstructure of the core was 550 HV0.1.
Rotational-bending fatigue tests were carried To determine the fatigue resistance, a static
out on the Schenck fatigue machine with a bending test was carried out. For the heat treated
constant distribution of the bending movement
along the length of the specimen. The frequency
was 100 Hz. Table 5 Chemical composition of specimens
One-point bending tests were carried out on prepared from steel grade 18HGT
the Amsler machine. In order to obtain the Specimen Chemical composition, %
No./diameter,
bending effect on this machine, a prototype mm C S Mn Cr Si Ni Cu Ti
addition was designed that, through a lever, 1 12 0.187 0.014 0.89 1.07 0.26 0.06 0.13 0.05
allows loading of the tested section of the speci- 2 12 0.185 0.015 0.87 1.12 0.28 0.06 0.12 0.05
3 12 0.189 0.014 0.89 1.12 0.32 0.06 0.12 0.05
men under a constant bending moment (Fig. 10). 1 14 0.216 0.009 1.00 1.00 0.35 0.13 0.13 0.07
The frequency was 150 Hz. Tests were carried 2 14 0.225 0.009 1.02 1.02 0.34 0.13 0.12 0.08
out to NG = 107 cycles. 3 14 0.215 0.009 0.98 1.02 0.33 0.12 0.14 0.08
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 Fig. 11 Microstructure of heat-treated-only 18HGT-grade
rotational-bending mode steel. Etched with nital. Original magnification: 500
(normalized)-only steel, it was not possible to In order to obtain a full characteristic of the
obtain a fatigue resistance value because of the properties of surface-enhanced materials and of
ductility of the material. Only the yield strength the reference version, a static tensile test was
was determined, and it amounted to 822.8 MPa performed. The average value of tensile strength
on specimens. Rm for specimens without a diffusion case was
The static bending test carried out on carbo- 599.8 MPa, and yield strength was determined
nitrided specimens is shown in Fig. 14. In this as Re = 430 MPa. Similarly, as in the case of the
case, it was not possible to determine the yield technological bending test, it was not possible to
strength, and only the relative bending strength determine the yield strength for the carboni-
was established as 2289.8 MPa. trided 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 carbo-
nitriding, 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. 14 Plot of static bending test of carbonitrided 18HGT-

grade specimen
Fig. 13 Microstructure of core of specimen made of carbo-

nitrided 18HGT-grade steel. Etched by nital. Orig- Fig. 15 Plot of static tensile test of carbonitrided 18HGT-
inal magnification: 500 grade specimen
bending, simplified Smith curves were plotted the characteristics of the steel achieved after
for the heat-treated-only material (Fig. 16) and thermochemical treatment, will be able to
for the carbonitrided material (Fig. 17). To plot reduce the amount of material used and to
the chart, values of unlimited fatigue resistance choose an optimal technological version, thus
were used from the Wohler curves. The upper ensuring high parameters and longer life of the
limit of the chart for the heat-treated-only designed component.
material is the yield strength obtained from the
bending test.
The Smith curve for materials after thermo- Summary
chemical treatment differs from that for the
reference heat-treated-only material, because its Testing of properties of structural materials
upper limit is determined by the bending yield is essential for development and manufacturing
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 fati-
gue 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 com-
parison 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
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 Normalizing Carburizing Carbonitriding
Yield strength 430 ... ...
(Re), MPa
Tensile strength 599.8 1302.3 1303.2
(Rm), MPa
Relative elongation 30.45 ... ...
(A5), %
Necking (Z), % 70.06 ... ...
Bending yield strength 822.8 ... ...
(Rg0.2), MPa
Bending strength (Rg), ... 2604.1 2889.8
MPa
Bending fatigue limit 358.0 745.0 887.1
(s 1), MPa
Surface-strengthening 1 2.08 2.48
Fig. 16 Simplified Smith plot for 18HGT-grade steel after coefficient, m
normalizing
as well as for the practice of design and the subsequent nitriding temperature. Initial
construction of machines. A large number of hardening by quenching and tempering is cri-
assessment criteria for materials properties, as tical to core hardness and to the properties of the
well as methods and testing equipment, stem nitrided case and affects the fatigue resistance of
from the multitude of different mechanisms for the material after nitriding.
failure of structural materials from which com- An analysis of fractured surfaces of nitrided
ponents, assemblies, and whole machines are specimens, exposed to service in conditions of
made. Among the most important of these are rotational bending, revealed that the weakest
failures due to the action of static loading. There location on the specimen cross section is the
is, however, an entire spectrum of different types zone of transition of the nitrided case to core.
of dynamic loading, as well as volume fatigue, The method of designing surface cases enables
impact loading, contact stresses, destruction by an explanation of the root cause of fatigue crack
wear of mating surfaces in friction, and others initiation under the nitrided case. The fatigue
that affect the performance of engineering limit in the cross section of the specimen was
components. described as a function of microhardness and
Processes connected with the calculation and residual stresses. The initiation of fatigue cracks
design of machine components and assemblies takes place in the location where stresses from
call for a database of materials properties. extraneous sources exceed the value of the
Modern computational methods and their con- fatigue limit, which is obviously in agreement
stant development force the necessity of deter- with conditions of strength. The favorable effect
mining an ever growing number of parameters of core hardness on fatigue resistance was
that describe the properties of materials. A good observed.
example of this is the attempt to describe the The investigations showed good fatigue
mechanism of material failure through the properties of cases obtained by carbonitriding,
coefficient of stress intensity (KIC), both when as well as a lowering of ductility of these cases.
the failure takes place under variable loads Results of investigations, presented in the form
(fatigue) as well as when it occurs due to wear in of Smith plots, confirm the necessity of further
processes involving friction. Computation of the pursuing investigations in this field. In designing
life of machine components in conditions of practice, the application of obtained results is the
variable loading calls for information not only least expensive method of lowering material
about the value of the fatigue limit but also about consumption and enhancing the life of machine
the angle of inclination of the straight line in the components.
range of limited fatigue strength, and also about It was established that fatigue resistance is
the parameters of the bend point of the fatigue significantly affected not only by compressive
curve. stresses but also by tensile stresses.
Modern structural material does not need to Among the parameters describing the state of
be homogeneous throughout its section. A great residual stresses, the distribution of stresses was
number of steels, plastics, and other metallic of more significance, followed by the value of
materials call for enhancement of the surface, the residual stresses.
due to the constant quest for decreasing material Models of surface layers described in litera-
and energy consumption as well as increasing ture are difficult to implement in industrial
properties. Surface layers are, in the majority of practice. There appears to be a need for the
cases, superficially hardened layers formed by creation of such a model of the surface layer.
thermal and thermochemical treatment or other This could be described by parameters that can
enhancement technologies, such as surface work be used in strength calculations and that would
hardening and anticorrosion coatings. allow its application in instances of different
Testing of materials properties after heat types of extraneous loading, depending on the
treatment shows that the achievement of desired type of service of the component. This model
service properties is connected with the appro- could become the basis for predicting the state of
priate selection of parameters not only in the the surface layer, based on required usable
final thermochemical treatment but also in the properties of machine components. Work on
prior volume heat treatment. In the case of such a model is carried out in two directions:
nitriding of machine components, this technol-
ogy is usually preceded by quenching and tem- Based on experimental description of the
pering at a minimum temperature 20 C above state of the surface layer through hardness
traverses, residual-stress distributions, and Therefore, the development of failure criteria,

distributions of element concentrations, for taking into account the joint effect of an accu-
example, carbon and nitrogen mulation of damage due to the working of
Based on a description of the material alternating loads, wear by friction, and the action
through the theory of elasticity and solving of corrosion, is a very important task, because
constitutive equations by numerical methods the determination of the criteria for failure will
enable proper selection of surface layers for the
Surface layer design for the criterion of fati- given service condition.
gue 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 REFERENCES
be characterized by required life and reliability,
featuring a sufficient life between overhauls, 1. A. Nakonieczny, Podwyzszenie wytrzy-
depending on the type of service, while at the maosci zmeczeniowej czesci maszyn
same time fulfilling the requirements for ecol- przez obrobke cieplna i powierzchniowa
ogy and ergonomics. Such parameters should obrobke plastyczna (Enhancement of Fa-
be attained concurrently with a reduction of tigue Strength through Heat Treatment and
material and energy consumption during manu- Surface Work Hardening), Proc. XXIV
facture and service. This task may be achieved Seminar IMP, XI on Metallurgy and Heat
only when modern computational methods are Treatment, Oct 2324, 1984 (Warsaw), IMP
implemented along with modern technology and (translated from Polish)
proper service conditions at each stage of the 2. T. Babul, A. Nakonieczny, and J. Taci-
product life, that is, study phase, design, manu- kowski, Wpyw umocnienia podoza na
facture, service, and recycling. wytrzymaosc zmeczeniowa azotowanej
The implemented computational methods stali 40HM (The Effect of Core Strength-
enable the design of products according to ening on the Fatigue Resistance of Nitrided
strength criteria, somewhat less often according 40HM Grade Steel), Proc. III Polish Sci-
to tribological criteria, and least often pertaining entific Conference on Surface Treatment,
to corrosion. Czestochowa-Kule, Politechnika Lodz,
Contemporary machine components and 1996 (translated from Polish).
assemblies are subjected in service to the joint 3. A. Nakonieczny and J. Tacikowski, Analiza
interaction of strength, tribological, and corro- pekania zmeczeniowego stali azotowanych
sion hazards. On the other hand, the implemented (An Analysis of Fatigue Fracturing of
computational methods enable the design of Nitrided Steels), Proc. First Polish
products with one selected mode of failure. Scientific Conference on Modern Tech-
In the construction of machine components, nology in Surface Engineering, Sept
there are many parts (crankshafts, threaded 1994, Politechnika Lodz (translated from
joints, springs) that are concurrently exposed to Polish)
different types of failure hazards during service: 4. A. Nakonieczny, The Effect of Residual
mechanical, tribological, or corrosive. Similar Stresses and Hardness on Fatigue Behavior
elements of construction (bridges, masts, cables, of Surface Treated Materials, Proc. MAT-
earth-moving and mining machines) are ex- TEC 91, Technology Transfer Series,
posed to concurrent hazards of fatigue-type A. Niku-Lari, Ed., 1991
stresses and corrosion. 5. B. Winderlich, Das Konzept der lokalen
Classical strength or tribological calculations Deuerfestigkeit und seine Anwendung auf
do not take into account the factor of time. martensitische Randschichten, in benson-
During service, due to the processes of fatigue, dere Laserhartungsschichten, Mater. wiss.
tribological, or corrosive deterioration, there Werkst. tech., Vol 21, 1990, p 378389 (in
occurs a change in the properties of the system German)
being evaluated. Tribological and corrosive 6. V.P. Kogaev, N.A. Machutov, and A.P.
processes cause a change in the geometry and Gusenkov, Rascety detalej masin i kon-
surface condition of the component. This, in turn, strukcij na procnost i dolgovecnost, Masi-
causes a change in the state of stresses in working nostroenie, Moskva, 1985, p 150182
systems, affecting their life and reliability. (in Polish)
7. J. Tacikowski and A. Nakonieczny, Report 10. J. Wyszkowski, Nowoczesne tendencje w

114,01,0163, IMP, Warsaw, 1992 zakresie naweglania i wegloazotowania
8. A. Nakonieczny, Dissertation, Russian gazowego, (Modern Trends in the Field
Academy of Science, Moscow, 1991 of Gas Carburizing and Carbonitriding),
9. W. Olszanski, More Important Problems IMP, Warsaw, 1974 (translated from Polish)
Pertaining to the Austenitic Carbonitriding 11. W. Olszanski, I. Sukowski, J. Tacikowski,
Process, XVI Seminar, VII on Metallurgy and J. Zysk, Obrobka cieplno-chemiczna,
and Heat Treatment Book 2, IMP, Warsaw, (Thermochemical Treatment) Book 5, IMP,
1977 Warsaw, 1979
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 composition of that alloy. The transformation
heat treating processes, because it is so closely products formed are typically illustrated with the
related to dimensional control requirements and use of transformation diagrams that show the
control of residual stresses. Quenching is often temperature-time dependence of the micro-
attributed to many distortion and cracking pro- structure formation process for the alloy being
blems, whether the quenching process is the studied. Two of the most commonly used trans-
actual root cause or not. Approximately 20% of formation diagrams are the time-temperature
the problems in heat treating relate to heating transformation and continuous cooling trans-
processes, while as much as 80% of the pro- formation diagrams.
blems relate to cooling processes. This chapter Time-temperature transformation (TTT)
provides an overview of the fundamental diagrams, also called isothermal transforma-
material- and process-related parameters of tion diagrams, are developed by heating small
quenching on residual stress, distortion control, samples of steel to the austenite transformation
and cracking. This overview is followed by temperature, followed by rapid cooling to a tem-
various selected case histories of failures perature intermediate between the austenitizing
attributed to the quenching process. and the martensite start (Ms) temperature, and
then holding for a fixed period of time until the
transformation is complete, at which point the
Phase Transformations During Heating transformation products are determined. This is
and Quenching done repeatedly until a TTT diagram is con-
structed, such as that shown for an unalloyed
Properties such as hardness, strength, duc- steel (AISI 1045) in Fig. 1 (Ref 1). The TTT
tility, and toughness are dependent on the diagrams can only be read along the isotherms.
microstructural products that are present in steel. Continuous Cooling Transformation Dia-
Typically, the first step in the transformation grams. Alternatively, a given steel may be
process is to heat the steel to its austenitizing continuously cooled from the austenitizing tem-
temperature. The austenitized steel is then perature at different specified rates. The pro-
cooled rapidly to avoid the formation of pearlite, portion of transformation products formed after
which is a relatively soft transformation product, cooling to various temperatures intermediate
and to maximize formation of martensite, a between the austenitizing temperature and the
relatively hard transformation product, and to Ms temperature is used to construct a continuous
achieve the desired as-quenched hardness. cooling transformation (CCT) diagram, such as
The most common transformation products the one shown for an unalloyed carbon steel
that may be formed in quench-hardenable steels (AISI 1045) in Fig. 2 (Ref 1). The CCT curves
from austenite are, in order of formation with provide data on the temperatures for each phase
decreasing cooling rate, martensite, bainite, transformation, the amount of transformation
pearlite, ferrite, and cementite. The formation of product obtained for a given cooling rate with
these products and the proportions of each are time, and the cooling rate necessary to obtain
dependent on the time and temperature cooling martensite. The critical cooling rate is dictated
history of the particular alloy and the elemental by the time required to avoid formation of
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
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 tetra-
gonal. 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 transfor-
mation curve to maximize the martensite trans-
formation 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) aus-
tenite to a more densely packed body-centered
cubic structure of ferrite. At faster cooling rates,
the formation of ferrite is suppressed, and mar-
tensite, which is an even less densely packed
body-centered tetragonal structure than auste-
nite, is formed. Illustrations of these crystal
structures are provided in Fig. 3 (Ref 1). This re-
sults in a volumetric expansion at the Ms tem-
perature as shown in Fig. 4 (Ref 1).
Figure 5 shows that the crystal lattice of aus-
tenite expands with increasing carbon content
(Ref 2). It has been reported that typically when
a carbide-ferrite mixture is converted to mar- Fig. 4 Steel expansion and contraction upon heating and
tensite, the resulting expansion due to increasing cooling. Source: Ref 1
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 quenching of steel, and the potential micro-
increase in size when austenite is converted structural transformations that are possible for a
to martensite is approximately 0.36 mm/mm given steel are illustrated by their CCT or TTT
(0.014 in./in.) for eutectoid compositions. This diagrams. Furthermore, dimensional changes de-
illustrates the effect of carbon structure and steel pend on carbon content and the microstructural
transformation on residual stresses and distor- transformation product formed. Table 1 sum-
tion leading to dimensional changes. marizes the atomic volumes of various micro-
Estimation of Volumetric Change due to structural components as a function of carbon
Steel Transformation upon Quenching. content (Ref 3). Table 2 provides an estimate of
Various microstructures are possible upon volumetric changes as a function of carbon
content for various metallurgical transforma- one-dimensional heating or cooling could be

tions (Ref 4, 5). estimated from (Ref 7):
Thelning reported that volumetric expansion
occurring as a result of quenching could be s=E e=E 1/3 (DV=V)=E a DT (Eq 2)
estimated from (Ref 6):
where E (modulus of elasticity) = 2 105
DV=V 100=(100 Vc Va ) 1:68C N/mm2 and a (coefficient of thermal expan-
+Va ( 4:64+2:21C) (Eq 1)
sion) = 1.2 10 5. Relative volume changes
due to phase transformation are illustrated in
Fig. 6 (Ref 7).
where (DV/V) 100 equals the percentage
Kunitake and Susigawa (Ref 8) reported that
change in volume, Vc equals the percentage by
volume of undissolved cementite, (100 Vc the tendency for cracking decreases as the start
of the martensite transformation temperature
Va) equals the percentage by volume of mar-
(Ms) increases. The Ms temperature was ap-
tensite, Va equals the percentage by volume of
austenite, and C equals the percentage by weight proximated from:
of carbon dissolved in austenite and martensite.
Ms ( C)=521 353C 225Si 24:3Mn
Berns reported that if the value of (DV/V) is
known or can be computed, internal stresses 27:4Ni 17:7Cr 25:8Mo (Eq 3)
that are developed in a part due to tempera-
ture differences (DT) arising from either 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 Volumetric change
Pearlite?austenite 4.64+2.21 C(a)
Austenite?martensite 4.64 0.53 C(a)
Austenite?acicular lower bainite 4.64 1.43 C(a)
Austenite?feathered upper bainite 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 3
Apparent atomic volume, A
Ferrite 11.789
Cementite 12.769
Ferrite+carbides 11.786+0.163 C(a)
Pearlite 11.916
Austenite 11.401+0.329 C(a)
Martensite 11.789+0.370 C(a)
(a) Percent carbon. Source: Ref 3 Fig. 6 Specific volume (DV/V) of carbon steels relative to
room temperature. Source: Ref 7
correlation between grain size and quench A summary of the Ms and Mf values for some
cracking, as shown in Fig. 8 (Ref 8). common steels is provided in Table 3.
Kunitake and Sisigawa (Ref 8) developed a The correlation between cracking sensitivity
relationship to interrelate the combined effect and the transformation temperature range is due
of both carbon content and elemental composi- in part to the low Mf caused by high-carbon steels
tion on cracking propensity. This was designated (which expand more) and to the fact that wide
as the carbon equivalent (Ceq), and it is calcu- transformation ranges may result in cracking of
lated by: the brittle untempered martensite formed at
higher temperatures in the transformation range.
Ceq =C+Mn=5+Mo=5+Cr=10+Ni=10 (Eq 4) Fujio et al. (Ref 10) showed that the vol-
umetric expansion caused by martensite for-
Figure 9 shows a good correlation between mation can be estimated from the maximum
the carbon equivalent and steel cracking. In cooling rate in a particular type of steel, as
general, steels are classified as crack sensitive shown in Fig. 10. Similar correlations were
if the Ceq value is greater than 0.52 to 0.55% evaluated for both cooling time and cooling rate
(Ref 8). at the Ms temperature. However, these correla-
Another measure of cracking tendency is the tions were dependent on the cross-sectional
difference in the start and finish temperatures size and thus could not be used for gears or
of martensite formation (Ms Mf) (Ref 9). other parts with complex shapes. Volumetric
expansion can be estimated for various cross-
sectional sizes by a correlation between the
volume fraction of martensite versus the cooling
Fig. 7 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
Austenitizing
AISI No. temperature, C Ms, C Mf , C
1065 815 275 150
1090 885 215 80
1335 845 340 230
3140 845 330 225
4130 870 375 290
4140 845 340 220
4340 845 290 165
4640 845 340 255
5140 845 330 240
8630 870 365 280
8695 845 135 ...
Fig. 8 Relationship between quench cracking frequency and 9442 860 325 15
austenitic grain size. Source: Ref 8
rate at the Ms temperature for the steel (Fig. 11). be used to predict dimensional changes caused
It is well known that retained austenite by the total carbon concentration in the marten-
can substantially affect distortion. Geller and sitic transformation product and the amount of
Brimene (Ref 11) published a nomogram that can retained austenite. Steel chemical compositions
Fig. 10 Relationship between maximum cooling rate and volumetric fraction of martensite. Source: Ref 10
Fig. 11 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 transfor-
mation. Source: Ref 10
Table 4 Steel chemical compositions listed in Fig. 12

Russian
steel Composition, wt%
desig-
nation(a) C Si Mn Cr Ni Mo V Ti W Co Cu S P
U8 0.750.84 0.170.33 0.170.33 0.15 max ... ... ... ... ... ... ... 0.03 0.03
max max
KhVG 0.901.05 0.100.40 0.81.10 0.901.20 ... ... ... ... 1.201.60 ... 0.30 0.03 0.03
max max max
ShKh15 0.951.05 0.170.37 0.200.40 1.301.65 0.30 ... ... ... ... ... 0.25 0.02 0.03
max max max max
7KhG2VM 0.680.76 0.200.40 1.802.30 1.501.80 ... 0.500.80 0.100.25 ... 0.50.9 ... 0.30 0.03 0.03
max max max
7KhG3V 0.680.76 0.200.40 3.03.5 1.501.80 ... ... 0.100.25 ... 0.50.9 ... ... ... ...
Kh12M 1.451.65 1.101.40 0.150.45 11.012.5 ... 0.400.60 0.150.30 ... ... ... ... 0.03 0.03
max max
4Kh5V2FS 0.350.45 0.801.20 0.150.40 4.505.50 ... ... 0.600.90 ... 1.602.20 ... 0.30 0.03 0.03
max max max
3Kh2V8F 0.300.40 0.150.40 0.150.40 2.202.70 ... ... ... ... 7.508.50 ... 0.30 0.03 0.03
max max max
4Ch5W2FS 0.350.40 0.801.20 0.150.40 4.505.50 0.35 0.30 max 0.600.90 0.03 1.602.20 ... 0.30 0.03 0.03
max max max max max
3Ch2W8F 0.300.40 0.150.40 0.150.40 2.202.70 0.35 0.50 max 0.200.50 0.03 7.808.50 ... 0.03 0.03 0.03
max max max max max
R14M7K25 1.0 ... ... ... ... 7.0 ... ... 14.0 25.0 ... ... ...
N18K9M5 1.0 ... ... ... ... 5.0 ... ... ... 9.0 ... ... ...
(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 Volume changes due to transformational
will assist in interpreting the nomogram shown in phase changes will be contained as residual-
Fig. 12: stress systems until the yield strength is
exceeded.
The shaded line represents zero distortion.
Steels with martensitic carbon contents and When parts are heated during heat treatment,
retained austenite levels falling on the line a thermal gradient exists across the cross section
will exhibit essentially no distortion. of the component. If a section is heated so that a
Martensitic steels with carbon contents and portion of the component becomes hotter than
retained austenite levels that fall bellow the the surrounding material, the hotter material
shaded line will exhibit shrinkage upon expands and occupies a greater volume than the
quenching. adjacent material and will thus be exposed to
Martensitic steels with carbon contents and applied stresses that will cause a shape change
retained austenite levels that fall above the when they exceed material strength. These
shaded line will exhibit expansion. movements can be related to heating rate and
section thickness of the component.
This nomogram was developed for various
Volume Changes During Phase Transfor-
construction and tool steels. Therefore, it should
mations. When a steel part is heated, it trans-
be used with caution for other steel grades (e.g.,
forms to austenite with an accompanying
high-speed tool steels).
reduction in volume. When it is quenched, the
Basic Distortion Mechanism. Shape and
structure transforms from austenite to marten-
volume changes occurring during heating and
site, and its volume increases. If these volume
cooling can be attributed to three fundamental
changes cause stresses to be set up that are con-
causes (Ref 12):
strained within the strength of the material, a
Residual stresses will cause shape change residual-stress system is created. If the stresses
when they exceed the yield strength of the cannot be contained, material movement will
material. occur, which will cause cracking under extreme
Stresses caused by differential expansion conditions. The expansion is related to the com-
due to thermal gradients will increase with position of the steel. Figure 13 shows the relative
the thermal gradient and cause plastic volume increase of two steels as a function of
deformation as the yield strength is ex- austenitizing temperature and specimen dimen-
ceeded. sions (Ref 13).
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
While each of these phenomena is a well- compared the distortion obtained with quench-
known physical change, the situation is made ing of a 0.4% medium-carbon plain steel bar of
more complex when all three events occur 200 mm diameter by 500 mm long in water or
simultaneously. In addition, other events, such as oil from 680 C (Ref 15, 16). The results, shown
heating rate, quenching, and inconsistent mate- in Fig. 16(a and c), show essentially equivalent
rial composition, further complicate the process. variation in diameter and length with both
Relief of Residual Stresses. If a part has cooling processes, which was due to thermal
locked-in residual stresses, these stresses can be strains within the steel. Interestingly, the well-
relieved by heating the part until the locked- known diameter variations at the end of the bar,
in stresses exceed the strength of the material. known as the end effect, were observed, which
A typical stress-strain curve obtained from a is attributable to heat extraction from both the
tension test is shown in Fig. 14 (Ref 12). Initial sides and ends of the bar (Ref 1).
changes in shape are elastic, but under increased If the same steel bars of the same dimensions
stress, they occur in the plastic zone and are are heated to 850 C to austenitize the steel and
permanent. Upon heating, the stresses are gra- then are quenched in water or oil, the results
dually relieved by changes in the shape of the shown in Fig. 16(b and d), respectively, are
part due to plastic flow. This is a continuous obtained (Ref 15, 16). Considerably greater
process, and as the temperature of the part is dimensional variation and lengthening of the bar
increased, the material yield stress decreases, as (for the oil quench) was obtained due to both
shown in Fig. 15 (Ref 14). It is a function not thermal and transformational strains within the
only of temperature but also of time, since the steel.
material will creep under lower applied stresses. Thuvander and Melander modeled the
It is apparent that the stresses can never be dimensional changes of a 70 mm steel (0.15%
reduced to zero, because the material will always C, 1% Mn, 0.75% Cr, 0.85% Ni) cube after
possess some level of yield strength below which austenitizing and then quenching in water and
residual stresses cannot be reduced. 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
Effect of Materials and Quench Process is greater when quenched in water than when
Design on Distortion quenched in oil (Ref 1).
Various factors may affect distortion and
Quenchant selection and quenching condi- growth of steel during heat treating. These
tions are critically important parameters in include component design, steel grade and
quench system design. For example, one study 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 Fig. 15 Variation of yield strength with temperature for three
obtained from a tension test. Source: Ref 12 generic classes of steel. Source: Ref 14
component design. Poor component design materials with sufficient hardenability should be
promotes distortion and cracking by accentuat- oil or salt quenched.
ing nonuniform and nonsymmetrical heat Design symmetry is also an important vari-
transfer during quenching. Component design able to minimize distortion. For example, the
characteristics that are common to distortion and unsymmetrical gear design shown in Fig. 18(a)
cracking problems include (Ref 19, 20): may typically undergo distortion, as shown in
Fig. 18(b) (Ref 19). (The load on a gear tooth
Parts that are long (L) with thin (d) cross increases by the 4.3 power of the taper, Ref 19).
sections. Long and thin parts are defined as The solution to the gear design problem shown
greater than L = 5d for water quenching, in Fig. 18 is to provide greater symmetry, as
L = 8d for oil quenching, and L = 10d for shown in Fig. 19. If this is not possible, press
austempering, where L is the length of the quenching or tooth-by-tooth induction hard-
parts, and d is the thickness or diameter. ening may be the only solutions (Ref 19, 20).
Parts that possess large cross-sectional areas Another common design problem is parts with
(A) and are thin (t), which are defined as holes, deep keyways, and grooves. One illus-
A = 50t tration of this problem is hardening of a shaft
Parts that exceed these dimensions must often with a lubrication cross hole, as illustrated in
be straightened or press quenched to maintain Fig. 20 (Ref 19). Preferred alternative designs
dimensional stability (Ref 20). If possible, are also shown in Fig. 20. If a radial cross hole is
Fig. 16 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
mandatory, the use of a carburized steel with oil the slot, caused by vaporization of the quench-
quenching would be preferred. ant. Therefore, upon cooling, the metal on the
Kern and Suess have reported that the size of side with the shaft is too short, pulling the shaft
the tapped holes can be maintained by the out of alignment. A general rule for solving such
insertion of SAE grade 8 set screws or bolts quench distortion problems is that the short side
(Ref 21). Prequenching can be used to control is the hot side, which means that the inside of the
the taper of plain holes during heat treatment. bowed metal was quenched more slowly than
Some hole distortion problems may require oil the opposite side (Ref 20).
quenching (conventional or hot oil) or aus- Flat plates are also susceptible to distortion
tempering. upon quenching. If the material is flat and stress
The distortion encountered when quenching a free, round or nearly square, and free of
notched part, such as a shaft with a milled slot, is decarburization, Kern and Suess have reported a
illustrated in Fig. 21 (Ref 20). In this case, guide (Table 5) to maintain a flat surface (within
nonuniform heat transfer results. The metal 0.025 mm, or 0.001/in., of size) if parts are
within the notch is affected by the shrinkage of racked and quenched edgewise (Ref 21). Parts
the metal around it due to slower cooling within 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. 20 Design solutions to the quench-cracking problem

Fig. 18 Schematic of a gear that is difficult to harden without often encountered in shaft hardening over a cross
the distortion shown. Source: Ref 19 hole. Source: Ref 19
Steel Grade and Condition. Although series with the 8600 series of steels (Ref 20).
quench cracking is most often due to nonuni- Dirty steels, those containing greater than
form heating and cooling, material problems 0.05% S, such as 1141 and 1144, are more
may also be encountered. Some typical material prone to cracking. The reasons include:
considerations include (Ref 19): greater alloy segregation in dirty steels leads
to alloy-rich and alloy-lean regions; there are
The compositional tolerances should be typically more surface seams that act as
checked to assure that the alloy is within stress raisers with dirty steels; and steels
specification. with higher sulfur levels are often manu-
Some alloys are particularly problematic. factured to coarse-grain practice for im-
For example, some steel grades must be proved machinability, which also imparts
water quenched when the alloy composition greater brittleness and propensity for crack-
is on the low side of the specification limit. ing.
Conversely, if the alloy composition is on Decarburization of up to 0.064 mm/
the high side, cracking is more common. 1.59 mm (0.0025 in./1/16 in.) diameter may
Steel grades that exhibit this problem
be present.
include 1040, 1045, 1536, 1541, 1137, 1141,
and 1144. As a rule, steels with carbon It is well known that cracking propensity
contents and hardenability greater than 1037 increases with carbon content. Therefore, the
are difficult to water quench (Ref 19). carbon content of the steel is one of the deter-
Some steel grades with high manganese are mining factors for quenchant selection. Table 6
prone to microsegregation of manganese and summarizes some steel mean carbon content
gross segregation of chromium and are concentration limits for water, brine, or caustic
prone to cracking. These include 1340, 1345, quenching (Ref 22).
1536, 1541, 4140, and 4150. If possible, it is Regions containing high concentrations of
often a good choice to replace the 4100 coarse carbide microstructure as a result of
improper forging may become the initiation
point for subsequent quench cracking, particu-
larly with parts of complex shape (Ref 23). It is
important to provide a sufficient forging reduc-
tion ratio to allow the carbide formation to
become fine and uniform (Ref 24).
Since part manufacture, such as gear pro-
duction, often requires machining, the condition
of the steel that is going to be machined is
critically important. Some workers have re-
commended that normalized and subcritical-
annealed steel is the ideal condition (Ref 18).
Subcritical annealing is performed to relieve
stresses incurred during normalization without
softening or homogenizing the steel. The sub-
critical annealing process reduces the carbon
content and alloy carbide content in the auste-
nite, allowing the production of more lath
Table 6 Suggested carbon content limits for

Fig. 21 Distortion often encountered when quenching a
water, brine, and caustic quenching
notch. Source: Ref 20
Hardening method/shapes Carbon, max%
Furnace hardening
Table 5 Guide to maintain a flat surface General use 0.30
Simple shapes 0.35
Ratio (max) of Very simple shapes, e.g., bars 0.40
Quenchant perimeter/thickness
Induction hardening
Water 30 Simple shapes 0.50
Oil 80 Complex shapes 0.33
Austemper 125
Gas 150 Source: Ref 22
martensite in the microstructure, which provides dimensional classification of various distortion-

higher fracture toughness and higher impact sensitive shapes is provided in Fig. 23, Ref 18).
toughness (Ref 23). Proper support when heating is required to
Machining. Material removal during minimize out-of-flatness and ovality problems,
machining can result in high residual-stress which may result in long grinding times,
levels and ultimately unacceptable distortion excessive stock removal, high scrap losses, and
(Ref 18). When excessive machining stresses are loss of case depth (Ref 18). To achieve adequate
imparted, the process may require modification distortion control, custom supports or press
to include a rough machining, then stress quenching may be required.
relieving, followed by fine machining. Pinion shafts, as defined in Fig. 23, are sus-
Component Support and Loading. Many ceptible to banding along their length if they are
parts may sag and creep under their own weight improperly loaded into the furnace, as shown in
when heat treated, which is an important cause Fig. 24 (Ref 18). When this occurs, the pinion
of distortion. An example of a component that is shafts must then be straightened, which will add
susceptible to such distortion is a ring gear. Di- to the production cost.
mension limits by which ring gears are classified Surface Condition. Quench cracking may
are provided in Fig. 22 (Ref 18). (A general 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 struc-
ture, 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 dis-
Fig. 22 Dimensions of a ring gear shape. Shape limitation:
cussed: tight scale formation, decarburization,
length/wall thickness, j1.5; inside diameter/outside
diameter (ID/OD), 40.4. Minimum wall thickness
p (WT) is

and the formation of surface seams or non-
defined by: WT i2.25 module+0.4 mod L OD3 . metallic stringers.
5
Source: Ref 18 Tight scale problems are encountered with

forgings hardened from direct-fired gas furnaces
Fig. 23 Classification of shapes. Source: Ref 18

with high-pressure burners (Ref 20, 22). The a heavy scale (0.13 mm or 0.005 in. deep)
effect of tight scale on the quenching properties retards the cooling rate. A very light scale
of two steels, 1095 carbon steel and 18-8 (0.013 mm or 0.0005 in. deep) also increased
stainless steel, is illustrated in Fig. 25 (Ref 23). the cooling rate of the 18-8 steel over that
These cooling curves were obtained by still obtained with the specimen without scale.
quenching into fast oil. A scale of not more than In practice, the formation of tight scale will
0.08 mm (0.003 in.) increases the rate of cool- vary in depth over the surface of the part,
ing of 1095 steel as compared to the rate resulting in thermal gradients due to differences
obtained on a specimen without scale. However, in cooling rates. This problem may yield soft
spots and uncontrolled distortion and is parti-
cularly 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 de-
carburization, which may lead to increased dis-
tortion 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 de-
carburization. This leads to nonuniform hard-
ness, which may contribute to increased
distortion and cracking because (Ref 20):
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
Fig. 24 Typical pinion shaft distortion due to furnace load-
products to form early, nucleating additional
ing. Source: Ref 18
undesirable products in the core. The
Fig. 25 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
decarburized side will be softer than the side part surface, which is heated primarily by
that did not undergo decarburizing, which is radiation. Thus, as the hotter surface tries to ex-
harder. The greater amount of martensite pand, it will be restrained by the cooler steel,
leads to distortion.The solution to this pro- leading to a hot upsetting condition and possibly
blem is to restore carbon into the furnace significant distortion even if quenched uni-
atmosphere or machine off the decarburized formly (Ref 21). A similar condition exists if the
layer. tray of gears is placed near radiant tube heaters
or electric heating elements in the furnace wall
The third surface-related condition that may and the remainder of the gear surface is heated
lead to cracking or material weakening is the by radiation from the roof of the furnace.
formation of surface seams or nonmetallic Localized overheating is particularly a
inclusions, which may occur in hot rolled or cold potential problem for inductively heated parts
finished material. The presence of these defects (Ref 4, 26). Subsequent quenching of the part
prevent the hot steel from welding to itself dur- leads to quench cracks at sharp corners and areas
ing the forging process, for example, creating a with sudden changes in cross-sectional area
stress raiser. To prevent this problem with hot (stress raisers). Cracking is due to increases of
rolled bars, stock should be removed before heat residual stresses at the stress raisers during the
treatment. Recommendations made earlier by quenching process. The solution to the problem
Kern are provided in Table 7 (Ref 22). is to increase the heating speed by increasing the
Although not a published standard, Kern has power density of the inductor. The temperature
reported that a seam or nonmetallic depth of difference across the heated zone is decreased by
0.025 mm/3.3 mm (0.001 in./0.13 in.) diameter continuous heating or scanning of several pis-
maximum is usually acceptable for cold finished tons together on a single bar (Ref 26).
bars (Ref 22). If the seam depth is excessive, it is For heat treating problems related to furnace
recommended that the bars be magniflux in- design and operation, it is usually suggested that
spected prior to heat treatment. (Ref 19):
Heating and Atmosphere Control. An
important source of steel distortion and cracking The vestibules of atmosphere-hardening
is nonuniform heating and not using the appro- furnaces should be loaded and unloaded with
priate protective atmosphere. For example, if purging. Load transfer for belt and shaker
steel is heated in a direct-gas-fired furnace with hearth furnaces should only occur with
high moisture content, the load being heated thorough purging to minimize atmosphere
may adsorb hydrogen, leading to hydrogen contamination.
embrittlement and subsequent cracking that If the load being heated in the furnace is
would not normally occur with a dry atmosphere excessively large, either nonuniform heating
(Ref 19, 26). over the specified heating cycle or nonuni-
One source of distortion is when a part is in form cooling when quenched will result. In
contact with the furnace hearth during heating, such cases, either the production rate can be
which may produce sufficient nonuniform tem- increased or smaller loads should be pro-
perature distribution within the part. This will cessed.
occur because the portion of the part in contact
with the furnace hearth will be heated con- Retained Austenite. Dimensional changes
ductively much faster than the remainder of the may occur slowly or quickly and are due to the
volume composition of the transformation pro-
ducts formed upon quenching. One of the most
Table 7 Minimum recommended material important, with respect to residual-stress varia-
removal from hot rolled steel products to prevent tion, distortion, and cracking, is the formation
surface seam and nonmetallic stringer problems and transformation of retained austenite. For
during heat treatment example, the data in Table 8 illustrate the slow
Minimum material removal per side(a) conversion of retained austenite to martensite,
Condition Nonresulfurized Resulfurized which was still occurring days after the original
Turned on centers 3% of diameter 3.8% of diameter quenching process for the two steels shown (Ref
Centerless turned or ground 2.6% 3.4% 15, 16). This is particularly a problem when
(a) Based on bars purchased to special straightness, i.e., 3.3 mm in 0.04 m dimensional control and stability is one of the
(0.13 in. in 5 ft) maximum. Source: Ref 22
primary goals of heat treatment. Therefore,
microstructural determination is an essential additional general comments regarding quen-

component of any distortion control process. chant selection include (Ref 4, 20):
Most machined parts made from alloy steels
Quenching Process are oil quenched to minimize distortion.
Most small parts or finish-ground larger
Other than component design, the quenching
parts are free quenched. Larger gears, typi-
process itself is one of the most frequently
cally those over 20 cm (8 in.), are fixture
encountered problems in heat treating. When
(die) quenched to control distortion. Smaller
designing a quenching process, it is important to
gears and parts, such as bushings, are usually
consider quenchant selection, quench severity,
plug quenched on a splined plug typically
and quench uniformity.
constructed from carburized 8620 steel.
Quenchant Selection and Severity.
Although a reduction of quench severity
Quench severity is defined as the ability of a
leads to reduced distortion, it may also be ac-
quenching medium to extract heat from a hot
companied by undesirable microstructures,
steel workpiece; expressed in terms of the
such as the formation of upper bainite
Grossmann number (H) (Ref 27). A typical
(quenched pearlite) with carburized parts.
range of Grossmann H-values (numbers) for
Quench speed may be reduced by quenching
commonly used quench media is provided in
in hot (150 to 205 C, or 300 to 400 F) oil.
Table 9. Figure 26 provides a correlation be-
tween the H-value and the ability to harden steel, When hot oil quenching is used for carbur-
as indicated by the Jominy distance (J-distance) ized steels, lower bainite, which exhibits
(Ref 20). Although Table 9 is useful to obtain a properties similar to martensite, is formed.
relative measure of the quench severity offered
by different quench media, it is difficult to apply
in practice, because the actual flow rates for Table 9 Typical quenching conditions and
moderate, good, strong, and violent Grossmann H-values
agitation are unknown. Alternatively, the mea- Quenching medium Grossmann H-value
surement of actual cooling rates or heat fluxes Poor (slow) oil quenchno agitation 0.20
Good oil quenchmoderate agitation 0.35
provided by a specific quenching medium does Very good oil quenchgood agitation 0.50
provide a quantitative meaning to the quench Strong oil quenchviolent agitation 0.70
severity provided. Some illustrative values are Poor water quenchno agitation 1.00
Very good water quenchstrong agitation 1.50
provided in Table 10 (Ref 28). Brine quenchno agitation 2.00
Typically, the greater the quench severity, the Brine quenchviolent agitation 5.00
Ideal quench ...
greater the propensity of a given quenching
medium to cause increased distortion or crack- Note: It is possible with high-pressure impingement to achieve H-values greater
than 5.00.
ing. This usually is the result of increased
thermal stress, not transformational stresses.
Specific recommendations for quench media
selection used with various steel alloys is pro-
vided by standards such as AMS 2759. Some
Table 8 Dimensional variation in hardened

high-carbon steel with time at ambient
temperature
Change in length after time,
Tempering Hard-
% 103
temperature, ness,
Steel type C HRC 7 days 30 days 90 days 365 days
1.1% C None 66 9.0 18.0 27.0 40.0
tool steel, 120 65 0.2 0.6 1.1 1.9
790 C 205 63 0.0 0.2 0.3 0.7
quench 260 61.5 0.0 0.2 0.3 0.3
1% C/Cr, None 64 1.0 4.2 8.2 11.0
840 C 120 65 0.3 0.5 0.7 0.6
quench 205 62 0.0 0.1 0.1 0.1
260 60 0.0 0.1 0.1 0.1 Fig. 26 Quench severity in terms of Grossmann (H) values.
J, Jominy distance. Source: Ref 20
Excellent distortion is typically obtained can be water quenched if they are in the form
with austempering, quenching into a med- of simple shapes and have no drilled or
ium just above the Ms temperature. The punched holes.
formation of retained austenite is a signific-
Other guidelines of Kern and Suess for hard-
ant problem with austempering processes.
ening steels include (Ref 21):
Retained austenite is most pronounced
where manganese and nickel are major If the part has widely varying section sizes
components. The best steels for austemper- (ratio of 3 to 1), or if it has holes, keyways, or
ing are plain carbon and chromium and grooves, water quenching may produce
molybdenum alloy steels (Ref 20). cracking regardless of the carbon content.
Aqueous polymer quenchants may often be Designing with generous fillets in these
used to replace quench oils, but quench regions may resolve the problem.
severity is still of primary importance. If the distortion must be as low as possible,
Gas or air quenching will provide the least oil or salt quenching should be used with
distortion and may be used if the steel has appropriate qualification. More recent work
sufficient hardenability to provide the has shown that polymer quenching may be
desired properties. used in some cases.
Low-hardenability steels are quenched in If 100% bainite is required, austempering in
brine or vigorously agitated oil. However, molten salt should be performed. To assure
even with a severe quench, undesirable that no retained austenite remains, a final
microstructures, such as ferrite, pearlite, or temper slightly below the austempering
bainite, can form. temperature is recommended.
Kern and Suess have provided guidelines for Quenchant Uniformity. Quench nonuni-
hardening steels to achieve optimal micro- formity is one of the greatest contributors to
structural control (Ref 21). To minimize the quench cracking. Quench nonuniformity can
potential for cracking: arise from nonuniform flow fields around the
In carbon or alloy steels containing50.3% C part surface during the quench or nonuniform
maximum, use a water quench. wetting of the surface (Ref 20, 2932). Both lead
Steels with 0.3 to 0.38% C can be water to nonuniform heat transfer during quenching.
quenched if they are in the form of simple Nonuniform quenching creates large thermal
shapes such as round bars. gradients between the core and the surface of the
If the carbon content is 40.38%, an oil part.
quench should be used. (More current When there is nonuniform cooling within the
references, such as AMS 2759, would permit part between the Ms and Mf, there will be a
polymer quenching of some alloys if stretching or elongation in areas where the
appropriate quench bath maintenance pro- cooling is slow, which will act as a push stress,
cedures are used.) leading to push cracking (Ref 33, 34). Another
Exceptions are carbon steels with low alloy form of cracking is pull cracking, which occurs
content (maximum of 1% Mn). Carbon steels with nonuniform surface cooling between the
containing40.95% C and 0.30 to 0.50% Mn austenitizing temperature and Ms. Push cracking
and pull cracking are the opposite of each other,
although the cracking event takes place for both
Table 10 Comparison of typical heat-transfer between the Ms and Mf. Figure 27 provides illus-
rates trations of both push and pull cracking (Ref 32).
Quench medium Heat-transfer rate, W m 2
K 1
Poor agitation design is a major source
Still air 5080 of quench nonuniformity. The purpose of the
Nitrogen (1 bar) 100150 agitation system is not only to take hot fluid
Salt bath or fluidized bed 350500
Nitrogen (10 bar) 400500 away from the surface and to the heat exchanger,
Helium (10 bar) 550600 but it also provides uniform heat removal over
Helium (20 bar) 9001000 the entire cooling surface of all of the parts
Still oil 10001500
Hydrogen (20 bar) 12501350 throughout the load being quenched. The
Circulated oil 18002200 batch quench system in Fig. 28 illustrates a
Hydrogen (40 bar) 21002300
Circulated water 30003500
system where axial (vertical) quenchant flow
occurs throughout a load of round bars lying
horizontally in a basket (Ref 19). In this case, the the opposite direction (Ref 31). The solution
bottom surfaces of the bars experience greater to this problem was to minimize the quenchant
agitation than the top surfaces. Cracks form on flow to that required for adequate heat transfer
the upper surfaces because of the nonuniform during the quench and to provide agitation by
heat loss. Agitation produces greater heat loss at mechanically moving the part up and down in the
the bottom, creating a large thermal gradient quenchant. Identifying sources of nonuniform
between the top and the bottom surfaces. fluid flow during quenching continues to be an
If a submerged spray manifold is used to important tool for optimizing distortion control
facilitate more uniform heat removal, the fol- and minimizing quench cracking.
lowing design guidelines are recommended: Nonuniform thermal gradients during
quenching are also related to interfacial wetting
The total surface of the part should experi- kinematics, which are of particular interest with
ence uniform quenchant impingement. vaporizable liquid quenchants, including water,
The largest holes possible (2.3 mm or oil, and aqueous polymer solutions (Ref 32).
0.09 in. minimum) should be used. Most liquid vaporizable quenchants exhibit
The manifold face should be at least 13 mm boiling temperatures between 100 and 300 C at
(0.5 in.) from the surface of the parts being
atmospheric pressure. When parts are quenched
quenched.
in these fluids, surface wetting is usually time-
Repeated removal of hot quenchant and
dependent, which influences the cooling process
vapor should be possible.
and the achievable hardness.
Excessive distortion was also obtained with an Another major source of nonuniform
agitation system illustrated in Fig. 29 when the quenching is foaming and contamination. Con-
quenchant flow was either in the same direction taminants include sludge, carbon, and other
relative to the direction of part immersion or in 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
Not all quench failures occur immediately

following the quench; some failures that occur
during subsequent use may be due to unac-
ceptably high and/or nonuniform stresses that
Fig. 27 Two forms of quench cracking. Source: Ref 32 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. 28 Harmful effects of impeded vertical quenchant flow

through the load of a batch quench system. Source: Fig. 29 Effect of quenchant flow direction on distortion.
Ref 19 Source: Ref 31
the volumetric expansion that accompanies the When cracked parts are subsequently tem-
martensitic transformation. It is directly pro- pered, the intergranular morphology may form a
portional to carbon content and microstructural thick oxide scale from the tempering process.
factors. These cracks can be instantaneous upon The microstructure adjacent to the crack will not
quenching, or they may be delayed. Also, some be decarburized unless a specimen with an
components may be crack-free, whereas see- undetected quench crack is rehardened. In
mingly identical components may have cracked. quenched and tempered steels, proof of quench
Delayed quench cracks can be the result of cracking is often obtained by opening the crack
additional transformation of retained austenite in and looking (visually) for temper color typical
steel. This occurs when heavily stressed retained for the temperature used.
austenite continues to transform to martensite The following are some case studies showing
prior to tempering or even after tempering, if sources of cracking that are often attributed to
there is sufficient retained austenite. However, the quench but whose root cause is not the
as mentioned previously, surface damage and quench itself. The quench only exacerbates the
inadequate microstructure (decarburizing, band- problem. There is only one example (case
ing, inclusions, coarse grain size), among others, study 1) where the cracking root cause was the
may also cause the part to fail (Ref 3335). quench severity.
Quench cracking typically initiates at the
surface, particularly at positions where geome-
trical changes occur, such as at corners, defects,
and inclusions. Quench cracks always begin at Case Studies in Quench Cracking
the part surface and have characteristics that are
easily recognized. First, the fracture generally Case Study 1: As-Quenched 4340 Steel
runs from the surface toward the center of the (Ref 25, 36). A component (Fig. 30) from AISI
mass in a relatively straight line, with either a 4340 steel cracked during heat treatment. Che-
longitudinal or radial orientation unless located mical analysis of the component confirmed a
by a change in section size. The crack is also composition compatible with AISI 4340 steel.
likely to open or spread and may exhibit a shear As shown in Fig. 30, the crack passes straight
lip. Shear lips are ledges on the side of the spec- from the surface to the core. Quench cracks
imen that make a 45 angle to the plane of frac- always begin at the part surface and have charac-
ture and may be present on the edges of some teristics that are easily recognized. First, the
predominantly brittle fractures to form a picture fracture generally runs from the surface toward
frame around the surface (Ref 35). The fracture
surfaces of quench cracks are always inter-
granular. It is common to find secondary crack-
ing, 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, inclu-
sions, 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 Fig. 30 Macrograph of AISI 4340 quenched and tempered
steel illustrating macroetched pure quench crack.
service. Source: Ref 25, 36
the center of the mass in a relatively straight line, With respect to the case being discussed here,
with either a longitudinal or radial orientation the cause of cracking was identified as due to
unless located by a change in section size excessively high quench severity.
(Ref 35). Case Study 2: Cracking of 4140 Block
The component steel is considered a high- Forging after Quenching and Tempering.
hardenability steel because of alloying elements Cracking was observed to occur with an AISI
such as chromium, nickel, and molybdenum. It 4140 block forging subsequent to quenching and
is recommended that when using high-hard- tempering. Chemical analysis was performed on
enability steels, quenchants exhibiting lower the steel block and compared to the specification
quench severity and time should be used. This is range for this alloy (Table 11), which confirmed
important because cracking can occur during that the steel was nominally 4140.
quenching due to thermal contraction stresses To verify the presence or absence of inclu-
coupled with the volumetric expansion that sions (quantity, morphology, and distribution), a
accompanies the martensitic transformation. As metallographic examination in the unetched
already described, excessive cooling rates (high condition was performed. In this condition,
quench severity) will produce greater thermal although the microstructure is not revealed, it is
stresses in addition to greater transformation easier to identify inclusions. The steel was
stresses. If the total residual stresses in the part examined near the crack, and the results are
exceed the yield strength of the steel, distortion shown in Fig. 31(a). No evidence of nonmetallic
will occur. If the ultimate strength is exceeded, inclusions was found that could be attributed to
cracking will occur (Ref 4, 5). the observed crack formation.
Table 11 Chemical analysis

Chemical composition, wt%
Material C Mn Si P S Cr Ni Mo
AISI 4140 0.380.43 0.751.00 0.150.35 0.035 0.040 0.801.10 0.25 0.150.25
Block 0.39 0.88 0.15 0.013 0.028 0.86 0.06 0.16
Fig. 31 Representative view of the surface and crack profile from the block sample. (a) Unetched condition. (b) Etched with 2% nital.

Material C Mn Si P S Cr Ni Mo Cu
AISI 4140 0.380.43 0.751.00 0.150.35 0.035 0.040 0.801.10 0.25 0.150.25 ...
AISI 4130 0.280.33 0.400.60 0.150.35 0.035 0.040 0.801.10 0.25 0.150.25 0.35
Component 0.40 0.88 0.25 0.025 0.031 1.03 0.09 0.22 0.25
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 con-
clude 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
Fig. 32 Representative view of the crack propagating from
4130 steel shaft was brazed to an AISI 4130 steel
porosity or voids within the brazed joint. Unetched.
Original magnification: 100 body. After quenching and tempering, cracking
was observed at the brazed joint, which propa-
gated into the reamer body.
The surface was also examined, and no evi- The nominal range of chemical compositions
dence was observed of detrimental surface for AISI 4130 and 4140 steel are provided in
conditions, such as small cracks or defects due to Table 12. Chemical analysis of the component,
machining that could contribute to the cracking. which was thought to be AISI 4130, is also
This is important because quenching involves shown, which confirms that the wrong steel alloy
high levels of thermal and transformation was used. The chemical analysis of the compo-
stresses, and the presence of imperfections in the nent is consistent with that for AISI 4140.
microstructure can increase the risk of cracking The higher carbon content of AISI 4140
the part. Imperfections act like very small relative to AISI 4130 means greater hard-
cracks. Inclusions and other surface defects enability and therefore greater probability for
function as stress raisers. cracking and increased distortion to occur. In
In Fig. 31, it is possible to see evidence of such cases, quenching should be less severe.
secondary cracking. It is common to find sec- From these data, it would appear that the AISI
ondary cracking that forms around the main 4130 steel reamer body was most likely exposed
crack, indicating that the component was under to an excessively high cooling rate for this steel
high stress due to thermal contraction stresses alloy if heat treatment parameters were set for
coupled with the volumetric expansion that 4130 steel.
accompanies the martensitic transformation. To verify the presence or absence of inclu-
Metallographic examination in the etched sions (quantity, morphology, and distribution), a
condition is necessary to verify other micro- metallographic examination in the unetched
structural characteristics. For steels, the most condition was made. In this case, the examina-
common etchant is 2% nital (2 mL HNO3 + tion was also made near the crack, and the results
98 mL ethanol, 95%). Figure 31(b) shows the are shown in Fig. 32.
same region of the Fig. 31(a) but in the etched Figure 32 did not reveal evidence of non-
condition. metallic inclusions, although it is possible to
Examination of the crack profiles revealed no observe the presence of voids in the brazed joint.
evidence of decarburization in the crack, These are undesirable and should be avoided,
although tempering oxide was observed. The since voids are stress raisers by amplifying the
stress during quenching and are nucleation sites Case Study 4: Presence of a Seam Defect
for crack formation in the brazed joints. This (Ref 25, 36). An AISI 4140 steel bar cracked
occurs during fast cooling, leading to increased after austenitization and quenching in a 20 to
thermal stresses in the component. Figure 32 21% aqueous polymer quenchant solution.
shows that cracking initiated from porosity or Chemical analysis confirmed that the steel used
voids within the brazed joint and appears to have was compatible with AISI 4140. Metallographic
propagated into the reamer body from quench- examination in the unetched condition was
ing stresses. Secondary cracks are also observed. performed on the steel near the crack, which is
Examination of the brazed joint in the etched shown in Fig. 34. No evidence of nonmetallic
condition, using 2% nital (2 mL HNO3 +98 mL inclusions was observed. However, examination
ethanol, 95%), revealed the predominance of of the defect profile revealed the presence of
tempered martensite and a uniform micro- seam defects. A seam defect is an unbounded
structure (Fig. 33). fold or lap on the surface of the metal that
From these analyses, it can be concluded that appears as a crack and is usually the result of a
the presence of voids (stress raisers) within the seam that was formed but not closed during the
brazed joint has nucleated cracks that propa- working process (such as rolling, forging, etc.)
gated under quenching stresses. The incorrect of the material. These defects typically exhibit
steel grade increased the potential for cracking. 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
Fig. 33 Representative view of the brazed joint between the
obtained with an unthreaded AISI 4140 stud bolt
reamer body (bottom) and reamer shaft (top).
Microstructure is tempered martensite. Etched with 2% nital. (25.4 mm diameter).
Fig. 35 Aspect of the defect in the etched condition. In the

cross section, it is possible to see a uniform micro-
structure compounded by martensite. Decarburizing and high-
Fig. 34 View of the identified seam defects in a bar sample of temperature oxidation can be observed. Etched with 2% nital.
AISI 4140. Unetched. Original magnification: 100 Original magnification: 100

AISI 4140 0.380.43 0.751.00 0.150.35 0.035 0.040 0.801.10 0.25 0.150.25 0.35
Sample 0.39 0.81 0.22 0.014 0.027 0.92 0.15 0.18 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.
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 evi-
dent, 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
Fig. 37 Representative view of the crack surface profile. can increase the risk of cracking. Imperfections
Unetched. Original magnification: 100 such as inclusions and other surface defects act
as stress raisers.
Figure 38 shows the steel in the etched con-
Chemical analysis of this steel is shown in dition (2% nital), which reveals a uniform tem-
Table 13, together with the nominal composition pered martensite microstructure with a slag
range of AISI 4140 steel. These data confirm that inclusion stress raiser.
the steel is consistent with AISI 4140. From those observations, it can be concluded
Metallography results for the steel in the that cracking was caused by quenching stresses
unetched condition are shown in Fig. 36(a acting upon stress-concentration sites of large
slag inclusions adjacent to the crack and a small cracking appeared to be intergranular, showing
seam or lap defect at the surface. Cracks were tempering oxide within the profile. No evidence
also observed propagating from and/or through of decarburization was observed. As noted
an apparent surface seam or lap-type defect. before, when cracked parts are subsequently
Case Study 6: Presence of Chemical Seg- tempered, the intergranular morphology may
regation (Ref 25, 36). A press-formed steel form a thick oxide scale from the tempering
flange made from AISI 1035 steel produced process.
cracks after quenching in an aqueous polymer Quenching stresses associated with the non-
quenchant solution. Chemical composition for uniform microstructure, caused, in part, from a
this steel grade is provided in Table 14. Com- slack quench condition or inherent chemical
parison with the actual composition of the steel segregation, have contributed to the observed
from the flange confirmed that it was the correct cracking. Slack quenching is related to hard-
AISI 1035 steel grade. ening of steel from the austenitizing temperature
Cracking was observed at the press-formed at a rate slower than the critical cooling rate,
ring location. Examinations (unetched) of the resulting in incomplete transformation and the
press-formed ring location were performed. formation of one or more transformation pro-
Figure 39, shows that in the unetched condition, ducts in addition to or instead of martensite.
there is no evidence of surface imperfections or Case Study 7: Network Carbides and
inclusions that could be attributed to the crack- Coarse Grain Size. A low-alloy 17CrNiMo6
ing of the steel. (0.18% C, 0.25% Si, 0.50% Mn, 1.65% Cr,
However, examinations in the etched condi- 0.80% Mo, 1.55% Ni) carburized steel gear
tion revealed a microstructure that exhibited produced cracks after carburizing, quenching,
chemical segregation in the form of banding, as and tempering. The carburized case was 1.8 to
shown in Fig. 40(a). The microstructure is non- 2.0 mm, and the measured surface hardness was
uniform and consists of bainite and tempered 57 to 61 HRC. After carburizing, the gear was
martensite, as illustrated in Fig. 40(b). quenched from 840 C using an aqueous poly-
Examination of the outer radius of the press- mer quenchant and subsequently double tem-
formed ring revealed evidence of cracking. The pered 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 inter-
granularly (Ref 37). Quench cracking is con-
sidered 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
Fig. 38 Etched condition showing tempered martensite
microstructure and slag inclusion. Etched with 2% is heavily influenced by volumetric expansion
nital. Original magnification: 100 during transformation hardening and the

AISI 1035 0.320.38 0.600.90 ... 0.04 0.050 ... ... ... ...
Flange 0.33 0.73 0.18 0.012 0.003 0.04 0.02 50.01 50.01
temperature extremes of quenching. Causes of The microstructure of the carburized case was
intergranular brittle fracture include brittle carefully examined, which also showed the
second-phase particles and/or films in grain presence of brittle network carbides in the prior-
boundaries (Ref 37). austenitic 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 super-
saturated with carbon, that is, it contains carbon
in excess of the eutectoid composition (0.8% C),
the carbide will precipitate at the grain bound-
aries during slow cooling from the carburizing
temperature. Under equilibrium cooling condi-
tions, 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
Fig. 39 Representative view of the cracking associated with
the radius of the press-formed ring. Unetched. Orig- retained by the resultant martensite-austenite
inal magnification: 100 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 inter-
granular cracking. Etched with 2% nital
Carbides, being ceramic compounds, are Alloying elements such as aluminum, niobium,
brittle. The more accelerated the diffusion of vanadium, or titanium function as grain-
carbon, the coarser the carbides become. Over- growth inhibitors, and the steel grade used
sized carbides do not resist thermal fluctuations, in this case did not contain these elements
because they are ceramic compounds; therefore, (Ref 38, 39).
cracking initiates due to thermal shock or fast Therefore, it was concluded that the presence
ramp-up and ramp-down of a furnace. Because of network carbides and nonuniform grain size,
carbides in the network configuration possess a coupled with quenching stresses, was respon-
brittle nature, once a carbide segment of a net- sible for the observed cracking.
work starts cracking, crack propagation is very Case Study 8: Presence of Stringer Inclu-
fast (Ref 40). sions and Chemical Segregation (Ref 25).
During carburizing, it is necessary to carefully Pins of AISI 1144 steel (resulfurized steel grade)
control the carbon content of the surface case. If were through hardened prior to induction hard-
this layer becomes hypereutectoid, cementite ening of the pin tip, and cracking and soft spots
will be present in the boundaries of the grain, were obtained. The nominal chemical compo-
forming a network, as demonstrated in Fig. 42(a sition of AISI 1144 steel is provided in Table 15
and b). along with the chemical analysis of the com-
In addition to network carbides, nonuniform ponent. These results confirm that the steel
grain size (ASTM 1 and 2 ) is also observed in used for the component is consistent with AISI
Fig. 42(a and b). Nonuniform grain size is a 1144.
problem, since the hardenability of a carbon Examinations in the unetched condition
steel may increase as much as 50% with an revealed evidence of many long stringer inclu-
increase in austenite grain size from ASTM 8 sions, which were oriented in streaks or bands
(6 to 10) to ASTM 3 (1 to 4). This phenomenon that were parallel (or longitudinal) with respect
causes a nonuniform martensitic transformation, to the length of the pins. A stringer inclusion
contributing to increased stress during quench- occurs when an impurity, either metallic or
ing. The effect becomes more pronounced if nonmetallic, is trapped in the ingot and is elon-
the carbon content is increased at the same time gated in the direction of hot working. It appears
(Ref 39). These factors contributed to the as a narrow streak that is parallel to the direction
cracking observed during quenching. of hot working (Ref 38). Stringers were also
Excessive carbon content (more than 0.8%) observed extending to the pin tip surface within
in the carburized layer was related to the the induction-hardened case (Fig. 43). Stringers
high carburizing temperature that was used can, like other inclusions, act as stress raisers.
(950 C). Higher temperatures result in greater At positions that may be far removed from those
solubility of the carbon in the austenite defects, the applied stress (applied load/cross-
phase. Grain growth can also be related to high sectional area) may be normal and will not pose
process temperature and steel chemistry. any problem. However, if the applied stress is
Fig. 42 (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

AISI 1144 0.400.48 1.351.65 ... 0.04 0.240.33 ... ... ... ...
Pin 0.44 1.50 0.23 0.008 0.29 0.05 0.02 0.02 50.01
associated chemical segregation in the presence

of quenching stresses.
Case Study 9: Decarburization and Oxi-
dized Grain Boundary (Ref 25, 36). A heavy
wall tube section of AISI 4140 tube stock pro-
duced 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
Fig. 43 Aspect of the stringer inclusions observed within unetched condition revealed no evidence of a
the pin tip location. Unetched. Original magnifica- large number of nonmetallic inclusions. Exam-
tion: 100
ination in the etched condition of the cross-
sectional microstructure shows that it consists of
less than the elastic limit in the vicinity of small uniform tempered martensite (Fig. 45).
defects or cracks, the stress is amplified. There- However, examination of the surface in the
fore, defects such as stringer inclusions are etched condition revealed high-temperature
called stress raisers and are very important dur- grain-boundary oxidation (Fig. 46). High-
ing quenching as well as during use. temperature grain-boundary oxidation occurs
Examination in the etched condition showed when grain boundaries starting at the surface of
that the microstructures consisted of many areas the part are oxidized. Normally, when parts are
of chemical segregation. Unavoidable chemical being heat treated, such as during carburization
segregation of alloying elements occurs during or austenitization (hardening), if the furnace
the solidification of an ingot in the steel pro- contains free oxygen from air leakage (ingres-
duction process. If this occurs on a grain-sized sion) into the furnace or excessive vapor or
scale, it is called microsegregation. If chemical steam is present, oxygen will diffuse into the sur-
segregation occurs on a much larger ingot-sized face of the material, resulting in oxidation of the
scale, it is referred to as macrosegregation, and grain boundaries and degradation of the engi-
the inhomogeneous steel structure will possess neering properties at the surface (Ref 40). High-
nonuniform properties throughout, particularly temperature grain-boundary oxidation also acts
in the direction that is transverse to the hot as a stress-concentration site for crack initiation.
working direction (Ref 38). In the AISI 1144 The surface profile in the etched condition
steel sample being analyzed, the chemical seg- (Fig. 47) revealed evidence of partial decarbur-
regation appeared as ferrite bands and was ization and tempering oxide within the cracks.
associated with the stringer inclusions, which When cracked parts are subsequently tempered,
can be associated with the soft spots that also can the intergranular morphology may form a thick
contribute to cracking (Fig. 44a and b). It should oxide scale from the tempering process.
be noted that stringer inclusions are not un- Decarburization appears when steels are
common for resulfurized steel grades, because processed by forming, forging, heat treating, or
the distribution and shape of these inclusions are any other thermal treatments where the material
often difficult variables to control. temperature may exceed 760 C for some time
The fracture pattern of the pin tips sub- with no atmospheric protection. If this occurs,
sequent to induction hardening appears to have the steel may start losing carbon from the heated
propagated from the stringer inclusions and surfaces, leading to a decarburized surface. This
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

AISI 4140 0.380.43 0.751.00 0.150.35 0.035 0.040 0.801.10 0.25 0.150.25 0.35
Sample 0.44 0.97 0.24 0.012 0.016 1.06 0.12 0.19 0.18
Fig. 45 Cross-sectional microstructure showing uniform Fig. 46 Surface profile adjacent to the cracking. Evidence of
tempered martensite. Etched with 2% nital. Original high-temperature grain-boundary oxidation. Etched
magnification: 400 with 2% nital. Original magnification: 400
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,
ASM Handbook, ASM International, 1991,
p 3334
Fig. 47 View of the crack profile showing tempering oxide 15. W.T. Cook, A Review of Selected Steel-
within the crack and decarburization at the surface. Related Factors Controlling Distortion in
Etched with 2% nital. Original magnification: 50
Heat-Treatable Steels, Heat Treat. Met.,
Vol 26 (No. 2), p 2736
decarburized layer can be partial or full and can 16. Y. Toshioka, Heat Treatment Deformation
degrade the engineering properties of the surface of Steel Products, Mater. Sci. Technol.,
relative to the matrix of the material. Vol 1, Oct 1985, p 18831892
Based on this evidence, the identified crack- 17. A. Thuvander and A. Melander, Calculation
ing was attributed to quenching stresses acting of Distortion During Quenching of a Low
on the oxidized surface grain boundaries. Carbon Steel, First ASM Heat Treatment
and Surface Engineering Conference in
REFERENCES Europe (Proc.), Part 2, Trans. Tech. Pub-
lications, 1992, p 767782
1. Y. Toshioka, Heat Treatment Deformation 18. P.C. Clarke, Close-Tolerance Heat Treat-
of Steel Products, Mater. Sci. Technol., Vol ment of Gears, Heat Treat. Met., Vol 25
1, 1985, p 883892 (No. 3), 1998, p 6164
2. S. Mocarski, Carburizing and Its Control 19. R.F. Kern, Thinking Through to Successful
I. Basic Considerations, Ind. Heat., Vol 41 Heat Treatment, Met. Eng. Q., Vol 11
(No. 5), 1974, p 5870 (No. 1), 1971, p 14
3. A. Bavaro, Heat Treatments and Deforma- 20. R. Kern, Distortion and Cracking, II: Dis-
tion, Trait. Therm., Vol 240, 1990, p 3741 tortion from Quenching, Heat Treat., March
4. F. Legat, Why Does Steel Crack During 1985, p 4145
Quenching, Kovine Zlitine Technol., Vol 32 21. R. F. Kern and M. E. Suess, Steel Selection:
(No. 34), 1998, p 273276 A Guide for Improving Performance and
5. R.W. Bohl, Difficulties and Imperfections Profits, Wiley-Interscience, New York,
Associated with Heat Treated Steel, 1970, p 1550
MEI Course 10, Lesson 13, ASM Inter- 22. R. Kern, Distortion and Cracking, III: How
national to Control Cracking, Heat Treat., April
6. K.E. Thelning, Steel and Its Heat Treat- 1985, p 3842
ment, Butterworths, London, U.K., 1985 23. H.-J. Chen and Z.W. Jiang, Microstructure
7. H. Berns, Distortion and Crack Formation Improvement and Low-Temperature Quen-
by Heat Treatment of Tools, Radex ching of Dies Made of Cr12-Type Steel,
Rundsch., Vol 1, 1989, p 4057 Jinshu Rechuli, No. 8, 1992, p 3941
24. H.-H. Shao, Analysis of the Causes of Distortion Control, G.E. Totten, Ed., ASM
Cracking of a 12% Cr Steel Cold Die During International, 1992, p 275282
Heat Treatment, Jinshu Rechuli, No. 11, 32. H.M. Tensi, A. Stich, and G.E. Totten,
1995, p 43 Fundamentals of Quenching, Met. Heat
25. R.R. Blackwood, L.M. Jarvis, D.G. Hoff- Treat., Mar/April 1995, p 2028
man, and G.E. Totten, Conditions Leading 33. V.D. Kalner and S.A. Yurasov, Internal
to Quench Cracking Other Than Severity of Oxidation During Carburizing, Met. Sci.
Quench, Heat Treating Including the Liu Heat Treat. (USSR), Vol 12 (No. 6), June
Dai Memorial Symposium, Proc. of the 1970, p 451454
18th Conf., R.A. Wallis and H.W. Walton, 34. A.A. Polyakov, Quenching Properties of
Ed., ASM International, 1998, p 575585 Parts Having Stress Concentrators, Met.
26. X. Cheng and S. He, Analysis of Quenching Sci. Heat Treat., Vol 37 (No. 78), 1995,
Cracks in Machine-Tool Pistons Under p 324325
Supersonic Frequency Induction Harden- 35. W. Becker, Fatigue Failure, Failure
ing, Heat Treat. Met. (China), No. 4, 1991, Analysis and Prevention., Vol 11 ASM
p 5152 Handbook, ASM International, 2002,
27. J.R. Davis, ASM Materials Engineering p 700727
Dictionary, ASM International, 1992, p 407 36. L.C.F. Canale, G.E. Totten, R.R. Black-
28. P.F. Stratton, N. Saxena, and R. Jain, Re- wood, L.M. Jarvis, and D.G. Hoffman, An
quirements for Gas Quenching Systems, Overview of Non-Quench Related Pro-
Heat Treat. Met., Vol 24 (No. 3), 1997, blems Often Attributed to the Quenchant
p 6063 and Quenching Process, 59th Congresso
29. H.M. Tensi, G.E. Totten, and G.M. Anual da ABM-Internacional, July 1922,
Webster, Proposal to Monitor Agitation 2004 (Sao Paulo, SP, Brazil)
of Production Quench Tanks, Heat 37. S. Lampman, Intergranular Fracture, Fail-
Treating: Including the 1997 International ure Analysis and Prevention, Vol 11, ASM
Induction Heat Treating SymposiumProc. Handbook, ASM International, 2002,
of the 17th Conf., D.L. Milam, D.A. Poteet, p 641649
G.D. Pfaffmann, V. Rudnev, A. Muehl- 38. Manufacturing Techniques, http://www.
bauer, and W.B. Albert, Ed., ASM Inter- mslab.boun.edu.tr/Heat_treatment.doc (Ac-
national, 1997, p 423441 cessed April 24, 2007)
30. S. Owaku, Quench Distortion of Steel Parts, 39 G. Parrish, The Influence of Microstructure
Netsu Shori (J. Jpn. Soc. Heat Treat.), Vol and the Properties of Case-Carburized
32 (No. 4), 1992, p 198202 Components, ASM International, 1980,
31. R.T. Von Bergen, The Effects of Quenchant p 236
Media Selection on the Distortion of 40. A. Abada, Why Do Heat-Resistant Alloys
Engineered Steel Parts, Quenching and Fail?, Ind. Heat., Oct 2002, p 5559
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 metal- temperatures, resulting in a revision of all for-
smiths was a complex phenomenon for blade- mer data.
smiths. The repeated working, heating, and A very important technological mark was the
cooling could cause embrittlement, with sword development of the beneficial molybdenum
failure occurring in the most critical moments of effect on the embrittlement reduction phenom-
a battle. Likewise, some of the earliest cannons enon, through work by Greaves and Jones
would break apart after the first shots following (Ref 1). For some time, embrittlement due to the
manufacture. These problems occurred in the tempering process has been shown to be an
ferrous alloy application until the benefit of important failure related to heat treatment. In
tempering became recognized. In the middle of this chapter, the causes and cases associated with
the 18th century, the tempering process (and/or problems originated by tempering are reviewed.
stress relief) received attention as a fundamen- However, to provide background on this phe-
tally important process in the heat treatment of nomenon, a brief description of the martensite
the ferrous components of tools. Some ironsmith reactions and the steel heat treatment of tem-
tools were treated by the so-called process of pering is given to review the different stages of
water annealing, whereby steel was tempered in microstructural transformation.
the range of 300 to 600 C. The slow cooling
was substituted by water cooling.
At the beginning of the 20th century, Krupp Martensite
developed a great number of patents based on
water cooling after tempering of chromium- Before describing the solid-state reactions
nickel steel. This phenomenon received atten- resulting from the tempering process in the fer-
tion after the start of WWI, when large amounts rous matrix, it is important to define the mar-
of steel were used by the armament industry. In tensitic structure as a function of the alloying
1917, the term tempering embrittlement was elements, especially for the carbon effect. Fer-
introduced by Dickenson, having been pub- rous martensite is composed of a body-centered
lished in papers by Brarley, Hatfield, Philpot, tetragonal crystallographic structure, with lat-
and Grenet. Some investigators, such as Greves tice parameters (c and a) related to the carbon
and his collaborators, began a set of experi- contents of its chemical composition, as shown
mental methods using notched bars to determine in the expression (Ref 2):
the susceptibility of tempering embrittlement. c=a=1+0:0467 (wt% C)
A relationship between the energy absorbed
after water cooling and annealing was termed The lattice ratio for the tetragonal structure is
the steel susceptibility ratio. At that time, all approximately 1.0467, with 1 wt% C in solid
the experiments were performed at room tem- solution. As shown in Fig. 1, hardness varies
perature, because no one anticipated that tem- with carbon content, and that effect is strongly
perature may also have an effect on the results. related to the distortions caused by the carbon
The effect of test temperature received atten- atom in the body-centered tetragonal structure.
tion in the beginning of 1944, when Jolivet Martensite is extremely hard (maximum of 800
and Vidal introduced experiments at different to 900 HV) and brittle.
The transformation of martensite from aus- Martensite is a nonequilibrium structure and

tenite is a nonequilibrium (athermal) process thus does not appear on the iron-carbon equili-
that occurs during rapid cooling from the aus- brium phase diagram. The face-centered cubic
tenite phase. Unlike the phase transformation lattice shearing of austenite (Bains deforma-
from atomic diffusion at equilibrium tempera- tion) in a martensitic transformation is illu-
tures, the martensitic transformation occurs strated in Fig. 2. The deformation is large and
when many atoms together undergo a shear rapid over many atoms in the lattice, and the
displacement. This rapid shear displacement of change in the polycrystalline system is accom-
atoms results in a rapid change in crystal struc- modated by lattice deformation. This can occur
ture during cooling. Thus, the martensitic by slide, mechanical twinning, or even a mixture
transformation is referred to as a diffusionless of both mechanisms according to the steel che-
process, because the transformation involves a mical composition. The crystal change results in
nonequilibrium (athermal) crystal change from a an expansion of the polycrystalline system.
shear transformation during rapid cooling from The effect of temperature on martensite for-
austenite. mation is directly related to the transformation
temperatures of martensite start (Ms) and mar-
tensite 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 defor-
Fig. 1 Hardness of martensite as a function of carbon content. mation mechanism occuring by dislocation
Source: Ref 3 slide, although the low temperatures provoke a
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
Steel Failures due to Tempering and Isothermal Heat Treatment / 287
plastic deformation by mechanical twinning. Measurements of dislocation density found

This way, it is possible to establish a martensite in martensite are on the order of 0.3 to
morphology with respect to the alloy content, 0.9 1013 cm/cm3 of the crystal. This disloca-
giving special attention to carbon. tion density is higher than the maximum that can
Figures 3 to 5 show the morphological aspects be obtained by elevating the percentage of cold
of martensite as a function of carbon content for plastic deformation. Some observations, made
steels through optic microscopy. In the case of by transmission electron microscope, show a
lath martensite, the deformation mechanism of very small cellular structure (approximately 0.2
the lattice is dislocation slip. This kind of mar- to 0.3 mm) inside the lath (Ref 5).
tensite is also known as slipped martensite. Figure 4 shows the martensite with a high
Morphologically, this martensite presents lath carbon content, observed with an optical
packages, which are separated by low-angle microscope. With a high carbon content, the
boundaries (Ref 5). In the past, it was thought microstructure has twinned martensite or plate
that the transformation units happened as lath martensite. With higher carbon contents, higher
packages, although recently it became clear that volume of retained austenite (Fig. 5) occurs,
each lath is independently formed, and the evi- because a higher carbon content lowers the Ms
dence shows that an austenite film exists, which
can be seen in Fig. 6.
Fig. 3 Lath (low-carbon) martensite in SAE 8620 alloy steel Fig. 5 Microstructure of quenched 1.3% C steel. Dark nee-
(Fe, 0.2% C, 0.8% Mn, 0.55% Ni, 0.5% Cr, 0.2% Mo) dles of plate martensite and white areas of retained
after heat treatment (954 C, or 1750 F, for 1 h, water quench) austenite (white arrow)
Fig. 6 Illustration of the austenite film surrounding martensite

laths in a Fe-10Cr-0.2C steel. Source: Adapted from
Fig. 4 High-carbon martensite (0.8% C). Etched with nital Ref 2
temperature. Twinning density can be seen by martensite. In the first, the deformation has a
transmission electron microscopy, because volumetric character, while in the second, it is
density is high even though twins are very nar- more directional (Ref 7). This behavior can be
row (on the order of 10 A ). seen in Fig. 8.
The percentage of slip martensite and twinned It can be seen in Fig. 8 that the difference
martensite in carbon steel and tool steel can be in the specific volume between austenite and
experimentally determined, as shown in Fig. 7. martensite is approximately 15% from very low
As the carbon content increases, the amount of carbon content to very high carbon content
lath martensite decreases. The untransformed (2% C). It is also interesting to observe that for
austenite increases by the Ms temperature low carbon, the change of volume from an as-
(martensite start temperature) reduction. The annealed condition to an as-hardened condition
increase in the retained austenite volume frac- is practically nil. On the other hand, increasing
tion can reduce the as-quenched hardness carbon content raised that difference.
mainly in the higher content range. These observations are important during the
Martensitic transformation causes an increase component process design. As mentioned ear-
in volume and size variations, which contributes lier, the greater the carbon content, the greater
to the residual tension stresses that develop in the embrittlement of the martensite plates
the surface after the heat treatment of quenching, (Ref 8). However, since retained austenite also
when transformation takes place in all of the increases with carbon content, this fraction of
sample cross sections, and transformation be- retained austenite will bring some toughness to
tween surfaces and nucleus occurs anachroni- the matrix as well as reduce the volume varia-
cally. The volume variation measured during the tion as shown in the curve A-FC in Fig. 8.
transformation from austenite to martensite in a This fact results in a lower load to the cold sur-
1% C steel is approximately 4% (the transfor- face because of the incomplete transformation
mation to pearlite results in a 2.4% expansion) of the nucleus. It is important to remember
(Ref 2), decreasing as far as the carbon is added that the nucleus presents a low yield strength
in the matrix. This occurs because of the dif- when retained austenite has a low mechani-
ferent carbon effect in the austenite related to the cal resistance, decreasing the possibility of
Fig. 7 Effect of carbon content on the lath martensite volume, retained austenite volume fraction, and Ms temperature. Source:
Adapted from Ref 6
developing tension stress in the surface of the properties can be changed when the component
component. is held isothermally at a temperature where
It can be seen in Fig. 8 that for 2% C, theo- austenite cannot form.
retically it could be possible to obtain a complete It is important to emphasize that tempered
transformation to martensite, but, this is not martensite usually does not contain martensite.
the case. However, it is interesting to observe Instead, it is a structure of fine carbide particles
that the retained austenite line matches the aus- in ferrite, which has formed from martensite
tenite line for very high carbon content. This is during the tempering. This structure has a lower
the case for Hadfield steels (Ref 9), which have a hardness than the martensite, but by proper
high manganese content (approximately 12%) choice of temperature and time used, the struc-
that guarantees an austenitic microstructure, ture developed will be fine to give the desired
even though the carbon amount is approximately hardness. Table 1 lists the colors associated with
1.2%. In this situation, it is possible to quench the tempering heats, and Table 2 illustrates the
large components with complex geometry times required to reach furnace temperature
without the risk of developing cracks, even during tempering (Ref 10).
while increasing quenchant severity.
The ability to form martensite is described
in terms of hardenability, which is related to Effect on Mechanical Properties
the presence of other alloy elements besides As noted, martensitic structures are too brittle
carbon. For example, molybdenum and manga- for most practical applications. However, it
nese increase hardenability, while cobalt lowers is possible to enhance the structure tough-
the hardenability of steel. A higher hardenability ness through tempering. The toughness usually
allows martensite formation with a slower comes at the expense of a decrease in yield
cooling rate. This is beneficial for reducing
the tensile residual stresses in the component
surface.
Table 1 Colors of tempering heats
Temperature(a) Temperature(b)
Tempering C F Color of oxides C F
188 370 Faint yellow 238 460
Tempering is historically associated with the 199 390 Light straw 265 510
210 410 Dark straw 293 560
heat treatment of martensite in steels. The 221 430 Brown 321 610
resultant microstructure is called tempered 232 450 Purple 337 640
254 490 Dark blue 349 660
martensite. The main purpose for tempering is to 265 510 Light blue 376 710
develop a usable combination of hardness and (a) Temperature held for 1 h. (b) Temperature held for 8 min
toughness. The microstructure and mechanical
Fig. 8 Specific volume (DV/V) of carbon steels relative to room temperature. Source: Adapted from Ref 7
p
strength and hardness, as illustrated in Fig. 9 Tt =647 [S(60 C+20)=Ht 70:9]1=4 73:45 SHt
and 10. Figure 11 shows other modifications +(5377561S)(%C)+505S(%V)+219S(%Mo)
of mechanical properties that occur when an oil-
quenched AISI 4340 steel is tempered at tem- +75S(%Cr)+66S(%Si)751 [C ]
peratures above 200 C. where Ht is the hardness after hardening and
The interrelationship between tempering tempering (HRC), S is the degree of hardening,
temperature, steel chemistry, and hardness can Sj1.0, and the alloying elements are given in
be estimated by the equation (Ref 13): weight percent. This formula is valid for a
HB=2:84Hh +75(%C) 0:78(%Si) tempering time of 2 h.
+14:24(%Mn)+14:77(%Cr)+128:22(%Mo)
Tempering Reactions
54:0(%V) 0:55T t +435:66
Tempering is a process in which the micro-
where HB is the Brinell hardness after hardening structure approaches equilibrium under the
and tempering, Hh is the Rockwel (HRC) hard- influence of thermal activation. It follows that
ness after hardening, and Tt is the tempering the tendency to temper depends on how far
temperature in C. This equation was developed the starting microstructure deviates from
for the following conditions:
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 tem-
perature ( 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 composi-
tion 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 Heating time, min
Cubes or Squares or Average Cubes or Squares or Average
C F spheres(a) cylinders(a) flats(a) spheres(b) cylinders(b) flats(b)
121 250 30 55 80 15 20 30
149 300 30 50 75 15 20 30
177 350 30 50 70 15 20 30
204 400 25 45 65 15 20 30
260 500 25 40 60 15 20 30
316 600 25 40 55 15 20 30
371 700 20 35 50 15 20 30
427 800 20 30 45 15 20 30
482 900 20 30 40 15 20 30
(a) In hot air oven, without circulation. (b) In circulation air furnace or oil bath (can be used only in lower temperatures)
equilibrium. Martensite microstructure is the Crystallographic adjustment (coherence

farthest, followed by bainite, ferrite, and stresses) between the precipitated phases
cementite (Ref 14). and the ferrous matrix
When the martensitic structure is metastable, Thermodynamic stability of reactions
there is a natural tendency to transform it to a
structure with more stability, and those mod- During tempering, the martensitic structure is
ifications are accelerated by increasing the submitted to a sequence of reactions, often
temperature during the tempering. superimposed and defined as temper stages (Ref
The modification that occurs during temper- 2, 3, 4, 5, 12).
ing is complex, and the transformations that take Stage 1. In high-carbon steels, the pre-
place during the treatment conditions necessary cipitation of excess carbon begins with the for-
to produce the best mechanical properties com- mation of a transition carbide, such as e (Fe2.4C).
bination are a result of accumulated knowledge, The e-carbide can grow at temperatures as low
not just from the academic point of view but also as 50 C. Martensite is said to be supersaturated
the practical aspect of observation. Most of the with carbon when the concentration exceeds its
time, the structures developed during isothermal equilibrium solubility with respect to another
heat treatments are influenced by the low content phase. However, the equilibrium solubility
of other elements besides iron and carbon. depends on the phase. The solubility will be
larger when the martensite is in equilibrium with
Tempering stages a metastable phase such as e-carbide. Approxi-
Solid-state reactions follow a sequence of mately 0.25 wt% C is said to remain in solution
precipitation that is related to variables such as: after the precipitation of e-carbide is completed.
Although most textbooks will begin a discussion
Diffusivity of the involved element of tempering with this first stage of tempering,
Surface energy of interfaces produced by the involving the redistribution of carbon and pre-
reactions cipitation of transition carbides, cementite can
Fig. 10 Effect of carbon content on the hardness of tempered plain steels. Source: Ref 10
precipitate directly (Ref 14). This is the case for martensite of the steels with less than 0.5% C
the lath martensite structure, where the disloca- content has a retained austenite amount lower
tion density can be as high as 1012 to 1013/cm2 than 2%, reaching 6% for 0.8% C. There are
(Ref 15). Trapped carbon atoms will not pre- some indications that austenite decomposes,
cipitate as transition carbides, but cementite is turning into ferrite and cementite, but presently a
more stable than trapped carbon. This stage consensus does not exist about whether this
begins at room temperature and extends to structure can be correlated to lower bainite,
250 C. A fine adjustment between the e- typically from the isothermal decomposition of
carbides and the ferrous structure is attributed to austenite, in the temperature range of 230 to
the precipitation-hardening effect of martensite 300 C.
in high-carbon steels tempered between 50 and Stage 3. Tempering at even higher tem-
100 C. peratures leads to a coarsening of the cementite
Stage 2. Tempering at higher temperatures, particles, with those located at the plate bound-
in the range of 200 to 300 C, for 1 h induces the aries growing at the expense of the intraplate
retained austenite to decompose into a mixture particles. This precipitation is responsible for
of cementite and ferrite. When the austenite is the embrittlement phenomenon observed at the
present as a film, the cementite also precipitates temperature of 250 to 400 C. It can be avoided
as a continuous array of particles that have the by adding silicon, which is an insoluble element
appearance of a film (Ref 3, 5, 12, 14). The in cementite. This allows cementite formation at
Fig. 11 Changes in the mechanical properties of AISI 4340 steel with tempering temperature. Source: Ref 12
temperatures where silicon diffusion occurs, Effect of Temperature and Alloying. The
thus slowing the process. effect of the tempering temperature on steels
Silicon is a substitutional alloy element that with increasing carbon contents can be inferred
accumulates in the ferrous matrix adjacent to the from Fig. 13. During tempering, the continuous
Fe3C carbide in the growth process. It increases decomposition of martensite to ferrite and car-
the carbon activity in this region, decreasing the bides changes the state of stress because of
carbon flow to the growing particle and conse- continuous dimensional changes. At low tem-
quently reducing its growth velocity (Ref 5, 16, peratures (first stage), a volume contraction
17). This silicon effect resulted in the develop- takes place as a consequence of e-carbide pre-
ment of alloy 300M, which substitutes for 4340 cipitation. In the second stage, with the trans-
steel in those applications where it is necessary formation of retained austenite (approximately
to use tempering temperatures that cause 300 C), the volume is increased. In stage 3, the
embrittlement of tempered martensite, which is progressive decomposition of martensite leads
soon defined. This reaction begins to occur at to a volume decrease.
temperatures on the order of 100 C. Cementite It is important to observe that the austeniti-
can also be observed during quenching when the zation temperature, which determines the
Ms temperature is elevated, as is the case of steel amount of carbon dissolved and the amount of
with a low carbon content, mainly in the mar- retained austenite, has a strong influence on the
tensite formed just below the Ms temperature. expected volume changes (Ref 2, 4, 12, 13).
This phenomenon is known as self-tempering Table 3 shows the changes in length for various
(Ref 5). steels as a function of tempering temperature.
Stage 4. In carbon steels, stage 3 marks the Alloyed steels can also have another stage
end of the tempering process. Spheroidization of with the precipitation of alloy carbides, includ-
Fe3C occurs as cementite coalesces. This phe- ing M2C (molybdenum), M7C3, M6C, M23C6
nomenon is sometimes called stage 4 of tem- (chromium rich), V4C3, TiC, and so on, where
pering (Ref 2). The lath boundary maintains the M refers to a combination of metal atoms.
stability up to approximately 600 C. Intense However, all of these carbides require long-
rearrangement occurs between the lath and its range diffusion of substitutional atoms. They
low-angle boundaries above 600 C. This can only precipitate when the combination of
recovery process is replaced by recrystallization time and temperature is sufficient to allow this
and coarsening (Fig. 12) at temperatures be- diffusion. The alloy carbides grow at the
tween 600 and 700 C (Fig. 13). expense of the less stable cementite. If the con-
centration 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 car-
bide, 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
Hardness decreases with increasing temper-
ing temperature (Fig. 10, 11). Consequently,
yield strength and tensile strength decrease as
Fig. 12 Fe-0.17C alloy quenched in water from 900 C and
tempered at 650 C for 5 h. Microstructure shows well. On the other hand, elongation and ductility
ferrite grains and spheroidized Fe3C increase. In this general context, a failure related
to tempering may be attributed to an incorrect brittleness is inherent to some extent in all steels,
choice of temperature (and/or time of temper- including carbon grades. For that reason, med-
ing), resulting in an incorrect hardness or low ium-temperature tempering is, as a rule, not
toughness. employed in practice, although it can ensure a
However, most failures are related to high yield limit. According to Krauss (Ref 12),
embrittlement phenomena. Quenched and tem- TME may or may not be associated with
pered steels are susceptible to a number of dif- impurity atom segregation to prior-austenitic
ferent types of embrittlement. Some of them grain boundaries, but the most common factor,
are due to structural modifications during tem- at least for medium-carbon steels, is the phe-
pering, as previously described. However, nomenon that takes place due to decomposition
there are some due to the interaction of the of retained austenite to cementite in the interlath
environment with the quenched and tempered
microstructures, such as hydrogen embrittle- Table 3 Length variations related to
ment and liquid metal embrittlement. Examples metallurgical reactions as a function of
of the first type of embrittlement are tempered tempering temperature ranges
martensite embrittlement and temper embrittle- Temperature Expansion (E) or
Stage range, C Metallurgical reactions contraction (C)
ment, which are described as follows.
Tempered Martensite Embrittlement. It is 1 0200 Precipitation of e-carbide; C
loss of tetragonality
well known that tempered martensite embrit- 2 200300 Decomposition of retained E
tlement (TME) is related to tempered martensite austenite
3 230350 e-carbides decompose to C
of specimens tempered between 250 and cementite
370 C, as shown in Fig. 16. The impact 4 350700 Precipitation of alloy E
toughness after tempering at this temperature carbides; grain
coarsening
range is lower than that obtained on tempering Source: Ref 18
at temperatures below 250 C. This type of
Fig. 13 Hardness as a function of tempering temperature for plain carbon steels. Source: Ref 6
region of martensite plates along the prior-

austenite 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 tem-
pering, 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 sec-
ond mode of fracture associated with TME is
intergranular. This kind of fracture is quite
common and has been related to phosphorus Fig. 15 Hardness and toughness of a tool steel as a function
of tempering temperature. Charpy V-test performed
segregation to the austenite grain boundary at room temperature in the short-transverse direction
Fig. 14 Effects of titanium, vanadium, chromium, and molybdenum on tempering hardness behavior. Source: Ref 5
during austenitization. However, data on the Investigations using AISI 4140 steels, have
effect of phosphorus on the fracture surfaces in shown that austenitizing temperature has an
the untempered condition indicate only that the influence on the TME phenomenon. High aus-
presence of phosphorus at the prior-austenite tenitizing temperature was found to be asso-
grain boundary is not sufficient for the devel- ciated with more pronounced TME, favoring
opment of TME. So, the interaction between brittle failure modes, even in specimens showing
phosphorus and cementite is necessary for the virtual absence of phosphorus segregation.
intergranular mode of TME (Ref 20, 21). These investigations found that high austenitiz-
It is important to understand that the fracture ing temperatures increase carbide dissolution in
is occurring along a prior-austenite grain austenite, apparently due to more intensive car-
boundary, which is now a high-angle ferrite bide precipitation and growth during tempering
grain boundary (Ref 22). Krauss (Ref 12) (Ref 23).
describes another type of transgranular fracture When TME appears, there is no heat treat-
mode associated with TME, which is observed ment to reverse the effect, other than to reaus-
in 4340-type steels. It is interlath cleavage tenitize and quench the steel, then temper
induced by cracking parallel to the cementite in a temperature range where TME does not
formed from the retained austenite transforma- occur. TME is also called one-step embrittle-
tion. Transgranular fractures in TME may be ment.
related to the interlath carbide thickness; thinner Temper Embrittlement. The second type
carbides cause interlath fracture, and thicker of embrittlement from tempering at high tem-
carbides promote translath cleavage. In some peratures is known as temper embrittlement
low-carbon steels, embrittlement is associated (TE). Temper embrittlement occurs when tem-
with peculiar carbide morphology that provides pering in the high-temperature range of 450 to
numerous sites for microcrack initiation, grow- 600 C. It is not a major problem because it may
ing by microvoid coalescence and then fracture, be avoided simply by quenching from the tem-
with little gross plastic deformation. pering temperature. This embrittlement can be
Silicon additions to carbon steels raise the reversible under high-temperature tempering.
temperature range in which TME occurs (as When steel that has undergone TE is heated
mentioned earlier), because silicon delays both to a temperature above 600 C and then cooled
the conversion of the transition carbide (e) to very quickly, its impact is restored. Therefore,
cementite within the martensite laths as well such brittleness is termed reversible. Temper
as cementite coarsening at boundaries at embrittlement is also called two-step embrittle-
higher tempering temperatures (Ref 16, 17). The ment, because two tempering treatments or a
effect of silicon content on the impact prop- heating step and a cooling step are sometimes
erties of 0.6%C-0.47%Mn-0.52%Cr-1.77%Ni- required to induce embrittlement. Figure 18
0.19%V-1.0%-2.5%Si steels can be seen in shows different situations where embrittlement
Fig. 17 (Ref 16). may or may not appear.
Fig. 16 Illustration of toughness loss after tempering in the embrittlement range. Source: Ref 17
Temper embrittlement of this category of Molybdenum is one of the main alloying

steel takes place due to impurity segregation at elements in many low-alloy steels and is an
the grain boundaries and, finally, decohesion of effective method of alleviating TE. Small addi-
the grain boundary. This leads to intergranular tions of molybdenum (0.2 to 0.3%) can diminish
fracture morphology. It is signaled by a material TE, while greater additions enhance the effect.
toughness loss. It is pointed out that grain- The mechanism of actuation is related to
boundary segregation depends on the alloying molybdenum segregation (equilibrium and non-
elements of the steel. equilibrium) during quenching, and the quen-
Carbon steels with less than 0.5% Mn are not ched-in vacancies play a role in the temper
prone to reversible TE. The phenomenon can process (Ref 28, 29).
only appear in alloy steels. Alloying elements As mentioned earlier, in chromium-
may have a different effect on steel after tem- molybdenum steel, phosphorus is the major
pering at the steel propensity to TE. Unfortu- embrittling element. However, the kinetics
nately, the most widely used alloying elements,
such as chromium, nickel, and manganese,
promote TE. When taken separately, they pro-
duce 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 Fig. 18 Schematic diagram showing thermal cycles leading
to the reversal of temper embrittlement and the de-
brittle after tempering as can be seen in Fig. 19. embrittlement heat treatment
Fig. 17 Charpy tests at room temperature as a function of tempering temperature for 0.6% C steel with different silicon contents.
Source: Adapted from Ref 16
of phosphorus segregation in this steel are dif- phosphorus is free to segregate, and the material
ferent from that in steels due to molybdenum. thus becomes embrittled (Ref 28, 29).
Molybdenum has a strong effect in lowering the Failure analysis of high-temperature studs
solubility of phosphorus in iron. This is inter- concluded that the failure related to TE could be
preted as due to a molybdenum-phosphorus delayed due to the presence of molybdenum and
scavenging reaction; so, if molybdenum is free vanadium. However, reductions of phosphorus
in solution, it appears to either prevent phos- contents in the steel and a grain size of
phorus from segregating to the grain boundaries approximately 10 mm could reduce the tendency
or to reduce the brittleness potency of phos- for brittle fracture (Ref 30).
phorus at the grain boundary. Although molyb- Table 4 shows elements that can segregate to
denum is an effective element to reduce the the former austenite boundaries that are now
susceptibility for TE, the precipitation of ferrite boundaries (Ref 2).
molybdenum as carbide must be taken care of. This segregation was shown in a conclusive
To avoid that, it is observed that vanadium is form through Auger electron microscopy per-
added in this grade of steel. Vanadium is a strong formed on intergranular fracture surfaces. This
carbide former compared to molybdenum and technique allowed the exact determination of
chromium. Vanadium initially forms MC-type atomic specie concentration segregated in con-
carbides; this changes the molybdenum-to- fined boundaries with approximately 10 A of
carbon and chromium-to-carbon ratios. The depth. This fraction varies from 0.3 to 2.0 for
increase in molybdenum-to-carbon ratio is steels when these matrix elements are lower than
favorable for Mo2C-type carbides, and that of 0.1%.
the chromium-to-carbon ratio is favorable for Although there are similarities in the effects
Cr7C3 carbides in this grade of steel. These of the two types of embrittlement from a prac-
changes in carbide formation sequences basi- tical standpoint, TME and TE are separable into
cally slow down the precipitation of molybde- two different phenomena because they occur in
num as carbides. When the molybdenum in solid two different ranges of temperature and also
solution in the ferrite matrix is fully removed, 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 inde-
pendent 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 devel-
opment of TE when aging tempered martensite,
Table 4 Chemical elements that can segregate

to the grain boundary
Group
IV B VB VI B
C N O
Si P S
Fig. 19 Influence of phosphorus and antimony on room- Ge As Se
temperature impact energy as a function of temper- Sn Sb Te
ing temperature in a Ni-Cr-Mo steel. Arrow shows the laboratory ... Bi ...
alloy. Source: Ref 19
will develop TE upon aging a bainitic micro- embrittlement was investigated. It was shown
structure (Ref 22). that martensitic microstructures are more sus-
In fact, bainitic microstructures, present in ceptible to intergranular fracture than bainitic
many kinds of low- and medium-alloy structural microstructures (Ref 35).
carbon steels, can also develop embrittlement Mechanical Tests for TE Determination.
when tempered (TBE). Tempering at low As previously mentioned, TE is characterized
temperature has only a small effect on the by decreasing the impact resistance by heating
cementite size and morphology. Consequently, and maintenance in the critical temperature
the low-temperature embrittlement phenomena where the phenomenon occurs (450 to 600 C)
are not found in conventional bainitic micro- or by slow cooling through this temperature
structure (Ref 31). range. The impact resistance can be recovered
However, for higher temperatures, embrittle- by heating up the embrittlement temperature
ment can be developed. The tempering tem- range (4600 C), followed by a quick cooling.
perature range for TBE depends on the chemical The TE apparently does not have an influence on
composition and microstructure of the steel hardness, yield strength, and elongation mea-
(Ref 32). sured in conventional tension tests. The same
Mechanisms that provoke TBE are similar can be said about fatigue resistance, although
to TME, that is, precipitation of cementite this phenomenon produces a drastic reduction in
formed by the decomposition of retained auste- toughness fracture and also increases the tran-
nite film at boundaries of bainitic ferrite laths or sition temperature, mainly for steels with nickel,
even around the martensite-austenite islands chromium, and magnesium additions.
during soaking at the tempering temperature Experimental methods that are more adequate
(Ref 33, 34). for TE studies consider the kind of load, the test
Secondary quenching (formation of marten- temperature, and also the velocity of the
site during cooling from the tempering tem- imposed deformation. Figure 20 shows the test
perature) and transformation of retained results of AISI 1340 steel, performed in different
austenite into martensite because of plastic tempering temperatures.
deformation after tempering are other possible Figure 21 indicates that the embrittlement
reasons for the occurrence of TBE. The influ- phenomenon of tempered martensite is just
ence of impurity segregation as well as the sensitive to determinate test conditions. The
as-quenched microstructure on the TE phe- room-temperature tensile test, using unnotched
nomenon in low-alloy steel was verified. The proof tests, does not present changes in terms of
susceptibility of lower bainite or martensite to reduction of section and tension rupture in the
Fig. 20 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
embrittlement temperature range. It is possible (approximately 40 C). Impact toughness is

to make an analogy with the embrittlement controlled by the intergranular fracture below
phenomenon that occurs in ferritic stainless steel this critical temperature, while it is controlled
having chromium above 15% when exposed by transgranular fracture (i.e., the matrix
to temperatures at approximately 475 C. In toughness) above this critical temperature.
some cases, this embrittlement is not detected Temperature is an important test parameter to
in tension tests, and almost no modifications determine embrittlement during tempering
occur in elongation during the test, but a strong (Ref 37).
variation in impact resistance occurs, decreasing Temper embrittlement is not detected by
the value compared to the same unembrittled simple plots of impact energy versus tempera-
alloy steel, as seen in Table 5 (Ref 36). ture. However, it is detected by more extensive
In examining Fig. 21, it is possible to once impact testing that measures variation of the
again realize the importance of the kind of test impact transition temperature with the temper-
used to verify the embrittlement phenomenon. ing temperature. In order to determine the
As seen in Fig. 20, a common method of maximum embrittlement temperature and the
detecting loss of toughness during tempering is boundaries of the brittleness to reversible TE,
the impact test, which measures the energy the secondary tempering method is recom-
absorbed in fracturing a sample in a specified mended (Ref 38).
impact loading for a specified specimen size and It is often possible to tell if a steel has failed
geometry. Results can be reported as impact because of one of these embrittlement problems
energy (for fracture) at a specific temperature or by examining the fracture surface. A grain-
impact energy as a function of the test tem- boundary fracture mode is characteristic of both
perature. It is possible also to give the impact types of embrittlement. However, in steels with
transition temperature, which is the temperature
where the impact energy-temperature curve Table 5 Mechanical properties of ferritic
changes from a relatively high value to a rela- stainless steel at room temperature under two
tively low value. This temperature is often taken different heat treatment conditions
as the temperature at the inflection point into the Annealed at 800 C
impact energy-temperature curve, as shown in Mechanical
properties
Annealed
at 800 C
and tempered at
450 C for 4 h
Fig. 21 (Ref 22). sR, Kgf/mm2 58.4 59.7
In Fe-Ni-C steel, the TME associated with sE, Kgf/mm2 33.0 40.0
the formation of grain-boundary cementite was A, % 22.8 23.6
Impact, 12 (94.5) 1.4 (11.0)
observed only when the impact test was per- Kgm/cm2 (J)
formed below the critical test temperature
Fig. 21 Transition of the fracture behavior of two hypothetical steels in two situations. A, tough. B, embrittled. Source: Adapted from
Ref 22
extremely low levels of phosphorus and sulfur, trapping were found. Lower susceptibility to
TME can display a cleavage surface. HEM was found when this steel was tempered
Although traditionally, TME is usually to a lower strength level, due to delayed onset of
detected by Charpy tests, a series of experiments brittle intergranular fracture (Ref 41).
were carried out on three commercial steels The fracture mechanism produced by this
to explore the possibility of characterizing TME embrittlement is not simple, because it is related
by macro- and microhardness tests. Results to the imposed load as well as the hydrogen
indicate distinct hardness peaks in two steels and amount in the sample. Tests to determine this
an inflexion in the other at approximately the embrittlement process are frequently static tests,
TME temperature. These experiments are based taking many hours or even many days to obtain
on the fact that TME is associated with im- the results. This kind of test is sometimes called
purity segregation, and microhardness measure- static fatigue. Figure 22 shows the relation
ments have elucidated such segregation effects between the crack propagation mode for differ-
(Ref 39). ent load levels.
Interaction of the TE phenomenon with Hydrogen embrittlement susceptibility is
Hydrogen Embrittlement. Hydrogen embrit- related to the hardness of the ferrous matrix,
tlement (HEM) occurs when high-strength steels and a very small amount of hydrogen can be
absorb an excessive amount of hydrogen in a deleterious in the case of high hardness of the
variety of environments. Hydrogen can be matrix. Under the influence of hydrogen, the
introduced into the material in service or during resistance to fracture or crack growth of steels
materials processing, for example, in the pre- is greatly reduced in gaseous hydrogen or
sence of hydrocarbons or hydrogen sulfide or hydrogen-containing environments. Thus, some
during pickling in acids, plating, welding, and experiments with 4140 steel under different heat
heat treatment (Ref 40).
When tensile stresses are applied to a hydro-
gen-embrittled component, it may fail prema-
turely. Hydrogen embrittlement failures are
frequently unexpected and sometimes cata-
strophic. 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, cata-
strophic 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 micro-
voids. In the case of carbides, their size is related
to the hydrogen trapping effect. For the boron-
bearing steel tempered to 1050 and 1300 MPa,
Fig. 22 Illustration of a crack growing at different load levels.
fine cementites having an effect on hydrogen Source: Adapted from Ref 19
treatments were carried out. Experiment results under a higher load than the steel subjected to
indicated that the acceleration of the crack standard quenching and tempering (Ref 44).
growth in hydrogen was more pronounced for Interaction of the TE Phenomenon with
quenched and tempered conditions when the Liquid-Metal Embrittlement. The presence of
tempered temperature range coincided with the select metallic specimens on the surface of
TME range of the 4140 steel (230 to 370 C). various alloys can provide for a very detrimental
At higher tempered temperature (approximately reaction under load, known as liquid metal
550 C), the influence of hydrogen became embrittlement (LME), causing brittle fracture by
insignificant. Specimens in the austempered intergranular cracking. LME is the reduction in
condition have the best performance on the tests, elongation to failure that can occur when nor-
as seen in Fig. 23 (Ref 42). mally ductile metals are stressed while in contact
The effect of the microstructure on HEM was with liquid metals. Failure of components rela-
investigated for a low-carbon (Mn-Si-Cr) steel. ted to LME is less common than failures caused
Microstructure formed by bainite and martensite by other processes, such as fatigue, HEM and
has better behavior than tempered martensite in stress-corrosion cracking, but a significant
intermediate-temperature tempering. At the number of industrial failures related to LME do
same strength level, the impact energy of the occur (Ref 4547).
mixed microstructure is 17% higher than only LME should depend on time of contact with
tempered martensite. It is attributed to the fact liquid metal while the solid is stressed. When
that bainite-martensite delays the TME onset this occurs, the solid metal fails instantly,
(Ref 43). because the flow of liquid metal into the crack tip
Similar work was performed by Lantsman during crack growth plays a significant role (Ref
et al. Experiments with cadmium-plated 48). The liquid metal affects the fracture beha-
65S2VA steel springs were carried out. In this vior at the tip of the crack, reducing the critical
process, the steel absorbs hydrogen, which leads stress intensity for fracture and altering the
to HEM. Results show that the susceptibility of micromechanism of fracture at the crack tip
HEM has a strong dependence on the pre- (Ref 49).
liminary heat treatment and structure. With an Sources of the aggressive elements vary,
identical hydrogen content, the austempered including unintentional or accidental exposures,
steel, with lower internal stresses, will fracture 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. Even-
tually, 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 com-
mon structural materials (Ref 45).
Table 6 Examples of embrittling metal

environments for some structural materials
Structural materials Embrittling environments
High-strength martensitic steel Hg, In, Sn, Pb, Cd, Zn, Li, Cu
Austenitic stainless steel Zn, Cu, Li
Titanium alloys Hg, Cd, Ag, Au
Aluminum alloys Hg, Ga, In, Sn, Pb, Cd, Zn, Na
Copper alloys Hg, Ga, Bi, Zn, Li, Sn, Pb, In
Zirconium alloys Hg, Cd, Cd-Cs, Zn
Nickel alloys Hg, In, Li, Zn, Ag
Magnesium alloys Na, K, Rb, Cs, Zn
Fig. 23 Impact values of 4140 specimens after various
types of tempering treatments. Source: Adapted from Source: Ref 45
Ref 42
Industrial environments provide diverse when the stresses occur at lower temperatures
sources of embrittling metals, for example, than the melting point. LME is not typical, but it
during application of coatings or during service is an important failure mechanism, and indus-
if temperatures and stress are sufficiently high trial awareness of potential problems is still
and there is close contact with the substrate, as limited.
mentioned previously.
During soldering, brazing, or welding, LME
can be developed if the residual or assembly Case Studies
stresses are sufficiently high. Metal lubricants
and overheated bearings are other examples that Case Study I: Grinding Cracks. The origin
can provoke LME (Ref 45). Literature presents of grinding cracks may be related to low effi-
various cases of failures associated with LME, ciency of the cooling system, microstructure and
such as failures in gas turbines, presented by material cleanliness, and also excessive rate of
D.W. Cameron (Ref 50). In this case, high material removal per pass. The heat generated
temperatures inherent in the gas turbine aggra- due to these parameters can produce visual
vated the phenomenon. characteristics on the worn surface, as seen in the
Research has shown that the martensitic steel cam shaft surface shown in Fig. 24. This com-
91 (9% Cr, 1% Mo) is prone to LME by liquid ponent (AISI 5160) was induction hardened and
lead when some conditions are fulfilled (Ref 52 tempered.
54). Quenched and tempered steels are suscep- The surface cam shaft heat effect can be
tible to lead embrittlement, and the conditions to studied by metallography and microhardness
develop this phenomenon are: determinations (100 g load) in a cross section
Presence of either external or internal lead in containing the worn surface. Figure 25 presents
the steel this microstructure with a low-depth white layer
Tensile loading (approximately 50 mm) containing the lower
Temperature between 200 and 480 C indenter marks, which means fresh martensite
(indentations 1 to 3). The fresh martensite has
If any one of these conditions is not observed, different chemical reactivity than tempered
LME is avoided (Ref 12). martensite or even other austenite decomposi-
There are a great number of possible tion products. Because of this the overheated
mechanisms for this embrittlement. In a general surface can be identified by special chemical
way, it occurs in metals when they are in contact etching methods, for example, cooper sulfide.
with low-melting metals and can happen even Table 7 presents the hardness values.
Fig. 24 Worn surface with excessive generation of heat (arrow)

Fig. 25 Microstructure close to the worn surface. Etched with 2% nital
Table 7 Microhardness values of the worn cam

shaft cross section
Identation Depth, mm Hardness, HV0.1 Average
1 40 782 ...
2 35 803 789
3 35 782 ...
4 95 433 ...
5 92 433 440.3
6 90 455 ...
7 155 520 ...
8 150 493 525.3
9 160 563 ...
10 220 592 ...
11 220 606 613.3
12 225 642 ...
13 405 642 ... Fig. 26 Grinding cracks on AISI 5160 steel cam shaft after
14 405 782 704.7 induction hardening and low tempering (high-hard-
15 405 690 ... ness tempering). Original magnification: 200
16 610 858 ...
17 605 690 770
18 625 762 ... observed grinding cracks are small and difficult
19 1000 724 ... to detect by nondestructive tests as well as
20 1000 772 766.3
21 1000 803 ... optical techniques classified as macrographic
22 Base material 256 ... techniques. In this case, they could be seen only
23 Base material 230 245.3
24 Base material 251 ...
through microscopy techniques applied directly
on the worn surface or using replica methods.
These defects must be avoided, because they
Going deeper in the case (right side of Fig. 25) may increase the precore failure probability.
close to the white layer, a darker etched zone Machining operations must be carefully per-
(indentations 4 to 9) has lower hardness (bigger formed, but surface compressive residual stres-
indentations), which means that the temperature ses may contribute to avoiding the grinding
was sufficient to cause tempering and reduce the cracks.
local hardness. Indentations 13 to 21 (400 to Case Study 2: Transgranular and Inter-
1000 mm) show higher hardness (Table 7) rela- granular Crack Path. Intergranular cracks are
ted to the process tempering operation, where frequently related to the TE phenomenon.
hardness values are close to the martensite as- However, during low-temperature tempering,
quenched condition. very close to the embrittlement range, a pre-
This condition is deleterious because worked dominantly intergranular crack path was found
surfaces are more prone to grinding cracks. in an AISI 5160 commercial steel after a tem-
One example can be seen in Fig. 26. The pering operation at 2000 C (Ref 55).
In a recent failure analysis of a compact 0.9%) shows an intergranular overload fracture

vehicle suspension spring (chemical composi- path, and the nucleus low-carbon structure
tion shown in Table 8) with a specified hardness (approximately 0.2%) presents a dimpled
range of 53 to 55 HRC, a fatigue crack nucleated transgranular crack pattern (Fig. 31a, b).
from a surface defect that had grown quickly The higher carbon content increases hardness
(radial marks), as seen in Fig. 27. A scanning values, which can promote stresses higher than
electron microscopy image of the radial fracture the cohesive grain-boundary strength or even
surface (Fig. 28) shows transgranular and tempering outside of the embrittling temperature
intergranular fracture paths. range, as mentioned previously.
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 ACKNOWLEDGMENT
TE range. However, a significant intergranular
crack path was found in the fracture surface. The authors would like to acknowledge J.C. Ven-
This case shows that the fracture can assume an dramim from ISOFLAMA Ind. Com Equip. Ltd.,
Brazil, for helpful suggestions in writing this chapter.
intergranular path even outside of the embrit-
tlement 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
Table 8 Vehicle suspension spring analysis

C Si Mn Cr S Ni P Cu
0.49 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 Fig. 29 Tempered martensite with some retained austenite
magnification: 6 (approximately 20%). Etched with 2% nital
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
REFERENCES Dekker Inc., New York, NY, 1997, p 527

662
1. M. Szczepanski, The Brittleness of Steels, 14. H.K.D.H. Bhadeshia, Tempered Marten-
John Wiley & Sons, Inc., New York- site, University of Cambridge, http://www.
London, New York, 1963 msm.cam.ac.uk/phasetrans/2004/Tempered.
2. R.W.K. Honeycombe, Steel Microstructure Martensite/tempered.martensite.html (ac-
and Properties, Edward Arnold Ltd., 1981, cessed March 2008)
p 78, 205 15. G. Krauss, Deformation and Fracture in
3. E.C. Bain and H.W. Paxton, Alloying Ele- Martensitic Carbon Steels Tempered at Low
ments in Steel, American Society for Temperatures, Metall. Mater. Trans. B,
Metals, 1966 Vol 32, 2001, p 205221
4. A.V. Sverdlin and A.R. Ness, Chapter 2 16. W.J. Nam and H.C. Choi, Effect of Silicon,
The Effects of Alloying Elements on the Nickel and Vanadium on Impact Toughness
Heat Treatment of Steel, Steel Heat Treat- in Spring Steels, Mater. Sci. Technol.,
ment Handbook, G.E. Totten and M.A.H. Vol 13, 1997, p 568574
Howes, Ed., Marcel Dekker Inc., New 17. W.J. Nam and H.C. Choi, Effect of Si on
York, NY, 1997, p 4591 Mechanical Properties of Low Alloy Steels,
5. G.R. Speich and W.C. Leslie, Tempering of Mater. Sci. Technol., Vol 15, 1999, p 527
Steel, Metall. Trans., Vol 3 (No. 5), May 530
1972, p 10431054 18. C. Jensen, Dimensional Changes during
6. G.R. Speich, Trans. Metall. Soc. Am. Inst. Heat Treatment of Hot Work Steel, Ind.
Mech. Eng., Vol 245 (No. 12), 1969, Heat., March 1993, p 3338
p 25532564 19. R.W. Hertzberg, Chapter 10: Micro-
7. M. Narazaki, G.E. Totten, and G.M. Web- structural Aspects of Fracture Toughness,
ster, Hardening by Reheating and Quench- Deformation and Fracture Mechanics of
ing, Handbook of Residual Stress and Engineering Materials, 4th Ed., John Wiley
Deformation, G.E. Totten, M. Howes, and & Sons, Inc., 1996, p 460463
T. Inoue, Ed., ASM International, 2002, 20. H. Ohtani and C.J. McMahon, Jr. Modes of
p 499 Fracture in Temper Embrittlement Steels,
8. A. Kelly and R.B. Nicholson, Strengthen- Acta metall., Vol 23, p 377386
ing Methods in Crystals, Materials Science 21. S.K. Banerj, C.J. McMahon, Jr., and H.C.
Series, Applied Science Publishers Ltd., Feng, Intergranular Fracture in 4340-
London, 1971, Printed in Great Britain Type Steels, Effect of Impurities and
by Galliard Limited, Great Yarmouth, Hydrogen, Metall. Trans. A, Vol 9, 1978,
England p 237247
9. J. Vatavuk and H. Goldenstein, Acos Had- 22. C.R. Brooks, Chapter 5Tempering,
field Contendo Carbonetos de Niobio e Principles of the Heat Treatment of Plain
Titanio, Anais do II Encontro de Tecnologia Carbon and Low Alloy Steels, ASM Inter-
e Utilizacao dos Acos Nacionais, May 18 national, 1996, p 127204
22, 1987 (Rio de Janeiro), ABM, COPPE 23. F.A. Darwish, L.C. Pereira, C. Gattis, and
(UFRJ), p. 151169 M.L. Graca, On the Tempered Martensite
10. Heat Treating Data Book, 8th ed., Seco/ Embrittlement in AISI 4140 Low Alloy
Warwick Corporation, 2000, p 96 Steel, Mater. Sci. Eng., Vol A132, 1991,
11. T. Swarr and G. Krauss, Effect of p L5L9
Structure on the Deformation of As- 24. I. Olefjord, Temper Embrittlement, Review
Quenched and Tempered Martensite in a 231, Int. Met. Rev., Vol 23, 1978, p 149163
Fe-0.2%C Alloy, Metall. Trans. A, Vol 7, 25. G. Krauss, Tempered Martensite Embrit-
1976, p 4148 tlement in AISI 4340 Steel, Metall. Trans.
12. G. Krauss, Chapter 8Tempering of Steel, A, Vol 10, 1979, p 16431649
Principles of Heat Treatment of Steel, 26. M.A. Grossman and E.C. Bain, Chapter 5
American Society for Metals, 1980, p 187 Tempering After Quench Hardening, Prin-
227 ciples of Heat Treatment, American Society
13. B. Liscic, Chapter 8Steel Heat Treat- for Metals, 1964, p 129175
ment, Steel Heat Treatment Handbook, G.E. 27. G.E. Totten, C.E. Bates, and N.A. Clinton,
Totten and M.A.H. Howes, Ed., Marcel Chapter 1Introduction to the Heat
Treating of Steel, Handbook of Quenchant 41. M. Wang, E. Akiyama, and K. Tsuzaki,

and Quenching Technology, ASM Interna- Hydrogen Degradation of a Boron-Bearing
tional, 1993, p 133 Steel with 1050 and 1300 MPa Strength
28. V.V. Zabilskii, Temper Embrittlement Levels, Scr. Mater., Vol 52, 2005, p 403
of Structural Alloy Steels Review, Met. 408
Sci. Heat Treat., Vol 29 (No. 1), 1987, 42. J.H. Chuang, L.W. Tsay, and C. Chen,
p 3242 Crack Growth Behaviour of Heat-Treated
29. S.-H. Song, R.G. Faulkner, and P.E.J. 4140 Steel in Air and Gaseous Hydrogen,
Flewitt, Quenching and Tempering Int. J. Fatigue, Vol 20 (No. 7), 1998, p 531
Induced Molybdenum Segregation to 536
Grain Boundaries in a 2.25%Cr-1Mo 43. D. Liu, B. Bai, H. Fang, W. Zhang, J. Gu,
Steel, Mater. Sci. Eng. A, Vol 281, 2000, and K. Chang, Effect of Tempering Tem-
p 2327 perature and Carbide Free Bainite on
30. S.G. Chowdhury, P. Kumar, S.K. Das, D.K. the Mechanical Characteristics of a High
Bhattacharya, and N. Parida, Failure Ana- Strength Low Alloy Steel, Mater. Sci.
lysis of High Temperature Studs, Eng. Fail. Eng. A, Vol 371, 2004, p 4044
Anal., Vol 8, 2001, p 521528 44. P.Sh. Lantsman, G.G. Vernovskaya, and
31. Y. Ohmori, H. Ohtani, and T. Kunitake, V.N. Kudryavtsev, Use of Austempering to
Met. Sci., Vol 8, 1974, p 357366 Reduce Hydrogen Embrittlement of Steel
32. G.L. Hu, Z.T. Liu, P. Wang, W.J. Hwa, and 65S2VA, Metalloved. Term. Obrab. Met.,
M.K. Kang, Mechanism of Tempered Vol 5, May 1973, p 7374 (translated from
Bainite Embrittlement in Some Structural Russian)
Steels, Mater. Sci. Eng. A, Vol 141, 1991, 45. S.P. Lynch, Metal Induced Embrittlement
p 221227 of Materials, Mater. Charact., Vol 28, 1992,
33. H.K.D.H. Bhadeshia and D.V. Edmonds, p 279289
Met. Sci., Vol 17, 1983, p 411419 46. P.L.J. Fernandez, R.E. Clegg, and D.R.H.
34. K.D.H. Bhadeshia and D.V. Edmonds, Met. Jones, Failure by Liquid Metal Induced
Sci., Vol 17, 1983, p 420425 Embrittlement, Eng. Fail. Anal., Vol 1
35. S. Raoul, B. Marini, and A. Pineau, Effect (No. 1), 1994, p 5163
of Microstructure on the Susceptibility of a 47. R.E. Clegg and D.R.H. Jones, Liquid Metal
533 Steel to Temper Embrittlement, J. Nucl. Embrittlement in Failure Analysis, Mater.
Mater., Vol 257, 1998, p 199205 Sci., Vol 27 (No. 5), 1992, p 453459
36. J. Vatavuk, Estudo Sobre os Efeitos do 48. K. Ina and H. Koizumi, Penetration of
Niobio na Resposta ao Tratamento Termico Liquid Metals into Solid Metals and Liquid
de Ligas Fe-17Cr, Fe-17Cr-Ni, e do Cr e do Metal Embrittlement, Mater. Sci. Eng. A,
Ni na Solubilidade do Nb na Ferrita, Dis- Vol 387389, 2004, p 390394
sertacao de Mestrado apresentada a` Escola 49. R.E. Clegg, A Fluid Flow Based Model to
Politecnica da Universidade de Sao Paulo, Predict Liquid Metal Induced Embrittle-
1986 ment Crack Propagation Rate, Eng. Fract.
37. H. Kwon and C.H. Kim, Tempered Mar- Mech., Vol 68, 2001, p 17771790
tensite Embrittlement Fe-Ni-C Steel, 50. D.W. Cameron, Failures in Large Gas
J. Mater. Sci., Vol 18 (No. 12), 1983, Turbines due to Liquid-Metal Embrittle-
p 36713678 ment, Mater. Charact., Vol 33, 1994, p 37
38. A.V. Taran, Determining Susceptibility to 43
Reversible Temper Embrittleness, Met. Sci. 51. Liquid Metal Embrittlement, Failure Ana-
Heat Treat., Vol 15 (No. 10), 1973, p 865 lysis and Prevention, Vol 10, Metals
868 Handbook, 8th ed., American Society for
39. A. Ghosh, K. Bhaduri, and K. Ray, Metals, 1975, p 228229
Characterization of Tempered Martensite 52. G. Nicaise, A. Legris, J.B. Vogt, and J.
Embrittlement Using Hardness Measure- Foct, Embrittlement of the Martensitic
ments, Mater. Charact., Vol 33, 1994, Steel 91 Tested in Liquid Lead, J. Nucl.
p 321327 Mater., Vol 296 (No. 1), July 2001, p 256
40. N.A. Tiner, Brittle Fracture in Equipment 264
Failure, Mech. Eng.-CIME, Vol 112 (No. 6), 53. A. Legris, G. Nicase, J.B. Vogt, and
1990, p 6064 J. Foct, Liquid Metal Embrittlement of the
Martensitic Steel 91: Influence of the Che- of a Martensitic Steel by Liquid Lead and
mical Composition of the Liquid-Metal Other Liquid Metals, J. Mater. Sci., Vol 40
Experiments and Electronic Structure (No. 910), 2005, p 24592463
Calculations, J. Nucl. Mater., Vol 301 55. A Reguly, T.R. Strohaeker, G. Krauss, and
(No. 1), 2002, p 7076 D.K. Matlock, Quench Embrittlement of
54. A. Legris, J.B. Vogt, A. Verleene, and I. Hardened 5160 Steel as a Function of
Serre, Wetting and Mechanical Properties, Austenitizing Temperature, Metall. Mater.
A Case Study: Liquid Metal Embrittlement Trans. A, Vol 35, Jan 2004, p 153162
Failure Analysis in Tool Steels

Rafael Agnelli Mesquita and Celso Antonio Barbosa,
Villares Metals
TOOL STEELS are an important class of selection of heat treatment or tool design.
steels due to their distinct applications and, A helpful way to think of this synergistic relation
especially, their specific heat treating issues. is to think of the factors as simple multiplying
Tool steels are used in various industrial appli- factors. If one of them is zero, regardless which
cations that require some kind of mold, die, or one, the whole result would be zero. Of course,
mechanical device. Tool steels are also a com- the factors are not so distinctly separate.
plex class of steels with compositions close to
carbon steel or more highly alloyed grades. Tool
steels are generally classified according to three Classification of Tool Steels
main definitions (Ref 1, 2):
They are used in some forming process or Due to the large number of grades and appli-
forming operation for metal, ceramic, or cations of tool steels, several possible classifi-
plastic shaping. cations have arisen. This division is important
Tool steel properties are only attained after for the discussion of heat treating. One of the
heat treating, normally performed by hard- most well-known classifications is made by the
ening (quenching) followed by tempering. American Iron and Steel Institute (AISI), divid-
This class of material is produced according ing tool steels into several classes according to
to rigorous melting and processing controls. application, composition, or heat treatment. This
Therefore, even for grades with chemical classification is shown in Table 1.
compositions very close to low-alloy carbon Another possible classification for tool steels
or engineering steels, the production prac- is their division into four groups according to the
tices of tool steels lead to substantial prop- final application: hot work, cold work, plastic
erty improvements. mold, and high-speed tool steels. The advantage
of this division is to deal with fewer groups and
According to this definition, heat treating is a to group the diverse grades within common
key issue for tool steels. In fact, heat treating aspects of each application, such as sizes, hard-
does have a strong effect on tool life, as des- ness, operating conditions (chocks, wear, or
cribed in this chapter. In some applications, plastic deformation), and surface-finishing re-
three major points are equally related to tool per- quirements. This chapter follows such a divi-
formance (supposing that operation is constant): sion, but the grade nomenclatures used here are
Design and manufacture finishing primarily from AISI.
Steel composition and its quality
Heat treatment applied Table 1 AISI classification for tool steels (Ref 1)
Group Symbol
Even though this approach shows the impor-
Water-hardened tool steels W
tance of heat treating, it may be too simple for Shock-resistant tool steels S
determining the variables related to tool per- Oil-hardening tool steels O
formances. That is, all these aspects interact with Air-hardening tool steels A
High-carbon and high-chromium die steels D
each other and may influence the final result. Tool steel for application in plastic molds P
For example, if a good design and proper heat Cr-, Mo-, or W-alloyed hot work tool steels H
Tungsten-alloyed high-speed steels T
treatment are applied to a tool produced with an Molybdenum-alloyed high-speed steels M
imporperly selected steel grade, the result can be Adopted from Ref 1
awful. The same may happen for an incorrect
Cold work tool steels are employed for Cold work tooling is also applied to shock
tooling operations in temperatures below operation applications, such as knives for shear
200 C, typically at room temperature, and for cutting thick plates (normally thicker than
steel forming (Fig. 1 shows some examples). 13 mm, or 1/2 in.), chisels, and some powder
The most common tools of this class are punches pressing molds or cold forging dies. In these
and dies for cold drawing and stamping, knives, situations, high toughness is also important,
thread-rolling dies, and coining or cold forging even if the wear resistance is reduced. Grades
tools. In these situations, mechanical strength of AISI class S are the most important example
and wear resistance are the main performance- in these applications. Toughness is also impor-
related properties, which are only obtained tant in the previously discussed application,
through high hardness after heat treating, nor- preventing several catastrophic failures or, more
mally at the 60 HRC level, and the dispersion of commonly, adhesive wear failures, in which
coarse carbides in the tool steel microstructure. microchipping is a very important issue. In these
Grades of AISI class D are the most important situations, the advent of new materials with
examples for such tools, but A and O class lower carbon and chromium contents has shown
grades are also employed. interesting results.
Hot work tool steels are used for applica-
tions in which process temperature is an im-
portant 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 em-
ployed for aluminum alloys and for producing
Fig. 1 Examples of cold work tools. Courtesy of Villares construction or automotive parts, respectively.
Metals All applications normally employ AISI grades H
Fig. 2 Examples of hot work dies for (a) press forging and (b) die casting. Courtesy of Villares Metals
Failure Analysis in Tool Steels / 313
steel as core tools. However, for large forging selection is different, and corrosion or wear
dies and tool holders, low-alloyed materials, resistance should be emphasized more than
such as AISI grade 6F3 (DIN 1.2714), may be manufacturing-related properties.
employed. AISI P20 or P20-modified grades (mainly
The most important properties for hot work DIN 1.2738) are the most employed steels in
tooling are hot strength and toughness. Hot the mold-making industry, delivered in the pre-
strength is normally related to the operating hardened condition with 32 HRC. Nevertheless,
conditions, specifically, how long and at what more specialized applications normally require
temperature the tools are exposed. This gives higher-alloyed steels, such as H13, modified
rise to an important concept called tempering martensitic stainless steel (e.g., DIN 1.2083), or
resistance, which describes the steel resistance even highly alloyed powder metallurgy grades.
to hardness reduction after exposure to high Currently, the advance of manufacturing tech-
temperatures. Toughness, which is related to nologies, especially those related to high-speed
crack end-life situations, is also very important machining technologies, has increased the appli-
in hot work tooling. Toughness requirements are cation of mold steels prehardened to high hard-
so important that hot work tool steels normally ness, such as 40 HRC or higher. This new level
have carbon contents close to 0.40% and hard- improves mold quality, especially the polishing
ness below 52 HRC, both values much lower characteristics.
than that of cold work grades. High-speed steels are also tool steels, but
Plastic mold steels have become a very they have important differences from other tool
important division of tools due to the increase in steels. They are usually used in cutting tools,
plastic material applications (Fig. 3). Mecha- whereas the other classes are mainly employed in
nical requirements for strength and toughness forming tools. Figure 4 shows some examples of
are less important than in the previous applica- high-speed tools. Regarding the manufacturing
tions. In general-application molds, the mold- characteristics, high-speed steel tools are mainly
manufacturing-related properties are the most produced by single companies and in smaller
important, because the end life of these molds is sizes but larger quantities than molds or dies.
rarely attained. Therefore, steel machinability, From the metallurgical point of view, high-
polishability, heat treating response, and weld- speed tool steels are definitely part of the tool
ability are of special interest. Although less steel groups. The AISI classes for such grades
common, there are high-demand applications, include the AISI M and T series, which corres-
such as the processing of corrosive polymers pond to the main alloy element present in their
or reinforced plastics, where the reason for steel composition: molybdenum (M) or tungsten (T).
Fig. 3 (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.
The first high-speed steels were the T type, with High-speed tool steel properties are normally
T1 being the pioneer grade. During WW II, the aimed at high wear resistance, which is the main
supply of tungsten, mainly from Germany or demand in cutting operations. Wear conditions
eastern Europe, was interrupted, leading to the are complex in high-speed steel tools, with
development of molybdenum-rich grades. The important aspects being high hardness, normally
first developed grade was M1, which is still close to 65 HRC, and the homogeneous dis-
employed in the United States, but M2 has tribution of undissolved carbides. Also impor-
become much more important. M2 has a com- tant is the material tempering resistance, in
bination of molybdenum and tungsten, in almost order to retard hardness reduction at the high
equal parts, which leads to very useful properties. temperatures developed in cutting processes.
This grade is used for several applications. Toughness should be of an acceptable level and
Today (2008), it is the most important high- reduce chipping and cracking, but this property
speed composition. Recently, the increase in is low when compared to other tool steels due
consumption of molybdenum, tungsten, and to extremely high hardness and the amount of
vanadium has increased the cost of such alloys, carbides in high-speed steel microstructures.
providing another driving force to the develop- In the following sections, several failures and
ment of new grades. In China, due to the large their relation to heat treatment are described.
amounts of tungsten ferroalloys in that country, For better understanding, the aforementioned
the use of tungsten-rich compositions has gained division of tool steels is kept in two main groups,
new attention. Another important element is cold work and hot work, which are the main
niobium (formerly known as columbium). Its classes where failures are observed. High-speed
large reserves in Brazil motivated the devel- steels and mold steels are also discussed in some
opment of niobium-modified grades in that specific cases. However, many aspects of a
country. 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.
Heat Treating Failures of Cold

Work Tools
Chemical Composition and Main Charac-

teristics of Cold Work Tool Steels. Typical
chemical compositions for the most common
cold work tool steels are presented in Table 2.
Fig. 4 Examples of high-speed tools. Courtesy of Villares AISI nomenclature is used in most cases. How-
Metals
ever, some newly developed grades, not yet
Table 2 Typical chemical compositions of cold work tool steels

Composition, wt%
AISI DIN UNS C Si Mn Cr Mo W V Fe Others
D2 1.2379 T30402 1.50 0.3 0.3 12.0 1.0 ... 0.9 bal ...
D3 1.2080 T30403 2.25 0.3 0.3 12.0 ... ... ... bal ...
D6 1.2436 ... 2.10 0.3 0.3 11.5 ... 0.7 0.2 bal ...
8%Cr-0.8%C; ... 0.85 1.0 0.3 8.5 2.1 ... 0.5 bal Nb = 0.15
e.g., VF800AT(a)
A2 1.2363 T30102 1.00 0.3 0.3 5.0 1.0 ... ... bal ...
O1 1.2510 T31501 0.95 0.3 1.3 0.5 ... 0.5 0.1 bal ...
S1 1.2542 T41901 0.45 1.0 0.3 1.4 0.2 2.0 0.2 bal ...
M2 reg. C 1.3343 T11302 0.89 0.4 0.3 4.2 5.0 6.1 1.9 bal ...
PM M3 : 2(b) ~1.3344(b,c) ~T11323(b,c) 1.28 0.4 0.3 4.2 5.0 6.3 3.0 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
standardized, are referred to in the commercial high temperatures, as discussed in later sections
grade. The high carbon level of almost all these for hot work steels.
grades is evident (except for S1 due to its use in Although important or even essential for
shock-resistant applications). These tool steels some applications, high wear resistance is nor-
have high as-quenched hardness and, when mally accompanied by low toughness. High
associated with high alloy content, high harden- hardness naturally causes a reduction in tough-
ability. ness due to its effect on material fracture tough-
The combination of high carbon content and ness. Carbides normally act as crack initiation
alloy elements, especially chromium, leads to sites; therefore, the higher the volume fraction
the formation of large (also called blocky) car- of large carbides, the lower the tool steel
bides within the microstructures of cold work toughness.
tool steels. Along with a high matrix hardness, The previous discussion can be illustrated
normally 60 HRC, these carbides are the main by materials properties and microstructures.
reason for the wear resistance of cold work tool Typical microstructures of tool steels used in
steels. A simple mechanism for abrasive wear cold work tooling are shown in Fig. 6. Coarse
is shown in Fig. 5, with an abrasive particle carbides, larger than 10 mm, are present within
grooving away portions (chips) of tool steel the microstructures. Such large carbides are
material. Hardness increase reduces the particle typically the M7C3 type, with M representing
penetration and the wear. However, steels can- the metal element that is primarily chromium.
not be much harder than 65 HRC (approximately They are typical for AISI D grades and are pre-
900 HV), which is much less than a typical sent in lower amounts in 8%Cr-0.8%C new
abrasive particle (between 1500 and 2500 HV). materials, as shown for VF800AT steel (the
Therefore, another mechanism takes place, this commercial brand name). This is the main
being performed by the presence of carbides. advantage of the higher toughness of such
They can be as hard as or even harder than grades. A combination of toughness and abrasive
abrasive particles, conferring to the material a wear resistance leads to important improvements
higher resistance to wear losses, which means in die life for 8% Cr steels, especially in metal-
high wear resistance. In typical cold tooling, the forming operations where adhesive wear is
wear mechanism is more complicated, being the main operating mechanism. Lower-alloyed
a combination of abrasive and adhesive wear. grades, such as O1, have almost no carbides
The aforementioned mechanism is still valid, in the microstructure. This is caused by the
but toughness is also shown to be important for lack of chromium content to combine with
situations where microchipping or microcrack- carbon. Therefore, they are normally used in
ing is present. less demanding applications in terms of wear
High-speed steels are similar to cold work resistance.
steels in this sense, with high hardness and dis- Besides the microstructural aspects, a dis-
persion of nondissolved carbides. The difference cussion about tempering curves is also important
is that high-speed steels present higher hot in order to understand cold work steel properties.
resistance promoted by higher alloy content, Figure 7 shows the tempering curves for some
which leads to intense secondary hardening at of these grades, selected because they enable
Fig. 5 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
a discussion of all materials. For almost all is known to dislocate temper embrittlement to
applications, the high hardness of cold working higher temperatures (Ref 4). A distinct behavior
tool steels is normally achieved from low tem- is presented by the newcomer 8% Cr steels.
pering treatments. AISI D and O steels normally Besides lower carbon and chromium, this class
have low secondary hardening. In fact, the cold also has higher molybdenum contents than AISI
work application is dictated by this situation. D grades. This enhances secondary hardening,
Because the materials show a low potential which enables tempering at higher than 500 C
for hardness retention at hot conditions, their and obtains hardness as high as 62 HRC. A
application is limited to low temperatures, substantial improvement in toughness and sur-
typically room temperature. face treatment behavior is obtained through this
In practical terms, the aforementioned curves alternative tempering, which is discussed sub-
show that tempering must normally be per- sequently.
formed at temperatures near 200 C, if a hard- Some heat treating failures are observed
ness of approximately 60 HRC is desired. AISI S immediately after heat treatment and appear as
grades, herein represented by S1 steel, are also small or (usually) large catastrophic cracks.
tempered in low temperatures, approximately However, other failure types related to heat
300 C. The chemical composition of these treatment are only observed during tool use,
alloys is specially designed for this condition, when one notices premature failure or a lower-
with a high silicon level, because this element than-normal performance.
Fig. 6 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
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-treating-

related failures are divided by topics that rep-
resent 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 pre-
sented information, keeping in mind the possi-
bility for interaction when solving or analyzing a
specific troubleshooting case.
Design-Related Failures. The previous dis- Fig. 8 Comparison of longitudinal Charpy V-notched impact
cussion of cold work tool steel metallurgy and toughness for various tool steel specimens taken from
89 mm square stock and tested at working hardness. Source:
characteristics explains why this class of mate- Ref 6
rials is so prone to fracture and cracking. Except
for AISI S grades, all other materials are very of tools. However, several tools are still
brittle. This is due to their intrinsic naturethe designed based only on previous experience.
combination of high hardness and primary Design faults may cause failures in heat
carbidesand also because cold work steels treatment but also during tool use, leading to
are used predominantly at room temperature, short service life. Failures just after heat treat-
where fracture toughness of steels is naturally ment normally occur in the presence of some
reduced (Ref 5). This fact is illustrated in Fig. 8, of the following features: heavy sections adja-
where the lower toughness of A, D, and O grades cent to light sections, sharp corners, stamp
in comparison to H or S steels is obvious. marks, blind holes, and improperly spaced holes
Cold work tool steels are thus prone to failure (Ref 7). Several of these faults are illustrated in
under stress concentrators, also called stress Fig. 9. Large section-size variation caused the
raisers, that are imposed by tool design or failures shown in Fig. 9(a) and 9(b), while the
machining. Today (2008), modern software is presence of sharp corners or closely spaced
able to calculate stresses and tool working con- holes is shown in Fig. 9(c).
ditions and can help to reduce stresses and Other stress-concentration effects can also
especially localized stresses under some regions cause or even facilitate cracking. One example
is presented in Fig. 9(d), showing cracking design. However, they can be minimized if heat
produced by stamp marks. It is not evident treatment and service failure issues are con-
in Fig. 9(d), but this failure also had a contri- sidered prior to tool design. In other words,
bution of poor machine finishing, because the the designer should foresee possible heat treat-
deep tool mark also acts as an important stress ment or service problems at the beginning of
concentrator. Figure 9(e) shows a typical sharp- tool design. As a result, several failures can
corner crack. In this case, the corner was filled, be avoided, and service life may be enhanced.
but there was a nick in the corner where the In this context, some basic advice is given in
cracking began. The shape of this fixture is also Fig. 10.
poor for steel that must be oil quenched. As in Another possibility for solving heat treatment
the case of Fig. 9(a), thinner outer regions cool or service cracking is tool steel selection. Instead
more rapidly, forming martensite first, while the of using water-hardening grades, oil-hardening
more massive central regions cool more slowly. ones are preferred in situations sensitive to
In some cases, it is not possible to eliminate quench cracking. In some circumstances, it is
all the stress-concentration effects from a tool possible to apply air-hardening grades, such as
Fig. 9 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 stress-
concentration 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
series A or the new grade type 8% Cr. For the macroscopic stress raiser effect due to two
example, this could be a solution for the failure major factors, described as follows.
shown in Fig. 9(e). Even D grades can be air First, both grinding and EDM cause local
quenched, depending on die section. However, heating in the tool surface that, depending on
the continuous cooling transformation diagram operational conditions, induces local tempering
of the cooling material should be analyzed for or, far worse, reaustenitizing, quenching, and
the hardening condition and also the possibility hardening. The high carbon of these grades,
for carbide precipitation on the grain boundaries. normally more than 1%, promotes high hard-
Before finishing this subject, one further point ness, more so than in other lower-carbon tool
should be considered. Stress raisers increase steels (Fig. 11). As a result of heating and mar-
failures in cold work tool steel, mainly due to tensite transformation, small cracks, normally
the intrinsically low toughness of such grades. hard to see with the unaided eye, may also be
However, almost any tool displays some loca- formed, acting as stress raisers during tooling
lized stress. Depending on working conditions, and thus enabling premature cracking. Sec-
one or more cracks can be initiated and propa- ondly, after such a metallurgical transformation
gate throughout, fracturing or spalling the tool. on the tool surface, the microstructure will be
Therefore, cold work steels are sensitive to over- predominantly untempered martensite (also
load failure. In many situations, no problems known as fresh martensite). This microstructure
exist in the steel, the design, or even in the heat is very brittle, especially in high-carbon steels
treatment; the only cause may be excessive such as the cold work grades. The pre-existing
stressing of the tool due to its incorrect use. cracks or other cracks formed during tool
Surface Damage by Grinding or Electrical operation are much more prone to propagate,
Discharge Machining. In the previous section, thus accelerating tool failure.
the intrinsic brittleness of cold work tool steels Grinding and EDM cause heating and actually
was discussed, as well as the correlation of tool act as a heat treatment applied to tool steel. In the
failures to poorly designed tools (regarding stress next sections, grinding and EDM are treated
raisers). This section discusses surface defects separately, with some advice for avoiding
introduced in tool manufacturing by grinding or problems.
electrical discharge machining (EDM). Never- Incorrect Grinding. Hardenable steels are
theless, these processes can introduce more than more prone to grinding cracks than low-carbon,
low-alloy steels. Cold work tool steels and high-
speed steels are the most sensitive tool steel
grades to such problems due to the high hardness
Fig. 11 Effect of carbon content on the hardness of differ-

ent 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 high-
Fig. 10 Simple possibilities for avoiding (a) sharp corners carbon grades such as cold work steels, leading to tool failure.
and (b) large variation in section. Source: Ref 10 Source: Ref 1
of as-quenched martensite and its low toughness. permanent deformation of the ground part when
Examples of grinding cracks are shown in grinding thin materials. This may be accom-
Fig. 12. panied by retained austenite formation, which
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 tem-
pering 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, toge-
ther 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 untem-
pered 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, result-
ing 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 grind-
ing wheels that cut the material with sharp cut-
ting edges instead of simply generating heat Fig. 12 Examples of grinding cracks. (a) Two views of an S1
tool cutter die cracked and spalled after grinding. As-
through friction (Ref 11). received (left) and after magnetic particle testing (right), accent-
uating the cracks Source: Ref 9. (b) A D2 die that cracked due to
The majority of grinding operations leave incorrect grinding (arrow indicates grinding marks) Failure was
residual stresses in the ground surface, usually also assisted by closely spaced holes and electrical discharge
being at a maximum close to the surface. The machining procedures. Generally, grinding cracks are not as easy
to see as this. It is usually necessary to examine the part under a
first and most common effect of such stresses microscope or with magnetic powder inspection in order to see
is the occurrence of cracks. Stresses can cause the cracks. Source: Ref 8
enhances deformation and the possibility for work tools for various reasons. Cold work tools
cracking. have an intrinsic high wear resistance and are
Three avenues are available to reduce grinding normally difficult-to-cut materials under regular
assisted failures. First is to control the heat machining processes, such as milling. Finishing
emerging from the grinding operation by use of the die making with EDM may be an interesting
a proper cooling process. Secondly, the effect solution, especially for complex-shaped tools.
of grinding stresses and problems can be reduced However, new developments in high-speed
by stress-relief tempering after grinding. This machining that feature low stock removal and
also involves the tempering of some regions of high frequency have been used on dies as hard as
untempered martensite, if present. The treatment 60 HRC.
temperature should be approximately 30 C The use of EDM on hardened steels, however,
below the previous tempering temperature to can produce a shallow, rehardened layer of
avoid any risk of reducing the hardness of the rapidly quenched as-cast structure and untem-
workpiece. Third, another way of reducing pered martensite at the surface, beneath which
grinding stresses is to tumble or blast the ground is a layer of tempered martensite (Fig. 14). The
parts (Ref 11). Obviously, if the heat damage is EDM surface layer is known as the white layer
too high, that is, cracks, stress relief may not help. because of its lighter appearance under optical
Incorrect EDM. Electrical discharge mach- microscope observation of etched samples (this
ining is often used in the production of cold 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 resi-
dual stresses; the origins of such stresses are
the thermal and phase transformation dimen-
sional 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 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
Fig. 13 (a) White layer on a tool surface rehardened by an caused by the EDM process and the presence of small cracks in
incorrect grinding procedure. (b) Typical hardness this layer. Original magnification: 500 . Courtesy of Villares
profile in regions close to cracks. Source: Ref 11 Metals
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
coarse (as-cast) microstructure, and frequent high-speed steels, where the surface white layer
precracked regions. has high brittleness and the base material (i.e.,
Some typical examples of die failures assisted the tool steel) has low resistance to crack pro-
by incorrect EDM are shown in Fig. 15. In pagation. However, EDM-assisted failures are
Fig. 15(a), a 63 mm (21/2 in.) diameter by also observed in hot work dies.
13 mm (0.5 in). long blanking die for a small Also important to mention here is the effect of
part cracked from the corners of the EDM cavity incorrect EDM in plastic molds. Although the
after the die had produced 20,000 pieces. The mechanical stressing is normally low, surface
die was made of heat treated A2 tool steel. In finishing (by polishing or texturing) is crucial in
Fig. 15(b), the surface of an AISI A4 cup plate is this application because plastic injected parts are
shown, with spalling at one of the holes, which able to reproduce any problems on the mold
were made by EDM. A laboratory investigation surface. Thus, EDM defects may cause serious
of this failure led to the typical appearance of quality problems to injected parts, impairing the
EDM-assisted failures: a coarse, white surface mold application.
layer that is very high and brittle, due to the Failures due to the Heat Treating Proce-
presence of an untempered martensitic matrix dure of Cold Work Steels. The heat treating
and net carbides. Below, unquenched martensite procedure can itself deeply change the micro-
is observed, followed by a region of overtem- structure and properties of all tool steels, not
pered martensite, after which the normal (core) only the cold work grades. This may occur even
microstructure is observed. if the specified hardness is obtained. This section
In many situations, the affected regions are not deals with failures caused by improper heat
exposed as clearly as in Fig. 15(c). However, the treating procedure and is divided into the three
typical white layer is always present on the tool most common causes in cold work tool steels:
surface, as shown in Fig. 14. The coarser this the use of incorrect temperatures, the use of
layer, the higher the probability for tool failure, excessively short tempering times (or even no
due to its brittleness and the fact that the EDM tempering at all), and the formation of excessive
white layer likely possesses cracks. amounts of retained austenite, caused either by
Three major practices are recommended for improper hardening or incorrect tempering.
avoiding premature cracking caused by EDM: Incorrect Hardening or Tempering Tem-
peratures. As for other tool steels, the same
Reduce the stock removal when finishing the
class of cold work tool steels may present im-
EDM process (if low-frequency EDM was
used for roughing, high-frequency should be portant differences in the indicated heat treat-
ing temperatures. If the temperature is higher
used for finishing). This is helpful for mini-
or lower than that indicated for a certain grade,
mizing the depth of the rehardened white
layer. mechanical properties may be altered, especially
for toughness. Thus, the die performance is also
The white layer should be eliminated or
strongly influenced.
minimized by light grinding or lapping. This
This section describes this effect in a specific
procedure is time-consuming but, in many
gradethe 8% Cr tool steel, which has been
situations, can lead to an impressive exten-
highly employed in tools that traditionally use
sion of tool life, especially when white layers
grades from the D or O series. As discussed
are thick and the tool is crack-sensitive.
previously, the 8%Cr-0.8%C steels have a dis-
For relieving stresses in EDM-processed
tinct combination of toughness and wear resis-
dies and increasing the toughness of the re-
tance that makes these grades very suitable for
maining white layer, a new tempering treat-
ment should be performed. Its temperature cold work tooling. However, their heat treating
temperatures are considerably different from
should be 30 to 50 C below the maximum
that used in the usual grades.
tempering temperature used in the heat
To illustrate this effect, an 8% Cr steel was
treatment, to avoid hardness loss. Normally,
chosen (commercial name VF800), and various
this procedure is easy to apply and therefore
temperatures were used for its heat treatment.
is highly recommended.
The composition of this grade is shown in
Although both grinding and EDM can damage Table 2. Such conditions were analyzed in
the tool surface, EDM problems are much the laboratory in terms of microstructure and
more common in industrial tool failures. This mechanical properties (measured by a bend
occurs in particular for cold work tool steels and test, Ref 12). Four conditions were applied, as
follows, and summarized in Fig. 16. For all, the This reduces the toughness to the observed
hardness was maintained at 60 HRC: levels. The last condition, 4, produced the lowest
toughness values. This is due to the intense
Condition 1: standard-condition VF800AT grain growth produced by the high hardening
grade, with hardening temperature approxi- temperatures, which are typical for high-speed
mately of 1030 C and tempering at high steels but not applied for this grade. In this
temperatures, 540 C (twice, for 2 h) microstructure (Fig. 16e), the coarse martensite
Condition 2: low hardening and low tem- plates reflect coarse austenite grains. The 8% Cr
pering temperatures, 970 and 200 C, cold work steels, as with other cold work steels,
respectively. This condition is typical for have much lower alloy content than high-speed
high-chromium and high-carbon D grades, steels. This reduces the pinning effect of carbide
such as D3 and D6. precipitation on grain boundaries, thus causing
Condition 3: typical hardening but lower rapid grain growth when high-speed steel hard-
tempering temperature. Although not indi- ening temperatures are used.
cated for 8% Cr steels, this condition is Excessive high hardening temperatures are
typical for D2 steel, a well-known grade for also common problems in heat treating high-
heat treaters. This is used in D2 for attaining speed steels. Hardening temperatures for these
the 60 HRC level, due to the weak secondary steels are close to the solidus temperature (less
hardness of this grade. However, the 8% than 50 C, 90 F), above which liquid forma-
Cr grades normally have higher alloy con- tion starts within the microstructure (in carbide
tent in terms of molybdenum or vanadium, rich areas), leading to expressive embrittlement.
allowing 60 HRC to be obtained after Hardening temperature control is thus very
high-temperature tempering. In the case of important. For example, M2 high-speed steel
VF800AT, up to 63 HRC is possible, de- hardening temperature is about 1200 C
pending on the hardening condition (Ref 13). (2192 F), but exceeding 1220 C (2228 F)
Condition 4: hardening temperature higher may cause loss of toughness without benefits
than normal, using a condition typical to hardness, and above 1240 C (2264 F),
for high-speed steels (when treated to liquation is likely to occur (Ref 38).
60 HRC)hardening at 1150 C and Figure 17 shows two cases of an incorrect heat
tempering at 570 C treatment procedure applied to an 8% Cr tool
steel, for two punches and a cutting blade, that
The results of mechanical properties and cracked prematurely. The microstructure ob-
microstructures for all conditions are shown served was close to condition 3 of Fig. 16, but
in Fig. 16. A substantial reduction is observed the microstructure was very difficult to observe
for conditions 2 to 4 compared to the material after regular (nital) etching. A stronger etching
treated under normal conditions (1). condition was applied. Besides tempering at low
This difference in mechanical properties may temperature, there were evidence of hardening
be understood based on the relative micro- overheating (coarse austenite grain sizes). This
structure for each condition. The first condition is therefore a combination of two incorrect
has a relatively dark martensitic matrix and situationsconditions 3 and 4.
dispersion of primary carbides, undissolved Excessive Retained Austenite Content.
during the hardening treatment. This is typical Hardening of tool steels involves the transfor-
for this material. The dark matrix indicates high- mation of an initial phase, austenite, formed
temperature tempering, where stress relief of during heating (austenitizing treatment). The
martensite transformation is well performed; the following transformations are directly depen-
dispersion of primary carbides is important for dent on austenite composition. Martensite for-
wear resistance. mation is of particular interest, because this is
In the other conditions, the microstructures the expected phase after quenching. Martensitic
show a different aspect. In conditions 2 and 3, transformation is distinct from the usual solid-
tempering at low temperatures is denoted by less phase transformations, because it does not
intense etching, converting to a lighter matrix. occur by the diffusion process. The formation
In these cases, hardness is produced by a highly of martensite depends on the temperature
unstable and stressed martensitic structure attained; therefore, two important temperatures
instead of the secondary hardening of high- are defined: the start and finish of martensite
temperature tempering (adequate condition). transformation, determined by the Ms and Mf
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
temperatures, respectively. If no transformation Retained austenite content is directly depen-

takes place during quenching, martensite is dent on the chemical composition, because it
formed gradually after the Ms temperature is determines the Ms and Mf values. The following
reached, and total transformation of austenite equation (Ref 14) quantifies this dependence
to martensite takes place when the part reaches for Ms, since it is also a similar rate for Mf. It is
the Mf temperature. important to observe that all the alloy elements
In tool steels, most compositions have Mf (with the exception of cobalt) reduce the Ms
values below room temperature; as a result, part values, especially carbon, thus increasing the
of the austenite is not transformed to martensite. amount of retained austenite:
This austenite fraction is called retained auste-
nite. It is distributed within the material micro- Ms ( C)=539 423%C 30:4%Mn 12:1%Cr
structure; after etching, it is observed as a light 17:7%Ni 7:5% Mo
matrix crossed by plates or laths of martensite (all elements in weight percent)
(Fig. 18). As is discussed later, higher harden- (Eq 1)
ing temperatures (overheating) lead to larger
amounts of retained austenite. In the micro- Consider again the phase transformation tak-
structure, it becomes more evident due to the ing place, keeping in mind the effect of alloy
presence of large martensite plates (resulting composition on Ms values. As explained, aus-
from larger austenite grain sizes) crossing the tenite will be transformed to martensite because
austenite phase within the matrix (Fig. 18b). of the rapid cooling of the quenching process.
Fig. 17 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. Micro-
structures etched with Villellas reagent. Courtesy of Villares Metals
Therefore, the composition that affects the Ms composition. For monophase steels, such as
value is the austenite composition at the mo- low-alloy, low-carbon grades, the austenite
ment of martensite transformation, not the alloy composition is practically the alloy composi-
tion. 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 car-
bide dissolution, which depends on time and
particularly on the austenitizing temperatures
involved. The higher the austenitizing tempera-
ture (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 tem-
perature 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 un-
desirable (Ref 17). First, it is softer than mar-
tensite, due to the crystallographic nature of
austenite (face-centered cubic lattice). How-
ever, the effect of the amount of retained aus-
tenite 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 tough-
ness. Due to its unstable nature, austenite at
room temperature tends to transform to mar-
tensite 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 transfor-
Fig. 18 Retained austenite in two cold work tool steels after
mation to martensite may deeply embrittle
hardening and tempering. (a) D2, with 60 HRC. (b)
O1, with 54 HRC. For both, the retained austenite content is working regions, causing cracks or fracture
higher than expected (due to overheating in the hardening
treatment). (c) O1 punch from which the microstructure in (b) was
failures. Third, this transformation of retained
obtained. Cracked in service after short life. Retained austenite austenite also causes a volume increase, due
and carbides are lighter in the microstructure after etching, but to the lattice difference between austenite
they can be differentiated; carbides are either faceted or round,
but retained austenite has no delineated area, being within the (more compact structure) and martensite. This
matrix. In both images, martensite plates that cross retained impairs dimensional stability, which is critical
austenite regions are clearly observed. In (a), the retained auste- for several tools, especially those that work
nite content is high enough to reduce the desired hardness from
60 to 54 HRC. Compare to the usual microstructure of O1 and with small clearances, such as fine blanking
D2 in Fig. 6. dies.
(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
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 stabiliza-
tion of retained austenite (Ref 15) and a second
tempering must always be applied. At the second
tempering temperature, this untempered mar-
tensite (also known as fresh martensite)
or bainite is tempered, being the final micro-
structure 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 elimi-
nate retained austenite. However, in AISI O or
D steels, tempering to 60 HRC or more is nor-
mally done at lower temperatures (5350 C),
where some amount of retained austenite will
exist after tempering; in these cases, the control
Fig. 20 Hardness and retained austenite as a function of
of retained austenite content should be done
tempering in A2 tool steel. Source: Ref 18
by not exceeding the hardening temperatures.
Nevertheless, in some cases, it is possible to lose
some hardness by using higher tempering tem-
Besides these three undesirable effects of peratures to reduce the amount of retained
retained austenite, the cause of excess retained austenite. This can be done for D2 steel if tem-
austenite should also be considered when ana- pered at approximately 520 C (Fig. 7); hard-
lyzing a failure. As discussed and shown in ness would be up to 58 HRC, but the amount of
Fig. 19, the increase in retained austenite content retained austenite decreases from approximately
in cold working tools is normally caused by 15% to less than 5% (Ref 18), considerably
exceeding the recommended hardening tem- improving toughness (Ref 19).
perature. However, the retained austenite con- Application of cryogenic treatments is also
tent may also be reduced when it is converted to a possibility to reduce retained austenite
martensite (or bainite) after the first tempering. content, especially in cold work tool steels that
are tempered at low temperatures. In cryogenic is not so obvious. After austenitizing and
treatments, tools are led to very low tempera- quenching, the steel is hardened to a very high
tures, approaching or crossing the final Mf hardnessin many cases, the highest hardness
temperature. They are normally performed after possible to attain for a given steel. After tem-
a stress relief (at about 150 C) to avoid tool pering, no significant differences can be ob-
cracking. After the cryogenic treatment, tem- served in hardness measurements, especially for
pering is also necessary to avoid brittleness from cold work steels, which have a work hardness
the just formed fresh martensite. very close to the as-quenched hardness. This can
In summary, the presence of retained auste- lead to several problems regarding the embrittle-
nite in unusually large amounts is an indication ment of a tool caused by poor tempering prac-
that either the hardening or tempering treatment tice.
has been inadequately conducted. Other sources During tempering, several solid-state pheno-
of brittleness can emerge in such situations, mena occur simultaneously in the steel micro-
besides the retained austenite itself. For structure. Depending on the alloy content of a
example, higher hardening temperatures pro- tool steel, the tempering curve presents a dif-
mote coarse grain sizes and also may increase ferent aspect, as shown in Fig. 22 by a division in
the potential for carbide precipitation on auste- classes. Class 1 is typical for high-carbon, low-
nite grain boundaries; both effects cause intense alloy tool steels, class 2 for high-chromium cold
embrittlement (condition 4, Fig. 16). Incorrect work steels, class 3 for high-speed steels with
tempering, with shorter times (see next section) strong secondary hardening, and class 4 for hot
or incorrect temperatures (previous section), work tool steels. Typical cold work steels, from
changes the stress relief of the martensite AISI D or O series, will have curves close to
structure and the whole strengthening mecha- classes 1 or 2.
nism, thus affecting material toughness as well As shown and discussed in Fig. 22 and
(condition 2, Fig. 16). quantitatively in Fig. 7, both AISI O and D
Therefore, several examples are observed grades must be tempered at lower temperatures
of industry failures assisted by retained to attain the highest hardness levels60 HRC
austenite that emerged from incorrect proce- or higher. Tempering is typically conducted
dures. Figure 21 presents two examples. The at temperatures of approximately 200 C. In
first, (Fig. 21a, b) shows an AISI O6 tool that practical situations, this is not that simple. The
cracked after limited service. Retained austenite heat flux at such low temperatures is also low,
is clearly observed in its microstructure. The and thus, tempering of dies may take several
second tool in Fig. 21 is an AISI S7 die. This hours. As a consequence, it is common to find
grade has lower carbon (~0.50%) and much examples of tools that were insufficiently tem-
lower undissolved carbides in comparison to the pered. Two cases are shown in Fig. 23.
O- or D-series steels. It should be less prone to Avoiding these kind of failures is not techni-
retained austenite formation and to the effect of cally difficult, but it is time (and money) con-
incorrect heat treating conditions. However, this suming. A basic rule should be observed: Any
tool was carburized, and a surface pickup of operation should not be conducted with the steel
carbon took place, leading to a reduction of Ms in the as-quenched condition. If the tempering
and Mf and thus causing the high amount of time is unknown, a rule of thumb to observe
retained austenite. In both cases, overaustenitiz- is that a steel is hardly ever overtempered (if
ing conditions were employed, enabling the the time is exceeded) but can be easily under-
existence of such high retained austenite content tempered (if the time is too short). This happens
and leading to embrittlement. because the tempering parameter, as described
Excessively Short or Absent Tempering. in Fig. 22, has the time in log scale. However, the
Tempering, the final heat treatment step most desirable situation is to have good control
applied to tool steels, is defined as the heating of the tool temperature during heat treating.
of a martensitic or hardened steel to some tem- One example is monitoring tool heating and
perature below A1 temperature (initial tempera- treatment time with thermocouples attached
ture of austenite formation); this step produces to the part; ideally, all tool regions should be
the final structure and mechanical properties of maintained at temperature for approximately
a hardened steel. This citation, from Ref 1, 2 h.
briefly explains the importance of tempering This discussion of tempering times is im-
treatment. However, in practical situations, this portant, due to the intense microstructural
transformations that occur in the initial temper- Heat Treating Failures of Hot Work Tools
ing stages. Carbon rearranges into martensite
crystals, and transition carbides are precipitated Chemical Composition and Main Charac-
that are only 2 to 4 nm in size. These transfor- teristics of Hot Work Tool Steels. Heat
mations considerably enhance material tough- treating of hot work tools is usually more critical
ness, but hardness decreases only slightly from than for cold work tools. Hot work tools are
the initial as-quenched hardness. In conclusion, normally larger and have higher machining
low-temperature tempering must not be sup- costs, besides being applied to high-demand
pressed based on hardness but kept due to the applications. Although cold work tools may also
necessary changes in steel microstructure that be critical, such as for drawing dies and indus-
it promotes. trial cutting blades, they usually do not exceed
Fig. 21 (a) AISI O6 graphitic tool steel punch machined from centerless-ground bar stock that cracked prematurely. (b) Micro-
structural 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
more than 500 kg. As discussed in the previous Toughness in hot work tools is an essential
section, it is useful to first describe the heat property for avoiding cracks and fractures that
treating conditions of the common grades and can be common in normal working conditions.
then to discuss failure analysis. For this reason, Unexpected and very unstable in cold work
Fig. 24 shows tempering curves for the most tools, cracks may be encountered after a given
common grades, whose chemical compositions operation time for hot work dies. The most
are shown in Table 3. important example is fine crack networks, large
By analyzing chemical compositions, one can in number but small in length, observed on a tool
anticipate several characteristics of hot work surface after the tooling operation. This kind of
tool steels and their differences from cold work crack is known as heat checking and is typically
steels. First, hot work steels normally have lower found in die-casting dies (Fig. 25) as well as in
carbon, leading to lower as-quenched hardness; forging dies or other hot work tools. The cause
also, undissolved carbides will be much lower or is thermal fatigue at the tool surface, caused by
even nonexistent in these grades. Thus, abrasive repeated temperature fluctuations (heating and
wear resistance, which depends on hardness cooling) during tool operation. A full explana-
and coarse primary carbides, is much lower in tion of this behavior is found in the literature,
hot work steels. Final hardness in hot work for example, Ref 20 to 22. Depending on the
tool steels is normally determined by desired intensity, heat checking can lead to tool failure,
toughness instead of wear resistance. In most which is typically the case in die-casting cavities
situations, hardness is between 40 and 50 HRC. (Fig. 25). In such situations, the casting metal,
Other properties, in addition to mechanical normally aluminum, enters into the heat check-
strength at room temperature and wear resis- ing cracks, making part extractions difficult or
tance, are important for hot work tool steels. The impairing part surface quality. In cases of a lack
main metallurgical properties are toughness of a toughness, gross cracking may also occur,
and strength at elevated temperatures; this last leading to complete loss of the working tool.
property is dependent on the hardness stability High mechanical strength at high temperature
at high temperatures. 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 dis-
located to the right (Fig. 24).
As shown schematically in Fig. 26, such high-
temperature 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 pre-
cipitation), the higher the tempering resistance
of hot work tool steels. Such precipitation
intensity depends on the amount of alloy ele-
Fig. 22 Schematic diagram of hardness versus tempering
ments in solid solution, which is related to the
temperature (assuming constant time at each tem-
perature) or versus a time-temperature tempering parameter for alloy composition and heat treating practice.
four major types of tempering response in tool steels. Classes 1 to Figure 27 explains this effect by comparing the
4 are typical, respectively, for low-alloy tool steels, cold work
chromium steels, high-speed steels, and hot work tool steels. addition of different alloy contents to carbon
Source: Ref 1 steels; a stronger precipitation hardening is
obtained by molybdenum, vanadium, or tung- boundary embrittlement by carbide precipita-

sten alloying. tion. Tempering, on the other hand, should
Therefore, the desired mechanical properties eliminate retained austenite and promote
of hot work tool steels are only attained if a adequate precipitation of the alloy carbide. In
proper heat treatment is applied. For example, summary, tool steel properties and the expected
adequate secondary hardening will only be performance are only possible after a quality
present if alloy elements are in solid solution in heat treatment. Otherwise, failures may occur,
the as-quenched structure (usually martensite). reducing die life and increasing tooling costs.
This is only possible through an adequate aus- The following sections describe some typical
tenitizing treatment for hardening, capable of failures of tool steels after heat treatment and
dissolving the alloy elements present in the form the main mechanism that caused failure. The
of carbides in the initial (annealed) state. In main mechanism is described and divided sys-
addition, a proper hardening procedure should tematically, but in practical situations, several
avoid excessive grain growth and grain 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
investigated to discover the root cause of a tool steel. In Fig. 24, this can be observed for
failure. high tempering temperatures, but the variation
In specialized literature, it is less common to is even higher if the whole tempering curve is
discuss failures of hot work tools than of cold observed (see the tempering curve for H13,
work. This is because of the intrinsic lower Fig. 28, for example). Normally, it is possible to
brittleness of hot work tool steels. Nevertheless, attain values between 30 and 58 HRC in the most
the subject is very important for these materials, common hot work steels. However, the steels are
due to the high value of a tooling set for hot not used within this hardness range. Although
working and also because of the large pro- some variation may exist, hardness levels higher
duction performance expected from it. Even if than 50 HRC or lower than 40 HRC are not
no catastrophic failure occurs, the reduction of typical.
hot work tool performance can cause serious To improve wear resistance, one can propose
damage in terms of cost and time. an increase in hardness. However, two aspects
Premature Cracking Caused by Excessive should be considered before such a decision.
Hardness. A wide range of hardness levels can First and most obvious is the toughness nece-
result from hardening and tempering a hot work ssary for a given application. Reduction of
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 low-
silicon 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) DIN UNS C Si Mn Cr Mo W V Fe Others
~H10 1.2365 T20810 0.32 0.30 0.3 2.9 2.8 ... 0.50 bal ...
H11 1.2343 T20811 0.37 1.0 0.3 5.0 1.3 ... 0.90 bal ...
H13 1.2344 T20813 0.40 1.0 0.3 5.0 1.3 ... 0.45 bal ...
Low-Si H11; ... 0.36 0.3 0.3 5.0 1.4 ... 0.45 bal ...
TENAX300(b)
~6F3 or ~L6 1.2714 ... 0.56 0.3 0.7 1.1 0.5 ... 0.15 bal Ni = 1.7
H21 1.2581 T20821 0.32 0.3 0.3 3.5 ... 9.0 0.50 bal ...
(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.
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
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
Fig. 26 Schematic of hardness after tempering and the effect showed a strong hardness decrease in the work-
of secondary hardening in high-alloy steels. Observe ing regions, indicating that a higher tempering
that high-temperature hardness is only possible through pre-
cipitation hardening caused by alloy carbides (secondary hard- resistance was necessary instead of higher
ening). At low temperatures, hardness is less than martensite due hardness. Such an alteration was done, sub-
to the presence of retained austenite, which is eliminated after
high-temperature tempering.
stituting H13 with a higher-molybdenum
grade, the commercial brand VHSUPER (not
standardized), which led to a 50% increase in the
toughness may accelerate several mechanisms tool performance.
of tool damage, such as heat checking, gross It is commonly found that excessive hard-
cracking, and wear by chipping, in addition to ness assists failures in hot work applications.
the risk of catastrophic cracking. Figure 30 shows further examples. For all, low
The relationship between hardness, tough- tempering temperature, short tempering time,
ness, and tempering temperature can be quanti- or even specification changes led to an increase
tatively evaluated in Fig. 28 for AISI H13. When in hardness and premature failure due to the
the tempering temperature is far from the hard- resulting low toughness.
ness peak, toughness rapidly increases with While excessively low hardness can lead to
decreasing hardness. Temperatures of approxi- problems, hardness also should not be too low.
mately 500 C can be considered to cause tem- The adequate value, as mentioned previously,
per embrittlement, with toughness being rather is normally between 40 and 50 HRC in hot work
low (Ref 2426). Such temper embrittlement tools; in some cases for warm forging, it
has been studied in relation to silicon content is possible to use tools up to 58 HRC. An inter-
(Ref 2730). It has been shown, for example, esting effect of hardness and toughness in final
that reducing the silicon content from 1% (usual performance can be discussed based on Fig. 31.
in H-series grades) to approximately 0.3% This graph shows the effect of these two prop-
causes a strong toughness increase, close to peak erties on heat checking, focused on die-casting
hardness but also for higher tempering tem- applications. The higher the hardness and
peratures. toughness, the lower the heat checking damage,
Secondly, the type of wear should be con- measured by an appropriate chart. However, as
sidered. In cold work tooling, wear is directly discussed and shown in Fig. 28, hardness and
related to hardness, but in hot work tooling, the toughness are also associated. Establishing ideal
hardness thus depends on several conditions of work steels has several important parameters,
tool design and use and is also based on previous but incorrect practice does not always produce
experience. As a starting value, 45 HRC would failures observable just after heat treating. Some
be recommended, with increases or decreases examples were shown in Fig. 29 and 30; several
depending on the results experienced. tools were improperly heat treated (to higher
Inadequate Heat Treating Procedures in hardness), but the problem was only observed
Hot Work Tool Steels. Heat treating of hot during tool use.
Fig. 27 Effect of vanadium molybdenum, tungsten, and chromium additions on secondary (high-temperature) hardness of medium-
carbon steels. Source: Ref 23
This is common for hot work tools due to the

low as-quenched hardness and the high hard-
enability of this class of tool steels. For example,
the use of improper temperatures may lead to the
specified hardness, which can even be homo-
geneous, but the mechanical properties, such
as toughness or tempering resistance, can be
deeply affected. A poor practice during quench-
ing can lead to high brittleness with no observed
changes in hardness or strength. These two as-
pects are described in this section.
Before continuing, a useful recommendation
should be made. Due to the high value of die-
casting 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 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
Fig. 28 Impact toughness and hardness as a function of
decrease. The substitution lead to 50% longer tool life. Courtesy
tempering temperature. Retained austenite content is
also shown. Notice the hash-marked area, indicated as a temper of Villares Metals
embrittlement region, where very low toughness is observed; this
region coincides with the peak hardness. Source: Ref 24
information for quality assurance of tool steel that reach up to 900 C. For this condition,
and also for applied heat treating. It is very tempering should be reduced to 250 C.
helpful for die-casting applications as well as for Condition 3: higher hardening temperature,
quality analysis of AISI H-series steels applied increasing from 1020 C to 1150 C. Such
to other processes. temperature is currently used for high-speed
Incorrect Hardening and/or Tempering steels heat treated to lower hardness. For 45
Temperatures. To analyze the effect of incor- HRC, tempering was slightly increased, to
rect temperatures on hot work steel properties, 640 C, also twice for 2 h.
several studies were conducted in the labora-
tory (Ref 12). Toughness was evaluated by Figure 32 evaluates the reduction in toughness
unnotched impact testing, according to the promoted by inadequate heat treating conditions,
NADCA procedure (Ref 32). Three conditions as well as the respective microstructures. The
were simulated. To show the effect of low low-tempering situation, condition 2, causes a
hardening temperature on tempering resistance, substantial reduction in toughness (40%) as well
conditions 1 and 2 were evaluated: as a loss in tempering resistance (the hardness
decrease was six times higher than expected).
Condition 1: standard condition for H13 Toughness reduction in this condition, is caused
grade, with hardening temperature at by incomplete austenite transformation, causing
1020 C and double tempering at 610 C a heterogeneous microstructure (Fig. 32c), as
(for 2 h at temperature), leading to 45 HRC well as by the low tempering temperature, which
Condition 2: Low hardening temperature (at does not promote adequate martensite stress
890 C), simulating a condition for furnaces relief. On the other hand, the decrease in
Fig. 30 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
tempering resistance is caused by the insufficient the dislocation in tempering curve indicates a
dissolution and reprecipitation of alloy carbides substantial increase in tempering resistance,
(poor secondary hardening), which is the main because temperature effect is exponential to
strengthening mechanism in hot work tool steels time effect in tempering conditions. This phe-
at high temperatures. nomenon also explains some advantages found
The most intense embrittlement was pro- in specimens austenitized at 1100 C compared
duced by condition 3too high hardening to 1020 C (Ref 22). However, modifications in
temperature. In this situation, very coarse aus- hardening temperatures are rarely possible in
tenite grains are produced (Fig. 32d), leading to practical (industrial) conditions. Increasing
increased grain-boundary embrittlement. When the hardening temperature deeply affects the
high hardening temperatures are used, precipi- precipitation behavior on grain boundaries dur-
tation of proeutectoid carbides on grain bound- ing quenching, causing intense embrittlement.
aries intensifies, causing a great reduction in Figure 33 shows this effect (note the dashed
toughness, as observed (90% lower). lines), but a full explanation is given in the fol-
Some sources studied the use of higher hard- lowing section (especially regarding Fig. 36).
ening temperature as one way to improve ther- Before continuing, it is interesting to show a
mal fatigue (Ref 22, 33). In fact, the increase case of failure caused by excessive hardening
in hardening temperature leads to more dis- temperature. Figure 34 presents such a casea
solution of alloy carbides, rich in vanadium and tool that cracked after low performance. It is
molybdenum, which increases the content of easy to see the grain-boundary crack propaga-
alloy elements in solid solution and enhances tion, caused by the coarse grain size as well as by
secondary hardening. A simple observation of precipitation of carbides on grain boundaries.
this is shown in condition 3, where tempering This is denoted by the preferential and strong
should be increased 30 C to attain the same etching of austenite grain boundaries.
hardness as the usual hardening condition. Slow Cooling during Quenching. Reach-
In terms of tempering resistance, 30 C (54 F) ing the final hardness in tool steels is quite a
Fig. 31 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
simple task for quenching. Although there are quenching practices, even using air quenching.
some exceptions, tool steels normally have However, two further important aspects should
high hardenability, with the as-quenched hard- be considered. First, the dashed lines in the CCT
ness attained even if improper procedures are diagram (Fig. 33) indicate formation of proeu-
applied. However, hardness, although necessary, tectic carbides. As mentioned in the previous
is not sufficient for the high required performance section, if austenitizing temperature increases,
of tool steels, as already shown. In this context, more alloy elements go into solid solution by
quenching should be considered as a process for carbide dissolution. In cooling, the process is
promoting the required mechanical properties, reversed, and such carbides tend to form again.
not only for attaining a specified hardness. This happens by precipitation in high-energy
For this reason, consider Fig. 33 again. For areas, the most important being the grain
both hardening temperatures, hardness higher boundaries. The result is a film of carbides
than 500 HV (~49 HRC) is obtained within a between grains, which weakens the interface
wide range of the continuous cooling tranfor- and promotes failure (Ref 35). Such a phenom-
mation (CCT) curve. Considering the hardness enon is marked by two characteristics: a strong
reduction after tempering, it is therefore quite etching at austenite grain boundaries (because
simple to obtain the final hardness with different the interfaces of carbides and steel are regions
(b) condition 1 (c) condition 2 (d) condition 3
Fig. 32 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
more prone to corrosion) and, in stronger cases, cooling also affects the previous microstructure,
intergranular failure. As shown in Fig. 33, an forming bainite instead of martensite. Although
increase in austenitizing temperature causes the precipitation necessary for secondary hard-
dislocation of the dashed lines to the left, ening is practically not affected (Ref 33),
indicating stronger precipitation, even when modification of the initial microstructure, from
high cooling rates are applied. Secondly, a slow martensite to bainite, also reduces toughness
Fig. 33 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
(ref 29, 33 and 36). The effect of decreasing the Therefore, both bainite and carbide pre-
cooling rate is thus embrittlement for both cipitation on grain boundaries must be avoided
mechanisms. 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 bound-
aries (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 treat-
ing 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 embrittle-
ment increases as the section size of the tool
increases. Figure 36 shows the effect of section
size and austenitizing temperature on the tough-
ness 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
Fig. 34 Die failure caused by excessive hardening tem- unavoidably sensitive to embrittlement in core
perature. Two tools were analyzed: one made of regions. Therefore, heavy-section tools with
VHSUPER steel (commercial brand name) and the other of AISI
H13. (a) One of the tools cracked in the position denoted by the deep engravings should be heat treated only after
arrow, where a sample was cut for analysis. (b) Typical micro- rough machining to avoid embrittlement of tool
structure from H13 tool and (c) from VHSUPER tool with 54 HRC working regions. The most important example
(45 HRC was expected). Note the coarse grain size, approxi-
mately ASTM 3 to 4. For these grades, grain size is expected to be in this field is die-casting dies. They are usually
approximately ASTM 7 to 10. Courtesy of Villares Metals heat treated only after machining to improve
(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
toughness, because die resistance to heat check- Nitriding is currently used in several tools and
ing is directly related to this property (Fig. 33). dies to improve wear resistance. However, as
Even if the quench is applied after machining, hardness increases, surface toughness decreases,
it is still important to control the Quench cooling and for deep tools, especially those prone to
rate (avoiding too slow conditions), due to the cracks, nitriding is prevented. In die casting,
possibility of embrittlement of surface regions, nitriding is used in some cases, claiming that the
which are the working areas of die-casting dies increase in surface hardness tends to increase the
(Ref 32). initiation of heat checking cracks; however, as
Failures of Nitrided Tools. Nitriding is cracks cross the nitriding layer, it has no effect at
commonly used for several tools and dies. For all. Depending on the nitriding layer condition, it
example, hot extrusion dies are typically nitrided can be harmful to crack initiation.
for all uses. They normally work for production Normally, the nitride layer in tool steels is 0.1
of parts in aluminum or other nonferrous alloys, to 0.3 mm thick. Hardness is higher than in
mainly for construction applications. Different carbon steels, due to the formation of alloy
from other hot working processes, extrusion is nitrides. In hot work steels from the H series,
continuous and involves constant flow between chromium is very important for this effect, and
the conformed alloy, and the tool steel. This the maximum surface hardness approaches 1100
enhances the wear condition, which is the usual HV. Typical nitrided microstructures present a
end-life mechanism. Nitriding considerably diffusion layer and a fine white layer on the tool
improves surface hardness and, consequently, surface (Fig. 37a). In H-series tool steels, the
wear resistance, this being the reason for its
application in virtually all extrusion dies. The
same approach is applied in hot forging tools.
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 dif-
ferent austenitizing temperatures. Toughness reduction is evident Fig. 37 Tool steel surface after nitriding. (a) White and dif-
at higher austenitizing temperatures and larger sizes, both related fusion layers (b) Coarse nitrides precipitated on grain
to lower cooling rates during quenching. Source: Ref 33 boundaries. See text for discussion. Courtesy of Villares Metals
nitriding layer is darker after etching, due to the coarse nitrides commonly leads to premature
depletion of chromium content from the steel crack initiation, with damage to several tools,
matrix, thus causing a reduction in corrosion especially forging tools prone to cracking or
resistance. In carbon or low-alloy steels, the for die-casting dies. Some examples of failures
nitride layer appears lighter in the micro- assisted by this phenomenon are shown in
structure and can increase corrosion resistance. Fig. 38. It is thus recommended, for these
In hot work tool steels, it is rather common to applications and in general, that the nitriding
observe coarse nitrides precipitated on grain process be controlled in order to avoid both a
boundaries inside the diffusion layer, as shown thick white layer and grain-boundary nitrides.
in Fig. 37(b). Depending on the intensity, this Today (2008), this control is usually performed
embrittles the tool and forms an exceptional in computer-controlled gas nitriding and plasma
route for crack propagation (Fig. 42). Preci- nitriding processes.
pitation of coarse nitrides on grain boundaries Excessive Heating Causing Tool Fail-
can also be accompanied by a thick white layer ure. As discussed previously, hardening of hot
on the surface. This layer is extremely hard work tool steel mainly results from precipitation
and brittle, because it is uniquely composed of of alloy carbides during the tempering treatment.
nitrides. It is therefore a common region for This phenomenon is also known as the fourth
crack initiation. The combination of a coarse stage of tempering and occurs after the mod-
white layer and grain boundary precipitation of ification of martensite and the formation
Fig. 38 Examples of undesirable microstructures encountered on the surface of nitrided tools. For both cases, the core micro-
structures 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. Never-
theless, an unexpected failure occurred (after less than 100 shots), caused primarily by improper tool use but also assisted by the nitriding
layer condition.
of cementite. At low temperatures, typically example, consider a forging tool that reaches
below 500 C, the thermodynamic driving force 750 C (~1380 F) for a half-second during each
(free energy) is insufficient for such carbide stroke. After producing 1000 parts, this tool will
precipitation (mainly molybdenum or vanadium be exposed to 750 C for a total time of 8.3 min.
carbides), and no hardness peak is observed Such heating is equivalent to a tempering para-
(Fig. 27). As the temperature increases, carbide meter of 35,260 or 2 h at 691 C (1275 F). In
formation starts and very small carbides (a few H11, such heating is capable of reducing the
nanometers) precipitate within martensite laths. hardness from 46 to less than 30 HRC.
Because they are fine and large in number, these It is in this context that the failures described
carbides are proper for obstructing dislocation in this section should be considered. In some
movement, causing substantial improvement in situations, tools are correctly heat treated, but,
strength. This occurs only at high temperatures, during hot working operation, heating is so
with precipitation of alloy carbides pertaining high that strength is reduced and failure is
to materials used in hot working, in this case, accelerated. One example, shown in Fig. 40, is
hot work tool steels. For higher temperatures, a hot forging punch that usually exhibits low
carbides tend to transform to more stable types or life if compared to the whole set of tools. It is
to coarsen, which increases the size and thus re- made of DIN 1.2885, a highly temper-resistant
duces the total number of precipitated carbides. hot work steel that has high amounts of
This results in loss of the dislocation blocking molybdenum and cobalt for improvement of
effect, leading to strength reduction. The total this property. Even though the analysis showed
effect is schematically shown in Fig. 26. a continuous reduction of hardness approaching
Temperature and time affect the precipitation the punch tip, which contacted the hot infor-
behavior. From a thermodynamic point of view, ging part. Besides the hardness values, it is
the lowest energy is obtained after alloy carbide interesting to note the microstructural behavior.
precipitation (reducing the free energy of ele- Regions far from the hot working areas are
ments in solid solution) and if their sizes are typically tempered martensite, becoming dar-
as large as possible (reducing the surface free ker closer to the punch tip. As previously
energy). Consequently, the higher the time explained, secondary carbides are not visible
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 rela-
tion, it can be seen that temperature actually
has the highest effect; however, if excessively
long times are used at appreciably high tempera-
tures, 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 Fig. 39 Tempering curve as a function of the time-tempera-
effects. This is the situation for several tool ture parameter for H11 steel containing 0.40% C,
0.92% Si, 5.09% Cr, 1.34% Mo and 0.52% V. For this curve,
applications, where tools are in contact at very t = time in hours, and T = temperature in F+460. Source:
high temperatures but for very short times. For Ref 37
under optical microscope; however, their effect was deeply influenced by thermal fatigue crack-
on corrosion resistance (and thus on etching ing (Fig. 41c, d). By comparing heated and
behavior) is evident. High tempering tempera- not heated areas, two further problems were
tures cause an increase in precipitation and observed. First, some regions on the die surface
carbides, leading to reduction of corrosion had been heated to high temperatures, leading to
resistance due to the depletion of alloy ele- reaustenitizing, rehardening, and the formation
ments in the steel matrix and to the increased of untempered martensite (Fig. 42a). Second, the
sites (interfaces) for corrosion. This explains tool surface had serious indications of nitriding
why the microstructure becomes darker where problems, with coarse nitride formed on grain
heating is more intense and hardness is lower boundaries and crack propagations through
(this will always happen if the hardening these regions (Fig. 42c). The combination of
practice is not changed). In the punch in intense surface heating and nitriding led to a
Fig. 40, heating is so high that secondary crazy pattern of hardness, differing substan-
carbides have become large enough to be tially from that encountered in non-heat-affected
observable by optical microscope. This is areas (Fig. 42b). This was caused by the com-
typically the annealed state for this alloy, not bination of rehardening and nitrogen diffusion to
obtained through transformation annealing but the core regions. The end result, of course, was a
by temper annealing. deep embrittlement of the die, as observed by its
The effect shown in Fig. 40 occurs in several intense cracking.
hot working tools. A further example is shown in In cases similar to those in Fig. 40 to 42, two
Fig. 41 and 42 of a forging die for automotive possibilities are possible to solve or reduce the
valves. After some time in production, wear of problem. The first is to employ one steel grade
the forging die produced grooves on the surface with a higher tempering resistance, normally a
of forged products (Fig. 41a), leading to die higher-alloyed grade; however, this approach
substitution. The analysis revealed that wear has some limitations. If the situation is only an
Fig. 40 Microstructures of hot forging punch made of DIN 1.2885. Etched with 4% nital for the same time in all microstructures
increase in tool resistance to softening, changing However, in some situations, heating is so

the steel grade may be a solution. For example, intense during tool application that it is im-
this was the situation described in Fig. 29. possible to solve the problem only by modifying
However, it should be noted that highly alloyed the tool steel employed; thus, there is a second
grades can also be more brittle, leading to possibility concerning cooling applied during
problems if gross cracking or heat checking hot work tooling. Coolants and lubricants are
are important issues for tool failure. Also, both important in hot work tooling, in many cases
the increased alloy is directly related to steel cost, being the same product. In cases where appre-
with an analysis of the cost and benefits of steel ciable regions of tools are reaching more than
substitution being important. The best way is to 700 C or even reaustenitizing, the best practice
consider the total tool cost for each produced is to work on the lubricating/cooling involved in
part, which includes the increase in steel cost the process, before considering changes to the
but also its higher performance as well as the tool steel used. It would be helpful to contact the
reduction of setup times for tool substitution. lubricant manufacturer and obtain updated
Fig. 41 (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
(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
information on the possibilities for a specific Mater., Vol 3 (No. 2), Oct/Dec 2005, p 1218
application. Besides the type of lubricant, a very (Abstract in English; Article in Portuguese)
important issue is the way the lubricant is 4. C.J. Altstetter, M. Cohen, and L. Averbach,
applied. In modern processes, robots may be Effect of Silicon on the Tempering of AISI
used for homogeneous application of lubricants. 43XX Steels, Trans. ASM, Vol 55, 1962,
It is also important to evaluate the possible p 287
changes in tool design to avoid heat during hot 5. G. Dieter, Mechanical Metallurgy, 3rd ed.,
working. The answer is complex and should McGraw-Hill Book Co., 1986
consider tool life, process productivity, and the 6. R.M. Hemphill and D.E. Wert, Impact and
necessary investments, always aiming for the Fracture Toughness Testing of Common
minimum cost per produced part. Grades of Tool Steels, Tool Materials for
Molds and Dies, G. Krauss and H. Nord-
berg, Ed., Colorado School of Mines Press,
Conclusion 1987, p 6669
7. Failure Analysis and Prevention, Vol 10,
All theory and data discussed in this chapter Metals Handbook, 8th ed., American
may be summarized by the following points: Society for Metals, 1975, p 500507
Heat treating of tools and dies may be con- 8. Met. Eng. Q., Feb 1973, p 3141
sidered crucial for their performance. Sev- 9. J.R. Davis, Ed., Tool Materials, ASM Spe-
eral examples show that success cases as cialty Handbook, ASM International, 1995
well as premature failures are often related to 10. Heat Treatment of Tool Steel, Technical
heat treating quality. brochure, Uddeholm Tooling, http://www.
Hardness measurements alone are usually uddeholm.com/files/heattreatment-eng-
not the best indication of proper quality in lish.pdf (accessed Dec 2006)
too steel heat treating. Other properties such 11. Grinding of Tool Steel, Technical bro-
as strength, toughness, and wear resistance, chure, Uddeholm Tooling, http://www.
as well as the microstructural features, are uddeholm.com/files/grinding-english.pdf
better indicators as to whether a heat treat- (accessed Dec 2006)
ment was done correctly or not. 12. R.A. Mesquita, D.R. Leiva and C.A. Bar-
Assessing these properties is usually difficult bosa, Estudos de Tratamento Termico nos
in tools and dies because it would be a Acos Ferramenta VH13ISO E VF800AT
destructive (and expensive) test. Therefore, (Heat Treating Studies of VH13 and
guaranteeing the correct procedure is the VF800AT Tool Steels), Proceedings of
best way to assure that the heat treatment of a Third Encontro da Cadeia de Ferramentas,
tool was performed correctly and thus Moldes e Matrizes, ABM, 2005, p 3040,
enables adequate tool performance. (Abstract in English; article in Portuguese)
In order to apply such correct conditions, the 13. Cold Work Tool Steel, VF800-AT Data-
best approach is to follow the specifications sheet, Villares Metals Company, http://www.
for each grade which is normally provided villaresmetals.com.br/ (accessed Dec 2006)
by the tool manufacturer. For several spe- 14. K.W. Andrews, Empirical Formulae for
cific points such as design, tool use, and the Calculation of Some Transformation
surface treatments, the examples given with Temperatures, J. Iron Steel Inst., Vol 203,
this chapter may be helpful. 1965, p 721727
15. W.J. Harris, Jr. and M. Cohen, Stabilization
REFERENCES of the Austenitic-Martensite Transforma-
tion. Trans AIME, Vol 80, 1949, pp. 447
1. G. Roberts, G. Krauss, and R. Kennedy, Tool 470.
Steels., 5th ed. ASM International, 1998 16. B.L. Averbach, S.A. Kulin, and M. Cohen,
2. Tool Steels, Steel Products Manual, Iron and The Effect of Plastic Deformation on Solid
Steel Society, April. 1988 Reactions, Part II: The Effect of Applied
3. R.A. Mesquita and C.A. Barbosa, Avaliacao Stress on the Martensite Reaction, Cold
das Propriedades de Desgaste e Tenacidade Working of Metals, American Society for
em Acos para trabalho a Frio (Evaluation Metals, 1949
of Wear and Toughness Properties of Cold 17. Heat Treating, Vol 4, ASM Handbook,
Working Tool Steels). Tecnol. Metal. ASM International, 1991
18. Cold Work Tool Steels, Datasheets, Mater. Sci. Eng. A, Vol 394, 2005, p 435
Uddeholm Tooling, Hagfors, Sweden 444
19. A. Mendanha, H. Goldenstein and C.E. 29. M. Umino, T. Sera, K. Kondo, Y. Okada and
Pinedo, The Role of Microstructure on the H. Tubakino, Effect of Silicon Content on
Toughness Behaviour of AISI D2 Cold Tempered Hardness, High Temperature
Work Tool Steel, Proceedings of the Seventh Strength and Toughness of Hot Working
International Tooling ConferenceTooling Tool Steels, Tetsu-to-Hagane (J. Iron Steel
Materials and Their Applications from Inst. Jpn.), Vol 89, (No. 6), June 2003,
Research to MarketTool 06, Vol 2, p 673679
(Politecnico de Torino, Torino, Italy), 2006, 30. R.A. Mesquita and C.A. Barbosa, Effect of
p 813819 Silicon and Phosphorous on the Toughness
20. M.A.H. Howes, Heat Checking in Die of G11 Hot Work Tool Steel, Proceedings
Casting Dies, Die Cast. Eng., March-April of 61st Brazilian Metallurgical and Mate-
1969, p 1216 rials Science Congress, ABM, 2006 (in
21. T.C. Benedyk, D.J. Moracz, and J.F. Wal- Portuguese)
lace, Thermal Fatigue Behaviour of Die 31. L. Eliasson and O. Sandberg, Effect of
Materials for Aluminum Die Casting, Paper Different Parameter on Heat-Checking
111, Proceedings of Sixth SDCE Inter- Properties of Hot-Work Tool Steels, New
national Die Casting Congress, 1970 Materials and Processes for Tooling, H.
(Cleveland, OH), The Society of Die Cast- Berns, H. Nordberg and H.-J. Fleischer, Ed.
ing Engineers, Inc., 1970, p 120 (Bochum, Germany), Verlag Schurmann
22. J. Sjostrom and J. Bergstrom, Thermal and Klagges KG, 1989, p 17
Fatigue Testing of Chromium Martensitic 32. Premium Quality H13 Steel Acceptance
Hot-Work Tool Steel after Different Criteria for Pressure Die Casting, 207/
Austenitizing Treatments. J. Mater. Pro- 2003, North American Die Casting Asso-
cess. Technol, Vol 153154, 2004, p 1089 ciation (NADCA)
1096 33. M.L. Schmidt, Effect of Austenitizing
23. W. Crafts and J.L. Lamont, Secondary Temperature on Laboratory Treated and
Hardening of Tempered Martensitic Alloy Large Section Sizes of H-13 Tool Steel,
Steel, Vol 180, TMS-AIME, 1949, p 741 Tool Materials for Molds and Dies., G.
24. Hot Work Tool Steels, Datasheets, Krauss and H. Nordberg, Ed. (Illinois)
Uddeholm Tooling, Hagfors, Sweden Colorado School of Mines Press Center,
25. J.R.T. Branco and G. Krauss, Toughness of 1987, p 118164
H11/H13 Hot Work Tool Steel, New 34. K.-E. Thelming, Steel and Its Heat Treat-
Materials Processes Experiences for Tool- ing, 2nd ed., Butterworths, London, 1984
ing. H. Berns, M. Hofmann, L.-A. Nor- 35. C.L. Briant and S.K. Banerji, Intergranular
strom; K. Rasche, and A.-M Schiner, Ed., Failure in Steel: The Role of Grain Bound-
Materials Search (Interlaken, Switzerland), ary Composition, Int. Met. Rev., (No. 4),
1992, p 121134 1978, p 164199
26. H. Berns, Strength and Toughness of Hot 36. T. Okuno, Effect of Microstructure on the
Working Tool Steels, G. Krauss and Toughness of Hot Work Tool Steels, AISI
H. Nordberg, Ed. Tool Materials for Molds H13, H10 and H19, ISIJ, Vol 27, (No. 1),
and Dies: Application and Performance, 1987
(Ilinois), The Colorado School of Mines 37. J.C. Hamaker, Jr., Die Steel Useful for Ultra
Press, 1987, p 4565 High-Strength Structural Requirements,
27. W.M. Garrison, Jr., Influence of Silicon Met. Prog., Dec 1956, p 93
on Strength and Toughness of 5wt-%Cr 38. R.A. Mesquita, C.S. Goncalves, and C.A.
Secondary Hardening Steel, Mater. Sci. Barbosa, Effect of Hardening Conditions on
Technol., Vol 3, April 1987, p 256259 the Mechanical Properties of High Speed
28. D. Delagnes, P. Lamesle, M.H. Mathon, N. Steels, Proceedings of the European Con-
Mebarki and C. Levaillant, Influence of ference on Heat Treatment 2008 Inno-
Silicon Content on the Precipitation of vation in Heat Treatment for Industrial
Secondary Carbides and Fatigue Properties Competitiveness, Verona, Italy, 2008, CD-
of a 5% Cr Tempered Martensitic Steel, ROM
Case Studies of Steel Component

Failures in Aerospace Applications
STEEL has limited use in modern aircraft, but Failure Analysis of a Catapult
it is used in landing gear systems, arrestment Holdback Bar
systems for naval aircraft, and for support
equipment. Because of the high strength levels
required (1790 to 2070 MPa, or 260 to 300 ksi, This investigation analyzed the failure of a
yield) to maintain justification for the use of repeatable-release holdback bar. This bar con-
steel, these very high-strength steels are prone to sisted of a failed strain bar and a failed T-head
environmental effects (stress-corrosion cracking (rod end connector). These components failed
and hydrogen embrittlement), flaws created during characterization fatigue testing. The
during manufacturing (laps, seams, machining strain bar and the T-head (rod end connector) had
gouges, and grinding), and heat treatment (dis- been subjected to spectrum loading. This fatigue
tortion, decarburization, and quench cracking). spectrum consisted of two 365 MPa (53 ksi)
Care is needed to protect and inspect these high- tensile loads followed by a single 469 MPa
value and safety-critical steel components. The (68 ksi) tensile load application. The strain bar
following case histories illustrate typical failures failed at 4875 cycles. The T-head failed at 3235
experienced by these high-strength steels. cycles; the desired lifetime for the strain bar and
The case histories in this chapter illustrate the T-head was 4500 cycles.
a variety of failure mechanisms. The causative The failed AISI 4340 steel strain bar and the
reasons vary from manufacturing to operational failed cadmium-plated AISI 4330V steel T-head
to environmental. Because of the many different were submitted for analysis.
types of root causes, the failure engineer must be Figure 1 shows the as-received failed strain
aware of the manufacturing process, the assem- bar. The fracture occurred at the aft radius of
bly process, as well as the environment to which the circumferential retainer ring groove located
component is exposed to effectively determine near a slot in the strain bar. Figure 2 shows
the primary cause of failure. the as-received failed T-head and the location
Fig. 1 As-received strain bar from fatigue testing of a hold-back bar assembly
of the fracture. Fracture occurred in the run-out crack lengths were 22 and 25 mm (0.85 and
area of internal threads at a distance of appro- 0.99 in.). No cracks were detected in the T-head.
ximately 50 mm (2 in.) from the forward end. The radius at the primary fracture of the strain
All of the parts were subjected to a magnetic bar was determined to be 1.3 mm (0.050 in.) At
particle inspection. In the strain bar, two cracks the secondary cracks, the radius was 1.14 mm
were found approximately 180 apart in the re- (0.045 in.). At present, there is no dimensional
tainer ring groove location shown in Fig. 3. The requirement for these radii.
Fig. 2 As-received T-head connector from fatigue testing of hold-back bar assembly
Fig. 3 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
Case Studies of Steel Component Failures in Aerospace Applications / 353
The fracture surfaces were examined visually techniques. Figure 8 shows the microstructures
and at up to 50 magnification using a stereo- of the strain bar and a typical T-head.
microscope. Figure 4 shows macrographs of the The inclusion contents of the metallographic
fracture surfaces of both the strain bar and the specimens appeared to be relatively high, as
T-head. The strain bar primary fracture surface is observed in air-melted steels. An energy-
exhibited multiple fatigue origins around the dispersive spectroscopy analysis identified man-
outside periphery of the fracture. The T-head ganese sulfide inclusions as being present.
fracture exhibited fatigue origins primarily near Manganese sulfide inclusions are typically pres-
internal thread roots. ent in air-melted steels and can severely affect
A scanning electron microscope (SEM) was fracture toughness (Ref 1). Inclusions can act as
used to examine the fracture surfaces and to initiation sites for fatigue failures; however, the
document the mode of failure. Figure 5 shows failure of these parts could not be associated
the fracture topographies of the primary fracture with the presence of inclusions.
in the strain bar. The secondary fractures in the Banding was also present in both the strain
strain bar are shown in Fig. 6. Fatigue striations bar and T-head metallographic specimens.
emanated from fatigue origins located on the Microhardness measurements made on the light
outside periphery of the fractures. Fracture rid- and dark bands in the strain bar and T-head
ges also gave evidence of fatigue origin loca- yielded hardness readings of 44.1 HRC (light),
tions on the outside periphery of the fracture. 42.6 (light), 49.8 (dark), and 49.1 (dark), respec-
Figure 7 shows topographic features of the tively. This microstructure is typical of an AISI
fracture surface of the T-head failures. Fatigue 4340 component that has been improperly
striations emanated from multiple origins loca- normalized (Ref 2). The light etching areas are
ted on the inside peripheries of the fractures. generally untempered or lightly tempered mar-
Hardness measurements were made on the tensite, while the darker etching regions
strain bar and T-heads to verify heat treat con- are martensite that has been tempered more
ditions. The hardness of the strain bar was thoroughly. The lighter etching regions contain
determined to be 41.6 HRC, and this met the higher hardenability than the darker etching
hardness requirement of 40 to 43 HRC. The regions. This is typically due to segregation of
hardness of the T-head was determined to be chromium, generally as the result of inadequate
48.8 HRC. This hardness value exceeded the normalizing prior to heat treatment. This type of
requirement of 46 to 48 HRC. nonuniform structure is not optimal. The harder
Metallographic sections were prepared regions tend to have lower fracture toughness
through the strain bar and a typical T-head at and have higher notch sensitivity than the dar-
the locations shown in Fig. 4. The specimens ker, more tempered regions. Because of this, it
were prepared using standard metallographic was thought that the premature fracture occurred
Fig. 4 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
because of improper heat treatment. The pres- After the chromium plating had been
ence of a higher level of inclusions could also removed, the flange area was subjected to a
contribute to notch sensitivity reduction. temper-etch operation of 1% nitric acid in
methanol. Numerous areas on the flange etched
dark, indicating that these regions had been
Cracking in a Main Landing Gear overtempered (Ref 3). A micrograph of the
Attach Pin temper-etched inspected flange is shown in
Fig. 11.
Multiple pins exhibited cracks in the flanges A typical crack was opened, and the fracture
during magnetic particle inspection as part of surface was examined visually and with a stereo-
the manufacturing process. Some parts were microscope. The fracture surface was discolored
returned to the vendor for rework, while others (Fig. 12). The discoloration appeared to be simi-
were reworked in-house. lar to a temper color in steel (straw or tan). After
The pins were machined from 300M steel. the chrome plating was ground to final dimen-
The pins were then heat treated to a strength sions, the attach pin had received a stress-relief
range of 1930 to 2070 MPa (280 to 300 ksi). operation that could have produced this dis-
The pin was ground and chromium plated. coloration.
Figure 9 shows the main landing gear (MLG) An SEM was used to examine the fracture
lever attach pin as received for examination. The surface. Figure 13 shows the SEM fractographs
magnetic particle inspection indications in the of the surface. The fracture topology was pri-
flange are shown in Fig. 10. marily intergranular.
Fig. 5 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)
A hydrogen analysis was conducted on the Based on the results of this investigation,
MLG lever attach pin before and after baking for it was concluded that the MLG lever attach pin
24 h at 190 C (375 F). The results of the contained numerous probable grinding cracks in
analysis indicated that before baking, the lever the flange.
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
MLG Linear Actuating Rod and Cylinder
considered to be low and will not produce a
hydrogen embrittlement failure (Ref 4). Also, The part was removed from the aircraft and
fractures caused by hydrogen embrittlement are
was reported to have accumulated 50 flight
generally not discolored.
hours. The failed part was from an early pro-
Metallographic sections were prepared
duction lot manufactured with electroless nickel
through the magnetic particle indications using
on the inner diameter surface rather than the
standard metallographic techniques. Shown in
drawing requirement of electrolytic nickel. The
Fig. 14 are areas where the chromium plating
manufacturing sequences for the MLG linear
had been reduced by possible grinding.
actuating piston rod cylinder were:
Figure 15 shows micrographs of a typical crack.
This crack extended through the chromium A 300M steel bar was machined to the re-
plating and into the base metal to a depth of ap- quired outside diameter and was then bored,
proximately 0.25 mm (0.010 in.). There was honed, and nickel plated.
no visual indication of untempered or over- The bar was swaged in an approximate
tempered martensite. 76 mm (3 in.) length on the end opposite
Fig. 6 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)
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 sup- The bar was then finish machined and heat
posedly swaged at 650 C (1200 F) and treated to a tensile strength range of 1930 to
then stress relieved. 2070 MPa (280 to 300 ksi). The part was
then magnetic particle inspected. Electro- intergranular-appearing origin on the inside

deposited nickel was specified for plating on diameter surface of the piston rod cylinder. The
the inside diameter surface. origin was located at approximately midlength
of the longitudinal crack, 200 mm (8 in.) from
This plating is relatively soft, ductile, and the mono-ball end. An arrest mark was present
has a relatively high melting point. However, on the fracture surface at approximately 85% of
electroless nickel, which has high phosphorus the wall thickness.
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 actu-
ating piston rod cylinder components as re-
ceived 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 frac-
tures 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 stereomicro-
scope. 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. 10 Magnetic particle inspection indications on the

flange of the main landing gear lever attach pin
Fig. 9 Macrograph documenting the appearance of the Fig. 11 Overtempered (darkened areas) in the flange
as-received main landing gear lever attach pin revealed by temper etching (2.5 mm)
An SEM was used to examine the fracture

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).
Hardness measurements were made on the
Fig. 12 Discoloration of the fracture surface (330 mm) piston rod cylinder. The hardness was 54 HRC,
and this met the hardness requirement of 53 to 55
HRC.
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)
Fig. 15 Micrographs showing the morphology of the cracks. (a) Overall view (50 mm). (b) Closeup of crack (20 mm)
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
Fig. 16 As-received main landing gear linear actuating most likely diffused to and concentrated at the
piston rod cylinder
crack tip, which caused the intergranular arrest
A hydrogen analysis was conducted on the region to occur. The crack then propagated to
piston rod cylinder. The hydrogen content was failure by ductile rupture.
determined to be 2.3 ppm. This amount of hy- Based on the results of this investigation, it is
drogen is considered sufficient to cause hydro- concluded that the MLG linear actuating piston
gen embrittlement (Ref 5). rod cylinder failure was most likely due to
A metallographic specimen, section 1-1, was hydrogen embrittlement.
removed at the origin location shown in Fig. 19.
The specimen was prepared using standard
metallographic techniques. The microstructure Failure Analysis of AISI 420 Stainless
at the origin is shown in Fig. 20. The micro- Steel Roll Pin
structure of the steel, which consisted of tem-
pered martensite, was normal. Several failures of AISI stainless steel roll
The plating at the origin consisted of three pins were reported. This pin is a standard part
distinct layers. The layers were subjected to an and is used to hold pin components together by
energy-dispersive x-ray analysis. The outermost expanding after compression. The pin is manu-
two layers consisted essentially of nickel and factured from AISI 420 stainless steel and is heat
phosphorus, whereas the inner layer consisted of treated to 46 to 55 HRC.
nickel and iron. The outermost two layers con- Figure 21 shows the as-received failed roll
tained cracks, and the inner layer did not. Also pin. The fracture extended the length of the pin.
evident in the micrograph is plating that had An SEM was used to examine the fracture
covered a small part of the fracture surface. This surface of the service failure and a laboratory-
gave evidence that a crack was present prior to created overload failure from the same lot of
plating or prior to the heating of the plating, material. Figure 22 shows the results of the SEM
which occurred during swaging. examination. In this figure, the topology of
A single intergranular origin was associated the fractures was intergranular, with some dim-
with the fracture surface, and this origin was ples present. This failure mechanism indicated
relatively clean. These factors suggest a delayed that the failure was due to overload (static
mode of failure, as occurs in a hydrogen rupture).
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)
Hardness measurements were made on the of tempered martensite, with carbides out-
roll pin to verify the heat treated condition. lining the prior-austenitic grain boundaries.
The hardness of the roll pin was measured to This is not a typical microstructure for quen-
be 49 HRC, which met the hardness requirement ched and tempered AISI 420 stainless steel
of 46 to 55 HRC. (Ref 2).
A metallographic specimen was prepared Generally, because of their high harden-
through the fracture surface. The specimen was ability, martensitic stainless steels can be quen-
prepared using standard metallographic tech- ched in either oil or air. Oil quenching ensures
niques. The microstructure (Fig. 23) consisted maximum ductility and corrosion resistance. Air
Fig. 19 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)
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 inter-
granular 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
The MLG lever was removed from service

after a hard carrier landing. Multiple cracks
developed during removal of the ion vapor-
deposited (IVD) aluminum with sodium hydro-
xide to analyze for residual stresses by x-ray
diffraction.
Fig. 20 Microstructure at fracture origin (12.5 mm) 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 frac-
ture 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 appear-
ances of origins (some of which were dis-
colored), 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 crack-
ing 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.
Fig. 21 As-received 420 roll pin. (a) Visual view of
The wall thickness at the origin location was
as-received roll pin (2.5 mm). (b) SEM view of
as-received roll pin (200 mm) measured to be 5.8 mm (0.229 in.). The drawing
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) Laboratory-
produced overload failure showing intergranular fracture (20 mm). (d) Laboratory-produced failure showing intergranular fracture with
dimples (5 mm). Compare to (b)
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.).
Hardness measurements were made on the
lever to verify heat treatment. The hardness of
the lever was 54.2 HRC, and this met the hard-
ness 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 hydro-
gen 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, low-
alloy steel.
Fig. 23 Microstructure of failed roll pin. Microstructure
consists of tempered martensite with carbides de- Based on the results, it is thought that the
corating the prior-austenite grain boundaries (10 mm) cracks formed in the MLG lever as a result of
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 Failure Analysis of an Inboard Flap

the IVD coating in addition to the residual Hinge Bolt
stresses present as a result of the hard carrier
landing. It is likely that accelerated stress- An inboard flap hinge bolt was found to be
corrosion cracking occurred because of the high failed after 286.5 flight hours. The bolt had been
residual stresses and electrolyte. machined from 4330V-modified steel bar and
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 por-
tion, which was 23 mm (0.9 in.) in length with
thread type 0.5000-20 UNJF-3A THD in accor-
dance with MIL-S-8879. The unthreaded por-
tion 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 sur-
face 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 resi-
dual 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 fea-
tures. 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
Fig. 25 Appearance of primary crack removed from part. point dimples indicative of ductile rupture were
(a) Region of primary crack where ion vapor-
deposited coating had been removed (13 mm). (b) Primary crack present.
showing branching (2 mm). (c) Fracture surface of crack after A laboratory-produced failure yielded a struc-
opening (3 mm)
ture that consisted only of dimples.
The corrosion products on the service failure
heat treated to the 1515 to 1655 MPa (220 to fracture surface were subjected to an energy-
240 ksi) tensile strength range. A nut with the dispersive x-ray analysis. Calcium, potassium,
bolt was machined from 4140 steel bar heat magnesium, and chlorine were identified as
treated to the 1240 to 1380 MPa (180 to 200 ksi being present.
tensile strength range. Rockwell hardness measurements were made
Figure 28 shows the inboard flap hinge bolt on the inboard flap hinge bolt to verify heat
and nut as received for examination. The bolt treatment condition. The hardness of the bolt
was approximately 102 mm (4 in.) long. The averaged 46 HRC. This met the hardness
Fig. 26 Typical fracture features. (a) Overall view (500 mm). (b) Intergranular fracture apparent at fracture origin (7 mm)
A hydrogen analysis conducted on the hinge

bolt yielded a value of 1.5 ppm, which is con-
sidered 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 nor-
mal. No untempered martensite was present.
Corrosion products were observed on the frac-
ture surface at the origin (Fig. 32). However,
there was no evidence of pitting or other corro-
Fig. 27 Metallographic examination of primary crack. Frac- sion processes that would have produced the
ture path followed prior-austenite grain boundaries.
No precipitates in grain boundaries (50 mm) were observed. corrosion products. No aluminum IVD was
present on the surface at the origin, because the
IVD had been removed prior to measuring
requirement of 46 to 48 HRC for 4330V- residual stresses.
modified steel heat treated to the 1515 to Based on the results of this investigation, it is
1655 MPa (220 to 240 ksi) tensile strength concluded that the inboard flap hinge bolt failed
range. The hardness of the nut was 45 HRC, in a delayed mode of failure, which was most
which was above the hardness requirement of likely stress-corrosion cracking. No material
40 to 43 HRC. The high hardness value did anomalies were observed that would account for
not appear to be detrimental. the failure of the inboard flap hinge bolt.
Fig. 28 As-received flap hinge bolt and nut
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 orig-
inated at the discolored region. It was deter-
mined that the discolored region, from which the
failure appeared to originate, was located at the
bottom of the axle, at approximately the 6:30
Fig. 29 Fracture surface of the failed flap hinge bolt 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 stereomicro-
Failure Analysis of a Nose Landing scope at up to 30 magnification confirmed that
Gear Piston Axle the failure of the axle originated at the discolored
region. Chevron marks were evident that indi-
This investigation determined the mode and cated the crack propagated away from the ends
initiating cause of failure of a nose landing gear of the discolored region, terminating at an area
(NLG) piston axle. Failure of the piston axle diametrically across from it, as shown in
occurred on the aircraft during taxiing after the Fig. 35(b). The discolored region extended
aircraft had accumulated approximately 60 h in along the circumference of the axle for a cord
service. When failure occurred, the nose wheel distance of 22.15 mm (0.872 in.) and a maxi-
completely separated from the piston strut. The mum depth from the inside diameter surface
NLG piston assembly was fabricated from 300M of 3.38 mm (0.133 in.). Actual wall thickness
steel and heat treated to a 1930 to 2070 MPa (280 at that location was 3.81 mm (0.150 in.). The
to 300 ksi) ultimate strength range. The axle plane of the discolored region was approxi-
was then shot peened and low-embrittlement mately 25 from being perpendicular to the
cadmium plated. White paint was applied to the longitudinal axis of the axle, along machining
inside and outside surfaces of the axle, with marks. These dimensions were shown in Fig. 33
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 the SEM. The locations where SEM examin-
showed a color variance ranging from a gold ations were performed are shown in Fig. 35(a).
color to reddish brown. Also, small patches of Prior to SEM examination, the discolored region
blue were evident. This observed color variation was cleaned with acetone and replication in an
is a characteristic typically observed on steel attempt to remove the scale from the fracture
surfaces that have been exposed to elevated surface. After several attempts, only a small
temperature, such as during heat treatment amount of the scale could be eliminated, indi-
(Ref 1). The machining marks indicate a surface cating the scale was firmly attached to the frac-
much rougher than is normal. ture surface. The SEM fractographs adjacent to
The hardness measurements were made at the inside diameter surface (area 1) were pri-
several locations around the circumference of marily intergranular and appeared to be covered
the axle adjacent to the fracture surface. The with scale. At areas beyond the inside diameter
hardness measured 53 to 55 HRC, which con- surface, the SEM fractographs revealed a mixed
formed to the drawing requirement of 53 to 55 intergranular with transgranular features and
HRC for 300M steel heat treated to a 1930 to patches of scale. These are shown in Fig. 37.
2070 MPa (280 to 300 ksi) condition. Along the periphery of the discolored region,
A section containing the discolored region fatigue striations could be observed at a higher
was removed from the axle and examined with magnification, as shown in Fig. 38. The depth of
the fatigue growth region was approximately indicates that the fracture did not initiate sub-
0.05 mm (0.002 in.). Rapid fracture extended surface, as would be the case for delayed failure
beyond the fatigue region. The SEM fractograph resulting from hydrogen embrittlement.
in Fig. 39 shows dimpled features that char- In an attempt to determine the composition
acterize a ductile mode of rapid fracture. When of the observed scale or to identify any con-
examining the discolored region with SEM at taminant that may be associated with the fracture
low magnification, there was no evidence of surface, an energy-dispersive x-ray (EDX) ana-
shear lip along the inner diameter surface, which lysis was performed on the discolored region.
The EDX analysis revealed no other element
except the ones common to the base metal
composition.
Fig. 32 Micrograph showing the microstructure at the frac-

ture origin. Microstructure consists of quenched and Fig. 33 As-received nose landing gear piston assembly and
tempered martensite (25 mm). IVD, ion vapor deposited the failed axle
Fig. 34 Discolored region of the fracture

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 dis- secondary crack, the crack was opened and the
colored region, a section was removed (section fracture surface examined with a light micro-
A-A, Fig. 35b) for metallographic examination. scope and the SEM. The fracture depth was
The microstructure was tempered martensite, relatively small (0.25 mm, or 0.01 in.), which
which is typical for 300M steel heat treated to made it difficult to determine with the light
the 1655 to 2070 MPa (280 to 300 ksi) ultimate microscope if a high-temperature scale was
strength range (Fig. 40). A secondary crack was present. The SEM examination revealed inter-
observed starting at the inner diameter surface granular and ductile transgranular features
(Fig. 40b), at a location showing large surface (Fig. 41). These features suggest a secondary
irregularities. In an attempt to investigate this heat treat crack.
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 spectro-
scopy method. The results of the analysis are
shown in Table 1. These values met the require-
ments 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 Fig. 39 Rapid fracture at location outside the discolored
piston axle during or prior to the heat treatment region (5 mm)
process. Because of the machining marks and machined from 4330V-modified steel that had
the initiation of cracking at these asperities, it been heat treated to a 1240 to 1380 MPa (180 to
is possible that cracking occurred during the 200 ksi) tensile strength range.
machining operation, in a fashion similar to the Figure 42 shows the adapter part of the mul-
formation of grinding cracks. tiple-leg aircraft-handling sling as received for
Based on the results of this examination, it is examination. Also shown in Fig. 42 is the loca-
concluded that the failure of the NLG piston tion and appearance of the crack that was at a
axle was due to the pre-existing crack on the clevis. The crack penetrated completely through
axle. This crack was created prior to or during the wall of the clevis.
heat treatment of the part. The fracture surface was opened, and the
fracture surface was examined visually and at up
Multiple-Leg Aircraft-Handling Sling surface. Fracture ridges indicated that origins
were present on the outside surface of the clevis.
A new multiple-leg aircraft-handling sling The fracture surface appeared faceted, which is
that had just received proof testing was found to characteristic of an intergranular, delayed mode
have a cracked clevis. The failed adapter was 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)
An SEM was used to examine the fracture low in relation to producing a hydrogen em-
surface and the side of the clevis at the origin. brittlement failure.
Figure 44 shows SEM photographs document- A metallographic specimen was prepared
ing topographic features of the fracture surface through a fracture origin. The specimen was
and corrosion pitting on the side of the clevis prepared using standard metallographic techni-
at the fracture origin. The pits were coated with ques. Figure 46 shows the microstructure, which
primer, which indicated they were present was normal (tempered martensite).
prior to painting. Also documented in Fig. 44 Also shown in Fig. 46 is a corrosion pit that
are intergranular topography and corrosion pro- was coated with paint, indicating the pit was
ducts at the fracture origins. These features are present prior to painting.
all characteristic of a stress-corrosion failure Based on the results of this investigation,
(Ref 4). it was concluded that the adapter failed due to
A laboratory overload fracture produced in stress-corrosion cracking. Corrosion pits at the
the clevis was examined in the SEM. Figure 45 fracture origin were present prior to painting.
shows an SEM fractograph documenting dim-
ples that are characteristic of an overload mode
of failure.
Hardness measurements were made on the Failure Analysis of an Aircraft Hoist
adapter to verify the heat treat condition. The Sling during Static Test
hardness of the adapter was 41.4 HRC, and this
met the hardness requirement of 40 to 43 HRC. An aircraft hoist sling was successfully tested
The hydrogen content of the adapter was to an ultimate load of 136,000 kg (300,000 lb).
determined to be 1 ppm, which is considered 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.
Table 1 Chemical analysis of nose landing
Figure 47 shows the as-received failed portion
gear piston axle
of the test fixture. The failure occurred in a
Chemical composition, %
consumable electrode weld that traversed
Si Mn Ni Cr Mo V Fe
around the length of the fracture, which was
1.42 0.72 1.79 0.68 0.49 0.008 bal
approximately 69 cm (27 in.). A dark, dis-
colored 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. 42 As-received failed multiple-leg aircraft-handling Fig. 43 Appearance of the fracture surface of the multiple-
sling leg aircraft-handling sling
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 pat-
terns, at a typical location. The side of the
fracture containing location C had been placed
in tension due to bending. Portions of the frac-
ture at locations A, B, and C shown in Fig. 47
were excised and examined on an SEM. Fig-
ure 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 loca-
tion 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. Fig. 45 Dimpled rupture indicating overload failure in a
An energy-dispersive spectrographic (EDS) laboratory-produced failure. Original magnification:
5000
analysis conducted on the fractures at A and B
basically detected iron, chromium, nickel, and
copper. Nickel was not detected at location C. In conclusion, a different consumable weld-
The EDS spectra are shown in Fig. 51. The base ing electrode was used at location C (no nickel
metal also showed the presence of iron, chro- present) than at locations A and B, where nickel
mium, nickel, and copper, which are present in was present. It is recommended that 17-4PH
17-4PH stainless steel. filler metal be used when welding 17-4PH steel.
This was apparently done at locations A and B Figure 52 shows the internal gear as received
but not at C, where no nickel was present. Nickel for examination. Also shown in Fig. 52 is the
gives ductility to the weld and helps to prevent appearance of typical cracks, which were loca-
cracking during cooling from the welding tem- ted on the inside of the internal gear on areas
perature (Ref 7). It appears that at location C, next to the gear spline. The cracks were located
where no nickel was present, a crack was completely around the circumference on both
formed, and the fracture surfaces at C became sides of the gear. A crack was opened, and the
discolored and pebbled in appearance at an fracture surface was examined visually and at up
elevated temperature. Lack of fusion due to in- to 30 magnification using a stereomicroscope.
sufficient heating was a general condition in the Figure 53 shows macrographs of the fracture
weld, and this also contributed to the failure. In surface. The fracture surface was discolored
summary, the weld, when placed under tension,
failed due to a crack and lack of fusion.
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 sur-
faces next to the gear spline. The typical manu-
facturing sequence is forge, machine, heat treat,
grind, and plate.
Fig. 46 Microstructure consisting of tempered martensite at

fracture origin. Original magnification: 200 . Cor- Fig. 47 As-received portion of the failed aft right-hand fitting
rosion pit coated with paint is evident in the micrograph. for the hoist sling
with shades of black and brown, as can be pro- diameter surface of the gear. There were no
duced by heating the steel. The color brown shear lips evident around the periphery of the
corresponds to the temper color, which is pro- fracture surface. Fracture ridges emanated from
duced by heating steel to approximately 238 C origins located on the inner diameter edge, and
(460 F) in air and cooling to room temperature. the fracture surface exhibited a faceted appear-
Chromium plating was observed on the outer ance indicative of a brittle intergranular failure.
Fig. 48 Typical lack of fusion area
Fig. 50 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
Fig. 49 Direction of fracture propagation magnification: 1200

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 ob-
served 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) Loca-
tion B weld metal. (c) Location C weld metal
Fig. 53 Macrographs of (a) magnetic particle indications of

Fig. 52 As-received internal spur gear showing location of cracks and (b) opened crack revealing the fracture
cracks surface
Main Landing Gear Axle

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 stereo-
microscope. 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 sur-
face. Multiple origins were located around the
periphery of the axle. Faceted areas character-
istic 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
Fig. 54 SEM fractographs showing the topographic features
irregular shape of the intergranular area are
of a typical fracture origin. (a) Fracture surface. Ori-
ginal magnification: 20. (b) Location A showing intergranular indicative of stress corrosion.
fracture. Original magnification: 3000 An SEM was used to examine the fracture
Shown in Fig. 59 are SEM fractographs doc-
umenting topographic features, that is, inter-
morphology of the cracks is typical of grinding granular topography and fatigue striations, on
cracks (Ref 8). the fracture surface. From the ends of the inter-
A hydrogen analysis conducted on the intergranular area, fatigue propagated through the
nal gear yielded a value of 1.85 ppm, which is wall thickness. Also documented in Fig. 59 are
not considered to be high enough to cause the lack of plating at a fracture origin and cor-
hydrogen embrittlement. rosion products, which were present at the
Hardness measurements were made on the intergranular fracture origins. The lack of plat-
internal gear to verify the heat treat condition. ing and the presence of corrosion products are
The hardness of the gear near the observed indicative of a stress-corrosion failure. Also
cracking was 44 HRC. This hardness value was shown in Fig. 59 is an area where the chromium
below the drawing hardness requirement of 50 was intact.
to 53 HRC. The bulk hardness of the gear was The hardness of the axle was determined to be
51 HRC. 53.9 HRC. This met the hardness requirement of
The lower-than-expected hardness at the 53 to 55 HRC.
crack initiation sites indicates that there was A chemical analysis by atomic absorption
some event that caused excessive local heating verified that the MLG axle was fabricated from
of the microstructure. Because of the nature of 300M steel.
the part and typical manufacturing sequence, it A metallographic specimen was prepared
is likely that the grinding cracks occurred as a through a typical fracture origin on the axle. The
result of aggressive grinding. Based on the specimen was prepared using standard metallo-
results of this investigation, it is concluded that graphic techniques. As determined from the
the internal spur gear failed due to aggressive metallographic specimen, the thickness of the
grinding. chromium plating was approximately 0.06 mm
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 magni-
fication: 100. (d) Intergranular crack along prior-austenite grain boundaries. Original magnification: 200
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 chro-
mium runout to the origin, no plating was pre-
sent. 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 micro-
graph also shows the chromium plating that had
been penetrated. Documented in Fig. 62 is the
Fig. 56 As-received failed axle 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
(0.0024 in.). Shown in Fig. 60 is the micro- relatively high and considered high enough
structure of the axle at a typical fracture origin. to cause hydrogen embrittlement (Ref 4). This
The chromium plating ended approximately hydrogen is in the form of atomic hydrogen and
1.6 mm (0.062 in.) from the origin. The hard- can be generated in corrosion reactions and then
ness of the light, discolored area at the absorbed by steel.
Based on the results observed, it was deter- a minimum tensile strength of 1100 MPa
mined that the MLG axle failed due to stress- (160 ksi) minimum. The threads and the head-
corrosion cracking. The protective cadmium to-shank fillet radius are rolled after heat treat-
plating had been worn away, allowing corrosion ment. Machining of the fillet radius is not
to occur at locations where the cadmium was no permitted. The specified dimensions are shown
longer present. in Fig. 64 and Table 3.
The failed bolt and the additional bolts were
examined using nondestructive testing, visual
examination, scanning electron microscopy,
Nondestructive Testing and Failure
metallography, and analytical chemistry. The
Analysis of Fin Attach Bolts after hardness of the parts was also measured.
Full-Scale Fatigue Testing Nondestructive testing was performed on
the submitted bolts to determine if ultrasonic
During the inspection, after the first lifetime inspection is a suitable field inspection method.
of fatigue testing, one aft fin attach bolt was Magnetic particle inspection was used as a
found to have broken. Five additional bolts were confirmation of the ultrasonic inspection
found to have crack indications. The location of method.
the bolts and the nomenclature used for identi- Ultrasonic reference standards were fabri-
fication of these bolts are shown in Fig. 63. The cated from AN-8 and AN-6 pan stock bolts.
failed bolt (bolt A) and the additional five bolts, These bolts were machined so that the head
identified as B, C, G, H, and I, were submitted to thickness and lengths were similar to the test
the laboratory to determine if ultrasonic testing articles. Notches were cut into the bolts at the
would be a suitable field inspection technique head radius and at the midpoint length of the grip
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. Fig. 58 Overall view at origin. (a) Worn cadmium plating at
Typical fracture origins are shown at arrows. origin. (b) SEM photograph of fracture surface
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 the manufactured notches at the head-to-shank
standards and notches are noted in Fig. 65 and fillet radii and the shank identical, it was not
Table 4. While every attempt was made to make always possible because of manufacturing
tolerances. The ultrasonic inspection was con-

ducted 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 stan-
dard (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). Sig-
nal 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-to-
head fillet radius of bolt C. No crack indications
were found in the other bolts.
The submitted bolts were inspected using
standard magnetic particle inspection techni-
ques. The results indicate that bolts C, G, and I
had circumferential crack indications at the
shank-to-head fillet radius. Bolt I had a cir-
cumferential crack indication around the shank
of the bolt. The magnetic particle indications are
shown in Fig. 68.
Fig. 61 Secondary branch cracking observed. (a) Overall
After nondestructive testing, bolts C, G, and B
etched microstructure. (b) Location E, showing
cracking on prior-austentite grain boundaries. (c) Location F, were inspected visually. The as-received bolts
showing voids in chromium plating and associated cracking are shown in Fig. 69. After inspection, the cracks
Fig. 62 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)
indicated by magnetic particle inspection were of the drill bit was inserted into the hole of the
exposed. This was accomplished by centering bolt to support the bolt shank. Using the drill bit,
the threaded portion of the bolts in the chuck of a the bolt was bent away from the crack indi-
lathe and drilling a small pilot hole. A larger drill cation, keeping the crack faces in tension. The
bit was then used to drill the center of the bolt, exposed crack was then analyzed.
leaving a wall thickness of approximately Bolt B showed no crack indications, using
0.76 mm (0.030 in.). Using a hacksaw, a small ultrasonic inspection or magnetic particle
saw cut was made opposite to the crack. The inspection. This bolt showed significant cir-
head of the bolt was held in a vise, and the shank cumferential scoring around the periphery of the
Fig. 63 Schematic of bolt location and nomenclature used

shank, at approximately the midpoint of the Bolt C showed a highly polished shank sur-
shank. Minor deformation at the corners of the face, with a circumferential scoring line in the
bolt head was detected. Because no crack in- same approximate location as bolt B. Some
dication was seen by nondestructive testing rounding of the head corners was evident. No
(NDT), no additional work was performed on identifying mark or part numbers were stamped
this bolt. on the head of the bolt. After exposing the crack,
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 0.38 0.43 0.36 0.41 0.38 0.43 0.22 0.29
Si 0.20 0.35 0.15 0.35 0.20 0.35 0.15 0.35
Mn 0.75 1.00 0.60 0.90 0.65 0.85 0.50 0.80
P ... 0.025 ... 0.025 ... 0.025 ... 0.020
S ... 0.020 ... 0.025 ... 0.020 ... 0.015
Cr 0.40 0.60 0.50 0.80 0.70 0.90 0.90 1.20
Mo 0.20 0.30 ... ... 0.20 0.30 0.15 0.25
Ni 0.60 0.70 1.10 1.50 1.65 2.0 ... 0.30
Fig. 64 Schematic of bolt dimensions
Table 3 Allowable bolt dimensions

Thread C D F H R
Size code Diameter Pitch A tolerance B min Max Min Max Min Min G +0.0 0.3 +0.0 0.2 J
03 M3 0.5 5.5 4.88 0.4 0.2 2.990 2.965 6.08 6.0 2.0 0.4 ...
04 M4 0.7 7.0 6.38 0.5 0.2 3.990 3.965 7.74 7.6 2.5 0.4 1.0
05 M5 0.8 8.0 7.38 0.5 0.2 4.990 4.965 8.87 8.5 3.0 0.5 1.2
06 M6 1.0 10.0 9.28 0.5 0.2 5.990 5.965 10.95 9.5 3.5 0.7 1.6
08 M8 1.25 13.0 12.28 0.5 0.2 7.987 7.962 14.26 12.4 4.5 0.7 2.0
10 M10 1.5 17.0 16.08 0.6 0.3 9.987 9.962 18.90 14.8 5.0 0.8 2.0
12 M12 1.5 19.0 18.17 0.6 0.3 11.984 11.959 21.10 15.5 6.0 0.9 2.0
14 M14 1.5 22.0 21.17 0.6 0.3 13.984 13.959 24.49 17.5 7.0 1.0 2.0
16 M16 1.5 24.0 23.17 0.6 0.3 15.984 15.959 26.75 19.5 8.0 1.1 2.0
Note: All dimensions in millimeters
the resulting fracture surface was examined Table 4 Dimensions of reference standards
(Fig. 70). The fracture surface of crack C had the used for ultrasonic testing
characteristic ratchet marks or shear ridges that Specimen
A B
indicate multiple initiation sites along the outer identification mm in. mm in.
edge of the bolt. The shape and appearance of HL1 12.4 0.487 5.66 0.223
the fracture surface was indicative of fatigue. HL2 10.1 0.399 2.39 0.094
HL3 7.70 0.303 1.37 0.054
Bolt G showed circumferential scoring HL4 5.99 0.236 0.81 0.032
(Fig. 69), and some rounding of the head corners HS1 9.47 0.373 3.96 0.156
HS2 7.82 0.308 1.98 0.078
was evident. After the crack was exposed, HS3 6.50 0.256 1.17 0.046
the resulting fracture surface was examined HS4 4.16 0.164 0.69 0.027
HS5 2.46 0.097 0.15 0.006
SL1 12.3 0.485 5.87 0.231
SL2 11.2 0.440 3.40 0.134
SL3 8.13 0.320 1.52 0.060
SL4 5.82 0.229 0.81 0.032
SS1 8.97 0.353 3.18 0.125
SS2 7.65 0.301 2.24 0.088
SS3 5.16 0.203 0.76 0.030
SS4 4.57 0.180 0.46 0.018
SS5 2.82 0.111 0.18 0.007
Table 5 Ultrasonic inspection techniques

equipment settings
Tune 0
Reject Off
Gain 53 dB
Frequency 25 MHz
Video filter FW
Damping Minimum
Delay 0
Range 8.3
Velocity 1
Fig. 65 Sketch of reference standards used during ultrasonic Repetition rate 3
inspection
Fig. 66 Ultrasonic signal responses from the manufactured reference standards

Fig. 67 Ultrasonic signal responses for each bolt inspected
(Fig. 71). The fracture surface of crack G had the inspection. This bolt showed significant cir-
characteristic ratchet marks or shear ridges that cumferential scoring around the periphery of
indicate multiple initiation sites along the outer the shank, at approximately the midpoint of the
edge of the bolt. The shape and appearance of shank. More general wear was seen on the shank
the exposed crack was indicative of fatigue. than bolt B. Because no crack indication was
Bolt H (Fig. 72) showed no crack indications, seen by NDT, no additional work was performed
using ultrasonic inspection or magnetic particle on this bolt.
Fig. 68 Ultraviolet light photographs of magnetic particle indications for each bolt inspected. NDT, nondestructive testing
Bolt I (Fig. 72) showed scoring along the indicate multiple initiation sites along the outer
circumference of the shank. In this bolt, crack edge of the shank fillet radius (Ref 9). The shape
indications were found by NDT in both the and appearance of the exposed crack was indi-
shank and the head. The exposed shank crack is cative of fatigue.
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 indi-
cative 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 head-
to-shank fillet radius fracture surface had the
characteristic ratchet marks or shear ridges that
Fig. 70 Fracture surface of bolt C with fracture schematic
Fig. 69 As-received condition of bolts B, C, and G Fig. 71 Fracture surface of bolt G with fracture schematic
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. 74 Fracture surface of the exposed crack in the head

fillet radius of bolt I
Fig. 72 As-received condition of bolts H and I
Fig. 75 As-received condition of the failed bolt
Fig. 73 Fracture surface of the exposed crack in the shank of

bolt I Fig. 76 Fracture surface of the failed bolt
fracture surface shows ratchet marks that are observed. A large fatigue zone was apparent.
characteristic of fatigue origins. Multiple fatigue The final fracture zone was badly damaged. The
origins, located to one side of the bolt, were 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 Machining marks and polishing evident on the shank

of the failed bolt
Fig. 80 SEM micrograph showing initiation of cracking on
the shank of bolt I at localized surface wear (400 mm)
Fig. 78 SEM micrograph of representative multiple fatigue

origins associated with machining marks on exposed Fig. 81 SEM micrograph of representative fatigue striations
cracks (100 mm) found on the bolt fracture surfaces (2 mm)
with more evidence of machining marks in the initiation of cracking in the shank on bolt I
(Fig. 77). (Fig. 80). Striations, characteristic of fatigue,
The fracture surfaces on bolts A, C, G, and I were seen on the fracture surface of bolt A
were examined using the SEM. The fracture (Fig. 81). Machining marks found in the head-
surfaces of the exposed cracks in the bolts were to-shank fillet radius were evidence that the
characterized by multiple fatigue origins, initi- head-to-shank fillet radius was machined
ating along machining marks (Fig. 78). In the (Fig. 82). This is unacceptable according to the
fractured bolt, initiation of cracking began at a design specification.
region of localized wear and scoring on the A metallographic specimen was prepared
shank of the bolt (Fig. 79). This was also evident using standard techniques. This specimen was
used to examine the microstructure and micro-
hardness 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 head-
to-fillet radius (Fig. 85). Neither magnetic par-
ticle nor ultrasonic NDT detected this crack.
The failed bolt was submitted to a micro-
hardness survey to determine the extent of de-
carburization. The microhardness survey was
taken with a Knoop indenter, with a 500 g load.
Indentations were taken at 0.0318 mm
Fig. 82 SEM micrograph of machining marks at the head-
to-shank fillet radius of bolt I (1 mm)
Fig. 84 Micrograph of representative quenched and tem-

pered 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 Fig. 85 Micrograph of the crack found in the head-to-shank
location of additional micrographs. There is no evidence of a fillet radius of the failed bolt. Etched with 2% nital
cold-formed head. Etched with 2% nital (667 mm) (25 mm)
(0.00125 in.) intervals until the hardness read- Determination of the composition of the
ings began to stabilize. The point at which the failed bolt was performed by arc-spark spec-
hardness stabilized was called the core hardness. trometry. The carbon content of the steel was
For the failed bolt, the core hardness was 404 determined by Leco carbon analyzer. Results are
KHN. In accordance with the specification, the shown in Table 6. The chemical results indicate
point at which the hardness is within 20 KHN of that this bolt does not meet the composition
the core hardness is defined as the depth of requirements of Table 2 for any of the specified
decarburization. For the failed bolt, this was at bolt alloys.
0.064 mm (0.0025 in.) deep, meeting the spe- The hardness of each bolt was measured using
cification requirements. The hardness indentions standard procedures. The calibration of the
are shown in Fig. 86. The results of the microhardness tester was verified, using two calibra-
hardness survey indicated that the surface tion test blocks (54.1+1.0 HRC and 34.6+1.0
hardness (Knoop), converted to Rockwell C, is HRC). The measured hardnesses on the test
26 HRC. The Knoop core hardness, converted to blocks were found to be within the range of the
Rockwell C, is 40 HRC. While the partial de- hardness test block calibration. A section of
carburization of the bolt satisfies process speci- the bolt was taken by cutting perpendicular to
fication requirements, the steep gradient of the the long axis of the bolt, in the threaded region.
hardness profile and the surface softness mea- The hardness of each bolt was determined by
sured indicated that there would be degradation taking a hardness measurement on the end of the
in the potential fatigue life of the bolts. cut section. Results are shown in Table 7.
Based on this investigation, it was conclu-
ded that:
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, initiat-
ing at localized wear on the bolt shank. The
soft surface hardness from partial decarbur-
ization aggravated the fatigue failure.
The exposed cracks in bolts C, G, and I
initiated at machining marks at the head-to-
Fig. 86 Micrograph of Knoop hardness (500 g load) survey of
shank fillet radius, and cracking propagated
failed bolt (100 mm)
by fatigue. Evidence of machining at this
radius was found. Bolt I had a second
initiation site on the shank.
Table 6 Measured chemistry of the failed bolt A The composition of the failed bolt did not
Percent, meet the chemical requirements of S105
Element wt%
bolts. However, the chemistry of the bolts is
C 0.42
Si 0.32
not thought to have contributed to the failure
Mn 0.82 of bolt A.
P 0.010 The bolts were machined instead of cold
S 0.014
Cr 0.49 formed. This results in lower fatigue strength
Mo 0.15 and proof load.
Ni 0.45
REFERENCES
Table 7 Measured hardness of submitted bolts
Hardness, 1. G.E. Dieter, Mechanical Metallurgy, McGraw-
Bolt Rockwell C
Hill, Inc., 1986
C 43.4 2. Metallography and Microstructures, Vol 9,
B 36.4
H 41.6 Metals Handbook, 9th ed., American Society
G 40.4 for Metals, 1985
I 39.2
Failed bolt A 36.9
3. D.A. Ryder et al., General Practice in Failure
Analysis, Failure Analysis and Prevention,
Vol 11, Metals Handbook, 9th ed., American Handbook, 9th ed., American Society for
Society for Metals, 1986 Metals, 1986
4. B.E. Wilde, Stress-Corrosion Cracking, 7. A.G. Glover et al., Failures of Weldments,
Failure Analysis and Prevention, Vol 11, Failure Analysis and Prevention, Vol 11,
Metals Handbook, 9th ed., American Society Metals Handbook, 9th ed., American Society
for Metals, 1986 for Metals, 1986
5. W.J. Jensen, Failures of Mechanical Fast- 8. E. Alban, Failures of Gears, Failure Analysis
eners, Failure Analysis and Prevention, Vol and Prevention, Vol 11, Metals Handbook,
11, Metals Handbook, 9th ed., American 9th ed., American Society for Metals, 1986
Society for Metals, 1986 9. R.J. Parrington, Fractography of Metals and
6. Threaded Steel Fasteners, Failure Ana- Plastics, Pract. Fail. Anal. Vol 2 (No. 5),
lysis and Prevention, Vol 11, Metals 2002, p16
Failure Analysis of Powder Metal

Steel Components
S. Ashok and Sundar Sriram, Sundram Fasteners Ltd.
POWDER METALLURGY (PM) technol- This chapter briefly introduces the processing
ogy provides a cost-effective method of produ- steps involved in PM (Fig. 1). These heat treat-
cing near-net shape products, especially when ments and the PM process steps are essential
a large number of the same or similar products to understanding failure analysis of PM parts.
are required. While the initial powder used The methods used for analyzing the failures are
is expensive compared to wrought steels, the then discussed. Methods of failure analysis itself
manufacturing process can avoid machining have recently become more systematic in terms
steps and provide more uniform composition of prevention and ensuring more robust process
control. Total cost of producing PM parts can be and high-quality products. Some case studies
less expensive than conventional metalshaping are given that illustrate different failures and the
(casting, forging) in many cases. It was first methods of prevention of these failures.
adopted by the auto industry. Since this initial
inception, other industries, such as the refri-
gerator, lawnmower, and hand tools industries, Powder Metallurgy Process
have also extensively adopted PM parts (Ref 1).
Manufacture of powder metal parts involves Powder Production. Solid-state reduction
pressing of metal powder in a die, sintering of iron ore and atomization of molten iron are
the green compact so that metallurgical bonding the most common methods of producing iron
takes place, sizing or coining for densification powders for structural parts. The powder thus
and dimensional correction, followed by finish- produced should be favorable for compaction
ing operations. The finishing operations include such that:
machining, case and through hardening, steam
treatment, and so on (Ref 1). It fills the die cavity repeatably, which is
This chapter reviews failure aspects of struc- measured in terms of the apparent density
tural ferrous PM parts, which form the bulk of and flow rate of the powder.
the PM industry. Focus is on conventional PM It can be compressed to the desired density
technology of parts in the density range of 6 to with the lowest pressure, which is measured
7.2 g/cc. This constitutes 90% of the parts pro- in terms of compressibility.
duced by PM technology. It has sufficient strength at that density so
The PM part undergoes multiple heat treat- that the green part can be handled, which is
ments, some of which are unique to PM while measured in terms of its green strength.
others are similar to that used by conventional
The filling characteristic of a powder is
manufacturing processes, such as forging and
measured by its apparent density and its
casting. These include:
flow rate. The compressibility of a powder is a
Powder annealing function of the particle morphology, processing
Sintering history, and purity. Water-atomized iron pow-
Case hardening ders can be compressed to higher densities
Carbonitriding compared to sponge iron powder. The higher
Case carburizing the impurity content (typically oxides), the
Through hardening poorer the compressibility. As a process, all iron
Induction hardening powders are generally annealed in a reduc-
Steam treatment ing atmosphere. Special high-compressibility
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 pre-
vents 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 pow-
der, 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.
Fig. 1 The powder metallurgy process Compacting in the higher-density range gradu-
ally increases the rate of plastic flow of the
metal. This results in the following:
powders are produced by double-annealing Increased work hardening of powder parti-

operations. cles
Blending. Alloying elements are typically Increased friction against the die walls
mixed with the iron powder by blending in Structural ferrous powder metal parts typi-
a double-cone blender. Elemental addition of cally have sections of varying thickness. With
alloying elements, that is, graphite, copper, or respect to heat treatment, it is very important that
nickel, has its drawback, because these elements all sections in the parts have uniform density
tend to segregate in the product as the alloying throughout. Low sectional densities result in
element content increases. A recent develop- through hardening and wide variation in the case
ment in this regard is the introduction of diffu- depth within the PM part. Various methods of
sion-bonded iron powder and binder-treated compaction have been developed to produce
premixes. These processes are proprietary. For parts with minimal difference in density. These
example, the Distaloy process is diffusion include:
bonding of nickel, molybdenum, and copper to
iron powders developed by Hoganas AB, Swe- Double-action compaction
den. The Ancorbond and Starmix methods are Spring floating die compaction
binder treated premixs to reduce graphite- Compaction by withdrawal process
lubricant segregation. In the case of diffusion- Multilevel compaction using multiple mov-
bonded iron powders, the alloying elements, ing punches in special multiplaten presses
Failure Analysis of Powder Metal Steel Components / 397
Sintering. During sintering of ferrous PM with density. In parts that have been repressed
parts, the following reactions occur: and resintered, this condition may be tolerable or
even negligible, but for parts of lesser density,
Metallurgical bonds are developed from the the depth of carbon penetration may be so great
mechanical interlocks between the powder that these sections of quenched parts will be
metal particles in the compact. brittle.
Metal oxides in the powder compact are Another reason that conventional gas carbur-
reduced by reaction with the carbon from the izing enjoys only limited use is because it does
blended graphite or with the combustible not increase hardenability. Thus, plain carbon
constituents of the atmosphere. grades usually must be quenched in an aqueous
Desired final level of combined carbon in the medium. This may result in cracking, especially
microstructure is obtained by diffusion of if carbon penetration is excessive. As with
carbon from graphite in the powder mix. wrought parts, the depth of carburized case of the
Densification of PM compacts can be PM parts depends on time and temperature.
achieved in sintering. The degree of densific- Carbonitriding is a modified form of car-
ation depends on the sintering parameters and the burizing. The principal process modification
alloys involved. Generally, higher sintering consists of introducing ammonia into the gas
temperatures and longer sintering times promote carburizing atmosphere, which results in the
densification of ferrous sintered parts. addition of nitrogen to the carburized case as
For conventional sintering, which generally is it is produced. Nascent nitrogen forms at
conducted at 1120 C, mesh belt conveyor fur- the workpiece surface by dissociation of the
naces are widely used. These furnaces can be ammonia in the furnace atmosphere. Nitrogen
used up to temperatures of 1150 C, primarily diffuses into the steel surfaces simultaneously
due to the temperature limitation of the belt and with the carbon, where the austenite stability is
metallic muffle material. greatly enhanced by nitrogen in solution. This
For stainless steels, proper sintering is critical way, the necessary quenching rate to form
for proper corrosion resistance (Ref 2). martensite is reduced, and a martensitic micro-
structure is obtained without expensive alloying
elements. Typically, carbonitriding of PM parts
Case Hardening is carried out at 790 to 880 C for a duration of
30 to 60 min.
Powder metallurgy steel parts can be case Carbonitriding is widely used for case hard-
hardened by several processes, although various ening of PM parts made of ferrous powders.
available processes are not equally suited to Densities of the sintered compacts vary from
every application. A clear case/core relationship approximately 6.8 to 7.9 g/cc. Parts may be
can be obtained only with parts having a density infiltrated with copper prior to carbonitriding.
of at least 7.2 g/cc. Carbonitriding is extremely effective for case
Carburizing. Powder metallurgy parts with hardening high-density (7.2 g/cm3) parts made
relatively low combined carbon contents of up to from sintered iron compacts (Fig. 2). Addition-
0.20 wt% can be carburized by conventional ally, it is reasonably effective for case hardening
pack or gas methods. Liquid carburizing is parts of lower density.
not recommended because of the difficulty of Equipment and Techniques. Procedures for
washing the parts free of salt. carbonitriding PM parts are essentially the same
Gas carburizing is more practical for PM parts as those used for similar wrought parts. Control
than pack carburizing. For this process to be of temperature and time is generally more cri-
successful, however, density as well as the pre- tical than for wrought parts because of porosity.
cise composition of the parts should be known. Lower temperatures are avoided to minimize the
Low-density parts should not be subjected to potential danger of explosion, and higher tem-
gas carburizing, because the carburizing gases peratures are avoided because case depth control
penetrate the voids. Consequently, a distinct is more difficult.
case is not achieved compared with the case The processing cycle, including composition
developed on wrought parts under the same of the atmosphere, is critical. The ammonia con-
conditions. Instead, the carbon penetration in tent (usually 1 to 5% of carrier gas by volume)
PM parts is generally deeper and relatively increases hardenability and affects dimensional
nonuniform. The extent of this condition varies stability. Because dimensional changes in heat
due to the porosity are the causes of the different

induction-hardening responses of PM parts. The
results of hardening are more sensitive to chem-
ical 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
Fig. 2 Effect of density on the case depth as measured through
to high carbon content or large pores aggravate
a hardness traverse from the surface. Courtesy of
P. Beiss, University of Aachen, Germany cracking. Segregation of alloying elements,
foreign inclusions, or large pores can serve as
stress raisers, making the powder metal part
treating are often crucial to the economic justi- susceptible to cracking.
fication of producing parts by PM, gas compo- It is quite common for PM parts to absorb
sition, temperature, and quenching medium oil. Thus, intensive ventilation is required, and
must be closely controlled. one must ensure that the reused quenchant pro-
When processing to a new specification, vides the required quench severity. Water-based
establishment of processing parameters is polymer fluids are the most common quenchants
usually made on small production quantities, used for induction hardening of PM parts.
thereby requiring sacrifice of only a few parts to Steam Treatment. Sintered parts are sub-
arrive at optimal conditions. Such adjustments jected to steam treatment, wherein controlled
are then recorded so they can be used when oxidaton of the ferrous part is carried out in
processing the next lot of similar parts. Tem- an atmosphere of superheated steam. As a result,
pering of carbonitrided parts requires special a layer of Fe3O4 forms on the surface as well
consideration, largely because the quenching oil as in the pores. This oxide is hard and has
they contain will partially evaporate and pollute excellent wear resistance (hardness of 450
the environment. Toughness of PM parts in HV0.05 kg), increases density, and improves
the hardened condition (either by oil quenching corrosion resistance. The iron oxide also serves
or through sinter hardening) is significantly to seal the pores, which results in parts that
improved on tempering. are impermeable to gases. This characteristic of
Induction Hardening. Induction heating is
a method of heating electrically conductive
materials by the application of a varying mag-
Table 1 Effect of low density on properties and
netic field whose lines of force are intersected induction hardening process parameters
by the workpiece. In this process, the varying
Effect on induction
magnetic field induces an electric potential, Property Change hardening parameters
which in turn results in generation of electric Thermal Decreases with Inefficient conduction
current depending on the geometry, the fre- conductivity decrease in heat transfer
density Larger temperature
quency, and the electrical characteristics of gradients(a)
the workpiece. The induced current, termed Electrical Increases with Larger current
eddy current, generates heat that makes it resistivity decrease in penetration depth
density
amenable for use in many different heating Magnetic Decreases with Larger current
applications, of which the induction hardening permeability decrease penetration depth
of steels and cast irons is one of the most pre- in density Lower coil electrical
efficiency
dominant. Structural Decreases with Wide scatter in
Induction hardening of PM parts has several homogeneity decrease in apparent hardness
density Wide scatter in case depth
differences compared to hardening wrought
steels and cast irons. Chemical composition, (a) The penetration of the quenchant into open porosity overcompensates the
effect of lower conductivity. As a result, low-density parts cool faster.
microstructural heterogeneity, and low density
steam-treated PM products has enabled its 9000:2000. The engineer anticipates the failures
widespread use in pistons and valve plates for of the product on the basis of known failure
compressors. The sealing of pores also helps the modes and designs the process and product
PM part to be amenable to corrosion-preventive parameters accordingly.
coatings (galvanic as well as barrier coatings) However, in practice, failures are observed in
and, in some cases, to gas and salt bath nitriding. spite of carrying out the measures that are
The formation of Fe3O4 is exothermic, and thus, recommended by an FMEA. This can be due to
a further increase in product temperature by the fact that assumptions which go into carrying
10 to 20 C is observed. Steam treatment is out the FMEA may change over a period of time,
typically carried out at 500 to 550 C for a time or the basis of the assumptions is incorrect, but
of 30 to 300 min. The upper limit of 550 C more commonly, the failures occur due to the
is chosen so that the products do not exceed breakdown in one or more so-called 4M para-
570 C, which in turn would result in FeO for- meters (man, machine, method, material). Thus,
mation (Fig. 3). 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
Failure Analysis Techniques relevant to heat treated products. In the case
of heat treated powder metal parts, the failure
Failure analysis techniques can be broadly modes and their causes are limited in number
classified as preventive methods or corrective and can be quantified easily. The common fail-
methods. Preventive methods of failure analysis ure modes of heat treated powder metal parts
are typically carried out during product or pro- are:
cess development. These refer to the standard
techniques for failure mode and effects analysis Wear
(FMEA), which are specified in the various Fracture
certifications, such as TS16949 and ISO Dimensional instability because of plastic
deformation
Dimensional instability because of phase
transformation
Corrosion
These can be related to deficiencies in mechan-
ical or metallurgical properties. Properties that
play a role in failure of heat treated PM steel
include:
Metallurgical properties: microstructure;
case depth; coating adhesion, hardness, and
thickness for wear-resistant coatings; salt
spray life in the case of coatings for corro-
sion 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
Fig. 3 Fe-O-H2O diagram indicating the zones in which
the cause of the failure may be associated with
various oxides of iron are stable. Courtesy of P. Beiss,
University of Aachen, Germany more than one stage of component production.
Failure Analysis Tools. The objective of the cause of the failure. Two tools are invaluable
failure analysis is to relate the failure to a pro- aids in this regard. These are process maps and a
duct property attributed to one or more 4M cause-and-effect matrix.
parameters of processing. The steps that are Process Maps. In charting a process map,
typically carried out in the failure analysis of the entire manufacturing process is considered.
heat treated powder metal parts are shown in The implicit assumption is that the cause of
Fig. 4. Examination of the product involves one failure can be related to not only the final
or more of the following techniques: operation but also to prior operations. Every
operation in the manufacturing process is
Visual or any other method of non- regarded as a process with inputs and outputs.
destructive examination The inputs consist of the 4M conditions and the
Fractography or examination under a low- output of the product requirements. All possible
magnification stereomicroscope 4M conditions that can result in a nonconform-
Hardness and related mechanical property ing product are listed as inputs, with the non-
measurements conforming properties of the products as output.
Metallography under an optical or scanning Table 2 is an example of a detailed process
electron microscope map for the PM process. This exercise is carried
Chemical analysis out for all stages in the manufacturing process.
The failure can occur because of a non- Once the process map is charted, the potential
conforming property in the product even though causes at different manufacturing steps are
the product meets the specifications. The cause identified.
of the former is related to the manufacturing Cause-and-Effect Matrix. The process map
process, and the latter is related to the design. is an effective tool that eliminates the possibility
However, both failures involve determining the of ignoring an operation that can result in failure
4M conditions that cause the failure and cor- of the product. However, information presented
recting the same. In the case of failure of a heat by the process map requires sorting for easy and
treated powder metal part, the failure mode as meaningful interpretation. The cause-and-effect
well as the causes of failure are limited. As (CE) matrix ensures this. In the CE matrix, the
mentioned previously, a principal activity is an 4M conditions that can cause the failure are
operation-wide 4M analysis to determine the listed in the y-axis and the defective properties
Fig. 4 Steps in a failure analysis

Table 2 Detailed powder metallurgy process map

Inputs Outputs Inputs Outputs
Compaction Carburizing/carbonitriding (continued)
Single platen press ... High time of cabon ...
No pressway Low density potential attainment
No double action Density difference within High soaking time Low surface hardness
compact exceeding 0.2 g/cc Oxidizing atmosphere Decarburization
High apparent density Carbon and lube segregation Decarburizing atmosphere Oxidation
Low apparent density Alloy segregation Carburizing atmosphere Excess growth
Segregation of alloys Dimensions not conforming High nitrogen potential Shrinkage
to specification Masking of products ...
Segregation of carbon and ... Delayed quench Carbide network in case
lubricant Hot oil quench Nonmartensitic transformation
Sintering product in case
Low cooling rate ... High-viscosity quench oil ...
High cooling rate Low hardness Delayed tempering Low mechanical properties
Low temperature Decarburization Poor hardenability High retained austenite
High temperature Oxidation Low density Low case depth
Low soaking time Excess growth Carbon segregation High case depth
High soaking time Shrinkage Alloy segregation High core hardness
Oxidizing atmosphere Carburized microstructure Low sectional density Low core hardness
Decarburizing atmosphere Low mechanical properties Induction hardening ...
Carburizing atmosphere Blistered product Too high frequency ...
High rapid burnoff Pinholes Too low frequency Cracked product
temperature High heating rate High case depth
Low rapid burnoff Low sintered density Low heating rate Low case depth
temperature High heating time Low surface hardness
Low density ... Delayed quench Localized zones of melting
Carbon-lubricant ... Severe quench ...
segregation High temperature ...
Alloy element ... Low temperature ...
segregation High carbon ...
Low sectional density ... Low density ...
Poor sinter hardenability ...
Steam treatment
Carburizing/carbonitriding
Low carbonitriding ... Continous mesh belt Low hardness
temperature furnace
High carbonitriding ... Batch furnace Leaky product (pores not well sealed)
temperature Low temperature Low oxide layer
Low soaking time High surface hardness Insufficient time Poor surface appearance
Insufficient steam Loose rust
Oil in pores Red rust
on the x-axis. These are usually transferred from but the bushes were observed to fail in the field.
the process map. A coordinate in this matrix The parameters that can cause the failure
links the potential cause, that is, the 4M con- on listed in the CE matrix for wear of sinter-
ditions, to the failure, that is, the defective pro- hardened bushes, as follows:
duct characteristic. Once all the coordinates are
filled, all the potential causes of a failure are Operation 4M parameter Observations
listed in a user-friendly format. An example of Powder Low carbon Yes
a process map and CE matrix is illustrated in chemistry
Fig. 5. Blending Segregation of carbon and No
alloy elements
Compaction Low density No
Sintering Low sintering temperature No
Low soaking time No
Case Studies of PM Steel Failures Oxidizing atmosphere No
Decarburizing atmosphere No
High rapid burnoff temperature No
Case Study 1: Wear after Sinter Hardening Bainite in microstructure Yes
Overloading of parts No
Sinter-hardened bushes were developed for
an application that required high wear and
impact fatigue resistance. The bushes were sin-
ter hardened to a martensitic-bainitic micro- Figures 6(a) and (b) show the failed bush.
structure using a prealloyed Fe-Cr-Mo powder. Figure 7 shows the microstructure of the failed
Accelerated rig testing for 24 h yielded no wear, bush, which consists of martensite and upper and
Carbide network in case

Low apparent hardness
High apparent hardness
Poor surface apearance

High retained austenite
transformation product
High core hardness
Low core hardness
Free ferrite in core
Thick oxide layer
Blistered product
Poor mechanical
Cracked product
High case depth
Thin oxide layer

Decarburization
Low case depth
Nonconforming
Nonmartensitic
dimensions
Melt zones
Loose rust
properties
Oxidation
Pinholes
Red rust
Process step Process inputs
Blending Excess time X
Insufficient time X X X
Elemental alloy additions X X X
High carbon X X X
High lubricant X X
Compaction
Single platen press
No pressway X
No double action X
High apparent density X
Low apparent density X X
Segregation of alloys X
Segregation of carbon and
X X X
lubricant
Sintering Low cooling rate X

High cooling rate X
Low temperature X X X
High temperature X
Low soaking time X X X
High soaking t ime X X
Oxidizing atmosphere X X X X X
Decarburizing atmosphere X
Carburizing atmosphere X
High rapid burnoff
temperature X X X X X
Low rapid burnoff
X
temperature
Low d ensity X X
Carbon-lubricant
X X X
segregation
Alloy element segregation X
Low sectional d ensity X
Poor s inter hardenability X
Low carbonitriding
Carburizing/ temperature
X X X X X
carbonitriding High carbonitriding X
X X X X
temperature
Low soaking time X X
High time of carbon
X X X
potential attainment
High soaking t ime X X
Oxidizing atmosphere X X
Decarburizing atmosphere X X X
Carburizing atmosphere XX X X X
High nitrogen potential X X X
Masking of products X X
Delayed quench X X X
Hot oil quench X
High-viscosity quench oil X
Delayed t empering X
Fig. 5 Cause-and-effect matrix derived from powder metallurgy process map

Carbide network in case

Low apparent hardness
High apparent hardness
Poor surface apearance

High retained austenite
transformation product
High core hardness
Low core hardness
Free ferrite in core
Thick oxide layer
Blistered product
Poor mechanical
Cracked product
High case depth
Thin oxide layer

Decarburization
Low case depth
Nonconforming
Nonmartensitic
dimensions
Melt zones
Loose rust
properties
Oxidation
Pinholes
Red rust
P ro c e s s s te p Process inputs
Carburizing/ Poor hardenability X X X X
Low density X X X
carbonitriding
Carbon segregation X
Alloy segregation X
Low sectional density X X X
Induction Too high frequency X

Too low frequency X
hardening
High heating r ate X
Low heating rate
High heating time X X
Delayed quench X X X X
Severe quench X
High t emperature X X
Low t emperature X X
High carbon X
Low density X X
Steam Continous mesh belt

furnace
t reatment Batch furnace
High t emperature X X X
Low temperature X
Insufficient time X X X
Insufficient steam X X X X X
Oil in pores 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 Corrective Measures. An improved chem-
bainite being worn out, resulting in eventual istry was derived, where the base iron powder
accelerated fatigue failure of the bush. was prealloyed with nickel and molybdenum,
Bainite is an essential component for and the carbon content was increased to 0.9%.
toughness. A fully martensitic microstructure This resulted in a predominantly martensitic
obtained by increasing the cooling rate and the microstructure. The presence of nickel ensured
carbon content has poor toughness. Thus, the the toughness requirements, and the increased
decision was made to use a different material carbon ensured the wear requirement. Figure 8
that can give the required hardness and tough- shows the new microstructure, wherein mar-
ness. tensite with some retained austenite is observed.
Fig. 7 Micrographs of the failed bush. (a) Martensite and lower bainite. (b) Upper bainite
Results. The improved bushes were assem-

bled 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, se-
gregations of these elements are not uncommon.
The parameters that can cause pinholes in valve Fig. 8 Martensite and some retained austenite in the modified
chemistry
seats one:
Operation 4M parameter Observations
Blending Segregation of carbon and lubricant Yes
Copper agglomeration No
No inspection for agglomerates Yes
Scanning electron microscopy/electron dis-
in powder persive x-ray analysis was carried out to deter-
Compaction Not applicable ... mine the cause. Figures 10 to 13 reveal
Low soaking time No high carbon content near the pinhole. Thus, the
Machining Pullout of nonmetallic inclusions No likely cause of the pinholes was concluded to
in machining be carbon-lubricant agglomeration during
blending.
Corrective Measures. Binder-treated pre-
In all of these above cases, the pinholes that mixes that minimize carbon-lubricant segre-
are observed after sintering or machining are in gation were recommended for the products.
the regions where the agglomerates were pre- An inspection procedure was evolved to check
sent, prior to sintering or machining. The copper the premix for the presence of agglomerates and
melts and diffuses into the iron matrix, carbon coarse particles, wherein the premix was sieved
diffuses in the matrix, and the lubricant burns and the +150 mm fraction was inspected for
off, leaving the pinhole in each case. In the case agglomerates under a stereomicroscope. This
of nonmetallic inclusions, they are pulled out ensured that a premix with agglomerates is not
during machining, thus leaving a pinhole. issued for compacting the products.
Fig. 9 Comparison of sintered versus regular bush in 12B model chain
Fig. 11 SEM micrograph of the pinhole of Fig.10 indicating a

Fig. 10 Pinhole in the valve seat at a magnification of 40 : 1 suspected graphite particle
Results. The rejections due to pinholes The parameters that can cause blistered sintered
decreased significantly as a result of using bin- products are:
der-treated premix. Operation 4M parameter Observations
Powder chemistry High nickel content Yes
Blending Segregation of nickel powder No
Elemental nickel additions Yes
Case Study 3: Blistered Sintered Products Compaction Not applicable ...
Sintering Low rapid burnoff temperature No
Synchronizer keys used in an automobile gear H2 in hot zone Yes
box were produced with Fe-3%Ni-0.5%C. Ele-
mental nickel and carbon were mixed with the
iron powder. Blistering of the products after It was observed that delubing the keys was not
sintering was observed in ~5% of the products. completed in rapid burnoff, and the lubricant
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 atmo- completely ensured. Thus, it was resolved that
sphere consists of N2-10%H2. In the presence reduction in catalytic activity of nickel could be
of H2, nickel acts as a catalyst favoring the an effective solution. Sulfur effectively poisons
cracking of the lubricant. This results in the nickel and limits its catalytic properties, so
lubricant exploding, and consequently, blistered 0.03% S was added to the blend.
products. Results. The addition of sulfur to the blend
Corrective Measures. Two possible solu- effectively reduced the rejections from 5%
tions to the problem were considered: to nil.
Ensure complete removal of the lubricant
prior to contact with H2 Case Study 4: Dimensional Instability
Reduce the catalytic activity of nickel during Shrink Fitting
Due to the limitations of sintering in a con- Exhaust valve seats with tool steel powder
tinous mesh belt furnace, the former could not be as a major constituent were shrink-fitted in a
cylinder head of a sport utility vehicle. It was parameters that can cause wear of a free-graphite
observed that the valve seats were expanding bush are:
during shrink-fitting instead of contracting.
This resulted in interference between the
valve seat and the cylinder head, causing rejec- Powder chemistry Not applicable ...
Blending Not applicable ...
tion of the cylinder heads. The parameters that Compaction High density No
can cause dimensional instability during shrink- Sintering High sintering temperature Yes
High sintering time No
fitting are: Cementite network Yes
Insufficient free graphite Yes
Operation 4M parameter Observations Oil impregnation Low oil content No
Powder chemistry Not applicable ...
Compaction Not applicable ...
Sintering Not applicable ... Metallography of the worn sample revealed a
Tempering High retained austenite Yes
cementite network in the product (Fig. 14). This
is a clear indication that excess carbon had gone
into solution, which resulted in the solid lubri-
A significant amount of retained austenite cant being depleted.
was observed even after tempering. This Corrective Measures. Sintering was carried
retained austenite was transforming to marten- out at 1080 C for 20 min, which resulted
site during shrink-fitting, where the products in excess combined carbon. The temperature
were cooled to 60 C, which resulted in a was reduced so that the combined carbon
volume expansion. was 50.8%. The sintering temperature for the
Corrective Measures. A complete trans- product was fixed at between 1040 and 1060 C
formation of retained austenite can be effected Results. Bushes sintered at 1060 C
by: for 20 min resulted in a pearlitic-steaditic
microstructure with no cementite network
Multiple tempering (Fig. 15).
Cooling to below the martensite finish tem-
perature (in this case, a deep cryogenic
treatment)
Combination of tempering and deep cryo- Case Study 6: Fracture of Steam-Treated
genic treatment Part
Valve plates for compressors were subjected
Trials revealed that a combination of subzero to grinding after steam treatment. Fracture of
treatment and tempering gave the best results a steam-treated valve plate was observed during
with respect to dimensional stability as well as grinding after steam treatment. The parameters
metallurgical properties.
Results. The valve seats subjected to a
combination of subzero treatment and tem-
pering 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.
Case Study 5: Wear after Sintering

Bushes were produced with Fe-0.45%P
2%Cu-2.5%C and supplied in the oil-impreg-
nated condition. These bushes were deliberately
undersintered so that some free graphite was
retained, which acted as a solid lubricant that Fig. 14 Pearlite-steadite microstructure with cementite net-
in turn minimized wear in application. The work
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
Fig. 15 Pearlite-steadite microstructure free from cementite parts, which consequently resulted in under-
network sintering.
Corrective Measures. Suitable one-point
lessons were imparted to the sintering operators
that can cause the fracture of a steam-treated to ensure that the right process plans were fol-
part are: lowed.
Powder chemistry Not applicable ... Case Study 7: Oxidation after Sintering
Compaction Low density No Exhaust valve seats for internal combustion
Sintering Low sintering temperature No engines are produced with a chromium-rich
Low soaking time No
Oxidizing atmosphere No hard-phase alloy. The hard-phase alloy is
Decarburizing atmosphere No essential for wear resistance at elevated tem-
High dewaxing temperature No peratures, to which the exhaust valve seat
Overloading of parts Yes
Steam treatment High temperature No is subjected. Oxidation of the hard phase was
Insufficient steam No observed after sintering. The parameters that can
Low soaking time No
Overloading of parts No
cause oxidation after sintering are:
Masking of parts No
Powder chemistry Chromium-rich alloy Yes
Hardness measurements revealed very low Compaction Rusted compacts No
Sintering High rapid burnoff temperature Yes
hardness. Product with hardness conforming to High cooling zone dewpoint Yes
specification and product with low hardness
Fig. 16 Low-hardness sample with almost no combined Fig. 17 Typical microstructure with ~0.7% combined
carbon. Only ferrite and free copper are observed. carbon
Chromium-rich particles are susceptible to density of 7 g/cc. Subsequently, the gear was
oxidation when the product is heated and cooled sintered, and a sizing operation was then carried
in the range of 650 to 1050 C. Oxidation of the out for dimensional correction. Induction hard-
hard-phase particles is favored by: ening of the gear was carried out after sizing, and
cracks were observed after induction hardening.
High rapid burnoff temperature
The parameters that can cause cracks after
High cooling zone dewpoint
induction hardening are:
Figure 18 shows the microstructure with
oxidized hard-phase particles. The oxide film
Powder chemistry High carbon Yes
around the hard phase does not allow the hard Blending Not applicable ...
phase to interact with the matrix. As a result, the Compaction Low density No
matrix will not have chromium carbides, which Sintering Low sintering temperature No
Low sintering time No
helps the valve seats to retain wear resistance at Sizing Not applicable ...
high temperatures. The rapid burnoff temper- Induction hardening High heating time No
ature was observed to be 650 C, and the cooling Severe quench No
zone dewpoint was between 20 to 30 C.

These are potential causes for oxidation.
Corrective Measures. The furnace was Figure 20 shows the crack. This crack origi-
allowed to stabilize for a period of 4 to 16 h. The nates in the middle of the root of the gear and
gas ratio was maintained at 70% N2 and 30% H2
during this period, and the gas volume was
increased by ~25%. The rapid burnoff tem-
perature 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 para-
meters resulted in the production of oxidation-
free valve seats (Fig. 19).
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 Fig. 19 No oxide layer observed in the hard-phase particles
after implementation of corrective measures
Fe-Cu-C chemistry and was compacted to a
Fig. 18 Exhaust valve seat with oxidized hard-phase Fig. 20 Crack in the middle of the root of the primary driven
particles gear
does not originate at any sharp corners. Hardness carbon content was reduced in the blend to 0.6%,
profiles from the root indicated that the root has which resulted in a lesser volumetric strain due
been hardened to a very high hardness (Fig. 21), to martensitic transformation in the product.
compared to the earlier sample in which no crack Induction hardening with this chemistry yielded
was observed. Microstructure near the root no cracks, and the hardness traverse matched
revealed untempered martensite, which gradu- that obtained in the rapid prototype samples.
ally became a mixture of martensite and bainite Results. Products with a lower combined
and subsequently ferrite and pearlite. This is in carbon content did not crack when induction
contrast to the earlier sample where the root was hardened.
not hardened and had a bainitic structure. Sec-
ondly, the combined carbon estimated from the Case Study 9: Cracks after Quenching
pearlite content in the sintered microstructure
was 0.7% compared to 0.6% in the product A flyweight for a governor assembly in a
that was successfully induction hardened. The diesel engine was carbonitrided to meet the
strain caused by the martensitic transformation product requirements for strength. This is a
increases in proportion to the combined carbon complex part wherein multiple punches are used
in the microstructure. Thus, the possible causes to form the sections of different thicknesses.
for the crack were: The product was observed to crack after carbo-
nitriding. The parameters that can cause cracks
Too high a soaking time, which resulted in a after quenching are:
severe quench
High combined carbon in the sintered
microstructure Blending Not applicable ...
Compaction High density difference No
Corrective Measures. The formation of Underfilling of powder Yes
martensite in the root was a result of quenching Low sintering time No
from very high temperatures. The time was Carbonitiriding High hardening temperature No
reduced, and the coil design and quench were Severe quench No
modified. However, the martensitic micro-

structure could not be avoided in the root, and
cracking was still observed. It was concluded The parts were subjected to magnetic particle
that the product was unable to withstand the inspection prior to heat treatment. No cracks
stresses associated with the transformation. The were observed. Metallography was then carried
Fig. 21 Hardness traverse from root

out at the location where the crack occurred in Carbonitriding was carried out at 840 C for
heat treatment. This is shown in Fig. 22. Insuf- 45 min with a carbon potential of 0.8% in the
ficient powder filling was observed (Fig. 23). atmosphere. In the case of the vendor, better
The underfill results in a localized region of low furnace gas sealing ensured that this case depth
strength, which gives way during subsequent was achieved in 20 min. The hardness traverse
quenching and results in crack formation. indicated that the longer soaking time had a
Corrective Measures. The localized under- higher carbon martensite in the case as well as
filling was due to the presence of a sharp corner a higher core hardness, which in turn resulted in
in the tool. A trial was conducted wherein the a case depth at the lower end of the specification
sharp corner was smoothed by filing the parts. for a lower soaking time. This is shown in
No crack was observed after carbonitriding. It Fig. 24.
was resolved to increase the radius in the tool so Corrective Measures. Higher shrinkage is
that a sharp corner and consequently an underfill ensured by a higher amount of martensitic
are avoided. Products thus produced were heat transformation. This was ensured by increasing
treated, and no cracks were observed. the soaking time from 20 min to 45 min. The
Results Products produced with an improved increased soaking time resulted in better expo-
radius in the tool resulted in elimination of sure of the product to carbon and nitrogen
product rejections due to quench cracks. enrichment, which in turn resulted in higher
martensite content and the desired level of
dimensional shrinkage.
Case Study 10: Dimensional Change Results. The increased martensite content
in Carbonitriding resulted in increased volume expansion. Thus,
the shrinkage in the bore is higher with a
An angle lever for a diesel injection pump was
soaking time of 45 min. Products soaked at
produced with FD-02-00 chemistry and carbo-
840 C for 45 min were found to shrink suffi-
nitrided to a case depth of 0.2 to 1.2 mm. When
ciently, and dimensions conformed to specifi-
the carbonitriding was established at a new heat
cations.
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:
Case Study 11: Low Surface Hardness
after Carbonitriding
Blending Not applicable ... Low surface hardness was observed in
Compaction Bore undersized due to tool wear No
Sintering High sintering temperature No an iron-copper part after carbonitriding. The
High sintering time No
Carbonitiriding Low soaking time Yes
Insufficient carbon and nitrogen No
enrichment
High quench oil temperature No
Insufficient transformation Yes
Fig. 23 Metallography of the region where the crack occur-

red in the sintered sample, showing a pre-existing
Fig. 22 Quench crack in flyweight defect caused by poor powder filling
Fig. 24 Hardness traverse in bell crank lever
parameters that can cause low surface hardness due to dimensional distortion. On a permanent
after carbonitriding are: basis, a periodic preventive maintainence
schedule for the oxyprobe was drawn up and
adhered to.
Results. Correcting the oxyprobe yielded
Blending Not applicable ... products with a martensitic/510% retained
Compaction Low density No austenite structure that met the requirements for
surface hardness.
Carbonitiriding Low soaking time No
Insufficient carbon and No
nitrogen enrichment
Excess carbon and nitrogen Yes
Case Study 12: Variation in Bore Diameter
enrichment after Heat Treatment
Delayed quench No
Pearlite/bainite in case No A valve retainer for a shock absorber was
High retained austenite Yes
produced from an FC-02-08 blend that was
hardened for wear resistance. High ovality was
observed after heat treatment. This resulted in
After ensuring that the density conformed ~10% of the products being out of specification.
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 exam-
ined 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 cor-
rective measure adopted to salvage the parts
was to carry out a deep freeze, because repro-
cessing may result in scrapping of the product Fig. 25 High retained austenite in the case
The parameters that can cause a wide variation Low breaking load indicated poor toughness
in bore diameter after heat treatment are: of the product. Poor toughness is related to both
Operation 4M parameter Observations microstructure and density. Lower densities
Powder chemistry Excess copper content No and a through-hardened microstructure with
Blending Elemental additions No martensite in the core can cause premature
Compaction Low density No fracture. Metallography was carried out in dif-
Low sintering time No ferent sections of the failed product. Metallo-
Carbonitiriding Volume expansion due to Yes graphy of the sections indicated that the thinner
martensitic transformation
Thermal contraction during Yes
sections in the product were through hardened.
quenching Figure 26 shows the through-hardened micro-
structure. Sectional density measurements indi-
The ovality was checked after sintering and cated that the thicker sections had a density of
after heat treatment. The data revealed that 47.2 g/cc, and the thinner sections had a density
hardening contributes significantly to distortion, of 7.1 g/cc.
and the contribution of sintering is negligible. However, this through hardening of thinner
Thermal contraction due to quenching and the sections was not observed in the initial samples
accompanying martensitic phase transformation that were tested. To determine the cause of this
contributed significantly to the distortion. variation across different heat treated batches,
Corrective Measures. To minimize the the carbon potential attainment was studied for
distortion, it was resolved to eliminate the various batches.
hardening operation and achieve the required During establishment of the carbonitriding
hardness in sintering. Prealloyed powder con- process, the process parameters were fixed for
taining nickel and molybdenum, blended spe- ~500 parts. The same was subsequently fixed
cially to produce products with negligible for the full batch quantity of 3500. In this par-
distortion after sintering, was adopted to achieve ticular case, the carbon potential pickup in the
both the hardness and dimensions after sintering. furnace was completely different for the trial
Results. Changing the process from con- and bulk. Figure 27 shows the difference in
ventional hardening to sinter hardening reduced carbon potential pickup. There is a significant
the dimensional spread to 50% of the variation difference in the time for attainment of a carbon
observed in products with a separate heat treat- potential of 0.8%. Longer times between 0.6 and
ment operation. 0.8% in the bulk lots result in additional carbon
pickup, which resulted in through hardening of
Case Study 13: Low Breaking Load after the thinner sections.
Carbonitriding Corrective Measures. The carbonitriding
temperature and time were reduced. The batch
A component for a gearbox application pro- quantity was fixed. The density in the compo-
duced with a chemistry of FD-02-00 and having nent was increased to 7.2 g/cc min in all
an overall density of 7.2 g/cc was subjected
to carbonitriding to meet the product require-
ments for wear and mechanical strength. During
assembly, the product failed by fracture. The
parameters that can cause low breaking load
after carbonitriding are:
Compaction Low density No
Low sectional density Yes
Carbonitiriding High carbonitriding temperature Yes
High time of carbon potential Yes
attainment
Variation in batch quantity Yes
Martensite in core Yes
No case depth inspection Yes
in thin section
Fig. 26 Martensite in the core of thin sections in the product
Fig. 27 Difference in carbon potential attainment for change in batch quantity
sections, which ensured closure of intercon- prior to steam treatment, metallography was
nected pores. Inspection standards for sectional carried out on the samples with the undersized
density and case depth measurements in the bore. An oxide layer of up to 15 mm was
thinner section were introduced so that any observed (Fig. 28). Typically, Fe3O4 layers are
nonconformance would be detected and not 3 to 5 mm, because the lattice of Fe3O4 does not
reach the customer. allow further diffusion of oxygen molecules
Results. The corrective measures resulted in once this layer is formed. Such a thick layer is
a pearlitic-bainitic microstructure in the core of due to the formation of FeO. The steam treat-
the thin section, and the toughness and, conse- ment set temperature was 560 C for this pro-
quently, the breaking load were achieved conduct. Because the Fe3O4 reaction is exothermic,
sistently. the temperature of the products exceeded
570 C, resulting in the formation of FeO.
Corrective Measures. The steam treat-
Case Study 14: Dimensional Change in ment temperature was subsequently reduced to
Steam Treatment 530 C, where an oxide layer of 4 to 6 mm was
obtained (Fig. 29). This resulted in the bore
Pistons having a density of 7 g/cc were steam
dimensions meeting the specification.
treated for wear resistance and corrosion resist-
ance. 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 para-
meters that can cause dimensional change in
steam treatment are:
Compaction Not applicable ...
Sintering Not applicable ...
Machining Bore undersized No
Steam treatment High temperature Yes
Thick oxide layer Yes
Fig. 28 Thick oxide layer indicative of FeO formation that

After ensuring that the machined dimensions was obtained at a steam treatment temperature of
of the bore were in accordance with specification 560 C
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
Fig. 29 Oxide layer of 6 mm, indicative of Fe3O4 obtained on steam penetration. An oil that has a higher
reduction of the steam treatment temperature to volatile content and a lower organic residue
530 C content was used for rust prevention, and the
same was used as a lubricant in sizing. The
Case Study 15: Low Core Hardness after change of oil resulted in a core hardness con-
Steam Treatment forming to specification. However, the hardness
obtained ranged from 145 to 190 HV. The pre-
Crankshaft sprockets with an Fe-Cu-C sence of products with hardness marginally
chemistry and a density of 6.6 to 6.8 g/cc were above 140 HV10 indicated that the hardness
supplied in the steam-treated condition. The must exceed 160 HV to avoid the chances
customer demanded that the hardness in the of core hardness going below specification.
surface and core meet the specification of This additional improvement in hardness
140 HV10 min. The parameters that can cause was achieved by having a completely pearlitic
low core hardness after steam treatment are: microstructure instead of a ferrite+pearlite
microstructure, which was effected by a blend
Operation 4M parameter Observations addition of 0.8% C.
Powder chemistry Low carbon Yes Results. Implementation of the corrective
Compaction High density No
measures resulted in the core hardness being
Sintering Low sintering temperature No achieved consisently in accordance with speci-
Low sintering time No fication, as follows:
High ferrite content Yes
Sizing Closed surface pores No
Properties Specification Before After
Oil in pores Yes
Steam treatment Insufficient time Yes Surface 140 152208 185235
Low temperature Yes hardness HV10 min
High temperature (4570 C) No Core hardness 140 110173 170207
Masking of products Yes HV10 min
Nonuniform steam circulation No Sintered ... Ferrite+ 100% pearlite
microstructure pearlite
Steam-treated ... Incomplete Complete steam
microstructure steam penetration with
penetration oxide layer of
It was observed that within a batch, some with low-oxide 46 mm in core
products had a core hardness greater than 140 layer52 mm
HV10. Low and high core hardness samples were in core
subjected to metallography. It was observed that
higher-hardness samples had better steam
penetration compared to low-hardness samples. REFERENCES
Corrective Measures. To ensure higher
core hardness, the products required greater 1. Asm Handbook, Volume 7, Powder Metal
exposure to steam. This was achieved by chan- Technologies and Applications, 1998
ging the steam treatment temperature and time 2. E. Klar and P. Samel, Powder Metallurgy
from 530 C/30 min to 550 C/2 h. Masking of Stainless Steels: Processing, Micro-
products can also potentially prevent uniform Structures, and Properties, ASM Interna-
exposure to steam. To eliminate masking, the tional, 2007
Induction Hardening
Janez Grum, University of Ljubljana
INDUCTION HEATING, in most applica- the action of external forces and moments. To
tions, is used to selectively heat only a portion of ensure a long life of the machine part, knowl-
the workpiece that requires treatment. This edge of the residual stresses in the machine part
usually means that the process can be accom- and how to adjust the size and distribution of the
plished in a relatively short time and with high residual stresses by means of the selection of an
efficiency, because energy is applied to the appropriate production technology are very
workpiece only where it is needed. important.
Induction surface hardening is applicable to In surface hardening, compressive residual
axisymmetric or near-axisymmetric machine stresses always occur in the thin surface layer
parts in steel or cast iron that are produced in due to martensite transformation. The size and
substantial volumes. There are two basic tech- variation of the residual stresses depend pri-
niques for induction hardening machine parts: marily on carbon content and less on the type
single-shot and scanning. The former employs and content of alloying elements in heat treat-
selective heating and quenching to harden a ment and surface-hardening steels. The variation
specific area or areas of the machine part in one of residual stresses in the surface layer can be
operation. The latter is usually applied to harden modified by varying the induction-heating con-
progressively long, continuous sections, such as ditions and by a quenching method.
shafts and spindles. In this instance, the scanning Induction surface hardening creates very
inductor traverses the length of the section, desirable residual stresses in the hardened sur-
heating only a relatively small area at any given face layer. Residual stresses are always of a com-
time, and is followed closely by the quench pressive nature and are usually present to the
arrangement, which is often an integral part of depth of the induction-hardened layer. Residual
the inductor. internal stresses, that is, the so-called residual
These advantages make it possible for in- stresses, are the stresses present in a material
duction hardening to be fully automated and are or a workpiece when there is no external force
especially suitable for a large series of work- and/or external moment acting on it. The resi-
pieces. dual stresses in metallic machine parts have
The induction-hardening procedure enables attracted the attention of technicians and engi-
an engineer, by simply adapting the shape of the neers only after manufacturing processes im-
induction coil, to ensure the desired shape of the proved to the level at which the accuracy of the
hardened profile of the surface layer. Likewise, manufacture exceeded the size of deformation,
the engineer can surface harden only that part of that is, distortion, of a workpiece/product.
the surface (local hardening) on which a certain Thus, it was almost 150 years ago that the
increased level of hardness and wear resistance effect of internal stresses on plasticizing, that is,
are desired. One of the main advantages of in- destruction, distortion, and plastic deformation,
duction hardening is the ability to harden a sur- of workpieces was already known. It was then
face layer only in certain places at a defined that experts introduced measurement of indivi-
penetration depth and shape. dual dimensions of products. For a given type of
For dynamically loaded machine parts, it is machining process, they connected the influence
very important to ensure the total compressive of the selected machining conditions with the
stresses in the thin, most-loaded surface layer. size of dimensional deviations. This was also the
The total stresses are a sum of residual stresses in beginning of an expert approach to the selection
a machine part and of load stresses produced by of the most suitable machining and/or heat
treating conditions based on the criterion of since manufacturers are frequently required to
minimum dimensional deviations, that is, mini- produce machine parts that, among other surface
mum workpiece distortion. Currently, measure- properties, must have quite specific residual-
ment of individual workpiece dimension is a stress distribution along the depth of the hard-
very practical, uncomplicated, and reliable ened layer. It has become a proven fact that high
method of product quality assessment. Through compressive stresses ensure high fatigue
the appropriate selection of the grinding wheel strength of machine components and reduce the
and grinding conditions and taking into account danger of crack occurrence and growth on the
the physical and mechanical properties of the surface of components.
workpiece material, very favorable compressive In thermal hardening, the surfaces of suitable
residual stresses in the hardened surface layer materials, usually plain carbon, low-alloy steels,
can be retained. or cast irons, are austenitized and then quenched
How is it possible to assure a desirable surface to produce a hard martensitic case that is usually
and surface layer quality after induction hard- tempered in a subsequent operation. Case depths
ening and fine grinding? Finding an answer to are normally in the range of 0.5 to 5 mm. Case
this question requires a very good knowledge of hardness is typically approximately 700 HV on
the grinding process on the microlevel as well as hardening and 600 to 650 HV after tempering at
knowledge of mechanical and heat effects acting 200 C (Ref 1).
on the layer of the workpiece, including the type Heating processes include electrical induction
of grinding wheel and the grinding conditions. and resistance (Ref 24), and direct impinge-
An all-inclusive consideration of the numer- ment methods using flames, (Ref 5), lasers
ous influences of the tool type and condition on (Ref 611), and electron beams (Ref 12). Of
the changes on the surface and in the surface these, induction heating is the most widely used.
layer of the workpiece in the given machining Laser and electron beam heating have recently
conditions is described by the term surface become established in a number of applications,
integrity. This is a scientific discipline providing mainly where distortion was a problem with the
an integral assessment of the surface and sub- induction method.
surface layer. It was defined at the beginning of Thermal treatments are mainly employed
the 1960s. For high-quality machine parts and when only local areas on machine parts require
parts subjected to heavy thermomechanical hardening and wear resistance. Consequently,
loads, different levels of description of the sur- such methods are energy efficient compared
face integrity were defined. A basic level of the with thermochemical methods where bulk heat-
surface-integrity description includes measure- ing of batches of components is undertaken.
ment of roughness and analysis of the micro- Normally, the surface-hardening process also
structure and microhardness in the thin surface introduces compressive stresses into the surface
layer resulting from the machining process layers, leading to an improvement in fatigue
under given machining conditions. The second properties. For example, the drive shafts of
level of the surface-integrity description in- heavy lorries and buses are induction surface
cludes studies of residual stresses in the surface hardened to improve their fatigue properties
layer and of mechanical properties of the given (Ref 10, 11).
material. The third level of the surface-integrity The skills and experience required for manual
description includes tests making clear the be- flame hardening have been largely superseded
havior of the given part under the operating by automated flame techniques or by the in-
conditions. duction method. However, for one-off machine
As essential advantage of induction surface parts, the simplicity of manual flame hardening
hardening is that it is possible to achieve a suf- and the extremely low capital investment ensure
ficient repeatability of the hardened layer thick- that the method is still used on a regular basis
ness on the workpiece as well as a desirable or (Ref 5, 13).
even prescribed hardened layer profile, ensuring From a heat treatment point of view, laser can
sufficient hardness and favorable distribution be considered as a versatile and flexible high-
of residual stresses in the hardened layer. A intensity heat source that can operate in air. It is
variety of steels and a whole range of induction- capable of undertaking a range of processes,
hardening methods provide the possibilities essentially simultaneously, since the laser beam
for very accurate planning of residual stress size can be directed through air by metal mirrors and
and distribution. This is of growing importance, switched and shared among a number of work
Induction Hardening / 419
stations. Manipulative techniques using mirrors annealed), they are more suitable for machining,
allow the beam to be directed to areas not while their strength properties and hardness can
accessible by other techniques, for example, the be subsequently refined by additional heat
bores of tubes. Set against these considerations treatment. Soft steels have good machinability,
are the high capital cost and low energy effi- while additional heat treatment improves their
ciency of the technique (Ref 1, 6, 10). mechanical properties to a desirable level.
Figure 1 presents a schematic of the heat treat-
ment methods used to achieve the desired
Steels for Surface Hardening mechanical properties (Ref 15, 16). Besides the
mentioned surface hardening, other thermo-
Heat treatable steels contain between 0.2 and chemical methods can be used to create a de-
0.6% C and can be carbon steels or low-alloyed sirable wear resistance. The steels for surface
steels (Table 1) (Ref 14, 15). hardening are from the group of heat treatable
In practice, engineers choose these steels steels, except that they usually contain between
because, in the soft condition (normalized or soft 0.35 and 0.5% C. Heat treatable steels can be
Table 1 Suitable steels for induction surface hardening

Chemical composition %
Type of steel C Si Mn P, max S, max Cr Mo Ni
1 0.330.39 0.150.40 0.500.80 0.035 0.035 ... ... ...
2 0.380.44 0.150.40 0.500.80 0.035 0.035 ... ... ...
3 0.430.49 0.150.40 0.500.80 0.035 0.035 ... ... ...
4 0.480.55 0.150.40 0.600.90 0.035 0.035 ... ... ...
5 0.500.57 0.150.40 0.400.70 0.035 0.035 ... ... ...
6 0.420.48 0.150.40 0.500.80 0.035 0.035 0.400.60 ... ...
7 0.340.40 0.150.40 0.600.90 0.035 0.035 0.901.20 ... ...
8 0.380.44 0.150.40 0.600.90 0.035 0.035 0.901.20 ... ...
9 0.380.44 0.150.40 0.500.80 0.035 0.035 0.901.20 0.150.30 ...
10 0.380.44 0.150.40 0.701.00 0.035 0.035 0.400.60 0.150.30 0.400.70
11 0.370.43 0.150.40 0.500.80 0.035 0.035 0.600.90 0.150.30 0.701.00
Source: Ref 14, 15
Fig. 1 Heat treatment methods for carbon and low-alloyed steels. Source: Ref 15
supplied in soft or refined conditions. Carbon surface to the austenitic zone above TA3 is of
steels with less than 0.35% C are in the nor- major importance, since not only the case depth
malized condition, and the products made from but also the right through-depth residual-stress
them can be in the soft or refined condition. For and microhardness profiles of the machine part
improvement of the wear resistance of these are to be provided. The through-depth micro-
steels, additional surface hardening can be hardness profile of the surface layer depends,
applied (Ref 14, 17, 18). with the given steel, on the induction-hardening
Below 0.35% C, the treated surface is insuf- conditions. Induction heating may be controlled
ficiently hard, and above 0.6% C, there is a by infrared thermometers. An infrared thermo-
danger of surface cracking after quenching. To meter is placed close to the induction coil, which
obtain a satisfactory hardening response in means that only the maximum surface temper-
induction surface hardening, it is necessary for ature may be measured. This technique is quite
the carbides to dissolve when the steel is heated. simple and practical. It allows a user to determine
Thus, steels that have been quenched and tem- the required, that is, optimal, power density for
pered and have small carbide particles respond heating. When a multiturn coil is used, for
better than steels with large spheroidized car- example, with longer workpieces, several infra-
bides. Similarly, the rate of solution of carbides red optical-fiber thermometers may be em-
in steels other than plain carbon and low-alloy ployed. Thus, with sufficient spacing between
steel is too slow for them to be satisfactorily two adjoining turns, the temperature may be
induction hardened. Steels must also have suf- measured at different locations on the machine
ficient hardenability to achieve the required part. A difficulty may arise due to nonuniform
hardness at the specified depth (Ref 17). heating along the coil movement. By collecting
The procedures are similar for fabrication of and processing the gathered measurement data,
machine parts made of carbon steels with more the process quality may be assessed. In scientific
than 0.35% C and alloyed steels, expect that research, temperature measurement with
because of the higher machinability required, thermocouples may be applied too. In induction
they must be in the soft-annealed condition. surface heating, temperature measurement using
For induction surface hardening, it is recom- thermocouples is quite exacting, since they are
mended to use steels with an appropriate carbon embedded in the machine part. The hardened
content (0.35 to 0.45%) and a careful selection state of steel is to be obtained at an appropriate
of alloyed elements. Usually, these steels have depth. The thermocouples embedded at a parti-
an increased silicon (0.15 to 0.40%) and man- cular location in the workpiece permit tem-
ganese (0.50 to 1.00%) content and other perature measurements at this location only. In
alloying elements, such as chromium, moly- progressive hardening, however, the momentary
bdenum, and nickel. More highly alloyed tool temperature is measured. Thus, the so-called
steels (O1, D2, D3, A1, and S1) and some mar- temperature cycles in heating, as well as in
tensitic stainless steels (AISI 416, 420, and cooling, are obtained. The temperature cycles
440C) are also sometimes induction hardened. make it possible to predict the efficiency of sur-
Among alloyed steels, chromium, chromium- face hardening. The choice of energy input, that
molybdenum, and chromium-molybdenum- is, power density, in an appropriate nomogram,
nickel steels prevail. With given combinations of that is, the choice of power and heating time with
alloying elements in induction-hardened mach- reference to the shape of the induction coil and
ine parts, the internal stresses in heating as well case depth, applies only to stationary heating and
as in cooling can be controlled. The steels are hardening of machine parts. This means that in
normally quenched in water. In certain cases, the progressive hardening the data obtained on the
alloy steels can be cooled by means of an oil energy input are of informative character only.
emulsion. Steels can also be oil quenched. Hav-
ing selected the right shape of the product and the
right choice of technology, it is possible to Main Features of Induction Heating
expect only minimal distortions of the workpiece
after completion of the heat treatment (Ref 14). Induction heat treatment is a segment of the
If the starting points are the hardness and much larger technical field of induction heating,
residual-stress profiles, then the induction-heat- which combines many other industrial processes
ing conditions should be adapted to the selected using the phenomenon of heating by induction
steel. Induction heating of a machine-part (Ref 1, 16, 1822).
Induction heating is often one of the most high-frequency alternating current, the choice of
effective heat treatment processes available for a frequency depending on heating requirements.
variety of applications, including: Induction heating power supplies are frequency
changers that convert the available utility line
Surface hardening frequency power to the desired single-phase
Through hardening power at the frequency required by the induction
Tempering and stress relief heating process. They are often referred to as
Annealing and normalizing converters, inverters, or oscillators, but they are
Grain refinement generally a combination of these. The converter
Precipitation hardening or aging portion of the power supply converts the line
Sintering of powdered metals frequency alternating current input to direct
In most of these applications, induction current, and the inverter or oscillator portion
heating is used to selectively heat only the por- changes the direct current to single-phase alter-
tion of the workpiece that requires treatment nating current of the required heating frequency.
(Ref 2). This usually means that the process can Many different power supply types and
be accomplished in a relatively short time and models are available to meet the heating require-
with high efficiency, because energy is applied ments of a nearly endless variety of induction
to the workpiece only where it is needed. heating applications (Ref 21, 22). The specific
One of the main features of induction heating application will dictate the frequency, power
compared with conventional heating procedures level (Fig. 2), and other inductor parameters
is that heat is generated in the workpiece itself. such as coil voltage, current, and power factor
In conventional heating procedures, the heat (cos Q) or Q factor (Ref 15, 19).
input achieved is only 5 to 200 kJ/m2s energy, Advantages in Surface Hardening of
whereas in induction heating this energy input Machine Components. Induction hardening
is 300 MJ/m2s. In induction heating, heat is most often used for surface hardening of
penetrates into the workpiece by the aid of machine components and has the following
Fig. 2 Typical power-frequency regions of induction heat treatment applications. Source: Ref 15, 19
advantages over other procedures (Ref 2, quenching agents that move the actual
1922): cooling rates close to the critical cooling
rate. These requirements can be met with the
Relatively short heating times right selection of quenching oils or polymer
Heating procedure is not strictly governed water solutions. Practical experience has
by hardening temperature. All that matters is shown that polymer water solutions are very
that the heating process does not end at too suitable for quenching of induction-heated
low a temperature, because sometimes it is surfaces, since optimal quenching can be
necessary for the transformation into auste- ensured with the right choice of concentra-
nite. The heating temperature is limited by tion of the polymer water solution.
the solidus-line temperature, since the pro- The induction-hardening procedure enables
cess is carried out while the material is in the the engineer, by simply adapting the shape
solid condition. Due to a short heating time, of the induction coil, to ensure the desired
there is no danger that the austenite grains shape of the hardened profile of the surface
would grow at higher austenitization tem- layer. Likewise, the engineer can surface
peratures, which also means there is no harden only that part of the surface (local
danger of coarse and brittle martensite for- hardening) on which a certain increased level
mation. of hardness and wear resistance is desired.
The quenching procedure is easy to perform, One of the main advantages of induction
contributing to short surface-hardening hardening is the possibility of hardening a
times. In progressive induction hardening, a surface layer only in certain places, at a
spray coil is located directly below the in- defined penetration depth and shape.
ductor and quenches the heated surface. In These advantages make it possible for
single-shot hardening, the inductor is de- induction hardening to be fully automated
signed to perform the function of heating as and are especially suitable for a large series
well as quenching. The coil around the work- of workpieces.
piece functions as an inductor in the heating Induction hardening always leaves com-
phase. After the austenitization temperature pressive residual stresses in the surface layer,
has been reached, the current is interrupted, which makes machine components more
and the coil functions as a spray for resistant to dynamic loads. Compressive
quenching. residual stresses in the surface layer after
Induction hardening is a short procedure that induction hardening prevent the occurrence
does not require any additional protection of cracks in dynamically loaded components
against oxidation. Thus, compared with and prevent the growth of existing cracks
other similar procedures, such as cementa- on the workpiece surface, if these are present
tion, it does not require much subsequent due to hardening or grinding.
machining. Induction hardening is appropriate for small-
Due to the nature of the procedure, the sized workpieces, since, by a well-chosen
workpieces, especially if symmetrically technology of heating and cooling or quen-
shaped, are less susceptible to undesirable ching, a hardened surface layer and a refined
deformations after induction hardening. The core can be ensured. Thus, the required wear
volume changes in the workpieces after hard- resistance of the machine component at a
ening the surface layer can be very accurately certain location as well as the required load-
predicted or estimated. The volume changes bearing capacity of the component can be
after induction hardening of thin layers are so created, experiencing only a slight loss in
small that quite often the function of the toughness of the core.
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 Induction Hardening of Machine Parts
cooling by self-cooling in air alone, that is,
by heat conduction from the heated surface Working Coil and Procedure during
layer into the remaining cold part of the Induction Hardening. Methods of induction
workpiece. With thicker layers and work- hardening of thin surface layers are always
pieces of greater mass, it is necessary to use adapted to the product size and shape and the
requirements of location, size, and shape of the always combined with a rotating motion of the
hardened layers profile. workpiece, or, on long workpieces, progressive
Designing an induction coil, however, is very hardening is used.
specific to a given application and can be con- Figure 3 shows different types of coils for
sidered the most important aspect of the engi- induction hardening designed for external sur-
neering of the system. There are several major faces (Fig. 3a), internal surfaces (Fig. 3b), and
functions that an induction coil must perform to front surfaces (Fig. 3c) of the workpiece. In all
make a job successful: three cases, the shape of the induction coil is
adapted to the size and shape of the work-
Induce current in the load so that the proper piece surface that is to be hardened (Ref 2, 15, 16,
heating pattern is obtained 1922).
Accomplish the proper heating pattern with Workpiece sizes depend on the manufacturing
as great an efficiency as possible possibilities of high-frequency coils and their
Provide an impedance match to the gen- types. This restricts the use of induction hard-
erator so that adequate power can be trans- ening primarily to small-sized machine compo-
ferred to the load nents. Induction hardening of internal surfaces
Have a geometry that will accomplish the requires a great deal of knowledge and experi-
aforementioned three major functions and ence in the design and manufacturing of induc-
permit easy loading and unloading of the tion coils of small diameters. The smallest
part being heated internal diameter of a machine component that is
First, the outer and inner surface on machine to be induction hardened depends on the manu-
components must be distinguished. The size and facturing possibilities of making small-diameter
shape of the induction coil must be adapted to coils, the capacity of the high-frequency gen-
the size and shape of the product, and during the erator (power, frequency), and the positioning
heat treatment process, the distance should be accuracy of the induction coil inside the
adapted to the workpiece. It is important that the workpiece.
design of the machine component be planned for The induction coils are made of materials with
induction-hardening specifics. A uniform gap the highest possible electrical conductivity.
size between the induction coil and the work- Consequently, copper or silver is usually chosen
piece surface is very important, since the energy to produce the induction coils. The induction
penetration depends on it. This is why single- coils are adapted to the shape of the machine part
shot hardening, where the workpiece and the coil at the location where heating, that is, surface
are at a standstill, is rarely used. To ensure a hardening, is performed. The choice of the size
uniform gap, single-shot hardening is almost and shape of the induction coil depends on the
Fig. 3 Significant types of coils for induction heating. Source: Ref 15, 16
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. Be-
cause the electrical conductivity of the material
is reduced due to plastic deformation of the
material, it is necessary to adequately anneal the Fig. 5 (a) Influence of multiturn coil on hardened-pattern
coils to soften the material. Thus, the highest profile. (b) Influence of shape and gap size of induction
coil turns on hardened-pattern profile. (c) Influence of number of
possible electrical conductivity is ensured. It is induction coil turns on hardened-pattern depth
also necessary to prevent the formation of cop-
per oxide at the coil surface to preserve highest inductor turns along the specified hardened
possible electrical conductivity of the material. pattern. This means different divisions among
This will provide a high efficiency of induction the individual coil turns. Under the same heating
heating. conditions, considerable difference in the
In single-shot induction heating, that is, hardened-pattern depth will be obtained with
induction surface hardening, it is important to the same heating time.
provide the same gap size all around the cylin- Figure 5(b) shows different examples of
drical machine part. Because such an arrange- shaping a multiple-turn induction coil around
ment can be very exacting, the heating process is the cylindrical machine part. The figure indi-
usually performed with the rotation of the cates that the gap size along the cylindrical part
machine part. Figure 5(a) shows single-shot changes, which produces different hardened-
hardening performed with a different number of pattern profiles. Figure 5(b) shows the influence
of a multiple-turn coil with too-large divisions of Quenching techniques are an important

the individual coil turns; therefore, the same heat- design feature of induction-hardening equip-
ing depth, that is, hardened-pattern depth, along ment. The important questions to be answered
the coil length is not achieved. Thus, individual when determining quenching systems include
coil areas showing a smaller hardened-pattern (Ref 2, 17, 23, 24):
depth are obtained at the machine part.
Figure 5(c) shows the influence of the cross Workpiece size and geometry
section of a single-turn coil on the hardened- Hardenability of steel
pattern profile. Type of austenitizing operation (surface or
For hardening of axles, shafts, and similar through hardening)
lengthy machine parts, the scan-hardening pro- Type of heating method (single-shot or
cess with a single-turn coil of an adapted cross scanning)
section is chosen. Such a hardening process is Type of quenchant
necessary, otherwise, high-energy densities and
The two most common types of systems
very long multiple-turn induction coils would be
consist of spray quench rings and immersion
needed. Figure 6 shows a diagram of heating, that
techniques. When quench rings are used for
is, through hardening, of a cylindrical rod of
round bars, their shape, like the coil, is generally
different diameters (D, in centimeters) with dif-
round (Ref 2, 23, 25). Figure 7 shows different
ferent inductor powers (P, in kilowatts), and a
ways to single-shot harden the surface layer.
given frequency ( f, kilohertz) and inductor
Common to all of them is that heating is
movement (v, in cm/s). The diagram shows a
performed along the length of the layer, so that
linear dependence between the connected power
the induction coil embraces the particular part of
at the coil and the depth of hardening, which
the workpiece where heat treatment is to be
means that the choice of a higher power produces
performed (Ref 15, 16). The ring may be located
hardening at a greater depth, the movement
concentric with the coil (Fig. 7c) or directly
velocity being the same. It can also be established
underneath or alongside it (Fig. 7b), as in the
which diameters of the axles or shafts can be
single-shot induction hardening setup.
through hardened if the same power is available
In any of these ways, it is important to ensure a
and different movement velocities are chosen.
constant gap size between the induction coil and
It is possible to distinguish between:
the workpiece surface. This is usually achieved
Single-shot hardening by a rotating motion of the workpiece. In the first
Scanning or progressive hardening case, shown in Fig. 7(a), the cooled induction
Tooth-by-tooth hardening or gap-by-gap coil embraces the object at a certain height. After
hardening heating to the austenitization temperature is
Fig. 6 Selection of heating conditions in scan hardening to provide through hardening of a cylindrical rod
completed, the workpiece is quenched in a heat treatment, the workpiece is rotating, and in
quenching bath, or quenching is done by moving this way, uniform heating and quenching are
the workpiece into the quenching spray located ensured. The heating and quenching conditions
beneath the induction coil (Fig. 7b). Figure 7(c) created in this way produce a uniform hardened
shows a method of heating and cooling in the surface layer and ensure repeatability of thick-
induction coil, which is made so that it enables ness and shape of the hardened profile.
heating of the workpiece surface for a certain In induction scanning or progressive hard-
period, and then, on switching the current off, the ening, workpieces move through the quench ring
same coil is activated as a spray for quenching. and coil, with quenching occurring immediately
The induction coil is usually made of copper after heating (Fig. 8). For nonsymmetric work-
tubing with additions of alloying elements pieces, the quenching system, like the coil, is
that do not affect the electrical conductivity of generally the same shape as the workpiece (Ref
copper but contribute to good machinability 15, 16).
of the material. Good machinability is required The second method of induction hardening
because it is necessary to make the openings involves a group of procedures of progressive
for the quenching spray. Thus, during the entire or progressive-rotating hardening. This type of
Fig. 7 Single-shot induction surface hardening of a cylindrical workpiece. Source: Ref 15, 16
hardening is employed when the required length Figure 9(a) shows single-shot hardening of tooth
of the hardened surface is longer than the length tips, where the coil encircles the whole gear.
(size) of the coil. In this case, the correct heating This method is very simple and suitable for high-
mechanism can be ensured by the optimal in- frequency heating of gears with a module up
duction coil design (number of coil turns, length to 3 mm and for smaller gear diameter. In this
of the coil) and suitable energy inputs for a given case, it must be ensured that the tooth flanks are
size and profile of the hardened layer (Ref 23, hardened at least along the length that is other-
25, 26). To ensure the uniformity and repeat- wise activated in the mesh, since the thickness of
ability of the hardened layer, progressive motion the hardened layer varies, being highest at the
of the induction coil must be provided, or the tooth tip and gradually decreasing toward the
workpiece is fed progressively through the coil tooth root. This method of hardening is suitable
while rotation of the workpiece takes place at the only for gears subjected to low loads, just to
same time. This method of induction hardening increase the wear resistance of the gear (Ref 16,
is presented in Fig. 8. It is possible to design the 23, 26, 27).
induction coil separately from the spray (Fig. 8a) Figure 9(b) shows induction surface hard-
or combined so that the induction coil has ening, where the whole tooth and a certain area
downward-directed spray openings for quench- below the tooth root are hardened. This method
ing in its bottom part (Fig. 8b). of induction surface hardening is appropriate for
Gear wheels belong to very demanding gears with a module up to 5 mm. Heating is
machine elements, from the point of view of achieved by two systems: first, with a current of
mechanical machining as well as heat treatment. medium frequency, and then in the final phase
In hardening smaller gears, that is, gears with for a relatively short period by heating with a
small modules or small diameters, the method of high-frequency current. By quenching, a desir-
single-shot hardening can be used to ensure a able hardness of the tooth flank surface and
varying thickness of the hardened layer (Ref 15, increased strength of the gear tooth can be
16). The thickness is maximum on the tooth tip achieved. In the root of the tooth, due to a refined
and then decreases toward the tooth root. microstructure, an increased material fatigue
Fig. 8 Progressive induction hardening of a cylindrical workpiece. Source: Ref 16

strength can be noted. Gears heat treated in this A number of graphs and nomographs are
way are suitable for highly loaded machine parts available for this purpose, offering a selection of
(Ref 2830). heat treatment conditions for heating and cool-
Figure 9(c) shows hardening of individual ing. The most important data are the starting
gear teeth, where the coil encircles the tooth and point data on energy input and frequency of the
the individual teeth are heated in turn and current, and the temperature to achieve in
quenched directly after heating. In the case of induction hardening. From these data, the time
large-module gears, simultaneous heating of the necessary for heating, in the case of single-shot
left and right flank of the same gear tooth can be hardening, or the feed rate of the workpiece or
performed (Fig. 9d). the rate at which the coil should move along the
Induction hardening of all gear teeth (Fig. 10) workpiece can be defined (Ref 2, 1820, 23, 27,
can be done only for gears with a module smaller 31, 32).
than 3 mm and a gear diameter up to 100 mm. Figure 11(a) illustrates induction surface
This kind of hardening requires high-power heating with a medium-and high-frequency
frequency generators, for example, 200 kW. current. The procedure is known as double-
For gears exposed to higher loads, where frequency heating (Ref 27). Here, the gear is first
refinement of the whole tooth is desirable, a placed into the coil fed by the medium-
power of 30 to 40 kW and longer heating times frequency current. Then, the gear is moved into
can be selected. the high-frequency coil, where only the surface
To ensure a desired penetration depth and layer of the gear tooth is reheated with a high-
profile of the induction-hardened layer, the fol- frequency current. When both phases are com-
lowing heat treatment parameters can be varied pleted, the gear is dropped or moved into a
in a single-shot or progressive induction hard- quenching bath. In this method of heating with a
ening by choosing a suitable power density and double frequency, the progressive motion of the
feed rate: workpiece can be combined with additional
rotation.
The size and shape of the induction coil Additional rotating motion ensures a uniform
adapted to the workpiece reheating of the surface layer and results in
Kind of steel and its thermal properties uniform microstructural changes on the left and
Size and mass of the workpiece on the right side of the tooth. In this way, undesirable
location where induction hardening is to be and nonuniform dimensional deviations be-
performed tween the left and right tooth flanks are pre-
Quenching agent and method of quenching vented. The same method of induction surface
Fig. 9 Different methods of induction surface hardening of gear wheels. Source: Ref 15, 16
heating is shown in Fig. 11(b), where the same the desirable hardness of the adjacent tooth. To
coil is used for heating with a medium- and high- avoid this, special protection made of thin cop-
frequency current (Ref 27). per sheets is used to prevent heating of the
The second method is hardening of individual adjacent tooth flanks. The conditions are pre-
gear wheels or individual gear gaps. Figure 12 sented in Fig. 15(ac) (Ref 27). In addition to
shows induction heating of a gear wheel tooth- protecting adjacent tooth flanks, the shape of
by-tooth and Fig. 13 shows induction heating the coil around the tooth has other effects. By
with the gap-by-gap method. In both methods, changing the shape and position of the coil
the tooth surface is first heated with the induc- around a particular tooth, it is possible to achieve
tion coil and then quenched with a specially equal hardened layer profiles, as in the case of
adapted spray system. single-shot hardening. The thickness of the
Figure 14 shows the shape and position of the hardened layer is at a maximum at the tooth
massive inductor placed around the gear tooth tip and then gradually decreases toward the tooth
(Ref 27). The coil is shaped so that the gap root. In Fig. 15(c), the coil is placed slightly
between the coil and the tooth surface varies. lower, heating only the tooth flank along
Only in this way is it possible to ensure a uni- the entire height from the root to the tip.
form thickness of the layer in the middle and on Figure 15(b) shows the lowest position of the
the edge of the tooth (Ref 3337). coil while still reaching below the tip of the
Difficulties with this method of gear wheel tooth. The coil positioned in this way does
hardening occur when the tooth gaps are too not heat the tip of the tooth but only the surface
small and the coil heats up the adjacent flank. of the tooth flank from the tip to the root and
This method of induction surface heating of gear yields maximum thickness of the hardened layer
teeth is often not appropriate, since after heating in the middle of the tooth. In this case, too, small
the adjacent tooth, the next step does not ensure gears are heated with a high-frequency current,
Fig. 10 Single-shot surface induction hardening of gear wheel tooth tips

Fig. 11 Induction heating system for gear wheel heating by double frequency (medium/high). Source: Ref 27
Fig. 12 Tooth-by-tooth induction surface heating process of Fig. 13 Gap-by-gap gear wheel induction surface heating by
installing the induction coil around the gear wheel moving the induction coil for heating and moving or
tooth for heating, followed by shifting spray quenching installing spray quenching
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 hard-
ening 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 uni-
form gap between the coil and the flanks of
Fig. 14 Relative position of the massive induction coil to the two adjacent teeth. The tooth-gap hardening
gear wheel tooth at induction heating. Source: Ref 27
method is very demanding and requires much
Fig. 15 Influence of induction coil height on profile of induction heating surface layer in an individual tooth. Source: Ref 27
experience and knowledge to achieve the be ensured to permit austenitizing. In fast heat-
desirable properties of the gear. ing, austenitizing can be accomplished only
This method is also known as contour hard- by heating the surface and subsurface to an
ening. It is an ideal method for heat treatment of elevated temperature. For example, with a
gears, because it increases the hardness on the heating time, t, of 1 s, for total homogenizing,
tooth surface only slightly while decreasing the a maximum surface temperature, Ts, of 880 C
load-bearing capacity in the root of the tooth. should be ensured in the first example and a
Gears heat treated in this way exhibit very good much higher surface temperature, 1050 C, in
behavior in operation, because compressive resi- the second example. This indicates that ap-
dual stresses are present in the root of the tooth. proximately 170 C higher surface temperature,
Gears with induction-hardened flanks, given that DTs, should be ensured in the second example
the dimensioning is carefully carried out, can (normalized state) than in the first example
achieve the highest fatigue strength. To verify (quenched and tempered state).
the results of induction surface hardening, it is Figure 17 shows a space TTA diagram
necessary to take certain measures for control- including numerous carbon steels with different
ling the quality of the hardened layer. For this carbon contents. The TTA diagram gives par-
purpose, hardness and microhardness measure- ticular emphasis to the characteristic steels, that
ments, supported by microstructural analysis, is, 1015, 1035, 1045, and 1070 steels, and their
are commonly used. A disadvantage of this pro- variations of the transition temperature, TA3 ,
cedure is that, due to the method of heating and with reference to the given heating rate and the
quenching (nonuniformly overheated left and corresponding heating time (Ref 38). Such a
right tooth flank), slightly higher dimensional temperature difference ensures, with regard to
deviations may be obtained than in the case of the heating and cooling conditions of the speci-
simultaneous hardening of both flanks of the men, the time required for austenite homo-
same tooth (Fig. 15d) (Ref 27). genizing, tA, in the given depth.
Metallurgical Aspect of Induction Surface Figure 18 shows a shift of the transformation
Heating. Prior to transformation hardening, temperature, which ensures the formation of
an operator should calculate the processing inhomogeneous and homogeneous austenite
parameters for the given power system. The pro- within the selected interaction times (Ref 39).
cedure is as follows. Some of the processing A shorter interaction time will result in a
parameters are chosen, some calculated. The slightly higher transformation temperature, TA1 ,
choice is usually left to the operator and his and also a higher transformation temperature,
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 hard-
ened layer, a temperature a little higher than the
transition temperature, TA3 , should be ensured.
Because of a very high heating rate, the equili-
brium 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 time-
temperature-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 Fig. 16 Time-temperature-austenitizing diagram for steel
microstructure, a sufficiently long time should 1053 in various states. Source: Ref 38
TA3 . To ensure the formation of homogeneous Figure 18(a) shows a temperature-time diagram
austenite with shorter interaction times, con- for austenitizing of Ck 45 steel.
siderably higher temperatures are required. The isohardnesses obtained at different in-
teraction 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 hyper-
eutectoid alloyed steel.
The diagram indicates that with short inter-
action times, which in laser hardening vary
between 0.1 and 1.0 s, homogeneous austenite
cannot be obtained; therefore, the micro-
structure 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
Fig. 17 Influence of heating rate and carbon content stresses and reduce wear resistance of such a
on austenitic transformation temperature. Source:
Ref 38 material.
Fig. 18 Temperature-time-austenitizing diagrams with lines of resulting hardness for various steels. Source: Ref 39
The distribution of residual stresses in heat that the second zone has not shown up. A similar
treatment procedures where only the surface of heat treatment can be applied to the spline inside
the workpiece is heated (induction hardening, the gear. A gear heat treated in this way is more
flame hardening) differs greatly from the pro- resistant to wear and corrosion and should have
cedures where heating is performed throughout high resistance to fatigue in bending because of
the entire volume (nitriding, cementation). In a smaller thickness of the layer in the second
nitriding and cementation, the aforementioned zone.
second layer in the subsurface does not appear at Many induction-hardened gears are treated in
all, because the direction of the heat flow is the tooth gap, that is, in simultaneous heating
opposite to the direction of the heat flow in and subsequent quenching of two adjacent
induction and flame hardening. The resultant flanks of the left and right tooth of the gear with a
operating tensile stresses on the surface or in the hardened root area. In view of the variety of
surface layer can thus be considerably smaller. methods to induction harden gears and the other
Due to the surface hardness, induction and flame possible ways of gear hardening, it is unwise to
hardening lowers the fatigue strength of mach- make a hasty decision in selecting the procedure.
ine components; therefore, care should be taken It is necessary to make a thorough analysis,
to diminish all detrimental effects in the surface including answers about the expected quality of
layer. the hardened layer and analysis of the operation
A typical example of induction surface hard- loads of the machine components.
ening is surface hardening of gears that are Highly loaded gears can be successfully
heated with a low heating rate and relatively low induction hardened if a high-frequency current
current frequency. The outer hardened zone in- and high-input power are used. High frequency
cludes almost the entire height of the gear teeth, is necessary in order to obtain a sufficient thick-
whereas the second zone is in the tooth root area. ness of the hardened layer on the tooth flanks and
A gear heat treated in this way will meet the a fine hardened layer in the tooth root area. A
wear-resistance requirements expected of the high-input power is necessary to increase the
gear tooth, while the strength of the other part of heat gradient, which makes the size of the sec-
the tooth is of minor importance. The fatigue ond zone smaller, resulting in a thinner second
strength in this case will be relatively low due to layer with tensile residual stresses.
high tensile residual stresses in the tooth root, On gears with a small diameter, the induction-
that is, in the second zone where the operating or hardening equipment should be able to harden
load tensions and the tensile residual stresses are the gear throughout its volume. Induction hard-
summed up. Figure 19 shows another example ening of entire gears has some advantages, such
of an induction-hardened gear where the fre- as rapid heating, no danger of decarbonization
quency of the current was so high that the gear and oxidation, high productivity, and repeat-
tooth is heated along the flank surface and tip, as ability in the gear quality. Thus, the hardened
is the case in cementation. The energy input in layer is only defined by the first zone to the depth
heating a gear tooth or the whole gear was such that is greater than the height of the gear tooth.
Fig. 19 Typical examples of induction surface hardening of (a) carbon steel and (b) alloyed steel gears produced from carbon steel (a)
and alloyed steel (b)
In induction hardening and quenching, lowering Figure 21 shows the dependence between
or even the disappearance of compressive resi- carbon content and the two martensite transfor-
dual stresses is achieved in the tooth root area, mations (Ref 38).
causing a considerable decrease in fatigue Consequently, an increase in carbon content
strength and a higher propensity for fatigue. in steel results in the selection of a lower critical
In flame heating, lower temperature gradients cooling rate. In general, the microstructures
are reached than in induction heating. This formed in the surface layer after transformation
results in increased thickness of the layer in the hardening can be divided into three zones:
second zone. Possible harmful effects due to the
A zone with completely martensitic micro-
disappearance of the second zone of the hard-
structure
ening layer can be avoided by heating the entire
gear, which is possible to do in commonly used A semi-martensitic zone or transition
microstructure
furnaces.
A quenched and tempered or annealed zone
When the induction coil has stopped heating,
an austenitic microstructure in the surface layer with reference to the initial state of steel
should be obtained. Then, the cooling process Transformation hardening of steel starts
for the austenitic layer begins. To accomplish from its initial microstructure, which is ferritic-
martensite transformation, it is necessary to pearlitic, pearlitic-ferritic. or pearlitic. In steel
ensure a critical cooling rate that depends on heating, transformation into a homogeneous
the material composition. Figure 20 shows a austenitic microstructure should be ensured.
continuous cooling transformation diagram Figure 22 shows the dependence of the maxi-
for EN19B steel, including the cooling curves mum surface temperature obtained in induction
(Ref 38). heating with a machine part made of steel with
Because carbon steels have different carbon 0.8% C and a pearlitic microstructure (Ref
contents, their microstructures also show differ- 40, 41). How the through-depth heating of the
ent contents of pearlite and ferrite. An increased machine part will proceed depends on the maxi-
carbon content in steel decreases the tem- mum surface temperature obtained and the
perature of the beginning of martensite trans- power density, which means that the through-
formation, TMS as well as of its finish, TMF depth hardness profiles will differ.
Fig. 20 Continuous cooling transformation diagram of EN19B steel. Source: Ref 38

With the same power density, this means that grains have little time left to grow. It can be
a higher maximum surface temperature will be assumed that the grains remain fine and do not af-
accomplished with a longer heating time. Since fect through-hardenability, that is, the through-
the heating times are usually short, austenitic depth hardness profile. The diagram shows the
heating conditions that provide, in all cases, a
completely homogeneous austenitic microstruc-
ture to a certain depth and, consequently, con-
stant hardness. Then follows a transition zone
consisting of homogeneous austenite and some
inhomogeneous austenite and pearlite. Conse-
quently, 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 sur-
face temperature, a maximum hardness of ap-
proximately 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 max-
imum 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
Fig. 21 Influence of carbon content in steel according to start decreases to that of the parent metal. With higher
and finish temperature of martensitic transformation maximum surface temperatures, the lowest
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
hardness is obtained at the surface, and then, to adjust the shape of the induction coil.
with a greater depth, it slowly decreases. With The heating temperature is higher than the
steel having a pearlite-ferrite or ferrite-pearlite hardening temperature, so that after quenching,
microstructure, a microstructure consisting not unwanted increased surface hardening is ob-
only of inhomogeneous martensite but also of tained. In cases of treatment of two local surface
pearlite-ferrite grains will occur in the transition areas on the workpiece that are in direct pro-
zone. Such a microstructure results in a stronger ximity, if heat is applied two times in sequence,
decrease in hardness than with the steel having a the heating of the second area may result in
pearlite microstructure. It is essential for effi- tempering of the previously treated area. In these
cient induction surface hardening that constant cases, the area hardened first may have a tem-
hardness is obtained to a sufficient depth and that pered microhardness with a lower hardness.
the hardness profile of the transition zone is Therefore, research has been done on how to
adequate. It is only in this way that notch defects form and adapt the coil to offer a more con-
in the hardened layer may be prevented and centrated magnetic flux. By adjusting the con-
better operation of the machine part under centration of the magnetic flux, it is possible to
dynamic load may be ensured. 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
Magnetic Flux Concentrators concentration of magnetic flux (Ref 19, 21). The
purpose of magnetic flux concentration is to
In induction surface heating, unwanted areas
improve the efficiency of surface heating and
on the workpiece are often heated or even hard-
reduce heat losses. The use of a magnetic flux
ened. This type of problem appears when the
concentrator enables selective local heating on
shape of the product surface is very complex,
workpiece/product areas with complex geo-
and it is therefore difficult to adjust the coil for
metry. Figure 23 shows a straight conductor with
local heating and quenching. For more de-
current density distribution in points A and
manding shapes, shields from materials with
B for three cases (Ref 19, 42):
good heat conductivity, for example, copper, are
often used on the product to prevent heating of Current density distribution in a straight
the workpiece. Heating the copper shield pre- conductor (Fig. 23a)
vents undesirable heating of certain areas as well Current density distribution in a straight
as loss of heat. Such an example can be seen in conductor when the conductive material
Fig. 15, which shows the protection of two ad- (workpiece) is approached (Fig. 23b). The
jacent heat flanks while heating the tooth in the current density is greater in point A, closer
middle with an induction coil (Ref 15, 1921, to the workpiece material, than in point B.
28, 42). As a result, the workpiece is locally heated
Heat losses occur due to local heating of the over a longer lengthlonger than the width
workpiece surface that is not to be hardened; of the conductor. This is referred to as current
however, this is done because it is not possible redistribution due to the proximity effect.
Fig. 23 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
By placing the conductor into the magnetic reflected in the different heating/workpiece
flux concentrator made of powder metals depth profiles created by the differences in the
based on iron, nickel, cobalt, and other power density distribution (Ref 19, 42).
powder materials, it is possible to achieve a Figure 25 shows the same area and shape, but
higher density of the magnetic field heating was performed with a magnetic flux
(Fig. 23c). concentrator (Ref 19, 42). This is placed on the
left and right sides of the surrounding coil, which
The concentrated magnetic field results in contributes to higher concentration of the mag-
localized heating on only those areas that are netic flux to prevent heating of the sides of the
meant to be hardened. The magnetic flux con- workpiece at this place. A redistribution of
centrator surrounds the conductor, so that the power density took place, and a desirable local
current density on the conductor surface is heating profile of the workpiece was achieved.
redistributed, as is the heating of the workpiece Advantages offered by the magnetic flux
material. concentrator are:
The effects of the magnetic flux concentrator
depend on: Smaller consumption of power
Improved efficiency of heating
Workpiece material and shape of the local Better exploitation of equipment due to
area to which heat is applied shorter heating times
Workpiece material and concentrator shape More selective heating of the workpiece
The current density in this case is the highest areas
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 con-
ductor, but, due to the high density of the mag-
netic flux, a considerably greater depth of the
heated area is achieved.
As for concentrator material, different appli-
cations 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 Fig. 24 Heating profile on rotational workpiece with
induction coil. Source: Ref 42
Resistance to high temperatures
Cooling abilities
Resistance to cooling effects of fire extin-
guishers
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 con-
centrator 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 Fig. 25 Heating profile on rotational workpiece with flux
differences in the current density, which are concentrator. Source: Ref 42
Achievement of desirable heating profiles of is incurred at the point where teeth or other
hardened layer profiles machine parts are most strongly loaded, which is
High repeatability of the procedure in terms at the transition from the hardened subsurface to
of hardened profile size and hardened layer the nonhardened core, where an increased stress
microstructure concentration will occur. Such heating circum-
Efficient protection from unwanted heating stances may be avoided by choosing lower-
of adjacent areas and successful prevention energy inputs or power densities. In this case,
of reheating and softening the already hard- however, the case depth is more difficult to con-
ened layers trol. If a hardened-and-tempered microstructure
Elimination of detrimental effects on the is to be provided at the tooth inside as well,
operators health due to exposure to the heating with two frequencies is applied. The
magnetic field hardened-and-tempered microstructure may be
Improved operation lifetime of the heating accomplished with only one frequency in heat
equipment and higher productivity treatment of smaller gear wheels when the induc-
Less workpiece distortion and lower costs in tion coil encompasses the entire gear wheel.
straightening and/or final grinding sub- The second procedure is tooth-by-tooth
sequent to hardening heating (Fig. 26b). In this case, the induction coil
Reduction of the quantity of rejected parts in is moved against a tooth so that an appropriate
terms of required size and shape of the hard- gap is provided between the concentrator and the
ened layer and workpiece distortion and surface of both flank profiles of the same tooth.
cracking The difference between the two techniques is
that in gap-by-gap heating, two adjacent tooth
Figure 26 shows two hardening procedures profiles and the root section of a gear wheel are
for tooth profiles of gear wheels using appro- heated, whereas in tooth-by-tooth heating the
priate induction coils with a ferritic core con- entire gear tooth is heated. Thus, in the first
centrating the magnetic field (Ref 19). example, only hardened gear wheel tooth pro-
The first procedure is known as gap-by-gap files are obtained, whereas in the second ex-
hardening (Fig. 26a), because the induction coil ample, hardened tooth profiles of the gear wheel
with the ferritic core is moved in the tooth space and quenched and tempered tooth inside are
so that a suitable gap between a tooth profile and obtained.
the inductor is provided. The advantage of using Figures 27(a to d) show various automobile
magnetic flux concentrators rather than the parts that were induction surface hardened using
conventional coil is that, with the same energy the single-shot or scan-hardening technique (Ref
input, heating time is shortened. Unfortunately, 19). They were prepared for macroscopic and
the use of concentrators also shows some defi- microscopic examinations of the hardened layer.
ciencies, since the transition zone between the The segments of the individual specimens of
hardened and nonhardened microstructures, various characteristic and exacting automobile
which is very important for hardening, is lost due parts after grinding, polishing, and macroetch-
to rapid heating. This results in less favorable ing permit identification of the profile of the
residual-stress and microhardness profiles of the hardened pattern, for example, in the cross-
hardened layer. Thus, a greater risk of failure sectional and longitudinal direction of the part,
respectively, as well as an analysis of the micro-
structure 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 single-
turn 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 mea-
Fig. 26 Gap-by-gap and tooth-by-tooth induction hardening surement, then the part was not subjected to
of gear wheels. Source: Ref 19 distortion. A sufficiently high initial hardening
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 statis-
tical 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 uninterrup-
ted production provided. These types of non-
destructive testing are long-standing; therefore,
nondestructive methods of testing the material
condition after induction hardening are grad-
ually being introduced. Of all the methods, mag-
netic 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
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 intro-
duced 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
Fig. 27 (a) Single-shot inductors used for both track (lobes)
in the thin surface layer, whereas the inner
and shaft of this automotive component. The part is
sectioned and acid etched to show the hardness pattern. All tracks core remains cold or is only slightly heated
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
(low-mass workpieces).
cam shafts that have been selectively induction surface hardened. The depth of current penetration depends on
The cam shaft was forged, heat treated, then final ground. No workpiece permeability, resistivity, and the
premachining was necessary. The equipment used a dual-spindle
transfer mechanism; the coil was a four-turn coil that heated four alternating-current frequency. Because the first
lobes per spindle at a time. Power was applied for each set of four two factors vary comparatively little, the great-
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 est variable is frequency. Depth of current
crankshaft journals resulting from the stationary induction- penetration decreases as frequency increases.
hardening process. (d) (Left) An unacceptable nonuniform hard-
ness pattern due to the nonuniformity of the workpiece, scanned High-frequency current generally is used when
with a single-turn inductor. (Right) An acceptable hardness shallow heating is desired; intermediate and low
profile achieved with a single-shot inductor. Source: Ref 19.
Courtesy of Inductoheat, Inc.
frequencies are used in applications requiring
deeper heating.
temperature provides an adequate through-depth Most induction surface-hardening applica-
microhardness of the hardened profile. The tions require comparatively high power den-
analysis is considered a destructive method, sities and short heating cycles to restrict heating
to the surface area. The principal metallurgical where t is time, measured in seconds; and dHC is
advantages that may be obtained by surface the depth of penetration for heat conduction,
hardening with induction are the same as for measured in millimeters.
flame hardening. The drop in magnetic perme- The total depth of penetration is obviously
ability of steels depends on the temperature line, dT = dCS +dHC. It should be stressed that these
TA2 , where steel transforms from magnetic into expressions give only a rough estimate of the
nonmagnetic ferrite. The larger the effect of the depth of penetration, and they have been inclu-
magnetic permeability change on the tempera- ded here only to show the fundamental effects of
ture line, TA2 , the smaller the carbon content in frequency and time.
the steel (the larger the proportion of ferrite in In flame heating, the temperature achieved on
the steel) and vice versa. Due to rapid heating, the surface at equal energy input is considerably
phase transformation moves upward toward higher than in induction heating, the overheating
higher temperatures. The temperature-time and the hardened layer thickness being depen-
curves of heating along the depth of a cylindrical dent on the heat conduction of the workpiece
component depend on the kinetics of the mag- material. Figure 28(a) shows the temperature-
netic transformation, TA2 , and the effects of other time variation over the cross section of the
phase transformations during induction heating. workpiece in flame heating (Ref 15). Charac-
A thickness of 1.0 to 1.5 mm is reached with a teristic of this variation is that the temperature
medium-frequency current. The temperature- rapidly changes with time, and therefore, the
time variation over the cross section of the steel conditions for the formation of a homogeneous
workpiece is a function of the following factors: austenitic microstructure are not fulfilled.
Figure 28(b) shows the temperaturetime
Penetration depth of eddy currents variation over the cross section of the workpiece
Heat conduction of the material in induction heating. The temperature variation
Heating rate of the surface is very similar to that in flame heating up to
Initial temperature of the surface magnetic transformation, that is, to line A2. At
Size and shape of the workpiece temperatures higher than line A2, eddy currents
grow characteristically, and the rate of heating
The depth of penetration of the heat is gov-
decreases sharply. This slows down the heating
erned mainly by the power and frequency em-
above temperature line A2. A reduced rate of
ployed. The normal power density is 0.1 to 2 kW/
heating on the surface provides the conditions
cm2 of the heated surface. The relationship
for faster heating into the depth of the work-
between depth of penetration and frequency can
piece. This figure shows that a relatively thin
be calculated approximately by using simplified
layer is heated up, but the layer has a rather
expressions, which are valid for the temperature
homogeneous austenitic microstructure. The
rise in steel up to the hardening temperature
temperature-time variation on the workpiece
(Ref 16):
cross section, or the temperature field, depends
on the workpiece size and shape. Thus, in heavy-
20
dCS = p cold state (20 C) mass workpieces, faster heat abduction into the
f remaining cold part of the workpiece is achieved,
and that is why the actual variation of tempera-
500
dHS = p hot state (800 C) ture over the cross section is steeper. This means
f that in heavy-mass workpieces, a higher surface
temperature than in low-mass workpieces must
where dCS is the depth of penetration in the cold be ensured to grant the same penetration depth.
state, measured in millimeters; dHS is the depth The microstructural changes in induction
of penetration in the hot state, measured in hardening depend to a large extent on the rate of
millimeters; and f is the frequency, measured in heating and subsequent cooling. The rates of
hertz. heating range from one to a few seconds, which
Due to heat conduction in the material during means that the diffusion processes may become
heating, the overall depth of penetration is larger. jeopardized. In steel, transformation of pearlite
It is possible to calculate the additional penetra- into austenite takes place in induction heating at
tion due to heat conduction from the expression: almost the same temperature as in conventional
p heating. In subeutectic steels suitable for surface
dHC =0:2 t hardening, it is important that the induction
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 heat-
ing rates (Ref 15, 27). The graph shows that
austenitization is clearly influenced by the
heating rate, especially when carbon con-
centrations 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 hard-
ening 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
Fig. 29 Influence of induction surface heating rate on hard-
higher with faster surface heating and smaller ening temperature for subeutectic steels. Source:
carbon content in the steel. Ref 15, 27
In surface hardening alloyed steels, a better
homogenization of the austenite is achieved in
the heating phase, and when quenching is com- hardness in the transition area with a hardened
pleted, a very homogeneous martensite with a and nonhardened microstructure.
uniform microhardness along the depth of the A very important heating requirement in the
hardened layer is derived. Unfortunately, there hardening procedure, besides the austenitization
is a transition temperature range with a smaller temperature, is the time necessary for austeniti-
ferrite content, which causes a sharp drop in zation, since both of these heating parameters
affect the size of austenitic grains. Austenitic range moves to higher temperatures. This means
grain size is, on the other hand, dependent on the that quenching from lower temperatures leads to
martensite formed subsequent to quenching and imperfect hardening, while higher temperatures
is reflected in the toughness of the surface layer. yield medium or rough needles of martensite.
That is why heat treatment conditions are sought The heating rate of 50 C/s is sufficient to reach
that would ensure the finest and most homo- a hardening temperature range between 850 and
geneous austenitic microstructure in the heating 925 C. In the case of a higher-energy input that
and overheating phases, leading to the formation heats up the thin surface layer with a rate higher
of very fine martensite with the highest possible than 140 C/s, the required temperature range
toughness of the hardened layer after quenching becomes 870 to 970 C. In both cases, a surface
(Ref 27, 29). This microstructural condition and hardness of 60 HRC is reached subsequent to
the achieved mechanical properties ensure good quenching.
wear resistance of machine components (gears) Figure 32(ae) shows the entire process of
and their good response under dynamic loads. induction hardening a cylindrical component
Figure 30 shows the growth of austenitic with a small diameter or cross section, which,
grains for different rates of induction heating for subsequent to hardening, leads to self-tempering
a steel with 0.4% C that can be used for induc- (Ref 15, 27).
tion hardening (Ref 15, 27). With increasing Figure 32(a and b) show induction heating of
rates of heating, the austenitic transformation a thin surface layer to the austenitic temperature
moves toward higher temperatures. A higher range, ensuring, a sufficient thickness of the
heating temperature creates a higher formation austenite layer d2 subsequent to quenching. The
rate of austenitic crystal and therefore fine grains process of quenching or self-tempering is shown
of austenite. Subsequent to quenching, these fine in Fig. 32(c and d), where, due to heat conduc-
grains of austenite ensure a very fine martensitic tion into the remaining cold part of the work-
microstructure. piece, the temperature on the workpiece surface
Figure 31 shows the relationship between layer increases, and thus, the thickness of the
the hardened layer hardness and the heating austenite layer increased to d3 or d4, respec-
rate and temperature for a steel with 0.45% C tively.
(Ref 15, 27). Due to a small workpiece cross section or low
For each steel, there exists a certain tem- workpiece mass causing heat conduction, the
perature range after hardening that yields the temperature in the middle of the workpiece rose
best microstructural condition and thus the best to a point corresponding to the tempering tem-
properties. With higher rates of heating, this perature 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

Fig. 30 Influence of surface heating rate on austenitic grain at various heating rates in steel with 0.45% C.
size. Source: Ref 15, 27 Source: Ref 15, 27
Fig. 32 Individual phases in induction heating and spray quenching in the workpiece surface layer and corresponding temperature-
diameter diagrams. Source: Ref 15, 27
with hardness measurements on the surface or

microhardness measurements in the cross sec-
tion 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 Fig. 33 Influence of carbon content on steel hardness after
various heat treatments. Source: Ref 15, 27
stresses present in the induction surface-
hardened layer (Ref 2, 27, 46).
type and shape of the induction coil used, the
alternating-current frequency, f (Hz), the and
power density Q (W/cm2). Power density is
Time-Temperature Dependence in defined by the selected power of the high-
Induction Heating frequency generator and the surface layer of
the workpiece. Surface heating depends on the
The time variation of temperature in induction coupling between the induction coil and the
heating of a thin surface layer depends on the workpiece. Figure 34 shows the influence of
the selected power density and frequency on the and 3 kHz can be selected in the single-shot
reference depth of the skin effect in a ferro- surface-hardening technique, in which case
magnetic material (Ref 2, 20). appropriate power densities between 2 and
A higher power density results in a greater 50 MW/m2 are obtained. With lower high fre-
reference depth of the skin effect and a greater quencies, such as 10 kHz, the same depth of
depth of the heated layer with the same maxi- surface-hardened layer, that is, 2.0 mm, can be
mum temperature obtained at the workpiece ensured only when the power density is changed
surface. Figure 35 shows the interdependence to 50 MW/m2. The lowest generator frequency,
between the heating parameters, that is, power 3 MW/m2, shown in Fig. 35 cannot ensure the
density and generator frequency, as a function of depth of a hardened layer smaller than 2.5 mm.
the specified depth of the hardened layer and the Immediately after tempering, an intensive in-
heating time required for single-shot techniques verse heat flow is expected as well.
of surface induction hardening (Ref 2, 20). The power density and frequency used in
The data supplied by the diagram can have a induction hardening are usually based on the
character of information only, but they make the shape and size of the machine part to be surface
selection of an optimal surface induction con- hardened; the case depth is specified also. In
dition easier. With the scanning technique of addition to the depth of hardening, the case
surface induction hardening, however, the speed pattern along the entire length of the machine
of the workpiece movement, rather than time, part is important.
assures the depth of the required hardened layer Regardless of the complexity of a workpiece
and should be defined. Generally, longer heating shape, case depth and transition that are as uni-
times are required with smaller power densities form as possible as well as regular-shaped hard-
and vice versa. For the same depth of hardened ened ends should be provided. Inadequate
layer, longer heating times with lower current pattern transitions may produce high stress
frequencies are also required. With regard to the concentrations related to the given loads in its
depth of hardened layer selected between 0.5 vicinity, so the material cannot resist dynami-
and 10.0 mm, generator frequencies of 450, 10, cally loaded parts.
Fig. 34 Reference depth of skin effect as a function of power Fig. 35 Influence of high-frequency generator on selection
density and selected generator frequency for ferro- of power density and heating time with given thick-
magnetic steel. Source: Ref 2, 20 ness of surface induction-hardened layer. Source: Ref 2, 20
Various heating conditions that are defined by such heating conditions, a maximum tempera-
the power density and frequency provide dif- ture of nearly 1000 C was attained, while
ferent pattern depths. The third influencing heating above a temperature of 800 C was
parameter is heating time. Thus, different somewhat slowed down. The data in the diagram
hardness and residual-stress profiles may be show that the time required for heating the speci-
achieved. The steel grade to be hardened, the men from the ambient temperature to 800 C is
loop shape, and the gap size between the equal to the time required for heating from the
inductor and the workpiece should be con- latter to the maximum temperature obtained at
sidered. The data found in the diagram apply the surface, that is, 1.6 s. A temperature cycle at
only to stationary hardening; therefore, with the surface takes 3.2 to 3.3 s. The temperature
scanning surface hardening, these values differences between the surface and the core in a
should be suitably corrected. For progressive given moment are the greatest during the heating
hardening, the hardening conditions should process, that is, DTmax 600 C. During the
be slightly corrected to allow for the loop quenching process, however, they can reach up
movement. to 360 C. Temperature gradient changes are
With the surface heat treatment processes, much stronger in heating than in quenching. In
studies are often conducted on the influence of material heating, there is also a great difference
the selected heating and quenching conditions in yield stress of the material, which can produce
on the depth of the hardened layer and the size of plastic deformation. Another very important
the transition zone between the hardened and finding (Ref 47) is that the theoretical model
nonhardened microstructures. One simple and is appropriate, since the results obtained
practical procedure to control surface heat were confirmed by the standard experimental
treatment is to measure the time variation of methods such as temperature measurement with
temperature from the beginning to the end of the thermocouples, diamond pyramid hardness test,
heating process and also from the beginning to and measurement of residual stresses with x-ray
the end of the quenching process. The heat diffraction. The difference between the mea-
process may be changed by changing the power sured and the calculated temperature cycles is
density and the generator frequency, whereas the very small. It occurs mainly in heating and
quenching process may be changed by the reaches up to 60 C at maximum, not taking into
selection of different quenching agents and account the losses due to eddy currents.
quenching processes. Figure 37 shows the variation of temperature
Figure 36 shows the measured and calculated from the surface toward the core with various
temperature cycles for the surface, the core, and volume power densities, Q, that is, 0.4
in a radius, r, of 7 mm in a depth of 1 mm in a 109, 1.2 109, and 2.4 109 W/m3 (Ref 48).
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. 37 Calculated variation of temperature through speci-

Fig. 36 Comparison of variations of calculated and mea- men cross section in induction heating up to
sured temperature cycles for a cylindrical specimen hardening temperature with different volume power densities.
16 mm in diameter. Source: Ref 47 Source: Ref 48
With smaller volume power densities, the maxi- selected so that the temperature of the diameter,
mum temperature differences between the sur- TA1 , was exceeded to a depth of 5.0 mm. This
face and the core also become smaller and the means that a change of microstructure and hard-
austenitizing times longer. Because of higher ness was expected even to the depth of 5.0 mm,
temperatures attained in the core, the inverse heat where only partial austenitization was obtained.
flow may even be so high as to produce a change Induction-heating conditions were chosen that
in microstructure and, consequently, a reduction subjected the surface layer to heating for up to
of hardness in the surface layer. 35 s. Time variations of temperature at the sur-
The induction heating industry has standar- face of the cylindrical specimen and in its sub-
dized power supply frequencies, and probably surface at depths of 2.0, 4.0, and 10.0 mm are
99% of the power used will be at the frequencies shown in Fig. 38 (Ref 49, 50). Time-temperature
listed in Table 2. Also included in Table 2 is the diagrams differ from the previous ones, since a
type of equipment used to a change 50 Hz (60 in distinctive deviation occurs in heating the speci-
some countries) to a higher frequency and the men material when the temperature of magnetic
conversion efficiency (Ref 23). domain, TA2 , is reached and exceeded.
Melander (Ref 49, 50) first treated single-shot The course of surface heating indicates that
surface induction hardening of low-alloy steel the transformation temperature TA2 was ob-
with 0.4% C, 0.7% Mn, and 1.1% Cr for tem- tained in 10 s. In spite of the same power den-
pering and hardening as well as surface hard- sity, further heating of the surface up to a
ening. For an analysis, a representative-sized temperature of 850 C was very slow due to the
machine part that is most often used, that is, nonmagnetic character of the steel surface layer;
a cylindrical specimen 40 mm in diameter, it took another 28 s. It is difficult to assess which
was chosen. Induction-heating conditions were models of induction heating are more suitable
than others. However, with the heating process
suggested, it is possible, after quenching, to
Table 2 Power sources, frequencies, efficiency,
ensure a homogeneous, fine austenitic micro-
and power for induction heating equipment
structure exhibiting the finest martensite with
Power (P), Efficiency (g), Frequency
Type kW % ( f ), kHz the highest possible hardness of the given steel.
Vacuum tube 5600 5060 200450 Due to the presence of alloying elements and
oscillators high cooling rates of the surface layer, in addi-
Motor generators 7.5500 7580 1, 3, 10 tion to fine martensite, up to 3% residual auste-
Frequency multipliers 1001000 9095 180 and 540
Frequency inverters 501500 8595 0.5, 1, 3, 10 nite also appears. Based on the time-temperature
Source: Ref 23
variation of heating, the depth of the hardened
layer ranges between 2.0 and 4.0 mm. From the
Fig. 38 Time-temperature cycle during single-shot surface induction heating and quenching. Source: Ref 49, 50
time-temperature variation at a depth of 10 mm, TA3 , for equilibrium heating are plotted (Ref 51).
it can be assumed that the specimen was heated Induction heating is a very fast process; there-
through the entire volume, that is, to the very fore, the temperature transformations shift to
core of the specimen. Because of a different time higher temperatures. To obtain homogeneous
variation of temperature in the fourth tempera- austenite in the surface layer, it is necessary to
ture cycle, it may be concluded that the maxi- heat the surface layer to the hatch-marked tem-
mum temperature obtained in the depth of perature range. The results shown in the four
10 mm is lower than the magnetic transforma- diagrams make it possible to draw the following
tion temperature of the given steel. Because of conclusions:
strong overheating of the cylindrical specimen
The heating curves differ strongly.
toward its center, lower temperature gradients
Only with the power of 40 kW does the tem-
occur, which result in a reduction of thermal
perature not reach the hatch-marked tem-
stresses during the heating process. According to
perature range in 30 s.
a comparatively high temperature in the core,
With all other powers higher than 60 kW,
temperature gradients between the surface and
the hatch-marked temperature range is
the core are generally lower, which produce a
reached in a shorter heating time.
decrease in axial internal stresses generated
With all powers higher than 60 kW, more
during quenching and also a decrease in axial
intensive heating of the surface layer occurs,
residual stresses. which is shown in a steeper hardness curve in
Figure 39 shows the time variations of
the hatch-marked temperature range.
through-thickness temperature in single-shot
induction heating with high-frequency generator The first question raised by an engineer would
powers of 40, 60, 100, and 180 kW and a be how to evaluate the through-thickness var-
medium-frequency current (Ref 51). Charac- iations of hardness and residual stresses with
teristic temperature transformations, TA1 and reference to the time variation of temperature.
Fig. 39 Time-temperature variations in single-shot heating at various powers. Source: Ref 51

The second question would be how to select not reached in spite of a very high power,
induction-heating parameters to obtain optimal 151 kW.
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 en- Quenching Systems for Induction
sure a cooling rate in the surface layer equal to Hardening
or a bit higher than the critical cooling rate
required. Control of the quenching process of a gear
Figure 40 shows the time and through-thick- wheel from high hardening temperatures
ness temperature variations of the specimen on ensures martensite microstructure. The cooling
heating with different powers and different rate should be high enough to prevent formation
velocities of workpiece movement, that is, of unwanted softer microstructures, such as
v1 = 140 mm/min (Fig. 40a), v2 = 220 mm/ pearlitic or bainitic ones; therefore, it is very
min (Fig. 40b), v3 = 370 mm/min (Fig. 40c), important that in the development of new sys-
and v4 = 680 mm/min (Fig. 40d), in scan hard- tems of induction hardening, quenching systems
ening (Ref 51). In the first case, with v1, the are designed properly.
hardened layer is obtained with the powers of 74 Process parameters should be accurately
and 56 kW. In the second case, with a higher controlled to ensure permanent and reproducible
velocity of movement, v2, the hatch-marked results after hardening of machine parts. It is also
temperature range is reached only in heating indispensable to define individual parameters
with the powers of 78 and 59 kW. In the third and determine their permissible deviation in
case, with v3, the hatch-marked temperature operation, presuming that the deviations have a
range is reached only in heating with the power negligible influence on the results of heat treat-
of 112 kW, whereas in the fourth case, with ment. The quenching system for induction hard-
the highest velocity, v4, this temperature range is ening is defined by eight parameters (Ref 17,
Fig. 40 Time-temperature variations in scan hardening at various scan speeds. Source: Ref 51
52): heat time/scan rate, power level, power quenching process after 0.1, 1.0, and finally 10 s
frequency, part position/rotation, quenchant of cooling (Ref 52).
concentration, quenchant flow, quenchant tem- In quenching, it is very important that the
perature, and quenching time. temperature gradients are high enough to pre-
Important simultaneous changes of one or vent plastification, that is, distortion, of the gear
more of these parameters can produce unwanted wheel. Such quenching conditions, including
effects on the workpiece, which, in extreme high-temperature gradients, result in a decrease
cases, result in an unsuitable microstructure, de-
viations 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 differ-
ent 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 com-
paratively thin hardened surface layer, the mar-
tensite microstructure can be obtained without
the application of a quenchant. In such cases,
heat sinks from the surface into the cold work-
piece 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 mea-
suring 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 var-
iation of temperature at the same measuring Fig. 41 Temperature profiles of tooth gear at given measur-
ing points during cooling process. (a) Measured. (b)
points of the gear-wheel tooth during quenching, Calculated. Source: Ref 52
taking into account selected heat-transfer coef-
ficients, 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 quen-
ching process and the distortion of gear-wheel
teeth or a tooth is very difficult.
The authors focused on an analysis of the con-
ditions 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 Fig. 42 Temperature distributions during the cooling
shows the variation of isotherms during the process. Source: Ref 52
in favorable compressive residual stresses after Spray quenching out of location after
hardening. It follows that a sufficiently high heating
cooling rate must be ensured at the tooth surface
Spray quenching is carried out immediately
as well as at the required depth of hardening. That
after heating with a short pause. It is used with
means that temperature gradients in the gear-
machine parts made of steels with good hard-
wheel tooth must be attained so that gear-wheel
enability and with a comparatively thin hard-
teeth are hardened with the minimum required
ened layer. Immersion quenching is carried out
cooling rate, ensuring the required martensite
after the inductor has been disconnected from
transformation. A sequence of images showing
the high-frequency generator. Gradual hard-
isotherms in very short time intervals from the
ening is very efficient and prevents the influence
beginning of quenching allows assessment of the
size of distortion of the tooth and residual stresses of inverse heat from the core toward the surface,
so that no reverse heat flow and no heating of the
to be expected after quenching.
already cooled surface occur. This problem is
Figure 43(a) shows the measured volume-
fraction distribution of martensite and Fig. 43(b) characteristic of large parts, where the work-
piece surfaces should be hardened selectively.
the calculated volume-fraction distribution of
In the past, straight oils and water-soluble oils
martensite when different values of the heat-
were used for quenching after surface induction
transfer coefficient, a, that is, 5000, 7500, and
heating. Straight oils and water-soluble oils
10,000 kcal/m2 Ch, are taken into account (Ref
produce mild quenching effects, which reduce
52). The main difficulty encountered in the cal-
difficulties due to distortion and/or crack initia-
culation is how to determine the heat-transfer
tion.
coefficient to obtain a description of real quen-
Straight oils require immersion quenching
ching conditions. Quenching is very intensive,
to minimize the risk of oil ignition, whereas
since it is carried out under a pressurized water-
water-soluble oils are more suitable for spray
jet; the gear wheel has been induction heated,
quenching. Other quenchants used are water,
whereas the core is cold.
polymeric water solutions of different con-
Two common quenching methods that use a
centration, water-soluble oils, saltwater, and
quenchant are spray quenching and immersion
so on.
quenching. Particularly popular spray quenching
Polymeric water solutions are inflammable
techniques that offer different possibilities are:
quenchants. They are prepared in various con-
Spray with progressive scan heating (scan centrations to be suitably adapted to various
hardening) cooling rates. For spray quenching, various flow
Spray after heating in position (single-shot rates of quenchants can be selected. They
hardening) depend on the size of inlet openings and the
Fig. 43 Martensite distribution of the hardened tooth gear. Source: Ref 52

number of outlet nozzle boreholes in one, two, or control. An appropriate machining technology
more rows at the quenching ring. The fluid flow and appropriate heat treatment processes should
rate can also be regulated by pressure. Technical be selected, and internal stresses lower than the
literature provides data on shaping of quenching yield stress should be ensured at any moment
nozzles as well as on the selection of the quen- and any location of a machine part during heat-
chant flow rate required. It is very important that ing and/or cooling. Analytical methods provide
a sufficient flow rate of the quenchant is chosen an insight into heat treatment conditions if the
to ensure sufficient heat removal from the work- time variation of internal stresses is monitored
piece surface layer during quenching. and the dependence between the cooling time
The flow rate is also controlled by the number and the specimen temperature at each point is
of boreholes for the flow and spraying of the known. With regard to the specimen tempera-
quenchant, respectively. An appropriate arrange- ture determined in this way, the specimen yield
ment of the boreholes in the ring permits a stress at each point can be determined as well.
uniform heat removal after spray and proper Consequently, with different heat treatment con-
and uniform cooling. The borehole cross sec- ditions, different values of physical quantities
tions should be only 15% of the total available can be selected. They are reflected in the chan-
ring surface and of the inductor in the single- ged conditions in the material, which make it
shot-type application, respectively. Rotation possible to study distortion of the machine part
velocities of the workpieces in surface induction during cooling and to determine the magnitude
heating are relatively high and can be selected of the residual stresses.
between 800 and 1000 rotations per minute, Figure 45 shows the time variation of axial
whereas the rotation velocity of the workpieces stresses at the surface and in individual depths,
during quenching is considerably lower, that is, that is, from 2.0 to 4.0 mm under the surface
40 to 60 rotations per minute. Very high rotation and in the middle of a cylindrical specimen
velocities of the workpieces are required when a (Ref 50).
uniform and reproducible depth of hardened In surface heating, tensile stresses occurred
layer is to be ensured, and particularly with at the surface due to thermal extension of the
relatively short heating times.
Figure 44 shows the dependence of cooling
rate changes on the momentary temperature at
the workpiece surface for four different con-
centrations of polymeric water solutions (Ref
24). The selection of an appropriate concentra-
tion 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, con-
sidering the thickness of the heated layer, should
be ensured by an appropriate quenching system.
Time Variation of Stresses and

Residual Stresses
With a certain heat treatment performed, the

required microstructural changes and an appro-
priate 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 Fig. 44 Polymer additive ratio effects on the cooling rate.
is very important that distortion be kept under Source: Ref 24
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 micro-
structure 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 surface- Fig. 46 Residual stress profile below the surface after
hardened layer are always of a compressive induction surface hardening. Source: Ref 15
nature, are relatively high, and have a good
effect on dynamically loaded components. The
existence of residual stresses in the radial as gentle as possible, lessening the effect of
direction, that is, into the depth of the hardened stress concentration in loaded components. This
layer, is very important, as is the absolute value contributes to the fact that a machine component
of residual stress on the surface and the stress is less susceptible to overloads in operation. It
profile in the transition from compressive into has been shown that residual stresses are closely
tensile stresses (Fig. 46) (Ref 15). linked to hardness variation and microstructure
In the case of induction hardening, a maxi- in the transition zone of the hardened layer, that
mum compressive residual stress in the surface is, in the narrow range between the hardened and
layer is achieved, which is very desirable for nonhardened microstructure.
dynamically loaded components. The transition Some examples of the dependence between
from compressive into tensile stresses should be microhardness and residual stresses are shown in
Fig. 47 (Ref 15). Figure 47(a) shows a very steep more difficult if the thickness of the layer is
microhardness profile in the transition zone and above 2 mm. Due to intensive cooling, the
the highest compressive stress in the surface, internal stresses in the surface layer may become
and, related to this, a very steep transition of so high that they cause failure of the component.
residual stresses into the tensile range. The This failure is very typical, because the sur-
change from compressive into residual stresses face layer separates from the core due to high
happens at the transition between the hardened radial stresses. Important features on the curve
and nonhardened areas. Figures 47(b and c) are:
show graphs of broader transition zones, which
Maximum value of compression residual
produce a change in the microhardness and also
stress in the hardened surface layer
the residual stresses (Ref 15). Thus, Fig. 47(c)
Maximum value of tension residual stress in
shows a very slight drop in hardness in the
the hardened surface layer
transition zone for the same microhardness on
Transition width of compression to tension
the surface.
of residual stress in the hardened surface
This is reflected in lower compressive stresses
layer
in the surface, accompanied by a slight change
Transition steepness of compression to ten-
of stresses in the transition zone into lower
sion of residual stress profile
tensile residual stresses. In induction surface
Layer depth with transition microstructure
hardening, the engineer should choose the kind
of heat treatment conditions that will result in The examples shown in Fig. 46 and 47 were
the microhardness and residual-stress profiles based on a presumption that the maximum
shown in the last two examples in Fig. 47. hardness and the maximum value of the residual
Investigations of residual stresses after induc- stresses obtained were independent from the
tion surface hardening have confirmed that when heating power density and the heating time
the hardened layer is 2 mm thick, the change of (Ref 15). In the case of induction surface hard-
compressive into tensile stresses happens in ening, a maximum compressive residual stress
compliance with the transition zone, that is, in the surface layer is achieved. The transition
the achieved depth of the hardened layer. When from compressive into tensile residual stresses
the thickness of the hardened layer is greater should be as gentle as possible, lessening the
than 2 mm, the transition from compressive effect of stress concentration in loaded compo-
into tensile stresses happens in the hardened nents. This contributes to the fact that a mach-
zone, that is, the martensite microstructure. This ined component is less susceptible to overloads
means that induction surface hardening is much in operation. It has been shown that residual
Fig. 47 Various residual-stress and hardness profiles below the surface. Source: Ref 15
stresses are closely linked to hardness variation obtain this kind of residual-stress profile. It
and microstructure in the transition zone of the should be emphasized that because of their
hardened layer and base material, that is, in the lower heat conductivity, alloyed steels for sur-
narrow range between the hardened and non- face hardening are very difficult to treat, and it is
hardened microstructure. difficult to ensure desirable residual stresses in
On the basis of the heating temperature cycle, the transition area. This is why special alloyed
microstructural changes in the surface layer steels are available that display a more favorable
after quenching can be predicted. Consequently, behavior in heating and quenching. Because of
the temperature cycles at the surface, that is, a these difficulties, delamination of the surface to
through-thickness heated layer, have a decisive the very core may occur. Figure 48(a) shows that
influence on the variation of residual stresses that the energy input was too weak, since only a
after quenching. The variation of the tempera- partial transformation of the steel matrix into
ture cycles in heating can be adjusted by the austenite occurred (Ref 15, 54). After quench-
selection of adequate energy input with refer- ing, this will produce only a small portion of
ence to the steel microstructure and its thermal martensitic microstructure, which is followed by
properties. a measured variation of residual stresses. The
Figure 48 shows three examples of induction second example, shown in Fig. 48(b), indicates
surface hardening, each with different energy that heating with a high-power density was very
input during heating, whereas the quenching intensive. The result of a too-high power density
process was the same in all cases (Ref 15, 54). is a distinct influence of heat conduction from
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 trans-
formation and must be as close as possible to
the critical cooling rate. The quenching process
must be carried out very carefully in the mar-
tensitic 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 Fig. 48 Residual-stress profiles after induction surface hard-
often confronted with the question of how to ening at various input energies. Source: Ref 15, 54
the surface to the subsurface, which, in turn, The progress of induction surface heating can
resulted in a higher temperature measured in the be supervised by measuring the temperature at
subsurface than at the surface after the delayed the workpiece surface using a pyrometer. The
interruption of heating at the beginning of the temperature measurement is a determination of
quenching process. Because of such a heating temperature cycles in individual depths during
regime and energy input at the lower limit of induction heating and subsequent cooling. There
the required energy, considerably lower com- are three basic and important metallurgical
pressive residual stresses were achieved at the conditions that will affect a successful induction
surface. In this case, the highest value of com- surface-hardening operation:
pressive residual stress was obtained at a depth Lower critical temperature, where metal-
of 1.0 mm. The variation of the stress was not lurgical phase transformation begins
the most adequate, since induction surface hard- Upper critical temperature, where the for-
ening can provide considerably higher com- mation of austenitic grain microstructure is
pressive residual stresses in the surface layer. complete
The anticipated variation of the residual stresses, Surface temperature, which affects the
that is, very high compressive residual stresses resultant grain size of the heat treated micro-
ranging between 800 and 1000 N/mm2, is structure
shown in Fig. 48(c) (Ref 15, 54).
A very favorable variation of residual stresses For a typical AISI 140 material, the steel
can be found next to the transition zone between reference books will list TA1 =750 C,
the quenched and unquenched surface layers, TA3 =795 C, and Tsmax =900 C. However,
which ensures favorable behavior of a machine because of the shorter heating times with
part under dynamic loads. induction heating, the actual values required for
In practice, the progress of heating and a satisfactory microstructure and process may be
quenching can be monitored by physical mod- 30 to 60 C higher than for conventional furnace
elling, in which case the temperature cycle at the heat treatment processes.
workpiece surface as well as the temperature Figure 49 shows the temperature cycles in the
cycles in the individual depths are obtained. The surface and in the individual depths during in-
variations of the temperature cycles in heating duction heating and quenching (Ref 15, 54).
and the known temperatures, that is, temperature The temperature range of the pearlite-ferrite
ranges, make it possible to determine changes transformation into austenite, which is a con-
of a pearlite-ferrite microstructure into an dition for the formation of the martensitic micro-
austenitic one. The latter makes it possible to structure after quenching, is plotted also.
predict the microstructure after surface hard- On the basis of the heating temperature cycle,
ening. microstructural changes in the surface layer after
Fig. 49 Temperature cycles during single-shot induction surface hardening, at various depths. Source: Ref 15, 54
quenching can be predicted. Consequently, the times, it is much more difficult to ensure the
temperature cycles at the surface, that is, a same depth of the hardened layer, which indi-
through-thickness heated layer, have a decisive cates a worse repeatability of the results. The
influence on the variation of residual stresses shorter heating times with the same cooling
after quenching. The variation of the tempera- times contribute to a shorter hardening cycle and
ture cycles in heating can be adjusted by the result in higher productivity. This higher pro-
selection of adequate power density with refer- ductivity, however, entails a high consumption
ence to the steel microstructure and its thermal of electric energy for heating and less favorable
properties. through-thickness microhardness and residual-
Figure 50 shows graphic representations of stress variations in the hardened layer. Figure 51
temperature cycles at the surface at various shows microhardness and residual-stress varia-
power densities in single-shot induction hard- tions with the highest power density (Q1) and
ening (Ref 15, 54). Each temperature cycle different heating times, tH1 . . . tH6 (Ref 15, 54).
includes induction heating, which is followed by The influence of heating time affects the
quenching a short while after the interruption microhardness and residual-stress variations.
of heating. Four temperature cycles of induction With a correctly chosen condition of induction
heating and quenching measured at different heating of the thin surface layer, adequate values
heating power densities using high-frequency of the maximum compressive residual stresses
current were plotted. With the same inductor are obtained at the surface. The variation of
having the same gap width between the inductor residual stresses across the hardened surface
and the workpiece surface, but with different layer, that is, in the transition zone between
power densities, different temperature cycles of the hardened and the unhardened material, is
heating were obtained. With a higher power more important. The variation should not be
density, the heating curve was steeper and the steep, which means that the gradient of residual
heating period shorter (Ref 54). stresses in the transition zone between the hard-
In all cases of heating, however, the same ened and unhardened material should be very
depth of the hardened layer was ensured. With gently sloping. The microhardness variation in
the given power density, however, it is more the transition zone, with a gradient of residual
difficult to ensure the same depth of the hard- stresses as gently sloping as possible, can be
ened layer, since the heating times are con- accomplished by selecting the given power
siderably reduced. With very short heating density and an appropriate heating time, since
Fig. 50 Temperature cycles at the surface in induction surface heating and quenching at various power densities. Source: Ref 15, 54
full use of compressive residual stresses in dyna- hardened layer. Compressive residual stresses
mically loaded parts can be made in this way. prevail in the remaining part of the nonhardened
Figure 52 shows a cylindrical workpiece or surface layer. A residual-stress profile is un-
sample with initial diameter DI and initial height favorable in the transition zone between the
HI. Its purpose is to show the process of residual hardened layer and the rest of the nonhardened
stresses after induction surface treating in the part of the workpiece.
central part of the workpiece (Ref 15, 27). With The transition from tensile to compressive
energy input to the thin surface layer, the initial residual stresses presents a serious danger for
diameter increases to the value of DA. Change of catastrophic failure. From the aforementioned
diameter is associated with thermal expansion of it can be understood that the workpiece must
the material and is due to ferrite and/or pearlite be slowly heated and also quenched with the
transformed into austenite. correct cooling rate. The actual cooling rate is
After quenching, a thin surface layer with a very important throughout the martensitic trans-
martensite microstructure is obtained that has a formation and must be as close as possible to the
different specific volume than the initial micro- critical cooling rate. The quenching process
structure. The hardened surface layer is ready to must be carried out very carefully in the mar-
receive a greater diameter (DH), which is resis- tensitic transformation temperature range to
ted by the initial microstructure. Of interest are ensure internal stresses lower than the yield
the residual stresses in the radial and axial point.
directions. Residual stresses are compressive in Figure 53 presents a crankshaft formed by hot
the hardened layer, then they change into tensile. forging, and Fig. 54 shows the manufacturing
Residual stresses in the axial direction are procedure from blank to crankshaft (Ref 15, 36).
compressive in the middle part of the hardened The procedure of forming should be carefully
layer and turn into tensile below the hardened prescribed, including the initial and final tem-
layer. In this way, very high tensile residual peratures of forging and the uniform plastic
stresses are achieved at the bottom of the deformation rate for the entire volume.
Fig. 51 Microhardness and residual-stress profiles at various heating times, tH1tH4. Source: Ref 15, 54
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 rela-
tionship between strength and toughness of the
product. The forging procedure is followed by
annealing to remove the residual stresses in-
curred 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 Fig. 53 Schematic presentation of a crankshaft with marked
main bearing locations. Source: Ref 15, 36
to stress relief as well as the effects of nonuni-
form cooling of the product, it is necessary for
the forgings to be subjected to straightening hardening was greater than the depth of the
prior to mechanical treatment. If necessary ad- damaged surface layer; therefore, machining
ditional annealing can be prescribed to remove could immediately be followed by induction
internal stresses induced by straightening. This surface hardening. After induction surface
is followed by turning and rough grinding to hardening, the size and distribution of residual
approach the final dimensions of the product. stresses contribute to toughness and fatigue
The technology of manufacturing the crankshaft strength of the material. In these tests, induction
involves careful selection of the conditions of surface hardening was followed by finish grind-
turning and subsequent rough grinding to avoid ing and nondestructive magnetic inspection of
the occurrence of internal stresses that would the surface to reveal the possible existence
remain in the material even after induction sur- of cracks on the product surface.
face hardening and grinding, thus reducing the Crankshafts were taken from production after
fatigue strength of the material. Induction hard- induction surface hardening with the heat treat-
ening may be preceded by stress annealing if the ment and machining conditions as specified in
depth of the surface hardening is smaller than the the technology sheet. The residual stresses on
depth of the damaged layer, since, in this way, it the main crankshaft bearings were measured on
is possible to change the unfavorable stress state the bearing location in the middle (sample A in
in the surface layer induced by machining. In Fig. 53), on the extreme left side (sample C) and
this case, the depth of the induction surface on the extreme right side (sample G).
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 stres-
ses on location A is very similar on both sam-
ples, 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 induc-
tion surface hardening on bearing location C is
very similar to that in the central bearing loca- Fig. 55 Residual-stress profile after induction surface hard-
ening on sample A of the mean bearing location in
tion A, except that its absolute values are slightly the middle of the crankshaft and on sample C on the extreme left
lower, and a distinct decrease in the residual side. Source: Ref 15, 20, 44
stresses can be noted as early as a depth of ap-
proximately 3 mm, reaching its minimum value
at a depth of approximately 5.0 mm.
The residual-stress distribution is just as data analysis of the measured values of the resi-
favorable as in the central location, except that dual stresses through the hardened-layer depth.
its absolute values are slightly lower. The dif- Figure 57 shows a zone of the scattered values
ference in the residual-stress distribution can of residual stresses, the calculated variation of
be related to the period of overheating in the mean values in the individual depths, and the
austenitization temperature, which resulted in a zone determined by the upper and lower
thinner layer in austenitization and also a thinner confidence limits for the hardened-layer depth,
hardened surface layer. to a depth of 5.5 mm (Ref 54).
Figure 56 shows a zone of the measured The zone between the upper and lower con-
values of residual stresses at four bearing loca- fidence limits is important and requires a very
tions of the knee shaft (Ref 54). The arithmetic definite gradient of residual stresses. Where the
mean value for characteristic depths of the har- limit, that is, the depth, should be set and where
dened layer was determined, and the highest and the gradient should be controlled depends on the
lowest values of residual stresses at the given chosen induction-heating conditions. The heat-
depth were established, respectively. ing conditions applied in the study indicated that
The upper and lower values of the residual the specified gradient of residual stresses should
stress with the individual depths of the hardened be ensured at a depth ranging between 3.0 and
layer were defined by the range of scatter of 5.0 mm.
residual-stress values. The upper and lower con- A practical application showed that the stress
fidence limits were defined using a statistical gradient between the surface and a depth of
3.0 mm was not problematic, since the values microstructural phases from the surface to the
obtained lay in this range in all cases. Difficul- center of the cylindrical specimen at the end of
ties may be encountered in induction heating cooling (Ref 20, 47).
and induction hardening, respectively, if the The initial microstructure was preserved to a
residual-stress value obtained at the surface lies radius, r, of 5.5 mm. In the radii between 5.5 and
at the lower limit of the scattered results. This 6.6 mm, the pearlitic-ferritic microstructure and
means that the measured compressive residual low-carbon martensite appeared. With the radii
stresses are at the lower limit as well, which exceeding 6.6 mm, a fine martensitic micro-
may result in a steeper gradient to the depth of structure with approximately 6.0% residual
3.0 mm. Consequently, the required residual- austenite was obtained due to intensive cooling.
stress variation cannot be achieved through the With the selected conditions of induction heat-
entire hardened layer. ing and fast cooling of the specimen, no homo-
The cylindrical specimen was surface induc- geneous martensite was formed in the surface-
tion heated to a temperature of 980 C and hardened layer. Figure 59 shows the calculated
then quenched in saltwater. Figure 58 shows and measured variations of individual com-
the calculated distribution of individual ponents 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
calculated variations of residual stresses repre- A particular problem with steels having
sent high compressive stresses at the surface, ferritic-pearlitic and pearlitic-ferritic micro-
that is, the axial component of residual stresses, structures, respectively, is that heating of short
sz, equals 803 N/mm2, and the tangential one, duration does not ensure complete homo-
sT, is 588 N/mm2. On the contrary, the tensile genization of austenite.
residual stresses were calculated after hardening Figure 60 shows the influence of inhomo-
in the core, with the preserved pearlitic-ferritic geneity of austenite on the level of residual
microstructure. Thus, the axial component of stresses, which is particularly noticeable in the
residual, stresses, sz, calculated for the core martensite zone (Ref 20, 47). The initial inho-
equalled to +370 N/mm2 and the tangential mogeneous austenitic microstructure resulted
one, sT, was +62 N/mm2. The diagram in in the appearance of a martensitic transfor-
Fig. 59 indicates that the maximum tensile mation with a small fraction of residual auste-
stresses were attained in the transition zone nite. With regard to the volume fraction of
between the hardened layer and the unhardened martensite and residual austenite, plastification
one (Ref 47). At a greater depth, very low stress of the material occurred, which produced in-
gradients occurred, and in the opposite direction, ternal stresses and the variation of residual
that is, in the thin surface layer to a depth of stresses, particularly in the thin surface-
2.5 mm, very high gradients of residual stresses hardened layer.
occurred. The variations of residual stresses Figure 61 shows the calculated distribution of
were determined experimentally by the x-ray residual stresses due to heating of the specimen
diffraction method. The large gradient changes with heating rates of 200 and 800 C/s to a
of the measured residual stresses in the temperature of 1050 C, followed by cooling
thin surface layer can also be confirmed by with a cooling rate of 1500 C/s (Ref 20, 47).
measurement. The results of the measured and With the high heating rate, 800 C/s, a very steep
calculated variation of residual stresses in the transition of residual stresses from the com-
surface-hardened layer sufficiently agree with pressive to the tensile zone was obtained, which
small local deviations. To determine the local resulted in a decrease in fatigue strength. With
deviations of the variation and of residual the considerably lower heating rate, 200 C/s,
stresses, numerous calculations were made with the variation of residual stresses in the thin
varying physical parameters of the material as surface layer was essentially more favorable than
well as different process parameters. that with the high heating rate. The variation of
Fig. 58 Calculated distribution of microstructures along

the cylinder radius at the end of cooling. Source: Fig. 59 Calculated and measured residual-stress profiles
Ref 20, 47 after induction surface hardening. Source: Ref 20, 47
the tangential and axial components of residual depth of 2.6 mm due to a small stress
stresses permits the following observations: gradient.
With the lower heating rate, the residual The radial component of residual stresses is
stresses at the surface are lower by 100 to 0 at the surface. In the subsurface, it is of the
200 N/mm2. tensile character. Thus, it equals approximately
The stress gradient for the tangential and 50 N/mm2 with the higher heating rate, 800 C/
axial components, st and sz, is very small in s, and is 400 N/mm2 with the lower rate,
the subsurface, from a depth of 0.7 to 200 C/s.
4.5 mm. Figure 62 shows simulation of the variation of
The transition from compressive to tensile residual stresses with a favorable rate of surface
residual stresses does not occur before a 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
maximum temperature obtained at the surface, Residual stresses due to surface induction
that is, 1050 C (Ref 20, 47). This was followed hardening of the gear wheels were calculated
by quenching with two cooling rates, 1500 and along the tooth surface during the heating pro-
300 C/s. cess as well the quenching process.
The variation of the tangential and axial The series of images in Fig. 63(a) show dis-
components of residual stresses permits the tributions of internal stresses during the heating
following observations: process after 20, 60, 100, and 187 s (Ref 52).
The profile variation of internal stresses indi-
With the higher cooling rate, compressive cates that:
residual stresses are obtained at the surface,
that is, the axial component of 800 N/mm2 In the initial heating phase, the internal
and the tangential component of 1050 stresses in the tooth root are of the com-
N/mm2. pressive character and reach up to 700 N/
With a low gradient, the axial and tangential mm2; with further heating, they change into
components of the stresses vary; they change tensile internal stresses ranging from +200
their sign to the tensile zone only at a depth to +300 N/mm2.
between 2.6 and 2.8 mm. They are considerably lower in the zone
With the lower cooling rate, considerably reaching from the root to the pitch circle.
lower compressive residual stresses are Up to a heating time of 60 s, they are of
obtained at the surface, that is, an axial the tensile character and reach up to 700
component of 275 N/mm2 and a tangen- N/mm2. With further heating, they gradually
tial component of 390 N/mm2. change to compressive stresses attaining
A comparison of the axial and tangential 100 N/mm2.
components of residual stresses indicates They are obviously very low and insigni-
that because of the considerably reduced ficant in the upper part of the gear-wheel
cooling rate, the latter is lower in the thin tooth, from the pitch-circle diameter to the
surface layer by a factor of 3. tip of the tooth, therefore, they are not plot-
The gradients of residual stresses with a ted. For this reason, deformations are the
higher or lower cooling rate are favorable, greatest at the tip of the gear-wheel tooth.
since a slow decrease of compressive to
tensile residual stresses results in a minor The series of images in Fig. 63 show changes
susceptibility of a machine part to fatigue of internal stresses during the heating pro-
under dynamic loads. cess (Fig. 63a) and the cooling or quenching
Fig. 62 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
process (Fig. 63b) (Ref 20, 52). It is charac- The magnitude of residual stresses was also
teristic of the quenching process that con- measured with the x-ray diffraction method and
siderably higher internal stresses occur and are strain gages. Figure 64 shows the results of stress
more important in the tooth root. The figures measurements in the pitch circle, that is, at the
show the magnitude and variation of internal middle of the tooth surface for both teeth and in
stresses after the cooling times of 0.1, 1, 10, and the tooth root where critical residual stresses
300 s, when the gear wheel is finally cool. The occur (Ref 20, 52). By means of the x-ray dif-
series of graphic representations of the internal fraction method, it was found that:
stresses at the tooth surface indicate that:
The highest residual stresses occur in the
In the initial quenching phase, at 0.1 and 1 s, root. At the surface they equal approximately
the internal stresses are of the tensile char- 540 N/mm2, then increase to 750 N/
acter along the entire tooth height. They are mm2 at a depth of 30 mm to reach 870 N/
the highest in the tooth root and gradually mm2 at a depth of 60 mm.
decrease toward the tip of the tooth. The residual stresses are a bit lower in the
Between the quenching times of 1 and 10 s, middle of the tooth surface, that is, at the
the sign of the internal stresses changes, pitch circle. At the surface, they equal ap-
becoming gradually of the compressive proximately 90 N/mm2, then gradually in-
character and ranging between 100 and crease with a greater depth to reach 400 N/
150 N/mm2 in the upper part of the tooth and mm2 at a depth of 60 mm.
up to 600 N/mm2 in the tooth root.
A general conclusion can be drawn that the
At the end of quenching, after the cooling
experimental as well as theoretical results agree
time of 300 s, the internal stresses change
very well and provide useful information,
considerably only in the tooth root and at-
especially to a technologist in the manufacture
tain up to 1500 N/mm2.
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 distri-
bution typical of induction hardening is ob-
tained, that is, compressive residual stresses in
the hardened layer with a martensite micro-
structure, 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
20 s 60 s 100 s 187 s density.
(a)
0.1 s 1s 10 s 300 s
Stress
daN/mm2
0 50
(b)
Fig. 63 Principal stress during (a) induction surface

heating process and (b) quenching process. Source: Fig. 64 Residual-stress measurements on the root below the
Ref 20, 52 tooth surface. Source: Ref 20, 52
A number of authors have done research into place the induction coil so that symmetric
the effects of power density or energy input used heating and quenching is achieved, resulting in
in heating, with the purpose of determining op- uniform thickness of the hardened layer with a
timal induction-hardening conditions that lead martensitic microstructure.
to good mechanical properties of gears. Lower Figure 66 shows induction surface hardening
power density at the same current frequency of a cylindrical rod and tube. When the gap
requires longer heating times at the same depth between the coil and the workpiece varied from
of the hardened layer and results in higher com- the maximum gap, r1, to the minimum gap, r2,
pressive residual stresses with a moderate tran- that is, r14r2, the hardened layer thickness
sition of residual stresses from the compressive varied also (Ref 15, 27). This means that the side
into the tensile region. of the workpiece with a thicker hardened layer
In induction hardening of gears, it is necessary will suffer greater changes in volume than the
to ensure the most uniform depth of the hardened side with the thin layer. A changing thickness of
layer to achieve a symmetric distribution of the hardened layer will result in distortion or
residual stresses in the gear tooth cross section. 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. Asymme-
Workpiece Distortion in Induction trical placement of the coil inside the tube (hole,
Surface Hardening bore) causes differences in the gap between the
coil and the bore surface inside the workpiece.
Dimensional changes of the workpiece are The conditions of heating are very similar to
closely related to the internal stresses occurring those mentioned earlier, and the cylindrical rod
in heating and cooling. During heat treatment, bends in the direction of the thicker hardened
when internal stresses in the workpiece are layer.
higher than the yield point, distortion of the Similar distortion happens on prismatic bod-
workpiece takes place. When the workpiece has ies where the specimen dimensions are a h l.
an axisymmetric shape, it is very important to Distortion of workpieces subsequent to induc-
tion surface hardening on rectangular cross-
sectional 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). Dis-
tortion 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 com-
parison, 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 micro-
structure in the transition zone. Figure 67(b)
shows the conditions in induction surface heat-
ing when the temperature transformation is not
Fig. 65 Residual-stress distribution in the induction surface- exceeded, TA1 . This means that on the comple-
hardened layer of the gear tooth. Source: Ref 15, 20 tion of quenching, the hardened layer has not
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
been reached (d0B = 0), and this is reflected on stresses and then by mechanical straightening
only a partial depth of the overheated layer, d00B. and finally grinding. The procedures of sub-
Despite surface quenching, volume changes in sequent treatment of machine components are
the surface layer are not affected, which means chosen mainly with respect to the size of the
that there is no distortion of the workpiece, so the bending deformation.
size of bending is f = 0. Bending distortion Figure 67(e) shows the conditions after
may be noted quite rarely, and only if very in- induction surface hardening of the specimen
tensive cooling is applied with temperature dif- on both sides. Several different procedures can
ferences between the surface and subsurface be chosen to ensure two-sided hardening. The
defined by the depth at which the temperature first procedure involves hardening of the top
in the specimen is equal to the temperature of surface to a certain depth, d0E , and then of
the environment. However, such cases can be the bottom surface to the same depth. In the
neglected, since the goal is to achieve the aus- second procedure, both top and bottom surface
tenitization and homogenization temperature, are hardened at the same time. The third proce-
which is much higher than the temperature of the dure involves hardening of the top surface
ferrite/pearlite-to-austenite transformation, TA3 . from the left to the right side and the bottom
In practice, the cooling rate is adapted to the surface from the right to the left. The most sui-
quenching medium, which ensures the forma- table procedure is the second one, where the
tion of a hardened microstructure with minimum top and bottom surfaces are hardened simulta-
internal stresses. neously to the same depth. The achieved depth
00
Figure 67(c) illustrates conditions after of heating is equal to dE , whereas the depth of
induction surface hardening achieves a very the hardened layer is equal to d0E . When equal
small depth of the hardened layer, d0C , and a heat treatment conditions are ensured on both
correspondingly small depth of the heat-affected sides of the specimen, the depth of the hardened
zone. The achieved depth of penetration (con- layer on the top surface, dEU, is equal to the
duction) of heat after induction hardening is depth of the hardened surface on the bottom
equal to d0C +d00C =dC . The volume changes in surface, dED.
the specimen are influenced only by the hard- This induction surface-hardening procedure
ened layer, that is, the volume of the hardened ensures that the bending deformation in the
layer versus the entire volume of the specimen. middle of the specimen is equal to zero ( f = 0).
If the volume changes are small, then the Figure 68 shows two crankshafts that differ in
bending of the specimen is negligibly small. For the position of the crankpin with respect to the
a given hardened layer depth, it is possible at main journal. The first example presents a
each point during the cooling process to ensure a crankshaft where the distance, a, between the
higher yield point of the material (Rp, 0:2 ) than axes of the crankpin and the main journal is
the internal stress in the axial direction of the greater than the sum of the radii of both the
specimen. When the depth of the hardened layer crankpin and the journal:
is equal to or smaller than one-twentieth of the
specimen height, dC 0 41/ h, there is no bend-
20 dGL dKL
ing deformation of the specimen, only a slight a4 +
2 2
increase in residual stresses in the axial direc-
tion, despite a minimum volume change of the In the second example, the crankshaft is
hardened surface layer. slightly different, having the distance between
Figure 67(d) illustrates the conditions after the two axes equal to the sum of the radii of the
induction surface hardening of a specimen with a crankpin and journal:
slightly greater depth of the hardened layer, dC0 ,
than on the specimen in Fig. 67(b), dC4dB. The dGL dKL
achieved depth of penetration (heat conduction) a= +
2 2
after induction heating is also slightly greater
and is dC or dC4dB. If the achieved hardened- This must be considered in the analysis of the
layer thickness is dC41/10 h, then hardening is deformation of the crankshaft after hardening
accompanied by deformation of the specimen, cylindrical parts that alternate in following one
described as bending deformation in the middle another (crankpin and main journal).
of the specimen ( f40). Surface hardening For the conditions after induction surface
is followed first by annealing to remove the hardening of the crankshaft described in the first
example, Fig. 68(a) shows induction surface and distortions are smaller in progressive hard-
hardening of the crankpin, which causes the ening than in single-shot hardening. The only
deformation of the main journal (Ref 15). Then problem is whether the size and shape of the
follows induction surface hardening of both workpiece allow the application of progressive
main journals on the left and right side. Due to hardening.
tensile residual stresses in the web between both Fujio et al. (Ref 52), in their third report on
main journals, the web sections come closer to induction hardening, focused on studies of dis-
one another, and the length of this crankshaft tortion of gear-wheel teeth and residual stresses
section shortens by 2Dl. Considering the number in gear wheels. The authors measured the outer
of crankpins or the number of main journals, the diameter and the root diameter of the gear wheel
entire length of the crankshaft decreases by Dl. across two opposite teeth and the root parts of
This value depends on the number of the main the teeth with a micrometer before and after
journals, N, and the size of contraction per one quenching. The same dimensions were also
embedment (Dl). The entire contraction of the calculated theoretically, taking into account the
crankshaft is thus DLC = N Dl. volume changes due to phase transformations,
In the second example (Fig. 68b), the crank- and serve as a basis for determination for
shaft is designed differently, with the axial dis- deviations.
tance between the crankpin and the main journal Figure 69 shows a change of the outer dia-
being equal to the sum of both radii (Ref 15). As meter of the gear-wheel tooth for gear wheels 2,
in the previous case, induction surface hardening 3, and 4 (Ref 20, 52). The gear wheel has 26
of the crankpin induces tensile stresses in the teeth, which means that the deviation was mea-
web, causing distortion of the main journal. Both sured between the first and twenty-sixth tooth,
adjacent main journals come closer to one the third and sixteenth, and so on, so that thirteen
another, but the distortion is smaller than in the measurements were made in total. The results of
first case. After induction surface hardening of the measurements are shown as points in the
both adjacent main journals, any temporary diagram. The dotted line represents the theore-
distortion that may have occurred vanishes, and tically calculated increase in diameter due to
the crankshaft has equal length, just as prior to heat treatment.
heat treatment. Research has confirmed that the Figure 70 shows a change of tooth height for
stress state in the workpiece is more favorable gear wheels 2, 3, and 4, as in the previous
Fig. 68 Crankshaft distortion after surface induction hardening of individual journal locations. Source: Ref 15
example (Ref 20, 52). The dotted lines in the unchanged for gear wheels 2 and 3 and de-
individual diagrams represent the theoretical creased by 0.1 mm maximum for gear wheel 4.
deviations of height of the individual teeth by A comparison of the outer tooth diameter and
0.18 mm and the deviations found by measure- the tooth root diameter, also taking into account
ment, which are mainly greater than the theo- the tooth height, shows that distortion of indi-
retical values. The greatest deviation can be vidual gear wheels is considerably more com-
found with gear wheel 4. It exceeds 0.2 mm. plicated and cannot be described by the selected
Figure 71 shows changes of the root diameter. measurement methods; therefore, changes of
Calculations show that after heat treatment, tooth profiles along with tooth height were
the root diameter decreases by approximately measured. Figure 72 shows tooth profile error
0.02 mm (Ref 20, 52). The measured values curves after induction surface heating and after
of the same diameter, however, remained 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 milli-
meters, 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 pro-
gressive 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. 69 Change of tip gear-wheel diameter after quenching.

Source: Ref 20, 52
Fig. 70 Change of whole depth of gear wheel after quench- Fig. 71 Change of root gear-wheel meter after quenching.
ing. Source: Ref 20, 52 Source: Ref 20, 52
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 treat-

ment 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 mea- Fig. 73 Distortion of individual tooth shape after induction
hardening caused by volume changes. Source:
surements supported by microstructural analysis Ref 15, 20, 55
are commonly used. A disadvantage of this
procedure is that, due to the method of heating
and quenching (nonuniformly overheated left volume changes result in a slightly smaller width
and right tooth flanks), slightly higher dimen- of the tooth above the pitch circle and slightly
sional deviations may be obtained than in increased tooth width below the pitch circle
simultaneous hardening of both flanks of the of the gear. These dimensional deviations are
same tooth (Ref 27). relatively small and negligible in the case of
Figure 73 shows the deviation in the dimen- gears with small diameter or small module, but
sions of the tooth after induction surface hard- may become more important in gears with larger
ening by heating with a coil that encircles the modules and greater tooth width (Ref 55, 56).
gear tooth (Ref 15, 20, 55). Surface hardening of gears is one of the most
Measurements of gear teeth and gear gaps frequent applications of induction surface hard-
after induction surface hardening show an ening. In the first part of this chapter, different
increased volume in the tooth root and thus a methods of induction heating of gears was dis-
smaller gap and increased volume at the tip of cussed, depending on gear size but also on the
the tooth (increased gear diameter). These properties expected of gears after heat treatment.
Gears are axisymmetric machine elements; surface induction hardening, the martensitic
therefore, volume changes due to the trans- transformation occurs in the whole tooth. Con-
formation into a martensitic microstructure in- sequently, the volume change of the tooth
duced by heat treatment are to be expected. is greater and so is the distortion. In dual-
Teeth themselves are symmetrical, which means frequency surface induction hardening, only
that careful application of uniform heating of the tooth contour becomes hardened. Con-
particular gear teeth and uniform quenching, sequently, the volume change of the tooth is
regardless of the method applied, should not smaller and so is the distortion, that is, it equals
cause distortion after heat treatment. In induc- only half of that obtained after single-frequency
tion surface hardening, compressive residual hardening.
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. Residual Stresses after Induction Surface
Figure 74 shows shape distortion after single- Hardening and Finish Grinding
frequency and dual-frequency induction surface
hardening (Ref 20, 57). Figure 74(a) shows a The last phase in the manufacturing of
simulation of distortion by showing the initial crankshafts is fine grinding, where achieving the
shape of the tooth half-cut prior to hardening desirable condition of the surface and the surface
(dotted line) and the modified shape after dual- layer requires:
frequency induction surface hardening (con- Suitable dimensions of the particular bearing
tinuous line). Figure 74(b) shows a similar locations with respect to the allowable
simulation of distortion but for conventional deviations
single-frequency induction surface hardening. Suitable surface roughness
In Fig. 74, the calculated actual size of dis- Grinding stresses are compressive or the
tortion is 50 mm, with an equivalent geometrical lowest tensile to maintain the favorable
scale of 2.0 mm. A comparison of both cases of stress profile obtained by induction surface
hardening from the viewpoint of tooth distortion hardening of the surface layer
clearly shows that distortion in single-frequency Smallest changes possible in the micro-
surface induction hardening is twice that structure and also smallest changes in the
obtained after dual-frequency hardening. hardness and microhardness profiles in the
This difference in tooth distortion can be heat-affected zone after grinding
attributed to the fact that in single-frequency
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 all-
inclusive consideration of the numerous influ-
ences 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 con-
ditions have been selected:
Different kinds of grinding
Different grinding conditions (gentle,
normal, abusive)
Fig. 74 Distortion after induction surface hardening with
Because of thermomechanical loads in the
(a) dual frequency and (b) single frequency. Source:
Ref 20, 57 thin surface layer during the grinding process,
very complex physical and chemical processes grinding grains is a result of mechanical and
take place on the microlevel. For an accurate thermal effects, which are reflected in reduced
description of the conditions in the contact zone cuttability. Figure 75 shows the basic forms of
for machining a given material, it is necessary to wear and damage on the grinding wheel grains,
consider the grinding method as well as the kind expressed by characteristic changes on the
of material from which the grinding wheel is grinding grains (Ref 15, 64). Due to mechanical
made, its structure, and the grinding conditions. loads on particular grinding grains, short-lived
The heat conditions in grinding are a result of the but intensive heat effects are created in the con-
contact conditions between the individual grind- tact or on the friction surfaces with the work-
ing grains and the workpiece material as well as piece. In these cases, blunting of the grinding
deformation work in chip formation in the shear grains (Fig. 75a), breaking off of the grind-
zone. The amount of generated heat is strongly ing grains (Fig. 75b), or splitting of the grinding
dependent on the chosen machining conditions grains (Fig. 75c) may happen. Requirements are
and is abducted mainly through the chip, while a often set that a worn-out grain should fall from
smaller part of the generated heat is transferred the grinding wheel at a certain moment. A worn-
through heat transfer into the thin surface layer out grain typically has an increased contact
of the workpiece. Heating of the chip does not surface, which causes the forces on the grinding
cause any particular difficulties, but heating of grain during the cutting process to be higher than
the thin surface layer of the workpiece creates the binding forces between the grains, and there-
the conditions for different mechanical and fore, the grain falls out (Fig. 75d). It is expected
thermokinetic processes, which cause micro- that the manufacturing engineer will carefully
chemical changes. Heating up the thin surface choose all the parameters of the process so that
layer of the workpiece can leave certain unde- the worn-down grains fall out. This ensures
sirable effects that change the properties of the more efficient machining and reduced thermo-
part surface layer and thus harm its operational mechanical effects in the thin surface layer of the
abilities. The generated friction heat is trans- workpiece material. To create a suitable quality
ferred through the heat-transfer phenomenon of the workpiece surface and surface layer, it
from the contact between the grinding grain and is necessary to ensure that grinding grains will
the workpiece into the grinding grain. fall out, and that the activation of new, sharp
The increased amount of heat on the grinding grinding grains is made possible.
grain or grinding wheel intensifies the wear It is very important to choose a cutting con-
processes and damage of particular grinding dition that would make the grains fall out only
grains, which follow very different and complex when their cuttability is reduced. Figure 75(e)
mechanisms and, in the final phase, affect the illustrates characteristic blunting of the grinding
serviceability and operational life of the tool. grain due to chemical reactions at high tem-
The grinding tool consists of grinding grains, peratures. Chemical reactions at high tempera-
each representing a process. Grinding grains are tures are frequently followed by filling the pores
interconnected with an appropriate binder, with overheated, highly plastic chips of the
which is defined by different degrees of porosity workpiece material (Fig. 75f). When the pores
of the grinding wheel structure. Therefore, it is
necessary to also know the force needed to break
the grinding grain from a given structure, ex-
pressed by the grinding wheel hardness. Simul-
taneous changes in volume proportions of the
grinding grains and binder can create different
structures on the grinding wheel that do not be-
have 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 pos-
sible to achieve a longer life and wear resistance
of the wheel in equal kinematic conditions of the Fig. 75 Basic forms of wear and damage on the grinding
wheel and the workpiece. The wear of the wheel grain. Source: Ref 15, 64
on the wheel surface become filled, the tem- different grinding conditions, different tem-
perature in the contact zone rises and changes perature cycles were obtained on the surface and
the progress of the temperature cycle into the in the depth of the heat-affected layer, which has
depth of the workpiece. The maximum tem- effected microstructural changes and changes
perature of the temperature cycle on the surface in the microhardness and residual stresses
rises, as does the temperature in the particular (Ref 6466).
depths of the workpiece material. This results in Figure 77 presents the temperature cycles on
greater depth of the heat-affected zone in the the specimen surface as a function of depth in
material, which may have fatal consequences in the hardened steel at given grinding conditions
terms of surface layer properties of the work- (Ref 15, 65). The temperature cycles can be
piece material. treated separately as a heating phase and then a
Generally, blunting of the grinding grains cooling phase.
makes the contact surfaces between the grains The maximum temperatures achieved at the
and the workpiece larger, which creates the surface and in the surface layer, respectively, are
conditions for increased mechanical effects in also very important (Ref 64). A distinction can
the contact zone and higher heat input, accom- be made between three characteristic cases of
panied by stronger heat effects, in the surface temperature cycles:
layer of the workpiece. A plastically deformed
The maximum temperatures at the surface
layer is created because of interaction between
and in the surface layer, respectively, are
the grinding tool and the workpiece on the place
higher than the melting temperature of the
where the chip separates from the base.
specimen material, depending on the tem-
A result is hardening of the thin surface layer
perature cycles (Fig. 78a) (Ref 65). Such
of the workpiece material and occurrence of
conditions may occur due to very sharp
internal stresses that may lead to failure of the
thin surface layer and/or deformation of the grinding conditions or the selection of an
inappropriate grinding wheel with regard to
workpiece, with the presence of residual stresses
the specimen material. The depth of the re-
at the end of machining.
Macro- and microanalysis with optical and/or melted layer is only a few micrometers and
makes a very fine ledeburite microstructure
electronic microscopes shows microcracks and/
containing fine cementite spread in residual
or other damage on the surface caused by an
austenite. The newly formed microstructure
inadequate grinding method or procedure.
has a slightly lower hardness than marten-
Microscopic assessment of the surface state and
site. The residual stresses in the thin surface
damage on the surface quite often points to
layer will be tensile, due to plastic defor-
inadequate grinding conditions. The most fre-
mation of the surface layer in grinding
quent surface damage includes hollows, mars,
caused by tensile forces in the contact zone
torn-off areas, built-up edges of the workpiece or
of the specimen material. To this should be
the tool, and so on. It is often necessary to con-
sider the generated heat effects that cause micro-
structural and/or chemical changes accompanied
by dimensional changes. The damage on the
workpiece surface should be taken very ser-
iously, since this may give rise to very detri-
mental 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 dis-
cussed steel, the depths of the remelted layer and Fig. 76 Maximum temperature drop as a function of depth in
the hardened steel during grinding with various work
the heat-affected layer can be defined. Under speeds, Vw. Source: Ref 15, 65, 66
added the tensile stresses induced by the The maximum temperature in the contact
occurrence of residual austenite. zone is lower than the temperature required
The maximum temperature in the contact for the beginning of austenitization and
zone is lower than the temperature required higher than the lower temperature that is
for the beginning of melting of the given limited by the temperature of steel temper-
material and higher than the austenitizing ing. It is equal to approximately 200 C
temperature. The lower temperature, that is, (Fig. 78c) (Ref 65). The grinding conditions
the austenitizing temperature, will shift are very mild, so that with the selection of
because of a high rate of heating of the the correct type of grinding wheel, no
specimen toward higher temperatures, as important changes are expected in the sur-
known from transformation diagrams face layer. Only martensitic tempering may
(Fig. 78b) (Ref 65). Provided the previous occur in the surface layer if it was not per-
microstructure of the surface layer was formed during the induction surface hard-
martensite-cementite-carbide, a finer mar- ening of the specimen.
tensitic microstructure with a higher carbon
The engineer should be aware that surface
content, obtained at the expense of cemen-
integrity depends on the tribological conditions
tite-carbide phases and with the possibility
in the operation of a component/assembly.
of a lower content of residual austenite, may
Therefore, adequate knowledge for the assess-
be expected in the thin surface layer after
ment of tribological conditions of a component
grinding. The modified content of the
in operation is very important for the prescrip-
cementite-carbide phase depends on the
tion of machining that would result in the desired
heating conditions, whereas the content of
surface condition and subsurface layer. In gen-
the residual austenite depends on the cooling
eral, there are two tribological systems. The first
conditions.
is the one present during machining, and the
second is acting in operation. In both tribologi-
cal systems, however, it is the properties of the
workpiece material and its condition prior to and
after machining that play a key role. Under dif-
ferent machining conditions of grinding, differ-
ent temperature cycles were obtained on the
surface and in the depth of the heat-affected
zone, which has effected microstructural chan-
ges and changes in the microhardness and resi-
dual 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
Fig. 77 Grinding temperature cycles in different depths in microstructure containing fine cementite spread
the hardened steel at given grinding conditions.
Source: Ref 15, 65 in residual austenite. The newly formed
Fig. 78 Maximum temperature drop as a function of depth in the induction surface-hardened steel at various speed, Vw. Source:
Ref 65
microstructure has a slightly lower hardness than surface. It is only in the subsurface that the
martensite. The residual stresses in the thin sur- tensile influence can be felt.
face layer will be tensile, due to plastic defor- The relative grinding stresses are low and
mation of the surface layer in grinding caused practically negligible, although they exert a
by tensile forces in the contact zone of the positive influence on the original residual
workpiece material as well as the tensile stresses stresses after induction hardening. The zone was
induced by the occurrence of residual austenite. defined by the upper limiting values character-
The relative grinding stress is obtained by istic of greater depths under normal grinding
measuring the residual stress after induction conditions and by the lower limiting values
surface hardening, followed by measuring the characteristic of gentle grinding conditions.
same spot after induction hardening and grind- Generally, care should be taken when choosing
ing, then calculating their difference. Figure 79 grinders and grinding conditions. The chosen
shows the measured absolute residual-stress grinder should ensure regeneration of the abra-
profile after induction surface hardening and the sive grains, so that the worn-out grains may fall
measured residual-stress profile after induction off and the new ones grind the surface. The
surface hardening and grinding (Ref 15). selected grinding conditions should ensure that
A relative grinding stress represents the dif- a suitable force be applied to the worn-out
ference between measured residual stress on abrasive grains so that they may fall off and new,
specimens that were induction surface hardened sharp abrasive grains become active. Figure 82
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 con-
ditions (Ref 54).
With gentle grinding conditions, compressive
relative grinding stresses will prevail at the
Fig. 80 Subsurface profile of relative grinding stress on

bearing location A. Source: Ref 15
Fig. 79 Subsurface residual-stress profile after induction

surface hardening and grinding (absolute stress) on Fig. 81 Relative grinding stress region through thin surface
bearing location A. Source: Ref 15 layer after gentle grinding conditions. Source: Ref 54
shows a zone of relative grinding stresses Grinding conditions can be chosen so that
occurring under abusive or normal grinding con- the melting temperature of the workpiece
ditions (Ref 54). The zone of grinding stresses material (gentle grinding conditions) is not
was determined from the profile of residual exceeded. Then, the favorable compressive
stresses after induction hardening and also after stresses after induction surface hardening
induction hardening and grinding (Ref 64). are lowered due to plastic deformation of the
By deducting the two residual stresses, the so- workpiece material during the process, and
called relative grinding stresses were deter- thus, relatively low tensile residual stresses
mined. The abusive and normal grinding con- are obtained. However, this will signi-
ditions produce remelting of the thin surface ficantly lower the productivity.
layer, that is, its rehardening. With the abusive Special attention should be paid to the
and normal grinding conditions, the tensile selection of the type of grinding wheel in
relative grinding stresses are always predomi- terms of grinding wheel material, binding
nant at the surface. At greater depths, there is agent, hardness, and pore density, since a
still the influence of remelting under the abusive correct selection can contribute to higher
grinding conditions, whereas rehardening oc- cutting efficiency concerning the plastic de-
curs under normal grinding conditions. formation of the workpiece material. In this
For the sake of clearness, the zone was defined way, the grinding tensile stresses are kept as
by the upper limit of residual-stress values, low as possible and make the compressive
which is characteristic of the abusive grinding residual stresses induced by induction sur-
conditions, and the lower limit, which is char- face hardening of the prevailing variety.
acteristic of the normal grinding conditions
Induction surface hardening creates a very
(Ref 64).
desirable residual stress state. Residual stresses
The results confirm as predominant the resi-
are always of a compressive nature and are
dual stresses, stresses induced by the plastic
usually present to the depth of the induction
deformation of the material, and as a lesser
surface-hardened layer.
influence, tensile stresses caused by the forma-
However, a major difficulty in induction sur-
tion of residual austenite. On the basis of the
face hardening is ensuring a very slight/slow
measurements of residual stresses after induc-
variation in microhardness and the existence of
tion surface hardening or induction surface
compressive residual stresses in transition areas
hardening and grinding, it can be concluded:
to the microhardness of the core material. By
For residual stresses after induction surface gently grinding and varying the hardness and
hardening and grinding, the conditions of existence of compressive stresses in the transi-
abusive grinding are a more favorable tion area, it is possible to diminish the notch
choice. They lower to a lesser extent the effect induced by stress concentration. Addi-
desirable compressive residual stresses after tional grinding of an induction surface-hardened
induction surface hardening. 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 favor-
able residual-stress state after induction surface
hardening (Ref 44, 59).
Hardness Profiles in the Induction

Surface-Hardened Layer
The induction surface-hardened layer was
analyzed by measuring the hardness and
microhardness and their relation to the micro-
Fig. 82 Relative grinding stress region through thin surface
layer after abusive or normal grinding conditions. structure (Ref 44). To establish some relation-
Source: Ref 54 ships between microstructural changes versus
hardness across the entire hardened layer and the Figure 83 presents the hardness profile in
microhardness in the very thin surface layer, the an induction surface-hardened layer to a depth
latter were measured by the Vickers method. of 5.0 mm and the microhardness profile to
Thus, Fig. 83 and 84 show the hardness profiles a depth of 150 mm on bearing location A
versus the induction-hardened layer depth (Ref 44).
according to Vickers at a load of 20 N or mea- The hardness measurements show that the
surements of microhardness according to Vick- quenched and tempered steel has a hardness of
ers in a very thin surface layer to a depth of approximately 220 to 260 HV2.0, and the sur-
150 mm at a load of 2 N (Ref 44). face-hardened layer has a hardness of approxi-
The hardness of the surface layer after mately 540 HV2.0. In the surface-hardened
induction surface hardening is very uniform in layer, a slight increase in hardness as a function
all the investigated main bearing locations and of depth is evident, which is conditioned by
equals approximately 520 to 550 HV2.0. The microstructural differences due to varied cool-
hardness profile highly conforms to the residual- ing rates of the surface layer. The surface is
stress profile, which is confirmed by a decrease cooled under the effects imposed by the cooling
in hardness in the transition zone. The hardness medium, yet at a greater depth, the effects of the
profile in the transition area is likewise very medium are accompanied by a more expressed
steep and points to high stress concentrations in effect of the cold mass of the core, resulting in
this location when the crankshaft is in the loaded the formation of very fine martensite and greater
state. 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
Kosel and Kosec (Ref 67) conducted an Fig. 87 shows a forging made of low-quality slab
investigation on the cracks formed at the ingot (Ref 67).
surface-hardened and ground bearing location The main purpose of the investigation was to
of a crankshaft. Crankshafts are made of high- find the cause of the weblike cracks after
quality chromium-molybdenum heat treatable induction surface hardening and the final
steel. The shafts were forged from square shafts grinding on the bearing location on the crank-
with a cross-sectional area of 120 by 140 mm. In shaft. Grinding was studied in the same way as at
the course of mechanical and heat treatments, the cut-out bearing locations. From the steel-
the workpieces were tested with various non- works slab ingots prepared for forging, speci-
destructive methods, for example, visual mens of a suitable length were cut out and turned
inspection, and, if required, magnetic particle to size.
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 steel-
works 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 cross-
sectional chemical composition of steel
Inadequate conditions of bearing-location
grinding
Fig. 86 Macrograph of etched favorable billet cross section.
After macrostructural and microstructural Source: Ref 67
examinations as well as a microchemical ana-
lysis 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 Fig. 87 Macrograph of etched unfavorable billet cross
crankshaft. Source: Ref 67 section. Source: Ref 67
The specimens prepared in this way were Figure 91 shows two temperature cracks, with
hardened and tempered at 560 C and then sur- a pore at the crack location being visible as well.
face hardened, with a maximum surface tem- An analytical model was elaborated to clearly
perature ranging between 820 and 900 C. The demonstrate the causes of crack formation at the
case depth ranged between 3 and 4 mm. hardened surface after grinding (Ref 67).
The macrograph in Fig. 88 clearly identifies The stress condition, sR, occurring at the
the bearing location and shows a considerable hardened-layer surface after heat treatment and
difference in case depth. Generally, it was found grinding was known. In several cases, the sur-
that: face-hardened layers took the shape of an
eccentric ring, which means a nonuniform thick-
There were insignificant differences in che-
ness of the hardened layer. The cracks occurred
mical compositions. at the location of the smallest layer thickness. As
There were segregations in the central part of
expected, residual stress occurs at the specimen
the crankshaft.
surface where cracks initiate when stress oc-
There were shrinkage cavities mainly on the curs due to phase transformation to martensite,
inside and also on the outside part of the
thermal stresses during hardening, stresses due
forging.
to martensitic tempering during grinding, and
There were microsegregations.
thermal stress after grinding (Ref 67).
Some of the hardened bearing locations
In the phase transformation from austenite to
showed dimensional variations with refer-
martensite, steel volume will increase. The vol-
ence to the anticipated thickness.
ume strain can be calculated using the equation:
The microhardness obtained at the surface
was within the expectations and somewhat VM 7VA
above a value of 700 HV2.0. eV =
VA
The depth of the hardened bearing locations
ranged between 2.0 and 2.5 mm. where VM is the volume of an elementary cell of
No differences in the microstructures could martensite, which can be calculated for the given
be found using a common optical micro- steel:
scope analysis or a scanning electron micro-
scope. VM =(2:861 0:013% C)2 (2:861+0:116% C)
The microhardness at the cracked cylindrical
specimens was somewhat lower, approxi- =23:7262 103 nm3
mately 600 HV2.0.
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 Fig. 89 Hardness characteristic at cross section of weblike
section of crankshaft bearing. Source: Ref 67 cracked neck as a function of depth, z. Source: Ref 67
and VA is the volume of an elementary cell of Using the proposed analytical model, the
austenite of the same steel: researchers (Ref 67) evaluated the magnitude of
stresses in steel after surface hardening and
VA =0:5(3:548+0:044% C)3=22:68242 103 nm3 grinding. They found a mutual dependence of
the hardened-layer profile and the resistance
then eV = 0.046. against material cracking during grinding.
A stress condition with such a volume strain Cracks will appear at the surface of those
would exceed the material strength. Table 3 bearings in which short overheating of steel
indicates that the smallest hardened-layer thick- occurred and the hardness in a narrow zone
ness required is obtained under the condition decreased strongly during grinding. In such
that the resulting stress shall be higher than the cases, through hardenability of a material
breaking stress, according to Hooks model, and reaches a depth of 150 mm, where tempered
no cracks occur at the workpiece surface after martensite will form. The thickness of the
grinding (zmin = 2.873 mm) (Ref 67). 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
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 hard-
ening (57 HRC). The steel is characterized by
good hardenability and is thus suitable for manu-
facturing 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 re-
quired. This steel is also suitable for surface
hardening (flame surface hardening, induction
surface hardening) and displays a very good
resistance to wear.
Fig. 90 Hardness characteristic of neck cross section without However, special attention must be paid to the
cracks as a function of depth, z. Source: Ref 67 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
Temperature Resultant
Yield
prior to stress after
stress of
Excentricity quenching hardening
of hardened used in martensite and grinding
layer Depth Hooks model RK ,
0:2, (T K ) (sR(T0)),
(e), mm (z), mm (TK), C N/mm2 N/mm2
0.594 3.156 343 1123 1306
0.709 3.0412 336 1144 1327
0.877 2.873 325 1177 1360
1.146 2.6045 307 1232 1414
Fig. 91 Thermal cracks in surface-hardened neck section. Source: Ref 67
Source: Ref 67
temperatures, and it preserves high toughness Surface roughness: In general, the smoother
even at low temperatures (Ref 44, 68, 69). the surface finish on the metal sample, the
Impact tests provide useful quantitative com- higher the fatigue strength.
parative data with relatively simple test speci- Surface condition: Since most fatigue fail-
mens and equipment. However, these tests do ures originate at the metal surface, any major
not provide property data for design purposes for change in the surface condition will affect
material sections containing cracks or defects. the fatigue strength of the metal.
Data of this type are obtained from fracture Environment: If a corrosive environment is
mechanics, in which theoretical and/or experi- present during the cyclic stress of a metal,
mental ways determine the fracture of structural the chemical attack greatly accelerates the
materials containing pre-existing cracks and rate at which fatigue cracks propagate.
defects.
The number of cycles to failure (N ) obtained During machining processes, various defects
by the S-N curve under load or a stress-control occur on the surface, such as small scratches and
condition is related to the total life of the part up grooves, and are introduced into the workpiece
to failure. Fatigue cycles throughout the life of surface. Typical failures are moving machine
the machine part consist of crack initiation and parts, such as shafts, connecting rods, and gears.
its propagation. Crack propagation refers to It is estimated that failures of machine parts in
stable crack growth up to the crack instability. machines contribute approximately 80% of
The S-N curve approach does not separate the fatigue failures. These surface detects can limit
crack initiation phase from propagation. Indus- the fatigue life. Improving the surface finish by
trial experts assume that the crack has already polishing will increase fatigue life significantly.
initiated in the machine part, and only the total One of the most effective methods of in-
number of cycles associated with the propaga- creasing fatigue life is the existence of residual
tion are of interest. Existing or initiated cracks compressive stresses in a thin surface layer.
assumed in the machine part are the result of Thus, applied surface tensile stress will be par-
dynamic load cycles induced during manu- tially reduced in magnitude by the residual
facturing, prior to its use. The size of the pre- compressive stress. The net effect is the prob-
existing crack can be assumed based on the ability of crack formation and a consequent re-
capability of suitable inspection. The expert may duction in fatigue failure.
assume an initial surface crack after testing and According to the AISI standard, the heat
makes a decision about the application of the treatable structural steel 4140 contains between
part. 0.38 and 0.45% C, 0.90 and 1.2% Cr, and 0.15
Using the available initial defects in the and 0.30% Mo. It has very high hardenability,
material, the total life of the part can be assessed contributing to high strength values in products
by an appropriate fracture mechanics method. with high mass. Molybdenum yields a desirable
Basic factors affecting the shape of the S-N fine microstructure after hot working as well as
curve are: heat treatment, contributing to a good strength-
to-toughness ratio. Due to its fine-grained
Materials selection and heat treating or cold microstructure, it also reaches a relatively high
working conditions toughness in the heat treated condition.
Various types of loading on the specimen, The strength of the steel as well as its sur-
such as tension, compression, torsion, or a face hardness and wear resistance may be in-
combination creased by heat treatment and thermochemical
Loading conditions described by medium treatment.
stress, amplitude stress, and frequency Mechanical properties of steel having a dia-
Influences on the environment carried by meter of up to 40 mm and between 40 to
temperature, corrosion, and other factors 100 mm are given in Table 4.
Some major factors that affect the strength of Tensile strength of the steel varies between
a metal include: 880 and 1080 N/mm2, and a minimum tough-
ness value, r3, equals approximately 41 J. The
Stress concentration: Fatigue strength is steel is very sensitive to notch and transition
greatly reduced by the presence of stress on machine parts subjected to fatigue loading.
raisers, such as notches, holes, keyways, or Fatigue strength of the material is lowest under
sharp changes in cross sections. torsional load, sT, and varies for the diameters
mentioned, that is, 16 to 40 mm, so that Figure 92 presents curves for different modes
sT = 285 N/mm2, and for diameters from 40 to of dynamically loaded specimens made from
100 mm, sT = 255 N/mm2. Fatigue strength various steels that had been heat treated in
under torsional load is three or four times lower assorted ways (Ref 15). From among six curves,
than the tensile strength of steel under static four represent specimens made from heat trea-
load, according to the data in the table. table steel and two for cementation steel speci-
In professional literature, data on fatigue mens. The heat treatable steel is Cr-Mo-Ni steel
strength of materials are usually presented for a with 0.37% Ni, and the specimens were heat
prescribed specimen shape and size that has treated in two different ways:
been adjusted to the testing device. Specimens
for fatigue strength are usually of cylindrical Surface hardening applied to specimens with
shape, with a smaller diameter in the middle part a smooth shape (curve 1) and specimens
and a rounded transition into the larger-diameter with a slot (curve 2)
part. The latter is then usually clamped for Quenching or tempering applied to smooth
testing. Modes of loading the specimens vary but specimens (curve 4) and specimens with a
are usually either torsion, bending, or tension/ slot (curve 6)
compression. The highest fatigue strength is
displayed by a specimen subjected to bending The cementation steel is a chromium-nickel
loads (Ref 27, 36). For other modes of loading, steel with 0.15% C, where the specimens were
the relation with bending fatigue strength, swb, smoothly shaped (curve 3) and slotted (curve 5).
is expressed empirically, that is, fatigue strength Axle shafts used in cars, trucks, and farm
in torsion, tw, is 0.58 swb, or fatigue strength in vehicles are, with few exceptions, surface hard-
tension/compression, swz, is 0.70 swb. ened by induction. Although a portion of the
hardened surface is used as a bearing in some
Table 4 Mechanical properties of heat treated axles, the primary purpose of induction hard-
structural steel 4140 ening is to put the surface under a state of com-
Diameter Tensile strength Yield point Extension Toughness pressive residual stress (Ref 45). By this means,
(D), mm (Rm), N/mm2 (Rp0.2), N/mm2 (A5), % (r3), J
the bending and torsional fatigue life of an axle
16540 9801180 769 11 41
405100 8801080 635 12 41
may be increased by as much as 200% over that
Source: Ref 15
for parts conventionally heat treated (Fig. 93)
(Ref 45).
Fig. 92 Graphs of the fatigue strength of surface-hardened and carburized specimens. Source: Ref 15
Induction-hardened axles consist of a hard include both single-shot and scanning techni-
surface, high-strength core, and tough outer case ques.
with good torsional strength and a tough, ductile The hardness achieved on the surface of the
core. Many axles also have a region in which the hardened specimens was 56 to 59 HRC and on
case depth is kept very shallow, so that the part the cemented specimens, 58 to 59 HRC. To test
can be readily straightened following heat the effects of the slot on fatigue strength, a slot
treatment. In addition to substantially improving of equal size and shape was made on all the
strength, induction hardening is also very cost- specimens, whether they were quenched or
effective. This is because most shafts are made tempered, hardened or cemented. It was made
of inexpensive, unalloyed medium-carbon steel in the middle of the cylindrical specimen to an
that is surface hardened to case depths of 2.5 to equal depth of 0.4 mm. The depth on the sur-
8 mm, depending on the cross-sectional size. As face-hardened and cementation specimens was
with crankshafts, typical hardness (after tem- 1.5 mm. The results of testing showed that there
pering) is approximately 50 HRC. Such hard, are significant differences in terms of heat
deep cases improve yield strength considerably treatment methods and that the highest fatigue
as well (Ref 45). strength was found in surface-hardened speci-
Modern transmission shafts, particularly mens.
those for cars with automatic transmissions, are A comparison of the fatigue testing results
required to have excellent bending and torsional showed that:
strength, as well as surface hardness for wear
resistance. Under well-controlled conditions, In surface-hardened specimens with a
induction-hardening processes are able to satisfy smooth cylindrical shape and with a slot, the
these needs, as shown by the data in Fig. 94, difference in the achieved fatigue strength is
which compares the fatigue resistance of minimal. This can be attributed to a very de-
through-hardened axles (Ref 45). The induction- sirable distribution and size of compressive
hardening methods employed are varied and residual stresses throughout the hardened
Fig. 93 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
layer. Since the depth of the slot reaches only oscillations (Ref 36). Figure 95(a) shows the
one-quarter of the hardened surface layer, bending stress for induction surface hardening
the size of compressive residual stresses at of adjacent flanks of two teeth with a coil
the slot is still very high, so that the weak- reaching into the tooth gap. In the process, the
ening due to the slot and stress concentration tooth flank as well as the tooth root are hardened
along the slot does not cause any essential (Ref 36). This kind of heat treatment of gears
drop in fatigue strength. from steels for induction surface hardening
In quenched and tempered specimens with provides a fatigue bending strength in the range
no earlier prestressing of the surface layer, of 320 to 490 N/mm2. Figure 95(b) shows the
the fatigue strength was considerably lower same relationship when the induction coil
than in surface-hardened specimens. The encircles an individual gear tooth. In this pro-
fatigue strength of quenched and tempered cess, the tooth flank is hardened, and the micro-
specimens with a slot was also remarkably structure and hardness in the tooth root are
lower. The results show that the difference in preserved (Ref 36). A result of this method of
fatigue strength in slotted induction surface- hardening is that the fatigue strength is drasti-
hardened specimens and slotted quenched cally lowered to values ranging from 200 to
and tempered specimens is 5 to 1. 300 N/mm2 for the entire range of steels suitable
Smooth cemented specimens displayed 25% for induction surface hardening. This is a con-
lower fatigue strength than the surface- siderable drop in fatigue strength for the material
hardened specimens of the same shape, in the tooth root (Ref 45, 70).
whereas the cemented specimens with a slot
displayed 50% lower fatigue strength than
the same surface-hardened specimens. Stress Profiles in Machine Parts in the
The question arises about what the fatigue Loaded State
strength is in those specimens where the slot
reaches deeper than the hardened layer. A Heat treatment engineers must be very careful
regular problem in these cases is crack in choosing the conditions of induction surface
occurrence and propagation of cracks start- hardening in order to benefit from the distribu-
ing from the slot. Due to the shape of the tion of residual stresses achieved in dynamically
specimen and the slot, stresses start con- loaded parts. In industrial practice, induction
centrating at these places, depending on the surface hardening should satisfy the requirement
type and size of external loads. It should not of fatigue resistance of machine components.
be forgotten that there are no compressive The main reason for this worsening of the
residual stresses along the slot, and the size properties of the machine part is attributed to
of tensile stresses along the slot plays a tensile residual stresses in the hardened layer
decisive role in crack occurrence. and undesirable hardness distribution in the
Figure 95 shows bending stress in a tooth root transition zone from the hardened into the
subjected to dynamic load versus the number of unhardened part of the subsurface. These effects
Fig. 94 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
are quite natural, and in the first phase are a of fatigue, especially if the latter is assessed only
result of very rapid local heating of the thin from the point of view of surface hardness. To
surface layer, while in the second phase this is successfully estimate the quality of the hardened
accompanied by forced quenching, which layer, one must select the optimal synergetic
ensures a critical cooling rate and the occurrence effects between the input electric energy and the
of a martensitic microstructure. Both phases in interdependence between the induction coil and
induction surface hardening can increase the risk the workpiece surface, connected with the
Fig. 95 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
occurrence of eddy currents in the workpiece be treated from the point of view of heating,
surface layer that leads to heating. Due to com- overheating, and cooling/quenching as well as
plex synergetic effects in induction heating the internal stresses created at a certain point
or hardening, it is necessary to carefully study during the treatment. During heat treatment,
each influence on the properties of the hardened internal stresses are created by the temperature
surface layer (Ref 27, 45). Fatigue strength differences and phase transformations between
in machine components that have been induction the core and the surface, which are a result of the
surface hardened is increased if the total sum volume differences between the core and the
of the load tensions and residual stresses in surface. The created volume differences
the surface layer is of a compressive nature. between the core and the surface then give rise to
To ensure the highest fatigue strength of a com- internal stresses. During the process of heating
ponent, it is necessary to provide the follow- and cooling, internal stresses may produce the
ing: following effects:
In dynamically loaded components, the When internal stresses are lower than that of
surface is prone to fatigue occurrence, so the the yield point, higher residual stresses are
surface must have the highest compressive induced by heat treatment in the workpiece,
stresses. but these would not cause distortions,
If the total sum of stresses, that is, load ten- cracks, or failure.
sion plus residual stresses on the surface, is During a certain moment in heat treatment,
always of a compressive nature, then there is internal stresses exceed the yield point,
no chance for the occurrence of cracks and which leads to distortions and lower residual
crack growth. stresses in the workpiece.
To ensure good behavior of the surface and During very detrimental conditions in heat
the hardened surface layer in the loaded treatment, internal stresses are higher than
condition, it is necessary to induce a suitable the tensile strength of the material, causing
prestressing in the surface layer. This can be the workpiece to crack and creating larger
achieved by a carefully selected heat treat- distortions and high residual stresses.
ment method that would create the highest
compressive residual stresses on the surface Numerous changes that take place in the
and a desirable profile of the latter in the hardened surface layer of the workpiece are
hardened subsurface layer. always a result of the heating and quenching
Induction surface hardening offers opportu- conditions. Therefore, it is necessary to study the
nities to ensure a considerable amount of events taking place in the workpiece directly
compressive stresses in the machine com- after the hardening temperature is reached.
ponent surface and to ensure restrained Three zones are distinguishable in a workpiece
transformation of compressive surface heated to the hardening temperature (Fig. 96a):
stresses into tensile residual stresses in the the first zone, where the outer layer is heated to
subsurface layer. the hardening temperature; the second zone,
which is heated below the hardening tempera-
The endurance of machine components sub- ture between the temperatures TA1 and TA3 for
jected to bending and torsion loads can be suc- rapid heating; and the third zone, where the
cessfully increased by ensuring sufficiently high temperature is lower than TA1 (Ref 15).
compressive residual stresses. A manufacturing Heating to the hardening temperature at a
goal is to create a sufficient amount of com- certain depth is followed by quenching. Quen-
pressive residual stresses with a favorable dis- ching results in the occurrence of compressive
tribution, since this is the only way to increase residual stresses (Fig. 96b), when the familiar
the reliability of components in operation. An transformations in the hardened layer take place.
early failure of a component in operation may The second layer does not suffer the same dis-
cause catastrophic damage on a machine and tortions as the surface layer, although the heat-
thus a loss in profit. A decisive role in the oc- ing there has been sufficient enough to improve
currence of residual stresses is played by the the properties of the material. In the second
synergetic effects between the heat treatment layer, hardening is incomplete, which, in com-
method, the type of material, and the shape of the parison with the first layer, results in lower
workpiece. For this reason, heat treatment must hardness and strength of the material.
When a machine component that has been Since fatigue of a machine component is a
surface induction hardened in this way is sub- very delicate problem connected with the total
jected to an external load (Fig. 96c), additional sum of tensile stresses in the second zone, there
tensile residual stresses can be noted in the first is a danger that the effects of fatigue are trans-
and second layers (Ref 15). ferred to the surface, due to the size, shape, and
Fig. 96 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
location of the hardened trace. When a loaded the workpiece is heated (induction hardening,
layer is exposed to tensile external stress up to flame hardening) differs greatly from the pro-
the surface, the surface becomes very sensitive cedures where heating is performed throughout
to the occurrence of cracks (Fig. 96d) (Ref 15). the entire volume (nitriding, cementation). In
This propensity to cracking is further increased nitriding and cementation, the aforementioned
if defects are present in the surface of the second layer in the subsurface does not appear at
workpiece material. all, because the direction of the heat flow is
A critical state in the second zone of the hard- opposite to the direction of the heat flow in
ened surface layer occurs on locally hardened induction and flame hardening. The resultant
workpieces, where the surface was overheated operating tensile stresses on the surface or in the
or nonuniformly heated. The conditions are surface layer can thus be considerably smaller.
quite the opposite in surface heating of work- Due to the hardness of the surface, induction and
pieces, such as in nitriding and cementation flame hardening lowers the fatigue strength of
or intensive but shallow quenching of steel machine components. Therefore, care should be
(Fig. 97) (Ref 15). In these surface heat treat- taken to diminish all detrimental effects in the
ment procedures, heating is carried out through- surface layer.
out the whole machine component, followed by A typical example of induction surface hard-
slow cooling, for example, nitriding or quench- ening is surface hardening of gears that are
ing such as in cementation. Residual stresses heated with a low heating rate and relatively low
in nitrided and cemented surface layers are current frequency. The outer hardened zone
compressive, while tensile residual stresses includes almost the entire height of the gear
occur in the central part with the refined pearlite teeth, whereas the second zone is in the tooth
microstructure. root area. A gear heat treated in this way will
The distribution of residual stresses in heat meet the wear resistance requirements expected
treatment procedures where only the surface of 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 carbon-
concentration gradient in the thin surface layer.
As this definition clearly indicates, two factors
may control carburizing. Either the carbon-
absorption reaction at the surface or the diffu-
sion of carbon in the steel will determine the
rate of carburizing. Carburizing is done at ele-
vated temperature, generally in the range of
850 to 950 C, although occasionally at tem-
Fig. 97 Stress profile in a round bar in the loaded state, where
residual stresses after carburizing or nitriding and peratures as low as 790 C and as high as
loading stresses add up. Source: Ref 15 1095 C.
With the hardened steel, that is, hardened In incomplete hardening, the volume changes at
machine part, the following was studied: the surface are greater than in the core, which
produces transformation strains. The residual-
Effective case depth
stress variation thus depends on:
Depth at which martensitic microstructure is
present Quenching/cooling conditions
Hardness variation in the transition zone Temperature difference between the surface
between martensite and matrix and the core during the quenching process
Microstructural composition in the transi- Temperature interval between the beginning
tion zone with bainite-pearlite-ferrite and end of martensitic transformation and
the cooling rate in this zone
The characteristics obtained are essentially
affected by: An image of the magnitude of internal stresses
during the quenching process can be provided by
The grade, chemical composition, and a test of quenching cylindrical specimens from a
microstructure of steel in the soft state temperature lower than the transformation tem-
The size, that is, mass, of the machine part perature. During the quenching process, tem-
Considering that certain properties described peratures are measured at the specimens, that is,
by the previously mentioned characteristics are at their surface and in their core. Maximum
specified for the machine part, adequate condi- thermal stresses can thus be calculated (Ref 1,
tions for quenching from the austenitizing tem- 20, 71, 72). The maximum thermal stresses
perature should be specified too. The quenching found in the cylindrical specimens with the
conditions are described with reference to the smallest diameter, 25 mm, and further specimen
mass of the machine part. diameters in a geometrical ratio with a factor of 2
Thus, in steel quenching, the following are are given in Table 5.
studied: As an approximate orientation to the cooling
rates obtained, a calculation was made for a
Volume, that is, dimensional changes due to temperature of 500 C. Comparative data on
the differences in microstructure between maximum thermal stresses refer to air cooling
the initial material condition and the hard- and oil quenching. With air cooling, the stresses
ened condition of the same material obtained in the specimen with the smallest dia-
Distortion of the machine part due to incor- meter, 25 mm, equal only 7 MPa, whereas for
rect shaping of the machine part and/or an the specimen with the largest diameter, 800 mm,
improper quenching procedure the stresses are as much as 200 MPa. The maxi-
Residual stresses due to volume changes mum thermal stresses, however, are con-
occurring between the hardened zone and siderably higher with oil quenching; they range
unhardened zone of the machine part between 230 MPa with the 25 mm diameter
Residual stresses may result from a defor- specimen and 620 MPa with the 800 mm dia-
mation of the machine part during the quenching meter specimen. The latter represents an extre-
process. Internal forces due to temperature mely high internal thermal stress that may result
stresses during quenching may exceed the yield in material plastification. The data exclude
stress of the material, which results in plastic stresses due to phase changes and stresses due to
deformation during the quenching process and inhomogeneity, which additionally increase
residual stresses after cooling. The magnitude of their value.
residual stresses is related to the yield stress of Results of the measured residual stresses with
the material at the temperature at which the surface-hardened steels are given in Table 6.
deformation occurred. The stresses were measured by the x-ray dif-
Thus, in the case of through hardening in fraction technique just below the surface, that is,
which a complete and homogeneous austenitic 0.05 mm. The results refer to steel grades
microstructure turns into a martensitic one, a 4% 832M13, 805A20, 805A17, 897M39, 905M39,
volume change of the machine part or a 1.3% and cold rolled steel after induction hardening,
linear increase in the machine-part size is with the longitudinal residual stresses being
obtained: measured. The values of longitudinal stresses
after carburizing to 1.0 to 1.5 mm case with
DV 0.8% surface carbon and direct quenching
100=4%
V without tempering range between 190 and
400 MPa, which depends on the steel grade 1000 MPa, is obtained in nitriding to the same
selected. With steel 832M13, testing of speci- depth. The highest residual stresses were mea-
mens after quenching and by undercooling to sured in cold-rolled steel subjected to additional
80 and 90 C with tempering and without induction surface hardening without tempering.
tempering was performed. Tempering con- A maximum value obtained was 1000 MPa.
tributes to obtaining tempered martensite and With additional tempering at 200 C, martensite
possible transformation of retained austenite, will become tempered and the residual stresses
which produces a reduction, that is, a compen- reduced to 650 MPa. By additional tempering at
sation, of the existing residual-stress profile. an even higher temperature, at which martensite
Thus, the lowest residual stress is obtained in disintegrates, the maximum measured stresses
steel quenching by undercooling and subsequent decrease. For example, at a tempering tem-
tempering. In nitriding of steel 897M39 to a perature of 400 C, they decrease to 170 MPa or
depth of 0.5 mm, however, a residual stress a factor of 6 lower than those obtained imme-
ranging between 400 and 600 MPa is obtained. diately after steel quenching.
In steel 905M39, however, a considerably Figure 98 shows the results of carbonitriding
higher residual stress, ranging between 800 and SAE 1118 steel with reference to the contents of
carbon and nitrogen at the surface and the gra-
dient of the two in the subsurface (Ref 71). The
Table 5 Cooling rate thermal stresses in simple austenite content, that is, the austenite gradient
rounds with no transformation
of the two in the subsurface, and the way a
Air cooled Oil quenched
complementary residual-stress variation pro-
Diameter Cooling rate Maximum Cooling Maximum
of round at 500 C thermal rate at 500 C
ceeds in the subsurface are very important.
TH stress thermal
TH stress
(D), mm (v500), C/s smax , MPa (v500), C/s smax , MPa With a gradual reduction of austenite content
25 0.662 7 20.0 230 to a depth of 0.5 mm, the residual stresses
50 0.312 15 6.12 290 decrease as well, and the maximum pressure
100 0.146 28 1.88 370
200 0.070 54 0.59 450 achieved is 200 MPa. To a depth of 45 mm,
300 0.0445 73 0.29 510 the stresses gradually increase to tensile values,
400 0.0326 100 ... (540)
800 0.0158 200 ... (620)
and at a depth of 1.5 mm, they are as much as
Source: Ref 1, 71
+100 MPa.
Input and Output Control of Steel for

Table 6 Residual stresses measured in surface
heat treated steels Induction Surface Hardening of Gears
Residual stress
(longitudinal), Production automation necessitates an ever-
Steel Heat treatment MPa increasing demand for greater uniformity of the
832M13 Carburized at 970 C to 1.0 mm 280
case with 0.8% surface carbon
Direct quenched, 80 C subzero 340
treatment, no temper
Direct quenched, 90 C subzero 200
treatment, tempered
805A20 Carburized and quenched 240340(a)
805A20 Carburized to 1.11.5 mm case 190230
805A17 at 920 C, direct oil quench, 400
no temper
805A17 Carburized to 1.11.5 mm case 150200
at 920 C, direct oil quench,
tempered at 150 C
897M39 Nitrided to case depth of 400600
905M39 approximately 0.5 mm 8001000
Cold Induction hardened, untempered 1000
rolled Induction hardened, tempered 650
steel at 200 C
Induction hardened, tempered at 350
300 C
Induction hardened, tempered at 170
400 C Fig. 98 Residual stress, carbon, nitrogen, and retained aus-
(a) Immediately subsurface, that is, 0.05 mm. Source: Ref 71 tenite through a carbonitrided case on SAE 1118
steel. Source: Ref 71
properties of heat treated steel parts (Ref 73). from the tooth top and to a depth of 9/6 of the
Therefore, a general market demand is for a tooth height. The experiments on heat treatable
certain prescribed quality that is usually the carbon steel C35 with a different history have
result of various properties (Ref 41, 7482). One
of the most important properties is hard-
enability, which should be a uniform and con-
stant 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 pre-
scribed for a worm gear tooth. The success of the
induction hardening was verified by the micro-
hardness measurement along the tooth profile
centerline (Ref 83, 84).
Figure 99 shows Jominy curves for the hard-
ness profiles of the specimen front of AISI 1050,
4150, and 4340 steels (Ref 40, 41). It generally
applies that Jominy curves for common test-
specimen 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 auste-
nitizing temperature, the through-height hard-
ness profile of the specimen increases so that at a
temperature of 1040 C, there is negligible dif-
ference in the hardness profile of the Jominy
specimen, regardless whether it was furnace
heated or induction heated.
The prescription for the verification of the Fig. 99 Jominy curves for end-quenched bars of (a) AISI
1050, (b) 4150, and (c) 4340 steels, austenitized
effects of induction hardening was carried out conventionally and by short-time induction heating. Source: Ref
for the upper- and lower-limit microhardness 40, 41
shown that with the existing heat treatment con- For the selected heat treatable carbon steel
ditions, it is not possible to achieve the micro- C35, a suitable acceptance control procedure
hardness profile required by the user (Ref 79). had to be prescribed, with special emphasis on
First, it was necessary to find a Slovenian hardenability.
substitute for the French steel CC35. This was Based on the corresponding criteria, the
done with the help of a cooperative industry. acceptance conditions should enable the classi-
Steel CC35 is a heat treatable steel that has been fication of steels into quality classes in accor-
mechanically machined and induction hardened dance with the different heat treatment
to ensure a uniform quality. prescriptions. The chemical composition of the
Steel C35, chosen as a substitute, is classified steel is presented according to International
among the high-grade, unalloyed, heat treatable Organization for Standardization (ISO) stan-
carbon steels with a maximum phosphorus and dards and that of the French according to the
sulfur content lower than 0.035%. It is intended Association Francaise de Normalisation
for parts smaller than 100 mm, subjected to (AFNOR). The chemical composition is given
lower loads, and with high requirements on by quoting the upper and lower content limit for
homogeneity of material after heat treatment each particular part. The limits of the Slovenian
and mechanical machining. 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 pro-
ducer, Ravne Ironworks.
The given confidence limits are presented by
the hatched area on the extreme right of the
figure. Since testing enables a successful selec-
tion of steel as well as optimal heat treatment
conditions, it was decided to test the upper and
lower confidence limits of a small-sized, hard-
ened 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 induc-
tion hardening. The success of the heat treatment
was verified by Vickers microhardness mea-
surements at a loading of 0.3 daN acting along
the centerline of the tooth profile. The verifica-
tion 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 measure-
ment 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
Fig. 100 Effect of (a) time at an 870 C austenitizing tem- ordering the gear, the customer specified
perature and (b) maximum surface temperature
on the Jominy curves for induction-hardened AISI 4150 steel.
the prescribed allowable hardness for forming
The curve for conventional furnace-heated 4150 is also shown the upper and lower confidence limits along the
in (b). Source: Ref 40, 41 height of the gear tooth.
The microhardness limit values were: French steel CC35 displays a very favorable
hardenability according to the Jominy tests.
On the tooth top at h = 0, the microhardness
Even to a depth of 7 mm, the hardenability
is 680 to 800 HV0.3.
curve shows 40 HRC, then hardness begins to
In the depth of 4/6 h or 1.45 mm, the micro-
decrease rapidly and, at a depth of 10 mm, is
hardness is 520 HV0.3 maximum. only 22 HRC.
In the depth of 8/6 h or 2.90 mm, the micro- The Slovenian steel C35 produced by the
hardness is 520 HV0.3 maximum. Ravne Ironworks has a slightly lower con-
In the depth of 9/6 h or 3.30 mm, the micro- tent of carbon and a considerably lower
hardness is 220 to 275 HV0.3. content of manganese. For example, for steel
On the basis of hardenability testing, the fol- of charge B, the carbon content is lower by
lowing conclusions can be drawn: 0.06% and the manganese content by 0.35%,
according to the French steel. On the basis of
Due to a higher content of carbon, manga- the data in the literature, it was determined
nese, and some traces of chromium, the 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
content reduces the effect of hardenability product, which was also limited by the upper and
by a factor of 1.5. The relative effect of lower confidence limit. The answer to which
silicon on hardenability is practically negli- charges of the steel are more suitable for use can
gible, although the content may vary within be obtained by associating the corresponding
0.1 to 0.4%. microhardness confidence limits on the product
and also on the Jominy specimen. The desired
The results of the hardenability testing of the
product characteristics were defined on the basis
discussed steels are presented in Fig. 101 (Ref
of the following conditions of induction hard-
83). Based on the mean values and their devia-
ening:
tions, it was determined that the curve of hard-
enability or hardness shows solid progress into Generator power22.5, 24.0 and 25.5 kW
the depth in the French steel CC35 and steel C35 Heating time3.9, 4.4, and 4.9 s
of charge C and D. On the other side, the Time/pause between the end of heating and
hardenability of Slovenian steel C35 of charge the beginning of quenching0.1 s
C and D is much lower, so that at a depth of Quenching time4.0 s
4 to 5 mm, the lower confidence limit is excee-
ded. The input control of steel should be focused The results of induction hardening on the gear
mainly on the lower confidence limit of hard- are presented in Fig. 103 for the French steel
enability, because the progress of hardness has a CC35 (Ref 83), and in Fig. 104 for the Slovenian
decisive influence on the strength properties of steel C35 at the given power values of a high-
the gear teeth. It was determined that the upper frequency generator and heating times (Ref 83).
confidence limit corresponds to the maximum The diagrams show the upper and the lower
hardness achieved, and that it cannot be ex- confidence limits for microhardness along the
ceeded in any way. It was therefore decided that tooth height. The results of microhardness dis-
the place of the corrected lower confidence limit tribution along the tooth height must fall within
would be defined using Jominy hardenability the mentioned limits. However, it can be seen
tests, determining the actual progress of hard- that it is possible to reach only a lower micro-
ness. As a criterion for the correction of the hardness on the tooth that results from a lower
lower confidence limit, the microhardness along hardenability of the steel or from unsuitable heat
the tooth height (9/6 h), limited by the maximum treatment conditions. The lower confidence
and minimum microhardness, was used. This limit is defined only by the microhardness of the
helped to establish the primary criteria con- basic material at a depth of 4/6 to 9/6 h and
ditioning the microhardness distribution on the represents only a theoretical limit. The data on
Fig. 103 Hardness distribution along the tooth symmetry line after heat treatment of the French steel CC35. Source: Ref 83
the steel properties and heat treatment condi- familiar with electromagnetic phenomena and
tions provide various actual measured micro- eddy currents, and have some experience in the
hardness distributions, where the microhardness right choice of energy inputs necessary for
of the basic material is reached in the depth from heating. The energy needed for heating can be
h to 9/6 h in the French steel CC35 and from 11/12 provided by changing the generator power as
h to h in the Slovenian steel C35. From all these well as the frequency of the current. In pro-
data, it can be stated that it is possible, with gressive hardening, to achieve a suitable energy
adequate heating and quenching conditions and input, the workpiece feed rate or the rate at
adequate hardenability, to reach the desired which the coil is moved must be adjusted,
microhardness distribution in the tooth. From whereas in single-shot hardening, suitable
Fig. 103, it can be noted that in the case of the energy input is achieved by adjusting the heating
French steel CC35 only three out of nine heat time with a high-frequency current. An essential
treatment conditions fall out of the confidence advantage of induction surface hardening is that
limit at the minimum power P = 22.5 kW (Ref it is possible to achieve a sufficient repeatability
83). Quite the contrary can be found from of the hardened layer thickness on the workpiece
Fig. 104 presenting Slovenian steel C35. Here, as well as a desirable or even prescribed hard-
a desirable distribution of microhardness ened-layer profile, ensuring sufficient hardness
along the tooth height was achieved only in and favorable distribution of residual stresses in
the conditions of maximum power (Ref 83). This the hardened layer. A variety of steels and a
means that for an equal efficiency of induction- whole range of induction-hardening methods
hardened gears from the French steel CC35, a provide the possibilities for very accurate plan-
significantly shorter heating time is necessary ning of the size and distribution of residual
than with steel C35. This time is estimated to be stresses. This is of growing importance, since
for even a second shorter and represents a 25% manufacturers are frequently required to pro-
shorter heat treatment cycle, contributing to duce machine components that, among other
lower costs of manufacturing. surface properties, must possess quite specific
Induction surface hardening of machine residual-stress distribution along the depth of the
components and especially gears is a very hardened layer. It has become a proven fact that
complex process involving a whole range of high compressive stresses ensure high fatigue
possible heat treatment methods, which are all strength of machine components and reduce the
reflected in either good or bad serviceability of danger of the occurrence and growth of cracks
machine components. The heat treatment engi- on the surface of components. As far as induc-
neer must be aware of the different effects of tion surface hardening is concerned, it is also
particular design shapes of induction coils, be quite important to choose the right quenching
Fig. 104 Hardness distribution along the tooth symmetry line after heat treatment of the Slovenian steel C35. Source: Ref 83
medium and method of quenching. For this ASM Handbook, ASM International, 1991,
reason, engineers must direct their attention not p 164202
only to the method of heating and possible 3. R.E. Haimbaugh, Practical Induction Heat
overheating of the surface layer but also to the Treating, ASM International, 2001
methods of quenching and the right choice of the 4. S. Lampman, Introduction to Surface
medium for quenching. Hardening of Steels, Heat Treating, Vol 4,
With increasing surface heating power in ASM Handbook, ASM International, 1991,
the final phase, an increase in microhardness p 264265
along the tooth symmetry line is also reached, 5. T. Ruglic, Flame Hardening of Steels, Heat
especially in the depth from 0.5 to 2.5 mm, Treating, Vol 4, ASM Handbook, ASM
whereas by making the heating time longer, a International, 1991, p 268
more pronounced effect directed into the depth 6. K. Sridhar and A.S. Khanna, Laser Surface
can be achieved. In this way, an adequate heat Heat Treatment, Chapter 3, Lasers in
treatment can ensure the required surface hard- Surface Engineering, N.B. Dahotre, Ed.,
ness (52 HRC) as well as an increased strength of Surface Engineering Series, Vol 1, ASM
the gear in the tooth-root part, which is of International, 1998, p 6970
extreme importance for certain operating con- 7. J. Grum and R. Sturm, Characteristics of
ditions. Laser Surface Melt-Hardening and of
Tests have shown that it is necessary to move Optimizing the Process, Second Int. Conf.
the lower hardenability limit at least on the level on Quenching and the Control of Distortion
indicated by the scattered microhardness values (Cleveland, OH), 1996, p 193200
of the French steel CC35. The new lower con- 8. J. Grum and R. Sturm, Properties of Laser
fidence limit is named the corrected limit, Surface Melt Hardening on Cast Iron,
ensuring the prescribed microhardness dis- Proceedings of the Sixth European Con-
tribution in the subsurface at a power P = ference on Laser Treatment on Materials,
24.0 kW and P = 25.5 kW. ECLAT96 (Stuttgart, Germany), 1996,
On the basis of the tests, it can be concluded p 501508
that the criterion of hardenability can be very 9. J. Grum and P. Zerovnik, Residual Stresses
successfully applied in the input control of in Laser Heat Treatment of Plane Surface,
steels. Hardness decreases with the distance Proc. of the First Int. Conf. on Quenching
from the face of the Jominy specimen and must and Control of Distortion (Chicago, IL),
fall within the confidence limits. In some cases, 1992, p 333341
that is, with Slovenian steel C35 of charge A 10. O.A. Sandven, Laser Surface Hardening,
and B, when the hardenability exceeds the Heat Treating, Vol 4, ASM Handbook,
lower confidence limit, an adequate selection of ASM International, 1991, p 286
induction-hardening conditions (power in kilo- 11. D. Bialod, Ed., Electromagnetic Induction
watts and heating time in seconds) can ensure and Electric Conduction in Industry, Centre
the desired microhardness along the tooth sym- Francais de LElectricite, Diffusion: CFE
metry line. The results confirm that heat treat- and TEC and DOC, Paris, 1997
ment conditions can be successfully determined 12. S. Schiller, S. Panzer, and B. Furchheim,
by relatively simple experiments that also make Electron Beam Surface Hardening of Steels,
the procedure more economical. Heat Treating, Vol 4, ASM Handbook,
ASM International, 1991, p 297
13. K.E. Thelning, Chapter 6.7: Flame Hard-
ening, Steel and Its Heat Treatment, Bofors
REFERENCES Handbook, Butterworth, London and
Boston, 1975, p 451465
1. H.C. Child, Surface Hardening of Steel, 14. K.E. Thelning, Chapter 6.6.2: Steel Grades
Engineering Design Guides 37, Published for Induction Hardening, Steel and Its Heat
for the Design Council, The British Stan- Treating, Bofors Handbook, Butterworth,
dards Institution and the Council of Engi- London and Boston, 1975, p 434435
neering Institutions, Oxford University 15. J. Grum, Induction Hardening, Materials
Press, Oxford, U.K., 1980 Science and Technology Series, Vol 1,
2. P.A. Hassell and N.V. Ross, Induction Heat Faculty of Mechanical Engineering,
Treating of Steel, Heat Treating, Vol 4, Ljubljana, 2001
16. K.E. Thelning, Chapter 6.6: Induction 28. R.S. Ruffini and V. Nemkov, Induction
Hardening, Steel and Its Heat Treating, Heating Systems Improvement by Appli-
Bofors Handbook, Butterworth, London cation of Magnetic Flux Controllers, Proc.
and Boston, 1975, p 432451 of the Int. Induction Heating Seminar,
17. K.E. Thelning, Chapter 6.2: Quenching and Padna, 1998, p 133139
Tempering of Constructional Steels, Steel 29. C. Durban, D. Durand, and P. Chevre,
and Its Heat Treating, Bofors Handbook, Determination of Austenitic Transforma-
Butterworth, London and Boston, 1975, tion During Fast Heat Treatment, 17th ASM
p 319325 Heat Treating Society Conference Pro-
18. P.G. Simpson, Induction HeatingCoil and ceedings, Including the First International
System Design, McGraw-Hill, New York, Induction Heat Treating Symposium, D.L.
Toronto, London, 1960 Milan, D.A. Poteet, G.D. Pfaffmann, V.
19. V. Rudnev, D. Loveless, R. Cook, and Rudnev, A. Muehlbauer, and W.A. Albert,
M. Black, Handbook of Induction Heating, Ed., ASM International, 1998, p 671676
Marcel Dekker, Inc., New York, Basel, 30. K.Z. Shepelyakovski and F. Bezmenov,
2003 Advanced Steels for Modern Induction
20. J. Grum, Induction Hardening, Handbook of Through and Surface Hardening, 17th ASM
Residual Stress and Deformation of Steel, Heat Treating Society Conference Pro-
G.E. Totten, M.A.H. Howes, and T. Inoue, ceedings, Including the First International
Ed., ASM International, 2002, p 220247 Induction Heat Treating Symposium, D.L.
21. V.I. Rudnev, R.L. Cook, D.L. Loveless, and Milan, D.A. Poteet, G.D. Pfaffmann, V.
M.R. Black, Induction Heat Treatment, Rudnev, A. Muehlbauer, and W.A. Albert,
Basic Principles, Computation, Coil Con- Ed., ASM International, 1998, p 651654
struction and Design Considerations, 31. E. Hohne, Inductionsharten, Werk-
Chapter 11A, Steel Heat Treatment Hand- stattbucher fur Betriebsfachlente, Kon-
book, G.E. Totten and M.A.H. Howes, strukteure und Studierende, Heft 116,
Ed., Marcel Dekker, Inc., New York, 1997, Herausgeber, H. Haake, Ed., Springer-
p 765767 Verlag, Berlin, Gottingen, Heidelberg, 1955
22. V.I. Rudnev, R.L. Cook, D.L. Loveless, and (in German)
M.R. Black, Introduction Heat Treatment, 32. G. Benkowsky, Induktionserwarmung,
Modern Power Supplies, Load Matching, VEB Verlag Technik, Berlin, 1985 (in
Process Control, and Monitoring, Chapter German)
11B, Steel Heat Treatment Handbook, 33. G.D. Pfaffmann, Selective Surface Treat-
G.E. Totten and M.A.H. Howes, Ed., ment of Gears by Induction Profile Hard-
Marcel Dekker, Inc., New York, 1997, ening, Mater. Sci. Forum, Vol 102104,
p 873874 1992, p 345364
23. G. Pfaffmann, Introduction to Induction 34. B. Crique, F. Rukcksuhl, O. Longeot,
Heating, Fundamentals of Induction Heat- C. Delalean, S. Plano, B. Ottosson, G.
ing, Society of Manufacturing Engineers, Bonzano, and C. Pichard, Multi Frequency
Anaheim, CA, 1998 Induction Hardening of Gears to Replace
24. D.J. Williams, Quench Systems for In- CarburizingMethodology of Develop-
duction Hardening, Met. Heat Treat., 1995 ment, 17th ASM Heat Treating Society
25. K.E. Thelning, Chapter 6.6.3, Equipment Conference Proceedings, Including the
for Induction Hardening, Steel and Its First International Induction Heat Treating
Heat Treatment, Bofors Handbook, Butter- Symposium, D.L. Milan, D.A. Poteet, G.D.
worth, London and Boston, 1975, p 435 Pfaffmann, V. Rudnev, A. Muehlbauer, and
437 W.A. Albert, Ed., ASM International, 1998,
26. V.K. Rudnev, D.L. Loveless, M.R. Black, p 887894
and P.J. Miller, Progress in Study of 35. AGME Standard, American National Stan-
Induction Surface Hardening of Carbon dard, Gear Materials and Heat Treatment
Steels, Gray Irons and Ductile (Nodular) Manual, American Gear Manufacturers
Irons, Ind. Heat., March 1996 Association, Virginia, 1989
27. M.G. Lozinskii, Industrial Applications 36. G. Parrish and D.W. Ingham, The Sub-
of Induction Heating, Pergamon Press, merged Induction Hardening of Gears, Heat
Oxford, 1969 Treat. Met., (No. 2), 1998, p 4350
37. H. Sandinger, The Submerged Induction 48. U. Bruckner, W. Schuler, and H. Walter,
Hardening of Gears, Heat Treat. Met., Untersuchungen zum Eigenspannungszu-
Vol 21 (No. 2), 1996, p 111123 beim inductiven Randschnitharten, Eigen-
38. W. Amende, Transformation Hardening of spannungen: Entstehung-MessungBewer-
Steel and Cast Iron with High-Power tung, Band 1, E. Macherauch and V. Hank,
Lasers, Chapter 3, Industrial Applications Ed., Deutsche Gesselschaft fur Metallkunde
of Lasers, H. Koebner, Ed., John Wiley & E.V., Oberursel, 1983, p 293308 (in
Sons, Chicheste, 1984, p 7999 German)
39. J. Meijer, R.B. Kuilboer, P.K. Kirner, and 49. M. Melander, Computer Calculations of Re-
M. Rund, Laser Beam Hardening: Trans- sidual Stresses due to Induction Hardening,
ferability of Machining Parameters, Pro- Eigenspannungen Entstehung-Messung-
ceedings of the 26th International CIRP Bewertung, Band 1, E. Macherauch and V.
Seminar on Manufacturing Systems Hank Ed., Deutsche Gesselschaft fur Metal-
LANE94, Laser Assisted Net Shape Engi- lkunde E.V., Oberursel, 1983, p 309328
neering, M. Geiger and F. Vollertsen, Ed., 50. A.J. Fletcher, Thermal Stress and Strain
Erlangen Meisenbach-Verlag, Bamberg, Generation in Heat Treatment, Chapter 9.2,
p 243252 Induction Hardening, Elsevier Applied
40. J.F. Libsch, W.P. Chuang, and W.I. Science, London and New York, 1989,
Murphy, The Effect of Alloying Elements p 182187
on the Transformation Characteristics of 51. G. Seulen and H. Voss, Oberflachenhartung
Induction-Heated Steels, Trans. ASM, Vol mil Induktionserhitzung bei mittleren Fre-
42, 1950, p 121 quenzen, Stahl Eisen, Vol 63 (No. 51),
41. C.R. Brooks, The Metallurgy of Induction 1943, p 929935 (in German)
Surface Hardening, Adv. Mater. Process., 52. H. Fujio, T. Aida, Y. Masumoto, and T.
2000, p H19H23 Tsuruki, Paper 169-14, Distortion and Re-
42. V. Rudnev and R. Cook, Magnetic Flux sidual Stresses of Gears Caused by Hard-
Concentrators: Myths, Realities, and Pro- ening, Third Report, Induction Hardening of
fits, Met. Heat Treat., March/April 1995 Gears Made of 535 C Carbon Steel, Bull.
43. W. Brunst, Die Inductive Warmebe- JSME, Vol 22 (No. 169), 1979, p 10011008
handlung, Springer-Verlag, Berlin, 1957 53. G.H. Ledl, Why Quenching Is Important
44. J. Grum and D. Ferlan, Residual Internal in Induction Heating, Met. Prog., 1967,
Stresses after Induction Hardening and p 200204
Grinding, 17th ASM Heat Treating Society 54. J. Grum, How to Select Induction Surface
Conference Proceedings, Including the Hardening and Finished Grinding Condi-
First International Induction Heat Treating tions in Order to Ensure High Compressive
Symposium, D.L. Milan, D.A. Poteet, G.D. Residual Stresses on Machine Parts Surface,
Pfaffmann, V. Rudnev, A. Muehlbauer, and Mater. Sci. Forum, Vol 426432, 2003,
W.A. Albert, Ed., ASM International, 1998, p 25992604
p 629639 55. G. Parrish, D.W. Ingham, and M. Chaney,
45. S.L. Semiatin and D.E. Stutz, Induction The Submerged Induction Hardening of
Heat Treatment of Steel, 2nd printing, Gears, Part 1: Processing and Application
American Society for Metals, 1987 Aspects, Heat Treat. Met., Vol 25 (No. 1),
46. N. Stevens, Induction Hardening and Tem- 1998, p 18
pering, Heat Treating, Vol 4, Metals 56. G. Parrish, D.W. Ingham, and M. Chaney,
Handbook, 9th ed., American Society for The Submerged Induction Hardening of
Metals, 1981, p 451483 Gears, Part 2: Properties Imparted and
47. S. Denis, M. Zandona, A. Mey, and S.A. Problems Solved, Heat Treat. Met., Vol 25
Boufoussi, Calculation of Internal Stresses (No. 2), 1998, p 4350
during Surface Heat Treatment of Steels, 57. T. Inoue, H. Inoue, F. Ikuda, and T. Horino,
Proceedings of European Conference on Simulation of Dual Frequency Induction
Residual Stresses, V. Hauk, H.P. Hougardy, Hardening Process of a Gear Wheel, Proc.
E. Macherauch, and H.D. Tietz, Ed., of Third International Conf. on Quenching
(Frankfurt, Germany), 1992, DEM Infor- and Control of Distortion, G.E. Totten,
mationsgesselschaft mbH, Oberursel, 1993, B. Lisicic, and H.M. Tensi, Ed., ASM
p 10111020 International, 1999, p 243250
58. M. Field, J.F. Kahles, and J.T. Cammet, 69. V. Nemkov, Role of Computer Simulation
Review of Measuring Method for Surface in Induction Heating Technique, Proc. of
Integrity, Ann. CIRP, Vol 21 (No. 2), 1971, the Int. Induction Heating Seminar (Padua),
p 219237 1998, p 301308
59. E. Brinksmeier, A Model for the Develop- 70. P.K. Braisch, The Influence of Tempering
ment of Residual Stresses in Grinding, and Surface Conditions on the Fatigue
Advanced In-Surface Treatments, Tech- Behaviour of Surface Induction Hardened
nologyApplicationsEffects, Vol 5, A. Parts, Mater. Sci. Forum, Vol 102104,
Niku-Lari, Ed., Published in Cooperation 1992, p 319334
with the Institute for Industrial Technology 71. H.C. Child, Residual Stress in Heat-Treated
Transfer International, Pergamon Press, Components, Heat Treat. Met., Vol 4, 1981,
Oxford, 1987, p 173189 p 8994
60. R. Snoeys, M. Maris, and J. Peters, Ther- 72. H.C. Child, The Heat Treatment of Tools
mally Induced Damage in Grinding, Key and DiesA Review of Present Status and
Note Papers, Ann. CIRP, Vol 27 (No. 2), Future Trends, Tools and Dies for Industry
1978, p 571581 Book, The Metals Society, p 185, 291321
61. P.G. Althaus, Residual Stresses in Internal 73. Electromagnetic Induction and Electric
Grinding, Ind. Diamond Rev., (No. 3), 1985, Conduction in Industry, Chapter 9, Heat
p 124127 Treatments by Induction, Centre Francais de
62. M. Moris and R. Snoeys, Heat Affected IElectricite, 1997
Zone in Grinding Operations, Proceed- 74. R. Jonsson, Toleranzen bei der Warmebe-
ings of the 14th International Machine handlung von Stahlen in Harterein, TZ Met.
Tool Design and Research Conference, bearb., Vol 80 (No. 5), 1986 (in German)
F. Koenigsberger and S.A. Tobias, Ed., 75. K.E. Thelning, Chapter 4: Hardenability,
Manchester, 1973, p 569669 Steel and Its Heat Treatment, Bofors
63. J. Triemel, Grinding with Cubic Boron Handbook, Butterworths, London and
Nitride, Proceedings of the 14th Inter- Boston, 1975, p 127181
national Machine Tool Design and 76. H. Arend and W. Neuhaus, Die Hartbarkeit
Research Conference, F. Koenigsberger des Stahles, Verlag W. Girardet, Essen,
and S.A. Tobias, Ed., Manchester, 1973, 1955 (in German)
p 671676 77. D.H. Breen and G.H. Walter, Computer-
64. J. Grum, A Review of the Influence of Based System Selects Optimum Cost
Grinding Conditions on Resulting Residual SteelI, Source Book on Materials Selec-
Stresses after Induction Surface Hardening tion, Vol I, American Society for Metals,
and Grinding, J. Mater. Process. Technol., 1977, p 182185
Vol 114 (No. 3), 2001, p 212226 78. D.H. Breen, G.H. Walter, and C.J. Keith,
65. J. Grum, How to Select Induction Surface Computer-Based System Selected Opti-
Hardening and Finished Grinding Condi- mum Cost SteelsII, Source Book on
tions in Order to Ensure High Compressive Materials Selection, Vol II, American
Residual Stresses on Machine Parts Surface, Society for Metals, 1977, p 186192
Mater. Sci. Forum, 2002, p 404, 407, 623, 79. D.H. Breen, G.H. Walter, and C.J. Keith,
628 Computer-Based System Selected Opti-
66. J. Grum, Measuring and Analysis of Resi- mum Cost SteelsIII, Source Book on
dual Stresses after Induction Hardening and Materials Selection, Vol III, American
Grinding, Mater. Sci. Forum, 2002, p 347, Society for Metals, 1977, p 193197
349, 453458 80. D.H. Breen, G.H. Walter, and J.T. Sponzilli,
67. F. Kosel and L. Kosec, Internal Stresses in a Computer-Based System Selected Opti-
Surface Hardened and Ground Steel, Mech. mum Cost SteelsIV, Source Book on
Eng. J., Vol 31 (No. 910), 1985, p 225230 Materials Selection, Vol IV, American
68. H. Staudinger, Induktive Ober- Society for Metals, 1977, p 198201
flachenharung und Eigenspannungen, 81. D.H. Breen, G.H. Walter, and J.T. Sponzilli,
Untersuchungen uber Stahlauswahl und Computer-Based System Selected Opti-
geeignete Warmevorbehandlung, Hart.- mum Cost SteelsV, Source Book on
Tech. Mitt., Vol 2 (No. 21), 1966, p 111123 Materials Selection, Vol V, American
(in German) Society for Metals, 1977, p 202206
82. H. Matthes, Novel Process of Quality Con- Heat Treating Society Conference Pro-
trol during Inductive Hardening Process, ceedings, Including the First International
17th ASM Heat Treating Society Con- Induction Heat Treating Symposium,
ference Proceedings, Including the First D.L. Milan, D.A. Poteet, G.D. Pfaffmann,
International Induction Heat Treating V. Rudnev, A. Muehlbauer, and W.A.
Symposium, D.L. Milan, D.A. Poteet, G.D. Albert, Ed., ASM International, 1989,
Pfaffmann, V. Rudnev, A. Muehlbauer, and p 763768
W.A. Albert, Ed., ASM International, 1989, 84. P. Birk, Hartenscnutzmassen bei der Ein-
p 727733 stazhartung, Hart.-Tech. Mitt., Herausgeg-
83. J. Grum, Input and Output Control of Steel eben von Reibensahm, Vol 10 (No. 3),
Intended for Surface Hardening, 17th ASM 1957, p 919 (in German)
Failure Analysis of Steel Welds

J.H. Devletian, Portland State University
D. Van Dyke, MEI-Charlton, Inc.
FAILURE ANALYSIS OF STEEL WELDS butt welds are listed in Table 1 and illustrated in
may be divided into three categories. They in- Fig. 1. In this figure, discontinuities 1 through 7
clude failures due to: design deficiencies, weld- are caused by poor workmanship. These include
related defects usually found during inspection, (1) porosity, (2) inclusions, (3) lack of fusion,
and failures in field service. Failures due to (4) lack of penetration, (5) undercut, (6) under-
design deficiencies usually result in a ductile fill, and (7) overlap. Discontinuities 8, 9, and 10
overload due to overstressing the component or are associated with the rolling operation in the
exposing the component to service conditions steel mill. These include (8) laminations, (9)
well beyond expected performance require- delaminations, and (10) seams and laps in the
ments. Weld-related defects, which fall outside steel plate.
the quality-acceptance criteria specified by the The most insidious discontinuities are those
applicable welding code, contribute to pre- that cause brittle cracking due to metallurgical
mature failures upon testing as well as in service. origins, such as HAC, solidification cracking,
Failures in the field may be insidious, because an or stress-corrosion cracking. Such brittle fail-
unanticipated fracture mechanism has been ures are characterized by fracture surfaces
causing problems without the designers knowl- exhibiting intergranular, interdendritic, or cleav-
edge. In this chapter, the failures due to various age modes of fracture, which occur at stress
discontinuities in the steel weldment are em- levels well below the yield stress. Addition-
phasized. These include poor workmanship, a ally, even failures by quasi-cleavage and/or
variety of hydrogen-assisted cracking (HAC)
failures, stress-corrosion cracking, fatigue, and
solidification cracking in steel welds. Table 1 Common types of weld discontinuities
illustrated in Fig. 1
No. Discontinuity Location(a)
1 Porosity(b) W
Discontinuities in Steel Welds Cluster(1b in Fig. 1)
Piping(1d in Fig. 1)
2 Inclusions W
By far, the most common steel welding prob- Non-metallic slag(2b)
lems are associated with poor workmanship, 3 Lack of fusion(b) W
such as lack of fusion, lack of penetration, por- 4 Lack of joint penetration(b) W
5 Undercut(b) BM
osity, undercut, arc strikes, and others. The num- 6 Underfill(b) W
ber and size of such discontinuities may be cause 7 Overlap(b) W
8 Laminations(b) BM
for rejection by most codes and can act as stress- 9 Delaminations(b) BM
concentration sites to reduce weld joint strength 10 Seams and laps(b) BM
and promote cracking, such as fatigue failures. A 11 Lamellar tears BM
12 Cracks(b)
description of the wide variety of discontinuities a. Longitudinal(12a) W, HAZ, BM
that can occur in steel weld metal and base metal b. Transverse(12b) W, HAZ, BM
is provided by many welding codes, one of c. Crater(12c) W
d. Throat(12d) W
which is the American Association of State e. Toe(12e) HAZ
Highway Transportation Officials/American f. Root(12f)
g. Underbead and HAZ(12g)
Welding Society (AASHTO/AWS) D1.5 Bridge
Welding Code (Ref 1). From this code, reject- (a) W, weld; HAZ, heat-affected zone; BM, base metal. (b) See Fig. 1. Source:
Ref 1
able discontinuities that can occur in common
microvoid coalescence can occur at reduced inspection indicated extensive discontinuities in

levels of ductility. Such cracking is illustrated in the weld bead. The first pass of the weld was
discontinuities 11 and 12a through 12g in Fig. 1. sectioned, polished, and etched in 5% nital.
These types of brittle failures are especially The macrograph of the welded plate is shown
insidious, because the fracture occurs at stress in Fig. 2. Extensive rejectable porosity (in ac-
levels and ductility levels well below the design cordance with AWS D1.1 Structural Welding
parameters. Code) developed within the weld metal but was
Two examples of brittle failure in steel welds not readily visible on the surface.
are hydrogen-assisted cracking and solidifi- The welding system was inspected for possi-
cation cracking. Among the discontinuities illus- ble sources of porosity. Inspection revealed that
trated in Fig. 1, HAC can commonly take the the mating flanges were covered with thick mill
form of lamellar tears, longitudinal cracks in the scale and were neither ground clean nor pre-
heat-affected zone (HAZ) (12a), transverse heated prior to welding. The welding electrode
cracks (12b), toe cracks in the HAZ (12e), and and flux combination was F7A2-EM12K. Nor-
underbead cracks in the HAZ (12g). Solidifi- mally, this flux/wire combination is designed to
cation cracking can commonly take the form be tolerant of light rust. However, in this case,
of longitudinal cracks along the weld centerline the porous mill scale was so thick that it acted as
(12a), crater cracks (12c), and fissures and a sink for moisture, oil, and other volatile con-
microfissures in the weld metal. taminants.
Example 1: Porosity in Weld Metal. Two- Excessive porosity was caused by welding on
pass submerged arc welding was used to butt heavily rusted steel. Simply rough grinding the
weld two 19 mm (3/4 in.) thick A709-grade 250 faying surfaces to bare metal prior to welding
plates. The joint design was a simple square completely eliminated the porosity problem.
groove butt joint with 1.6 mm (1/16 in.) root If rough grinding is not possible, the welding
opening. It was designed to be welded in two engineer can choose a more active flux that is
passes, with a single pass on each side of the designed for welding rusted plate. Active fluxes
plate. After the first side was welded, the weld contain strong deoxidizers, such as titanium
bead exhibited an unusually large amount of and aluminum, but should only be used in a
reinforcement. Subsequent magnetic particle single pass.
Fig. 1 Schematic illustration of common types of discontinuities in welds. Reproduced with permission of the American Welding
Society. See also Table 1. Source: Ref 1
Failure Analysis of Steel Welds / 505
Example 2: Lack of Penetration. A36 a rejectable discontinuity along the same edge of
structural steel was butt welded by automatic the weld.
flux-cored arc welding. A double V-groove joint Visual inspection of the discontinuity in the
preparation was used for a two-pass welding bend test showed that the base metal and weld
operation on 6.4 mm (1/4 in.) thick plate. By metal were not fused at the toe. Subsequent
visual examination, the weld appeared satis- metallographic examination showed that the
factory. Full penetration of the butt joint was lack of side-wall fusion existed throughout the
required. Transverse-to-weld tensile testing length of the weld.
revealed inadequate yield strength. The tensile Clearly, the robot and fixturing were slightly
test failure occurred in the weld metal. Normally, misaligned. The robot welder directed the
previous tensile tests always failed in the unaf- welding heat too far on one side of the joint and
fected base metal. not enough on the other side. The robot was re-
Radiographic examination of the weld joint programmed to provide adequate weave to melt
and the mandatory transverse-to-weld tensile both faying surfaces sufficiently to prevent a
test showed the weld to have inadequate pene- recurrence of the problem.
tration. Inspection of the fracture surface of the
tensile specimen revealed unfused metal at the
plate midthickness. The fractured steel exhibited Fatigue of Welded Joints
a ductile slant fracture. By measuring the ap-
proximate area of fused cross section and area of Fatigue cracking is a result of repetitive
unfused cross section, it was determined that fluctuating stress causing fracture well below the
approximately 80% of the joint thickness carried yield strength of the steel. Factors required for
the load. fatigue fracture include a sufficiently high ten-
Even though the weld fracture was ductile, the sile stress, a sufficiently large variation in ap-
reduced cross-sectional area of fused metal (due plied stress, and a sufficient number of cycles of
to lack of penetration) did not provide enough applied stress. Cracks resulting from fatigue
sound metal to meet the required joint strength. account for more than half of all failures. Fatigue
Welding parameters were changed to ensure full fractures are insidious because there is no visible
penetration. plastic deformation preceding fracture. Fatigue
Example 3: Lack of Side-Wall Fusion. DH- fractures are promoted by stress concentrators
36 structural steel for shipbuilding was butt such as notches, sharp fillets, corners, holes,
welded by robotic gas metal arc welding. A threads, splines, keyways, dents, gouges, laps,
double V-groove joint preparation was used for a folds, flakes, and delaminations in plates, sheets,
two-pass welding operation. By visual exam- and forgings. Tensile residual stresses arise
ination, the weld appeared satisfactory. Dye- from punched holes, heat treatment, welding,
penetrant testing revealed a discontinuity along and so on.
one edge of the weld. Face bend testing revealed 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 inter-
rupted from time to time, for example, constant-
speed 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 initi-
ation sites, particularly in rotating shafts subject
to bending loads with stress concentrations.
Multiple initiation sites are easily recognized be-
cause 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
Fig. 2 Wormhole or piping porosity in weld metal deposited
by submerged arc welding. Plate is 19 mm thick. meet, they are connected by a ridge of deformed
metal called a ratchet mark. For example, three metal and HAZ. There was no evidence of weld
adjacent propagating fatigue cracks will have defects such as undercut or porosity present in
two ratchet marks connecting them. the section.
When viewing a fatigue fracture surface at Scanning electron microscopy showed fine
high magnification of approximately 500 to striations within the beach mark areas of the
5000 under the scanning electron microscope pipe. The SEM showed a ductile dimpling struc-
(SEM), striations are usually visible. In ductile ture within the brightly colored band of the frac-
structural materials, crack propagation is by ture surface. Energy-dispersive spectroscopy of
microvoid coalescence, and the crack path be- several areas of the fracture surface revealed
comes perpendicular to the tensile axis. The only iron, manganese, and oxygen.
correlation between striation spacing and tensile The pipe fractured due to fatigue cracks that
load has been well established, where one stria- initiated at several locations at the toe of the
tion represents one fatigue load cycle. Striation weld. The fatigue cracks propagated until there
spacing increases with increasing maximum was insufficient cross section to support the
tensile loading. Secondary cracks sometimes service loads. The morphology of the fracture
occur alongside striations, which better resolves suggests the cracking propagated under reversed
striation spacing. bending stresses.
For brittle materials, crack propagation occurs
in the cleavage or intergranular mode. Since
brittle cracks can grow unstably, the crack path Hydrogen-Assisted Cracking Theory
provides less clear correlation between striation
spacing and loading. Although striations are less In addition to discontinuities resulting from
well defined, secondary cracks are helpful in poor workmanship, a very common cause of
resolving some striations. An absence of observ-
able 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 heat-
recovery 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 dia-
meter by 4.8 mm, or 3/16 in., wall thickness) Fig. 3 Overview of pipe section. Cracking is visible on right
end of the pipe at the toe of the weld. Courtesy of MEI-
section 127 mm (5 in.) long with flanges welded Charlton, Inc.
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 cen-
ter, 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 Fig. 4 Overview of pipe end. Ratchet marks and beach marks
weld and proceeded in between the weld filler are clearly visible. Courtesy of MEI-Charlton, Inc.
failure in welded steel plates, castings, and Thus, modern structural steel plate and filler
structural components is HAC. Although HAC metals contain reduced carbon for increased
of steel weldments has been studied for over resistance to hydrogen cracking and, coinci-
a half-century, the mechanism of cracking is dently, for increased fracture toughness. Many
still uncertain. Fortunately, judicious hydrogen low-carbon steels, such as the U.S. Navys
management has permitted the crack-free HSLA-65, HSLA-80, and HSLA-100, are
welding of high-strength steels primarily by designed to be welded without preheating (Ref
preheating the weld to a temperature that must 2022).
be increased with increasing: The presence of hydrogen in welds can cause
failure at levels of strength and plastic strain well
Diffusible hydrogen content
below design yield strength and ductility. The
Yield strength current models of HAC involve pre-existing
Hardness
defect sites in the weld metal. These initiation
Carbon equivalent
sites include inclusions, minor phase particles,
Thickness microscopic cracks, and other discontinuities.
Restraint
Many theories attempt to explain the brittleness
Susceptible microstructure
caused by the presence of as little as a few parts
Cracking can take place in either the weld per million (ppm) of hydrogen in steel. The
metal or the HAZ, depending on which is more decohesion theory of Oriani (Ref 2329) states
susceptible. Generally, the peak hardness, that dissolved hydrogen migrates to regions of
strength, and carbon equivalent values of the triaxial tensile stress, and that the cohesive for-
HAZ are greater than those of the weld metal. ces between atoms in the iron lattice are reduced
So, cracking usually takes place in the HAZ. in proportion to the interstitial hydrogen con-
This is the most common form of HAC and is centration. This theory provides for the observed
called underbead cracking. Because HAC is increase in hydrogen solubility at the tip of a
dependent on time for hydrogen atoms to diffuse crack (Ref 30). Diffusible hydrogen has the
at room temperature to crack nucleation sites, it additional effect of supplying internal mechan-
is also called delayed cracking. Such cracking ical stresses that produce an apparent softening
can take place anytime, from minutes to several of the steel. Hydrogen also causes growth of
days after the weld has cooled. Several other microvoids in undeformed specimens and the
forms of hydrogen cracking are also possible initiation of microcracks in specimens deformed
and are discussed. to the critical strain (Ref 29). During cooling and
After over a half-century of research on the shrinkage of a weld, these defect sites become
deleterious effects of hydrogen in steel, the localized regions of triaxial tensile stress con-
fundamental mechanism of HAC is still uncer- centration. Atomic hydrogen diffuses inter-
tain. Preheating, however, is the most reliable stitially to these regions of the expanded lattice.
method to prevent weld cracking. Unfortu- As the concentration of hydrogen increases, the
nately, preheating is labor-intensive and costly. cohesive strength between iron atoms decreases
More recent studies have shown that preheating below the local intensified stress level, at which
temperatures can be either reduced or eliminated point cracking occurs. A definitive reason why
if the composition of the steel base metal and the cohesive strength between iron atoms
filler metal is low in crack-promoting elements. decreases in the presence of hydrogen atoms
(Ref 219). For example, an empirical carbon has not been fully explained. In 1960, Troiano
equivalent formula (CEN), developed by Yur- (Ref 31) suggested that certain sites act to con-
ioka et al. (Ref 2), clearly reflects the fact that centrate stress to promote cracking, and that
alloying elements such as carbon and boron hydrogen reduces the cohesive strength between
produce high CEN values and have a strong iron atoms.
detrimental effect on hydrogen crack resistance: In a classic 1972 paper (Ref 32), C.D. Bea-
chem presented a model suggesting that the
presence of sufficiently concentrated hydrogen
CEN= C+ A( C) fSi=24+ Mn=6+ Cu=15 dissolved in the lattice just ahead of the crack
+ Ni=20+( Cr+ Mo+ Nb+ V)=5+ 5Bg tip aids whatever deformation processes the
microstructures will allow. Intergranular, quasi-
where A(C) = 0.75+0.25 tanh{20(C0.12)}, cleavage, or microvoid coalescence fracture
and all elements are in weight percent. modes are dependent on the microstructure, the
crack-tip stress intensity, and the concentration given material. Management of hydrogen traps
of hydrogen. That is, hydrogen reduces the stress can have a profound effect on the susceptibility
intensity needed for cracking regardless of the to HAC. In theory, decreasing the amount of
mode of fracture, as shown in Fig. 5. Even the diffusible hydrogen by introducing permanent
stress intensity needed for HAC by microvoid or irreversible traps has the beneficial possibility
coalescence is reduced in the presence of hydro- of reducing the susceptibility to HAC.
gen. Beachems model unifies several theories Olson et al. (Ref 3740) have shown that
but shows how the stress-sorption and lattice em- effective hydrogen traps can include micro-
brittlement models are unnecessary. The model scopic particles of Ce2O3, TiC, Y2O3 VC, and
shows that the planar pressure effects are neces- NbC. Although these traps would decrease
sary at low stress intensities and are necessary the amount of diffusible hydrogen in the weld,
only to augment the driving force from the ap- large quantities of such particles would have a
plied loads. The basic hydrogen-steel interaction deleterious effect on ductility and fracture tough-
appears to be an easing of dislocation motion or ness. Steel can be embrittled by hydrogen if it is
dislocation generation, or both. stressed at temperatures high enough to allow
In 2003, Albert et al. (Ref 33) found that hydrogen to diffuse to potential embrittlement
hydrogen traps increased with alloying ele- sites, but also low enough that the hydrogen
ments. Measurements of diffusible hydrogen in is not depleted from embrittlement-producing
weld metal decreased in steel welds containing traps. With increasing temperature, steel is
increasing amounts of chromium and molyb- expected to exhibit mechanical behavior de-
denum alloying elements. The HAC in steel termined largely by the nature of hydrogen
is dependent on hydrogen trapping, which is transport among the populations of various
the interaction between diffusible hydrogen hydrogen traps (Ref 34).
and lattice defects. Several investigators have Increasing the tensile strain or residual tensile
examined hydrogen trapping in steel (Ref 34 stress also increases the ability of the steel to
40). These traps include point defects (vacan- retain hydrogen in body-centered cubic (bcc)
cies, solute atoms), dislocations, interfaces and iron. Normally, the solubility (s) of hydrogen in
surfaces (grain boundaries, particle-matrix inter- bcc iron is given by:
faces, cracks, external surfaces), and volume
defects (voids, second-phase particles). s=47:66 p1=2 exp (72:72 107=RT)
Gibala and DeMiglio (Ref 34) experimentally
measured binding energies (Table 2) and satur- where s is in ppm by mass, R is the gas constant,
abilities of hydrogen traps in AISI 4340 steel, T is the absolute temperature, and p is the pres-
which led them to predict probable suscept- sure in atmospheres. For example, at room
ibilities to HAC. Both primary and secondary temperature and pressure, the solid solubility of
crack paths were explained in terms of the types hydrogen in bcc iron is only 6.7 10 4 ppm.
of microstructural traps that predominate in a 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

Binding energy
Hydrogen trap type (Ed), kJ/mol
Interstitial solutes 315
Ti atom 26
Vacancy 46
Screw dislocation core 2030
Mixed dislocation core 59
Hydrogen vapor/void 29
Grain boundary 59
Free surface 7095
AlN interface 65
Fig. 5 Effect of hydrogen content on hydrogen-assisted Fe3C interface 84
cracking (HAC) for microvoid coalescence (MVC), TiC interface 96
quasi-cleavage (QC), and intergranular (IG) fracture modes. Source: Ref 34
Adapted from Beachem. Source: Ref 32
factor given by: that the coarse-grainced HAZ would always be

the location of maximum susceptibility to
c=cn =exp(Asys 1=2 Bsys ) cracking. Using the advantages of modern low-
carbon steels and consumables, Nippon Steel
(Ref 4) designed a series of commercial low-
A={2(1+n)V=3RT}{2E=p}1=2 carbon steels for line pipes that could be welded
while maintaining high strength and toughness
B={2(1+n)V=RT}{2=p} with equally low-carbon filler metals. These
pipeline steels included the X-65, X-70, and
where c is the hydrogen concentration at the X-80 grades, which contain very low carbon
crack tip, cn is the hydrogen concentration in the (50.03%), high manganese for strength and to
bulk steel, n is Poissons ratio, V is the partial control the bainite transformation, and 0.001%
molar volume of hydrogen in iron, E is the boron to suppress the proeutectoid ferrite nucle-
elastic modulus, and sys is the yield strength. ation at austenite grain boundaries. It is desirable
For a high-yield-strength steel having sys = to have a large amount of acicular ferrite in the
1000 MPa, the solubility of hydrogen in steel weld metal for optimal strength and toughness as
increases by a c/cn factor of 6.5. The tip of a well as good resistance to HAC (Ref 43, 44).
crack is an excellent site for hydrogen accumu- Microconstituents detrimental to weld metal
lation, because the strain is greater than that toughness and possibly increased susceptibility
associated with sys. to HAC include grain-boundary ferrite, mar-
Although the mechanism for HAC is still tensite, and side-plate ferrite, because these
debated, equations to prevent HAC have been structures provide a continuous path for cleav-
developed empirically. For example, the preheat age crack propagation.
parameter (Pw) and the hydrogen accumulation
parameter (PHA), were developed from experi-
mental work in Japan by Yurioka et al. (Ref 15,
41) as practical tools to empirically predict a
Types of Hydrogen-Assisted Cracking
preheating temperature to prevent HAC:
Hydrogen-assisted cracking can appear in
four common forms:
Pw =Pcm +HD =60+R=40,000
Underbead or delayed cracking
where Pcm is the Ito-Bessyo (Ref 42) carbon Weld metal fisheyes
equivalent given as: Ferrite vein cracking
Hydrogen-assisted reduced ductility
P cm =C+Si=30+(Mn+Cu+Cr)=20+Ni=60 As mentioned earlier, the mechanism of HAC
is not clear, but management of hydrogen and
+Mo=15+V=10+5B
the prevention of HAC are well established.
Preheating the weld area prior to and during
where HD is the diffusible hydrogen content welding provides the most reliable resistance to
in mL/100 g, elements are in weight percent, HAC. There are many empirically-derived
and R is the restraint intensity in MPa. The HAC methods to calculate preheat temperatures to
can be avoided if the value of Pw50.3. In the prevent HAC. All of the various types of HAC
Pw equation, the effects of alloy composition, can be avoided by good welding practice. The
hydrogen content, and restraint are taken into forms of HAC are discussed in the following
account empirically. sections.
Consumables manufacturers have recognized
that filler metals containing less carbon than the
steel plate would produce not only enhanced Underbead or Delayed Cracking
weld metal toughness at required strength levels By far, the most common form of HAC is
but also greater resistance to HAC. In fact, vir- underbead or delayed cracking, schematically
tually all of the older algorithms for determin- illustrated as discontinuity 12g in Fig. 1 and
ing preheating temperatures were based on the described in Table 1. Typically, this form of
composition of the base metal and not the weld cracking occurs in the coarse-grained HAZ up to
metal. Cracking of the HAZ by HAC was so 72 h after the weld has cooled. This is because
common in structural steels that it was assumed the HAZ typically has higher carbon content and
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 dif-
fusion 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, fol-
lowed 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 Fig. 6 Underbead cracking at the toe of the fillet weld on the
flange
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. Cross-
country 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 Fig. 7 Toe cracking on the flange side of the flange-to-pipe
fillet weld, showing the weld metal, heat-affected
was 203 mm (8 in.) outside diameter by 6.4 mm zone, and unaffected base metal. Cracking occurred in the mar-
(0.25 in.) wall thickness, and the flange was tensitic (white) heat-affected zone of the flange.
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 con-
fined 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 Fig. 8 Fracture surface of flange failure in the as-received
condition. Intergranular fracture is shown as well as
incorrect steel was used for the flange. The debris retained from the field.
correct flange steel was supposed to be a low- all-weld-metal tensile specimens that fail due to
carbon steel. Instead, the flange was a 0.8% HAC. In tensile testing, fisheyes reduce the weld
(high) carbon steel, which was very susceptible metal ductility measurements, such as percent
to HAC in the HAZ when welded without pre- elongation and percent reduction of area. Fish-
heating. eyes are local areas within the weld that are more
Cracking in the HAZ on the flange side of the hardenable due to solute banding, cellular, or
fillet weld was due to the mistaken use of a high dendritic segregation of alloying elements (Ref
0.8% C steel instead of the specified low-carbon 45). These initiation areas may also be richer in
steel. Quality control measures need to be fol- localized hydrogen due to their proximity to
lowed to prevent mixed steels from being used. hydrogen traps such as inclusions. Since these
alloy-rich segregated areas are more susceptible
to brittle HAC, small, localized brittle-fracture
Weld Metal HAC and Fisheyes zones appear visually on a tensile test fracture
Welding of modern low-carbon steels often surface as bright round spots surrounded by gray
results in HAZs with greater resistance to HAC. ductile fracture. The bright round spot may
Thus, the weld metal composition is now as consist of a local region of typically inter-
susceptible as the HAZ. If the weld metal con- granular or possibly cleavage fracture sur-
tains sufficient diffusible hydrogen content, has rounded by ductile dimpled failure. Both
high yield strength, and is in a highly stressed intergranular and cleavage failures are brittle
condition, the susceptibility of such weld metal fracture modes that appear much more brightly
to HAC is very possible. For example, in than the surrounding material, which is ductile
the line-pipe industry, new thermomechanical- dimpled and gray-appearing.
controlled processing steels achieve high yield Example 6: Fisheyes on Fracture Sur-
strength through thermal processing in the roll- face. A high-strength steel, HSLA-100, was
ing mill, so that the carbon content and carbon butt-welded with a matching-strength filler
equivalent levels for a given yield strength have metal using gas metal arc welding (GMAW) and
dropped substantially. With this reduction in argon-5%CO2 shielding gas at a heat input of
carbon equivalent, the susceptibility to HAZ 1.1 kJ/mm (28 kJ/in.) without preheating. The
cracking has also declined significantly. Since filler metal contained Fe-0.03%C-1.4%Mn-
the as-deposited weld metal achieves strength 3%Ni-0.7%Mo. Because of the very low carbon
primarily through alloying, the weld metal is content, the weld metal hardness did not exceed
now very susceptible to HAC. Often, field 24 HRC. Tensile test results showed inadequate
welding of X-65 and X-70 line pipe is performed ductility of only 8% elongation.
with high-hydrogen E8010G cellulosic elec- Upon examining the fracture surface of the
trodes. In this case, the weld metal yield strength tensile specimen, multiple fisheyes were
is greater than both the HAZ and unaffected base observed, as shown in Fig. 9. An SEM image of
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
Table 3 Chemical analysis of the flange and pipe

Chemical
element Flange Pipe
C 0.80 0.07
Mn 0.67 1.10
Si 0.25 0.24
Ni 0.01 0.01
Cr 0.23 0.01
S 0.13 0.005
P 0.018 0.017
Al ... 0.038
Nb ... 0.02 Fig. 9 Brittle fisheyes appear as bright spots in a gray ductile
matrix.
the center of a bright fisheye exhibited brittle strength and low hardness. It was also found that
intergranular fracture, as shown in Fig. 10. The nickel alloying additions tended to promote
SEM images outside of the fisheyes showed HAC in the form of ferrite vein cracking, while
ductile dimpled microvoid coalescence. Welds an equivalent amount of molybdenum resisted
were repeated to determine the amount of dif- cracking. Both nickel and molybdenum are
fusible hydrogen in similar welds in accordance essential alloying elements for enhancing frac-
with AWS A4.3 (a standard welding test for ture toughness in both the weld metal and base
diffusible hydrogen). Despite the fact that metal. The mechanism by which nickel and
GMAW is known as a very low-hydrogen pro- molybdenum appear to have virtually opposite
cess, the value obtained for diffusible hydrogen effects on susceptibility to HAC is not known.
was 9 mL/100 g, which was far greater than Example 7: Ferrite Vein Cracking in
expected. It was found that the filler metal High-Heat-Input Welds. Single-pass full-
manufacturer used excess hydrocarbon-base penetration electroslag welds were deposited on
lubricant during the wire-drawing operation 50 mm (2 in.) thick ASTM A588 steel using a
because of the high strength of the wire. heat input of 42 kJ/mm (1070 kJ/in.) for bridge
Fisheyes were caused by excessive amounts applications. The ASHTO/AWS D1.5 Bridge
of diffusible hydrogen in the weld metal due to Welding Code required both radiographic and
the lubricant residue on the filler metal. ultrasonic testing (UT) of the completed welds.
The UT revealed possible indications of crack-
ing around the weld center. The weld metal was
Ferrite Vein Cracking sectioned for metallographic examination, and
A very unexpected form of HAC is ferrite ferrite vein cracking was found, as shown in
vein cracking, which can occur in slowly cooled Fig. 11.
electroslag welds. In recent studies of electro- Clearly, cracking was confined to the grain-
slag welding of 50 and 75 mm thick low-carbon boundary ferrite, which was nucleated at the
steel at Portland State University (Ref 46), prior-austenitic grain boundaries.
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 prior-
austenite 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. 11 Ferrite vein crack occurring in the prior-austenite

Fig. 10 SEM image of center of fisheye showing intergranular grain boundaries of weld metal deposited on A709-
fracture grade 50W
Review of the welding procedure revealed exhibited an intergranular mode of fracture.

that the flux and tubular wire used in this process Scanning electron microscopy at 50 revealed
were not baked prior to welding. The unbaked that the crack propagated intergranularly along
flux was a source of moisture. Despite its ex- the prior-austenite grain boundaries, as shown in
tremely high weld heat input and the use of mild Fig. 13. However, at approximately 4000 , the
steel electrodes, electroslag weld metal has been intergranular fracture surface exhibited shallow
shown to be susceptible to HAC. The fabricated dimples, as shown in Fig. 14. This dimpled
metal-cored tubular wire was also unbaked. intergranular mode of fracture is due to the
Under the welding heat, moisture in the flux weakness of the ferrite envelopes surrounding
and filler metal produces atomic oxygen and each prior-austenite grain, as shown in Fig. 15.
hydrogen. Since electroslag welds are large The presence of diffusible hydrogen caused a
single-pass deposits, the weld center is under reduction of ductility of grain-boundary ferrite
substantial tensile stress. The combination of sufficient to fail the bend test. The Beachem
ample diffusible hydrogen and high-shrinkage diagram in Fig. 5 shows that microvoid coales-
tensile residual stress at the weld center provides cence can also be adversely affected by diffu-
the necessary ingredients for ferrite vein crack- sible hydrogen. Using low-hydrogen practices,
ing. Subsequent welds were made with flux that such as baking the flux prior to use, eliminated
was baked to 204 C (400 F) and a new metal- the cracking problem during bend testing.
cored wire that was baked at an elevated tem-
perature prior to shipment. The resulting welds
have since been free of ferrite vein cracking.
Although the mechanism of HAC of grain- Stress-Corrosion Cracking of Steel
boundary ferrite is unknown, elimination of
cracking was achieved by reducing the sources Stress-corrosion cracking of steels is possible
of moisture or hydrogen. 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 Fig. 12 Side-bend test failure of weld
diffusible hydrogen. Even the ductility asso-
ciated with ductile microvoid coalescence is
substantially reduced in the presence of diffu-
sible 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 fol-
lowing 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 ex-
cept 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 op- Fig. 13 Scanning electron micrograph of fracture surface of
tical microscopy of the cracked bend specimen bend failure
stress (but below the yield strength) and an Although many theories for stress-corrosion
aggressive environment. The absence of one of cracking have been suggested, only two appear
these elements will eliminate stress-corrosion to be the basis for such cracking. These are the
cracking. Welds are particularly good sites for stress-sorption theory and the electrochemical
stress-corrosion cracking, because substantial theory.
tensile residual stresses are always present. The The stress-sorption theory states that dam-
shrinkage stresses associated with the solidifi- aging substances in the environment are chemi-
cation and cooling of welds produce near-yield cally absorbed onto the surface of the steel,
tensile residual stresses in and around the weld. causing a reduction in the cohesive bonding
Relatively mild chemical environments can force between iron atoms. Only an atom-thick
activate the stress-corrosion cracking process. surface layer is needed to seriously affect the
Environments known to cause stress-corrosion bonding forces between surface atoms. There
cracking of plain carbon and alloy steels include is a threshold stress necessary to initiate stress-
liquefied ammonia, hydrogen sulfide, molybde- corrosion cracking. In some ways, this mechan-
num disulfide, sodium hydroxide sour gas, high ism resembles HAC in steels (discussed earlier),
pH values, nitrate solutions, and many other where only a few ppm of hydrogen in concert
corrosive environments. Stress-corrosion crack- with tensile stress are needed to reduce the
ing of carbon steel can even take place in pure cohesive bonds between iron atoms and cause
water under high temperature and pressure. 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 em-
itted 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 micro-
structure 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
Fig. 14 Higher-magnification image of fracture surface in evidence of the electrochemical nature of this
Fig. 13 showing dimpled intergranular fracture
cracking process, stress-corrosion cracking can
be stopped by applying cathodic protection. As
soon as cathodic protection is removed, stress-
corrosion 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 spray-
ing boiler feedwater into the line.
When cracks were initially discovered, they
were ground out and rewelded. The repair welds
Fig. 15 Optical microscopy of grain structure of electroslag reportedly cracked after only a few days in
weld metal. Original magnification: 50 service.
A visual examination revealed that neither the attemperator process. It is likely that an upset in
original welds nor the repair welds penetrated the chemistry of the boiler feedwater at some
through to the inside of the pipe. The cracking time in the past contaminated the downstream
consisted primarily of circumferential cracks lines and led to the SCC in this instance.
down the center of the weld. In addition, there The consequences of additional steam line
were a few locations where short longitudinal failures need to be evaluated. Given the nature of
cracks emanated from the primary circumfer- the cracking, inspection methods capable of
ential crack. Macroscopically, the cracks were detecting the cracks in early stages are limited. It
generally irregular and branching. may be most cost-effective to replace the steam
Sections of the cracked region were prepared lines downstream of the attemperator and re-
for metallographic examination of the weld evaluate the methods used for ensuring the
geometry and crack morphology, as shown in proper chemistry of the boiler feedwater.
Fig. 16. A microhardness scan of the base metal Future repairs should ensure complete crack
revealed a hardness of 79 HRB, with a hardness removal and full penetration welds using proper
of 87 to 94 HRB in the original weld. The repair preheat and interpass temperatures to minimize
weld was significantly harder, with a hardness of hardness gradients.
22 to 34 HRC.
Metallographic investigations revealed an
intergranular branching morphology, as shown
in Fig. 17. Scanning electron microscopy also Solidification Cracking of Steel
showed the presence of several secondary
branching cracks emanating from the primary Solidification cracking is one of several forms
crack. These cracks also proceeded in an inter- of hot cracking. Solidification cracking in steel
granular fashion. Energy-dispersive spectro- and steel alloys occurs near the end of the soli-
scopy did not reveal the presence of foreign dification process and is caused by two dominant
materials at the fracture surface in a measurable factors: tensile stress acting on the weld during
quantity. solidification, and a large temperature range
The pipe fractured due to stress-corrosion between the solidus and liquidus temperatures or
cracking (SCC) precipitated by the presence of the presence of low-melting impurities such as
geometric stress concentrations and high resi- sulfer and phosphorus. The tensile stress acting
dual tensile stresses in the weld. The SCC is an on the weld can arise from either shrinkage
environmentally induced cracking mechanism tensile stresses produced during solidification
that can occur in a susceptible material in the and cooldown, or externally applied tensile
presence of tensile stress and an aggressive stress or tensile restraint stress. The effect of the
chemical agent. Although the particular agent in- liquidus-to-solidus temperature range has been
volved 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. 17 Micrograph of the crack near the weld root. Original

Fig. 16. Cross section of weld at butt joint. Etchant: 2% nital. magnification: 100 . Etchant: 2% nital. Courtesy of
Courtesy of MEI-Charlton, Inc. MEI-Charlton, Inc.
dealt with generally by several empirical soli- However, recent research has shown that the
dification cracking equations. For example, effect of carbon on solidification cracking of
Matsuda (Ref 47) developed a very popular low-carbon steels is far more complex and
parameter for solidification cracking of steels nonlinear than predicted by the Lt formula.
called Lt, where increasing Lt increased sus- For example, Masumoto (Ref 48) showed that
ceptibility to cracking: solidification cracking was enhanced for carbon
contents 40.1%. Conversely, Ohshita et al. (Ref
Lt =70(C Si=12 Mn=9+3P+4S+Ni=23 41) reported that cracking was enhanced for
+Cr=35+Mo=70) carbon 50.1% and that nickel additions were
beneficial in reducing the cracking effect of
where all elements are in weight percent. carbon, Karjalainen et al. (Ref 49) surveyed the
Clearly the effects of carbon and alloying technical literature and reported that there was a
elements such as manganese, molybdenum, least-susceptible range of carbon contents
chromium, and nickel were assumed to be linear, between 0.1 and 0.17%. Within this range, the
as shown in the Lt equation. As the Lt equation cracking susceptibility was minimized. Ichi-
suggests, decreasing carbon content has always kawa et al. (Ref 50) reported peak solidification
been assumed to decrease solidification cracking cracking susceptibility at 0.035% C, followed by
susceptibility in steel weld metal. Prediction enhanced cracking when the carbon content
formulas for solidification cracking show exceeded 0.1%. Most recently, Kim et al. (Ref
the effect of carbon on cracking to be linear. 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 soli-
dification 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 transvar-
estraint tests. Maximum crack distance and
Fig. 18 Maximum crack length (MCL) as a function of carbon
maximum crack length were measured at a 4%
content in iron weld metal obtained in transvar-
estraint tests at 4% augmented strain. Source: Ref 53, 54 augmented strain.
Fig. 19 Solidification cracking in weld metal

The data (schematically plotted in Fig. 18) with increasing carbon content because of the
clarify some of the conflicting issues in the decreasing solidus/liquidus temperature range.
literature. Solidification cracking susceptibility In region 4, the solidification cracking suscept-
is controlled by the brittle temperature range ibility increased due to the increasing solidus/
(BTR), the d?c transformation stress, and the liquidus temperature range.
iron-carbon peritectic reaction. According to Example 10: Solidification Cracking of
Shankar and Devletian (Ref 53, 54), there are Steel Weld. Welds were deposited by flux-
four distinct %C ranges that produced char- cored arc welding on 12 mm thick AISI/SAE
acteristic solidification behavior. These include: 1020 steel plate at high travel speeds for maxi-
mum cost-effectiveness. Weld joints were
Region 1: less than ~0.09% C (maximum highly restrained during welding to prevent
solid solubility of carbon in d-iron) distortion. Visible longitudinal centerline cracks
Region 2: ~0.09 to 0.11% C (maximum were observed, as shown in Fig. 19. The portions
solidification cracking) of the weld seam that were not visibly cracked
Region 3: ~0.11 to 0.16% C failed the root bend test.
Region 4: greater than 0.16% C (iron-carbon The cracked specimen was broken open and
peritectic point) observed under the SEM. Rejected bend-test
In region 1, there was no cracking below specimens were also broken open for examina-
0.01% C in transvarestraint tests. This is because tion by SEM. The crack surface clearly showed
the solidus/liquidus temperature range was that the mode of fracture was solidification
negligible. However, as the carbon content in- cracking, as shown in Fig. 20. Spectrographic
creased, the cracking susceptibility increased analysis of the weld metal admixture revealed a
rapidly up to approximately 0.06% C. The carbon content of 0.1%. Subsequent welds were
cracking dropped slightly at 0.075% and then deposited with reduced welding travel speed in
continued to increase with carbon content up to order to reduce the length of the teardrop shape
0.09%, due to the increasing solidus/liquidus of the weld puddle. Subsequent welds deposited
temperature range. In region 2, a large peak in at the reduced welding speed were crack-free.
solidification cracking was observed. This cri- Centerline cracking failure was caused by
tical cracking peak, centered at approximately solidification cracking. Decreasing welding
0.1% C, was found to be due to the simultaneous speed reduced susceptibility to solidification
action of three factors: the maximum solidus/ cracking in the weld metal. Reducing restraint
liquidus temperature range, the d?c trans- during welding and reducing the weld metal
formation stresses, and the occurrence of the carbon content (or Matsudas Lt factor, men-
BTR. At 0.1% C, cracking occurred with a tioned previously) would have also decreased
minimum critical strain and low fracture stress. the occurrence of solidification cracking.
In region 3, solidification cracking decreased
REFERENCES
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
Fig. 20 Scanning electron micrograph of fracture surface Strength Pipeline Steels, Weld. J., Vol 65
showing solidification cracking (No. 7), 1986, p 175s183s
6. N. Yurioka and T. Kasuya, A Chart Method Micro- and Transverse Cracking in Weld
to Determine Necessary Preheat in Welding, Metals, Nippon Steel Corporation, Oct
Weld. World, Vol 35 (No. 5), 1995, p 3138 1983
7. N. Yurioka and T. Kasuya, A Chart 20. R.J. Wong and R.D. Hayes, Arc Welding
Method to Determine Necessary Preheat in Consumables for HSLA Steels with Yield
Welding, IIW Doc. II-1230-94 and IIW Strengths of 80 ksi and Above, The Metal-
Doc. IX-1740-94, International Institute of lurgy of Welding and Qualification of
Welding, 1994, p 12 Microalloyed Steel Weldments, J.T. Hickey
8. T. Kasuya and N. Yurioka, Determination et al., Ed., American Welding Society,
of Necessary Preheat Temperature to Avoid 1990, p 450489
Cold Cracking under Varying Ambient 21. P.W. Holsberg and R.J. Wong, Welding of
Temperature, ISIJ Int., Vol 35 (No. 10), HSLA-100 Steel for Naval Applications,
1995, p 11831189 Weldability of Materials, R.A. Patterson
9. M. McParlan and B.A. Graville, Hydrogen and K.W. Mohin, Ed., ASM International,
Cracking in Weld Metals, Weld. J., Vol 55 1990, p 219239
(No. 4), 1976, p 95s102s 22. J.J. DeLoach, C. Null, S. Fiore, and
10. J. Brozda, A Comparison Between the P. Konkol, The Right Welding Wire Could
Levels of Preheat Necessary to Prevent Help the U.S. Navy Save Millions, Weld. J.,
Cold Cracking during Welding of Low Vol 78 (No. 6), June 1999, p 5558
Alloy High Strength Steels, Weld. Int., 23. R.A. Oriani and P.H. Josephic, Equilibrium
Vol 11 (No. 2), 1997, p 2532 Aspects of Hydrogen-Induced Cracking of
11. B.A. Graville, Interpretive Report on Steels, Acta Metall., Vol 22 (No. 9), 1974,
Weldability Tests for Hydrogen Cracking of p 10651074
Higher Strength Steels and their Potential 24. R.A. Oriani, Effects of Hydrogen on the
for Standardization, Bulletin 400, Welding Plastic Response of Steels, Hydrogen
Research Council, New York, 1995 Effects in Metals, I.M. Bernstein and A.W.
12. T. Terasaki, G.T. Hall, and P.L. Harrison, Thompson, Ed., Metallurgical Society of
Predictive Equation for Cooling Time of AIME, 1981, p 235245
CTS Test Welds, Trans. Jpn. Weld. Soc., 25. R.A. Oriani, Hardening and Softening
Vol 21 (No. 2), Oct 1990, p 5155 Induced by Hydrogen in Carbon Steels,
13. N. Yurioka, H. Suzuki, H. Okumura, S. NATO Conf. Series, Mater. Sci., Vol 5,
Ohshita, and S. Saito, Carbon Equivalents 1983, p 795798
to Assess Cold Cracking Sensitivity and 26. J.K. Lin and R.A. Oriani, Effect of Hydro-
Hardness of Steel Welds, Nippon Steel gen on the Initiation of Shear Localization
Technical Report 20, Dec 1982, p 6173 in Plain Carbon Steels, Perspectives in
14. F.R. Coe, Welding Steels without Hydrogen Hydrogen in Metals, Pergamon Press, 1986,
Cracking, The Welding Institute, Abington, p 615621
U.K., 1973 27. R.A. Oriani, HydrogenThe Versatile
15. N. Yurioka, H. Suzuki, S. Ohshita and Embrittler, Corrosion, Vol 43 (No. 7), July
S. Saito, Determination of Necessary Pre- 1987, p 390397
heat Temperature in Steel Welding, Weld. 28. R.A. Oriani, Effects of Hydrogen on the
J., Vol 62 (No. 6), 1983, p 147s153s Plastic Properties of Medium Carbon Steels,
16. C.L.M. Cottrell, An Improved Predictive Metall. Trans. A, Vol 11 (No. 11), 1980,
Method for Avoiding Hydrogen Cracking, p 18091820
Weld. Met. Fabr., April 1990, p 178183 29. J.K. Lin and R.A. Oriani, Effect of Hydro-
17. N. Yurioka, Studies on Delayed Cracking gen on the Initiation of Shear Localization
in Steel Welds (Report 1); Prevention in Plain Carbon Steels, Acta Metall., Vol 31
of Root Cracking in Root-Pass Welds, (No. 7), 1983, p 10711077
Nippon Steel Corporation, Dec 1981 30. H.A. Wriedt and R.A. Oriani, Effect of
18. N. Yurioka, Studies on Delayed Cracking Tensile and Compressive Elastic Stress
in Steel Welds (Report 2); Prevention of on Equilibrium Hydrogen Solubility in a
Cracking in Restraint Multi-Run Welds, Solid, Acta Metall., Vol 18 (No. 7), 1970,
Nippon Steel Corporation, Dec 1981 p 753760
19. N. Yurioka, Studies on Delayed Cracking 31. A.R. Troiano, The Role of Hydrogen
in Steel Welds (Report 3); Prevention of and Other Interstitials in the Mechanical
Behavior of Metals, Trans. ASM, Vol 52, Cracking in Steel Construction: Report 1,
1960, p 5480 J. Jpn. Weld. Soc., Vol 40 (No. 11), 1971,
32. C.C. Beachem, New Model for Hydrogen- p 11171127
Assisted Cracking (Hydrogen Embrittle- 43. O. Grong and D.K. Matlock, Micro-
ment), Metall. Trans., Vol 3 (No. 2), Feb structural Development in Mild and Low
1972, p 437451 Alloy Steel Weld Metals, Int. Met. Rev.,
33. S.K. Albert, V. Ramasubbu, N. Parva- Vol 31, 1986, p 2748
thavarthini, and T.P.S. Gill, Influence of 44. P.L. Harrison and R.A. Farrar, Acicular
Alloying on Hydrogen-Assisted Cracking Ferrite in C-Mn Weld Metal, J. Mater. Sci.,
and Diffusible Hydrogen Content in Cr-Mo Vol 22 (No. 11), 1987, p 38123820
Steel Welds, Sadhana, Vol 28 (No. 3, 4), 45. G.L.F. Powell, I.H. Brown, V.M. Linton,
June/Aug 2003, p 383393 and J.L. Davidson, Cold Cracking and
34. R. Gibala and D.S. DeMiglio, Hydrogen in Segregation in Multi-Pass Welds of Quen-
Steels: Interactions, Traps and Crack Paths, ched and Tempered Steel, Trends in Weld-
Hydrogen Effects in Metals, I.M. Bernstein ing Research, S.A. David et al., Ed., ASM
and A.W. Thompson, Ed., Metallurgical International, 2002, p 575580
Society of AIME, 1981, p 113122 46. W. Wood, J.H. Devletian, and R. Turpin,
35. G.M. Pressouyre, Hydrogen Traps, Repel- NGI-ESW Consumables Development,
lers, and Obstacles in Steel: Consequence Final Report, FHWA Contract DTFH61-97-
on Hydrogen Diffusion, Solubility, and 0030, Aug 2003
Embrittlement, Metall. Trans. A, Vol 14 47. F. Matsuda, Hot Cracking Susceptibility of
(No. 10), 1983, p 21892193 Weld Metal, First U.S.-Japan Symposium
36. G.M. Pressouyre and I.M. Bernstein, on Weldability of Materials, AWS, JWS,
Example of the Effect of Hydrogen Trap- and JWES, 1990, p 1935
ping on Hydrogen Embrittlement, Metall. 48. I. Masumoto, Solidification, Blow Hole For-
Trans. A, Vol 12 (No. 5), 1981, p 835844 mation and Cracking of Steel Weld Metal,
37. C.A. Lensing, Y.D. Park, I.S. Maroef, and Weld. World, Vol 31 (No. 3), p 211221
D.L. Olson, Yttrium Hydrogen Trapping to 49. L.P. Karjalainen, V.P. Kujanpaa, and N.
Mangage Hydrogen in HSLA Welds, Weld. Suutala, Hot Cracking in Iron Base Alloys
J., Vol 83 (No. 9), 2004, p 254s266s Effect of Solidification Mode, Advan-
38. I.S. Maroef and D.L. Olson, Evaluation of ces in Welding Science and Techno-
Hydrogen Trapping for Hydrogen Man- logy, American Society for Metals, 1986,
agement in Steel Welding, Proc. of Sixth p 145149
International Trends in Welding Research, 50. K. Ichikawa, H.K.D.H. Bhadeshia, and
ASM International, 2002, p 557562 D.J.C. MacKay, Model for Solidification
39. D.L. Olson, I. Maroef, C. Lensing, D. Cracking in Low Alloy Steel Weld Metal,
Smith, T. Wildeman, and M. Eberhart, Sci. Technol. Weld. Joining, Vol 1 (No. 1),
Hydrogen Management in Steel Weldments, p 4350
J.L. Davidson and D.L. Olson, Ed., Organ- 51. J.G. Kim, K.T. Rie, and J. Ruge, Contribu-
izing Committee of the Joint Seminar on tion to Hot Cracking Susceptibility of Sub-
behalf of Defense Science and Technology merged Arc Steel Weldments, Steel Res,
Assoc. and Welding Technology Institute of Vol 56 (No. 12), p 639644
Australia, 1996, p 119 52. Y.M. Won, H.N. Han, T.J. Yeo, and K.H.
40. I. Maroef, D.L. Olson, M. Eberhart, and Oh, Analysis of Solidification Cracking
G.R. Edwards, Hydrogen Trapping in Fer- Using the Specific Crack Susceptibility,
ritic Steel Weld Metal, Int. Mater. Rev., ISIJ Int., Vol 40 (No. 2), p 129136
Vol 47 (No. 4), 2002, p 191223 53. J.H. Devletian, Final Report, Award
41. S. Ohshita, N. Yurioka, N. Mori, and 0096290, National Science Foundation,
T. Kimura, Prevention of Solidification 2004
Cracking in Very Low Carbon Steel, Weld. 54. V. Shankar and J.H. Devletian, Solidifi-
J., Vol 62 (No. 5), p 129s136s cation Cracking in Low Alloy Steel Welds,
42. Y. Ito and K. Bessyo, Determination of Sci. Technol. Weld. Joining, Vol 10 (No. 2),
Preheating Conditions to Avoid Weld 2005, p 236243
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_521-523.pdf/Appendix_1/ 18/8/2008 4:24PM Plate # 0 pg 521
APPENDIX 1
Metric Conversion Guide
THIS APPENDIX is intended as a guide for Institute, 1430 Broadway, New York, NY
expressing weights and measures in the Systeme 10018
International dUnites (SI). The purpose of SI The International System of Units, SP 330,
units, developed and maintained by the General 1986, National Institute of Standards and
Conference of Weights and Measures, is to Technology. Order from Superintendent of
provide a basis for worldwide standardization of Documents, U.S. Government Printing
units and measure (Table A1.1). For more infor- Office, Washington, DC 20402-9325
mation on metric conversions (Table A1.2), the Metric Editorial Guide, 4th ed. (revised),
reader should consult the following references: 1985, American National Metric Council,
1010 Vermont Avenue NW, Suite 1000,
Standard for Metric Practice, E380, Washington, DC 20054960
Annual Book of ASTM Standards, American ASME Orientation and Guide for Use of SI
Society for Testing and Materials, 1916 (Metric) Units, ASME Guide SI 1, 9th ed.,
Race Street, Philadelphia, PA 19103 1982, The American Society of Mechanical
Metric Practice, ANSI/IEEE 268 Engineers, 345 East 47th Street, New York,
1982, American National Standards NY 10017
Table A1.1 Base, supplementary, and derived SI units

Measure Unit Symbol Measure Unit Symbol
Base units Electric charge density coulomb per cubic meter C/m3
Electric field strength volt per meter V/m
Amount of substance mole mol
Electric flux density coulomb per square meter C/m2
Electric current ampere A
Electric potential, volt V
Length meter m
potential difference,
Luminous intensity candela cd
electromotive force
Mass kilogram kg
Electric resistance ohm V
Thermodynamic kelvin K
Energy, work, quantity joule J
temperature
of heat
Time second s
Energy density joule per cubic meter J/m3
Entropy joule per kelvin J/K
Supplementary units
Force newton N
Plane angle radian rad Frequency hertz Hz
Solid angle steradian sr Heat capacity joule per kelvin J/K
Heat flux density watt per square meter W/m2
Derived units
llluminance lux lx
Absorbed dose gray Gy Inductance henry H
Acceleration meter per second squared m/s2 Irradiance watt per square meter W/m2
Activity (of becquerel Bq Luminance candela per square meter cd/m2
radionuclides) Luminous flux lumen lm
Angular acceleration radian per second squared rad/s2 Magnetic field strength ampere per meter A/m
Angular velocity radian per second rad/s Magnetic flux weber Wb
Area square meter m2 Magnetic flux density tesla T
Capacitance farad F Molar energy joul per mole J/mol
Concentration (of amount mole per cubic meter mol/m3 Molar entropy joule per mole kelvin J/mol K
of substance) Molar heat capacity joule per mole kelvin J/mol K
Conductance siemens S Moment of force newton meter Nm
Current density ampere per square meter A/m2 Permeability henry per meter H/m
Density, mass kilogram per cubic meter kg/m3 Permittivity farad per meter F/m
(continued)
Table A1.1 (continued)

Measure Unit Symbol Measure Unit Symbol
Power, radiant flux watt W Specific entropy joule per kilogram kelvin J/kg K
Pressure, stress pascal Pa Specific volume cubic meter per kilogram m3/kg
Quantity of electricity, coulomb C Surface tension newton per meter N/m
electric charge Thermal conductivity watt per meter kelvin W/m K
Radiance watt per square meter W/m2 sr Velocity meter per second m/s
steradian Viscosity, dynamic pascal second Pa s
Radiant intensity watt per steradian W/sr Viscosity, kinematic square meter per second m2/s
Specific heat capacity joule per kilogram kelvin J/kg K Volume cubic meter m3
Specific energy joule per kilogram J/kg Wavenumber 1 per meter l/m
Table A1.2 Conversion factors

To convert from to multiply by To convert from to multiply by
Angle Force per unit length
degree rad 1.745 329 E02 lbf/ft N/m 1.459 390 E+01
lbf/in. N/m 1.751 268 E+02
Area
Fracture toughness
in.2 mm2 6.451 600 E+02 p p
in.2 cm2 6.451 600 E+00 ksi in: MPa m 1.098 800 E+00
in.2 m2 6.451 600 E04
Heat content
ft2 m2 9.290 304 E02
Btu/lb kJ/kg 2.326 000 E+00
Bending moment or torque cal/g kJ/kg 4.186 800 E+00
lbf in. Nm 1.129 848 E01 Heat input
lbf ft Nm 1.355 818 E+00
kgf m Nm 9.806 650 E+00 J/in. J/m 3.937 008 E+01
ozf in. Nm 7.061 552 E03 kJ/in. kJ/m 3.937 008 E+01
Length
Bending moment or torque per unit length
A nm 1.000 000 E01
lbf in./in. N m/m 4.448.222 E+00 min. mm 2.540 000 E02
lbf ft/in. N m/m 5.337 866 E+01 mil mm 2.540 000 E+01
in. mm 2.540 000 E+01
Current density
in. cm 2.540 000 E+00
A/in.2 A/cm2 1.550 003 E01 ft m 3.048 000 E01
A/in.2 A/mm2 1.550 003 E03 yd m 9.144 000 E01
A/ft2 A/m2 1.076 400 E01 mile km 1.609 300 E+00
Mass
Electricity and magnetism
oz kg 2.834 952 E02
gauss T 1.000 000 E04
lb kg 4.535.924 E01
maxwell mWb 1.000 000 E02
ton (short, 2000 lb) kg 9.071 847 E+02
mho S 1.000 000 E+00
ton (short, 2000 lb) kg 103(a) 9.071 847 E01
Oersted A/m 7.957 700 E+01
ton (long, 2240 lb) kg 1.016 047 E+03
V cm Vm 1.000 000 E02
V circular-mil/ft mV m 1.662 426 E 03 Mass per unit area
oz/in.2 kg/m2 4.395 000 E+01
Energy (impact, other) oz/ft2 kg/m2 3.051 517 E01
oz/yd2 kg/m2 4.882 428 E+00
ft lbf J 1.355 818 E+00
Btu (thermochemical) J 1.054 350 E+03 lb/ft2 kg/m2 4.882 428 E+00
cal (thermochemical) J 4.184 000 E+00 Mass per unit length
kW h J 3.600 000 E+06
Wh J 3.600 000 E+03 lb/ft kg/m 1.488 164 E+00
lb/in. kg/m 1.785 797 E+01
Flow rate Mass per unit time
ft3/h L/min 4.719 475 E01 lb/h kg/s 1.259 979 E04
ft3/min L/min 2.831 000E+01 lb/min kg/s 7.559 873 E03
gal/h L/min 6.309 020 E+02 lb/s kg/s 4.535 924 E01
gal/min L/min 3.785 412 E+00
Mass per unit volume (includes density)
Force
g/cm3 kg/m3 1.00 000 E+03
lbf N 4.448 222 E+00 lb/ft3 g/cm3 1.601 846 E02
kip (100 lbf ) N 4.448 222 E+03 lb/ft3 kg/m3 2.767 846 E+01
tonf kN 8.896 443 E+00 lb/in.3 g/cm3 2.767 846 E+04
kgf N 9.806 650 E+00 lb/in.3 kg/m3 2.767 990 E+04
(continued)
(a) kg 103 = 1 metric ton or 1 megagram (Mg)
Appendix 1: Metric Conversion Guide / 523

To convert from to multiply by To convert from to multiply by
Power Thermal conductivity
Btu/s kW 1.055 056 E+00 Btu in./s ft2 F W/m K 5.192 204 E+02
Btu/min kW 1.758 426 E02 Btu/ft h F W/m K 1.730 735 E+00
Btu/h W 2.928 751 E01 Btu in./h ft2 F W/m K 1.442 279 E01
erg/s W 1.000 000 E07 cal/cm s C W/m K 4.184 000 E+02
ft lbf/s W 1.355 818 E+00
ft lbf/min W 2.259 697 E02 Thermal expansion
ft lbf/h W 3.766 161 E04 in./in. C m/m k 1.000 000 E+00
hp (550 ft lbf/s) kW 7.456 999 E01 in./in. F m/m k 1.800 000 E+00
hp (electric) kW 7.460 000 E01
Velocity
Power density
2 2 ft/h m/s 8.466 667 E05
W/in. W/m 1.550 003 E+03
ft/min m/s 5.080 000 E03
Press capacity ft/s m/s 3.048 000 E01
in./s m/s 2.540 000 E02
See Force
km/h m/s 2.777 778 E01
Pressure (fluid) mph mm/h 1.609 344 E+00
atm (standard) Pa 1.013 250 E+05 Velocity of rotation
bar Pa 1.000 000 E+05
in. Hg (32 F) Pa 3.386 380 E+03 rev/min (rpm) rad/s 1.047 164 E01
in. Hg (60 F) Pa 3.376 850 E+03 rev/s rad/s 6.283 185 E+00
lb/in.2 (psi) Pa 6.894 757 E+03 Viscosity
torr (mm Hg, 0 C) Pa 1.333 220 E+02
poise pa s 1.000 000 E01
Specific heat stokes m2/s 1.000 000 E04
Btu/lb F J/kg K 4.186 800 E+03 ft2/s m2/s 9.290 304 E02
cal/g C J/kg K 4.186 800 E+03 in.2/s mm2/s 6.451 600 E+02
Stress (force per unit area) Volume

2
tonf/in. (tsi) MPa 1.378 951 E+01 in.3 m3 1.638 706 E05
kgf/mm2 MPa 9.806 650 E+00 ft3 m3 2.831 685 E02
ksi MPa 6.894 757 E+00 fluid oz m3 2.957 353 E05
lbf/in.2 (psi) MPa 6.894 757 E03 gal (U.S. liquid) m3 3.785 412 E03
MN/m2 MPa 1.000 000 E+00
Volume per unit time
Temperature
ft3/min m3/s 4.719 474 E04
F C 5/9 ( F32) ft3/s m3/s 2.831 685 E02
R K 5/9 in.3/min m3/s 2.731 177 E07
Temperature interval Wavelength
F C 5/9
A nm 1.000 000 E01
(a) kg 103 = 1 metric ton or 1 megagram (Mg)
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 C F C F C F C
... 458 272.22 ... 368 222.22 ... 278 172.22 306.4 188 122.22
... 456 271.11 ... 366 221.11 ... 276 171.11 302.8 186 121.11
... 454 270.00 ... 364 220.00 ... 274 170.00 299.2 184 120.00
... 452 268.89 ... 362 218.89 457.6 272 168.89 295.6 182 118.89
... 450 267.78 ... 360 217.73 454.0 270 167.78 292.0 180 117.78
... 448 266.67 ... 358 216.67 450.4 268 166.67 288.4 178 116.67
... 446 265.56 ... 356 215.56 446.8 266 165.56 284.8 176 115.56
... 444 264.44 ... 354 214.44 443.2 264 164.44 281.2 174 114.44
... 442 263.33 ... 352 213.33 439.6 262 163.33 277.6 172 113.33
... 440 262.22 ... 350 212.22 436.0 260 162.22 274.0 170 112.22
... 433 261.11 ... 348 211.11 432.4 258 161.11 270.4 168 111.11
... 436 260.00 ... 346 210.00 428.8 256 160.00 266.8 166 110.00
... 434 258.89 ... 344 208.89 425.2 254 158.89 263.2 164 108.89
... 432 257.78 ... 342 207.78 421.6 252 157.78 259.6 162 107.78
... 430 256.67 ... 340 206.67 418.0 250 156.67 256.0 160 106.67
... 428 255.56 ... 338 205.56 414.4 248 155.56 252.4 158 105.56
... 426 254.44 ... 336 204.44 410.8 246 154.44 248.8 156 104.44
... 424 253.33 ... 334 203.33 407.2 244 153.33 245.2 154 103.33
... 422 252.22 ... 332 202.22 403.6 242 152.22 241.6 152 102.22
... 420 251.11 ... 330 201.11 400.0 240 151.11 238.0 150 101.11
... 418 250.00 ... 328 200.00 396.4 238 150.00 234.4 148 100.00
... 416 248.89 ... 326 198.89 392.8 236 148.89 230.8 146 98.89
... 414 247.78 ... 324 197.78 389.2 234 147.78 227.2 144 97.78
... 412 246.67 ... 322 196.67 385.6 232 146.67 223.6 142 96.67
... 410 245.56 ... 320 195.56 382.0 230 145.56 220.0 140 95.56
... 408 244.44 ... 318 194.44 378.4 228 144.44 216.4 138 94.44
... 406 243.33 ... 316 193.33 374.8 226 143.33 212.8 136 93.33
... 404 242.22 ... 314 192.22 371.2 224 142.22 209.2 134 92.22
... 402 241.11 ... 312 191.11 367.6 222 141.11 205.6 132 91.11
... 400 240.00 ... 310 190.00 364.0 220 140.00 202.0 130 90.00
... 398 238.89 ... 308 188.89 360.4 218 138.89 198.4 128 88.89
... 396 237.78 ... 306 187.78 356.8 216 137.78 194.8 126 87.78
... 394 236.67 ... 304 186.67 353.2 214 136.67 191.2 124 86.67
... 392 235.56 ... 302 185.56 349.6 212 135.56 187.6 122 85.56
... 390 234.44 ... 300 184.44 346.0 210 134.44 184.0 120 84.44
... 388 233.33 ... 298 183.33 342.4 208 133.33 180.4 113 83.33
... 386 232.22 ... 296 182.22 338.8 206 132.22 176.8 116 82.22
... 384 231.11 ... 294 181.11 335.2 204 131.11 173.2 114 81.11
... 382 230.00 ... 292 180.00 331.6 202 130.00 169.6 112 80.00
... 380 228.89 ... 290 178.89 328.0 200 128.89 166.0 110 78.89
... 378 227.78 ... 288 177.78 324.4 198 127.78 162.4 108 77.78
... 376 226.67 ... 286 176.67 320.8 196 126.67 158.8 106 76.67
... 374 225.56 ... 284 175.56 317.2 194 125.56 155.2 104 75.56
... 372 224.44 ... 282 174.44 313.6 192 124.44 151.6 102 74.44
... 370 223.33 ... 280 173.33 310.0 190 123.33 148.0 100 73.33
(continued)
F C F C F C F C
144.4 98 72.22 +89.6 +32 +0.00 323.6 162 72.22 557.6 292 144.44
140.8 96 71.11 +93.2 +34 +1.11 327.2 164 73.33 561.2 294 145.56
137.2 94 70.00 +96.8 +36 +2.22 330.8 166 74.44 564.8 296 146.67
133.6 92 68.89 +100.4 +38 +3.33 334.4 168 75.66 568.4 298 147.78
130.0 90 67.78 +104.0 +40 +4.44 338.0 170 76.67 572.0 300 148.89
126.4 88 66.67 +107.6 +42 +5.56 341.6 172 77.78 575.6 302 150.00
122.8 86 65.56 +111.2 +44 +6.67 345.2 174 78.89 579.2 304 151.11
119.2 84 64.44 +114.8 +46 +7.78 348.8 176 80.00 582.8 306 152.22
115.6 82 63.33 +118.4 +48 +8.89 352.4 178 81.11 586.4 308 153.33
112.0 80 62.22 +122.0 +50 +10.00 356.0 180 82.22 590.0 310 154.44
108.4 78 61.11 +125.6 +52 +11.11 359.6 182 83.33 593.6 312 155.56
104.8 76 60.00 +129.2 +54 +12.22 363.2 184 84.44 597.2 314 156.67
101.2 74 58.89 +132.8 +56 +13.33 366.8 186 85.56 600.8 316 157.78
97.6 72 57.78 +136.4 +58 +14.44 370.4 188 86.67 604.4 318 158.89
94.0 70 56.67 +140.0 +60 +15.56 374.0 190 87.78 608.0 320 160.00
90.4 68 55.56 143.6 62 16.67 377.6 192 88.89 611.6 322 161.11
86.8 66 54.44 147.2 64 17.78 381.2 194 90.00 615.2 324 162.22
83.2 64 53.33 150.8 66 18.89 384.8 196 91.11 618.8 326 163.33
79.6 62 52.22 154.4 68 20.00 388.4 198 92.22 622.4 328 164.44
76.0 60 51.11 158.0 70 21.11 392.0 200 93.33 626.0 330 165.56
72.4 58 50.00 161.6 72 22.22 395.6 202 94.44 629.6 332 166.67
68.8 56 48.89 165.2 74 23.33 399.2 204 95.56 633.2 334 167.78
65.2 54 47.78 168.8 76 24.44 402.8 206 96.67 636.8 336 168.89
61.6 52 46.67 172.4 78 25.56 406.4 208 97.73 640.4 338 170.00
58.0 50 45.56 176.0 80 26.67 410.0 210 98.89 644.0 340 171.11
54.4 48 44.44 179.6 82 27.78 413.6 212 100.00 647.6 342 172.22
50.8 46 43.33 183.2 84 28.89 417.2 214 101.11 651.2 344 173.33
47.2 44 42.22 186.8 86 30.00 420.8 216 102.22 654.8 346 174.44
43.6 42 41.11 190.4 88 31.11 424.4 218 103.33 658.4 348 175.56
40.0 40 40.00 194.0 90 32.22 428.0 220 104.44 662.0 350 176.67
36.4 38 38.89 197.6 92 33.33 431.6 222 105.56 665.6 352 177.78
32.8 36 37.78 201.2 94 34.44 435.2 224 106.67 669.2 354 178.89
29.2 34 36.67 204.8 96 35.56 438.8 226 107.78 672.8 356 180.00
25.6 32 35.56 208.4 98 36.67 442.4 228 108.89 676.4 358 181.11
22.0 30 34.44 212.0 100 37.78 446.0 230 110.00 680.0 360 182.22
18.4 28 33.33 215.6 102 38.89 449.6 232 111.11 683.6 362 183.33
14.8 26 32.22 219.2 104 40.00 453.2 234 112.22 687.2 364 184.44
11.2 24 31.11 222.8 106 41.11 456.8 236 113.33 690.8 366 185.56
7.6 22 30.00 226.4 108 42.22 460.4 238 114.44 694.4 368 186.67
4.0 20 28.89 230.0 110 43.33 464.0 240 115.56 698.0 370 187.78
0.4 18 27.78 233.6 112 44.44 467.6 242 116.67 701.6 372 188.89
+3.2 16 26.67 237.2 114 45.56 471.2 244 117.78 705.2 374 190.00
+6.8 14 25.56 240.8 116 46.67 474.8 246 118.89 708.8 376 191.11
+10.4 12 24.44 244.4 118 47.78 478.4 248 120.00 712.4 378 192.22
+14.0 10 23.33 248.0 120 48.89 482.0 250 121.11 716.0 380 193.33
+17.6 8 22.22 251.6 122 50.00 485.6 252 122.22 719.6 382 194.44
+21.2 6 21.11 255.2 124 51.11 489.2 254 123.33 723.2 384 195.56
+24.8 4 20.00 258.8 126 52.22 492.8 256 124.44 726.8 386 196.67
+28.4 2 18.89 262.4 128 53.33 496.4 258 125.56 730.4 388 197.78
+32.0 +0 17.78 266.0 130 54.44 500.0 260 126.67 734.0 390 198.89
+35.6 +2 16.67 269.6 132 55.56 503.6 262 127.78 737.6 392 200.00
+39.2 +4 15.56 273.2 134 56.67 507.2 264 128.89 741.2 394 201.11
+42.8 +6 14.44 276.8 136 57.78 510.8 266 130.00 744.8 396 202.22
+46.4 +8 13.33 280.4 138 58.89 514.4 268 131.11 748.4 398 203.33
+50.0 +10 12.22 284.0 140 60.00 518.0 270 132.22 752.0 400 204.44
+53.6 +12 11.11 287.6 142 61.11 521.6 272 133.33 755.6 402 205.56
+57.2 +14 10.00 291.2 144 62.22 525.2 274 134.44 759.2 404 206.67
+60.8 +16 8.89 294.8 146 63.33 528.8 276 135.56 762.8 406 207.78
+64.4 +18 7.78 298.4 148 64.44 532.4 278 136.67 766.4 408 208.89
+68.0 +20 6.67 302.0 150 65.56 536.0 280 137.78 770.0 410 210.00
+71.6 +22 5.56 305.6 152 66.67 539.6 282 138.89 773.6 412 211.11
+75.2 +24 4.44 309.2 154 67.73 543.2 284 140.00 777.2 414 212.22
+78.8 +26 3.33 312.8 156 68.83 546.8 286 141.11 780.8 416 213.33
+82.4 +28 2.22 316.4 158 70.00 550.4 288 142.22 784.4 418 214.44
+86.0 +30 1.11 320.0 160 71.11 554.0 290 143.33 788.0 420 215.56
(continued)
Appendix 2: Temperature Conversion Table / 527
F C F C F C F C
791.6 422 216.67 1040.0 560 293.33 2210.0 1210 654.44 3380.0 1860 1015.6
795.2 424 217.78 1058.0 570 298.89 2228.0 1220 660.00 3398.0 1870 1021.1
798.8 426 218.89 1076.0 580 304.44 2246.0 1230 665.56 3416.0 1880 1026.7
802.4 428 220.00 1094.0 590 310.00 2264.0 1240 671.11 3434.0 1890 1032.2
806.0 430 221.11 1112.0 600 315.56 2282.0 1250 676.67 3452.0 1900 1037.8
809.6 432 222.22 1130.0 610 321.11 2300.0 1260 682.22 3470.0 1910 1043.3
813.2 434 223.33 1148.0 620 326.67 2318.0 1270 687.78 3488.0 1920 1048.9
816.8 436 224.44 1166.0 630 332.22 2336.0 1280 693.33 3506.0 1930 1054.4
820.4 438 225.56 1184.0 640 337.78 2354.0 1290 698.89 3524.0 1940 1060.0
824.0 440 226.67 1202.0 650 343.33 2372.0 1300 704.44 3542.0 1950 1065.6
827.6 442 227.78 1220.0 660 348.89 2390.0 1310 710.00 3560.0 1960 1071.1
831.2 444 228.89 1238.0 670 354.44 2408.0 1320 715.56 3578.0 1970 1076.7
834.8 446 230.00 1256.0 680 360.00 2426.0 1330 721.11 3596.0 1980 1082.2
838.4 448 231.11 1274.0 690 365.56 2444.0 1340 726.67 3614.0 1990 1087.8
842.0 450 232.22 1292.0 700 371.11 2462.0 1350 732.22 3632.0 2000 1093.3
845.6 452 233.33 1310.0 710 376.67 2480.0 1360 737.78 3650.0 2010 1098.9
849.2 454 234.44 1328.0 720 382.22 2498.0 1370 743.33 3668.0 2020 1104.4
852.8 456 235.56 1346.0 730 387.78 2516.0 1380 748.89 3686.0 2030 1110.0
856.4 458 236.67 1364.0 740 393.33 2534.0 1390 754.44 3704.0 2040 1115.6
860.0 460 237.78 1382.0 750 398.89 2552.0 1400 760.00 3722.0 2050 1121.1
863.6 462 238.89 1400.0 760 404.44 2570.0 1410 765.56 3740.0 2060 1126.7
867.2 464 240.00 1418.0 770 410.00 2588.0 1420 771.11 3758.0 2070 1132.2
870.8 466 241.11 1436.0 780 415.56 2606.0 1430 776.67 3776.0 2080 1137.8
874.4 468 242.22 1454.0 790 421.11 2624.0 1440 782.22 3794.0 2090 1143.3
878.0 470 243.33 1472.0 800 426.67 2642.0 1450 787.78 3812.0 2100 1148.9
881.6 472 244.44 1490.0 810 432.22 2660.0 1460 793.33 3830.0 2110 1154.4
885.2 474 245.56 1508.0 820 437.78 2678.0 1470 798.89 3848.0 2120 1160.0
888.8 476 246.67 1526.0 830 443.33 2696.0 1480 804.44 3866.0 2130 1165.6
892.4 478 247.78 1544.0 840 448.89 2714.0 1490 810.00 3884.0 2140 1171.1
896.0 480 248.89 1562.0 850 454.44 2732.0 1500 815.56 3902.0 2150 1176.7
899.6 482 250.00 1580.0 860 460.00 2750.0 1510 821.11 3920.0 2160 1182.2
903.2 484 251.11 1598.0 870 465.56 2768.0 1520 826.67 3938.0 2170 1187.8
906.8 486 252.22 1616.0 880 471.11 2786.0 1530 832.22 3956.0 2180 1193.3
910.4 488 253.33 1634.0 890 476.67 2804.0 1540 837.78 3974.0 2190 1198.9
914.0 490 254.44 1652.0 900 482.22 2822.0 1550 843.33 3992.0 2200 1204.4
917.6 492 255.56 1670.0 910 487.78 2840.0 1560 848.89 4010.0 2210 1210.0
921.2 494 256.67 1688.0 920 493.33 2858.0 1570 854.44 4028.0 2220 1215.6
924.8 496 257.78 1706.0 930 498.89 2876.0 1580 860.00 4046.0 2230 1221.1
928.4 498 258.89 1724.0 940 504.44 2894.0 1590 865.56 4064.0 2240 1226.7
932.0 500 260.00 1742.0 950 510.00 2912.0 1600 871.11 4082.0 2250 1232.2
935.6 502 261.11 1760.0 960 515.56 2930.0 1610 876.67 4100.0 2260 1237.8
939.2 504 262.22 1778.0 970 521.11 2948.0 1620 882.22 4118.0 2270 1243.3
942.8 506 263.33 1796.0 980 526.67 2966.0 1630 887.78 4136.0 2280 1248.9
946.4 508 264.44 1814.0 990 532.22 2984.0 1640 893.33 4154.0 2290 1254.4
950.0 510 265.56 1832.0 1000 537.78 3002.0 1650 898.89 4172.0 2300 1260.0
953.6 512 266.67 1850.0 1010 543.33 3020.0 1660 904.44 4190.0 2310 1265.6
957.2 514 267.78 1868.0 1020 548.89 3038.0 1670 910.00 4208.0 2320 1271.1
960.8 516 268.89 1886.0 1030 554.44 3056.0 1680 915.56 4226.0 2330 1276.7
964.4 518 270.00 1904.0 1040 560.00 3074.0 1690 921.11 4244.0 2340 1282.2
968.0 520 271.11 1922.0 1050 565.56 3092.0 1700 926.67 4262.0 2350 1287.8
971.6 522 272.22 1940.0 1060 571.11 3110.0 1710 932.22 4280.0 2360 1293.3
975.2 524 273.33 1958.0 1070 576.67 3128.0 1720 937.78 4298.0 2370 1298.9
978.8 526 274.44 1976.0 1080 582.22 3146.0 1730 943.33 4316.0 2380 1304.4
982.4 528 275.56 1994.0 1090 587.78 3164.0 1740 948.89 4334.0 2390 1310.1
986.0 530 276.67 2012.0 1100 593.33 3182.0 1750 954.44 4352.0 2400 1315.6
989.6 532 277.78 2030.0 1100 598.89 3200.0 1760 960.00 4370.0 2410 1321.1
993.2 534 278.89 2048.0 1120 604.44 3218.0 1770 965.56 4388.0 2420 1326.7
996.8 536 280.00 2066.0 1130 610.00 3236.0 1780 971.11 4406.6 2430 1332.2
1000.4 538 281.11 2084.0 1140 615.56 3254.0 1790 976.67 4424.0 2440 1337.8
1004.0 540 282.22 2102.0 1150 621.11 3272.0 1800 982.22 4442.0 2450 1343.3
1007.6 542 283.22 2120.0 1160 626.67 3290.0 1810 987.78 4460.0 2460 1348.9
1011.2 544 284.44 2138.0 1170 632.22 3308.0 1820 993.33 4478.0 2470 1354.4
1014.8 546 285.56 2156.0 1180 637.78 3326.0 1830 998.89 4496.0 2480 1360.0
1018.4 548 286.67 2174.0 1190 643.33 3344.0 1840 1004.4 4514.0 2490 1365.6
1022.0 550 287.78 2192.0 1200 648.89 3362.0 1850 1010.0 4532.0 2500 1371.1
(continued)
F C F C F C F C
4550.0 2510 1376.7 5090.0 2810 1543.3 5702.0 3150 1732.2 8402.0 4650 2565.5
4568.0 2520 1382.2 5108.0 2820 1548.9 5792.0 3200 1760.0 8492.0 4700 2593.3
4586.0 2530 1387.8 5126.0 2830 1554.4 5882.0 3250 1787.7 8582.0 4750 2621.1
4604.0 2540 1393.3 5144.0 2840 1560.0 5972.0 3300 1815.5 8672.0 4800 2648.8
4622.0 2550 1398.9 5162.0 2850 1565.6 6062.0 3350 1843.3 8762.0 4850 2676.6
4640.0 2560 1404.4 5180.0 2860 1571.1 6152.0 3400 1871.1 8852.0 4900 2704.4
4658.0 2570 1410.0 5198.0 2870 1576.7 6242.0 3450 1898.8 8942.0 4950 2732.2
4676.0 2580 1415.6 5216.0 2880 1582.2 6332.0 3500 1926.6 9032.0 5000 2760.0
4694.0 2590 1421.1 5234.0 2890 1587.8 6422.0 3550 1954.4 9122.0 5050 2787.7
4712.0 2600 1426.7 5252.0 2900 1593.3 6512.0 3600 1982.2 9212.0 5100 2815.5
4730.0 2610 1432.2 5270.0 2910 1598.9 6602.0 3650 2010.0 9302.0 5150 2843.3
4748.0 2620 1437.8 5288.0 2920 1604.4 6692.0 3700 2037.7 9392.0 5200 2871.1
4766.0 2630 1443.3 5306.0 2930 1610.0 6782.0 3750 2065.5 9482.0 5250 2898.8
4784.0 2640 1448.9 5324.0 2940 1615.6 6872.0 3800 2093.3 9572.0 5300 2926.6
4802.0 2650 1454.4 5342.0 2950 1621.1 6962.0 3850 2121.1 9662.0 5350 2954.4
4820.0 2660 1460.0 5360.0 2960 1626.7 7052.0 3900 2148.8 9752.0 5400 2982.2
4838.0 2670 1465.6 5378.0 2970 1632.2 7142.0 3950 2176.6 9842.0 5450 3010.0
4856.0 2680 1471.1 5396.0 2980 1637.8 7232.0 4000 2204.4 9932.0 5500 3037.7
4874.0 2690 1476.7 5414.0 2990 1643.3 7322.0 4050 2232.2 10022.0 5550 3065.5
4892.0 2700 1482.2 5432.0 3000 1648.9 7412.0 4100 2260.0 10112.0 5600 3093.3
4910.0 2710 1487.8 5450.0 3010 1654.4 7502.0 4150 2287.7
4928.0 2720 1493.3 5468.0 3020 1660.0 7592.0 4200 2315.5
4946.0 2730 1498.9 5486.0 3030 1665.5 7682.0 4250 2343.3
4964.0 2740 1504.4 5504.0 3040 1671.1 7772.0 4300 2371.1
4982.0 2750 1510.0 5522.0 3050 1676.7 7862.0 4350 2398.8
5000.0 2760 1515.6 5540.0 3060 1682.2 7952.0 4400 2426.6
5018.0 2770 1521.1 5558.0 3070 1687.8 8042.0 4450 2454.4
5036.0 2780 1526.7 5576.0 3080 1693.3 8132.0 4500 2482.2
5054.0 2790 1532.2 5594.0 3090 1698.9 8222.0 4550 2510.0
5072.0 2800 1537.8 5612.0 3100 1704.4 8312.7 4600 2537.7
APPENDIX 3
Steel Hardness Conversions

FROM A PRACTICAL STANDPOINT, it is Table A3.1 Examples of published hardness
important to be able to convert the results of conversion equations
one type of hardness test into those of a different Steels
test. Because a hardness test does not measure a 7300
well-defined property of a material and because HB=
1307HRB
(40100 HRB)
all the tests in common use are not based on the 3710
same type of measurements, it is not surprising HB= (30100 HRE)
1307HRE
that universal hardness conversion relationships 1,520,00074500 HRC
HB= (540 HRC)
have not been developed. Hardness conversions (1007HRC)2
instead are empirical relationships that are de- 25,000710(577HRC)2
HB= (4070 HRC)
fined by conversion tables limited to specific 1007HRC
categories of materials. That is, different con- HRB=1347
6700
(+7 HRB, 95% CL)
version tables are required for materials with HB
1=2
greatly different elastic moduli or with different 2:43 106
HRC=119:07 (2401040 HV)
HV
strain-hardening capacity. 1=2
6:85 105
The most reliable hardness-conversion data HRA=112:37 (2401040 HV)
HV
exist for steel that is harder than 240 HB. The 1=2
5:53 105
indentation hardness of soft metals depends on HR15N=117:947
HV
(2401040 HV)
the strain-hardening behavior of the material 1=2
1:88 106
during the test, which in turn depends on the HR30N=129:527
HV
(2401040 HV)
previous degree of strain hardening of the
3:132 106
1=2
material before the test. The modulus of elasti- HR45N=133:517
HV
(2401040 HV)
city also has been shown to influence conver- HB = 0.951 HV (steel ball, 200400 HV)
sions at high hardness levels. At low hardness HB = 0.941 HV (tungsten-carbide ball,
levels, conversions between hardness scales 200700 HV)
measuring depth and those measuring diameter Cemented carbides
1=2
are likewise influenced by differences in the 2:43 106
HRC=117:357 (9001800 HV)
modulus of elasticity. HV
1=2
Hardness conversions are covered in stan- 2:437106
2117
dards such as SAE J417, Hardness Tests and HRA=
HV
(9001800 HV)
Hardness Conversions; ISO 4964, Hardness 1:885
Rockwell from Knoop for steels
ConversionsSteel; and ASTM E140, Stan-
dard Hardness Conversion Tables for Metals. HRC = 64.934 log HK 140.38 (15 gf)
HRC = 67.353 log HK 144.32 (25 gf)
Conversion tables for nickel and high-nickel HRC = 71.983 log HK 154.28 (50 gf)
alloys, cartridge brass, austenitic stainless steel HRC = 76.572 log HK 163.89 (100 gf)
HRC = 79.758 log HK 170.92 (200 gf)
plate and sheet, and copper can be found in HRC = 82.283 log HK 176.92 (300 gf)
ASTM E140. Recently, ASTM committee E-28 HRC = 83.58 log HK 179.30 (500 gf)
on indentation hardness has developed mathe- HRC = 85.848 log HK 184.55 (1000 gf)
matical conversion formulas based on the White cast irons
conversion-table values fround in ASTM E140. HB = 0.363 (HRC)2 22.515
(HRC)+717.8
Over 60 conversion formulas are listed in the HV = 0.343 (HRC)2 18.132
appendix of ASTM E140, and these formulas (HRC)+595.3
can be used in place of the tables. A computer is HV = 1.136 (HB)2 26.0
helpful in performing the calculations quickly. Austenitic stainless steel
Other hardness conversion formulas for var- 1
=0:0001304(1307HRB) (6090 HRB, 110192 HB)
ious materials have also been published, and a HB
Stable alpha-beta titanium alloys
list of some other conversion formulas is given
HRC = 0.078 HV+8.1
in Table A3.1. The standard procedure for
reporting converted hardness numbers from ASTM E140, are for conversion among
indicates the measured hardness and test scale in Rockwell, Brinell, and Vickers hardness for heat
parenthesesfor example, 451 HB (48 HRC). treated carbon and alloy steels, almost all con-
The method of conversion (table, formula, or structional alloy steels, and tool steels in the as-
other method) should also be defined. forged, annealed, normalized, and quenched and
When making hardness correlations, it is best tempered conditions. The tables are also sum-
to consult ASTM E140. Tables A3.2 to A3.5, marized in graphical form in Fig. A3.1.
Table A3.2 Approximate Rockwell B hardness conversion numbers for nonaustenitic steels
Rockwell Superficial Rockwell
Knoop, Brinell, Tensile Brinell,
B, 100 kgf, A, 60 kgf, E, 100 kgf, 15T, 15 kgf, 30T, 30 kgf, 45T, 45 kgf, 500 gf 3000 kgf, strength 500 kgf,
1/16 in. ball diamond 1/8 in. ball 1/16 in. ball 1/16 in. ball 1/16 in. ball Vickers and over 10 mm ball MPa (ksi) 10 mm ball
100 61.5 ... 93.1 83.1 72.9 240 251 240 800 (116) 201
99 60.9 ... 92.8 82.5 71.9 234 246 234 787 (114) 195
98 60.2 ... 92.5 81.8 70.9 228 241 228 752 (109) 189
97 59.5 ... 92.1 81.1 69.9 222 236 222 724 (105) 184
96 58.9 ... 91.8 80.4 68.9 216 231 216 704 (102) 179
95 58.3 ... 91.5 79.8 67.9 210 226 210 690 (100) 175
94 57.6 ... 91.2 79.1 66.9 205 221 205 676 (98) 171
93 57.0 ... 90.8 78.4 65.9 200 216 200 648 (94) 167
92 56.4 ... 90.5 77.8 64.8 195 211 195 634 (92) 163
91 55.8 ... 90.2 77.1 63.8 190 206 190 620 (90) 160
90 55.2 ... 89.9 76.4 62.8 185 201 185 614 (89) 157
89 54.6 ... 89.5 75.8 61.8 180 196 180 607 (88) 154
88 54.0 ... 89.2 75.1 60.8 176 192 176 593 (86) 151
87 53.4 ... 88.9 74.4 59.8 172 188 172 579 (84) 148
86 52.8 ... 88.6 73.8 58.8 169 184 169 572 (83) 145
85 52.3 ... 88.2 73.1 57.8 165 180 165 565 (82) 142
84 51.7 ... 87.9 72.4 56.8 162 176 162 558 (81) 140
83 51.1 ... 87.6 71.8 55.8 159 173 159 552 (80) 137
82 50.6 ... 87.3 71.1 54.8 156 170 156 524 (76) 135
81 50.0 ... 86.9 70.4 53.8 153 167 153 503 (73) 133
80 49.5 ... 86.6 69.7 52.8 150 164 150 496 (72) 130
79 48.9 ... 86.3 69.1 51.8 147 161 147 482 (70) 128
78 48.4 ... 86.0 68.4 50.8 144 158 144 475 (69) 126
77 47.9 ... 85.6 67.7 49.8 141 155 141 469 (68) 124
76 47.3 ... 85.3 67.1 48.8 139 152 139 462 (67) 122
75 46.8 ... 85.0 66.4 47.8 137 150 137 455 (66) 120
74 46.3 ... 84.7 65.7 46.8 135 147 135 448 (65) 118
73 45.8 ... 84.3 65.1 45.8 132 145 132 441 (64) 116
72 45.3 ... 84.0 64.4 44.8 130 143 130 434 (63) 114
71 44.8 100 83.7 63.7 43.8 127 141 127 427 (62) 112
70 44.3 99.5 83.4 63.1 42.8 125 139 125 421 (61) 110
69 43.8 99.0 83.0 62.4 41.8 123 137 123 414 (60) 109
68 43.3 98.0 82.7 61.7 40.8 121 135 121 407 (59) 108
67 42.8 97.5 82.4 61.0 39.8 119 133 119 400 (58) 106
66 42.3 97.0 82.1 60.4 38.7 117 131 117 393 (57) 104
65 41.8 96.0 81.8 59.7 37.7 116 129 116 386 (56) 102
64 41.4 95.5 81.4 59.0 36.7 114 127 114 ... 100
63 40.9 95.0 81.1 58.4 35.7 112 125 112 ... 99
62 40.4 94.5 80.8 57.7 34.7 110 124 110 ... 98
61 40.0 93.5 80.5 57.0 33.7 108 122 108 ... 96
60 39.5 93.0 80.1 56.4 32.7 107 120 107 ... 95
59 39.0 92.5 79.8 55.7 31.7 106 118 106 ... 94
58 38.6 92.0 79.5 55.0 30.7 104 117 104 ... 92
57 38.1 91.0 79.2 54.4 29.7 103 115 103 ... 91
56 37.7 90.5 78.8 53.7 28.7 101 114 101 ... 90
55 37.2 90.0 78.5 53.0 27.7 100 112 100 ... 89
54 36.8 89.5 78.2 52.4 26.7 ... 111 ... ... 87
53 36.3 89.0 77.9 51.7 25.7 ... 110 ... ... 86
52 35.9 88.0 77.5 51.0 24.7 ... 109 ... ... 85
51 35.5 87.5 77.2 50.3 23.7 ... 108 ... ... 84
50 35.0 87.0 76.9 49.7 22.7 ... 107 ... ... 83
49 34.6 86.5 76.6 49.0 21.7 ... 106 ... ... 82
48 34.1 85.5 76.2 48.3 20.7 ... 105 ... ... 81
47 33.7 85.0 75.9 47.7 19.7 ... 104 ... ... 80
46 33.3 84.5 75.6 47.0 18.7 ... 103 ... ... 80
45 32.9 84.0 75.3 46.3 17.7 ... 102 ... ... 79
44 32.4 83.5 74.9 45.7 16.7 ... 101 ... ... 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
Appendix 3: Steel Hardness Conversions / 531
Table A3.3 Approximate Rockwell C hardness conversion numbers for nonaustenitic steels, according
to ASTM E140
C, 150 kgf, A, 60 kgf, D, 100 kgf, 15 N, 15 kgf, 30 N, 30 kgf, 45 N, 45 kgf, Knoop, 500 gf Brinell, 3000 kgf, Tensile strength,
diamond diamond diamond diamond diamond diamond Vickers and over 10 mm ball MPa (ksi)
68 85.6 76.9 93.2 84.4 75.4 940 920 ... ...
67 85.0 76.1 92.9 83.6 74.2 900 895 ... ...
66 84.5 75.4 92.5 82.8 73.3 865 870 ... ...
65 83.9 74.5 92.2 81.9 72.0 832 846 739(a) ...
64 83.4 73.8 91.8 81.1 71.0 800 822 722(a) ...
63 82.8 73.0 91.4 80.1 69.9 772 799 705(a) ...
62 82.3 72.2 91.1 79.3 68.8 746 776 688(a) ...
61 81.8 71.5 90.7 78.4 67.7 720 754 670(a) ...
60 81.2 70.7 90.2 77.5 66.6 697 732 654(a) ...
59 80.7 69.9 89.8 76.6 65.5 674 710 634(a) 2420 (351)
58 80.1 69.2 89.3 75.7 64.3 653 690 615 2330 (338)
57 79.6 68.5 88.9 74.8 63.2 633 670 595 2240 (325)
56 79.0 67.7 88.3 73.9 62.0 613 650 577 2158 (313)
55 78.5 66.9 87.9 73.0 60.9 595 630 560 2075 (301)
54 78.0 66.1 87.4 72.0 59.8 577 612 543 2013 (292)
53 77.4 65.4 86.9 71.2 58.6 560 594 525 1951 (283)
52 76.8 64.6 86.4 70.2 57.4 544 576 512 1882 (273)
51 76.3 63.8 85.9 69.4 56.1 528 558 496 1820 (264)
50 75.9 63.1 85.5 68.5 55.0 513 542 481 1758 (255)
49 75.2 62.1 85.0 67.6 53.8 498 526 469 1696 (246)
48 74.7 61.4 84.5 66.7 52.5 484 510 455 1634 (237)
47 74.1 60.8 83.9 65.8 51.4 471 495 443 1579 (229)
46 73.6 60.0 83.5 64.8 50.3 458 480 432 1524 (221)
45 73.1 59.2 83.0 64.0 49.0 446 466 421 1482 (215)
44 72.5 58.5 82.5 63.1 47.8 434 452 409 1434 (208)
43 72.0 57.7 82.0 62.2 46.7 423 438 400 1386 (201)
42 71.5 56.9 81.5 61.3 45.5 412 426 390 1344 (195)
41 70.9 56.2 80.9 60.4 44.3 402 414 381 1296 (188)
40 70.4 55.4 80.4 59.5 43.1 392 402 371 1254 (182)
39 69.9 54.6 79.9 58.6 41.9 382 391 362 1220 (177)
38 69.4 53.8 79.4 57.7 40.8 372 380 353 1179 (171)
37 68.9 53.1 78.8 56.8 39.6 363 370 344 1137 (166)
36 68.4 52.3 78.3 55.9 38.4 354 360 336 1110 (161)
35 67.9 51.5 77.7 55.0 37.2 345 351 327 1075 (156)
34 67.4 50.8 77.2 54.2 36.1 336 342 319 1048 (152)
33 66.8 50.0 76.6 53.3 34.9 327 334 311 1027 (149)
32 66.3 49.2 76.1 52.1 33.7 318 326 301 1006 (146)
31 65.8 48.4 75.6 51.3 32.5 310 318 294 972 (141)
30 65.3 47.7 75.0 50.4 31.3 302 311 286 951 (138)
29 64.8 47.0 74.5 49.5 30.1 294 304 279 930 (135)
28 64.3 46.1 73.9 48.6 28.9 286 297 271 903 (131)
27 63.8 45.2 73.3 47.7 27.8 279 290 264 882 (128)
26 63.3 44.6 72.8 46.8 26.7 272 284 258 861 (125)
25 62.8 43.8 72.2 45.9 25.5 266 278 253 848 (123)
24 62.4 43.1 71.6 45.0 24.3 260 272 247 820 (119)
23 62.0 42.1 71.0 44.0 23.1 254 266 243 806 (117)
22 61.5 41.6 70.5 43.2 22.0 248 261 237 792 (115)
21 61.0 40.9 69.9 42.3 20.7 243 256 231 772 (112)
20 60.5 40.1 69.4 41.5 19.6 238 251 226 758 (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
Table A3.4 Approximate equivalent hardness numbers for Brinell hardness numbers for steel
Rockwell superficial
Brinell hardness number(a) Rockwell hardness No. hardness No.,
3000 kgf load, 10 mm diamond indenter
A scale, B scale, C scale, D scale,
ball(a)
60 kgf 100 kgf 150 kgf 100 kgf 15 N 30 N 45 N Knoop
Brinell Tungsten- Vickers load, load, load, load, scale, scale, scale, hardness No.,
indentation Standard carbide hardness diamond 1/16 in. diamond diamond 15 kgf 30 kgf 45 kgf 500 gf load Scleroscope
diam, mm ball ball No. indenter diam ball indenter indenter load load load and greater hardness No.
2.25 ... (745) 840 84.1 ... 65.3 74.8 92.3 82.2 72.2 852 91
2.30 ... (712) 783 83.1 ... 63.4 73.4 91.6 80.5 70.4 808 ...
2.35 ... (682) 737 82.2 ... 61.7 72.0 91.0 79.0 68.5 768 84
2.40 ... (653) 697 81.2 ... 60.0 70.7 90.2 77.5 66.5 732 81
2.45 ... 627 667 80.5 ... 58.7 69.7 89.6 76.3 65.1 703 79
2.50 ... 601 640 79.8 ... 57.3 68.7 89.0 75.1 63.5 677 77
2.55 ... 578 615 79.1 ... 56.0 67.7 88.4 73.9 62.1 652 75
2.60 ... 555 591 78.4 ... 54.7 66.7 87.8 72.7 60.6 626 73
2.65 ... 534 569 77.8 ... 53.5 65.8 87.2 71.6 59.2 604 71
2.70 ... 514 547 76.9 ... 52.1 64.7 86.5 70.3 57.6 579 70
2.75 (495) ... 539 76.7 ... 51.6 64.3 86.3 69.9 56.9 571 ...
... 495 528 76.3 ... 51.0 63.8 85.9 69.4 56.1 558 68
2.80 (477) ... 516 75.9 ... 50.3 63.2 85.6 68.7 55.2 545 ...
... 477 508 75.6 ... 49.6 62.7 85.3 68.2 54.5 537 66
2.85 (461) ... 495 75.1 ... 48.8 61.9 84.9 67.4 53.5 523 ...
... 461 491 74.9 ... 48.5 61.7 84.7 67.2 53.2 518 65
2.90 444 ... 474 74.3 ... 47.2 61.0 84.1 66.0 51.7 499 ...
... 444 472 74.2 ... 47.1 60.8 84.0 65.8 51.5 496 63
2.95 429 429 455 73.4 ... 45.7 59.7 83.4 64.6 49.9 476 61
3.00 415 415 440 72.8 ... 44.5 58.8 82.8 63.5 48.4 459 59
3.05 401 401 425 72.0 ... 43.1 57.8 82.0 62.3 46.9 441 58
3.10 388 388 410 71.4 ... 41.8 56.8 81.4 61.1 45.3 423 56
3.15 375 375 396 70.6 ... 40.4 55.7 80.6 59.9 43.6 407 54
3.20 363 363 383 70.0 ... 39.1 54.6 80.0 58.7 42.0 392 52
3.25 352 352 372 69.3 (110.0) 37.9 53.8 79.3 57.6 40.5 379 51
3.30 341 341 360 68.7 (109.0) 36.6 52.8 78.6 56.4 39.1 367 50
3.35 331 331 350 68.1 (108.5) 35.5 51.9 78.0 55.4 37.8 356 48
3.40 321 321 339 67.5 (108.0) 34.3 51.0 77.3 54.3 36.4 345 47
3.45 311 311 328 66.9 (107.5) 33.1 50.0 76.7 53.3 34.4 336 46
3.50 302 302 319 66.3 (107.0) 32.1 49.3 76.1 52.2 33.8 327 45
3.55 293 293 309 65.7 (106.0) 30.9 48.3 75.5 51.2 32.4 318 43
3.60 285 285 301 65.3 (105.5) 29.9 47.6 75.0 50.3 31.2 310 42
3.65 277 277 292 64.6 (104.5) 28.8 46.7 74.4 49.3 29.9 302 41
3.70 269 269 284 64.1 (104.0) 27.6 45.9 73.7 48.3 28.5 294 40
3.75 262 262 276 63.6 (103.0) 26.6 45.0 73.1 47.3 27.3 286 39
3.80 255 255 269 63.0 (102.0) 25.4 44.2 72.5 46.2 26.0 279 38
3.85 248 248 261 62.5 (101.0) 24.2 43.2 71.7 45.1 24.5 272 37
3.90 241 241 253 61.8 100.0 22.8 42.0 70.9 43.9 22.8 265 36
3.95 235 235 247 61.4 99.0 21.7 41.4 70.3 42.9 21.5 259 35
4.00 229 229 241 60.8 98.2 20.5 40.5 69.7 41.9 20.1 253 34
4.05 223 223 234 ... 97.3 (19.0) ... ... ... ... 247 ...
4.10 217 217 228 ... 96.4 (17.7) ... ... ... ... 242 33
4.15 212 212 222 ... 95.5 (16.4) ... ... ... ... 237 32
4.20 207 207 218 ... 94.6 (15.2) ... ... ... ... 232 31
4.25 201 201 212 ... 93.7 (13.8) ... ... ... ... 227 ...
4.30 197 197 207 ... 92.8 (12.7) ... ... ... ... 222 30
4.35 192 192 202 ... 91.9 (11.5) ... ... ... ... 217 29
4.40 187 187 196 ... 90.9 (10.2) ... ... ... ... 212 ...
4.45 183 183 192 ... 90.0 (9.0) ... ... ... ... 207 28
4.50 179 179 188 ... 89.0 (8.0) ... ... ... ... 202 27
4.55 174 174 182 ... 88.0 (6.7) ... ... ... ... 198 ...
4.60 170 170 178 ... 87.0 (5.4) ... ... ... ... 194 26
4.65 167 167 175 ... 86.0 (4.4) ... ... ... ... 190 ...
4.70 163 163 171 ... 85.0 (3.3) ... ... ... ... 186 25
4.75 159 159 167 ... 83.9 (2.0) ... ... ... ... 182 ...
4.80 156 156 163 ... 82.9 (0.9) ... ... ... ... 178 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 quenched-
and-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

Brinell hardness number(a) Rockwell hardness No. hardness No.,
3000 kgf load, 10 mm diamond indenter
ball(a)
60 kgf 100 kgf 150 kgf 100 kgf 15 N 30 N 45 N Knoop
Brinell Tungsten- Vickers load, load, load, load, scale, scale, scale, hardness No.,
indentation Standard carbide hardness diamond 1/16 in. diamond diamond 15 kgf 30 kgf 45 kgf 500 gf load Scleroscope
diam, mm ball ball No. indenter diam ball indenter indenter load load load and greater hardness No.
4.85 152 152 159 ... 81.9 ... ... ... ... ... 174 ...
4.90 149 149 156 ... 80.8 ... ... ... ... ... 170 23
4.95 146 146 153 ... 79.7 ... ... ... ... ... 166 ...
5.00 143 143 150 ... 78.6 ... ... ... ... ... 163 22
5.10 137 137 143 ... 76.4 ... ... ... ... ... 157 21
5.20 131 131 137 ... 74.2 ... ... ... ... ... 151 ...
5.30 126 126 132 ... 72.0 ... ... ... ... ... 145 20
5.40 121 121 127 ... 69.8 ... ... ... ... ... 140 19
5.50 116 116 122 ... 67.6 ... ... ... ... ... 135 18
5.60 111 111 117 ... 65.4 ... ... ... ... ... 131 17
and-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
Brinell hardness No., Rockwell hardness No. (diamond pyramid)
3000 kg load, hardness No.,
10 mm ball diamond indenter
60 kgf 100 kgf 150 kgf 100 kgf Knoop
Vickers Tungsten- load, load, load, load, 15 N scale, 30 N scale, 45 N scale, hardness No., Scleroscope
hardness Standard carbide diamond 1/16 in. diamond diamond 15 kgf 30 kgf 45 kgf 500 gf load hardness
No. ball ball indenter diam ball indenter indenter load load load and greater No.
940 ... ... 85.6 ... 68.0 76.9 93.2 84.4 75.4 920 97
920 ... ... 85.3 ... 67.5 76.5 93.0 84.0 74.8 908 96
900 ... ... 85.0 ... 67.0 76.1 92.9 83.6 74.2 895 95
880 ... (767) 84.7 ... 66.4 75.7 92.7 83.1 73.6 882 93
860 ... (757) 84.4 ... 65.9 75.3 92.5 82.7 73.1 867 92
840 ... (745) 84.1 ... 65.3 74.8 92.3 82.2 72.2 852 91
820 ... (733) 83.8 ... 64.7 74.3 92.1 81.7 71.8 837 90
800 ... (722) 83.4 ... 64.0 73.8 91.8 81.1 71.0 822 88
780 ... (710) 83.0 ... 63.3 73.3 91.5 80.4 70.2 806 87
760 ... (698) 82.6 ... 62.5 72.6 91.2 79.7 69.4 788 86
740 ... (684) 82.2 ... 61.8 72.1 91.0 79.1 68.6 772 84
720 ... (670) 81.8 ... 61.0 71.5 90.7 78.4 67.7 754 83
700 ... (656) 81.3 ... 60.1 70.8 90.3 77.6 66.7 735 81
690 ... (647) 81.1 ... 59.7 70.5 90.1 77.2 66.2 725 ...
680 ... (638) 80.8 ... 59.2 70.1 89.8 76.8 65.7 716 80
670 ... (630) 80.6 ... 58.8 69.8 89.7 76.4 65.3 706 ...
660 ... 620 80.3 ... 58.3 69.4 89.5 75.9 64.7 697 79
650 ... 611 80.0 ... 57.8 69.0 89.2 75.5 64.1 687 78
640 ... 601 79.8 ... 57.3 68.7 89.0 75.1 63.5 677 77
630 ... 591 79.5 ... 56.8 68.3 88.8 74.6 63.0 667 76
620 ... 582 79.2 ... 56.3 67.9 88.5 74.2 62.4 657 75
610 ... 573 78.9 ... 55.7 67.5 88.2 73.6 61.7 646 ...
600 ... 564 78.6 ... 55.2 67.0 88.0 73.2 61.2 636 74
590 ... 554 78.4 ... 54.7 66.7 87.8 72.7 60.5 625 73
580 ... 545 78.0 ... 54.1 66.2 87.5 72.1 59.9 615 72
570 ... 535 77.8 ... 53.6 65.8 87.2 71.7 59.3 604 ...
(continued)
and-tempered conditions; less accurate for cold-worked condition and for austenitic steels. Source: ASTM E140

Brinell hardness No., Rockwell hardness No. (diamond pyramid)
3000 kg load, hardness No.,
10 mm ball diamond indenter
60 kgf 100 kgf 150 kgf 100 kgf Knoop
Vickers Tungsten- load, load, load, load, 15 N scale, 30 N scale, 45 N scale, hardness No., Scleroscope
hardness Standard carbide diamond 1/16 in. diamond diamond 15 kgf 30 kgf 45 kgf 500 gf load hardness
No. ball ball indenter diam ball indenter indenter load load load and greater No.
560 ... 525 77.4 ... 53.0 65.4 86.9 71.2 58.6 594 71
550 (505) 517 77.0 ... 52.3 64.8 86.6 70.5 57.8 583 70
540 (496) 507 76.7 ... 51.7 64.4 86.3 70.0 57.0 572 69
530 (488) 497 76.4 ... 51.1 63.9 86.0 69.5 56.2 561 68
520 (480) 488 76.1 ... 50.5 63.5 85.7 69.0 55.6 550 67
510 (473) 479 75.7 ... 49.8 62.9 85.4 68.3 54.7 539 ...
500 (465) 471 75.3 ... 49.1 62.2 85.0 67.7 53.9 528 66
490 (456) 460 74.9 ... 48.4 61.6 84.7 67.1 53.1 517 65
480 (448) 452 74.5 ... 47.7 61.3 84.3 66.4 52.2 505 64
470 441 442 74.1 ... 46.9 60.7 83.9 65.7 51.3 494 ...
460 433 433 73.6 ... 46.1 60.1 83.6 64.9 50.4 482 62
450 425 425 73.3 ... 45.3 59.4 83.2 64.3 49.4 471 ...
440 415 415 72.8 ... 44.5 58.8 82.8 63.5 48.4 459 59
430 405 405 72.3 ... 43.6 58.2 82.3 62.7 47.4 447 58
420 397 397 71.8 ... 42.7 57.5 81.8 61.9 46.4 435 57
410 388 388 71.4 ... 41.8 56.8 81.4 61.1 45.3 423 56
400 379 379 70.8 ... 40.8 56.0 80.8 60.2 44.1 412 55
390 369 369 70.3 ... 39.8 55.2 80.3 59.3 42.9 400 ...
380 360 360 69.8 (110.0) 38.8 54.4 79.8 58.4 41.7 389 52
370 350 350 69.2 ... 37.7 53.6 79.2 57.4 40.4 378 51
360 341 341 68.7 (109.0) 36.6 52.8 78.6 56.4 39.1 367 50
350 331 331 68.1 ... 35.5 51.9 78.0 55.4 37.8 356 48
340 322 322 67.6 (108.0) 34.4 51.1 77.4 54.4 36.5 346 47
330 313 313 67.0 ... 33.3 50.2 76.8 53.6 35.2 337 46
320 303 303 66.4 (107.0) 32.2 49.4 76.2 52.3 33.9 328 45
310 294 294 65.8 ... 31.0 48.4 75.6 51.3 32.5 318 ...
300 284 284 65.2 (105.5) 29.8 47.5 74.9 50.2 31.1 309 42
295 280 280 64.8 ... 29.2 47.1 74.6 49.7 30.4 305 ...
290 275 275 64.5 (104.5) 28.5 46.5 74.2 49.0 29.5 300 41
285 270 270 64.2 ... 27.8 46.0 73.8 48.4 28.7 296 ...
280 265 265 63.8 (103.5) 27.1 45.3 73.4 47.8 27.9 291 40
275 261 261 63.5 ... 26.4 44.9 73.0 47.2 27.1 286 39
270 256 256 63.1 (102.0) 25.6 44.3 72.6 46.4 26.2 282 38
265 252 252 62.7 ... 24.8 43.7 72.1 45.7 25.2 277 ...
260 247 247 62.4 (101.0) 24.0 43.1 71.6 45.0 24.3 272 37
255 243 243 62.0 ... 23.1 42.2 71.1 44.2 23.2 267 ...
250 238 238 61.6 99.5 22.2 41.7 70.6 43.4 22.2 262 36
245 233 233 61.2 ... 21.3 41.1 70.1 42.5 21.1 258 35
240 228 228 60.7 98.1 20.3 40.3 69.6 41.7 19.9 253 34
230 219 219 ... 96.7 (18.0) ... ... ... ... 243 33
220 209 209 ... 95.0 (15.7) ... ... ... ... 234 32
210 200 200 ... 93.4 (13.4) ... ... ... . . .. 226 30
200 190 190 ... 91.5 (11.0) ... ... ... ... 216 29
190 181 181 ... 89.5 ( 8.5) ... ... ... ... 206 28
180 171 171 ... 87.1 (6.0) ... ... ... ... 196 26
170 162 162 ... 85.0 (3.0) ... ... ... ... 185 25
160 152 152 ... 81.7 (0.0) ... ... ... ... 175 23
150 143 143 ... 78.7 ... ... ... ... ... 164 22
140 133 133 ... 75.0 ... ... ... ... ... 154 21
130 124 124 ... 71.2 ... ... ... ... ... 143 20
120 114 114 ... 66.7 ... ... ... ... ... 133 18
110 105 105 ... 62.3 ... ... ... ... ... 123 ...
100 95 95 ... 56.2 ... ... ... ... ... 112 ...
95 90 90 ... 52.0 ... ... ... ... ... 107 ...
90 86 86 ... 48.0 ... ... ... ... ... 102 ...
85 81 81 ... 41.0 ... ... ... ... ... 97 ...
and-tempered conditions; less accurate for cold-worked condition and for austenitic steels. Source: ASTM E140
1000 1000
900 900
800 800
Vickers hardness
Vickers hardness
700 700
45 N
600 600
500 500
30 N 15 N
400 400
300 300
200 200
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Rockwell C hardness Rockwell superficial hardness number
800 800
700 700
Brinell hardness number
Brinell hardness number
600 600
45 N 30 N
500 500
15 N
400 400
300 300
200 200
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Rockwell C hardness Rockwell superficial hardness number
1000 70
800 60
Rockwell C hardness
Vickers hardness
45 N
600 50
30 N 15 N
400 40
200 30
0 20
0 100 200 300 400 500 600 700 800 900 1000 0 10 20 30 40 50 60 70 80 90 100
Brinell hardness number Rockwell superficial hardness number
Fig. A3.1 Approximate equivalent hardness numbers for steel. Points represent data from the hardness conversion tables.
APPENDIX 4
Austenitizing Temperatures for Steels
TEMPERATURES RECOMMENDED for the furnace atmosphere all influence the rate of
austenitizing carbon and low-alloy steels prior heating of the steel part to the required tem-
to hardening are given in Table A4.1 (for perature selected from Tables A4.1 and A4.2.
direct-hardening grades) and Table A4.2 (for The difference in temperature rise within
carburized steels). Table A4.2 is applicable to thick and thin sections of articles of varying
carburized steels that have been cooled slowly cross section is a major problem in practical
from the carburizing temperature and are to be heat-treating operations. When temperature
furnace hardened in a subsequent operation. uniformity is the ultimate objective of the heat-
For most applications, the rate of heating to ing cycle, this is more safely attained by slowly
the austenitizing temperature is less important heating than by rapidly heating. Furthermore,
than other factors in the hardening process, such the maximum temperature in the austenite range
as maximum temperature attained throughout should not exceed that required to achieve
the section, temperature uniformity, time at the necessary extent of solution of carbide. The
temperature, and rate of cooling. The thermal temperatures listed in Tables A4.1 and A4.2
conductivity of the steel, the nature of the conform with this requirement. When heating
furnace atmosphere (scaling or nonscaling), with significant cross-section variations, provi-
thickness of section, method of loading (spaced sions should be made for slower heating to
or stacked), and the degree of circulation of minimize thermal stresses and distortions.
Table A4.1 Austenitizing temperatures for direct-hardening carbon and alloy steels (SAE)
Temperature Temperature
Steel C F Steel C F
Carbon steels Free-cutting carbon steels
1025 855900 15751650 1137 830855 15251575

1030 845870 15501600 1138 815845 15001550
1035 830855 15251575 1140 815845 15001550
1037 830855 15251575 1141 800845 14751550
1038(a) 830855 15251575 1144 800845 14751550
1039(a) 830855 15251575 1145 800845 14751550
1040(a) 830855 15251575 1146 800845 14751550
1042 800845 14751550 1151 800845 14751550
1043(a) 800845 14751550 1536 815845 15001550
1045(a) 800845 14751550 1541 815845 15001550
1046(a) 800845 14751550 1548 815845 15001550
1050(a) 800845 14751550 1552 815845 15001550
1055 800845 14751550 1566 855885 15751625
1060 800845 14751550
1065 800845 14751550 Alloy steels
1070 800845 14751550
1074 800845 14751550 1330 830855 15251575
1078 790815 14501500 1335 815845 15001550
1080 790815 14501500 1340 815845 15001550
1084 790815 14501500 1345 815845 15001550
1085 790815 14501500 3140 815845 15001550
1086 790815 14501500 4037 830855 15251575
1090 790815 14501500 4042 830855 15251575
1095 790815(a) 14501500(b) 4047 815855 15001575
(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).

Steel C F Steel C F
4063 800845 14751550 5155 800845 14751550
4130 815870 15001600 5160 800845 14751550
4135 845870 15501600 51B60 800845 14751550
4137 845870 15501600 50100 775800(c) 14251475(c)
4140 845870 15501600 51100 775800(c) 14251475(c)
4142 845870 15501600 52100 775800(c) 14251475(c)
4145 815845 15001550 6150 845885 15501625
4147 815845 15001550 81B45 815855 15001575
4150 815845 15001550 8630 830870 15251600
4161 815845 15001550 8637 830855 15251575
4337 815845 15001550 8640 830855 15251575
4340 815845 15001550 8642 815855 15001575
50B40 815845 15001550 8645 815855 15001575
50B44 815845 15001550 86B45 815855 15001575
5046 815845 15001550 8650 815855 15001575
50B46 815845 15001550 8655 800845 14751550
50B50 800845 14751550 8660 800845 14751550
50B60 800845 14751550 8740 830855 15251575
5130 830855 15251575 8742 830855 15251575
5132 830855 15251575 9254 815900 15001650
5135 815845 15001550 9255 815900 15001650
5140 815845 15001550 9260 815900 15001650
5145 815845 15001550 94B30 845885 15501625
5147 800845 14751550 94B40 845885 15501625
5150 800845 14751550 9840 830855 15251575
(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.
Steel C F Steel C F
Carbon steels Alloy steels
1010 760790 14001450 3310 790830 14501525
1012 760790 14001450 4320 830845 15251550
1015 760790 14001450 4615 815845 15001550
1016 760790 14001450 4617 815845 15001550
1017 760790 14001450 4620 815845 15001550
1018 760790 14001450 4621 815845 15001550
1019 760790 14001450 4626 815845 15001550
1020 760790 14001450 4718 815845 15001550
1022 760790 14001450 4720 815845 15001550
1513 760790 14001450 4815 800830 14751525
1518 760790 14001450 4817 800830 14751525
1522 760790 14001450 4820 800830 14751525
1524 760790 14001450 8115 845870 15501600
1525 760790 14001450 8615 845870 15501600
1526 760790 14001450 8617 845870 15501600
1527 760790 14001450 8620 845870 15501600
8622 845870 15501600
Free-cutting carbon steels
8625 845870 15501600
1109 760790 14001450 8627 845870 15501600
1115 760790 14001450 8720 845870 15501600
1117 760790 14001450 8822 845870 15501600
1118 760790 14001450 9310 790830 14501525
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_539-540.pdf/Appendix-5/ 18/8/2008 4:30PM Plate # 0 pg 539
APPENDIX 5
Temper Colors for Steels
IT IS COMMON and long-time knowledge Place a piece of bright copper strip, a bright
that steel heated in contact with air at tempera- carbon steel strip and a bright stainless steel
tures in the tempering range takes on various strip side by side on the belt and send them
temper colors due to the formation of a thin oxide through the furnace.
film. As an example, the 1948 ASM Metals Use the following chart to determine the
Handbook gives the following table: source of oxidation and the color chart to
determine its location:
Heating temperature, F Color
400 Faint straw Sample surface appearance
440 Straw
Copper Carbon steel Stainless steel Conclusions
475 Deep straw
520 Bronze Oxidized Oxidized Oxidized Air leak
540 Peacock Bright Oxidized Oxidized Water leak
590 Full blue Bright Bright Oxidized Very small air
640 Light blue Or water leak
Color Chart (back cover). Another exam- Time-Temperature Effect. A study at the
ple of temper colors is shown on the back cover Illinois Institute of Technology obtained de-
of this book. This example was developed to tailed information on temper colors for plain
illustrate the effect of oxidation with furnace carbon steel, especially on the effect of time and
leaks. Oxidation of steel in the front of a con- temperature. This was obtained using 16 mm
tinuous furnace results in a frosted or matt sur- (5/8 in.) diam hot-rolled bars of a standard SAE
face finish with a loss of surface brightness. 1035 steel. Samples that were 50 mm (2 in.)
Oxidation in the high heat section is evidenced long were cut from the bars and carefully
by a black oxide that is often scaled and flakes machined and cleaned to give smooth, bright
from the surface of the steel. If oxidation is surfaces. They were then heated for various
occurring in the cooling section of a continuous times at several temperatures in air-circulating
furnace, use the following test to determine the furnaces controlled to within +3 C (+5 F)
cause: of the desired temperatures. The results of this
Copper/Steel/Stainless Steel Test. study are shown in Fig. A5.1 for 1035 steel.
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
Fig. A5.1 Temper colors after heating 1035 steel in circulating air (atmospheric pressure)
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 20100 20200 20300 20400 20500 20600 20700
grade Treatment or condition (68212) (68392) (68572) (68752) (68932) (681112) (681292)
1008 Annealed 12.6(b) 13.1(b) 13.5(b) 13.8(b) 14.2(b) 14.6(b) 15.0(b)
1008 Annealed 11.6 12.5 13.0 13.6 14.2 14.6 15.0
1010 Unknown 11.9(c) 12.6 13.3 13.8 14.3 14.7 14.9
1010 Unknown ... ... ... 15.1(d) ... ... ...
1015 Rolled 11.9(b) 12.5(b) 13.0(b) 13.6(b) 14.2(b) ... ...
1015 Annealed 12.2(b) ... ... 13.4(b) ... 14.2(b) ...
1016 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1017 Unknown 12.2(b) ... ... 13.5(b) ... 14.5(b) ...
1018 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1019 Unknown 12.2(b) ... ... 13.5(b) ... 14.7(b) ...
1020 Annealed 11.7 12.1 12.8 13.4 13.9 14.4 14.8
1020 Unknown 12.2(b) ... ... 13.5(b) ... 14.2(b) ...
1021 Unknown 12.0(b) ... ... 13.5(b) ... 14.3(b) ...
1023 Unknown 12.2(b) ... ... 13.5(b) ... 14.4(b) ...
1025 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1026 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1029 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1030 Annealed 11.7(b) ... ... 13.5(b) ... 14.4(b) ...
1035 Annealed 11.1 11.9 12.7 13.4 14.0 14.4 14.8
1037 Annealed 11.1(b) ... ... 13.5(b) ... 14.6(b) ...
1039 Annealed 11.1(b) ... ... 13.5(b) ... 14.6(b) ...
1040 Annealed 11.3 12.0 12.5 13.3 13.9 14.4 14.8
1043 Annealed 11.3(b) ... ... 13.5(b) ... 14.6(b) ...
1044 Annealed 11.1(b) 12.0(b) ... 13.3(b) ... ... ...
1045 Annealed 11.2(e) 11.9(e) 12.7(e) 13.5(e) 14.1(e) 14.5(e) 14.8(e)
1046 Unknown 11.1(b) ... ... 13.5(b) ... ... ...
1050 Annealed 11.1(b) 12.0(b) ... 13.5(b) ... ... ...
1053 Unknown 11.1(b) ... ... 13.5(b) ... ... ...
1055 Annealed 11.0 11.8 12.6 13.4 14.0 14.5 14.8
(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).

AISI-SAE 20100 20200 20300 20400 20500 20600 20700
1064 Unknown 11.1(b) ... ... 13.5(b) ... ... ...
1065 Unknown 11.1(b) ... ... 13.5(b) ... ... ...
1070 Unknown 11.5(b) ... ... 13.3(b) ... ... ...
1078 Unknown 11.3(b) ... ... 13.3 ... ... ...
1080 Annealed 11.0 11.6 12.4 13.2 13.8 14.2 14.7
1080 Unknown 11.7(b) 12.2(b) ... ... ... ... ...
1085 Annealed 11.1(b) 11.7(b) 12.5(b) 13.2(b) 13.6(b) 14.2(b) 14.7
1086 Unknown 11.1(b) ... ... 13.1(b) ... ... ...
1095 Unknown ... ... ... ... ... ... 14.6(d)
1095 Annealed 11.4 ... ... ... ... ... ...
1095 Hardened 13.0(b) ... ... ... ... ... ...
1117 Unknown 12.2(f) ... ... 13.1(g) ... ... ...
1118 Unknown 12.2(f) ... ... 13.3(g) ... ... ...
1132 Unknown 12.6(f) ... ... ... ... ... ...
1137 Unknown 12.8 ... ... ... ... ... ...
1139 Unknown 12.6(f) ... ... ... ... ... ...
1140 Unknown 12.6(f) ... ... ... ... ... ...
1141 Unknown ... 12.6(b) ... ... ... ... ...
1144 Unknown 13.3 ... ... ... ... ... ...
1146 Unknown 12.8 ... ... ... ... ... ...
1151 Unknown ... 12.6(b) ... ... ... ... ...
1330 Unknown 12.0 12.8 13.3 ... ... ... ...
1335 Unknown 12.2 12.8 13.3 ... ... ... ...
1345 Unknown 12.0 12.6 13.3 ... ... ... ...
1522 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1524 Unknown 11.9 12.7 ... 13.9 ... 14.7 ...
1524 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1526 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1541 Annealed 12.0(b) ... ... 13.5(b) ... 14.4(b) ...
1548 Unknown 11.9(b) ... ... 13.3(b) ... 14.6(b) ...
1551 Annealed 11.7(b) ... ... 13.5(b) ... 14.6(b) ...
1552 Unknown 11.1(b) ... ... 13.5(b) ... ... ...
1561 Annealed 11.1(b) ... ... 13.5(b) ... 14.6(b) ...
1566 Annealed 11.5(b) ... ... 13.5(b) ... 14.7(b) ...
2330 Annealed 10.9(e) 11.2(e) 12.1(e) 12.9(e) 13.4(e) 13.8(e) ...
2515 Unknown 10.9(f)(h) ... 12.6(i) ... 13.5(j) ... ...
3120 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
3130 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
3140 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
3150 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
4023 Unknown 11.7(k) ... ... ... ... ... ...
4027 Unknown 11.7(k) ... ... ... ... ... ...
4028 Unknown 11.9 12.4 12.9 ... ... ... ...
4032 Unknown 11.9 12.4 12.9 ... ... ... ...
4042 Unknown 11.9 12.4 12.9 ... ... ... ...
4047 Unknown 11.9 12.4 12.9 ... ... ... ...
4130 Unknown 12.2 ... ... 13.7 ... 14.6 ...
4135 Unknown 11.7 12.2 12.8 ... ... ... ...
4137 Unknown 11.7 12.2 12.8 ... ... ... ...
4140 Oil hardened, tempered 12.3 12.7 ... 13.7 ... 14.5 ...
4142 Unknown 11.7 12.2 12.8 ... ... ... ...
4145 Oil hardened, tempered 11.7 12.2 12.8 ... ... ... ...
4147 Unknown 11.7 12.2 12.8 ... ... ... ...
4161 Unknown 11.5 12.2 12.9 ... ... ... ...
4320 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
4337 Unknown 11.3(k) ... ... ... ... ... 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: Physical Properties of Carbon and Low-Alloy Steels / 543

AISI-SAE 20100 20200 20300 20400 20500 20600 20700
4340 Oil hardened, tempered 12.3 12.7 ... 13.7 ... 14.5 ...
600 C(1110 F)
4340 Oil hardened, tempered . . .(m) 12.4 ... 13.6 ... 14.3 ...
630 C (1170 F)
4422 Unknown 11.7(k) ... ... ... ... ... ...
4427 Unknown 12.6 ... 13.8 ... ... 15.1 ...
4615 Unknown 11.5 12.1 12.7 13.2 13.7 14.1 ...
4617 Carburized and hardened 12.5 13.1 ... ... ... ... ...
4626 Normalized and tempared 11.70(f) ... 12.6(i) ... ... 13.8(j) ...
4718 Unknown 11.3 12.2 13.1 ... ... ... ...
4815 Unknown 11.5(f) 12.2(n) 13.1(o) ... ... ... ...
4820 Unknown 11.3(f) 12.2(n) 12.9(o) ... ... ... ...
5046 Unknown 11.9 12.4 12.9 ... ... ... ...
50B60 Unknown 11.9 12.4 12.9 ... ... ... ...
5117 Unknown 12.0 12.8 13.5 ... ... ... ...
5120 Unknown 12.0 12.8 13.5 ... ... ... ...
5130 Unknown 12.2 12.9 13.5 ... ... ... ...
5132 Unknown 12.2 12.9 13.5 ... ... ... ...
5135 Unknown 12.0 12.8 13.5 ... ... ... ...
5140 Annealed ... 12.6 13.4 13.9 14.3 14.6 15.0
5145 Unknown 12.2 ... 13.1( p) ... ... ... ...
5150 Unknown 12.8 ... 13.7( p) ... ... ... ...
5155 Unknown 12.2 ... 13.1(q) ... ... ... ...
52100 Annealed 11.9(b) ... ... ... ... ... ...
52100 Hardened 12.6(b) ... ... ... ... ... ...
6150 Annealed 12.2 12.7 12.3 13.7 14.1 14.4 ...
6150 Hardened, tempered 12.0 12.5 12.9 13.0 13.3 13.7 ...
205 C (400 F)
615 0 Annealed 12.4(e) 12.6(e) 13.3(e) 13.8(e) 14.2(e) 14.5(e) 14.7(e)
8115 Unknown 11.9 12.6 13.3 ... ... ... ...
81B45 Unknown 11.9 12.6 13.3 ... ... ... ...
8617 Unknown 11.9(k) ... 12.8(r) ... ... 14.0(s) ...
8622 Unknown 11.1 12.2 12.9 ... ... ... ...
8625 Unknown 11.1 12.2 12.9 ... ... ... ...
8627 Unknown 11.3 12.2 12.9 ... ... ... ...
8630 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
8637 Unknown 11.3 12.2 12.8 ... ... ... ...
8720 Unknown 11.3(k) ... ... ... ... ... 14.6(l)
8822 Unknown 11.3 12.2 12.9 ... ... ... ...
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).
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 Material
2.63.3 1.41.8 Pure Silicon (Si) 1112 6.16.9 Shock-resisting tool steel
2.26.1 1.23.4 Pure Osmium (Os) 12 6.5 Structural steel
4.54.6 2.52.6 Pure Tungsten (W) 1113 5.97.1 Air hardening medium-alloy cold work
0.68.7 0.34.8 Iron-cobalt-nickel alloys tool steel
4.85.1 2.72.8 Pure Molybdenum (Mo) 1113 6.27.0 High manganese carbon steel
5.6 3.1 Pure Arsenic (As) 1014 5.67.6 Malleable cast iron
6.0 3.3 Pure Germanium (Ge) 12 6.6 Mold tool steel
6.1 3.4 Pure Hafnium (Hf) 8.815 4.98.4 Nonresulfurized carbon steel
5.77.0 3.23.9 Pure Zirconium (Zr) 1114 5.97.5 Chromium molybdenum alloy steel
6.36.6 3.53.7 Pure Cerium (Ce) 9.415 5.28.2 Chromium alloy steel
6.26.7 3.43.7 Pure Rhenium (Re) 1213 6.57.0 Molybdenum/molybdenum sulfide alloy
6.5 3.6 Pure Tantalum (Ta) steel
4.98.2 2.74.6 Pure Chromium (Cr) 12 6.8 Chromium Vanadium alloy steel
6.8 3.8 Pure Iridium (Ir) 1114 5.97.6 Cold work tool steel
2.012 1.16.7 Magnetically soft iron alloys 1114 6.07.5 Ductile medium-silicon cast iron
7.1 3.9 Pure Technetium (Tc) 7.617 4.29.4 Nickel with chromium and/or iron,
7.27.3 4.04.1 Pure Niobium (Nb) molybdenum
5.19.6 2.85.3 Pure Ruthenium (Ru) 1114 6.27.5 Resulfurized carbon steel
4.511 2.56.2 Pure Praseodymium (Pr) 1213 6.47.4 High strength low-alloy steel (HSLA)
7.19.7 3.95.4 Beta and near beta titanium 4.820 2.711 Pure Lutetium (Lu)
8.38.5 4.64.7 Pure Rhodium (Rh) 1015 5.68.3 Duplex stainless steel
8.38.4 4.64.7 Purr Vanadium (V) 9.913 5.57.3 High strength structural steel
5.511 3.16.3 Zirconium alloys 9.016 5.08.9 Pure Promethium (Pm)
8.48.6 4.74.8 Pure Titanium (Ti) 1213 6.57.4 Pure Iron (Fe)
8.68.7 4.84.8 Mischmetal 1114 5.98.0 Metal matrix composite aluminum
7.69.9 4.25.5 Unalloyed or low-alloy titanium 1015 5.68.6 Cobalt alloys (including Stellite)
7.710 4.35.7 Alpha beta titanium 6.020 3.311 Pure Yttrium (Y)
4.014 2.27.8 Molybdenum alloys 1115 6.08.5 Gray cast iron
8.89.1 4.95.1 Pure Platinum (Pt) 9.017 5.09.6 Precipitation hardening stainless steel
7.611 4.25.9 Alpha and neat alpha titanium 13 7.4 Pure Bismuth (Bi)
9.39.6 5.25.3 High-chromium gray cast iron 7.020 3.911 Pure Holmium (Ho)
9.39.9 5.25.5 Ductile high-chromium cast iron 1116 6.18.6 Nickel copper
9.110 5.15.6 Pure Gadolinium (Gd) 13 7.4 Pure Nickel (Ni)
8.411 4.76.3 Pure Antimony (Sb) 14 7.5 Palladium alloys
8.611 4.86.3 Maraging steel 1214 6.87.7 Pure Cobalt (Co)
9.9 5.5 Protactinium (Pa) 1017 5.69.6 Cast austenitic stainless steel
9.810 5.45.8 Water-hardening tool steel 1315 7.08.2 Gold alloys
1011 5.65.9 Molybdenum high-speed too steel 8.119 4.511 High-nickel gray cast iron
6.814 3.87.8 Niobium alloys 14 7.8 Bismuth tin alloys
9.312 5.26.5 Ferritic stainless steel 7.020 3.911 Pure Uranium (U)
7.614 4.27.5 Pure Neodymium (Nd) 14 7.8 Pure Gold (Au)
11 5.9 Cast ferritic stainless steel 1019 5.311 Pure Samarium (Sm)
8.912 4.96.9 Hot work tool steel 7.921 4.412 Pure Erbium (Er)
9.512 5.36.6 Martensitic stainless steel 1316 7.09.0 Nickel chromium silicon gray cast iron
9.912 5.56.5 Cast martensitic stainless steel 14 7.8 Tungsten alloys
11 6.1 Cermet 1415 7.78.4 Beryllium alloys
1012 5.66.6 Ductile silicon-molybdenum cast iron 1218 6.710 Manganese alloy steel
1012 5.66.5 Iron carbon alloys 1020 5.611 Iron alloys
9.312 5.26.9 Pure Terbium (Tb) 9.719 5.411 Proprietary alloy steel
9.813 5.46.9 Cobalt chromium nickel tungsten 15 8.5 White cast iron
1012 5.86.7 High-carbon high-chromium cold work 1219 6.710 Austenitic cast iron with graphite
tool steel 8.822 4.912 Pure Thulium (Tm)
11 6.2 Tungsten high-speed tool steel 1418 7.59.8 Wrought copper nickel
8.514 4.77.8 Commercially pure or low-alloy nickel 1319 7.010 Ductile high-nickel cast iron
11 6.3 Low-alloy special purpose tool steel 4.527 2.515 Pure Lanthanum (La)
7.116 3.98.7 Pure Dysprosium (Dy) 1618 8.810 Wrought high copper alloys
9.313 5.27.2 Nickel molybdenum alloy steel 17 9.4 Cast high copper alloys
1112 6.16.6 Pure Palladium (Pd) 1519 8.311 Wrought bronze
11 6.3 Pure Thorium (Th) 1718 9.29.8 Cast copper
11 6.4 Wrought iron 1618 9.110 Wrought copper
1013 5.77.0 Oil-hardening cold work tool steel 17 9.6 Cast copper nickel silver
7.615 4.28.5 Pure Scandium (Sc) 9.825 5.414 Austenitic stainless steel
1112 6.16.8 Pure Beryllium (Be) 1619 8.911 Cast bronze
6.317 3.59.4 Carbide 1619 8.911 Wrought copper nickel silver
1013 5.77.3 Nickel chromium molybdenum alloy steel 18 10 Pure Barium (Ba)
(continued)
Source: Thermal Properties of Metals and Alloys, ASM, 2002


CTE CTE
10 6/K 10 6/F Material 10 6/K 10 6/F Material
18 10 Cast copper nickel 2526 1415 Cast magnesium aluminum zinc
18 10 Pure Tellurium (Te) 2526 1415 Pure Magnesium (Mg)
1820 9.911 Silver alloys 2526 1415 Wrought magnesium aluminum zinc
19 11 Pure Silver (Ag) 2526 1415 Cast magnesium aluminum manganese
1721 9.412 Wrought brass 26 15 Cast magnesium rare earth
1623 8.913 3xx.x series cast aluminum 1736 9.220 Commercially pure tin
silicon+copper or magnesium 2527 1415 Commercially pure magnesium
1624 8.913 2xxx series wrought aluminum copper 26 15 Pure Ytterbium (Yb)
1624 8.913 Zinc copper titanium alloys 2033 1118 Pure Indium (In)
1624 9.113 6xxx series wrought aluminum magnesium 2528 1416 Lead tin solder
silicon 2729 1516 Commercially pure or low-alloyed lead
20 11 Pure Strontium (Sr) 28 16 Tin silver
2021 1112 Cast brass 2324 1319 9xx.x series cast aluminum plus other
1824 1013 1xx.x series commercially pure cast elements
aluminum 29 16 Pure Lead (Pb)
2022 1112 4xx.x series cast aluminum silicon 2830 1617 Pure Thallium (Tl)
1923 1113 2xx.x series cast aluminum copper 2632 1418 Magnesium alloys
1232 6.418 Pure Gallium (Ga) 2040 1112 Unalloyed or low-alloy zinc
22 12 Manganese (Mn) 2240 1222 5xxx series wrought aluminum magnesium
1222 6.912 4xxx series wrought aluminum silicon 3032 1718 Pure Cadmium (Cd)
22 12 Pure Calcium (Ca) 3335 1819 Zinc copper
2124 1213 7xxx series wrought aluminum zinc 35 19 Pure Europium (Eu)
2224 1213 3xxx series wrought aluminum manganese 3749 2127 Pure Selenium (Se)
23 13 8xx.x series cast aluminum tin 56 31 Pure Lithium (Li)
23 13 Unalloyed aluminum ingot 64 36 Pure Sulfur (S)
2225 1214 1xxx series commercially pure wrought 6971 3839 Pure Sodium (Na)
aluminum 83 46 Pure potassium (K)
2325 1314 5xx.x series cast aluminum magnesium 90 50 Pure Rubidium (Rb)
2425 1314 7xx.x series cast aluminum zinc 14203 7.8113 Pure Plutonium (Pu)
2129 1216 Tin lead 125 70 Pure Phosphorus (P)
2228 1215 Zinc aluminum 97291 54162 Pure Cesium (Cs)
2327 1315 Zinc copper aluminum
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
Thermal conductivity Thermal conductivity
Btu/ Btu/
W/m K (h ft F) Material W/m K (h ft F) Material
23 12 Pure Tellurium (Te) 1315 79 Cobalt chromium tungsten alloys
0.25 0.13 Pure Selenium (Se) 14 8 Pure Europium (Eu)
3 2 Pure Sulfur (S) 1416 89 Manganese alloy steel
5 3 Pure Technetium (Tc) 25 15 Cast martensitic stainless steel
58 35 Pure Plutonium (Pu) 15 8 Pure Erbium (Er)
8 5 Permanent magnet iron alloy 15 9 Pure Promethium (Pm)
8 5 Pure Manganese (Mn) 1121 612 Austenitic stainless steel
611 37 Beta and near beta titanium 16 9 Pure Scandium (Sc)
811 47 Pure Bismuth (Bi) 16 9 Pure Gallium (Ga)
612 37 Alpha beta titanium 16 9 Pure Holmium (Ho)
810 56 Pure Mercury (Hg) 16 10 Pure Lutetium (Lu)
614 48 Cobalt chromium nickel tungsten alloys 133 0.619 Cobalt alloys
417 210 Alpha and near alpha titanium 17 10 Pure Neodymium (Nd)
11 6 Pure Gadolinium (Gd) 17 10 Pure Thulium (Tm)
11 6 Pure Dysprosium (Dy) 1224 714 Cast austenitic stainless steel
11 6 Pure Terbium (Tb) 18 11 Pure Barium (Ba)
11 7 Pure Cerium (Ce) 1522 913 Pure Titanium (Ti)
1113 67 Cermet 1522 913 Low-alloy titanium
915 59 Nickel molybdenum alloys 33 19 Pure Rubidium (Rb)
13 7 Austenitic cast iron with graphite 1128 616 Uranium alloys
13 7 Pure Praseodymium (Pr) 835 520 Nickel with chromium, iron, molybdenum
13 8 Mischmetal 1330 717 Duplex stainless steel
13 8 Pure Lanthanum (La) 22 13 Cast ferritic stainless steel
13 8 Pure Samarium (Sm) 1925 1115 Pure Rhenium (Re)
1116 610 Ferritic stainless steel 2025 1115 Maraging steel
1017 610 Pure Yttrium (Y) 1431 818 Martensitic stainless steel

Thermal conductivity Thermal conductivity
Btu/ Btu/
W/m K (h ft F) Material W/m K (h ft F) Material
1332 819 Precipitation hardening stainless steel 77 44 Pure Thorium (Th)
2224 1314 Pure Hafnium (Hf) 35121 2070 Carbide based material
1927 1116 Pure Zirconium (Zr) 47110 2764 Wrought magnesium aluminum zinc
24 14 Pure Indium (In) 80 47 Pure Indium (In)
840 523 Zirconium alloys 51113 3065 Cast rare earth magnesium alloys
1742 1024 Hot work tool steel 7292 4253 Pure Nickel (Ni)
2229 1317 Cast nickel-silver copper alloys 6999 4057 Pure Cobalt (Co)
1734 1020 Pure Antimony (Sb) 87 50 Pure Osmium (Os)
2230 1317 White cast iron 88 51 Pure Rhodium (Rh)
2032 1219 Cold work tool steel 88 50 Pure Gallium (Ga)
2429 1417 Shock-resisting tool steel 8791 5053 Pure Chromium (Ch)
1836 1121 Pure Cesium (Cs) 8497 4956 Pure Cadmium (Cd)
1342 824 Ductile cast iron 56142 3282 Magnesium alloys
2530 1517 Pure Uranium (U) 65138 3880 Tungsten alloys
2036 1221 High-speed tool steel 96108 5697 (2xx.x series) Cast aluminum copper
2434 1420 Lead antimony 100108 5863 Potassium (K)
1050 622 Nickel steel 25189 14109 Cast brass
1645 926 Nickel copper alloys 21195 12115 Wrought copper nickel
2345 1326 Cast copper-nickel 108113 6265 Low-alloy zinc
34 20 Low-alloy lead 105125 6172 Pure Zinc (Zn)
1355 732 Magnetically soft iron alloy 105125 6172 Zinc aluminum
3435 1920 Pure Lead (Pb) 116 67 Pure Ruhenium (Ru)
2941 1724 Pure Gallium (Ga) 88146 5185 (5xx.x series) Cast aluminum magnesium
2350 1328 Chromium alloy steel 125126 7273 Pure Calcium (Ca)
1163 637 Specialty stainless steel 26230 15133 Wrought bronze
38 22 Pure Thorium (Th) 100159 5892 (7xx.x series) Cast aluminum zinc
39 22 Pure Ytterbium (Yb) 92168 5397 (3xx.x series) Cast aluminum silicon plus
3643 2125 Chromium molybdenum alloy steel copper or magnesium
40 23 Tin antimony 117146 6885 Wrought magnesium zinc
2754 1631 Tantalum alloys 131138 7680 Pure Sodium (Na)
3557 2033 Lead tin 133138 7780 Cast magnesium zinc
2558 1433 Ultrahigh strength steel 138155 8090 (4xxx series) Wrought aluminum silicon
3748 2128 Molybdenum alloy steel 121163 7094 (4xx.x series) Cast aluminum silicon
3154 1831 Pure Vanadium (V) 141147 8285 Pure Iridium (Ir)
3946 2327 Pure Thallium (Tl) 134155 7790 Molybdenum (Mo)
3454 2029 Nickel chromium molybdenum alloy steel 138156 8090 Pure Silicon (Si)
44 26 Nickel molybdenum alloy steel 145151 8487 Pure Rhodium (Rh)
883 548 Nickel specialty alloys 130178 75103 Pure Tungsten (W)
1281 747 Cast bronze 112205 65118 (5xxx series) Wrought aluminum
4657 2733 Tin lead alloys magnesium
47 27 Pure Protactinium (Pa) 153167 8897 Pure Magnesium (Mg)
4550 2629 Niobium alloys 109212 63123 Beryllium alloys
4650 2729 High-manganese carbon steel 100234 58135 Wrought brass
4752 2730 Resulfurized carbon steel 154193 89112 (3xxx series) Wrought aluminum
2970 1740 Niobium alloys manganese
2080 1246 Gray cast iron 121227 70131 (7xxx series) Wrought aluminum zinc
5051 2930 High-strength low-alloy steel 167184 97106 (8xx.x series) Cast aluminum tin
2972 1742 Mold steel 146210 85121 Pure Beryllium (Be)
51 30 Malleable cast iron 112251 65145 (2xxx series) Wrought aluminum copper
1886 1150 Low-alloy nickel 85315 49182 Cast high copper alloys
5354 3031 Pure Niobium (Nb) 167244 97141 (1xxx series) Commercially pure wrought
3871 2241 Nonresulfurized carbon steel aluminum
5257 3033 Pure Tantalum (Ta) 142284 82164 (6xxx series) Wrought aluminum
5457 3133 Tin silver alloys magnesium silicon
58 34 Pure Rubidium (Rb) 209222 121128 (1xx.x series) Commercially pure cast
4275 2443 Cast magnesium aluminum manganese aluminum
4475 2643 Cast magnesium aluminum zinc 221247 128143 Pure Aluminum (Al)
4871 2841 Pure Rhenium (Re) 80410 46237 High copper alloys
5665 3238 Pure Germanium (Ge) 117431 21249 Molybdenum alloys
5274 3043 Pure Tin (Sn) 194391 112226 Wrought coppers
67 39 Pure Chromium (Cr) 292316 169182 Pure Gold (Au)
6076 3544 Pure Palladium (Pd) 346391 200226 Cast copper
21116 1267 Wrought nickel-silver copper alloys 360413 208239 Silver alloys
6973 4042 Pure Platinum (Pt) 417428 241247 Pure Silver (Ag)
7176 4144 Pure Lithium (Li) 415493 240285 Silver tungsten
6880 3947 Pure Iron (Fe) 633716 366414 Silver tungsten carbide
Table A6.4 Thermal conductivities of carbon and low-alloy steels

Conductivity, W/m K, atC (F) (a)
AISI-SAE Treatment or 0 100 200 300 400 500 600 700 800 1000 1200
grade condition (32) (212) (392) (572) (752) (932) (1112) (1292) (1472) (1832) (2192)
1008 Unknown 59.5 57.8 53.2 49.4 45.6 41.0 36.8 33.1 28.5 27.6 29.7
1008 Annealed 65.3(b) 60.3 54.9 ... 45.2 ... 36.4 ... 28.5 27.6 ...
1010 Unknown ... 46.7 ... ... ... ... ... ... ... ... ...
1015 Annealed 51.9 51.0 48.9 ... ... ... ... ... ... ... ...
1016 Annealed 51.9 50.2 47.6 ... ... ... ... ... ... ... ...
1018 Annealed 51.9 50.8 48.9 ... ... ... ... ... ... ... ...
1020 Unknown 51.9 51.0 48.9 ... ... ... ... ... ... ... ...
1022 Annealed 51.9 50.8 48.8 ... ... ... ... ... ... ... ...
1025 Annealed 51.9 51.1 49.0 46.1 42.7 39.4 35.6 31.8 26.0 27.2 29.7
1026 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1029 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1030 Annealed ... 51.0 ... ... ... ... ... ... ... ... ...
1035 Annealed ... 50.8 ... ... ... ... ... ... ... ... ...
1037 Annealed ... 51.0 ... ... ... ... ... ... ... ... ...
1039 Annealed ... 50.7 ... ... ... ... ... ... ... ... ...
1040 Annealed ... 50.7 ... ... ... ... ... ... ... ... ...
1042 Annealed 51.9 50.7 48.2 45.6 41.9 38.1 33.9 30.1 24.7 26.8 29.7
1043 Annealed ... 50.8 ... ... ... ... ... ... ... ... ...
1044 Annealed ... 50.8 ... ... ... ... ... ... ... ... ...
1045 Annealed ... 50.8 ... ... ... ... ... ... ... ... ...
1046 Unknown 51.2 49.7 ... ... ... ... ... ... ... ... ...
1050 Annealed 51.2 49.7 46.8 ... ... ... ... ... ... ... ...
1055 Unknown 51.2 49.7 ... ... ... ... ... ... ... ... ...
1060 Unknown 50.5 ... 46.8 ... ... ... ... ... ... ... ...
1064 Unknown 51.2 49.7 ... ... ... ... ... ... ... ... ...
1070 Unknown 49.9 48.4 ... ... ... ... ... ... ... ... ...
1078 Annealed 47.8 48.2 45.2 41.4 38.1 35.2 32.7 30.1 24.3 26.8 30.1
1078 Unknown 49.6 48.1 ... ... ... ... ... ... ... ... ...
1080 Unknown 50.5 ... 46.8 ... ... ... ... ... ... ... ...
1086 Unknown 49.9 48.4 ... ... ... ... ... ... ... ... ...
1095 Unknown ... 46.7 ... ... ... ... ... ... ... ... ...
1117 Unknown 51.9(b) ... ... ... ... ... ... ... ... ... ...
1118 Unknown 51.5(b) ... ... ... ... ... ... ... ... ... ...
1141 Unknown ... 50.5 47.6 ... ... ... ... ... ... ... ...
1151 Unknown ... 50.5 47.6 ... ... ... ... ... ... ... ...
1522 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1524 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1526 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1541 Annealed 51.9 50.1 48.4 ... ... ... ... ... ... ... ...
1548 Unknown 50.5 49.0 48.3 ... ... ... ... ... ... ... ...
1551 Annealed 50.7 49.3 48.4 ... ... ... ... ... ... ... ...
1561 Annealed 51.2 49.7 ... ... ... ... ... ... ... ... ...
1566 Annealed 51.2 49.7 ... ... ... ... ... ... ... ... ...
2515 Unknown 34.3(b) ... ... 38.2(c) ... ... ... ... ... ... ...
4037 Hardened and ... 48.2 45.6 ... 39.4 ... 33.9 ... ... ... ...
tempered
4130 Hardened and ... 42.7 ... 40.6 ... 37.3 ... 31.0 ... 28.1 30.1
tempered
4140 Hardened and ... 42.7 42.3 ... 37.7 ... 33.1 ... ... ... ...
tempered
4145 Hardened and 41.8(b) ... ... ... ... ... ... ... ... ... ...
tempered
4161 Hardened and 42.7(b) ... ... ... ... ... ... ... ... ... ...
tempered
4427 Unknown 36.8(b) ... ... ... ... ... ... ... ... ... ...
4626 Unknown ... 44.1 ... ... ... ... ... ... ... ... ...
5132 Unknown 48.6 46.5 44.4 42.3 38.5 35.6 31.8 28.9 26.0 28.1 30.1
5140 Hardened and ... 44.8 43.5 ... 37.7 ... 31.4 ... ... ... ...
tempered
8617 Unknown ... 43.3 ... ... ... ... ... ... ... ... ...
8622 Unknown ... 37.5(d) ... ... ... ... ... ... ... ... ...
8627 Unknown ... 37.5(d) ... ... ... ... ... ... ... ... ...
8637 Unknown ... 37.5(d) ... ... ... ... ... ... ... ... ...
8822 Unknown ... 37.5(d) ... ... ... ... ... ... ... ... ...
(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)
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
Specific heat capacity Specific heat capacity
J/kg K Btu/lb F) Material J/kg K Btu/lb F) Material
113117 0.0270.028 Pure Thorium (Th) 337376 0.0800.090 Cast beryllium copper nickel alloys
92142 0.0220.034 Tantalum alloys 335390 0.0800.093 Cast brass
116117 0.028 Pure Uranium (U) 377 0.090 Cast copper-nickels
104140 0.0250.033 Pure Mercury (Hg) 377.0 0.090 Cast nickel-silvers
121129 0.0290.031 Pure Bismuth (Bi) 375398 0.0900.095 Wrought copper-nickels
126130 0.0300.031 Pure Gold (Au) 374425 0.0890.102 Cobalt chromium nickel tungsten
112133 0.0270.032 Pure Lead (Pb) alloys
130132 0.0310.032 Pure Iridium (Ir) 360420 0.0860.100 Wrought high-copper alloys
130131 0.031 Pure Osmium (Os) 387393 0.0920.094 Pure Copper (Cu)
129134 0.0310.032 Low alloyed lead 317462 0.0760.110 Pure Selenium (Se)
128138 0.0310.033 Commercial gold 360420 0.0860.100 Wrought bronze
136 0.032 Pure Platinum (Pt) 372409 0.0890.098 Gallium (Ga)
134142 0.0320.034 Pure Tungsten (W) 382400 0.0910.096 Pure Zinc (Zn)
137140 0.033 Pure Rhenium (Re) 280502 0.0670.120 Zirconium alloys
111167 00270.040 Uranium alloys 371418 0.0890.100 Wrought brass
130150 0.0310.036 Pure Thallium (Tl) 398 0.095 Beryllium copper
134151 0.0320.036 Tungsten alloys 377419 0.0900.100 Wrought nickel silvers
130159 0.0310.038 Lead antimony alloys 377420 0.0900.100 Wrought coppers
131159 0.0310.038 Pure Plutonium (Pu) 383418 0.0910.100 Unalloyed or low-alloy zinc
144148 0.0340.035 Pure Hafnium (Hf) 377439 0.0900.105 Nickel molybdenum alloys
147 0.035 Pure Tantalum (Ta) 389420 0.0930.100 Cast high copper alloys
150 0.036 Pure Lutetium (Lu) 335481 0.0800.115 Nickel molybdenum alloy steel
154 0.037 Pure Ytterbium (Yb) 376440 0.0900.105 Cast bronze
160 0.038 Pure Thulium (Tm) 352473 0.0840.113 Cobalt alloys
162 0.039 Pure Cerium (Ce) 285545 0.0680.130 Nickel with chromium and/or iron,
165 0.039 Pure Holmium (Ho) molybdenum
168 0.040 Pure Erbium (Er) 335502 0.0800.120 Precipitation hardening stainless steel
171 0.041 Pure Dysprosium (Dy) 418421 0.1000.101 Cobalt chromium tungsten alloys
182 0.043 Pure Europium (Eu) 377565 0.0900.135 Austenitic stainless steel
182 0.043 Pure Terbium (Tb) 414452 0.0990.108 Pure Cobalt (Co)
188 0.045 Pure Promethium (Pm) 368502 0.0880.120 Nickel steel
190 0.045 Pure Neodymium (Nd) 439 0.105 Chrome-nickel-iron superalloy
193195 0.0460.047 Mischmetal 420460 0.1000.110 Air-hardening medium-alloy cold
150239 0.0360.057 Tin lead alloys work tool steel
195 0.047 Pure Lanthanum (La) 448 0.107 Ultrahigh-strength steel
195 0.047 Pure Praseodymium (Pr) 420481 0.1000.115 Maraging high strength steel
196 0.047 Pure Samarium (Sm) 431477 0.1030.114 Martensitic stainless steel
201 0.048 Pure Tellurium (Te) 456 0.109 Molybdenum alloy steel
204213 0.0490.051 Pure Antimony (Sb) 393525 0.0940.125 Zinc aluminum alloys
219 0.052 Tin silver alloys 444473 0.1060.113 Pure Iron (Fe)
213239 0.0510.057 Lead-tin alloys 460 0.110 Cast martensitic stainless steel
205238 0.0490.057 Pure Tin (Sn) 460 0.110 Mold steel
205247 0.0490.059 Commercially pure tin 460 0.110 High-carbon high-chromium cold
230 0.055 Cadmium work tool steel
202267 0.0480.064 Pure Cesium (Cs) 460 0.110 Molybdenum high-speed tool steel
236 0.056 Pure Gadolinium (Gd) 460 0.110 Gray cast iron
201272 0.0480.065 Niobium alloys 460 0.110 Cold work tool steel
235239 0.0560.057 Pure Silver (Ag) 460 0.110 Water-hardening tool steel
238 0.057 Pure Ruthenium (Ru) 420502 0.1000.120 Nonresulfurized carbon steel
201280 0.0480.067 Molybdenum alloys 460462 0.110 Hot work tool steel
236246 0.0560.059 Pure Palladium (Pd) 453473 0.1080.113 Pure Nickel (Ni)
234251 0.0560.060 Commercial silver 442487 0.1060.116 Pure Chromium (Cr)
244248 0.0580.059 Commercial palladium 440494 0.1050.118 Chromium alloy steel
229264 0.0550.063 Pure Indium (In) 470 0.112 Permanent magnet iron alloys
233264 0.0560.063 Indium Semi-conductor Grade 440502 0.1050.120 Duplex stainless steel
254 0.061 Pure Rhodium (Rh) 460486 0.1100.116 High strength low alloy steel (HSLA)
246276 0.0590.066 Pure Molybdenum (Mo) 461486 0.1100.116 High-manganese carbon steel
268 0.064 Pure Niobium (Nb) 461486 0.1100.116 Austenitic cast iron with graphite
285 0.068 Pure Barium (Ba) 448502 0.1070.120 Chromium molybdenum alloy steel
281289 0.0670.069 Pure Zirconium (Zr) 418540 0.1000.129 Nickel copper alloys
298300 0.0710.072 Pure Yttrium (Y) 460502 0.1100.120 Austenitic cast stainless steel
308 0.074 Pure Strontium (Sr) 430532 0.1030.127 Commercially pure or low-alloy nickel
310322 0.0740.077 Pure Germanium (Ge) 460502 0.1100.120 Ductile cast iron
310328 0.0740.078 Pure Arsenic (As) 486 0.116 High strength structural steel
325331 0.0780.079 Gallium compounds 481502 0.1150.120 Resulfurized carbon steel
272418 0.0650.100 Zinc alloys 475508 0.1130.121 Pure Manganese (Mn)
330380 0.0790.091 Pure Rubidium (Rb) 460527 0.1100.126 Malleable cast iron
(continued)

Specific heat capacity Specific heat capacity
J/kg K Btu/lb F) Material J/kg K Btu/lb F) Material
500 0.119 Cast ferritic stainless steel 887963 0.2120.230 (6xxx series) Wrought aluminum
498506 0.1190.121 Pure Vanadium (V) magnesium silicon
420586 0.1000.140 Ferritic stainless steel 8001050 0.1910.251 Cast magnesium aluminum zinc
460572 0.1100.137 Alpha and near alpha titanium alloys 7951050 0.1900.251 (7xxx series) Wrought aluminum zinc
370674 0.0880.161 Alpha beta titanium alloys 900963 0.2150.230 (1xxx series) Commercially pure
460586 0.1100.140 Cast stainless steels wrought aluminum, 99.00% or
510553 0.1220.132 Manganese alloy steel greater
490580 0.1170.139 Beta and near beta titanium alloys 960963 0.2290.230 (3xx.x series) Cast aluminum
513599 0.1230.143 Pure Titanium (Ti) silicon+copper or magnesium
515597 0.1230.143 Unalloyed or low-alloy titanium 963 0.230 (4xx.x series) Cast aluminum silicon
448670 0.1070.160 Nickel chromium molybdenum alloy 963 0.230 (5xx.x series) Cast aluminum
steel magnesium
561567 0.1340.135 Pure Scandium (Sc) 963 0.230 (7xx.x series) Cast aluminum zinc
623655 0.1490.156 Pure Calcium (Ca) 963 0.230 (8xx.x series) Cast aluminum tin
691 0.165 Carbon 9661050 0.2310.251 Cast magnesium rare earth alloys
712728 0.1700.174 Pure Silicon (Si) 10101050 0.2410.251 Commercially pure magnesium
733 0.175 Pure Sulfur (S) 9621109 0.2300.265 Wrought magnesium aluminum zinc
741 0.177 Pure Phosphorus (P) alloys
770 0.184 Pure Potassium (K) 10001080 0.2390.258 Cast magnesium aluminum
864 0.206 (4xxx series) Wrought aluminum manganese
silicon 10251064 0.2450.254 Wrought magnesium zinc alloys
816963 0.1950.230 (2xxx series) Wrought aluminum 1050 0.251 Cast magnesium zinc
copper 1060 0.253 Pure Magnesium (Mg)
893920 0.2130.220 (3xxx series) Wrought aluminum 12221234 0.2920.295 Pure Sodium (Na)
manganese 12851620 0.3070.387 Pure Boron (B)
895933 0.2140.223 Pure Aluminum (Al) 15001630 0.3580.389 Beryllium alloys
875963 0.2090.230 (2xx.x series) Cast aluminum copper 18862070 0.4500.494 Pure Beryllium (Be)
920 0.220 (8xxx series) Wrought 33003515 0.7880.840 Pure Lithium (Li)
aluminum+other elements
879963 0.2100.230 (5xxx series) Wrought aluminum
magnesium
Table A6.6 Specific heats of carbon and low-alloy steels

Mean apparent specific heat, J/Kg K, at C (F)
AISI- Treatment 550600 650700 700750 750800 850900
SAE or 50100 150200 200250 250300 300350 350400 450500 (1022 (1202 (1292 (1382 (1562
grade condition (122212) (302392) (392482) (482572) (572662) (662752) (842932) 1112) 1292) 1382) 1472) 1652)
1008 Annealed 481 519 536 553 574 595 662 754 867 1105 875 846
1010 Unknown 450 500 520 535 565 590 650 730 825 ... ... ...
1015 Annealed 486 519 ... ... ... 599 ... ... ... ... ... ...
1016 Annealed 481 515 ... ... ... 595 ... ... ... ... ... ...
1017 Unknown 481(a) ... ... ... ... ... ... ... ... ... ... ...
1018 Annealed 486 519 ... ... ... 599 ... ... ... ... ... ...
1020 Unknown 486 519 ... ... ... 599 ... ... ... ... ... ...
1025 Annealed 486 519 532 557 574 599 662 749 846 1432 950 ...
1030 Annealed 486 519 ... ... ... 599 . . .. ... ... ... ... ...
1035 Annealed 486 519 ... ... ... 586 ... ... ... ... ... ...
1040 Annealed 486 519 ... ... ... 586 ... ... ... ... ... ...
1042 Annealed 486 515 528 548 569 586 649 708 770 1583 624 548
1045 Annealed 486 519 ... ... ... 586 ... ... ... ... . . .. ...
1050 Annealed 486 519 ... ... ... 590 ... ... ... ... ... ...
1060 Unknown 502 544 ... ... ... ... ... ... ... ... ... ...
1070 Unknown 490 532 ... ... ... ... ... ... ... ... ... ...
1078 Annealed 490 532 548 565 586 607 670 712 770 2081 615 ...
1086 Unknown 500 532 ... ... ... ... ... ... ... ... ... ...
1095 Unknown 461(b) ... ... ... ... ... ... ... ... ... ... ...
1117 Unknown 481 ... ... ... ... ... ... ... ... ... ... ...
1140 Unknown 461(b) ... ... ... ... ... ... ... ... ... ... ...
1151 Unknown 502(b) ... ... ... ... ... ... ... ... ... ... ...
1522 Annealed 486 519 ... ... ... 599 ... ... ... ... ... ...
1524 Annealed 477 511 528 544 565 590 649 741 837 1449 821 536
(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)

Mean apparent specific heat, J/Kg K, at C (F)
AISI- Treatment 550600 650700 700750 750800 850900
SAE or 50100 150200 200250 250300 300350 350400 450500 (1022 (1202 (1292 (1382 (1562
grade condition (122212) (302392) (392482) (482572) (572662) (662752) (842932) 1112) 1292) 1382) 1472) 1652)
1561 Annealed 486 519 ... ... ... ... ... ... ... ... ... ...
4032 Unknown ... 461(c) ... ... ... ... ... ... ... ... ... ...
4130 Hardened 477 515 ... 544 ... 595 657 737 825 ... 833 ...
and
tempered
4140 Hardened ... 473(d) ... ... ... 519(d) ... 561(d) ... ... ... ...
and
tempered
4142 Unknown ... 502(c) ... ... ... ... ... ... ... ... ... ...
4626 Normalized 335(e) ... ... ... ... ... ... ... 615(f) ... ... ...
and
tempered
4815 Unknown 481(b) ... ... ... ... ... ... ... ... ... ... ...
5132 Unknown 494 523 536 553 574 595 657 741 837 1499 934 574
5140 Hardened 452(d) 473(d) ... ... ... 519(d) ... 561(d) ... ... ... ...
and
tempered
8115 Unknown 461(b) ... ... ... ... ... ... ... ... ... ... ...
8617 Unknown 481(a) ... ... ... ... ... ... ... ... ... ... ...
8637 Unknown ... 502(c) ... ... ... ... ... ... ... ... ... ...
(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)
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 Fed. R. of Germany Japan United Kingdom France Italy Sweden
(SAE) (DIN) (JIS) (BS) (AFNOR NF) (UNI) (SS14)
Carbon steels
1005 1.0288, D5-2 ... 970 015A03 ... 5598 3CD5 1160
1.0303, QSt32-2
1.0312, D5-1
1.0314, D6-2
1.0393, ED3
1.0394, ED4
1.1012, RFe120
1006 1.0311, D7-1 ... 970 030A04 A35-564 XC6FF 5598 3CD6 1147
1.0313, D8-2 970 040A04 5771 C8 1225
1.0317, RSD4 970 050A04
1.0321, St23
1.0334, StW23
1.0335, StW24
1.0354, St14Cu3
1.0391, EK2
1.0392, EK4
1.1009, Ck7
1008 1.0010, D9 G34445 1449 3CR A35-551 XC10 5598 3CD8 1142
1.0318, St28 STKM11A 1449 3CS XC6 1146
1.0320, St22 (11A) 1449 3HR XC6FF
1.0322, USD8 1449 3HS
1.0326, RSt28 1717 ERW101
1.0330, St2, St12 3606 261
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
1010 1.0204, UQSt36 G4051 S10C 1449 40F30, 43F35, A33-101 AF34 5331 C10 1232
1.0301, C10 G4051 S9Ck 46F40, 50F45, CC10 6403 C10 1265
1.0328, USD10 60F55, 68F62, C10 7065 C10 1311
1.0349, RSD9 75F70 (available in HR, 7846 C10
1.1121, Ck10 HS, CS conditions) 5598 ICD10
1.1122, Cq10 1449 4HR, 4HS, 4CR, 4CS 5598 3CD12
970 040A10 (En2A, 5771 C12
En2A/1, En2B)
970 045A10, 045M10 7356 CB10FF,
(En32A) CB10FU
(continued)

Carbon steels (continued)

970 050A10
970 060A10
980 CEW1
1012 1.0439, RSD13 G4051 S12C 1449 12HS, 12CS A33-101 AF37 ... 1332
1501 141-360 A-35 551 XC12 1431
970 040A12 (En2A, C12
En2A/1, En2B)
970 050A12
970 060A12
1013 1.0036, USt37-2 ... 3059 360 A35-551 XC12 5869 Fe360-1KG, 1233
1.0037, St37-2 3061 360 CC12 Fe360-2KW 1234
1.0038, RSt37-2 3603 360 6403 Fe35-2 1330
1.0055, Ust34-1 7070 Fe34CFN
1.0057, RSt34-1 7091 Fe34
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 G4051 F15Ck 970 040A15 XC15 5331 C16 1370
1.1132, CQ15 G4051 S15C 970 050A15 7065 C16
1.1135, Ck16A1 970 060A15 7356 CB15
1.1140, Cm15 970 080A15, 080M15 7846 C15
1.1141, Ck15 970 173H16
1.1144
1.1148, Ck15AI
1016 1.0419, RSt44.2 ... 3059 440 ... ... 1370
1.0467, 15Mn3 3606 440 2101
1.0468, 15Mn3A1 970 080A15, 080M15
1.1142, GS-Ck16 970 170H15
970 17H16
1017 ... G4051 S17C 1449 17HS, 17CS A35-551 XC18 ... 1312
970 040A17 A35-552 XC18
970 050A17 A35-566 XC18
970 060A17 A35-553 XC18S
A35-554 XC18S
1018 1.0453, C16.8 ... 970 080A17 A33-101 AF42 ... ...
C20
1019 ... ... ... ... ... ...
1020 1.0402, C22 G4051 S20C 970 040A20 A35-551 XC18 5598 1CD20 1450
1.0414, D20-2 G4051 S20CK 970 050A20 A35-552 XC18 5598 3CD20
(En2C, En2D)
1.0427, C22.3 970 060A20 A35-566 XC18 6922 C21
1.0460, C22.8 A35-553 C20 7356 CB20FF
1.1149, Cm22 A35-553 XC18S
1.1151, Ck22 A35-554 XC18S
CC20
1021 ... ... 970 070M20 A35-551 21B3 5332 C20 ...
970 080A20 A35-552 21B3 7065 C20
A35-553 21B3
A35-557 21B3
A35-566 21B3
1022 1.0432, C21 ... 3111 Type 9 A35-551 20MB5 5771 20Mn4 ...
1.0469, 21Mn4 970 120M19 A35-552 20MB5
1.0482, 19Mn5 970 170H20 A35-553 20MB5
1.1133, 20Mn5, A35-556 20MB5
GS-20Mn5
(continued)
Appendix 7: AISI to Non-AISI Steel Cross Reference / 553


1.1134, Ck19 A35-557 20MB5
A35-566 20MB5
A35-566 20M5
1023 1.1150, Ck22.8 G4051 S22C 1449 2HS, 22CS ... 5332 C20 ...
1.1152, Cq22 970 040A22 (En2C, En2D) 7065 C20
970 050A22
970 060A22
970 080A22
1025 1.0406, C25 G4051 S25C ... A35-552 XC25 5598 ICD25 ...
1.0415, D25-2, D26-2 A35-566 XC25 5598 3CD25
1.1158, Ck25
1026 1.1155, GS-Ck25 ... 970 070M26 ... 7845 C25 ...
1.1156, GS-Ck24 970 080A25 7847 C25
970 080A27
1029 1.0562, 28Mn4 G3445 970 060A27 A33-101 AF50 ... ...
STKM15A 970 080A27 (En5A) CC28
(15A),
STKM15C C30
(15C)
G4051 S28C
1030 1.0528, C30 G4051 S30C 1449 30HS, 30CS A35-552 XC32 5332 C30 ...
1.0530, D30-2 970 060A30 A35-553 XC32 6403 C30
1.1178, Ck30 970 080A30 (En5B) 7065 C30
1.1179, Cm30 970 080M30 (En5) 7845 C30
1.1811, G-31Mn4 7874 C30
5598 3CD30
6783 Fe50-3
7065 C31
1035 1.0501, C35 G4051 S35C 1717 CDS105/106 A33-101 AF55 5333 C33 1550
1.0516, D35-2 970 060A35 A35-553 C35 5598 1CD35 1572
1.1172, Cq35 970 080A32 (En5C) A35-553 XC38 5598 3CD35
1.1173, Ck34 970 080A35 (En8A) A35-554 XC38 7065 C35
1.1180, Cm35 980 CFS6 XC35 7065 C36
1.1181, Ck35 XC38TS 7847 C36
C35 7356 CB35
1037 1.0520, 31Mn4 G4051 S35C 3111 type 10 ... ... ...
1.0561, 36Mn4 970 080M36
970 170H36
1038 No international
equivalents
1039 1.1190, Ck42A1 ... 970 060A40 40M5 ... ...
970 080A40 (En8C) A35-552 XC38H2
970 080M40 (En8) A35-553 38MB5
970 170H41 A35-556 38MB5
A35-557 38MB5
A35-557 XC38H2
XC42, XC42TS
1040 1.0511, C40 G4051, S40C 1287 A33-101 AF60 5598 ICD40 ...
1.0541, D40-2 1449 40HS, 40CS C40 5598 3CD40
1.1186, Ck40 3146 Class 1 Grade C 6783 Fe60-3
1.1189, Cm40 3146 Class 8 6923 C40
970 060A40 7065 C40
970 080A40 (En8C) 7065 C41
970 080M40 (En8)
1042 1.0517, D45-2 G4051 S43C 970 060A42 A35-552 XC42H1 ... ...
970 080A42 (En8D) A35-553 C40
CC45
XC42, XC42TS
1043 1.0558, GS-60.3 G4051 S43C 970 060A42 A35-552 XC42H2 7847 C43 ...
970 080A42 (En8D)
970 080M46
(continued)


1044 1.0517, D45-2 ... ... ... ... ...
1045 1.0503, C45 G4051 S45C 970 060A47 A33-101 AF65 3545 C45 1672
1.1184, Ck46 G5111 SCC5 970 080A47 A35-552 XC48H1 5332 C45
1.1191, Ck45, 970 080M46 A35-553 XC45 7065 C45
GS-Ck45
1.1192, Cq45 A35-554 XC48 7845 C45
1.1194, Cq45 XC48TS 7874 C45
1.1201, Cm45 C45 5598 1CD45
1.1193, Cf45 5598 3CD45
7065 C46
7847 C46
1046 1.0503, C45 ... 3100 AW2 45M4TS ... ...
1.0519, 45MnAl 970 080M46 A35-552 XC48H1
1.1159, GS-46Mn4 A35-552 XC48H2
XC48TS
1049 ... G3445 STKM17A 970 060A47 A35-552 XC48H1 6403 C48 ...
(17A)
G3445 STKM17C 970 080A47 A35-554 XC48 7847 C48
(17C)
XC48TS
1050 1.0540, C50 G4051 S50C 1549 50HS A35-553 XC50 5332 C50 1674
1.1202, D53-3 G4051 S53C 1549 50CS 7065 C50
1.1206, Ck50 970 060A52 7065 C51
1.1210, Ck53 970 080A52 (En43C) 7845 C50
1.1213, Cf53 970 080M50 (En43A) 7874 C50
1.1219, Cf54 5598 ICD50
1.1241, Cm50 5598 3CD50
6783 Fe70-3
7847 C53
1053 1.1210 Ck53 G4051 S53C 970 080A52 (En43C) 52M4TS 7847 C53 1674
1.1213 Cf53 A35-553 XC54
1.1219 Cf54
1055 1.0518, D55-2 G4051 S53C 3100 AW3 A33-101 AF70 5598 3CD55 ...
1.0535, C35 G4051 S55C 970 060A57 A35-552 XC55H1 7065 C55
1.1202, D53-3 970 070M55 A35-552 XC55H2 7845 C55
1.1203, Ck55 970 080A52 (En43C) A35-553 XC54 7874 C55
1.1209, Cm55 970 080A57 XC55 7065 C56
1.1210, Ck53 C55 7847 C53
1.1213, Cf53
1.1219, Cf54
1.1220, D55-3
1.1820, C55W
1059 1.0609, D58-2 ... 970 060A62 A35-553 XC60 ... ...
1.0610, D60-2
1.0611, D63-2
1.1212, D58-3
1.1222, D63-3
1.1228, D60-3
1060 1.0601, C60 G4051 S58C 1449 60HS A35-553 XC60 3545 C60 1678
1.0642, 60Mn3 1449 60CS 7064 C60
1.1221, Ck60 970 060A57 7065 C60
1.1223, Cm60 970 080A57 7845 C60
1.1740, C60W 7874 C60
5598 3CD60
7065 C61
1064 1.0611, D63-2 ... 970 060A62 ... 5598 3CD65 ...
1.0612, D65-2 970 080A62 (En43D)
1.0613, D68-2
1.1222, D63-3
1.1236, D65-3
1065 1.0627, C68 ... 970 060A67 XC65 ... ...
1.0640, 64Mn3 970 080A67 (En43E)
1.1230, Federstahldraht
FD
(continued)


1.1233
1.1240, 65Mn4
1.1250, Federstahldraht
VD
1.1260, 66Mn4
1069 1.0615, D70-2 ... ... A35-553 XC68 ... ...
1.0617, D73-2 XC70
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 ... 1449 70HS, 70CS XC70 3545 C70 1770
1.0643, 70Mn3 970 060A72
1.1231, Ck67 970 070A72 (En42)
970 080A72
1074 1.0605, C75 ... 970 070A72 (En42) A35-553 XC75 3545 C75 1774
1.0645, 76Mn3 970 080A72 XC70 7064 C75
1.0655, C74
1.122, D73-3
1075 1.0614, D75-2 ... ... A35-553 XC75 3545 C75 ...
1.0617, D73-2 XC70 7064 C75
1.0620, D78-2 5598 3CD70
1.1242, D73-3 5598 3CD75
1.1252, D78-3
1.1253, D75-3
1078 1.0620, D78-2 G4801 SUP3 970 060A78 XC80 5598 3CD80 ...
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
1080 1.1259 80Mn4 ... 1449 80HS, 80CS XC80 5598 3CD80 ...
1.1265 D85-2 970 060A78 5598 3CD85
970 060A83
970 070A78
970 080A78
970 080A83
1084 1.1830, C85W ... 970 060A86 XC85 ... ...
970 080A86 ... ... ...
1085 1.0647, 85Mn3 ... 970 080A83
1.1273, 90Mn4
1.1819, 90Mn4
1086 1.0616, C85, D85-2 ... 970 050A86 A35-553 XC90 5598 3CD85 ...
1.0626, D83-2 5598 3CD90
1.0628, D88-2
1.1262, D83-3
1.1265, D85-3
1.1269, Ck85
1.1272, D88-3
1090 1.1273, 90Mn4 ... 1449 95HS ... 3545 C90 ...
1.1819, 90Mn4 1449 95CS 7064 C90
1.1282, D95S3 970 060A96 5598 3CD95
1095 1.0618, D95-2 G4801 SUP4 1449 95HS A35-553 XC100 3545 C100 1870
1.1274, Ck101 1449 95CS 7064 C100
1.1275, Ck100 970 060A99
1.1282, D95S3
1.1291, MK97
1.1545, C105W1
1.1645, C105W2
(continued)

Carbon-manganese steels
1513 1.0424, Schiffbaustahl ... 1449 40/30 HR 12M5 ... ...
CS:DS
1.0479, 13Mn6 1449 40/30 HS A33-101 AF50-S
1.0496, 12Mn6 1449 40/30 CS A35-501 E35-4
1.0513, Schiffbaustahl 1453 A2 A35-501 E36-2
A32
1.0514, Schiffbaustahl 2772 150M12 A35-501 E36-3
B32
1.0515, Schiffbaustahl 970 125A15
E32
1.0549 970 130M15
1.0579 970 130M15 (En201)
1.0583, Schiffbaustahl
A36
D36
E36
1.0599
1.8941, QStE260N
1.8945, QStE340N
1.8950, QStE380N
1522 1.0471, 21MnS15 G4106 SMn21 1503 221-460 A35-551 20MB5 4010 FeG52 2165
1.0529, StE350-Z2 1503 223-409 A35-552 20M5 6930 20Mn6 2168
1.1120, GS-20Mn5 1503 224-490 A35-556 20M5 7660 Fe510
1.1138, GS-21Mn5 3146 CLA2 A35-552 20MB5
1.1169, 20Mn6 980 CFS7 A35-553 20MB5
1.8970, StE385.7 A35-556 20MB5
1.8972, StE415.7 A35-557 20MB5
1.8978 A35-566 20MB5
1.8979
1524 1.0499, 21Mn6A1 G4106 SMn21 1456 Grade A ... ... ...
1.1133, 20Mn5, G5111 SCMn1 970 150M19 (En14A,
GS-20Mn5 En14B)
1.1160, 22Mn6 970 175H23
980 CDS9, CDS10
1526 ... ... 970 120M28 A35-566 25MS5 2130
1527 1.0412, 27MnSi5 G5111 SCMn2 1453 A3 ... 4010 FeG60 ...
1.1161, 26Mn5 1456 Grade B1, Grade B2 7874 C28Mn
1.1165, 30Mn5 3100 A5
1.1165, GS-30Mn5 3100 A6
1.1170, 28Mn6 970 150M28 (En14A,
En14B)
1536 1.0561, 36Mn4 G4052 SMn1H 1045 A35-552 32M5 4010 FeG60 ...
1.1165, 30Mn5 G4052 SMn433H 3100 A5, A6 A35-552 38MB5
1.1165, GS-30Mn5 G4106 SMn1 970 120M36 (En15B) A35-553 38MB5
1.1166, 34Mn5 G4106 SMn433 970 150M36 (En15) A35-556 38MB5
1.1167, 36Mn5, G5111 SCMn2 A35-557 38MB5
GS-36Mn5
1.1813, G-35Mn5 G5111 SCMn3
1541 1.0563, E G4106 SMn2, 970 135M44 40M5 ... 2120
SMn438
1.0564, N-80 G4052 SMn2H, 970 150M40 45M5 2128
SMn438H
1.1127, 36Mn6 G4106 SMn3, A35-552 40M6
SMn443
1.1168, GS-40Mn5 G4052 SMn3H,
SMn443H
G5111 SCMn5
1548 1.1128, 46Mn5 ... ... ... ... ...
1.1159, GS-46Mn4
1551 1.0542, StSch80 ... ... 24M4TS ... ...
1552 1.0624, StSch90B ... ... 55M5 ... ...
1.1226, 52Mn5
(continued)

Carbon-manganese steels (continued)

1561 1.0908, 60SiMn5 ... ... ... ... ...
1566 1.1233 ... ... ... ... ...
1.1240, 65Mn4
1.1260, 66Mn7
Resulfurized carbon steels
1108 1.0700, U7S10 G4804 SUM12 ... A35-562 10F1 ... ...
1.0702, U10S10
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) 35MF4 4838 CF35SMn10 ...
970 216M36 (En15AM) A35-562 35MF6
970 225M36
1139 1.0726, 35S20 ... 970 212A37 (En8BM) 35MF4 ... 1957
970 212M36 (En8M) A35-562 35MF6
970 216M36 (En15AM)
970 225M36
equivalents
1141 ... G4804 SUM42 970 212A42 (En8DM) A35-562 45MF4 ... ...
970 216A42
1144 1.0727, 45S20 G4804 SUM43 970 212A42 (En8DM) A35-562 45MF6 4838 CF44SMn28 1973
970 212M44
970 216M44
970 225M44
970 226M44
1146 1.0727, 45S20 ... 970 212M44 45MF4 ... ...
1151 1.0728, 60S20 ... ... ... ... 1973
1.0729, 70S20
Resulfurized/rephosphorized carbon steels
equivalents
1212 1.0711, 9S20 G4804 SUM21 ... 10F2 4838 10S20 ...
1.0721, 10S20 12MF4 4838 10S22
1.1011, RFe160K S200 4838 CF9S22
1213 1.0715, 9SMn28 G4804 SUM22 970 220M07 (En1A) A35-561 S250 4838 CF9SMn28 1912
1.0736, 9SMn36 970 230M07 S250 4838 CF9SMn32
1.0740, 9SMn40 970 240M07 (En1A)
1215 1.0736, 9SMn36 G4804 SUM23 970 240M07 (En1B) A35-561 S300 4838 CF9SMn32 ...
4838 CF9SMn36
12L14 No international
equivalents
Alloy steels
equivalents
1335 1.5069, 36Mn7 ... ... ... ... ...
1340 1.5223, 42MnV7 ... ... ... ... ...
1345 1.0625, StSch90C ... ... ... ... ...
1.0912, 46Mn7
1.0913, 50Mn7
1.0915, 50MnV7
1.5085, 51Mn7
1.5225, 51MnV7
4023 1.5416, 20Mo3 ... ... ... ... ...
4024 1.5416, 20Mo3 ... ... ... ... ...
(continued)

Alloy steels (continued)

4027 1.5419, 22Mo4 ... ... ... ... ...
4028 ... ... 970 605M30 ... ... ...
4032 1.5411 G5111 SCMnM3 970 605A32 ... ... ...
970 605H32
970 605M30
970 605M36 (En16)
4037 1.2382, 43MnSiMo4 ... 3111 Type 2/1 ... ... ...
1.5412, GS-40MnMo4 3 3111 Type 2/2
1.5432, 42MnMo7 970 605A37
970 605H37
4042 1.2382, 43MnSiMo4 ... ... ... ... ...
1.5432, 42MnMo7
equivalents
4118 1.721, 23CrMoB4 G4052 SCM15H 970 708H20 ... 7846 18CrMo4 ...
1.7264, 20CrMo5 G4105 SCM21H 970 708M20
G4052 SCM418H
G4105 SCM418H
4130 ... G4105 SCM1 1717 CDS110 A35-552 30CD4 30CrMo4 2233
G4105 SCM432 970 708A30 A35-556 30CD4 6929 35CrMo4F
G4105 SCM2 A35-557 30CD4 7356 34CrMo4KB
G4105 SCM430 7845 30CrMo4
G4106 SCM2 7874 30CrMo4
4135 1.2330, 35CrMo4 G4054 SCM3H 970 708H37 35CD4 5332 35CrMo4 2234
1.7220, 34CrMo4 G4054 SCM435H 970 708H37 A35-552 35CD4 6929 35CrMo4F
1.7220, GS-34CrMo4 G4105 SCM1 A35-553 35CD4 7356 34CrMo4KB
1.7226, 34CrMoS4 G4105 SCM432 A35-556 35CD4 7845 35CrMo4
1.7231, 33CrMo4 G4105 SCM3 A35-557 34CD4 7874 35CrMo4
G4105 SCM435
4137 1.7225, GS-42CrMo4 G4052 SCM4H 3100 type 5 40CD4 5532 40CrMo4 ...
G4052 SCM440H 970 708A37 42CD4 5333 38CrMo4
G4105 SCM4 970 708H37 A35-552 38CD4 7356 38CrMo4KB
970 709A37 A35-557 38CD4
4140 1.3563, 43CrMo4 G4052 SCM4H 3100 Type 5 40CD4 3160 G40CrMo4 2244
1.7223, 41CrMo4 G4052 SCM440H 4670 711M40 A35-552 42CD4, 5332 40CrMo4
42CDTS
1.7225, 42CrMo4 G4103 SNCM4 970 708A40 A35-553 42CD4, 7845 42CrMo4
42CDTS
1.7225, GS-42CrMo4 G4105 SCM4 970 708A42 (En19C) A35-556 42CD4, 7847 41CrMo4
42CDTS
1.7227, 42CrMoS4 G4105 SCM440 970 708H42 A35-557 42CD4, 7874 42CrMo4
42CDTS
970 708M40
970 709A40
970 709M40
4142 1.3563, 43CrMo4 ... 970 708A42 (En19C) 40CD4 7845 42CrMo4 2244
1.7223, 41CrMo4 970 708H42 A35-552 42CD4, 7874 42CrMo4
42CDTS
970 709A42 A35-553 42CD4,
42CDTS
A35-556 42CD4,
42CDTS
A35-557 42CD4,
42CDTS
4145 1.2332, 47CrMo4 G4052 SCM5H 970 708H45 A35-552 45SCD6 ... ...
G4052 SCM445H A35-553 45SCD6
G4105 SCM5,
SCM445
4147 1.2332, 47CrMo4 G4052 SCM5H 970 708A47 A35-552 45SCD6 ... ...
1.3565, 48CrMo4 G4052 SCM445H A35-553 45SCD6
(continued)


1.7228, 50CrMo4 G4105 SCM5, A35-571 50SCD6
SCM445
1.7228, GS-50CrMo4
1.7230, 50CrMoPb4
1.7238, 49CrMo4
4150 1.3565,48CrMo4 ... ... A35-571 50SCD6 ... ...
1.7228, 50CrMo4
1.7228, GS-50CrMo4
1.7230, 50CrMoPb4
1.7238, 49CrMo4
4161 1.7229, 61CrMo4 G4801 SUP13 3100 BW4 ... ... ...
1.7266, GS-58CrMnMo4 3146 CLA12 Grade C
43
4320 ... G4103 SNCM23 ... 20NCD7 3097 20NiCrMo7 2523
G4103 SNCM420 A35-565 18NCD4 5331 18NiCrMo7 2523-02
G4103 SNCM420H A35-565 20NCD7 7846 18NiCrMo7
4340 1.6565, 40NiCrMo6 G4103 SNCM8 4670 818M40 ... 5332 40NiCrMo7 ...
G4103 SNCM439 970 2S.119 6926 40NiCrMo7
G4108 SNB23-1-5 7845 40NiCrMo7
G4108 SNB24-1-5 7874 40NiCrMo7
7356
40NiCrMo7KB
E4340 1.6562, 40NiCrMo7 3 ... 970 2S.119 ... ... ...
4422 1.5419, 22Mo ... ... 23D5 3608 G20Mo5
equivalent
4615 ... ... ... 15ND8 ... ...
4617 ... ... 970 665A17 ... ... ...
970 665H17
970 665M17 (En34)
4620 ... ... 970 665A19 2ND8 ... ...
970 665H20
970 665M20
4626 ... ... 970 665A24 (E35B) ... ... ...
equivalent
4720 ... ... ... 18NCD4 ... ...
equivalent
equivalent
equivalent
50B40 1.7003, 38Cr2 G4052 SMnC3H ... A35-552 38C2 7356 41Cr2KB ...
1.7023, 28CrS2 G4052 SMnC443H A35-556 38C2
G4106 SMnC3 A35-557 38C2
G4106 SMnC443 A35-552 42C2
G5111 SCMnCr4 A35-556 42C2
A35-557 42C2
50B44 ... ... ... 45C2 7847 45Cr2 ...
5046 1.3561, 44Cr2 ... ... ... ... ...

50B46 No international
equivalent
50B50 1.7138, 52MnCrB3 ... ... 55C2 ... ...
5060 1.2101, 62SiMnCr4 ... 970 526M60 (En11) 61SC7 ... ...
A35-552 60SC7
5115 1.7131, 16MnCr5, G4052 SCr21H 970 527A17 16MC5 7846 16MnCr5 2127
GS-16MnCr5
1.7139, 16MnCrS5 G4052 SCr415H 970 527H17 A35-551 16MCS
(continued)


1.7142, 16MnCrPb5 G4104 SCr21 970 527M17
1.7160, 16MnCrB5 G4104 SCr415
5117 1.3521, 17MnCr5 ... ... 18Cr4 ... ...
1.7016, 17Cr3 A35-551 16MC5
1.7131, 16MnCr5,
GS-16MnCr5
1.7139, 16MnCrS5
1.7142, 16MnCrPb5
1.7168, 18MnCrB5
5120 1.2162, 21MnCr5 G4052 SCr22H ... A35-551 20MC5 7846 20MnCr5 ...
1.3523, 19MnCr5 G4052 SCr420H A35-552 20MC5
1.7027, 20Cr4 G4052 SMn21H
1.7028, 20Cr5 4 G4052 SMn421H
1.7121, 20CrMnS3 3 G4104 SCr22
1.7146, 20MnCrPb5 G4104 SCr420
1.7147, GS-20MnCr5
1.7149, 20MnCrS5
5130 1.8401, 30MnCrTi4 G4052 SCr2H 970 530A30 (En18A) 28C4 ... ...
G4052 SCr430H 970 530H30
G4104 SCr2
G4104 SCr430
5132 1.7033, 34Cr4 G4104 SCr3 970 530A32 (En18B) A35-552 32C4 7356 34Cr4KB ...
1.7037, 34CrS4 G4104 SCr435 970 530A36 (En18C) A35-553 32C4 7874 34Cr4
970 530H32 A35-556 32C4
A35-557 32C4
5135 1.7034, 37Cr4 G4052 SCr3H 3111 Type 3 38C4 5332 35CrMn5 ...
1.7038, 37CrS4 G4052 SCr435H 970 530A36 (En18C) A35-552-38Cr4 6403 35CrMn5
1.7043, 38Cr4 970 530H36 A35-553 38Cr4 5333 36CrMn4
A35-556 38Cr4 7847 36CrMn4
A35-557 38Cr4 7356 38Cr4KB
7845 36CrMn5
7874 36CrMn5
7847 38Cr4
5140 1.7035, 41Cr4 G4052 SCr4H 3111 Type 3 A35-552 42C4 5332 40Cr4 2245
1.7039, 41CrS4 G4052 SCr440H 970 2S.117 A35-557 42C4 7356 41Cr4KB
1.7045, 42Cr4 G4104 SCr4 970 530A40 (En18D) A35-556 42C4 7845 41Cr4
G4104 SCr440 970 530H40 7874 41Cr4
970 530M40
5147 1.7145, GS-50CrMn4 4 ... 3100 BW2, BW3 50C4 ... ...
3146 CLA 12 Grade A
3146 CLA 12 Grade B
5150 1.7145, GS-50CrMn4 4 ... 3100 BW2 ... ... 2230
1.8404, 60MnCrTi4 3100 BW3
3146 CLA 12 Grade A
3146 CLA 12 Grade B
5155 1.7176, 55Cr3 G4801 SUP11 ... A35-571 55C3 ... ...
G4801 SUP9
5160 1.2125, 65MnCr4 G4801 SUP9A 970 527A60 (En48) ... ... ...
970 527H60
equivalent
E50100 1.2018, 95Cr1 ... ... A35-565 100C2 ... ...
1.3501, 100Cr2
E51100 1.2057, 105Cr4 ... ... ... 3160 G90Cr4 ...
1.2109, 125CrSi5
1.2127, 105MnCr4
1.3503, 105Cr4
E52100 1.2059, 120Cr5 ... 970 534A99 (En31) ... 100C6 2258
1.2060, 105Cr5 970 535A99 (En31) 3097 100Cr6
1.2067, 100Cr6
1.3505, 100Cr6
1.3503, 105Cr4
(continued)


1.3514, 101Cr6
1.3520, 100CrMn6
equivalent
6150 1.8159, GS-50CrV4 G4801 SUP10 970 735A50 (En47) A35-552 50CV4 3545 50CrV4 2230
970 S.204 A35-553 50CV4 7065 50CrV4
A35-571 50CV4 7845 50CrV4
7874 50CrV4
equivalents
equivalents
8615 ... ... ... 15NCD2 3097 16NiCrMo2 ...
15NCD4 5331 16NiCrMo2
7846 16NiCrMo2
8617 ... ... 970 805A17 18NCD4 ... ...
970 805H17 18NCD6
970 805M17 (En 361)
8620 1.6522, 20NiCrMo2 G4052 SNCM21H 2772 806M20 18NCD4 5331 20NiCrMo2 2506-03
1.6523, 21NiCrMo2 G4052 SNCM220H 970 805A20 20NCD2 6403 20NiCrMo2 2506-08
1.6526, 21NiCrMoS2 G4103 SNCM21 970 805H20 A35-551 19NCDB2 7846 20NiCrMo2
1.6543, 21NiCrMo2 2 G4103 SNCM220 970 805M20 (En362) A35-552 19NCDB2
A35-551 20NCD2
A35-553 20NCD2
A35-565 20NCD2
A35-566 20NCD2
8622 1.6541, ... 2772 806M22 23NCDB4 ... ...
23MnNiCrMo5 2
970 805A22 A35-556 23MNCD5
970 805H22 A35-556 23NCDB2
970 805M22 A35-566 22NCD2
8625 ... ... 970 805H25 25NCD4 ... ...
970 805M25 A35-556 25MNCD6
A35-566 25MNDC6
equivalents
8630 1.6545, 30NiCrMo2 2 ... ... 30NCD2 7356 ...
30NiCrMo2KB
8637 ... ... 970 945M38 (En100) 40NCD3 5332 38NiCrMo4 ...
7356 ...
38NiCrMo4KB
7845 39NiCrMo3
7874 39NiCrMo3
8640 1.6546, 40NiCrMo2 2 ... 3111 Type 7, 2S.147 40NCD2 5333 40NiCrMo4 ...
970 945A40 (En 100C) 40NCD2TS 7356
40NiCrMo2KB
40NCD3TS 7845 40NiCrMo2
40NCD3 7874 40NiCrMo2
7847 40NiCrMo3
equivalents
equivalents
equivalents
equivalents
equivalents
8660 ... ... 970 805A60 ... ... ...
970 805H60
(continued)


equivalents
8740 1.6546, 40NiCrMo2 2 ... 3111 Type 7, 2S.147 40NCD2 7356
40NiCrMo2KB ...
equivalents
equivalents
9260 ... G4801 SUP7 970 250A58 (En45A) 60S7 ... ...
970 250A61 (En45A) 61S7
E9310 1.6657, 14NiCrMo13 4 ... 970 832H13 16NCD13 6932 15NiCrMo13 ...
970 832M13 (En36C) 9335 10NiCrMo13
S.157
equivalents
equivalents
equivalents
APPENDIX 8
Non-AISI to AISI Steel Cross Reference
Table A8.1 Cross reference to steels by country

Country Designation AISI Country Designation AISI
FRANCE AFNOR FRANCE (continued) AFNOR
20 MC 5 5120 CC 55 1060
20 MC 5 5120H XC 10 1010
20 NCD 2 8617 XC 15 1015
20 NCD 2 8617H XC 15 1017
20 NCD 2 8620 XC 18 1015
20 NCD 2 8620H XC 18 1017
22 NCD 2 8617 XC 18 S 1023
22 NCD 2 8617H XC 25 1023
22 NCD 2 8620 XC 32 1034
22 NCD 2 8620H XC 35 1034
25 CD 4 (S) 4130 XC 38 1034
25 CD 4 (S) 4130H XC 38 TS 1038
32 C 4 5130H XC 38 TS 1038H
32 C 4 5132 XC 42 1045
32 DCV 28 H10 XC 42 1045H
35 CD 4 4135 XC 42 TS 1045
35 CD 4 4135H XC 42 TS 1045H
35 CD 4 TS 4135 XC 45 1045
35 CD 4 TS 4135H XC 45 1045H
35 M 5 1039 XC 48 1045
38 C 4 5132H XC 48 1045H
38 C 4 5135 XC 60 1064
40 CD 4 4137 XC 65 1064
40 CD 4 4137H XC 68 1070
40 CD 4 4140 XC 90 1086
40 CD 4 4140H
Z 2 CND 17.12 316L
40 M 5 1335
Z 2 CND 19.15 317L
40 M 5 1335H
Z 6 CA 13 405
42 C 2 5140H
Z 6 CN 18.09 304
42 C 2 5150
Z 6 CND 17.11 316
42 C 4 5135H
Z 6 CNN6 18.10 347
42 C 4 5140
Z 6 CNT 18.10 321
42 CD 4 4137
Z 6 CNU 17.04 431
42 CD 4 4137H
Z 8 C 17 430
42 CD 4 4140
Z 8 CD 17.01 434
42 CD 4 4140H
Z 10 C 13 410
45 C 2 5140H
Z 10 C 14 410
45 C 2 5150
Z 10 CF 17 430F
50 CV 4 6150
Z 10 CNF 18.09 303
50 CV 4 6150H
Z 12 C 13 410
55 C 3 5155
Z 12 C 13 M 403
55 C 3 5155H
Z 12 CN 17.08 301
55 S 7 9255
Z 12 CNS 25.20 310
55 WC 20 S1
Z 12 CNS 25.20 314
60 S 7 9260
Z 15 CN 16.02 431
60 S 7 9260H
Z 15 CN 24.13 309S
61 SC 7 9260
Z 18 N 5 A2515
61 SC 7 9260H
Z 20 C 13 420
90 MV 8 O2
Z 30 WCV 9 H21
100 C 6 E52100
Z 38 CDV 5 H11
CC 20 1020
Z 40 COV 5 H13
CC 35 1035
(continued)
Adapted from Engineering Properties of Steel, ASM International, 1982, p 509523


FRANCE (continued) AFNOR GERMANY DIN
(continued)
Z 80 WCV 1.1274 1095
18-04-01 T1 1.2080 D3
Z 80 WKCV 1.2330 P20
18-05-04-01 T4 1.2341 P4
Z 85 DCWV 1.2343 H11
08-04-02-01 H41 1.2344 H13
Z 85 DCWV 1.2363 A2
08-04-02-01 M1 1.2365 H10
Z 85 WDCV 1.2379 D2
06-05-04-02 M2 1.2510 O1
Z 90 WDCV 1.2550 S1
06-05-04-02 M3 (Class 1) 1.2581 H21
Z 100 CDV 5 A2 1.2606 H12
Z 110 WKCDV 1.2625 H23
07-05-04-04-02 M41 1.2735 P6
Z 110 WKCDV 1.2842 O2
07-05-04-04-02 M42 1.3202 T15
Z 120 WDCV 1.3246 M41
06-05-04-03 M3 (Class 2) 1.3246 M42
Z 130 WDCV 1.3249 M33
06-05-04-04 M3 (Class 2) 1.3249 M34
Z 200 C 12 D3 1.3255 T4
1.3265 T5
GERMANY DIN
1.3342 M3 (Class 1)
1.0204 1008 1.3343 M2
1.0402 1020 1.3344 M3 (Class 2)
1.0419 1016 1.3346 H41
1.0501 1035 1.3346 M1
1.0601 1060 1.3348 M7
1.0700 1108 1.3355 T1
1.0702 1109 1.3501 E50100
1.0711 1212 1.3503 E51100
1.0715 1213 1.3505 E52100
1.0718 12L13 1.4001 410S
1.0718 12L14 1.4002 405
1.0904 9255 1.4005 416
1.0909 9260 1.4006 410
1.0909 9260H 1.4016 430
1.0912 1345 1.4021 420
1.0912 1345H 1.4024 403
1.1121 1010 1.4057 431
1.1133 1022 1.4104 430F
1.1141 1015 1.4112 440B
1.1141 1017 1.4113 434
1.1151 1023 1.4125 440C
1.1157 1039 1.4301 304
1.1158 1025 1.4303 305
1.1165 1330 1.4303 308
1.1165 1330H 1.4305 303
1.1167 1335 1.4306 304L
1.1167 1335H 1.4310 301
1.1172 1030 1.4401 316
1.1176 1038 1.4404 316L
1.1176 1038H 1.4438 317L
1.1181 1034 1.4449 317
1.1186 1040 1.4510 430Ti
1.1191 1045 1.4512 409
1.1191 1045H 1.4532 632
1.1209 1055 1.4541 321
1.1210 1050 1.4546 348
1.1221 1064 1.4550 347
1.1226 1548 1.4568 631
1.1230 1065 1.4828 309
1.1231 1070 1.4833 309S
1.1269 1086 1.4841 310
1.1273 1090 1.4841 314
(continued)

Appendix 8: Non-AISI to AISI Steel Cross Reference / 565

GERMANY DIN ITALY UNI
(continued) 1.4935 422 9 SMn 23 1213
1.4971 661 9 SMnPb 23 12L13
1.4980 660 9 SMnPb 23 12L14
1.5069 1340H 10 S 20 1212
1.5419 4419 20 NiCrMo 8617H
1.5419 4419H 20 NiCrMo 2 8617
1.5419 4422 20 NiCrMo 2 8620
1.5680 A2515 20 NiCrMo 2 8620H
1.5711 3140 25 CrMo 4 4130
1.6511 9840 25 CrMo 4 4130H
1.6523 8617 25 CrMo 4 KB 4130
1.6523 8617H 25 CrMo 4 KB 4130H
1.6523 8620 30 NiCrMo 2 KB 8630
1.6523 8620H 30 NiCrMo 2 KB 8630H
1.6543 8622 34 Cr 4 KB 5130H
1.6543 8622H 34 Cr 4 KB 5132
1.6543 8720 34 CrMo 4 KB 4135
1.6543 8720H 34 CrMo 4 KB 4135H
1.6543 8822 35 CrMo 4 4135
1.6543 8822H 35 CrMo 4 4135H
1.6545 8630 35 CrMo 4 F 4135
1.6545 8630H 35 CrMo 4 F 4135H
1.6546 8640 38 Cr 4 KB 5132H
1.6546 8640H 38 Cr 4 KB 5135
1.6546 8740 38 CrB 1 KB 50B40
1.6546 8740H 38 CrB 1 KB 50B40H
1.6562 E4340 38 CrMo 4 4142H
1.6562 E4340H 38 CrMo 4 KB 4137
1.6565 4340 38 CrMo 4 KB 4137H
1.6565 4340H 38 CrMo 4 KB 4140
1.6755 4718 38 CrMo 4 KB 4140H
1.6755 4718H 38 NiCrMo 4 9840
1.7006 5140H 40 Cr 4 5135H
1.7006 5150 40 Cr 4 5140
1.7007 50B40 40 CrMo 4 4137
1.7007 50B40H 40 CrMo 4 4137H
1.7030 5130 40 CrMo 4 4140
1.7033 5130H 40 CrMo 4 4140H
1.7033 5132 40 NiCrMo 2 KB 8640
1.7034 5132H 40 NiCrMo 2 KB 8640H
1.7034 5135 40 NiCrMo 2 KB 8740
1.7035 5135H 40 NiCrMo 2 KB 8740H
1.7035 5140 40 NiCrMo 7 E4340
1.7138 50B50 40 NiCrMo 7 E4340H
1.7138 50B50H 40 NiCrMo 7 KB E4340
1.7147 5120 40 NiCrMo 7 KB E4340H
1.7147 5120H 41 Cr 4 KB 5135H
1.7176 5155 41 Cr 4 KB 5140
1.7176 5155H 50 CrV 4 6150
1.7218 4130 50 CrV 4 6150H
1.7218 4130H 55 Si 8 9255
1.7220 4135 58 WCr 9 KU S1
1.7220 4135H 88 MnV 8 KU O2
1.7223 4142H 100 Cr 6 E52100
1.7225 4137 C 20 1020
1.7225 4137H C 35 1035
1.7225 4140 C 60 1060
1.7225 4140H CB 10 FU 1008
1.7228 4147 CB 35 1030
1.7228 4147H G 22 Mn 3 1022
1.7228 4150 G 22 Mo 5 4419
1.7228 4150H G 22 Mo 5 4419H
1.7362 501 G 22 Mo 5 4422
1.7511 6118 G 40 CrMo 4 4137
1.7511 6118H G 40 CrMo 4 4137H
1.8159 6150 G 40 CrMo 4 4140
1.8159 6150H
(continued)


ITALY (continued) UNI JAPAN (continued) JIS
G 40 CrMo 4 4140H S 15 CK 1015
ICL 472 T 321 S 15 CK 1017
X 2 CrNi 18 11 304L S 17 C 1015
X 2 CrNi 18 11 S 17 C 1017
KG 304L S 20 C 1023
X 2 CrNi 18 11 S 20 CK 1023
KW 304L S 22 C 1023
X 2 CrNiMo 17 12 316L S 25 C 1025
X 3 Cr Ni 18 11 304L S 28 C 1025
X 5 CrNi 18 10 304 S 38 C 1034
X 5 CrNiMo 17 12 316 S 40 C 1040
X 5 CrNiMo 18 15 317 S 45 C 1045
X 6 CrAl 13 405 S 45 C 1045H
X 6 CrNi 23 14 309S S 48 C 1045
X 6 CrNiTi 18 11 321 S 48 C 1045H
X 6 CrNiTi 18 11 S 53 C 1050
KG 321 S 55 C 1050
X 6 CrNiTi 18 11 SCCrM 1 4130
KT 321 SCCrM 1 4130H
X 6 CrNiTi 18 11 SCCrM 3 4315
KW 321 SCCrM 3 4135H
X 8 Cr 17 430 SCM 1 4135
X 8 CrMo 17 434 SCM 1 4135H
X 8 CrNi 19 10 305 SCM 2 4130
8 CrNi 19 10 308 SCM 2 4130H
X 8 CrNiNb 18 11 347 SCM 4 4137
X 10 CrNiS 18 09 303 SCM 4 4137H
X 10 CrS 17 430F SCM 4 4140
X 12 Cr 13 410 SCM 4 4140H
X 12 CrNi 17 07 301 SCM 4 H 4137
X 12 CrS 13 416 SCM 4 H 4137H
X 16 CrNi 16 431 SCM 4 H 4140
X 16 CrNi 23 14 309 SCM 4 H 4140H
X 16 CrNiSi 25 20 310 SCM 5 4147
X 16 CrNiSi 25 20 314 SCM 5 4147H
X 20 Cr 13 420 SCM 5 4150
X 22 CrNi 25 20 310 SCM 5 4150H
X 22 CrNi 25 20 314 SCM 5 H 4147
X 28 W 09 KU H21 SCM 5 H 4147H
X 35 CrMo 05 SCM 5 H 4150
KU H11 SCM 5 H 4150H
X 35 CrMoV 05 SCMn 2 1330
KU H13 SCMn 2 1330H
X35 CrMoW 05 SCMn 3 1335
KU H12 SCMn 3 1335H
X 75 W 18 KU T1 SCPH 11 4419
X 78 WCo 1805 SCPH 11 4419H
KU T4 SCPH 11 4422
X 80 WCo 1810 SCr 2 5130H
KU T5 SCr 2 5132
X 82 MoW 09 KU H41 SCr 2H 5130H
X 82 MoW 09 KU M1 SCr 2H 5132
X 82 WMo 0605 SCr 3 H 5132H
KU M2 SCr 3 H 5135
X 150 CrMo 12 SCr 4 H 5135H
KU D2 SCr 4 H 5140
X 150 WCoV SCS 19 304L
130505 KU T15 SKD 1 D3
X 210 Cr 13 KU D3 SKD 5 H21
SKD 6 H11
JAPAN JIS
SKD 12 A2
S 9 CK 1010 SKD 61 H13
S 10 C 1010 SKD 62 H12
S 12 C 1010 SKH 2 T1
S 15 C 1015 SKH 3 T4
S 15 C 1017 SKH 4A T5
(continued)


JAPAN (continued) JIS SWEDEN SS14
SKH 9 M2 1370 1015
SKH 52 M3 (Class 2) 1370 1017
SKH 53 M3 (Class 2) 1450 1020
SMn 1 H 1330 1550 1035
SMn 1 H 1330H 1665 1064
SMn 2 1335 1672 1045
SMn 2 1335H 1672 1045H
SMn 2 H 1335 1678 1064
SMn 2 H 1335H 1770 1070
SMnC 21 1022 1778 1070
SNCM 8 4340 1870 1095
SNCM 8 4340H 1914 12L13
SNCM 21 8617 1914 12L14
SNCM 21 8617H 2090 9255
SNCM 21 8620 2120 1335
SNCM 21 8620H 2120 1335H
SNCM 21 H 8617 2225 4130
SNCM 21 H 8617H 2225 4130H
SNCM 21 H 8620 2230 6150
SNCM 21 H 8620H 2230 6150H
SUH 309 316 2234 4135
SUH 310 316L 2234 4135H
SUH 409 409 2242 H13
SUH 616 422 2244 4137
SUM 11 1109 2244 4137H
SUM 12 1109 2244 4140
SUM 21 1212 2244 4140H
SUM 22 1213 2260 A2
SUM 22 L 12L13 2302 410
SUM 22 L 12L14 2303 420
SUM 23 L 12L13 2320 430
SUM 24 L 12L13 2325 434
SUM 24 L 12L14 2332 304
SUP 4 1095 2337 321
SUP 10 6150 2338 347
SUP 10 6150H 2346 303
SUP 11 50B50 2347 316
SUP 11 50B50H 2348 316L
SUS 301 301 2352 304L
SUS 303 303 2367 317L
SUS 304 304 2383 430F
SUS 304 L 304L 2722 M2
SUS 305 305
UNITED KINGDOM B.S.
SUS 305 308
SUS 305 J1 305 040 A 20 1020
SUS 305 J1 308 060 A 35 1035
SUS 316 316 060 A 62 1060
SUS 316 L 316L 060 A 96 1095
SUS 317 317 070 M 20 1020
SUS 321 321 080 A 32 1035
SUS 347 347 080 A 35 1035
SUS 403 403 080 A 37 1035
SUS 405 405 080 A 40 1040
SUS 410 410 080 M 36 1035
SUS 410S 410S 2 S 93 1040
SUS 416 403 2 S 117 5135H
SUS 420 J1 420 2 S 117 5140
SUS 430 430 2 S 119 4340
SUS 430 430F 2 S 119 4340H
SUS 431 431 2 S 130 348
SUS 434 434 2 S 516 1345
SUS 440 B 440B 2 S 516 1345H
SUS 440 C 440C 2 S 517 1345
SUS Y 310 310 2 S 517 1345H
SUS Y 310 314 3 S 95 4340
SUS Y 316 316 3 S 95 4340H
(continued)


UNITED KINGDOM B.S. UNITED KINGDOM B.S.
(continued) (continued)
5 S 80 431 530 H 30 5130
120 M 36 1039 530 H 32 5130H
150 M 36 1039 530 H 32 5132
220 MO7 1213 530 H 36 5132H
250 A 53 9255 530 H 36 5135
250 A 58 9260 530 H 40 5135H
250 A 58 9260H 530 H 40 5140
302 S 17 304 530 M 40 5135H
303 S 21 303 530 M 40 5140
304 S 12 304L 534 A 99 E52100
304 S 14 304L 535 A 99 E52100
304 S 15 304 640 M 40 3140
304 S 16 304 708 A 37 4135
304 S 18 304 708 A 37 4135H
304 S 22 304L 708 A 42 4137
304 S 25 304 708 A 42 4137H
304 S 40 304 708 A 42 4140
310 S 24 310 708 A 42 4140H
310 S 24 314 708 M 40 4137
316 S 12 316L 708 M 40 4137H
316 S 14 316L 708 M 40 4140
316 S 16 316 708 M 40 4140H
316 S 18 316 708 A 40 4137H
316 S 22 316L 709 M 40 4137
316 S 24 316L 709 M 40 4140
316 S 25 316 709 M 40 4140H
316 S 26 316 735 A 50 6150
316 S 29 316L 735 A 50 6150H
316 S 30 316 805 A 20 8622
316 S 30 316L 805 A 20 8622H
316 S 31 316L 805 A 20 8720
316 S 37 316L 805 A 20 8720H
316S 40 316 805 A 20 8822
316 S 41 316 805 A 20 8822H
316 S 82 316L 805 H 20 8617
317 S 12 317L 805 H 20 8617H
321 S 12 321 805 H 20 8620
321 S 18 321 805 H 20 8620H
321 S 22 321 805 M 20 8617
321 S 27 321 805 M 20 8617H
321 S 40 321 805 M 20 8620
321 S 49 321 805 M 20 8620H
321 S 50 321 816 M 40 9840
321 S 59 321 817 M 40 4340
321 S 87 321 817 M 40 4340H
347 S 17 347 3111 Type 6 4340
347 S 17 348 3111 Type 6 4340H
347 S 18 348 ANC 1 Grade A 410
347 S 40 348 ANC 3 Grade B 347
403 S 17 410S BA 2 A2
405 S 17 405 BD 2 D2
409 S 17 409 BD 3 D3
410 S 21 410 BH 11 H11
416 S 21 416 BH 12 H12
420 S 29 403 BH 13 H13
420 S 37 420 BH 21 H21
430 S 15 430 BM 1 H41
431 S 29 431 BM 1 M1
434 S 19 434 BM 2 M2
530 A 30 5130 BM 34 M33
530 A 32 5130H BM 34 M34
530 A 32 5132 BO 1 O1
530 A 36 5132H BO 2 O2
530 A 36 5135 BT 1 T1
530 A 40 5135H BT 4 T4
530 A 40 5140 BT 5 T5
(continued)


UNITED KINGDOM B.S. UNITED STATES AMS
BT 15 T15 5376 661
CDS-18 420 5501 304
CDS-20 321 5502 501
CDS 105/106 1039 5503 430
CDS 110 4130 5504 410
CDS 110 4130H 5505 410
En. 44 B 1095 5506 420
En. 47 6150 5507 316L
En. 47 6150H 5508 615
En. 56 A 410 5510 321
En. 56 B 403 5511 304L
En. 58 B 321 5512 347
En. 58 C 321 5513 304
En. 58 E 304 5514 305
En. 58 F 347 5515 302
En. 58 G 347 5516 302
En. 58 H 316 5517 301
S. 139 E4340 5518 301
S. 139 E4340H 5519 301
S. 525 348 5520 632
S. 527 348 5521 310S
S. 536 304L 5522 314
S. 537 316L 5523 309S
Type 3 5132H 5524 316
Type 3 5135 5525 660
Type 7 8640 5526 651
Type 7 8640H 5527 651
Type 7 8740 5528 631
Type 7 8740H 5529 631
Type 8 E4340 5531 661
Type 8 E4340H 5532 661
5546 633
UNITED STATES AMS
5547 634
5010 D 1212 5548 633
5024 C 1137 5549 634
5032 1020 5554 633
5040 1010 5556 347
5042 1010 5557 321
5044 1010 5558 347
5045 1020 5559 321
5047 1010 5560 304
5053 1010 5561
5060 1015 5565 304
5069 1018 5566 304
5070 1022 5567 304
5075 1025 5568 631
5077 1025 5570 321
5080 1035 5571 347
5082 1035 5572 310S
5085 1050 5573 316
5110 1080 5574 309S
5112 1090 5575 347
5112 E 1086 5576 321
5115 1070 5577 310S
5120 D 1074 5578
5121 1095 5579 651
5122 1095 5585 661
5132 1095 5591 410
5331 4340 5592 330
5333 8615 5594 634
5342 630 5602 501
5343 630 5604 630
5344 630 5610 416Se
5354 615 5611 403
5355 630 5612 403
5369 651 5613 410
(continued)


UNITED STATES AMS UNITED STATES AMS
5615 414 5775 633
5615 615 5776 410
5618 440C 5780 634
5620 420F 5781 634
5620 420F(Se) 5794 661
5621 420 5795 661
5622 630 5804 660
5626 T1 5805 660
5627 430 5812 632
5628 431 5813 632
5630 440C 5817 615
5631 440A 5821 410
5632 (Type 1) 440F 5824 631
5632 (Type 2) 440F(Se) 5825 630
5636 302 6260 F E9310
5637 302 6265 B E9310
5639 304 6272 8617
5640 (Type 1) 303 6274 8620
5640 (Type 2) 303Se 6275 94B17
5641 303Se 6275 A 94B15
5643 630 6276 8620
5644 631 6277 8620
5645 321 6280 8630
5646 347 6281 8630
5647 304L 6290 4615
5648 316 6294 4620
5649 316F 6322 8740
5650 309S 6323 8740
5651 310S 6325 8740
5652 314 6327 8740
5653 316L 6342 C 9840
5654 347 6350 4130
5655 422 6355 8630
5657 632 6356 4130
5673 631 6358 8740
5674 347 6359 4340
5678 631 6360 4130
5680 347 6361 4130
5681 347 6362 4130
5685 305 6365 C 4135
5686 305 6370 4130
5688 302 6371 4130
5689 321 6372 C 4135
5690 316 6373 4130
5691 316 6381 4140
5694 310 6382 4140
5695 310 6390 4140
5697 304 6395 4140
5716 330 6414 4340
5720 651 6415 4340
5721 651 6437 H11
5722 651 6440 E52100
5731 660 6441 E52100
5732 660 6442 B E50100
5734 660 6443 E51100
5735 660 6444 E52100
5736 660 6444 E52100
5737 660 6446 E51100
5738 303Se 6447 E52100
5742 634 6448 6150
5743 634 6449 E51100
5744 634 6450 6150
5745 633 6455 6150
5768 661 6466 502
5769 661 6467 502
5774 633 6485 H11
(continued)


UNITED STATES AMS UNITED STATES ASME
6487 H11 SA240 (XM-8) S43035
6488 H11 SA240 (XM-21) 304HN
6530 8630 SA249 304
6550 8630 SA249 304L
7240 1060 SA249 304N
7301 6150 SA249 309
7304 1095 SA249 310
7470 615 SA249 316
SA249 316L
ASME
SA249 316N
5041 1006 SA249 317
SA182 304 SA249 321
SA182 304L SA249 347
SA182 304N SA249 348
SA182 310 SA249 (XM-19)
SA182 316 SA249 (XM-29)
SA182 316L SA268 329
SA182 316N SA268 405
SA182 321 SA268 409
SA182 347 SA268 410
SA182 348 SA268 430
SA182 430 SA268 446
SA193 305 SA268 (XM-8) S43035
SA193 316 SA268 (XM-27)
SA193 321 SA312 304
SA193 347 SA312 304L
SA194 303 SA312 304N
SA194 303Se SA312 309
SA194 305 SA312 310
SA194 316 SA312 316
SA194 321 SA312 316L
SA194 347 SA312 316N
SA194 416 SA312 317
SA194 416Se SA312 321
SA194 (Type 3) 501 SA312 347
SA194 (Type 6) 410 SA312 348
SA194 (Type 8) 304 SA312 (XM-19)
SA213 304 SA312 (XM-29)
SA213 304L SA320 303
SA213 304N SA320 303Se
SA213 310 SA320 316
SA213 316 SA320 321
SA213 316L SA320 (B8) 304
SA213 316N SA320 (B8C) 347
SA213 321 SA358 304
SA213 347 SA358 304N
SA213 348 SA358 309
SA240 302 SA358 310
SA240 304 SA358 316
SA240 304L SA358 316N
SA240 304N SA358 321
SA240 305 SA358 347
SA240 309S SA358 348
SA240 310S SA376 304
SA240 316 SA376 304N
SA240 316L SA376 316
SA240 316N SA376 316N
SA240 317 SA376 321
SA240 317L SA376 347
SA240 321 SA376 348
SA240 347 SA387 (Type 5) 501
SA240 348 SA387 (Type 5) 502
SA240 405 SA403 304
SA240 410 SA403 304L
SA240 410S SA403 309
SA240 430 SA403 310
(continued)


UNITED STATES ASME UNITED STATES ASTM
SA403 316 A29 1017
SA403 316L A29 1018
SA403 316N A29 1019
SA403 317 A29 1020
SA403 321 A29 1021
SA403 347 A29 1022
SA403 348 A29 1023
SA403 (XM-19) A29 1025
SA409 304 A29 1026
SA409 309 A29 1030
SA409 310 A29 1034
SA409 316 A29 1035
SA409 317 A29 1038
SA409 321 A29 1038H
SA409 347 A29 1039
SA409 348 A29 1040
SA412 A29 1044
SA412 201 A29 1045
SA412 (XM-19) A29 1045H
SA430 304 A29 1046
SA430 304N A29 1050
SA430 316 A29 1055
SA430 316N A29 1059
SA430 321 A29 1060
SA430 347 A29 1064
SA479 302 A29 1065
SA479 304 A29 1070
SA479 304L A29 1074
SA479 304N A29 1080
SA479 310S A29 1086
SA479 316 A29 1090
SA479 316L A29 1095
SA479 316N A29 1108
SA479 321 A29 1109
SA479 347 A29 1116
SA479 348 A29 1119
SA479 405 A29 1132
SA479 410 A29 1137
SA479 430 A29 1141
SA479 (XM-8) S43035 A29 1144
SA479 (XM-19) A29 1211
SA479 (XM-27) A29 1212
SA564 630 A29 1213
SA564 (XM-25) A29 12L13
SA638 660 A29 12L14
SA688 304 A29 1215
SA688 304L A29 1547
SA688 316 A29 1548
SA688 316L A29 15B48H
SA688 (XM-29) A29 9260
SA705 630 A36
SA705 631 A57 1064
SA705 (XM-12) A59 9260
SA705 (XM-13) A107 1117
SA705 (XM-25) A107 1118
SA737 (XM-27) A107 1141
A108 1008
ASTM
A108 1010
A26 1064 A108 1015
A29 1005 A108 1016
A29 1006 A108 1017
A29 1008 A108 1018
A29 1010 A108 1020
A29 1012 A108 1030
A29 1015 A108 1035
A29 1016 A108 1040
(continued)


UNITED STATES ASTM UNITED STATES ASTM
A108 1117 A193 347
A108 1118 A193 410
A108 1137 A193 501
A108 1141 A194 303
A108 1144 A194 303Se
A108 1211 A194 304
A108 1212 A194 316
A108 1213 A194 321
A108 12L14 A194 347
A108 1215 A194 410
A131 A194 416
A131 (A) A194 416Se
A131 (AH32, DH32, EH32) A194 501
A131 (AH36, DH36, EH36) A194 (grade 8F) 303Se
A131 (B) A213 304
A131 (CS, DS) A213 304L
A131 (D) A213 304N
A131 (E) A213 310
A135 A213 316
A139 A213 316L
A139 (B) A213 316N
A139 (C) A213 321
A139 (D) A213 347
A139 (E) A213 348
A167 301 A228 1086
A167 302 A229 1065
A167 302B A230 1064
A167 304 A231 6150
A167 304L A240 302
A167 305 A240 304
A167 308 A240 304L
A167 309 A240 304N
A167 309S A240 305
A167 310 A240 309S
A167 310S A240 310S
A167 316 A240 316
A167 316L A240 316L
A167 317 A240 316N
A167 317L A240 317
A167 321 A240 317L
A167 347 A240 321
A167 348 A240 348
A176 403 A240 405
A176 405 A240 410
A176 409 A240 410S
A176 410 A240 430
A176 410S A240 (XM-8) S43035
A176 430 A240 (XM-21) 304HN
A176 442 A249 304
A176 446 A249 304L
A177 301 A249 304N
A181 1034 A249 305
A182 304 A249 309
A182 304L A249 310
A182 304N A249 316
A182 310 A249 316L
A182 316 A249 316N
A182 316L A249 317
A182 316N A249 321
A182 321 A249 347
A182 347 A249 348
A182 348 A268 329
A182 430 A268 405
A193 304 A268 409
A193 316 A268 430
A193 321 A268 430Ti
(continued)


A268 443 A304 4142H
A268 446 A304 4145H
A268 (XM-8) S43035 A304 4147H
A269 304 A304 4150H
A269 316 A304 4161H
A269 316L A304 4320H
A269 317 A304 4340H
A269 321 A304 4419H
A269 347 A304 4620H
A269 348 A304 4626H
A270 304 A304 4718H
A271 304 A304 4720H
A271 321 A304 4815H
A271 347 A304 4817
A273 1026 A304 4817H
A274 9840 A304 4820H
A276 302 A304 50B40H
A276 302B A304 50B44H
A276 304 A304 5046H
A276 304L A304 50B46H
A276 304N A304 50B50H
A276 305 A304 50B60H
A276 308 A304 5120H
A276 309 A304 5130H
A276 309S A304 5132H
A276 310 A304 5135H
A276 310S A304 5140H
A276 314 A304 5150H
A276 316 A304 5155H
A276 316L A304 5160H
A276 316N A304 51B60H
A276 317 A304 6118
A276 321 A304 6150H
A276 347 A304 81B45H
A276 348 A304 8617H
A276 403 A304 8620H
A276 405 A304 8622H
A276 410 A304 8625H
A276 414 A304 8627H
A276 420 A304 8630H
A276 430 A304 86B30H
A276 431 A304 8637H
A276 440A A304 8640H
A276 440B A304 8642H
A276 440C A304 8645H
A276 446 A304 86B45H
A276 (XM-21) 304HN A304 8650H
A276 (XM-27) 304HN A304 8655H
A284 (C) 304HN A304 8660H
A295 E50100 A304 8720H
A295 E51100 A304 8740H
A304 1330H A304 8822H
A304 1335H A304 9260H
A304 1340H A304 94B15H
A304 1345H A304 94B17H
A304 15B48H A304 94B30H
A304 4027H A304 E4340
A304 4028H A304 E4340H
A304 4037H A304 E9310H
A304 4042H A311 1137
A304 4047H A311 1141
A304 4118H A311 1144
A304 4130H A312 304
A304 4135H A312 304L
A304 4137H A312 304N
A304 4140H A312 309
(continued)


A312 310 A322 4042
A312 316 A322 4047
A312 316L A322 4118
A312 316N A322 4130
A312 317 A322 4137
A312 321 A322 4140
A312 347 A322 4142
A312 348 A322 4145
A313 302 A322 4147
A313 304 A322 4150
A313 305 A322 4161
A313 316 A322 4320
A313 631 A322 4340
A314 202 A322 4419
A314 302 A322 4615
A314 302B A322 4620
A314 303 A322 4626
A314 303Se A322 4718
A314 304 A322 4720
A314 304L A322 4815
A314 305 A322 4817
A314 308 A322 4820
A314 309 A322 50B40
A314 309S A322 50B44
A314 310 A322 50B46
A314 310S A322 50B50
A314 314 A322 50B60
A314 316 A322 5120
A314 316L A322 5130
A314 317 A322 5132
A314 321 A322 5135
A314 347 A322 5140
A314 348 A322 5150
A314 403 A322 5155
A314 405 A322 5160
A314 410 A322 51B60
A314 414 A322 6118
A314 416 A322 6150
A314 416Se A322 81B45
A314 420 A322 8615
A314 430 A322 8617
A314 430F A322 8620
A314 430F(Se) A322 8622
A314 431 A322 8625
A314 440A A322 8627
A314 440B A322 8630
A314 440C A322 8637
A314 446 A322 8640
A314 501 A322 8642
A314 502 A322 8645
A320 303 A322 8650
A320 303Se A322 8655
A320 304 A322 8660
A320 316 A322 8720
A320 321 A322 8740
A320 347 A322 8822
A322 1330 A322 9255
A322 1335 A322 9260
A322 1340 A322 94B17
A322 1345 A322 94B30
A322 3140 A322 94B40
A322 4023 A322 9840
A322 4024 A322 E9310
A322 4027 A322 E51100
A322 4028 A322 E52100
A322 4037 A331 1330
(continued)


A331 1335 A358 316
A331 1340 A358 316N
A331 1345 A358 321
A331 3140 A358 347
A331 4023 A358 348
A331 4024 A368 302
A331 4027 A368 304
A331 4028 A368 305
A331 4037 A368 316
A331 4042 A376 304
A331 4047 A376 304N
A331 4118 A376 316
A331 4130 A376 316N
A331 4137 A376 321
A331 4140 A376 347
A331 4142 A376 348
A331 4145 A387 (5) 501
A331 4147 A387 (5) 502
A331 4150 A403 304L
A331 4161 A403 304N
A331 4320 A403 309
A331 4340 A403 310
A331 4419 A403 316
A331 4615 A403 316L
A331 4620 A403 316N
A331 4626 A403 317
A331 4718 A403 321
A331 4720 A403 347
A331 4815 A403 348
A331 4817 A409 304
A331 4820 A409 309
A331 50B60 A409 310
A331 5120 A409 316
A331 5130 A409 317
A331 5132 A409 321
A331 5135 A409 347
A331 5140 A412 201
A331 5150 A412 202
A331 5155 A429 201
A331 5160 A429 202
A331 51B60 A429 (XM-19)
A331 6150 A430 304
A331 8617 A430 304N
A331 8620 A430 316
A331 8622 A430 316N
A331 8625 A430 321
A331 8627 A430 347
A331 8630 A453 651
A331 8637 A453 660
A331 8640 A457 651
A331 8642 A458 651
A331 8645 A473 202
A331 8655 A473 302
A332 8660 A473 302B
A331 8720 A473 303
A331 8740 A473 303Se
A331 8822 A473 304
A331 9260 A473 304L
A331 94B17 A473 305
A331 94B30 A473 308
A331 E4340 A473 309
A331 E52100 A473 309S
A355 4135 A473 310
A358 304N A473 310S
A358 309 A473 314
A358 310 A473 316
(continued)


A473 316L A505 4150
A473 317 A505 4320
A473 321 A505 4340
A473 347 A505 4615
A473 403 A505 4620
A473 405 A505 4718
A473 410 A505 4815
A473 410S A505 4820
A473 414 A505 5130
A473 416 A505 5132
A473 416Se A505 5140
A473 420 A505 5150
A473 430 A505 5160
A473 430F A505 6150
A473 430F(Se) A505 8615
A473 431 A505 8617
A473 440A A505 8620
A473 440B A505 8630
A473 440C A505 8640
A473 446 A505 8642
A473 501 A505 8645
A473 502 A505 8650
A477 651 A505 8655
A478 302 A505 8660
A478 304 A505 8720
A478 304L A505 8740
A478 305 A505 9260
A478 316 A505 E4340
A478 316L A505 E51100
A478 317 A505 E52100
A479 302 A510 1005
A479 304 A510 1006
A479 304L A510 1008
A479 304N A510 1010
A479 310S A510 1012
A479 316 A510 1015
A479 316L A510 1016
A479 316N A510 1017
A479 321 A510 1018
A479 347 A510 1019
A479 348 A510 1020
A479 403 A510 1021
A479 405 A510 1022
A479 410 A510 1023
A479 430 A510 1025
A479 (XM-8) S43035 A510 1026
A492 302 A510 1030
A492 304 A510 1035
A492 305 A510 1038
A493 302 A510 1039
A493 304 A510 1040
A493 305 A510 1044
A493 321 A510 1045
A493 347 A510 1046
A493 384 A510 1050
A493 410 A510 1055
A493 430 A510 1060
A493 431 A510 1070
A493 440C A510 1080
A505 4118 A510 1090
A505 4130 A510 1095
A505 4137 A510 1547
A505 4140 A510 1548
A505 4142 A511 302
A505 4145 A511 304
A505 4147 A511 304L
(continued)


A511 305 A519 4340
A511 309 A519 4422
A511 309S A519 4427
A511 310 A519 4720
A511 310S A519 4817
A511 316L A519 4820
A511 317 A519 50B40
A511 321 A519 50B44
A511 329 A519 5046
A511 347 A519 50B46
A511 403 A519 50B50
A511 405 A519 50B60
A511 410 A519 5120
A511 414 A519 5130
A511 416Se A519 5132
A511 430 A519 5135
A511 440A A519 5140
A511 443 A519 5150
A511 446 A519 5155
A512 1025 A519 5160
A512 1030 A519 51B60
A513 1016 A519 81B45
A513 1017 A519 8630
A513 1018 A519 8637
A513 1019 A519 8640
A513 4130 A519 8642
A513 8620 A519 8645
A519 1008 A519 86B45
A519 1010 A519 8650
A519 1012 A519 8660
A519 1015 A519 8720
A519 1017 A519 8740
A519 1018 A519 8822
A519 1019 A519 9260
A519 1020 A519 94B15
A519 1021 A519 94B17
A519 1022 A519 94B30
A519 1025 A519 94B40
A519 1026 A519 9840
A519 1030 A519 E4340
A519 1035 A519 E9310
A519 1040 A519 E50100
A519 1045 A519 E51100
A519 1050 A519 E52100
A519 1330 A534 4023
A519 1335 A535 4320
A519 1340 A535 4620
A519 1345 A535 4720
A519 3140 A535 4820
A519 4023 A535 E52100
A519 4024 A544 1017
A519 4027 A544 1018
A519 4028 A544 1020
A519 4037 A544 1022
A519 4042 A544 1030
A519 4047 A544 1035
A519 4118 A544 1038
A519 4130 A545 1006
A519 4135 A545 1008
A519 4137 A545 1010
A519 4140 A545 1012
A519 4142 A545 1015
A519 4145 A545 1016
A519 4147 A545 1018
A519 4150 A545 1019
A519 4320 A545 1021
(continued)


A545 1022 A576 1019
A545 1026 A576 1020
A545 1030 A576 1021
A545 1035 A576 1022
A545 1038 A576 1023
A546 1030 A576 1025
A546 1035 A576 1026
A546 1038 A576 1030
A546 1039 A576 1035
A546 1040 A576 1038
A547 1335 A576 1039
A547 1340 A576 1040
A547 4037 A576 1044
A547 4137 A576 1045
A547 4140 A576 1046
A547 4142 A576 1050
A547 4340 A576 1055
A548 1016 A576 1060
A548 1018 A576 1070
A548 1019 A576 1080
A548 1021 A576 1090
A548 1022 A576 1095
A549 1008 A576 1547
A549 1010 A576 1548
A549 1012 A579 632
A549 1015 A579 634
A549 1016 A579 (grade 61) 633
A549 1017 A579 (grade 62) 631
A549 1018 A579 (grade 63) 431
A554 301 A580 302B
A554 302 A580 304
A554 304 A580 304L
A554 304L A580 305
A554 305 A580 308
A554 309 A580 309
A554 309S A580 309S
A554 309S(Cb) A580 310S
A554 310S A580 314
A554 316L A580 316L
A554 317 A580 317
A554 347 A580 347
A554 430 A580 348
A554 430Ti A580 403
A564 630 A580 405
A564 631 A580 410
A564 632 A580 414
A564 634 A580 420
A565 422 A580 430
A565 615 A580 431
A567 661 A580 440A
A575 1008 A580 440B
A575 1010 A580 440C
A575 1012 A580 446
A575 1015 A581 303
A575 1017 A581 303Se
A575 1020 A581 416
A575 1023 A581 416Se
A575 1025 A581 430F
A575 1044 A581 430F(Se)
A576 1008 A581 (XM-2) 303MA
A576 1010 A582 303
A576 1012 A582 303Se
A576 1015 A582 416
A576 1016 A582 416Se
A576 1017 A582 420F(Se)
A576 1018 A582 430F
(continued)


A582 430F(Se) A681 A10
A582 (XM-2) 303MA A681 D2
A600 M1 A681 D3
A600 M2 A681 D4
A600 M3 (Class 1) A681 D5
A600 M3 (Class 2) A681 D7
A600 M4 A681 H10
A600 M6 A681 H11
A600 M7 A681 H12
A600 M10 A681 H13
A600 M30 A681 H14
A600 M33 A681 H19
A600 M34 A681 H21
A600 M36 A681 H22
A600 M41 A681 H23
A600 M42 A681 H24
A600 M43 A681 H25
A600 M44 A681 H26
A600 M46 A681 H41
A600 M47 A681 H42
A600 T1 A681 H43
A600 T2 A681 O1
A600 T4 A681 O2
A600 T5 A681 O6
A600 T6 A681 O7
A600 T8 A681 P2
A600 T15 A681 P3
A632 304 A681 P4
A632 304L A681 P5
A632 310 A681 P6
A632 316L A681 P20
A632 317 A681 P21
A632 348 A681 S1
A633 347 A681 S2
A638 660 A681 S4
A639 661 A681 S5
A646 4130 A681 S6
A646 4140 A681 S7
A646 4340 A682 1030
A646 E52100 A682 1035
A651 304 A682 1040
A651 409 A682 1045
A651 430 A682 1050
A651 430Ti A682 1055
A651 434 A682 1060
A651 (XM-8) S43035 A682 1064
A659 1015 A682 1065
A659 1016 A682 1070
A659 1017 A682 1074
A659 1018 A682 1080
A659 1020 A682 1086
A659 1021 A682 1095
A659 1023 A682 W1
A666 201 A686 W5
A666 202 A686 304
A666 301 A688 304L
A666 302 A688 316L
A666 304 A688 630
A681 A2 A693 631
A681 A3 A693 632
A681 A4 A693 633
A681 A5 A693 634
A681 A6 A699
A681 A7 A705 630
A681 A8 A705 631
A681 A9 A705 632
(continued)


UNITED STATES ASTM UNITED STATES FED
A705 634 QQ-S-763 430
A711 3140 QQ-S-763 440A
A711 4135 QQ-S-763 440B
A711 4720 QQ-S-763 440C
A711 8660 QQ-S-763 446
A711 E9310 QQ-S-766 201
A711 E50100 QQ-S-766 202
A711 E51100 QQ-S-766 301
A711 E52100 QQ-S-766 302
B511 330 QQ-S-766 304
B512 330 QQ-S-766 304L
B535 330 QQ-S-766 309
B536 330 QQ-S-766 310
B546 330 QQ-S-766 316
QQ-S-766 316L
FED
QQ-S-766 321
QQ-S-633 QQ-S-766 347
(C12L13) 12L13 QQ-S-766 348
QQ-S-635 (C1030) 1030 QQ-S-766 420
QQ-S-635 (C1035) 1035 QQ-S-766 430
QQ-S-635 (C1045) 1045 QQ-S-766 446
QQ-S-635 (C1050) 1050 QQ-T-570 A2
QQ-S-637 1141 QQ-T-570 A3
QQ-S-637 1215 QQ-T-570 A4
QQ-S-637 (C1008) 1008 QQ-T-570 A5
QQ-S-637 (C1109) 1109 QQ-T-570 A6
QQ-S-637 (C1116) 1116 QQ-T-570 A7
QQ-S-637 (C1117) 1117 QQ-T-570 A8
QQ-S-637 (C1118) 1118 QQ-T-570 A9
QQ-S-637 (C1119) 1119 QQ-T-570 A10
QQ-S-637 (C1132) 1132 QQ-T-570 D2
QQ-S-637 (C1137) 1137 QQ-T-570 D3
QQ-S-637 (C1144) 1144 QQ-T-570 D4
QQ-S-637 (C1211) 1211 QQ-T-570 D5
QQ-S-637 (C1212) 1212 QQ-T-570 D7
QQ-S-637 (C1913) 1213 QQ-T-570 H10
QQ-S-698 (C1008) 1008 QQ-T-570 H11
QQ-S-698 (C1015) 1015 QQ-T-570 H12
QQ-S-700 (C1025) 1025 QQ-T-570 H13
QQ-S-700 (C1030) 1030 QQ-T-570 H14
QQ-S-700 (C1035) 1035 QQ-T-570 H19
QQ-S-700 (C1045) 1045 QQ-T-570 H21
QQ-S-700 (C1050) 1050 QQ-T-570 H22
QQ-S-700 (C1055) 1055 QQ-T-570 H23
QQ-S-700 (C1065) 1065 QQ-T-570 H24
QQ-S-700 (C1074) 1074 QQ-T-570 H25
QQ-S-700 (C1080) 1080 QQ-T-570 H26
QQ-S-700 (C1086) 1086 QQ-T-570 H41
QQ-S-700 (C1095) 1095 QQ-T-570 H42
QQ-S-763 202 QQ-T-570 H43
QQ-S-763 302 QQ-T-570 O1
QQ-S-763 304 QQ-T-570 O2
QQ-S-763 304L QQ-T-570 O6
QQ-S-763 305 QQ-T-570 O7
QQ-S-763 309 QQ-T-570 S1
QQ-S-763 310 QQ-T-570 S2
QQ-S-763 316 QQ-T-570 S4
QQ-S-763 316L QQ-T-570 S5
QQ-S-763 317 QQ-T-570 S6
QQ-S-763 321 QQ-T-590 M1
QQ-S-763 347 QQ-T-590 M2
QQ-S-763 403 QQ-T-590 M3 (Class 1)
QQ-S-763 405 QQ-T-590 M3 (Class 2)
QQ-S-763 410 QQ-T-590 M4
QQ-S-763 414 QQ-T-590 M6
QQ-S-763 420 QQ-T-590 M7
(continued)


UNITED STATES FED UNITED STATES MIL SPEC
QQ-T-590 M10 MIL-S-862 446
QQ-T-590 M30 MIL-S-866 1016
QQ-T-590 M33 MIL-S-866 8615
QQ-T-590 M34 MIL-S-980 E52100
QQ-T-590 M36 MIL-S-5000 E4340
QQ-T-590 M41 MIL-S-5059 301
QQ-T-590 M42 MIL-S-5059 304
QQ-T-590 M43 MIL-S-5059 316
QQ-T-590 M44 MIL-S-6049 8740
QQ-T-590 M46 MIL-S-7420 E52100
QQ-T-590 T1 MIL-S-7493
QQ-T-590 T2 (A4615) 4615
QQ-T-590 T4 MIL-S-7493
QQ-T-590 T5 (A4620) 4620
QQ-T-590 T6 MIL-S-8503 6150
QQ-T-590 T8 MIL-S-11310
QQ-T-590 T15 (CS1005) 1005
QQ-W-412 MIL-S-11310
QQ-W-412 (II) (CS1006) 1006
QQ-W-423 302 MIL-S-11310
QQ-W-423 304 (CS1008) 1008
QQ-W-423 305 MIL-S-11310
QQ-W-423 310 (CS1010) 1010
QQ-W-423 316 MIL-S-11310
QQ-W-423 321 (CS1012) 1012
QQ-W-423 347 MIL-S-11310
QQ-W-423 410 (CS1017) 1017
QQ-W-423 416 MIS-L-11310
QQ-W-423 420 (CS1018) 1018
QQ-W-423 430 MIL-S-11310
QQ-W-461 1006 (CS1020) 1020
STD-66 202 MIL-S-11310
STD-66 304 (CS1022) 1022
STD-66 310 MIL-S-11310
STD-66 430 (CS1025) 1025
MIL-S-11310
MIL SPEC
(CS1030) 1030
MIL-C-24111 630 MIL-S-11310
MIL-F-20138 304 (CS1040) 1040
MIL-S-862 302 MIL-S-11310
MIL-S-862 303 (ORD4150) 4150
MIL-S-862 303Se MIL-S-11713 (2) 1070
MIL-S-862 304 MIL-S-16788 (C10) 1095
MIL-S-862 304L MIL-S-16974 1015
MIL-S-862 309 MIL-S-16974 1050
MIL-S-862 310 MIL-S-16974
MIL-S-862 316 (Gr. 1060) 1060
MIL-S-862 316L MIL-S-16974 1080
MIL-S-862 317 MIL-S-16974 1330
MIL-S-862 321 MIL-S-16974 1335
MIL-S-862 347 MIL-S-16974 1340
MIL-S-862 403 MIL-S-16974 3140
MIL-S-862 405 MIL-S-16974 4130
MIL-S-862 410 MIL-S-16974 4135
MIL-S-862 416 MIL-S-16974 4140
MIL-S-862 416Se MIL-S-16974 4145
MIL-S-862 420 MIL-S-16974 4340
MIL-S-862 430 MIL-S-16974 8620
MIL-S-862 430F MIL-S-16974 8625
MIL-S-862 430F(Se) MIL-S-16974 8630
MIL-S-862 431 MIL-S-16974 8640
MIL-S-862 440A MIL-S-16974 8645
MIL-S-862 440B MIL-S-18411 1117
MIL-S-862 440C MIL-S-18411 12L13
MIL-S-862 440F MIL-S-18733 4135
MIL-S-862 440F(Se) MIL-S-20166
(continued)


UNITED STATES MIL SPEC UNITED STATES MIL SPEC
(CS1116) 1116 MIL-S-25043 631
MIL-S-22141 E52100 MIL-S-46042 651
MIL-S-22698 (B) MIL-S-46049 1065
MIL-S-23195 304 MIL-S-46049 1074
MIL-S-23195 304L MIL-S-46409 1065
MIL-S-23195 347 MIL-S-81506 630
MIL-S-23195 348 MIL-S-81591 630
MIL-S-23196 304 MIL-T-6845 304
MIL-S-23196 304L MIL-T-8504 304
MIL-S-23196 347 MIL-T-8506 304
MIL-S-23196 348 MIL-W-46078 631
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_585.pdf/Appendix_9/ 18/8/2008 4:38PM Plate # 0 pg 585
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
Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_587-599.pdf/F5113_Appendix-10/ 18/8/2008 4:39PM Plate # 0 pg 587
APPENDIX 10
Isothermal Diagrams of Selected Steels
A selection of isothermal diagrams for REFERENCE

Carbon steels (1019, 1030, 1050, 1060, 1080) 1. From Atlas of Isothermal Transformation
Cr-Mo steels (4130, 4140) and Cooling Transformation Diagrams,
Ni-Cr-Mo steels (4340, 8620) American Society for Metals, 1977
Ni-Mo Steel (4640)
Cr steel (5160, 52100)
Fig. A10.1 Carbon steels, 1019. Source: Ref 1

Appendix 10: Isothermal Diagrams of Selected Steels / 589




Fig. A10.6 Chromium-molybdenum steels, 4130. Source: Ref 1

Fig. A10.7 Chromium-molybdenum steels, 4140. Source: Ref 1

Fig. A10.8 Ni-Cr-Mo steels, 4340. Source: Ref 1

Fig. A10.9 Nickel-molybdenum steels, 4640. Source: Ref 1

Fig. A10.10 Chromium steels, 5160. Source: Ref 1

Fig. A10.11 Nickel-chromium-molybdenum steels, 8620. Source: Ref 1

Fig. A10.12 Chromium steels, 52100. Source: Ref 1

APPENDIX 11
Continuous Cooling Diagrams of

Selected Steels
Selected continuous cooling transformation Ni-Cr-Mo steels (1/2 Ni-Cr-Mo, 11/2 Ni-Cr-
diagrams for: Mo, 31/2 Ni-Cr-Mo)
Chromium steels (1/2 Cr, 1Cr, 11/4 Cr-Mo,
Carbon steels with nominal carbon contents
1Cr-V, 11/2 Cr-Al-Mo)
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-Cr-
Mo steels REFERENCE
Silicon steels (Si-Mn, Si-Cr, Si-Cr-Mo)
1/4 Mo steel 1. From Atlas of Continuous Cooling Trans-
Nickel steels (31/2 Ni, 11/2 Ni-Mn, 13/4 Ni-Mo, formation Diagrams for Engineering Steels,
11/4 NiCr) American Society for Metals, 1977
C Si Mn P S Cr Mo Ni Al Nb V
0.18 0.20 0.45 0.020 0.020
900
A Start
Ac3
10%
800
50%
Ac1 F
700 90%
P Finish
600
C B
500
400
300 M
200
100
1000 500 200 100 50 20 10 5 2 1
Cooling rate at 800 C C per min
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 50
HV HRC
400
40
300 30
20
200 As cooled
10
100
Fig. A11.1 0.18 C (10171022), analysis wt%, austenitized at 900 C, previous treatment rolled
Appendix 11: Continuous Cooling Diagrams of Selected Steels / 603
C Si Mn P S Cr Mo Ni Al Nb V Cu Sn
0.44 0.28 0.81 0.035 0.037 0.14 0.04 0.15 0.12 0.016
900
800
Ac3 A Start
Ac1 10%
700 F
50%
90%
Finish
P
600
C B
500
400
Ms
300
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardenability band Hardness after
BS 970 080H46 transformation
700 60
600 As cooled
T 500 C 1 h
500 T 550 C 1 h 50
HV T 600 C 1 h HRC
400 T 650 C 1 h
40
T 700 C 1 h
300 30
20
200 10
100
0.86 0.20 0.60 0.020 0.020
900
800
Ac1a
Ac1b A + C Start
700
P Finish
600
C
500
B
400
10% 50% 90%
300
Ms
200 10%
50%
M
100 90%
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar
diameter 5 10 20 50 100 150 200 300 500
mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 As cooled 50
HV HRC
400 40
300 30
20
200 10
100
C Si Mn P S Cr Mo Ni AI Nb V
0.96 0.20 0.60 0.020 0.020
900
800
Ac1a
A C Start
Ac1b
700
P
Finish
600
C B
500
10% 50% 90%

400
300
200 Ms
10%
50%
100 M
1000 500 200 100 50 20 10 5 2 1
90%
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
As cooled
500 T 400 C 1 h 50
HV T 500 C 1 h HRC
400
40
300 30
20
200 10
100
0.46 0.25 1.80 0.020 0.015
900
800
Ac3
Ac1 A
700
Start
10%
600 50%
F 90%
Finish
C
P
500
400
B
300 Ms
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
Hardenability band
800
SAE 1345H
Hardness after
transformation
700 60
600
500 50
HV HRC
400
40
As cooled
300 30
20
200 10
100
Fig. A11.5 13/4 Mn (1547) (1345), analysis wt%, austenitized at 850 C, previous treatment rolled
0.44 0.20 1.50 0.020 0.250
900
800
Ac3
A
Ac1 Start
700
10%
50%
F 90%
600
Finish
C P
500
B
400
300 Ms
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar
5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 50
HV As cooled HRC
400
40
300 30
20
200 10
100
Fig. A11.6 11/2 Mn+S (1139), analysis wt%, austenitized at 850 C, previous treatment rolled
0.38 0.25 1.50 0.020 0.020 0.45
900
800 Ac3
Ac1 Start
700 A
10%
F 50%
P
600
C
500
90%
400
B Finish
Ms
300
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
BS 970 608H37 As cooled transformation
700 60
T 550 C 1 h
600 T 600 C 1 h
T 650 C 1 h
500 T 700 C 1 h 50
HV HRC
400
40
300
30
20
200
10
100
Fig. A11.7 11/2 Mn Mo, analysis wt%, austenitized at 845 C, previous treatment rolled
0.20 0.25 1.25 0.025 0.015 1.15 0.02 0.15
900
Ac3
800
Start
A
Ac1 10%
700 50%
90%
F
Finish
600
P
C
500
B
400 M
s
300
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
Hardenability band
DIN 17210 20MnCr5
500 50
HV
HRC
400
40
300 30
As cooled
20
200 10
100
Fig. A11.8 11/4 Mn Cr, analysis wt%, austenitized at 870 C, previous treatment rolled
0.19 0.20 1.60 0.020 0.020 0.25 0.55
900
Ac3
800
A Start
700 10%
Ac1 50%
F P
600
C
500
B 90%
Ms
400
Finish
300
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 50
HV HRC
400 40
300 As cooled 30
20
200 10
100
Fig. A11.9 11/2 Mn Ni Mo, analysis wt%, austenitized at 870 C, previous treatment rolled
0.43 0.24 1.35 0.025 0.025 0.45 0.20 0.75
900
800
Ac3
Ac1
700 A
Start
10%
F
600 P
C
500
50%
90%
400 B
300 Ms Finish
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
As cooled
600 T 550 C 1 h
T 620 C 1 h
500 T 650 C 1 h 50
HV HRC
400 40
300 30
20
200 10
100
Fig. A11.10 11/2 Mn Ni Cr Mo, analysis wt%, austenitized at 850 C, previous treatment rolled and softened 650 C 1 h
0.40 1.75 0.85 0.030 0.030
900
Ac3
800 Start
Ac1 A 10%
50%
90%
700 Finish
F
P
600
C
500
B
400
Ms
300
200 M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
NF A-35 41S7 transformation
700 60
600
500 50
As cooled
HV HRC
400
40
300 30
20
200 10
100
Fig. A11.11 13/4 Si Mn, analysis wt%, austenitized at 910 C, previous treatment rolled
0.45 3.40 0.60 0.015 0.010 8.50
900 Ac3
Start
800
10%
A + C P 50%
700 Finish
600
C
500
400
300
B
200
Ms
100
M
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar
diameter 5 10 20 50 100 150 200 300 500
mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600 As cooled
500 50
HV HRC
400 40
300 30
20
200 10
100
Fig. A11.12 31/2 Si Cr, analysis wt%, austenitized at 1050 C, previous treatment rolled and softened 650 C 1 h
0.60 1.90 0.85 0.025 0.025 0.30 0.25
900
Ac3
800 Start
Ac1 10%
50%
A 90%
700 Finish
P
600
500
400
B
300
Ms
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 50
HV HRC
400 40
As cooled
300 30
20
200 10
100
Fig. A11.13 2 Si Cr Mo, analysis wt%, austenitized at 910 C, previous treatment rolled
0.40 0.30 0.80 0.025 0.020 0.26
900
800
Ac3
A
Ac1 Start
700
10%
F 50%
90%
600
P Finish
C
500 B
400
Ms
300
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
SAE 4042H transformation
700 60
600
500 50
HV As cooled HRC
400
40
300 30
20
200 10
100
Fig. A11.14 1/4 Mo (40374042), analysis wt%, austenitized at 810 C, previous treatment rolled
0.40 0.26 0.62 0.007 0.005 0.23 0.10 3.45
900
800
Ac3 A
700
Ac1
600 Start
10%
C F 50%
500
B
400 90%
Finish
300
Ms
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar
diameter 5 10 20 50 100 150 200 300 500
mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
As cooled
500 T 575 C 1 h 50
HV T 625 C 1 h HRC
400
40
300 30
20
200 10
100
Fig. A11.15 31/2 Ni, analysis wt%, austenitized at 860 C, previous treatment rolled
0.16 0.25 1.40 0.020 0.015 0.20 0.05 1.50
900
800 Ac3
A
Start
Ac1 10%
700
50%
600
F
C
500
90%
B
400
Ms h
Finis
300
200 M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 50
HV HRC
As cooled
400
40
300 30
20
200 10
100
Fig. A11.16 11/2 Ni Mn, analysis wt%, austenitized at 840 C, previous treatment rolled
0.40 0.15 0.48 0.016 0.040 0.15 0.25 1.75
900
800
Ac3
A
Ac1 Start
700
10%
50%
F 90%
600 Finish
P
C
500
B
400
Ms
300
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600 As cooled
T 500 C 1 h
500 T 600 C 1 h 50
HV HRC
400
40
300 30
20
200 10
100
Fig. A11.17 13/4 Ni Mo, analysis wt%, austenitized at 845 C, previous treatment rolled
0.40 0.23 0.75 0.020 0.020 0.65 1.30
900
800
Ac3
A
Ac1 Start
700
10%
50%
F 90%
600 Finish
P
C
500
400 B
Ms
300
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar
5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
As cooled
600 T 550 C 1 h
T 600 C 1 h
500 T 650 C 1 h 50
HV HRC
400 40
300 30
20
200 10
100
Fig. A11.18 11/4 Ni Cr, analysis wt%, austenitized at 850 C, previous treatment rolled
0.24 0.20 0.80 0.020 0.020 0.50 0.20 0.55
900
Ac3
800 Start
A
10%
Ac1 50%
700
90%
F
600 P
Finish
C
500
400 Ms
300
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
Hardenability band
600
BS 970 805H25
500 50
HRC
400
HV 40
300 30
20
200 As cooled
10
100
Fig. A11.19 1/2 Ni Cr Mo (86228627) (8720) (8822), analysis wt%, austenitized at 830 C, previous treatment rolled
0.40 0.25 0.60 0.020 0.020 1.20 0.15 1.50
900
800
Ac3
Ac1
700 A
F
600 P
C
Start
500 10%
50%
400 Ms 90%
B
300
Finish
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
As cooled
T 550 C 1 h
600
T 600 C 1 h
T 650 C 1 h
500 50
HV HRC
400
40
300 30
20
200
10
100
Fig. A11.20 11/2 Ni Cr Mo, analysis wt%, austenitized at 850 C, previous treatment rolled, softened 650 C 1 h
0.13 0.20 0.50 0.020 0.020 0.85 0.18 3.40
900
800 Ac3
Ac1 A
700
Start
10%
50%
600
F
C
500
90%
Ms B
400
300
Finish
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
Hardenability band
500 50
BS 970 832H13
HV HRC
As cooled
400
40
300 30
20
200
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.
0.59 0.25 0.60 0.025 0.025 0.65 0.20
900
800
Ac3 A
Ac1
Start
700
Finish
P
600
C
10%
500 50%
90%
400
B
300
Ms
200
100 M
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
Hardenability band
500 50
BS 970 527H60
HV HRC
400 40
As cooled
300 30
20
200 10
100
Fig. A11.22 21/2 Cr (5060) (51555160), analysis wt%, austenitized at 830 C, previous treatment rolled
0.39 0.20 0.70 0.020 0.020 1.05
900
800 Ac3
Ac1 A Start
700
Finish
F
600
P
C
500 10%
B
50%
90%
400
Ms
300
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
ISO R 683 VII 3 transformation
As cooled 60
700
T 500 C 1 h
600 T 550 C 1 h
T 600 C 1 h
500 T 650 C 1 h 50
HV T 700 C 1 h HRC
400
40
300 30
20
200 10
100
Fig. A11.23 1 Cr (5140), analysis wt%, austenitized at 870 C, previous treatment rolled
0.42 0.25 0.85 0.020 0.020 1.15 0.20
900
800
Ac3
Start
Ac1 A 10%
700 50%
90%
Finish
F
600
P
C
500
400
B
Ms
300
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
Hardenability band
800
SAE 4142H Hardness after
transformation
700 As cooled
60
T 450 C 1 h
600 T 500 C 1 h
T 550 C 1 h
500 T 600 C 1 h 50
HV T 650 C 1 h HRC
400 T 700 C 1 h
40
300 30
20
200 10
100
Fig. A11.24 11/4 Cr Mo (41404142), analysis wt%, austenitized at 860 C, previous treatment rolled
0.50 0.25 0.75 0.025 0.025 0.95 0.05 0.15 0.20
900
800 Ac3
Ac1 Start
700 A 10%
50% 90%
P
600 Finish
C
500
B
400
Ms
300
M
200
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300 500
diameter mm Oil
10 20 50 100 150 200 300 500
mm Water
800
Hardness after
transformation
700 60
600
500 Hardenability band 50

ISO R 683 XIV 13
HV HRC
400
40
300 30
As cooled
20
200 10
100
Fig. A11.25 1 Cr V, (6150), analysis wt%, austenitized at 875 C, previous treatment rolled
0.42 0.30 0.65 0.020 0.020 1.65 0.33 1.00
900
Start
800 Ac1
10%
50%
90%
A
700 P Finish
F
600
C
500
400 B
Ms
300
200
M
100
1000 500 200 100 50 20 10 5 2 1
0
mm 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 1000 2000 mm
Air
Bar 5 10 20 50 100 150 200 300
diameter mm Oil
10 20 50 100 150 200 300
mm Water
800
Hardness after
transformation
700 60
600
As cooled
500 50
Hardenability band HRC
HV
SIS 14 29 40
400 40
300 30
20
200 10
100
Fig. A11.26 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 pg 629

Index
A isothermal, 3(T)
powder, 395
abrasive erosion, 129 recrystallization, 3(T)
abrasive wear, 129, 130(F), 137, 195(T), 315(F) soft, 3(T)
acute-angled keyway, 11, 12(F) solution, 38, 3940
adhesion colloids, 138 spheroidized, 107(F)
adhesive wear, 129(F), 137, 315 stress, 459
aerospace applications (case studies) subcritical, 3(T), 16, 266267
cracking in a main landing gear attach pin, 354355, temper, 346
357(F), 358(F), 359(F) time range (normal), 40
failure analysis of a catapult holdback bar, 351354(F), transformation, 346
355(F), 356(F) water, 285
failure analysis of a main landing gear lever, 362364(F), anode, 131132
365(F), 366(F) anodic stress, 74(F)
failure analysis of a nose landing gear (NLG) piston axle, Association Francaise de Normalisation (AFNOR), 493
367372(F), 373(F) asymptotic, definition, 116
failure analysis of an aircraft hoist sling during static test, austenitic stainless steels, 4, 5(T), 36, 39
373375(F), 376(F), 377(F) austenitizing temperatures for steel, 537538(T)
failure analysis of an inboard flap hinge bolt, 364366(F), austenitizing temperatures for direct-hardening carbon
367(F), 368(F), 369(F) and alloy steels (SAE), 537538(T)
failure analysis of an internal spur gear, 375378(F), reheating (austenitizing) temperatures for hardening
379(F) of carburized carbon and alloy, 538(T)
fracture analysis of ASI 420 stainless steel roll pin, autofrettage, 66
359362(F), 363(F)
main landing gear (MLG) axle, 378380(F), 381(F),
382(F) B
MLG linear actuating rod and cylinder, 355359(F),
360(F), 361(F), 362(F) backscattered Z contrast, 113
multiple-leg aircraft-handling sling, 372373(F), 374(F), bainite, 2, 209, 210(F), 279(F), 341, 403404. See also lower
375(F) bainite; upper bainite
nondestructive testing and failure analysis of fin banding, 103(F), 278, 279(F), 353
attach bolts after full-scale fatigue testing, beach marks, 77, 78(F), 97, 113, 118, 505, 506(F)
380392(F&T) Biot modulus, 6
aging tendency, 3536 blind holes, 14, 33(T), 152(F)
Agricultural Ammonia Institute, 75 blowholes, 155, 156
AISI to non-AISI steel cross reference, 551562(T) boost-diffuse cycle, 184185
Aluminizing, 33 boriding, 33
American Association of State Highway Transportation brittle fracture, 30, 3334, 5359(F), 120122(F)
Officials/American Welding Society (AASHTO/ brittle temperature range (BTR), 517
AWS), 503 brittle transgranular fracture, 295
American Iron and Steel Institute (AISI), 311(T) buckling, 5, 30, 56, 88(F), 89, 128(F)
annealing bulk composition evaluation, 114
austenitic stainless steels, 39
definition, 3(T)
diffusion (homogenizing annealed), 3(T) C
double-annealing, 395396
full, 3(T) carbides
incorrect, 107(F) cementite, 222223(F)
intercritical ( partial), 3(T) film or flake, 227, 228(F)

630 / Index
carbides (continued) quenching cracks, 204207(F)

globular, 223225(F) residual stresses, 196200(F&T)
network, 225227(F&T), 228(F) retained austenite, 191196(F&T)
overview, 222 steel selection and hardenability, 181196(F&T)
spheroidal, 224 survey summary of sources of gear failures, 178(T)
carbon steels transition zone, 211, 213(F)
coefficients of linear thermal expansion, 541543(T) carburized steels
heat capacity, summary of, 548549(T) carbon potential, 216
physical properties of, 541550(T) case depth, 209
specific heats of, 549550(T) case structure, 218
thermal conductivities of, 547(T) corrosion resistance, 233
thermal conductivity, summary of, 545546(T) decarburization, 95(F)
thermal expansion, summary of, 544545(T) film or flake carbides, 227
carbonitrided components gas-carburized, 199200
carburizing and carbonitriding, comparison of, 178(T) incorrectly, 187(F)
case, definition, 183 intergranular cracking, 96, 118
case depth, 183, 209 intergranular oxidation, 219, 220(F)
case depth, improper, 208 network carbides, 225, 226
case hardness, 209 partial melting, 233234(T)
core hardness, 209 pitting, 226, 233
core microstructure, 210211, 212(F) puzzle piece, 108(F)
dimensional stability, 200204(F) retained austenite, 191196(F&T)
overcarbonitriding, 211, 213, 214(F), 215(F) spheroidized carbide particles, 224
overview, 177178(F&T) case crushing, 231232(F)
quenching, 191 casting process, failures
quenching cracks, 205 clean steel, 165166
residual stresses, 196 cold joints, 163165
carbonitriding component lifetime factors, 151
case hardening, 397398(F) decarburization during microfusion, 162163, 165(F)
case studies, 411412(F), 413414(F) improper cast design, 151153(F), 154(F)
dimensional change in, 411412(F) inclusions, 165175(F&T)
fatigue property characteristics, 246250(F&T) macroinclusions, 166167
process, 177178(F&T) microinclusions, 167168(T)
SAE 1118 steel, 491(F) porosity, effects due to, 154162(F), 163(F)
steel selection, 181 cathode, 131132
stress, 196 cause-and-effect (CE) matrix, 400401, 402403(F)
carburized components cavitation creep, 128
carbides, 222227(F&T) cementite, 222223(F)
carburizing and carbonitriding, comparison of, definition, 1
178(T) characteristics vs. defect, 88
carburizing process, 184185 Charpy impact test, 49, 55, 58, 295(F), 297, 297(F)
case, definition, 183 Charpy V-notch (CVN) impact toughness testing, 317(F),
case crushing, 231232(F) 343(F), 513
case depth, 183185(F), 186(F) chevron marks, 48, 5354(F), 119(T), 367, 370(F)
case depth, improper, 207208(F) chicken-wire pattern, 208
case hardness, 209, 210(F) chromizing, 33
core microstructure, 210211, 212(F) cladding, 39
decarburization, 213, 215217(F) clean steel, 165166, 229
design, 179181(F) clogging, 167
dimensional stability, 200204(F) coefficient of thermal expansion, 45
grain size, 217219(F) cold cracking, 36
internal oxidation, 219222(F) cold forming, 44
macropitting, 230231(F) cold joints, 163165
micropitting, 230 cold junction, 163
noncarbide inclusions, 228229(F) cold work tools, heat treating failures of, 314330(F&T)
overcarburizing, 211, 213, 214(F), 215(F) characteristics, 315317(F)
overview, 177178(F&T) chemical composition, 314317(F)
partial melting, 233234(T) design-related failures, 317319(F)
pitting corrosion, 232233(F) hardening temperatures, incorrect, 323324, 325(F),
quenching, 185, 187191 326(F)

Index / 631
heat treating processing, 323 crack

incorrect EDM, 321323(F) blunts, 116
incorrect grinding, 319321(F) brittle, 92, 121(F), 275, 506
large (blocky) carbides, 315 brittle fracture, 30
retained austenite content, excessive, 324, 326329(F), carbides, 315
330(F) circumferential, 360(F), 382
short or absent tempering, 329330, 331(F), 332(F) coloration, 60(F), 90
surface damage, 319323(F) crack path, 96
tempering temperatures, incorrect, 323324, 325(F), crack-free welding, 507
326(F) ductile, 92, 120(F)
component design ductile fracture, 33
design, overview, 1 extension, 56
design aspects faces, 44(F), 50, 63(F), 64(F)
general thermal stress, 7, 32 fatigue, 101, 171(F), 172(F), 173(F), 229(F), 242(F).
overview, 2, 4 See also fatigue crack initiation
peak stress, 7, 32, 38 fatigue failures, 30
primary stress, 7, 8, 32 forging, 142
residual stress, 78 formation, 228
secondary stress, 7, 32 formulation, 99, 103, 143(F), 153(F)
distortion, techniques for controlling, 1618 fracture, 9097(F), 98(F), 118
hardening, 2 grinding, 207208(F)
heat treat processing, introduction to, 12(T) growth mechanism, 51
heat treated components, 814(F) growth rates, 49, 50, 80
heat treatment design, 2931 heat checking, 334(F)
heat treatment failures, examples, 1829(F) intergranular path, 306(F)
heat treatment, modeling of, 3133(F&T) lateral oxide formation, 221
heat-transfer theory, 47(F&T) longitudinal, 169, 170(F), 357, 360(F)
quenching medium, 17 macrobrittle, 92
quenching techniques, 1718 macroductile, 92
risk-based approach to, 4041(F) mechanical, 348(F)
steel grade/condition, 1416 metallography, 5051
tempering, 2 microscale crack path, 97
welded components, 3336 networks, 33
contact fatigue pitting. See macropitting normalization, 153(F), 154(F)
contact loading, 95, 98(F), 206(F) nucleation, 128, 130, 159, 160, 168(F), 170, 171(F)
continuous cooling diagrams, 601627(F) optical micrograph, 27(F)
1/4 Mo steel, 615(F) order, 48
carbon steels, 602605(F) origin, 21, 25(F), 90, 162(F)
chromium steels, 623627(F) oxide filled, 104(F)
Mn steels, 606607(F) pearlitic matrix, 53
Mn-Cr steels, 609(F) pre-existing, 116
Mn-Mo steels, 608(F) profile, 101(F)
Mn-Ni-Cr-Mo steels, 611(F) propagation, 301, 482, 506. See also crack propagation
Mn-Ni-Mo steels, 610(F) quench, 21, 24(F), 25(F), 27(F), 102(F), 154(F), 411(F).
nickel steels, 616619(F) See also quench cracking
Ni-Cr-Mo steels, 620622(F) resistance, 107
silicon steels, 612614(F) resistance to propagation, measuring, 34
continuous cooling transformation (CCT) diagrams, 32, secondary, 27(F), 50, 101(F), 274(F), 370, 372(F), 382(F)
255257(F), 339340(F) shank, 388, 389(F)
contour hardening, 432, 470471 sharp corner, 318(F)
corrosion stage I, 78
crevice, 131 stage II, 78
definition, 131 stage III, 78
galvanic, 131132 surface, 482
pitting, 131 temperature, 480, 481(F)
stress-corrosion cracking (SCC), 131 through-the-wall, 76
types of, 131 tip, 78, 80, 112, 116, 126(F), 302
corrosion fatigue, 100101(F), 118, 120 transgranular, 162(F)
corrosive wear, 129, 138 transgranular path, 304305(F&T)
cotter holes, 152 vertical, 64(F)

632 / Index
crack extension, 38, 56, 118 D

crack initiation
carbides, 315 damage, definition, 87
classifying, 118 damage mechanism, definition, 113
electroplating, 45 damage mode, definition, 113
fatigue crack initiation, 122124(F) decarburization, 215217(F)
fatigue cycles, 482 on carburized steel, 95(F)
fatigue failures, structural features of, 78 deformation, 8890(F)
high-temperature grain-boundary oxidation, 281, buckling, 89
283(F) elastic, 89, 98(F)
hot work tool steels, 344 gradual onset, 8889(F)
inhibiting, 196 plastic, 89
internal oxidation formation, 221 sudden, 88(F), 89
microcracks, 206 die-casting dies, 331, 334(F), 341, 343
nitriding, 343 dimensional stability
retard, 123 distortion, 200202(F)
S-N curves, 482 isothermal-transformation diagrams, 202(F)
subsurface, 93, 94(F), 229(F) warpage, 200, 202
crack propagation, 124127(F) discontinuities, 133, 503505(F&T)
beach marks, 77, 78(F) distortion
brittle fracture, 121, 175 bending, 468
carbides, 280 buckling, 128(F)
cleavage crack propagation, 509 creep, 128129(F)
cleavage fracture, 54(F) reasons for, 127129(F)
definition, 482 residual stresses, 129
ductile fracture, 120(F) techniques for controlling, 1618
ductile structural materials, 506 yielding, 127128
fatigue crack propagation, 162, 170, 172(F), double-frequency heating, 428
173(F) ductile crack path. See microvoid coalescence (MVC)
grain-boundary, 338, 341(F) ductile dimples. See microvoid coalescence (MVC)
hydrogen embrittlement, 301(F) ductile fracture, 28(F), 5153(F), 91(F), 9293(F), 96,
intergranular brittle fracture, 121122(F) 120(F), 226227
nitrided tools, 344, 348(F)
rate of, 80
resistance, 34 E
SCC fractures, 73
stress analyses, 116 elastic collapse. See buckling
cracking electrical discharge machining (EDM), 2021
cold, 36 electrochemical theory, 514
delayed. See underbead cracking electron-dispersive x-ray (EDX), 169
fatigue, 506(F) electroplating, 2829, 45
ferrite vein, 512513(F) embrittlement, 293303(F&T)
flat, 82 definition, 38
HAC, 509513(F&T) temper (two-step embrittlement), 296
hydrogen-induced, 36 tempered martensite embrittlement, (TME), 294296(F)
intergranular, 59, 63(F), 96(F), 100(F), 205206(F), endogas, 177
279(F) energy-dispersive x-ray analysis, 113
microcracking, 183, 192, 206(F), 218 energy-dispersive x-ray spectroscopy (EDS), 26
premature, 333335, 336(F), 337(F), 338(F) environmentally assisted failure. See corrosion
quench, 5965(F), 273283(F&T) etching, 11(F), 64(F), 104(F), 210(F)
SCC, 28, 36, 70, 7276(F&T), 77(F) acid, 320
solidification, 515517(F) cooper sulfide, 303
toe, 503(t), 504(t) corrosion resistance during, 337(F)
underbead, 507, 510511(F) nital, 171, 174(F), 216, 324, 325(F)
creep, 128129(F) retained austenite, 326, 327(F)
creep rupture, 7677(F), 78(F) stain, 113
critical flaw size, 49, 116 temper, 45(F), 62(F), 357(F)
critical scuffing temperature, 230 Euler buckling, 30, 89
cyclic fracture propagation, 80, 82(F) exogenous inclusions, 165, 229
cyclic loading, 77, 118, 122(F) exothermic sleeves, 158

Index / 633
F fatigue resistance, 241244(F&T)

after nitriding/nitriding treatments, 242244(F&T)
failure, definition, 40, 87 fatigue striations
failure analysis, guidelines of aluminum alloy, 125(F)
background data, collection of, 111112 formation of, 80, 83(F)
chemical analysis, 114115 interstitial-free steel, 125(F)
conclusions, formulation of, 117 representative, 381(F), 390(F)
evidence, analysis of, 116117 stage II crack growth, 78(F), 81(F)
failed part, preliminary examination of, 112 typical, 354(F), 355(F), 356(F)
failure mechanism, determination of, 113114 feeders, 158
fracture mechanics analysis, 90, 116 ferrite vein cracking, 512513(F)
macroscopic examination/analysis, 112 fine grinding, 472477(F)
mechanical testing, 112 finite element analysis (FEA), 116
metallographic specimens, 113 fisheyes, 115, 243(F), 511512(F)
microchemical analysis, 114 flat cracking, 82, 83(F)
microscopic examination/analysis, 112113 forging, failure in
nondestructive testing, 112 case studies
photographic examination, 112 avoidance of flow through, lap, and crack, 145148(F)
the report, 117118 crankcase underfill, 138139(F)
specimens, 112 spade bit, 140142(F)
stress analysis, 114115 summarized, 133(T)
testing, 116 trim tear, 142143(F)
failure analysis, stages of tube bending, 139140(F), 141(F)
background information, collection of, 47 upset forging, 143145(F), 146(F)
macroscopic examination, 50 discontinuities, 133
mechanical testing, 49 factors in analysis of cold forging failures, 134(T)
metallography, 5051(F) factors in analysis of hot forging failures, 135(T)
microscopic examination, 50(F) forging process design, 134138(F&T)
nondestructive testing, 4849 forging tolerances, 135137
preliminary visual examination, 4748 lubricant performance, 138
specimens, collection/preservation of, 4950 lubrication, 137, 138
specimens, sectioning of, 50 wear types, 137138
failure mechanism, definition, 113 Fouriers first law, 5
failure mode and effects analysis (FMEA), 40, 399 Fouriers second law, 4
failure mode, definition, 113 fracture mechanics analysis, 115116
failures, mechanisms/causes of, 4386(F&T), 87109(F) fracture mechanism, 5183(F)
corrosion and environmental damage, 99101(F) brittle fracture, 5359(F)
deformation, 8890(F) ductile fracture, 5153(F)
design deficiencies, 43 fatigue, 7783(F)
failure analysis, 4751(F) intergranular brittle fracture, 5977(F&T)
fracture mechanism, 5183(F) fracture modes
fractures, 9097(F) brittle, 28(F)
heat treating process, 106108(F) ductile, 28(F)
heat treatment, poor response from, 101103(F), grain-boundary, 300
104(F) identification chart, 119(T)
manufacture, 4446(F) intergranular, 26, 28(F), 80, 82(F)
material issues, 4344(F) microvoid coalescence (MVC), 51, 52(F), 507508(F)
processing, 4446(F) TME, 65
service conditions, 4647(F) transgranular, 80, 82(F), 296
wear, 9799(F) fracture surface matching, 118
fatigue, 7783 fracture toughness, 49, 55, 185, 522523(T)
fatigue crack initiation, 122124(F) fracture toughness testing, 49
carburized steels, 185 fractures
inhibit, 177 brittle, 96. See also brittle fracture
structural surface anomalies, 199200 corrosion-fatigue cracks, 118, 120
fatigue fracture, 122127 cyclic loading, 118
fatigue crack initiation, 122124(F) examining, 9093(F)
fatigue crack propagation, 124127(F) fracture modes, 19(T), 118
nitrided layers, 241253(F&T) hydrogen embrittlement, 96. See also hydrogen
fatigue intrusions and extrusions, 80(F), 122(F) embrittlement (HEM)

634 / Index
fractures (continued) heat treated steel parts

intergranular fracture, 96. See also intergranular brittle component characteristics, 104106(F)
fracture corrosion damage, 99101(F)
macroscale brittle fracture, 91(F), 92(F) deformation, 8890(F)
macroscale ductile fracture, 91(F), 93(F) environmental damage, 99101(F)
macroscale features, 9193(F) fracture, 9097(F)
microscale features, 9597(F) heat treatment, poor response to, 101108(F)
origin of, 118 wear, 9799(F)
process of, 118 heat treater, 89, 91, 92, 104, 106, 107108
river line features, 97(F) heat treating process
slant, 56. See also brittle fracture atmosphere, improper, 107108(F)
stress versus strength, 9395(F) design engineers and, 106
stress-corrosion cracks, 118, 120 errors, 106107
striations, 118 heat treaters and, 106
fretting, 97, 98(F), 130131(F) heating rate, inadequate, 107108(F)
temperature, inadequate, 107108(F)
time at temperature, inadequate, 107108(F)
heat treatment
G component design process, 2930
design, 2931
galling or seizing, 129
failures, examples of, 1829(F)
galvanic corrosion, 131132
heat treating errors, 18
gap-by-gap hardening, 429, 430(F), 439(F), 470471
heating errors, 18
gas bubble, 155, 167
material behavior, 3031
gas holes (blowholes), 154, 155, 156
modeling of, 3133(F&T)
gas metal arc welding (GMAW), 511
phase 2 design review to avoid failures, 33(T)
gas porosity, 154156(F)
poor response to
glow discharge optical emission spectroscopy, 220
component characteristics, 104106(F)
gouging, 119(T)
raw material characteristics contributing to,
grain size, influence of, 217219(F)
101103(F), 104(F)
grain-boundary sliding, 128(F)
solid-phase transformation model, 3132
graphite products, 138
temperature errors, 18
gray cast iron, 121(F)
thermomechanical modeling, 3233
Great Boston Molasses Disaster, 5657(F)
welded components, 3640
grid crossings, 153(F)
heat-affected zone (HAZ), 35
grinder burn, 98
heat-transfer theory, 47(F&T)
grinding
Hertzian stresses, 97, 98(F), 206(F), 230
fatigue strength, 80
high-speed steels, 313314(F)
incorrect, 319321(F)
high-velocity oxy-fuel (HVOF) coating, 62
manufacture and processing, 44, 45(F)
holes, types of, 152(F). See also individual types
porosity, eliminating, 504
hot tops, 158
quench cracking, 62(F)
hot work tools, heat treating failures of, 330349(F&T)
residual stresses after, 472477(F)
characteristics of, 330333(T), 334(F), 335(F)
stock removal, 16
chemical composition, 330333(T), 334(F), 335(F)
surface damage, 319
cooling (slow) during quenching, 338343(F)
surface oxidation, removing, 222
heat treating procedures, inadequate, 335337(F)
thermal defects, preventing, 208
heating, excessive, 344349(F)
grinding burn, definition, 207
incorrect hardening temperatures, 337338, 339(F),
grinding burns, 207, 208
340(F), 341(F), 343(F)
grinding cracks, 207208(F), 303304(F)
incorrect tempering temperatures, 337338, 339(F),
grinding grains, 473474(F)
340(F), 341(F), 343(F)
Grossman H-values (numbers), 270(T)
nitrided tools, 343344(F), 348(F)
premature cracking, 333335, 336(F), 337(F),
338(F)
H hydrogen embrittlement (HEM)
electroplating, 2829, 45
halos (or fisheyes), 115. See also fisheyes intergranular brittle fracture, 59(F), 6870(F), 71(F),
heat checking, 331, 334(F) 72(F), 73(F)
heat checking cracks, 334(F) TE, interaction with, 301302(F)
heat treat processing, introduction to, 12 hydrogen traps, 508(T)

Index / 635
hydrogen-assisted cracking (HAC) machine parts, 440444(F)

examples power supplies, 421
failure to pass bend tests due to hydrogen, 513(F) rotation velocities, 452
ferrite vein cracking in high-heat output welds, supervising, 456(F)
512513(F) temperature cycles, 456457(F), 458(F)
fisheyes on fracture surface, 511512(F) time-temperature dependence in, 444449(F&T)
underbead cracking, 510511(F) use of, 417
ferrite vein cracking, 512513(F) induction scanning, 426, 427(F)
fisheyes, 511512(F) induction surface hardening
hydrogen-assisted reduced ductility, 508(F), 513(F), advantages of, 421422
514(F) gears, 491497(F)
underbead cracking, 503(T), 504(F), 509511 residual stresses, 472477(F)
weld metal HAC, 511 workpiece distortion in, 466472(F)
hydrogen-assisted cracking theory, 506509(F&T) induction surface heating, 430(F), 432437(F), 442(F), 456,
hydrogen-assisted reduced ductility, 508(F), 513(F), 514(F) 457(F), 465(F), 471(F)
hydrogen-induced cracking, 36 induction surface-hardened layer, 444, 477481(F&T)
inhomogeneity, 462, 463(F), 479
Instron TT-DM machine, 242
intergranular brittle fracture
I
causes of, 59, 279. See also individual causes
creep process and, 128(F)
impact testpieces, 121(F)
creep rupture, 7677(F), 78(F)
impact tests, usefulness of, 482
hydrogen embrittlement, 59(F), 6870(F), 71(F), 72(F),
impact toughness testing, 49
73(F)
inclusions
liquid metal embrittlement, 6667(F&T)
case studies
quench cracking, 5965(F)
failure in the axle of a reduced section in a rotating
SAE 3161, 122(F)
component, 170171, 172(F&T), 173
solid metal embrittlement, 6768(F&T), 69(F)
failure of a 52100 steel axle, 171, 173175(F)
stress-corrosion cracking (SCC), 70, 7276(F&T), 77(F)
failure of a steam turbine rotor blade, 168170(F&T),
temper embrittlement, 6566(F)
171(F)
TME and, 65(F), 295296
categories of, 53
intergranular cracking, 59, 63(F), 96(F), 100(F), 205206(F),
classifications, 165
279(F)
definition, 165
interlath cleavage, 296
noncarbide inclusions, 228229(F)
internal oxidation, 219222(F)
stringer, 166, 280281(F)
internal porosity, 158159(F)
indigenous inclusions, 165
International Organization for Standardization (ISO), 493
induction coils
ion vapor-deposited (IVD), 362, 364(F), 365, 367(F), 369(F)
materials for, 423424(F)
iron-carbon equilibrium diagram, 585(F)
multiple-turn induction coil, 424425(F)
isothermal diagrams
types of, 423
carbon steels, 588592(F)
induction hardening
chromium steels, 597(F), 599(F)
fatigue strength, 481485(F&T)
chromium-molybdenum steels, 593594(F)
fine grinding, 472477(F)
nickel-chromium-molybdenum steels, 598(F)
induction heating, 420422(F), 440444(F),
nickel-molybdenum steels, 596(F)
444449(F&T)
Ni-Cr-Mo steels, 595(F)
induction surface heating. See induction surface heating
isothermal-transformation diagrams. See time-temperature
induction surface-hardened layer, 477481(F&T)
transformation (TTT) diagrams
machine parts, 422432(F), 485491(F&T)
magnetic flux concentrators, 437440(F)
overview, 417419 J
quenching systems for, 449452(F)
steels, 419420(F&T) jig-bore grinding, 2(F), 20
stresses/residual stresses, time variation of, 452466(F) job shop heat treater, 87
surface hardening, 421422, 466472(F) Jominy distance (J-distance), 270
induction heating
coils for, 423(F)
definition, 398 L
features of, 420422(F)
gear wheels, 429, 430(F), 431(F) lap, definition, 145146, 277
Jominy curves, 4150 steel, 492, 493(F) lath martensite, 187, 188(F)

636 / Index
Liberty ships, 54 mechanical cracks, 348(F)

linear thermal expansion, 5(T), 541(T), 544(T) mechanism and mode, difference between, 113114
liquid metal embrittlement (LME), 6667(F&T), 302303(T) mechanism of failure, definition, 87
liquid shrinkage, 157(F) metric conversion guide
liquidation temperature, definition, 233(T) base SI units, 521(T)
low-alloy steels conversion factors, 522523(T)
coefficients of linear thermal expansion, 541543(T) derived SI units, 521522(T)
cold forging, strain-limiting criteria for, 142(F) supplementary SI units, 521(T)
core microstructure, 210, 212(F) microchemical analysis, 114
embrittlement, 302(T) microcracking, 183, 192, 206(F), 218
HAZ maximum hardness, 3536 microinclusions, formation of, 167168(T)
heat capacity, summary of, 548549(T) micropitting, 230
heat treatment procedures, 3637 microscale fracture features, 9597(F)
linear thermal expansion, 5(T) microscopic examination, 50(F), 112113
quench age embrittlement, 59 microvoid coalescence (MVC), 96
SCC and, 36, 73, 74(T) modulus of elasticity, 89, 529
specific heats, 549550(T)
thermal conductivities, 547(T)
thermal conductivity, summary of, 545546(T) N
thermal expansion, summary of, 544545(T)
thermal hardening, 418 National Safety Transportation Board, 75
lower bainite, 2, 210(F), 270, 292, 299, 403(F) nil-ductility temperature (NDT), 5556
lubricant performance, 138 niobium (formerly columbium), 39, 314
lubrication nitrided layers, fatigue fracture of
graphite products, 138 carbonitriding, 246250(F&T)
inadequate, 4647(F) fatigue evaluation, 244246(F&T), 247(F&T)
micropitting, 230 fatigue resistance, 241244(F&T)
wear and, 137138(F) nitrided steels, 244246(F&T), 247(F&T)
nitrocarburizing, 33
node dislocation, 152, 153(F)
M non-AISI to AISI steel cross reference, 563583(T)
noncarbide inclusions, 228229(F)
nondestructive testing, 4849
machine parts, induction hardening of
carburized parts, residual stresses in, 489491(F&T) dye-penetrant method, 48
contour hardening, 432 eddy-current methods, 48
double-frequency heating, 428 magnetic particle inspection, 48
ultrasonic testing (UT), 4849
gap-by-gap, 429, 430(F), 439(F), 470471
gear wheels, 427432(F), 439(F) North American Die Casting Association (NADCA),
induction scanning, 426, 427(F) 336337
progressive hardening, 426427(F) notch, 12(F), 55, 56, 70(F), 179(F), 266(F)
quenching, 442444(F)
quenching techniques, 425426(F) O
scan-hardening process, 425(F)
single-shot induction hardening, 422, 423, 424(F), 427, Occupational Safety and Health Administration (OSHA),
428(F), 440(F), 496 4041
single-shot induction heating, 424(F), 448(F) overcarbonitriding, 211, 213, 214(F), 215(F)
stress profiles, 485491(F&T) overcarburizing, 211, 213, 214(F), 215(F)
tooth-by-tooth, 430(F), 439(F)
tooth-gap hardening, 431432(F)
macroinclusions, formation of, 166167 P
macropitting, 230231(F)
macroscale brittle fracture, 91(F), 92(F) pancake (flattened disc), forging, 144(F)
magnetic flux concentrators, 437440(F) Paris regime, 126127(F)
martempering, 1718, 60 partial melting, 233234(T)
martensite, 285289(F) part-way downs, 146147
definition, 2 passing hole, 152(F)
diffusionless process, 286 pearlite
morphology of, 187(F) crack/void formation, 53
slipped martensite, 287 decarburization on carburized steel, 95(F)
tempered martensite, 289 definition, 1

Index / 637
phase transformations, 255263(F&T) low breaking load after carbonitriding, 413414(F)

CCT diagrams, 255257(F) low core hardness after steam treatment, 415
distortion mechanism, 261 low surface hardness after carbonitriding, 411412(F)
metallurgical crystal structure, 257(F), 258(F) oxidation after sintering, 408409(F)
residual stresses, relief of, 263(F) pinholes after sintering, 404405(F), 406(F)
TTT diagrams, 255, 256(F) variation in bore diameter after heat treatment,
volume changes during, 261263(F) 412413
volumetric change upon quenching, 257261(F&T), wear after sinter hardening, 401, 403404(F)
262(F) wear after sintering, 407(F), 408(F)
phosphorus, 55, 65, 115, 295296, 297298 powder metallurgy process
phosphorus segregation, 185, 295296, 297298 blending, 396
pinholes, 154, 155156(F), 404405(F), 406(F) compaction, 396
pitting, definition, 113114 powder production, 395396
pitting corrosion, 232233(F) sintering, 397
plastic blunting process, 78, 80 progressive hardening, 420, 426, 427(F), 446, 455, 469,
plastic deformation 496. See also gap-by-gap hardening
avoiding, 8, 30 puzzle piece carbide, 108(F)
brittle fracture and, 34, 53, 54(F), 120
buckling and, 128
creep and, 128 Q
ductile fracture and, 51, 96, 120
formation of, 7, 52, 286287, 331, 446 quench cracking, 5965(F)
fracture surface matching, 118 definition, 5960
shot peening and, 199(F) intergranular fracture, 96
plastic mold steels, 313(F) localized overheating, 62
plate martensite, 187, 188(F) martempering, 60
polishing, 51, 113, 313(F), 390(F), 482 mitigation of, 60
porosity stress raisers role in, 272273
blowholes, treating, 156 quench cracking, case studies
casting component feeding, 159 as-quenched 4340 steel, 273274(F)
effects due to, 154162(F), 163(F) cracking of 4140 block forging after quenching and
gas, 154156(F) tempering, 274275(F&T)
internal, from nucleation, 158159 decarburization and oxidized grain boundary,
internal, from the surface, 158, 159(F) 281283(F)
pinholes, treating, 156 network carbides and coarse grain size, 278280(F)
shrinkage pores, 156158(F) presence of a seam defect, 276(F)
postweld heat treatment (PWHT), 37 presence of chemical segregation, 278, 279(F)
powder metal steel components presence of slag inclusions and a lap
case hardening defect, 276278(F&T)
carbonitriding, 397398(F) presence of stringer inclusions and chemical
carburizing, 397 segregation, 280281(F&T)
induction hardening, 398(T) use of improper steel alloy and presence of voids in a steel
steam treatment, 398399(F) brazed joint, 275276(F&T)
definition, 395, 396(F) quench nonuniformity, 189
failure analysis techniques, 399401(F&T) quenching
CE matrix, 400401, 402403(F) carbonitrided components, 191
failure mode and effects analysis (FMEA), 399 carburized components, 185, 187191
process maps, 400, 401(T) component design, 263265(F), 266(F&T)
steps, 400(F) contamination, 190
tools, 400(F) dry die quenching, 18
powder metallurgy process, 395396, 397(F) failures due to, 255284(F&T)
powder metal steel failures, case studies immersion quenching, 451
blistered sintered products, 405406 induction hardening, 449452(F)
cracks after induction hardening, 409410(F) martempering, 1718
cracks after quenching, 410411(F) phase transformations, 255263(F&T)
dimensional change in carbonitriding, 411, 412(F) press quenching, 18
dimensional change in steam treatment, 414(F), quench nonuniformity, 189
415(F) quenchant nonuniformity, 271272(F)
dimensional instability during shrink fitting, 406407 quenchant selection, 270271(T)
fracture of steam-treated part, 407408(F) retained austenite, 191196(F&T)

638 / Index
quenching (continued) soft skin layers, 199200

slack quenching, 278 softening replica tape, 4950
spray quenching, 430(F), 444(F), 451452 solid metal embrittlement, 6768(F&T), 69(F)
steel grade/condition, 266270(F&T) solid shrinkage, 158
atmosphere control, 269 solidification cracking, 515517(F)
component support/loading, 267(F), 268(F) solidification shrinkage, 157158(F)
heating control, 269 solid-phase transformation model, 3132
machining, 267 spalling failure, 20(F), 23(F), 138, 231, 322(F), 323
retained austenite, 269270(T) spheroidization, 107(F)
surface condition, 267269(F&T) spray quenching, 430(F), 444(F), 451452
quenching cracks, 204207(F), 341 stainless steels, PWHT of, 3740
austenitic chromium-nickel, 3839
chromium steels, 37
duplex austenitic-ferritic chromium-nickel, 3940
R fully austenitic, 39
martensitic, 38
ratchet mark, definition, 505506
soft martensitic, 38
ratchet marks, 126(F), 388(F), 389(F), 506(F)
stabilized, 39
recognized and generally accepted good engineering practice
unstabilized austenitic, 39
(RAGAGEP), 40
static fatigue test, 301
rehardening burn, 207
static tests, 301
reoxidation, 166167
steady-state creep rate, 77, 139
residual stresses, 196200(F&T), 472477(F)
steel hardness conversions
retained austenite, 191196(F&T), 269270(T)
approximate equivalent hardness numbers for Brinell
risers, 158
hardness numbers for steel, 532533(T)
river patterns, 54, 55(F), 97(F), 376(F)
approximate equivalent hardness numbers for
rolling, 130
steel, 535(F)
hot rolling, 89
approximate equivalent hardness numbers for Vickers
surface rolling, 122, 123(F), 124
(diamond pyramid) hardness numbers for steel,
533534(T)
approximate Rockwell B hardness conversion numbers for
S nonaustenitic steels, 530(T)
approximate Rockwell C hardness conversion numbers for
scanning electron microscope (SEM), 21, 50(F) nonaustenitic steels, 531(T)
shank crack, 388, 389(F) examples of published hardness conversion equations,
sharp asperity, 130(F) 529(T)
shear lip, 50, 51, 118, 119(T), 273 steel susceptibility ratio, 285
Sherlock Holmes rule, 114 steel welds, failure analysis of
shot blasting, 198 discontinuities, 503505(F&T)
shot peening, 198199(F), 200(T) examples
shrinkage pores, 159, 160(F) fatigue cracking of welded pipe flange, 506(F)
examples: failure analysis of a mill gear, 159161(F&T), lack of penetration, 505
161162(F), 163(F), 164(F&T) lack of side-wall fusion, 505
silicon, 56, 183, 223, 293, 296, 297(F), 334 porosity in weld metal, 504, 505(F)
siliconizing, 33 SCC of a weld, 514515(F)
single-shot induction hardening, 422, 423, 424(F), 427, solidification cracking of steel weld, 516(F), 517(F)
428(F), 440(F), 496 fatigue, 505506(F)
single-shot induction heating, 424(F), 448(F) hydrogen-assisted cracking (HAC), 509513(F&T)
sintering hydrogen-assisted cracking theory, 506509(F&T)
ferrous PM parts, 397 solidification cracking, 515517(F)
oxidation after, 408409(F) stress-corrosion cracking (SCC), 513515(F)
pinholes after, 404405, 406(F) steels
PM process map, 401(T), 402(F) alloyed, tempering, 293, 294(F&T), 295(F)
wear after, 407, 408(F) carbon. See carbon steels
slack quenching, 278 carbon, isothermal diagrams
slant fracture, 56, 505 1019, 588(F)
sliding, 56, 128(F), 129, 130(F), 195, 196(F), 199 1030, 589(F)
slip band extrusions, 78, 80(F) 1050, 590(F)
slip band intrusions, 78, 80(F), 122 1060, 591(F)
Smith curve, 247248, 250(F) 1080, 592(F)

Index / 639
carbon steels with nominal carbon contents stress-corrosion cracking (SCC), 28

0.18 C (10171022), 602(F) austenitic stainless steels, 36
0.44 C (10391046), 603(F) intergranular brittle fracture, 70, 7276(F&T), 77(F)
0.86 C (10801090), 604(F) stress-sorption theory, 514
0.96 C (10901095), 605(F) striations, 78, 80, 97, 124, 506. See also fatigue striations
carburized. See carburized steels stringer inclusion, 166, 280281(F)
chromium, 3739 stringers, 25(F), 103(F), 280, 281(F)
1 Cr (5140), 624(F) structural flaw, 241
1 Cr V, (6150), 626(F) subcritical annealing, 16, 266267
11/4 Cr Mo (41404142), 625(F) sudden onset deformation, 89
11/2 Cr Al Mo, 627(F) sulfidizing, 33
21/2 Cr (5060) (51555160), 623(F) superposition principle, 127
5160, 597(F) surface (or skin) effect, 440
52100, 599(F) surface carbon content
chromium-molybdenum decarburization, 213, 215217(F)
4130, 593(F) overcarbonitriding, 211, 213, 214(F), 215(F)
4140, 594(F) overcarburizing, 211, 213, 214(F), 215(F)
clean, 165166, 229 surface fatigue wear, 138
cold work tool, 312(F) surface integrity, 418, 475
continuous cooling diagrams, 601627(F) surface rolling, 122, 123(F), 124
high-speed, 313314 surface structure anomalies, 199200
hot work tool, 312313(F) Systeme International dUnites (SI), 521
induction hardening, 419420(F&T)
interstitial-free, 124, 125(F)
isothermal diagrams, 587599(F) T
low-alloy steels. See low-alloy steels
Mn TBE, 299
11/2 Mn+S (1139), 607(F) temper colors for steels
13/4 Mn (1547) (1345), 606(F) color chart (back cover)
Mn-Cr, 11/4 Mn Cr, 609(F) time-temperature effect, 539, 540(F)
Mn-Mo, 11/2 Mn Mo, 608(F) temper embrittlement (TE), 296299(F)
Mn-Ni-Cr-Mo, 11/2 Mn Ni Cr Mo, 611(F) HEM, interaction with, 301302(F)
Mn-Ni-Mo, 11/2 Mn Ni Mo, 610(F) LME, interaction with, 302303(T)
Mo, 1/4 Mo (40374042), 615(F) mechanical tests for, 299301(F&T)
nickel mechanism of, 6566(F)
1/2 Ni Cr Mo (86228627) (8720) (8822), 620(F) molybdenum, effect on, 297298
11/4 Ni Cr, 619(F) phosphorus, effect on, 297298
11/2 Ni Cr Mo, 621(F) vanadium, effect on, 298
11/2 Ni Mn, 617(F) welded components, 3435
13/4 Ni Mo, 618(F) temperature conversion table, 525528
31/2 Ni, 616(F) tempered martensite embrittlement, (TME), 65(F),
31/2 Ni Cr Mo (9310), 622(F) 294296(F)
nickel-chromium-molybdenum, 8620, 598(F) tempering
nickel-molybdenum, 4640, 596(F) alloyed steels, 293, 294(F&T), 295(F)
Ni-Cr-Mo, 4340, 595(F) case study: grinding cracks, 303304(F)
nitrided, 244246(F&T), 247(F&T) case study: transgranular and intergranular crack path,
plastic mold, 313(F) 304305(F&T), 306(F)
as quenched 4340, 273274(F) colors of tempering heats, 289(T)
silicon embrittlement, 293303(F&T)
13/4 Si Mn, 612(F) heating times, 290(T)
2 Si Cr Mo, 614 martensite, 285289(F)
31/2 Si Cr, 613(F) mechanical properties, effect on, 289290(F), 291(F),
tool, 311314(F&T), 314330(F&T) 292(F)
stress reactions, 290291
analysis, 115116 stages, 291293(F)
concentrations, 3233, 45, 46(F), 98(F) tempering embrittlement, 285
machine parts, 485491(F&T) tempering resistance, 36(F), 313, 331, 334, 338, 339
versus strength, 9395(F) tensile, definition, 92
types of, 78 thermal expansion
stress raisers, 179(F), 272273 austenite versus ferrite, 5

640 / Index
thermal expansion (continued) U

austenitic stainless steels, 39
carbon and low-alloy steels, specific heats of, 549550(T) ultrasonic testing (UT), 512
coefficients of, 45, 6(F), 9(F), 32 underbead cracking, 507, 510511(F)
coefficients of linear thermal expansion for carbon and undercut, 20(F), 46(F), 503(T)
low-alloy steels, 541543(T) undercutting, 4546(F)
ferrous materials, 5(T) unnotched impact testing, 337
heat capacity, summary of, 548549(T) upper bainite, 2, 270, 306(F), 403(F)
induced strain, 7 upset forging, 143145(F&T), 146(F)
residual stress, influence on, 5, 32
summary of, 544545(T)
thermal conductivities of carbon and low-alloy steels,
547(T) V
thermal conductivity, summary of, 545546(T)
thermal shock, 56(T), 32 volume wear, 138
thermochemical modeling, 33
thermomechanical modeling, 3233
through hardening, 395, 396, 413, 421, 425(F),
W
442(F), 490
through-hardened, 413(F), 484, 485(F)
wear, 9799(F)
time-temperature transformation (TTT) diagrams, 2, 32,
wear, definition, 129
202(F), 255, 256(F)
wear-assisted failure, 129131(F)
time-temperature-austenitizing (TTA) diagram, 432(F),
welded components, failure aspects of
433(F)
aging tendency, 35
toe cracking, 503(T), 504(F)
austenitic stainless steels, SCC of, 36
tool steels, failure analysis in
brittle fracture, 3334
classification of, 311314(F&T)
cold cracking, 36
cold work, 312(F)
hardening tendency, 3536
cold work tools, 314330(F&T)
segregation tendency, 36
high-speed steels, 313314(F)
temper embrittlement, 3435
hot work, 312313(F)
welded components, heat treatment procedures
hot work tools, 330349(F&T)
austenitic-ferritic dissimilar joints, 40
plastic mold steels, 313(F)
carbon steels, 3637
tooth-by-tooth induction hardening, 430(F), 439(F)
low-alloy steels, 3637
tooth-gap hardening, 431432(F)
stainless steels, PWHT of, 3740
torsional fatigue, 177, 483
total oxidation potential (TOP), 221(F)
total-life fatigue analysis, 177
transition temperature, 35, 55, 66, 67(F) Y
transmission electron microscopy, 65, 220, 288
tribological systems, 475 yield point, 483(T)
twinning density, 288 yield strength, 144(T), 161(T), 250(T), 507
two-step embrittlement, 296 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
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
Customer Service, ASM International
Mail
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 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Failure Analysis of Heat Treated Steel Components PDF

Uploaded by

Copyright:

Available Formats

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_i.

pdf/Front Matter_Title-page/ 12/8/2008 8:00PM Plate # 0 pg i

2008 ASM International. All Rights Reserved.

2008 ASM International. All Rights Reserved.

2008 ASM International. All Rights Reserved.

This book is dedicated to our families, without whose continued

My husband, Antonio Carlos Canale,

To my lovely wife, Carla Mesquita, and my dear son, Rafael

2008 ASM International. All Rights Reserved.

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

2008 ASM International. All Rights Reserved.

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

2008 ASM International. All Rights Reserved.

Steel Heat Treatment Failures due to Quenching ........................................................... 255

Failure Analysis in Tool Steels ....................................................................................... 311

Case Studies of Steel Component Failures in Aerospace Applications ............................ 351

Failure Analysis of Powder Metal Steel Components ..................................................... 395

Induction Hardening ..................................................................................................... 417

2008 ASM International. All Rights Reserved.

Main Features of Induction Heating ................................................................................. 420

Failure Analysis of Steel Welds ...................................................................................... 503

Appendix 1: Metric Conversion Guide .......................................................................... 521

Index ............................................................................................................................. 629

2008 ASM International. All Rights Reserved.

Lauralice C.F. Canale, Ph.D.

ASM International, Materials Park, Ohio, USA

To order products from ASM International:

Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US)

DESIGN involves different creative aspects: involve a complex interrelationship of variables

2 / Failure Analysis of Heat Treated Steel Components

Table 1 Heat treating process for carbon steels

4 / Failure Analysis of Heat Treated Steel Components

expansion coefficient, and E is the elastic mod- Steel mm/m in./ft

6 / Failure Analysis of Heat Treated Steel Components

(90) and low thermal expansion

8 / Failure Analysis of Heat Treated Steel Components

10 / Failure Analysis of Heat Treated Steel Components

12 / Failure Analysis of Heat Treated Steel Components

Fig. 10 Design solutions to the distortion problem shown in

Fig. 8 Drive shaft pinion with fatigue fractures propagating

14 / Failure Analysis of Heat Treated Steel Components

Finally, part design recommendations to fillets wherever possible.

16 / Failure Analysis of Heat Treated Steel Components

poorest loading technique would be to pile them Fog or mists

18 / Failure Analysis of Heat Treated Steel Components

Fig. 17 (a) Front view of an AISI O1 tool steel die that

20 / Failure Analysis of Heat Treated Steel Components

electrical discharge machining (EDM) techni-

22 / Failure Analysis of Heat Treated Steel Components

(a) (b) (c)

24 / Failure Analysis of Heat Treated Steel Components

occurs at relatively low temperatures, de-

Fig. 24 (a) AISI S5 tool steel hammer head that cracked

(a) (b) (c)

Fig. 28 Schematic of wrist pin. Note stress raisers at A and

from the head of the screw. A crack was

26 / Failure Analysis of Heat Treated Steel Components

downtime. This, and other contributing factors, Nondestructive testing