Sie sind auf Seite 1von 301

Copyright © 2008 ASM International®. All rights reserved.

Stainless Steels for Design Engineers (#05231G)



ASM International®
Materials Park, Ohio 44073-0002
Copyright © 2008
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, December 2008

Great care is taken in the compilation and production of this book, but it should be made clear that NO WAR-
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 infor-
mation. 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
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 reproduc-
tion, in connection with any method, process, apparatus, product, composition, or system, whether or not cov-
ered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a de-
fense 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 (2007–2008), Lichun L.
Chen, Chair.
ASM International staff who worked on this project include Scott Henry, Senior Manager of Product and Service
Development; Steven R. Lampman, Technical Editor; Eileen De Guire, Associate Editor; Ann Britton, Editorial
Assistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production Coordinator; Diane
Grubbs, Production Coordinator; Patty Conti, Production Coordinator; and Kathryn Muldoon, Production Assistant
Library of Congress Control Number: 2008934669
ISBN-13: 978-0-87170-717-8
ISBN-10: 0-87170-717-9
SAN: 204-7586
ASM International®
Materials Park, OH 44073-0002
Printed in the United States of America

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

Chapter 1 Metallurgy........................................................................................................1


Chapter 2 Corrosion Theory............................................................................................11

Chapter 3 Corrosion Kinetics..........................................................................................19

Chapter 4 Corrosion Types..............................................................................................27

Chapter 5 Oxidation.......................................................................................................57


Chapter 6 Austenitic Stainless Steels ..............................................................................69

Chapter 7 Duplex Stainless Steels...................................................................................91

Chapter 8 Ferritic Stainless Steels .................................................................................109

Chapter 9 Martensitic Stainless Steels ..........................................................................123

Chapter 10 Precipitation-Hardening Stainless Steels ......................................................137

Chapter 11 Casting Alloys...............................................................................................147

Chapter 12 Melting, Casting, and Hot Processing...........................................................155

Chapter 13 Thermal Processing ......................................................................................161

Chapter 14 Forming........................................................................................................173

Chapter 15 Machining ....................................................................................................181

Chapter 16 Surface Finishing ..........................................................................................193

Chapter 17 Welding........................................................................................................201

Chapter 18 Architecture and Construction.....................................................................213

Chapter 19 Automotive and Transportation Applications................................................225

Chapter 20 Commercial and Residential Applications ....................................................233

Chapter 21 Marine Systems Applications........................................................................243

Chapter 22 Petroleum Industry Applications ..................................................................247

Chapter 23 Chemical and Process Industry Applications ................................................257

Chapter 24 Pulp-and-Paper Industry Applications ..........................................................265

Appendix 1 Compositions...............................................................................................269

Appendix 2 Physical and Mechanical Properties of Select Alloys....................................279

Appendix 3 Introduction to Thermo-Calc and Instructions for

Accessing Free Demonstration ....................................................................281

Index .................................................................................................................................285


The rate of growth of stainless steel has outpaced that of other metals and alloys, and by 2010 may
surpass aluminum as the second most widely used metal after carbon steel. The 2007 world produc-
tion of stainless steel was approximately 30,000,000 tons and has nearly doubled in the last ten years.
This growth is occurring at the same time that the production of stainless steel continues to become
more consolidated. One result of this is a more widespread need to understand stainless steel with
fewer resources to provide that information. The concurrent technical evolution in stainless steel and
increasing volatility of raw material prices has made it more important for the engineers and design-
ers who use stainless steel to make sound technical judgments about which stainless steels to use and
how to use them.
This book provides design engineers with an up-to-date source of information at a level useful for
both metallurgists and other engineers and technicians. It seeks to bridge the gap between the inter-
net where much current, but raw information is available and scholarly books and journals that pro-
vide theory that is difficult to put into practice. The content of the book is selected for utility for the
user of stainless steel. The first section gives elementary metallurgy and identification of constituents
of stainless, the effects of alloying elements and a significant section on corrosion. A second section
is oriented toward processes important to users of stainless steel. The third section is about each fam-
ily of stainless alloys and includes the most recent additions that have come to the market. The fourth
section deals in some depth with the major applications for stainless steel. This last part is presented
without the promotional bias which is found in many steel producers’, alloy producers’, and trade as-
sociations’ literature. While a number of steel producers have provided assistance to the author, there
has been no attempt to unfairly bias information in their favor. To the contrary, those producers re-
sponsible for generating factual, useful data for the user community are those who should benefit the
most by books such as this. The author is particularly indebted to Allegheny Ludlum and John
Grubb, and his many colleagues who assisted him, for technical assistance throughout the writing
and to Carnegie Mellon University for their support. The author also wishes to thank Professor Srid-
har Seetharaman at Carnegie Mellon University for his help in writing the corrosion chapter and oth-
ers who helped: Roy Matway of CMU, Vittorio Boneschi of Centro-Inox; Paul Mason of Thermo-
Calc; Bob Drab of Schmolz Bichenbach; Elisabeth Torsner and Chuck Turack Outukumpu, USA;
Scott Balliett of Latrobe Steel; Jim Halliday and Fred Deuschle of Contrarian Metals Resources; Pro-
fessors Tony DeArdo of Pitt and Gerhard Welsch of CWRU; the staffs of Centro-Inox, Euro-Inox,
SSNA, The Nickel Institute; and the editorial staff at ASM International, Scott Henry, Eileen
DeGuire, Charlie Moosbrugger and Steve Lampman. I would also like to thank the many members
of my forum at who have contributed much collective knowledge and perspective to
this book.
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
Stainless Steels for Design Engineers #05231G

To order products from ASM International:

Online Visit

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

Fax 1-440-338-4634
Customer Service, ASM International
9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
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)
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.
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
Stainless Steels for Design Engineers #05231G

To order products from ASM International:

Online Visit

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

Fax 1-440-338-4634
Customer Service, ASM International
9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA
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)
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.
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 1-10 All rights reserved.
DOI: 10.1361/ssde2008p001



Summary phases coexisting simultaneously. Stainless

steel is an exceptional alloy system in that it is
COMPARED TO ALLOY STEELS, stainless not a dilute solution. Alloy steels may contain
steels are chemically complex. The large number several percent of alloying elements, such as
of alloying elements makes possible a larger carbon, manganese, nickel, molybdenum,
range of possible phases or basic crystal struc- chromium, and silicon, in addition to the impu-
tures. The large amount of the alloying elements rities sulfur, oxygen, and phosphorus. Alloy
makes the deviation from the behavior of pure steels typically contain very small amounts of
iron greater; consequently, the calculations that titanium, niobium, and aluminum. The total
predict which phases will exist are more difficult. amount* of these alloying elements seldom ex-
The three basic phases of stainless steels are fer- ceeds 5%. The same is true for most aluminum
rite, austenite, and martensite. The wide variety alloys. In contrast, stainless steels contain no
of alloys that exist is based on: less than about 11% chromium alone. Most
stainless alloys have manganese, silicon, car-
• Combinations of these phases
bon, and nickel in thermodynamically mean-
• Altering the composition of these phases
ingful amounts as well as large concentrations
• Adding secondary phases for particular
of nickel and/or molybdenum.
purposes The result of the large number of alloying ele-
Metallurgy, as discussed in this chapter, fo- ments in relatively high concentrations is that
cuses on phases normally encountered in stain- stainless steel can have many stable phases con-
less steels and their characteristics. In subsequent currently. In almost every case, having phases
chapters on types of stainless steel, there are other than the principal one or two phases for
more detailed treatments of the alloys made of which the alloy was designed is undesirable be-
these phases and their properties. cause of the possibility of undesirable variations
in mechanical or corrosion performance. The
producer of stainless steel controls the chemical
composition and thermomechanical processing,
Introduction so that when the processor or end user receives
the product it is usually in the correct condition.
Most widely used alloy systems, such as car-
However, subsequent processing or service con-
bon steels, alloy steels, and aluminum alloys, are
ditions may alter the carefully established phase
relatively dilute solutions of several elements in
structure. Therefore, it is necessary to discuss
the parent matrix. Carbon and alloy steels, with
the phases that can exist in stainless steel and
very few exceptions, are principally of the mag-
the conditions under which they form so that the
netic body-centered cubic (bcc) phase or a
enlightened user will know which phases to
slightly distorted version of it. Aluminum alloys
avoid and how to avoid them.
share the face-centered cubic (fcc) structure of
It is possible to use thermodynamics to calcu-
pure aluminum. A given structure, which can
late which phases may exist at a given tempera-
have a certain range of compositions, is what is
meant by a phase, just as a gas or liquid is a * All compositions are given in weight percent unless
phase. In solid metals, there can be a number of stated otherwise.
2 / Stainless Steels for Design Engineers

ture for a given composition. It is not remotely not a small, pedantic point. Most stainless steels
feasible, however, to give an adequate treatment are used in the metastable condition. For exam-
of the thermodynamics required to do this. The ple, the common alloy 304 (also called 18-8) is
topic alone requires a book. The necessary normally used in the fully austenitic condition.
knowledge has been embedded in proprietary It would “rather” be partly ferritic, but the sub-
computer programs that will be used instead. stitutional diffusion of chromium in austenite
that is required to form a ferrite phase of a sepa-
rate composition is so slow that it cannot occur
Thermodynamics of Stainless Steel in terrestrial time frames. However, if energy is
applied by mechanical shear, the austenite can
Pure metals, from a practical viewpoint, are transform without diffusion to the lower free-
either liquid or solid depending on temperature, energy martensite phase, a quasi-bcc structure
with the possibility of some trivial small gas of lower free energy.
vapor pressure. A law of thermodynamics is that The calculation of which phases exist under
the number of possible condensed (i.e., solid) equilibrium conditions proves to be extraordi-
phases equals the number of elemental con- narily difficult in complicated alloy systems.
stituents plus one. The solid has a crystallo- This is because thermodynamic values can be
graphic structure that may vary with tempera- measured accurately only in the liquid state, so
ture. Many metals have a less-dense bcc the values for the solid state are extrapolations.
structure at high temperature and transform to a Also, the interaction between elements is very
denser fcc structure at lower temperatures. Iron important in nondilute alloys such as stainless
does this. Iron has the curious characteristic of steel. Consequently, most published phase dia-
transforming from fcc back to the low-density grams are experimentally derived. To determine
bcc at still lower temperatures. This is a result of which phases exist at a given composition and
the unpaired 3d orbital electrons (those that give temperature, a sample is made, equilibrated at
rise to ferromagnetism) that are not given up as the appropriate temperature, and quenched to
valence electrons, causing repulsive forces be- room temperature. It is assumed that the charac-
tween atoms and requiring a more widely spaced teristic equilibrium phases have been frozen and
structure. are then identified by various techniques for
All thermodynamic properties are based on structure, composition, and the like. This impor-
interatomic attractions. In metals, the metal tant work is obviously tedious and susceptible
atoms give up valence electrons to the entire to experimental error and applies only to spe-
mass. These electrons are of varying energy cific compositions. Any “what if” extrapolation
states and highly mobile. They are responsible to a different alloy composition carries the risk
for the ability of metals to conduct heat and of error.
electricity well. The attraction, the strength of A practical tool has been developed that per-
the bond, is proportional to the charge differ- mits phase diagrams to be calculated for arbitrary
ence and distance. The attraction determines compositions. These are computer simulated,
such macroscopic properties as melting temper- mathematical models that can perform the com-
ature, density, and elastic modulus. plex thermodynamic calculations. To do this with
In this book, the main concern with thermo- accuracy requires databases of thermodynamic
dynamics is predicting which phases are present values. These values must be derived from com-
both at equilibrium and in the quite frequent puter analysis of experimental phase equilibrium
metastable state. The prediction involves calcu- diagrams. They are expensive to derive and vali-
lating the free energy of the various possible date, and only a few exist. Hence, they are pro-
phases. The phase with the lowest energy is prietary. In Appendix 3, a license to one such
most favored, but others may have free energies program, Thermo-Calc, can be found. The ver-
that permit them to exist. The difference be- sion has a reduced three-element capability but
tween these two is that the equilibrium state, uses the same proprietary thermodynamic data-
that of the lowest free energy, may require base of the full version. The program allows de-
atomic rearrangements to occur for equilibrium termination of which phases can exist for any
compositions to be reached on an atomic scale. composition and temperature. Whether the
If diffusion is too sluggish for these rearrange- phases will form depends also on kinetic factors.
ments to take place, the structure may retain the First, however, it is good to become familiar with
prior metastable structure indefinitely. This is the principal phases found in stainless steel.
Chapter 1: Metallurgy / 3

Phases and oxygen, have no influence on which phase

is favored. Again, it must be emphasized that
Ferrite the influence of an alloying element on structure
has zero bearing on its influence on corrosion
The basis of stainless alloys is, of course, resistance.
iron. Iron, as stated, solidifies as a bcc alloy be- The elements that promote ferrite over austen-
fore transforming to the denser fcc austenite at ite also have the effect, at still lower tempera-
lower temperatures. At still lower temperatures, tures, of promoting intermetallic compounds
it reverts to the bcc structure. It is accurate to generally composed of iron, chromium, and
surmise from this that the free energy of both some of those alloying elements. These are dis-
structures is close. Alloying elements that pro- cussed separately.
mote one structure over the other can therefore Metals are effective solvents in both the liq-
change which one predominates. The element uid and solid states. An important part of steel-
that produces the ability to form the passive film making is refining the molten metal to remove
that makes stainless corrosion resistant, the undesired impurities dissolved in it. The
chromium, has the characteristic of stabilizing normal technique is to add elements that react
the bcc structure. As chromium is added to iron, selectively with the targeted impurities to form
the temperature range over which austenite is an immiscible reactant that can become part of
stable grows smaller until, at about 12% the slag and physically separated from the re-
chromium, ferrite is stable at all temperatures. fined alloy. This is done for the primary impuri-
This is, coincidentally, the approximate level of ties oxygen and sulfur. A third common impu-
chromium needed to keep alloys from rusting rity, phosphorus, is not so easily removed and
under ambient conditions, but this effect is not must be excluded from raw materials to be kept
related to whether the structure is bcc or fcc. under control.
The iron-chromium phase diagram (Fig. 1) In stainless steel, carbon and nitrogen can be
shows the composition and temperature regions detrimental impurities. Both are quite soluble
where ferrite (a), martensite (α' ), austenite (γ), in molten iron-chromium alloys and are fairly
and sigma phase (σ) are stable. soluble in ferrite at high temperatures. This sol-
While chromium is the principal ferrite-pro- ubility decreases exponentially with tempera-
moting alloying element, other elements have ture so that it is essentially zero at room tem-
similar effects, but none produces the quality of perature. These elements have small atomic
stainlessness. Silicon, aluminum, molybdenum, sizes compared to iron and chromium and,
tungsten, niobium, and titanium all favor ferrite. when dissolved, squeeze into interstitial sites
Carbon, nitrogen, manganese, nickel, and cop- within the bcc matrix. Such interstitial solute
per do not and expand the temperature range atoms profoundly distort the structure. They are
over which austenite exists. Elements that are much more soluble in the fcc structure, which,
insoluble in iron at austenite-forming tempera- while denser, has roomier interstitial spaces, so
tures, such as the impurities phosphorus, sulfur, they stabilize that structure. To preserve the fer-
rite structure, carbon and nitrogen must be
There are additional reasons to eliminate car-
bon and nitrogen. During cooling as these ele-
ments become less and less soluble, they must
precipitate. The most thermodynamically favor-
able form in which they can precipitate is as a
compound of chromium, with which they are
very reactive. This occurs at the grain bound-
aries, where nucleation is favored, and depletes
those regions of chromium, rendering them less
corrosion resistant. A second effect is a loss of
toughness due to these precipitates. The diffusion
rates of carbon and nitrogen in ferrite are too
high to prevent this precipitation by quenching.
Fig. 1 The iron chromium phase diagram. Courtesy of Modern refining methods can reduce carbon
Thermo-Calc Software plus nitrogen to under 0.020%, but even this is
4 / Stainless Steels for Design Engineers

too high. So, to avoid the detrimental effects of The bcc structure of ferrite allows more
chromium carbide and nitride formation in fer- rapid diffusion than does the fcc structure of
rite, other benign carbides and nitrides such as austenite. This is true for both the interstitial
those of titanium or niobium are allowed to form diffusion of the elements helium, boron, car-
preferentially. This approach is called stabiliza- bon, nitrogen, and oxygen and the substitu-
tion and is used for most ferritic alloys today. The tional diffusion of all other elements. The rate
older approach, as characterized by alloy 430, is of diffusion of all elements, both interstitial
to permit chromium carbides and nitrides to form and substitutional, in ferrite is about two or
but then to perform a subcritical anneal to reho- three orders of magnitude higher than in
mogenize the chromium and coarsen the pre- austenite. The practical implication of this is
cipitates so that they have only a small negative that precipitation reactions generally cannot be
effect on mechanical properties. suppressed by quenching in ferrite if they in-
Hydrogen and boron are other elements that volve interstitial elements, whereas they can
can be interstitially dissolved in ferrite. Boron is be in austenite. Intermetallic phases can form
normally found at levels of around 5 to 10 ppm. more rapidly in ferrite. This becomes an issue
At higher levels, boron substitutes for carbon in only when total chromium plus molybdenum
carbides. Hydrogen is soluble to several parts exceeds about 20%, above which the sigma
per million by weight. It does not cause hydro- phase appears. This is thus only an issue for
gen embrittlement in annealed ferrite. If the fer- superferritic (high chromium content) alloys or
rite is cold worked, the solubility of hydrogen for the ferrite phase of duplex (ferrite-austen-
increases as the defect structure accommodates ite) alloys.
hydrogen atoms. In this condition, ferrite may The mechanical properties of the ferrite phase
be embrittled by hydrogen, especially if it en- are discussed extensively in Chapter 8, “Ferritic
ters the metal through corrosion processes like Stainless Steels.” Here, it is only necessary to
pitting. This is one explanation of, and the most note that ferrite in stainless steel closely resem-
likely explanation for, stress corrosion cracking. bles low-carbon steel in mechanical behavior. It
While hydrogen is easily removed by argon shares the following characteristics:
oxygen decarburization (AOD), assuming ab- • A toughness transition that occurs around
solutely dry blowing gases and additions are
room temperature
used, it can be picked up during pickling, weld- • Notch sensitivity
ing, or annealing as well as by corrosion. • A yield point phenomenon
All stainless alloys rely on having a uniform • Pronounced crystallographic anisotropy of
level of chromium and the other element,
mechanical properties
molybdenum, which assists in corrosion resist- • High stacking fault energies and low work-
ance, distributed throughout the matrix. If there
hardening rates
are locally low levels of these elements, local-
ized resistance to corrosion is reduced, and lo- These issues are dealt with in the same way as
calized corrosion can occur. This can occur by in carbon steel when these characteristics be-
the precipitation of any phase that is richer in come an issue. The first two are controlled by
chromium or other corrosion-resisting ele- reduction of interstitial levels and refining of
ments. Because chromium is a reactive ele- grain size. The yield point is eliminated by
ment, its success depends to a great degree on slight elongation by temper rolling or elimina-
maintaining the homogeneity required for tion of interstitial carbon and nitrogen, whose
proper corrosion-resistant performance. Incor- interaction with dislocations causes the yield
rect thermal processing is the main way homo- point. The anisotropy is either utilized to advan-
geneity can be lost. Stabilizing makes it much tage by maximizing it, as in the case of deep-
easier to keep chromium from segregating in drawing alloys, or minimized by refining grain
ferritic alloys. size and randomizing grain orientation by spe-
A by-product of stabilization with titanium is cial thermomechanical processing.
that oxygen and sulfur are also eliminated as Ferrite has a greater thermal conductivity
compounds of titanium along with carbon and and lower thermal expansion than austenite. Its
nitrogen. These impurity elements would other- strength decreases with temperature more than
wise also precipitate as compounds containing that of austenite, but the good match in thermal
some chromium, potentially depleting chromium expansion between the ferrite and its oxide
in the vicinity of their precipitation. still makes it an excellent high-temperature
Chapter 1: Metallurgy / 5

material. Ferrite has very nearly the same cor- basis. Their use is limited by their solubility and
rosion resistance as austenite, but since ferrite their tendency to form precipitating compounds
can hold no nitrogen in solution, it cannot ben- with chromium. Manganese acts largely through
efit from this element. In duplex alloys, the its ability to promote nitrogen solubility. Super-
ferrite is generally the more corrosion resistant austenitic stainless steels, such as S34565, use 4
phase because it is richer in chromium and to 6 % manganese to permit nitrogen levels of
molybdenum. 0.4 to 0.6% to be achieved, resulting in higher
pitting corrosion resistance.
Since all stainless steels contain principally
iron and chromium, the addition of a substantial
The second major constituent phase of the amount of austenitizing elements is necessary to
stainless steel alloy system is austenite. Austen- transform the structure to austenite. As a rule of
ite has an fcc atomic structure. The fcc structure thumb, iron alloys require about 17% chromium
is common in many transition metals to the right and 11% nickel (or its equivalents) to remain
of iron in the periodic table. As stated, the fcc austenitic at room temperature. One percent
structure should be considered normal for metals nickel can be replaced by about 2% manganese
well below their melting temperature as it is a as long as nitrogen is present to maintain the
denser structure. The presence of the bcc struc- same phase stability. The omnipresent carbon
ture relates to the unpaired 3d electrons, which and nitrogen have an effect 30 times that of
provide ferromagnetism. Adding elements to nickel, so even in the small amounts in which
iron that causes pairing of the 3d electrons di- they are normally present, they have a signifi-
minishes ferromagnetism and promotes the fcc cant effect. These stabilizing factors are mapped
structure. Nickel and manganese are the most in the Schaeffler diagram of Fig. 2 (Ref 1),
prominent alloying elements that do this, but the whose purpose is to predict the phase makeup
interstitials carbon and nitrogen are the most of weld metal. Since welds solidify relatively
powerful austenite stabilizers on a percentage rapidly, no carbides or intermetallic phases

Fig. 2 Schaeffler-Delong constitution diagram showing phases present in as-solidified stainless steels at room emperature as a func-
tion of composition demonstrating carbon and nitrogen contributions to nickel effects. Adapted from A.L. Schaeffler,
Constitution Diagram for Stainless Steel Weld Metal, Met. Prog., Vol 56, Nov 1949, p 680–688; and W.T. Delong, A Modified Phases
Diagram for Stainless Steel Weld Metals, Met. Prog., Vol 77, Feb 1960, p 98
6 / Stainless Steels for Design Engineers

form, and only ferrite, austenite, and martensite and then frozen in a state of supersaturation in
will be present. Thus, they provide useful infor- the austenite when it forms on cooling. The sul-
mation about the compositional effects on phase fur and oxygen then precipitate during cooling or
development in nonequilibrium situations. The subsequent hot working as isolated inclusions.
nickel equivalent (vertical axis) summarizes The interface between these inclusions and the
how nitrogen, carbon, and other elements com- matrix is the locus of corrosion pit initiation,
bine to create a nickel-like effect. The horizon- quite probably because of chromium depletion
tal axis does the same for chromium and those occurring during and as a result of inclusion
elements that have a similar effect. growth. When an alloy solidifies as austenite,
In most common stainless steels, austenite is sulfur immediately segregates to the grain
normally present in the metastable state, for ex- boundaries because of its low solubility in
ample, the retained austenite in alloy steels. Those austenite, and it forms a low-strength film with a
with carbon above 0.02% would eventually break low melting temperature. This causes poor hot
down into austenite plus carbides, and those with workability and hot cracking of welds.
less than about 30% chromium plus nickel will The diffusion rates in austenite are quite low
form martensite if deformed sufficiently. But in compared to ferrite, so even interstitial elements
the annealed state, the austenite in standard cannot move quickly enough to precipitate below
austenitic stainless steels will remain indefinitely about 400 °C (750 °F). This permits carbon and
as fully austenitic without precipitates unless nitrogen to exist in very high degrees of supersat-
heated above 400 °C (750 °C) for protracted peri- uration if introduced below this temperature, as
ods of time or deformed extensively. is done by various proprietary processes. The
Interstitial elements are much more soluble in low diffusion rates restrict such colossally super-
austenite than in ferrite. Of these, only nitrogen saturated zones to thin surface layers, but they
is considered a beneficial alloying element. It can reach phenomenal hardness of over Rc 70.
both strengthens and improves the pitting corro- The austenite structure does not discourage the
sion resistance of austenite. Carbon has a paral- formation of intermetallic compounds such as
lel effect, but its tendency to form chromium sigma, but it does, fortunately, make their forma-
carbides limits its use and in fact leads to its tion very sluggish, as seen in Fig. 4. The differ-
minimization in most alloys. Before the AOD ence of three orders of magnitude for carbide
was developed and carbon levels in stainless formation reflects the difference between the dif-
steels were higher, austenitic stainless steels fusion of carbon and that of substitutional ele-
were sometimes stabilized by titanium or nio- ments. The formation of sigma in ferrite is about
bium to counter the effects of carbon. Both car- 100 times faster than in austenite. Sigma is al-
bon and nitrogen stabilize the austenite phase, most never seen in commercial 316 alloys.
permitting lower levels of nickel to be used in
austenitic alloys.
Interstitial atoms of carbon and nitrogen dis-
tort the fcc lattice, causing it to expand about
1% linearly per 1 wt% of solute (Fig. 3) (Ref 2).
This produces solid solution hardening of the
austenite. The work hardening of austenite is in-
creased by nitrogen. A third interstitial solute,
hydrogen, produces the same effect but to a
lesser degree. Austenite is not embrittled by hy-
drogen to the extent ferrite or martensite is, but
hydrogen does raise its flow stress and hardness
while lowering its work-hardening rate.
Sulfur and oxygen are considered impurities
because they form inclusions, usually chrome/
manganese silicates and sulfides. If present in
sufficient amounts, sulfur and oxygen precipitate
as primary inclusions before or during solidifica-
tion. In most austenitic stainless alloys, the re-
mainder of these elements are near saturation in
the as-solidified ferrite at very high temperatures Fig. 3 Lattice expansions due to carbon. Source: Ref 2
Chapter 1: Metallurgy / 7

The mechanical properties of austenite are The physical properties of austenite com-
quite different from those of ferrite. Austenite is pared to ferrite include lower thermal and elec-
characterized by: trical conductivity and greater thermal expan-
sion. It is also, of course, nonmagnetic.
• Low stacking fault energies leading to high
work-hardening rates
• Good toughness even at very low tempera- Martensite
tures Martensite is a phase that forms from the dif-
• Low notch sensitivity fusionless shear of austenite to a distorted cubic
• Lack of a sharp elastic limit or hexagonal structure. This transformation can
• Good high-temperature strength occur spontaneously on cooling or isothermally
• Fairly isotropic mechanical properties with externally applied deformation. It is essen-
tially ferrite that has been formed with a super-
While there is not a great deal of difference saturation of carbon. The resulting structure is
in the yield strengths of austenitic and ferritic very fine and highly faulted, making it quite
alloys of similar alloy levels, austenitic alloys hard. As in carbon steel, the hardness of the
are more ductile, have high work-hardening martensite increases dramatically with intersti-
rates, and therefore have higher tensile tial content because of the huge strain intersti-
strengths. Austenite can be cold worked to ex- tial elements impose on the bcc lattice, distort-
tremely high strengths, around a maximum of ing it into tetragonality.
2000 MPa (290 ksi). Chapter 3, “Austenitic Martensite in stainless steels is restricted to
Stainless Steels,” gives a more thorough and alloy levels at which austenite can form at
quantitative treatment of the mechanical prop- higher temperatures but at which the austenite is
erties of austenite. unstable at ambient temperatures. This gives
In duplex stainless steels, a secondary austen- martensite a fairly narrow composition range.
ite, γ2, can form from ferrite below 650 °C The lowest alloy level is that of the basic 12%
(1200 °F). At this temperature, it has the same chromium steels with 0.1 to 0.2% carbon. The
composition as the ferrite from which it forms most highly alloyed martensites are found in the
and is called type 1. In the 650 to 800 °C (1200 precipitation-hardening grades. Thus, marten-
to 1470 °F) range, a range that can be encoun- sitic stainless steels are inherently limited in
tered in the heat-affected zone (HAZ) at γ/δ corrosion resistance to a level no better than a
boundaries during welding, another type forms. 17 or 18% chromium alloy and often barely
This so-called secondary austenite, γ2, type 2, is qualify as stainless after the chromium tied up
somewhat enriched in nickel over the ferrite as chromium carbide is recognized as not con-
from which it forms but poorer in nitrogen than tributing to the corrosion resistance.
the primary austenite, giving it poorer corrosion The as-formed martensite to the degree it has
resistance. Secondary austenite can also coform significant carbon content is hard and requires
with sigma as γ/δ grain boundaries are depleted tempering to give it adequate toughness. The
of chromium. This secondary austenite is called tempering reaction is the precipitation of car-
type 3 and is also poor in chromium. bon in the form of carbides with the concurrent
loss of internal strain in the martensite lattice.
The complexities of tempering require its dis-
cussion in detail to be found in Chapter 3,
“Martensitic Stainless Steels.” It is worth noting,
however, that all tempering involves carbide for-
mation, thus losing some corrosion-fighting
There are two forms of martensite, the ε, ep-
silon, and the α', alpha prime. Epsilon is formed
in steels with low stacking fault energy, which
are primarily the leaner austenitic alloys. Thus, it
forms at cryogenic temperatures or by cold
working. It appears in martensitic alloys of the
precipitation-hardening type. It is nonmagnetic,
Fig. 4 Precipitation kinetics in 316 stainless steel. Source: Ref 3 has a hexagonal close-packed (hcp) structure,
8 / Stainless Steels for Design Engineers

and is very difficult to identify microscopically. chromium or molybdenum, causing localized

The a' martensite is the familiar magnetic vari- lower corrosion resistance. Intermetallic phases
ety known in alloy steels that forms both by form by diffusion of substitutional alloying ele-
quenching and by deformation. ments, which makes their precipitation slower
The mechanical properties of stainless than that of carbides, but they can form in a
martensite are parallel to those of alloy steels. matter of minutes in alloy-rich grades. Defor-
The high quantity of alloying elements in mation, which enhances substitutional diffu-
stainless give an extreme depth of hardening, sion, accelerates their formation. The principal
so there is no concern with ancillary phases intermetallic phases are described next.
such as bainite. The physical properties are Alpha Prime. Not to be confused with
very close to those of ferrite of the same com- martensite, alpha prime is an ordered iron-
position. chromium phase (i.e., iron and chromium atoms
occupy specific, rather than random, sites on
Intermetallic Phases two intersecting superlattices). This structure is
quite brittle. It forms at relatively low tempera-
The number of phases that can coexist in an tures, between 300 and 525 °C (570 and 980
alloy is proportional to the number of alloying °F). Before its true nature was understood, its
elements in the alloy. Table 1 lists data on the presence was known through its causing the
more common precipitates found in stainless phenomenon called 475 embrittlement, origi-
steel. It is not surprising that stainless steel with nally called 885 °F embrittlement. This is some-
iron, chromium, nickel, manganese, silicon, and times confused with temper embrittlement,
often molybdenum, titanium, and niobium which occurs in the same temperature range but
should have numerous ancillary phases. Inter- is caused by phosphide precipitation on prior
metallic phases are normally hard and brittle. austenite grain boundaries of martensite. Alpha
They can render the bulk alloy brittle when they prime precipitation can cause 475 embrittle-
form along grain boundaries. The other concern ment in ferritic or duplex stainless steels and
arising from intermetallic phase formation is the limits their use in this temperature range but not
depletion from the surrounding matrix of at higher temperatures, at which the phase dis-
solves. This phase forms at chromium contents
Table 1 Precipitated phases found in stainless as low as 15%, but fortunately it takes a rela-
steels tively long time to form, on the order of hours,
Precipitate Structure Parameter, A Composition so it will not occur inadvertently during thermal
processing such as welding or annealing.
NbC fcc(a) a = 4.47 NbC
Sigma. Sigma is a brittle tetragonal phase
NbN fcc a = 4.40 NbN richer in chromium and molybdenum than ei-
TiC fcc a = 4.33 TiC ther the ferrite or austenite matrix around it. It
TiN fcc a = 4.24 TiN forms preferentially at ferrite-austenite bound-
Z-phase Tetragonal a = 3.037 c = 7.391 CrNbN aries in the temperature range 600 to 1000 °C
M23C6 fcc a = 10.57–10.68 Cr16Fe5Mo2C (e.g.)
(1110 to 1470 °F) in alloys with more than
about 18% chromium plus molybdenum. Its
M23(C,B)6 fcc a = 10.57–10.68 Cr23(C,B)6
composition is sometimes given as (CrMo)35
M6C Diamond a = 10.62–11.28 (FeCr)21Mo3C; (FeNi)65, but examination of the iron-chromium
cubic Fe3Nb3C; M5SiC
phase diagram shows that it is archetypically an
M2N Hexagonal a = 2.8 c = 4.4 Cr2N equiatomic iron chromium compound. It is
MN Cubic a = 4.13–4.18 CrN strongly promoted by silicon and suppressed by
Gamma fcc a = 3.59 Ni3(Al,Ti) nitrogen. Stabilized alloy grades show more
prime rapid sigma formation than unstabilized alloy
Sigma Tetragonal a = 8.80 c = 4.54 Fe, Ni, Cr, Mo grades (e.g., 347 versus 304). In unstabilized al-
Laves Hexagonal a = 4.73 c = 7.72 Fe2Mo, Fe2Nb
loys the prior precipitation of carbides destabi-
phase lizes austenite, leading to subsequent sigma for-
Chi phase bcc(b) a = 8.807–8.878 Fe36Cr12Mo10
mation. This makes alloys like 310H,
essentially 25Cr-20Ni, especially prone to
G-phase fcc a = 11.2 Ni16Nb6Si7,
sigma formation.
(a) fcc, face-centered cubic. (b) bcc, body-centered cubic.
Sigma forms much more rapidly from ferrite
than from austenite because of the 100-fold
Chapter 1: Metallurgy / 9

higher diffusion rate of alloy elements in ferrite. The precipitation of the carbide from ferrite
This makes it a much larger issue in superfer- occurs at grain boundaries, is extremely rapid,
ritic and duplex alloys, which have high and cannot be suppressed by quenching. Less
chromium and/or molybdenum levels. Chapter than 20 ppm carbon content is required to pre-
7, “Duplex Stainless Steels,” contains an in- vent its precipitation from ferrite, although up
depth discussion of sigma. to 50 ppm can be effectively kept in solution by
Chi. Chi, χ, is similar to sigma except it con- very vigorous quenching. From austenite, car-
tains more molybdenum and less chromium and bide precipitation occurs below about 900 °C
has a cubic structure. It can coexist with sigma (1650 °F) for carbon levels under 0.10% and at
and forms in the same temperature range. It also 650 °C (1200 °F) for carbon levels below
precipitates at ferrite-austenite boundaries and 0.03%. For practical purposes, precipitation
has the same deleterious effects. ceases below 500 °C (930 °F) due to the slow-
Laves Phase. The laves phase has the struc- ing diffusion of carbon. While carbon is essen-
ture A2B where A is iron or chromium and B is tially insoluble in austenite at room tempera-
molybdenum, niobium, titanium, or silicon. It ture, quenching can easily preserve up to 0.10%
forms at 550 to 650 °C (1020 to 1200 °F) over in supersaturation, as is commonly seen in type
the course of hours. Thus, although its effect 301 stainless.
would be deleterious, it seldom becomes a prac- The carbide precipitation occurs first at grain
tical problem. It is possible for it to form at tem- boundaries. The chromium that combines with
peratures below sigma and above alpha prime, the carbon comes from the matrix in the imme-
but the long times for formation make it rare. diate vicinity and therefore decreases the
chromium content of that region, giving rise to
Carbides, Nitrides, Precipitation Hardening, the phenomenon of sensitization, which comes
from the original phrase “sensitization to inter-
and Inclusions
granular corrosion.” Nickel and molybdenum
Carbon and nitrogen are very important in all decrease the solubility of carbon and thus accel-
steels, but they take on a special significance in erate the precipitation. Nitrogen retards precipi-
stainless steel because chromium, the essential tation. Cold work accelerates precipitation. The
alloying element of stainless steel, reacts more carbide has a hardness of about Rc 72. This
vigorously with carbon and nitrogen than iron makes the phase a useful constituent in wear
does. Except for its role in hardening martensite resistance in martensitic alloys.
and strengthening austenite at high tempera- In higher carbon grades such as the marten-
tures, carbon is almost universally a detrimental sitic stainless alloys, additional, more carbon-
impurity from a corrosion point of view and is rich, carbides may form. These include M7C3
minimized. Its beneficial effect on corrosion re- and M3C. The latter carbide forms during the
sistance when it is in solution is negligible low-temperature tempering of martensite, while
because so little of it can be held in solution. the former precipitates at higher temperatures.
Nitrogen has a lesser tendency to form com- Stabilizing carbides are those that are formed
pounds with chromium, so it is considered a by the intentional addition of elements such as
beneficial alloying element in austenite but not titanium and niobium. These elements form car-
in ferrite, in which it has essentially zero solu- bides of the type MC (metal carbide). The car-
bility. Common carbide and nitride precipitating bon in these compounds may be replaced by ni-
phases are also listed in Table 1. trogen or, in the case of titanium, sulfur. These
Carbides. M23C6 is the main carbide found in carbides form preferentially over chromium car-
stainless steel. Its structure is orthorhombic, and bides and thus prevent sensitization. They pre-
it contains both iron and chromium. It can form cipitate in both the liquid and solid states. In the
at any temperature at which the host austenite or solid state, the precipitate normally forms within
ferrite becomes saturated with carbon. It is grains. The Ti(CN) appears as a cube of gold
mainly chromium carbide, but iron can substitute TiN surrounded by gray TiC. The Nb(C,N) is
for chromium up to about 50%. Other elements, less regularly shaped. They affect mechanical
such as tungsten, vanadium, and molybdenum, properties in ferrite both by their influence on
can also dissolve in this carbide. The ratio of recrystallization and by their ability to act as nu-
chromium to iron in the carbide increases with cleation sites for brittle fracture
time and temperature, as chromium diffusion Nitrides. At low levels, nitrogen can substi-
permits, up to a maximum of 4 or 5 to 1. tute for carbon in M23C6. At higher nitrogen
10 / Stainless Steels for Design Engineers

levels, Cr2N can form. This can occur in duplex tal as inclusions have been shown to be the initi-
alloys if they are heated to a solution annealing ation sites for corrosion pits, which have been
temperature at which the alloy has high solubil- linked to both their sulfur ions disrupting the
ity for nitrogen. Cooling from these tempera- passive layer and their chromium content caus-
tures can cause the excess nitrogen to precipi- ing slight local chromium depletion.
tate as needles of Cr2N. Another nitride CrN
can form in HAZs of welds.
Precipitation-Hardening Phases. Phases
that have a very similar lattice match to the par- Properties of Stainless Steels
ent phase can precipitate coherently, that is,
without changing the continuity of the crystal Physical and mechanical properties of repre-
lattice. In these cases, the slight mismatch sentative stainless steel alloys are summarized
causes a strain that can significantly restrict dis- in Appendix 2. Properties are also discussed in
location movement and thereby strengthen the chapters specific to each alloy family. The
matrix. One such precipitate is gamma prime, reader is referred to primary sources, such as
an intermetallic, ordered, fcc phase with the company web sites, such as Ref 4 and 5.
composition Ni3(AlTi). Copper forms the ep-
silon phase, essentially pure copper, which
causes precipitation hardening. The secondary REFERENCES
hardening of martensite due to the precipitation
of molybdenum nitride or carbide is also a pre- 1. D.J. Kotecki, Welding of Stainless Steels,
cipitation-hardening reaction. Welding, Brazing, and Soldering, Vol 6,
Inclusions. Inclusions are principally oxides ASM Handbook, ASM International, 1993,
and sulfides that form in the melt (type I), at the p 677–707
end of solidification (type II), or in the solid 2. G.E. Totten, M. Narazaki, R.R. Black-
(type III). Type I inclusions are the largest and wood, and L.M. Jarvis, Failures Related to
are globular. Except when they are deliberately Heat Treating Operations, Vol 11 ASM
kept to improve machinability, they are physi- Handbook, ASM International, 2002, p
cally removed by various steelmaking practices. 192–223
Type II inclusions form in interdendritic spaces 3. High Performance Stainless Steels, Refer-
as the solubility of oxygen and sulfur drop on ence Book Series 11 021, Nickel Develop-
solidification. Type III inclusions precipitate the ment Institute, p 16
remaining oxygen and sulfur, up to 100 ppm for 4. ASM Handbook, Vol 1, Properties and
normal manganese-silicon killed stainless Selection, ASM International, 1990
steels, in the solid state either on preexisting in- 5. ASM Speciality Handbook, Stainless Steels,
clusions or as micron-size particles. Inclusions ASM International, 1996
are mainly oxides and sulfides of silicon and
manganese. If more reactive elements, such as
aluminum or titanium, are present, their oxides SELECTED REFERENCES
and sulfides can also be present.
Sulfides and oxysulfides can be beneficial for • D.J. Kotecki and T.A. Siewert, WRC 1992
machining as solid-state lubricants and chip Constitution Diagram, Welding Journal, Vol
breakers. Otherwise, their presence is detrimen- 5, 1992, p 171s–178s
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 11-18 All rights reserved.
DOI: 10.1361/ssde2008p011


Corrosion Theory

Summary Electrochemical Reactions

THIS CHAPTER INTRODUCES THE funda- In electrochemical reactions, charge is trans-

mentals of electrochemical theory as it pertains ferred across interfaces of species of different
to corrosion. Topics covered include an overview chemistry. Consider, for example, the reaction:
of electrochemical reactions, Faraday’s law, the
Nernst equation, galvanic versus electrochemical 2Fe (s) + O 2 (g) + 2H 2 O → 2Fe 2+ + 4 OH − (Eq 1)
cells, and Pourbaix diagrams. The examples pro-
vided relate these fundamentals to the corrosion
resistance of stainless steels. An inspection of this reaction suggests that
three phases must be present for the reaction to
proceed: an ion-conducting phase (water-based
solution), a metallic phase (iron), and a gas
Introduction phase O2(g). Second, electrons have been trans-
ferred from the metallic phase, iron to O2 +
Corrosion—the environmental degradation H2O. Figure 1(a) shows the arrangement of an
of materials through electrochemical reac- experimental setup in which Reaction 1 could
tions—is a key subject for more or less all proceed.
classes of alloys that fall within the broad defi- On the left, iron is allowed to dissolve
nition of stainless steels because these alloys according to:
were developed with the intention of prevent-
ing corrosion. This chapter aims first to provide 2Fe (s) → 2Fe 2+ + 4 e − (Eq 2)
an introduction to the fundamentals of electro-
chemical theory as it pertains to corrosion. resulting in Fe2+ ions that dissolve in the
Thermodynamics are presented in light of elec- water-based solution and electrons that are car-
trochemical potentials as opposed to purely ried to the right side, where they participate in
chemical ones. Chapter 3 introduces the formal the reaction:
terms needed to describe electrode reaction ki-
netics. Chapter 4 describes the various forms of O 2 + 2H 2 O + 4 e − → 4 OH − (Eq 3)
corrosion and how they are related to alloy
metallurgy, chemistry, and structure. Chapter 5 Inside the water-based solution, ions (Fe2+,
focuses on oxidation. For an in-depth study of OH⫺, H+, or any others) migrate, thereby con-
electrochemical kinetics and electroanalytical stituting a so-called ionic current. This current
methods, Ref 1 is recommended. For a broader together with Reactions 2 and 3 and the trans-
study of corrosion, the reader is referred to port of electrons from left to right form a closed
texts by Jones (Ref 2), Uhlig and Revie (Ref 3), circuit called an electrochemical cell. The cell is
and Fontana (Ref 4) and to ASM Handbook, made up of four parts: the two electrodes where
Volume 13A (Ref 5). the charge transfer Reactions 2 and 3 take place
12 / Stainless Steels for Design Engineers

often are described as half cells, for example,

Fe / O 2 / OH − and Fe / Fe 2+ .

Faraday’s Law

If the cell in Fig. 1(a) was allowed to proceed

and thermodynamics favored to proceed ac-
cording to the direction in Reaction 1, then a
current i will flow from the anode to the cath-
ode, and the amount of charge passed per unit
time as a result of this current will be linked to
the amount of iron dissolved per unit time or the
amount of oxygen reacted per unit time by
virtue of Eq 2 and 3. This is given by Faraday’s

nNF = it (Eq 4)

Here, i * t is the charge passed (in coulombs);

N is the moles of consumed/produced specie
(e.g., moles consumed iron in Reaction 2); n is
the ratio of electrons to consumed/produced
species, which in the case of Reaction 2 will be
2; and F is Faraday’s constant, which is essen-
tially the charge in coulombs corresponding to
1 mole of electrons.

Fig. 1 Schematic illustration of (a) a differential aeration cell

The Nernst Equation
involving iron dissolution and (b) the same cell with a
variable resistor and voltmeter
Electrochemical reactions require a transfer
(the anode and cathode, respectively), an elec- of charge; hence, there is a coupling between
trolyte, and an electron pathway. It should be chemical and electrical energy. Consider the hy-
noted that electrodes are interfaces that require pothetical setup in Fig. 1(a) with the addition of
several phases to be in contact. Oxidation, Re- a variable resistor and a voltmeter, resulting in
action 2, occurs at the anode and reduction, Re- the arrangement shown in Fig. 1(b). Thermody-
action 3, occurs at the cathode. The electrolyte namically, the Gibbs free energy of the cell is
is the medium through which the ions migrate; that of Reaction 1:
in the case of corrosion reactions, this is most
commonly a water-based solution, but at high ΔG = ΔH − T ΔS

(a ) (a )
4 2
temperatures it could be a solid oxide. The final (Eq 5)
OH − Fe 2+
constituent of the electrochemical cell is a path- = ΔG 0
+ RT ln
way through which electrons can migrate from aH O PO
2 2

the anode to the cathode.

As a shorthand notation, electrochemical cells where ⌬G is the Gibbs free energy, H is the en-
are written by separating components within a thalpy, S is the entropy, R is the gas constant,
phase by a comma and separating phases by a and T is the absolute temperature. If this is neg-
slash; gaseous species are written next to their ative, the reaction would be expected to proceed
conducting electrode. For example, the cell de- spontaneously as written in Reaction 1. Let us
scribed in Fig. 1(a) would be recorded as assume that this is the case. The thermal heat
Fe / O 2 / OH − , Fe 2+ / Fe . This cell is an example produced by the system can be divided into two
of a differential aeration corrosion cell, which is parts: the thermal heat produced by the cell Qt
discussed later. Processes at a single electrode and the heat produced at the resistor QRes. QRes
Chapter 2: Corrosion Theory / 13

in this case is heat, but in essence it represents Here, pi and ri are the concentrations of reac-
the available energy or work, which in the case tant and products, respectively, and αi and βi are
of a resistance is given by the product of charge the numbers that are needed to balance the reac-
passed times potential difference. If the resist- tion stoichiometrically. In the case of Reaction
ance approached infinity ( R→ ), Reaction 1 1, Eq 10 would be:
would proceed through infinitesimal steps and

( ) ( ) ⎞⎟
4 2
can be considered thermodynamically reversible.
RT ⎜ aOH− aFe 2+
In this case, the thermal heat produced by the Erxn = Erxn
− ln (Eq 11)
cell is minimized and according to thermody- 4F ⎜ aH O PO ⎟
⎝ 2 2 ⎠
namics is given as Qt = Qrev = T⌬S1. On the
other hand, the net work gained QRes is maxi- If the emf according to Eq 11 is positive, this
mized and constitutes the rest of the free energy: means that the free energy is negative (according
to the Nernst equation); hence, the net reaction is
QRes = ΔG = ΔH − T ΔS (Eq 6) thermodynamically favored as it is written in
Reaction 1. By inspection of Eq 11, it can be
As mentioned, the energy dissipated through seen that it is the difference between two hypo-
the resistance is charge passed times potential thetical half reactions, ( Erxn = EO /OH− − EFe 2+ /Fe )
difference, and in this case the potential differ- defined as:

ence is the reversible potential difference E;

thus, in an absolute sense: ⎛
( )

RT ⎜ aOH− ⎟
EO = EO 0
− ln (Eq 12)
ΔG = nFE (Eq 7) 2 /OH − 2 /OH −
4 F ⎜ aH O PO ⎟
⎝ 2 2⎠

Here, n is the number of electrons passed per which corresponds to the reduction Reaction 3
atom of iron reacted, and F = 96,485 C per mole and:
electrons, is Faraday’s constant. The reversible
potential difference E represents the potential
difference between the two electrode reactions EFe 2+ /Fe = EFe 2+ /Fe −
0 RT ⎛ ( aFe ⎞
ln ⎜

) (Eq 13)
(cathode and anode), and as such they are asso- 4 F ⎜⎝ aFe 2+ ⎟⎠
ciated with Reaction 1 rather than a physical
cell. The potential difference is referred to as which corresponds to the reverse of Reaction
the electromotive force (emf) of the cell. It is 2, that is, if it was a reduction reaction. The
also referred to as the open circuit potential be- potentials as written in Eq 12 and 13 are
cause it is the potential measured by the volt- called reduction potentials, and because
meter in Fig. 1(b) when a negligible current Erxn = EO /OH− − EFe 2+ /Fe has to be positive for the
flows. It is defined here as Erxn. By convention, 2
reaction to be thermodynamically favored as
this potential is positive for a spontaneous reac- written in Eq 1, the reduction potential E O 2 /OH−
tion (as opposed to the chemical free energy, has to be larger than E Fe 2+ /Fe. If it was not, then
which is negative); hence, Eq 7 becomes: Reaction 1 would proceed in the reverse direc-
tion, which means that the electrode Reactions
ΔG = − nFErxn (Eq 8) 2 and 3 would be reversed and thus so would
the anode and cathode of the cell.
and if all elements have unit activities: It is useful to list reduction potentials for half-
cell reactions, just as it is useful to list free en-
ΔG 0 = − nFErxn
(Eq 9) ergy data. However, half-cell potentials (like
any electrical potentials) cannot be measured in
an absolute sense; only potential differences can
Equation 8 is the Nernst equation. By virtue be measured. ( Erxn = EO /OH – − EFe 2+ /Fe can be
of Eq 8 and 9 and the expression for Gibbs free 2
measured because it is a difference.) Therefore,
energy of a reaction (e.g., Eq 5), an expression
half-cell potentials are measured with respect to
for Erxn is obtained:
a reference electrode. Reference electrodes are
constructed such that they have a stable potential;
RT ⎛ p1 1 p2 2 p3 3 ... ⎞
α α α

Erxn = Erxn
− ln ⎜ ⎟ (Eq 10) this is discussed further in Chapter 3. A com-
nF ⎝ r1β1 r2β2 r3β3 ... ⎠ mon reference electrode in aqueous solutions is
14 / Stainless Steels for Design Engineers

the normal hydrogen electrode (NHE), also Galvanic versus Electrochemical Cells
known as the standard hydrogen electrode
(SHE), with a potential set (arbitrarily) as zero When reactions in a cell occur spontaneously
at all temperatures. The NHE is schematically in the direction dictated by the open-circuit po-
shown in Fig. 2. In shorthand notation, it is: tential of a cell that is positive ( Erxn > 0) , a cur-
Pt / H 2 (a = 1)/ H + (a = 1) , and the half-cell reac- rent flows as shown in Fig. 3(a). This is the case
tion is: in environmentally caused electrochemical cor-
rosion reactions. It also is the case in fuel cells
2 H + + 2 e− = H 2 (Eq 14) and batteries (under discharge), in which the
Table 1 (Ref 6) lists half-cell reduction stan- current is used as electricity. These types of
dard potentials ( EOx/Re) versus NHE that are a cells are called galvanic cells, in which chemi-
result of the emf of the following types of cells cal energy is converted to electrical energy.
(for Reaction 2, as an example): Most of the discussion in the following chapters
concerns these types of cells. In electrolytic
Pt / H 2 (a = 1)/ H + (a = 1), Fe 2+ (a = 1) / Fe cells (Fig. 3b), an imposed electrical potential
counters the “natural” cell potential to drive a
reaction in a desired direction. These types of
cells are used for many metallurgical processes,
such as electroplating, electrorefining and elec-
troextraction (e.g., the Hall-Heroult aluminum
smelting cell), and for other applications, such
as charging batteries. In the case of corrosion,
the principle is used for protection against cor-
rosion. In electrolytic cells, electrical energy is
converted to chemical energy.

Table 1 Standard half-cell reduction potentials

versus the normal hydrogen electrode
Standard half-cell reduction
Reaction potential vs. NHE(a) (V)

Fe 3+ + e− = Fe 2 + 0.771
O 2 + 2H 2 O + 4e − = 4 OH – (pH = 14) 0.401
2 H + + 2 e− = H 2 0.000
Ni 2 + + 2 e− = Ni –0.250
Fe 2+ + 2e− = Fe –0.447
Cr 3+ + 3e− = Cr –0.744
2 H 2 O + 2 e− = H 2 + 2 OH − (pH = 14 ) –0.828
Fig. 2 The normal hydrogen electrode (NHE) (a) NHE, normal hydrogen electrode. Source: Ref 6

Fig. 3 Schematic of (a) galvanic cell and (b) electrolysis cell

Chapter 2: Corrosion Theory / 15

Corrosion Tendency ing to Eq 13, which, assuming a Fe2+ activity of

10−6 (this is an arbitrary value but is usually
The tendency to corrode, that is, whether a taken to represent a low ion concentration), be-
system consisting of anode, cathode, and elec- comes at room temperature (using Table 1 for
trolyte can react thermodynamically, is deter- the standard potential):
mined by evaluating Erxn. If this is positive, then
there is thermodynamically a possibility for cor- EFe 2+ /Fe = −0.447 − 0.0295 log(aFe 2+)
rosion. The rate of corrosion, which is in most
cases determined by corrosion kinetics, is dis- = –0.624 V (vs. NHE) (Eq 16)
cussed in Chapter 3. Consider, for example, a
case of iron in aerated water. Figure 1 (with Figure 4(a) shows a schematic plot of the two
electrode Reactions 2 and 3) can be viewed as reduction potentials (Eq 15 and 16) versus pH.
an idealized equivalent cell for this situation. It Because a spontaneous reaction requires Erxn
should be noted, however, that the locations of to be positive, if the only pertinent reactions
anode(s) and cathode(s) on the iron surface can- were Eq 2 and 3, this means that corrosion (due
not be identified with ease. At room temperature to iron dissolution to Fe2+ and oxygen reduc-
(298 K), 1 atm oxygen partial pressure, and tion) is possible when the line representing
using Table 1, Eq 12 can be written by assuming EO2/OH− (Eq 15) lies above the line representing
unit activity for water and unit activity coeffi- EFe / Fe (Eq 16). This is indicated by the region

cient for OH−: shaded in gray in Fig. 4(a). Hydrogen reduction

is another possible cathode reaction in water:
EO – = 0.401 + 0.059 pOH
2 / OH
2 H + + 2 e− → H 2 (Eq 17)
= 0.401 + 0.059(14 − pH )
= 1..227 − 0.059 pH V (vs. NHE) (Eq 15) and its reduction potential is (using the defini-
tion of pH):
Here, the following definition of pH has been
used: pH = –log CH+, pOH = –log COH− and pH EH+ / H = EH0 + / H − ln
+ pOH = 14. Similarly, the iron dissolution Re-
2 2 nF aH2 +
action 2 will have a reduction potential accord- = 0 − 0.059 pH V vs. NHE (Eq 18)

Fig. 4 Reduction potential versus pH for iron and (a) oxygen gas reduction and (b) hydrogen ion reduction
16 / Stainless Steels for Design Engineers

Figure 4(b) shows the condition in which corro- Fe(OH)2 = HFeO 2− +H + ,

sion under deaerated conditions (due to iron dis-
solution to Fe2+ and hydrogen ion reduction) is (
pH = 14.30 + log aHFeO−
) (Eq 19b)
possible as a gray shaded region. In Fig. 4(a)
and (b), the regions where iron is stable are
denoted as immunity (corresponding to immu-
nity from corrosion). When comparing these Fe 3+ + 3H 2 O = Fe(OH)3 +3H + ,
two figures, it is noteworthy that hydrogen ions pH = 1.613 – (1/3) log(a Fe3+ ) (Eq 19c)
are able to cause corrosion only under relatively
low pH conditions, whereas oxygen gas is able
to corrode iron in the entire pH range. Since these are independent of potential, they
will appear as vertical lines (see lines 19a to 19c
in Fig. 5a).
The following pH-independent electrochemi-
The Construction of Pourbaix Diagrams cal reactions need to be considered, and they
will result in horizontal lines (Fig. 5a):
Figures 4(a) and (b) are types of phase dia-
grams that show the stable phases in an area
bounded by pH and potential. In reality, several Fe 2+ + 2 e− → Fe,
electrochemical and chemical reactions need to
be considered when constructing these types of EFe 2+ /Fe = −0.447 + 0.0295 log aFe 2+( ) (Eq 20a)
diagrams. Each reaction is represented by a line.
In the case of iron, the following chemical reac-
tions will have to be considered (the pH de- )
HFeO 2− + H 2 O = Fe ( OH 3 + 2e− ,

pendency of these reactions is listed next to
them [Ref 7] and since they are not electro-
chemical, they are evaluated from the equilib-
EFe(OH) / HFeO2– = −0.810 − 0.0591 log aHFeO–
( 2

rium constants): (Eq 20b)

The following electrochemical reactions will

Fe 2 + + 2 H 2 O = Fe(OH )2 + 2 H + ,
depend on pH and thus will be sloped depend-
pH = 6.65 − 0.5 log aFe 2 + ) (Eq 19a) ing on this dependence (Fig. 5a).

Fig. 5 Pourbaix diagram for iron. (a) Schematic matching Eq 19 to 21 in text to lines. (b) Actual complete diagram. Source: Ref 7
Chapter 2: Corrosion Theory / 17

Fe + 2H 2 O = Fe ( OH 2 + 2H + + 2e− ,)
EFe(OH) = −0.0470 − 0.0591 pH (Eq 21a)
2 / Fe

Fe + 2H 2 O = HFeO −2 + 3H + + 2e− ,
EHFeO- /Fe = 0.495 − 0.0886 pH

+ 0.0295 log aHFeO–
) (Eq 21b)

Fe 2+ + 3H 2 O = Fe ( OH 3 + 3H + + e− , )
EFe(OH) 2+ = 1.057 − 0.1773 pH
3 / Fe

( )
− 0.0591 log aFe 2+ (Eq 21c)

Fe(OH)2 + H 2 O = Fe(OH)3 + H + + e− , Fig. 6 Pourbaix diagram for chromium in water. Source: Ref 8
EFe(OH) = 0.271 − 0.0591 pH (Eq 21d)
3 /Fe(OH)2

For the pH-dependent reactions (chemical reduction are able to cause corrosion through
and electrochemical), one can readily label the the entire pH region. Unfortunately, Fe-OH cor-
regions depending on what iron species increas- rosion products are generally not passivating.
ing pH favors. If iron would be an anode and Iron or carbon steel alloys are therefore not par-
the tendency to corrode were to be evaluated, ticularly corrosion resistant in water solutions.
then the reduction potential for a possible cath- Figure 6 shows the Pourbaix diagram for
ode reaction would be placed on this diagram. If chromium (Ref 8). While chromium oxidizes
this point were to be, for example, in A in Fig. even more readily than iron, it forms Cr2O3 over
5(a), this means that the reduction potential for a significantly large region that is of relevance
this assumed cathode lies below any reduction to pH values in water solutions. Since Cr2O3 is
potential of iron, and hence under these condi- protective, it prevents further corrosion. When
tions iron is immune (since Erxn is negative). In chromium is added to iron as an alloying ele-
fact any Fe2+ ions present could plate as iron. ment, it corrodes selectively due to its low re-
On the other hand, if the reduction potential of duction potential, but this means that it also pro-
the assumed cathode reaction were to lie in tects the iron alloy due to the properties of
point B, then there is a tendency to dissolve iron Cr2O3. This is the basic design principle behind
to Fe2+ since Erxn is positive. Finally, if the re- iron-chromium-based stainless steels.
duction potential of the assumed cathode was at
point C, corrosion would occur, resulting in
Fe(OH)3, but when oxides or hydroxides are REFERENCES
formed there is a possibility that this product
could form a solid protective layer that kineti- 1. A.J. Bard and L.R. Faulkner, Electrochemi-
cally hinders further corrosion. These types of cal Methods: Fundamentals and Applica-
diagrams are called Pourbaix diagrams. Figure tions, 2nd ed., Wiley, 2001
5(b) shows the Pourbaix diagram for iron over- 2. D.A. Jones, Principles and Prevention of
laid with the common cathode reactions in Corrosion, 2nd ed., Prentice Hall, 1996
water, Eq 15 and 18 (Ref 8). The ionic activity 3. H.H. Uhlig and R.W. Revie, Corrosion and
was previously arbitrarily set at 10–6, but from Corrosion Control: An Introduction to Cor-
the Pourbaix diagram it can be seen that changes rosion Science and Engineering, 3rd ed.,
in ion activity do not have dramatic effects on Wiley, 1985
the boundaries. It can be seen that both the 4. M.G. Fontana, Corrosion Engineering, 3rd
oxygen gas reduction reaction and hydrogen ion ed., McGraw-Hill, 1986
18 / Stainless Steels for Design Engineers

5. ASM Handbook, Vol 13A, Corrosion: Fun- 8. S.A. Bradford, Corrosion Control, 2nd ed.,
damentals, Testing, and Protection, S.D. CASTI Publishing, Inc., 2001, p 41
Cramer and B.S. Covino Jr., Ed., ASM
International, 2003
6. Handbook of Chemistry and Physics, 71st SELECTED REFERENCE
ed., CRC Press, 1991
7. D.A. Jones, Principles and Prevention of • M. Pourbaix, Atlas of Electrochemical Equi-
Corrosion, 2nd ed., Prentice Hall, 1996, p. 59 libria in Aqueous Solutions, NACE, 1974
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 19-25 All rights reserved.
DOI: 10.1361/ssde2008p019


Corrosion Kinetics

Summary the penetration due to dissolution of element i

tions with equilibrium that is known through
t ⋅ Mi
thermodynamics. In practice, the rate at which r= j (Eq 1)
corrosion reactions occur is the most important ni ⋅ F ⋅ ρ
consideration. This chapter deals with corro-
sion kinetics, which allows us to understand the The penetration rates for iron and various
rates of corrosion. stainless steels are listed in Table 1 (Ref 1) in
units of mils (0.001 in.) per year, or mpy. In the
case of alloys, the ratio of Mi /ni is computed as
Introduction an equivalent weight (EW) according to:

Consider the differential aeration cell dis-

cussed in the Chapter 2 on corrosion theory, Fe/ 1
EW =
O2/OH–, Fe2+/Fe. If the thermodynamic condi- fn (Eq 2)
tions favor electrochemical corrosion of iron, ∑ Mi i
that is, Erxn = EO / OH – EFe / Fe is positive, then
– 2+

a net corrosion current i will flow, resulting in

iron dissolution and consumption of oxygen gas where fi, ni, and Mi are the weight fraction, va-
according to the net reaction, 2Fe (S) + O2 + lence, and molar mass of element i,

2H2O → 2Fe2+ + 4OH . The magnitude of this respectively.
current will determine the rate or iron dissolu- As stated, the amount of corroded (dis-
tion according to Faraday’s law, which was solved) iron is determined by the current i, and
introduced in Chapter 2: nNF = it. Because n = the magnitude of this current is determined by
2 and F = 95,485 C per mole electrons, the
moles of dissolved iron are given as a function
of time as N = i* t/(2* 95,485). Practically, this
Table 1 Penetration rates for a current of 1
can be readily converted to lost mass m, which
µA/cm2 (mpy)
in the case of iron loss becomes m = N* MFe =
MFe* i * t/(2* 95,485), or thickness reduction r, Density, Penetration
Alloy Element/oxidation state g/cm3 EW(a) rate, mpy
which in the case of iron becomes r = MFe* i*
t/(2* 95,485* A* ␳Fe). Here, MFe and ␳Fe are Fe
molar mass and density of iron, respectively. It 321 Fe/2,Cr/3,Ni/2 7.9 25.13 0.41
is often the thickness loss (referred to as 309 Fe/2,Cr/3,Ni/2 7.9 24.62 0.41
316 Fe/2,Cr/3,Ni/2,Mo/3 8.0 25.50 0.41
penetration per unit time) that is useful; 430 Fe/2,Cr/3 7.7 25.30 0.42
therefore, i/A is often replaced by j, which is 446 Fe/2,Cr/3 7.6 24.22 0.41
20Cb3 Fe/2,Cr/3,Mo/3,Cu/1 7.97 23.98 0.39
defined as current density and has the units
(a) Equivalent weight. Source: Ref 1
amperes/square meters. A general equation of
20 / Stainless Steels for Design Engineers

RT ⎛ ( aFe ) ⎞

EFe 2+ /Fe = EFe0 2+ /Fe − ln ⎜ ⎟

4 F ⎝ aFe 2+ ⎠

When a cell is not under open circuit (i.e., a

net current passes through it), the cathode and
anode potentials deviate from the half-cell po-
tentials, and the electrode states are then defined
as being polarized. The polarization is quantified
as overpotentials ␩, which are defined by the de-
viation from the equilibrium half-cell potentials,
that is, for the cathode, ηc = Ecathode – EO2 / OH and
for the anode, ηa = Eanode – EFe 2 + Fe. Effectively, the
Fig. 1 Schematic illustration of a differential aeration cell overpotential reduces the activation energy for
involving iron dissolution. Kinetic steps: (1) electrode the electrode Reactions 3 and 4. In the case of a
reactions, (2) ion conduction, (3) electron conduction
reduction reaction at a cathode, such as Reac-
tion 3, the overpotential is negative, and driving
the corrosion potential. The corrosion potential an electrode toward a lower potential drives
is determined by the reaction potential (which electrons from the electrode into the solution,
was discussed in Chapter 2) and the kinetics of resulting in a net cathodic current ic at this elec-
the various steps involved in completing the trode. Similarly, at the anode the overpotential
electrochemical circuit depicted in Fig. 1. is positive, which results in electrons that are fa-
These involve: (a) electrode reactions at the vored to be removed from the solution and
cathode and anode, (b) conduction of ions in transferred into the electrode, thus producing a
the electrolyte, and (c) conduction of electrons net anodic current ia.
from the anode to the cathode. The conduction If the magnitudes of cathode and anode polar-
of electrons is generally not a problem in stain- ization are large, as would be expected in a gal-
less steels because the corroding metal (iron) vanic cell, the relation between each electrode
and scale (Cr2O3) provide an easy path for elec- current/current density and overpotential is
trons. The other two kinetic processes are dis- given by the following equations (for a thor-
cussed briefly in this chapter. ough derivation of the current overpotential
equation, Ref 2 is recommended):
The Butler-Volmer Equation
ic CO (0, t ) ⎛ αnF ηc ⎞
jc = = j 0 ,c 2 exp ⎜ − (Eq 5)
For the case study 2Fe (S) + O2 + 2H2O Ac C O* ⎝ RT ⎟⎠
→ 2Fe2+ + 4OH–, the cathode and anode reac- 2

tions are:
O 2 + 2H 2 O + 4e − → 4 OH − (Eq 3)
ia ⎛ (1 − α )nF ηa ⎞
ja = = j0 ,a exp ⎜ ⎟⎠ (Eq 6)
Aa ⎝ RT

2 Fe ( s ) → 2 Fe 2 + + 4 e− (Eq 4) where the jo,i terms are the exchange current

densities and represent the equally large cath-
The Nernst equation predicts an open circuit ode and anode currents at equilibrium (zero
potential of Erxn = EO / OH – EFe / Fe overpotential) at the electrodes. The exchange
– 2+
2 current densities are a measure of the electrocat-
where alytic ability of the surface to promote/demote

( )

the electrode charge transfer reactions; as such,
RT ⎜ aOH – ⎟
− ln they can vary over many orders of magnitude
2 /OH – 2 /OH –
4 F ⎜ aH O PO ⎟ depending on the surface chemistry and struc-
⎝ 2 2⎠
ture and on electrode reaction. The α-terms are
and fractions that define the amount to which the
Chapter 3: Corrosion Kinetics / 21

activation energies are lowered. They do not electrode, and the rate of cathode reaction will
have to be the same for the anode and the cath- depend on how rapidly oxygen molecules dif-
ode, but due to the uncertainty in evaluating fuse to the electrode/electrolyte interface. As a
them, they are often taken as 0.5. The concen- limiting case, when the oxygen concentration is
tration terms represent the ratios between the re- actually zero at the interface, the corrosion cur-
actant concentration at the electrode/electrolyte rent can, through Faraday’s law, be coupled to
interface and bulk, which could deviate from the steady-state flux of diffusive oxygen supply
unity as a result of consumption/production of through a boundary layer δ. This limiting case
species at the interface. In an iron-based alloy, current is called the limiting current (iL or jL)
this ratio for the anode would be close to unity and can be expressed as:
because the reactant is iron itself, and no
concentration gradient would be expected as a DO nFCO*
result of the corrosion reactions. When a net jcorr = jl = 2 2 (Eq 11)
corrosion current flows, icorr = ia = ic. If the cath- δ
ode and anode areas are assumed to be equal,
then jcorr = ja = jc and Eq 5 and 6 can be rewrit- In a nonlimiting case, the corresponding
ten (using βc = 2.3RT/(αnF) and βa = 2.3RT/ equation would be:
((1 ⫺ α)nF) as:
DO nF (CO* − CO (0, t ))
j 0 ,c CO (0, t ) jcorr = 2 2 2
(Eq 12)
ηc = β c log + β c log 2 δ
(Eq 7)
jcorr C O*
Combining Eq 11 and 12, one obtains:
CO (0, t ) jcorr
= 1− (Eq 13)
ηa = β a log corr (Eq 8) CO * jl
j0 ,a
Thus, Eq 7 can be written:
Tafel Regime: Electrode-Kinetics Control.
If the electrode charge transfer reactions are
rate limiting, the supply of oxygen to the reac- j 0 ,c ⎛ j ⎞
ηc = β c log + β c log ⎜ 1 − corr ⎟ (Eq 14)
tion site would be rapid enough to maintain a jcorr ⎝ ji ⎠
concentration at the electrode close to that of
the bulk. In this case, Eq 7 and 8 would both re- The slower the diffusion (small d), the lower
sult in a linear dependence of the overpotentials the limiting current and thus a larger contribu-
versus log jcorr: tion from the mass-transfer-dependent second
term on the overpotential.
ηc = β c log j0 ,c − β c log jcorr (Eq 9)
Migration and Ionic Diffusion
The ionic transport in the electrolyte phase,
ηa = β a log jcorr − β a log j0 ,a (Eq 10) the flux of an ion i under an electric field φ
across a distance L, can be shown to be:
Mass Transfer Control. In Eq 7, the term:
∂ci zi F Δφ
CO (0, t ) J i = − Di − Dc (Eq 15)
2 ∂x RT i i L
In an electrolyte with many different ions, an
stands for the ratio of oxygen gas concentration ion current through an area A can be computed
at the electrode/electrolyte interface and the by multiplying Eq 15 with zi * A and summing
concentration in the bulk, sufficiently far away the contribution from all ions:
from the interface. If the electrode reaction ki-
netics are very fast, the depletion of oxygen will ∂Ci + ( x ) F2A
i = FA∑ zi Di + +
lead toward a zero oxygen concentration at the ∂x RT
22 / Stainless Steels for Design Engineers

and the respective equations describing the

× ∑ zi 2Ci Di Δφ / L (Eq 16) overpotentials will be:

j 0 ,c ⎛ j ⎞
Because the first term is important only at the ηc = β c log + β c log ⎜ 1 − corr ⎟ (Eq 21)
jcorr ⎝ ji ⎠
regions near the electrodes (where consump-
tion/creation of species occur), the current in
the majority region of the electrolyte can be es- and
timated as:
ηa = β a log (Eq 22)
F A j0 ,a
∑ zi 2Ci Di Δφ / L (Eq 17)
Now, the potentials of anode and cathode
Using Ohm’s law (R = U/i), the electrolyte re- when current is flowing are in each case the
sistance can be computed as: equilibrium potential plus overpotential, that is:

⎛ F2A ⎞ RT PH2
Relectrolyte = L / ⎜
⎝ RT
∑ zi 2Ci Di ⎠⎟ (Eq 18) Ecathode = EH0 + / H −
nF aH2 +
+ ηc

The resistivities of some test solutions are j0 ,c

= 0 − 0.059pH + β c log
shown in Table 2. jcorr

⎛ j ⎞
Mixed Potential Theory and + β c log ⎜ 1 − corr ⎟ (Eq 23)
⎝ ji ⎠
Polarization Diagrams

Viewing the electrochemical cell as an elec-

trical circuit, Kirchoff’s law can be used to
design a so-called polarization diagram. Con-
Eanode = −0.447 − 0.0295 log aFe 2+ + ηa ( )
sider, as a case study, a steel corroding under jcorr
deaerated conditions, in a water solution, as = −0.624 + β a log (Eq 24)
j0 ,a
shown in Fig. 2. Assume that the pH is such that
a passive layer does not form (see the discus-
sion of Pourbaix diagrams in Chapter 2). The A polarization diagram is now constructed by
cathode and anode reactions, respectively, are: plotting the anode and cathode potentials versus
log jcorr. Strictly speaking, to close the circuit,
2 H + + 2 e− → H 2 (Eq 19)

Fe ( s ) → Fe 2 + + 2 e− (Eq 20)

Table 2 Test solution resistivity

Ratio by Resistivity,
Test solution volume ohm-cm
Natural Seawater ... 25
Fresh (tap) water adjusted 28:1 500
with seawater
Fresh (tap) water adjusted 68:l 1,000
with seawater
Fresh (tap) water adjusted 950:1 3,000
with seawater
Deionized water adjusted 21:10 10,000
with fresh (tap) water
Fig. 2 Schematic polarization diagram
Chapter 3: Corrosion Kinetics / 23

the potential drop across the electrolyte needs to Passivation

be included, which simply equals icorr * Relectrolyte
(the electrolyte resistance is evaluated from Eq Theory. In Chapter 2, it was identified
18); however, in many cases, this term can be through the Pourbaix diagrams that there were
neglected. A schematic polarization diagram is conditions under which an alloy could be
shown in Fig. 3. The anode polarization is linear passive. In the case of stainless steels, the range
with decade current as predicted by Eq 24 be- of pH and other conditions under which this
cause the overpotential has only a Tafel regime would occur has been increased thanks to the
and no mass transfer dependence. On the other chromium content, which readily forms a Cr2O3
hand, the cathode polarization deviates from the scale. In general, a passive layer constituted of
Tafel behavior as a result of the effect of the adsorbed molecules or thin oxide/hydroxide
mass transfer (hydrogen ion supply), dependent layers decreases the corrosion current. Re-
on the limiting current in Eq 23. It is notewor- searchers (Ref 3) have reported that the con-
thy that the cell shown in Fig. 1 does not have a stituents of the passive film are alpha Cr2O3 and
macroscopic anode and cathode. Different mi- Cr(OH)3nH2O. The structure is reported to be a
croscopic regions on the surface are assumed to nanocrystalline spinel, epitaxial to the surface.
act as cathodes and anodes, and in the lack of The grain size may decrease with increasing
more detailed knowledge, the cathode and chromium content. This protection by
anode areas are assumed to be equal. The over- chromium requires a threshold level of 11 to
all mixed potential of the surface would be at a 12% chromium.
corrosion potential Ecorr, defined in Fig. 2. Effect on Polarization Diagrams. The polar-
In effect, the corrosion current resulting from ization diagram for a passive alloy is quite dif-
the cell depends on the equilibrium half-cell po- ferent from those discussed for active alloys. A
tentials (Ecathode and Eanode), the Tafel slopes (βc schematic of a typical polarization curve is
and βa), the exchange current densities (jo,c and shown in Fig. 5. When a passive alloy is anodi-
jo,a), and any limiting current density (jl). Figure cally polarized, it initially behaves like an active
3 shows schematically how decreasing any of the alloy (i.e., with a Tafel slope, etc., as the pas-
Tafel slopes and increasing an exchange current sive layer is building up). The building up is
density increases the corrosion rate. The effect of actually a selective dissolution of iron, which
the electrolyte resistance has been ignored; that causes a greater remaining surface concentra-
is, corrosion current is where the two polariza- tion of chromium and other alloying elements.
tion curves intersect. Figure 4 shows the effect of Once the passive layer is formed and offers
increased mass transfer, which would result in an protection against further dissolution, the po-
increase in the limiting current. In an active (non- tential-decade current relation drops to lower
passive) alloy, this results in an increased corro- currents. This happens at potentials beyond the
sion current up to a point. passivation potential Epp. At some high enough

Fig. 3 Corrosion rate and the effect of (a) Tafel slope and (b) exchange current density
24 / Stainless Steels for Design Engineers

polarization level, the passive layer breaks j0 ,c

down, and the metal becomes active again; this = 0 − 0.059pH + β c log (Eq 25)
region is called the transpassive regime. The
design of a structure involving a passive metal
should aim at forming a corrosion cell in which This will result in a straight line as shown in
the cathode polarization curve intersects the an- Fig. 6, which will be shifted vertically depend-
odic one in the passive regime. ing on the pH. The dashed circles indicate the
Consider, for example, an alloy that exhibits intersection between anode and cathode polar-
the behavior shown in Fig. 6 in deaerated ization curves that would yield the corrosion
acidic solutions with different pH. If mass current. At a sufficiently high pH (= pH1), the
transfer limitations due to hydrogen ion supply alloy is clearly not optimal because intersection
are neglected, then the cathode polarization is occurs in the active regime, and the passive
given by: properties are not utilized. This is what occurs
when a reducing acid is too strong for a given
stainless steel, such as with concentrated hy-
RT PH2 drochloric acid. At pH = pH2, on the other
Ecathode = EH0 + / H − ln + ηc hand, a low-corrosion current is obtained as a
2 nF aH2 +
result of intersection at the passive regime.
This is the benevolent case when stainless steel
is correctly matched to the environment, and
low rates of uniform corrosion occur. Finally, at
pH = pH3, the resulting corrosion current is
again high as a result of intersection occurring
at the transpassive region. This could occur
with some stainless steels exposed to a very
strong alkali solution. Using a similar argu-
ment, the readers can themselves deduce the
effects of cathode exchange current density and
Tafel slopes. In the discussion of active anode
polarization, it was found that increasing the
transport rate of cathode reactants through
agitation, for example, would increase the cor-
rosion rate up to a point but beyond that have
no further effect (see Fig. 4). In the case of a
passive/active behavior, the effect of mass
Fig. 4 Effect of increasing the limiting current by, for example,
increased agitation in the electrolyte. Beyond the
dashed line, increasing the limiting current would have no fur-
ther effect

Fig. 5 Schematic of a passive anode polarization curve Fig. 6 Effect of cathode polarization
Chapter 3: Corrosion Kinetics / 25

is the actual passivation; the iron removal is

really a chemical cleaning operation, which hap-
pens to be called passivation. During the produc-
tion of stainless steel, after a final anneal another
version of passivation is carried out. The oxide
from annealing in air is dissolved by a strong
mixture of nitric and hydrofluoric acids, which
does not allow passivation. This treatment,
called pickling, removes by dissolution both the
oxide layer and the chromium-depleted layer
below the oxide formed during annealing. The
depleted layer can extend a number of microns
in depth and would seriously degrade corrosion
resistance if not removed (Ref 4). This is then
followed by a straight nitric acid immersion,
Fig. 7 Effect of mass transport which ensures complete passivity. This is the
procedure that should be performed on the ox-
transport is somewhat different, as shown ides formed during welding if full corrosion re-
schematically in Fig. 7. sistance is to be restored. Simply removing the
Increasing mass transport, such that the limit- oxide through mechanical means leaves a
ing current increases, results initially in an in- chromium-depleted layer that corrodes more
creased corrosion current (e.g., increasing jl readily than is expected of the alloy.
from 1 to 2). It should be noted that there are
several intercepts possible (both in the active
and passive regime), but assuming there are REFERENCES
defects present, it is likely that there will be cor-
rosion corresponding to the higher current. In- 1. D.A. Jones, Principles and Prevention of
creasing the limiting current beyond the knee Corrosion, 2nd ed., Prentice Hall, 1996
corresponding to Epp, however, results in a drop 2. A.J. Bard and L.R. Faulkner, Electrochemi-
in the current because now the only corrosion cal Methods: Fundamentals and Applica-
potential possible is at the intersection in the tions, 2nd ed., Wiley, 2001
passive regime. This is the case for jL3. 3. M.P. Ryan et al., Critical Factors in Focal-
In the normal use of stainless steel, achieving ized Corrosion, Proc. Electrochem. Soc.,
passivity takes on several forms. What is often Vol 150, 2003, p 583–594
called passivation is actually a cleaning process 4. J. Grubb and J. Maurer, “Corrosion of the
in which contaminants, such as tramp iron, are Microstructure of a 6% Molybdenum Stain-
removed from the surface. Dilute nitric acid is less Steel with Performance in a Highly Ag-
an excellent vehicle to achieve this. This gressive Test Medium,” paper 300 pre-
medium has the additional benefit of forming a sented at Corrosion 95, NACE
passive film on an active stainless surface. This International, 1995
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 27-56 All rights reserved.
DOI: 10.1361/ssde2008p027


Corrosion Types

Summary plex in their behavior because the influence of

processing and alloying variables changes the
STAINLESS STEEL is unusual among alloy ability of this layer to form and remain stable in
systems in that its corrosion resistance derives the face of aggressive environments. The be-
from the passivating ability of a minor con- havior of stainless steel is further affected by its
stituent, chromium. Thus, while stainless steels microstructural complexity. Stainless steel al-
can be made to be essentially immune to corro- loys may have many constituent elements and
sion in many environments, it can also experi- many thermodynamically possible phases, and
ence various debilitating forms of localized cor- none of these are necessarily uniform in their
rosion, which stem from the failure of this composition. Yet, it is the composition of the
passive film. This chapter explores the behavior alloy in contact with the specific environment at
of stainless steel in media that promote uniform any microscopic point that determines the cor-
corrosion and the various mechanisms of local- rosion resistance of that particular point.
ized corrosion, such as pitting and crevice cor-
Uniform Corrosion
Introduction When all parts of a corroding surface have
equal access to the corroding atmosphere and
To most designers, the most recognized char- the structure of the corroding metal is relatively
acteristic of stainless steel is corrosion resist- uniform, a uniform thinning of the material is
ance. Stainless, unlike noble metals such as expected. Stainless steels are materials of choice
gold, does not obtain its excellent corrosion re- because, by virtue of their passive behavior, they
sistance from inertness. Instead, it is the reactiv- show very low rates of uniform corrosion in
ity of chromium that allows the surface layer of many environments. The metallurgy and pro-
corrosion product to become sufficiently adher- cessing of a particular grade are designed to pro-
ent and impenetrable, which effectively stops vide passivity in a given environment. The envi-
further corrosion by isolating the base material ronment can be too aggressive to allow passivity
from the environment. This resistance to corro- to be maintained either by being too reducing, as
sion is called passive behavior or passivity. with some acid media, so that passivating
Other metals, such as aluminum and titanium, species cannot form or by being too oxidizing so
form similar layers and also exhibit passivity. that the oxidized species that normally affect
The important difference in the case of stainless passivity are no longer stable. The former is
steel is that chromium is still a minor con- called dissolution in the active state, while the
stituent, never more than 30% by weight, some- latter is termed transpassive dissolution.
times little more than 10%. How much Intelligent design and knowledge of the envi-
chromium there is and how uniformly it is dis- ronmental variables for a stainless steel compo-
tributed have a profound effect on corrosion re- nent ensure that the alloy is used in the passive
sistance by virtue of its ability to concentrate state, at which uniform corrosion occurs at a
into the surface film. Stainless steels are com- very low rate.
28 / Stainless Steels for Design Engineers

Among the important media with which we sive, or transpassive; Fig. 1 illustrates the effect
encounter uniform, but acceptably controlled, of redox potential on a solution.
corrosion in stainless steel are atmospheric and Certain anions have strong effects in media
marine environments and chemical environ- through their well-known, if not well under-
ments such as sulfuric acid, phosphoric acid, ni- stood, disruption of the passive film. Halides are
tric acid, strong bases, and organic acids, such well known for this effect, but sulfides are
as acetic and formic. active. These anions seem to intervene in the
Pickling is an example of controlled, acceler- adsorption of the hydroxyl ions. In acid media,
ated uniform corrosion. This is typically done these anions accelerate uniform corrosion,
with 10 to 20% hot sulfuric acid or a mixture of while in neutral media they may result in local-
hydrofluoric and nitric acids. ized corrosion. Anions that form soluble com-
plexes with elements in stainless, such as
Environmental Variables Influencing amines, formates, or acetates, can also disrupt
the stability of the passive film and thus pro-
Uniform Corrosion
mote active corrosion.
The corrosion of stainless steels is usually the Of the physical variables of the environment,
result of contact with an electrolyte, allowing a it should be obvious that temperature is para-
complex set of partial electrochemical reac- mount since all the reactions are thermally acti-
tions, which may occur sequentially or concur- vated. Increasing temperature may speed the
rently. The corrosion rate depends on the formation of the passive film when thermody-
current exchanged between the negative and namic conditions are favorable, but in general
positive electrode (anode and cathode). These one expects increasing temperature to increase
may be on a macroscopic or microscopic level. corrosive attack. Access to passivating species,
The main consideration is normally ionic trans- such as oxygen, is important in establishing and
por tthrough the passive film, which after all maintaining passivity.
is what makes stainless so effective against Increased diffusion of reacting species in the
corrosion. liquid will normally accelerate the partial reac-
The chemical parameters that influence the tions, but if passivity is stable, the rate-limiting
media with respect to uniform corrosion rate are transport through the passive film will not be af-
the acidity and the oxidation-reduction (redox) fected. Therefore, increasing the flow rate of a
potential of the electrolytic medium, both of corrosive fluid does not automatically acceler-
which act through their influence on the stabil- ate corrosion. The reduction of concentration
ity of the passive film, rendering it active, pas- gradients can be beneficial against localized

Fig. 1 Reduction potential versus pH for iron and (a) oxygen gas reduction and (b) hydrogen ionreduction
Chapter 4: Corrosion Types / 29

corrosion, and flow can bring to the surface an stituent of sigma phase, which it promotes,
increased supply of passivating species. however, it combines with chromium. If this
Increased flow rate in a fluid medium is dele- happens at relatively low temperatures, the sur-
terious if it induces mechanical damage to the rounding matrix is depleted of both chromium
passive film by erosion, abrasion, or cavitation. and molybdenum, and the corrosion resistance
These are complex mechanisms, but it should in that region is diminished. Nitrogen also is ef-
be apparent that the success of a stainless steel fective when in solid solution in austenite but
to a given flow condition will depend mainly on can precipitate as a chromium nitride under cer-
its ability to form and re-form its passive film. tain conditions and cause depletion of the re-
Somewhat counterintuitively, thinner passive maining matrix. Local structure and composi-
films are more protective than thicker films tion are paramount. This becomes more
among stainless alloys. The tenacity of the thin important to localized corrosion, as discussed
passive films on stainless (and titanium) make later, but it should be remembered in examining
these alloys quite resistant to flow-accelerated uniform corrosion because corrosion will cease
corrosion, as contrasted to copper and alu- to be uniform when composition becomes
minum alloys, which have soft, thick corrosion nonuniform.
product films. The compositional material variables that in-
fluence uniform corrosion are not exactly the
same as those that will be seen to influence lo-
Material Variables
calized corrosion. The foremost element is, of
Stainless steels have a great variety of alloy- course, chromium. Researchers (Ref 1) have re-
ing elements and microstructures. As a general- ported that the constituents of the passive film
ization, we can say that corrosion resistance is a are alpha Cr2O3 and Cr(OH)3nH2O. The struc-
function of composition rather than structure. ture is reported to be a nanocrystalline and epi-
Then, we must quickly add the qualifiers to this taxial to the surface. The grain size may de-
statement. On an undisrupted, stress-free sur- crease with increasing chromium content. This
face, local composition does quite precisely de- protection by chromium requires a threshold
termine corrosion resistance. But, stainless level of 11 to 12% chromium. This threshold
steels are seldom homogeneous or at thermody- has been attributed most convincingly to the
namic equilibrium. Impurities such as oxygen minimum chromium content that permits
and sulfur are usually present, mainly as inclu- chromium atoms on surface sites to be linked by
sions since they have diminishingly small solu- adsorbed oxygen atoms (Ref 2). In any event,
bility at room temperature. At high tempera- the mechanism by which this thin, several-
tures after solidification, as in welds, they can nanometer-thick, film forms is the subject of on-
be present in supersaturation, ready to precipi- going debate, but we do know that it is enriched
tate as inclusions that alter local composition. in chromium, and that it is thinner for higher
The tendency of carbon and nitrogen to form chromium alloys. The critical current density j,
precipitates is controlled by diffusion rates, as measured during polarization, is also smaller
which if elevated by increasing temperature can as chromium content increases (Fig. 2). This is
cause debilitating, composition-altering precipi- consistent with the lower dissolution of noncon-
tation. The even more slowly diffusing substitu- tributing elements required to achieve a critical
tional alloying elements, such as chromium, surface chromium concentration. Increases in
molybdenum, and nickel, have strong tenden- chromium can also be seen to lower the current
cies to form phases that disturb their uniformity density in the passive region. This is manifest in
in the austenite or ferrite matrix in which they alloy performance as a reduction in the uniform
are intended to work. So, any discussion of the corrosion rate in a given medium. From an elec-
influence of alloying element on corrosion re- trochemical point of view, this is explained as a
sistance of a phase like austenite or ferrite must manifestation of the stability of the Cr(OH)3
recognize that alloying elements exert their ef- nH2O.
fect when they are in solution in that phase. The The role of molybdenum is less clear. The ob-
same element may under some conditions not served action of molybdenum is to greatly re-
be in solution and have a contrary effect. An ex- duce the critical current density required for
ample is molybdenum, which is obviously a passivation. This is also seen as accelerating the
great enhancer of corrosion resistance when in formation of the passive films and as increasing
solid solution. When it precipitates as a con- the resistance of the alloy to depassivation at
30 / Stainless Steels for Design Engineers

Fig. 3 Influence of alloying element on corrosion rate as ex-

plained by the effect on polarization.Source: Ref 6

Fig. 2 Schematic illustration of polarization behavior for a

passive alloy with and without pitting occurring

lower pH. The role of molybdenum is not to en-

rich in the passive film itself, although it can be
found in the film. Its potency is far more than its
presence can take into account. Pure molybde-
num is itself not passive. Its action does not ap-
pear to be via a product of reaction. Instead, it
seems to reduce the dissolution rate of elements
other than iron, which would promote a surface
richer in chromium (Ref 3). The action of
molybdenum as an alloying element is compli-
cated by the fact that molybdate ions are known
to impede pit growth as a separate effect from
Fig. 4 Influence of alloying elements on uniform corrosion
their action within the alloy matrix (Ref 4). rate in 20% sodium chloride solution with carbon
Copper has a similarly complicated effect, with dioxide pressure of 20 MPa. Source: Ref 7
copper ions gettering sulfide ions and redeposit-
ing as metallic copper (Ref 5). chromium in the passive layer and to decrease
Nickel also lowers the critical current density active dissolution of noniron alloying elements,
for passivation without contributing directly to thereby promoting both the formation and sta-
the passive film’s stability. This also may be the bility of the passive film. A summary of the
result of the stronger bond between nickel and known major alloying effects in acidic chloride
chromium reducing the anodic dissolution rate media is shown in Fig. 3 in acidic chlorides. Al-
of the alloy by permitting the anodic enriching loying elements provide benefits in the part of
of the surface by selection iron dissolution. the chart where they appear (Ref 6). From this,
Nickel does not actively help passive film for- it can be seen that chromium, molybdenum,
mation and can actually hinder film stability in nickel, copper, and nitrogen all assist in the ac-
highly acidic/oxidizing environments. tive region, while chromium, molybdenum, and
Nitrogen, however, appears to be more like nitrogen expand the region of passivity and di-
molybdenum in its effect. While nickel and cop- minish the corrosion current.
per provide no benefit to the stability of the pas- An example of the influence of these alloying
sive film once it is formed, both nitrogen and elements on the uniform corrosion rate of stain-
molybdenum do, and to a degree that cannot be less steels in a sodium chloride/carbon dioxide
explained by their presence in the film. This environment is shown in Fig. 4 (Ref 7). Note
may then relate to their thermodynamic action the alloying composition is measured by a
within the alloy itself. Molybdenum and nitro- crevice corrosion index (CCI), which is dis-
gen act both to enhance the enrichment of cussed in the section Localized Corrosion.
Chapter 4: Corrosion Types / 31

Fig. 5 Corrosion table for stainless steels and titanium in sulfuric acid plus copper sulfate. Corrosion rate legend: 0, < 0.1 mm/yr
(corrosion resistant); 1, 0.1–1 mm/yr (useful in certain circumstances); 2, > 1.0 mm/yr (material not recommended). Source:
Ref 8; see source for interpretation of data. Courtesy of Outukumpu Stainless

Unfortunately, it cannot be assumed that this less steels in a great number of environments
relationship is true for other environments, al- can be obtained. Figure 5 shows an example of
though other empirical relationships exist or can one such table. Many of the isocorrosion charts
be generated. Because the influence of alloying in this book are reprinted from this source,
element varies with environment, we need to
discuss some of the more commonly encoun- ments/start.asp (Ref 8).
tered severe environments. These tables are supplemented by isocorro-
sion diagrams such as that shown in Fig. 6.
Corrosion in Acids and Bases These diagrams show constant corrosion behav-
ior under varying environmental conditions such
The examples discussed in Chapter 3, “Cor- as temperature and solution composition. This
rosion Kinetics,” refer mostly to corrosion in information is available to guide the designer in
these aqueous solutions, in which the slow thin- selecting appropriate steels for various environ-
ning rate of the chosen alloy can be determined ments, and it is highly recommended that it be
through the mixed potential theory and polar- used. Free sites tend to promote proprietary al-
ization diagrams. In the case of stainless steels, loys, as these charts suggest. The serious engi-
the alloy chemistry is chosen such that the neer will consult multiple sources and unbiased
passive-active behavior favors corrosion in the sources before making alloy decisions.
passive regime. The corrosion rate of the vari- The influence of alloying element is by no
ous stainless steels in the myriad possible envi- means the same in all environments. So, while it
ronments has been measured in probably all is useful and necessary to have these experi-
practical cases. These data can be obtained from mental data, it is also helpful to understand the
a number of sources, such as the National Asso- peculiarities of some of the major alloy-envi-
ciation of Corrosion Engineers (NACE) and ronment pairings.
ASM Handbook volumes. None is more accessi- Sulfuric Acid. Stainless steels require more
ble than the Web site of Outukumpu, which than a minimum amount of alloying to resist sul-
contains a “Steel Professional Tool,” a lookup furic acid. Straight 16% chromium grades such
table in which the corrosion rate of many stain- as 430 fare poorly, while the nickel-containing
32 / Stainless Steels for Design Engineers

Fig. 6 Isocorrosion curves for 17-12-2.5 stainless steel and titanium in sulfuric acid plus copper sulfate.Source: Ref 8. Courtesy of
Outukumpu Stainless

304 has more than an order of magnitude better

corrosion rate in either dilute or concentrated
sulfuric acid at ambient temperatures. Figure 7
(Ref 9) shows the isocorrosion rate curves for
several common alloys. Alloying with molyb-
denum is also very effective, as is alloying with
copper. If passivity cannot be established, in-
creasing chromium content actually increases
corrosion rate.
The corrosion behavior of sulfuric acid varies
greatly with concentration. At low concentra-
tions, sulfuric is a classic reducing acid. It dis-
sociates in water to create hydrated hydrogen
ions (H3O+) that release hydrogen gas bubbles
during the corrosion reaction. As the acid con-
centration increases, the solutions become more
corrosive, and progressively more highly al-
loyed stainless steels are required to provide ad-
equate corrosion resistance. At about 50% acid,
only the most highly alloyed stainless alloys
(alloy 20, AL-6XN, C-276, etc.) can provide ac-
ceptable corrosion rates, and even these alloys
are restricted to use at near ambient tempera-
tures. As acid concentration increases beyond
50%, the solution begins to show oxidizing be-
havior. At acid concentrations above 80%,
nickel-molybdenum-copper-bearing stainless
steels begin to exhibit useful corrosion resist-
ance. In the 93 to 98% sulfuric acid concentra-
tion range, carbon steel can be used to hold sul- Fig. 7 Isocorrosion rates of various stainless steels in sulfuric
furic acid at ambient temperatures, although acid. Source: Ref 9
stainless steels provide better performance at el-
evated temperatures or if flow-erosion can occur. resistant high-chromium (type 310S) and high-
In the 96 to 100% sulfuric acid concentration silicon (MECS ZeCor UNS S38815 and Sand-
range, at elevated temperatures, the oxidizing vik SX S32615) stainless steels are frequently
character is quite pronounced, and oxidation- used, especially in sulfuric acid-manufacturing
Chapter 4: Corrosion Types / 33

Fig. 8 Influence of alloying element on corrosion rate in contaminated sulfuric acid. Source: Ref 11

non-molybdenum-bearing alloys. Their superi-

ority in deaerated solutions is much less
marked. Studies (Ref 10) have shown that in
sulfuric acid molybdenum is highly enriched in
the passive film, and when molybdenum is an
alloy, chromium also is enriched. This is a man-
ifestation of selective dissolution of other ele-
ments in the matrix.
Oxidizing impurities, such as ferrous ions, act
like aeration to diminish the corrosive attack,
but reducing impurities such as halides have an
extremely negative effect, as the corrosion ta-
bles will show. These effects are not linear and
underscore the value of these tables.
The uniform corrosion rate in contaminated
Fig. 9 Isocorrosion curves for various alloys in sulfuric acid sulfuric acid may be more important than in
pure acid since this represents a potentially
equipment. Sulfuric acid-containing dissolved likely failure mode because contamination is a
sulfur trioxide is called oleum, and such solu- constant hazard.
tions are often identified as sulfuric acid of Figure 8 shows the corrosion rate of various
greater than 100% concentration. High- alloys in sulfuric acid contaminated with chlo-
chromium stainless steels (i.e., type 310S) are rides and iron. These researchers (Ref 11) found
among the very few materials that exhibit corro- that the resistance to attack correlated to the
sion resistance in oleum. (See MTI Materials alloy content by the formula shown.
Selector Volume 3—Sulfuric Acid at www.mti- Figures 9 and 10 show how isocorrosion rates for more information.) vary with alloy and contamination level.
Aeration has a major influence on corrosion Hydrochloric acid is very destructive of the
rates because oxygen stabilizes the passive film. passive film on stainless. An alloy like 304 is
Molybdenum-alloyed stainless has dramatically not suitable even in a deaerated 1% HCl solu-
lower corrosion rates in aerated solutions than tion at room temperature. Chromium additions
34 / Stainless Steels for Design Engineers

Fig. 12 Isocorrosion curves for austenitic AL-6XN (UNS

N08367) and 904L (UNS N08394) stainless steels
Fig. 10 Isocorrosion curves for various alloys in sulfuric acid in hydrochloric acid. Source: Ref 12
with chlorides

Nitric acid is strongly oxidizing. This actu-

ally promotes the passive film formation; con-
sequently, even low-chromium alloys remain
passive at all concentrations at ambient temper-
ature (see Fig. 14) (Ref 8).
The addition of molybdenum, which is so
generally helpful, is deleterious in this case be-
cause it forms soluble compounds. It is useful
to keep carbon, silicon, and phosphorus as low
as possible.
Silicon is unusual in that normal levels (0.4 to
1.0 %) are worst, with very low (0.05%) or very
high levels (4.0%) beneficial. The low levels of
silicon contents of these alloys are useful for
their action in minimizing grain boundary seg-
Fig. 11 Isocorrosion curves for various stainless steels in
regation, which is the usual locus of attack.
hydrochloric acid. Source: Ref 8. Courtesy of Out-
okumpu Stainless High silicon levels contribute to a general pro-
tective silica surface layer in concentrated acid,
which augments the true passive layer. This su-
periority appears above the azeotropic composi-
are only modestly helpful, while nickel, copper, tion of about 67%, which is a common com-
and molybdenum are more beneficial. Stainless mercial concentration, as shown in Fig. 15 for
steels are not good materials for contact with high-silicon austenitic stainless steels (Ref 13).
hydrochloric acid. Figures 11 (Ref 8) and 12 Phosphoric Acid. This oxidizing acid be-
(Ref 12) show how even the most highly al- haves more like sulfuric acid in that simple
loyed grades can withstand only dilute concen- iron-chromium alloys have only moderate re-
trations and low temperatures. sistance to uniform corrosion in them, while al-
While a stainless steel vessel may not be in- loying with molybdenum and copper produces
tended to be used for hydrochloric acid, resist- major improvements. This can be seen in Fig. 16,
ance to lesser amounts of chlorides is important in which alloys with increasing nickel (18-10)
because of the possibility that an acidic environ- show clear benefits over a chromium-molybde-
ment may be contaminated with chlorides. When num alloy (18-2), and the added molybdenum
this is a possibility, then proper alloy selection in 317 (17-14-4) is better, while 904L with
must guard against it. Figure 13 shows a correla- nickel, molybdenum, and copper is even better
tion between alloy content and resistance to the (Ref 8).
general corrosion (GI) by sulfuric acid contami- In the commercial production of phosphoric
nated with hydrochloric acid (Ref 11). acid, halide impurities may be present, in which
Chapter 4: Corrosion Types / 35

Fig. 13 Influence of alloy content on corrosion rate in hydrochloric acid

case alloys with higher molybdenum, chromium,

copper, and nitrogen may be required.
Organic Acids. The weakly dissociating or-
ganic acids are normally not aggressive against
stainless steels. The exceptional dangerous en-
vironments are those that include high tempera-
ture and the presence of chloride contamination.
It should be noted that in formic acid, which
does dissociate more strongly, nickel is detri-
mental. This phenomenon is also seen in the
production of urea via the intermediary ammo-
nium carbamate. The difficulty lies in the high-
temperature solubility of nickel complexes and
is best addressed by the use of ferritic or duplex
Fig. 14 Isocorrosion curve for nitric acid. Courtesy of
Outokumpu Stainless Alloying with molybdenum seems to provide
the greatest resistance to uniform corrosion in
strong organic acids, as illustrated in Fig. 17
(Ref 8). If halides are present in organic acids
and liberated by contact with water, then pH
and chloride concentration will govern the cor-
rosive attack, which could then become nonuni-
Strong Bases. In strong bases, the stainless
steels are generally quite resistant to uniform
corrosion. Straight chromium (17%) alloys are
usable at any concentration up to 50 °C. Adding
molybdenum and nickel does little to further
improve performance as the underlying resist-
ance is due to chromium. Increasing chromium
levels provide increased resistance. Attack
when it does occur can be manifested as grain
Fig. 15 Corrosion behavior of high-silicon alloys in concen- boundary attack. Figure 18 shows isocorrosion
trated nitric acid. Courtesy of Outokumpu Stainless curves for sodium hydroxide (Ref 8).
36 / Stainless Steels for Design Engineers

Fig. 16 Isocorrosion curves in phosphoric acid: (a) 0.1 mm/yr for various stainless steels; (b) 0.1 mm/yr for titanium and 17-12-2.5
stainless steel. Courtesy of Outokumpu Stainless

Fig. 17 Isocorrosion curves in organic acids: (a) acetic acid; (b) formic acid. Source: Ref 8. Courtesy of Outokumpu Stainless

In the pulp-and-paper industry, chemical

pulping is called the kraft or sulfate process. In
the presence of sulfur, nickel can be quite detri-
mental, and ferritic or duplex alloys are
preferred. This again is caused by the solubility
of nickel complexes formed in the presence of
sulfur-containing compounds. This can be seen
in Fig. 19, which shows a 26-1 (chromium-
molybdenum) alloy significantly outperforming
higher alloys that contain nickel and molybde-
num (Ref 12).

Atmospheric Corrosion
Atmospheric corrosion is an example of uni-
Fig. 18 Isocorrosion curves for various materials in sodium
form corrosion that occurs when a thin layer of
hydroxide. SCC, stress corrosion cracking. Courtesy
of Outokumpu Stainless water condenses on a metal surface and as such
Chapter 4: Corrosion Types / 37

2) that atmospheric corrosion can in effect be

The most deleterious impurity in the atmos-
phere for stainless is the chloride ion. Chlorides
are pervasive. Borne from oceans by normal cli-
matological processes, they are found far inland.
In many colder climates, they are also seen in
high concentrations from road salts. Without
washing or the natural rinsing by rain, surface
chloride concentrations can become very high.
Thus, the rules of corrosion of aerated aqueous
solutions are followed by stainless with respect
to atmospheric corrosion. The difficulty is accu-
rately estimating the solution that constitutes the
aqueous solution. Much experience has shown
that if coastal and road salt effects are minimal,
then 18% chromium alloys such as 304 experi-
Fig. 19 Corrosion rates of various alloys in simulated evapo-
rator liquid. Source: Ref 12 ence such negligible visible corrosion that they
can be used for exposed, unrinsed architectural
depends on humidity, temperature, and other at- purposes. If the same alloy is used in an unrinsed
mospheric conditions. The rate of corrosion coastal environment, red rust stain will occur.
measured as defined in Chapter 3, “Corrosion This is the corrosion product from metastable
Kinetics,” as dissolution r of element i as and possibly stable pitting. In Japan, where
coastal conditions prevail throughout, much re-
t ⋅ Mi search has been done that has shown that a pit-
r= j ting resistance equivalent number (PREN) of 25
ni ⋅ F ⋅ρ
is necessary for freedom from corrosion (i.e.,
zero pitting) (Ref 14). This contrasts to a re-
where j is current density, t is time in seconds, quirement of about PREN 35 to resist pitting in
M is molar mass, n is valence, F is the faraday seawater. Pitting is a form of localized corro-
constant, and ρ is the density), and therefore has sion, and PREN is an index to pitting resistance.
two contradicting effects of temperature. In These concepts are examined in the next section.
general, temperature increases the exchange
current density and transport properties and thus
Localized Corrosion
the kinetic rates involved in corrosion. On the
other hand, increasing temperatures may re- Localized corrosion is in general more
duce the concentration of dissolved oxygen in damaging from a structural integrity point of
the electrolyte and eventually will dry the sur- view than uniform corrosion since the corro-
face and thus limit the electrochemical corro- sion current is limited to a small area and the
sion due to the access to an electrolyte. In steels, penetration distance is large. Often, localized
the corrosion products are (a) an outermost corrosion involves a large-area cathode and a
layer of porous rust (FeOOH) characterized by small-area anode, which means that for a given
low water content but easy access to oxygen corrosion current, the corrosion current density
and (b) an inner layer of magnetite (Fe2O3) in at the anode is very large. In localized corro-
which pores are filled with water. The access of sion, unlike uniform corrosion, the anode and
oxygen to the bare metal limits the cathode re- cathode are clearly identifiable locations, and
duction reaction rate, and in relatively pure at- the reason that certain structural features as-
mospheres, the corrosion rate is negligible due sume the roles of cathode and anode can be
the protective nature of the oxide. However, sul- used to categorize and exemplify different
fur dioxide impurities in the atmosphere react cases. Interestingly, the cathode and anodes,
with water to form sulfuric acid, which tends to while identifiable, can vary across scales, that
dissolve the protective oxide. In the case of is, from distinct macroscopic components or
stainless steels, the passive region is extended parts to microstructural features.
due to chromium, to a wide enough region in In Chapter 2, the tendency for corrosion
terms of pH (see Pourbaix diagrams in Chapter was introduced as a positive value for an
38 / Stainless Steels for Design Engineers

electrochemical cell potential (Erxn) correspon- arises in which access of oxygen is not the same
ding to a spontaneous electrochemical reaction to different areas of a sample. In effect, this re-
forming a galvanic cell. Erxn is obtained as: sults in that the cathode reaction:

RT ⎛ p1 1 p2 2 p3 3 ... ⎞
α α α
O 2 + 2H 2 O + 4e − → 4 OH − (Eq 3)
Erxn = Erxn
− ln ⎜ β β β ⎟ (Eq 1)
nF ⎝ r1 1 r2 2 r3 3 ... ⎠ is limited from proceeding in some areas but not
others. This gives rise to a differential aeration
Here, pi and ri are the concentrations of reac- cell. For example, consider Fig. 20(a), in which
tant and products, respectively, and αi and βi are a metal is partially immersed in water. Trans-
the numbers that are needed to balance the reac- port distance of oxygen increases with depth;
tion stoichiometrically. Any time that Erxn is the limiting current would then vary with depth,
positive, there is thermodynamically a tendency such as at locations 1 and 2, and cathode polar-
for an electrochemical reaction, in our case a ization curves as a result of this are schemati-
corrosion reaction. The rate of corrosion, as dis- cally plotted in Fig. 20(b). Near the surface of
cussed in Chapter 3, is dependent on the polar- the water, where oxygen is readily replenished,
ization behavior. passivation is likely to be fast, and thus anodic
iron dissolution is slow. This region assumes the
role of the cathode, and reaction 1 occurs. Suffi-
Dissimilar Metals and Differential ciently far away from the surface, if there are
Aeration Cells regions where passivation is incomplete (e.g.,
surface defects or scratches) or has broken
The case of dissimilar metals and differential down as a result of, for example, Cl− (see sec-
aeration cells is perhaps more important in ac- tion on pitting), repassivation does not readily
tive alloys than for stainless steels, which are occur since oxygen transport is too slow. These
generally passive, and occurs when two regions become anodes where the following re-
metals/alloys are in contact that have elements action occurs:
in them that are dissimilar in the electromotive
force (emf) series (see Chapter 2) and there is an 2Fe (s) → 2Fe 2+ + 4e − (Eq 4)
electrolyte present. For example, if nickel and
iron pipes are connected and water flows though
them containing some traces of Ni2+ ions, then: The distance at which this occurs is balanced
by being large enough to limit the rate of
RT ⎛ aFe 2 + ⎞
oxygen transport but not too long to be
Erxn = ENi
/ Ni
− EFe0 2 +/ Fe − ln ⎜ ⎟ (Eq 2) strongly influenced by ion transport that is
2 F ⎝ aNi 2 + ⎠
needed to complete the electrochemical cell.
Resulting corrosion currents are shown in Fig.
In this case, the corrosion tendency is prima- 20b. This type of degradation is called water-
rily caused by the first two terms on the right line corrosion.
side of Eq 2, the dissimilarity in the standard Crevice Corrosion. In stainless, the more
half-cell reduction potentials: significant occurrence of this type of cell occurs
when a crevice, from whatever cause, exists,
ENi 2+
/ Ni
− EFe0 2 + / Fe = − 0.250 + 0.447 = 0.197 V and reactions within the crevice or pit cause the
accumulation of iron ions by:
This tendency is caused by the galvanic dis-
similarity between the metals. This is normally 2Fe (s) → 2Fe 2+ + 4e − (Eq 5)
important for alloys joined to stainless that are
themselves less noble. Less-noble alloys, such The regions adjacent to the drop that main-
as carbon steel, can fail rapidly if coupled to tained their passive layer and have access to
stainless. A classic example is the use of carbon oxygen act as cathodes where the oxygen reduc-
steel fasteners for joining stainless sheets. Dif- tion reaction takes place:
ferent stainless steel alloys have minor differ-
ences when passive, but if the environment is O 2 + 2H 2 O + 4e − → 4 OH − (Eq 6)
such that one alloy is active while another is
passive, then the galvanic differential could be This reaction maintains an alkali solution. As
harmfully large. In many cases, a situation a result of the geometry, Fe2+ ions remain and
Chapter 4: Corrosion Types / 39

Fig. 20 Schematic illustration of (a) sample partially immersed in water; (b) resulting polarization behavior for two different passi-
vating alloys (A and B polarization curves)

enrich in the water-filled pit; to maintain charge Pitting corrosion is important to designers be-
neutrality, Cl– migrates into the pit. This causes cause it is corrosion under conditions at which
the following reaction: corrosion may not have been anticipated. Thus,
it is both a materials selection and an environ-
Fe 2+ + 2H 2 O + 2Cl − → Fe ( OH 2 + 2HCl (Eq 7) mental control problem. Its consequences may
be only cosmetic, such as on a building or appli-
which has several consequences: (a) Hy- ance facade, or potentially catastrophic, such as
drochloric acid further acidifies the pit and if leaks of toxic materials were to result from
increases the rate of iron dissolution since de- perforation. Stainless steels are designed to be
creasing pH increases cathode half-cell poten- passive, and localized corrosion is the local loss
tial, which increases corrosion rate (see polar- of passivity. Whether the consequences are
ization diagram construction in Chapter 3, major or not, it is always undesirable, and good
“Corrosion Kinetics.”) (b) The formation of design allows it to be avoided.
porous Fe(OH)2 further helps to isolate the pit, What do we know for certain about pitting?
thereby separating anode and cathode regions in We know quite a lot, really. Experts now con-
the differential aeration cell. (c) The presence of clude that since the early 1970s the local chem-
Cl– prevents repassivation. istry of pitting has been understood (Ref15).
As a result of increased acidification, the dis- The greatest contributions to this field have
solution rate becomes autocatalytic, and as a re- been electrochemical studies. The tools of elec-
sult the pit grows in depth. trochemistry have been especially successful in
At the outside, the reaction: elucidating the mechanism involved in pit
growth and pit stability (Ref 16). The local en-
2Fe ( OH 2 + O 2 + H 2 O → 2Fe ( OH )3
(Eq 8) vironment within pits has been sufficiently
measured and correlated with cavity geometry
further consolidates the isolation of the pit and that some experts can say, “In a sense, all pitting
impedes the ingress of oxygen. is crevice corrosion” (Ref 15). This is to say that
the electrochemistry of cavities such as pits and
crevices is quite similar and has been well mod-
Pitting Corrosion eled. These same tools, however, have been
Pitting corrosion is the most intensely studied much less successful in clarifying the mecha-
and debated form of corrosion of stainless steel. nism of pit initiation, which is still the subject
40 / Stainless Steels for Design Engineers

of debate, possibly indicating that the root Inclusions. The question of what causes the
causes are more metallurgical than electro- initial dissolution that causes both stable and
chemical. metastable pits focuses on inclusions, which
Figure 21 depicts a polarization curve for most authorities (Ref 18) have concluded are
stainless steel in a chloride-containing solution. associated in some way with pit initiation. In
Pitting occurs in the zone in which passivity is the absence of inclusions, metastable pitting
expected. As potential increases, small spikes in events are not noted, and the potential at which
corrosion current occur. These spikes measure pitting occurs is the beginning of the transpas-
local dissolution, called metastable pitting. sive regime.
Some such sites complete their dissolution and What are the typical inclusions in stainless
repassivate, while others continue to grow as steel? Inclusions in steel are normally the
stable pits. The potential at which stable pitting residue of normal deoxidation and desulfuriza-
occurs is the pitting potential, while metastable tion taken during steel refining usually done in
pitting can occur at much lower potentials. Pit- an argon oxygen decarburization (AOD). After
ting events, stable or not, cause the generation removal of the carbon, the subsequent objective
of iron ions and local pH reduction. To the ex- is to remove or render less harmful the dis-
tent these remain concentrated in a small vol- solved oxygen and sulfur, which if left in solu-
ume, they will affect subsequent events. The tion would later precipitate as low-melting-
dissolution during metastable pitting is located point iron compounds that would make the steel
at the matrix-inclusion interface. Different re- fragile and unworkable at high temperatures.
searchers assume dissolution of the inclusion, Inclusions in stainless steel are typically oxides
while others assume dissolution of the matrix. and sulfides. A key point to understand when
The dissolution parameters, as measured by considering inclusions as initiation sites for pit-
current transients, depend on variables not of ting is that inclusions are not simply inert debris
the inclusion chemistry but of the matrix com- but precipitates that are seeking thermodynamic
position, notably molybdenum and nitrogen equilibrium with the steel in which they have
levels (Ref 17), which is in keeping with the re- previously been dissolved. The reactions in
duction in dissolution of the matrix that these stainless steel differ thermodynamically from
alloying elements confer. those in carbon steel because of the presence of
high chromium concentrations. This lowers the
activity of oxygen and sulfur, making them
more soluble, as Table 1 indicates (Ref 19). It
also alters the efficiency of deoxidizing ele-
ments. Aluminum is a powerful deoxidant in
carbon steel but is less effective in stainless,
while titanium becomes a stronger deoxidizer in
stainless. Their effect on sulfur is similar to that
on oxygen.
The bottom line is that oxygen and sulfur are
generally removed by silicon/manganese deoxi-
dation, but that this process occurs in both the
liquid and solid states. That it carries over sig-
nificantly into the solid state means that diffu-
Fig. 21 Schematic of a passive anode polarization curve sion has a major role in determining if equilib-

Table 1 Typical values of activities and activity coefficients in liquid steels: activities in the 1 mass %
solution: ai = fi . %i
Metal Al C Mn P S Si Ti H N O Cr Ni
Carbon steel, %i ... 0.05 0.45 0.02 0.01 0.3 0.05 ... ... ... ... ...
1600 °C fi 1.05 1.06 1.0 1.1 1.0 1.15 0.93 1.0 0.97 0.85 ... ...
ai ... 0.053 0.45 0.022 0.01 0.345 0.046 ... ... ... ... ...
Stainless steel, %i ... 0.05 0.45 0.02 0.01 0.3 0.05 ... ... ... 18 8
1600 °C fi 3.6 0.49 1.0 0.32 0.66 1.24 9.4 0.93 0.17 0.21 0.97 1.0
ai ... 0.025 0.45 0.006 0.007 0.372 0.47 ... ... ... 17.5 8.0
Chapter 4: Corrosion Types / 41

rium reactions occur and whether they go to the liquid. At high sulfur and manganese concen-
completion. We will see that they do not. trations, some manganese sulfides can precipitate
Oxide inclusions also are common. They are during solidification interdendritically, while nor-
formed as the products of the reactions of sili- mal alloys with less than 100 ppm of sulfur form
con and manganese with dissolved oxygen. The their inclusions after solidification. The distinc-
thermodynamics of the reactions determine at tion is important because precipitation in the liq-
any time how much oxygen can be dissolved in uid state permits rapid diffusion, which results in
the steel at equilibrium. That equilibrium is eas- the most thermodynamically favorable species,
ily achieved in the molten state, in which diffu- manganese sulfide, to form. It may, and often
sion is very rapid, but achieved more slowly does, nucleate on a preexisting inclusion, such as
once the material has solidified. The inclusions silicate present from the deoxidation process. In
in the solid state grow by the diffusion of oxy- austenitic steels, manganese is generally present
gen to inclusion sites, where it precipitates as an at a level of around 1.5% as a deoxidant and as a
oxide of silicon or manganese to the extent that substitute for some nickel. An inclusion formed
these are locally present or of chromium when in the molten metal does not cause alloy deple-
its local concentration (or more properly, its ac- tion around it. One that forms or grows in the
tivity) makes it more favorable. These oxides solid state does cause depletion of the elements
are often the nucleation sites for manganese sul- that are precipitating, causing its growth.
fide inclusions. If manganese is lowered to very low levels,
Sulfur is a very surface active impurity that as- the supersaturation of sulfides is pushed to a
sists in weld penetration in stainless by virtue of lower temperature, at which lower diffusion
its effect on weld pool circulation. Otherwise, it rates hinder or prevent the precipitation. Thus,
is a detrimental impurity, forming low-melting low-manganese alloys can be free of manganese
oxysulfides that diminish hot workability. Man- sulfide inclusions even at somewhat high sulfur
ganese is a strong sulfide former, and it is the levels. Such alloys have elevated resistance to
main agent used to tie up sulfur. Manganese sul- pit initiation. Lower manganese levels also ther-
fide precipitates as an inclusion as a function of modynamically reduce the chromium sulfide co-
manganese and sulfur concentrations and tem- precipitation in inclusions, lowering chromium
perature. Inclusions form not only in the molten depletion around manganese sulfide/chromium
metal but also in the solidified metal. The solubil- sulfide inclusions.
ity, which is high in the liquid state, decreases on Elements more effective than silicon and
solidification, as seen in Fig. 22. Only resulfur- manganese are now in use for deoxidation and
ized free-machining stainless steels have suffi- desulfurization. These include aluminum, cal-
cient sulfur to precipitate manganese sulfide in cium, cerium, and other rare earth metals
(REMs), and titanium. The action of calcium is
notable. In a well-deoxidized and well-stirred
melt and with a basic slag, calcium dissolved in
the metal will react with dissolved sulfur to
form calcium sulfide, which will be incorpo-
rated into the slag phase. Aluminum, while a
potent deoxidizer, is less effective directly in
desulfurization, but it can act indirectly by re-
ducing a small amount of Ca2+ in the slag, al-
lowing the formation of calcium sulfide. Tita-
nium can sequester some sulfur as titanium
carbosulfide precipitates. The greatest amount
of sulfur removal is obtained by the addition of
cerium or other REMs, usually in the form of
the alloy mischmetal. These reactive elements
typically form oxysulfide particles in the melt
that may be trapped in the slag before metal
Oxygen is normally dissolved in solidifying
Fig. 22 Pseudo-binary-phase diagram for iron and sulfur at stainless steel, also at amounts in the neighbor-
1.8% manganese and 18% chromium hood of 100 ppm depending on deoxidation
42 / Stainless Steels for Design Engineers

methods. Inclusions based on oxygen and sulfur solidification as opposed to austenite first, AF),
formed in the liquid or during solidification are as is almost always the case with commercial
relatively large, greater than 1μ. As the alloy alloys, more sulfide precipitation happens in the
cools after solidification, precipitation continues solid state, pitting resistance is lowered propor-
since sulfur and oxygen are decreasingly solu- tionately to the sulfur level (Ref 22), and there
ble with temperature, to virtually nil at room is little negative effect from solidification segre-
temperature. This causes existing inclusions to gation, as is shown in Fig. 23 and 24 (Ref 23).
grow and new ones to nucleate. This precipita- Solidification can also occur in a mixed ferritic-
tion is similar to that which carbon undergoes in austenitic mode, in which case each microstruc-
stainless, except carbon is generally not super- tural component behaves according to the chart.
saturated until below 1200 °C at the highest in The ratio of chromium and chromium-like ele-
most alloys, whereas sulfur and oxygen are nor- ments molybdenum and silicon to nickel and
mally near saturation even at freezing or almost nickel-like elements carbon, nitrogen, man-
always when the solidifying ferrite transforms ganese determines the mode of solidification. It
to austenite. Thus, inclusions grow via diffusion
of oxygen and sulfur, which, as interstitials, dif-
fuse much more rapidly than the silicon or man-
ganese with which they have the greatest ther-
modynamic affinity. But precipitate they must,
even if the silicon and manganese in the vicinity
of their inclusion are exhausted. Thus, inclu-
sions can grow with chromium substituting for
either silicon or manganese as the precipitating
partner for oxygen and sulfur. The inclusion
growth necessarily depletes the surrounding re-
gion of reactants, silicon, manganese, and
chromium (Ref 20). Inclusions thus formed are
nonequilibrium in nature, and thermal cycles of
steel production are rarely sufficient for the
equilibrium to be attained. The chromium en-
richment of such inclusions and corresponding
chromium depletion of surrounding regions has
been measured (Ref 21) and corresponds to the Fig. 23 Influence of sulfur level on pitting resistance of
unannealed welds for different solidification modes.
depletion seen next to chromium carbide pre- Source: Ref 23
cipitates at grain boundaries in sensitized al-
loys. These zones are altered in size and shape
by thermomechanical processing in wrought al-
loys but exist fairly undistorted in welds. Hot
rolling and cold rolling followed by annealing
elongate manganese sulfide inclusions and flat-
ten them, allowing depleted zones around the
inclusion in the reduced dimension to be more
rapidly homogenized during annealing. Thus,
wrought material has better pitting resistance
than cast or welded material. Inclusions that
precipitate from the liquid, as is more the case
for alloys solidifying in an austenitic mode, are
at equilibrium with the surrounding matrix by
virtue of the faster diffusion in liquids, do little
to diminish the chromium content around them,
and have a small effect on lowering pitting re-
sistance. Pitting resistance is still affected to a
degree by alloy depletion due to solidification Fig. 24 Influence of sulfur level on pitting resistance of welds
without homogenizing anneal. FA, ferrite forming
segregation. However, if the alloy solidifies in a first on solidification as opposed to austenite first, AF. Source:
ferritic mode (FA, i.e., ferrite forming first on Ref 23
Chapter 4: Corrosion Types / 43

can also be altered by freezing rate. Faster cool- stress corrosion cracking (SCC). There have
ing favors austenitic solidification. been numerous proposed mechanisms for the
Long-term annealing of welds has shown breakdown of a passive film in chloride-con-
that sufficient time and temperature to achieve taining media; these have been summarized in
some rehomogenization the alloy result in bet- other publications (Ref 23). These hypotheses
ter pitting resistance (Ref 24), approaching that deal with how a passive film on a homogeneous
of the wrought alloy. Examination of the de- surface could break down. They include:
creasing solubility of sulfur in stainless in Fig. • Adsorption of chloride ions
22 indicates that the precipitation of sulfides • Penetration of the passive film by chloride
that cause chromium depletion occurs in delta
ferrite on freezing when sulfur exceeds 0.007% ions
• Film breakdown by electrostriction
and in austenite when sulfur exceeds 0.003%. • Formation of stable metallic chlorides
Oxygen behaves in a parallel manner and is • Coalescence of cationic vacancies
usually present in sufficient quantities, about • Random localized thinning of the passive
0.01% in manganese/silicon deoxidized steels,
to cause the same phenomenon. This funda- film
• Local variations in the composition of the
mentally is due to the high ratios of the diffu-
sivities of oxygen and sulfur to chromium, corrosive medium
which are about 10,000 and 680, respectively. By and large, these mechanisms presuppose
Whenever fast-diffusing elements such as oxy- a stainless steel surface that is homogeneously
gen, sulfur, carbon, and nitrogen, which have a passive and try to explain the observed inho-
strong affinity for chromium and a solubility mogeneous behavior of the passive film. How-
that decreases strongly with temperature, are ever, since it is clear that the surface is not ho-
present in steel, their precipitation will result in mogeneous, especially with regard to the
some degree of chromium depletion around the passive film, these hypotheses are not neces-
precipitation site because chromium diffuses sary to explain the behavior of everyday stain-
too slowly to be replenished. less steels, which unfortunately have abundant
The low chromium around inclusions is a suf- inclusions and chemical inhomogeneities capa-
ficient condition for the local dissolution meas- ble of locally diminishing the integrity of the
ured as metastable pitting, and if the depletion passive film. More research in understanding
zone shape and size are favorable, then stable the exact nature of the inhomogeneity of stain-
pitting would ensue. less steel surfaces is necessary for a complete
Certain other types of inclusions/precipitates understanding of pit nucleation and therefore
are less harmful in this regard. Titanium, for in- prevention.
stance, which is often added to form carbides Pitting Resistance. Pitting has been exten-
and nitrides, also forms sulfides and oxides sively correlated with environment and compo-
more strongly than manganese and therefore sitional variables. The most well-known and
does so at higher temperatures. Such precipi- useful correlations are between the PREN and
tates have a much lower tendency to allow the critical pitting temperature (CPT) and by ex-
chromium to join in the precipitation since the tension to the pitting potential.
higher the temperature of precipitation the more For austenitic alloys:
that diffusion allows the more favorable reac-
tion to occur. Rare earths also behave the same PREN = % Cr + 3.3 % Mo + 30 % N (Eq 9)
way. Metastable pitting is diminished by the
presence of these elements. For ferritic alloys, which hold no nitrogen in
The initiation of pitting is also affected by solution:
stress and inclusion orientation (Ref 25), which PREN = % Cr + 3.3 % Mo (Eq 10)
the researchers correlated to the dimensions of
the inclusion-derived cavity being able to sus- For duplex alloys, which have two phases,
tain a sufficiently low pH due to iron dissolu- neither of which matches the bulk composition:
tion to maintain stable pitting. The influence of
PREN = % Cr + 3.3 % Mo + 16 % N (Eq 11)
stress was to cause cracking at otherwise unfa-
vorably shaped inclusions, which then provided These equations are useful, if approximate,
a crevice capable of sustaining stable pitting. and their correlation is shown in Fig. 25 (Ref
This will be relevant to later discussions of 26) They do not include tungsten, which, if
44 / Stainless Steels for Design Engineers

alloy behavior predictions. It is noteworthy that

the elements copper and nickel, which are bene-
ficial against uniform corrosion and which slow
the growth of pits by this same action, do not
contribute to increasing the resistance to the
onset of pitting. This is another manifestation of
pitting initiated by the local stability of the pas-
sive film, which is primarily a function of local
chromium content. Molybdenum and nickel
thus seem to bolster local chromium content in
the passive film. Nitrogen seems to act by con-
centrating at the passive film-alloy interface
rather than by buffering the solution by ammo-
nia formation, which has been proposed (Ref
36). Research on very pure sputtered films of
iron-chromium alloys have demonstrated that
both titanium and niobium in solution diminish
active dissolution, assist repassivation, and im-
prove pitting resistance (Ref 29). In most practi-
cal cases, these elements are not found in solu-
tion because of their affinity for oxygen, sulfur,
Fig. 25 Variation of critical pitting temperature with pit- carbon, and nitrogen, with which they form
ting resistance equivalent number (PREN) of
austenitic steels in water plus 6% FeCle. Source: Ref 26 compounds.
It should also be noted that the critical PREN
values vary with crystallographic structure. Fer-
ritic alloys require somewhat lower PREN val-
ues to exhibit similar pitting resistance as
austenitic alloys of somewhat higher PREN.
While pitting is of great theoretical and prac-
tical interest, there are significant problems in
actually conducting good tests. Monitoring of
the electrochemical potential during the test is
considered mandatory by most researchers.
How is a metallic sample suspended in a solu-
tion without creating any crevices and without
exposure at the liquid-gas interface? The devel-
opment of the flooded gasket technique (used in
ASTM G150) was a milestone, but it also has
Fig. 26 Differential variation of critical pitting temperature some problems—most notably the potential for
of several stainless steel alloys for unwelded dilution of the test solution, especially during
wrought and welded material. Source: Ref 13
prolonged testing. FeCl3 testing benefits from
the fact that the solution creates a reproducible
present, has half the effectiveness of molybde- positive potential.
num. They neglect carbon, which seldom varies While the PREN approximates the pitting re-
enough to have a visible effect but has been sistance of an alloy, there is a standard test by
shown when in colossal supersaturation to have which the CPT is measured. Pitting in a given
a factor of about 10, not unlike nitrogen, an- medium capable of causing pitting does not
other interstitial that it resembles in solution occur below a temperature that is characteristic
thermodynamically (Ref 27). It also does not in- of the medium and the material, with the myriad
clude the negative influence of elements such as exceptions of stress state, surface finish, mi-
sulfur. Likewise, the equations cannot deal with crostructure, etc. The most commonly used test
inhomogeneity issues, so welded alloys have media are the unacidified 10% FeCl3, which is
different CPTs for the same PREN (Fig. 26) used in the ASTM G 48 practice B, and the 3.5%
(Ref 13). These equations are all-other-things- NaCl solution of the ASTM G 150. The latter, if
being-equal equations and are useful for gross modified to 0.1N NaCl, allows the ECPT, the
Chapter 4: Corrosion Types / 45

electrochemical pitting potential, of lower alloys Cl = %Cr + 4.1%Mo + 27%N (Eq 12)
such as 304 to be measured (Ref 30).
Since a crevice has a preexisting favorable
geometry for pit growth, any pitting event,
Crevice Corrosion metastable or stable, can initiate ongoing
In the case of pitting, the geometry that crevice corrosion. Crevices are thus incubators
makes up the pit is essential in creating the dif- for corrosion triggered by metastable pitting
ferential aeration cell and to cause the autocat- events. The dissolution of iron during passiva-
alytic dissolution process. In many cases, a tion itself as well as the differential oxygen cell
geometry that retains and acidifies water is al- created by the crevice contribute to the process.
ready present in crevices in different types of It is logical to think that alloying the elements
structures such as gaskets, under faulted coat- that contribute to lowering the critical current
ings, under bolt or screw heads, etc. Crevice density for passivation and the uniform corro-
corrosion occurs because zones have restricted sion rate, such as nickel, would reduce the cre-
access of reactants and restricted exit of corro- ation of the reactants that start the crevice corro-
sion products. It is especially the inhibition of sion process, but this presumed effect is not
the cathodic reaction inside the crevice by the strong enough to be reflected in this actual be-
dearth of oxygen, which sets up a more aggres- havior Eq 12 represents, although it is generally
sive environment within the crevice than with- acknowledged that austenitic steels perform bet-
out. The interior reactions become increasingly ter than ferritic steels in the absence of molyb-
anodic, and the aggressiveness of the environ- denum. Materials are characterized as having a
ment can reach a threshold at which active cor- critical depassivation pH. If crevice conditions
rosion occurs, while the situation exterior to the are such that the reactions over time allow the
crevice is safely passive. Crevice corrosion oc- pH to be reduced to this level, then active corro-
curs at lower temperature than pitting in the sion will begin within the crevice. Thus, passive
same environments, so it is a greater danger in film stability seems to be the critical factor
that sense. rather than corrosion rate after initiation.
The relationship between the alloy content, Preventing Crevice Corrosion. The coun-
given as the crevice corrosion resistance equiv- termeasures against crevice corrosion are ca-
alent number (CCREN), and critical crevice thodic protection, design, maintenance, and, of
corrosion temperature (CCT), shown in Fig. 27 course, alloy selection. Designing to avoid
(Ref 11), is similar to that of PREN (PI) to CPT crevices should include maximizing the volume
except for the molybdenum factor being more of unavoidable crevices, engineering flow to en-
important: hance transport in and out of crevices, and

Fig. 27 Variation of critical crevice corrosion temperature with alloy content

46 / Stainless Steels for Design Engineers

avoiding stagnation. Any maintenance or design chromium. In the various grades of stainless
procedure that prevents formation of deposits is steels, there are many intermetallic phases that are
beneficial. Welds are particularly vulnerable thermodynamically stable but kinetically slow to
surface sites, so any combination of welds and precipitate that are enriched in chromium. An ex-
crevices or crevices caused by poor weld geom- ample of such a phase is chromium carbide (Fe,
etry must be avoided. S32205 is a benchmark Cr)23C6. These phases tend to form at grain
alloy of sorts. It has just sufficient alloying to boundaries where nucleation is favored, resulting
resist pitting in seawater, but it is susceptible to in a depletion of chromium in the adjacent re-
crevice corrosion. gions, as shown in Fig. 28. Thus, the chromium-
As a practical matter, crevices are almost im- depleted regions near the grain boundaries are
possible to eliminate. Threaded fasteners and sensitized in that they behave as active anodes
joints represent severe crevices and should be compared to the larger interior of the grains that
avoided in aggressive environments if possible. are still passive. In an aerated corrosive environ-
Gasketed joints are another severe crevice loca- ment, the smaller chromium-depleted nonpassive
tion, and their usage should be curtailed to the anodes dissolve, whereas the larger cathodes re-
minimum practical extent. In these situations, duce oxygen, resulting in a localized corrosion
judicious use of very expensive, highly corrosion along grain boundaries. Any heat-treating or
resistant materials is justified. The use of smooth welding procedure of stainless steels should thus
welded joints is thus generally preferred. In a be tailored to avoid sensitization.
more general consideration, deposition and When a stainless steel is heat treated, there is
fouling create crevice sites, and design and op- a risk that the unwanted phases may form, de-
erational controls to preclude the formation of pending on the time-temperature history and
deposits and the prompt removal of sludge and precipitation kinetics of the unwanted phase.
the like are necessary. But in some situations, Figure 29 shows schematically the temperature
such as marine exposures, biofouling will create versus time due to welding and the resulting
crevice sites. This fouling may be macroscopic, sensitization. Figure 29 shows a TTT (time-
such as from shellfish and barnacles, or it may temperature-transformation) curve for precipi-
be microscopic. Microscopic biofouling causes tation of the unwanted phase. Near the weld
the special form of crevice corrosion called mi- (A), the time spent in the temperature region
crobiologically influenced corrosion (MIC) dis- where precipitation occurs is too short, whereas
cussed in a separate section). far away from the weld (C) the temperature ex-
perienced is too low. At location B, there is,
Sensitization/Grain Boundary Corrosion however, a risk for sensitization.
The maintenance of a passive layer in a wide Austenitic. Sensitization can occur at any
range of pH conditions in stainless steels is de- temperature at which carbon is supersaturated
pendent on the alloying elements, primarily in an alloy. Current austenitic stainless steels
have carbon levels of under 0.10% normally

Fig. 28 Schematic illustration of sensitization due to Fig. 29 Schematic illustration of how a heat treatment re-
chromium-rich precipitates that deplete adjacent lates to sensitization due to precipitation kinetics.
regions of chromium. GB, grain boundary TTT, time-temperature-transformation
Chapter 4: Corrosion Types / 47

and under 0.03% for low-carbon L grades. activity of carbon and make alloys more suscep-
Thus, normal grades sensitize below around tible. Nitrogen lowers the tendency for carbide
800 °C. The supersaturation increases with formation and slows sensitization. Nonthermo-
decreasing temperature, but below about 500 °C dynamic effects are those of austenite grain size
diffusion of carbon is too slow for carbon to and prior cold work. Decreasing grain size and
move to grain boundaries and cause the damag- therefore increasing grain boundary surface
ing combination with chromium that causes area decreases the amount of precipitate per unit
sensitization. Low-carbon grades avoid sensiti- area of grain boundary and therefore the amount
zation because they are not sufficiently super- of chromium depletion per unit area. Cold work
saturated at temperatures at which carbon is accelerates diffusion and makes precipitation
mobile enough to diffuse to grain boundaries. more rapid, thus aggravating sensitization.
Ferritic. Another situation exists in ferritic The thermodynamic affinity tool can be used
stainless steels, in which carbon is much less to prevent chromium carbide formation in an-
soluble but is much more mobile. Annealing other way. Introducing alloying elements that
over 900 °C can put enough carbon in solution to combine with carbon more strongly and rapidly
cause sensitization even at the lowest carbon lev- than chromium can exhaust the supply of car-
els attainable in an AOD and even at the fastest bon available to precipitate as chromium car-
possible quench rates. The damaging chromium bide. There are a number of candidate elements,
depletion caused by this very rapid precipitation zirconium, vanadium, tantalum, niobium, and
can be undone by a simple rehomogenization titanium, most prominently. Of these, the diffu-
anneal of the remaining chromium. This is theo- sivity and affinity for carbon of niobium and ti-
retically possible with austenitic alloys also, but tanium make them the best for this purpose.
the diffusion rates of chromium in austenite as Each forms stable carbides at much higher tem-
so slow that it is impractical in most real cases. peratures than chromium, starving chromium of
Duplex steels have a subtle near immunity to sufficient carbon to form damaging precipitates.
carbide sensitization. While they are typically The caveat with titanium is that it forms oxides,
low carbon anyway, the carbides that do form sulfides, and nitrides preferentially to carbides.
do so at ferrite-austenite grain boundaries. Here, Therefore, sufficient quantities must be used to
chromium is consumed from the chromium-rich accommodate the prior formation of these
ferrite phase, leaving the austenite intact. Their phases. Niobium tends more toward carbide
large grain boundary area keeps carbide con- than nitride formation but is a weaker carbide
centration per unit area low, and the fast diffu- former than titanium. The solubility products of
sion in the ferrite keeps austenite from becom- these precipitation reactions are:
ing depleted. However, the rapid formation of
intermetallic phases at the ferrite-austenite in- 6780
log [Ti ][C ] = 2.97 − (Eq 13)
terfaces can lead to a rapid loss of corrosion re- T
sistance and a severe loss of toughness if expo-
sure to temperatures within the intermetallic
precipitation range is not controlled. 9350
Martensitic steels are quenched as austenite log[ Nb][C ] = 4.55 − (Eq 14)
to and through the Ms temperature without time
for carbon to precipitate in austenite. The car- These equations follow the form of the gen-
bon in the martensite can precipitate and cause eral equation for precipitation reactions:
sensitization if reheated to the 300 to 700 °C re-
gion. Fortunately, heating to above 700 °C re- log[ M ][ X ] = A − H / RT (Eq 15)
homogenizes the chromium and eliminates sen-
sitization. in which A is a constant, H is the heat of disso-
Effect of Alloying. Besides determining lution, R is the gas constant, and T is the ab-
basic phase structure, alloying plays a role in solute temperature. If the amount of titanium or
susceptibility to sensitization. Those elements niobium is stoichiometrically sufficient, no car-
that reduce the tendency of chromium carbides bon will form chromium carbides under equilib-
to form also reduce the susceptibility to sensiti- rium conditions. It is possible to defeat the sta-
zation. This is a purely thermodynamic effect. bilization reactions by quenching the alloys
Molybdenum, silicon, and nickel promote car- from temperatures at which titanium carbide or
bide formation by increasing the thermodynamic niobium carbide is dissociated. If free carbon is
48 / Stainless Steels for Design Engineers

left free in the matrix by quenching, then on re- layer occur, and region II, where the protective
heating it may form carbides with the most lo- layer is not fully developed, suggesting an ap-
cally accessible favorable element, such as preciable electrochemical effect. The latter is a
chromium, rather than the most thermodynami- zone that exists in alloys that have zones of
cally favorable element, which would be tita- chromium depletion.
nium or niobium. This can occur when a stabi- Stress corrosion cracking has always been
lized alloy such as 321 is welded. A zone away among the most controversial subjects among
from the weld may experience a high enough metallurgists and electrochemists. The debate
temperature to put carbon into solution and then centers on whether the critical mechanism is
cool just rapidly enough to not form only the dissolution or fracture, and if a fracture, by what
equilibrium titanium carbide but also Cr23C6 at mechanism. Is the cracking zone locally soft-
grain boundaries, causing the type of sensitiza- ened, locally hardened, transformed, to a more
tion called knife-line attack. This problem has brittle phase or embrittled by hydrogen? As of
nearly ceased to exist as modern 321 has low this writing, there is no general agreement on
levels of carbon and nitrogen for economic rea- which type mechanism is the fundamental
sons; this effectively precludes this chromium cause, but there is room for convergence. Obvi-
carbide precipitation in most cases. ously, elements of many may come into play. It
Welding. Many of the most severe problems is likely, as in most prolonged arguments, that
of sensitization arise when stainless steels are no hypothesis is completely correct. We will try
welded to carbon or low-alloy steels. In these to fairly set out what is known and agreed on as
situations, construction code rules usually re- fact and then present researchers’ views in an
quire that the carbon steel component be given a unbiased manner, but since we concern our-
stress relief annealing (SRA) treatment. Such selves only with stainless steel, no attempt is
SRA treatments are typically in the sensitization made to address an all-encompassing theory.
temperature range for austenitic stainless steels. Crack Initiation. In stainless steels, cracks
Use of low-carbon or stabilized grades is neces- can be seen to initiate at surface defects and ir-
sary in such cases. Even then, use of the lowest regularities. In stainless steel, it must be agreed
allowable temperature SRA treatment for the by all that the preponderant initiation site is a
shortest allowable time is preferred. corrosion pit or, in some cases, a crevice. Inter-
granular corrosion sites, as are seen in sensi-
tized material, can also provide the conditions
Corrosion Combined with for SCC initiation. The interrelationship be-
tween pits and SCC cracks has been studied
Fatigue or Fracture
(Ref 25). Stress lowers the anodic potential at
which pitting occurs and permits metastable pits
Environmentally induced failure occurs when
brittle failure under tensile mechanical loading
occurs at a lower stress when a material is sub-
jected to a corrosive environment than what
would happen in a noncorrosive environment.
This introduces us to what is perhaps the most
controversial technical subject in all of stainless
steel research, SCC.

Stress Corrosion Cracking

The key cause for SCC is the cooperating ef-
fects of tensile stress and a corrosive environ-
ment. Such cases can be identified in most alloy
systems, and even pure metals, which were
thought to be more or less immune, also have
had cases of SCC reported.
In passive metals, two sensitive potential re-
gions for the occurrence of SCC have been
identified and are shown in Fig. 30: region I,
where pitting and breakdown of the passive Fig. 30 Zones of susceptibility to stress corrosion cracking
Chapter 4: Corrosion Types / 49

to become stable via the generation of cracks. environmental variables. The most common ex-
Cracks, once formed, presumably have favor- ample is that of sensitized 304 in high-tempera-
able geometry to duplicate pit internal chemical ture water or caustic media. The relevance of
reactions and must be considered to be de- this to the normal case of stainless steels must
scribed by the models that apply to pits and be questioned since, by definition, the sensitized
crevices. grain boundaries themselves can be depleted of
The stress at which SCC initiates has a chromium to a degree they are not stainless and
threshold, which has been reported as between have a much less stable austenitic structure,
25 and 50% of the yield strength in austenitic having their martensite start temperature Ms,
stainless steel. raised by the loss of chromium.
The temperature at which SCC is initiated Material Variables. Martensitic stainless
ranges from ambient to under 100 °C for steels and martensitic precipitation hardened
martensitic materials, while austenitic alloys stainless steels are quite susceptible to SCC.
begin their sensitivity above room temperature This susceptibility increases with hardness,
and increase in susceptibility with increasing yield strength, and embrittling heat treatments.
temperature. The ferritic steels, while consid- They will crack at threshold stresses equal to
ered nearly immune to SCC, have their maxi- 50% of yield strength. These alloys can be tem-
mum susceptibility in the same range as marten- pered at sufficiently high temperatures that they
sitic steels. In environments of mixed chlorides become soft and tough enough to have very
and sulfides, however, SCC can occur in all good resistance.
types of stainless at room temperature. This has Ferritic stainless steels of low and medium
been seen in the SCC of austenitic stainless chromium are generally not susceptible to SCC.
steel in swimming pool environments, in which Ferritic alloys, which can have a martensitic
chloride ions can condense on the stressed steel structure, should be considered martensitic for
and cause pitting and SCC. SCC purposes. If purely ferritic alloys are al-
Cracks propagate very slowly below spe- loyed with copper, molybdenum, and nickel,
cific certain stress intensity levels, but once that they can become susceptible. The presence of
intensity is reached, they have a plateau rate α'( or high-temperature embrittlement also in-
that is fairly constant until the stress level at creases susceptibility, as does cold work.
which catastrophic failure occurs at very high Despite the controversy surrounding the
propagation rates. Rates of crack propagation mechanism of SCC in austenitic stainless steels,
are exponentially increased by increasing tem- there is almost complete agreement that SCC of
perature. The crack propagation rate has been body-centered cubic (bcc) stainless steels,
seen across a range of alloys to be linearly pro- martensitic, ferritic, and pH is simply a mani-
portional to the average current density that festation of HE, with hydrogen provided by ei-
alloy experiences when its surface is strained, ther anodic (e.g., active corrosion within a pit)
indicating that reactions at the crack tip are or cathodic reactions.
strain sensitive, and overall rate limiting, but Duplex stainless steels have low susceptibil-
not necessarily the mechanism of cracking. ity to SCC. Their dual-phase microstructure en-
Crack growth is discontinuous with individual sures that under conditions that crack austenite,
steps of growth many times the average rate, ferrite remains as a crack-arresting phase, while
which is similar to that seen with gaseous hy- under conditions that cause SCC in highly al-
drogen embrittlement (HE). The crack growth loyed ferrite, the austenite is a crack arrester.
gives off acoustic emissions as cracking steps A second explanation of the resistance of du-
occur. These steps of growth are brittle and are plex alloys to SCC is that their two phases have
seen as facets on fractographs with cleavages different corrosion potentials, and that the
corresponding to crystallographic planes. The mixed potential that arises because they are in
crack facets match with high perfection, show- intimate contact is outside the potential range
ing almost no evidence of plastic deformation for SCC on either phase. This fits with the re-
or dissolution. sistance to SCC of wrought alloys with a lamel-
The propagation path may be intergranular or lar structure and the lesser resistance of cast al-
transgranular. Grain boundary propagation in loys that lack that structure.
stainless steels usually corresponds to condi- Austenitic stainless steels are the type of stain-
tions under which grain boundaries are less cor- less steel generally associated with SCC, and
rosion resistant because of either material or they vary in their degree of susceptibility to
50 / Stainless Steels for Design Engineers

SCC. All other things being equal, alloying which are probably the greater cause of resist-
elements that delay or prevent localized corro- ance to SCC.
sion do the same to delay SCC. This is simply Environmental Variables. There are three
the delay of initiation. However, if pitting can be key types of environments in which SCC occurs
delayed indefinitely, then SCC can also, assum- in stainless:
ing, of course, more harmful localized corrosion, • Chloride-containing solutions
such as that due to intergranular chromium de- • Caustic solutions
pletion, is not occurring. Molybdenum, which • Polythionate and thiosulfate solutions
we already know helps prevent pitting and
crevice corrosion, also increases the threshold The cases of polythionate and thiosulfate solu-
stress for SCC, as shown in Fig. 31 (Ref 31). tions are industrially important but can be ade-
But, if metastable or stable pitting is occurring, quately explained as simply the stress-assisted
the threshold stress has been reached, and the intergranular corrosion of sensitized material.
temperature is sufficient, then SCC will proceed. Hot caustic solutions are aggressive against
It is mitigated by material variables such as cold stainless steels. Certain combinations of concen-
work and by alloying elements that increase trations, temperature, impurity, and dissolved
austenite stability. Many publications cite nickel oxygen can cause SCC as well as other undesir-
as beneficial in enhancing resistance to SCC, able corrosive attack. Resistance to general cor-
often referring to the data from Fig. 32. However, rosion is proportional to nickel content, but fer-
its role seems mainly to be as an austenite stabi- ritics and duplex alloys are less prone to SCC.
lizer and as a retarder of active corrosion. The 304 has been reported to have no meaningful
minimum in the curve corresponds to the nickel threshold stress for SCC in hot caustic solutions,
level at which the structure is entirely austenitic, leading one to question whether such a failure
but least stably so. Lower nickel levels produce should even be classified with SCC of the typi-
better immunity through the duplex structure, cal chloride-induced type or belong with the pre-
while higher levels promote austenite stability vious polythionate and thiosulfate solutions.
and correspond to alloys having more alloying Chloride containing environments are the
elements, such as chromium and molybdenum, main ones that induce SCC. Water can cause

Fig. 31 Influence of molybdenum on resistance to stress Fig. 32 Variation of resistance to stress corrosion cracking
corrosion cracking (SCC) in austenitic steels with nickel (and other) content and structure
Chapter 4: Corrosion Types / 51

SCC at sufficiently high temperatures (i.e., does advance discontinuously, and after each
above 100 °C) if there are even very low com- advance there is fresh surface, which comes
bined concentrations of chloride (greater than into equilibrium with the solution within the
0.1 ppm) and oxygen (greater than 0.1 ppm) crack. So, any experiment, such as that shown
dissolved (see Fig. 33) (Ref 32). in Fig. 34 (Ref 33), that tests crack propagation
Failure times decrease exponentially with de- against electrochemical events will absolutely
creasing chloride content. Crack growth rate in- support this model.
creases by a factor of ten with each 30 °C rise in It is axiomatic that films must rupture and re-
temperature. Decreasing pH lowers the temper- form as cracks advance discontinuously. The
ature at which SCC occurs in a given time. weakness of this model is that it does not pro-
Mechanisms. There have been many mecha- vide a mechanism for brittle fracture, and the
nisms proposed for SCC in stainless steel. We very brittle features of transgranular SCC frac-
focus only on those that address the failure in ture surfaces do not show any supporting evi-
chloride-containing media, the main concern for dence of dissolution. Research (Ref 34) show-
users of stainless steel. ing that metal dissolution at the crack tip is
The models that have found some support isotropic rather than crystallographically ori-
are: ented make dissolution models incapable of
• Slip dissolution being reconciled with the crystallographic frac-
• Adsorption-enhanced plasticity ture surface facets.
• Adsorption-induced brittleness Adsorption-induced brittleness, also known
• Hydrogen embrittlement as stress-sorption, looks to the parallels be-
tween liquid metal embrittlement and SCC to
Slip dissolution (anodic dissolution) was the explain the mechanism of SCC as the action of
earliest proposed model for SCC. It simply pro- adsorbed species weakening atomic bonds on
poses that at a crack tip a passive film forms, and the crack tip surface. If the action is on the sur-
after time it fractures by an unspecified mecha- face, however, the mechanism cannot produce
nism. The fresh active surface may or may not the observed discontinuous, brittle cracks which
repassivate, after which the process repeats itself. characterize SCC. Only in alloys such as Fe-3Si
The strength of this model is that it actually are steps small enough to make this mechanism
does describe what is happening. The crack plausible.

Fig. 33 Variation of susceptibility to stress corrosion cracking (SCC) with media oxygen and chloride content for 304 stainless steel.
Source: Ref 32
52 / Stainless Steels for Design Engineers

Fig. 34 Crack propagation rates of various metals plotted versus current density. Source: Ref 33

Adsorption-enhanced plasticity/hydrogen em- actually take place, but none is specific enough
brittlement encompasses a number of models to have been tested by critical experiments to
that observe that adsorbed species enter the lat- prove or disprove it.
tice in the vicinity of the crack tip and then It has been demonstrated that hydrogen is ab-
cause failure by one of several mechanisms: sorbed into the material at the crack tip. The
• Dealloying and porosity main question is whether it causes damage by
• Adsorption-induced brittleness creating porosity, altering dislocation mobility,
• Coalescence of voids formed by cross slip or causing lattice decohesion. There is support
for each.
enhanced by the adsorbed species It has been observed that where SCC occurs
Since hydrogen is the only species that is pro- there is a large concentration of vacancies. This
duced in quantity and is capable of diffusing into has led to speculation that porosity is a weaken-
the lattice, HE is implicit in all these models. ing mechanism responsible for SCC (Ref 35). It
All of the models have support in that they has been proposed and supported by calcula-
have some experimental observations that show tions that hydrogen lowers the energy required
that the phenomena they propose as causal or vacancy formation. The lowest energy
Chapter 4: Corrosion Types / 53

configuration is calculated as two hydrogen

atoms per vacancy. This pairing of hydrogen
solute atoms to dislocations is very reasonable
given the major distortion the interstitial hydro-
gen causes to the lattice, so there is no basis to
challenge the enhanced vacancy formation.
Whether the effect is large enough to cause fail-
ures has not been demonstrated. The largest
measurable effect of hydrogen has been a slight
acceleration of stress relaxation in martensite.
The relevance of hydrogen-induced vacancy ag-
glomeration as the principal cause of failure must
be considered questionable until some further
critical experiments link the vacancies to the ob-
served instances of failure quantitatively, and Fig. 35 Stress-strain curve for single crystals of stable
austenitic stainless steel with and without hydrogen.
more important, to show how this mechanism Source: Ref 36
could account for the temperature and stress de-
pendence observed. The major influence of va- in Fig. 35 (Ref 36). All these interstitials strain
cancy formation due to hydrogen may be to en- the lattice and therefore harden in proportion to
hance the volume expansion due to hydrogen. their atomic size. Hydrogen, as the smallest of
Adsorption-induced brittleness, also known them, has about half the distorting effect and
as stress sorption, looks to the parallels be- half the hardening effect. But, its small size
tween liquid metal embrittlement and SCC to makes it mobile at ambient temperatures, so it
explain the mechanism of SCC as the action of can diffuse to sites where it can alter mechani-
adsorbed species weakening atomic bonds on cal properties.
the crack tip surface. If the action is on the sur- But, while hydrogen causes dislocation mo-
face, however, the mechanism cannot produce tion and lower work hardening, it does not
the observed discontinuous, brittle cracks that weaken austenite, so this theory by itself cannot
characterize SCC. Only in alloys such as Fe- account for the role of hydrogen in SCC and, by
3Si are steps small enough to make this mech- inference, in HE in the more general case.
anism plausible. In stainless steels, there The quandary of hydrogen finally having
seems to be nothing to support this proposed been shown to have a clear effect on mechanical
mechanism. properties but having that not account for either
Adsorption-enhanced plasticity has become SCC or HE may be put to rest by the additional
known recently as HELP or hydrogen-enhanced observations of hydrogen’s role as a lattice dis-
localized plasticity. The underlying mechanism torter (Ref 37). While not formalized as a pro-
at work in this model is the hydrogen-induced posed hypothesis for SCC, the role of hydrogen
shielding between microstructural defects. This as a generator of very high stresses has been
has been observed distinctly in single crystals of pointed out as a factor that cannot be neglected
austenitic stainless alloys. The Cottrell atmos- when evaluating other proposed mechanisms.
phere of hydrogen around dislocations causes Hydrogen has been shown to distort the lattice
mutual repulsion, causing strain to be localized in proportion to its concentration. The effect is
on certain slip systems. This has been observed not small, accounting for about 1% strain per
to occur and has caused deformation to become 0.1% concentration by weight, as shown in
concentrated in Luders bands in austenitic al- Fig. 36 (Ref 38). At hydrogen levels of over
loys, which of course do not show such behav- 1000 ppm, which are thought to exist around
ior without hydrogen (Ref 36). This also pro- growing SCC crack tips, there could therefore
duces ε-martensite in austenitic alloys, which be hydrogen concentration gradients capable of
would be considered stable without hydrogen producing additional tri- or biaxial stresses on
and deformation. the matrix ahead of the crack tip that may ap-
This theory encounters a problem, however, proach the yield stress and account for some or
with the fact that the same studies showed that all of the difference between the normal fracture
hydrogen actually strengthens the matrix by toughness KI and the KISCC, that for SCC. This
solid solution hardening. It acts in much the also precludes the necessity of hypothesizing
same way as carbon and nitrogen do, as shown hydrogen-induced phase changes, although
54 / Stainless Steels for Design Engineers

Fig. 36 Dilation of austenite due to hydrogen in solution. Source: Ref 38

were they to exist, they would result in the same volved distinguishing among the same mecha-
lattice expansion. In both cases, the failure nisms, namely:
would occur at a region ahead of the crack tip • Decohesion
and beyond the highest hydrogen concentration, • Enhanced local plasticity
which is what is observed to occur. The growth • Adsorption embrittlement
of this stress over time with increasing hydro- • Void coalescence
gen-producing corrosion would account for the
observed kinetics, locus, and stress dependence The identification of the operative mechanism
of SCC. for HE involves again distinguishing what role
If nothing else, the main models for SCC and each of the above contributes to HE in a given
the experimental results on which they are situation since all are known to be real metallur-
based should be reexamined in view of the fact gical phenomena.
that the stresses induced by hydrogen are not The main difference between HE and SCC in
negligible and, in fact, may account for much of stainless steel is that HE is limited to ferrite,
the observed SCC behavior of stainless steels. which is hardened by cold work or alloying, and
The next few years may finally see the resolu- martensite. Austenite is somewhat diminished
tion of the lengthy debate over the causes of in ductility by hydrogen, but not subject to the
SCC. If it comes, it will be from critical experi- completely brittle, discontinuous cracking of
ments, which can quantitatively differentiate bcc stainless. The observations that make a
among the above effects and measure the con- given model plausible as a mechanism for SCC
tribution of each. lack traction for the same materials in HE. It
hard to envision enhanced plasticity involved in
the completely brittle fracture of high-strength
Hydrogen Embrittlement
martensitic stainless steels, whereas void coa-
Like SCC, there has been debate about HE lescence by vacancy creation seems more likely
that has produced more heat than light. This in- to account for the observed behavior.
Chapter 4: Corrosion Types / 55

The resolution of mechanism here also must Biocorrosion and Microbiologically

account for the contribution of hydrogen-in- Induced Corrosion
duced stress as well as hydrogen effects on me-
chanical processes, especially since the ob- There are many cases for which biological or-
served susceptibility to HE is proportional ganisms contribute to initiating or enhancing
hardness, therefore to the amount of hydrogen a rates of corrosion. This can occur in natural en-
given material can hold both in normal intersti- vironments such as ground or seawater as well
tial solution and the amount it can trap at lattice as domestic and industrial environments such as
defects (Ref 39), especially the dislocations the nuclear and chemical processing industries,
within the plastic zone at the crack tip, which for example. This is called biocorrosion or
provide enhanced hydrogen solubility where it MIC, microbiologically induced corrosion.
can aggravate the applied crack opening with a The bacteria that are known to influence cor-
wedge effect from hydrogen dilation. rosion can be sorted as aerobic bacteria that lie
in aerated water and anaerobic bacteria. Among
Corrosion Fatigue the anaerobic bacteria that are known to [16]
Just like SCC, corrosion fatigue causes brittle (Ref 40) affect stainless steels can be counted:
failure under a combined environment of corro- Desulfibrio and Desulfotomaculum. Both of
sion and a tensile stress component. The stress, these are so-called sulfate-reducing bacteria
however, is cyclic and in a test of stress versus (SRB), which means that they promote the reac-
number of cycles (S vs. N), failure will occur at a tion:
lower N under the corrosive environment. The SO 24− → S2− + 4 O (Eq 16)
cracks are transgranular, and the collaborative ef-
fect of corrosion and fatigue is that corrosion ac-
celerates the plastic deformation that accompa- which in turn accelerates the cathode reaction:
nies the evolution of extrusions and intrusions.
In corrosion fatigue, an obvious pit corrosion 2H = + 2e − → 2H (Eq 17)
site may not be necessary because of the com-
bined action of cyclic stresses and the environ- Aerobic bacteria flourish under oxygen (Ref
ment. However, an initiation site that is the 40). Examples are the iron-oxidizing Gal-
weakest link in a combined mechanical and lionella and Sphaerotilus, which increase the
metallurgical sense will be the initiation point anode dissolution reaction:
after which conditions that may not cause SCC
can help propagate fatigue cracking at lower Fe → Fe 2+ = 2e − (Eq 18)
stresses than would be expected in more benign
environments and in environments that may not by converting the ferrous iron-ion product
cause SCC or pitting under static loads. (Fe2+) to less soluble ferric (Fe3+). Due to this,
The importance of the environmental interac- macroscopic so-called tubercules form that can
tion is reflected in the sensitivity to frequency of cause crevice-type shelters where differential
stress application. High-frequency loading aeration and pit initiation can occur.
gives less time for corrosive attack and brings Countering MIC with biocides can cause
crack propagation rates down closer to those in problems in manganese-containing waters. Oxi-
air. In some materials, crack propagation rates dizing biocides, such as ozone, chlorine, or per-
are elevated above those in air at all stress lev- oxide, can cause manganese to be oxidized to
els, while in others a threshold stress intensity manganese dioxide. The precipitated deposits of
must be reached before an acceleration is noted. manganese dioxide can accelerate pitting corro-
Some materials show a combination of both. sion even in low-chloride waters in which al-
The first case seems to be merely fatigue as- loys such as 316 would otherwise be safe from
sisted by corrosion, while the last two seem to pitting attack.
indicate an SCC–type behavior. Biocorrosion is most commonly encountered
The same uncertainties that cloud our under- in ambient aqueous environments, which are the
standing of SCC necessarily disguise the pre- environments in which most microorganisms
cise mechanism of corrosion fatigue, which have evolved to thrive. So, it tends to be a prob-
must be viewed as a combination of SCC and lem for the medium-alloyed steels such as 304
fatigue. and 316, which are used in these environments.
56 / Stainless Steels for Design Engineers

The more chemically or thermally hostile envi- 15. R.C. Newman, Corrosion, Dec 2001,
ronments in which higher alloyed grades are p 1030–1041
used are also hostile to bioorganisms and thus 16. N.J. Laycock and R.C. Newman, Corros.
minimize the problem. Sci., Vol 39, 1997, p 1771
The development of microbiological consortia 17. Y. Kobyashi, S.Virtanen, and H. Bohni,
allow anaerobes to flourish under biofilms that Proc. Electrochem. Soc., 1999, p 533–540
form in an aerated environment. These represent 18. Z. Szlarska-Smialowska, Pitting Corrosion
a differential aeration cell that acts just like a se- of Metals, NACE, Houston, TX, 1986
vere crevice. Also, the action of microbes in 19. E.T. Turkdogan, Fundamentals of Steelmak-
raising the corrosion potential is key to under- ing, Institute of Materials, 1996
standing why natural seawater is so much more 20. H.S. Kim and H. Lee, Met. Trans. A,
corrosive than sterile sodium chloride or syn- Vol 32A, June 2001, p 1519
thetic seawater solutions. And, macrofouling or- 21. M.P. Ryan, D.E. Williams, et al., Nature,
ganisms are important. They create crevices and Vol 415, Feb 2002, p 770–777
sites where microfouling can start early. At the 22. A.J. Grekula et al., Corrosion, 40, 1984,
same time, they are sources of turbulence in p 569
flowing systems, and this turbulence can cause 23. Stainless Steels Les Editions de physique, 1993
flow erosion in copper materials, making use of 24. N. Suutala and M.Kurkela, Stainless Steel
stainless steels more attractive. ‘84, Metals Institute, 1985, p 240–247
25. T.Suter, E.G. Webb, H. Bohni, and
REFERENCES R.C. Alkire, J. Electrochem. Soc., Vol 148
(No. 5), 2001, B174–B185
1. M.P. Ryan et al., Critical Factors in Local- 26. M.O. Spiedel, Stainless Steels ‘87, Institute
ized Corrosion, Proc. Electrochem Soc., Vol of Metals, London, 1988, p 247–252
150, 2003, p 284–294 27. Y. Cao, F. Ernst, and G.M. Michal, Acta
2. K. Sieradski and R.C. Newman, J. Elec- Mater., Vol 51, 2003, p 4171.
trochem. Soc., Vol 133, 1986, p 1980 28. G. Lothongkum et al., Corros. Sci., Vol 48,
3. L. Brewer, Science, Vol 161, 1968, p 115 2006, p 137–153
4. W.J. Tobler and S.Virtanen, Critical Factors 29. S. Fujimoto , Sci. Technol. Adv. Mater., Vol
in Localized Corrosion, Proc. Electrochem 5, 2004, p 195–200
Soc., 2003, p 583–594 30. J.D. Fritz, J.F. Grubb, B.W. Parks, and
5. B. Baroux et al., Corros. Sci., Vol 47 (No. C.P. Stinner, Stainless Steel World, KCI,
5), 2005, p 1097–1117 P01488, 2001
6. 31. M.O. Spiedel, Met. Trans. A, Vol 12A,
products/xq/asp/T.1/qx/productCategory. 1981, p 779
html 32. A.J. Sedricks, Corrosion of Stainless Steels,
7. K. Kimura et al., High Cr Stainless OCTG Wiley, 1979, p 158
with High Strength and Superior Corrosion 33. R.N. Parkins, Br. Corros. J., Vol 14, 1979, p 5
Resistance, JFE Technical Report 7, Jan 2006 34. S. Tahtinen, H. Hahhinen, and T. Hakkarainen,
8. Stainless ‘84, Metals Institute, 1985, p 143–148
documents/start.asp 35. M. Nagumo et al., Met. Trans. A, Vol 32A,
9. J.E. Truman, Corrosion: Metal/Environ- Feb 2001, p 332
ment Interaction, Vol 1, Newness-Butter- 36. H. Hanninen et al., Hydrogen Effects on
worths, 1976, p 352 Materials Behavior, TMS, 2003, p 201–210
10. J.P. Audouard, Stainless Steels, Les Editions 37. V.J. Gadgil, Scr. Metal., Vol 28, 1993,
de physique, 1993, p 268 p 1489–1494
11. H. Mimura et al., Nippon Steel Tech. Report 38. M. Hoelzel et al., Mater. Sci. Eng. A,
90, July 2004, p 94–99 Vol 384, 2004, p 255–261
12. 39. B.G. Pound, Hydrogen Effects on Materials
Documents/AL-6XN_sourcebook.pdf Behavior, TMS, 2003, p 93–103
13. 40. S.C. Dexter, Microbiologically Influenced
Documents/al610_611.pdf Corrosion, Corrosion: Fundamentals, Test-
14. F. Tagashi et al., Kawasaki Technical Re- ing, and Protection, Vol 13A, ASM Hand-
port 31, 1994 books, 2003, p 398–413
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 57-68 All rights reserved.
DOI: 10.1361/ssde2008p057



Summary is negative. In Eq 1, the free energy G is decreased

by a lower nobility of the metal (or a higher activ-
STAINLESS STEEL, often considered mainly ity a of a metallic alloying element), a lower tem-
as a corrosion-resisting material, plays an impor- perature T, and a higher partial pressure P of the
tant role as a heat-resisting material. This is oxidizing gas according to:
partly due to its ability to retain strength at higher
temperatures at which many otherwise useful al- ⎛ aMeX ⎞
loying systems, such as aluminum, copper, and ΔG = ΔG 0 + RT ln ⎜ y
⎟ (Eq 2)
⎜⎝ aM PX0.5 y ⎟⎠
even titanium, soften. Stainless steel retains 2

strength and has excellent oxidation resistance

from room temperature to nearly 1000 °C, at In the case of alloy oxidation, for which tem-
which other economical alternatives are lacking. peratures are high enough to form mixed oxides
or spinels, the activities of the oxide species
also need to be considered.
The standard Gibbs free energy ΔG0 is often
presented in Richardson-Jeffes (Gibbs free
energy-temperature) diagrams such as the one
High-temperature oxidation is a form of envi-
shown in Fig. 1 (Ref 4).
ronmental degradation of metals and alloys that
It is evident from Fig. 1 that the major alloying
results from the following chemical reaction in
element in stainless steels, chromium, forms a
which metal atoms M react with gaseous oxidants:
thermodynamically significantly more stable
M (s) + 0.5 yX 2 ( g) → MX y (Eq 1) oxide, Cr2O3, than those of the base alloy iron
(FeO, Fe3O4, and Fe2O3) or the major ternary el-
Due to the high temperatures involved, these ement nickel (NiO), and to a great extent, the
reactions are generally rapid and thus are a con- chromium content determines the oxidation be-
cern for high-temperature applications such as havior of stainless steels.
components for power generation. The elec- The Effect of Chromium. The oxidation of
tronegative gaseous oxidant X could be sulfur, multicomponent alloys is a complex process
chlorine, etc., but the discussion here mainly is from both thermodynamic and kinetic points of
limited to oxidation by oxygen or water vapor view. A range of oxides may form with various
(in the latter case, hydrogen would be added as degrees of thermodynamic stabilities and stoi-
a product in Eq 1. For a thorough study of oxi- chiometries (including complex ones with dif-
dation, referred to Ref 1 to 3). ferent cations), and there might be degrees of
solubilities of oxides in one another. Kinetics of
their growth is complex because metal solute
diffusion in the metal phases varies, as do metal
Thermodynamics of Oxidation and oxygen ion mobilities in the different oxide
As discussed in Chapter 2, Corrosion Theory, a Birks, Meier, and Pettit distinguished between
reaction will be possible when the net free energy two basic types of behavior: (a) a noble matrix
58 / Stainless Steels for Design Engineers

Fig. 1 Standard Gibbs free energy of formation of some metal oxides as a function of temperature. Source: Ref 4

metal with less-noble alloying elements and (b) spinels Fe1.5Cr1.5O4 (with a solid solubility with
both matrix element and alloying elements are Fe3O4) and FeCr2O4 form. The progressive
nonnoble. The concept of nobility is decided by change in oxidation behavior as chromium is
the thermodynamic conditions of Eq 2; that is, added to iron has been described in the litera-
an element for which the free energy defined by ture (Ref 1).
Eq 2 is negative is nonnoble. We first discuss the At lower chromium contents and above a min-
more common case (b), in which oxidation takes imum temperature, an iron-chromium alloy
place under significantly oxidizing conditions, would behave as pure iron, where FeO would
such as air. In such a situation, it can be seen form next to the metal, then gradually Fe3O4 and
from Fig. 1, that the matrix iron is nonnoble and Fe2O3 would form toward the gas as oxygen po-
so are many of the solutes (chromium, molybde- tential increases. Isolated pockets of spinel may
num, aluminum, silicon, manganese, etc.). form within the FeO layer. The oxidation of iron
A high-temperature Fe-Cr-O phase diagram is proceeds predominantly due to the rapid ionic
shown in Fig. 2. It can be seen here that Fe2O3 diffusion of Fe2+ cations on the FeO layer,
and Cr2O3 are soluble in each other, and that the which leads to growth of this layer. If chromium
Chapter 5: Oxidation / 59

Fig. 3 Parabolic rate constants for the growth of several oxides.

Source: Ref 6

Here, m is the added mass, A is the area

exposed to the oxidizing atmosphere, t is the
time exposed, and km is the parabolic rate con-
stant. The subscript “m” is added here to denote
that the reaction is measured as added mass (it
can also be defined for oxide thickness X). As
Fig. 2 The iron-chromium-oxygen phase diagram at 1300 °C. shown in Fig. 3, the parabolic rate constant for
Source: Ref 5 the oxides of chromium, silicon, and aluminum
are low compared to others, and this is the rea-
in the base alloy is increased, the spinel pockets son that these elements are used as alloying ele-
increase, and the mobility of Fe2+ decreases. As ments to reduce oxidation rates for alloys in
chromium content is increased further, a mixed- high-temperature applications.
spinel scale is formed. Iron diffusion through the The solubility of Fe3O4 in the spinel will
mixed spinel is significant, and thus the scale is eventually result in continued iron oxide for-
not yet protective. As chromium content is in- mation, as the iron-chromium system is not an
creased further, an outer layer of Cr2O3 is optimal basis for high-temperature oxidation
formed, and the oxidation behavior becomes resistance, although it might be an option from
similar to that of chromium. A chromium limit an economical standpoint compared to other
of roughly 20% is needed to achieve a perma- alloy systems such as superalloys. The high
nent Cr2O3 scale. This amount decreases if mobility of both iron and manganese in the
nickel is added. While the thermodynamic driv- spinel structure is also an important factor.
ing force is important, the chromium content of The major stainless steels used for oxidation
the alloy can override because the supply of resistance fall into two categories: the ferritic
chromium to the interface becomes dominant. stainless steels and the austenitic. Table 1 lists
At chromium contents less than about 16 wt%, some of the more significant alloys commonly
the oxidation rate is influenced by the rate of encountered in applications for which oxida-
supply of chromium from the alloys beneath the tion resistance is paramount. The value of fer-
oxide. Above 16%, the supply of chromium is ritic alloys (such as 409, 439, and 446) is that
fast enough that chromium gradients are low they are relatively inexpensive, and that they
enough that instead transport in the oxide layer have a thermal expansion coefficient that is
controls the rate. The rate of oxidation then fol- closer to that of the oxide than do austenitic al-
lows a so-called parabolic law (this is explained loys (such as 302B, 309, and 310). This gives
in the next section), by which the mass change them an advantage in cyclic oxidation applica-
per unit area due to oxidation (incorporation of tions even though their strength at high temper-
oxygen) is given by: atures does not rival that of austenitic alloys.
The ferritic stainless steels are the most widely
( m / A)
= km t (Eq 3) used alloys based on their low cost, which has
60 / Stainless Steels for Design Engineers

Table 1 Oxidation-resisting grades of stainless steel in common use

Composition, %

UNS Name C N Cr Ni Mn Si Ti Nb Other

S40900 409 0.08 10.5–11.75 0.5 1 1 6X(C + N) to ... ...
11Cr-Cb(a) 0.01 0.015 11.35 0.2 0.25 1.3 ... 0.35 ...
12SR(a) 0.02 0.015 12 ... ... ... 0.3 0.6 1.2 Al
S43935 439 0.07 0.04 17.0–19.0 0.5 1 1 0.20 + 4X(C + ... ...
N) to 0.75
18Cr-Cb(a) 0.02 ... 18 ... 0.3 0.45 0.25 0.55 ...
18SR(a) 0.015 ... 17.3 0.25 0.3 ... 0.25 ... 1.7 Al
4742(a) 0.08 ... 18 ... 0.7 ... ... ... 1.0 Al
S44600 446 0.2 0.25 23.0–27.0 0.6 1.5 1 ... ... ...
S30215 302B 0.15 ... 17.0–19.0 8.0–10.0 2 2.0–3.0 ... ... ...
S30415 153MA 0.04–0.06 0.12–0.18 18.0–19.0 9.0–10.0 0.8 1.0–2.0 ... ... 0.04 Ce
S38150 253MA 0.05–0.10 0.14–0.20 20.0–22.0 10.0–12.0 0.8 1.4–2.0 ... ... 0.04 Ce
S30900 309 0.2 ... 22.0–24.0 12.0–15.0 2 0.75 ... ... ...
S31000 310 0.25 ... 24.0–26.0 19.0–22.0 2 1 ... ... ...
Note: All compositions include Fe as balance. Single values are maximum, unless otherwise specific (a) Indicates typical analysis

made them the standard alloys for automotive scale-gas surface. Similarly, the electric charge
exhaust systems. can be carried by either n-type (electrons) or p-
type (electron holes) electronic defects. The
case will be determined by the equilibrium de-
Transient Oxidation fect structure of the oxide, which depends on
temperature and oxygen partial pressure. In the
The oxidation of a clean metal surface on ex- case of Cr2O3, chromium cations are the pre-
posure to an oxidizing environment will ini- dominantly mobile defects (Fig. 4b) as a result
tially lead to all the nonnoble components of the of a very small degree of deviation from stoi-
alloy being oxidized together, forming mixed chiometry in the cation lattice, that is, Cr2-5O3,
oxides having composition similar to the base leading to metal deficiency. The contribution of
alloy. In stainless steels, these initial oxides are chromia grain boundary diffusion is large and
typically Fe-Cr-Ni-Mn mixed oxides. This is probably dominates the process at temperatures
called transient oxidation. As these oxides of interest. The defect can be described as an in-
thicken, the partial pressure of oxygen at the teraction with oxygen, at high oxygen poten-
scale-metal interface falls until only the most tials, through Kroger-Vink notations as:
reactive element present in high concentration
can be oxidized. For stainless steels, this means 3 3
O ( g) = VCr''' + OOx + 3h •
that a layer of Cr2O3 is eventually established in 2 2 2
contact with the alloy.
−3 3
⎡VCr''' ⎤ p3 = K 2 pO3/ 4 → ⎡VCr''' ⎤ = K12 K 2 Po4 (Eq 4)
⎣ ⎦ ⎣ ⎦
The Electrochemical Nature of 2 2

The electron holes that form as charge-com-
pensating defects serve as the “electron lead” in
Once an inner scale of Cr2O3 is formed, as
the electrochemical cell in Fig. 4b. The free en-
shown in Fig. 4(a), the oxidizing gas is reduced
ergy of Reaction 4 determines the concentration
at the gas-scale interface, and the chromium is
of mobile defects and thus the diffusion coeffi-
oxidized at the metal-scale interface. The Cr2O3
cient and electrochemical mobility (Be) of the
scale serves as both electrolyte, through which
cation according to:
ions are transported, and electron lead, through
which electronic defects are transported. In
principle, either or both metal or oxygen ions ⎛ ΔGm ⎞
DCr 3+ = ⎡⎣VCr 3+ ⎤⎦ γΛ 2 υ ⋅ exp ⎜ −
can migrate. If oxygen ion mobility dominates, ⎝ RT ⎟⎠
then the scale would continue to grow at the
⎛ ΔH m ⎞
oxide-metal interface, whereas if chromium ion = ⎡⎣VCr 3+ ⎤⎦ ⋅ cons tan t ⋅ exp ⎜ − (Eq 5)
mobility dominates, the oxide will grow at the ⎝ RT ⎟⎠
Chapter 5: Oxidation / 61

Fig. 4 Metal with oxide scale. (a) A protective scale that prevents gas access. (b) Schematic of electrochemical oxidation through a
protective oxide scale that serves as electrolyte and electron lead. The case is for mobile cations

3FDCr 3+
effects are ignored (i.e., scales are relatively
3+ =
BCr (Eq 6) thick compared to range of space charge ef-
RT fects). As a case study, let us assume that the
mobile ion defect is cations due to metal vacan-
where zi is the ion charge, F is Faraday’s con- cies in the scale.
stant (96,457 C.eq–1). Compared to wustite The molar flux J (moles/m.s) of a particle i in
(FeO), the equilibrium constant of Eq 4 is quite an electrolyte subjected to an electrochemical
low, resulting in a low degree of nonstoichiome- potential gradient was shown to be:
try in Cr2–δO3 compared to Fe1–δO (where δ can
be as large as 0.05), and thus the transport of
Ji = − ci Bi
∂ ( μ i + zi Fφ ) (Eq 7)
Cr3+ through its scale is much slower than the ∂x
transport of Fe2+ through FeO and thus the dif-
ference in parabolic rate constants in Fig. 3.
where zi is the ion charge, F is Faraday’s con-
stant (96,457 C/gram equivalent), and φ is the
electric field (V). The electronic or ionic con-
Kinetics and Oxidation Rates: Wagner’s ductivity κ in an electrolyte can be computed
Theory through:

κ = F 2 ∑ zi Bi ci
The parabolic oxidation rate was introduced (Eq 8)
without explanation in Eq 3. It was first de-
scribed in terms of oxide defect structure and re- The contribution of a given ion specie type or
sulting transport properties by Wagner (Ref 6), electron defect type to this conductivity is de-
and the theory is explained in most of the mono- noted as the partial conductivity and computed
graphs on oxidation, such as Chapter 3 in Ref 3 as:
and Chapter 4 in Ref 1. This treatment follows
the derivation in Ref 1. Consider a general case, 2
as shown in Fig. 4 under the assumptions that κ i = F 2 zi Bi ci (Eq 9)
(a) the scale is compact and adherent, (b) elec-
trode reactions are rapid enough to be in equi- Inserting Eq 9 into Eq 7 yields:
librium at the interface and surface, (c) nonstoi-
chiometry is small and uniform throughout the
Ji = −
κi ∂ ( μ i + zi Fφ )
scale (i.e., defects are in thermal equilibrium (Eq 10)
F zi
throughout the scale), and (d) double-layer
62 / Stainless Steels for Design Engineers

Now, if the mobile particles are a single type assumed, that is, a linear drop across X, at all
of metal cations (e.g., Cr3+) and electrons, then times:
two fluxes are present:
dx k
Jc = −
κc ∂ ( μ c + zc Fφ ) =
dt X (t ) (Eq 18)
2 2
∂x (Eq 11)
F zc
The constant k is the parabolic rate constant.
and A mass balance can be written where the flux of
cations for a period of time dt is equated to the
Je = −
κe ∂ ( μ e + ze Fφ ) amount of metal being accumulated as cations
2 2
∂x (E q 12) inside the scale of thickness dx:
F ze
Jc ⋅ dt = CC dx (Eq 19)
Electrical neutrality requires that:

Jc Z c + Je Z e = 0 By combining Eq 18 and 19 and inserting Eq

(Eq 13)
16 for the flux, an expression for the parabolic
rate constant is obtained:
And at the oxide-scale/metal interface, the
anode reaction is in equilibrium, that is: μ 'M
z + 1 κ cκ e
M = M + zce − and therefore,
k= 2 2 ∫κ dμ (Eq 20)
zc F CC μ ''M c
+ κe M

μ M = μ c + zc μ e (Eq 14) If the mobility and thus partial conductivity

of electrons is significantly higher than that of
Combining Eq 11 to 14, the potential gradient the ions (a reasonable assumption), then Eq 20
is eliminated, and the chemical potential gradi- can be simplified as:
ents can be replaced by the metal (M) potential
gradient, and the following equation results: μ 'M
κ cκ e ∂μ M
k= 2 2
zc F CC ∫ κ dμ
c M (Eq 21)
Jc = −
zc2 F 2 ( κ c + κ e ) ∂x (Eq 15)
Since from diffusion theory we know that
To obtain an explicit function for the cation Dc = BcRT, and inserting this in Eq 9, one
flux, Eq 15 needs to be integrated after variable obtains
separation, keeping in mind that conductivities
and metal chemical potential may vary within 2 Dc
κ c = F 2 zc c
the scale. Integrating from the gas-scale surface RT c
(x = 0, μM = μ′′M) to the scale-metal interface
(x = X, μM = μ′′M) one obtains. and inserting this into Eq 21 results in:

μ" μ 'M
1 M
κκ 1
Jc = − 2 2 ∫ c e dμ M (Eq 16) k= ∫ D dμ (Eq 22)
zc F x μ ' c + κ e
κ RT μ ''M
c M

Now, if the growth of the oxide scale is con- It is often more convenient to express Eq 22
trolled by the flux of cations: in terms of oxygen potentials rather than the
metal potentials. It was assumed at the onset of
dx ∂C this analysis that the deviation from stoichiom-
∝ JC = − DC C (Eq 17)
dt ∂x etry is small and constant throughout the scale.
Therefore, the oxide potential is constant and:
The concentration drop across the scale is
constant since interface and surface reactions zc
μ + μ M = μ MO = constant (Eq 23)
are at equilibrium. If quasi steady state is 4 O2 zC / 2
Chapter 5: Oxidation / 63

And thus, dμ O = dRT ln PO ∝ − dμ Me. Therefore, 2

thickness loss due to evaporation, which is de-
scribed by a first-order reaction kinetics expres-
Eq 22 can be written:
sion with rate constant ke. Thus, the thickness
ln PO'' PO'' change becomes:
2 2
k∝ ∫ Dc d ln PO =
2 ∫ dP
PO O2 (Eq 24) dx k
ln PO'
2 = −k (Eq 27)
dt X (t ) e

In the case of Cr2O3, if bulk diffusion is dom-

inating, Eq 4 and 5 inserted into the diffusion This results in a so-called paralinear (as op-
coefficient in Eq 24 for C = Cr3+, results in: posed to parabolic) rate for the oxide thickening
(Fig. 6), and at a critical oxide scale thickness
X( the rate of thinning due to evaporation equals
PO'' PO''
P 3/ 4 2 the rate of thickening due to oxidation. In Eq
k∝ ∫ PO
dPO = ∫ P −1/ 4 dPO
2 2
27, this means that dx/dt = 0; consequently, X =
2 P' O2 k/ke. While at first this seems to suggest that it

( ) − (P )
3 ⎡ '' 3/ 4 3/ 4
⎤ does not affect the oxidation process in that the
= P '
(Eq 25)
4 ⎢⎣ O2 O2 ⎥⎦ rate of oxidation does not increase, the forma-
tion and evaporation of chromium oxides re-
sults in greater chromium consumption in the
Thus, the parabolic rate constant would be pre- alloy compared to what would be the case if
dicted to vary with the power of three-quarters of evaporation did not occur. As a result of evapo-
the external oxygen partial pressure. ration losses, stainless steels that depend on a
Grain boundary diffusion has however been protective chromium oxide layer are limited in
identified to be important in the case of Cr3+ use to temperatures up to 900 to 1000 °C.
transport (Ref 8). The observed growth rate of The presence of water vapor promotes the
Cr2O3 polycrystalline films is far too fast to be formation of even more volatile oxyhydroxides
accounted for by bulk diffusion of chromium (e.g., CrO2(OH)2 ) (Ref 11, 12).
ions; instead, grain boundary diffusion would
be expected to dominate (Ref 9).
Spalling and Cracking of the Scale

The Volatile Nature of Cr2O3 At elevated temperatures or during tempera-

ture cycling, there are multiple ways in which
At high enough temperatures and high stresses can develop that may crack and blister
enough oxygen partial pressures, the formation the scale, rendering it nonprotective. The differ-
of a gaseous hexavalent chromium oxide CrO3* ent causes of stress generation are described in
could lead to thinning of the Cr2O3 scale ac- Chapter 5 in Ref 1. So-called growth stresses
cording to the following reaction: arise due to changes caused by the oxidation
process itself. These include differences in lat-
3 tice mismatch, alloy depletion in the metal,
Cr2 O3 (s) + O 2 ( g) = 2CrO3 ( g) (Eq 26) point-defect gradients in scales containing ox-
ides such as FeO, with large deviation from sto-
Figure 5 shows the vapor pressure of the su- ichiometry, recrystallization, and volume differ-
peroxide as a function of temperature and par- ences between the oxide and metal. The last is
tial pressure of O2. perhaps the most commonly mentioned and is
The effect of this reaction on the oxidation ki- characterized by the Pilling-Bedworth ratio, ab-
netics can be described as follows: The thick- breviated as PBR (Ref 13).
ness change described through the parabolic
VOxide VmOxide
rate constant in Eq 18 is corrected for by the PBR = = Metal (Eq 28)
VMetal Vm * ν

* Hexavalent chromium is now considered a human car-

cinogen and is rigorously regulated by both the Occupa- Here, the subscript m stands for molar volume,
tional Safety and Health Administration (OSHA) and the and υ is the number of metal atoms needed to
U.S. Environmental Protection Agency (EPA). form a stoichiometric unit of the oxide (in the
64 / Stainless Steels for Design Engineers

Fig. 5 Chromium-oxygen system species volatility as a function of temperature and oxygen pressure. Source: Ref 10

case of Cr2O3, υ is 2, and in the case of FeO, it

is 1). When PBR is greater than 1, then the
oxide is expected to be in compression and is
likely to be protective, whereas if it is less than
1, the oxide is in tension and thus nonprotective.
There are, however, many exceptions to this,
partly because the stress state often depends
more on the mechanisms and conditions of the
oxidation process rather than the properties of
metal and oxides.
Thermal, stresses are caused by differences in
thermal expansion between the oxide and metal,
and the stresses generated in oxide scales can be
estimated: Fig. 6 Schematic of paralinear oxidation as a result of evapo-
ration of chromium superoxide

σ Oxide =
− EOxide ( α Oxide − α Metal ΔT
ity, t is thickness, and ΔT is the temperature
(Eq 29)
⎛ ⎞
( ) t E
1 − ν p ⎜ 1 + 2 Oxide Oxide ⎟
⎝ t Metal E Metal ⎠
change. In general, α is larger for the metal than
the oxide; thus, during cooling the stresses are
expected to be compressive and during heating
The equation is written for a case shown tensile. Thermal stresses can cause spalling of
schematically in Fig. 7, where both sides on a the protective oxide layer, and it is most severe
metal undergo oxidation. Here, σ is the stress, under cyclic conditions.
υp is Poisson’s ratio (it has been assumed that Effect of Silicon, Aluminum, and Molyb-
there is no mismatch), α is the coefficient of denum. Due to concerns about the cost of
thermal expansion, E is the modulus of elastic- chromium and its (former) classification as a
Chapter 5: Oxidation / 65

internal Cr2O3 layer forms and thickens. Even-

tually, the surface aluminum-oxide layer flakes
off. Also, due to the low oxygen potential
needed to form Al2O3, internal oxidation may
result below the metal-scale interface in alloys
tox tm tox in which formation of a continuous alumina
scale film does not occur.
Molybdenum is suggested to strain the lat-
tice due to its larger size and consequently in-
Oxide Metal Oxide crease the rate of bulk diffusion of elements
(Ref 11), which can enhance the rate of initial
Cr2O3 formation. Molybdenum is usually con-
sidered detrimental for oxidation resistance.
Molybdenum normally forms MoO2 oxide, but
Fig. 7 Schematic of a cross section of oxidized sample indi- this can oxidize further to form the low-melt-
cating dimensions in Eq 29 for predicting thermal ing and volatile MoO3. If the MoO3 evapo-
rates, there is little problem, but if its
volatilization is inhibited by low atmosphere
circulation, liquid MoO3 can accumulate and
strategic material, there were efforts to try to dissolve the protective Cr2O3 scale, leading to
substitute less-expensive elements such as alu- catastrophic oxidation.
minum and silicon that also are known to form Effect of Rare Earth Additions. Cerium, lan-
protective layers (Ref 14), even though they thanum, and yttrium additions are known to im-
have significant metallurgical and mechanical prove oxidation resistance of high temperature
drawbacks. nickel- and iron-based alloys (Ref 6). Rare earth
Silicon additions of 1.5 wt% or more have the additions have been suggested to have a multi-
effect of forming a continuous amorphous sub- tude of beneficial effects, such as reducing the
surface layer in iron-silicon and Fe-Cr-Si alloys growth kinetics of Cr2O3 scales, stabilizing
that is relatively impervious to transport of ions. Cr2O3 scales at lower chromium levels, increas-
The mechanism for the evolution of such a layer ing adhesion, and preventing spalling of the
is as follows: (1) The more readily available oxide scale during thermal cycling. The expla-
iron or chromium first forms a surface layer, nation for any of this does not seem clear, but
and this causes an enrichment of silicon at the some hypotheses have been suggested. The ef-
oxide-metal interface. (2) As sufficient silicon is fect on the growth kinetics could be because the
enriched, the SiO2 layer is formed. It has been reactive element ions collect at grain boundaries
reported that alloys with chromium content as and block fast path diffusion.
low as 6 wt% and silicon content of 1.5 wt% The improved adherence could be because
perform in terms of oxidation as well as com- these elements getter tramp elements such as
mercial stainless steels. Also, an addition of 4 sulfur and suppress void formation at the inter-
wt% Si to a Fe14wt%Cr14wt%Ni alloy resulted face. Furthermore, they might form so-called
in a 200-fold reduction in weight gain at 900 °C oxide pegs at the interface (Ref 16). The pre-
However, this SiO2 layer seems to promote cise role of the rare earth additions to Cr2O3
oxide spalling, especially in cyclic service. oxide protection and the mechanism by which
Aluminum forms a very stable thin outer they are incorporated into the scale during the
layer of Al2O3 that initially reduces the oxida- surface treatment processes remain unknown.
tion rate. Alumina is among the most stable and An understanding of these fundamental issues
defect-free oxides, giving it an extremely low would help to develop optimum alloy
diffusion rate. In consequence, if sufficient alu- chemistries for selected high-temperature and -
minum is present to maintain the protective alu- pressure applications and to further develop the
mina scale, the aluminum-bearing alloys pro- surface infusion process. The lack of funda-
vide the greatest oxidation resistance attainable mental understanding of how rare earths im-
in engineering alloys. If, however, the alu- prove oxidation resistance has not stopped the
minum content is not sufficient to force alumina development of several alloys that benefit from
scale formation at the scale-metal interface, an the effect.
66 / Stainless Steels for Design Engineers

Oxidation Under Less-Oxidizing quasi-steady-state situation as shown in Fig. 9,

Atmospheres the flux can be written:

dm N X − N OS NS
The two types of alloy oxidation behaviors, JO = = − DO O = DO O (Eq 30)
(a) a noble matrix metal with less-noble alloy- dt Vm X Vm X
ing elements and (b) both matrix element and
alloying elements are nonnoble, were men- Within the depth X, the oxygen solubility is
tioned. When designing against oxidizing envi- such that Eq 2 is negative enough that Cr2O3
ronments, case b is perhaps the most relevant, forms. Beyond X, it is not. At the distance X, the
and most of the discussion has been devoted to oxygen concentration is negligibly low com-
this. However, during annealing for microstruc- pared to the surface composition, which is in
tural control, steels are exposed to furnace gases equilibrium with the gas phase. The molar vol-
at high temperatures that have relatively low ume Vm is used to obtain the flux in units of
oxygen or steam contents, for which case a will moles per square meter. Within the layer 0 < x <
apply, that is the atmosphere does not cause iron X, all the chromium is assumed to be oxidized;
(or nickel) to oxidize but chromium (and alu- therefore, the accumulated mass due to oxygen
minum, silicon, molybdenum, etc.) does. addition:
For simplicity, assume a binary system A-B
of “noble” iron and “reactive” chromium. In o 3

this case, depending on the concentration of N B0 υX N Cr 2 X

m= = (Eq 31)
the reactive element and atmosphere, the Vm Vm
oxide of the reactive element can, in principle,
form either on the surface (as has been dis- Differentiating Eq 31 with time, equating to
cussed so far) or internally as discrete oxide Eq 3, and separating variables results in:
particles in a metal matrix through oxygen
diffusion into the metal. Both cases are shown N OS DO N OS DO
schematically in Fig. 8. Let us discuss the XdX = dt = (Eq 32)
υN BO 3 0
conditions that promote one or the other of N
2 Cr
these by starting with a situation in which (a)
no surface oxide exists and (b) the oxygen at-
mosphere is such that the solubility of oxygen Integrating Eq 32 from x = 0 to x = X results in
within a distance X is enough to thermodynam- an expression for the internal oxidation depth X:
ically render Cr2O3 stable according to Eq 2 1/ 2
but none of the iron oxides. The derivation is 1/ 2
⎛ ⎞
⎛ 2N S D ⎞ ⎜ 2N S D ⎟
done in terms of both a generic system A-B X =⎜ O0O t⎟ =⎜ O O t⎟ (Eq 33)
causing an oxide BOν and for iron-chromium ⎝ υN B ⎠ ⎜ 3 N Cr

causing Cr2O3. Assume for a start that the DO ⎝ 2 ⎠
>> DCr, and thus while oxygen diffuses into
the alloy, chromium does not counterdiffuse.
The flux of oxygen inward into the metal is
then the cause of increased mass. Assuming a

Fig. 8 Schematic of two cases in a less-oxidizing atmosphere. Fig. 9 Quasi-steady-state approximation of the moving
(a) Adsorption of oxygen leading to internal oxidation boundary problem of internal oxidation. Counterdif-
and (b) external oxidation as the B element migrates. fusion of B is assumed to be negligible
Chapter 5: Oxidation / 67

This expression predicts a parabolic depend-

ence of X with time, just as Eq 3 did for the ex-
ternal oxidation. For a more rigorous derivation
(without assuming quasi steady state), refer to
Appendix B in Ref 1. Now, let us see what
causes this to transition into an external scale. It
was assumed in the derivation of Eq 33 that
counterdiffusion of chromium does not occur.
When considering Eq 33, is clear that the rate of
penetration of the internal oxidation front will
decrease with (a) increasing NB0, (b) decreasing
NOS, and (c) decreasing DO. If DCr was not negli-
gible, there would be a gradual change in oxida-
tion morphology if the ratio (NB0DB)/(NOSDO).
Gradually, if the ratio were increased there
would be a slowing of the inward penetration of
the internal oxide front and an enrichment of Fig. 10 Temperature dependence of metal dusting of iron.
BOυ in the internally oxidized zone. Wagner de- Source: Ref 18
veloped a model (Ref 16), based on that at some
point, when the volume fraction of BOυ versus trogen absorption, but if the oxygen is depleted
volume metal in the internally oxidized zone before all surfaces are oxidized, the remaining
reaches a critical value g*, there is a transition material can be rapidly nitrided by the residual,
from internal to external oxidation; specifically, essentially pure, nitrogen atmosphere.
this happens when:
1/ 2
N >⎢
NO ⎥ (Eq 34)
⎣ 2ν
DBVOx ⎦ 1. N. Birks, G.H. Meier, and F.S. Pettit, Intro-
duction to the High-Temperature Oxidation
If more than one reactive elements were pres- of Metals, 2nd ed., 2006, Cambridge Uni-
ent (such as is the case in stainless steels in versity Press, New York
which aluminum, silicon, molybdenum, nio- 2. P. Kofstad, High Temperature Corrosion,
bium, etc., may be present), this will decrease 1988, Elsevier Applied Science, London
the inward flux of oxygen, and thus the transi- 3. K. Hauffe, Oxidation of Metals, Plenum
tion to external may occur at a lower solute Press, New York, 1965
(CCr0) concentration than what is predicted by 4. F.D. Richardson and J.H.E. Jeffes, J. Iron
Eq 34. Steel Inst., Vol 160, 1948, p 261
Metal Dusting. Under reducing conditions, 5. C. Wagner and K. Grünewald, Z. Phys.
in products of combustion atmospheres, oxida- Chem., Vol 40B, 1938, p 455
tion and carburization may occur simultane- 6. J.H. Park, W.E. King, N.L. Peterson, and
ously and at a higher rate than exhibited in pure S.J. Rothman, The Effect of Reactive Ele-
oxidation. Under even more reducing condi- ment on Self-Diffusion in Cr2O3, Norman
tions, the condition called metal dusting may L. Peterson Memorial Symposium, Oxida-
occur. Metal dusting is often characterized by tion of Metals and Associated Mass Trans-
the generation of large, smooth pits that look as port, edited by M.A. Dayananda, S.I. Roth-
if metal had been scooped from the surface. The man, and W.E. King, AIME, Warrendale,
underlying phenomenon is the formation of PA, 1998, p 103–107
metal carbides, which manifests itself as the 7. C. Wagner and K. Grünewald, Z. Phys.
breakup of bulk metal to metal powder. This oc- Chem., Vol 40B, 1938, p 455
curs at temperatures at which the carbide is 8. D. Caplan and G.I. Sproule, Effect of Oxide
most stable (Fig. 10). Grain Structure on the High Temperature
During oxidation in air, if large surface area is Oxidation of Cr, Oxid. Met., Vol 9, 1975,
present under conditions of restricted air supply, p 459–472
oxygen can be depleted to the point that oxida- 9. B.B. Ebbinghaus, Combust. Flame, Vol 93,
tion essentially ceases. Oxide films inhibit ni- 1993, p 119–137
68 / Stainless Steels for Design Engineers

10. K. Hilpert et al., JECS, Vol 143/11, 1996, p of Metals and Associated Mass Transport,
3642–3647 edited by M.A. Dayananda, S.I. Rothman,
11. N.B. Pilling and R.E. Bedworth, J. Inst. and W.E. King, AIME, Warrendale, PA,
Met., Vol 29, 1923, p 529 1998, p 323–340
12. J.K. Tien and J.M. Davidson, Oxide Spalla- 14. E.J. Felten, J. Electrochem. Soc., Vol 108,
tion Mechanisms, Stress Effects and the Ox- 1961, p 490
idation of Metals, ed. J.V. Cathcart, AIME, 15. C. Wagner, J. Electrochem. Soc., Vol 103,
New York, 1975, p 200 1956, p 571
13. J. Rawers, Understanding the Oxidation 16. C.M. Chun, J.D. Mumford, and T.A. Rama-
Protection of Fe-Cr-Si Alloys, Norman L. narayanan, J. Electrochem. Soc., Vol 147,
Peterson Memorial Symposium, Oxidation 2000, p 3680
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 69-90 All rights reserved.
DOI: 10.1361/ssde2008p069


Austenitic Stainless Steels

Summary because their greater thermal expansion co-

efficient tends to cause the protective oxide
AUSTENITIC STAINLESS STEELS are the coating to spall.
most common and familiar types of stainless 2. They can experience stress corrosion crack-
steel. They are most easily recognized as non- ing (SCC) if used in an environment to which
magnetic. They are extremely formable and they have insufficient corrosion resistance.
weldable, and they can be successfully used 3. The fatigue endurance limit is only about
from cryogenic temperatures to the red-hot tem- 30% of the tensile strength (vs. ~50 to 60%
peratures of furnaces and jet engines. They con- for ferritic stainless steels). This, combined
tain between about 16 and 25% chromium, and with their high thermal expansion coeffi-
they can also contain nitrogen in solution, both cients, makes them especially susceptible to
of which contribute to their high corrosion re- thermal fatigue.
sistance. Were it not for the cost of the nickel However, the risks of these limitations can be
that helps stabilize their austenitic structure, avoidable by taking proper precautions.
these alloys would be used even more widely.

Introduction Alloy Families in Perspective

Austenitic stainless steels have many advan- The fundamental criterion in the selection of
tages from a metallurgical point of view. They a stainless steel is generally that it can survive
can be made soft enough (i.e., with a yield with virtually no corrosion in the environment
strength about 200 MPa) to be easily formed by in which it is to be used. Good engineering
the same tools that work with carbon steel, but practice sometimes requires that materials be
they can also be made incredibly strong by cold selected for sufficient, but finite, service life.
work, up to yield strengths of over 2000 MPa This is especially true for high-temperature
(290 ksi). Their austenitic (fcc, face-centered service, for which creep and oxidation lead to
cubic) structure is very tough and ductile down limited life for all materials. The choice among
to absolute zero. They also do not lose their the stainless steels that can be used in that envi-
strength at elevated temperatures as rapidly as ronment is then based on the alloy from which
ferritic (bcc, body-centered cubic) iron base al- the component can be produced at the lowest
loys. The least corrosion-resistant versions can cost, including maintenance, over the intended
withstand the normal corrosive attack of the service life. The ferritic stainless steels are less
everyday environment that people experience, expensive for the same corrosion resistance but
while the most corrosion-resistant grades can sometimes are found lacking because of:
even withstand boiling seawater.
• Lack of toughness, as is the case at subambi-
If these alloys were to have any relative
weaknesses, they would be: ent temperatures or in thicknesses greater
than about 1.5 mm
1. Austenitic stainless steels are less resistant • Lack of great ductility, specifically if more
to cyclic oxidation than are ferritic grades than about 30% elongation is needed
70 / Stainless Steels for Design Engineers

• Susceptibility to high-temperature embrit- The austenitic alloys can have compositions

tling phases when moderately alloyed anywhere in the portion of the Delong diagram
labeled austenite shown in Fig. 1 (Ref 1). This
The less-expensive martensitic grades are diagram was designed to show which phases are
used instead of austenitic when high strength present in alloys in the as-solidified condition,
and hardness are better achieved by heat treat- such as found in welds. Thus it also applies to
ing rather than by cold work, and mechanical castings and continuously cast products. As a
properties are more important than corrosion re- practical matter of castability, the composition
sistance. This is also the case for the more ex- of most commercial alloys falls along the zone
pensive PH grades, which can achieve corrosion of several percent ferrite as cast. The salient fea-
resistance only matching the least corrosion re- ture of austenitic alloys is that as chromium and
sistant of the austenitic alloys. molybdenum are increased to increase specific
Duplex grades match austenitic grades in cor- properties, usually corrosion resistance, nickel
rosion resistance and have higher strength in the or other austenite stabilizers must be added if
annealed condition but present the designer the austenitic structure is to be preserved.
with challenges with regard to embrittling The traditional way of displaying the
phases that can form with prolonged exposure austenitic stainless steels is to present 302 as a
to elevated temperatures and only moderate base. Figure 2 shows one such diagram. Dia-
ductility like the ferritic alloys. grams such as these treat alloys as an evolution-
So, the austenitic grades are the most com- ary family tree and subtly mislead. Many alloys
monly used grades of stainless mainly because, were pushed toward obsolescence because of
in many instances, they provide very predictable advances in processing. For instance, 321 was
levels of corrosion resistance with excellent me- developed as an alloy in which the detrimental
chanical properties. Using them wisely can save effects of carbon were negated by addition of ti-
the design engineer significant costs in his or her tanium. The widespread adoption of the argon
product. They are a user-friendly metal alloy oxygen decarburization (AOD) in the 1970s
with life-cycle cost of fully manufactured prod- made this alloy unnecessary, except for special
ucts lower than many other materials. circumstances, since carbon could be cheaply

Fig. 1 Schaeffler-Delong stainless steels constitution diagram. Adapted from Ref 1, 2

Chapter 6: Austenitic Stainless Steels / 71

Fig. 2 The austenitic stainless family. Source: Ref 3

removed routinely. Likewise, 302 gave way to nickel equivalents (manganese, nitrogen, carbon,
the lower-carbon 304, for which the even lower- etc.) must also be added in matching amounts,
carbon 304L is commonly substituted and du- austenite stability is also increased. If molybde-
ally certified to qualify as either grade. While num, a chromium equivalent, is added, corro-
low carbon prevents sensitization, stabilized sion resistance but not oxidation resistance is
grades may still be preferred for special applica- enhanced. And, if nitrogen is the austenite stabi-
tions such as type 321 in aerospace and type lizer added to balance increases chromium or
347 in refinery service. Similar inertia keeps the molybdenum, then corrosion resistance is also
higher-nickel 300 series as the de facto standard increased. With small exceptions, that is the
when the more cost-efficient high-manganese rationale of austenitic grade design. Silicon
200 series is the logical basic grade. The rele- is used as an alloy to promote oxidation resist-
vant types of austenitic alloys can nonetheless ance and resistance to corrosion by oxidizing
be rationalized with this diagram. acids. Copper is used to promote resistance to
As chromium is added, oxidation resistance sulfuric acid. Rare earths make a more stably
and corrosion resistance increase. Because oxidation-resisting scale. Niobium increases
72 / Stainless Steels for Design Engineers

creep resistance. Sulfur and selenium increase Lean Alloys

In this chapter, austenitic alloys are classified Lean austenitic alloys constitute the largest
into three groups: portion of all stainless steel produced. These are
• Lean alloys, such as 201 and 301, are gener- principally 201, 301, and 304. Alloys with less
ally used when high strength or high forma- than 20% chromium and 14% nickel fall into
bility is the main objective since the lower, this unofficial category. Since they are stainless,
yet tailorable, austenite stability of these al- it is generally taken for granted that these alloys
loys gives a great range of work-hardening will not corrode, and these alloys have suffi-
rates and great ductility. Richer alloys, such cient corrosion resistance to be used in any in-
as 305, with minimal work hardening are the door or outdoor environment, excluding coastal.
high-alloy, lowest work-hardening rate These grades are easily weldable and formable
grade for this purpose. The general-purpose and can be given many attractive and useful sur-
alloy 304 is within this group. face finishes, so they are very much general-
• Chromium nickel alloys when the objective purpose alloys. Table 1 lists some typical com-
is high temperature oxidation resistance. positions of the most commonly used lean
This can be enhanced by silicon and rare austenitic alloys. These typical compositions
earths. If the application requires high-tem- vary with end use, raw material cost factors, and
perature strength, carbon, nitrogen, niobium, the preference of a given manufacturer. The
and molybdenum can be added. 302B, 309, compositions of standard alloys are often fine-
310, 347, and various proprietary alloys are tuned to the intended end use. In this table, the
found in this group. word drawing indicates higher nickel for lower
• Chromium, molybdenum, nickel, and nitro- work hardening, while tubing indicates alloys
gen alloys when corrosion resistance is the with higher sulfur to facilitate gas tungsten arc
main objective. Alloys such as silicon and welding (GTAW) penetration. Tensile indicates
copper are added for resistance to specific lower alloy levels to increase the work-harden-
environments. This group includes 316L, ing rate for material that is intended to be used
317L, 904L, and many proprietary grades. in the cold-worked, high-strength condition.
316L is included in its most common tubing end
Wrought alloys generally have cast counter- use chemistry even though it is a corrosion-re-
parts that differ primarily in silicon content. sisting alloy because it is so pervasively used as
Versions that require enhanced machinability a service center sheet item.
have a high content of controlled inclusions, The main difference among the lean
sulfides, or oxysulfides, which improve machin- austenitic alloys lies in their work-hardening
ability at the expense of corrosion resistance. rate: the leaner the alloy, the lower the austenite
Carbon is kept below 0.03% and designated an stability. As unstable alloys are deformed, they
L grade when prolonged heating due to multi- transform from austenite to the much harder
pass welding of heavy section (greater than martensite. This increases the work-hardening
about 2 mm) or when welds requiring a post- rate and enhances ductility since it delays the
weld stress relief are anticipated. onset of necking since greater localized

Table 1 Typical compositions of the most commonly used lean austenitic alloys
Alloy Designation C N Cr Ni Mo Mn Si Other Other Other
201 S20100 0.08 0.07 16.3 4.5 0.2 7.1 0.45 0.001 S 0.03 P 0.2 Cu
201 drawing S220100 0.08 0.07 16.9 5.4 0.02 7.1 0.5 0.001 S 0.30 P 0.6 Cu
201LN S20153 0.02 0.13 16.3 4.5 0.2 7.1 0.45 0.001 S 0.03 P 0.5 Cu
301 tensile S30100 0.08 0.4 16.6 6.8 0.2 1.0 0.45 0.001 S 0.03 P 0.3 Cu
301 drawing S30100 0.08 0.04 17.4 7.4 0.02 1.7 0.45 0.007 S 0.03 P 0.6 Cu
303 S30300 ... ... ... ... ... ... ... ... ... ...
304 S30400 0.05 0.05 18.3 8.1 0.3 1.8 0.45 0.001 S 0.03 P 0.3 Cu
304 drawing S30400 0.05 0.04 18.4 8.6 0.3 1.8 0.45 0.001 S 0.03 P 0.3 Cu
304 extra drawing S30400 0.06 0.04 18.3 9.1 0.3 1.8 0.45 0.001 S 0.030 P 0.4 Cu
304L tubing S30403 0.02 0.09 18.3 8.1 0.3 1.8 0.45 0.013 S 0.030 P 0.4 Ci
305 S30500 0.05 0.02 18.8 12.1 0.2 0.8 0.60 0.001 S 0.02 P 0.2 Cu
321 S32100 0.05 0.01 17.7 9.1 0.03 1.0 0.45 0.001 S 0.03 P 0.4 Ti
316L S31603 0.02 0.0 16.4 10.5 2.1 1.8 0.50 0.010 S 0.03 P 0.4 Cu
Chapter 6: Austenitic Stainless Steels / 73

deformation is more than offset by greater lo- such as occurs when they are sensitized or when
calized strain hardening. solute segregation occurs, as from welding, then
These grades are best viewed as a continuum the equation applies on a microscopic scale.
with a lower boundary at 16%Cr-6%Ni and an Sensitized zones (i.e., the regions near grain
upper boundary at 19%Cr-12%Ni. This repre- boundaries where chromium carbides have pre-
sents the range from minimum to maximum cipitated) will have a much higher tendency to
austenite stability. Since that is the main distinc- transform to martensite. Figures 3(a) and (b)
tion within this grade family, let us examine its show the changes in phase structure as a func-
basis. tion of composition over ranges that encompass
Martensite and Austenite. Stability. The these alloys.
formation of martensite at room temperature Martensite can be present in two different
may be thermodynamically possible, but the forms. The α′-form is the bcc magnetic form,
driving force for its formation may be insuffi- while ε is a nonmagnetic, hcp (hexagonal close-
cient for it to form spontaneously. However, packed) version. The formation of ε versus α′ is
since martensite forms from unstable austenite related to the stacking fault energy of the alloy,
by a diffusionless shear mechanism, it can occur which is given by (Ref 6):
if that shear is provided mechanically by exter-
nal forces. This happens during deformation, Y300SF (mJ m-2) = Y0SF + 1.59Ni – 1.34Mn
and the degree to which it occurs varies with + 0.06Mn2 – 1.75Cr + 0.01Cr2
composition according to (Ref 4): + 15.21Mo – 5.59Si
Md30 (°C) = 551 – 462(%C + %N) –60.69(C + 1.2N)1/2
– 9.2(%Si) – 8.1(%Mn) – 13.7(%Cr) + 26.27(C + 1.2N)
– 29(%Ni + Cu) – 18.5(%Mo) (Cr + Mn + Mo)1/2
– 68(%Nb) – 1.42 (GS – 8) (Eq 1) + 0.61[Ni•(Cr + Mn)]1/2 (Eq 2)

This is the temperature at which 50% of the Epsilon martensite formation is favored in
austenite transforms to martensite with 30% alloys of lower stacking fault energy. The fcc
true strain (Ref 5). It should be noted that even structures deform by slip between (111)
elements that are chromium equivalents in pro- planes. Viewed from these planes, the structure
moting ferrite are austenite stabilizers in that is a series of ABCABC atom arrangements.
they impede martensite formation. This temper- Slip between planes can result in an
ature is the common index of austenite stability. ABCA/CAB structure. This so-called stacking
This regression analysis was generated for ho- fault generates an hcp structure. With lower
mogeneous alloys. If alloys are inhomogeneous, stacking fault energies, these are more readily

Fig. 3 (a) Iron-chromium phase diagram at 8% nickel; (b) iron-nickel phase diagram at 18% chromium
74 / Stainless Steels for Design Engineers

Fig. 4 Variation of martensite formation with temperature and true strain for 304. Source: Ref 7

formed, and ε predominates. The stacking fault and tensile strength, respectively, are reported
can also be viewed as two partial dislocations (Ref 10) to follow the equations:
with the material between them faulted. These
partial dislocations, when generated in abun- YS( MPa) = 15.4[4.4 + 23(%C) + 32(%N)
dance, cannot readily slip past one another + 0.24(%Cr ) + 0.94(%Mo)
and thus pile up, increasing work-hardening + 1.3(%Si ) + 1.2(%V)
rates. + 0.29(%W ) + 2.6(%Nb)
As in carbon and alloy steels, the martensite
+ 1.7(%Ti ) + 0.82(%Al )
transformation can take place simply by cool-
ing, but in the lean austenitic alloys the temper- + 0.16(%Ferrite )
atures are well below ambient. The more stable + 0.46(d −1/1/ 2 )
(Eq 3)
alloys do not transform even with cryogenic
treatment. Figure 4 shows the variation of TS ( MPa) = 15.4[29 + 35(%C) + 55(%N)
martensite formation with temperature and true
+ 2.4(%Si ) + 0.11((%Ni ) + 1.2(%Mo)
strain for 304. Martensite formed in these alloys
is quite stable and does not revert until heated + 5.0(%Nb) + 3.0(%Ti ) + 1.2(%Al )
well above the temperatures (Fig. 5) at which it + 0.14(%Ferrite ) + 0.82(d −1/ 2 ) (Eq 4)
was formed. The carbon levels of austenitic
stainless steels are always relatively low, so In each case, d is the grain diameter in mil-
strain-induced martensite is self-tempering and limeters.
not brittle. Another researcher (Ref 11) gave the rela-
Martensite has been found to form in unstable tionships as:
austenite due to the electrochemically induced
supersaturation by hydrogen (Ref 9). Under YS ( MPa ) = 120 + 210 N + 0.02 + 2 Mn + 2Cr
conditions of cathodic charging, superficial lay- + 14 Mo + 10 Cu + (6.15 − 0.054δ)δ
ers were found to transform to ε under condi- + (7 + 35(N + 0.2))d −1/ 2 (Eq 5)
tions of intense hydrostatic compression. Dur-
ing subsequent outgassing, α′ was found to
form due to reversals in the stress state. Marten- TS = 470 + 600( N + 0.2)
site thus formed is, of course, susceptible to + 14 Mo + 1.5δ + 8d −1/ 2 (Eq 6)
hydrogen embrittlement.
Mechanical Properties. The tensile proper- Again, d is grain diameter in millimeters, and
ties in the annealed state not surprisingly relate δ is percent ferrite. The claimed accuracy for
well to composition. The 0.2% yield strength the latter set of equations is 20 MPa and is said
Chapter 6: Austenitic Stainless Steels / 75

Fig. 6 Variation of impact strength with temperature for (a)

austenitic, (b) duplex, and (c) ferritic stainless steels

form more stable dislocation arrays that break

loose at a higher and distinct yield point.
The tensile properties of austenitic stainless
steels with unstable austenite, that is, those with
Md30 temperatures (Eq 1) near room tempera-
ture, are very strain rate dependent. This is sim-
Fig. 5 Reversion of martensite formed by cold work. Source:
ply due to the influence of adiabatic heating
Ref 8
during testing increasing the stability of the
to apply to both austenitics and duplex stainless austenite. Tests run under constant temperature
steels, but clearly the tensile strength relation- conditions, either by slow strain rates or use of
ship must break down for leaner alloys, such as heat sinks, produce lower tensile strengths.
301, in which tensile strength increases with de- Thus, reported tensile strengths should not be
creasing alloy content because of the effect of taken as an absolute value but a result that can
increasing alloying causing less transformation be significantly changed by changes in testing
to martensite, which inarguably produces higher procedure, even with accepted norms and
tensile strengths in austenitic stainless steels. standards.
Equation 3 must also be favored over Eq 5 in Highly cold-worked austenitic stainless
that it accounts for carbon explicitly. steels are often used for their robust mechanical
One other hardening mechanism is possible properties. Few metallic materials can match
in austenitic stainless steels, and that is precipi- the very high strengths they can achieve. Very
tation hardening. Most precipitation-hardening lean 301 can be cold worked to yield strengths
stainless steels are unstable austenite, which is on the order of 2000 MPa because of its unsta-
transformed to martensite before the precipita- ble austenite transforming to martensite. When
tion hardening takes place. One commercial cold worked to lower degrees, it can provide
alloy, A-286, is entirely austenitic and employs very high strength while keeping impressive
the precipitation within the austenite matrix of ductility.
Ni3 (titanium, aluminum) for strengthening. Austenitic stainless steels have exceptional
This is dealt with in a separate section. toughness. The ambient temperature impact
Austenitic stainless steels do not have a clear strength of austenitic stainless steels is quite
yield point but can begin to deform at as little as high. This is not surprising in view of their high
40% of the yield strength. As a rule of thumb, tensile strengths and high elongations. What is
behavior at less than half the yield strength is most remarkable is the absence of a transition
considered fully elastic and stresses below two- temperature, which characterizes ferritic and
thirds of the yield strength produce negligible martensitic materials. Figure 6 shows impact
plastic deformation. This quasi-elastic behavior strength of the various stainless steel types ver-
is a consequence of the many active slip sys- sus temperature. This again is due to the multi-
tems in the fcc structure. Even highly cold- plicity of slip systems in the fcc structure and
worked material exhibits this phenomenon, al- the fact that they do not require thermal activa-
though stress-relieving cold-worked material tion. This makes the austenitic stainless steels,
will cause dislocations to “lock in place” and especially the 200 series, the optimal cryogenic
76 / Stainless Steels for Design Engineers

material, surpassing the 9% nickel martensitic tion in austenite, and diffusion rates are suffi-
steels in cost, toughness, and, of course, corro- cient for carbon and chromium to segregate into
sion resistance. precipitates. The solubility of carbon in austen-
Precipitation of Carbides and Nitrides. ite is over 0.4% at solidification but decreases
Carbon is normally considered as an undesir- greatly with decreasing temperature. The solu-
able impurity in austenitic stainless steel. While bility is given by (Ref 12):
it stabilizes the austenite structure, it has a great
thermodynamic affinity for chromium. Because 6272
log (C ppm ) = 7771 − (Eq 7)
of this affinity, chromium carbides, M23C6, form T (°K )
whenever carbon reaches levels of supersatura-
The equilibrium diagram for carbon in a basic
18%Cr10%Ni alloy is shown in Fig. 7. At room
temperature, very little carbon is soluble in
austenite; even the 0.03% of L grades is mostly
in a supersaturated solution. The absence of car-
bides in austenitic stainless is due to the slow
diffusion of carbon and the even slower diffu-
sion of chromium in austenite. At a carbon level
of 0.06%, which is found in most 304, supersat-
uration is reached below about 850 °C. Below
this temperature, supersaturation increases ex-
ponentially, while diffusion decreases exponen-
tially. This results in precipitation rates that vary
with temperature and carbon level as shown in
Fig. 8. At these temperatures, grain boundary
diffusion is much more rapid than bulk diffu-
sion, and grain boundaries provide excellent nu-
cleation sites, so precipitation occurs along
grain boundaries. And, because carbon diffuses
several orders of magnitude more rapidly than
chromium, carbon diffuses to and combines
Fig. 7 Carbon solubility in 18–10 austenitic stainless. Source: with chromium essentially in situ, depleting the
Ref 13 grain boundaries of chromium in solution.

Fig. 8 The precipitation rates for Cr23C6 as a function of carbon content

Chapter 6: Austenitic Stainless Steels / 77

Fig. 10 Variation of carbide precipitation locus with time.

Source: Ref 16

Much longer term heat treatment is required to

eliminate these depleted zones by rehomoge-
nization of slowly diffusing chromium than the
short time required to form them. This is very
evident for carbides, but also true for oxides.
Underneath chromium-rich oxide scales is a
layer depleted in chromium and lower in corro-
sion resistance. This is why not only scale from
welding must be removed, but also the underly-
ing chromium-depleted zone.
Other precipitation processes that give rise to
chromium depletion are α and χ and the solid-
state precipitation of oxides, nitrides, and sul-
fides. Chromium precipitates that form in the
liquid alloy do not cause depletion of chromium
Fig. 9 Depletion of chromium from the austenite near grain
locally because no chromium gradients are set
boundaries due to chromium carbide precipitation.
Source: Ref 14 up around them during precipitation as diffu-
sion in the liquid is very rapid. Thus, primary
Figure 9 shows that the local chromium de- carbides, oxides, and sulfides are not per se
pletion is such that the chromium level can be- harmful to corrosion resistance. But, if the same
come low enough that it has not even enough to compounds form and grow in the solid state,
be stainless and certainly much lower corrosion chromium depletion occurs (Ref 15).
resistance than the surrounding area. This zone, Alloying elements can have a major influence
because it is lower in chromium, also has very on carbide precipitation by their influence on the
unstable austenite and is quite prone to marten- solubility of carbon in austenite. Molybdenum
site formation. Figure 10 shows how the locus and nickel accelerate the precipitation by dimin-
of precipitation changes with time and tempera- ishing the solubility of carbon. Chromium and
ture. Carbon relatively far from grain bound- nitrogen increase the solubility of carbon and
aries in the interior of grains remains in super- thus retard and diminish precipitation. Nitrogen
saturation until much longer times and much is especially useful in this regard (Fig. 11).
greater supersaturation since bulk diffusion is Increasing austenite grain size accelerates
required for the nucleation and growth of these precipitation, as does cold work, especially in
precipitates. the interior of grains, where diffusion is en-
The key observation is that any solid-state hanced by increased defect density.
precipitation of a chromium-rich precipitate Nitrogen is much more soluble than carbon
will necessarily cause local chromium depletion and does not give rise to sensitization phenom-
and a resulting loss of corrosion resistance. ena as does carbon even though Cr2N can be a
78 / Stainless Steels for Design Engineers

Fig. 11 Delay in carbide precipitation induced by nitrogen

level. Source: Ref 17

stable phase when the solubility limit is ex-

ceeded. The solubility is over 0.15% in austen-
ite, so its precipitation seldom has the possibil-
ity of occurring, but it does become an issue in Fig. 12 Variation of hardness with depth and therefore car-
ferritic stainless steels in this regard, for which bon content in colossal supersaturation
solubility is much lower. Manganese and
chromium increase the solubility of nitrogen in to ignore the titanium-consuming capacity of
austenite. oxygen and sulfur unless they have been mini-
Stabilization. Before carbon was easily low- mized by refining, which can be done quite
ered to harmless levels, it was found that adding readily.
more powerful carbide formers than chromium Even if sufficient titanium or niobium is pres-
could preclude the precipitation of chromium ent to combine with all carbon, kinetic consid-
carbides. Titanium and niobium are the most erations may result in that not occurring. High
useful elements in this regard. They form car- temperatures, such as encountered in welding,
bides with solubility that follows the following dissociate carbides. If quenched from this state,
equation type: carbon can be free to form Cr23C6 if it is re-
heated to temperatures above 500 °C.
log [ M] [ X] = + A − H (Eq 8) Carbon has always been considered totally
RT undesirable from a corrosion point of view be-
cause of its tendency to form chromium car-
For titanium carbide and niobium carbide, the bides. Recently, however, new processes have
respective solubilities are: been developed to supersaturate carbon in
6780 austenite below the temperatures at which it has
log [Ti ] [C] = 2.97 − (Eq 9) sufficient mobility to form carbides. This so-
called colossally supersaturated austenite re-
sults in very high hardness (Fig. 12) and corro-
9350 sion resistance over limited depths. From this,
log [ Nb] [C] = 4.55 − (Eq 10)
T however, we can see that carbon, like nitrogen,
is actually beneficial to corrosion resistance in
Oxides and sulfides are more energetically fa- solid solution, although this is not observed at
vorable than are carbides and nitrides of these normal concentrations. It is possible to see that
metals. Thus, any additions made to form car- if it could be kept in solution it would be appro-
bides must be sufficient to account for the prior priate to give it a factor of around 10 in the pit-
formation of these compounds. Nitrogen also ting resistance equivalent number (PREN)
competes with carbon for available titanium or equation:
niobium. Thus, for successful gettering of all
carbon, there must be sufficient titanium or nio- PREN = %Cr + 3.3(%Mo)
bium to combine stoichiometrically with all
these species present. + 30(%N) + 10(%C) (Eq 11)
This requires in rough terms that titanium ex-
ceed four times the carbon plus nitrogen, or that This is consistent with the similar thermody-
niobium exceed eight times, after accounting namic interaction coefficients that carbon and
for the oxygen and sulfur. It would be a mistake nitrogen share with regard to chromium.
Chapter 6: Austenitic Stainless Steels / 79

High-Temperature Alloys

The austenitic stainless steels can have an

exceptional combination of strength and corro-
sion resistance at temperatures above 500 °C.
They are often called on to resist attack by
oxygen, sulfur, carburizing, nitriding, halogens,
and molten salts. Austenitic stainless steels are
the most creep resistant of the stainless steels.
Alloying with carbon, nitrogen, and niobium
produces the greatest strength at elevated tem-
peratures. Refer to the properties database for
comparisons among the grades. This discus-
sion concentrates on their resistance to high-
temperature environments, which is their salient
Oxidation Resistance. Resistance to oxida-
tion comes from the protective Cr2O3 scale that
forms on the surface of the material. Above
about 18% chromium, a continuous scale forms.
The scale acts as a barrier to oxygen and greatly
slows further oxidation of metal below the
scale. Below the composition at which com-
plete Cr2O3 coverage occurs, the film will also Fig. 13 Variation of parabolic oxidation rate with chromium
level and temperature. Source: Ref 18
contain the less-protective spinel FeCr2O4. The
Cr2O3 scale is more protective because it better the less-protective FeCr2O4, and the scale
restricts the diffusion of oxygen to the interface growth rate will increase beyond the parabolic
between the scale and the base metal, which is relationship, leading to breakaway oxidation.
where the oxidation reaction occurs. As the The breakaway temperature increases with in-
oxide grows, the path to the interface lengthens, creasing chromium level.
and the rate of oxidation slows. This generates The austenitic alloys benefit over the ferritic
the parabolic-type oxide growth rate that char- alloys from the presence of nickel. For a given
acterizes these alloys. The rate of oxide growth chromium level, oxidation rates decrease with
is expressed simply as: increasing nickel content. Figures 14, 15, and
16 display this relationship. The optimal range
R = kt (Eq 12) for the iron base stainless steels, shown in Fig.
14, is reached by the commercial alloy 310,
in which the rate is in units of mass gained per with 25Cr-20Ni composition.
unit of surface area and time. This rate is a Alloying can alter the oxidation-resisting per-
strong function primarily of chromium level, as formance of the austenitic stainless steels. Some
can be seen in Fig. 13. elements form more protective oxide layers
The rate increases exponentially with temper- than Cr2O3. Aluminum and silicon are most use-
ature since diffusion governs this phenomenon. ful in this regard. Aluminum forms a layer of
The rate drops dramatically as chromium Al2O3 that is more restrictive to oxygen diffu-
reaches the concentration necessary to generate sion than is Cr2O3, as does silicon through the
the protective Cr2O3 layer. Above this suffi- formation of SiO2. The alloys 302B, 153MA,
ciency level, further increases in chromium are and 253MA all use elevated silicon levels. Alu-
not as beneficial; they mainly provide a reser- minum’s strong ferrite-promoting tendency re-
voir of chromium to re-form the Cr2O3. stricts its utility in austenitic grades, however.
As long as chromium can diffuse to the inter- While the gains from using under 3% silicon
face at a sufficient rate to satisfy the incoming are impressive, rare earths can yield even
flux of oxygen, the parabolic rate is maintained. greater benefits from mere trace additions.
If there is insufficient chromium flux, then the 153MA (UNS S30415) is a variation on 304
oxygen penetrates beyond the interface and using silicon and cerium. Cerium appears to act
forms Cr2O3 in situ. The oxide will change to at the metal-scale interface such that the oxides
80 / Stainless Steels for Design Engineers

Fig. 14 Influence of nickel on oxidation of iron-chromium

alloys. Source: Ref 19 Fig. 16 Corrosion rates for various stainless steels and nickel
base alloys. Source: Ref 20

Fig. 15 Isooxidation curves. Source: Ref 20

formed are thinner, tougher, more adherent, and ritic alloys, they stress their scale more during
more protective, although there is no consensus thermal cycling. This can fracture the scale,
on the mechanism. Figure 17 shows the im- causing spalling and rapid subsequent oxidation
provement quantitatively (Ref 21). of the underlying metal. This serious perform-
Because austenitic stainless steels have a ance flaw also is remedied by rare earths, as
greater thermal expansion coefficient than fer- shown in Fig. 18 (Ref 21).
Chapter 6: Austenitic Stainless Steels / 81

Fig. 17 Comparison of rare earth-alloyed stainless alloys to conventional stainless alloys, 4833 = 309S, 4845 = 310S.
Source: Ref 21

The formation of an oxide on stainless steel

should be understood to imply de facto the de-
pletion of chromium from the underlying
metal surface. Whether the scale is formed in
service, during heat treating, or during welding
the surface, once the oxide is removed, there
will be less chromium than the bulk alloy,
often by a very significant amount, and there-
fore the corrosion resistant will be less. This is
why welds must have not only their heat tint
removed, but also the underlying metal which
is depleted in chromium, to a depth on the
order of 10 μ (Ref 24).
Fig. 18 310S (4845) versus rare earth-modified 253MA in Microstructure can also affect oxidation re-
cyclic oxidation. Source: Ref 21
sistance. As a generalization, it can be said that
changes that promote the diffusion of chromium
While the mechanism by which rare earths assist in the formation of a protective scale and
make the scale tougher and more adherent are improve oxidation resistance. Thus, cold work
vague, their effect in making austenitic alloys and finer grain size are positive factors via their
much better at resisting high-temperature oxida- enhancement of chromium diffusion.
tion, especially cyclic oxidation, are undeniable. At the very highest temperatures, Cr2O3 can be
Alloying elements can also be detrimental to further oxidized to CrO3, which has significant
oxidation resistance. Manganese, an even more vapor pressure above about 1100 °C. The com-
potent oxide former than chromium, forms a positions of some of the main high-temperature
manganese-chromium spinel that is less protec- austenitic alloys mentioned here are given in
tive than the Cr2O3. Molybdenum and tungsten, Table 2.
which are refractory metals and are beneficial to Other Environments. The most common ad-
aqueous corrosion resistance, form volatile, low- dition species that aggravates high-temperature
melting oxides (MoO3 and WO3) that promote oxidation is water vapor. At 10%, water vapor
catastrophic oxidation (Ref 22, 23). Vanadium will increase oxidation by a factor of ten. It acts
also forms an oxide, V2O5, which melts at 660 by increasing the porosity of the oxide scale and
°C and can also cause catastrophic oxidation. by promoting formation of the volatile CrO2
82 / Stainless Steels for Design Engineers

(OH)2 species. As a rule of thumb, maximum In lean austenitic alloys used in high-temper-
service temperatures should be reduced by 50 to ature, 600 to 1000 °C service, formation times
100 °C in the presence of steam. are relatively long, on the order of 100 h or
Halogens can attack oxide scales and cause more. In richer alloys, such as 310, times can be
their degradation or volatilization. as short as several hours. If the temperature at
Carburization and nitriding are best prevented which the alloy is to be used is one in this tem-
by an oxide layer that forms at very low oxygen perature range, then some σ is a foregone con-
partial pressures as chromium and silicon con- clusion, and while σ will have little detrimental
tents are increased. Austenitic alloys have no effect on short-term properties at these tempera-
advantage over ferritic in this regard. tures, long-term properties such as creep, stress
Intermetallic Phases. Transition elements rupture, and especially rupture ductility are de-
may combine to form intermetallic phases in graded by σ. For alloys, σ is an even greater
which the formula can vary from B4A to A4B. concern as these are prone to its formation and
Table 3 lists the most common secondary can inadvertently form some during processing.
phases encountered in austenitic stainless steels, If such alloys are intended for use near room
i.e., apart from austenite and ferrite. Sigma for- temperature, then their toughness will be seri-
mation is retarded by nitrogen, so alloys such as ously reduced by the brittle σ, which will form
153MA are less prone to it. Lower chromium first at triple points and then throughout grain
and higher nickel are beneficial. Silicon and boundaries. Because of this morphology, just a
aluminum are detrimental, as is molybdenum. few percent intermetallic phase can cause
The most relevant is σ. It can contain as little toughness to decrease by an order of magnitude.
as four iron to one chromium or molybdenum in High-Temperature Mechanical Properties.
a tetragonal structure. Thus, it can exist in many Above about 500 °C, yield strength is less ap-
conventional austenitic alloys. Other relevant propriate than creep strength in assessing the
phases are χ and Laves. The greatest risk from adequacy of an austenitic stainless steel for
these phases is the loss of room temperature structural purposes. The resistance of a material
toughness, followed by some loss of corrosion to creep is generally measured by the creep rup-
resistance. ture strength, which is the stress that causes

Table 2 Notable high-temperature austenitic alloys

Alloy Designation C N Cr Ni Mo Mn Si Other Max temp, °C
302B S30215 0.15 0.07 17.8 8.1 ... 1.8 2.5 ... 950
304H S30409 0.08 0.08 18.8 8.1 ... 1.8 0.50 ... 820
321H S32109 0.06 0.03 17.8 9.1 ... 1.8 0.50 0.6 Ti 820
153MA S30415 0.05 0.15 18.5 9.5 ... 0.6 1.3 0.05 Ce 1000
309S S30909 0.08 0.07 23.0 12.2 ... 1.7 0.50 ... 1040
309Si DIN 1.4828 0.08 0.07 19.8 11.1 ... 1.8 2.0 ... 1040
253MA S30815 0.08 0.17 21.0 10.5 ... 0.6 1.5 1.0 Al, 0.05 Ce 1100
310S S31008 0.05 0.03 24.6 19.2 ... 1.6 0.60 ... 1090
353MA S35315 0.05 0.15 25.0 35.0 ... 1.5 ... 0.05 Ce 1200+
330 S33000 0.06 ... 18.0 35.0 ... 1.7 0.90 ... 1200
332Mo S35125 0.04 0.04 21.0 34.5 2.4 1.1 0.40 0.40 Nb 1200

Table 3 Secondary phases in austenitic stainless steel

Precipitate Structure Parameter, (Å) Composition
NbC fcc(a) a = 4.47 NbC
NbN fcc a = 4.40 NbN
TiC fcc a = 4.33 TiC
TiN fcc a = 4.24 TiN
Z-phase Tetragonal a = 3.037, c = 7.391 CrNbN
M23C6 fcc a = 10.57–10.68 Cr16Fe5 Mo2C (e.g.)
M6C Diamond cubic a = 10.62–11.28 (FeCr)21Mo3 C; Fe3Nb3C; M5SiC
σ-phase Tetragonal a = 8.80, c = 4.54 Fe,Ni,Cr,Mo
Laves phase Hexagonal a = 4.73, c = 7.72 Fe2Mo, Fe2Nb
χ-phase bcc(b) a = 8.807–8.878 Fe36Cr12 Mo10
G-phase fcc a = 11.2 Ni16Nb6 Si7, Ni16Ti6 Si7
(a) fcc, face-centered cubic; (b) body-centered cubic
Chapter 6: Austenitic Stainless Steels / 83

rupture after 10,000 or 100,000 h. If deforma- preferred. Substitutional elements have limited
tion is a greater concern, however, the creep de- effect, but interstitial solid solution elements,
formation strength, that is, the stress that results such as carbon and nitrogen, are quite useful.
in a strain of 1% after 10,000 or 100,000 h, can Nitrogen is the better addition in this regard,
be used as a basis for design. plus it has the collateral benefit of strongly
Cold work and precipitates tend to be ineffec- retarding intermetallic phase precipitation.
tive strengtheners at temperatures that produce Figures 19 to 22 compare mechanical properties
solution annealing and precipitate coarsening of the major high-temperature austenitic alloys
(overaging). Thus, solid solution hardening is (Ref 25).

Fig. 19 Charpy V toughness after 200 hr aging

Fig. 20 Relative 100,000-h creep strength

84 / Stainless Steels for Design Engineers

Fig. 21 100,000-h creep rupture strength

Fig. 22 High-temperature, short-time yield strength

Corrosion-Resistant Austenitic Alloys temperature, thus giving the common perception

that they are superior in corrosion resistance.
Stainless steels are almost always chosen at The main advantage austenitic alloys have is
least in part for their corrosion resistance. In their ability to utilize the powerful and inexpen-
normal atmospheric conditions, alloys with sive alloying element nitrogen. That is the key
more than 10.5% chromium do not rust. aspect of the more modern austenitic stainless
Austenitic alloys require higher levels than steels that have come into use in the last two
this to stabilize the austenitic structure at room decades.
Chapter 6: Austenitic Stainless Steels / 85

The ion that is most aggressive against stain- The main factors in the resistance of austenitic
less steel is one of the most pervasive in our alloys to pitting attack are generally given by:
environment. The chloride ion is found in
abundance over the entire earth. It is, of PREN = %Cr + 3.3(%Mo) + 30(%N) (Eq 13)
course, in seawater, but also in the rain, on
roads, in food, and even in our bodies. Chlo- Pitting resistance is measured by ASTM G 48
rides destabilize the passive film. If the condi- (practice C) in which the lowest temperature at
tions of chloride concentration, temperature, which pitting occurs in a 6% FeCl3 is measured.
and acidity are sufficiently aggressive to break This is the critical pitting temperature, CPT. The
down the film, then active corrosion ensues. If relationship between PREN and CPT is shown
this is highly localized because of a local in Fig. 23 (Ref 26).
weakness in the passive film, then pitting oc- The function of chromium in the passive film
curs. Such a pit may be unstable and repassi- is intuitively clear. As the chromium content of
vate, or it may grow without limit. Other an alloy increases, the ready reservoir of
halides have the same effect, but they are less chromium to form the chromium-rich layer is fa-
ubiquitous. cilitated. The roles of molybdenum and nitrogen
Because of the specific virulence of the chlo- are subtler and are still subject to controversy. It
ride ion and because of its universal presence, is the subject of much research, which has been
corrosion-resistant austenitic stainless steels all summarized in reviews. The obvious paradox is
look like they were designed to resist chloride- how can elements that are not active in the pas-
pitting attack. Pitting in stainless steels is in sive film be so effective in maintaining its in-
most instances the threshold level of corrosion. tegrity. We do know that the essential chromium
Crevice corrosion is, however, more severe and in the matrix of stainless steels is quite reactive
usually design limiting vis-à-vis corrosion. and will form compounds with carbon, oxygen,
Crevices can exist not only in deliberate joints sulfur, and other transition elements. When it
but also under environmental deposits, paint does, it is no longer effective as a passive film
films, weld splatter, etc. There are other envi- former. The regions from which the chromium
ronments in which the resistance follows differ- diffused to form the chromium-rich phase will
ent rules, such as oxidizing acids, bases, and or- be poor in chromium unless subjected to a
ganic acids, but these are best regarded as lengthy homogenizing anneal. Most theories of
exceptions. pitting founder at the start because they assume

Fig. 23 Critical pitting temperature versus pitting resistance equivalent number (PREN); SUS 329J4L = S31260, YUS 270 = S31254.
Source: Ref 26
86 / Stainless Steels for Design Engineers

a homogeneous passive film, which is an impos- The discovery that nitrogen was beneficial
sible goal in reality. The chapter on corrosion against corrosion permitted a breakthrough in
deals with this topic in more depth. alloy development by the 1980s. Nitrogen was
The passive layer is extremely thin compared increased to around 0.20% from nominal levels
to oxide layers. It is on the order of 1 to 10 nm of 0.05%. This was found to increase PREN by
thick. Its formation does not cause chromium another five units, but more importantly, also
depletion beneath it, as oxide layers do. As the suppressed σ formation to times that permitted
alloy content of chromium and molybdenum in- thicker sections to be welded without embrittle-
crease, the film is thinner, and the current den- ment. Research into the thermodynamics of
sity required to form the film is correspondingly nitrogen in austenite showed that manganese in-
reduced. creased the solubility of nitrogen appreciably.
The corrosion-resistant austenitic stainless This permitted even higher levels of total alloy-
steel grades range from 316 to the various high- ing to be achieved. This was exploited in the
molybdenum, high-nitrogen alloys commercial- alloys UNS S34565 and S32654, which contain
ized in the last ten years, the most notable of 3 to 6% manganese and about 0.50% nitrogen.
which are listed in Table 4 with their typical The PRENs of these alloys are over 50, which
analyses. gives them a critical pitting temperature around
Early grades were based on alloying with 100 °C. Table 5 lists the performance of the var-
chromium and molybdenum with sufficient ious popular corrosion-resistant austenitic stain-
nickel added to preserve the austenitic structure. less steels.
Each of these elements facilitates the formation The advances are quite appreciable and made
of the passive film and reduces the corrosion stainless steel a viable material for many appli-
rate in the active state. Further experimentation cations for which previously nickel base or tita-
showed that molybdenum was not beneficial nium alloys had been required.
under highly oxidizing conditions, but that sili- The obvious question in view of the success
con was helpful under such conditions. Copper of the use of high manganese levels in conjunc-
was beneficial against sulfuric acid. tion with high nitrogen levels in the most highly
Alloy development came in stages. First, 317 alloyed austenitic stainless steels is when this
was the most significant corrosion-resistant approach will be used for the medium-alloyed
alloy. Then, more chromium and molybdenum austenitics to make alloys superior to 316, 317,
were added, and the class of alloys known col- and 904 without the high levels of nickel and
lectively as the 6%Mo alloys was commercial- molybdenum that render these alloys so expen-
ized. Allegheny’s AL-6X™ is representative of sive. It does not take much imagination to envi-
this group. With PRENs of around 40, these al- sion alloys equal to 904L in PREN with less
loys were resistant to seawater at ambient but nickel and molybdenum than 316L that would
not at elevated temperatures. This left a great be almost totally resistant to intermetallic phase
deal wanting in corrosion resistance. These al- precipitation and have much greater resistance
loys were also very difficult to process, at least to SCC because of higher austenite stability.
in part because they rapidly formed brittle grain The same case could be made for a 317-equiva-
boundary σ-phases in as little as several minutes lent alloy in corrosion resistance with less than
at some temperatures. This limited chromium 7% nickel and 0.5% molybdenum, essentially a
and molybdenum levels to a total of about 30%. 301 in alloy cost. In view of these trends in

Table 4 Typical compositions of corrosion-resistant austenitic stainless steels

Alloy Designation C N Cr Ni Mo Mn Si Other Other
316L S31603 0.02 0.03 16.4 10.5 2.1 1.8 0.5 ... ...
316Ti S31635 0.02 0.03 16.4 10.5 2.1 1.8 0.5 0.40 ...
317L S31703 0.02 0.06 18.4 12.5 3.1 1.7 0.5 ... ...
317LM S31725 0.02 0.06 18.4 13.7 4.1 1.7 0.5 ... ...
904L N80904 0.02 0.06 19.5 24.0 4.1 1.7 0.5 1.3 Cu ...
JS700 N08700 0.02 0.06 19.5 25.0 4.4 1.7 0.5 0.4 Cu 0.3 Nb
254SMO S31254 0.02 0.20 20.0 18.0 6.1 0.8 0.4 0.8 Cu ...
4565 S34565 0.01 0.45 24.0 18.0 4.5 6.0 ... ... ...
654SMO S32654 0.01 0.50 24.0 22.0 7.2 3.0 ... 0.5 Cu ...
AL6-XN N08367 0.02 0.22 20.5 24.0 6.2 0.4 0.4 0.2 Cu ...
Al6-XN Plus N08367 0.02 0.24 21.8 25.3 6.7 0.3 0.4 0.2 Cu ...
Chapter 6: Austenitic Stainless Steels / 87

alloy development, the future use of 316 and thermomechanical history. As a rule of thumb,
317 should be numbered. There is no justifica- the environment to initiate SCC must be suffi-
tion for the use of scarce and expensive re- ciently severe to cause localized corrosive at-
sources such as nickel and molybdenum when tack. The most dangerous situation is one in
cheap, abundant replacements like manganese which the expectation of pitting is marginal.
and nitrogen are available. The use of 316 as a The mechanism of SCC has been a subject of
standard alloy should in the future be eroded by intense academic controversy for many years.
more cost-effective alloys such as the lean du- Because of this, much of the literature has fo-
plex alloys like 2003. cused on arguing a case rather than clarifying
The same environments that cause pitting the phenomenon. What can be said about SCC
corrosion also cause crevice corrosion. A in austenitic stainless steels with consensus?
crevice is a volume in and out of which diffu-
• Risk of SCC is low at room temperature and
sion is restricted to a degree that corrosion prod-
increases exponentially with temperature.
ucts accumulate and cause the contained envi- • SCC is preceded by localized corrosive at-
ronment to become increasingly aggressive in
tack, which has an incubation time, and then
pH and [Cl–]. Conditions that are below the
threshold for pitting can cause crevice corro- proceeds in a discontinuous manner.
• Fracture may be transgranular, intergranular,
sion. The critical temperature for crevice corro-
sion is also measured in 6% FeCl3 (ASTM G-48 or both. It is almost entirely lacking in plas-
B or D). It is the lowest temperature at which tic deformation with little, if any, metal loss.
crevice corrosion occurs. This temperature, the • Alloying or treatments that delay localized
CCT (critical crevice temperature), is lower attack or stabilize austenite can delay SCC
than the CPT. GTAW a wrought alloy also low- up to the point of virtual immunity.
ers the CPT to about the level of the CCT. The • SCC is aggravated by increased chloride
reason for this lowering of resistance to local- concentration and acidity, but also exists in
ized attack has been thought to be related to caustic environments.
alloy depletion caused either by dendrite coring • Stress must exceed a threshold for a given
during solidification or chromium depletion set of conditions for SCC to occur.
around inclusions. The relation to crevices • Anodic or cathodic polarization may prevent
would thus seem to be that surfaces contain nu- SCC under conditions at which it would oth-
merous flaws with respect to corrosion resist- erwise be expected, or it may cause SCC in
ance, which, while not capable of sustaining pit- environments in which it would not other-
ting, can in a crevice dissolve and alter the wise occur.
environment sufficiently that the new harsher Austenitic stainless steels are not alone in their
environment can generally destabilize the pas- susceptibility to SCC. All stainless steels suffer
sive film and proceed autocatalytically. from SCC under some set of conditions of envi-
Stress corrosion cracking is the bane of ronment and material thermomechanical history.
austenitic stainless steels. SCC occurs when The key is to choose an alloy that is resistant
there is both a tensile stress of a sufficient mag- under the conditions of use. To a first approxima-
nitude and a sufficiently aggressive environ- tion, this means using an alloy that will not pit
ment. The threshold stress varies with alloy and under the conditions of use, then designing its

Table 5 Corrosion resistance ratings of various austenitic stainless steels, using 30 factor for
Alloy Designation PREN(a) CPT oC CCT oC
316L S31603 24 15 –3
316Ti S31635 23 15 –3
317L S31703 30 25 0
317LMN S31726 34 30 4
904L N80904 35 40 15
JS700 N08700 36 43 15
254SMO S31254 46 73 38
4565 S34565 53 90 50
654SMO S32654 63 105 75
AL6-XN N08367 50 78 43
AL-6XN Plus 50 min 95 60
(a) PREN, pitting resistance equivalent number.
88 / Stainless Steels for Design Engineers

use to be below the threshold stress completes a Nitric acid is not particularly aggressive
sound design approach if residual stresses can be against stainless steels. Resistance to it is pro-
accurately known. Otherwise, assuming that the portional to chromium content. So, attack,
metal will have residual stresses equal to 100% should it take place, is preferably at grain
of the yield strength is the prudent engineering boundaries, where segregation of elements such
approach. Figure 24 shows threshold stress for a as carbon, phosphorus, and silicon can lower
number of alloys. chromium locally. These elements are kept low
Special Corrosive Environments. Knowl- for nitric acid service. Pitting is not a risk. Stan-
edge of the ability of the various stainless steels dard usage is:
to resist specific environments is essential to the
design process. This information is extensive • Below 50% concentrations and below 100
since it must correlate many environments and °C, 304L and 17% Cr ferritics are used.
temperatures for many materials. Hence, refer • Around the 65% aziotrope, 310 is most re-
to the publications of organizations such as the sistant, especially a low carbon version, but
National Association of Corrosion Engineers 304 NAG with low carbon, phosphorus, and
(NACE) or to the Web sites of companies such silicon is more often used.
as Allegheny Ludlum or Outukumpu, where • For 98% solutions or for lower concentra-
such information is available freely. The more tions that contain other stronger oxidizers,
reputable producers will give assistance on spe- alloys with 4% Si, 18% Cr, and 15% Ni or
cific questions. Engineering forums on the In- 5% Si, 17% Cr, and 17% Ni (UNS S30600
ternet, such as, should also and S30601) have been developed.
be considered a resource. The following discus- Hydrochloric acid, not surprisingly, is quite
sion presents just the principles of the resistance aggressive against stainless steel. It is very ef-
of austenitic stainless steels to specific, more fective in destabilizing the passive film. Thus,
common environments. resistance to hydrochloric acid is simply an ex-
Sulfuric acid is common, aggressive, and must treme case of resistance to pitting in chlorides
be contained. Figure 25 shows the isocorrosion with resistance given by Eq 12. Only the most
curves for several alloys in pure sulfuric acid. highly alloyed austenitic alloys, such as AL-
Alloy 20, 904L, and alloy 825 were devel- 6XN®, should be considered and then under
oped specifically for sulfuric acid service. Each conditions that are tolerable, such as those
contains several percent copper that, while not shown in Fig. 26.
beneficial against pitting, concentrates in the Strong bases such as NaOH and KOH are not
passive film and diminishes general corrosion. especially aggressive against stainless. The 17%
Molybdenum and tungsten are also very benefi- chromium alloys can be used up to 50 °C, while
cial for resistance to sulfuric acid. Phosphoric 304L can be used to 90 °C. As is the case with
acid is similar to sulfuric acid in its effect on nitric acid, chromium and nickel are beneficial,
austenitics but somewhat less aggressive. while molybdenum is counterproductive. The
25% chromium alloys such as 310 or an equal

Fig. 24 Threshold stress for stress corrosion cracking (SCC)

for various alloys. Source: Ref 27 Fig. 25 Isocorrosion in pure sulfuric acid. Source: Ref 28
Chapter 6: Austenitic Stainless Steels / 89

contributing causes. Powder injection-molded

stainless components often have porosity that is
generally spherical. When exposed to the sur-
face, such pores act as crevices and lower the
pitting potential also. All of these factors are op-
erative and can act in unison.
Very fine abrasive polishing causes little
residual stress and has very minimal crevice
creation. Thus, mirror-type polished finishes do
not degrade corrosion resistance, but they do
not enhance corrosion resistance as does elec-
tropolished mirror finishes, which remove
chromium-depleted sites, which can initiate
Fig. 26 Resistance to hydrochloric acid. Source: Ref 28 pitting.

chromium duplex can be used to 150 °C, above REFERENCES

which temperature nickel base alloys are re-
quired. High-chromium ferritic stainless steels 1. A.L. Schaeffler, Constitution Diagram for
are also very good choices. Stainless Steel Weld Metal, Met. Prog.,
Organic acids are generally less aggressive Vol 56, Nov 1949, p 680–688
against stainless than are mineral acids since 2. W.T. Delong, A Modified Phases Diagram
they are less dissociated in solution. They be- for Stainless Steel Weld Metals, Met. Prog.,
come hazardous when they contain chloride Vol 77, Feb 1960, p 98
ions, at high temperatures, or when they disso- 3. Design Guidelines for Selection and Use of
ciate strongly, such as with formic acid. Stainless Steel, SSINA,1998, p 3
Because of the large number of organic com- 4. K.-J. Blom, “Press Formability of
pounds that may be considered, refer to the var- Stainless Steels,” paper presented at Stain-
ious corrosion tables. less steels ‘77
Surface Finish. The corrosion resistance of 5. F.B. Pickering, “Physical Metallurgical De-
austenitic stainless steels is quite dependent on velopments in Stainless Steel,” paper pre-
surface condition, as are other stainless steels. sented at Stainless Steel ‘84, Goteborg
Treatments that enhance surface concentrations 6. Q.-X. Dai et al., Chin. Phys., Vol 11, 2002, p
of beneficial elements or remove detrimental 596–600, doi:10.1088/1009-1963/11/6/315
constituents can greatly alter performance. 7. Aciers Inoxidables, Les Editions de
Oxide formation depletes surface chromium, so Physique, Les Ulis, Paris, 1993, p 564
strong pickling or electropolishing of the 8. Aciers Inoxidables, Les Editions de
descaled surface is especially important. Studies Physique, Les Ulis, Paris, 1993, p 565
have shown chromium depletion of a maximum 9. P. Marshall, Austenitic Stainless Steels, Mi-
of 6% extending 10 μ before reaching bulk crostructure and Mechanical Properties,
chromium levels. This is equivalent to the de- Elsevier, 1984
pletion seen in sensitization. The increase in at- 10. Aciers Inoxidables, Les Editions de
tack rate from this depletion is huge. A 1000- Physique, Les Ulis, Paris, 1993, p 579
fold increase in weight loss in the ASTM G 48 11. H. Nordberg, Mechanical Properties of
B test has been seen by a superficial loss of 6% Austenitic and Duplex Stainless Steels, In-
chromium. novation in Stainless Steels ‘93 (Firenze),
Likewise, surface abrasion, especially coarse 1993, p 2.217
abrasion, has a major detrimental effect. The 12. Aciers Inoxidables, Les Editions de
120 grit #3 finish often seen on stainless reduces Physique, Les Ulis, Paris, 1993, p 566
pitting resistance by as much as the equivalent 13. S.J. Rosenberg and C.R. Irish, J Res. Nat.
of 5 PREN, that is, equal to a reduction in Bar. Stand., Vol 48, 1952, p 40
chromium content of 5%. Rolled finishes are 14. Aciers Inoxidables, Les Editions de
much preferred. The mechanism for this has not Physique, Les Ulis, Paris, 1993, p 410
been clearly established; exposure of MnS in- 15. M. McGuire, “A Diffusion Model for the
clusions, the microcrevices abrasion produces, Influence of Oxygen and Sulfur on the Non-
and residual stress have been cited as possible Equilibrium Distribution of Chromium in
90 / Stainless Steels for Design Engineers

Austenitic Stainless Steel Welds and Slabs,” 23. N. J. Grant, Accelerated Oxidation of Met-
paper presented at Proceedings MS&T als at High Temperature, Trans. ASM, Vol
‘04,2004 44, 1961, p 128–137
16. R. Stickler and A. Vinckier, Trans. ASM, 24. J.F. Grubb, paper 04291 presented at Corro-
Vol 54, 1961, p 362 sion 2004, NACE, 2004, p 1–15
17. Aciers Inoxidables, Les Editions de Phy- 25. ACOM Files, High Temperature Stainless
sique, Les Ulis, Paris, 1993, p 568 Steels,
18. Aciers Inoxidables, Les Editions de 26. J. Okamoto et al., A Super-Austenitic Stain-
Physique, Les Ulis, Paris, 1993, p 448 less Steel for Tubing and Piping Applica-
19. Aciers Inoxidables, Les Editions de Phy- tions, Nippon Steel Technical Report 90,
sique, Les Ulis, Paris, 1993., p 453 July 2004
20. Aciers Inoxidables, Les Editions de 27. ACOM Files, High Temperature Stainless
Physique, Les Ulis, Paris, 1993, p 454 Steels,
21. 28. Allegheny Technologies, “AL6-XN®
22. W.C. Leslie, Mechanism of Rapid Oxida- Alloy”
tion at High Temperature, Trans. ASM, Vol
41, 1958, p 1213–1219
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 91-107 All rights reserved.
DOI: 10.1361/ssde2008p091


Duplex Stainless Steels

Summary they are exceptional materials. Whether these

duplex alloys will grow to the full extent of
THE NEWEST FAMILY of stainless steels is their potential depends on several factors:
the duplex alloys. The mixture of ferrite and
austenite in their structure gives them higher • Will high nickel and molybdenum prices be
strength than either phase by itself. Duplex al- sufficient motivation to drive designers to
loys have at least 20% chromium, so they are explore alternatives to traditional austenitic
considered as highly corrosion-resistant alloys grades?
but not high-temperature alloys because of em- • Will producers overcome their inhibition to
brittling phases. Their low nickel content makes aggressively market these grades through
them more economical than austenitic alloys of their cost-saving potential?
the same level of corrosion resistance, espe- • Will producers perfect the techniques to pro-
cially when their greater strength can be utilized duce these grades reliably so that their avail-
to reduce the amount of material required. They ability is unquestioned?
should largely replace alloys such as 316L • Will design codes change to correctly reflect
and317L in the future. the duplex materials’ higher ratio of yield
strength to tensile strength?
Why are there such issues with a family of al-
Introduction loys that has been successfully used for 20 years?
The concept of duplex stainless steels is simple:
Duplex stainless steels are the newest and islands of austenite in a continuous matrix of
fastest-growing alloy group in the stainless steel highly alloyed ferrite, as seen in Fig. 1. This
family. They are called duplex because at room
temperature they consist of two phases, ferrite
and austenite. Discovered in the 1920s, they
languished in a suboptimized and underutilized
state until recently. They possess excellent
strength, toughness, and corrosion resistance.
They also display exceptional resistance to
stress corrosion cracking (SCC) and corrosion
fatigue. The leaner grades, such as 2304, corre-
spond to 316L in corrosion resistance but have
double the yield strength, while the higher alloy
grades like 2507 compete with the 6% molyb-
denum superaustenitics in corrosion resistance
while still possessing much greater strength.
Their limitations lie in their lack of cryogenic
toughness and their inability to withstand tem-
peratures much above 300 °C without forming
embrittling phases. But between –100 and 300 °C Fig. 1 Wrought 2205 duplex microstructure
92 / Stainless Steels for Design Engineers

combination in principle offers high strength be- in the ferrite matrix, and each phase would have
cause of the possibility of refining the dual-phase equal corrosion resistance despite having differ-
grain structure and thereby raising yield strength ent compositions. It took a long time for that to
according to the Hall-Petch relationship as well be accomplished.
as by solid solution hardening, especially with Figure 2 shows a simple Fe-Cr-Ni constitu-
nitrogen. In addition, the absence of a continuous tional diagram (Ref 1). The salient points are
austenite phase provides relief from SCC by hav- that the typical successful alloys nearly bisect
ing any propagation of cracks in austenite ar- the two-phase field for austenite and ferrite. It is
rested by the ferrite phase. also obvious that the composition of the austen-
The optimization of the alloy system had to ite and the ferrite must be quite different.
wait for two events, both related to nitrogen. Ferrite contains a great deal more chromium
First, the control of nitrogen in the refining by than austenite; hence, its pitting corrosion re-
the argon oxygen decarburization (AOD) process sistance contribution from chromium is much
allowed the control nitrogen content up to the greater than the resistance of the austenite be-
solubility limit. Second, the understanding of cause in duplex grades:
the thermodynamics of the alloy system became
understood and reduced to a computer model. PREN = %Cr + 3.3 × %Mo + 16 × %N (Eq 1)
At this point, the alloys developed over the first
50 years of development became obsolete, and If one were to add molybdenum to increase
new grades with higher nitrogen vastly improved pitting resistance, it would preferentially parti-
the performance and user friendliness. Why this tion to the ferrite, further exacerbating the dif-
was so important can be seen by studying the ferential between the two phases.
structure of these alloys. This is where nitrogen saves the day. Addi-
tions of nitrogen concentrate nearly entirely in
the austenite. This lowers the activity of
Structure and Alloy Design chromium and thereby effectively attracts more
chromium to the austenite phase than would
The ideal structure of a duplex grade would otherwise be present. This stabilizes the austen-
be a stable 50-to-50 ratio of austenite to ferrite ite, keeping the ratio of ferrite to austenite more
at all temperatures at which it is to be used with- nearly constant with temperature. The pitting
out other phases. The austenite would be islands resistance of the austenite increases signifi-

Fig. 2 The Fe-Cr-Ni phase diagrams. The shaded area results from nitrogen additions
Chapter 7: Duplex Stainless Steels / 93

cantly to approximately that of the ferrite. In ad- (PREN) of about 35 and fills a niche in corrosion
dition, the nitrogen solid solution strengthens resistance where austenitics and ferritics are
the austenite and retards the formation of inter- lacking, greater than 317L stainless, PREN = 30,
metallic phases, which is not bad for an element and below the 6% molybdenum grades, such as
that costs nothing. AL-6XN alloy, with PRENs of around 45. Fer-
ThermoCalc, developed by the Swedish Royal ritics have a gap between 442 (18Cr-2Mo) and
Academy, has been an especially valuable tool in the super ferritics (28Cr-4Mo). By varying the
helping us understand and design better duplex chromium, nickel, and molybdenum, leaner al-
stainless steels. Without being able to computer loys can be devised that save cost based on re-
model the thermodynamics of the system, it duced molybdenum and nickel. Conversely,
would be impossible to project the partitioning of more corrosion-resistant alloys with higher
the various potential alloying elements. Figure 3 PRENs can also be mapped, such as Fig. 4, with
shows isopleth diagrams for a basic 2205 compo- the same diagrams varying nickel for a higher
sition in which nickel is varied. molybdenum level. This composition includes
The 2205 is the workhorse grade of duplex. It the important 2507 alloy.
has a pitting resistance equivalent number Partitioning of elements (Fig. 5) between
austenite and ferrite is an important issue. The
partitioning tendency is a strong function of
temperature. Figure 6 shows that as temperature

Fig. 3 The iron-nickel diagram for 22% Cr, 3% Mo, 0.15% N Fig. 5 Partitioning tendencies of various elements between
ferrite and austenite. Source: Ref 2

Fig. 4 The iron-nickel diagram for 25% Cr, 4% Mo, and

0.25% N: N is a nitride, χ is chi, σ is sigma, α is fer- Fig. 6 Variation of partitioning ratio with temperature.
rite, and γ is austenite Source: Ref 2
94 / Stainless Steels for Design Engineers

increases, the partitioning diminishes until at cleation and diffusion. The areas around the
just above 1300 °C it approaches unity for all newly formed σ are naturally somewhat di-
normal substitution-alloying elements (Ref 2). minished in chromium and molybdenum, so
For nitrogen, however, the tendency is to in- the alloy’s resistance to localized corrosion is
creasingly segregate to austenite as temperature compromised also.
increases (Ref 2). Figure 9 shows the TTT (time-temperature-
A danger in these alloys is that austenite transformation) diagram for various high-alloy
formed from ferrite on heating, such as during stainless steel, including austenitic, ferritic, and
welding or annealing, will contain only the low duplex. Alloys of all structures, ferritic,
amount of nitrogen that was in the ferrite from austenitic, and duplex, with high chromium and
which it was formed, until diffusion can restore molybdenum encounter the σ problem fairly
equilibrium. If the heating time does not permit equally and in proportion to their alloy content
this, this so-called secondary austenite will (Ref 2). This is the reason that the use of nitro-
have low nitrogen and therefore low pitting gen instead of molybdenum is so beneficial to
corrosion resistance, as shown in Fig. 7(e). Ni- the leaner alloys, not just in cost, but for the
trogen alters the phase stability, making austen- major reduction in rate of formation of sigma.
ite stable to higher temperatures. This helps Figure 10 shows the large reduction in σ forma-
keep welds from becoming excessively ferritic tion enjoyed by the lean alloy AL 2003™ mate-
and disturbing the desirable 50-to-50 ratio of rial compared to the higher molybdenum 2205
austenite to ferrite. Secondary austenite with alloy (Allegheny Ludlum).
low nitrogen is remedied by diffusion if the There are other intermetallic phases in addi-
phase forms at higher temperatures at which tion to σ. They include χ, R, π, and τ. These are
diffusion of nitrogen can rehomogenize the ni- of more research than practical interest because
trogen level. σ, with its bad consequences, forms sooner and
A crucial aspect of alloy design in the duplex in greater quantity under the same conditions
alloys involves the avoidance of unwanted compared to the others.
phases. The duplex stainless steels have all the Carbides and nitrides can also form in duplex
potential problems with embrittling phases of alloys. The nitride Cr2N can form when satu-
the ferritic and austenitic stainless steels com- rated ferrite is quenched from a high tempera-
bined since they contain both as phases. Ferrite ture, as can occur in the welding process. It is
forms two main embrittling phases, α′ and σ. possible that this would result in nearby
The α′ is generally believed to be a result of chromium depletion and a decrease in corrosion
the miscibility gap that exists in the iron- resistance. Carbide formation does not as easily
chromium system, by which ferrite undergoes cause chromium depletion in duplex alloys be-
spinodal decomposition into the iron-rich α, cause the precipitation at the ferrite-austenite
normal ferrite, and the chromium-rich α′, grain boundary does not deplete the austenite as
which is a brittle ordered alloy. Higher levels greatly in chromium locally because of the
of chromium or the presence of copper or neighboring ferrite having a much higher diffu-
molybdenum exacerbate this reaction, which sivity for chromium. The point is generally
has a formation that follows an Arrhenius-type moot since all modern duplex grades contain
curve with a maximum at around 400 °C. Fig- less than 0.030% carbon.
ure 8 shows the α′ formation kinetics for five Table 1 lists the duplex grades currently
duplex alloys (Ref 2). While duplex grades available commercially. Figure 7 shows a series
have good oxidation resistance and high-tem- of duplex photomicrographs.
perature strength, the α′ problem restricts their
use to below about 315 °C. Ferrite and austen-
ite both form intermetallic phases, of which
the most prominent and dangerous is σ, a Mechanical Properties
tetragonal phase richer in chromium and
molybdenum than the ferrite from which it In many ways, the duplex stainless alloys
forms. It is brittle and forms at grain bound- represent a best of both worlds in combining
aries, so its precipitation has the immediate ef- traits from the austenitic and ferritic alloys.
fect of lowering toughness. Cold work acceler- They offer high as-annealed strength with
ates the precipitation process by up to an order good toughness and ductility. Table 1 lists
of magnitude by virtue of its dual effect on nu- the major grades of duplex stainless steels;
Chapter 7: Duplex Stainless Steels / 95

Fig. 7 (a) As-cast duplex structure, austenite in a ferrite matrix. (b) 2205 annealed; austenite phase contains twins.
(c) 2507 as-welded; weld is highly ferritic because of rapid cooling rate. (d) Same weld as (c) after homogenization anneal.
(e) 7-Mo Plus with ( (dark areas) that has induced the formation of secondary austenite (arrows)

Table 2 lists typical and minimum properties The most striking and unexpected character-
for the major duplex alloys and those of some istic of the duplex grades is their high yield
comparable ferritic and austenitic alloys for strength, more than double that of comparable
comparison. austenitic grades.
96 / Stainless Steels for Design Engineers

Fig. 8 Kinetics of ( formation

The strength of the duplex grades is driven by

the strength of the continuous ferrite phase. It
owes its strengthening primarily to:
• Solid solution hardening by nickel, molyb-
denum, chromium, copper, and manganese
• Interstitial solid solution hardening by car-
bon and nitrogen
• Strengthening by grain refinement
These components have been related to the
mechanical properties by the following equa-
tions (Ref 3):
Fig. 9 Sigma formation kinetics at various alloy levels Rp0.2 = 120 + 210 N + 0.02
+2( Mn + Cr ) + 14 Mo + 10 Cu
+ (6.15 − 0.54δ)δ
+ (7 + 35( N + 0.02))d −1/ 2 (Eq 2)

Rp1.0 = Rp0.2 + 40 ± 9 (Eq 3)

Rm = 470 + 600( N + 0.02)

+14 Mo + 1.5δ + 8d −1/ 2 (Eq 4)

where δ is the ferrite content in percent, d is the

lamellar spacing, and results are in megapas-
The influence of nitrogen is interesting in that
at lower levels (e.g., below 0.1% nitrogen)
Fig. 10 Delay in ( precipitation in lean duplex 2003 austenite is the weaker phase, but additional ni-
Chapter 7: Duplex Stainless Steels / 97

Table 1 Duplex compositions

UNS Name C N Cr Ni Mo Mn Si Cu W P S
S32900 329 0.08 ... 23.0–28.0 2.5–5.0 1.0–2.0 1.0 0.75 ... ... 0.040 0.030
S31200 44LN 0.03 0.14–0.20 24.0–26.0 5.5–6.0 1.2–2.0 2.0 1.0 ... ... 0.040 0.030
S31260 DP3 0.03 0.10–0.30 24.0–26.0 5.5–7.5 2.5–3.5 1.0 0.75 0.2–0.8 0.1–0.5 0.030 0.030
S31500 3RE60 0.30 0.05–0.10 18.0–19.0 4.25–5.25 2.5–3.0 1.2–2.0 1.4–2.0 ... ... 0.030 0.030
S31830 2205(old) 0.03 0.08–0.20 21.0–23.0 2.5–3.5 2.5–3.5 2.0 1.0 ... ... 0.030 0.020
S32001 19 D 0.03 0.05–0.17 19.5–21.5 1.0–3.0 4.0–6.0 1.0 ... ... 0.040 0.030
S32003 2003 0.03 0.14–0.20 19.5–21.0 3.0–4.0 1.5–2.0 2.0 1.0 ... ... 0.040 0.030
S32101 2101 0.04 0.20–0.25 21.0–22.0 1.35–1.70 0.1–0.8 4.0–6.0 1.0 0.1–0.8 ... 0.040 0.030
S32205 2205 0.03 0.14–0.20 22.0–23.0 4.5–6.5 3.0–3.5 1.0 2.0 ... ... 0.030 0.020
S32304 2304 0.03 0.05–0.20 21.5–23.5 3.0–5.0 2.5 1.0 0.05–0.6 ... 0.040 0.040
S32520 Uranus 0.03 0.20–0.35 24.0–26.0 5.5–8.0 3.0–5.0 1.5 0.8 0.5–3.0 ... 0.035 0.020
S32550 255 0.04 0.10–0.25 24.0–27.0 6.0–8.0 2.9–3.9 1.5 1.0 1.5–3.0 . . . 0.040 0.030
S32750 2507 0.03 0.20–0.30 24.0–26.0 6.0–8.0 3.0–5.0 1.2 0.8 0.5 ... 0.035 0.020
S32760 Zeron 100 0.03 0.20–0.30 24.0–26.0 6.0–8.0 3.0–5.0 1.0 1.0 0.5–1.0 0.5–1.0 0.030 0.010
S32906 2906 0.03 0.30–0.40 28.0–30.0 5.8–7.5 1.5–2.6 0.8–1.5 0.5 0.8 ... 0.030 0.030
S32950 7–Mo Plus 0.03 0.15–0.35 26.0–29.0 3.5–5.2 1.0–2.5 2.0 0.6 ... ... 0.035 0.010
S39274 DP3W 0.03 0.24–0.32 24.0–26.0 6.0–8.0– 2.5–3.5 1.0 0.8 0.2–0.8 1.5–2.5 0.030 0.020
S39277 AF 918 0.025 0.23–0.33 24.0–26.0 6.5–8.0 3.0–4.0 0.8 0.8 1.2–2.0 0.8–1.2 0.030 0.020

Table 2 Duplex mechanical properties

Grade Name Rp0.2 Rm A5 HB RC –40 °C, J
S31200 44LN 450 690 25 293 31 ...
S31260 DP3 485 690 20 290 31 ...
S31830 2205(old) 450 62 25 293 31 ...
S32003 2003 450 620 25 290 30 40
S32001 19D 450 640 25 290 25 ...
S32101 2101 450 650 25 290 32 40
S32205 2205(new) 460 640 25 290 32 40
S32304 2304 400 600 25 290 31 40
S32520 Uranus 52N+ 550 770 25 ... 28 ...
S32550 Ferralium 550 760 15 302 32 ...
S32750 2507 550 795 15 310 32 40
S32760 Zeron 100 550 750 25 270 ... ...
S32960 7-Mo Plus 485 690 15 293 32 ...

trogen strengthens the austenite so that above above 300 °C is not recommended, no higher-
0.2% nitrogen, the austenite becomes the temperature properties are shown. Since they
stronger phase. have a ductile-to-brittle transition, they also are
The two phases are elongated parallel to the not well suited to cryogenic use.
major strain axis of working such as from hot or Impact Strength. Toughness is a significant
cold rolling. As working increases, the mi- consideration when using duplex alloys to re-
crostructure and properties become increasingly place the extremely tough austenitic alloys.
anisotropic, with the austenite taking on a (110) Duplex alloy low-temperature toughness is in-
[223] texture and the ferrite (100) [011] to (211) termediate to that of ferritic and austenitic al-
[011] (Ref 2). loys. This having been said, it should be noted
Because the ferrite phase controls mechani- that the duplex alloys can have excellent tough-
cal properties, the dependence of these proper- ness levels, such as 100 J at –100 °C in the so-
ties on temperature is significant since flow in lution-annealed condition. As would be ex-
body-centered cubic (bcc) structures is ther- pected, toughness improves with decreasing
mally activated. Figure 11 shows the variation grain size and deteriorates with cold work. The
of yield and tensile strengths of various grades most deleterious effect on toughness comes
along with that of austenite and ferrite of simi- from the precipitation of intermetallic phases,
lar composition. Because use of these alloys such as α′ and σ, which cause a sharp decrease
98 / Stainless Steels for Design Engineers

Fig. 11 Variation of ferrite, austenite, and duplex with temperature. Source: Ref 4

in toughness level and a concurrent increase in

transition temperature. The combined effect of
cold work and α′ can be seen in Fig. 12. Lean
alloys such as 2001, 2003, and 2101 have a
much slower rate of formation of α′ and are
much less at risk for loss of toughness from ex-
posure in the 300 to 600 °C range, as was
shown in Fig. 11.
Fatigue. Fatigue tests on duplex stainless
steels indicate that they possess a fatigue limit
of about 50% of the yield strength when tested
in air (Ref 4). The ratio of the fatigue strength in
a hostile environment to that in air is a useful
measure of the complementary strong points of
the duplex grades (i.e., strength and corrosion
resistance). Figure 13 shows that ratio for vari-
ous alloys plotted versus their PREN. As an
alloy’s resistance to corrosive attack increases,
its fatigue limit in a given environment ap-
proaches that in air, indicating simply that cor-
rosion plays an increasingly small role in fa-
tigue crack propagation as corrosion resistance
increases. While this is intuitively reasonable, it
is not diminished because the duplex reward the
user with a higher level of yield strength and fa-
tigue strength in air, so the net useful strength
Fig. 12 Increase in transition temperature with α′ formation
under cyclic loading is much greater than that of with aging for (a) annealed 2705 and (b) cold-
equivalent-PREN austenitic alloys. worked 2205. Source: Ref 4
Chapter 7: Duplex Stainless Steels / 99

Fig. 13 Influence of pitting resistance equivalent number (PREN) to fatigue strength in NaCl solution versus in air. Source: Ref 2

Forming and Machining Heavily formed sections should be fully

annealed, not just stress relieved, whenever
The higher strength and lower ductility of the there is a potential for SCC in the service
duplex grades compared to austenitics gives environment.
them correspondingly less ability to be cold
formed. Duplex alloys have sufficient ductility
to be cold drawn; they behave like ferritics or Corrosion Resistance
austenitics of similar alloy level. This, however,
is an alloy level at which excellent formability Because duplex alloys are made up of two
is seldom expected. Nevertheless, duplex alloys phases, ferrite and austenite, each must carry its
can be cold formed like austenitic alloys. Oper- own weight in resisting corrosion. Early alloys
ations such as bending, drawing, and pressing that were lacking in nitrogen generally had a fer-
can readily be performed. Bend radii should be rite phase that, because of the greater partition-
at least twice sheet thickness. Tubing can be ex- ing of the chromium and molybdenum to the fer-
panded into tube sheets, but care must be taken rite, had higher corrosion resistance than the
to produce tight roller-expanded joints. Tubing austenite. As nitrogen is added, it enriches the
bend radii should be at least twice tubing out- austenite phase preferentially until the corrosion
side diameter (OD). resistance of the austenite phase reaches that of
100 / Stainless Steels for Design Engineers

the ferrite. This approach is common to all more the aggressiveness of these media is enhanced
recently developed alloys starting with the revi- by contamination.
sion of 2205 from UNS S31803 to S32205, Sulfuric Acid. Figure 14 shows the behavior
which has primarily higher nitrogen. The net re- of S32304 compared to 304 and 316 in sulfuric
sult is a type of alloy that has most of the highly acid. Figure 15 shows additional, more highly
desirable corrosion resistance characteristics of alloyed duplex grades. The use of copper as an
superferritic grades without their limiting lack of alloying element in S32550 (1.5%) and S32760
mechanical properties, mainly toughness. (0.5%) gives them much better performance
than the otherwise similar S32750.
In real-life situations, such as seen in flue gas
General Corrosion
desulfurization, sulfuric acid can be contami-
The duplex alloys offer important advantages nated with chlorides. While this contamination
in performance over the austenitic grades in a is deadly to 316 and 317, it has only a minor ef-
number of significant aggressive media, includ- fect on the copper-alloyed duplexes (Fig. 16).
ing sulfuric acid, hydrochloric acid, sodium hy- Hydrochloric Acid. Historically stainless
droxide, phosphoric acid, and organic acids. steels have had their poorest performance when
This performance extends to situations in which confronted by hydrochloric acid. Here again, the

Fig. 14 The 0.1 mm isocorrosion curves. Source: Ref 5 Fig. 15 The 0.1 mm isocorrosion curves. Source: Ref 5

Fig. 16 Isocorrosion (0.1 mm/yr) performances of several austenitic and duplex alloys. Source: Ref 6
Chapter 7: Duplex Stainless Steels / 101

copper/tungsten-alloyed duplexes show excep- considered for use with nitric acid, and even
tionally good performance, as seen in Fig. 17. then no advantage can be claimed.
This extends the usefulness of stainless steels to Sodium Hydroxide. Much of the older pub-
an environment that had previously been off lished data on the behavior of stainless steel has
limits. Indeed, the duplex stainless steels in gen- seemed to promote the notion that higher nickel
eral can be said to be relatively indifferent to the levels were beneficial in strong bases. There
pH of chloride solutions and are affected rather seems now to be little to support that notion.
by the chloride concentration and temperature. Figures 18 and 19 clearly indicate, respectively,
Nitric Acid. It is fairly well known and ac- that the duplex alloys with their relatively low
cepted that resistance to nitric acid, which was nickel levels significantly outperform the higher
one of the first uses of stainless steel, depends nickel 304L and 316L, with performance im-
almost entirely on the chromium content. proving with increasing chromium content. The
Molybdenum, in all other instances a very bene- advantage is magnified when the environment is
ficial alloying element, has a strongly negative contaminated with chlorides, as is the case of
influence on resistance to this highly oxidizing the white liquors of kraft digesters.
acid. Consequently, only the leanest-molybde- Phosphoric Acid. While pure phosphoric
num duplex alloys, such as S32304, should be acid is not a very corrosive medium for stainless

Fig. 17 Isocorrosion (0.1 mm/yr) performance of duplex in HCl compared to 316L. Source: Ref 6

Fig. 19 Corrosion rates in white liquors plus chlorides.

Fig. 18 Corrosion rates in boiling NaOH. Source: Ref 7 Source: Ref 8
102 / Stainless Steels for Design Engineers

Fig. 20 Minimum temperatures for wet phosphoric acid (WPA) with an isocorrosion rate of 0.127 mm/yr. Source: Ref 9

In combinations of acetic and formic acid, the

superiority of duplex alloys is quite evident, as
seen in Fig. 22. S32750 shows virtual immu-
nity, while in mixtures contaminated with
halides its performance ranks very closely to
expensive nickel-based superalloys such as
N06625 and N06455. Even the lower alloyed
S32205 can offer an order of magnitude im-
provement over S31703 in hot contaminated
acetic acid.

Pitting Corrosion
The different analysis of the two main phases
in duplex alloys means that each has its own pit-
ting resistance equivalent number, PREN. The
Fig. 21 Isocorrosion (0.1 mm/yr) performances of various
ferrite phase has the relationship common to
alloys. Source: Ref 9
ferritic grades:
steel, contaminants again can render it so. PREN + %Cr + 3.3%Mo (Eq 5)
Halides are particularly common and aggressive
contaminants. Figure 20 shows the substantial while the austenite obeys the more familiar:
improvement in performance of the duplex al- PREN + %Cr + 3.3%Mo + 30%N (Eq 6)
loys over 316L when contaminants are present.
The duplex grades partition these critical ele-
Performance again improves with increasing
ments in such a way that the overall PREN of
chromium, molybdenum, and nitrogen levels.
most alloys comes out to be approximately Eq 1.
Organic Acids. Duplex alloys perform partic-
If one has the actual analysis of each phase,
ularly well in organic acids and have an excellent
then the proper relationship to use is Eq 2. These
record in industrial plants. In acetic acid, 304L
relationships are incomplete in that they only ad-
handles lower temperatures and concentrations.
dress the major alloying elements. Tungsten has
Alloys such as S32205 perform well. In
half the value of molybdenum and is frequently
formic acid, the most aggressive organic acid,
S32750 is resistant at all concentrations almost
to the boiling point, outperforming even tita- PREN = %Cr + 3.3(%Mo + 0.5 × %W) + 16%N
nium (see Fig. 21). (Eq 7)
Chapter 7: Duplex Stainless Steels / 103

Fig. 22 Corrosion rates for various alloys in 50% acetic plus formic acid, boiling. Source: Ref 10

If nonwrought material is involved, as in as- temperatures. This austenite, which forms from
cast and welded alloys, these relationships ferrite, has very little nitrogen, which clearly
greatly overstate PREN. This is because non- lowers its pitting corrosion resistance.
equilibrium-diminished concentrations of The duplex alloys stand up very well in com-
chromium are often found around precipitates, parison to corresponding superaustenitic alloys.
especially (manganese, chromium) S inclusions Figure 23 shows how CPT varies with PREN.
(Ref 11, 12) and because of lower alloy content This ranking is not always linear, as Fig. 24
locally due to solidification segregation, princi- shows, with pitting potential dropping fairly
pally of molybdenum. This is most significant rapidly with temperature and at different rates
in welded tubing, which can have higher sulfur for different alloys. Figures 25 and 26 show the
levels to increase weld penetration. Tube welds influence of pH and chloride concentration, re-
can be reequilibrated by high-temperature an- spectively. In 3% NaCl (Fig. 26), the rankings
nealing, but field girth welds will show dimin- show a minor variation with pH and a rational
ished corrosion resistance if unannealed. So, relationship to alloy content. The influence of
untreated welds can have PREN’s 5 to 15 lower chloride concentration is strong over a wide
than the parent alloy, which equates to the local- range of concentrations.
ized lowering of chromium levels. The critical These tests are best for judging relative per-
pitting temperature (CPT) of welds often de- formance of alloys and must be used cautiously
creases to near the critical crevice corrosion when extrapolating lab results to service per-
temperature (CCT) of the parent metal. formance. The degree to which short-term tests,
The precipitation of chromium- or molybde- whether potentiostatic or strictly immersion, re-
num-rich second (third, in this case) phases, flect long-term performance has not been well
such as σ or α′ inevitably results in diminish- established.
ment of these key alloying elements in the re-
gion surrounding the precipitate, which will
make it more prone to localized corrosion. Crevice Corrosion
This can also occur when secondary austen- Crevices exist both by design and inadver-
ite forms during the heating of alloys to high tently. Crevices are occluded volumes of liquid
104 / Stainless Steels for Design Engineers

Fig. 23 Critical pitting temperature in seawater measured potentiostatically versus pitting resistance equivalent number (PREN).
Source: Ref 13

Fig. 24 Variation of pitting potential with temperature. SCE, saturated calomel electrode. Source: Ref 14

in which oxygen and corrosion products reach The difference increases with total alloy con-
levels quite different from the exterior environ- tent, as can be seen in Fig. 27. Interestingly, the
ment and become highly corrosive. Thus, the difference is approximately the same as is the
tighter the crevice is, the greater the restriction difference in CPT between the wrought alloy
of diffusion between the crevice and the bulk and the welded alloy.
and therefore the greater the chance of crevice
corrosion occurring. An alloy’s susceptibility to
Stress Corrosion Cracking
crevice corrosion is proportional to its resist-
ance to pitting corrosion under the same condi- Stress corrosion cracking (SCC) has long
tions. The CCT is lower than the CPT by about been the Achilles’ heel of stainless steels. Only
10 to 30 °C. soft ferritic stainless steels are immune to it. It
Chapter 7: Duplex Stainless Steels / 105

occurs at temperatures and in environments

where stainless would be a perfect material if
only it did not stress corrosion crack. The
arrival of duplex stainless steels has to a very
large degree ameliorated, if not solved, that
SCC is unfortunately poorly understood. Like
pitting, whose initiation mechanism has not
been identified, SCC has both its initiation and
propagation mechanisms still open to debate.
But the duplex alloys have good strength
mainly through fine grain size and solid solution
hardening, which seems to avoid the hydrogen-
trapping dislocation types that seem to be asso-
ciated with hydrogen failures. So, while we can-
Fig. 25 Variation of critical pitting temperature (CPT) with not state the mechanism for SCC, we can map
pH. Source: Ref 14 out the conditions under which duplex alloys
are susceptible to SCC. The major environmen-
tal factors that affect SCC are chloride concen-
tration and temperature. Figure 28 shows the
remarkable advantage the duplex alloys have
over the comparable austenitic alloys with re-
gard to the temperatures at which they may be
used without SCC. The duplex alloys in this re-
gard are governed in their behavior by their fer-
rite matrix, through which cracks must propa-
gate (Ref 16). Ferritic stainless steels are known
for their resistance to SCC in the annealed con-
dition. The advantage of the duplex lies in their
composite-type microstructure with the crack-
arresting austenite phase and the toughening
fine grain structure. The duplex alloys show a
higher threshold stress for SCC as a percentage
of their yield strength (Fig. 29) than austenitic
alloys. This is in spite of their higher yield
strength, again giving these alloys more usable
Fig. 26 Critical pitting temperature (CPT) as a function
In ferrite, SCC susceptibility is a maximum
of NaCl concentration. SCE, saturated calomel
electrode. Source: Ref 5 below 100 °C, while in austenite susceptibility
appears to begin around 50 °C and increase
monotonically with temperature. The tempera-
ture at which SCC occurs at the fastest rate in-
creases with nickel content. This is also char-
acteristic of ferritic and martensitic materials
and mirrors their hydrogen embrittlement be-
havior. H2S also accelerates failure in chloride
environments (Fig. 30), and cold work acceler-
ates failure and lowers threshold stress values.
While duplex alloys behave in many regards
like ferritic alloys in their SCC or hydrogen
embrittlement response, they do not have the
same relationship between susceptibility and
bulk hardness. Other ferritic and martensitic
Fig. 27 Critical crevice temperature (CCT) and critical pit- alloys display pronounced susceptibility to
ting temperature (CPT). Source: Ref 15 these failures modes when their hardness
106 / Stainless Steels for Design Engineers

Fig. 28 Stress corrosion cracking (SCC) in neutral aerated NaCl. Testing duration 1000 hr. Source: Ref 5

Fig. 30 Suggested chloride and pH limits for cold-worked

duplex alloys. Source: Ref 17

exceeds Rc 22. The duplex alloys have an-

nealed hardness over Rc 30 without being in
danger. This probably simply indicates that
hardness as a measure of susceptibility is valid
only insofar as it reflects a certain yield
strength threshold as it does in tempered
martensite and is not valid for ferrite/austenite
composite structures. Thus, it is very important
to understand duplex SCC behavior as a sepa-
Fig. 29 Constant load stress corrosion cracking (SCC) tests in rate study and not interpret it in terms of
aerated MgCl2 at 150 °C. Source: Ref 5 austenitic or martensitic SCC.
Chapter 7: Duplex Stainless Steels / 107

REFERENCES 9. Avesta Sheffield, Corrosion Handbook for

Stainless Steels, 1994
1. P. Lacombe, B. Baroux, and G. Beranger, 10. B. Walden et al., Stainless Steel ’93,
Stainless Steels, Les Editions de Physique, Florence, AIM, 1993, p 3.47
2003 11. M.F. McGuire, MS&T Conf. Proc., 2004, p
2. R.N. Gunn, Duplex Stainless Steels, Abing- 831–846
ton Publishing, 1997, p 28 12. M. Ryan, D. Williams, R. Chater, B.
3. H. Nordberg H, Innovation of Stainless Hutton, and D. McPhail, Why Stainless
Steel, Conf. Proc., AIM, Florence, 1994, p Steel Corrodes, Nature, Vol 412, p 770
2.217–2.229 13. C.V. Roscoe et al., Duplex Stainless Steels
4. Charles, Duplex Stainless Steels ’91, Vol 1, ’86, The Hague, Nederlands Instituut voor
Beaune, Les Editions de Physique, 1991, p Lasteckniek, 1986, p 126–135
3–48 14. J.M. Drugli et al., Paper 270 presented at
5. S. Bernhardsson, Duplex Stainless Steels Corrosion ’90, Las Vegas, NACE, 1990
’91, Vol 1, Beaune, Les Editions de 15. S. Bernhardsson, Paper 164 presented at
Physique, 1991, p 137–150 Corrosion ’90, Las Vegas, NACE, 1990
6. J. Nichols J, 12th International Corrosion 16. T. Kudo, H. Tsuge, and A. Seki, Stainless
Congress, Houston, NACE, p 1237 Steel ’87, The Institute of Metals, 1988, p
7. E.-M. Horn, Werkstoffe und Korrosion, Vol 168–175
42, 1991, p 511–519 17. R. Francis, Duplex Stainless Steels ’94, Vol
8. J.P. Audouard, Stainless Steel Europe, April 3, Glasgow, TWI, 1994, paper KIV
1992, p 45
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 109-122 All rights reserved.
DOI: 10.1361/ssde2008p109


Ferritic Stainless Steels

Summary technology was documented long before AOD

was invented (Ref 1). It was not until carbon
THE FERRITIC STAINLESS STEELS are and nitrogen levels were brought down to AOD
the lowest-cost highly corrosion- and oxidation- levels that it became truly practical for ferritic
resisting alloys in existence. They are useful alloys. The level of carbon plus nitrogen was
mainly as light-gauge sheet since their tough- lowered from around 0.10% to around 0.04%,
ness drops off rapidly for heavier sections. Even and less-expensive high-carbon ferrochromium
as they have grown in use more than any other could be used instead of expensive low-carbon
type of stainless, they could still economically versions. Thus, there exist two types of ferritics:
displace the popular but expensive 304 for the early high-carbon types such as 430, 434,
many routine applications. 436, and 446 and the more modern stabilized al-
loys led by 409 and 439.
The older, unstabilized grades are not always
Introduction fully ferritic. Their carbon levels cause them to
form some high-temperature austenite, which
Ferritic stainless steels are simplest, lowest- transforms to martensite if quenched. This
cost stainless steels. In their minimal form, they makes their welds brittle. To be used, they are
contain simply enough chromium to overcome normally in the annealed condition, which re-
their inherent level of carbon impurity and hit quires a lengthy subcritical anneal to avoid
the 11% chromium in solution required for martensite and to evenly distribute chromium
“stainlessness.” Early in the 20th century, 430 after all carbides have stably formed. The newer
came into being, and the attainable levels of stabilized alloys behave as if they are interstitial
carbon removal required 16% chromium for free. They are ferritic at all temperatures (ex-
this to occur. So much extra chromium was re- cluding for the moment the possibility of extra-
quired because during annealing, to develop the neous phases such as (α' and σ) and can be eas-
fully ferritic structure, carbon combines with ily welded without fear of unwanted phases.
chromium, rendering it useless as a corrosion Stabilization does not preclude excessive grain
fighter. In October 1967, the first commercial growth in the fusion or heat-affected zone
use of argon oxygen decarburization (AOD) (HAZ) of welds, which can render them brittle.
changed the world for ferritic stainless steel. The mechanical properties of ferritic stainless
This process, in which argon and oxygen are steels appear similar to austenitics strengthwise,
blown through the molten metal to selectively but they lack the ductility of austenitics, and
remove carbon without removing chromium they are limited at low temperatures by brittle-
(described in detail elsewhere in this book), re- ness and at high temperatures by softness.
duced the carbon plus nitrogen levels suffi- The lower thermal expansion coefficient of
ciently that their effect could be nearly negated ferritics makes their scale more compatible with
by small additions of titanium or niobium, the base alloy and provides them with a lesser
which combine strongly with carbon and nitro- tendency to spall. This makes them excellent
gen and effectively remove them from solution. for high-temperature applications with thermal
This process is called stabilization, and the cycles, provided their strength is adequate.
110 / Stainless Steels for Design Engineers

The corrosion resistance of ferritics is ham- A similar predecessor was 410S, a low-carbon
pered by their inability to utilize nitrogen. The version of 410 to which some understabilizing
absence of nickel, which characterizes these al- amount of titanium is added but that still re-
loys, is not a problem since nickel adds little to quires annealing for full ferritic properties. The
corrosion resistance. The titanium stabilization key issue of the 11% chromium ferritics is how
of the modern alloys has quite a beneficial effect to deal with carbon and nitrogen. The 405 and
since titanium is a powerful deoxidizer and 410S take the approach of minimizing it and live
desulfurizer, both of which can cause local with annealing. The 409 uses full titanium stabi-
chromium depletion and pitting. Ferritics, more- lization. The hidden problem with using only ti-
over, are essentially free from stress-corrosion tanium is that unless nitrogen levels are made
cracking (SCC) since they are below the thresh- very low, the amount of titanium required to
old hardness for hydrogen embrittlement in combine with it can reach levels at which the
body-centered cubic (bcc) ferrous alloys. There first TiN precipitates in the molten metal. This
are a few exceptions. slaggy precipitate agglomerates, causing casting
The main attraction of ferritic stainless steels problems and surface defects. This gave 409 a
over austenitics is their cost. The old compari- reputation as a grade unsuitable for applications
son of 430 versus 304 is a bit unfair since 304 is that required good appearance because the tita-
richer in chromium. A fair comparison might be nium streaks were difficult to avoid and greatly
between 439 and 304. The corrosion resistance highlighted by polishing. This has largely been
of these two alloys is barely distinguishable overcome by better refining techniques to reduce
under normal ambient conditions. They are both carbon plus nitrogen to levels below 0.02% and
very formable and weldable. The vast majority the use of dual stabilization by titanium and nio-
of the objects made commercially from 304 bium; 468 (UNS S40930) is such an alloy.
could be switched to 439 with no adverse con- The historical archetype of ferritic stainless
sequence. But, if nickel is selling for $7 per steels was 430, which has existed since the
pound, then the total cost of 304 versus 439 is 1920s and is still widely used. Its drawbacks are
doubled by its presence. No design engineer can lack of weldability, relatively poor corrosion re-
afford to ignore this level of incentive to learn sistance because so much of its chromium is
to use ferritic stainless steels. tied up as carbides, and modest formability. The
new archetype for this medium-chromium level
is 439. With 17% chromium and single (439) or
Ferritic Stainless Alloys dual stabilization (468), this alloy overcomes
the problems of 430 and can readily replace 304
The ferritic stainless alloys generally group in in most applications with significant cost sav-
low (10.5 to 12.0%), medium (16 to 19%), and ings. In North America, 439 is mainly used as a
high (greater than 25%) chromium. They can be higher-temperature automotive exhaust alloy,
stabilized or not. These distinctions are some- but in Europe 430Ti is used extensively in more
what imposed after the fact. Rather than giving visually challenging applications, such as appli-
them an order that they truly do not possess, the ances. There, it is generally used instead of 439
most significant alloys are all listed in Table 1 whenever the part can be designed to be formed
with their compositions. from it.
The low-chromium ferritic stainless steels Now, 434 and 436 are little used as their his-
began with the development of MF-1, the prede- torical application in automotive trim finds little
cessor of 409, in the 1960s. Its excellent corro- place in today’s automotive styling. A modern
sion resistance, compared to carbon steel; rela- offshoot of these alloys, which are basically
tively low cost; good welding; and formability molybdenum enhanced 430, is 444. This alloy
permitted it to replace aluminized carbon steel has roughly the corrosion resistance of 316L but
and cast iron in automotive exhaust systems, is fully resistant to SCC in the welded or an-
opening up what eventually became the largest nealed condition. This makes it especially useful
single market for stainless steel. It was made for applications such as hot water heaters, heat
possible by the very low carbon plus nitrogen exchangers, and food- and beverage-processing
levels the AOD process provided and the use of equipment.
stabilization. Thus, 409 was an improvement on Both the nominally 11 and 18% chromium
405 in which aluminum performed a quasi stabi- alloys are sometimes modified to enhance
lization, and low carbon suppressed martensite. their high-temperature strength or oxidation
Table 1 Ferritic stainless compositions
Alloy Designation C N Cr Ni Mn Si Mo P S Ti Nb Other

405 S40500 0.08 ... 11.5–14.5 0.6 1 1 ... 0.04 0.03 ... ... 0.10–0.30 Al
400 AK alloy 0.05 ... 12.0–13.0 ... 1 1 ... 0.03 0.03 ... ... 0.25 Al
409 S40900 0.08 ... 10.5–11.75 0.5 1 1 ... 0.45 0.45 6x(C + N) to 0.75 ... ...
409 S40910 0.03 0.03 10.5–11.7 0.5 1 1 ... 0.04 0.02 6x(C + N) to 0.5 0.17 ...

409 S40920 0.03 0.03 10.5–11.75 0.5 1 1 ... 0.04 0.02 8x(C + N) to 0.15–0.50 ... ...

409 ultraform AK alloy 0.02 0.02 10.5–11.7 0.5 0.75 1 ... 0.04 0.01 8x(C + N) ... ...

466 S40930 0.02 0.02 10.5–11.75 0.05 1 1 ... 0.04 0.01 0.8 + 8x(C+N) Ti + Nb ... ...
409Cb S40940 0.06 ... 10.5–11.7 0.5 1 1 ... 0.04 0.04 10xC to 0.75 Nb ... ...

409Ni S40975 0.03 0.03 10.5–11.7 0.5–1 1 1 ... 0.04 0.03 ... ... ...

11 Cr-Cb AK alloy 0.01 0.015 11.35 0.2 0.25 1.3 ... ... ... ... 0.35 ...
12 SR AK alloy 0.02 0.015 12 ... ... ... ... ... ... 0.3 0.6 1.2 Al
Alfa I ATI alloy 0.025 ... 13 ... 0.035 0.03 ... ... ... 0.4 ... 3 Al
Alfa II ATI alloy 0.025 ... 13 ... 0.035 0.03 ... ... ... 0.4 ... 4 Al
4724 Outukumpu 0.08 ... 13.5 ... 0.7 1 ... ... ... ... ... 1 Al
429 S42900 0.12 ... 14.0–16.0 0.75 1 1 ... 0.04 0.03 ... ... ...
430 S43000 0.12 ... 16.0–18.0 0.75 1 1 ... 0.04 0.3 ... ... ...
430F S43020 0.12 ... 16.0–18.0 ... 1.25 1 ... 0.06 0.15 ... ... ...
430Se S43023 0.12 ... 16.0–18.0 ... 1.25 1 ... 0.06 0.06 ... ... 0.15 Se
430Ti S43036 0.1 0.04 16.0–19.5 1 1 1 ... 0.04 0.03 0.20 + 4x(C+N) to 1.10 ... 0.15 Al

439 S43035 0.07 0.04 17.0–19.0 0.5 1 1 ... 0.04 0.03 0.20 + 4x(C+N) to 1.10 ... ...
439LT S43932 0.03 0.03 17.0–19 0.5 1 1 ... 0.04 0.03 0.20 + 4x(C+N) to 0.75 Ti+Nb ... 0.15 Al
439 HP ATI, AK 0.01 0.01 17.5 0.2 0.35 0.45 ... 0.02 0.001 0.35 ... ...
439 ultraform alloys
468 S46800 0.03 ... 18.0–20.0 0.5 1 1 ... 0.04 0.03 Ti + Nb: 0.20 + 4x(C + N) to 1.10 ... ...
18 Cr-Cb AK alloy 0.02 ... 18 ... 0.3 0.45 ... ... ... 0.25 0.55 ...
18SR AK alloy 0.015 ... 17.3 0.25 0.3 ... ... ... ... 0.25 ... 1.7
typical Al
4742 Outukumpu 0.08 ... 18 ... 0.7 1.3 ... ... ... ... ... 1
typical Al
Chapter 8: Ferritic Stainless Steels / 111
Table 1 Ferritic stainless compositions
Alloy Designation C N Cr Ni Mn Si Mo P S Ti Nb Other
434 S43400 0.12 ... 16.0–18.0 ... 1 1 0.75–1.25 0.04 0.03 ... ... ...
436 S43600 0.12 ... 16.0–18.8 ... 1 1 0.75–1.25 0.04 0.03 ... Nb + Ta 5xC 0.7 . . .
441, S44100 0.03 ... 17.5–18.5 ... 1 1 ... 0.04 0.015 0.1-0.6 9xC 0.3–1.0 . . .
442 S44200 0.2 ... 18.0–23.0 0.6 1 1 ... 0.04 0.04 ... ... ...
436S ATI alloy 0.01 0.015 17.3 0.3 0.2 0.4 1.2 0.02 0.001 8x(C + N) min ... ...
112 / Stainless Steels for Design Engineers

444, YUS S44400 0.025 0.035 17.5–19.5 1 1 1 0.75–1.25 0.04 0.03 Ti + Nb: 0.20 + 4x(C + N) to 0.80 ... ...

433 ATI alloy 0.01 ... 20 0.25 0.3 0.4 ... 0.02 0.001 ... 10x(C + N) . . .
4762 Outukumpu 0.08 ... 24 ... 0.7 1.4 ... ... ... ... ... 1.5
typical Al
453 ATI alloy 0.03 ... 22 0.3 0.3 0.3 ... 0.02 0.03 0.02 ... 0.60 Al
typical 0.1
E-Brite, S44627 0.01 0.015 25.0–27.5 0.5 0.4 0.4 0.75–1.25 0.02 0.02 ... 0.5–0.20 0.2 Cu
26-1 0.5
Cu + Ni
Monit S44635 0.025 0.035 24.5–26.0 3.5–4.5 1 0.75 3.5–4.5 0.04 0.03 Ti + Nb: 0.20 + 4x(C + N) to 0.80 ... ...
Sea-cure S44660 0.025 0.035 25.0–27.0 1.5–3.5 1 1 2.5–3.5 0.04 0.03 Ti + Nb: 0.20 + 4x(C + N) to 0.80 ... ...
29-4C S44735 0.025 ... 28.0–30 0.5 1 0.75 3.5–4.5 0.04 0.03 Ti + Nb: 0.20 + 4x(C + N) to 0.80 ... ...
446 S44600 0.2 0.25 23.0–27.0 0.6 1.5 1 ... 0.04 0.03 ... ... ...
CC-50 Cast alloy 0.5 ... 26.0–30 4 1 1.5 ... ... ... ... ... ...
Chapter 8: Ferritic Stainless Steels / 113

resistance. Again, the driving force has been changers and extensively in condensing por-
the requirements of the hot end of exhaust tions of high-efficiency residential furnaces.
systems (e.g., exhaust manifolds). Alloying The lower-alloyed Seacure had a slight tough-
with niobium and molybdenum adds to the ness advantage that permitted it to be used at
high-temperature strength, while additions of wall thicknesses of 1/16 in. when AL29-4C®
chromium, silicon, and aluminum increase ox- was too brittle. As with other ferritics, these al-
idation resistance. There exists an array of loys are generally only suitably tough when
proprietary alloys as shown in Table 1; these used in thin section size (i.e., less than several
are usually developed for specific automotive millimeters).
needs and employ all or some of these alloy- It is difficult to say ferritic stainless steels are
ing variations. The use of silicon and alu- underutilized since they account for about half
minum decreases formability and can acc- the world’s production of stainless, but there are
elerate (formation, so their use involves many applications in which more expensive
trade-offs. austenitic stainless steels are used needlessly.
Alloys with more than 20% chromium are Ferritic stainless steels are a viable alternative
used specifically for high-oxidation or corrosion to nickel-bearing austenitics when thickness is 2
resistance. Despite the relative lack of high- mm or less and drawing and bending instead of
temperature strength, these alloys are particu- stretch forming is permitted. There are many
larly useful because of their high-oxidation re- applications where the longer corrosion life of
sistance, which they derive from the tight low-chromium ferritics should economically re-
adherence of their oxide scale. The close match place carbon steel, as they have in automotive
between the thermal expansion coefficient of exhaust systems. There are no technical barriers
the scale and the alloy prevents spallation of the to obtaining these savings; design engineers
oxide, which would lead to breakaway oxida- need to learn how to use these alloys.
tion. This was the purpose of the earliest high-
chromium ferritic stainless, 446. The perform-
ance of 446 has been exceeded by lower alloyed Metallurgy of Ferritic Stainless Steels
grades, such as the aluminum-alloyed ferritics.
A prime example of the state of the art is 453, Chromium stabilizes the ferritic structure at
which has not only 22% chromium and 0.6% high temperatures. Thus, above about 11%
aluminum but also rare earths in trace amounts chromium, austenite does not exist at any tem-
(i.e., 0.1%). perature in pure iron chromium alloys, as seen
As in austenitic alloys, rare earths act as very in Fig. 1.
powerful oxide and sulfide formers that concen- However, iron-chromium alloys devoid of
trate at the metal-oxide interface and stabilize it, carbon are not practical, so early metallurgists
again preventing spallation. This type of alloys saw the diagram shown in Fig. 2 with the level
finds use in high-temperature applications such
as planar oxide fuel cells.
The high-chromium alloys, when used for
corrosion resistance, are usually called super-
ferritics. In the 1960s, E-Brite® was developed.
To obtain high toughness, it was vacuum refined
to very low carbon plus nitrogen levels. It was
followed by the more capable 29-4®. Later, this
alloy was stabilized and became the still-popu-
lar AL 29-4C®. (E-Brite now has a new life as a
fuel cell material based on its oxidation resist-
ance and very low thermal expansion coeffi-
cient.) These alloys saw success as replace-
ments for 316L when SCC was a problem. This
alloy and its close neighbor SeaCure® are used
primarily in tubing where corrosion resistance
is most important. It was developed for welded
condenser tubing where seawater or brackish
water is involved. It is also used in heat ex- Fig. 1 Iron-chromium phase diagram from Thermocalc
114 / Stainless Steels for Design Engineers

of carbon at 0.20%, which represented the pu- sufficiently long homogenization anneal at a
rity level attainable in arc furnace refining. Car- low enough temperature that carbon and nitro-
bon is essentially insoluble in ferrite at ambient gen have very little solubility. This is standard
temperatures, and carbides of chromium and in the processing of unstabilized ferritic stain-
iron will form to the extent carbon is available. less steels, such as 430. Rapid cooling of unsta-
Since carbon diffuses interstitially much more bilized alloys causes carbon and nitrogen to pre-
rapidly than chromium can substitutionally, cipitate within grains. This severely embrittles
chromium is combined in situ, especially along the material and does not avoid sensitization.
grain boundaries, which are fast-diffusion This is called high-temperature embrittlement
paths. This locally depletes chromium, and the because it comes from putting carbon and nitro-
alloy is sensitized. This can be eliminated by a gen into solution at a high temperature and then
causing it to precipitate in a harmful manner.
These alloys were only ferritic at room tem-
perature if they were given a subcritical anneal
to transform austenite to ferrite. Otherwise, at
room temperature they would be ferrite plus
martensite. There are alloys that are intended to
use a mixed ferrite/martensite structure, but
they are treated later as a variation from the nor-
mal ferritic alloys.
The introduction of AOD refining permitted
much lower levels of carbon, as seen in Fig. 3,
opening the door for fully ferritic stainless steels.
Carbon and nitrogen added together produce
about the same effect as carbon alone. So, un-
stabilized fully ferritic alloys are not feasible
below 20% chromium without extreme refining
techniques, such as electron beam refining,
which are not commercially viable for low-cost
Fig. 2 Iron-chromium phase diagram at 0.20% carbon alloys. Thus, nearly all modern ferritic alloys

Fig. 3 Iron-chromium diagram at low carbon levels Source: Ref 2

Chapter 8: Ferritic Stainless Steels / 115

are “stabilized.” This means that a strong car- factors prevented TiC formation. The latter ef-
bide former such as titanium or niobium is fect was real in early austenitic alloys, such as
added in sufficient quantity to combine with all 321, leading to knife-line corrosion attack after
the carbon plus nitrogen, removing them from welding, but does not exist in low interstitial
solution. ferritic alloys, which have much greater diffu-
These reactions are simply: sion rates than austenitic alloys. But, since car-
bon mobility is quite high, it is not practical to
Ti + C = TiC (Eq 1) quench alloys quickly enough to prevent car-
Ti + N + TiN (Eq 2)
bide precipitation as is possible in austenitics
(detailed in the Chapter 6, “Austenitic Stainless
Nb + C = NbC (Eq 3) Steels”). Figure 4 shows the time-temperature-
transformation (TTT) curve for an unstabilized
Nb + N = NbN (Eq 4) 430-type alloy with carbon plus nitrogen of
Titanium is the stronger getter for carbon and 0.08% (Ref 3).
nitrogen. The thermodynamic driving force for Stabilization causes nonchromium carbides to
carbide and nitride formation is given by form at high temperatures, precluding chromium
carbide precipitation. The net effect is that
−7700 modern stabilized ferritic alloys behave as inter-
(Ti )(C) = + 2.75 (Eq 5)
stitial free and can be mapped using the pure
iron-chromium diagram shown in Fig. 1.
The rate of diffusion of carbon in ferrite is
−15790 around 100 times greater than that of carbon in
(Ti )( N) = + 5.40 (Eq 6)
T austenite. The solubility of carbon in ferrite is
vastly lower than it is in austenite. Because of
It must be noted that titanium has an even these factors, the heat treatments to avoid sensi-
higher affinity for oxygen and sulfur than for tization are essentially reversed. Carbon in
carbon, so that the removal of carbon from so- austenite can be retained in supersaturation for
lution is preceded by the removal of oxygen, extended periods of time. This is why austenitic
nitrogen, and sulfur in that order. This will be L grades do not sensitize even though they are
seen to have a major influence on corrosion re- slightly supersaturated. Sensitization occurs
sistance as the MnS inclusions generally asso- at higher levels of carbon by prolonged heating
ciated with the initiation of pitting are not at 600 to 850 °C. In ferritics, carbon cannot
found in titanium-stabilized grades of nor- be kept in supersaturation even by the most
mally low sulfur. In practice, the removal of rapid quenching, and sensitization is alleviated
oxygen begins in the molten state with the for- by prolonged heating in the 600 to 850 °C
mation of titanium sulfide and nitride and next range to allow chromium to equalize where
in the molten or solid state, depending on con- carbide precipitation has previously made it
centrations. It is desirable to keep sulfur and inhomogeneous.
nitrogen low enough that precipitation is in the
solid state so that precipitates do not agglom-
erate and cause large primary inclusions that
become unsightly surface defects. TiCS forms
in the solid state if sulfur is present; if not, TiC
forms. Essentially all carbon is removed from
solution below 1250 °C if carbon and nitrogen
are kept as low as possible and a stoichiomet-
ric amount of titanium is available (i.e., greater
than about four times the carbon plus nitrogen).
The stabilization formula in various specifi-
cations is more than four times the carbon plus
nitrogen because experimentally it has been
found that sometimes understabilization oc-
curs. This is due to the influence of oxygen and
sulfur having prior compound formation with
Fig. 4 430 time-temperature-transformation (TTT) curve. K,
the titanium and less importantly that kinetic carbideSource: Ref 3
116 / Stainless Steels for Design Engineers

Ferritic alloys, like austenitic alloys, can form temperature with the same structure as ferrite
intermetallic phases. The most prominent is σ, but with the chromium and iron atoms in an or-
which can be seen to form in higher-chromium dered bcc matrix in which iron and chromium
stainless steels (i.e., those with chromium plus occupy sites equivalent to two interlocking sim-
molybdenum of 20% or more). Formation of σ ple cubic matrices. Because the lattice so
occurs when such alloys are held between 500 closely matches that of ferrite, the precipitate is
and 800 °C; it is a hard, brittle tetragonal phase coherent and causes hardening. The α' embrit-
with equal parts iron and chromium. Thus, its tlement causes an extreme loss of toughness as
formation causes chromium depletion of the ad- well as hardening. It also causes a loss in corro-
joining ferrite. Formation requires substitutional sion resistance via the chromium depletion of
diffusion of chromium so is slower to form than that part of the matrix that surrenders chromium
carbides, minutes rather than seconds. Since to the α'-phase.
cold work enhances substitutional diffusion, it Figures 5 and 6 show the hardening effect of
accelerates σ formation. The σ forms preferen- α' and the resulting loss of toughness, respec-
tially along grain boundaries for diffusion rea- tively (Ref 6).
sons, and this causes it to have a major embrit-
tling effect. The σ may be redissolved by
solution annealing, but regaining full homo- Mechanical Behavior
geneity is not immediate.
Another embrittling phenomenon is the for- Ferritic stainless steels are quite similar in
mation of α'. This was named 885 °F or 475 °C their mechanical behavior to carbon steel. The
embrittlement before its cause was understood. main influence of chromium is to produce some
Before the nature of α' was known, it was con- solid solution hardening. Let us review the
fused with temper embrittlement, which occurs strengthening mechanisms of bcc iron. Pure
in martensitic alloys at the same temperature. iron is an extremely soft material with a yield
Temper embrittlement is the segregation of strength well under 10,000 psi. This softness is
phosphorus to prior austenitic grain boundaries not seen in practice because steel is never pure.
and does not occur in fully ferritic alloys. The α' Carbon has an extremely powerful effect on
is the ordered equiatomic chromium iron phase hardening, as does nitrogen.
that forms by spinodal decomposition; it has the The influence of substitutional alloying ele-
same composition as σ but exists at a lower ments is also quite significant. According to

Fig. 5 Influence of α' formation on hardnessSource: Ref 4

Chapter 8: Ferritic Stainless Steels / 117

Fig. 6 Influence of α' formation on toughness Source: Ref 5

Paxton (Ref 7), the misfit of solute atoms causes

lattice strains proportional to the amount dis-
solved and provides strengthening through the
lattice friction term. This mechanism also in-
creases the impact transition temperature unfa-
vorably. Elements that produce a refining of
grain size are the exception to this general rule in
carbon steel, but the lack of an austenite-to-fer-
rite transformation in stabilized ferritic stainless
steels negates this benefit for them. Figure 7
shows that fairly common ingredients and impu-
rities have strong hardening effects (Ref 6). Man-
ganese and silicon are normally deoxidizers, but
in titanium-stabilized alloys, titanium takes over
the deoxidizing role so their presence can be lim- Fig. 7 Influence of substitutional elements on hardness of
iron alloys
ited. Phosphorus is virtually impossible to refine
from stainless steel, so its presence at around
0.02% is normally a given unless low-phospho- The softest ferritic stainless alloys are the
rus raw materials are used as a starting point. 409 variations made for highly formed exhaust
The worst toughness-inhibiting effects come system components. They contain as little
from interstitial elements to grain boundaries: manganese, silicon, nickel, and other substitu-
oxygen, carbon, and nitrogen. The effect of car- tional elements as possible and have a mini-
bon plus nitrogen on transition temperature is mum of carbon plus nitrogen, so that the re-
profound, as seen in Fig. 8 (Ref 8). sulting precipitate fraction after titanium
Stabilizing removes the interstitial carbon and addition is as low as possible. To maximize
nitrogen, along with oxygen and sulfur, from softness and formability, titanium and niobium
solution. This does not produce a major soften- in excess of that required for stabilization must
ing, however, because the precipitate itself has a also be minimized as they will cause solid so-
hardening effect. lution hardening.
118 / Stainless Steels for Design Engineers

Fig. 9 Corrosion of titanium-stabilized 29% Cr plus 4% Mo

alloys in ASTM A 763 Y test. Source: Ref 11

Fig. 8 Influence of interstitial carbon and nitrogen on tough-

ness transition temperature Source: Ref 8


Stabilization is essential to ferritic stainless

steels to avoid the precipitation of grain bound- Fig. 10 Corrosion of niobium-stabilized 29% Cr plus 4% Mo
alloys in ASTM A 763 Y test. Source: Ref 11
ary carbides. Combined carbon plus nitrogen
levels below 100 ppm are necessary to avoid
both sensitization and embrittlement, but with- As titanium and niobium are added to alloys,
out proper heat treatment even alloys of this pu- their corrosion resistance is improved (Figs. 9
rity can incur debilitating loss of toughness due and 10) (Ref 11). Maximum improvement in
to carbide and nitride precipitates (Ref 9). These corrosion resistance levels off once full stabi-
levels are not economically attainable for com- lization is reached. Excess amounts of the stabi-
mercial alloys, so stabilization is the correct en- lizing elements have negligible effect, but tita-
gineering answer. nium-stabilized alloys have a lower rate of
Stabilization is generally considered as the corrosion than niobium-stabilized alloys. This is
simple gettering of carbon and nitrogen by a probably due to titanium’s ability to eliminate
suitable carbide and nitride former. It was not sulfur and oxygen from solution.
known until about 1980 just what the mecha- Toughness improves for niobium-stabilized
nisms of embrittlement were in the ferritic alloys up through full stabilization and then be-
stainless steels, however. The distinguishing of gins to decline. This is a result of excess stabi-
α' from those related to interstitials and their lizing alloy acting as a solid solution hardener
stabilizers (Ref 10) permitted stabilizing ele- and therefore a toughness reducer. This tough-
ments to be optimized. ness reduction is more pronounced with tita-
Titanium combines with carbon and nitrogen nium, which is a stronger solid solution hard-
stoichiometrically by: ener (Figs. 11 and 12) (Ref 11).
The upshot of this understanding was the in-
Ti = 4 × C + 3.4 × N (Eq 7) troduction of dual stabilization, through which
both weld and base metal toughness and corro-
Niobium requires a greater weight percentage: sion resistance are optimized. The same study
recommended that dual stabilization follow the
Nb = 7.7 × C + 6.6 × N (Eq 8) following formula:
Chapter 8: Ferritic Stainless Steels / 119

(Ti + Nb) ≥ 6 × (C + N) (Eq 9)

The toughness of these alloys has a broad op-

timum that takes advantage of the corrosion-re-
sisting benefits of titanium (Fig. 13) (Allegheny
Other strong carbide formers such as zirco-
nium and vanadium are ineffective stabilizers
because their mobility at the temperatures at
which they are thermodynamically capable of
forming sufficiently large percentages of car-
Fig. 11 Charpy V-notch impact ductile to brittle transition bides and nitrides is too low to rid the matrix of
temperature (DBTT) of titanium-stabilized 29%Cr these elements. They also have too great a ten-
plus 4%Mo alloys test. Source: Ref 11
dency to form intermetallic compounds.
Toughness in ferritic stainless steels is a
major consideration. If ferritic alloys enjoyed
the same toughness as austenitic alloys, there
would be few instances when the use of the
much more expensive nickel-bearing grades
would be justified. Because stabilized alloys
are ferritic at all temperatures, there is no
automatic grain-refining transformation as ex-
ists in carbon steel. If grains grow large from
annealing at high temperatures or welding, then
the transition temperature increases. Section
size also has an effect. Stabilized ferritic stain-
less steels are seldom used in thicknesses of
Fig. 12 Charpy V-notch impact ductile to brittle transition over several millimeters because of decreasing
temperature (DBTT) of niobium-stabilized 29%Cr toughness. Figure 14 shows how transition
plus 4%Mo alloys test. Source: Ref 11

Fig. 13 Toughness of dual-stabilized low-alloy ferritic stainless. AL 466 is recognized as S40930

120 / Stainless Steels for Design Engineers

tion. But, the anisotropy does result in remark-

able drawing characteristics, with ferritic
stainless steels with elongations in tensile tests
in the mid-30% range being nearly equal to
austenitic stainless steels with over 50% elon-
The measure of anisotropy is the Lankford
ratio. It is expressed as:

r0 + 2r45 + r90
R= (Eq 10)

When this expression equals 1, then a mate-

rial is isotropic. As the value increases from 1,
the drawability increases, as measured by the
Fig. 14 Change in transition temperature with thickness for
limiting drawing ratio (LDR), the ratio of the
29Cr-4Mo-2Ni alloy. Source: Ref 12
diameter of a disk to that of the deepest cylinder
temperature can increase with thickness (Ref into which it can be drawn. The ferritic stainless
12). This effect is due simply to stress states steels in sheet form have LDRs of around 2.2
transitioning from biaxial to the more embrit- compared to 2.0 for 304.
tling triaxial with increasing thickness. The good formability of ferritic stainless
steels has some drawbacks. They are subject to
ridging, which is the formation of visible ridges
Texture and Anisotropy parallel to the direction of elongation. This is an
artifact of texture in the material. A combination
The deformation of ferritic bcc materials is of careful chemistry design and thermomechan-
characterized by limited slip systems, high ical processing is required to keep it under con-
stacking fault energy, and lattice anisotropy. trol. The approach centers on variables that in-
So, when ferritic stainless are deformed, dis- crease stored energy from deformation to
locations tend not to dissociate as they do in promote recrystallization over recovery during
austenitic stainless steels. The lack of dissoci- annealing.
ation of dislocations encourages cross slip. The ferritic stainless steels even carry for-
This minimizes dislocation tangles and work ward some of the preferred grain orientation
hardening. that come from initial solidification when
When ferritic stainless steels are deformed, growth of dendrites is along preferred crystallo-
certain crystallographic slip systems predomi- graphic directions. Hot working merely reori-
nate, so that large deformations mechanically ents these similarly oriented grains en masse.
bring different grains via rotation into closer Without phase changes or enough stored energy
crystallographic alignment. This preferred de- to provoke full recrystallization, randomness of
formation along easier slip planes results grain orientation is never achieved.
macroscopically in overall mechanical proper- Titanium-stabilized steels show more texture
ties varying with direction with respect to the and recovery versus recrystallization than do
prior deformation. Thus, ferritic stainless niobium-stabilized alloys. This is because tita-
steels, like low-carbon steels, have pronounced nium carbides and nitrides form at higher tem-
mechanical anisotropy. This is manifest in perature and are therefore coarser. They thus
their deep drawing characteristics. Heavily present less obstruction to dislocation motion
cold-rolled and annealed ferritic stainless than finer niobium precipitates. Furthermore,
steels draw quite well. They resist thinning. niobium precipitates tend to dissociate to a
When elongated, they contract in the width di- greater degree than those of titanium. This puts
rection while keeping virtually the same thick- niobium in solution during hot working where it
ness. This same phenomenon means that they can interact with dislocations. Thus, alloys at
cannot be stretch formed since plain strain least partially stabilized with niobium can
quickly results in fracture because of the re- achieve greater recrystallization, which can
sistance to deformation in the thickness direc- translate to finer grain size and less anisotropy.
Chapter 8: Ferritic Stainless Steels / 121

Boron additions to ferritic stainless steels chromium and molybdenum. Likewise, other
result in the formation of grain boundary car- elements can have a negative effect. Any ele-
bides, M23(C, B)6. If added to titanium-stabi- ment that can combine with chromium or
lized steels, the carbides form on preexisting molybdenum can detract from corrosion resist-
TiN particles and result in coarser overall pre- ance by their removal of these essential ele-
cipitate arrays since finer, lower-temperature ments from solution. The most notorious of
precipitating TiC or TiCS precipitates are at these is carbon, whose tendency to form
least partially precluded. The net result is chromium carbides causes areas around such
coarser grain size and no major improvement carbides to be partially depleted of chromium.
in mechanical properties over the use of tita- However, nitrogen, oxygen, and sulfur can also
nium alone (Ref 13). Additions of boron to form chromium compounds and cause localized
niobium-stabilized steels does cause finer pre- loss of corrosion resistance. Manganese sul-
cipitates and grain size than would niobium fides, for instance, are almost always seen to be
alone (Ref 14). the locus of pitting corrosion (Ref 15). More
careful examination has shown that such sul-
fides grow in the solid state as chromium/man-
High-Temperature Properties ganese sulfides and deplete their very close sur-
roundings of chromium, inviting corrosion to
begin at the inclusion-matrix interface, where
High-temperature mechanical properties of
chromium levels in solution are reduced (Ref
ferritic stainless steels are often important to
their successful use because their oxidation re-
Other factors that lead to loss of localized cor-
sistance is excellent and better than austenitics,
rosion resistance are the formation of
but their high-temperature strength is lower
chromium-rich phases such as α' and σ. Either
than that of austenitics. This has led to consider-
of these with about 50% chromium will cause
able development of high-temperature proper-
adjoining ferrite to have lower chromium levels.
ties, primarily for the automotive market. Re-
Because ferrite has a non-close-packed struc-
search has determined that high-temperature
ture, diffusion rates, both substitutional and in-
strength and creep resistance are best served by
terstitial, are about two orders of magnitude
stabilizing grain size and having niobium in
higher than in austenite. That means that any
solid solution. Adding titanium to niobium-sta-
deleterious chromium-depleting reaction can
bilized steels stabilizes the type of carbide, es-
happen more rapidly. Alloys cannot be
pecially preventing the formation of the coarse
quenched rapidly enough to forestall sensitiza-
M6C, whose growth decreases strength. The
tion, the precipitation of chromium carbides
relatively high insolubility of TiC causes this.
that depletes grain boundary regions of
Niobium is concurrently made available for
chromium. Instead, carbon must be neutralized
high-temperature solid solution strengthening.
by stabilization, or the chromium depletion
must be removed by homogenization in long-
box anneals. Note that the latter technique is
Corrosion and Oxidation Resistance also possible in austenitics but would require
annealing for excessively long times, 102 h or
Corrosion resistance is chemistry dependent so.
rather than structure dependent, so ferritic stain- The ferritic stainless steels are valued for
less steels behave just as do other stainless their resistance to SCC. Even in environments
steels of the same crucial alloy content. that cause pitting, the normal initiation step for
The main alloying elements that provide re- SCC, annealed ferritic stainless steels do not
sistance to localized corrosion, general corro- undergo SCC as long as alloying elements such
sion, and crevice corrosion are chromium, as nickel, copper, and cobalt are kept below
molybdenum, and nitrogen. Since nitrogen is 0.5% in aggregate. Cold work sufficient to raise
essentially insoluble in ferrite, it cannot con- their hardness above Rc 20 to 22 can make them
tribute to the corrosion resistance of ferritic susceptible to both SCC and its cousin, hydro-
stainless steels as it can in austenite. Other al- gen embrittlement. The more highly alloyed su-
loying elements, such as copper and nickel, can perferritic alloys are even susceptible to hydro-
add to corrosion resistance in special cases, but gen embrittlement in the annealed condition
they are of secondary importance compared to (Ref 17). As with martensitic stainless steels,
122 / Stainless Steels for Design Engineers

this susceptibility is a maximum near room tem- 10. J. Grubb, R. Wright, and P. Farrar, “Micro-
perature and declines with increasing tempera- mechanisms of Brittle Fracture in Titanium-
ture, as opposed to austenitics, which see their Stabilized Stainless Steels,” Special Publi-
maximum susceptibility above room tempera- cation 706, ASTM, 1980
ture. This limits these alloys’ ability to employ 11. J. Grubb, Stabilization of High-Chromium
cathodic protection safely to –0.80 Vsce, at Ferritic Stainless Steels, Proc. Int. Conf.
which point corrosion in seawater is, if not Stainless Steels, ISIJ, Chiba, 1991
eliminated, reduced to very low levels (Ref 18). 12. M.A. Streicher, Stainless Steel ‘77, p 27
13. E. El-Kashif, K. Asakura, T. Koseki, and
K. Shibata, ISIJ Int., Vol 44, 2004, p
REFERENCES 1568–1575
14. N. Fujita, K. Ohmura, E. Sato, and A.
1. F.M. Beckett and R. Franks, Trans AIME, Yamamoto, Nippon Technical Report 71,
Vol 113, 1934, p 126–143 Oct 1996
2. Stainless Steel, Les Editions de Physiques, 15. T. Suter, E. Webb, H. Bohni, and R. Alkire,
1992, p 483 Pit Initiation in I M NaCl With and Without
3. Stainless Steels, Les Editions de Physique, Mechanical Stress, J. Electrochem. Soc.,
2003 Vol 148 (No. 5), 2001, B174
4. H.D. Newell, High Chromium Irons, Met. 16. M. Ryan, D. Williams, R. Chater, B. Hutton,
Prog., April 1947, p 617–626 and D. McPhail, Why Stainless Steel Cor-
5. P.J. Grobner, The 885 °C (475 °C) Embrit- rodes, Nature, Vol 412, 2002, p 770
tlement of Ferritic Stainless Steels, Metall. 17. J. Grubb, “Hydrogen Embrittlement of Su-
Trans., Vol 4, 1973, p 251–260 perferritic Stainless Steels,” paper presented
6. Handbook of Stainless Steels, Peckner and at 1984 ASM Int’l Conference on New De-
Bernstein, McGraw Hill, 1977, p 5–9, 5–12 velopments in Stainless Steel Technology,
7. H.W. Paxton, Alloying, ASM, 1998, p 213 Detroit, September 1984
8. H. Abo et al., Stainless Steel ‘77 18. J. Grubb and J. Maurer, “Use of Cathodic
9. J. Grubb and R. Wright, The Role of C and Protection With Superferritic Stainless
N in the Brittle Fracture of Fe-26Cr, Met. Steels in Seawater,” paper presented at Cor-
Trans. A, Vol 10A, Sept 1979, p 1247–1255 rosion 84, New Orleans, April 1984
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 123-135 All rights reserved.
DOI: 10.1361/ssde2008p123


Martensitic Stainless Steels

Summary which restricts the temperature and composition

ranges over which it is possible to obtain a fully
THE SMALLEST CATEGORY of stainless austenitic structure from which to form marten-
steels in usage volume is the martensitic stain- site. The presence of ferrite in a martensitic
less steels. This is mainly because these alloys structure is detrimental to strength, hardness,
are limited in corrosion resistance because of and toughness. Ferrite can appear in the as-cast
the necessity of keeping alloy levels low to pro- structure and be formed during austenitizing or
duce the martensite structure. Even so, they fill tempering. All the usual concerns inherent in
an important niche as a strong, hard, and tough any martensitic alloys are still present; temper
alloy of fairly good corrosion resistance and as embrittlement, retained austenite, etc.
a strong, stable, high-temperature alloy. Martensitic stainless steels are the most mar-
ginally corrosion resistant of all the stainless al-
loys. The requirement that they be fully austeni-
Introduction tizable limits the amount of corrosion-resisting
chromium and molybdenum they can contain.
Nearly 100 years ago cutlery was first sold in Much of the carbon in them detracts from the ef-
Great Britain with a composition of 13% fective chromium content by forming chromium
chromium and 0.25% carbon. This was the first carbides. In addition they are always susceptible
commercial use of stainless steel and cutlery to stress corrosion cracking (SCC) when their
with the same basic analysis is still sold today. hardness exceeds about Rc 22. These limitations
The useful alloys of martensitic stainless steel combine to make their excellent properties us-
contain from roughly 11 to 18% chromium and able in only mild environments compared to
up to 1.0% carbon. Relatively small amounts of other stainless steels. Their high strength and
nickel, molybdenum, tungsten, vanadium, and hardness for their relatively low cost ensure their
niobium are also added at times for specific pur- place as a very useful engineering material.
poses explained in this chapter. Those marten- Table 1 lists the most significant of the
sitic stainless steels in which elements such as martensitic stainless steel alloys. The reader
copper and titanium are added to produce addi- should be aware that some alloys which are
tional hardening through precipitation are dis- quite similar are discussed primarily in other
cussed in Chapter 4, “Corrosion Types.” chapters dealing with specifically PH stainless
The designers and engineers already familiar steels or primarily ferritic stainless steels. The
with martensitic carbon and alloy steels will distinction between martensitic stainless steels
find nothing confusing about martensitic stain- and some other stainless alloy families is some-
less steels. There is no aspect of martensitic times vague. Nearly all the precipitation-hard-
steels that does not apply directly to stainless ening stainless steels are used in the martensitic
martensitic steels. The additional concerns one state, but their special hardening mechanism of
must have with stainless martensite relate precipitation within a martensitic matrix causes
mainly to those that are due to the strong ferri- them to be categorized separately somewhat ar-
tizing influence of chromium. Chromium bitrarily. By this conventional logic, some of the
strongly promotes the formation of ferrite, martensitic alloys containing molybdenum or
Table 1 Compositions (wt%) of martensitic stainless steels
Alloy UNS Form C Mn S Si Cr Mo Ni Other Other
403 S40300 Wrought 0.15 max 1 0.03 0.5 11.5-13.5 ... ... ... ...
410 S41000 Wrought 0.15 max 1 0.03 1 11.5-13.5 ... ... ... ...
410S S41003 Wrought 0.03 1 0.03 ... 10.5- ... 1.5 0.03 ...
12.5 N
410 S41008 Wrought 0.08 1.5 0.03 1 11.5- 0.6 ... ... ...
410Cb S41040 Wrought 0.18 max 1 0.03 1 11.5- ... ... 0.05-0.30 ...
13.5 Nb
412 S41003 Wrought 0.030 max 1.5 0.03 1 10.5- ... 1.5 0.030 max ...
12.5 N
414 S41400 Wrought 0.15 max 1 0.03 1 11.5-13.5 ... 1.25-2.50 ... ...
414 mod S41425 Wrought 0.05 0.5-1.0 0.005 0.6 12.0- 1.5- 4.0- 0.06-0.12 0.30 Cu
124 / Stainless Steels for Design Engineers

15 2 7 N
415 S41500 Wrought 0.05 max 0.50-1.00 0.03 0.6 11.5-14.0 0.50-1.00 3.50-5.50 ... ...
416 S41600 Wrought 0.15 max 1.25 0.15-0.30 1 12.0-14.0 ... ... ... ...
416Se S41623 Wrought 0.15 max 1.25 0.06 1 12.0- ... ... 0.15 min ...
14 Se
418 S41800 Wrought 0.15-0.20 0.5 0.03 0.5 12.0-14.0 ... 1.80-2.20 2.50-3.50 W ...
420 S42000 Wrought 0.15 min 1 0.03 1 12.0-14.0 ... ... ... ...
4116 DIN Wrought 0.5 ... ... ... 14.5 0.65 ... ... 0.15 V
1.4116 Nominal
420F S42020 Wrought 0.15 min 1.25 0.15 min 1 12.0-14.0 0.6 ... ... ...
420FSe S42023 Wrought 0.15 min 1.25 0.06 1 12.0-14.0 0.6 ... 0.15 min Se ...
422 S42200 Wrought 0.20-0.25 1 0.03 0.75 11.0-13.5 0.75-1.25 0.50-1.00 0.75-1.25 W 0.15-0.30V
424 S42400 Wrought 0.06 max 0.50-1.00 0.03 0.30-0.60 12.0-14.0 0.30-0.70 3.50-4.50 ... ...
425 S42500 Wrought 0.08-0.20 1 0.01 1 14.0-16.0 0.30-0.70 1.00-2.00 ... ...
425mod ... Wrought 0.50-0.55 1 0.03 1 13.0-14.0 0.80-1.20 0.5 ... ...
Trinamet ... Wrought 0.30 max 1 0.03 1 12.0 1.00 ... 2.00 ...
-14 -3 -3.00 Cu
HP13Cr-1 JFE Wrought 0.025 0.45 ... ... 13 1 4 ... ...
HP13Cr-2 JFE Wrought 0.025 0.45 ... ... 13 2 5 ... ...
NT-CRS Nippon Wrought 0.03 1.45 ... ... 12.7 1.4 4.5 1.5 Cu 0.040 N
Table 1 (continued)
Compositions (wt%) of martensitic stainless steels
Alloy UNS Form C Mn S Si Cr Mo Ni Other Other
NT-CRSS Nippon Wrought 0.02 2 ... ... 12.3 2 5.8 1.5 Cu 0.015 N
KL-12Cr JFE Wrought 0.01 ... ... ... 11 ... 2.4 0.5 Cu 0.010 N
KL-HP JFE Wrought 0.01 ... ... ... 12 2 5.5 ... 0.010 N
12Cr Nominal
431 S43100 Wrought 0.20 max 1 0.03 1 15.0-17.0 ... 1.25-2.50 ... ...
440A S44002 Wrought 0.60-0.75 1 0.03 1 16.0-18.0 0.75 ... ... ...
440B S44003 Wrought 0.75-0.95 1 0.03 1 16.0-18.0 0.75 ... ... ...
440C S44004 Wrought 0.95-1.20 1 0.03 1 16.0-18.0 0.75 ... ... ...
440F S44020 Wrought 0.95-1.20 1.25 0.10-0.35 1 16.0-18.0 0.40-0.60 0.75 ... ...
440FSe S44023 Wrought 0.95-1.20 1.25 0.03 1 16.0-18.0 0.6 0.75 0.15 min Se
BG-42 Nominal Wrought 1.15 ... ... 0.3 14.5 4 ... 1.2 V ...
ATS-34 Nominal Wrought 1.05 0.4 ... 0.35 14 4 ... ... ...
14-4 Nominal Wrought 1.05 0.5 ... 0.3 14 4 ... ... ...
154 CM Nominal Wrought 1.05 0.45 ... 0.3 14 4 ... ... ...
CPM Nominal PM 1.45 ... ... ... 14 2 ... 4.0 V ...
CPM Nominal PM 2.15 0.4 ... ... 17 0.4 ... 5.5 V ...
CPM Nominal PM 2.2 ... ... ... 13 1 ... 9.0 V ...
CA-15 J91150 Cast 0.15 max 1 0.04 1.5 11.5-14.0 0.5 1 ... ...
CA 15M J91151 Cast 0.15 max 1 0.04 0.65 11.5-14.0 0.15-1.00 1 ... ...
CA-40 J91153 Cast 0.20-0.40 1 0.04 1.5 11.5-14.0 0.5 1 ... ...
CA-40F J91154 Cast 0.20-0.40 1 0.020-0.040 1.5 11.5-14.0 0.5 1 ... ...
CB-6N J91650 Cast 0.06 max 0.5 0.02 1 10.50-12.50 1 ... ...
CB-6MN J91540 Cast 0.06 max 1 0.03 1 11.50-14.00 0.40-1.00 3.50-4.50 ... ...
CA-28MVW J91422 Cast 0.20-0.28 0.50-1.00 0.03 1 11.00-12.50 0.90-1.25 0.50-1.00 0.90-1.25 W 0.20-0.30 V
Chapter 9: Martensitic Stainless Steels / 125
126 / Stainless Steels for Design Engineers

tungsten should also be considered precipitation- on temperature. The amount is given by the
hardening alloys, but they customarily are not Koistinen and Marburger equation (Ref 1):
and will not be in this work.
The ferritic alloys often have compositions 1− Vα ′ = exp{β (Ms − T )}
(Eq 1)
that allow them to be partially martensitic under
some conditions. 430 (UNS S43000) and The martensite is coherent with the parent
3CR12 (UNS S41003) can contain some austenite and resembles the passage of slip dis-
martensite if their heat treatment is such that locations through the crystal. The sum of many
austenite is allowed to form and is followed by such dislocations is shear, and this can be
rapid cooling. Even 409 (UNS S409XX) can macroscopically visible as in Fig. 1.
form some austenite if chromium is at the high The formation of martensite is essentially me-
end of its possible range and nickel and man- chanical (i.e., via deformation, not diffusion).
ganese residual levels are high. The martensitic The shear and volume expansion, about 4%,
alloys themselves can be made to be partially which accompanies the transformation, in-
ferritic by forcing their carbon contents to low volves a great deal of strain energy that must be
levels as is customarily done with 410S (UNS taken into account. This is shown diagrammati-
S41003). Not understanding these alloys can cally in Fig. 2 (Ref 2).
lead to unexpected consequences in mechanical
properties or corrosion performance.

Martensite Formation

Martensite as a phenomenon deserves a brief

review. Martensite forms as result of the diffu-
sionless transformation of austenite. The
austenite may be supersaturated with carbon or
nitrogen, but that is not necessary for the trans-
formation. The driving force for the transforma-
tion is simply the much lower free energy of the
ferrite phase over the austenite phase, which
can be attributed largely to large mutual repul-
sion between iron atoms that possess unpaired
outer electrons with the same quantum number
and magnetic polarity. This free-energy differ-
ential increases with decreasing temperature. At
a certain temperature, the martensite start tem- Fig. 1 Martensite platelets emerging from the surface.
Source: Ref 2
perature Ms, the transformation occurs sponta-
neously via the coordinated movement of atoms
in a shearing-type mode at very high speeds ap-
proaching the speed of sound in the material.
The composition of the martensite is identical to
that of the parent austenite.
There is regularity to the relationship between
the parent austenite and the martensite.
Greninger and Troiano determined that the
close-packed planes of the austenite {111} var-
ied from the {011} of the martensite by only
0.2°. Further, the direction of the <101 bar> of
the austenite was only 2.7° from the <1 bar 11
bar> of the martensite. These relationships de-
fine the habit plane that constitutes the austenite
martensite boundary.
Martensite forms essentially independent of Fig. 2 The martensite reaction ab contrasted to the nucleation
time and the fraction transformed depends only and growth-type transformation of austenite to ferrite, ac
Chapter 9: Martensitic Stainless Steels / 127

This energy differential between ferrite and stitial sites and forms various carbides, leaving
martensite is stored in the high-strain energy the parent martensite less strained, softer, and
matrix. Applied strains affect the transforma- tougher. Figure 4 shows that the large strain en-
tion. Indeed, metastable austenite can readily be ergy in martensite varies with the carbon con-
transformed to martensite by deformation. tent, and Fig. 5 (Ref 4) shows how hardness
However, the untransformed austenite is hin- varies with carbon content.
dered from transforming by the compression it Nitrogen behaves similarly to carbon in both
receives from the already-formed martensite. austenite and martensite, but its solubility is
Thus, some residual austenite is commonly lower, and it is less significant as an alloying el-
found between lathes of martensite. ement accordingly. Hydrogen and boron, as in-
At the Ms temperature, the body-centered terstitials, also raise hardness.
cubic (bcc) phase becomes preferable energeti-
cally, but this temperature is too low for diffu-
sion transformation, and a slight shear in the Phase Structure
austenite lattice causes a rearrangement of the
atoms from a face-centered cubic (fcc) to a dis- Figure 6(a) to (h) shows a series of photomi-
torted bcc structure. The amount of distortion is crographs of various martensitic alloys (Ref 5).
proportional to the amount of carbon in the in- A stainless martensitic alloy should have the
terstices of the structure. These interstices are following characteristics:
considerably smaller in the bcc structure even
though it is expanded from the fcc. The octahe- • It must have at least 10.5% chromium to
dral sites change from 2.86 by 3.56 A to 2.86 by qualify as stainless and even more for better
2.86 A, as shown in Fig. 3 (Ref 3). The distor- corrosion resistance.
tion is accommodated by accommodation from • It should be fully austenitic at some temper-
site to site at low carbon levels, but above about ature.
0.018% carbon this can no longer be accommo- • The temperature at which austenite forms on
dated and a tetragonal distortion occurs (Ref 3). heating should be sufficiently high to permit
The carbon is in a state of supersaturation in the tempering above the temper embrittlement
as-formed martensite. When the martensite is range.
tempered, the carbon diffuses from these inter- These criteria are somewhat challenging. Fig-
ure 7(a) shows that at low-carbon (0.05%) lev-
els austenite is stable up to about 12%
chromium, above which some δ-ferrite tends to
be stable at all temperatures below the melting
point. Increasing carbon slightly expands the
chromium level at which full austenitization can
occur (Fig. 7b) (Ref 3).

Fig. 3 Change in size of the octahedral interstitial site with

the change from face-centered cubic (fcc) to body- Fig. 4 Strain energy of martensite dependence on carbon
centered cubic (bcc). Source: Ref 2 content. Source: Ref 2
128 / Stainless Steels for Design Engineers

Fig. 5 Variation in martensite hardness with carbon content

The interplay between chromium and carbon The variety of martensitic stainless steels
is further explained in Fig. 8(a) and (b), in would be very limited if only chromium and
which it becomes clear that for higher- carbon were available as alloying elements, but
chromium alloys the range over which full fortunately nickel again can make an important
austenitization can occur is further restricted. contribution. Nickel greatly expands the
Chapter 9: Martensitic Stainless Steels / 129

Fig. 6 (a) Annealed 410 showing carbides within an equiaxed ferrite matrix. (b) 410 quenched and tem-
pered. (c) 416 quenched and tempered: white ferrite and gray sulfides in a martensite matrix. (d) 420
quenched and tempered showing fine carbides in a martensite matrix. (e) 420 quenched and tempered showing
surface decarburization. (f) 440A annealed displaying primary and smaller secondary carbides in a ferrite ma-
trix. (g) 440B quenched and tempered displaying both primary and secondary carbides. (h) 440C quenched and
tempered displaying significant primary carbides plus finer secondary carbides in a martensite matrix
130 / Stainless Steels for Design Engineers

Fig. 7 Iron-chromium phase diagrams at two low-carbon levels

1800 1800

1600 Liquid 1600 Liquid

1400 1400
Temperature, °C

Temperature, °C

γ γ
1200 1200 α+γ
γ + carbide
1000 1000
γ + carbide

800 800
α + carbide α + carbide
600 600
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mass, %C Mass, %C

Fig. 8 (a) Iron-chromium phase diagram at 12% chromium; (b) iron-chromium diagram at 17% chromium

chromium levels and temperatures at which amount of total alloy that can be used and in the
austenite is stable as is shown in Fig. 9 (Ref 3). end puts an upper limit on the ability of marten-
Table 2 quantifies the influences of the vari- sitic stainless steels to achieve high corrosion re-
ous possible alloying elements on the key prop- sistance. This is because as the main corrosion
erties of martensitic stainless steels. fighters, chromium and molybdenum, which are
It can be seen that the elements that promote ferritizers, are increased, so must austenitizers
austenite, with the exception of cobalt, all de- such as nickel. The coordinated increase in these
press the Ms temperature. This puts a limit on the elements lowers the Ms to such a degree that the
Chapter 9: Martensitic Stainless Steels / 131

alloys become stably austenitic before much Austenitizing is complicated in martensitic

higher corrosion resistance is obtained. stainless steels because many grades contain
The ability to temper without austenite rever- carbon at levels intended to produce carbides
sion is an important trait. Obviously, if trans- for wear resistance purposes. Since carbon solu-
forming martensite to austenite during temper- bility varies strongly with temperature at
ing caused subsequent untempered martensite austenitizing temperatures (Fig. 7a and b), con-
or other undesirable phases, this would limit trol of temperature is vital to have the correct
one’s ability to temper at a high enough temper- balance of carbon in solution versus carbon as
ature to achieve desired toughness. This limits carbide since carbon in solution has such a
the use of nickel while encouraging the use of strong influence on ferrite content, Ms, and me-
elements like molybdenum. chanical properties.
Copper has become an important alloying el- Austenitizing temperature also determines
ement in martensitic stainless steels because it austenite grain size. This affects Ms, but more
greatly improves corrosion resistance in certain importantly it influences subsequent toughness.
environments without diminishing an alloy’s Phosphorus precipitates at prior austenite grain
ability to be tempered. boundaries during tempering with a maximum
effect at 475 °C. This is the infamous temper
embrittlement. Figure 10 (Ref 3) shows the sig-
Thermal Processing nificant toughness change that occurs as in-
creasing austenitizing temperature increases
The main concerns with processing marten- austenite grain size and permits greater phos-
sitic stainless steels are austenitizing, quench- phorus concentrations at grain boundaries.
ing, tempering/stress relieving, and annealing. Refining phosphorus from any chromium-
containing steel is quite challenging thermody-
namically, so achieving low phosphorus levels
depends mainly on restrictions on raw materials
for melting. Because this is difficult or costly,
grain size control is the main tool for control-
ling temper embrittlement.
The higher-carbon grades, those above 0.20%
carbon, should be heated gradually through
stage heating to avoid cracking due to thermal
stresses. Soaking at 800 °C until uniform tem-
perature is achieved minimizes this risk.
Another concern during austenitizing is su-
perficial carbon loss, an example of which is
shown in Fig. 6(e). Heating in air to 1050 °C
can cause surface carbon to decrease by approx-
imately 0.10% per hour, resulting in much
lower surface hardness. This loss increases with
base carbon level and austenitizing temperature.
Carbon or nitrogen pickup could also occur if
the atmosphere was rich in these elements. The
carbon potential of the furnace atmosphere must
be controlled to avoid potentially serious prob-
Fig. 9 The expansion of the range of austenite stability with lems. If hydrogen atmospheres are used the
nickel content

Table 2 Influence of alloying elements on ferrite, Ms, and austenite start

Element N C Ni Co Cu Mn Si Mo Cr V Al
Lowering of % –220 –210 –20 –7 –7 –6 6 5 14 18 54
ferrite per % element
Lowering of MS –475 –475 –17 0 to 10 –17 –30 –11 –21 –17 –46 ...
per % element
Change of AC 0 to 280 0 to 250 –30 to –115 0 0 –25 to –66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750
per % element
132 / Stainless Steels for Design Engineers

P: 0.047%

[ ]
P: 0.035%

[ ]
P: 0.021%
Transition temperature, °C


P: 0.003-0.004%
[ ]
0 [ ]

[ ]

−40 [ ]

5 10 20 50 100 200

Austenite grain size, μm

Fig. 10 Influence of austenite grain size and phosphorus level on toughness

danger of embrittlement after quenching must as the transformation of M3C into M7C3 can re-
be recognized. Stress relief without delay would sult in a secondary hardening, a true precipita-
be mandatory. tion-hardening effect. In the presence of strong
The high chromium content of these alloys carbide-forming alloying elements such as
renders them very deep hardening. Air harden- molybdenum, vanadium, and tungsten, the M2X
ing is generally sufficient. Oil quenching which carbide can become the more stable species and
is faster may be slowed by heating the oil. be responsible for the secondary hardening. At
Avoiding quench cracking and excessive 500 °C, coarser M23C6 and M7C3 begin to grow
warpage is almost always a greater concern than at grain boundaries. This is accompanied by a
depth of hardening so air quenching is standard. pronounced softening. The hardening reduction
Because the quenching and the transforma- with stress relief and tempering for a 12% Cr
tion it causes are inevitably accompanied by alloy is shown in Fig. 11 (Ref 6).
residual stresses in a brittle material, stress re- Separately at the 475 °C range, the previously
lieving should be immediate to avoid cracking. mentioned phosphorus segregation to prior
Higher-carbon grades should not even be al- austenitic grain boundaries occurs. This effect
lowed below room temperature before stress begins to disappear above 550 °C. Thus true
relief. Pickling should never be done on as- tempering is conducted above this temperature.
quenched material because this could easily re- The microstructural changes at these tempera-
sult in hydrogen uptake and delayed cracking tures are the above-mentioned loss of carbon
by hydrogen embrittlement. from solid solution, carbide precipitation and
Heating as-quenched material to between 150 coarsening, and, of course, stress relief. The re-
and 400 °C produces stress relieving. Besides sult is a pronounced softening and toughening.
the normal flow on a microscopic scale, which If the material contains retained austenite, it
we understand as stress relieving, there is a may decompose to ferrite and carbide with a
slight growth in the number of fine cementite negative effect on toughness.
particles and a corresponding decrease in the The molybdenum, vanadium, and tungsten-
amount of carbon in solid solution. This results alloyed grades will resist softening during tem-
in a slight decrease in hardness. At 400 °C, a pering because of the strength of the secondary
further precipitation of M2X and M7C3 as well hardening they undergo due to precipitation
Chapter 9: Martensitic Stainless Steels / 133

Initial hardness

450 300°C
300°C 350°C
400 450°C 450°C

300 700°C

200 700°C

11 12 13 14 15 16 17 18 19 20 21 22 23

T (20 + LOG t) × 10−3

Fig. 11 Influence of tempering on hardness

hardening of carbides and nitrides. Nickel most sophisticated martensitic stainless alloys.
seems to amplify this action by its influence on While 420 is the common alloy and is quite
diminishing the solubility of carbon in the ma- serviceable, much more wear- and corrosion-re-
trix. Thus, the tempering of the higher-alloy sistant alloys exist. At one time, 440C was the
martensitic stainless steels can truly be consid- maximum step up from 420; however, further
ered a precipitation-hardening reaction. alloying with molybdenum for corrosion resist-
The higher-carbon, higher-chromium grades ance and vanadium for hardness of the carbide
are typically only stress relieved because the re- phase has led to improvements. The wear resist-
moval of chromium from solution by carbide ance of a blade is largely determined by the
formation at higher temperatures causes an un- hardness and amount of carbides while the
acceptable loss of corrosion resistance. toughness is governed by the matrix properties.
These alloys are used at very high hardness
levels, so cleanliness is very important to tough-
Applications ness, which measures the ability to withstand
chipping in use. Electroslag remelting (ESR) or
High-Temperature Use. The basic 12 % Cr vacuum induction melting-vacuum arc remelt-
martensitic alloy has been the basis of alloying ing (VIM-VAR) provides the cleanliness re-
improvements that were done to produce better quired, while powder metallurgy is optimal for
high-temperature performance, especially for obtaining very fine carbide size and uniformity.
turbines. The nominal analyses of some prominent
The addition of vanadium and niobium, both grades are shown in Table 3.
of which form much more stable carbides than The martensitic alloys have a tendency to-
chromium, results in alloys that have vastly im- ward centerline segregation during solidifica-
proved creep resistance in the 550 °C range, as tion as well as toward the formation of primary
shown in Fig. 12 (Ref 3). carbides. This has produced limitations in the
Tool and Cutlery Alloys. A high-profile use amount of highly wear-resistant constituents
of martensitic stainless steels is in cutlery. Hunt- such as vanadium carbide (hardness Rc 75),
ing knives, sport knives, and chefs’ cutting tools which can be introduced into the matrix in con-
are highly valued items and contain some of the ventional production. Powder metal techniques
134 / Stainless Steels for Design Engineers


Applied stress, kg/mm2



1 (0.2C-10.5Cr)
2 (0.2C-10.5Cr-0.1Nb)
3 (0.2C-10.5Cr-0.1V)
4 (0.2C-10.5Cr-0.1V-0.1Nb)

3 10 30 100 300 1000 3000 10,000

Rupture life, h

Fig. 12 Influence of vanadium and niobium on high-temperature properties

Table 3 Tool and cutlery martensitic stainless steels alloy compositions

Alloy UNS Form C Mn S Si Cr Mo Ni Other Other
420 S42000 Wrought 0.15 min 1 0.03 1 12.0-14.0 ... ... ... ...
4116 DIN 1.4116 Wrought 0.5 ... ... ... 14.5 0.65 ... ... 0.15 V
440A S44002 Wrought 0.60-0.75 1 0.03 1 16.0-18.0 0.75 ... ... ...
440C S44004 Wrought 0.95-1.20 1 0.03 1 16.0-18.0 0.75 ... ... ...
BG-42 Nominal Wrought 1.15 ... ... 0.3 14.5 4 ... 1.2 V ...
ATS-34 Nominal Wrought 1.05 0.4 ... 0.35 14 4 ... ... ...
14-4 Nominal Wrought 1.05 0.5 ... 0.3 14 4 ... ... ...
154 CM Nominal Wrought 1.05 0.45 ... 0.3 14 4 ... ... ...
CPM Nominal PM 1.45 ... ... ... 14 2 ... 4.0 V ...
CPM Nominal PM 2.15 0.4 ... ... 17 0.4 ... 5.5 V ...
CPM Nominal PM 2.2 ... ... ... 13 1 ... 9.0 V ...

are not subject to the same limitations as contin- good toughness without tempering and mini-
uous casters and have alloys the production of mize the loss of chromium to carbides, main-
alloys with high volume content of VC. One taining it in solution for corrosion resistance.
such alloy is Crucible CPM 90V with 14% Cr, The addition of nickel and molybdenum yields
9% V, 1% Mo, and 2.3% C. This alloy has equal full austenite and martensite transformation and
or better toughness and corrosion resistance as improves corrosion resistance. Table 4 lists sev-
440C but has ten times the wear resistance at eral such alloys by JFE: the first two can be
the same macrohardness. made to meet L80 specifications and produced
Oil Country Tubular Good and Line Pipe. as seamless. JFE reports production of over
The need for corrosion resistance in oil produc- 100,000 tons per year of this product (Ref 7).
tion tubulars has grown as the quality of petro- The third alloy is a near match for the precipita-
leum deposits has become less optimal. Use of tion-hardening stainless Custom 450 (UNS
stainless can eliminate for corrosion inhibitors S45000) (Ref 8). Like other precipitation-hard-
in H2S and CO2 environments. This has led to enable steels, it shows excellent resistance to
the use of low-carbon martensitic stainless SCC at high strength levels. Figures 13 and 14
steels. Low carbon and nitrogen levels give show the improvements in corrosion resistance
Chapter 9: Martensitic Stainless Steels / 135

Table 4 JFE Steel/Nippon Steel oil country tubular goods and line pipe alloys
Alloy UNS Form C Mn S Si Cr Mo Ni Other Other
HP13Cr-! JFE Wrought 0.025 0.45 ... ... 13 1 4 ... ...
HP13Cr-2 JFE Wrought 0.025 0.45 ... ... 13 2 5 ... ...
NT-CRS Nippon Wrought 0.03 1.45 ... ... 12.7 1.4 4.5 1.5 Cu 0.040 N
NT-CRSS Nippon Wrought 0.02 2 ... ... 12.3 2 5.8 1.5 Cu 0.015 N
KL-12Cr JFE Wrought 0.01 ... ... ... 11 ... 2.4 0.5 Cu 0.010 N
KL-HP JFE Wrought 0.01 ... ... ... 12 2 5.5 ... 0.010 N
12Cr Nominal

ronments to about Rc 22 to avoid SCC by the hy-

drogen embrittlement mechanism. The stainless
can resist this failure mode at higher strengths. As
specifying bodies such as the American Petroleum
Institute (API) approve the use of stainless tubu-
lars at higher strength levels than carbon steel
tubulars are safely capable of handling, then the
strength improvement, coupled with the orders of
magnitude improvement in corrosion resistance,
will cause a great increase in their use.
Lower carbon levels permit the use of field
welds without tempering so that similar alloys
can be used for line pipe. These are corrosion
Fig. 13 Corrosion rates of stainless versus carbon steel resistant and yet meet X70 and X80 class speci-
fications. These alloys are the last two in Table
2. These uses of martensitic stainless steels for
oil production represent possibly the greatest
growth area for any kind of stainless steel in the
first decade of the 21st century.


1. D.R. Koistinen, R.E. Marburger, “A General

Equation Prescribing the Extent of the
Austenite/Martensite Transformation in
Pure Iron,” Acta Met, Vol 7, 1959, p 59
3. Bletton, Aciers Inoxidables, Les Editions de
Physique les Ulis, Paris, 1993, p 481
4. ASM Handbook Desk Edition, 1985, p 28–9
Fig. 14 Corrosion rates for stainless oil country tubular goods 6. K.J. Irvine et al., JISI, Vol 195, ISIJ Interna-
(OCTG) alloys under severe operating conditions
tional 1960, p 386–405
over carbon steel L80 oil country tubular goods 7. S. Deshimaru et al., “Steels for Production,
(OCTG) under test conditions representative of Transportation and Storage of Energy, JFE
difficult real-use environments (Ref 7). Technical Report (No. 2), March 2004
The improvements in martensitic steels for p 55–67
these applications are hardly more than a thor- 8. M. Kimura et al., “High CR Stainless
ough revisiting of the developments of the OCTG with High Strength and Superior
1950s and 1960s. This does not diminish their Corrosion Resistance,” JFE Technical Re-
importance. Carbon steels are limited in sour envi- port (No. 7), Jan 2006, p 7–13
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 137-146 All rights reserved.
DOI: 10.1361/ssde2008p137


Stainless Steels

Summary Introduction

THE PRECIPITATION-HARDENABLE (PH) The PH stainless steels exploit the low austen-
grades are a highly specialized family of stain- ite stability possible in the chromium/nickel
less steels whose existence derives from the stainless steels by making the alloys so lean in
need for very high-strength materials with good composition that they can be made to trans-
corrosion resistance. The workhorse alloys are form nearly entirely to martensite by thermal
the martensitic PH grades, which are used in or mechanical treatment. This martensite can
many forms. Primarily used as forgings, bar, then be further hardened by the coherent pre-
and other hot-worked forms, they can also be cipitation of intermetallic compounds, ele-
obtained in cold-rolled sheet and strip, although mental copper, nitrides, or even phosphides.
not with the flatness expected from non-PH This precipitation hardening can also be made
stainless. The semiaustenitic alloys are more to occur in a fully austenitic matrix, and this
amenable to production as sheet, strip, and wire also provides a commercial PH alloy. But, the
and are designed for applications that require martensitic PH grades are by far the more
extensive forming before hardening. The fully common. The border between the more highly
austenitic PH alloys fill a small niche where alloyed martensitic stainless steels, which un-
high mechanical properties are required at tem- dergo secondary hardening during tempering,
peratures above or below which the other PH and the PH alloys is indeed vague. Some au-
grades are found lacking, when a nonmagnetic thors have astutely treated them as a single
material is required, or when the higher thermal group. Here, we treat them separately because
expansive coefficient of an austenitic material is they are traditionally considered as separate
desired. In no case is corrosion resistance better alloys.
than that of normal 304 found in PH stainless The advantage of the PH alloys over the
steels. If enhanced strength and very high corro- strictly martensitic stainless steels is that they
sion resistance are required, then the designer attain great strength with higher toughness and
should look to duplex stainless steels for the op- corrosion resistance than can be obtained
timal material. If cost is a greater concern than through the hardening of martensite through
corrosion resistance or toughness, then marten- carbon. In addition, they can be fabricated in a
sitic stainless steels should be considered for relatively soft state and then hardened with very
applications where strength and hardness over little dimensional change.
that of annealed ferritic and austenitic stainless The PH grades were developed at the begin-
is required. ning of World War II, with Stainless W (UNS
The increased use of titanium alloys and ad- S17600) by U.S. Steel generally acknowledged
vanced composite materials may occur at the as the first. The later-developed grades are dis-
expense of the stainless PH alloys and at the tinguished from the first by their more uniform,
same time may create some new niche applica- and therefore tougher, microstructure through
tions for them. the elimination of residual δ-ferrite and retained
138 / Stainless Steels for Design Engineers

austenite and by more astute alloy design and movement. As time and temperature of precipi-
chemistry control. tation increase, the zones can grow to sizes that
The mechanism of precipitation hardening is cannot accommodate the small size differential;
parallel to that used to strengthen aluminum al- coherency is lost, and with it the hardening ef-
loys in which the precipitation of a coherent sec- fect diminishes. The precipitation has the dual
ond phase from a supersaturated solid solution is function of stress relieving the martensite while
produced by an aging heat treatment. The coher- further hardening the matrix through the precipi-
ent precipitate strains the lattice and impedes the tation of the coherent precipitate. The mechani-
motion of dislocations, producing strengthening. cal properties of the final microstructure depend
Overaging causes the precipitates to lose co- on the initial strength of the matrix before aging,
herency, and softening follows. The precipitate the amount of precipitate, and the coherency of
that causes the hardening is normally nickel the precipitate. The ideal microstructure for the
(aluminum/titanium) (Ref 1). Figure 1 shows the initial matrix is 100% martensite. To the extent
compounds that can form from the precipitation there is δ-ferrite or retained austenite, properties,
of supersaturated aluminum and titanium in an especially yield strength and toughness in the
iron alloy matrix. transverse direction, are compromised. The
It is also possible to produce a hardening reac- aging temperatures can also be high enough that
tion by the precipitation of elemental copper reversion of martensite to austenite occurs,
(Ref 2). In nitrogen-bearing alloys, a hardening which also lowers subsequent tensile properties.
may be produced by the precipitation of Cr2N While the presence of persistent, large bands
(Ref 3). The precipitation begins with the diffu- of either δ-ferrite or γ-austenite is undesirable,
sion of the precipitating species to sites on the but both also have benefits. The presence of
existing matrix. These enriched zones are called some fine bands of δ-ferrite promotes easier and
Guinier-Preston (GP) zones. Close dimensional more reproducible precipitation of chrome car-
matchup between the precipitating species and bides at the δ/γ interface during the “austenite
the parent matrix is required. The differential conditioning” or “trigger anneal” heat treatment
should be on the order of a percent. This allows step for semiaustenitic alloys (17-7, AM350,
not only coherency but also strain. The coherent etc.). Although bands of stable austenite are
precipitate is a effective barrier to dislocation undesirable, it is the presence of residual interlath

γ + Ni(AlTi) + Ni3AlTi Ni(AlTi)

Ni(AlTi) + Ni2AlTi
4 Ni2AlTi
Ni2(AlTi) + Ni3AlTi
γ + Ni2AlTi Ni3Ti
γ + Ni(AlTi)
Aluminum, wt%

Limit of austenite
ferrite region in
solution treated
2 conditions

γ + Ni2(AlTi) + Ni3AlTi

γ + Ni3(AlTi)

γ γ + Ni3Ti + Ni3(AlTi)

γ + Ni3Ti

0 1 2 3 4

Titanium, wt%

Fig. 1 Possible aluminum/titanium precipitates

Chapter 10: Precipitation-Hardening Stainless Steels / 139

austenite that provides the work-hardening abil- strength than that of which the martensitic or
ity in many of these PH alloys. It is this work semiaustenitic alloys are capable. The austenitic
hardening that gives the PH alloys, especially the PH strength is better above 750 °F.
semiaustenitic ones, their unusual combination
of high strength plus ductility and toughness in
the fully hardened state. Martensitic Precipitation-Hardenable
The complexity of PH steels comes from the Stainless Steels
processing involved in producing the martensitic
structure in which the precipitation will occur. The martensitic PH alloys are, as stated, fully
The most straightforward alloys are the marten- martensitic at room temperature. Their marten-
sitic, also called the martensitic PH alloys. These site is a relatively soft, low-carbon (less than
steels are supplied in the fully martensitic condi- 0.05%) martensite as opposed to the higher car-
tion with hardness in the low Rc 30s. This is con- bon found in the martensitic stainless steels. The
fusingly called the annealed condition, or condi- early alloys of this type, 17–7 PH and 17-4 PH,
tion A, even though the matrix is untempered contained up to 10% δ-ferrite stringers, which
martensite. After the material is fabricated, it is caused poor through-thickness toughness. This
subjected to an aging treatment designated by would be expected from the Schaeffler-Delong
the aging temperature in Fahrenheit (e.g., diagram, but this is asking too much of the Scha-
H-950). These aging temperatures range from effler-Delong diagram, which was developed for
950 °F (510 °C) to 1150 °F (620 °C). welds, to predict the phase composition of alloys
A second major group of PH grades is the that have been homogenized by hot working.
semiaustenitic. These grades in the normally The inaccuracy of the diagram for more com-
furnished condition A are fully austenitic. This plex systems was overcome, and alloys were de-
is accomplished by adding elements that lower signed that had minimal δ-ferrite and still trans-
the martensite start temperature, such as more formed entirely to martensite, if not at room
chromium, molybdenum, and nickel. The temperature, at least at a reasonably attainable
austenite is more formable than martensite, and subzero temperature. This was done first by trial
it has the possibility of superior corrosion resist- and error and more recently by using thermody-
ance because of higher chromium content. This namic computer models, such as ThermoCalc, to
is balanced by the need to use either cold work, predict equilibrium phase composition. This de-
cryogenic treatment, or a destabilizing anneal to velopment was very significant for making the
cause the matrix to become martensitic before alloy family useful as a high-strength/high-
its precipitation aging treatment. toughness material for demanding applications
Last, if the austenite is made very stable by requiring high mechanical properties and corro-
further alloying additions, a precipitation reac- sion resistance. The most advanced PH alloys
tion can still be made to occur by the same type are martensitic PH grades by Cartech, Custom
of aging treatment without martensite ever form- 465 and 475.
ing. The precipitation takes place in austenite Table 1 shows the more significant of these al-
and therefore results in lower room temperature loys compared on a strength basis; Table 2 shows

Table 1 Mechanical properties of martensitic precipitation-hardenable alloys

Alloy UNS Condition Yield, MPa Tensile, MPa Elongation, % HRC CVN ft-lb
Stainless W S17600 H-950 ( 510) 1240 1340 14 42 ...
17-4 PH S17400 H-925 (495) 1210 1310 14 41 40
15-5 PH S15500 H-925 (495) 1210 1300 15 41 20
H1100 (595) 930 1025 17.50 34 70
13-8 PH S13800 H-950 (510) 1450 1550 12 47 25
H1050 (565) 70
Custom 450 S45000 H-900 (480) 1270 1350 14 42 60
H1100 (595) 460 970 23 ... 180
Custom 455 S45500 H-950 (510) 1515 1585 10 48 8
H1050 (565) 1205 1310 14 40 25
Custom 465 S46500 H-950 (510) 1650 1765 11 49 13
H1000 (535) 1500 1600 13 48 28
Custom 475 ... H-975 (525) 1855 2005 5 54 ...
H 1100 (595) 1315 1572 13 48
Ferrium S53 ... ... 1565 1985 14–16 54 18
140 / Stainless Steels for Design Engineers

Table 2 Composition of martensitic precipitation-hardenable alloys

Alloy Designation C Mn Si Cr Ni Mo Al Cu Ti Other

Stainless W S17600 0.1 0.5 0.5 17 6.3 ... 0.2 ... 1 P 0.3
17-4 PH S17400 0 0.6 0.6 16 4.3 ... ... 3.2 ... ...
15-5 PH S15500 0 0.6 0.6 15 4.3 ... ... 3.2 ... ...
13-8 PH S13800 0 0.1 0.1 13 8.5 2 1.1 ... ... ...
Custom 450 S45000 0 0.3 0.3 15 6 0.8 ... 1.5 ... 0.3 Nb
Custom 455 S45500 0 0.3 0.3 12 8.5 ... ... 2.5 1 0.3 Nb
Custom 465 S46500 0 0.2 0.2 12 11 1 ... ... 2 ...
Custom 465 (275) ... 0 0.2 0.2 12 11 1 ... ... 2 0.2 Nb
Custom 475 ... 0 0.4 0.4 11 8 5 1.2 ... ... 8.0 Co
Ferrium S53 ... 0.2 0.1 0.1 10 5.5 2 0 ... ... 1W
... 0.3 V
... 14 Co

Fig. 2 Typical microstructures of precipitation-hardenable (PH) stainless steels: (a) 15-5PH as-quenched martensite; (b) 13-8 PH so-
lution treated and aged displaying fine martensite; (c) 17-7 PH displaying ferrite stringers in a martensite matrix; (d) 17-7 PH
showing residual ferrite stringers and inclusions

their nominal compositions. Figure 2 shows a se- less W to the latest in Custom 475 besides the
ries of photomicrographs of PH alloys. elimination of δ-ferrite is in the volume fraction
The main advancement metallurgically in of the precipitating phase and the elimination of
these alloys from the top, and earliest, in Stain- retained austenite. Alloy designers found that to
Chapter 10: Precipitation-Hardening Stainless Steels / 141

reduce δ-ferrite they also tend to stabilize • Molybdenum to offset loss of corrosion re-
austenite, which also happens to reduce the sistance by minimization of chromium, to
temperature at which martensite forms, MS. increase the temperature at which austenite
Last, the higher levels of molybdenum reduce forms, and to form another hardening pre-
the tendency to form secondary austenite during cipitate in the presence of cobalt
aging. To minimize δ-ferrite requires reducing • Cobalt to stabilize austenite while raising
chromium or molybdenum, which also reduces MS
corrosion resistance. As a result, as alloy • Aluminum or titanium to form intermetallic
strength increases, corrosion resistance is com- precipitates with nickel or copper to precipi-
promised. The alloys with the greatest strength tate as elemental copper
potential, Custom 465 and 475, barely qualify
It is possible to quantify these various influ-
as stainless with around 11% chromium. But,
ences on phases. This is summarized in Table 3
from a utility point of view, these alloys are de-
in terms of the influence of the element on dif-
signed to have maximum mechanical properties
ferent factors measured in degrees Centigrade
with adequate corrosion resistance, so this is
for a 12% chromium alloy.
viewed as an acceptable compromise.
Rapid quenching of these alloys is not re-
The newest alloy is Ferrium S53, one of the
quired. They are air hardenable. But, the cool-
recent alloys designed by computer-assisted
ing of these alloys must be completed expedi-
thermodynamic calculations. It was designed to
tiously through the final stages of martensite
replace 300M, 4340, and AerMet 100 on an
formation with minimal delay. During delays
equal mechanical properties basis but also pro-
after the start of martensite transformation has
vide the corrosion resistance necessary to be
occurred, the remaining austenite tends to stabi-
used in aircraft components without cadmium
lize, and full transformation to martensite does
plating. Its composition superficially seems de-
not occur. When this happens, the higher levels
ficient in chromium to provide “stainlessness,”
of austenite reduce subsequent mechanical
but the cobalt level raises the thermodynamic
properties after aging.
activity of chromium sufficiently that the equiv-
As with any alloy used at such high strength
alent of 12% chromium in a non-cobalt-contain-
levels, microstructural cleanliness is essential,
ing alloy is achieved. The precipitation harden-
but air melting and argon oxygen decarburiza-
ing mechanism is the precipitation of Mo2C. It
has been established that this hardening mecha- tion (AOD) refining are quite adequate.
nism optimizes resistance to stress corrosion Corrosion Resistance. The martensitic PH
cracking (SCC) for a given strength level. stainless steels obey the same rules as other
The alloying characteristics of these grades stainless steels with regard to corrosion resist-
are: ance. The martensite carries no nitrogen in solu-
tion, so the resistance to pitting is given by:
• Low carbon, nitrogen, silicon, and man-
ganese because these elements lower MS PREN = %Cr + 3.3%Mo (Eq 1)
without contributing to age hardening
• Low chromium to suppress δ-ferrite In the martensitic PH alloys, no chromium is
• Sufficient nickel to suppress δ-ferrite and rendered ineffective by the formation of Cr23C6
provide for precipitates without excessively since carbon is either held low or stabilized by
depressing MS titanium or niobium. Thus, the corrosion

Table 3 Influence of alloying elements on key transformations

Element N C Ni Co Cu Mn Si Mo Cr V Al
Lowering –220 –210 –20 –7 –7 –6 6 5 14 18 54
of % ferrite
per % element

Lowering of MS –475 –475 –17 0 to 10 –17 –30 –11 –21 –17 –46
per % element
Change of AC 0 to 280 0 to 250 –30 to –115 0 0 –25 to –66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750
per % element
142 / Stainless Steels for Design Engineers

resistance will equal that of stabilized ferritic al- ments, the austenite could be made stable at
loys of the same pitting resistance equivalent room temperature. This would make the alloy
number (PREN) for which voluminous data are softer and more fabricable and, most impor-
available. tantly, permit them to be manufactured as cold-
A greater concern is the risk of SCC in these rolled sheet and strip. If the austenite could then
alloys. While the mechanism of SCC in be transformed to martensite by cryogenic treat-
austenitic alloys is still debatable, it has long ment, cold work, or special heat treatment, then
been clear that, for martensitic alloys, SCC is it could be age hardened just like the martensitic
simply a manifestation of hydrogen embrittle- PH grades. This has been accomplished for a
ment in which the hydrogen is provided by local group of alloys called the semiaustenitic PH
corrosion. The existence of pitting is a suffi- grades. The “semi” signifies that the austenite
cient, if not necessary, condition for SCC to in these alloys is metastable rather than stable at
occur if the temperature is within the range of ambient temperatures. Also, it should be noted
susceptibility and the material is inherently sus- that these semiaustenitic alloys usually contain
ceptible. The material susceptibility is largely a some δ-ferrite in their predominantly austenitic
function of resistance to crack propagation in microstructure after annealing.
any given alloy that is measured by fracture These alloys are complex metallurgically be-
toughness. The martensitic PH grades have ex- cause of the technique used to make the austenite
cellent toughness and low rates of crack propa- stable at room temperature after a full solution
gation, but none should be considered immune anneal. The austenite is rendered stable by fairly
to SCC since their hardness is never below the high levels of carbon, a powerful austenite stabi-
Rc 22 level, which is considered to be the lizer, in solution. The amount of carbon that can
threshold hardness for susceptibility to SCC in be held in solution is a function of annealing tem-
body-centered cubic (bcc) ferrous alloys. perature. The 1050 °C anneal of the condition A
The suitability of high-strength alloys for use mill anneal puts all the carbon in solution, giving
in potential SCC-provoking environments con- the austenite the stability of a normal 301-type
taining H2S is regulated in many locales by Na- alloy. This permits extensive forming. The key is
tional Association of Corrosion Engineers to apply a subsequent lower-temperature anneal
(NACE) International Standard MR01-75. In it, so that less carbon goes into solution. Some of it
the use of S17400 is permitted if it is double will thus form M23C6. This is in a sense deliber-
tempered at 620 °C and its hardness is 33 HRC ately sensitizing the alloy, but the sensitization
or less, while S45000 can be used if it has been takes place at such a high temperature that
aged at 620 °C for 4 h and its hardness is 31 chromium deficits around precipitated carbides
HRC or less. These permissible hardness levels are minimized by diffusion. This causes a higher
are significantly higher than allowed in non-PH Ms temperature because of the lower amount of
martensitic alloys, 22 HRC, which reflects the carbon, and chromium, in solution in the austen-
fact that the martensitic matrix has the tough- ite. Depending on the temperature at which the
ness of a lower-hardness martensite. anneal is done, the Ms temperature can be con-
In marine environments, the PH alloys are trolled so that a transformation to martensite can
susceptible to SCC if used at a high strength be raised to either room temperature or some at-
level. S17400 aged at 480 °C with a yield tainable cryogenic temperature. Figure 3 shows a
strength of 1240 MPa is susceptible to SCC, chart of these heat treatment options.
while higher aging temperatures (above 540 °C) The lower strength levels achieved by the
producing lower strengths renders the materials condition T route in Fig. 3 reflect the lower car-
immune at stresses near the yield strength, ap- bon content of the martensite, while the highest
proximately 1170 MPa. The threshold strength strength of the condition C route reflects the
for ordinary martensitic stainless steels would compound influence of cold work and marten-
be 1030 MPa (Ref 4). site hardness with similar subsequent contribu-
tion from age hardening.
The main alloys of this group are listed in
Semiaustenitic Precipitation- Table 4. Examination of the chemistries in this
Hardenable Stainless Steels table shows that the first two alloys rely on the
precipitation of Ni3Al for the hardening, while
If a martensitic stainless steel were alloyed the last two have no apparent precipitating
more strongly with austenite-stabilizing ele- components. Their hardening is a more subtle
Chapter 10: Precipitation-Hardening Stainless Steels / 143

Fig. 3 Processing routes for S15700 Source: Ref 5

Table 4 Compositions of semiaustenitic precipitation-hardenable alloys

Alloy Designation C Mn Si Cr Ni Mo Al N
17-7 PH S17700 0.1 0.5 0.3 17 7.1 ... 1 0
15-7 PH S15700 0.1 0.5 0.3 15 7.1 2.2 1 0
AM-350 S35000 0.1 0.8 0.4 17 4.3 2.8 ... 0.1
AM-355 S35500 0.1 0.9 0.4 16 4.3 2.8 ... 0.1

secondary hardening from the tempering of titanium would preferentially deplete the alloy
martensite rather than the classic precipitation of carbon and nitrogen, precluding the action of
hardening via precipitation of intermetallic the conditioning heat treatment, which relies on
compounds. In 17-7 and 15-7, aluminum rather manipulating the amount of carbon in solution.
than titanium is the precipitating agent because In AM350 and AM355, it is the precipitation of
144 / Stainless Steels for Design Engineers

Table 5 Mechanical properties of semiaustenitic precipitation-hardenable alloys

Yield, Tensile,
Alloy UNS Condition MPa MPa Elongation, % HRC
17–7 PH ... TH 1050 (565) 1100 1310 10 42
RH 950 (510) 1380 1520 9 46
CH 900 (480) 1585 1655 2 49
15–7 PH ... TH 1050 (565) 1380 1450 7 45
RH 950 (510) 1550 1650 6 48
CH 900 (510) 1720 1790 2 50
AM-350 ... SCT 850 (450) 1210 1420 12 46
SCT1000 (540) 1020 1165 15 40
AM-355 ... SCT 850 (450) 1250 1510 13 48
SCT 1000(540) 1035 1124 22 38

(Cr,Fe)2N within the martensite phase that is that all carbon is present as chromium carbide,
responsible for age hardening. In addition, and that the chromium content is diminished by
molybdenum produces a secondary hardening that amount before applying Eq 1. In addition,
in carbon-bearing martensite. These two early carbon can remove some molybdenum in the
alloys did not possess the hardening potential form of carbides, and nitrogen can remove
that alloys employing copper-, titanium-, or chromium as a nitride. The effective corrosion
aluminum-based precipitates enjoy. resistance of these alloys thus is similar to fer-
The mechanical properties of these alloys ritic 430. The designer is thus advised to consult
are not greatly different from the martensitic al- with producers about corrosion resistance de-
loys, as can be seen in Table 5. Their separate pending on the thermal processing that will be
existence is due to the need for alloys that are used, especially if double aging is performed,
more fabricable at room temperature than are the which can cause some degree of chromium de-
alloys that are martensitic at room temperature. pletion at grain boundaries.
This benefit is offset by the necessity to condi- While the general and pitting corrosion re-
tion anneal before age hardening. If corrosion sistance of the semiaustenitic PH alloys are
resistance is a high concern, then the alloy and never quite as good as most austenitic stainless,
heat treatment that yields the greatest amount of they have very good resistance to SCC com-
chromium in solution should be chosen. Thus pared to ordinary martensitic stainless steels.
condition CH is better than RH, which is better The δ-ferrite and the generally well-tempered
than TH, the order of ascending solution anneal- martensitic matrix provide a crack-arresting
ing temperature and ascending Ms. feature and good inherent toughness that resist
The martensitic PH grades have somewhat SCC at higher strength levels than in straight
better strength because they have a more uni- martensitic stainless steels. AM-355 in the SCC
formly martensitic structure, and they employ a (850 °F) condition can withstand stresses of
tougher, lower-carbon martensite. The fact that 75% of 0.2% offset yield strength in salt spray
the semiaustenitic alloys are typically sheet without SCC failure.
products generally makes their service tough-
ness requirements less onerous, so that their re-
tained δ-ferrite is not a crippling drawback be-
cause of its detrimental effect on through Austenitic Precipitation-Hardenable
thickness toughness. Stainless Steels
Corrosion Resistance. The semiaustenitic
PH alloys tend to higher values of PREN than The austinitic PH class consists of just one
the martensitic alloys inherently since they are important commercial alloy, A-286. The impor-
alloyed to have lower Ms temperatures. The tance of the alloy is that it is entirely stable
thermal processing of these alloys causes a sig- austenite in both the solution-annealed and the
nificant portion of the chromium to be removed age-hardened condition. This means it is very
from solution as chromium carbide. This lowers formable and nonmagnetic. And, because the
the corrosion resistance from what would be ex- precipitation takes place in an austenite matrix,
pected based on the bulk composition. From an the precipitation takes place at a higher temper-
engineering point of view, it is best to assume ature, around 700 °C. This gives the alloy the
Chapter 10: Precipitation-Hardening Stainless Steels / 145

Table 6 Austenitic precipitation-hardenable composition

Alloy UNS C Mn Si Cr Ni Mo Al V Ti
A-286 S66286 0.05 1.5 0.5 15 25.5 1.3 0.15 0.3 2.15
Discalloy S66220 0.04 1.6 0.5 14 26 3 ... ... 1.7

Fig. 4 A-286 properties as a function of test temperature. Source: Ref 5

potential to be used to near 700 °C without fear PH alloys, makes it attractive for a variety of
of overaging. Thus, austenitic PH stainless rep- aerospace and nonaerospace uses.
resents a way to strengthen the austenite matrix, The hardening mechanism of A-286 is the
which has the following advantages: precipitation of Ni3 (aluminum, titanium). Dif-
fusion, even at the higher temperatures, is
• High ductility and therefore high formability slower in austenite, so aging treatments are typ-
in the soft, unaged condition ically 16 h. Table 6 gives the typical composi-
• High toughness at all temperatures and tion of A-286, and Fig. 4 shows some typical
strength levels properties as a function of temperature for a
• Excellent creep and stress rupture properties standard 980 °C solution treatment followed by
• Excellent oxidation, corrosion, and SCC re- a 720 °C, 16 h aging.
sistance The mechanical properties can be greatly en-
This alloy is rightly considered an iron- hanced by cold working prior to aging, as is
based superalloy and is the root of the group shown in Fig. 5.
that succeeds it in properties, the nickel- and The toughness of the austenitic matrix is
cobalt-based superalloys. These last alloys can abundant and quite temperature insensitive.
attain greater strength and creep resistance Charpy V-notch values of over 60 J are typical
than A-286, effectively ending further devel- from –200 to 800 °C.
opment of austenitic PH alloys. The lower cost The corrosion resistance of A-286 is compa-
of the A-286 alloy, compared to nickel-base rable to that of 304 and 316. It has slightly
146 / Stainless Steels for Design Engineers

Fig. 5 The influence of cold work on aging response in A-286. DPH, diamond pyramid hardness. Source: Ref 5

better resistance to SCC despite its higher allurgical and Materials Transactions A,
strength level. Vol 30A, Feb 1999.
3. G. Aggen, Ph.D. thesis, Carnegie Mellon
4. E.E. Denhard, “Stress Corrosion Cracking
1. F.B. Pickering, Physical Metallurgical De- of High Strength Stainless Steels in Atmos-
velopments of Stainless Steel, Stainless ’84, pheric Environments”, paper presented at
Goteborg, Sept 3–4, 1984, p 2–28. the Twenty-fourth Meeting of the AGARD
2. M. Murayama, Y. Katayama, and K. Hono, Structures and Materials Panel (Turin,
Microstructural Evolution in a 17-4 PH Italy), April 17–20, 1967.
Stainless Steel After Aging at 400 °C, Met- 5. Allegheny Technology Blue Sheets
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 147-154 All rights reserved.
DOI: 10.1361/ssde2008p147


Casting Alloys

Summary essential element. Thus, the casting alloys listed

in Tables 1 and 2 (Ref 1) are recognizable as ap-
WITH TYPICAL ALLOY SYSTEMS, cast- proximate counterparts of the co-listed wrought
ing is often the most convenient method by alloys (AISI grade). This cross reference to
which to produce components. This is true for wrought equivalents is helpful when looking for
stainless steels—both for corrosion-resisting data about an alloy that may be more easily
and for heat-resisting applications. This chapter found for wrought alloys than for cast.
discusses primarily the alloys used for stainless The High Alloy Product Group of the Steel
steel castings and their metallurgy. Foundry Founder’s Society of America employs a nam-
methods are discussed to the degree they are ing system (ACI, the Alloy Casting Institute)
specific to the stainless alloys. for cast alloys that is significant; these designa-
tions are currently assigned by ASTM as grades
and are added to ASTM specifications. The first
Stainless Steel Casting Alloys letter, “C” or “H,” indicates corrosion resisting.
The second letter indicates the relative amount
Essentially any wrought stainless alloy compo- of nickel, from a minimum of 0 to 1% for “A”
sition can be modified to be made as a cast alloy. up to 30% nickel for “N” alloys. The number
The systemic difference between cast alloys and following the hyphen for “C” alloys designates
their wrought equivalents is that cast alloys gen- the maximum carbon in hundredths of a percent.
erally contain between 1.0 and 2.5% silicon. As The suffix letters designate additional alloying
with other ferrous alloys, this is done to increase elements, such as Cu for copper, M for molyb-
the fluidity of the melt to make it cast more effec- denum, N for nickel or nitrogen, F for free ma-
tively. Silicon has strong metallurgical effects, chining, and C for columbium (niobium). The
both beneficial and detrimental, which should be heat-resisting, “H,” alloys have generally only a
understood by the user of cast stainless steels. second letter designating relative nickel level on
These are explained. A second general observa- the same scale as “C” alloys but going past stain-
tion is that the stabilized ferritic stainless steel al- less steels all the way to nickel-based alloys. The
loys, which constitute almost half the tonnage of inclusion of a number after the first two letters
all stainless steel used, are notably absent from indicates the center of the carbon range ex-
the cast alloys. This is because these alloys are pressed in hundredths of a percent by weight.
single phase at all temperatures in the solid state To learn more about the influence of alloying
and because they have large as-cast grain sizes elements, refer to the chapters on the individual
that can only be refined by heavy cold work fol- alloy families; see Section 3. Here, we briefly
lowed by annealing. This makes them quite lack- summarize:
ing in toughness as cast. Since heavy cold work • Pitting and crevice corrosion resistance, as
defeats the purpose of casting to achieve a near- well as general corrosion resistance, are en-
net shape, stabilized ferritic stainless steels are hanced by chromium, molybdenum, tung-
seldom used as castings. Also, the standard stabi- sten, and nitrogen and carbon in solution.
lizing alloy, titanium, is too readily oxidized for • Localized corrosion is caused by chromium
normal foundry practice to avoid the loss of this depletion, which occurs when precipitates
148 / Stainless Steels for Design Engineers

Table 1 Compositions of cast stainless corrosion resisting alloys

Composition(a), wt%—maximum or range
ACI designation AISI grade UNS %C %Mn %Si %Cr %Ni %Mo %Other
Chromium alloys
CA-15 410 J91150 0.15 1.00 1.50 11.5–14.0 1.0 0.50
CA-15M J91151 0.15 1.00 0.65 11.5–14.0 1.0 0.15–1.00
CA-40 420 J91153 0.40 1.00 1.50 11.5–14.0 1.0 0.50
CA-40F 420F J91154 0.2–0.4 1.00 1.50 11.5–14.0 1.0 0.20–0.40 Ss
CB-30 431,442 J91803 0.30 1.00 1.50 18.0–22.0 2.0
CC-50 446 J92613 0.30 1.00 1.50 26.0–30.0 4.0
Chromium-nickel alloys
CA-6N J91650 0.06 0.50 1.00 10.5–12.5 6.0–8.0
CA-6NM S41500 J91540 0.06 1.00 1.00 11.5–14.0 3.5–4.5 0.4–1.0
CA-28MWV 422 J91422 0.20–0.28 0.50–1.00 1.00 11.0–12.5 0.5–1.0 0.9–1.25 0.9–1.25 W,
0.2–0.3 V
CB-7Cu-1 17-4PH (AISI 630) J92180 0.07 0.70 1.00 15.5–17.7 3.6–4.6 2.5–3.2 Cu,
0.2–0.35 Nb,
0.05 N max
CB-7Cu-2 15-5 PH (XM-12) J92110 0.07 0.70 1.00 14.0–15.5 4.5–5.5 2.5–3.2 Cu,
0.2–0.35 Nb,
0.05 N max
CD-3MN 2205 (S32205) J92205 0.03 1.50 1.00 21.0–23.5 4.5–6.5 2.5–3.5 1.0 max Cu,
0.10–0.30 N
CD-3MCuN 255 (S32550) J93373 0.03 1.20 1.10 24.0-26.7 5.6–6.7 2.9–3.8 1.4–1.9 Cu,
0.22–0.33 N
CD-3MWCuN (S32760) J93380 0.03 1.00 1.00 24.0–26.0 6.5–8.5 3.0–4.0 0.5–1.0 Cu,
0.5–1.0 W,
0.20–0.30 N
CD-4MCu J93370 0.04 1.00 1.00 24.5–26.5 4.75–6.0 1.75–2.25 2.75–3.25 Cu
CD-4MCuN J93372 0.04 1.00 1.00 24.5–26.5 4.7–6.0 1.75–2.25 2.75–3.25
Cu, 0.10–0.25 N
CD-6MN J93371 0.06 1.00 1.00 24.0–27.0 4.0–6.0 1.75–2.25 1.75–2.5 Cu,
0.15–0.25 N
CE-3MN 2507 (S32750) J93404 0.03 1.50 1.00 24.0–26.0 6.0–8.0 4.0–5.0 0.10–0.30 N
CE-8MN J93345 0.08 1.00 1.50 22.5–25.5 8.0–11.0 3.0–4.5 0.10–0.30 N
CE-30 312 J93423 0.30 1.50 2.00 26.0–30.0 8.0–11.0
CF-3 304L J92500 0.03 1.50 2.00 17.0–21.0 8.0–12.0
CF-3M 316L J92800 0.03 1.50 2.00 17.0–21.0 8.0–12.0 2.0–3.0
CF-3MN 316LN J92700 0.03 1.50 1.50 17.0–21.0 9.0–13.0 2.0–3.0 0.10–0.20 N
CF-8 304 J92600 0.08 1.50 2.00 18.0–21.0 8.0–11.0
CF-8C 347 J92710 0.08 1.50 2.00 18.0–21.0 9.0–12.0 Nb
CF-8M 316 J92900 0.08 1.50 2.00 18.0–21.0 9.0–12.0 2.0–3.0
CF-10 304H J92590 0.04–0.10 1.50 2.00 18.0–21.0 8.0–11.0
CF-10M 316H J92901 0.04–0.10 1.50 1.50 18.0–21.0 9.0–12.0 2.0–3.0
CF-10MC 316H J92971 0.10 1.50 1.50 15.0–18.0 13.0–16.0 1.75–2.25 (10xC)–1.2 Nb
CF-10SMnN NITRONIC™ 60 J92972 0.10 7.00–9.00 3.50–4.50 16.0–18.0 8.0–9.0 0.08–0.18 N
CF-12M 316 0.12 1.50 2.00 18.0–21.0 9.0–12.0 2.0–3.0
CF-16F 303 J92701 0.16 1.50 2.00 18.0–21.0 9.0–12.0 1.5 max 0.2–0.35 Se
CF-20 302 J92602 0.20 1.50 2.00 18.0–21.0 8.0–11.0
CG-6MMN NITRONIC™ 50 J93790 0.06 4.00–6.00 1.00 20.5–23.5 11.5–13.5 1.5–3.0 0.1–0.3 Nb,
0.1–0.3 V,
0.2–0.40 N
CG-8M 317 J93000 0.08 1.50 1.50 18.0–21.0 9.0–13.0
CG-12 308 J93001 0.12 1.50 2.00 20.0–23.0 10.0–13.0
CH-8 309S J93400 0.08 1.50 1.50 22.0–26.0 12.0–15.0
CH-10 309H J93401 0.04–0.10 1.50 2.00 22.0–26.0 12.0–15.0
CH-20 309 J93402 0.20 1.50 2.00 22.0–26.0 12.0–15.0
CK-3MCuN 254SMO™ J94653 0.025 1.20 1.00 19.5–20.5 17.5–19.5 6.0–7.0 0.5–1.0 Cu,
0.18–0.24 N
CK-20 310 J94202 0.20 2.00 2.00 23.0-27.0 19.0–22.0
CN-3M 904L J94652 0.03 2.00 1.00 20.0–22.0 23.0–27.0 4.5–5.5
CN-3MN AL-6XN® J94651 0.03 2.00 1.00 20.0–22.0 23.0–27.0 6.0–7.0 0.18–0.24 N
CN-7M 320 N08007 0.07 1.50 1.50 19.0-22.0 27.5–30.0 2.0–3.0 3.0–4.0 Cu
CN-7MS J94650 0.07 1.50 3.50 18.0–20.0 22.0–25.0 2.5–3.0 1.5–2.0 Cu
CT-15C N08151 0.05–0.15 0.15–1.50 0.50–1.50 19.0–21.0 31.0–34.0 0.5–1.5 Nb
(a) Balance Fe for all compositions. Source: Ref 1
Chapter 11: Casting Alloys / 149

Table 2 Compositions of cast heat-resistant stainless and nickel base alloys

Composition(a), wt%—maximum or range
ACI designation Nearest AISI grade UNS %C %Cr %Ni %Si max
HA 504 J82090 0.20 max 8–10 1.00
HC 446 J92605 0.50 max 26–30 4–max 2.00
HD 327 J93005 0.50 max 26–30 4–7 2.00
HE 312 J93403 0.20–0.50 26–30 8–11 2.00
HF 302B J92603 0.20–0.40 19–23 9–12 2.00
HH 309 J93505 0.20–0.50 24–28 11–14 2.00
HI J94003 0.20–0.50 26–30 14–18 2.00
HK 310 J94224 0.20–0.60 24–38 18–22 2.00
HK-30 J94203 0.25–0.35 23–27 19–22 1.75
HK-40 J94204 0.35–0.45 23–27 19–22 1.75
HL N08604 0.20–0.60 28–32 18–22 2.00
HN J94213 0.20–0.50 19–23 23–27 2.00
HP N08705 0.35–0.75 24–28 33–37 2.00
HP-50WZ 0.45–0.55 24–28 33–37 2.50
HT 330 N08605 0.35–0.75 13–17 33–37 2.50
HT-30 N08603 0.25–0.35 13–17 33–37 2.50
HU N08005 0.35–0.75 17–21 37–41 2.50
HW N08006 0.35–0.75 10–14 58–62 2.50
HX N06050 0.35–0.75 15–19 64–68 2.50
(a) Balance Fe for all compositions. Manganese content: 0.35–0.65% for HA, 1% for HC, 1.5% for HD, 2% for the other alloys. Phosphorus and sulfur contents: 0.04
(max) for all but HP-50WZ. Molybdenum is intentionally added only to HA: 0.90–1.2%. Maximum molybdenum for other alloys is 0.5%. HH contains 0.2% N
(max). HP-50WZ also contains 4–6% W, 0.1–1.0% Zr, and 0.035% S (max) and P (max). Source: Ref 1

form in the solid state. These precipitates are minimize solidification hot cracking or to allow
carbides, oxides, and sulfides as well as inter- weld repair of cracks that do form. It has been
metallic phases richer in chromium, molyb- shown that the existence of ferrite can increase
denum, or nitrogen than the matrix. the resistance to stress corrosion cracking.
• General corrosion resistance follows the
above guidelines but is also helped by cop-
per and nickel, which do not assist in pitting
resistance. Metallurgy of “C” Alloys
• High-temperature oxidation resistance is en-
hanced by increasing chromium and silicon. The corrosion-resistant “C” series have
Wrought alloys employ aluminum and rare wrought counterparts from which they differ es-
earths to help oxidation resistance, but the sentially only in silicon content. This silicon has
difficulty of keeping these elements from no significant influence on corrosion resistance
being oxidized requires special techniques or mechanical or physical properties, so an un-
such as vacuum induction melting and inert derstanding of these alloys by approximating
refractories for molds. them to their wrought counterparts is justified.
• Iron-chromium (ferritic) alloys have better The main difference between the cast and
thermal fatigue resistance but poorer creep wrought product forms of these alloys is the
resistance than iron-chromium-nickel grain structure. In wrought grades, the grain
(austenitic) alloys. structure can be manipulated by deformation
and heat treatment. The use of deformation is
The alloy designation system largely ignores not an option in cast alloys; consequently, the
the wrought alloy distinctions by microstructure opportunities for grain refinement in cast alloys
(i.e., ferritic, austenitic, duplex, PH [precipita- are limited. To counteract the problems of lower
tion hardening], and martensitic). One reason is corrosion resistance of cast grades, a homoge-
that the most widely used wrought-stabilized nizing solution heat treatment is necessary to
ferritics (e.g., 409, 439) do not exist as common counteract the chromium depletion that occurs
casting alloys, and nominally austenitic alloys due to solidification segregation and precipitating
in the cast form contain enough ferrite to be sig- phases. Representative mechanical properties
nificantly magnetic. Thus, the distinctions based for “C” alloys are listed in Table 3 (Ref 1).
on phase are not as well defined for casting al- Martensitic. CA alloys are martensitic. The
loys. The high ferrite content in the nominally metallurgy is straightforward and equivalent to
austenitic casting alloys is to avoid or at least their wrought counterparts. The mechanical
Table 3 Room temperature mechanical properties of corrosion resisting cast stainless alloys

Tensile strength Yield strength, 0.2% offset Charpy toughness

Elongation % Reduction Hardness,
Alloy Heat treatment(a) MPa ksi MPa ksi in 50 mm, 2 in. in area, % HB J ft-lb Specimen
CA-6NM >955 oC (1750 oF), AC, T 827 120 689 100 24 60 269 94.9 70 V-notch
CA-15 980 oC (1800 oF), AC, T 793 115 689 100 22 55 225 27.1 20 Keyhole notch
CA-40 980 oC (1800 oF), AC, T 1034 150 862 125 10 30 310 2.7 2 Keyhole notch
CB-7Cu 1040 oC (1900 oF), OQ, A 1310 190 1172 170 14 54 400 33.9 25 V-notch
CB-30 790 oC (1450 oF), AC 655 95 414 60 15 ... 195 2.7 2 Keyhole notch
CC-50 1040 oC (1900 oF), AC 669 97 448 65 18 ... 210 ... ... ...
CD-4MCu 1120 oC (2050 oF), FC to 745 108 558 81 25 ... 253 74.6 55 V-notch
1040 oC (1900 oF), WQ
150 / Stainless Steels for Design Engineers

1120 oC (2050 oF), FC to 896 130 634 92 20 ... 305 35.3 26 V-notch
1040 oC (1900 oF), A
CE-30 1095 oC (2000 oF), WQ 669 97 434 63 18 ... 190 9.5 7 Keyhole notch
CF-3 >1040 oC (1900 oF), WQ 531 77 248 36 60 ... 140 149.2 110 V-notch
CF-3A >1040 oC (1900 oF), WQ 600 87 290 42 50 ... 160 135.6 100 V-notch
CF-8 >1040 oC (1900 oF), WQ 531 77 255 37 55 ... 140 100.3 74 Keyhole notch
CF-8A >1040 oC (1900 oF), WQ 586 85 310 45 50 ... 156 94.9 70 Keyhole notch
CF-20 >1095 oC (2000 oF), WQ 531 77 248 36 50 ... 163 81.4 60 Keyhole notch
CF-3M >1040 oC (1900 oF), WQ 552 80 262 38 55 ... 150 162.7 120 V-notch
CF-3MA >1040 oC (1900 oF), WQ 621 90 310 45 45 ... 170 135.6 100 V-notch
CF-8M >1065 oC (1950 oF), WQ 552 80 290 42 50 ... 170 94.9 70 Keyhole notch
CF-8C >1065 oC (1950 oF), WQ 531 77 262 38 39 ... 149 40.7 30 Keyhole notch
CF-16F >1095 oC (2000 oF), WQ 531 77 276 40 52 ... 150 101.7 75 Keyhole notch
CG-8M >1040 oC (1900 oF), WQ 565 82 303 44 45 ... 176 108.5 80 V-notch
CH-20 >1095 oC (2000 oF), WQ 607 88 345 50 38 ... 190 40.7 30 Keyhole notch
CK-20 1150 oC (2100 oF), WQ 524 76 262 38 37 ... 144 67.8 50 Izod V-notch
CN-3MN 1150 oC (2100 oF), WQ 770 112 365 53 50 185 190 140 V-notch
CN-7M 1120 oC (2050 oF), WQ 476 69 214 31 48 ... 130 94.9 70 Keyhole notch
(a) AC, air cool; FC, furnace cool; OQ, oil quench; WQ, water quench; T, temper; A, age. Source: Ref 1
Chapter 11: Casting Alloys / 151

properties are governed by the thermal pro- high level of nickel negates much of the potential
cessing, and strength, hardness, and toughness cost savings duplex alloys offer. All castings are
can be varied over a wide range. The CB 30 solution annealed and quenched to eliminate em-
and CC 50 alloys are ferritic and, as such, have brittling intermetallic phases.
negligible toughness but effectively deliver The cast duplex alloys may offer a better engi-
corrosion resistance. The toughness of CB 30 neering approach than the equivalent austenitic
can be improved by balancing the chromium cast alloys because they have greater strength
and silicon to a lower part of the range and the and lower alloy cost for the same level of corro-
carbon and nickel to the higher end to render sion resistance. They do not have the same prob-
the microstructure partly martensitic. lems of hot cracking that make casting austenitic
Precipitation Hardening. The cast PH alloys steels difficult. The poor hot workability of
include CB-7Cu-1, which behaves in a similar duplex steel is not an issue for castings. It is im-
way to 17-4PH, which has an overlapping com- portant for designers to understand that cast du-
position range. Note that most other major plex steels are totally compatible galvanically
wrought PH grades rely on titanium and alu- with wrought or cast austenitic alloys of the
minum to form coherent strengthening precipi- same corrosion resistance. Mixing components
tates and so do not have cast counterparts. Cop- with different microstructures does not create a
per, which can harden ferrite but not austenite, galvanic differential when corrosion resistance
is thus the only strengthener available. There is levels are similar. Reluctance to mix alloys for
one cast PH alloy that has no wrought counter- galvanic reasons can be an expensive error when
part. It is CD-4MCu; however, it is rarely used their similar corrosion resistances makes them
in the precipitation-hardened condition and is compatible, even if they are quite different
most commonly classified as a duplex stainless microstructurally.
steel in which the nitrogen level is closely con- Austenitic-Ferritic. The typical CF alloys,
trolled. This is a highly alloyed duplex grade which make up about two-thirds of U.S. stain-
that contains copper to precipitation harden the less steel castings, are nominally austenitic but
ferrite phase. Oil field CO2 corrosion is resisted always contain ferrite. This is not detrimental
by alloys that resemble the martensitic PH and improves resistance to stress corrosion
grades. These alloys are discussed in Chapter cracking and sensitization. Homogenization an-
22, “Petroleum Industry Applications” and can nealing can reduce the amount of ferrite and re-
be considered castable alloys. sult in lower yield and tensile strength and
Duplex. The cast equivalents of alloys 2205 higher elongation and toughness. The composi-
and 2507, J 92205, and J 93380 have similar tion balance is the main determinant of ferrite
properties and corrosion resistance. Modern level. Increasing the nickel, nitrogen, manganese,
wrought duplex alloys rely on nitrogen to parti- or carbon content decreases ferrite. Increasing
tion the alloy with uniform corrosion resistance chromium, silicon, or molybdenum content in-
in each phase and to suppress intermetallic phase creases ferrite. Increasing the solidification rate
formation. Cast alloys are effectively limited to will increase the ratio of austenite to ferrite in
0.25% nitrogen before gas porosity becomes ex- duplex or austenitic-ferritic alloys. The predom-
cessive. Porosity can be reduced by replacing inantly austenitic matrix has a very high tough-
some nickel with manganese, which increases ni- ness even at cryogenic temperatures. Ferrite, if
trogen solubility. Doing so would expand the continuous, decreases toughness. Fortunately, it
most promising area of stainless steel develop- is seldom present as a continuous phase. The
ment, lean duplex alloys such as 2101 and 2003, loss of toughness associated with high ferrite
to the cast grades. Alloy 2101 with 4 to 6% man- content can be aggravated by heating the ferrite
ganese provides the corrosion resistance of CF- above 475 oC (885 oF) for a sufficient time for
8M or 316L with total nickel plus molybdenum the ferrite to decompose to the brittle α and α'.
of only 2% versus the 12% required for the At higher temperatures, development of the σ
austenitic alloy. The duplex alloys also have phase would have a similar embrittling effect.
greater strength and are nearly immune to stress These phases thus formed are quickly redis-
corrosion cracking. These alloys represent signif- solved and removed by annealing. Note that
icant cost-savings potential for the foundry and sometimes copper is added to austenitic alloys
for its customers. CE-30 is duplex steel, which is to improve corrosion resistance in sulfuric acid
fairly simple metallurgically and uses only environments. It has no precipitation hardening
chromium for corrosion resistance. However, its effect in austenite, as it does in ferrite. When
152 / Stainless Steels for Design Engineers

used as a precipitating hardening agent, copper has higher thermal expansion and lower thermal
does not increase corrosion resistance. conductivity than ferrite, which aggravates ther-
Virtually any non-titanium-bearing, corro- mal fatigue and oxide spalling. Nevertheless,
sion-resistant, austenitic, wrought alloy can the better high-temperature strength of austenite
have a cast counterpart. Curiously, the 2xx low- generally is the predominant consideration, and
nickel alloys are not found in most cast alloys most “H” alloys are austenitic. Tables 4 and 5
lists. If a specific wrought alloy cannot be found list properties of “H” alloys (Ref 1 to 3).
to have a published cast counterpart, the de- Ferritic HA, HC, HD. Of the ferritic alloys
signer should not avoid requesting a producer HA, HC, and HD, HA with less than 10%
to supply a version that the foundry is confident chromium is not quite stainless but is useful to
of making. The designer must thoroughly un- 650 oC (1200 oF) for petroleum refinery applica-
derstand the design of the alloy desired so that tions. HC and HD are very high chromium
any alterations to its composition necessary to ferritics that have very low toughness and creep
allow castability will not compromise expected resistance but are quite oxidation and sulfida-
performance. tion resistant. They can be cost-effective materi-
als when high-temperature strength is not an
overriding concern.
Metallurgy of “H” Alloys Austenitic HE-HP. The predominant high-
temperature grades are the austenitic HE
The heat-resisting “H” alloys are principally through HP, after which come the nickel alloys,
austenitic. Alloying elements and impurities which are not generally classified as stainless
diffuse more slowly through the face-centered steels because they contain less than 50% iron.
cubic (fcc) austenitic structure than the bcc The high material cost of the nickel base alloys
ferrite structure, making the austenite more re- restricts their use to those specific environments
sistant to diffusion-controlled creep. Austenite where maximum carburization or nitriding

Table 4 Mechanical properties of heat-resistant cast stainless alloys at room temperature

Tensile strength Yield strength

Alloy Condition MPa ksi MPa ksi Elongation, % Hardness, HB

Standard grades
HA N + T(a) 738 107 558 81 21 220
HC As-cast 760 110 515 75 19 223
Aged(b) 790 115 550 80 18 ...
HD As-cast 585 85 330 48 16 90
HE As-cast 655 95 310 45 20 200
Aged(b) 620 90 380 55 10 270
HF As-cast 635 92 310 45 38 165
Aged(b) 690 100 345 50 25 190
HH, type 1 As-cast 585 85 345 50 25 185
Aged(b) 595 86 380 55 11 200
HH, type 2 As-cast 550 80 275 40 15 180
Aged(b) 635 92 310 45 8 200
HI As-cast 550 80 310 45 12 180
Aged(b) 620 90 450 65 6 200
HK As-cast 515 75 345 50 17 170
Aged(c) 585 85 345 50 10 190
HL As-cast 565 82 360 52 19 192
HN As-cast 470 68 260 38 13 160
HP As-cast 490 71 275 40 11 170
HPNb(d) As-cast 450 220 8 ...
HPNbTi(e) As-cast 450 220 8 ...
HT As-cast 485 70 275 40 10 180
Aged(c) 515 75 310 45 5 200
HU As-cast 485 70 275 40 9 170
Aged(f) 505 73 295 43 5 190
HW As-cast 470 68 250 36 4 185
Aged(g) 580 84 360 52 4 205
HX As-cast 450 65 250 36 9 176
Aged(f) 505 73 305 44 9 185
(a) Normalized and tempered at 675 °C (1250 °F). (b) Aging treatment: 24 h at 760 °C (1400 °F), furnace cool. (c) Aging treatment: 24 h at 760 °C (1400 °F), air cool.
(d) ISO 13583-2 specification minima. (e) ISO 13583-2 specification limits for microalloyed grade. (f) Aging treatment: 48 h at 980 °C (1800 °F), air cool. (g) Aging
treatment: 48 h at 980 °C (1800 °F), furnace cool. Source: Ref 1
Chapter 11: Casting Alloys / 153

Table 5 High-temperature mechanical properties of “H” alloys

Tensile Creep rate Stress to Stress to
Yield strength strength 0.0001%/h 1% in 100,000 rupture rupture
Alloy Temp ksi, MPa ksi, MPa psi, MPa h psi, MPa in 1000 h in 10,000 h
HA 1400 oF 16 27
1800 oF
HC 1400 oF 8.7 10.5 1.3 1.3
1800 oF 2.1 2.5 3.6 0.6
HD 1400 oF 36 3.5 7.0
1800 oF 15 1.0 2.5
HE 1400 oF 3.5 11.0
1800 oF 1.0 2.5
HF 1400 oF 35 6.0 4.4 9.1
1800 oF 1.6 (est)
HH 1400 oF 17 33 3.0
TYPE 1 1800 oF 6.3 9 1.1
HH 1400 oF 18 35 7.0 2.0 8.0
TYPE 2 1800 oF 7 11 2.1 1.6 0.9
HI 1400 oF 6.6 8.5
1800 oF 1.9 2.6
HK 1400 oF 6.8 6.3 12.0
1800 oF 2.7 0.9 2.8 1.7
HL 1400 oF 50 7.0 15
1800 oF 18.7 2.8 (est) 5.2
HN 1400 oF
1800 oF 2.4 2.1
HP 1400 oF
1800 oF 2.1
HPNb(a) 800 oC ... ... 51 MPa ... 55 MPa
HPNbTi(a) 800 oC ... ... 54 MPa ... 64 MPa
HT 1400 oF 26 35 8.0 12
1800 oF 8 11 2.0 2.7 1.7
HU 1400 oF 40 8.5
1800 oF 6.2 10 2.2 2.9 1.8
HW 1400 oF 23 32 6.0 7.8
1800 oF 8 10 1.4 2.6
HX 1400 oF 19.5 42 6.4
1800 oF 6.9 10.7 1.6 2.2

(a) Data from ISO 13583-2. Source: Ref 2, 3

resistance is mandatory. For oxidation and sulfi- alloying also are discussed in the oxidation sec-
dation resistance, the iron base alloys are pre- tion of Chapter 6.
ferred. These cast stainless steels derive their The major problem that all producers of stain-
oxidation resistance from their chromium level. less steels face is that of transferring molten
The chromium near the surface acts as a reser- metal from the furnace to the mold cavity. This
voir to replenish the protective iron/chromium problem is heightened when the foundry makes
oxide scale as explained in Chapter 6 in the sec- complex shapes. Methods developed to protect
tion on oxidation. Silicon, another stable oxide the molten stream from exposure to air to prevent
former, assists in forming this protective scale reoxidation have shown great promise and have
and resistance to carburization. Other typical been demonstrated by the wrought alloy produc-
alloying elements do not aid in oxidation resist- ers who tend to produce much simpler shapes
ance. If it were possible to cast these alloys with than the foundry. Protection of the molten stream
aluminum or rare earth additions without them could result in castings with much better high-
being lost to oxidation before solidification, temperature performance that could be used in-
there could be some impressive benefits. Such stead of some use of higher nickel alloys.
alloys exist in wrought form, for example, High-temperature strength is modestly im-
153MA and 253MA. The metallurgical basis of proved by higher levels of chromium and
the benefits from aluminum and rare earth nickel. Molybdenum improves high-temperature
154 / Stainless Steels for Design Engineers

strength, but its detrimental effect on oxidation higher-performance grades. It is also possible to
resistance and its promotion of intermetallic use AOD-refined master melt stock to achieve
precipitation limits its use. Carbon is very effec- the same benefits as AOD refining while using
tive for promoting high-temperature strength induction melting.
and suppression of intermetallic phase forma- Welding of cast stainless alloys is a common
tion. All “H” alloys, therefore, employ much practice and does not present problems when
higher carbon levels than the “C” alloys. This using approved weld procedures and qualified
does, however, directly imply that the corrosion welders. Chapter 17 describes joining methods
resistance of “H” alloys, should it be an issue, is in detail. The same precautions about sensitiza-
significantly degraded over otherwise similar tion apply to castings. Welding of non-niobium-
“C” alloys. stabilized “C” alloys with carbon levels above
The HP grades have undergone significant 0.03% will require postweld annealing to redis-
development over the last 30 years. This devel- solve chromium carbides, which will otherwise
opment has come about through the addition of make the alloy susceptible to corrosive attack in
niobium to increase creep and rupture proper- the chromium-depleted regions of the heat-
ties. The use of microalloying additions has de- affected zone.
livered creep and rupture properties some 30% Iron and nickel base “H” alloys that are fully
higher than the HP grade without niobium mi- austenitic can suffer from hot shortness due to
croalloying. It is unfortunate that these HP sulfide films that precipitate along grain bound-
grades have not at this time found their way into aries even at low bulk sulfur levels. This makes
ASTM standards; however, work is under way them susceptible to hot cracking of welds. Al-
to remedy this omission. Currently, the most up- loys with some ferrite are less susceptible to hot
to-date collection of these grades can be found cracking, so most “C” alloys are highly resistant
in ISO 13583-2. to this problem.

Foundry Practice REFERENCES

1. Cast Stainless Steels, Metals Handbook,

While the scope of this book does not extend
desk ed., J.R. Davis, Ed., ASM Interna-
to the production of castings, certain aspects are
tional, 1988, p 386–390
important to the user of castings. For the last 50
2. International Organization for Standardiza-
years, virtually all stainless has been refined in
tion,, ISO 13583-2
argon oxidation decarburization (AOD) vessels
3. Steel Founders Society of America, online
or versions thereof. For “C” alloys, this refining
method should be considered a basic require-
ment for good quality where the carbon levels
are restricted to low levels (e.g., CF3M). “H”
alloys are less refined inherently and can be SELECTED REFERENCE
simply arc or induction melted; however it may
be necessary to use refining techniques for the •
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 155-160 All rights reserved.
DOI: 10.1361/ssde2008p155


Melting, Casting, and Hot Processing

Summary austenitic alloys, impossible to produce. The

advent of AOD, continuous casting, ladle metal-
THE PRIMARY PRODUCTION PROC- lurgy, and powerful hot rolling mills has led to
ESSES of melting, casting, and hot processing stainless steels of much higher quality produced
are invisible to the end user. The vast majority of at lower cost. Ironically, the low processing cost
stainless steel is made by arc furnace melting fol- of stainless steel has spurred demand and made
lowed by argon oxygen decarburization (AOD) some of its ingredients, such as molybdenum
refining and continuous casting. It is not normal, and nickel, which are relatively scarce and ex-
and it is seldom beneficial for the end user to pensive commodities, even more costly, forcing
specify processing paths. The end user should, the cost of many alloys to spike even higher
however, be knowledgeable and require the pro- than in earlier years.
ducer to document the process and the producer’s
control of it.
Melting and Refining

Introduction The arc furnace is nearly universally used for

the first step in the production of stainless steel.
The manner in which stainless steel is made The arc furnace is quite flexible in the types of
at the producing mill can have a great impact on charge materials it can accept. Since the charge
its final properties. These production methods materials for stainless steel are typically carbon
have undergone a major evolution over the last steel and stainless steel scrap, this flexibility al-
50 years and are mainly responsible for stain- lows scrap of all types to be used. The necessary
less steels becoming the practical, widespread chromium is added as ferrochromium, whose
engineering materials they are today. Traditional cost is inversely related to its carbon content.
carbon and alloy steel-making methods are not The carbon content of the heat of steel is roughly
suitable for stainless steels. The fundamental 1.5 to 2.5% when it is melted and ready to
difference is that the basic decarburization step, charge into the separate refining vessel.
which is common to all steel making, is thermo- It is this carbon whose removal is the primary
dynamically very difficult in stainless steel be- focus of refining. In the 1960s, Union Carbide
cause the essential element, chromium, reacts engineers perfected a method, the previously
more strongly with the purifying agent, oxygen, mentioned AOD process, of removing nearly all
than does carbon. Thus, early stainless steel mak- the carbon from molten stainless steel without
ing, done in an arc furnace, was a lengthy process significant loss of chromium. This process is
that necessarily involved high chromium losses based on the following chemical reaction:
to the slag as carbon was removed. This process
Cr3O4 (Solid) + yC = yCO (gas) + Cr (Eq 1)
was not only very expensive, the carbon levels
that could be achieved were not much below The equilibrium for this reaction is:
0.10%, making most of today’s stainless steels,
whose carbon levels range from 0.010% in sta- −ΔG
Ln (K ) = (Eq 2)
bilized ferritic alloys to about 0.07% in normal 4575T
156 / Stainless Steels for Design Engineers

where K is the equilibrium constant, and G is steel combine to reduce the oxygen content of
the Gibbs free energy. the steel to around 100 ppm. This could be fur-
Working through the thermodynamics yields ther reduced by aluminum, but aluminum-based
the relationship that summarizes the important inclusions are generally undesirable. The ther-
relationship among carbon, chromium, and CO modynamic activity of aluminum is consider-
(Ref 1): ably reduced in iron as chromium levels in-
crease, so its role as a deoxidizer is less valuable
in stainless steels. Titanium, on the other hand, is
%Cr 13,800
Log = =− + 8.76 − 0.925 log p CO enhanced as a deoxidizer in chromium-iron al-
%C T loys, and consequently small amounts of it are
(Eq 3) sometimes used as a supplementary deoxidant in
alloys even though an alloy specification may
Thus, increasing the temperature works to in- not call for any. Titanium is believed to reduce
crease the elimination of carbon as CO, which hot working defects. More active deoxidants,
evolves from the melt. This is similar in principle such as calcium and magnesium, can be used
to the basic oxygen furnace (BOF) process for when required. Also note that even if no inten-
carbon steel in which oxygen is injected into tional addition of metallic calcium is made,
molten steel to remove carbon by oxidizing it. strong deoxidation with aluminum or titanium
The key to the AOD process, though, is the in- can reduce small amounts of calcium from the
jection of oxygen and argon into the bath to keep CaO in the slag, producing measurable calcium
the partial pressure of CO (pCO) very low. This is content in the metal.
done at a temperature consistent with economic Besides carbon and oxygen, other impurities
refractory life. The injection is done through can be removed from the molten stainless. Once
tubes called tuyeres in the bottom of the barrel- the steel has been deoxidized, sulfur can be
shaped vessel. The injection and the reaction readily removed by contact with a basic slag.
cause extremely thorough mixing, which would Sulfur can be reduced to less than 0.001% in
never happen in the flat, stagnant, arc furnace the AOD, and this excellent purity level is com-
bath. This mixing not only allows the CO-pro- mercially furnished without additional price
ducing reaction to reach equilibrium, but also premium. Sulfur, although a harmful impurity
the mixing of the slag and metal also permits from a corrosion standpoint, is often deliber-
desulfurization. By increasing the ratio of argon ately kept at moderate levels (0.008 to 0.015%)
to oxygen in the injected gas as the refining pro- for tungsten inert gas (TIG) welding penetration
ceeds, the carbon is selectively oxidized with- (see Chapter 17) and at high levels (0.15%+) for
out concurrent chromium oxidation. A typical machinability (see Chapter 15). These trade-offs,
starting ratio is 3 to 1 oxygen to argon/nitrogen which are beneficial to processors, should be
by volume. The ending ratio can be as low as 1 viewed with skepticism by end users, whose
to 9, oxygen to argon/nitrogen. The choice of product integrity is compromised. There are
which inert gas to use, argon or nitrogen, is based processing methods for which higher levels of
on cost and final nitrogen content desired. Stabi- sulfur are not necessary that are preferable to
lized stainless steels require low carbon and ni- the end user while not compromising welding
trogen levels, for instance, so the more expensive or machining costs. For example, machinability
argon must be used. can be improved by calcium additions that pro-
It is possible to use a vacuum system to keep duce malleable oxides to replace the deleterious
the partial pressure of CO low when refining sulfides (see Chapter 15), and welding methods,
with injected oxygen. This is the vacuum oxy- such as laser welding, can be used in many
gen decarburization (VOD) process. The VOD cases to eliminate the need for the weld penetra-
process can achieve slightly lower carbon levels tion enhancement of sulfur while increasing
but does not achieve cleaner steel as some welding speeds.
believe. Phosphorus is an impurity for which no prac-
In both processes, after final carbon content tical removal technology exists in stainless
has been achieved ferrosilicon is added to reduce steel. Any known process to remove it first re-
the chromium in the slag and have it return to the moves chromium. Thus, it exists in almost all
molten steel. The excellent mixing of the slag stainless steel at levels close to its normal speci-
and metal in the AOD permits this to be done fication limit, about 0.030% in austenitic alloys
efficiently. The silicon plus the manganese in the and 0.020% or less in ferritic alloys, which are
Chapter 12: Melting, Casting, and Hot Processing / 157

made from a higher percentage of low-phospho- on interstitial solubility. The higher solubility
rus carbon steel scrap. The deleterious effects of of carbon, nitrogen, and oxygen in stainless
phosphorus on corrosion are not avoided unless steels is significant. A manganese/silicon deox-
much lower levels are achieved. Consequently, idized stainless steel will still have about 100
its presence is tolerated since it has no differen- ppm of dissolved oxygen at the freezing tem-
tial effect over the range in which it is found. perature as opposed to the less than 10 ppm of
Heavy metals are eliminated by high-temper- oxygen found in aluminum-killed carbon steel.
ature AOD blowing, as is hydrogen. Care must This oxygen precipitates as oxides in the solid
be taken not to reintroduce such impurities after state.
refining, which is a risk when using damp or Vacuum induction melting (VIM) is another
contaminated scrap for coolant. method of melting stainless steels. This is a
Alloy adjustment can be done in the AOD or nearly slag-free process, and little refining is
preferably in a treatment-and-transfer ladle. The possible. Melt purity is largely controlled by the
tapped molten steel generally has excess heat purity of the starting material, and use of AOD
from the highly exothermic refining process. master melt stock for VIM remelting is com-
This allows the composition to be measured and mon. Limited decarburization is possible via in-
adjusted before it must be cast. This can be done jection of oxides such as Fe3O4 or SiO2 to create
very precisely by wire feeding of alloying ele- CO evolution inside the vessel. Using this tech-
ments through the slag into the heat, which can nique, very low carbon levels (less than 50
be stirred by argon bubbling via porous plugs. ppm) are achievable commercially. Use of VIM
This technique is very effective for the fine- is generally limited to high-value, high-purity,
tuning of reactive elements such as titanium. or low-tonnage melts.
The refining treatments used for carbon steel
and stainless steel are very similar, but there are
subtle differences because of the difference in Remelting
the thermodynamics of dilute solutions like car-
bon steel and highly alloyed, nondilute solutions Some stainless steels and related alloys are
like stainless steel. Table 1 shows the factors by remelted to refine composition or ingot struc-
which additions of various elements to stainless ture. There are two principal remelt processes:
steel (j) alter the thermodynamic activity of vacuum arc remelting (VAR) and electroslag
other alloying elements (i). remelting (ESR).
Equation 4 is used to calculate the activity of In VAR, the material to be remelted is cast
elements in steel. The activity coefficient γ into a cylindrical electrode and placed inside a
varies with the concentration of alloying ele- cylindrical water-cooled vacuum chamber. A
ment x by: high-current direct current (dc) arc is estab-
lished between the electrode and a starter plate
δ ln γ i at the bottom of the chamber. The end of the
RT lnγ i = RT lnγ i0 + ∑
j =1... n
δxj (Eq 4) electrode is melted, and the molten drops fall
through the intervening vacuum. Volatile con-
stituents escape from the molten drops, and the
This calculation is best left to computer pro- purified drops collect to form a molten pool on
grams such as Thermo-Calc that have been per- top of the starter plate. VAR parameters are ad-
fected for these lengthy procedures. It should justed to maintain a shallow pool, which solidi-
be noted that chromium, which is always pres- fies in a bottom-up fashion. The shallowness of
ent in nondilute quantities, has a powerful effect the molten pool produces a refined grain

Table 1 Influence of alloying elements on the thermodynamic activity of carbon, nitrogen,

sulfur, and oxygen
i Al C Cr Mn Mo N Ni O S Si Ti W
O .04 .14 –.02 –.01 –.01 .11 .01 –.34 .05 .08 ... –.005
N –.03 .13 –.05 –.02 –.01 0.0 .01 .05 .01 .05 –.53 –.001
S ... .11 –.01 –.03 .003 .01 0.0 –.27 –.03 .06 –.07 .01
0 –.39 –.45 –.04 –.02 .003 .06 .006 –.20 –.13 –.13 –.6 –.01
158 / Stainless Steels for Design Engineers

structure with less solidification segregation ceramic tube, called the submerged entry
than found in typical cast product. nozzle, into the mold, which is covered with a
In ESR, the material to be remelted is cast consumable protective and lubricating slag
into an electrode of similar shape, but slightly cover, called a mold powder. The mold powder,
smaller than the water-cooled mold. A gap be- which melts in the mold as it is added, contains
tween the electrode and a starter plate at the bot- ceramics, fluxes, and carbon. The level of the
tom of the mold is filled with a prepared slag. molten metal should be carefully controlled by
Typically, this slag is calcium fluoride-based ultrasonic measurement, or other methods, to
with high lime (CaO) content. Additional ingre- prevent fluctuations in level that may entrap
dients control the basicity, fluidity, oxidizing slag in the slab surface. The entire water-cooled,
potential, and other properties of the slag. A copper alloy mold oscillates in a precise pattern
high current is used to melt the slag, which in as the solidifying strand of steel is withdrawn
turn melts the end of the electrode, and the from the mold bottom by pinch rolls and
molten drops fall through the slag. Reaction of sprayed with water to cool it. The pinch rolls
the molten drops with the slag removes sulfur apply enough pressure to slightly deform the
and some other impurities, and the purified slab. This deformation has a crucial, seldom-rec-
drops collect to form a molten pool on top of the ognized effect. It causes a beneficial recrystal-
starter plate. ESR melting typically is done at a lization that improves hot working characteris-
higher rate than VAR, and the molten pool is tics of austenitic and duplex alloys. In ferritic
deeper. This deeper pool produces a grain struc- alloys, it can cause excessive grain growth,
ture between that of VAR and typical cast prod- which detracts from hot workability. The initial
uct, with commensurate intermediate segrega- portion of slab cast in a sequence is seldom
tion patterns. of adequate quality to be used because of
exogenous inclusions, entrapped mold powder,
and non-steady-state solidification structure.
Casting The defective portion must be identified and
scrapped or diverted to low-quality requirement
Continuous slab, billet, and bloom casting end uses.
have become the standard methods of making The strand is bent from an initial slightly
stainless steel primary products, replacing the curved shape to flat and cut into slabs. More
obsolete ingot method. There are some alloys than one heat of steel may be cast sequentially
that cannot be continuously cast, but these repre- without restarting the process. This is ideal eco-
sent a miniscule percentage of stainless produc- nomically and for quality reasons since initial
tion. Continuous casting produces slabs directly, and final segments of a casting can contain
thus removing the costly soaking and slab- more inclusions and aberrant structure. Some
rolling processes. In a well-executed continuous end users stipulate that no first slabs be applied
casting operation, slabs are of sufficient quality to their orders. Producers generally apply first
that they require no surface conditioning before slabs to less-critical uses or discard suspect sec-
being hot rolled. Slabs range in thickness from tions of them. If casting conditions are not opti-
13 to 63 cm (5 to 15 in.). The segregation in con- mal, the result can be slabs with poor surface
tinuous casters is less than in ingots because of quality that must be surface ground.
the smaller section size. It is not eliminated, Slabs are sometimes quenched to avoid pre-
however, and certain alloying elements concen- cipitation of phases; however, they may be held
trate at the centerline, where they defy homoge- at high enough temperatures prior to hot rolling
nization. Carbon and molybdenum are examples to stay above the temperature range in which
of alloying elements with this tendency. embrittlement can occur or to stay above the
In properly executed continuous casting, the temperature at which an embrittled slab can
ladle feeds by a slide gate, or preferably a stop- fracture. Ferritic and martensitic alloys are es-
per rod gate, into a ceramic tube into the large pecially prone to these problems.
tundish situated over the caster mold. The metal There has been great interest for decades in
in the tundish is covered with a protective slag producing stainless steel coils directly from
cover, and flow patterns within the tundish are the melt in so-called strip casters. Elimination
designed to minimize dead spots and encourage of hot rolling could be quite valuable in stain-
removal of inclusions by impingement with the less steel, whose hot rolling from slab can be
slag cover. The metal feeds through another both expensive and problematic. There are a
Chapter 12: Melting, Casting, and Hot Processing / 159

number of such machines in pilot or limited mills that routinely roll carbon steel to 1.5 mm
production. They have not had sufficient com- (0.06 in.) can struggle to attain 4.5-mm (0.18-in.)
mercial or technical success to have become a thickness for 316 stainless.
factor in the industry. Since their development The high separating forces on the hot rolling
is only being undertaken by those large stain- mill stands also cause greater roll deflection and
less steel producers who already have the hot compression, which if not countered by roll
rolling assets that strip casting would replace, bending or roll shifting schemes can lead to sig-
it seems unlikely that strip casting will soon nificant variation in thickness across the sheet,
become a major factor even if it is perfected as much as 0.25 mm (0.01 in.). This variation as
technically. a percentage of thickness is not reduced by cold
Another method of shortcutting the casting/ rolling and is a major cause of tolerance loss in
ingot step has been perfected: the powder met- sheet and strip. Hot-rolled bands vary in thick-
allurgy approach. In powder metallurgy, the re- ness along the length of the coil because the tail
fined molten metal is atomized by gas or liquid end of the slab is colder and harder to roll. Coil
and made to freeze into small particles. These boxes (on reversing mills) address this problem
particles, having been quenched extremely to a degree by permitting the semirolled coil to
rapidly, are quite homogeneous. Powder tech- equalize in temperature.
nology methods allow for the design of alloys Hot strip tandem mills powerful enough to
that would otherwise freeze with too much successfully roll high-quality stainless steel hot-
segregation and too coarse a structure with rolled bands are massively expensive and are
conventional production methods. Traditional seldom justified for the tonnage of stainless steel
powder metallurgy production methods are rolling a given melt shop produces, although
used to make small near-net shape compo- rolling stainless on hot tandem mills used pri-
nents, avoiding most of the costly machining marily for carbon steel can be an excellent pro-
steps. More impressively, powder technology duction method.
is also used to produce massive components. Hot Steckel mills have become the favored
For example, very high carbon/vanadium method of hot rolling stainless steel because
stainless tool steel components can be made by their throughput better matches stainless steel
encapsulating powder in an evacuated canister melt shop production outputs. This permits the
in which it can be sintered and hot worked to melt shop and caster to be adjacent to the hot
100% density and virtually complete homo- mill, which permits energy-saving hot charging
geneity. Chapter 9 on martensitic alloys dis- of slabs. In hot Steckel mills, typically a four-
cusses these materials. high reversing rougher rolls slabs to about 3-cm
(1.2-in.) thick. Then, the transfer band is rolled
to final gauge on a separate reversing four-high
Hot Rolling finishing mill with coil boxes to preserve tem-
perature. The economy of having only two mill
Hot rolling remains an essential process for stands makes these mills ideal for typical stain-
the vast majority of stainless steel used. Hot less production quantities and permits the cost
rolling characteristics of stainless steels vary of sophisticated mill capabilities, such as roll
greatly. Ferritic stainless steels are extremely shifting, roll crossing, or roll bending, not to
easy to hot roll since they have a soft, single- have to be duplicated among many stands. This
phase structure at hot rolling temperatures. is the same justification for using Sendzimir
Martensitic stainless steels roll like their carbon mills to cold roll stainless. In both cases, the
and alloy steel counterparts since their mi- logic applies more to austenitic alloys than to
crostructure during hot rolling is a moderately the easily rolled ferritic stainless alloys.
alloyed austenite similar to alloy steels. The mi- In either case, the hot-rolled band carries a
crostructure during hot rolling is the crucial fac- heavy, embedded scale that must be removed
tor. Austenitic stainless steels have high strength from the surface before further processing in
at hot rolling temperatures. Furthermore, the most cases. Some alloys can be cold rolled in the
low diffusion rates in austenite slow recrystal- “black band” state at a cost of coarser surface
lization so that the steel does not always soften finish and greater rolling mill roll wear. If nor-
between stands in tandem mills. This increases mal cold rolling or use as hot-rolled coil is fore-
mill loads, and lower reductions must be taken seen, the hot-rolled band must then be annealed
than for alloy steels. Powerful hot strip tandem and pickled since the as-rolled hot-rolled band
160 / Stainless Steels for Design Engineers

has poor corrosion resistance, poor mechanical rily as edge cracks and slivers. Edge cracks are
properties, residual cold work and hardness vari- simply a lack of ductility at the colder strip edge.
ations, as well as a heavy oxide layer. Stainless hot ductility often has a narrow tem-
perature window, and many factors can affect
the size of that window depending on alloy type.
Defects The most inherently challenging alloys for hot
working are the duplex alloys and the alloys that
Stainless steel hot-rolled bands can contain solidify in the fully austenitic state. The former
many types of defects. These are seldom seen has a mixed-phase structure, and the phases can
by the end user because they are removed exhibit mechanical incompatibility at certain
when they are not prevented. They do have temperatures. The latter alloys reject sulfur and
repercussions on delivery. The major cate- oxygen during solidification and slab reheating
gories are: to the grain boundaries, where they form very
• Hot mill defects weak films. But, even alloys such as 304 and
• Inclusion-related defects 316 can have very poor hot ductility if they con-
• Hot ductility-related defects tain much sulfur and oxygen or if they are re-
heated for long times or at temperatures above
Stainless steel is less forgiving of hot mill 1250 °C (2280 °F), which facilitates diffusion of
faults because its surface is not removed by oxi- sulfur and oxygen to the grain boundaries and
dation to the degree carbon steel’s surface is. also encourages very large grains. This poor hot
Thus, a skid mark from a slab-heating furnace ductility manifests itself as “slivers,” which can
will remain through the hot rolling, annealing, require grinding of the entire hot band surface.
and cold rolling processes. This is true of all hot These tendencies are fought by low oxygen and
mill scratches, gouges, digs, etc. Rolling stain- sulfur levels and minimal slab-reheating temper-
less requires a different mindset than rolling atures and times, as well as slab surface working
carbon steel, which argues against the benefits in the caster pinch rolls. Sometimes, very poor
of rolling stainless on a mill built and primarily hot working alloys are given a single hot reduc-
used for carbon steel. tion pass on a hot mill to produce a full recrystal-
Inclusion-related defects are all essentially lization that disperses grain boundary-weakening
avoidable by using state-of-the-art technology. elements on subsequent reheat.
Protecting metal from reoxidation and keeping
precise mold-level control in the continuous
caster prevents all inclusions of a size that can REFERENCE
produce a defect.
Hot ductility defects are more subtle. They 1. D. Peckner and I.M. Bernstein, Handbook of
arise from many causes and are manifest prima- Stainless Steels, McGraw-Hill, 1977, p 3–13
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 161-171 All rights reserved.
DOI: 10.1361/ssde2008p161


Thermal Processing

Summary desirable, and some are potentially very detri-

mental. Readers are encouraged to review the
THE THERMAL PROCESSING of stainless earlier chapters on phases in stainless steel
steel is a topic the end user should approach (Chap. 6–10) to familiarize themselves with
with great respect. It is not simple in concept or these phases.
in practice. Before attempting to carry out any Each of the alloy groups of stainless steels has
thermal processing on stainless steel, the practi- radically different thermal processing objectives
tioner must understand the alloy design, compo- and requirements; therefore, each is discussed
sition, and processing history of the material in separately.
question. The thermal processing then must be
designed and executed in a planned, controlled
manner. The consequences of failure in thermal Austenitic Stainless Steels
processing can become catastrophic to mechan-
ical properties and corrosion resistance. Thermal processes applied to austenitic stain-
less steels include:
• Soaking for homogenization and preparation
Introduction for hot working
• Annealing to remove the effects of cold
The thermal processing of stainless steels can work and to put alloying elements into solid
have many purposes. Normally, the objectives solution (solution annealing)
are simple: heating for hot working, annealing to • Stress relieving
soften after cold working, solution annealing
The temperatures at which these processes
to homogenize, heating to temper martensite, or
are carried out are shown in Table 1 for typical
to stress relieve. However, even if the objective
austenite compositions.
is simple, the processes that occur are anything
but simple. Variations in temperature, times at
temperature, heating and cooling rates, and at- Soaking
mosphere can have complex and easily unin- Because virtually all stainless steel is continu-
tended consequences. There is no substitute for ously cast, the older soaking process of holding
understanding the processes that are occurring ingots in soaking pits for many hours is rarely
when stainless steels are heated for successful used. The soaking had two functions. The obvi-
heat treating to be achieved. ous one was to equilibrate at the right tempera-
Stainless steels have many alloying elements ture for hot working. The second, less-obvious,
in large amounts. Many of these elements are one was to achieve greater chemical homogene-
highly reactive thermodynamically. The practi- ity. The lack of homogeneity comes from the
cal consequence of this is that many phases are solute segregation that occurs as solute ele-
thermodynamically possible at different temper- ments are rejected from the material that was
atures. Stainless also reacts with its environment first to freeze. Solute segregation was exagger-
at high temperatures, causing changes in surface ated by the slow solidification of ingots, and
alloy content. Some of the resulting phases are continuous casting helped make the stainless
162 / Stainless Steels for Design Engineers

Table 1 Recommended thermal processing temperatures for austenitic alloys

Annealing Annealing ASTM A480(a) Stress Stress
Alloy temperature, °C temperature, °F 2006, °F relieving, °F relieving, oC
Standard alloys ... ... ... 1500–1600 non-L, 815–870 non-L,
1000–1600 L grades 540–870 L grades
201, 202, 201LN 1010–1120 1850–2050 1900 min ... ...
301, 301LN, all versions 1010–1120 1850–2050 1900 min ... ...
304, 304L, 305, all versions 1010–1120 1850–2050 1900 min ... ...
316, 316L, 316N, 317, 317L 1040–1175 1900–2150 1900 min ... ...
308, 309, 309S, 310, 310S 1040–1175 1900–2150 1900 min ... ...
Stabilized alloys ... ... ... 1000–1600 540–870
321 955–1065 1750–1950 1900 min ... ...
347, 348 980–1025 1800–1950 1900 min ... ...
20Cb-3 925–955 1700–1750 ... 925–1010 ...
Moderately alloyed, Creq<26 1120–1175 2050–2150 Various 1500–1600 815–870
S31725, N08028, JS700
Highly alloyed, lower sigma 1120–1175 2050–2250 Various Not recommended ...
alloys, Creq>30, high N
AL6-XN, 4565, 654SMO,
Highly alloyed, sigma-prone 1205–1230 2200–2250 Various Not recommended ...
alloys, Creq>30, low N
(a) Standard specification for general requirements for flat-rolled stainless and heat-resisting steel plate, sheet, and strip.

steel more homogeneous. Nevertheless, cast rounding matrix by sufficient soaking. Welds that
slabs and blooms must be soaked to eliminate are unannealed have such precipitated inclusions
as-cast segregations. This process, to the extent in an unequilibrated state, and the result is dimin-
it is done, occurs as they are reheated to the ap- ished chromium concentration and poorer pitting
propriate temperature for hot working. Soaking resistance.
dissolves the few percent of residual delta fer-
rite that are present on slab solidification. It is
important to soak at the highest temperature at Annealing
which delta ferrite is not a stable phase so that it Annealing serves two main functions in
will dissolve, about 1250 °C (2280 °F) for most stainless steel: It removes the effect of cold
austenitic stainless steels. work by replacing strained microstructure with
Soaking at higher temperatures causes ferrite new strain-free grains, that is, recrystallization.
levels to increase, negating the homogenization New grains nucleate and grow. If stored strain
and causing very poor hot workability. Longer energy is insufficient, as happens often with
times at temperature than the minimum required ferritic stainless steels, true recrystallization is
for thermal uniformity also cause problems as difficult to achieve, and the annealing process
any sulfur and oxygen impurities are rejected may only produce recovery without recrystal-
from austenite and can diffuse to grain bound- lization, leaving the same grains relieved of
aries, where they form weak, plastic films that stress. This leaves the surviving grains with the
also degrade hot workability. Grain growth, by same crystallographic orientation that deforma-
reducing grain boundary area, exacerbates this tion produced and may or may not be the de-
effect. Thus, soaking times are best minimized sired outcome. Second, annealing returns into
and closely controlled. Alloys are therefore de- solution solute that has been precipitated as un-
signed to have only a slight amount of delta fer- wanted phases, principally carbides, but also
rite to be redissolved during soaking. Ferrite is intermetallic phases. Annealing also may help
useful because it has a high solubility for oxygen to reduce solute segregation remaining from
and sulfur. Having none would result in impurity the casting process, making the composition
rejection of these elements to grain boundaries more homogeneous. The homogenization
during solidification, which results in the worst- process is accelerated by the reduction in di-
possible hot working characteristics. The oxygen mensions from hot and cold working. A reduc-
and sulfur trapped in the ferrite during solidifica- tion in dimension by a factor of two reduces the
tion precipitates in the solid state as inclusions, time to achieve a given degree of homogeniza-
which also must be equilibrated with the sur- tion by a factor of four.
Chapter 13: Thermal Processing / 163

Annealing to recrystallize is fairly rapid. To a carbides can be redissolved and carbon dif-
first approximation, it is instantaneous, and the fused away from the carbide, but this does not
results are merely a function of the maximum mean that all composition gradients have been
temperature attained. This may not be the case reduced to zero. This means that precipitates
for continuous annealing lines, in which transit may re-form more rapidly in such a material
time can be short enough, less than a minute at than they would in a completely homogeneous
temperature, to limit the grain size attained. The alloy.
driving force for recrystallization is the strain The annealing temperature for a given alloy is
energy stored in the lattice from deformation. chosen based on the temperature required to put
The strain energy in a given material is propor- all alloying elements into solution. Higher car-
tional to the square of the material’s flow stress. bon levels, for instance, require higher tempera-
As the material is heated, recovery occurs first. tures to dissolve all the carbon. This is one of
This is the change in physical and mechanical the principle values of accurate phase diagrams.
properties associated with dislocation annihila- If it were simply a consideration of recrystal-
tion and polygonalization that occurs before the lization, all alloys could be annealed at similar
nucleation and growth of new grains. temperatures at the low end of the recom-
The nucleation of new grains occurs at high- mended range. Within the recommended range,
angle grain boundaries and proceeds by the the temperature selected should be determined
movement of roughly hemispherical growth by the desired grain size. End use determines
fronts into strained areas. The percentage re- whether a fine or coarse grain size is preferable.
crystallized, once a sufficient temperature is Table 1 lists recommended annealing tempera-
reached, grows sigmoidally. It is normal for tures for austenitic stainless steels.
the time to fully recrystallize to be rather less The overall interplay between prior cold work
than the time to attain that temperature. Even and annealing temperature on mechanical prop-
at the lower range of annealing temperatures, erties of annealed material can be summarized
times are generally less than 1 min. Recrystal- as (Ref 3):
lization will not occur if the lattice contains in-
sufficient strain energy. Thirty percent cold • Grain size of a given alloy is the most im-
work should be used as a rough threshold for portant parameter in characterizing mechan-
the required amount. Annealing after lower ical properties.
amounts of cold work is characterized by • Yield and tensile strength are essentially
scarce nucleation sites and can result in very constant for a given grain size regardless of
large and nonuniform grain size. Hot-worked the amount of prior cold work; however, the
material often has a composite structure that elongation depends on the prior reduction.
may have already had some recrystallization • Yield strength, tensile strength, and hardness
depending on the final reduction temperature. are essentially linear functions of grain size.
Annealing may not produce a clear recrystal- • Elongation decreases with finer grain size
lized structure in this case. and at an increasing rate as grain size be-
The relative rapidity of recrystallization an- comes finer as long as cross-section size of
nealing is due to the fact that it is rate con- the test specimen is not extremely small.
trolled by short-range diffusion. Solution an- This is not true for very coarse-grained ma-
nealing requires longer-range diffusion and terial.
thus can require much longer times. Some stud- • Maximizing elongation comes from maxi-
ies have shown that welds, for instance, do not mum prior cold work and medium annealing
recover completely from their loss of corrosion temperatures
properties that arise from local alloy depletion • Anisotropy coefficients, or plastic strain ra-
unless they have been annealed for times on the tios r are constant up to about 40% reduction
order of 1 h (Ref 1). Others have seen homoge- after, which r45 and rn increase sharply, while
nization in as little as 10 min (Ref 2). Wrought rt decreases. This leads to earing during
materials can require much shorter times be- drawing.
cause reductions during hot working have re- • The increase in properties for a one ASTM
duced diffusion distances. It should be noted grain size increment is:
that precipitates can be redissolved and not ap- a. 13 MPa (2 ksi) for tensile strength
parent in the annealed microstructure without b. 20 MPa (3 ksi) for yield strength
full homogeneity being achieved. For example, c. 2 HRB for hardness
164 / Stainless Steels for Design Engineers

Atmospheres for annealing are important. metallic phases. Use of high-chromium and-
Austenitic stainless steels heated in air, of molybdenum alloys without enhanced nitrogen
course, form oxide scales. Beneath this oxide, is no longer recommended, and the use of lower-
the metal matrix becomes significantly depleted nitrogen alloys should be reexamined and ques-
of chromium (Ref 4), often more than 5% lower tioned if specified.
in chromium and to a depth of as much as 10 µ Last, stainless surfaces should be scrupu-
(395 µin.). So, not only must any oxide be re- lously clean before annealing. Even hard water
moved, so must the chromium-depleted layer. deposits can cause differential oxide growth,
This requires aggressive pickling, which while which can cause etched spots on the surface,
done commonly, may not be practical for many where the postanneal pickling attacks the differ-
stainless users. The chromium-depleted zone, ent oxide more strongly. Carbonaceous materi-
however, does pickle rapidly precisely because als left on the surface are even more objection-
it does have less chromium. To avoid oxide able because they can cause carburization and
scale formation, vacuum, hydrogen, or inert gas subsequent loss of corrosion resistance.
atmospheres may be used. Stabilizing anneals are sometimes conducted
If vacuum is used, it should be less than on stabilized alloys such as 321 and 347. This is
2 × 10–3 torr (0.3 Pa). If an inert gas or hydrogen useful when carbon levels are sufficiently high
is used, the key consideration is moisture con- that significant dissociation of carbides occurs
tent. The dew point must be –40 °C (–40 °F) or at annealing temperatures. A second anneal at
lower. More stringent levels may be required if lower temperature, about 900 °C (1650 °F),
mirror finishes are desired. Cool down must be then is done to permit the carbon to combine
rapid as oxidation potential increases as tempera- with the stabilizing element rather than leaving
ture decreases. Vacuum or inert gas is preferable it available to form chromium carbides. Current
to hydrogen for alloys containing stable oxide preferred practice for these alloys is to maintain
formers such as aluminum or titanium or for al- carbon and nitrogen below 0.03% for corrosion-
loys containing boron. resistant service, which renders this stabilizing
Austenitic alloys that are subject to sensitiza- unnecessary. Alloys used for high-temperature
tion must be cooled rapidly enough from anneal- service benefit from the creep-resisting contri-
ing temperatures to avoid carbide precipitation butions of higher carbon levels.
during cooling. If forced air or water quenching
are impractical or if section size prohibits rapid
cooling, then using stabilized or low-carbon Stress Relieving
grades is indicated. Austenitic stainless steel weldments often con-
Superaustenitic stainless steels, and even al- tain residual stresses, which can cause distortion
loys like 317, present a special problem because or lead to stress corrosion cracking in service.
these alloys have significant sigma-forming ten- They are commonly stress relieved at tempera-
dencies. Sigma forms initially because solidifi- tures slightly below the annealing temperature,
cation segregation causes local enrichment of so that residual stresses may be relieved by
sigma-promoting elements, such as molybde- creep. One hour at 900 °C (1650 °F) reduces
num. It can also form from slow cooling of slabs residual stress by about 85%. Lower tempera-
or hot bands. This latter sigma forms at grain tures require exponentially longer times for the
boundaries and will cause embrittlement and re- same stress relief, with times doubling for each
duced corrosion resistance, so it must not only 100 °C (180 °F) decrement as decreasing diffu-
be redissolved, but also the alloy must be ho- sion rates, which govern creep, are encountered.
mogenized to remove the residual concentration Cold-worked austenitic stainless steels have a
gradients from the sigma. If this is not done, markedly diminished proportional limit, partic-
chromium- and molybdenum-depleted regions ularly in compression. This Bauschinger effect,
will still exist, and sigma will re-form much which arises from the easy mobility of disloca-
more rapidly during subsequent exposure to tions, can be eliminated by stress relieving at
high temperatures. For this reason, the higher around 350 °C (660 °F) for 2 h, which provides
ends of the annealing ranges are recommended, the thermal energy for dislocation interactions
and annealing times should be generous. Newer to lock into place. This produces a sharp yield
alloys have higher nitrogen contents to suppress point without premature nonproportional elastic
formation of sigma and other deleterious inter- deformation.
Chapter 13: Thermal Processing / 165

Ferritic Stainless Steels phases that require dissolution. Alloys with high
chromium and molybdenum contents can form
Ferritic stainless steels, from an annealing σ and/or α', the brittle, ordered body-centered
point of view, must be discussed in two cate- cubic (bcc) phase, at temperatures below an-
gories. First are the modern, stabilized alloys, nealing temperatures, so rapid cooling is pru-
which are ferritic at all temperatures. These al- dent when chromium plus molybdenum ex-
loys behave as interstitial-free (IF) alloys be- ceeds 20%.
cause the interstitial carbon and nitrogen are re- The driving force for recrystallization in these
moved from solution as a stable precipitate. In alloys is limited by the lower stored energy
the second category are the older ferritic steels, from deformation inherent in the bcc structure.
which have enough austenitizing elements, usu- In addition, the pronounced deformation texture
ally carbon, in solution to cause them to form leads to annealing responses that are more accu-
austenite at what would otherwise be a good an- rately characterized as recovery and grain
nealing temperature. This makes them truly growth with diminished recrystallization. These
quasi-martensitic alloys, and they must be alloys retain this texture after annealing, and
treated accordingly. Table 2, which lists heat- this characteristic anisotropy is exploited for
treating temperatures for ferritic stainless alloys, good drawability. The major concern is to avoid
also shows which grades fit into which category. excessive annealed grain size, which greatly re-
duces toughness. Anneal at the higher end of the
Soaking range only if the loss of toughness associated
Heating of ferritic stainless for hot working is with large grain size is not a concern. Stabiliz-
straightforward. Whether stabilized or not, ing anneals are normally unnecessary for stabi-
these alloys are heated to the 1000 to 1100 °C lized ferritics as their high diffusion rates ensure
(1830 to 2010 °F) range for hot working. The freedom from knife-edge attack due to sensiti-
superferritics can be heated to up to 1300 °C zation from free unbonded carbon combining
(2370 °F). At this temperature, no debilitating with chromium at grain boundaries. The stabi-
phases occur, and ductility is good. The high lizing additions of titanium and/or niobium tie
diffusion rate inherent to the ferritic structure up the carbon as stable TiC or NbC, which does
makes homogenization easy. As long as hot not redissolve during annealing.
working is completed at temperatures above The interstitial-bearing ferritic stainless
that at which austenite forms, good hot ductility steels must be annealed subcritically, or the for-
is expected. This is not a concern with IF alloys, mation of austenite at higher temperatures
which do not form austenite. would make martensite formation on cooling
virtually unavoidable. Thus, a typical primary
Annealing anneal cycle for a typical alloy such as 430
would be nearly 24 h at 750 °C (1380 °F), the
The IF ferritics do not undergo any phase majority of which is thermal equilibration of
change with temperature during the course of the large coil mass. The actual time at tempera-
properly executed heat treatment. The objective ture required is less than 1 h. Continuous an-
of annealing is generally simply to remove the nealing is not practical because the diffusion of
effects of cold work. This is because they do not carbon is too slow to occur in the dwell time at
need to have carbon put into solution and, ex- temperature typical in continuous annealing
cept in rare cases, do not have intermetallic lines. This cycle also precipitates essentially all
the carbon and nitrogen as mixed Cr/Fe car-
Table 2 Recommended annealing temperatures bides and nitrides and homogenizes chromium
for ferritic alloys content. This necessarily slow process permits
Annealing Annealing
Alloy temperature, oC temperature, oF
subsequent subcritical annealing for mechani-
Stabilized, Cr+Mo<20
cal properties (to alleviate the effects of cold
409, 439,18 SR 870–925 1600–1700 work) to be done in a few minutes since carbon
Unstabilized, Cr+Mo<20 has been eliminated from solution by the for-
405, 430, 434, 436 705–790 1300–1450
Stabilized, Cr+Mo>20 mation of fairly stable carbides. Since the ma-
29-4C, Monit, Seacure, 444 1010–1065 1850–1950 terial is generally purchased in the annealed
Unstabilized, Cr+Mo>20, condition, the user need never be concerned
446 760–830 1400–1525
with such lengthy anneals.
166 / Stainless Steels for Design Engineers

Stress relieving is rarely a concern for any martensite cannot be avoided by furnace cool-
type of ferritic stainless. Unstabilized grades ing from austenitic temperatures, then only sub-
should not be welded, and if they are, full sub- critical annealing is feasible. But, even for
critical annealing is required. Stabilized grades nickel-free alloys the hardenability is so great
have no need for postweld heat treatment. Low- that annealing by slow cooling is quite difficult.
temperature heat treatment runs the risk of α' Martensitic alloys are put into the annealed con-
formation and is best avoided. dition for processing before they are quenched
and tempered for their final use. Thus, the more
economic subcritical anneal is the predominant
Martensitic Stainless Steels annealing heat treatment.
The nickel-bearing alloys have such high
The martensitic stainless steels resemble the hardenability that annealing in the critical range
unstabilized ferritic stainless steels described. cannot produce softening by any practical cool-
The martensitic stainless steels form essentially ing rate, so subcritical annealing is always rec-
100% austenite on heating and have very high ommended for these alloys. Nickel reduces the
hardenability, so their ability to be softened by temperature at which austenite is stable as
annealing is limited. The traditional martensitic shown in Chapter 9, Fig. 9. Other additions like
stainless steels are iron/chromium/carbon al- vanadium, molybdenum, and tungsten promote
loys, sometimes with a small amount of nickel secondary hardening and tempering resistance,
and/or molybdenum. More recently, alloys have and subcritical annealing of these alloys be-
been developed for petroleum applications that comes a slow, difficult process. This is a charac-
contain high copper, nickel, and/or molybde- teristic of the so-called super 12Cr alloys.
num and low carbon. The principles of heat Martensitic alloys have lower corrosion re-
treatment of the two alloy categories are the sistance in the annealed condition than in the
same. The more highly alloyed newer alloys hardened condition because in this state they
are, in fact, simpler to heat treat because their have the maximum amount of chromium tied up
low carbon and nitrogen levels alleviate the as chromium carbide.
need to temper.
Soaking Table 3 lists the austenitizing and tempering
Hot working should be carried out in the ranges for martensitic stainless steels. Full
austenitic range. Temperatures for this are austenitizing is crucial to producing a micro-
listed in Table 3. Forging and hot working structure that is fully martensitic. Only austenite
should always be followed by annealing to transforms to martensite. If other constituents,
avoid stress cracking due to the deep hardening such as δ ferrite or carbides, exist during the
of these alloys. austenitizing heat treatment before quenching,
they will not transform to martensite. Some
alloys, such as the 440 group, have enough car-
bon that the austenitizing temperature deter-
Martensitic stainless steels can be annealed mines how much carbon is put into solution.
by subcritical anneal and sometimes by full an- The carbon in solution in the austenite will be-
neal depending on alloy level. If the alloy level come the carbon level in the martensite, which
is such, as in the nickel-containing grades, that directly determines strength and corrosion

Table 3 Recommended annealing, austenitizing, and tempering temperatures for martensitic alloys
Alloy Subcritical anneal, Full anneal, Austenitizing, Tempering, low Tempering, high
C (oF) o
C (oF) o
C (oF) range, oC (oF) range, oC (oF)
Straight 650–760 830–885 925–1010 205–370 400–700
Cr, C<0.20, 410, (1200–1400) (1525–1625) (1700–1850) (565–605) (1050–1125)
Ni/Mo, C<0.20, 620–705 Not recommended 980–1065 205–370 400–700
414, 431, 415, (1150–1300) (1800–1950) (565–605) (1050–1125)
425, C>0.20
440A,B,C F, 420 675–760 845–900 1010–1065 150–300 Not recommended
(1245–1400) (1555–1650) (1850–1950) (300–700)
Chapter 13: Thermal Processing / 167

resistance because undissolved carbides contain precipitate with the abundant iron atoms first. At
chromium, which diminishes that available for higher temperatures and longer times, more ther-
corrosion resistance. Austenitizing temperature modynamically stable carbides, such as Cr23C6,
and holding time become most significant when form. Carbide formation is a complex function
carbon exceeds 0.20%, where its solubility is a of temperature, time, and composition. The
steep function of temperature. growth of carbides reduces strain and hardness.
The ␦ ferrite is a generally undesirable phase There are exceptions, such as the precipitation of
that can be produced by temperature excursions Mo2C, whose morphology produces a precipita-
or composition variations. Excessive austenitiz- tion hardening (PH), called secondary harden-
ing temperatures can cause its formation, as can ing. Niobium and vanadium also form carbides
low levels of carbon, which may be originally that result in higher hardness at all tempering
present in the alloy or arise from decarburiza- temperatures. Had this been understood early
tion. It will cause lower hardness and toughness on, these steels could have been correctly in-
if present. cluded in the PH stainless group. The PH steels,
Heating rates should be such that a uniform AM-350 and AM-355, both derive their PH
temperature is attained before the allotropic from the precipitation of Mo2C and Mo2N. In all
transformation from bcc to face-centered cubic other cases, higher tempering temperatures lead
(Fcc), which involves a more than 1% linear di- to lower hardness.
mension change and can cause distortion or The nickel-bearing alloys have a restricted
cracking. Oxidation during austenitizing can upper tempering temperature because of the
cause serious carbon loss on the surface, which danger of re-forming austenite, which would
will result in serious loss of surface hardness. then transform to untempered martensite during
Heating 410 in air for 10 min at 1100 °C (2010 cooling, requiring a second tempering operation.
°F) can cause surface carbon to decrease by Intermediate temperatures can lead to the
one-half, lowering hardness from HRC 45 to phenomenon of temper embrittlement, which is
under 20. caused by the precipitation of phosphorus and
Quenching rate is not a significant issue for other species, such as, but not limited to, car-
the martensitic stainless steels since they have bides, at prior austenitic grain boundaries. This
such high hardenability, but some, especially phenomenon is distinct from the precipitation of
those with higher carbon levels, may have re- α', which causes the so-called 475 °C embrittle-
tained austenite, which can lower hardness and ment, which occurs more severely in alloys
cause problems with dimensional stability. The with higher chromium levels. Because the for-
quenching rate must be sufficient, however, to mation of martensite is diffusionless, the
avoid precipitation of carbides in the austenite austenite boundaries maintain the microcompo-
during quenching since the sensitization would sitions they have at high temperatures. Auste-
persist in the final microstructure. If this oc- nite has low solubility for impurities such as
curs, a subzero treatment at below –75 °C phosphorous, so phosphorous is more highly
(–100 °F) should be undertaken immediately concentrated in the grain boundary regions.
after quenching. This compositional inhomogeneity can be elim-
inated by higher tempering temperatures or by
the addition of molybdenum, which combines
Tempering with the phosphorus and prevents the embrittle-
Untempered martensite has insufficient tough- ment. The existence of temper embrittlement is
ness to be a useful engineering material. During the reason that Table 3 recommends avoiding
tempering, carbon is precipitated from the super- certain temperature ranges for tempering.
saturated state it is in when it is quenched into Lower tempering temperatures and higher
the bcc martensite structure during the diffusion- austenitizing temperatures are best for corrosion
less transformation. The strain energy associated resistance because both minimize the amount of
with the lattice strain of the bcc martensite chromium tied up as carbide. Quenching from
caused by the poor fit of the carbon in the tetrag- higher temperatures also enhances toughness.
onal interstices is very large. Heating to even
low temperatures allows carbon enough mobil-
Stress Relieving
ity to diffuse and precipitate as carbide. Since
carbon diffuses 106 times as fast as iron, If quenched martensitic steels are not imme-
chromium, or other carbide formers, it tends to diately tempered, then they should be promptly
168 / Stainless Steels for Design Engineers

stress relieved. Otherwise, the residual stress eliminating retained austenite, enhances dimen-
from quenching could result in stress corrosion sional stability but diminishes toughness. Table 4
cracking even in seemingly benign environ- lists the solution treatments for all PH grades.
ments. Stress relieving is simply a low-tempera- The as-quenched state is called condition A.
ture tempering operation, but some elimination This is the normal condition in which the mate-
of residual stress does occur. Higher tempera- rial is supplied from the mill and is intended to
tures and longer times produce greater stress be soft enough for machining and some forming.
relief and maximize elastic properties, but opti- If softer material is required, the H-1150M con-
mal toughness is obtained at intermediate times dition exists in which the material is first highly
and temperatures. overaged at 760 °C (1400 °F), allowing some
austenite to re-form. The subsequent aging then
overages that martensite while retaining some
Precipitation Hardening Stainless Steels stable austenite. The result is a very tough
There are three classes of PH steels, and each Aging. The time and temperatures required
requires totally different heat treatment. The to produce this precipitation are also given in
classes are martensitic, semiaustenitic, and Table 4. The condition code itself tells the tem-
austenitic. The most straightforward alloys are perature at which the aging is conducted in
the martensitic PH grades. Like the plain marten- that the code numbers are based on thermal
sitic alloys, the martensitic PH alloys are care- processing temperatures expressed in degrees
fully designed to produce a nearly fully austenitic Fahrenheit (oF), for example, TH 900 means
structure at high temperature that quenches to a transformed to martensite (T) and aged at 900
nearly fully martensitic structure on cooling. The degrees Fahrenheit. The final properties are a
martensite is low in carbon, so it is relatively soft function of both aging time and temperature.
and not prone to brittleness. So, the hardness and Lower temperatures result in higher possible
strength of these alloys is derived from a subse- hardness but lower toughness. The precipi-
quent tempering-type heat treatment during tates, as mentioned, are optically invisible and
which various constituent elements form ex- cause very little dimensional change. Contrac-
tremely fine coherent precipitates that greatly tion on the order of 0.0005 in./in. from aging is
strain and therefore strengthen the matrix. There typical, often permitting machining to final di-
are numerous precipitates that can provoke this mensions in condition A. All aging treatments
effect, and they are described in detail in the PH are above the temper embrittlement range to
chapter (see Chapter 10). All require the short- which these alloys are susceptible. Service
range diffusion of substitution elements to form temperatures in this range would result in em-
these optically invisible precipitates. brittlement, so use above 350 °C (660 °F)
should be avoided. Molybdenum-bearing
grades should be selected to minimize this
Martensitic PH Grades phenomenon if high-temperature use is con-
Solution treatment of these alloys is con- templated.
ducted to achieve a fully austenitic structure. Solution Heat Treatment and Conditioning.
The constituent elements are easily dissolved, The semiaustenitic grades are more complicated
so excessive temperature or time is unnecessary than the martensitic PH alloys. These alloys are
and could be counterproductive if it were to re- designed so that they are austenitic when
sult in ferrite formation or surface oxidation, quenched from the solution heat treatment tem-
which would be detrimental to final mechanical perature. This is also called condition A, and it
properties. The presence of retained ferrite is permits them to be highly formable. This stabi-
mainly a function of alloy and composition lization of the austenite comes mainly from
within the allowed range. Earlier grades such as higher chromium and carbon levels. These al-
17–4 and the obsolete stainless W intrinsically loys essentially resemble a lean 301 austenitic
contained some ferrite. The subsequent alloys alloy, many with some molybdenum substitut-
are substantially ferrite free. Most alloys may ing for part of the chromium. The key to these
retain some austenite after quenching to room grades is making them behave as a martensitic
temperature, in which case subzero treatment alloy. This is done by precipitating some of the
should be done within 24 h to avoid further sta- carbon as chromium carbide at a temperature at
bilizing the austenite. Subzero treatment, by the high end of what would normally be consid-
Chapter 13: Thermal Processing / 169

Table 4 Recommended annealing and stress-relieving temperatures for martensitic PH grades

Alloy Condition code Solution anneal Conditioning Aging
13-8 A 925 °C 15 min, oil or air cool below 15 °C –75 °C 8 h
RHxxx 925 °C 15 min, oil or air cool below 15 °C –75 °C 8 h xxx °F for 4 h
Hxxx 925 °C 15 min, oil or air cool below 15 °C ... xxx °F for 4 h
15-5 A 1035 °C 30 min, oil or air cool below 15 °C ...
Hxxx 1035 °C 30 min, oil or air cool below 15 °C ... xxx °F for 4 h
17-4 A 1035 °C 30 min, oil or air cool below 15 °C ...
Hxxx 1035 °C 30 min, oil or air cool below 15 °C ... xxx °F for 4 h
450 A 1035 °C for 1 h, water quench ...
Hxxx 1035 °C for 1 h, water quench ... xxx °F for 4 h
455 A 830 °C for 1 h, water quench ...
Hxxx 830 °C for 1 h, water quench ... xxx °F for 4 h
465, 275 A 980 °C for 1 h, cool rapidly –75 °C 8 h
Hxxx 980 °C for 1 h, cool rapidly –75 °C 8 h xxx °F for 4 h
475 A 925 °C 1 h, air cool –75 °C 8 h
Hxxx 925 °C 1 h, air cool –75 °C 8 h xxx °F for 4 h

17-7, 15-7 A 1065 °C for 30 min, air cool ... ...
T 1065 °C for 30 min, air cool 760 °C (1400 °F), 90 min, ...
air cool to RT for 30 min.
C 1065 °C for 30 min, air cool Cold reduce ...
R 1065 °C for 30 min, air cool 955 °C (1750 °F) 10 min, air ...
cool, chill to –75 °C for 8 h
THxxx 1065 °C for 30 min, air cool 760 °C (1400 °F) xxx °F for 4 h
90 min, air cool to RT for 30 min
CHxxx 1065 °C for 30 min, air cool Cold reduce xxx °F for 4 h
RHxxx 1065 °C for 30 min, air cool 955 °C (1750 °F) 10 min, air xxx °F for 4 h
cool, chill to –75 °C for 8 h
AM-350 A 1010–1065 °C ...
L (equivalent to T) 1010–1065 °C 930 °C for 90 min, air cool ...
SC (equivalent to R) 1010–1065 °C 930 °C for 90 min, air cool, ...
180 min at –75 °C
SCTxxx 850 °F or 1010–1065 °C 930 °C for 90 min, air cool, xxx °F for 180 min
1000 °F 180 min at –75 °C

DA (double aged) ... 930 °C for 90 min, air cool, 450–540 °C 180 min
730–760 °C 180 min
Am-355 A 1025–1040 °C ... ...
L (equivalent to T) 930 °C for 90 min, air cool ...
SC (equivalent to R) 930 °C for 90 min, air cool. ...
180 min at –75 °C
SCTxxx 850 °F or 930 °C for 90 min, air cool, xxx °F for 180 min
1000 °F 180 min at –75 °C
DA (double aged) 930 °C for 90 min, air cool. 440–470 °C
730–760 °C 180 min for 180 min
Equalized and 930 °C for 90 min, 540–590 °C for
overtempered air cool, 730–760 °C 180 min 180 min
A-286 ST1650 900 °C 120 min, oil/water quench ... ...
ST1650A 900 °C 120 min, oil/water quench 730 °C 16 h
ST1650DA 900 °C 120 min, oil/water quench 730 °C 16 h,
650 °C 8 h
ST1800 980 °C 120 min, oil/water quench
ST1800A 980 °C 120 min, oil/water quench 730 °C 16 h
PH, precipitation-hardenable; RT, room temperature.

ered sensitization, 760 °C (1400 °F). The car- finishes (Mf) near room temperature. This
bide precipitation occurs at the interfaces of the process is called austenite conditioning. The
small amount of residual ferrite these alloys heat treatment scheme just described would be
have and also at grain boundaries. The deletion condition A-1400. This material, after quench-
of carbon and chromium from the matrix ing to room temperature, would be said to be
changes the matrix composition sufficiently that condition T.
the temperature for the start of the martensitic The higher the temperature of the condition-
transformation (Ms) of the depleted austenite in- ing, the less carbon is precipitated and the lower
creases from below zero to about 65 to 100 °C the resulting Ms. The highest conditioning tem-
(18 to 212 °F). The martensitic transformation peratures of 955 °C (1750 °F) cause a sufficiently
170 / Stainless Steels for Design Engineers

low Ms that subzero treatment is required to for a normal solution anneal of non-PH austen-
obtain the fully martensitic structure required for ite. The subsequent precipitation aging requires
age hardening. This would be called A-1750. higher temperatures and longer times because
After subzero treatment at –73 °C (–100 °F), it diffusion is much slower in austenite.
would be called condition R-100.
It is even possible to obtain the fully marten-
sitic structure by cold work from full condition Duplex Stainless Steels
A, which is quenched from 1065 °C (1950 °F).
This is called condition C and requires heavy Duplex stainless steels are both ferritic and
cold rolling to accomplish. Since there is no heat austenitic, so their heat treatment combines the
treatment to precipitate carbon from the austen- same elements and principles as their principle
ite matrix, the resulting martensite is the hardest. phases.
The aging treatments of the semiaustenitic
alloys are identical to those for the martensitic
alloys because the treatments are standardized.
The resulting mechanical properties vary in a Duplex alloys are multiphase at all useful
complex fashion with the alloy composition and working temperatures, making their hot worka-
thermomechanical treatment history before bility quite poor. It is extremely important to
aging. The principles at work that determine the drive sulfur to the lowest possible levels, less
mechanical properties, besides the strain in- than 0.001%, to achieve satisfactory hot ductil-
duced by the precipitate phase, are: ity. Otherwise, soaking should proceed the same
• Carbon level of the quenched martensite as for ferritic stainless alloys since ferrite con-
• Amount of cold work of austenite prior to stitutes the continuous phase to be worked.
Because these alloys always contain high
aging chromium and generally high molybdenum,
These factors do not come into play with the they should be cooled as rapidly as possible
martensitic PH grades because all of the carbon from high temperatures to avoid formation of
is in solution, and they are not significantly cold sigma or other intermetallic phases.
worked prior to aging. The semiaustenitic PH
alloys, because of the variable amount of car- Annealing
bide precipitation, have an equally variable
chromium content, so corrosion resistance will The function of annealing in the duplex alloys
vary accordingly. The highest-temperature solu- is generally to:
tion and carbide precipitation treatments pro- • Remove the effects of cold work
vide the best corrosion resistance, as do the • Restore the balance between the volume
lowest aging temperatures. fraction of ferrite and austenite
• Achieve equilibrium composition within
Austenitic PH Alloys both the austenite and ferrite
• Dissolve unwanted intermetallic phases
In contrast to the martensitic and semi-
austenitic PH alloys, the austenitic PH alloys The annealing range of duplex alloys is some-
are austenitic under all conditions. Their what restricted, approximating the overlap be-
strengthening reaction comes from precipitation tween what each of the two constituent phases
of titanium- and aluminum-nickel intermetallic would be annealed at separately. Table 5 lists
phases within the austenite matrix as occurs in the normal annealing temperatures for these al-
superalloys. Solution treatment is very typical loys. The use of nitrogen as a key alloying

Table 5 Recommended annealing and stress-relieving temperatures for duplex alloys

Annealing Annealing
Alloy temperature, °C temperature, °F ASTM A480 2006 Stress relieving, °F Stress relieving, °C
Lean duplex, Cr+Mo<23, 1010–1100 1850–2010 Various Not recommended
2003-2101, 19-D-2304
Medium alloy, Cr+Mo<26, 1040–1100 1900–2010 1040 min Not recommended
Cr+Mo>26, 2507, 52N+, 1050–1150 1925–2100 Various Not recommended
Zeron 100, 255
Chapter 13: Thermal Processing / 171

element has improved the annealing behavior of tions, either from ␣' beginning at 350 °C (660
these alloys since its diffusion is quite rapid, °F) or from ␣/␹, which takes over at 600 °C
causing ferrite-austenite equilibrium to be at- (1110 °F), would occur and so is not indicated
tained very rapidly. Nitrogen also hinders (for- for these alloys.
mation and facilitates the dissolution of second-
ary austenite, which can form after quenching
from welding temperatures and cause regions of REFERENCES
poor corrosion resistance.
These alloys are not very susceptible to car- 1. A. Garner, The Effects of Autogenous
bide sensitization and normally have very low Welding on Pitting Corrosion in Austenitic
carbon content. Thus, the guiding principle in Stainless Steel, Corrosion, Vol 35 (No. 3),
annealing is simply to achieve phase balance 1979, p. 108–114
and avoid cooling so slowly that intermetallic 2. J.F. Grubb, personal communication, June
phases may form. 4, 2006
The strengthening of duplex is normally 3. Data courtesy J&L Specialty Steel, October
achieved by the strong grain refinement and 2, 1998
solid solution hardening. No strengthening heat 4. J.F. Grubb, Proceedings of the International
treatments are used. Stress relief would have to Conference on Stainless Steels, 1991,
occur at temperatures at which embrittling reac- Chiba, ISIJ
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 173-180 All rights reserved.
DOI: 10.1361/ssde2008p173



Summary shown that it is important to understand and

exploit these characteristics to optimize forming
STAINLESS STEELS RANGE in formability of stainless steels.
from the extremely formable austenitic alloys to
the hard-to-form martensitic alloys. For steels
with equivalent corrosion resistances, formabil- Flat, Rolled Stainless Steel
ity increases with the level of austenitizing al-
loying elements. The ferritic alloys are the least The vast majority of carbon steel and, espe-
alloyed, least expensive, and least formable; cially, stainless products are flat products. These
duplex steels are next, and austenitic steels are steels are formed by bending, roll forming, spin-
the most formable but most expensive. How- ning, hydroforming, and deep drawing. Roll
ever, if the specific structural anisotropy of forming is most commonly used to produce
ferritic alloys, which gives particularly good welded pipe and tubing and is simply bending
deep-drawing characteristics for a given level of done in a continuous manner. Bending is a sim-
ductility, can be exploited, the best forming ple operation, and there is no meaningful change
economies can be gained. in thickness of the sheet during the operation.
The higher yield strength and work-hardening
rates of most stainless steels will result in greater
Introduction springback than would be experienced in carbon
steel. Tooling must be adjusted to compensate
The technology for forming stainless steel is for this. In neither of these forming methods is
quite similar to that for forming carbon steels. there large motion of the formed material across
The primary difference is the higher strength of the die surface, so lubricant is not normally
all types of stainless steels compared to drawing- used. The reader is referred to the Forming and
type carbon steels. This higher strength requires Forging in Volume 14 of the ASM Handbook for
greater sophistication in tooling and lubricants, detailed charts on bend radii and springback re-
and it requires more powerful forming equip- lated to bending.
ment. The higher initial strength is also Other forming techniques employ more com-
accompanied by a higher work hardening rate in plex deformation processes. Deep drawing is
austenitic stainless steels, which further distin- the foremost of these. Figure 1 (Ref 1) shows
guishes them from carbon steel. Galling also schematically what occurs during drawing. A
must be recognized as a danger and prevented. round blank is held between dies over a cavity,
Stainless steel has lower thermal conductivity and a punch pushes the material into the cavity
than carbon steels, which can cause it to retain to produce the part. If the dies pinch the blank
heat from deformation and friction, thereby to only a small degree, the process is normal, or
decreasing lubricity. Last, stabilized stainless al- ordinary, drawing. If the dies significantly
loys contain abrasive carbide microconstituents. restrain the periphery of the blank so it cannot
The various types of stainless steel have very move, stretch forming occurs. Material proper-
different deformation characteristics in terms of ties determine whether a material is most suc-
strain hardening and anisotropy. It will be cessfully formed by stretch forming or drawing.
174 / Stainless Steels for Design Engineers

rate. If it work hardens faster than it becomes

thinner, the strain is distributed, and local fail-
ure is prevented. Austenitic materials have the
face-centered cubic (fcc) crystal structure with
many slip systems and low stacking fault ener-
gies. This means that they can generate many
complex arrays of tangled dislocations, which
cause strain hardening. They can also transform
during stretching to the much harder martensite,
for which the deformation is greatest, again
redistributing the deformation away from the
potentially thinning area. This makes austenitic
stainless steels particularly suitable for stretch
Deep Drawing. Deep drawing, referred to
simply as drawing, without stretching requires a
different material characteristic. For drawing, a
low work-hardening rate is desirable so that the
material can compress in the circumferential
direction while elongating in the radial direction.
Obviously, a high ability to elongate is always
useful regardless of any other characteristic.
But, ferritic material has one other advantage:
Body-centered cubic (bcc) alloys have more
slip systems than fcc alloys. When bcc alloys
are rolled to become flat stock, they may retain
a preferred crystallographic orientation, called
texture, as a result of the deformation. This non-
random crystal structure can cause the material
to have higher strength in the through-thickness
direction. This directional variation in proper-
ties is called anisotropy.
When a material with desirable texture is
stretched, it flows in the stretching direction and
contracts laterally at lower stresses than are
required to initiate plastic flow in the through-
thickness direction. As long as the work-hard-
ening rate keeps the flow stress below the
through-thickness yield strength, there will be
no thinning. The geometry of deep drawing
with constraint fits such materials’ capabilities.
This is why carbon steels and ferritic stainless
steels deep draw well. If the material were con-
strained from contracting while being stretched,
the tensile strength would be exceeded before
through-thickness flow occurred and the mate-
rial would fracture with little deformation. So,
Fig. 1 Deep drawing schematic. Source: Ref 1 these materials cannot be stretch formed. Even
without stretching, some hold-down pressure is
Stretch Forming. In stretch forming, the required to prevent wrinkling of the blank
material is constrained from moving wholly before it is pulled into the die. This is more pro-
into the die. Thus, the section that enters the die nounced for thinner blanks and for material
is stretched more and must become thinner. with higher work-hardening rates. In many
Whether the material becomes so thin locally forming situations, adjustments to the drawing
that it fails is governed by its work-hardening process (i.e., hold-down force adjustment, draw
Chapter 14: Forming / 175

bead contour, blank size, lubrication, die radii, and 2.0 are more common. Figure 2 shows the
etc.) may be more important than material prop- LDR as a function of the Lankford ratio. It
erties in determining whether the desired part should noted that the very ductile 304 fares no
can be made successfully. better than carbon steel because of the advantage
The material properties that are important to of a beneficial anisotropy, which the ferrous bcc
formability are ductility, as measured by tensile structure has. Indeed, the best-performing deep-
test elongation; the work hardening rate, which drawing stainless steels are low interstitial ferritic
is the instantaneous slope of the true stress, true steels with boron added. A number of stainless
strain curve and is called n; and the anisotropy. steels are compared in Table 1 (Ref 4).
The measure of anisotropy is the Lankford ratio It shows that when comparing different types
(Ref 2), expressed as: of materials, some tests are not good predictors
of deep-drawing performance. Optimizing the
material/drawn component combination is far
r0 + 2r45 + r90 from simple, especially when other considera-
R= (Eq 1)
4 tions, such as cost and material performance in
service must be factored in. End users are
where R is the average strain ratio, r0 is the strain encouraged to deal directly with the producing
ratio in the longitudinal direction, r45 is the strain mill early in the design stage of any new high-
ratio measured at 45° to the rolling direction (of production, deep-drawn component. The pro-
the sheet metal-forming operation), and r90 is the ducing mills, while not necessarily exhaustive
strain ratio in the transverse direction. R deter- sources of information, are certainly reservoirs
mines the average depth (that is, the wall height) of knowledge of current practice.
of the deepest draw possible. When this expres- The most widely used summary of a mater-
sion equals 1, then a material may be considered ial’s formability is contained in its forming limit
isotropic, that is, the material properties are the diagram (FLD), developed by Keeler and Back-
same for all crystallographic orientations. As a ofen (Ref 5). This diagram shows the locus of
first approximation, the Lankford ratio equals failure under varying strain states. Figure 3
the ratio of the lateral strain to the through-thick- shows a comparison of the FLDs for austenitic
ness strain during the tensile test of a sheet spec- stainless steel and carbon steel. The FLD tells
imen. As the value increases from 1, the drawa-
bility increases because the material tends to
maintain a constant thickness while changing
shape from a flat blank to a cup shape. The abil-
ity to be deep drawn is measured by the limiting
drawing ratio (LDR), the ratio of the diameter of
a disc to that of the deepest cylinder into which
it can be drawn.
The ferritic stainless steels in sheet form have
LDRs of around 2.2 compared to 2.0 for 304. For
austenitic steel, the ratio is about 1.0, while for
flat-rolled carbon steel and ferritic stainless steel,
Fig. 2 Limiting drawing ratio variations with Lankford ratio.
it can be greater than 2.0, but values between 1.5 Source: Ref 3

Table 1 Deep-drawing materials comparison

0.2% proof Tensile Erichsen Conical cup
Lankford value
stress, strength, Elongation, Hardness value, value,
Steel N/mm1 N/mm2 % HV n value ro r45 r90 r̄ mm mm
YUS 190 343 497 33.8 173 0.20 1.60 1.47 2.10 1.66 9.5 26.7
YUS 436S 275 459 34.8 135 0.21 1.67 1.63 2.12 1.76 9.8 26.9
YUS 436S 284 483 34.5 137 0.22 1.49 1.90 2.01 1.83 9.8 26.8
YUS 4O9D 239 424 37.2 116 0.24 1.51 1.77 2.11 1.79 9.7 26.7
SUS 430 308 472 31.8 159 0.21 0.94 0.92 l.50 1.07 8.9 28.5
SUS304 281 705 64.0 172 0.44 0.91 1.19 0.83 1.03 12.5 27.0
Source: Ref 2
176 / Stainless Steels for Design Engineers

Fig. 4 Forming limit diagrams for categories of stainless

steels. A, austenitic stainless steel; F, ferritic stainless
Fig. 3 Forming limit diagram of carbon steel compared to steel; HAS, high-strength austenitic stainless steel; HSF, high-
austenitic stainless steel. Source: Ref 3 strength ferritic stainless steel; FA(50), ferritic-austenitic stainless
steel. Source: Ref 3

the point of failure for a given sheet material

with a given thermomechanical history over a
full range of combined strain states. These dia- total interstitial content (i.e. carbon plus nitro-
grams are generated by examining circle grids gen) and by thermomechanical working to give
printed on material that is deformed to failure in a fine-grained, fully recrystallized, yet benefi-
various modes. The single most important value cially anisotropic, microstructure. Figure 5
on the curve is the intersection of the curve with shows how the FLD of an enhanced 409 ferritic
the major strain axis at zero minor strain. This stainless steel, 409 Ultra Form, compares to the
can be used as an index of formability and is the already highly evolved 409.
elongation possible for plane strain conditions. For austenitic stainless steel, maximum
This value within a given class of materials is drawability is obtained by low work-hardening
proportional to the strain-hardening exponent rates coupled with maximum elongation, as
because a higher work-hardening rate causes exemplified by 305 and high-nickel 304. This
higher localized resistance to thinning, which is comes with a cost penalty as the easiest way to
the precursor of failure. improve formability is to increase the nickel
There is much more variety within the fami- level, although replacing the expensive nickel
lies of stainless steel than within carbon steel. with copper or manganese, as in 204Cu, has
Figure 4 shows generalized FLDs for austenitic, been shown to be effective. Many austenitic
ferritic, high-strength ferritic, high-strength components are made by stretched deep draw-
austenitic, and duplex stainless steels. ing. In this case, the preferable alloys are the
Deep drawing of components is seen as a way leaner austenitics, 201 and 301. These alloys
to obtain near-net shape. Since tooling is costly, it form martensite more rapidly than do 304 and
is necessarily a high-volume application. Quite 305 during stretching. Martensite has a 4%
often, designers push component design to the greater volume than the austenite from which it
limit of a material’s ability to be formed. There forms and a much greater strength. This gives
are various drivers that cause this. One is to elim- 201 and 301 the ability to redistribute deforma-
inate extra operations or components by consoli- tion from thinning areas elsewhere and stretch
dating them into one more complex deep-drawn extensively, making them an optimal material
part. Another is to use the least-expensive alloy. for objects such as sinks.
In some industries, such as the household appli- The specific alloy composition is often finely
ance industry, as many components as possible tuned for a given part and tooling design, and
are deep drawn from ferritic stainless alloys and small deviati ons can dramatically increase
the more costly austenitic alloys are used only breakage rates. Even such minor process
when the part cannot be made from a ferritic. changes as blank temperature variations due to
There has been much research to develop fer- ambient temperature can alter work-hardening
ritic stainless alloys with improved formability. rates enough to cause breakage problems. Cer-
This has been accomplished by reducing the tainly, this can occur when designs push the
Chapter 14: Forming / 177

Fig. 5 Optimized 409 for forming versus normal 409. Source: Ref 5

envelope of a material’s capability, but cost A last key variable in which practice and ma-
pressures generally drive designers to this terial interact is strain rate. Ferritic steels flow
extreme. more easily at lower strain rates and are thus
On very severe forming, intermediate more formable. Austenitic steels experience the
annealing may be required to either enhance opposite effect if they are susceptible to marten-
ductility or reduce required pressure. Stainless site formation. Adiabatic heating can retard the
steel in the as-drawn condition will have resid- martensitic transformation and reduce the work-
ual stress and may have sufficient hardness to hardening rate, changing their forming charac-
be susceptible to delayed failure if placed in a teristics. When drawing, this is good, but for
corrosive environment. Bright-annealed alloys stretching it may not be.
with high martensite levels from forming can Hydroforming, a variation on deep drawing
fail by hydrogen embrittlement with just the in which hydrostatic pressure forces a blank
residual hydrogen from annealing. Therefore, into the die cavity, can improve the degree to
the use of bright-annealed lean alloys such as which stainless steels can be deep drawn. The
301 is not recommended. hydroforming process avoids friction between
178 / Stainless Steels for Design Engineers

the blank and the tool. Deformation is spread forming difficulties and lack of compliance with
more evenly across the blank, and the material mechanical property requirements. They also
forms close to its theoretical best. Productivity have significant rolling anisotropy, which causes
using this technique is relatively low, so its use the yield strength transverse to the rolling direc-
is justified mainly when it is not otherwise pos- tion to be consistently higher than it is in the
sible to make a certain design in one drawn rolling direction. This behavior is contrary to the
component. An example of this may be the pro- general behavior of single-phase alloys. The dif-
duction of a complex exhaust manifold that re- ference in yield strength is sufficient, reportedly
quires a higher-alloyed ferritic stainless with up to 15% lower in the longitudinal direction
relatively low formability. than in the transverse direction required for ten-
Besides failure by breakage, there are other sile tests (Ref 7), that it is both serious design
less-severe defects found on deeply drawn and forming considerations.
parts. Austenitic steels can develop a surface Ferritics undergo a more specialized surface
condition known as orange peel, the result of relief because their anisotropic grain structure
slip planes within a grain disrupting the sur- can yield in a more concerted fashion and give
face. Orange peel is prevented by keeping even greater surface relief, called ridging and
grain size fine so that the surface relief is too roping. This is minimized by refining grain
small to be seen. Austenitic stainless steels can size, achieving full recrystallization versus just
also develop anisotropy, which while less recovery during annealing, and temper passing
severe than ferritic steels, can cause “earing,” (i.e., elongation of about 1% on a cold-rolling
in which round blanks deform nonuniformly in mill) to suppress the yield point phenomenon
a four-, six-, or eight-fold symmetry, causing that is characteristic of ferrous bcc materials.
excess ear-shaped material to extend beyond End users should always make their use of the
the intended dimensions of the component. material known to the producing mill so that
Material is wasted because larger blanks have the correct thermal processing path can be
to be used. The steel producer can minimize employed for the manufacturing process the
the phenomenon by keeping cold roll reduc- material will undergo.
tions above about 60%. One measure of the One of the most important material consider-
earing tendency is derived from the Lankford r ations for deep drawing is surface finish. Flat-
measurements: rolled stainless should be fully annealed and
pickled so that the surface holds lubricant well
and yields as readily as possible. Temper pass-
r0 + r90 + 2 r45 ing will drastically reduce the drawability of
Δr = (Eq 2)
2 stainless by smoothing the surface (rather than
increasing the yield strength). Temper passing
The left side of Eq 2, ⌬r, is a measure of the with roughened rolls does not significantly
variation of plastic strain ratio r with direction harm drawability. The surface finishes that are
in the plane of a sheet. Values of ⌬r near zero produced by temper rolls with special finishes,
generally indicate minimal tendency toward such as Koolline, retain lubricant well and can
earing, while ⌬ values significantly above or be drawn with minimal distortion.
below zero indicate increased tendency toward Tooling for stainless must be strong and wear
earing. A combination of a high R value from resistant. Traditional tooling materials are D2
Eq 1 and a low ⌬r value provides optimal tool steel and high-strength aluminum bronze.
drawability. D2 tool steel must be hardened to HRC 60 to
It should be noted that deformation always 62 and must have smooth surfaces. The use of
produces some surface relief, so highly reflec- powder metal techniques to produce tool steels
tive surfaces become spectrally diffuse, or for dies has permitted much higher volume frac-
cloudy, after plastic deformation. This has been tions of ultrahard microconstituents such as
an issue for items such as automotive bright vanadium carbide to be introduced, thereby
trim. Mechanical buffing can restore the luster, vastly improving wear resistance without harm-
but the time and expense of buffing increase ing toughness or even raising overall hardness.
dramatically if orange peel or roping (a similar The benefits of cast aluminum bronzes are low
surface defect) is excessive. friction, high thermal conductivity, and low ten-
Duplex stainless steel flat products exhibit dency to gall. They are preferred when finished
significant in-plane anisotropy that can cause part surface appearance is more important than
Chapter 14: Forming / 179

Table 2 Suitability of various lubricants for use in forming of stainless steel

Blanking Press- Drop Contour
and brake Press Multiple-slide Deep hammer roll
Lubricant piercing forming forming forming drawing Spinning forming forming Embossing
Fatty oils and C B C A C A C B B
Soap-fat NR NR C A B B C B C
Wax-base B B B A B B C B A
Heavy-duty B NR B A B B NR A B
Dry film (wax B B B NR B A B NR A
or soap
plus borax)
Pigmented B NR A B A C NR NR NR
Sulfurized or A A B+ A C NR A B A
Chlorinated oils or A(h) NR A NR A NR A(i) A NR
waxes(f) high-
viscosity types(g)
Chlorinated oils or B+ A A A B NR A(i) A A
waxes(f) low-
viscosity types(j)
Graphite or NR (l) (l) NR (l) NR (l) NR NR

A, excellent; B, good; C, acceptable; NR, not recommended; SUS, Say-bolt universal second. Ratings consider effectiveness, cleanliness, ease of removal, and other
suitability factors. (a) Vegetable or animal types; mineral oil is used for blending. (b) May be diluted with water. (c) Water emulsions of soluble oils; contain a high con-
centration of extreme pressure (EP) sulfur or chlorine compounds. (d) Chalk (whiting) is most common pigment; others sometimes used. (e) EP types; may contain
some mineral or fatty oil. (f) EP chlorinated mineral oils or waxes; may contain emulsifiers for ease of removal in water-base cleaners. (g) Viscosity of 4,000 to 20,000
SUS. (h) For heavy plate. (i) For cold forming only. (j) Viscosity (200 to 1000 SUS) is influenced by base oil or wax, degree of chlorination, and additions of mineral oil.
(k) Solid lubricant applied from dispersions in oil, solvent, or water. (l) For hot forming applications only.

absolute die life and forming pressures are medium-sized hardware items, such as screws,
moderate. bolts, nuts, rivets, and specialized fasteners.
Lubricants for stainless steel forming must be As with flat-rolled forming operations, the
able to prevent metal-to-metal contact under primary difference between carbon/alloy steels
higher pressures than those seen with carbon and stainless steels comes from the higher yield
steel. The ASM Metals Handbook, Desk Edi- strength and higher work-hardening rates of
tion, lists common lubricants as shown in Table stainless. Anisotropy is not a significant consid-
2 (Ref 1). Not listed in the table are the newer eration for long products. The ferritic stainless
thermoplastic acrylic polymers that, when ap- steels are the most easily cold headed. The use
plied to the surface at a density of around 1 g/m2 of the most formable stainless alloys for flat
(0.004 oz/ft2), provide a dry film with lubricat- products, the stabilized ferritic alloys, is limited
ing properties surpassing any of those listed in for long products because of the severe lack of
Table 2. toughness these alloys show for cross sections
greater than about 2 mm (0.08 in.). But, the
nonstabilized ferritics, the martensitic, precipi-
tation-hardenable (PH), austenitic, and duplex
Stainless Long Products grades are all cold formable.
In cold-heading terminology, the maximum
Cold heading, one of the most important possible deformation an alloy can tolerate is ex-
forming operations conducted on stainless long pressed in terms of the length of long product
products, is a forming process that increases the exposed beyond the die that can be successfully
cross-sectional area of a room temperature forged into the upset. This is measured in the
blank at one or more points along its length. number of diameters of initial stock. So, an op-
Cold heading is typically a high-speed process timal ferritic such as 430 can tolerate upsets up
in which the blank is progressively moved to about 2.25 diameters, while a very low work-
through a multistation machine. The process is hardening austenitic, such as 384, can tolerate
widely used to produce a variety of small- and 3.0. The martensitic, PH, and richer duplex
180 / Stainless Steels for Design Engineers

their work hardening is not instantly removed

by annealing. For example, the initial pressure
requirements for a given deformation range for
304 can be three times what is required to de-
form carbon steel and up to five times for higher
molybdenum alloys. For greater deformation,
this multiple increases (see Fig. 6a and b).
Austenitic stainless steel loses ductility when
heated above 1280 °C (2335 (F) because of low
melting phases in the grain boundaries. As
much as possible, all alloys should be hot
worked in a single phase field of the phase dia-
gram to avoid mixtures of ferrite and austenite
since the great difference in their great strengths
can cause failures. Duplex stainless steels and
other alloys (e.g., 17-7 PH) that have high lev-
els of ferrite in austenite or austenite in ferrite at
the hot-working temperature exhibit reduced
hot ductility compared to either fully austenitic
or fully ferritic stainless steels and are more dif-
ficult to hot work.


1. Forming of Sheet, Strip, and Plate: Deep

Drawing, in Metals Handbook, desk ed.,
J.R. Davis, Ed., ASM International, 1998, p
2. U.F. Kocks, C. Tomé, H.-R. Wenk, Ed., Tex-
ture and Anisotropy, Cambridge University
Press, Cambridge, UK, 1998
3. E. Schedin, “Forming Stainless Steel,”
Fig. 6 Forces required for hot working. Source: Ref 1 ACOM Technical Paper, www.outokumpu
4. H. Sumitomo and T. Tanoue, Nippon Steel
alloys are in the 1.5 to 2.0 range of formability. Technical Report 71, October 1996
The lean duplex, when they find their way into 5. S.P. Keeler and W.A.A. Backofen, ASM
wider use, should resemble 430 with 2.25 diam- Trans. Q, Vol. 56 (No. 163), 1963, p 25–48
eters maximum. 6. “409 Ultra Form Stainless Steel,” Product
Hot forming of stainless steel is done as an in- Data Bulletin,
tegral part of their production; therefore, all 7. R. Cordewener et al., “Duplex Stainless
stainless alloys can be forged. The main issue is Steels,” Paper 109, TWI conference, Glas-
that the high hot strength of stainless requires gow, 1994
much more force than would be required for
carbon steels. Martensitic stainless steels re- SELECTED REFERENCE
quire 10 to 100% more force than 4340 alloy
steel, while the austenitics require much more • ASM Handbook, Vol. 14, Forming and
because of their high hot hardness and because Forging, ASM International, 1988
Stainless Steels for Design Engineers Copyright © 2008 ASM International®
Michael F. McGuire, p 181-191 All rights reserved.
DOI: 10.1361/ssde2008p181



Summary From a more focused viewpoint, the machin-

ability of a material is further described by:
complex operation. Not only does a shop need 1. Consistency: Does the material machinabil-
the correct supporting equipment and supplies, ity stay the same when bundles are changed?
a better understanding of the metal itself is ad- 2. Tool life/wear: How long does the tool last
vantageous. Technology in the production of a in the machining operation? This could be
more machinable stainless steel is advancing. minutes, hours, shifts, or days.
The incorporation of complex oxides has led to 3. Productivity: How many parts were made
the development of materials that allow higher in an hour, shift, or day?
machining speeds and increased productivities, 4. Cost per part: What is the cost of the final
both of which are reducing machining costs and geometry?
keeping shops competitive. 5. Cycle time: How fast can a part be com-
6. Surface finish: How smooth or shiny is the
Introduction 7. Chip control: Are the chips manageable?
8. Maximum cutting speed: How fast can the
Stainless steel forgings, castings, plate, and part be cut without affecting tool life?
long products all are frequently machined. This 9. Maintaining tolerances: How long can the
fundamentally involves the removal of a layer machining operation continue before ad-
of material from the workpiece with a cutting justments are made?
tool one or multiple times until a finished or 10. Minimal operator intervention: Does the
semifinished part is produced. Machining, in it- operator need to constantly adjust setup?
self, is a complex topic with many variables.
Rather than attempt to understand all aspects of
machining, it is helpful to consider a material’s
machinability, that is, its ability to be machined
and the factors that affect its ability to be ma-
chined. In Fig. 1, a macroview shows how the
machinability of a material is influenced by the
interaction of humans, machine, methods, mate-
rial, and management. Some of the variables
can affect the appearance of the material, while
others affect the performance of the piece, mak-
ing machining an art as well as a science. Opti-
mum machinability is obtained when each of
these sectors come together, providing the best
possible conditions for efficient machining. Any
change in one of these sectors can change the
behavior or efficiency of a machining job. Fig. 1 The 5 M’s of machinability
182 / Stainless Steels for Design Engineers

This list is somewhat empirical or job re- ƒ(chemistry, cleanlinesss,

lated, but it provides guidelines for defining a Machinability of
job since cutting conditions can be very differ- stainless steels = structure, processing,
cross section)
ent for each material and part. For example, if
the surface finish of a part is very important, it
may be necessary for chip control and tool life Each variable contributes uniquely to machin-
to be sacrificed. Clearly, machining involves ability. Machine shops and users generally have
much more than simply cutting a piece of very little influence on these material variables.
metal. Because no two mills are exactly identical, there
Machining is a very empirically mature sub- will be differences in machinability of a steel
ject. The recommended feed rate, depth of cut, grade provided by different mill suppliers. How-
tool material, and cutting fluid for a given mate- ever, having an understanding of how these vari-
rial/material condition (thermomechanical his- ables contribute to machinability is invaluable.
tory) can be found in readily available pub- Armed with an understanding of the material and
lished tables. Books such as the ASM how it is made, one can determine the tooling,
Handbooks; Machinery’s Handbook, published coolant, and setup of the machining job.
by Industrial Press; Marks’ Standard Hand- Let us take a closer look at these variables.
book for Mechanical Engineers, published by
McGraw-Hill Book Company; or the Machin-
ing Data Handbook, 3rd edition, by the
Machinability Data center at the Institute of Ad- The role of chemistry is to define not only
vanced Manufacturing Services (IAMS; for- the different grades of stainless steel (ferritic,
merly known as Metcut Research Associates martensitic, etc.), but also how the grade
Inc.) in Cincinnati, OH; include much of the is chemically balanced within the specific
data used in industry today. Material manufac- grade; for example, the amount of carbon in a
turers are also a source of valuable machining martensitic stainless can change tool wear char-
data. A typical guide from ASM is shown as acteristics, or a change in nickel content within
Table 1 (Ref 1). specification limits can alter the stringiness of a
Rather than simply reproducing data, the chip. Combined, both will be the basis the mate-
focus of this chapter is the metallurgical factors rial’s machinability.
governing the machinability of stainless steels. Each of the elements used to produce stain-
Most of the information regards machining less steels will contribute some general machin-
stainless bar products; however, many of the ing attributes. The effects of the elements as de-
concepts could be applied to forgings as well as scribed next are general, and slight deviations
castings. may be encountered depending on the stainless
grade. However, for the more common stainless
grades used today, these effects of these alloy-
ing elements are fairly accurate.
Physical and Mechanical Properties Iron is the base element in a stainless steel. It
is a soft, gummy material that has high work-
The machinability of stainless steels is very hardening characteristics. Iron is characterized
difficult to characterize in definitive terms be- by surface finishes that are difficult to obtain
cause of the broad nature of these materials. A and chips that are stringy, and it has a high ten-
ferritic stainless steel, such as type 430,