ASM Handbook, Volume 1: Properties and Selection: Irons, Steels, and High-Performance Alloys Copyright © 1990 ASM International®

ASM Handbook Committee, p 689-736 All rights reserved.

DOI: 10.31399/asm.hb.v01.a0001039 www.asminternational.org

Embrittlement of Steels
G e o r g e F. V a n d e r V o o r t , C a r p e n t e r T e c h n o l o g y Corporation

IRON-BASE ALLOYS are susceptible selection of scrap for electric furnace melt- and testing programs must be properly
to a number of embrittlement phenomena. ing is based on the scrap impurity level, and planned and executed.
Some of these affect a wide range of com- every effort is made to keep residual levels Service and environmental conditions
positions, while others are specific to a of phosphorus, antimony, tin, and arsenic can also cause a number of embrittlement
rather narrow range of compositions. These as low as possible. Subsequent impact test- problems. Engineers working with steel
problems promote brittle service failures ing of coupons from the forgings is used to components that are susceptible to such
that may be catastrophic in nature or may verify the initial toughness. Such informa- operating-induced problems must be aware
reduce the useful service life of a compo- tion is compared to specification require- of the potential problems and must establish
nent. This article reviews these embrittle- ments and historical data base information regular inspection programs to detect prob-
ment problems and presents examples of to maintain quality. Special melting practic- lems before they become critical. An excel-
their influence on mechanical properties. es, such as vacuum carbon deoxidation, lent example of such programs is the on-
These problems arise from compositional, have also been adopted for critical compo- site, in situ examination of steam piping in
processing, or service conditions, or com- sition control. Furthermore, additional electric power generation plants; these ex-
binations of these three conditions. knowledge is obtained by surveillance test- aminations make extensive use of field met-
If the embrittlement occurs during pro- ing and by postmortems on retired forgings. allography and replication to assess creep
cessing at the mill, it may be detected However, not all potential problems can damage and predict remnant life.
during routine testing depending on the de- be detected, or prevented, at the mill. Some
gree of embrittlement and the nature of the arise from handling or fabrication problems.
testing program. The steelmaker is general-
Embrittlement of Iron
For example, low-carbon sheet steels that
ly aware of the potential problems that are not aluminum killed are roller leveled at Before discussing the embrittlement of
particular grades are susceptible to and will the mill prior to shipment to suppress the steels, this article will first examine the
check the various well-known parameters yield point and prevent strain-age embrittle- embrittlement of iron because the number
that can promote such problems. This starts ment. However, if this steel is not formed of such studies are few compared to those
with an examination of chemical composi- within a certain time period, the yield point for steels. Grain-boundary segregation of
tion, such as gas contents or impurity ele- will return, and discontinuous yielding may elements such as oxygen, sulfur, phospho-
ments that are known to cause problems. occur when the sheet is cold formed, result- rus, selenium, and tellurium is known to
For example, in the melting of ingots for ing in cosmetically damaging L~iders bands produce intergranular brittle fractures in
pressure vessels or rotors where temper on the product. For this and other fabrica- iron at low temperatures. Studies of the
embrittlement is a potential problem, the tion problems, postfabrication inspection effects of such impurities in pure iron have
been greatly aided by the development of
Auger electron spectroscopy (AES).
Temperature, °F Intercrystalline fractures in iron at low
0 200 400 600 800 1000 1200 temperatures occur when the carbon con-
300 ~ 1~ ~ T r~ tent is low. It has been assumed that the
absence of carbon is more important than
I I I I I O*yg°.oonten,,°,o - 200 the presence of embrittling grain-boundary
I Yr ~ I i 1 0 0.001
--, 225 ~ ~ _ Q 0.0037 __ impurities. However, impurities must be
I Ill .Z I I Ao.oo4o
• 0.007 - 150 .: present, and the role of carbon appears to
be one of a competitor for grain-boundary
I III II ~ I . I n 0.011
I Ill II I I I I ~_ • 0.057 sites when such impurities are present.
Oxygen. A study of the toughness of
.E_ ~I00 g iron-oxygen alloys found that intergranular
~_ o , .E_ embrittlement and a rise in impact transition
temperature in iron were produced by oxy-
75 ~0 0 gen levels of 30 ppm or more (Ref l). Figure
1 shows the Charpy V-notch impact transi-
tion curves from this study for a series of
0 0 iron-oxygen alloys with increasing oxygen
- I O0 0 I O0 200 300 400 500 600 700 contents (from l0 to 2700 ppm). However,
Temperature, °C the sulfur levels of the heats were rather
high, from 30 to 54 ppm for the alloys shown
Fig. I Charpy V-notch impact energy curves for iron-oxygen alloys with varying oxygen content. Source: Ref I in Fig. l, except for the 0.011% O heat,

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 690 / Service Characteristics of Carbon and Low-Alloy Steels

350 300

201
? 4O
• Electrolytic iron o~ e~

300 © 0 Electrolytic iron 600 o 0 High-purity iron - 500 _ Oxygen content, p p m ~ 80

~ 0 • Purified iron
250
Z 0 2000
250 500 eo • 400 E
- 40 o~
- 400 200
E 200 Q • ._g • 125 -- g
":- 200 ~0 400 ,--:. [] 30 "~:
g |
~ © ~ .~_
.~_ • 1 ~-0

150 300
150 300 .~- 0) -20 ~ L o)

.b m
a~ £ -40 -40 6
10O ~ .~ 100
- 2oo .g
g g
E3 -60 " ~ ~ a
=~ 50 0 10 30 50 70 90 110
• 100 m
cm •~ 50
0 -loo ~ Carbon content, ppm
(3 CI
0 '~I 0~"! "~ ''~ 0
0 v io, f- 0 Ductile-to-brittle transition temperatures of
10 100 103 104 Fig. 4 high-purity iron as a function of carbon con-
-50 ~ tent and oxygen content. Source: Ref 6
Oxygen content, ppm
-100
-100 Ductile-to-brittle transition temperatures
0 20 40 60 80 100 Fig. ] (from tests using Charpy U-notch specimens)
Carbon content, ppm as a function of oxygen content for a decarburized tent with the very low bulk carbon content.
electrolytic iron and a high-purity iron with 10 ppm C. Another specimen with 235 at. ppm O ex-
"qSl::|~" 2 Transition temperature, Tt, versus carbon
content for two different high-purity irons,
Source: Ref 6 hibited about 15% cleavage fracture, along
each containing 2000 ppm O. Source: Ref 4 with areas that were predominantly inter-
granular. The study found that for the ma-
on the brittleness of iron (Ref 6). The terial with higher oxygen content, quench-
which had 16 ppm S. Because this work was Charpy U-notch transition temperatures of ing specimens from temperatures of 1103
conducted long before the introduction of electrolytic iron and high-purity iron with and 1123 °C (2017 and 2053 °F), near the
AES, the true nature of the embrittling varying oxygen and carbon contents were solubility limit, gave predominantly inter-
grain-boundary element(s) is open to ques- determined, as in Ref4. Figure 3 shows that granular low-temperature fractures. When
tion. the DBTTs for these two irons were con- the temperature was lowered from 973 to
Work on high-purity iron found intergran- stant over a wide range of oxygen contents 773 °C (1783 to 1423 °F), below the solubil-
ular brittleness at low test temperatures (1 to 2000 ppm). The DBTT of the electro- ity limit, the amount of intergranular frac-
when specimens were quenched from tem- lytic iron (210 °C, or 410 °F) was consis- ture decreased. Also, if the specimen was
peratures where carbon was in solution (Ref tently much higher than that of the high- slow cooled from temperatures near the
2, 3). It was believed that this brittleness purity iron (20 °C, or 70 °F). Also, fractures solubility limit, low-temperature fractures
was due to oxygen. Lowering the quench for the electrolytic iron specimens in the were fully by cleavage. Therefore, the study
temperature reduced the embrittlement. Ti- brittle range were fully intergranular, concluded that intergranular brittle fracture
tanium additions were not found to be help- whereas those for the high-purity iron were of iron was due to the oxygen in solution.
ful in reducing the intergranular embrittle- by cleavage. Carbon and sulfur contents Sulfur. The embrittlement of iron by sul-
ment. were 30 and 35 ppm, respectively, for the fur has been demonstrated by several au-
Reference 4 reviews the study from Ref 3 electrolytic iron and 10 and less than 3 ppm, thors. Prior to Auger analysis, such results
and reports on similar work performed in respectively, for the high-purity iron. Man- were assumed, but without direct proof.
France that found that sulfur, not oxygen, ganese contents were 12 ppm for the elec- Auger work has, indeed, confirmed that
was the cause of the embrittlement. Tests trolytic iron and less than 0.5 ppm for the sulfur is a potent embrittler of iron, even at
were done using electrolytic iron containing high-purity iron, too little to tie up the sulfur bulk levels as low as 10 ppm. Furthermore,
35 ppm S and a purified iron containing less completely in the electrolytic iron. Figure 4 the displacement of sulfur on the grain
than 1 ppm S; each contained 2000 ppm O shows the results when carbon and oxygen boundaries when carbon is added has been
and varying carbon contents. The purified were both varied. When the carbon level proved by Auger analysis. Low-tempera-
iron was found to be much less brittle than was increased to 30 ppm, there was a large ture impact tests performed on three heats
the electrolytic iron (Fig. 2). At 10 ppm C, improvement in the DBTT, irrespective of of relatively pure iron obtained intergranu-
the purified iron was free of intergranular the oxygen level. This study also demon- lar fractures, depending on the heat treat-
fracture; the electrolytic iron fractured in- strated that the addition of elements that ment used (Ref 8). Sulfur contents ranged
tergranularly below the ductile-to-brittle form sulfide precipitates improved the from 14 to 100 ppm, carbon contents from
transition temperature (DBTT). DBTT. less than 10 to 70 ppm, and oxygen contents
Reference 5, on the other hand, shows On the other hand, a study using irons from 8 to 420 ppm. Auger analysis revealed
Auger analysis of intergranular fractures of with less than 2 ppm S and 0.5 ppm C and heavy segregation of sulfur to the grain
the pure iron specimen from Ref 2 that Auger analysis demonstrated oxygen grain- boundaries. No clear evidence of oxygen on
contained 60 ppm S. Sulfur was not detect- boundary enrichment of intergranular frac- the intergranular fractures was obtained.
ed on the intergranular fracture, while car- tures broken by impact within the AES Other studies have shown the detrimental
bon, nitrogen, and oxygen were. The au- chamber (Ref 7). Segregation of elements influence of small sulfur additions (Ref 9,
thors did state that the oxygen peak could such as sulfur, phosphorus, and nitrogen 10). Figure 5, from these studies, shows the
possibly be due to oxygen contamination was not observed on the fracture, but oxy- DBTT as a function of bulk sulfur content
after fracture in the high vacuum used for gen was detected. For a specimen with a (-<60 ppm) and the beneficial influence of
Auger work (fracture made inside the evac- bulk oxygen content of 430 at. ppm, there carbon additions (~10 to 30 ppm) for iron
uated chamber). were two monolayers of oxygen at the in- containing 2000 ppm O. F o r less than or
It has been demonstrated that large vari- tergranular fracture surface. Carbon was equal to 10 ppm C, the transition tempera-
ations in oxygen content have no influence not present at the fracture surface, consis- ture increased from 0 to over 600 °C (30 to

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 691

700 600 from the austenite region, the fracture sur-

face exhibited a layer of nearly pure phos-
/

/r
--1000
o 500
u_ phorus at the grain boundary with a thick-
g
o
ness of I to 1.5 nm (10 to 15 A). The ternary
- 800 alloys containing sulfur exhibited DBTTs of
500 ol - lOOO~ ~ 400 about 350 °C (660 °F). The study concluded
E E that sulfur, even at much lower concentra-
-~LI_Q~/ - 800 E --600 tions than phosphorus, is a more potent
,oo g 300 c
o
embrittler of iron.
Metalloids such as phosphorus, arsenic,
- 400
=
'~ 300
- 600 i~ 200 antimony, and tin do not produce embrittle-
ment of pure iron containing minor amounts
6 100 ellurium / 200 of carbon in the same manner as sulfur,
•~ 200 -- 400 although they do in alloy steels (Ref 16). It
6 has been demonstrated that such elements
o 0 - 0 produce embrittlement of carbide/ferrite
and surrounding ferrite/ferrite interfaces
(Ref 17). This appears to be a nonequilib-
o~@ /Z~99~ G
-o
o -100
0 20 40 60
Impurity content, ppm
80 rium segregation problem, however.
The influence of tin on high-purity iron
Influence of sulfur, tellurium, and selenium
and low-carbon steel has been examined
-I00
=15" 6 on the transition temperature of purified iron
r;n
0 20 40 60 80 (Ref 18). Detailed chemical analyses of the
Sulfur, ppm containing up to 10 ppm carbon and approximately pure iron specimens used were not given in
2000 ppm O. Source: Ref 10 the study, although it was stated that the
Fig. 5
Influence of sulfur on the transition temper- base metal had a carbon content of 20 ppm
ature of purified iron containing 2000 ppm O,
and the influence of carbon content on sulfur embrit- Researchers have also studied phospho- and an oxygen content of 400 ppm. The
tlement. Increasing carbon content has the beneficial rus segregation in pure iron (Ref 14). Again, addition of 0.5% Sn to the pure iron reduced
effect of decreasing sulfur embrittlement. Source: Ref the specimens contained a significant the impact strength in the ductile region to
9, 10 amount of sulfur, but mechanical properties such a degree that the absorbed energy was
were not determined. However, a small constant up to 70 °C (160 °F). Specimens
>1110 °F) with increasing sulfur content. amount of manganese was present that water quenched from 650 °C (1200 °F) ex-
However, with 25 or 30 ppm C, the influ- should precipitate the sulfur as a sulfide. hibited impact results similar to those of
ence of sulfur was small. Also demonstrated The carbon content was reduced to below tin-free pure iron, while slowly cooled spec-
was the scavenging influence of a 0.5% AI 10 ppm. Specimens were austenitized, wa- imens were embrittled. The addition of
addition, which suppressed embrittlement ter quenched, tempered at 850 °C (1560 °F) 0.15% C to the Fe-0.5Sn alloy did reduce
(constant DBTT) for the full range of sulfur for 1 h, and then furnace cooled to the aging the embrittling influence of tin, and the
(-<60 ppm) tested. temperature. Specimens were fractured alloy had better toughness than the pure
Selenium and tellurium are similar to sul- within the Auger chamber. Phosphorus was iron specimen when water quenched from
fur and are potential embrittlers of pure observed on the surface of intergranular 650 °C (1200 °F). The addition of 0.15% P to
iron. One study (Ref 10) showed that they fractures, but not on cleavage fractures. the Fe-0.5Sn-0.15C alloy raised the DBTT
do cause embrittlement, but to a lesser Auger analysis showed that the amount of about 20 °C (36 °F) and lowered the upper-
degree than sulfur (Fig. 6). In this work, the phosphorus on the intergranular fractures shelf energy when water quenched from 650
carbon content was less than 10 ppm, and increased with bulk phosphorus content. °C (1200 °F). The examination of fractured
2000 ppm O was present. Although seleni- Also, as the aging temperature decreased, the specimens showed a change from transgran-
um and tellurium had less influence on the grain-boundary phosphorus content in- ular cleavage to intergranular fracture as the
DBTT, these two elements reduced the creased, and the fracture became more inter- tin content increased, particularly for the
absorbed energy values much more than did granular. When carbon was added (-<80 slowly cooled specimens.
sulfur. Therefore, in terms of impact ener- ppm), the grain-boundary phosphorus con- This survey of the influence of impurities
gy, the elements can be placed in order of centration decreased. A deep-drawing steel on the embrittlement of pure iron has demon-
increasing influence as follows: sulfur, sele- containing 7 ppm C, 310 ppm P, and 0.36% strated that the design of experiments and the
nium, and tellurium. Others have also re- Mn fractured intergranularly in the drawing interpretation of results are difficult. Many of
ported the embrittlement of Fe-0.04% C by direction, and phosphorus was detected on the early studies did not recognize the signif-
0.02% Te (Ref 11). the grain boundaries (Ref 14). Similar steels icance of relatively minor amounts of sulfur in
Other Impurity Elements.Phosphorus has with 14 ppm C, 80 ppm P, and 0.38% Mn did the high-purity irons used. It is clear that
also been reported to embrittle pure iron (Ref not fracture during deep drawing. Auger analysis is required to determine the
12, 13). However, both of these studies used A study of the embrittlement of iron by embrittling species. When sulfur is present in
materials with a significant sulfur content, phosphorus, phosphorus and sulfur, and the absence of sulfide-forming elements, it
and they were performed prior to the devel- antimony and sulfur demonstrated that the has a dominating influence on properties and
opment of Auger analysis. However, radioac- embrittlement was different from that of behavior. The addition of carbon above about
tive tracer analysis demonstrated the segrega- temper embrittlement in that it was not 10 ppm will reduce or eliminate embrittlement
tion of phosphorus at the grain boundaries of reversible (Ref 15). Segregation occurred at effects.
an Fe-0.09% P alloy (Ref 12). Phosphorus was all temperatures in ferrite but was negligible
reported to be 50 times as prevalent at the or limited in austenite. Quenching from the Embrittlement in Carbon
grain boundaries as in the grain interiors. austenite region produced specimens that Steels and Alloy Steels
Phosphorus does substantially increase the fractured by cleavage. When quenched
strength of ferrite. Again, the addition of from the two-phase region, fractures did Several forms of embrittlement can occur
carbon was shown to reduce the influence of exhibit phosphorus at the grain boundaries. during thermal treatment or elevated-tem-
phosphorus on embrittlement. When an Fe-0.2P alloy was furnace cooled perature service of carbon and alloy steels.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 692 / Service Characteristics of Carbon and Low-Alloy Steels

-v 140
g
150 E
/ x c 0.005
"7 ~ 120
Hardness
140
\ 1.6 E
1.5 lOO ~ 0.004 ~
\ ~.4
m { \ Particle size [ I
> 130 dness~ 1.3 ~. ~ 80
~ 120
-o
Har
as-quenchld/ \ 1.2 --~
._o
1.1 ~
o_
~ 6o
\
110 1.0 ~
0.9 ~ 40
IO0 0.8
~ _ _ ~ 0 001 o_
0.7 ~r 2O

1 10 100 103 1 10 100 103

Aging time at 60 °C (140 °F), h Aging time at 60 °C (140 °F), h

Influence of aging time at 60 °C (140 °F) after quenching from 500 °C (930 °F) on the hardness, particle size, and interparticle spacing for an Fe-O.02N alloy. Source:
Fig. 7 Ref 25

These forms of embrittlement (and the not produce quench-age embrittlement. • Body-centered cubic (bcc) Fe~6N2 ni-
types of steel that some forms specifically Carbon steels with a carbon content of 0.04 trides with a circular disk shape precip-
affect) are: to 0.12% appear to be most susceptible to itate at low temperature on dislocations
this problem.
• Blue brittleness (carbon steels) Quench aging is caused by the precipita-
• Quench-age embrittlement (low-carbon This latter type of nitride causes strength-
tion of carbide and/or nitride from solid
steels) ening during quench aging in iron-nitrogen
solution. One study has reviewed the
qi Strain-age embrittlement (low-carbon alloys (Ref 25). Iron-carbon alloys also have
quench aging of iron-nitrogen, iron-carbon,
steels) two types of carbides that can form during
and iron-carbon-nitrogen alloys (Ref 25).
• Aluminum nitride embrittlement (carbon quench aging. Cementite forms at high tem-
Figure 7 shows the change in hardness,
and alloy steels) peratures, and a low-temperature carbide
interparticle spacing, and particle size of
• Graphitization (carbon and alloy steels) that is identical in morphology and habit to
precipitates in an Fe-0.02N alloy quenched
from 500 °C (930 °F) and aged at 60 °C (140
Blue Brittleness °F) up to 500 h. The precipitates grew from
Carbon steels generally exhibit an in- about 30 nm (300 A) in diameter after 2 h to
crease in strength and a reduction of ductil- about 80 nm (800 A) in diameter after 500 h. 700 [ - ~ ~ 1 0 0
ity and toughness at temperatures around With aging, the hardness increased rapidly 90 ~
300 °C (570 °F). Because such temperatures to about 150 HV after 10-h aging at 60 °C
produce a bluish temper color on the sur- 80
(140 °F) and then decreased to about 120
face of the specimen, this problem has been HV with aging to 500 h. 500 70
called blue brittleness (Ref 19-22). It is Low-Carbon Steels. For low-carbon
generally believed that blue brittleness is an steels, quench aging is due mainly to carbon
accelerated form of strain-age embrittle- because the nitrogen level is usually too low
ment. Deformation in the blue-heat range to have a substantial effect. Results for a } 3oo 4o ~
followed by testing at room temperature rimming steel containing 0.03% C (Fig. 8)
produces an increase in strength that is show the increase in hardness with aging l 1 10 100 10
greater than when the deformation is per- time at 60 °C (140 °F) for specimens Aging time at60°C (140 °F),h
formed at ambient temperature. Blue brit- quenched from 725 °C (1335 °F). The tensile As-quenched
(a)
tleness can be eliminated if elements that tie strength increased rapidly to a maximum
up nitrogen are added to the steel, for value after 10 h at 60 °C (140 °F) and then o~
example, aluminum or titanium. decreased slowly with further aging. The 8opt, , , I f r i 1
yield strength also increased rapidly and ~___EIongation ___ ~_ Reduction _ _
Quench-Age Embrittlement reached a maximum in about the same / ! ,narea
If a carbon steel is heated to a tempera- time but remained constant with continued :24
ture slightly below its lower critical temper- aging. The elongation and reduction of .-- 50 ' 20 c_
ature and then quenched, it will become
harder and stronger but less ductile (Ref
area decreased as the strength increased;
they reached a minimum after 10 h and
N 4o 16 }
m 30 ~ 12 "~
23-28). This problem has been called then increased somewhat with continued 1 10 100 103 o
quench aging or quench-age embrittlement. aging. Aging time at 60 °C (140 °F), h
The degree of embrittlement is a function of Iron-Nitrogen and Iron-Carbon Alloys. As-c Jenched
time at the aging temperature. Aging at For iron-nitrogen alloys, two types of ni- (b)
room temperature requires several weeks to trides can precipitate during quench aging:
reach maximum embrittlement. Lowering Influence of aging time at 60 °C (140 °F) after
F i g . 8 quenching from 725 °C (1335 °F) on the tensile
the quenching temperature reduces the de- • Face-centered cubic (fcc) Fe4N platelets properties of an Fe-0.03C rimming alloy. (a) Tensile
gree of embrittlement. Quenching from tem- form at high temperatures and generally and yield strength. (b) Elongation and reduction in
peratures below about 560 °C (1040 °F) does are found at grain boundaries area. Source: Ref 25

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 693

about 4 to 6% reduction. However, if the

material is not formed within the safe
period, discontinuous yielding will eventu-
ally return and impair formability.
Results of one study illustrate the influ-
ence of strain aging on mechanical proper-
ties (Ref 32). Three steels made by different
processes were evaluated: Steel A, silicon
and aluminum deoxidized steel; Steel B,
capped open hearth steel; and Steel C,
capped Bessemer steel. Steel C had the
highest nitrogen content. Steels B and C
had low aluminum contents, while steel A
had sufficient aluminum to tie up the nitro-
gen. Strips of each were normalized and
loaded in tension to a permanent strain of
10%. The strips were held at 25, 230, 480,
and 650 °C (75, 450, 900, and 1200 °F) for
various lengths of time (-<25 000 h at 25 and
230 °C, or 75 and 450 °F; -<10 h at 480 °C, or
900 °F; and 2 h at 650 °C, or 1200 °F).
Hardness, tensile properties, and impact
properties (half-width Charpy V-notch
specimens) were determined at different
aging times.
Figure 10 shows the impact test results for
steels A, B, and C strained 10% in tension and
aged at room temperature up to 25 000 h. The
impact curves are shifted with aging at room
temperature for all three steels; steel A exhib-
its the best aged toughness, and steel C the
poorest. Figure 11 shows the increase in
hardness for steels A, B, and C aged for times
up to 25 000 h at 25 °C (75 °F) and 230 °C (450
°F). Room-temperature aging produced a
gradual increase in hardness with time. The
maximum hardness was about the same and
was reached quickest by steel C and slowest
Fig. 9 Example of stretcher-strain marks (LCiders bands) on a cold-formed steel part by steel A. The hardness increase with aging
at 230 °C (450 °F) was constant for steel A and
slowly decreased with aging for steels B and
FeI6N 2 and may be isomorphous with it can problem, the part can be annealed before C.
also form. With aging, the low-temperature galvanizing. Alternatively, the use of sheet In low-carbon steels, strain aging is
carbide will gradually be replaced by Fe3C steels containing elements that tie up nitro- caused chiefly by the presence of interstitial
(Ref 25). The phase changes during aging of gen, for example, aluminum, titanium, zir- solutes (carbon and nitrogen), although hy-
iron-nitrogen and iron-carbide quench aged conium, vanadium, or boron, will prevent drogen is known to produce a lesser effect
steels are discussed in the literature (Ref strain-age embrittlement. at low temperatures. These interstitial sol-
28). Strain aging may also lead to stretcher- utes have high diffusion coefficients in iron
strain marks (Liiders bands) on cold- and high interaction energies with disloca-
Strain-Age Embrittlement formed low-carbon sheet steel compo- tions. The change in mechanical properties
Strain aging occurs in low-carbon steels nents. These marks are cosmetic defects, of low-carbon rimming steels with different
deformed certain amounts and then aged, rather than cracks, but their presence on carbon and nitrogen contents that were pre-
which produces an increase in strength and formed parts is unacceptable (Fig. 9). Dur- strained 4% and aged various times at 60 °C
hardness but a loss in ductility (Ref 25-27, ing tensile loading, sheet steel that is sus- (140 °F) has been demonstrated (Fig. 12)
29-34). The degree of deformation, or cold ceptible to this defect will exhibit nonuni- (Ref 34). This work clearly demonstrates
work, is important. Generally, about a 15% form yielding followed by uniform the detrimental influence of higher carbon
reduction in thickness provides the maxi- elongation. The elongation at maximum and nitrogen contents on strain aging. The
mum effect. The resulting brittleness varies load and the total elongation are reduced, solubilities of carbon and nitrogen in iron
with the aging temperature and time. Aging decreasing cold formability. In a nonalu- are quite different. Nitrogen solubility is
at room temperature is very slow, requiring minum-killed sheet steel, a small amount high in the temperature range where rapid
several months to obtain maximum embrit- of deformation, typically 1% reduction, precipitation occurs; the solubility of car-
tlement. As the aging temperature is in- will suppress the yield point for several bon, in equilibrium with cementite, is very
creased, the time for maximum embrittle- months. This process is referred to as low. Therefore, strain aging that is due to
ment decreases. Certain coating treatments, roller levelling or temper rolling (Ref 31). carbon at temperatures below 100 °C (210
such as hot dip galvanizing, can produce a This process is more effective in eliminat- °F) is insignificant. However, above 100 °C
high degree of embrittlement in areas that ing the sharp yield point and preventing (210 °F), • carbide can redissolve and pro-
were cold worked the critical amount; this strain aging than stretching the steel duce substantial strain aging (Ref 35). Strain
can lead to brittle fractures. To prevent this through the L~iders strain, which requires aging attributable to nitrogen is caused by

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 694 / Service Characteristics of Carbon and Low-Alloy Steels

Test temperature, °F Test temperature, °F curs in the temperature range of 100 to 300
-100 -50 0 50 100 150 200 -100 -50 0 50 100 150 200 °C (210 to 570 °F) (which includes tempera-
70 [I I I I I ~- 50 70 tures at which blue brittleness occurs). If
II I 3"1 y I~'~£:~/ A ~-50
60 Steel A ~ 60 - Steel B the interstitial solute content is substantial,
o 50
--40 ~ ~ dynamic strain aging may be observed at
~ 5.= ~ 50 5-~" E room temperature. At very high strain
= = 4o - 3 0 ~_~ ~_~ I~[ --30 > ' ~
- N°rm~'ize/d~
2 ~ /~ ' Z ~ l " - 40 fff E
rates, as in impact testing, dynamic strain
O
~~ 30 --20 ~
O ¢c = ' ~ 30
O~5
Normaliz.ed\ ~ o
--20 o ' ~
aging is observed at temperatures above 400
°C (750 °F), up to about 670 °C (1240 °F).
N~ N N t~
._~ 20 ,~ ~_ ~.~_ 20 Reference 33 presents dynamic strain aging
=
cO 10 10 fro- 10 ~.E_ results for five carbon steels. Carbon and
10
0 0 0 0
nitrogen are, again, the most important ele-
-75 -50 -25 0 25 50 75 100 -75 -50 -25 0 25 50 75 100 ments in dynamic strain aging. Nitrogen is
Test temperature, °C Test temperature, °C more important than carbon because of the
lower solubility of carbon.
(a) (b)

Test temperature, °F Aluminum Nitride Embrittlement

-100 -50 0 50 100 150 200 300 350 It is well known that aluminum nitride
70 j] I I J I 50 precipitation in aluminum-killed steels can
60 Steel C ..__--- --() Symbol A g i n g interval, h cause embrittlement and fracture. Several
o~ 50
© Normalized types of problems due to aluminum nitride
• 5 precipitation have been observed: inter-
100
c 40 ~- Normalized
o. ~ Q-CJ~ • 1000 granular fractures in castings (Ref 36-43),
O'5 30
[] 10000 panel cracking in ingots (Ref 44-47), and
/ • 25000 reduced hot ductility (Ref 48-53).
2O Intergranular fracture in castings, in both
- to the as-cast and heat-treated conditions,
~ have been sporadically observed for many
o
-50 -25 0 25 50 75 100 125 150 175 years. The fractures occur at the primary
Test temperature, °C austenite grain boundaries formed during
solidification. In as-cast specimens, ferrite
(e) films are generally observed at these grain
Influence of straining in tension and aging at 24 °C (75 °F) on the Charpy V-notch (half width) impact boundaries. The incidence of such cracking
F i g . 1 0 strength for three steels. (a) Steel A, silicon and aluminum killed, 0.25% C with 0.013% AI and 0.011% has been shown to increase with increases
N. (b) Steel B, capped open hearth steel, 0.07% C with 0.005% AI and 0.005% N. (c) Steel C, capped Bessemer in aluminum and nitrogen contents and with
steel, 0.08% C with 0.006% AI and 0.016% N. All three steels were strained 10% and aged. Source: Ref 32
slower cooling rates after casting.
It has been claimed that additions of
Aging time, h titanium, zirconium, boron, sulfur, molyb-
0.5 1 2 3 5 10 50 100 103 104 denum, or copper decrease the tendency for
30 f I I I I 0 © cracking (Ref 36). The cooling rate between
0 C, J 1150 and 700 °C (2100 and 1290 °F) is
20 C , O 230oC [ -- [ 1
! L ~ ..o..._ . l r . ~ D.-
important in controlling the amount of alu-
10 minum nitride precipitation. The minimum
¢ Steel A amount of aluminum nitride required to
> 0
"1- produce intergranular fracture is lower for
30
I I I - rl II z I I . . I A r
alloy steels than for plain carbon steels
c (0.002 versus 0.004%) (Ref 37). Minimizing
"~ 20 230 oc I &
the nitrogen content, using the lowest pos-
sible amount of aluminum for deoxidation,
Steel B and increasing the cooling rate after solidi-
A
fication are recommended, and it has been
c 3O J rl I r , demonstrated that cracking can be prevent-
ed by deoxidizing with titanium or zirconi-
20 •_ = . ,m
230 °C
, I m /~-
E
.
,~
m
rq
J ~'-"~. um or by combined titanium and aluminum
10 • " " ~" I ' (Ref 37).
/ 25 °C Steel C Some researchers have claimed that high-
0 i i
103 104 105 106 107 10e er levels of phosphorus and sulfur reduce
the susceptibility to aluminum nitride inter-
Aging time, s
granular fractures (Ref 38). Nitrogen con-
Increase in hardness for steels A, B, and C from Fig. 10 after straining in tension (10%) and aging at 24 tent has been shown to be more important
F i g . 11 and 230 °C (75 and 450 °F) for up to 25 000 h. Source: Ref32
than aluminum content because aluminum
is always present in amounts greater than
nitrogen that is not tied up with strong curs simultaneously with straining. In this that required to tie up all of the nitrogen
nitride formers, for example, aluminum, case, the effective strain rate, that is, the (Ref 41). Higher aluminum contents do,
titanium, zirconium, vanadium, or boron. dislocation velocity, controls the extent of however, increase the solubility tempera-
Dynamic Strain Aging. Strain aging can aging of a particular steel. For normal ten- ture of aluminum nitride and provide an
also occur dynamically, that is, aging oc- sile strain rates, dynamic strain aging oc- additional driving force for precipitation.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 695

60
I F I predominantly pearlitic matrix structures.
8 The susceptibility to panel cracking varies
0 Specimen 1
45 • Specimen 2 - with melt practice: electric arc furnace
6 steels are most prone; basic open hearth,
~. Specimen 5 r"~ / ...._..4 .~-'- ~ L ' - - - '''~-
~-" 30
basic oxygen furnace, and acid open hearth
4 ~ steels are less prone. Large ingots are more
susceptible than small ingots. Stripping of
15 Change in 2 the ingot at as hot a temperature as possible
lower yield stress (AY)
reduces susceptibility. Again, aluminum
0 I I I I 0 and nitrogen contents are very important.
2.5 Panel cracking is not observed with less
than 0.015% AI but occurs with increasing
o~
2.0 frequency with increasing aluminum con-
tent above this level. Also, for a given
1.5 aluminum content, increasing the nitrogen
co above 0.005% increases panel cracking.
..=, t.o Crack surfaces are oxidized but not decar-
© burized. Because the intergranular cracks
generally propagate along ferrite/pearlite in-

Y
0.5 I
L~ders strain
terfaces, it has been suggested that cracking
I occurs at relatively low temperatures, prob-
0 I I I
30
ably below 850 °C (1560 °F) (Ref 44).
O ~ • -4 A statistical study of panel cracking
found that the only significant variable was
~ 15 ~" [ --2 ~ the level of the aluminum addition (Ref 45).
< It was suggested that cracking began inter-
O-- 0 ~ nally and propagated to the surface because
Change in ultimate
tensile strength (AU) of cooling-induced stresses. In a study of
t i i i -- -2
-15 panel cracking in two alloy steels, one con-
5 taining 0.025% AI and 0.008% N and the
/
other having 0.004% A1 and 0.006% N, the
0 former exhibited panel cracking, and the
latter did not. Additions of aluminum and
o
-5 titanium to the crack-prone steel composi-
< z
tion resulted in the elimination of grain-
-10 boundary ferrite networks and freedom
Change in elongation
to fracture (AE) from cracking (Ref 46).
-15 i h A statistical study of panel cracking in
low-carbon plate steels found that the steels
0.06 °r
~ did not exhibit grain-boundary ferrite net-
0.04 ! / 'q
works, but rolled bloom surfaces were
heavily cracked in some cases. Some of the
observations in this study differ from those
<~ 0.02 - z~ I I
O.o- of other authors who have studied panel
Change in cracking. However, the statistical analysis
o ~ ~ --.c ~ ~ ~; work-hardening
showed that cracking increased with alumi-
coefficient (An)
num content and track time (time between
-0.02 I I I ingot stripping and charging into the soaking
1 10 100 103 104 105
pit). Cracking was minimized by holding the
Aging time at 60 °C (140 °F), min
aluminum contents to 0.030% or less (Ref
Influence of grain size and aging time at 60 °C (140 °F) on the mechanical property changes caused by 47).
Fig. 1 2 strain aging. A 0.038C-0.0042N-0.001AI steel was quenched from 200 °C (390 °F), prestrained 4%, and Reduced Hot Ductility. Numerous studies
tested after different aging times. Grain sizes, in grains/mm 2 (ASTM number), were: specimen 1, 50 (2.7);
specimen 2, 195 (4.7); specimen 5, 1850 (7.9). Source: Ref 34 have demonstrated that increasing alumi-
num and nitrogen contents degrade hot duc-
tility; this influence is most pronounced in
Aluminum nitride is known to precipitate Panel Cracking. Panel cracks are longitu- the temperature range where aluminum ni-
with one of two morphologies: plates or dinally oriented surface cracks on the side tride precipitation is greatest (Ref 48-53).
dendritic arrays (Ref 38). The dendritic face of an ingot that generally form near the One author found that 3.4% Ni raised the
form of aluminum nitride found on the in- center of the face and can extend to the solution temperature of aluminum nitride in
tergranular fracture surfaces of aluminum midradius of the ingot (Ref 44). Such cracks En 36 alloy steel by about 100 °C (210 °F)
nitride embrittled castings precipitates from can occur on more than one ingot face and compared to plain carbon steels with the
the liquid near the conclusion of solidifica- can run the entire length of the ingot. Panel same aluminum and nitrogen levels (Ref
tion (Ref 42). These aluminum nitride den- cracks are observed in aluminum-killed 48). Another study evaluated the hot ductil-
drites may be nucleation sites for platelike steel ingots, particularly those with 0.4 to ity of chromium-molybdenum-vanadium
aluminum nitride that precipitates after so- 0.7% C (plain carbon steels) or those with turbine rotor steels with 0.002 to 0.066% AI
lidification. The platelike aluminum nitride somewhat lower carbon contents for alloy and 0.007 to 0.014% N. This study showed
produces the small, shiny fracture surface steel ingots. These carbon contents lead to that nitrogen and residuals (copper, tin,
facets that are generally observed (Ref 43). ferrite grain-boundary network films with antimony, and arsenic) reduced the hot duc-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 696 / Service Characteristics of Carbon and Low-Alloy Steels

tility. The addition of titanium and/or boron Test temperature, ° F tal than manganese, and tin and antimony
improved the hot ductility in the test range 1400 1600 1800 2000 are extremely detrimental to copper-
(800 to 1000 °C, or 1470 to 1830 °F) where induced hot shortness (Ref 58). Chromium
high nitrogen contents were detrimental decreases the solubility of copper in austen-
(Ref 49). ite and increases the susceptibility to cop-
In hot ductility tests on carbon-manga- per-induced hot shortness, although its in-
nese steels containing 0.032 to 0.073% AI fluence is small (Ref 59).
and 0.0073 to 0.0105% N, aluminum nitride 80! ~ --
was found to reduce deformability and in- Graphitization
crease the resistance to the deformation of In the early 1940s, several failures of
austenite. These trends were enhanced as c 70 welded joints in high-pressure steam lines
the aluminum nitride particle size decreased occurred because of graphite formation in
(Ref 50). Other hot ductility tests on low- ~ 60
the region of the weld heat-affected zone
carbon steels also showed a large decrease that had been heated during welding to the

)
in hot ductility because of aluminum nitride critical temperature of the steel used (Ref
precipitation, depending on the volume 61-65). Extensive surveys of carbon and
fraction and size of the precipitates (Ref 51). o A carbon-molybdenum steel specimens re-
Five steels were tested containing different • 8 moved from various types of petroleum
levels of soluble aluminum (aluminum not 40--
• D
- ~ refining equipment revealed graphite in
tied up with oxygen) and with nitrogen o E about one-third of the 554 specimens exam-
present as aluminum nitride. Figure 13 30 ined (Ref 61, 64). In most cases, graphite
shows the reduction in area for hot tensile 700 800 900 1000 110[ formation did not occur until about 40 000 h
tests over a range of temperatures for five Test temperature,°C or more at temperatures from 455 to 595 °C
steels. Steels A and B showed a large de- (850 to l l00 °F). Aluminum-killed carbon
crease in percent reduction in area with I ~ Composition ~ 1- Particle size steels were susceptible, but silicon- killed or
increasing temperature; however, steel C Low-carbon steel " N, ppm(a) Ti, %" nm ~, low-aluminum killed steels were not. The
did not exhibit such embrittlement. The Curve A .............. 80 ... 90 900 O0.5Mo steels were more resistant to
difference between steels A, B, and C, Curve B .............. 70 ... 80 800 graphitization than carbon steels, but they
which all had high levels of nitrogen in the Curve C .............. 72 ... 210 2100 were similarly influenced by the nature of
Curve D .............. 2 ... (b) (b)
form of aluminum nitride, was the particle Curve E .............. I 0.06 (b) (b)
the deoxidation practice. Chromium addi-
size. The smaller aluminum nitride particles tions and stress relieving at 650 °C (1200 °F)
(a) As A I N . [ b ) N o data
in steels A and B were detrimental to hot both retarded graphitization.
ductility, while the larger particles in steel C An examination of the graphitization sus-
Elevated-temperature tensile test results for
did not cause embrittlement; this agrees Fig. 13 five plain carbon steels containing various ceptibility of a number of alloy steels
with the findings in Ref 50. Steel E con- amounts of aluminum nitride. The nitrogen content showed that chromium-molybdenum steels
tained 0.06% Ti and exhibited the best hot (in ppm) of the steels in the form of aluminum nitride used for steam piping, either annealed or
ductility. was: A, 80; B, 70; C, 72; D, 2; E, 1. Source: Ref 51 normalized, were resistant to graphitiza-
An evaluation of the hot ductility of low- tion. Nickel and nickel-molybdenum steels
carbon killed steels found a substantial re- did graphitize during high-temperature ex-
duction in fracture strain between 700 and layer at the surface of steels heated for posure. Chromium-bearing steels did not
800 °C (1290 and 1470 °F). This reduction forging and rolling in oxidizing atmo- graphitize and appeared to be quite stable
was most pronounced between 750 and 775 spheres. It has been shown that tin, antimo- (Ref 65).
°C (1380 and 1425 °F), where ferrite formed ny, and arsenic residuals alter the solubility The deoxidation practice used in making
along the prior-austenite grain boundaries. of copper in austenite during high-tempera- the steels is the most important parameter
The intergranular fracture surfaces exhibit- ture soaking. Because the hot-working tem- influencing graphitization. High levels of
ed aluminum nitride and manganese sulfide perature is usually above the melting point aluminum deoxidation strongly promote
precipitates (Ref 53). of elemental copper, liquid copper is pro- graphitization for both carbon and carbon-
Hot ductility is also impaired by high duced that will penetrate the austenite grain molybdenum steels used for steam lines.
residual impurity contents, chiefly those of boundaries and cause cracking by liquid While nitrogen appears to retard graphitiza-
copper and tin (Ref 54-60). Examinations of metal embrittlement. Nickel and molybde- tion, high levels of aluminum remove nitro-
forging and rolling defects have frequently num concentrate with copper and raise the gen and thus promote graphitization (Ref
revealed concentrations of elemental cop- melting point of copper; tin, antimony, and 66). Molybdenum additions (0.5%) help sta-
per, high levels of copper and tin, or copper arsenic also concentrate at the scale/metal bilize cementite but do not fully offset the
in the scale. Tin residual levels are normally interface and lower the melting point of influence of high aluminum additions. Man-
much lower than copper residual levels, but copper. If copper is not present, tin, anti- ganese and silicon both affect graphitiza-
there appears to be a synergistic effect mony, and arsenic have little detrimental tion, but their influence is small at the levels
between copper and tin that enhances em- effect on hot workability (Ref 56). used in these alloys. Chromium appears to
brittlement (Ref 54). An examination of It has been shown that tin reduces the be the best alloy addition for stabilizing
longitudinal cracks in medium- and high- solubility of copper in austenite, which is carbides.
carbon steels found that cracking occurred probably more important than its influence Tensile tests of affected steam piping
between 700 and 500 °C (1290 and 930 °F) on the melting point of copper (Ref 56). Tin indicate that the graphite present did not
and depended on high copper and tin con- reduces the solubility of copper in austenite affect the tensile strength. Charpy V-notch
tents. Aluminum was not added to these by a factor of three; therefore, when tin is impact strength, however, was reduced
steels, but grain-boundary ferrite networks present, molten copper can form at the substantially. Localized graphitization near
were present. The copper and tin were surface at lower bulk copper contents. a welded joint appears to be much more
segregated to the ferrite networks (Ref 55). Nickel reduces copper-induced hot short- damaging to pipe behavior than general,
Nickel, copper, tin, antimony, and ar- ness, manganese slightly increases hot uniform graphitization. The localized graph-
senic often become enriched in the subscale shortness, arsenic is slightly more detrimen- itization apparently produces notches that

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 697

concentrate stress and reduce load-bearing Testing temperature, °F nally applied coolant during hot working, will
capability. -330 -150 32 212 390 570 control this cooling rate.
5o I I Steels containing less than l0 ppm S do
35
950-1200 °C -

not overheat, regardless of the heating tem-

Overheating
45 / I I perature up to burning. However, this level
The overheating of steels occurs when
they are heated to excessively high temper-
atures prior to hot working (Ref 67-76).
40 // Overheated,

oC
1320 °C ~ 30
of sulfur is difficult to obtain. Steels with
relatively low sulfur contents, (for example,
in the range of 0.001 to 0.005%) are being
Heating at even higher temperatures will 35 - 25
produced in greater quantities today be-
cause incipient grain-boundary melting, a cause of the detrimental influence of sulfur
problem known as burning. Thus, overheat- 30 ] [ Overheated,1340 °C on properties. However, it has been dem-
ing occurs in the temperature range be- onstrated that the minimum temperature at
tween the safe range normally used prior to which overheating occurs in low-sulfur
25 steels is lower than that for high-sulfur
hot working and the higher range where / ~ / Overheated'1390°C - 1 5 E
liquation begins. Hot working after burning steels (>0.010% S). Overheating problems
generally results in the tearing or rupture of
the steel due to the liquid in the grain
boundaries. Hot working after overheating
/[ thus have been experienced in low-sulfur
steels heated at temperatures that usually
do not cause overheating (Ref 73-75). Addi-
generally does not result in cracking; if tions of rare earth elements have been
sufficient hot reduction occurs, the influ- shown to prevent overheating by modifying
ence of overheating may be minor or negli- the solubility of the sulfide formed. High-
gible. If the degree of hot reduction is small, sulfur steels appear to require a greater
mechanical properties, chiefly toughness degree of overheating to cause fracture
and ductility, will be affected. o o faceting and impaired properties than do
Fracture surfaces of overheated steels -200 -100 0 100 200 300 low-sulfur steels.
given limited hot reduction often exhibit a Testing temperature, °C The problem of overheating is complex.
coarse-grain faceted appearance. Such fea- Sulfide dissolution and reprecipitation at
Impact energy values versus test tempera-
HIS" 14 ture for En 111 alloy steel specimens heated the prior-austenite grain boundaries causes
E~n
tures are most evident after quenching and
tempering to develop optimum toughness. to the indicated temperature for 1 h, oil quenched, fracture faceting and impairment of proper-
Other problems, such as aluminum nitride and tempered for 1 h at 675 °C (1245 °F). Source: Ref 68 ties. However, when faceting is observed,
embrittlement, may also produce a faceted mechanical properties may not be signifi-
fracture surface. Additional tests are need- cantly affected. Overheating and its influ-
ed to distinguish among these problems (Ref this article). The use of temperatures up to ence on properties depend on the sulfur
76). 1200 °C (2190 °F) produced no change in content, the soaking temperature, grain
Although the mechanical properties of impact energies, but temperatures above size, the cooling rate through the overheat-
burnt steels are always very poor, the me- 1200 °C (2190 °F) produced a decrease in ing range, and the degree of hot reduction.
chanical properties of overheated steels upper-shelf energy. The burnt specimens Furthermore, the amount of faceting ob-
show considerable scatter. For tensile tests, displayed a substantial loss in toughness. served on the test fracture depends on the
the elongation and reduction of area are Because the pieces were not forged after the heat treatment (particularly the tempering
most affected by overheating and decrease burning treatment, they did not fracture. temperature), the test temperature, the test
with increasing heating temperature. Frac- Presence of Facets. Many of the studies of specimen orientation, and the amount of
ture faceting and substantial decreases in overheating have been conducted without deformation after sulfide reprecipitation.
tensile ductility normally are observed after subsequent hot working after overheating
severe overheating. and thus do not simulate actual commercial Thermal Embrittlement
Impact properties are usually more sensi- experience. These studies are of limited of Maraging Steels
tive to overheating than is tensile ductility. value. Numerous theories have been pro-
In examining impact test results, several posed to explain overheating. The examina- Maraging steels will fracture intergranu-
interrelated features should be examined: tion of facets on fractures of overheated lady at low impact energies if improperly
steels reveals that the facet surfaces are processed after hot working. This problem,
• Change in upper-shelf energy covered with fine ductile dimples, and small known as thermal embrittlement, occurs
• Impact strength transition temperature manganese sulfides can be found within the when maraging steels that have been heated
• Presence of facets dimples (because two fracture surfaces are above 1095 °C (2000 °F) are slowly cooled
Upper-shelf energy appears to be particu- formed, a manganese sulfide particle will be through, or held within, the temperature
larly sensitive to overheating. Figure 14 found in either half of the mating dimples range of 980 to 815 °C (1800 to 1500 °F) (Ref
shows impact energy curves for En Ill alloy after fracture). The facet surfaces corre- 77-81). The embrittlement is caused by the
steel specimens heated to a variety of tem- spond to prior-austenite grain surfaces precipitation of TiC and/or Ti(C,N) on the
peratures from the normal soaking range to formed during overheating. austenite grain boundaries during cooling
the burning range; no hot working was As the soaking temperature is raised, man- through, or holding within, the critical tem-
performed. The specimens were first heated ganese sulfide in the steel dissolves (that is, perature range. The degree of embrittle-
to 950 °C (1740 °F) for l0 min, transferred to the sulfur goes into solution in the austenite). ment increases with time within the critical
the high-temperature zone for 7 min, then Dissolved sulfur diffuses toward the austenite range. Increased levels of carbon and nitro-
transferred back to the furnace at 950 °C grain boundaries, where it reprecipitates. gen render maraging steels more susceptible
(1740 °F) and held for 50 min. They were oil Therefore, overheating changes the size and to thermal embrittlement. Auger analysis
quenched, tempered 1 h at 675 °C (1245 °F), distribution of sulfides in the steel. The cool- has shown that embrittlement begins with
and water quenched to minimize temper ing rate through the overheating range also the diffusion of titanium, carbon, and nitro-
embrittlement (more information about affects the size and dispersion of the inter- gen to the grain boundaries, and observa-
temper embrittlement is given in the section granular sulfides. In commercial practice, the tion of TiC or Ti(C,N) precipitates repre-
"Temper Embrittlement in Alloy Steels" in size of the overheated piece, and any exter- sents an advanced stage of embrittlement.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 698 / Service Characteristics of Carbon and Low-Alloy Steels

Results of a study on thermal embrittle- Plate finishing temperature, °F

short time periods in which cracking prob-
1600 1700 1800 1900 2000
ment demonstrate its influence on an 110 lems occur frequently and then stop for no
i I l ...._.~L.....-ol 100
18Ni(250) maraging steel (Ref 77). Plates 100 apparent reason. Evidence also indicates
12.7 mm (0.5 in.) thick were rolled with ~ 90 - 90 that quench cracking can be more frequent
finishing temperatures in the range of 1080 80 for certain heats of steel, again for no obvi-
to 870 °C (1980 to 1600 ° F) and then cooled £ 80 ous reason. Instances have also been docu-
-"~- 70 ~
by three different methods (water, air, and ~ 70 mented (Ref 86) in which extensive quench
vermiculite). Test results showed that the - 60 ~- cracking has occurred in material processed
finishing temperature and cooling rate from 80 --Hot rolled and aged-- " ~ - - 50 from the bottom portion of ingots.
the finishing temperature had minor effects 50 I I Quench crack fractures are always inter-
850 900 950 1000 1050 1100
on the tensile properties but a significant Plate finishing temperature, °C
granular. In quenched and tempered steels,
effect on fracture toughness. Figure 15 (al proof of quench cracking is often obtained
shows the plane-strain fracture toughness by opening a crack (if necessary) and visu-
results as a function of finishing tempera- Plate finishing temperature, °F ally looking for the temper color typical for
ture (temperature at the end of rolling) and 140
1600 1700 1800 1900 2000 the temperature used (Ref 88, 89). If the
cooling rate for hot-rolled and aged material 120 crack occurs during or after quenching but
(Fig. 15a) and for annealed and aged mate- 130 'Cooling
['1 medium from
'1 hot '1 '
rolling _ before tempering, and if the crack is open to
rial (Fig. 15b and 15c). In general, vermic- ~ 120 110 c
O Water the furnace atmosphere, a thin oxide layer
ulite cooling (slow) and a high finishing g_ 110- • Air • 100[~ will form on the surface. The color of the
LX Vermiculite
temperature produced the lowest tough- ~ 100 oxide layer depends on its thickness, which
ness, except for specimens water quenched 9O in turn depends on the tempering tempera-
after rolling (Fig. 15b and 15c). Detailed ~ 90 8O ture and the steel composition. Quench
information on maraging steels is available 80 cracks begin at the surface and propagate
in the article "Maraging Steels" in this 70
Annea, at 87a, °C and 7O
inward. They are usually oriented longitu-
Volume. 850 900 950 1000 1050 1100 dinally or radially unless located by a
Plate finishing temperature, °C change in section size, by surface imperfec-
Quench Cracking
(b) tions, or by changes in surface microstruc-
ture (such as an interface between hardened
Plate finishing temperature, °F
The production of fully martensitic mi- and nonhardened areas).
1600 1700 1800 1900 2000
crostructures in steels requires a heat treat- 125 I I ] I [ i.. 11o
ment cycle that involves quenching after
austenitization. The composition of the
[~115
105
o _ I00~.~T e m p e r Embrittlement
steel, its size, and the desired depth of in Alloy Steels
hardening dictate the required quench me- >E 95 90 -~
dium. Certain steels are known to be sus- - 80
Temper embrittlement--also known as
85 - - temper brittleness*, two-step temper em-
ceptible to cracking during or slightly after Annealed,at 815, °C a n d , a g ~ " ~ " ' ~
quenching. This is a relatively common
75 70 brittlement, or reversible temper embrittle-
problem for tool steels, particularly those
850 900 950 1000 1050 1100 ment--is a problem associated with tem-
that require liquid quenching (Ref 82). (c)
Plate finishing temperature, °C pered alloy steels that are heated within, or
Numerous factors can contribute to slowly cooled through, a critical tempera-
cracking susceptibility (Ref 82-87): i:|a Influence of mill finishing temperature and ture range, generally 300 to 600 °C (570 to
- , b . 1 5 manner of cooling on the plane-strain frac-
ture toughness (K=c) of 18Ni(250) maraging steel heat i 110 °F) for low-alloy steels. This treatment
• Carbon content treated three ways. (a) Hot rolled and aged. (b) An- causes a decrease in toughness as deter-
• Hardenability nealed at 870 °C (1600 °F) and aged. (c) Annealed at 815 mined with Charpy V-notch impact speci-
• M~ temperature (the temperature at °C (1500 °F) and aged. Source: Ref 77 mens (Ref 90-114). It is a particular problem
which martensite starts to form) for heavy-section components, such as
• Part design pressure vessels and turbine rotors, that are
ity steels. Quench crack sensitivity also
• Surface quality slowly cooled through the embrittling range
increases as the severity of the quench rate
• Furnace atmosphere after tempering and also experience service
increases. Control of the austenitizing tem-
• Heat treatment practice at temperatures within the critical range.
perature is also important, particularly for
As the carbon content is raised, the M s and high-carbon tool steels. Excessive retained Temper embrittled steels exhibit an in-
Mf (temperature at which martensite forma- austenite and coarse-grain structures pro- crease in their DBTT and a change in fracture
tion ends) temperatures decrease, and the mote quench cracking. Quench uniformity mode in the brittle test temperature range
volumetric expansion and transformation is important, particularly when liquid quen- from cleavage to intergranular. The DBTT
stresses accompanying martensite forma- chants are employed. When high-carbon can be assessed in several ways, the most
tion increase. In general, steels with less steels are quenched to form martensite, common being the temperature for 50% duc-
than 0.35% C are free of quench cracking they are in a highly stressed condition. tile and 50% brittle fracture (50% fracture
problems. Such low-carbon steels have Therefore, tempering must be done imme- appearance transition temperature, or
higher M~ and Mf temperatures, which al- diately after quenching to relieve these FATT), or the lowest temperature at which
low some stress relief to occur during the stresses and minimize the risk of cracking. the fracture is 100% ductile (100% fibrous
quench. Also, transformation stresses are Surface quality is also very important be- criterion). Transition temperatures based on
lower, and the lower strength of the martens- cause seams, laps, tool marks, stamp absorbed energy values are not normally em-
ite formed (low-carbon lath martensite) can marks, and so on, act as stress concentra- ployed. Temper embrittlement is reversible;
accommodate the strains more readily than tors to locate and enhance quench cracking that is, the toughness of embrittled steels can
can a higher-carbon steel. susceptibility (Ref 82). be restored by tempering them above the
Alloy steels with ideal critical diameters Quench cracking is a problem that often
of 4 or greater are more susceptible to defies prediction and can be difficult to *The term temper brittleness was first used to describe
quench cracking than are lower-hardenabil- diagnose. Heat treaters have experienced this problem by J.H.S. Dickenson in 1917.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 699

120
after heat treatment (870 °C, or 1600 °F, for
1 h, oil quench; 625 °C, or 1155 °F, for 1 h,
100 q
J
180 water quench) and after step-cool embrittle-
ment. Figure 18 shows the results for addi-
80 tions of antimony to plain carbon and alloy
,,<
60 f, /
-- 135 I---
steels of various analyses. The bars show
the 50% FATT after tempering at 625 °C
t'
o= O Phosphorus
._¢ / • Antimony - - 90 (1155 °F) (left end--not embrittled) and af-
4O /
Arsenic
ter step cooling (right end--embrittled); the

Y
• Tin value under the bar is the shift in 50%
20
- - 45 FATT. The addition of 600 to 800 ppm Sb to
the 0.4C-3.5Ni-I.7Cr steel caused a shift in
0 transition temperature of 695 °C (1285 °F).
0.05 0.1 0.15 0.2 0.25 The same steel, but without carbon, exhib-
(a) Impurity level, wt% ited a shift of 222 °C (432 °F), but its
hardness was much lower (-80 HRB versus
250 27 HRC). The steels with only nickel and
450
carbon or chromium and carbon and 600 to
L)
800 ppm Sb exhibited much less embrittle-
o
.<1
200
360 ou_ ment, while the plain carbon steel was not
<1 embrittled by antimony.
I--
< Figure 19 shows the results for additions of
150 J
270 ~
J about 500 ppm each of phosphorus,
LL

/ tin, and antimony to the nickel-chromium-

O j.,.....,.o o

_J
._c 100 ~// 180 carbon, nickel-carbon, and chromium-carbon
.E steels. The nonembrittled toughnesses of
#, the nickel-chromium-carbon-phosphorus and
50 I / s0 N chromium-carbon-phosphorus alloys were
poorer than those of the other alloys shown,
probably because of the segregation of phos-
0.05 0.1 0.15 0.2 0.25 phorus in austenite. Phosphorus also embrit-
(b) Impurity level, wt% tled the chromium-carbon-phosphorus alloy
Influence of phosphorus, antimony, arsenic, and tin impurity elements on the temper embrittlement much more than the nickel-carbon-
Fig. 16 susceptibility of nickel-chromium experimental steels based on the change in (a) 50% FATT and (b) phosphorus alloy. This is due to an interac-
100% fibrous FA'FI- after aging at 450 °C (840 °F) for 1000 h. Source: Ref 91 tion between chromium and phosphorus. Tin
appears to embrittle the nickel-chromium-
carbon alloy more than phosphorus in that the
critical region followed by rapid cooling, for the late 1950s (Ref 90, 91). The change in change in fracture appearance transition tem-
example, water quenching. This decreases 50% FATT and 100% fibrous FATT with perature (AFATT) was greater. However, the
the DBTT and changes the low-temperature isothermal aging at 450 °C (230 °F) for up to grain size of the nickel-chromium-carbon-
(that is, below the 50% FATT) intergranular 1000 h for nickel-chromium and nickel- phosphorus alloy was ASTM No. 8, while
brittle appearance back to the cleavage mode. chromium-molybdenum laboratory heats that of the nickel-chromium-carbon-tin alloy
Temper embrittlement does not occur in with controlled compositions and impurity was ASTM No. 6. Also, the nonembrittled
plain carbon steels, only in alloy steels. levels was determined. Figures 16 and 17 toughness of the alloy containing phosphorus
Also, the degree of embrittlement varies show some test results for the influence of was much poorer. The 50% FATT values for
with alloy steel composition. Therefore, the antimony, phosphorus, tin, and arsenic and these two compositions are nearly identical
alloying elements present, and their combi- aging at 450 °C (840 °F) for 1000 h (Ref 90, and would be even closer if the grain sizes
nations and levels, are important. However, 91). Embrittlement was greater for the nick- were the same. Otherwise, it appears that tin
certain impurity elements must be present if el-chromium steels than for the nickel-chro- embrittled the nickel-carbon alloy more than
temper embrittlement is to occur. The em- mium-molybdenum steels because of the phosphorus, while phosphorus embrittled the
brittling impurities are, in decreasing order beneficial influence of molybdenum. The chromium-carbon alloy more than tin. Ar-
of influence on a weight percent basis, nickel-chromium steels also showed sub- senic was a much weaker embrittler. The
antimony, phosphorus, tin, and arsenic. Of stantially greater embrittlement from the results in Fig. 18 and 19 clearly show that the
these elements, phosphorus is most com- manganese addition than did the nickel- combination of nickel and chromium re-
monly present in alloy steels and it has chromium-molybdenum steels. The addi- sulted in much greater embrittlement, par-
captured the most attention in research tion of about 0.7% Si had a smaller and ticularly for additions of antimony and tin.
studies. Manganese and silicon also in- similar embrittling influence for both The results for phosphorus show that a
crease the susceptibility to embrittlement. grades. strong interaction exists between chromi-
Although alloy steels are ferritic in the The important role of alloying elements um and phosphorus, while phosphorus
tempered condition, fracture below the has been clearly demonstrated in tests that causes little embrittlement in nickel steels
DBTT occurs along prior-austenite grain also used heats of controlled composition not containing chromium.
boundaries where both alloying elements (Ref 92). The tests were performed using The beneficial influence of molybdenum
and impurity elements are concentrated. 0.4% C alloy steels containing nickel, chro- on phosphorus-induced temper embrittle-
mium, or nickel and chromium, as well as a ment has been known for many years (Ref
Effect of Composition on AFATT plain carbon steel. Controlled additions of 93, 94). It has also been known that carbon
The proof that antimony, phosphorus, tin antimony, phosphorus, tin, and arsenic steels are immune to temper embrittlement,
and/or arsenic is an essential ingredient(s) were made to these compositions. The 50% but that substantial additions of manganese
for temper embrittlement was obtained in FATT was evaluated for each composition cause susceptibility to this problem. In ad-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 700 / Service Characteristics of Carbon and Low-Alloy Steels

<
60/
80

40 ~o
o Phosphorus
• Antimony
A Arsenic
• Tin
i
-- 125

-- 100

75
u_
o
-300 -100

27
I

°C
100
I
50%FATT,°F
300
I
S00
[
700
I
900 1100

u_ u_ A F.-J
f -- 50 g
20 t i B F--
695°C
._= ._c
~ - - 25
222 °C
0 C F
50 °C
-- -25 D H
-20
0.02 0.04 0.08 0.08 0.1 0.12 0.14 48 °C
(a) Impurity level,wt% E
-7 °C
140
F J-r

120 225 -200 200 400 600

o 50%FATT,°C
180
k-
< Composition, wt% I
kL 8O Steel N i ~ C r C Sb
o~ <~
00 100 i . . O ¢
135
60 g Alloy
.E o

40 ,0
90 ~ Line A .......... 3.5 1.7 0.4 • . .
Line B .......... 3.5 1.7 0.4 0.064).08
f ---- •.--. m Line C .......... 3.5 1.7 . . . >0.08
20 f Line D .......... 3.5 . . • 0.4 0.06-0.08

0 ~ J 0
Line E .......... • . • 1.7 0.4 0.06-0.08

Plain carbon
0.02 0.04 0.06 0.08 0.1 O.12 O.14
(b) Impurity level, wt% Line F .......... - . . . . . 0.4 >0.08

Influence of phosphorus, antimony, arsenic, and tin impurity elements on the temper embrittlement
I:;,I 1 Influence of alloying elements on the tem-
Fig° 1 7 susceptibility of nickel-chromium-molybdenum experimental steels based on the change in (a) 50% "'5" 8 per embrittlement of steels (compositions
FATr and (b) 100% fibrous FAn after aging at 450 °C (840 °F) for 1000 h. Source: Ref 91 given in accompanying table) containing 600 to 800
ppm Sb. The left end of bar gives the nonembrinled
DBTI-; the right end of bar gives the DBTI- after
embrittlement (except for line F, which is reversed).
dition, high levels of manganese in alloy Value between bar ends is the shift in 50% FATI-.
500 °C (885 and 930 °F). The influence of Source: Ref 92
steels have been known to render them molybdenum on the prevention of temper
more susceptible to temper embrittlement. embrittlement appears to depend on how
Reference 95 includes an evaluation of much of it is dissolved in the matrix as concentration profile from the grain bound-
the addition of 0.6% Mo, along with con- opposed to how much is tied up in carbides. ary into the grain interior was much shal-
trolled antimony, phosphorus, tin, arsenic, As more molybdenum becomes tied up in lower than for the impurity elements. Fig-
and other elements, to AISI 3340 (3.5Ni- the carbides, the beneficial influence of ure 20 shows an example of Auger analysis
1.7Cr-0.4C). The addition of molybdenum molybdenum decreases. Therefore, de- of antimony, sulfur, and phosphorus segre-
eliminated or greatly reduced embrittlement pending on the temperatures experienced gated to either fracture or free surfaces (Ref
due to step cooling for additions of antimo- and the presence of other strong carbide 99, 100). These results were obtained by
ny, tin, and arsenic, but not for additions of formers, molybdenum may or may not be alternate argon ion sputtering (depth profil-
phosphorus. The addition of 0.7% Mn to able to suppress temper embrittlement. ing) and analysis.
this steel produced substantial embrittle- A major advancement in understanding The degree of embrittlement also depends
ment, which was largely eliminated when temper embrittlement was brought about by upon the time at temperature within the
0.6% Mo was added. In this work, it was the development of Auger electron spec- critical region. Extensive isothermal embrit-
shown that phosphorus segregates to the troscopy and its application to embrittle- tlement studies were performed on a heat of
austenite grain boundaries during austeniti- ment studies in 1969 (Ref 97-99). This work AISI/SAE 3140 alloy steel (0.39C-0.79Mn-
zation; antimony does not do this. This permitted the direct chemical analysis of 0.30Si-0.028S-0.015P-1.26Ni-0.77Cr-0.02Mo,
work also clearly showed that manganese is segregants on the intergranular fracture sur- ASTM No. 8 grain size) (Ref 101). Specimens
an embrittling element, not merely an en- faces of embrittled specimens. Such work were austenitized at 900 °C (1650 °F) for I h,
hancer of embrittlement. has shown that the embrittling impurity water quenched, and tempered at 675 °C
Later work was conducted to clarify elements are segregated to within the first (1245 °F) for I h, and water quenched. This
these results (Ref 96). The earlier work had few monolayers of the embrittled grain resulted in a hardness of 23 HRC and a
employed specimens with a very coarse boundaries. The degree of enrichment of transition temperature (100% fibrous criteri-
grain size (Ref 95). Results with somewhat these elements may be 100 to 10 3 times the on) of -83 °C (-117 °F). Specimens were
finer grain sizes, although still rather bulk concentration. Alloy element segrega- aged at temperatures from 325 to 650 °C (615
coarse, showed that 0.5 to 0.6% Mo addi- tion at these boundaries was also detected. to 1200 °F) for times ranging from 4 min to 240
tions would prevent temper embrittlement However, the concentration of these alloy- h. Figure 21 shows the time-temperature-
caused by phosphorus in 3.5Ni-I.7Cr-0.2C ing elements was found to be only 2 to 3 embrittlement diagram developed from this
steels for aging times up to 1000 h at 475 and times that of the bulk concentration, and the work.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 701

50%FATT,°F 50% FATT,°F 50% FATT,°F

-300 -100 100 300 -300 -100 100 300 -300 -100 100 300
T I I 1 I

255 °C °C
r

°C 84 oc 22 °C
J r - ~

40oc
r J

-200 -100 100 200 -200 -100 0 100 200 -200 -100 100 200
50% FATT,°C 50%FATT, °C 50% FATT,°C

(a) (b) (e)

- - Composition, wt% l r ~ Composition, wt% - - ]

Composition, wt% I Ni[N~Cr C Sn
Steel ~ C r C P Steel Steel 'Ni Cr C As

Line A ........ 3.5 1.7 0.4 0.05-0.06 Line D ....... 3.5 1.7 0.4 0.046--0.048 Line G ........ 3.5 1.7 0.4 0.05-0.053
Line B ........ 3.5 • - . 0.4 0.05-0.06 Line E ........ 3.5 • . • 0.4 0.046-0.048 Line H ........ 3.5 -.. 0.4 0.05-0.053
Line C ........ • • • 1.7 0.4 0.05-0.06 Line F ........ • . . 1.7 0.4 0.046-0.048 Line I ........ • • • 1.7 0.4 0.05-0.053

Influence of alloying elements on the temper embrittlement of steels (compositions given in accompanying tables). (a) Steel containing 500 to 600 ppm P. (b)
Fig. 19 Steel containing 460 to 480 ppm Sn. (c) Steel containing 500 to 530 ppm As. The left end of bar gives the nonembrittled DBFr; the right end of bar gives the
DB]q- after embrittlement. Value between bar ends is the shift in 50% FATr. Source: Ref 92

Because of some minor differences in tion temperatures to account for hardness tlement (AFATT = 100 °C, or 210 °F) was
hardness and apparent inconsistencies in differences (Fig. 22). Only limited tests obtained at 475 to 500 °C (885 to 930 °F)
the results for aging at 525 °C (975 °F), were done at temperatures above 575 °C after 240 h aging (the longest time used).
additional work was performed (Ref 102). (1065 °F), but the work performed indicated Tests done using step cooling showed that
Minor corrections were made in the transi- that temper embrittlement occurred at tem- the degree of embrittlement that occurred
pering temperatures up to 675 °C (1245 °F), was greater than would have been predicted
Distance, in. x 10 6 which is close to the lower critical temper- from the isothermal data.
ature. The diagrams exhibit the classic C- The above data was analyzed to predict
0.02 0.04 0.06
1.0 curve appearance. The nose of the curve is the degree of grain-boundary phosphorus
at 550 °C (1020 °F), but maximum embrit- segregation in this steel (analysis for anti-
© S onAISI 5140
• S on Fe-0.6Sb
A PonAISl 5140 650 1200
I
• Sb on AISI 3340
0.8 ~ n [] Sb on Fe-2.2Sb I
I

• Sb on Fe-0.6Sb
600 - 1100
>,

0.6
? o
550
1000

0.4 ~~0
500 100% fibrous
Z
FATT, °C
- 900 .~
O -60
[] • -5O
450 A -40 E
0.2 • -30
W
[] -20 --..q ~ - o 8OO
W

• -10
400 V 0
----_____
• 10
0 ¢ 20 )
700
0 0.5 1.0 1.5
350 I I [
Distance, nm 0.1 1 10 100 103
17;,rl 20 Normalized intensities of Auger peaks (as a Embrittling time, h
"qS" function of depth below the surface) from
antimony, sulfur, and phosphorus segregated to grain Time-temperature diagram for isothermally temper-embrittled AISI/SAE 3140 alloy steel showing
boundaries or free surfaces (depth profiling by argon "hiS" 21 constant embrittlement levels (100% fibrous FA'I-I-) for quenched and tempered (675 °C, or 1245 °F, for
ion sputtering). Source: Ref 99 1 h) specimens. Source: Ref 101

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 702 / Service Characteristics of Carbon and Low-Alloy Steels

700 have compared results for a variety of mi-

crostructures as a function of hardness. In
general, tempered martensite is more sus-
-- 1200
650 O ceptible than tempered bainite to temper
embrittlement, but tempered bainite is more
susceptible than pearlitic-ferritic structures.
600 - 1100 This analysis is somewhat misleading, how-
o LL ever, because nonembrittled tempered mar-
tensite is much tougher than nonembrittled
550 bainite at the same hardness, and after
embrittlement, tempered martensite is still
E
tougher than tempered bainite. The same
500
( # holds true when comparing bainitic and
- 900 pearlitic-ferritic microstructures. This has
E
LU C o r r e c t e d transition t e m p e r a t u r e been demonstrated for chromium-molybde-
(100% fibrous FATT), °C ~ ~ tu num-vanadium steels, as shown in Fig. 25
450 --
O-60 , -25 , 5 ~"~ ~ ~ (Ref 106). The alloy composition was 0.3C-
• -as (} -20 ~ 10 ~_~__. -800 0.91Mn-0.27Si-0.15Ni-l.3Cr-I.2Mo-0.31V-
z& -50 ~ -15 15
400 -- • -45 ~ -10 ® 20 0.025P-0.0045S-0.005As-0.0008Sb-0.027Sn.
[] -40 ~ -5 ~ 25 ~ ~ The toughness of embrittled tempered mar-
[ ] 30
•
V -30
-35 [] 0
~ - 700
tensite was better than that of nonembrit-
tied tempered bainite over the hardness
350 I I I I
0.1 1 10 100 10 3
range evaluated. Also, the shift in 50%
F A T T for tempered martensite increased
Embrittling time, h
with hardness and was greater than that for
F i g . '~2 Revised time-temperature diagram for temper-embrittled AlSl/SAF 3140 alloy steel. Source: Ref 102
tempered bainite. Only one hardness level
was obtained for the ferrite-pearlite condi-
tion, and the shift in 50% F A T T because of
mony, tin, and arsenic was not performed, as the grain size becomes larger. Neverthe- embrittlement (step cooling) was only 5 °C
but the amount of these elements was as- less, coarse-grain steels are recognized to (9 °F).
sumed to be very low) (Ref 103). Auger be more severely embrittled than fine-grain Embrittlement Predictive Equations. Much
analysis of similar steels was used to calcu- steels. Figure 24 shows the results of aging of the research on temper embrittlement has
late the monolayers of phosphorus segregat- a 0.33C - 0.59Mn - 0.03P - 0.031S - 0.27Si - concentrated on the influence of composi-
ed to the prior-austenite grain boundaries 2.92Ni - 0.87Cr steel for various times at 500 tion on susceptibility to temper embrittle-
(Fig. 23). °C (930 °F) followed by water quenching. ment for fixed embrittlement conditions. In
Microstructure and Grain Size. It is well Prior to aging, the specimens had been general, these studies have concentrated on
known that microstructure influences the austenitized (at 850 °C, or 1560 °F, for two basic steel grades, nickel-chromium-
susceptibility to temper embrittlement and fine-grain specimens; at 1200 °C, or 2190 °F, molybdenum-vanadium and 2V4Cr-IMo,
the resulting degree of embrittlement. Be- for coarse-grain specimens) and tempered which are used for rotors and pressure
cause the impurity and alloying elements at 650 °C (1200 °F) for I h; they were oil vessels, respectively.
segregate to the prior-austenite grain quenched after tempering. The coarse-grain In a report on an ASTM study of vacuum
boundaries, grain size has an influence. As specimens were embrittled to a greater ex- carbon deoxidized nickel-chromium-molyb-
the grain size becomes larger, the grain- tent than the fine-grain specimens. Similar denum-vanadium rotor steels isothermally
boundary surface area decreases. There- results have been obtained by others (for embrittled at 400 °C (750 °F) for 10 000 h
fore, for a fixed level of impurities and example, Ref 105). (Ref 107), the shift in F A T T (2~FATT) in
constant embrittling temperatures and time, Matrix microstructures are also impor- degrees centigrade was correlated to the
there will be greater coverage of the grain tant because they control toughness, for impurity content and molybdenum concen-
boundaries in a coarse-grain steel than in a both nonembrittled and embrittled steels. tration (all in weight percent) by:
fine-grain steel. However, the distance over Most studies have evaluated temper embrit-
which the impurities must diffuse increases tlement of martensitic specimens, but a few AFATr = 7524P + 7194Sn + 1166As - 52Mo
- 450 000(P × Sn) (Eq 1)

No significant influence was found for anti-

mony. Equation 1 states that phosphorus,
tin, and arsenic increased embrittlement,
while molybdenum decreased it. Also, a
phosphorus-tin interaction that decreased
embrittlement was observed.
A correlation between the 50% FATT
and impurity content (J factor) for both
nickel-chromium-molybdenum-vanadium and
2 V4Cr-1Mo steels has been demonstrated (Ref
001 01 1 10 100 103 104 105 106 108). The J factor equation is:
Embrittling time h
J = (Mn + Si)(P + Sn) × 104 (Eq 2)
Time-temperature diagram for the segregation of phosphorus in temper-embrittled AISI/SAE 3140
F i g . 2 3 alloy steel. The numbers next to the curves describe the degree of phosphorus segregated during the
embrittlement treatment (not including the 006 monolayers of phosphorus segregated prior to the isothermal where all concentrations are in weight per-
aging treatments) Source: Ref 103 cent.

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 703

250 degree of embrittlement is greater for mar-

(J
o tensite than bainite and least for ferrite-
200 - - 400 pearlite. However, embrittled martensite is
Coarse grain ,~ ~ ~ ~ ~ still tougher than nonembrittled bainite,
while embrittled bainite is tougher than non-
150
f
i ( -- 300 ~
0a
embrittled ferrite-pearlite. Additional infor-
f ~_ marion may be found in Ref 110 to 114.
100 l
r
u_
F -- 200
Tempered
o i j Fine grain
x~ 50
-- 100
Martensite Embrittlement
o o Tempered martensite embrittlement (also
o 0
known as 350 °C, or 500 °F, embrittlement
--0
and as one-step temper embrittlement) of
-50 high-strength alloy steels occurs upon tem-
- - -100 pering in the range of 205 to 370 °C (400 to
-100 700 °F) (Ref 115-132). It differs from temper
4 8 10 embrittlement in the strength of the material
(a) Time at 500 °C (930 °F), h and the temperature exposure range. In
temper embrittlement, the steel is usually
250 450 tempered at a relatively high temperature,
producing lower strength and hardness, and
embrittlement occurs upon slow cooling af-
200 360 o_ ter tempering and during service at temper-
atures within the embrittlement range. In
~ _ _ _ . . . .

p_" Coarse grain . ~

tempered martensite embrittlement, the
u_ ,~ r- ~ " " ~ < steel is tempered within the embrittlement
150 /I _ ~ 270 u_
range, and service exposure is usually at
Fine grain room temperature. Therefore, temper em-
brittlement is often called two-step temper
o 100 180 ,~, embrittlement, while tempered martensite
embrittlement is often called one-step tem-
w

E
"-- .E
per embrittlement.
50 s y 90 Because alloy steels are heat treated to
form martensite and then may be tempered
within the embrittlement region, this prob-
0, f 0 lem has been widely referred to as tempered
0 2 10 martensite embrittlement (TME). While
(b) Time at 500 °C (930 °F h some have claimed that structures other
than martensite are not embrittled by tem-
Influence of prior-austenite grain size on the temper embrittlement of a nickel-chromium alloy steel that
Fig. 24 was heat treated to produce t w o levels of grain size. The alloy was tempered at 650 °C (1200 °F) and aged
pering in this region, it is well established
various times at 500 °C (930 °F). (a) Actual 100% fibrous FAI-F. (b) Change in 100% fibrous FAll'. Source: Ref 104 that lower bainite is embrittled when tem-
pered in this region. Other structures, such
as upper bainite and pearlite/ferrite, are not
A more detailed correlation has been pro- The above discussion shows that carbon embrittled by tempering in this region.
vided for nickel-chromium steels doped steels are immune to temper embrittlement While temper embrittlement is evaluated
with manganese, phosphorus, and tin (Ref and that only alloy steels are susceptible to by the change in the ductile-to-brittle tran-
109). The equation combines the grain- it. Both bulk composition and impurity lev- sition temperature, most studies of tem-
boundary phosphorus and tin concentra- els are important, although without the lat- pered martensite embrittlement have evalu-
tions, the prior-austenite grain size, and the ter, temper embrittlement will not occur. ated only the change in room-temperature
hardness level. Equation 3 was extended to The most potent impurity elements are an- impact energy. In general, when an as-
a nickel-chromium-molybdenum-vanadium timony, phosphorus, tin, and arsenic (in quenched alloy steel is tempered, the tough-
steel with hardnesses of 20 and 30 HRC, order of decreasing potency). Manganese ness at room temperature increases with
ASTM grain sizes of No. 3 and No. 7, and and silicon also promote embrittlement. Al- tempering temperature up to about 200 °C
isothermal embrittlement at 480 °C (895 °F) loys containing nickel and chromium in (390 °F). However, with further increases in
for 6000 h (Ref 110). The resulting equation combination are more susceptible than tempering temperature, the toughness de-
was: those that contain nickel or chromium sep- creases. Then, with increasing tempering
AFATT = 4.8P + 24.5Sn + 13.75(7 - GS) arately. Molybdenum additions are effec- temperatures above about 400 °C (750 °F), the
tive in retarding or eliminating temper em- toughness increases again. This change in
+ 2(HRC - 20) + 0.33(HRC - 20) brittlement when impurities are present. toughness with tempering temperature is not
(P + Sn) + 0.036(7 - GS)(HRC - 20) However, to be effective, the molybdenum apparent when examining hardness or tensile
(P + Sn) (Eq 3) must be dissolved in the ferritic matrix, not strength data, which generally decrease with
tied up as carbides. Embrittlement occurs increasing tempering temperatures.
where AFATT is in degrees centigrade, the because of tempering or service time within To illustrate the influence of tempering
concentrations of phosphorus and tin are a temperature range of 300 to 600 °C (570 to temperature on properties, Fig. 26 shows
the Auger peak-height ratios with respect to 1110 °F) or because of slow cooling through tension, hardness, and Charpy V-notch
iron, HRC is the Rockwell hardness, and this range. Coarse-grain material is more room-temperature tests for AISI 4340 alloy
GS is the grain size number. susceptible than fine-grain material. The steel (Ref 115). Room-temperature impact

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 704 / Service Characteristics of Carbon and Low-Alloy Steels

250 observation that TME did not occur in

high-purity steels but did occur in steels
O
200 400 containing elements such as phosphorus
o
(Ref 117, 118). Nitrogen was also observed
150 to cause embrittlement, as were large
- 300 amounts of manganese and silicon.
E, F/P NE After these developments, several studies
E 100 E
- 200 suggested that impurities segregated to pri-
t.-
I.- NE, F/P or-austenite grain boundaries and cementite
<
LL 5O ,,< precipitation along grain boundaries were
.-------~ E
-- 100
both required for TME. The segregation of
-~ 0 f LO
impurities was seen as a necessity; cement-
"~ Martensite -0 ite precipitation was cited as an important
factor, but not one that would cause TME
-5O
by itself (Ref 120-122). At about the same
.-.......
6,- '--'-'- - -100 time, transmission electron microscopy
-100
showed that martensite in as-quenched me-
10 30 40
20
dium-carbon alloy steels has small amounts
(a) Hardness, HRC
of retained austenite films present between
the laths. Several studies attributed TME to
120 the decomposition of interlath-retained aus-
tenite to intedath cementite. It was suggest-
ed that alloying elements that promote the
100 180
decomposition of retained austenite to ce-
o
"~
F-
80
M a r t e n ~
J 135
LL
o
<I
mentite at lower temperatures, such as
manganese and chromium, enhance TME,
while alloying elements that do not promote
<
u-
austenite decomposition at low tempera-
6O tures, such as silicon, nickel, and alumi-
90 num, shift TME to higher temperatures.

J
~ 40
Most of these studies exhibited transgranu-
g, lar test fractures of controlled-composition
alloys.
Other studies have observed cementite
-- 45
20
precipitation from interlath-retained austen-
0 ~ F/P ite but have described a more complex
process for TME. For example, a study of
0
10 20 30 40
(b) Hardness, HRC TME in AISI 4340 and 300M alloy steels
stated that the magnitude of TME increased
Influence of microstructure on the temper embrittlement susceptibility of a chromium-molybdenum- significantly when the volume fraction of
Fig. 2 5 vanadium alloy steel as a function of hardness. (a) Actual 50% FATI-. (b) Change in 50% FAIq-. F/P,
ferrite-30% pearlite structure; E, embrittled; NE, nonembrittled. Source: Ref 106 interlath-retained austenite was high. A par-
tial decomposition of the retained austenite
to interlath cementite upon tempering was
energy decreases with tempering in the crit- many studies have shown that fractures are observed. Subsequent deformation during
ical region (205 to 370 °C, or 400 to 700 °F). partly or substantially intergranular, partic- testing transformed the balance of the re-
The tensile and hardness results show no ularly with tempering at about 350 °C (660 tained austenite that became destabilized
influence at room temperature. Figure 27 °F), other studies have observed only trans- because of the depletion of carbon during
shows room-temperature Charpy V-notch granular fractures. This difference may in- cementite precipitation. The study further
impact energy for AISI 4340 at two different fluence the respective interpretations of the concluded that the segregation of impurity
phosphorus levels, 0.003 and 0.03%, as a TME mechanism. Also, studies have been elements at the prior-austenite grain bound-
function of tempering temperature (Ref made using controlled impurity and compo- aries contributes to the embrittlement pro-
128). The low-phosphorus specimens have sition alloys or commercial alloys, and the cess by providing an alternate weak path for
consistently higher impact energies at all factors that cause TME may differ for these fracture to follow. The study thus account-
tempering temperatures, and both suffer a two types of alloys. ed for either transgranular or intergranular
reduction in impact energy with tempering The earlier studies of TME noted that fracture modes.
in the critical region. As a final example, embrittlement coincided with the initiation The concept that a dual role played by
Fig. 28 shows plane-strain fracture tough- of cementite precipitation (Ref 115). Such interlath carbides is the primary problem
ness and Charpy V-notch impact energy at low-temperature precipitated cementite is and impurity segregation is a secondary
room temperature for 300M alloy steel (a long and platelike and is present at the grain influence has been verified by others (Ref
high-silicon modification of AISI 4340) as a boundaries. Therefore, the shape and loca- 128-130). In an evaluation of TME in AISI
function of tempering temperature (Ref tion of the cementite produced at low tem- 4340 containing 0.003 and 0.03% P, both
127). Both test procedures show somewhat pering temperatures was cited as the cause steels exhibited TME (Ref 128). For the
similar results, although the tempering tem- of TME (Ref 116). low-phosphorus steel, interlath cementite
perature ranges for embrittlement are slight- The emphasis then shifted to the study of initiated cleavage across the martensite
ly different. the influence of impurity elements, chiefly laths, while for the high-phosphorus steel,
Activating Mechanism. The mechanism phosphorus (which segregates to the prior- intergranular fracture resulted from the
causing TME is not as well defined as that austenite grain boundaries during austeniti- combined influence of carbide formation
which causes temper embrittlement. While zation). This shift was brought about by the and phosphorus segregation. However, the

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 705

Tempering temperature, °F Tempering temperature, °F

0 300 600 900 1200 100 250 400 550 700 850
70
I 50 I I I II I I I - 35
-~ 40 - - O 0.003 wt% P ~, , ~ }30 ~'
O 60 O- • 0.03 wt% P
cc 125 Z,
I

~ 5o

"1- 40
- q ) 5 _E

0 0
30 0 100 200 300 400 500
Tempering temperature, °C
2500
-- 360
I:-'a Influence of phosphorus on the room-tem-
nn~, 27 perature Charpy V-notch impact energy of
-- 340 AISI 4340 alloy steel as a function of tempering tem-
2300 perature. Source: Ref 128

-- 320
k
2100 ~Tensile strength
-- 300 a 350 °C (660 °F) temper (Ref 130). Cement-
ite begins to precipitate at 250 °C (480 °F).
~. 1900 -- 280 .- With higher tempering temperatures, the
number of precipitated carbides increases,
-- 260 ~
and their thickness and length increase.
Phosphorus segregates to the prior-austen-
UO 1700 ite grain boundaries during austenitization
-- 240 but does not diffuse during these low-tem-
Yield strength ~'3 perature tempering treatments. This study
1500 N.,~
\ -- 220 showed that the initial drop in impact ener-
gy with tempering at 250 to 300 °C (480 to
-- 200 570 °F) was caused by the influence of
cementite. However, with tempering at 350
1300
°C (660 °F), the amount and size of the
-- 180
cementite at the grain boundaries was suf-
ficient to cause intergranular fracture at the
1100 160 grain boundaries that were already weak-
60 ened by the presence of impurities.
Instrumented Charpy V-notch testing of
o~ 40 AISI 4130 alloy steel containing 0.002 and
D~ j
Elongation 0.02% P showed similar losses in impact
c~ 20 energy for both steels when tempered in the
critical region (Ref 131). As would be ex-
0
l
7 7 pected, the higher-phosphorus steel had
lower impact strength at all tempering tem-
70!
50 peratures. However, the fractures were all
60 transgranular, irrespective of phosphorus
40 "~
content or tempering temperature. The pre-
50 maximum-load energy decrease with in-
:>,
creasing tempering temperature in the range
40 30 ~ of 200 to 400 °C (390 to 750 °F) was attrib-
oJ
uted to a change in the work-hardening rate.
E
3o
20 ~ A recent study of TME in Fe-0.25C-
-- 20 E 10Cr-base martensitic steels concluded that
10 TME was not due to the decomposition of
10
interlath-retained austenite (Ref 132). This
0 0 study instead suggested that TME was the
-20 150 315 480 650 result of the coarsening of interlath and
Tempering temperature. °C intralath carbides.
As this review has shown, our under-
Influence of tempering temperature on the properties of AISI 4340 alloy steel that was oil quenched standing of tempered martensite embrittle-
Fig. 26 and tempered 1 h. Source: Ref 115
ment is not complete. Fractures of test
specimens tempered within the critical re-
plane-strain fracture toughness test results levels of manganese, silicon, phosphorus, gion do vary with transgranular fractures
displayed TME only in the high-phosphorus and sulfur. These fractures were attributed observed with tempering between 200 and
steel; the Charpy V-notch results, on the to carbide precipitation and impurity segre- 300 °C (390 and 570 °F), and intergranular
other hand, showed TME in both steels. A gation at the prior-austenite grain bound- fractures usually, but not always, observed
study of AISI 4340 steel observed intergran- aries (Ref 129). Another study of 4340 found with tempering at about 350 °C (660 °F).
ular fractures in specimens with various substantial intergranular fracture only after These differences may be due to differences

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 706 / Service Characteristics of Carbon and Low-Alloy Steels

Tempering temperatu re, ° F M23C 6 carbides on the austenite grain was greater with aging at 650 °C (1200 °F)
200 400 600 800 1000 1200 boundaries• The mechanism for sensitiza- than at 700 °C (1290 °F). Increasing the
120 tion was initially proposed as being due to
I I I ] holding time caused the width of the deplet-
the depletion of chromium at the grain ed zone to increase. The width of chromium
/I
- 100
boundaries when the carbides form (Ref depletion along the grain boundaries around
133, 134). The chromium content adjacent
[•100 to the grain-boundary carbides was thought
individual carbides was much greater than
the width of the depletion zone into the
to drop below some critical limit, which grain interior ( - 3 ~m, or 120 i~in., versus
---: 80 rendered the alloy susceptible to severe -0.15 ~m, or 6 i~in., for samples aged 100 h
2 ~g localized attack by the corrosive environ- at 700 °C, or 1290 °F) (Ref 156). Evidence of
ment. molybdenum depletion has also been ob-
-0o :~ This theory for the sensitization mecha- tained. These direct measurements are in
g 6o nism was widely accepted, and consider- relative agreement with theoretical calcula-
able indirect evidence for its validity was tions of chromium levels in equilibrium with
L
subsequently developed• However, when growing carbides during sensitization (Ref
40 ~ the electron microprobe was developed in 160-163) and with empirical modifications of
~ 40
O Valid K'ic results the 1950s and researchers tried to detect the such theories (Ref 164).
• Invalid Kic results impoverished chromium zone, they were Transmission electron microscopy stud-
unable to do so (Ref 135,136), or they could ies of the precipitated carbides have dem-
2o I I I I 20
0 100 200 300 400 500 600 700
show chromium depletion only when the onstrated that the susceptibility to inter-
steel was carburized (Ref 137), or they granular corrosion with sensitization
Tempering temperature, °C
(a) detected only small regions of possible temperature and time correlates well with
chromium depletion (Ref 138). Therefore, it the morphology of the grain-boundary car-
Tempering temperature, °F was concluded that if such a zone exists, it bide precipitates (Ref 165-169). These stud-
200 400 600 800 1000 1200 must be less than l Ixm wide (40 i~in.) and ies have demonstrated that the grain-bound-
60 i I I the concentration gradient must be rather ary precipitate due to sensitization is always
steep• This prompted a number of other (Cr,Fe)23C 6, that is, M 2 3 C 6 carbide• The
40
theories to be developed to explain sensiti- preferred sites for the nucleation of M23C6
50
zation, for example, electrochemical con- are, in decreasing order of occurrence (Ref
sideration of the nobility of the carbides 166):
3o (Ref 139, 140), strains at the carbide/
4O
austenite interface (Ref 139), and grain- • Delta ferrite-austenite phase boundaries
boundary strain energy acting as the driving • Austenite grain boundaries
force for intergranular attack (Ref 141). • Incoherent twin boundaries
~= 30
The picture was further complicated by a • Coherent twin boundaries
20 ~
number of observations of intergranular The carbides grow in the plane of the grain
- 2oj/ E corrosion in austenitic stainless steels under or twin boundaries. The morphology of the
conditions unfavorable for sensitization, precipitated M23C6 depends on the type of
10 and in nonsensitized stainless steels with no boundary where precipitation occurs and
detectable grain-boundary precipitation the temperature (Ref 166). Precipitates that
(Ref 142-149). The latter occurred in highly form at ~-ferrite-austenite phase boundaries
0 0 oxidizing solution, such as the nitric dichro- or at austenite grain boundaries are den-
0 ~ 100 200 300 400 500 600 700 mate solution. Because of these difficulties, dritic or geometric in shape; the shape de-
/ Tempering temperature,°C attention was focused on the influence of
As-quenched pends on the boundary orientation and mis-
impurity elements, such as phosphorus, fit between the grains, the temperature, and
(b) that are segregated to the austenite grain the time at temperature. Those that form at
I:|t, 28 Influence of tempering temperature on (a)
boundaries. It was subsequently proposed incoherent twin boundaries look like rib-
"~" the room-temperature plane-strain fracture that intergranular corrosion of austenitic bons of connected trapezoids while those
toughness and (b) the Charpy V-notch impact energy stainless steels was due to the presence of a that form at coherent twin boundaries have
of 300M alloy steel that was austenitized 1 h at 870 °C
(1600 °F), oil quenched, and tempered for I h. Source:
continuous grain-boundary path of either an equilateral thin triangular shape. Grain-
Ref 127 carbides or solute segregated regions (Ref boundary precipitates are classified into
142). Direct evidence for impurity segrega- three categories:
tion, chiefly of phosphorus, has been ob-
in carbon, alloy, and impurity content, as tained from intergranular fracture surfaces • Dendritic shapes
well as in strength level, test temperature, of both sensitized and nonsensitized auste- • Separate geometric shapes
nature of the test, and grain size. Impurities nitic stainless steels (Ref 148-152). • Sheets of interconnected geometric parti-
appear to influence coarse-grain steels to a cles
The development of scanning transmis-
greater degree than fine-grain steels• sion electron microscopes (STEM) with The sheets form at the lower temperatures;
electron beam sizes of about l0 nm (100 A) dendrites form at higher temperatures with-
Sensitization in diameter, coupled with energy-dispersive in the sensitization range, and the small,
x-ray analysis, has provided direct proof of isolated geometric particles can form over
Austenitic stainless steels become suscep- chromium depletion due to carbide precipi- the entire sensitization range and above, up
tible to intergranular corrosion when sub- tation at the grain boundaries (Ref 153-159). to about 980 °C (1800 °F) (Ref 166). Sensitiza-
jected to temperatures in the range of 480 to These studies have demonstrated that sig- tion is most severe for specimens sensitized at
815 °C (900 to 1500 °F), generally from nificant chromium depletion occurs after temperatures in the lower portion of the range
welding or service conditions• This suscep- sensitization at grain boundaries adjacent to where sheets of interconnected geometrically
tibility has been termed sensitizatior and precipitated carbides• For equal times, the shaped carbides are formed. It is well known
has been attributed to the precipitation of degree of chromium depletion for T316LN that healing occurs with very long times in the

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 707

sensitization range; that is, long holding times Sensitization temperature, °F water, forming grain-boundary precipitates
result in reduced chromium depletion and 1000 1150 1300 1450 1600 of (Ti,Cr)(C,N) (Ref 177). The chromium-
reduced intergranular attack. When this oc- 1.13 j I I I Heat1 I I I to-titanium (Cr/Ti) ratio in these precipi-
curs, the sheets of interconnected particles tates is approximately 1/3. Aging at 480 to
number wt% P
gradually separate into arrays of thick geo- 550 °C (895 to 1020 °F) causes these precip-
1.01- - - (3 1 0.025--
metric particles (Ref 166). itates to grow, and the Cr/Ti ratio increases
Several approaches have been taken to ~ A 7 0.070 to approximately 1/2. This depletes the
minimize or prevent the sensitization of .E 0.89 grain-boundary zone around the precipi-
austenitic stainless steels. If sensitization tates of chromium, thereby increasing the
results from welding heat and the compo- susceptibility to intergranular corrosion.
nent is small enough, solution annealing will .~ 0.77 Again, long times at 480 to 550 °C (895 to
dissolve the precipitates and restore immu- 1020 °F) reduce the chromium gradient
nity. However, in many cases this cannot around the particles and restore corrosion
be done because of distortion problems or 0.65 resistance. Aging above 600 °C (1110 °F)
the size of the component. In these cases, a also produces resistance to intergranular
low-carbon version of the grade or a stabi- corrosion because the chromium in the
lized composition should be used. Complete °o 0.53 (Ti,Cr)(C,N) precipitates is replaced by ti-
immunity requires a carbon content below tanium, that is, the CrFFi ratio decreases. A
about 0.015 to 0.02% (Ref 170). Additions of titanium-stabilized 12% Cr ferritic grade
niobium or titanium to tie up the carbon are "~ 0.41 was also found to sensitize under similar
also effective in preventing sensitization as heat treatment conditions (Ref 178).
long as the ratio of these elements to the Duplex stainless steels are resistant to
carbon content is high enough. Stabilizing
~ 0.29
intergranular corrosion when aged in the
heat treatments aimed at producing inter- region of 480 to 700 °C (895 to 1290 °F). It
granular carbides are not very effective. 0.17 has been recognized for some time that
Few studies have been conducted on the duplex grades with 20 to 40 vol% ferrite
influence of sensitization on mechanical exhibit excellent resistance to intergranular
properties. In general, carbide precipitation 0.05 corrosion (Ref 179). A study of intergranu-
produces a slight increase in tensile strength 525 600 675 750 825 900
lar corrosion in AISI 308 stainless steel that
and a minor reduction in tensile ductility Sensitization temperature, °C was heat treated to produce 15% ferrite
(Ref 134). Heats of AISI 304 containing Influence of phosphorus content and aging found that aging at 600 °C (1100 °F) caused
various levels of phosphorus and sulfur H~5" 29 temperature on the relative loss in Charpy the precipitation of M23C 6 at austenite-fer-
after sensitization were tested using half- impact energy (tested using half-size specimens bro- rite boundaries (Ref 180). When this occurs,
ken after cooling in liquid nitrogen) for sensitized AISI
size Charpy V-notch specimens fractured at 304 stainless steel. Sensitization time, 5100 s. Source: most of the chromium in the M23C 6 c o m e s
liquid nitrogen temperature. Longitudinally Ref 171 from the ferrite grains; only a very small
oriented specimens were sensitized at 550 amount comes from the austenite grains. A
to 850 °C (1020 to 1560 °F) for 15 to 105 min. chromium-depleted zone is formed at the
The decrease in impact energy, relative to tergranular attack in ferritic grades than is austenite/carbide interface, which is very
the results before sensitization, were much the case for austenitic grades. In the case of narrow, compared to those in fully austen-
greater in phosphorus-doped steels than in welds, the attacked region is usually larger itic sensitized stainless steels. Aging for 7 h
sulfur-doped steels. Fractures were inter- for ferritic grades than for austenitic grades at 600 °C (1110 °F) replenished the chromi-
granular for the phosphorus-doped steels because temperatures above 925 °C (1700 um-depleted zone and stopped the localized
and transgranular for the sulfur-doped °F) are involved in causing sensitization. intergranular attack (ASTM A 262E test).
steels. Figure 29 shows the reduction in Ferritic grades with less than 15% Cr are Therefore, the healing of depleted zones is
half-size Charpy V-notch impact energy for not susceptible, however. One study dem- much more rapid in duplex grades.
phosphorus-doped steels aged at different onstrated that ferritic grades with 16 to 28% The aging of duplex stainless steels pro-
temperatures for 5100 s (85 min); the spec- Cr were susceptible to intergranular corro- duces a variety of phases in the ferrite (Ref
imens were tested after cooling in liquid sion when rapidly cooled from above 925 °C 180). Aging at 480 °C (895 °F) does not
nitrogen. For heats 1 and 6, embrittlement (1700 °F). This susceptibility was due to produce M23C6, but it will produce an ex-
was greatest at 750 °C (1380 °F); somewhat solution of carbides and nitrides followed tremely small, finely distributed, chromi-
higher or lower temperatures produced near- by their reprecipitation in the grain bound- um-rich c~' phase in the ferrite. Aging at 600
normal toughness. Embrittlement was greater aries. Subsequent annealing at 650 to 815 °C to 700 °C (1110 to 1290 °F) produces a
for high-phosphorus heats, but these levels (1200 to 1500 °F) restored corrosion resis- complex series of transformations of the
are greater than those encountered in com- tance (Ref 172). Therefore, the thermal pro- ferrite phase.
mercial heats. The loss in toughness for the cesses causing intergranular corrosion in Another study of AISI 308 showed that
phosphorus-doped heats increased with hold- ferritic stainless steels are different from for a given carbon content there is a critical
ing times at temperatures between 650 and those for austenitic stainless steels. Reduc- amount and distribution of the ferrite/
825 °C (1200 and 1515 °F); somewhat higher ing the carbon and nitrogen interstitial lev- austenite interfacial boundary area (Ref
or lower temperatures again produced near- els improves the intergranular corrosion re- 181). Above this critical level, the alloy is
normal toughness. The modified Strauss test sistance of ferritic stainless steels. immune to intergranular corrosion with ag-
produced higher corrosion rates for heats Sensitization can occur in titanium-stabi- ing between 480 and 700 °C (895 and 1290
sensitized at 650 °C (1200 °F), and the corro- lized ferritic stainless steels (Ref 177, 178). °F) for up to 1000 h. If the amount and
sion rate increased linearly with increasing The thermal treatment that causes sensiti- distribution of these boundaries is below the
time and increasing impurity content. zation, however, is altered by the addition critical level, two types of sensitization be-
Ferritic Stainless Steels. Sensitization and of titanium. First, high-temperature expo- havior can occur. The amount and distribu-
intergranular corrosion also occur in ferritic sure requires a temperature in excess of tion of the ferrite/austenite interfaces may
stainless steels (Ref 172-178). A wider range 1050 °C (1920 °F) to dissolve the Ti(C,N) be adequate to produce the rapid healing of
of corrosive environments can produce in- that reprecipitates upon cooling, even with chromium-depleted regions that form at the

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 708 / Service Characteristics of Carbon and Low-Alloy Steels

540 550

•I•'•
I I I [ 1000
Hardness, HV 220 /" ( c°u280
f - 300 500 56% Cr
320
520 f 340 450 ,,,,O" -- oo
360!
-- 950 > 400 ,.,,..,X3~ ~,'q['~
380 T ,..,O .,q-'
5OO ~_ f"
? "o, do
350 "1
N 480 -- 900
~" 2 5 0 - y
E
200
460 ~N
-- 850 150 .-~ -.~.0-...-0-
Time required for 100 15~/°Cr I
440 first stage of hardening 0 1 2 5 10 20 50 100 200 500 103
(increase of 10-15 HV)
,,Q Aging time, h
\ -- 800
420 Influence of aging time at 475 °C (885 °F) on
0.1 1 10 100 103 104 F i g . 31 the hardness of iron-chromium alloys with
15, 20.6, 23, 37, and 56% Cr. Source: Ref 192
Aging time, h

Time-temperature-constant hardness curves for Fe-30Cr (Alloy 90) after aging done between approx-
F i g . 3 0 imately 430 and 540 °C (805 and 1005 °F), around the region of 475 °C embrittlement. Specimens rolled (1020 °F), forming disk-shaped a' (Ref 196).
at 900 °C (1650 °F); starting hardness, 195 to 205 HV. Source: Ref 192
An Fe-20Cr alloy will decompose by nucle-
ation and growth at 470 °C (880 °F), while
austenite/carbide interface during aging. If A 1953 study using transmission electron Fe-30Cr, Fe-40Cr, and Fe-50Cr alloys will
not, the alloy will behave as a fully austen- microscopy observed that 475 °C embrittle- decompose spinodally at 470 °C (880 °F).
itic grade. Additional information on austen- ment caused the precipitation of a coherent Even for a severely embrittled alloy, 475
itic, ferritic, and duplex stainless steels is chromium-rich bcc phase with a lattice pa- °C embrittlement is reversible. Properties
available in the articles "Wrought Stainless rameter only slightly greater than the iron- can be restored within minutes by reheating
Steels" and "Cast Stainless Steels" in this rich bcc ferritic matrix phase. These precip- the alloy to 675 °C (1250 °F) or above (Ref
Volume. itates were extremely small, °for example, 188, 194). The degree of embrittlement in-
about 15 to 30 nm (150 to 300 A) in diameter creases with chromium content; however,
475 °C Embrittlement for an Fe-27% Cr alloy aged from 10 000 to embrittlement is negligible below 13% Cr.
34 000 h at 480 °C (900 °F). The precipitates Carbide-forming alloying additions, such as
Iron-chromium alloys containing 13 to were nonmagnetic and contained about 80% molybdenum, vanadium, titanium, and nio-
90% Cr are susceptible to embrittlement Cr. The rate of growth of the precipitates bium, appear to increase embrittlement,
when held within or cooled slowly through was very slow, and they did not appear to particularly with higher chromium levels.
the temperature range of 550 to 400 °C (1020 over age (Ref 191). Other studies have con- Increased levels of carbon and nitrogen also
to 750 °F). This phenomenon has been firmed these findings (Ref 192-199). The enhance embrittlement and, of course, are
called 475 °C (885 °F) embrittlement and it 1953 study was unable to explain its obser- detrimental to nonembrittled properties as
produces an increase in tensile strength and vations based on the existing iron-chro- well. Cold work prior to 475 °C (885 °F)
hardness and a decrease in tensile ductility, mium phase diagram. Later work (Ref 192) exposure accelerates embrittlement, partic-
impact strength, electrical resistivity, and concluded that 475 °C embrittlement was a ularly for higher-chromium alloys.
corrosion resistance (Ref 182-203). Micro- precipitation-hardening phenomenon result- Figure 30 demonstrates the C-curve na-
structure effects are minor: Grain bound- ing from the presence of a miscibility gap in ture of the increase in hardness due to aging
aries etch more widely, and the grain inte- the iron-chromium system below 600 °C for an Fe-30% Cr ferritic stainless steel (Ref
riors darken. (I 110 °F). The location of the miscibility gap 192). The nose of the curve decreases with
Numerous theories have been proposed to was later refined (Ref 200). time. Figure 31 shows the results for aging
account for 475 °C embrittlement. The prob- Aging at 475 °C (885 °F) has been shown to at 475 °C (885 °F) for up to 1000 h for
lem occurs with iron-chromium ferritic and cause a rapid rate of hardening with aging iron-chromium alloys with 15, 20.6, 23, 37,
duplex ferritic-austenitic stainless steels, and between about 20 and 120 h because of ho- and 56% Cr. As the chromium content in-
not with austenitic grades. The earliest theo- mogeneous precipitation. The rate of harden- creased, the time to the initial increase in
ries suggested that embrittlement was due to ing is much slower with continued aging from hardness decreased. Over 200 h were re-
precipitation of second phases such as 120 to 1000 h. During this aging period, the quired for the 15% Cr alloy, only about 4 h for
phosphides (Ref 183), carbides or nitrides precipitates grow. Aging beyond 1000 h pro- the 20.6% Cr alloy, and less than 1 h for the
(Ref 184), or oxides (Ref 188). Others suggest- duces little increase in hardness because of 23% Cr alloy. The initial increase in hardness
ed that embrittlement was due to cr phase, the stability of the precipitates, which do not was nearly instantaneous for the 37 and 56%
which does form in iron-chromium alloys, or grow larger than about 30 nm (300 ,~). Cr alloys. Accompanying the increase in
some transitional phase that precedes ~r for- Precipitation of the chromium-rich (~' hardness with aging at 475 °C (885 °F) are an
mation (Ref 185-187). However, tr forms at phase in iron-chromium alloys can occur increase in tensile and yield strength and a
higher temperatures and has never been de- either by spinodal decomposition or by nu- decrease in tensile ductility and impact ener-
tected in 475 °C embrittled specimens. Japa- cleation and growth, depending on the aging gy. Examples of such data can be found in
nese researchers (Ref 189, 190) suggested that temperature and alloy composition. For ex- Ref 185, 186, 199, and 201 to 203.
475 °C embrittlement was due to ordering, ample, an Fe-30% Cr alloy will decompose
that is, from the formation of Fe3Cr, FeCr, to chromium-rich precipitates in an iron- Sigma Phase Embrittlement
and FeCr3 superlattices. However, subse- rich matrix (chromium-depleted) inside the
quent experiments using neutron diffraction spinodal at 475 °C (885 °F), forming spheri- The existence of cr phase in iron-chro-
failed to detect evidence of ordering. cal ec', or outside the spinodal at 550 °C mium alloys was first detected in 1907 by

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 709

the observation of a thermal arrest in cool- Temperature, ° F

ing curves (Ref 204). It was suggested that 1000 1200 1400 1600 1800
90 I I I I
the thermal arrest could be due to the
formation of an F e C r intermetallic com- 60
pound. The first actual observation of cr in 80
iron-chromium alloys was reported in 1927
(Ref 205). The phase was referred to as the
brittle constituent (B constituent) and was 50
reported to be corundum hard. In a discus-

-J
sion of this paper, it was suggested that the
B constituent was the F e C r intermetallic 40 •
-. ....._
phase detected in 1907. The cr phase was

/
identified by x-ray diffraction in 1927 (Ref
50 ~ t
206) and in 1931 (Ref 207). Numerous earli-
er investigators, however, had failed to de- 30 g
tect the presence of or; this failure caused E
E
considerable confusion.
After the existence of ~r was firmly estab- Aging time, h 20
lished, numerous studies were conducted to o 48
define the compositions and temperatures 20 - - • 100 % ~r~.~
zx 500
over which cr could be formed. This pro- • 1000
duced a series of refinements to the iron- E] 3000
10
chromium equilibrium diagram, as in the 10
case of 475 °C embrittlement. Successive
studies demonstrated that cr could form in 0 0
alloys with lower and lower chromium con- 500 600 700 800 900 1000
tents under the proper conditions; one Temperature, °C
study demonstrated the formation of ~r at
Influence of aging time and temperature on the room-temperature impact energy of an Fe-25Cr-20Ni
480 °C (900 °F) in an alloy with less than Fig. 32 alloy. Annealed value, 89 J (66 ft • Ibf). Source: Ref 225
12% Cr. In general, ,r forms with long-time
exposure in the range of 565 to 980 °C (1050
to 1800 °F), although this range varies some-
what with composition and processing. Sig- retard cr formation. Carbon additions de- tion of 1 lMo-29Cr-55Fe-5Ni in a specimen
ma formation exhibits C-curve behavior crease cr formation by forming chromium aged 3000 h at 815 °C (1500 °F) (Ref 220). A
with the shortest time for formation (nose) carbides, thereby reducing the amount of WDS analysis using the electron micro-
generally occurring between about 700 and chromium in solid solution. Additions of probe of A I S I 316 heated 60 h at 870 °C
810 °C (1290 and 1490 °F); the temperature tungsten, vanadium, titanium, and niobium (1600 °F) obtained a cr composition of
that produces the greatest amount of cr with also promote cr formation. As might be 26.4Cr-3.3Ni-53.7Fe-8.5Mo (Ref 221). An
time is usually somewhat lower. expected, cr forms more readily in ferritic analysis of cr in a failed A I S I 316 super-
The general characteristics of cr phase than in austenitic stainless steels. Coarse heater tube using S T E M - E D S of thin foils
have been reviewed extensively (Ref 209- grain sizes from high solution-annealing obtained a composition of 52.7Fe-37Cr-
216). The name sigma stems from work temperatures retard cr formation, and prior 3.7Mo-4.8Ni-0.7Si-0.4Mn (average of sev-
done in 1936 (Ref 217). Sigma-type phases cold working enhances it. The influence of eral results).
have since been identified in over fifty bi- cold work on cr formation depends on the Reference 223 presents an analysis of 0r in
nary systems and in other commercial al- amount of cold work and its effect on re- three versions of A I S I 310 (a standard ver-
loys (for example, nickel-base superalloys). crystallization. If the amount of cold work sion, a low-carbon version, and a high-
Sigma phase has a tetragonal crystal struc- is sufficient to produce recrystallization at silicon version) and titanium-stabilized
ture with 30 atoms per unit cell and a c/a the service temperature, cr formation is en- A I S I 316. The analysis was done using the
ratio of approximately 0.52 (Ref 216). Sigma hanced. If recrystallization does not occur, electron microprobe. F o r the low-carbon
in iron-chromium alloys has a hardness the rate of a formation may not be affected. A I S I 310S, the chromium-to-iron (Cr/Fe)
equivalent to approximately 68 HRC (940 Small amounts of cold work that do not ratio in cr was constant and equal to 1 for all
HV). Because of its brittleness, cr often promote recrystallization may actually re- temperatures and times; the composition
fractures during indentation. At room tem- tard cr formation (Ref 218). was 46Cr-46Fe-8Ni. F o r the AISI 310 and
perature, cr is nonmagnetic. Embrittlement The composition of cr in austenitic stain- A I S I 310Si, the composition of cr varied
effects due to cr are greatest at room tem- less steels is more complex than it is for with temperature and time, and the Cr/Fe
perature. simple iron-chromium ferritic grades. Sev- ratio of 1 was obtained after a certain time
Austenitic and Ferritic Stainless Steels. In eral studies, particularly for AISI 316, have at any temperature used. The higher the
commercial alloys, silicon, even in small analyzed the composition of or. These stud- temperature, the shorter the time required
amounts, markedly accelerates the forma- ies have used a traditional wet chemical to obtain a Cr/Fe ratio of one. When the
tion of or. In general, all of the elements that analysis of bulk-extracted or, wavelength- Cr/Fe ratio stabilized at 1 for 310 and 310Si,
stabilize ferrite promote ~r formation. Mo- dispersive spectroscopy (WDS) with an ~r had the same composition as in the low-
lybdenum has an effect similar to that of electron microprobe, or energy-dispersive carbon version. The work with 316Ti
silicon; aluminum has a lesser influence. spectroscopy (EDS) with either a scanning showed that the molybdenum content in cr
Increasing the chromium content, of transmission electron microscope (for thin increased with temperature. As the molyb-
course, also favors cr formation. Small foils) or a scanning electron microscope denum content in cr increased, the iron
amounts of nickel and manganese also in- (SEM) (for bulk specimens) (Ref 219). content increased, and the chromium and
crease the rate of ,r formation, although An analysis of cr in AISI 316 by the nickel contents decreased. The ( C r + M o ) /
large amounts, which stabilize austenite, SEM-EDS approach obtained a composi- ( F e + N i ) ratio was constant, and the formu-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 710 / ServiceCharacteristicsof Carbonand Low-AlloySteels
1000

\ \
18
- 13
371°C 900 ~ . ~ " / - = 1600
12
,, L3
o
800
- 1400 °
10' P 7OO

o vz
\ X X
9S
8~
E
6OO " " ~.,,,.., , . , ~ °C
100°C-
- 1200 ~.

1000
I--

7~ 5O0

\ \
400 ~::::::~ - 800
6 ~
427 °C 0.3 1.0 2.0 10 30 100
5 3 8 °C 5 ~- Time, h

6
r;~
HS" 34 Time-temperature relationships to produce
25 and 100 °C (75 and 212 °F) DBTI-s for a
-
36 29Cr-4Mo ferritic stainless steel as a function of aging
times that cover both the 475 °C (885 °F) embrittlement
4 ~'~ ~ -- ---- 482 °C ~ ' ' - ' 2 range and the ~ phase embrittlement range. Source:
2
Ref 226
1
816°C
0 0
0.2 0.4 0.6 1 2 4 6 10 20 40 60 100 200
Aging time, h intergranular cr films formed, producing in-
tergranular tensile and impact fractures.
Influence of aging time and temperature on the room-temperature Charpy impact energy of a low
Fig. 33 interstitial content 29Cr-4Mo ferritic stainless steel. Source: Ref 226
While some studies have demonstrated
only a minor increase in hardness and
strength because of cr formation, studies of
la for ~r was expressed as (Cr,Mo)35(Fe,Ni)65. rite-stabilizing elements such as molybde- some steels have demonstrated more sub-
Although the chemical composition of num and titanium; the rate of formation can stantial changes. Tensile ductility, like
varied with temperature and time, the c/a be quite rapid in these alloys. toughness, is generally substantially re-
ratio of the tetragonal unit cell was constant The most sensitive room-temperature duced. One study has demonstrated that
at 0.519. property for assessing the influence of cr is high-chromium ~-hardenable alloys are use-
The influence of molybdenum, which is the impact strength. A study of the influ- ful in applications involving high-tempera-
known to promote ~r formation, was exam- ence of cr on the toughness of AISI 310 ture erosion or wear, for example, exhaust
ined using wrought 25Cr-20Ni alloys with shows the dramatic loss in toughness due to valves (Ref 227). Such steels generally con-
additions of up to 8.2% Mo (Ref 224). Sigma cr (Fig. 32). With increasing time at temper- tain from 20 to 30% Cr, about 0.25 to 0.45%
was analyzed with the electron microprobe ature, particularly in the range of 760 to 870 C, and additions of manganese and nickel to
using specimens aged at 850 °C (1560 °F) for °C (1400 to 1600 °F), the toughness de- produce a duplex structure; they also gen-
525 h (the 8.2% Mo specimen was aged for creased by about 85%. erally contain additions of elements that
350 h). As the molybdenum content in the The influence of high-temperature expo- promote cr formation, such as silicon and
25Cr-20Ni alloys increased to 8.2%, the chro- sure on the toughness of a low-interstitial molybdenum. Such steels can be hardened
mium content in ~ decreased from 42.6 to 29Cr-4Mo ferritic stainless steel has been to about 40 HRC by ~ and retain their
31.0%, the iron content decreased from 43.3 examined (Ref 226). Figure 33 shows the strength at high temperatures. Although the
to 38.8%, the nickel content remained con- room-temperature impact strength trends toughness of such steels is reduced about
stant at about 10%, and the molybdenum for this alloy as a function of aging temper- 35% by the presence of or, they do perform
content increased to 14.3%. In addition, as ature and time. Aging at 371 °C (700 °F) well as long as the extent of or-phase embrit-
the molybdenum content of the steels in- produced no loss in toughness. However, tlement is not severe. The toughness and
creased, the volume fraction of cr increased aging at 427,482, and 538 °C (800, 900, and ductility of or-containing steels at high tem-
from 3 to 60% for these aging treatments. 1000 °F), in the range of 475 °C embrittle- peratures is considerably better than at
Sigma formation in pure iron-chromium ment, produced a loss of toughness that was room temperature; however, such steels are
alloys is rather sluggish, which accounts for most pronounced at 482 °C (900 °F). Aging not useful at high temperatures if shock
much of the confusion in early studies con- at higher temperatures, where tr was resistance is required.
cerning its existence. Subsequent work formed, produced more pronounced embrit- An examination of the high-temperature
showed that the formation of cr was dramat- tlement, which was greatest at 816 °C (1500 properties of a 25Cr-20Ni-2Si (type 314)
ically accelerated by prior cold work and by °F). Figure 34, which is a summary of this austenitic stainless steel aged between 650
silicon additions (Ref 217). In ferritic stain- data, shows the time at aging temperatures and 980 °C (1200 and 1800 °F) showed that
less steels, the addition of even minor between 371 and 978 °C (700 and 1790 °F) with the proper amount and distribution of
amounts of other alloying elements expands required to produce a DBTT of 25 and 100 ~r, substantial increases in yield and tensile
the compositional range over which ~r may °C (77 and 212 °F). This produces a C-curve strengths result for test temperatures up to
form and increases the rate of formation. presentation of the time for embrittlement 760 °C (1400 °F) (Ref 228). For conditions
All of the ferritic stabilizing elements pro- as a function of aging temperature. For cr involving slow strain rates, cr phase reduces
mote ~ formation. formation, embritttement was most rapid at creep resistance. Variations in austenite
Sigma will also form in austenitic alloys. about 775 °C (1425 °F); 475 °C embrittle- grain size, however, can exert an even
In fully austenitic alloys, cr forms from the ment was slower with a maximum rate at greater effect on high-temperature tensile
austenite along grain boundaries. If ~-ferrite about 480 °C (900 ° F). Chi phase was also and creep properties. Fine grain sizes in-
is present in the austenitic alloy, cr forma- observed, along with or, after aging in the crease short-time high-temperature tensile
tion is more rapid and occurs in the 8- high-temperature range. Sigma formed over strength but reduce long-time creep
ferrite. Sigma will form more readily in the range of 595 to 925 °C (1100 to 1700 °F). strength. The room-temperature ductility of
austenitic alloys containing additions of fer- Embrittlement was most pronounced when these alloys is poor but is restored at tem-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 711

peratures above 540 to 650 °C (1000 to 1200 Table 1 Classificationof hydrogen experience only a decrease in tensile elon-
°F). A fine distribution of~ is detrimental to damage processes gation and reduction in area. The phenom-
creep strength but minimizes loss of ductil- Hydrogen embrittlement enon is sensitive to strain rate and hydrogen
ity at room temperature. Sigma has also Hydrogen environmental embrittlement content. It is increased by decreases in
been found to be detrimental in thermal- Hydrogen stress cracking strain rate and increases in hydrogen.
fatigue situations (Ref 229). High-tempera- Loss in tensile ductility
Hydrogen attack Hydrogen
ture exposure can produce a variety of Blistering
phases, and embrittlement is not always due Shatter cracks, flakes, and fisheyes Environmental Embrittlement
solely to ~ formation (Ref 230). Therefore, Microperforation Hydrogen attack occurs at high tempera-
each situation must be carefully evaluated Degradation in flow properties tures rather than at or near room tempera-
Metal hydride formation
to determine the true cause of the degrada- ture. Hydrogen reacts with carbon within
tion of properties. Source: Ref 235,236 the steel to form methane gas. This may
Duplex Stainless Steels. Sigma phase is simply decarburize the steel, lowering its
known to form quite rapidly in duplex stain- strength, or may produce cracks or fissures.
less steels. One study, for example, ob- Hydrogen has no known beneficial effects This problem does not occur below 200 °C
served ~ after 15 min at 750 °C (1380 °F) and on iron and steel, only detrimental effects. (390 °F). Blistering is observed in low-
2 min at 850 °C (1560 °F) in a 20Cr-10Ni- While the influence of hydrogen on high- strength steels and other metals. It is caused
3Mo duplex alloy (Ref 231). Another study strength steels is well known, even the by atomic hydrogen diffusing to internal
found that (r formed after 2 min at 900 °C softest irons are not immune from hydrogen defects or inclusions where it precipitates as
(1650 °F) in a 21Cr-7Ni-2.4Mo-l.3Cu (UNS damage. molecular (diatomic) hydrogen. This gener-
$32404) duplex stainless steel containing Despite the thousands of research articles ates substantial pressure, which produces
33% ferrite (Ref 232). Sigma formed in a published on this subject, there is still a blisters. Shatter cracks, flakes, and fisheyes
C-curve manner along with a number of great deal that is not fully understood and are terms that describe cracks or the surface
other phases, but sigma was the worst em- considerable controversy over the mecha- appearance of a fracture in castings, forg-
brittler and led to massive pit initiation in nisms by which hydrogen exerts influence. ings, wrought alloys, or weldments. These
corrosion tests. Sigma formed in the ferrite, The damage produced by hydrogen has defects are caused by the presence of ex-
and pitting occurred in the chromium-mo- been known since at least 1875 (Ref 234). cessive hydrogen in the liquid melt prior to
lybdenum-depleted ferrite-(r regions. The general term hydrogen damage has solidification. The solubility of hydrogen in
An evaluation of the effects of alloying been used to be consistent with a phenom- the melt is much greater than after solidifi-
elements on (r formation in duplex stainless enological classification scheme covering cation; cracking can result if the hydrogen
steels found that increasing chromium and the various hydrogen degradation processes level is above a critical level and is not
molybdenum contents caused an increase in (Ref 235,236). Hydrogen can cause a wide permitted to diffuse out of the material
the rate of (r formation and in the maximum variety of problems, and it is helpful to before cooling to ambient temperature. Mi-
amount produced (Ref 233). Increasing the categorize them (Table 1). Other reviews on croperforation occurs mainly when steels
nickel content decreased the maximum hydrogen damage can be found in Ref 237- are exposed to very high-pressure hydrogen
amount of (r that could form but increased 251. near room temperature. As the term im-
the rate of (r formation. Sigma formation Hydrogen environmental embrittlement, plies, small fissures are formed in the metal.
occurred primarily by the decomposition of hydrogen stress cracking, and loss of tensile A degradation in flow properties occurs in
ferrite into cr and austenite. ductility are grouped under hydrogen em- steels at room temperature. Metal hydride
The influence of cr on corrosion character- brittlement. Hydrogen environmental em- formation does not occur in steels, but it is
istics is rather complex. In many instances, brittlement refers to the degradation of me- a problem in metals such as titanium, nio-
little influence is observed in environments chanical properties in metals such as steels bium, and zirconium.
normally used with a particular alloy. Large (r when they are deformed in a hydrogen- The fact that hydrogen embrittlement in-
particles appear to be rather harmless; a fine containing environment. Degradation is en- creases with decreasing strain rate and in-
distribution of particles, particularly if present hanced by slow strain rates and high hydro- creasing temperature above ambient, as
at the grain boundaries and in highly oxidizing gen pressures. well as the fact that these are mutually
solutions, is more harmful. Hydrogen stress cracking is character- dependent variables, is directly opposite to
ized by the brittle fracture of an alloy that is the general trend for bcc metals that under-
Hydrogen Damage normally ductile while being stressed below go a ductile-to-brittle transition behavior
its yield strength in a hydrogen-containing with increasing strain rate and decreasing
Of all the embrittlement problems dis- environment. This problem is also called temperature below ambient. This unique
cussed in this article, that due to hydrogen "static fatigue" and hydrogen-induced behavior arises because the phenomenon is
is the most widespread and influences the cracking. In general, there is a certain stress controlled by the diffusion of hydrogen
behavior and properties of nearly all ferrous level (threshold) below which such cracking through the lattice. Hydrogen effects are
alloys and many other metals. Indeed, the does not occur. Hydrogen stress cracking further complicated by the chemical analyt-
ubiquitous nature of hydrogen and the wide involves the absorption of hydrogen, fol- ical limitations associated with hydrogen
variety of forms of damage that it can pro- lowed by hydrogen diffusion to highly analyses. Although a bulk hydrogen analy-
duce make this subject difficult to address stressed regions, particularly those associ- sis can be made, hydrogen in the lattice
and summarize. Hydrogen can come from ated with notches. Thus, there is a delay cannot be differentiated from hydrogen
many sources. It can be retained in the time, or incubation time, before cracking within voids or traps.
steel: occurs. H2S stress cracking is a specific Figure 35 demonstrates the influence of
type of hydrogen stress cracking. hydrogen on the tensile reduction of area as
• Upon solidification (supersaturated) The loss of tensile ductility, one of the a function of strength level (Ref 237). The
• From acid cleaning (pickling) oldest forms of hydrogen damage, occurs in AISI 4340 specimens, quenched and tem-
• From electroplating steels and other alloys exposed to hydro- pered to three different strength levels,
• From contact with water or other hydro- gen. It generally affects lower-strength al- were electrolytically charged with hydrogen
gen-containing liquids or gases loys. Fracture does not occur, as it does using an aqueous 4% H z S O 4 solution.
• From pure hydrogen gas with higher-strength alloys; these alloys Charging times were selected that lowered

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 712 / Service Characteristics of Carbon and Low-Alloy Steels

50 50
I
- - Uncharged

40
Uncharged
I 80 MPa strength level -
40
/
o~
35 30
/
~ 30 20 O/ ©
._~
g 25 1880 MPa strength level
10 ( ) / / / Hardness: 53 HRC
~ 20
I I I

C
r¢-
©'-----'-'----O ~ '

/
15 ' ©

10 50

~
- - Uncharged o....._d.---o-- -O--o--
5
[ 40 1585 MPa strength level
0 Hardness: 48 HRC
0 1 2 3 4 5 6 7 8
Charging time, min ~ 30
._=
I : | a 35 Effect of charging time on the tensile reduc- g
""b ° tion in area of AISI 4340 steel quenched and ~ 20
tempered to different strength levels. Source: Ref 237
n- • O - - - -

10
the tensile ductility, but not to 0. Figure 35
shows that with increasing charging time,
that is, with increased hydrogen pickup, the
reduction of area decreased. As the tensile ~V

Uncharged ~ ---O--
strength increased, the same degree of
charging had a greater influence on decreas- 40
ing the reduction in area. f 1380 MPa strength level
Hardness: 43 HRC
Hydrogen embrittlement, in classic form,
is reversible in that the damage should be 30
eliminated by removing the hydrogen. o.....
However, hydrogen can also cause perma- 20
nent damage. Figure 36 shows the recovery
of tensile reduction in area for AISI 4340
specimens charged for 5 min and then aged 10
for various times at room temperature be-
fore tensile testing (Ref 237). For all three
strength levels (the same strength levels in 0.01 0.1 1.0 10 100 103
Fig. 35), the tensile ductility returned to Aging time at room temperature, h
normal levels with time. About 100 h were Effect of aging time on the tensile ductility of heat-treated AISI 4340 at three different strength levels.
required; however, there was no clear trend Fig. 36 Source: Ref 237
in recovery time with strength level.
Figure 37 shows the combined effects of
reversible and permanent damage due to lower than that in the uncharged steel. The caused by hydrogen is difficult to separate,
hydrogen (Ref 252). The AISI 1020 carbon hydrogen-charged specimens that were particularly in higher-strength materials.
steel tensile specimens were tested both aged at 250 °C (480 °F) exhibited the same
uncharged and cathodically charged with high ductility as the uncharged specimens Hydrogen Stress Cracking
hydrogen, and each group was further test- down to - 1 6 0 °C (-255 °F), but at -196 °C and Loss of Tensile Ductility
ed after aging for 20 h at 250 °C (480 °F) in (-321 °F) they exhibited the same very low Another classic form of hydrogen embrit-
air. Tensile tests were performed over a percentage of reduction in area as the non- tlement, hydrogen stress cracking (also
wide range of temperatures, from - 1 9 6 °C aged charged specimens. Aging apparently called "static fatigue" or hydrogen-induced
(-321 °F) (liquid nitrogen temperature) to removed the hydrogen that caused the loss cracking) involves delayed fracture at
room temperature. The uncharged speci- of ductility from 20 to -160 °C (70 to -255 stresses below the yield strength. Figure 38
mens, with or without aging, exhibited a °F), but it did not remove the low ductility illustrates the nature of these failures for
gradual decrease in tensile reduction of area at -196 °C (-321 °F). Further testing of high-strength hydrogen-charged steels (Ref
with decreasing test temperature, as would aged, charged specimens where the surface 238). The notched tensile strength may be
be expected. The aging treatment had no layer was removed resulted in the same lower than for noncharged specimens. A
influence on the uncharged specimens. The -196 °C (-321 °F) ductility as the un- wide range of applied stresses can cause
charged specimens that were not aged ex- charged specimens. It was shown that ca- delayed fractures, with the applied stress
hibited much lower tensile ductility from thodic charging produced blisters and having only a minor influence on the time to
room temperature to - 1 6 0 °C (-255 °F), cracks at the surface (permanent damage) failure. Below some level of applied stress,
where they had about the same ductility as that made the charged specimens notch fracture does not occur. This critical level
the uncharged specimens. However, at the sensitive at -196 °C (-321 °F) whether of stress is referred to as the "static endur-
liquid nitrogen temperature ( - 1 9 6 °C, or hydrogen was present or not. The effects of ance limit." Additional work has demon-
-321 °F), the reduction in area was much combined reversible and permanent damage strated that the time between the incubation

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Fmbrittlement of Steels / 713

lF
Temperature, °F
-320 -280 -240 -200 -160 -120 -80 -40 0 40 80
N°tchten: ]
80 I I I 1 I I I I I [.~..~_,~per critical stress
I l \/'°cu, bat'°n time t
70 Z~
~[ ft~//F ractuc: t:~:!itical stress

60 f ,~

/ -
\
o~ 50 \ •
0.1 1 10 100 103
Fracture time, h
"~ 40 ~3 I ~ --~ ....
.o_ I "'" --~ .... ~ ........ Failure characteristics of a hydrogen-charged
s= I • t Fig. 38 high-strength steel. Source: Ref 238
~ 30
h- I
I
I 0 Uncharged was 1585 MPa (230 ksi) before charging and
20 I • Uncharged plus plating. For any aging (baking) time, the
I 20 h at 250 °C in air
I A Charged plus time to fracture increased slightly with de-
1011 20 h at 250 °C in air creasing load until the critical applied stress
• Charged at which fracture did not occur was
o I I t reached. Also, as the aging time increased,
-200 -180 -160 -140 -120 -100 -80 -60 -40 -20 0 20 40 the critical stress increased, and the times
Temperature, °C
to fracture at loads above this level in-
creased. With aging at 150 °C (300 °F), the
Fig. 3 7 Influence of test temperature and reversible versus permanent damage caused by hydrogen charging critical stress level increased from 515 MPa
on the tensile ductility of AISI 1020 steel, with and without aging. Source: Ref 252 (75 ksi) to about 1655 MPa (240 ksi), indi-
cating a recovery brought about by the
of cracking and failure decreases with de- (cathodically charged 5 min in aqueous 4% removal of hydrogen.
creasing test temperature until, at - 4 6 °C H2SO 4 and then cadmium plated, 25 min) Work has been done that demonstrates
(-50 °F), they are coincident (Ref 253). that were aged at 150 °C (300 °F) for various the combined influence of strain rate and
Resistivity measurements also showed that times and then loaded in tension until frac- test temperature on the tensile ductility of
crack growth was discontinuous. ture (Ref 238, 239). The notched tensile AISI/SAE 1020 carbon steel (Ref 254, 255).
Figure 39 shows test results for hydrogen- strength of the material was 2070 MPa (300 Specimens were tested in tension at strain
charged notched specimens of AISI 4340 ksi), and the smooth bar tensile strength rates from 0.05 to 19 000 min -~ over the
temperature range -196 to 100 °C (-320 to
212 °F). The test specimens were either
2200 i i spheroidize annealed before testing (for un-
Normal notch strength = 2070 MPa charged specimens) or cathodically
300 charged. Figure 40 shows the results of
2 0 0 0 ~ ~ ~. .,~....__0 I U~harged --
these tests. For the uncharged specimens
(Fig. 40a), the tensile ductility (natural log
275
of the original cross-sectional area, ao, to
the final cross-sectional area, af) decreased
250 to zero when the test temperature reached
-195 °C (-320 °F). The variation in strain
rate had little effect on the uncharged spec-
Bake18h 225
imens. For the charged specimens (Fig.
,4oo 200
40b), a marked reduction in tensile ductility
is evident, with decreasing strain rate in the
center of the test temperature range produc-
1200 ~ Bake 12 h 175 ~ ing two intersecting surfaces (coded a and d
in Fig. 40b). Surface a, in the lower-temper-
ature test region, shows that ductility de-
1000 150 creases with increasing temperature and
decreasing strain rate. Surface d, in the
800 ~V~ Bake 7 h .~ 125 higher-temperature test region, shows that
ductility decreases with decreasing temper-
Bake 3 h [~" ature and decreasing strain rate. As the
100
strain rate increases, the temperature for
6O0
Bake 0.5 h .L lowest ductility (the intersection of surfaces
75 a and d) increases from about - 4 0 to 50 °C
400 ~I (-40 to 120 °F).
0.01 0.1 1 10 100 103 Figure 41 shows the test results for
Fracture time, h charged specimens as a function of charging
Time to fracture as a function of applied stress and aging (baking) time at 150 °C (300 °F) for notched time at the lowest strain rate, 0.05 min -~.
I:|o 3 9
"'~° tensile specimens of AISI 4340 that were hydrogen charged and electroplated with cadmium. Source: The graph shows the loss in fracture ductil-
Ref 238 ity with increased charging time and the

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 714 / Service Characteristics of Carbon and Low-Alloy Steels

0!81 "Duc,ty 06

1 0 3 1 ~ ~ -

Strainrate, mii~ ~ (130) (0) Strain rate, mi~O

0.t -20 (200)
(0)
Temperature, °C (°F) -240 (-200) Temperature, °C (°F)
(-400) (-400)
(a) (b)
Tensile ductility of AISI/SAE 1020 carbon steel as a function of strain rate and test temperature for (a) spheroidize-annealed specimens and (b) cathodically
Fig, 4(1 charged specimens. Curve i bounds the range of strain rates and temperatures where embrittlement was observed. Source: Ref 255

increase in hydrogen content. The loss in midpoint specimen aged at 175 °C (350 °F) Hydrogen-induced fractures are not al-
ductility becomes relatively constant after for 15 h had a 64% reduction in area and ways intergranular. In high-strength steels,
about 30 min charging time. contained 1.4 ppm hydrogen. Fisheyes were cracking is often intergranular, but trans-
A correlation has also been demonstrated observed on the nonaged specimen fractures granular cleavage is also observed. For duc-
between total hydrogen content and tensile but not on those that were aged. Eight months tile low-strength steels, the fracture mode is
reduction in area (Ref 256). Test specimens later, additional tests were made with better ductile with changes in the dimple size.
were machined from the center, midpoint, percentage reduction in area values and less Prior temper embrittlement of high-strength
and edge of a 127 x 127 mm (5 x 5 in.) of a ductility gradient across the billet. Figure alloy steels enhances hydrogen stress
square billet of AISI 1010 carbon steel 42 shows the results from these two test cracking (Ref 257-261). The threshold stress
known to have a rather high hydrogen con- periods. The values on the graph indicated by intensity for crack growth is lowered, and
tent. Reduction in area results were 13.4, the solid dots were from similar tests made on intergranular fractures occur. The influence
23.0, and 33. !% with corresponding hydro- a 127 x 127 mm (5 x 5 in.) billet from another of temper embrittlement on hydrogen stress
gen contents of 5.7, 5.2, and 3 ppm. A heat of the same grade. cracking is particularly pronounced as the
yield strength is decreased. Very low impu-
rity content steels do show high threshold
1.6 ); 16 stress intensity values and freedom from
intergranular fracture (Ref 258).
Hydrogen-assisted ductile fractures have
1,4(F~k ,4 been reported for a variety of steels, chiefly

1.2 ~ 12
80
\

'\\
Nk~ Hydrogen o~
content o 7O
-~o 1.0 10 ~.
60
~ 0.8 8 o~
8 2 ¸ 50
Fracture
o 0.6 / ~ strain o~ .c_
if_ 6 .~ 40
O T
O ~ 3o
O \

%0
n-
0.4 D / 4
20
/
0.2 - 2

J
10

o,~ 0 1 2 3 4 5 6 7
0 0.1 1 10 100
Hydrogen content, ppm
Time, h
Iz;¢l Relationship between hydrogen content
I:;o 41 Fracture strain and hydrogen content of AISI/SAE 1020 steel as a function of charging time for tensile "15" 4 2 and tensile reduction in area for two heats
"b ° tests conducted at room temperature, with a strain rate of 0.05 min -1. Source: Ref 255 of AISI 1010 steel. Source: Ref 256

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 715

18.5 has been examined by several researchers.

/
- 0.72 One study determined the critical stress
18 intensity for slow crack growth in hydro-
- 0.70
gen-precharged 18% Ni maraging steels and
17.5
Wate~) for 300M alloy steel (Ref 273). The maraging
steel was tested at three yield strength lev-

~
- 0.68 els: 1750, 1880, and 2020 MPa (254,273, and
17 293 ksi); the 300M was tested at a yield
ied argon pul>"O-K>
:"O-~s - 0.66 strength of 1705 MPa (247 ksi). Specimens
E 16.5
E
equal volumes) were charged with different levels of hydro-
- 0.64 sz" gen, and the critical stress intensity for slow
16
Humidified argon crack growth was determined (Fig. 44).
-- 0.62 ~
Each of the four steels contained about 0.5
ppm H prior to charging, and the K~c values
o 15.5 (.5
for the noncharged specimens are plotted at
y Pure hydrogen -- 0.60 this value (dashed line). The data points for
15 the charged specimens are the critical stress
-- 0.58 intensities where slow crack growth began.
14.5 Note that the lowest-strength maraging
Humidified argon plus
oxygen (equal volumes) -- 0.56 specimens, with the highest K~c, had the
14
highest critical stress intensities after charg-
ing. The critical stress intensities for the
-- 0.54
13.5
two higher-strength maraging steels were
5 10 15 20 25 30 the same after charging to the same hydro-
Time, min gen levels despite their substantial differ-
ences in Kj~. The 300M alloy, with the
1:|o 43 Influence of oxygen on crack growth in AISI H l l tool steel quenched and tempered to a yield strength lowest yield strength (1705 MPa, or 245 ksi)
"e" of 1585 MPa (230 ksi). Source: Ref 272
and a Kic only slightly greater than the
highest yield strength (2020 MPa, or 292 ksi)
maraging grade, had much lower critical
spheroidized carbon steels and austenitic containing environments. Studies have stress intensities and was more susceptible
stainless steels (Ref 262-266). The fractures demonstrated that high-strength steels are to hydrogen embrittlement.
exhibit ductile dimples (microvoid coales- embrittled by exposure to hydrogen gas Similar tests were performed using 18%
cence), and the dimple sizes in charged even at, or below, atmospheric pressure Ni maraging steel at two different yield
specimens have been reported to be either (Ref 271). For example, a study was con- strengths: 1650 and 1915 MPa (239 and 278
smaller or larger than those in uncharged ducted on the effect of hydrogen gas at ksi) in high-purity hydrogen at pressures
specimens. The change in dimple size can atmospheric pressure on subcritical crack from 11.5 to 133.3 kPa (86 to 1000 torr) (Ref
be correlated with the loss in tensile reduc- growth using AISI H11 tool steel (Ref 272). 274). The crack growth rate as a function of
tion of area (Ref 262). Although argon, helium, and nitrogen are the crack tip stress intensity, K, was deter-
Austenitic Stainless Steels. For many inert environments, if they are humidified, mined for each steel as a function of tem-
years, it was believed that only bcc metals subcritical crack growth occurs. The study perature and hydrogen pressure. This infor-
were susceptible to hydrogen embrittle- also demonstrated that the addition of oxy- mation was used to determine the critical or
ment. However, it has since been shown gen to the environment, even in amounts as threshold stress intensity, Kth, at which
that austenitic stainless steels, and several low as 0.7% by volume, would terminate crack growth began. At K values above Kth,
other fcc metals and alloys, can become crack growth. The removal of the oxygen the crack growth increased (stage I). At
embrittled by hydrogen, and their behavior would restart crack growth. Oxygen dis- intermediate K values, crack growth be-
is similar in nature to the bcc metals (Ref solved in water, however, had no influence came relatively constant over a range of K
242,267). A number of studies have focused on the crack growth rate. Figure 43 shows values (stage II). At higher K values, crack
on hydrogen-embrittled AISI 304 or 304L. results for H11, quenched and tempered to growth became rapid (stage III) up to spec-
These studies have demonstrated that hy- a yield strength of 1585 MPa (230 ksi), using imen failure. Figure 45 shows the data for
drogen charging partially transforms the a precracked specimen under load in vari- the 1650 MPa (239 ksi) yield strength
austenite to or' (bcc) and e (hexagonal close- ous environments. For the first 5 min, the 18Ni(250) maraging steel tested in hydrogen
packed) martensites (Ref 268). These phas- crack grew slowly in humidified argon. at 133.3 kPa (1000 torr) for stage I and II
es can also be formed in 304 or 304L, and in Then, an equal volume of oxygen was add- crack growth. Over the temperature range
other low-stability austenitic stainless ed to the environment, and cracking of - 6 0 to 23 °C (-75 to 73 °F), the apparent
steels, by cold working (uncharged) speci- stopped for the next 5 min. Next, this threshold stress intensity, Kth , varies only
mens at low temperatures. Tests of more environment was removed and replaced by slightly from about 11 to 16.5 MPa V ~ (10
stable austenitic grades, such as AISI 310, pure hydrogen, and crack growth was very to 15 ksiV'i-n~.). No detectable crack growth
have produced conflicting results. While rapid. After about 1 min, the pure hydrogen was observed at 50 or 80 °C (120 or 175 °F)
some studies showed no embrittlement (Ref was replaced by humidified argon, and the for K levels up to 79 and 92 MPaX/m (72 and
269), others detected embrittlement. The crack growth rate decreased. Again, an 84 ksiX/~.), respectively. In the stage II
unstable 304 alloy exhibits transgranular equal volume of oxygen was added to the regime, where the crack growth rate (da/dt)
cleavage fractures, and the stable 310 alloy humidified argon, and crack growth ceased. was constant, the highest crack growth rate
exhibits ductile fractures, when tested in This atmosphere was removed and replaced occurred at 23 °C (73 °F). Figure 46 shows
hydrogen gas environmental cells (Ref 270). by water, and the crack again began to grow crack growth data for stage II crack growth
Tool Steels. Fracture mechanics ap- at a constant rate. as a function of test temperature and hydro-
proaches have also been used to study Maraging Steels. The influence of hydro- gen pressure. The upper curve shows data
crack nucleation and growth in hydrogen- gen on slow crack growth in maraging steels for hydrogen at 133.3 kPa (1000 torr). At

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 716 / Service Characteristics of Carbon and Low-Alloy Steels

90 i I 80 K, ksi i~Ji~. Temperature, °C

it,/ 0 10 20 30 40 50 60 40 20 0 -20 -40 -60
KIt data © HeatA 0.1 I I I I i ii ii 1
II J 1 0 -4

80 •
A
•
Heat B
Heat C
Heat D
- 70 23 °C
2 °C_
'~/m:~',,~~ 1°°°
B Pressure, terr

• 430
/" 210
-20 °C 10 3 86 10=3
70 ~ 60
-30 °C
E
E -42 °C ._E e3
80 ~- o.ol E f"
/f'4. . . . . . . E 10-5 ._E

,\
~-
-5o
.~ .~
-60 °C
5o
1 0 -4
.~ - 40 .-~
~ X ~ ~ ~ 10 4
40
1 0 -3
0 10 20 30 40 50 60 70
-30 ~E K, MPa Vm- 10 .6
o 30 o 3.0 3.4 3.8 4.2 4.6 5.0
Ic;. 4 5 Crack growth rates of 18Ni(250) maraging Temperature, 103/T, K 1
"b ° steel (1648 MPa, or239 ksi, yield strength) in
20 I 20 hydrogen at 133 kPa (1000 torr) as a function of test r.', Effect of test temperature on stage II crack
temperature and stress intensity range. Source: Ref H ~ . 4 6 growth rates for 18Ni(250) maraging steel in
274 gaseous hydrogen at different pressures (133, 57, 28,
I k~_ 10 and 11.5 kPa, or 1000, 430, 210, and 86 torr). Source:
10 I Ref 274
I
I
temperature than it does for the higher-
strength grade. Temperature, °C
0 I 0
0 1 2 3 4 Line Pipe Steels. Blisters form in steels 60 40 20 0 -20 -40 -60
Hydrogen, ppm when atomic hydrogen enters the steel from IE 90 ~ l I [ I I
the surrounding environment. The hydro- ½
i

s0 t¢
I : | . 4 4 Influence of hydrogen content on the crit- gen diffuses through the steel to an internal ~; 80 -- PH2 = 133 kPa _ 5
"~" ical stress at and below which cracks do not defect or inclusion, where it builds up as 70 ~
grow. Alloys A, B, and C are three different 18% Ni
maraging steels at increasing levels of yield strength molecular hydrogen. This process produces 70 "6
(1740, 1870, and 2020 MPa, or 252, 271, and 293 ksi); high pressure, which causes a blister to 60
alloy D is 300M alloy steel at 1705 MPa (247 ksi). form. The blister will grow as more hydro- '~ 6 0 - -
Source: Ref 273 o) 50 ~
gen diffuses to it, deforming the surround-
ing metal until it ruptures. Blister formation .E_' ~ 50 ~ O
r~- I / I18Ni(200)
40 ~
is observed in environments containing hy- 40
low temperatures, the crack growth rate drogen sulfide or where hydrogen is electro- 30 ~
increased with increasing temperature (re- lytically charged. Blister formation is com-
gion A) until a temperature range was mon in the more ductile steels, such as sour --20 ~
reached where the crack growth did not gas transmission line pipe steels (Ref 276- 20 18Ni(250) g
vary much with temperature (region B), and 278). ~ lO lO o_
the maximum crack growth rate was ob- Hydrogen-induced blistering of line pipe ,<
tained. With further increases in tempera- steels results from contact with wet envi- 0
ture, the crack growth rates decrease rapid- 3.0 3.4 3.8 4.2 4.6 5.0 5.4
ronments containing HzS. Hydrogen diffus-
Temperature, 103/7, K-~
ly (region C). Lowering the hydrogen es to inclusions and precipitates at the
pressure shifts the curves to lower crack inclusion/matrix interface. A blister-crack Effect of test temperature on the apparent
growth rates, and the temperature for the array forms; cracking of the metal around Fig. 47 threshold stress intensity factor for two
maraging steels tested in gaseous hydrogen at 133 kPa
maximum crack growth rate appears to de- the blisters is caused by the molecular hy- (1000 torr). Source: Ref 275
crease slightly. drogen pressure and applied or residual
Another study also tested 18% Ni marag- stresses. These crack arrays are internal,
ing steels at two different yield strengths rather than at the surface. Elongated man- (Ref 279-289). Flakes are caused by an
(18Ni(200) at 1270 MPa, or 184 ksi, and ganese sulfides are the most common nucle- excessive hydrogen content, that is, hydro-
18Ni(250) at 1650 MPa, or 239 ksi) in low- ation sites, but cracking can also occur at gen above some safe level or threshold
pressure hydrogen (Ref 275). Figure 47 manganese silicates and at large complex level. This safe level varies somewhat with
shows a plot of the apparent threshold carbonitrides. steel composition, section thickness, cool-
stress intensity level for the onset of slow It is well known that high sulfur contents ing rate, inclusion content, and segregation
crack growth, Kth, for these two grades in are highly detrimental for sour gas applica- and is generally in the range of 1 to 2 ppm.
hydrogen at 133 kPa (1000 torr) and over the tions. High-quality line pipe steels require Some steel compositions, particularly low-
temperature range of - 6 0 to 60 °C ( - 7 5 to very careful processing, which involves carbon steels, appear to be rather free of
140 °F). The K,h is quite low at low temper- desulfurization to very low levels, sulfide flaking problems.
atures (stage I.I) and increases slowly with shape control, controlled rolling, grain re- The solubility of hydrogen in liquid steel
increasing temperature until a temperature finement, and segregation control. decreases substantially when it solidifies (to
is obtained at which Kth increases rapidly Formation of Flakes in Steels. Flaking or about 7 ppm in ~ iron). With continued
and Klc is reached. The lower-strength steel hairline cracking in steels, particularly in cooling, it decreases to about 5 ppm and
has higher Kth values at low temperatures, forgings, heavy-section alloy plate steels, then at the ~ to ~ transformation it abruptly
and its critical temperature (where Kth in- and railroad rails, has been an ongoing drops to about 2.5 ppm. With continued
creases rapidly) occurs at a much lower problem since the beginning of steelmaking cooling, it decreases to less than 1 ppm at

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 717

700

600 - -
10
I

Slag
cooled
t
Cooling rate from 900 to 400 °F, °F/h

Hood
cooled
100
I

Air
cooled
I
103

/
/
/
I

Water
quenched
104
I
400

3O0
/
Air cooled
"~ 5.6 °C/h 22 °C/h 86 °C/h / 600 °C/h
-- 500-- ; ,," - _ Hood cooled
-- - - -
m Slag cooled

~_ 400 200 ~=~

m
to
o
# o
P
"6 300 .g /
/ T ~3"5 ppm H2 E Ni-Mo-V ~
/
E 200 // \\ 1 1"12,ppm 100
~" ~ .,> . . . . I 1080 / 4.5
/ .... ,o-v I \ //40
20
100 e,,,,,,~ \ ~ ~ -- Zx~_ Z Z --'//4"0 1045
~_.~ ~ ~- -- / I~ 3:8: Ni-Cr-Mo-V 0~
10 100 10 3 10 4 0 1 2 3 4 5 6
Hydrogen content, ppm
Cooling rate from 480 to 205 °C, °C/h
Effect of hydrogen content and cooling rate
Effect of cooling rate after hot working on the number of flakes in etch disks of AISI 1045 and 1080 Fig, 4 9 after hot working on the number of flakes
Fig. 4 8 carbon steels and nickel-molybdenum-vanadium and nickel-chromium-molybdenum-vanadium alloy
in etch disks of AISI 1080 carbon steel and nickel-
steels. Source: Ref 287 molybdenum-vanadium alloy steel. Source: Ref 287

room temperature. Some of the hydrogen from each block, macroetched, and exam-
present in excess of the solubility limit at ined. The total number of flakes were vanadium steels (indicated by the lower
room temperature can be accommodated at counted, except for disks that were quench curve for nickel-molybdenum-vanadium).
microstructural traps, such as sulfide inclu- cracked. Hydrogen analyses of the blocks F o r the 1080 steel, the flaking index was 8 at
sions, and rendered harmless. However, if were made immediately after cooling to a 4.0 ppm; at 3.7 ppm it flaked only when
the hydrogen content is too high, cracking safe handling temperature. water quenched, and at 2.2 ppm it did not
will occur after a certain delay time, usually The study showed that the number of flake even when water quenched. These
less than 2 weeks, unless the steel is reheat- flakes increased with the cooling rate (Ref results for nickel-molybdenum-vanadium
ed to a temperature where it can safely 287). The flakes were generally located in a and 1080 as a function of hydrogen content
diffuse out of the steel. In the years prior to circumferential pattern between the surface and cooling rate are better shown in Fig. 49,
the development of vacuum degassing (even and the center of the block and as the which clearly demonstrates the presence of
today for some steels, such as rails), cooling rate decreased, they were closer to a critical threshold level of hydrogen in
wrought or forged products susceptible to the center. In general, as the number of these flake-sensitive steels. The threshold
flaking were held at temperatures between flakes decreased, their size increased. levels were approximately 3.5 ppm for 1045
about 500 and 650 °C (930 and 1200 °F) for a Figure 48 shows data for the relationship and nickel-chromium-molybdenum-vana-
time that depended on the product thick- between the number of flakes on the etch dium, 3.3 ppm for 1080, and 1.6 ppm for
ness; this was done before the products disks as a function of the cooling rate and nickel-molybdenum-vanadium. The high
were cooled to ambient temperature, and it hydrogen content for each grade. Quench susceptibility of the nickel-molybdenum-va-
allowed the hydrogen content to decrease to cracking, which occurred in some of the nadium steel was attributed to the high
a safe level. With the drive to produce steels water-quenched blocks, made the analysis transformation stresses associated with the
with lower sulfur contents, rail producers more difficult by requiring the use of an formation of martensite from enriched aus-
have discovered that these safe controlled- estimated value of the number of flakes tenite; the high susceptibility of 1080 was
cooling practices have had to be extended (dashed lines). In general, water quenching attributed to its low fracture toughness.
for further hydrogen removal because of the produced about twice as many flakes as did
decrease in hydrogen traps available. Simi- air cooling. The 1045 and nickel-chromium- Metal-Induced Embrittlement
lar flaking problems have occurred in low- molybdenum-vanadium steels had low sus-
sulfur plate steels that were free of such ceptibilities to flaking at hydrogen levels of The term metal-induced embrittlement
problems at the former higher-sulfur levels. 4 and 3.8 ppm. The 1080 and nickel-molyb- (MIE) applies to cases in which contact
An interesting study was done of flake denum-vanadium steels were much more with a liquid metal embrittles a solid metal
formation in plain carbon steels (AISI 1045 susceptible at hydrogen levels of 4.0 and 3.5 (liquid metal embrittlement, LME) or in
and 1080) and in nickel-molybdenum-vana- ppm. Based on the number of flakes, a which contact with a solid metal somewhat
dium and nickel-chromium-molybdenum- relative susceptibility of 1 was assigned for below its melting point (solid metal embrit-
vanadium forging steels melted and teemed 1045 and nickel-chromium-molybdenum- tlement, SME) embrittles a solid metal.
to produce a range of hydrogen contents vanadium, and values of 8 and 75 were Only recently has a distinction been made
(Ref 287). Five ingots of each grade were assigned for air-cooled blocks of 1080 and between these two forms of MIE. Histori-
produced and rolled to 305 mm (12 in.) nickel-molybdenum-vanadium. The figure cally, L M E was the term used to describe
square blooms, and each was hot sheared to also shows the influence of hydrogen con- all MIE problems.
produce four 610 mm (24 in.) long sections. tent on the nickel-molybdenum-vanadium
These sections were cooled at different steel blocks with hydrogen contents of 3.5 Liquid Metal Embrittlement
rates and by different methods (water and 2.0 ppm. Reducing the hydrogen con- Liquid metal embrittlement is a phenom-
quench, air cool, hood cool, and slag cool). tent of the steel from 3.5 to 2.0 reduced its enon in which the ductility or fracture stress
After a 2-week incubation period, to permit susceptibility to a level similar to that of the of a solid metal is reduced by surface con-
flaking to occur, transverse disks were cut 1045 and nickel-chromium-molybdenum- tact with a liquid metal. The first recorded

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 718 / Service Characteristics of Carbon and Low-Alloy Steels

recognition of this problem was made in with the form that causes instantaneous Tensile stresses, either applied or residual,
1914 when the embrittlement of [3 brass by failure of the solid metal, the most common are required, while the solid metal is wet by
mercury was observed (Ref 290). Since form. This type of L M E has been referred the liquid metal. Again, a certain critical
then, L M E has been identified as the cause to as adsorption-induced embrittlement. stress (threshold) can be identified above
of numerous failures, and numerous re- Liquid metal embrittlement is often consid- which cracking occurs and below which it
views of L M E have been published (Ref ered a special type of brittle fracture, and does not. Subsequent studies focused on
291-301). fracture mechanics approaches have been the identification of specific liquid-solid
There are at least four distinct forms of applied to its analysis. metal couples that undergo LME.
liquid metal embrittlement (Ref 292): Many different metals and alloys besides Studies of specific L M E couples have
ferrous alloys can fail because of LME. empirically identified certain trends that are
• Type 1 : Instantaneous fracture of a metal Only specific liquid metals are known to usually, but not always, obeyed. For exam-
under an applied or residual tensile stress cause LME of each specific metal or alloy. ple, most L M E couples have very little
when in contact with certain liquid met- Indeed, this is one of the more perplexing mutual solid solubility and exist as immis-
als. This is the most common type of aspects of LME, and some investigators cible liquids in the liquid state. Solid metals
LME refer to it as specificity (Ref 302). However, that are highly soluble in the liquid metal
• Type 2 : Delayed failure of a metal when studies have generally used tests involving a and solid-liquid metal combinations that
in contact with a specific liquid metal limited number of conditions, for example, form intermetallic compounds are usually
after a certain time interval at a static load a single temperature, applied stress, grain immune to LME. Such studies have dem-
below the ultimate tensile stress of the size, and so on, and until a much wider onstrated that good wetting of the substrate
metal. This form involves grain-boundary range of variables has been tested, the va- is required. It is interesting that the factors
penetration by the liquid metal and is less lidity of this concept cannot be proved. that promote the lowest interfacial energy
common than type 1 Also, many L M E tests have been of the and therefore the best wetting are high
• Type 3: Grain-boundary penetration of a go/no go type rather than quantitative mea- mutual solid solubility and intermetallic
solid metal by a specific liquid metal, surements of the degree of embrittlement. compound formation, both of which pro-
which causes the solid metal to eventual- Certain conditions are recognized as be- mote low susceptibility to LME. A number
ly disintegrate. Stress does not appear to ing required for, but not necessarily suffi- of studies have examined the effect of add-
be a prerequisite for this type of L M E in cient to cause, embrittlement. These condi- ing various solute elements to particular
all observed cases tions include (Ref 291): liquid metals, but no clear trends have been
• Type 4 : High-temperature corrosion of a identified.
solid metal by a liquid metal, causing • Adequate wetting of the solid-metal sub-
Several researchers have attempted to
embrittlement, an entirely different prob- strate by the liquid metal
correlate the occurrence and severity of
lem from types 1 to 3 • An applied or residual stress present in
the solid metal L M E for particular couples with the elec-
In some respects, L M E is similar to stress- tronegativities of the metals. For some cou-
• A barrier to plastic flow existing at some
corrosion cracking (SCC), which always ples, maximum embrittlement arises when
point in the solid metal that is in contact
requires a measurable incubation period be- the solid and liquid metals have similar
with the liquid metal
fore fracture occurs. Stress-corrosion electronegativities. As the difference in
cracking is more similar to type 2 L M E than Of course, for failure to occur, the liquid electronegativity increases, the degree of
metal must be of the correct composition to embrittlement decreases. While many ex-
to the other types. This review will concen-
cause L M E of the solid metal. amples support this relationship, exceptions
trate on the influence of MIE on the me-
If the solid metal is not notch sensitive, as also exist (Ref 294).
chanical properties of ferrous alloys.
in the case of fcc metals, the crack will The presence of notches (stress concen-
A number of observations have been
propagate only when the liquid metal feeds trators) increases the severity of LME, and
made of interactions between liquid metal
the crack. However, for a notch-sensitive this has prompted the use of fracture me-
and solid metal. These observations can be
classified as (Ref 291): metal, such as bcc metals Like iron, the chanics concepts in the study of LME. As
nucleated crack may become unstable and stress raisers are well known to be a detri-
• No observed interaction propagate ahead of the liquid metal. Crack mental influence in brittle fracture, their
• Dissolution of the solid metal into the propagation can be extremely rapid; rates effect on L M E is understandable.
liquid metal. Diffusion of the liquid metal between 500 and 5000 mm/s (20 and 200 Reference 301 reports the results of a
into the solid metal in./s) have been recorded (Ref 293). In most very extensive literature survey concerning
• Formation of intermetallic compounds at cases, crack paths are intergranular, but the nature and occurrence of LME. The
the liquid/solid interface transgranular cracking has been observed, following liquid metals have been reported
• Intergranular penetration of the liquid and a few cases of ductile fracture have to cause L M E of ferrous alloys: aluminum
metal into the solid metal without the even been reported. Factors that influence (minor L M E for pure iron, no LME for a
presence of applied or residual stress the fracture stress and ductility of metals 0.2% C steel), antimony, bismuth (both
• Brittle premature failure of the solid met- subject to L M E include: positive and negative results have been re-
al caused by intergranular penetration of ported), cadmium, copper, gallium (both
• Composition of the solid and liquid met-
the liquid metal into the solid metal under positive and negative results have been re-
als
the influence of applied or residual tensile • Temperature ported), indium, lead, lithium, mercury, so-
stress. Failure occurs after a finite incu- dium, tellurium, tin, and zinc. Solders,
• Strain rate
bation period • Grain size bearing metals, and brazing alloys have also
• Brittle instantaneous failure of a stressed been shown to produce LME in certain
• Thermal-mechanical history of the solid
metal when wetted by a specific liquid metal tested ferrous alloys. As might be expected,
metal. Grain boundary penetration is not for each of the above-mentioned liquid met-
necessarily observed Some of the earliest studies of L M E did not als only a limited range of solid ferrous
recognize the important role of stress in alloys has been evaluated. Liquid metals
The last three observed interactions are the L M E (Ref 303). Some early studies, how- that have not been found to cause LME of
prime forms of liquid metal embrittlement. ever, did demonstrate the significance of ferrous alloys include cesium, rubidium,
Most studies of L M E have been concerned stress in producing L M E (Ref 304-306). selenium, silver (although silver-containing

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement o f Steels / 7 1 9

50 17OO shows test results for AISI 4130 steel at 44

#.
300
=
ILII II II I 45 1600
(4
1
- 240
HRC immersed in molten lithium at 200 °C
(390 °F). As the applied stress decreased
4o ~, 1500 / - 220
~; 250 3~ g from 310 to 138 MPa (45 to 20 ksi), the time
200
11111111 30 ~
1400
1300
~/~ 200
to failure increased from about 1 s to 1 to
200 min.
25 ~ 1200 ~ 180
The strength of the material also influenc-
150 20 ~.
~- lfOO .4 ~/ 16o es L M E results. Figure 51 demonstrates
15 <
< 100 this by showing the normal tensile strength
I I11
d~
1000 ,~ ~ . . ~ ~,~ 140 ~:~
of AISI 4130 alloy steel quenched and tem-
To 900
pered to hardnesses from 20 to 55 HRC and
0.01 0.1 1 10 100 1000 o~ 800
Time to fracture, mm
the corresponding fracture strength of spec-
imens wet with liquid lithium at 205 °C (400
Time to fracture as a function of applied 600 k"~ ~ k S0 °F). At tensile strengths above 1035 MPa
H g , 50 load for AISI 4130 alloy steel specimens (at
44 HRC) wetted with molten lithium at 200 °C (390 °F). (150 ksi), catastrophic embrittlement occurs
500 ~ .'~"~ ,k - 60
Source: Ref 291 400 ~1.~\ (Ref 291). Tensile ductility, however, was
reduced at all of the strength levels.
300 40
While the majority of LME studies have
brazing alloys do cause LME), and thal- 200 been conducted using tensile specimens,
lium. Reference 301 contains source data O Ultimate tensile strengt ;20
100 • Wetted fracture strength ,.."~ some work has been done using fatigue load-
for these observations and should be con- 0 0 ing conditions. Figure 52 shows axial load
sulted for further information. 20 25 30 35 40 45 50 55
fatigue tests of AISI 4340 with a tensile
The effect of test temperature, T, is com- Hardness, HRC
strength of 1310 MPa (190 ksi) tested at a
plex and is not predictable. Generally, tem- -,~s""", 51 Plot of hardness versus tensile strength for stress ratio of 0.75 with or without a coating
perature increases above the melting point AISI 4130 alloy steel with and without wet- with liquid mercury (Ref 291). As the maxi-
of the embrittling liquid decrease the degree ting by molten lithium at 205 °C (400 °F). Source: Ref
mum applied stress decreased, the coated
of embrittlement. Maximum embrittlement 291
specimens exhibited a proportionally lower
often occurs near the melting point of the fatigue life than did the uncoated specimens.
liquid, Tin. In many couples, embrittlement Fatigue crack growth rate measurements
occurs below the melting point of the liquid. °F), which is well above the melting point of
were performed using liquid lithium and
For example, the embrittlement of AISI zinc (419 °C, or 786 °F) (Ref 308).
AISI 304L stainless steel at temperatures
4140 alloy steel starts at temperatures below Liquid metal embrittlement reduces ten-
ranging from 473 to 973 K (200 to 700 °C)
the melting point for cadmium (T/Tm= 0.75) sile ductility and stress at fracture if the (Ref 309). Figure 53 shows the crack growth
and for lead and tin (T/Tm= 0.85) (Ref 307). failure occurs at a stress below the usual data from these tests as a function of tem-
This phenomenon is called solid metal em- yield strength of the metal. The time to
perature for 304L specimens in argon and in
brittlement or solid metal induced embrit- fracture is usually directly related to the liquid lithium using two different test fre-
tlement. In other cases, the liquid tempera- level of the applied stress. At some high quencies and a stress intensity range of 14
ture must be above Tm, as for the level of applied stress, fracture can occur
MPa~/-m (12.7 ksi~/i-m.). The crack growth
embrittlement of austenitic stainless steel immediately after the specimen is wetted by
rates in lithium were much greater than
by zinc. It has been reported that the zinc the liquid metal. As the level of applied those in argon at all temperatures and load
embrittlement of austenitic stainless steels stress is reduced, the time to failure in-
frequencies. In lithium, the crack growth
has not been observed below 750 °C (1380 creases. This is shown in Fig. 50, which
rate decreased as the temperature increased
from 473 to about 700 K (200 to 700 °C) and
then increased strongly with further in-
1000 creases in temperature.
o -- 140 Crack propagation in argon was trans-
granular for all test conditions. At 573 K
Uncoated (300 °C) in lithium, a mixture of transgran-
ular cleavage and intergranular fracture was
90O 13o
Z; observed that varied with load frequency
and stress intensity range (Fig. 54). The
O percentage of intergranular fracture in-
--120 creased with decreasing stress intensity
O-~ range and increasing frequency. The influ-
~ 8o0
/ O-~
ence of load frequency at 573 and 873 K
Coated -- 110 (300 and 600 °C) is shown in Fig. 55. The
with mercury effect of frequency on the amount of inter-
E granular fracture at 873 K (600 °C) is oppo-
700 site to that at 573 K (300 °C). This difference
O-~ -- 100
was also accompanied by a difference in the
appearance of the intergranular facets; at
873 K (600 °C), the facets were rounded,
- 90 with numerous dimples; at 573 K (300 °C),
600 the facets were sharp, and occasional tears
104 105 106 107 and slip bands were observed.
Number of cycles to failure The liquid metal embrittlement failures of
.,~:"". 52 Fatigue life under axial loading for AISI 4340 alloy steel (1310 MPa, or 190 ksi, tensile strength) wetted steels caused by molten copper and copper-
with mercury. Source: Ref 291 base alloys are common, and numerous

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 720 / Service Characteristics of Carbon and Low-Alloy Steels

Temperature, °C Stress intensity factor range (AK), ksi i~T~.

200 300 400 500 000 700 10 12 14 16 18 20
10 e I I I I I I 1.0
I I I

Loading frequency
-- 10-6 O 20 Hz
0.8 • 5Hz
--~ O • 20 Hz A1Hz
~ & 0.2 Hz -~
"~ 10-' ~) • o
--~ 0.6
/ -B

$ O --10 7~
"
-~ O Lithium O~--'~ ~ 0.4
A
10-8

Argon
0.2 ~ ~ A
. O
--10 e

10-9
0 • ~ "_~,"~ ~-"
400 500 600 700 800 900 1000 10 12 14 16 18 20 22
Temperature, K
Stress intensity factor range (AK), MPa~m
Influence of test temperature and environ-
I:|,r, 53 Influence of loading frequency and stress intensity factor range on the percentage of intergranular
''b ° ment (argon versus molten lithium) on the
crack growth rate using two different loading frequen-
Fig. 54 fracture for AISI 304L in molten lithium at 573 K (300 °C). Source: Ref 309
cies and a stress intensity factor range of 14 MPaV'm
(12.7 ksiN/]-m.) for AISI 304L stainless steel. Source: Ref
3O9 notched tensile creep specimens at 1100 to strain field in the heat-affected zone while
1130 °C (2010 to 2065 °F) and stresses of 8.3 the steel was austenitic. Heat affected zone
studies have been made. Such failures are and 11.0 MPa (1.2 and 1.6 ksi) (Ref 312- areas that remained ferritic were not pene-
characterized by grain-boundary penetra- 315). Liquid copper significantly altered the trated (Ref 317). In another study involving
tion of liquid copper, which can be quite creep behavior of pure iron, causing prema- copper deposition by a gas metal arc, alloy
rapid. One of the earliest studies of L M E of ture failure. Embrittlement was of the de- steels such as AISI 4340, 4140, and 304
steel, in this case by molten brass, was layed type controlled by grain-boundary were found to be extremely sensitive to
made in 1927 (Ref 310). An evaluation of the copper penetration. The depth of surface copper penetration. In comparison, pene-
influence of various liquid metals on plain cracking was controlled by the depth of tration depths for AISI 1340, 1050, Armco
carbon steel, silicon steel, and chromium copper penetration. Dihedral-angle mea- iron, and carburized Armco iron were only
steel found that these steels were embrittled surements for liquid copper in the steel about one-third that of the alloy steels. A
at 1000 to 1200 °C (1830 to 2190 °F) by liquid grain boundaries revealed that 34 ° was the ferritic stainless steel, AISI 430, was almost
tin, zinc, antimony, copper, 5% tin-bronze, most frequent angle for exposure at 1100 or completely immune to copper LME. The
and 10% zinc-brass; liquid bismuth, cadmi- 1130 °C (2010 or 2065 °F). study concluded that, when stresses were
um, lead, and silver caused little or no Liquid metal embrittlement of steels by absent, the ease of penetration by molten
embrittlement (Ref 311). copper has also been encountered during copper was a function of the alloy content
A series of studies of L M E of iron by welding and joining processes. An examina- of the steel. Grain-boundary copper pene-
liquid copper was done using tests per- tion of copper weldment deposition on fer- tration occurred under an applied stress,
formed under nonoxidizing conditions with ritic and austenitic stainless steels in duplex forming partially filled or open cracks in the
ferrite-austenite grades found that stable steel. Steels that remained ferritic at the
1.0
ferrite reduced the penetration of copper; if melting point of copper were not penetrated
ferrite was in excess of 30%, penetration did (Ref 318).
~ 0.8 not occur. Also, copper deposited on fully As reviewed earlier in the section "Alu-
.z-~ o ~ (60O °C)
ferritic grades did not cause cracking, but it minum Nitride Embrittlement," residual
~ 0.6 r v did in austenitic grades. Wetting by molten copper can detrimentally influence the hot
copper was also examined. A contact (dihe- workability of steel by internal LME. Ref-
c~ 0.4 dral) angle of 92 to 100° was observed for erence 321 summarizes work on the influ-
ferritic grades (no wetting of the grain ence of copper on the hot workability and
c 0.2 boundaries) and 22 to 28 ° for the austenitic scaling characteristics of steels. Elements
grades (wetting) at 1100 °C (2010 °F) (Ref such as tin, arsenic, and antimony dissolved
0.01 0.1 1 10 100
316). in the steel increase the sensitivity of sur-
Loading frequency, Hz A series of studies has examined LME of face cracking that is due to copper.
steels by copper during welding (Ref 317- Perhaps the most c o m m o n LME failure
I:;,,T
55Influence of loading frequency and temper- 320). A study of HY-80 steel welded with or is the fracture of railroad axles caused by
"~" ature of lithium on the percentage of inter-
granular fracture for AISI 304L. Stress intensity factor
without copper-nickel filler metal found in- c o p p e r penetration from overheated fric-
range, ,3K, was 14 MPaV'm (12.7 k s i ~ . ) . Source: Ref filtration of the grain boundaries by the tion bearings. These failures are relatively
3O9 copper-nickel deposit under an applied common, even t o d a y , and have been ob-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 721

Test temperature, °F point of cadmium (321 °C, or 610 °F). Crack

200 400 600 800
propagation rates increase with temperature
(below the melting point, Tm), and a thresh-
I old stress level is required for cracking (Ref
328).
1.0
\\ \
Leaded Alloy Steels. Solid metal embrittle-
,\
ment of leaded alloy steels, that is, internal
rather than external embrittlement, has
............. " -". been studied extensively (Ref 329-333). Ref-
uO
erence 329 contains the first discussion of
the differences between liquid metal and
"0 solid metal embrittlement. Short-time high-
temperature tensile specimens of AISI 4140
CD
~@ 0.8
c
0
c
/' and 41L40 (leaded) were tested between
room temperature and approximately 480
/, °C (900 °F). The melting point of lead is 327
/ ! °C (621 °F). Tensile test results for the
._c
/ /] leaded and nonleaded steels were com-

~ 0.6 \\ \ / //I pared. Figure 56 shows the ratio of the

percentage reduction in area of the leaded
steel to the unleaded steel for tensile
strength levels from 825 to 1515 MPa (120 to
220 ksi) and temperatures from ambient to
approximately 480 °C (900 °F). The ductility
of the leaded steels was substantially re-
duced at temperatures well below the melt-
"6 0.4 ing point of lead. The loss in ductility in-
creased with tensile strength and was
" il
\
.c greatest at temperatures just above the
/I
g
melting point of lead. By 480 °C (900 °F),
most of the ductility has been recovered.
rr

0,2
/i Similar tests of leaded and nonleaded
Charpy V-notch impact specimens revealed
that embrittlement effects were apparent
, /i beginning at about 28 °C (50 °F) below the
melting point, suggesting a sensitivity to
\"'~ ....... "'/: I strain rate. Also, when comparing the ulti-
mate tensile strength for leaded and non-
leaded specimens as a function of hardness
at the maximum embrittlement temperature
0
0 100 200 300 400 500
(370 °C, or 700 °F), results were quite sim-
ilar to those shown in Ref 291 for AISI 4130
Test temperature, °C
in liquid lithium at 205 °C (400 °F) (Fig. 51).
Relative decrease in the percent reduction of area for leaded versus nonleaded tensile specimens of Catastrophic failure occurred for leaded
H ~ , 56 AISI 4145 alloy steel as a function of test temperature and ultimate tensile strength (825 to 1655 MPa, steels at tensile strengths above 965 to 1140
or 120 to 240 ksi). Source: Ref 329
MPa (140 to 165 ksi).
A fractographic examination was done of
these test fractures (Ref 330). Five types of
served ever since the development of rail- Solid Metal Embrittlement fractures were observed that related to the
roads. A friction bearing consists of a test temperature. Tensile specimens broken
bronze shell lined with babbit metal bond- Solid metal embrittlement occurs when at temperatures up to 205 °C (400 °F) exhib-
ed to the inner surface. The bearing is the temperature is somewhat below the ited the classic cup-and-cone fracture and
lubricated through an opening in the shell, melting point of the embrittling metal (Ref failed by microvoid coalescence (ductile).
which is packed with cotton waste. Oil 297, 300, 307, 324, 326). Its characteristics, Specimens tested between 205 and 315 °C
from a reservoir is carried to the bearing except for this distinction, are quite similar (400 and 600 °F) exhibited small shear lips
through the cotton waste by a wicklike to that of LME. Crack propagation rates for and a mixture of intergranular fracture re-
mechanism. In 1914, a study presented SME, however, are much slower than for gions within a transgranular fracture. This
results for a Krupp axle that failed in LME, and multiple cracking is often ob- was the range where embrittlement (SME
service (Ref 322). The surface of the axle served in SME, but not in LME. only) began and became progressively
had been exposed to very high tempera- Iron-Base Alloys. Solid metal embrittle- worse. Tensiles broken between 315 and
tures, as evidenced by the microstructural ment of iron-base alloys can be caused by a 345 °C (600 and 650 °F) showed severe
changes, and cracking was present. The number of low melting point metals. Solid embrittlement and numerous coarse facets
heated region was the part of the axle metal embrittlement of steels by cadmium (called "fisheyes" by the authors of the
within the support bearing. Bronze bearing (Ref 307,326, 327) was studied in depth as a study). Fracture was a mixture of transgran-
metal was observed to be associated with result of the fracture of cadmium-plated ular and intergranular modes. Tensile spec-
the cracks and grain-boundary penetra- nuts (Ref 327). The failures occurred in the imens broken between 345 and 480 °C (650
tion. Such failures still occur and have temperature range of 200 to 300 °C (390 to and 900 °F), covering the region of maxi-
recently been reviewed (Ref 323). 570 °F), which is well below the melting mum embrittlement to recovery, exhibited

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 722 / Service Characteristics of Carbon and Low-Alloy Steels

Test temperature, °F Test temperature, °F a weight percent basis, it enhanced lead

200 400 600 800 1000 200 400 600 800 1000 embrittlement less than antimony addi-
80 I ~ I I 2200 I I 320
tions did.
/
~ ~\ ii,! I
60
2000 0 " " - ---'-'- = Combined SME and LME has been dem-
40 ~C.~. ___04 J Pure lead --280 onstrated for AIS14140 alloy steel (1380 MPa,
20 1800 0.004% Sb
or 200 ksi, tensile strength) in contact with
0 1600 - 240 pure zinc, lead, cadmium, tin, and indium
80 ? I I A over the temperature range from ambient to
6o I I - the melting point of the pure, low melting
40 0.004%s0 1800 point alloy (Ref 307). Figure 58 shows test
20 0.01%s0 240 results for 4140 in contact with these metals.
0 1600
All of the data have been normalized for ease
60~,-.-__.<>.~
40 N/0.01% Sb 1800 --
of comparison. Temperature is expressed in
homologous fashion (that is, by the ratio of
the test temperature used to the melting point

,80 -..\,..
20 ~ ~'~1 1600 ~, ~ - 240
of the embrittler):
~ o ~ X Jo.1%sbj T
TH = - - (Eq 4)
---
60 ~ . m C k ~ 1800
..~
i rm
40 0.1% Sb ~ 1600 - 240
where TH is the homologous temperature.
8o2°
o %__ I~ I 1800
" ~ X,,_ .3% Sb
The true fracture strength (TFS) ratio is the
ratio of the true fracture stress of the em-
60 ~ I1~ - brittlement couple to the true fracture stress
of uncoated 4140 at the same temperature;
[ ~'.~.u_~ O-----~
40 " 4 ~ ~.rk... ~ 0 . 3 % Sb 1600 I~'~ 240
the reduction in area (RA) ratio is defined in
20 280 the same way, based on the percentage
0 1800 reduction in area values. As the melting
80 I I It
60 ~ --. 1600 ~,i ~ Sb 240 point of the embrittler is approached, the
40
20 .... 1%Sb 1400 • - 200
TFS ratio and RA ratio decrease. For tem-
peratures at the melting point of the embrit-
tier, RA ratios vary from 0 to approximately
0 ~ ~ 1200
80 I [ ~ 160 0.2, indicating severe embrittlement. All of
60 ~ 1000 ,,,,, the elements in this test were potent embrit-
tiers when the test temperature was increased
40 800 120
20 ~ j S b to about 75% of the melting temperature.
0 600 Times to failure are reported in Ref 326
0 100 200 300 400 500 600 0 100 200 300 400 500 600 for AISI 4140 (1380 MPa, or 200 ksi, tensile
Test temperature, °C Test temperature, °C strength) in contact with solid zinc, lead,
cadmium, tin, and indium at temperatures
(a) (b)
below the melting points of the embrittlers.
Effect of adding antimony to lead on (a) the tensile reduction in area and (b) fracture strength as a The time to failure decreased as the temper-
Fig. 57 function of test temperature for externally embrittled AISI 4145 steel heat treated to a yield strength ature increased toward the melting point
of 1380 MPa (200 ksi). Source: Ref 333
and decreased as the applied stress was
increased toward the tensile strength. Crack
propagation was intergranular along the pri-
large shear lips and a large faceted origin. Nonleaded AISI 4145 tensile specimens at or-austenite grain boundaries.
The fractures were completely intergranular a tensile strength of 1380 MPA (200 ksi) As with other embrittlement mechanisms,
and were moving away from the origin were externally wetted with high-purity the segregation of embrittling impurities
toward the opposite shear lip. Inclusions lead and lead-antimony alloys. F o r the (phosphorus, antimony, arsenic, and tin) to
were always observed to be associated with high-purity lead, embrittlement began at the prior-austenite grain boundaries can influ-
the intergranular facets. Above 480 °C (900 lower temperatures, and the recovery of ence MIE susceptibility. In an examination of
°F), the ductility was recovered, and there embrittlement occurred at a lower temper- the influence of prior temper embrittlement of
was a return to a cup-and-cone fracture. ature (370 °C, or 700 °F), than for the AISI 3340 alloy steel containing controlled
Leaded Carbon and Alloy Steels. An eval- internally embrittled leaded steels. Addi- additions of phosphorus, antimony, tin, and
uation of a variety of leaded carbon and alloy tions of up to 2% Sb were made to the pure arsenic, the segregation of tin and antimony
steels found that all were subject to SME and lead, and similar tests were run that to the grain boundaries increased the suscep-
LME (Ref 331). Embrittlement began at showed that the onset of embrittlement, tibility to MIE by lead or tin; however, the
about the same temperature and was most the temperature range for severe embrit- segregation of phosphorus and arsenic to the
severe in the same temperature range, 315 to tlement, and the temperature for the re- grain boundaries produced a minor decrease
370 °C (600 to 700 °F), but the recovery of covery of ductility were all affected by the in MIE susceptibility in tests from room tem-
ductility at higher temperatures varied some- addition of antimony (Fig. 57) (Ref 332, perature to 425 °C (800 °F).
what, from about 400 to 480 °C (750 to 900 333). As the antimony content increased to
°F). It was thought that the composition of the 0.3%, Tr gradually increased. However, Neutron
lead could be affecting this brittle-to-ductile for the 1 and 2% Sb specimens, little or no
transition temperature, Tr.
Irradiation Embrittlement
evidence for recovery was obtained up to
Subsequent work c o m p a r e d externally 480 °C (900 °F). Similar work was done for The neutron irradiation of nuclear reactor
applied liquid lead and lead alloys to the lead containing up to 9% Sn (Ref 333). components can produce an increase in
previous internal lead work (Ref 332, 333). Results were similar for tin except that on strength and in the ductile-to-brittle transi-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 723

I dramatically, and the rule appears to be

Melting
point [
invalid. Also, in most cases the chemical
o
1.0 species causing SCC need not be present in
large amounts or high concentrations in the
environment.
The caustic cracking of riveted boiler
== steels was also observed late in the nine-
0.8
teenth century. Caustic cracking involves a
localized high concentration of free alkali at
high temperatures, as would exist in a boil-
er. At the beginning of this century, SCC of
0.6 ~ aluminum alloys due to moisture was ob-
served. Also, SCC of martensitic steels was
t0, D ~ ~--==~
n I encountered, but the nature of such cracks
I was not recognized until much later (addi-
tional information is available in the article
I "Wrought Stainless Steels" in this Vol-
0.8 I ume). Cracking of plain carbon steels (mild
steels) by nitrates was encountered in the
chemical industry at the beginning of this
._o [] Zn century.
0.6
z~ Pb Stainless steels, first used extensively in
0 Cd
(> Sn the 1930s, were also observed to be suscep-
"6 '7 In tible to such failures, which were due to
chlorides or caustic environments at high
temperatures. Also in the 1930s, magnesium
~ 0.4 aircraft alloys were found to be susceptible
cr
to SCC caused by moisture. Cracking prob-
lems with titanium alloys were observed
beginning in the 1950s.
0.2 Considerable research has been conduct-
ed on stress-corrosion cracking. Much of it
has been centered on gaining an under-
standing of the conditions that cause it for
i commercial alloys and defining the influ-
0.4 0.5 0.6 0.7 0.8 0.9 1.0 ence of alloy composition, processing, and
Homologous temperature, TH microstructure. Research has also focused
on defining the mechanism(s) responsible
I:|.w Normalized true fracture strength ratio and reduction of area ratio as a function of the homologous
"'15. 58 temperature (TH = T/Tm) for heat-treated AISI 4140 alloy steel in external contact with solid lead,
for SCC and developing remedies to mini-
cadmium, zinc, tin, and indium. Source: Ref 307 mize or prevent SCC failures. This article
concentrates on SCC of iron-base alloys,
with emphasis on practical aspects rather
than mechanisms. References 335 to 361
tion temperature. The degree of irradiation- Stress-Corrosion Cracking provide a wealth of additional information
induced embrittlement depends on the neu- on stress-corrosion cracking, particularly
tron dose, neutron spectrum, temperature, Stress-corrosion cracking is a generic with respect to iron-base alloys.
steel composition, and heat treatment. term describing the initiation and propaga-
Tempered martensite is less susceptible to tion of cracks in a metal or alloy under the Properties and Conditions
neutron irradiation embrittlement than tem- combined action of tensile stresses (applied Producing Stress-Corrosion Cracking
pered bainite or ferrite-pearlite microstruc- and/or residual) and a corrosive environ- A number of general characteristics of
tures. Impurity elements in the steels can ment. The history of SCC failures begins in stress-corrosion cracking should be summa-
influence embrittlement; for example, phos- the nineteenth century with the develop- rized before dealing with specific topics.
phorus levels above 0.015% and copper ment of cold-drawn brass cartridges for Alloy Susceptibility to SCC. First, the alloy
levels above 0.05% are detrimental. ammunition. Similar failures of brass con- must be susceptible to stress-corrosion
Radiation produces swelling and void for- denser tubes brought additional attention to cracking. Although it was formerly believed
mation. Void density decreases as the irra- this problem, which soon gained the name that pure metals are immune to SCC, recent
diation temperature increases, but the aver- season cracking. These failures were due to experience has shown instances in which
age void size increases. The examination of residual tensile stresses, and the benefits of pure metals containing certain impurity el-
fractures of irradiated ferritic materials test- stress-relieving treatments were subse- ements were susceptible. It could be argued
ed below their DBTT reveals a change in quently shown. Later, the detrimental influ- that such pure metals were not really pure.
fracture mode from transgranular cleavage ence of ammonia on the stress-corrosion Also, tests of high-purity alloys have dem-
to mixed cleavage and intergranular frac- cracking of brass was observed, which lead onstrated greatly enhanced resistance to
ture. The examination of fracture made to the empirical rule of the specificity of the SCC even for alloys that are normally quite
above the DBTT reveals a change in dimple environment causing SCC for each alloy. susceptible. Therefore, because pure is a
size and depth. Additional information on This rule was cited for many years, but with relative term, and because the influence of
this subject can be found in the article subsequent research, particularly in the last impurity elements on SCC requires further
"Effect of Neutron Irradiation on Proper- 30 years, the list of environments that can study, it can be concluded that improving
ties of Steels" in this Volume. cause SCC for specific alloys has grown the purity of metals and alloys is very

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 724 / Service Characteristics of Carbon and Low-Alloy Steels

Table 2 Partial list of aqueous 600

environments known to cause
stress-corrosion cracking in steels -- 80

Low-carbon steel and Austenitic

low-alloy ferritic steels stainless steels
500 I Type ~ -- 70
Nitrates Chlorides
Phosphates Hydroxides
Sulfates Fluorides
Carbonates Bromides T ~ e ? ~ ~,
X 309
316 - 60
400 - ' ~ ' '
Carbon monoxide/carbon dioxide Water/oxygen
Hydrogen sulfide Sulfates #_ 347 .~
Ferric chloride Thiocyanates
Ammonia Thiosulfates
Organic liquids Tetrathionates _

Water/oxygen Polythionates
Cyanides Sulfur dioxide
Hydroxides Sulfurous acid
<
Source: Ref 358

~ - 20
beneficial, but it is not yet a commercially
viable solution to the SCC problem.
Static Tensile Stresses. The metal or alloy "/////, /// -- 10
must not only be susceptible to SCC, it
must also be subjected to static tensile
0 Iu
stresses for SCC to occur. These stresses 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 103 (5X 103)
may be applied service loads, or they may Time to fracture, h
result from residual stresses, or both.
Stresses can also arise from the wedging I=;n 59 Composite curves showing the relative stress-corrosion cracking resistance of standard austenitic
"b ° stainless steels in boiling 42% MgCl 2, Source: Ref 362
action of corrosion products growing in a
crack. These stresses need not exceed the
yield strength of the material and can be SCC in iron-base alloys (Table 2). As Table can cause SCC, the process may begin with
relatively low in comparison. Generally, but 2 indicates, a great many damaging environ- the formation of a corrosion pit. Such pits
not always, some stress level will exist ments are known (and this is only a partial could be present on a part before it is placed
below which cracking does not occur but list); further research will probably add to in service if, for example, the part were
above which it does (that is, a threshold the list. Also, as previously mentioned, the previously excessively pickled (that is, acid
stress level discussed previously in the sec- damaging species does not need to be pres- cleaned). Pits may also form at inclusions
tions " H y d r o g e n D a m a g e " and "Metal- ent in large quantities or high concentra- that intersect the free surface of the part. Pit
Induced Embrittlement" in this article). tions. Indeed, even parts-per-million levels geometry is quite important; the pit depth
Compressive residual stresses, such as from have been known to cause SCC failures. must be much greater than the pit width by
shot peening, are beneficial in reducing or Only small quantities are necessary because about a factor of ten, at a minimum. The
preventing SCC. the damaging species can become concen- environment within the pit must change if it
SCC-lnducing Chemical Species. The alloy trated in the affected area. Also, an envi- is to act as a crack initiator. Not all pits
in question must not only be under the ronment that causes SCC in one alloy may initiate stress-corrosion cracks. It is possi-
influence of an applied or residual tensile not produce SCC in another alloy (unless ble that a preexisting pit may not develop
stress of sufficient magnitude, it must also they are similar in nature and composition). the same electrochemical conditions as one
be in contact with an environment that Similarly, a particular environment may that forms during exposure in the environ-
contains a chemical species that will cause cause SCC of an alloy only under specific ment. The local chemistry in the pit appears
SCC. An interesting aspect of SCC is that conditions, such as a particular tempera- to be important in SCC.
the alloy in question is normally nearly ture, degree of aeration, or pH. A particular Intergranular Corrosion or Slip Dissolu-
inert, from the standpoint of general corro- heat treatment or strength of the alloy could tion. Stress-corrosion cracks can also ini-
sion, in the environment that causes SCC. also have an impact. tiate in the absence of pitting, for example,
This anomalous behavior was one of the by slip dissolution (film rupture) processes
biggest surprises to early researchers who Parameters Affecting or by intergranular corrosion. In intergran-
studied SCC of stainless steels because Stress-Corrosion Cracking ular corrosion-initiated SCC, the grain-
these are alloys noted for their superior Variables such as temperature, pit geom- boundary composition must differ from
corrosion resistance. Indeed, SCC failures etry, intergranular corrosion, and slip dis- the bulk composition. This can occur be-
generally show very few corrosion effects. solution processes can also promote stress- cause of the segregation of elements to the
The damaging environment, however, may corrosion cracking, depending on their grain boundaries (or because of sensitiza-
produce a film on the surface of the alloy. degree or severity. tion in austenitic stainless steels). The slip
Such films can be tarnish, passivating lay- Temperature is often a critical parameter dissolution process is associated with lo-
ers, or dealloyed layers, depending on the in SCC. For some systems, failures occur calized corrosion on slip planes at the
system. These films retard general corro- only at temperatures above some specific surface of low stacking fault energy auste-
sion. level, and the severity or rapidity of failure nitic alloys.
As mentioned, it was believed for many increases with temperature above this limit.
years that only a few such species could Chloride-induced SCC of austenitic stain- SCC Verification Procedures
cause SCC for certain alloys or families of less steels is known to follow this behavior. Stress-corrosion cracking is a delayed
alloys. A partial list has been compiled of Pit Geometry. F o r a smooth specimen or failure process in which a certain incubation
species in aqueous environments that cause part under tension in an environment that time is required to initiate cracking. The

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 725

cracks propagate slowly, and propagation 103

O
rates can vary from 0.1 to 1 0 - 9 mm/s (0.004
to 4.0 × 1 0 - " in./s) (Ref 358). Cracking O

/
continues until the stress exceeds the frac- Cracking
ture strength of the remaining noncracked > oO o
O
cross section. Crack initiation and propaga- D
tion are usually divided into three stages: O o
C O@ O
• Crack initiation and stage I propagation
(increasing growth rate with increasing
stress intensity)
• Steady-state crack propagation (stage II: 0
X
growth constant over intermediate stress X C
©
intensities)
©
• Crack propagation to final fracture (stage 8
100
III: rapid increase in crack growth rate at D
high stress intensities) C
O

The threshold stress intensity necessary to

produce SCC is called K i s c o O
l G % l
0
Historically, stress-corrosion cracking
tests have been conducted on statically
loaded smooth specimens at various stress

1 o
Minimum time
levels, and the time to fracture is deter- J to cracking
mined for specific environmental condi- rn O O
O
tions. The resulting data are generally plot- O ~.
O
ted on semilog paper, with the time to 0
failure located on the log scale. Tests are
conducted in relatively severe environ- 0
ments under the assumption that if cracking ~×
10

/
does not occur under the test conditions, x0
the alloy is safe to use at the same stress
level and temperature in a less severe envi- 0
ronment. 0

No cracking
Chloride cracking tests of stainless steels, O
for example, have been performed using
O
boiling magnesium chloride (MgCI2) solu-
O
tions, which constitute a very severe envi-

/
D
ronment. It is well known that the boiling O
point of aqueous magnesium chloride varies do
with the amount of magnesium chloride.
This test was introduced in 1945 and is
detailed in ASTM G 36. In this test, the
× indicates commercial wire
composition is adjusted so that the solution indicates did not crack in 30 days
boils at 155 +- 1 °C (311 -+ 2 °F). The test
must be carefully controlled because results
do vary for different magnesium chloride
contents and temperature.
0 10 20 30 40 50 60 70 80
U-Bend Testing. U-bend specimens have Nickel content, %
also been widely used to test for SCC. They
exhibit large elastic and plastic strains and Effect of n i c k e l c o n t e n t on the time to cracking f o r Fe-18-20Cr-Niwiresin b o i l i n g 4 2 % M g C l 2. Source:
Fig. 60 Ref363
are a very severe test arrangement for a
smooth specimen. ASTM G 30 describes
the use of U-bend specimens for SCC test-
ing. crack growth does not occur, K,scc. Kiscc reduction in areas, or the ultimate tensile
varies as a function of alloy chemistry, heat strengths. While these are useful tests for
Fracture mechanics methods are much treatment, strength level, and environment. comparing different materials or environ-
more recent in development but are now ments, they are less useful than the frac-
widely used in research studies. Precracked Slow strain rate tests are also used to ture mechanics approach for predicting
specimens containing a sharp notch are evaluate the SCC tendency of alloys. Either in-service behavior.
used with either a constant applied load or a precracked or smooth specimens are used
fixed crack opening displacement. Crack in these tests, which are conducted by
growth rates, da/dt, are measured as a func- slowly increasing the load or strain on the SCC Evaluation of Stainless Steels
tion of the stress intensity factor, K. Tests alloy. Generally, results in the detrimental Austenitie Stainless Steels. Figure 59
are conducted to define the crack growth environment are compared to those in an shows the time to failure by SCC in boiling
rate in stages I, II, and III and the critical inert environment, for example, by plot- magnesium chloride as a function of applied
threshold stress intensity level below which ting the ratios of the strains to failure, the stress for a variety of austenitic stainless

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 726 / Service Characteristics of Carbon and Low-Alloy Steels

IliA IVA VA Stress intensity, ksi

10 20 30 40 50 60
10 5 - -
IIA VIA VIA VIA Solution annealed, 1060 °C 10-6 m
10-6 Specimen type: double cantilever
beam, constant load
10 7
>.
10 7

IVB VB VIB VIIB ~1 VII [ IB liB

A • X ;10-6 "~
1 0 -8

Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As b 109 go
1 0 -9
p,
X X A ~ Base % • • • X 8 Klscc 10_10 o

1 ~ Single
10-10 cracks - - o~
Zr Cb Mo Tc Ru Rh Pd Ag Cd In Sn Sb 03 Branched cracks 03
1 0 11

X X A X X X • A X 10-11 I [ I
10 20 30 40 50 60
Stress intensity, MPax/~-
Hf Ta W Re Os Ir Pt Au Hg TI Pb Bi Influence of stress intensity on crack
Fig, 6 2 growth rate and branching of stress-corro-
X X X X × X X X sion cracks for AISI 304L in 42% MgCI2 at 130 °C (265
°F). Threshold stress intensity level, Ktscc, was 8
MPa~/'m (7.3 k s i ~ ) . Source: Ref 265
Rare earths
~ Beneficial A Variable ['~ Neeffect ~ Z 100

X Detrimental -- Not investigated 50 ~,

o / \\
/ -,
-'bl:|~ 61 Segment of periodic table showing influence of chemical element additions to austenitic stainless /
steels on their SCC behavior in chloride solutions. Source: Ref 350, 354 \
20 / \
/o
o h\
/
steels (Ref 362). The families of curves shift carbon and molybdenum. As these ele- 10 ~, Oi "b,.
to greater SCC resistance with increased ments increased in concentration, the time / o \ O
nickel content. The 304 and 304L alloys / °/
to fracture passed through a minimum at
contained 9% Ni; the 305,309, 316,347, and about 0.065 to 0.085% C and !.5 to 2.1% ~" 5 / 9/ \ "o\
347L alloys had nickel contents ranging Mo. Titanium additions had only a minor -= /o /
\
from 10.5 to 13.9%; and the 310 and 314 influence on the effect of carbon. For the
alloys had nickel contents of 24.5 and rather small nickel range studied, nickel
o ; o//
o~ O \
19.7%, respectively.
The influence of nickel on stress-corro-
additions had a minor beneficial influence
on the time to fracture.
.E
~" I / ¢ %,,
sion cracking of iron-chromium-nickel al- References 350 and 354 summarize the / o
loys has been demonstrated (Fig. 60) (Ref influence of alloying-element additions to /
363). Alloys with 8 to 77% Ni, 18 to 20% Cr, austenitic stainless steels on their SCC be- /
and a balance of iron were produced as havior in chloride solutions (Fig. 61). Many /
wires (annealed, partially or fully hardened alloying elements are detrimental to chlo-
0. l
by drawing) and loaded in tension in boiling ride cracking, but some are beneficial; a few /
42% MgCI2 to determine the time to crack- have variable effects. Nickel, cadmium,
ing with loads of 230 or 310 MPa (33 or 45 zinc, silicon, beryllium, and copper are list- 0.2
ksi). Cracking occurred most readily for ed as being beneficial. Nickel is beneficial
alloys with about 8 to 12% Ni, and some of for resisting chloride cracking in solutions
the 8% nickel wires broke within 1 h. The other than magnesium chloride. However, 0.1
time to cracking increased with increasing 0 20 40 60 80 100
for ferritic stainless steels, small additions
nickel content; alloys with more than ap- of nickel are detrimental to SCC resistance Ferrite, %
proximately 45% Ni were immune to crack- in chlorides. Cadmium and zinc are listed as r.*n Effect of ferrite content on the time to
rag. being beneficial, but they have not been H ~ . 63 failure for duplex stainless steels in boiling
42% MgCI 2 under an applied stress of 240 MPa (34.8
The effect of composition on SCC of studied systematically. The influences of ksi). Source: Ref 368
austenitic stainless steels, particularly in silicon and beryllium have been studied in
magnesium chloride solutions, has been some depth. Silicon additions improve the
widely studied (Ref 344, 345,349, 350, 352, resistance to SCC in magnesium chloride, appear to be detrimental, but larger
355,363,364). Multiple regression analysis but results in sodium chloride (NaC1) at amounts appear to be helpful. Tin and man-
was used to detect compositional influences higher temperatures showed little benefit. ganese additions have demonstrated either
for 19 austenitic heats tested in boiling 42% The beneficial influence of copper is minor. no influence or detrimental or beneficial
MgCI2 (Ref 364). The study concluded that Additions of boron, aluminum, and co- effects, depending on the amounts present.
the most important alloying elements were balt have variable effects. Small amounts Like carbon and molybdenum, chromium

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 727

103 60

O
\,
CaCI2
g
100
d
Unbroken
after
50 [ I. ! i
o 300 h

~ 10

40

1.0
2.0 3.0 4.0 5.0
pH
o
~.0
Ir;a Effect of solution pH on the time to failure 30
H S " 64 for AISI 304 in MgCI 2 and CaCI2 at 125 °C ~$17400
(255 °F) with an applied load of 345 MPa (50 ksi). & $17700
Source: Ref 369

20 1 ~
/
additions exhibit a minimum in chloride AISI 410
[] Z
cracking resistance, apparently at about AISI 416
AISI 431
20%.
Crack growth rates for AISI 304L in !
0 AISI410
magnesium chloride at 130 °C (265 °F) have 10
• AISI416
/
been measured as a function of stress inten- AISI431 0 0 • /
sity, K (Ref 365). Figure 62 shows the • $17400
[] $17700
results using precracked specimens. The
o,.o/
threshold stress intensity level,,__K3Kscc, was
about 8 MPaX/-~ (7.3 ksiVin.). Single
cracks were observed in stage I, and 20 25 30 35 40 45 50
branching cracks were observed in stage II. Hardness, HRC
Temperature has a very important influ-
Relationship between specimen hardness and time to failure in aqueous 0.5% acetic acid saturated
ence on the chloride cracking of austenitic Fig. 6 5 with hydrogen sulfide for martensitic and precipitation-hardenable stainless steels. Source: Ref 370
stainless steels. Experience has shown that
it does not occur at room temperature.
Intergranular attack was observed in se-
verely sensitized alloys and in welded AISI of the ferrite must be controlled in order to Heat-treated martensitic and precipita-
301 after a 5-year ambient temperature ex- achieve such benefits. tion-hardenable (PH) stainless steels can
posure in a marine atmosphere at Kure An evaluation of the influence of small also fail by SCC. For these grades, the
Beach, NC (Ref 366). Normally, chloride amounts of 8-ferrite (< 10%) on SCC in 18% strength level has a strong influence on
cracking is transgranular; however, these Cr-8% Ni grades in boiling 42% MgCI2 cracking tendency. Figure 65 shows this
alloys cracked intergranularly, apparently found an increase in the time to fracture influence in a distilled water solution con-
because of the sensitization. Therefore, the with increasing B-ferrite content (Ref 367). taining 0.5% acetic acid (CH3COOH) satu-
failures may have been a form of intergran- The beneficial effect of ferrite in duplex rated with hydrogen sulfide (HzS), using
ular stress-assisted attack, rather than true stainless steels exposed to boiling magne- U-bend specimens (Ref 370). For the stan-
SCC (Ref 354). Chloride cracking does oc- sium chloride has been demonstrated (Ref dard martensitic grades, cracking occurred
cur at temperatures above ambient, gener- 368). Figure 63 shows these results, which readily at hardnesses above 25 HRC; for the
ally at temperatures above 60 °C (140 °F). indicate a maximum in the time to fracture PH grades, cracking occurred readily above
As the temperature increases above this at about 40% ferrite. 40 HRC.
limit, the time to fracture decreases, other Solution pH also has an influence on An evaluation of AISI 410 and three PH
factors being constant. Additional informa- stress-corrosion cracking. The pH in the grades (UNS S17400, S13800, and S15500)
tion on the exposure of austenitic steels to crack can be quite different from that of the was conducted using the National Associa-
marine atmospheres is available in the arti- bulk environment. Acidic conditions are tion of Corrosion Engineers (NACE) sulfide
cle "Wrought Stainless Steels" in this Vol- required for chloride cracking, and an in- stress cracking test (5% NaCI and 0.5%
ume. crease in acidity enhances SCC. acetic acid solution purged with argon to
Duplex stainless steels in the annealed Austenitic Stainless Steels. Figure 64 remove oxygen and saturated with hydro-
condition are more resistant to SCC than shows the influence of pH on the time to gen sulfide) at loads of 172,345, or 515 MPa
are the common austenitic grades AISI 304 fracture for AISI 304 in magnesium chloride (25, 50, or 75 ksi) in tension (Ref 371).
and 316. It is known that 8-ferrite in austen- and calcium chloride (CaC12) at 125 °C (255 Figure 66 shows the time to fracture as a
itic grades improves the resistance to chlo- °F). The specimens were loaded in tension function of yield strength for specimens
ride cracking. This has been attributed to at 345 MPa (50 ksi). Variations in the pH loaded at 345 MPa (50 ksi). The tempering
the blocking of crack propagation by the had a greater effect on magnesium chloride temperature (for AISI 410) and aging treat-
~-ferrite. Relatively high amounts (as in than on calcium chloride, and the former ments (for the PH grades) are shown for
duplex grades) are required to be effective. was a much more severe environment than each strength level and alloy. The numbers
Naturally, the morphology and distribution the latter. in parentheses are the amount of austenite

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 7 2 8 / Service Characteristics of C a r b o n and L o w - A l l o y Steels

0.2% yield strength, ksi 4000

68 80 100 120 140 160 180 200 220
] I I
103 I I I I I I Grain size
2000 o ASTM No. 5
o c • ASTM No. 9
& 1150M(36%) 1000
k
\
\ 600
\
\
•~ 1 5 0 (30%) .c_ 400
E
o. ,}
©•
\
100
200
1150M (19%) \ . . S13800
J~z o c

_.= ~ NACE (14%) \~. o I- 100

o
\ \
\1~,,817400 \
60 N

iz "~I~NACE Nz~50 (8°/o) \\ 40

900 (1%)
o N : CE " \ . 1150"~ 00 (3%) \ \ \ ec ~
10

,, OOlO
'O.12 I @, ""-.J050 (1%) \ 2O
815500 \ J go0 • o"-1

/~--1050 (7%)
,\ \ 10
0 10 20 30 40 50 60 70 80
$41000~ 400 and 600C ~ 8 0 0 ' \ ~ / ~ 950 (1%)"\
Applied stress, % of yield strength
\ \950 \ ~1000(4%)
"1ooo (1%) \ "bl:"a* 6 7 Effect of austenite grain size and applied
1000 stress as a percentage of the yield strength
~,950 (3%/
1000 ~A900 (311.7 and 358.5 MPa, or 45.2 and 52 ksi) for specimens
with grain sizes ASTM No. 5 and No. 9, respectively,
on the time to failure for AISI 302 wires in boiling 42%
MgCI 2. Source: Ref 372
400 600 800 1000 1200 1400 1600
0.2% yield strength, MPa
tercrystalline Embrittlement of Pure
Time to failure in an aqueous solution (5% NaCI and 0.5% acetic acid deaerated and saturated with Iron, Mem. Sci. Rev. Met., Vol 70
Fig. 66 hydrogen sulfide) as a function of yield strength for UNS $41000, $15500, $17400, and $13800 stainless (No. i), 1973, p 13-22
steels loaded to 345 MPa (50 ksi) in tension. Tempering temperatures, in degrees Fahrenheit, or temper condition are
shown by data points; values in parentheses are the approximate austenite contents. Source: Ref 371 11( J.R. Rellick et al., The Effect of Tel-
lurium on Intergranular Cohesion in
Iron, Metall. Trans., Vol 2, May 1971,
present. The alloys exhibited increased Addition of Scavengers, Metall. p 1492-1494
cracking resistance in the order AISI 410, Trans., Vol 1, April 1970, p 929-937 12. M.C. Inman and H.R. Tipler, Grain-
SI5500, S17400, SI3800. 4. P. Jolly, Discussion of "The Elimina- Boundary Segregation of Phosphorus
Grain size differences can influence tion of Oxygen-Induced Intergranular in an Iron-Phosphorus Alloy and the
stress-corrosion cracking resistance. Stud- Brittleness in Iron by Addition of Effect Upon Mechanical Properties,
ies have shown that increasing the grain size Scavengers," Metall. Trans., Vol 2, Acta Metall., Vol 6, Feb 1958, p 73-84
accelerates SCC, although the effect is not Jan 1971, p 341-342 13. B.E. Hopkins and H.R. Tipler, The
large, except at low applied stresses (Ref 5. J.R. Rellick et al., Further Informa- Effect of Phosphorus on the Tensile
372). Figure 67 shows the time to failure for tion on Oxygen Induced Intergranular and Notch-Impact Properties of High-
AISI 302 wires in boiling magnesium chlo- Brittleness in Iron, Metall. Trans., Vol Purity Iron and Iron-Carbon Alloys, J.
ride as a function of the applied stress, 2, Jan 1971, p 342-343 Iron Steel Inst., Vol 188, March 1958,
expressed as a percentage of the yield 6. C. Pichard et al., The Influence of p 218-237
strength and for grain sizes of ASTM No. 5 Oxygen and Sulfur on the Intergranu- 14. H. Erhart and H.J. Grabke, Equilibri-
and 9. The influence of grain size appears to lar Brittleness of Iron, Metall. Trans. um Segregation of Phosphorus at
be most important with regard to crack A, Vol 7A, Dec 1976, p 1811-1815 Grain Boundaries of Fe-P, Fe-C-P,
initiation, particularly for the slip dissolu- 7. A. Kumar and V. Raman, Low Tem- Fe-Cr-P, and Fe-Cr-C-P Alloys, Met.
tion mechanism. perature Intergranular Brittleness of Sci., Vol 15, Sept 1981, p 401-408
Iron, Acta Metall., Vol 29, 1981, p 15. P.V. Ramasubramanian and D.F.
REFERENCES 1131-1139 Stein, An Investigation of Grain-
8. B.D. Powell et al., A Study of Inter- Boundary Embrittlement in Fe-P, Fe-
I. W.P. Rees and B.E. Hopkins, Inter- granular Fracture in Iron Using Auger P-S and Fe-Sb-S Alloys, Metall.
granular Brittleness in Iron-Oxygen Spectroscopy, Metall. Trans., Vol 4, Trans., Vol 4, July 1973, p 1735-1742
Alloys, J. Iron Steel Inst., Vol 172, Oct 1973, p 2357-2361 16. C.J. McMahon, Jr., Strength of Grain
Dec 1952, p 403-409 9. P. Jolly and C. Goux, Influence of Boundaries in Iron-Base Alloys, in
2. C.J. McMahon, Jr., Intergranular Brit- Certain Impurities on Intercrystalline Grain Boundaries in Engineering Ma-
tleness in Iron, Acta Metall., Vol 14, Embrittlement of Iron, Mem. Sci. terials, Claitor, 1975, p 525-552
July 1966, p 839-845 Rev. Met., Vol 66 (No. 9), 1969, p 17. J.R. Rellick and C.J. McMahon, Jr.,
3. J.R. Rellick and C.J. McMahon, Jr., 605 -617 Intergranular Embrittlement of Iron-
The Elimination of Oxygen-Induced 10. C. Pichard et al., Influence of Sulfur Carbon Alloys by Impurities, Metall.
Intergranular Brittleness in Iron by Type Metalloid Impurities on the In- Trans., Vol 5, Nov 1974, p 2439-2450

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 729

18. C.J. Thwaites and S.K. Chatterjee, bon Steels, Metall. Trans. A, Voi 9A, 48. L.A. Erasmus, Effect of Aluminum
Effect of Tin on the Impact Behavior Dec 1978, p 1765-1775 Additions on Forgeability, Austenite
of Alloys Based on High-Purity Iron 34. D.V. Wilson and B. Russell, The Con- Grain Coarsening Temperature, and
and Mild Steel, J. Iron Steel Inst., Vol tribution of Precipitation to Strain Ag- Impact Properties, J. Iron Steel Inst.,
210, Aug 1972, p 581-587 ing in Low Carbon Steels, Acta Met- Vol 202, Jan 1964, p 32-41
19. R.L. Kenyon and R.S. Burns, Testing all., Vol 8, July 1960, p 468-479 49. R. Harris and L. Barnard, Experienc-
Sheets for Blue Brittleness and Stabil- 35. E.T. Stephenson and M. Cohen, The es of Hot Shortness in the Forging of
ity Against Changes Due to Aging, Effect of Prestraining and Retemper- Certain Low-Alloy Steels, in Defor-
Proc. ASTM, Vol 34, 1934, p 48-58 ing on AISI Type 4340, Trans. ASM, mation Under Hot Working Condi-
20. E.O. Hall, The Deformation of Low- Vol 54, 1961, p 72-83 tions, SR 108, Iron and Steel Institute,
Carbon Steel in the Blue-Brittle 36. C.H. Lorig and A.R. Elsea, Occur- 1968, p 167-177
Range, J. Iron Steel Inst., Vol 170, rence of Intergranular Fracture in Cast 50. F. Vodopivec, Influence of Precipita-
April 1952, p 331-336 Steels, Trans. AFS, Vol 55, 1947, p tion and Precipitates of Aluminum Ni-
21. G. Mima and F. Inoko, A Study of the 160-174 tride on Torsional Deformability of
Blue-Brittle Behavior of a Mild Steel 37. B.C. Woodfine and A.G. Quarrell, Ef- Low-Carbon Steel, Met. Technol.,
in Torsional Deformation, Trans. JIM, fects of AI and N on the Occurrence of Vol 5, April 1978, p 118-121
Vol 10, May 1969, p 227-231 Intergranular Fracture in Steel Cast- 51. G.D. Funnell and R.J. Davies, Effect
22. T. Takeyama and H. Takahashi, ings, J. Iron Steel Inst., Vol 195, Aug of Aluminum Nitride Particles on Hot
Strength and Dislocation Structures of 1960, p 409-414 Ductility of Steel, Met. Technol., Vol
a-Irons Deformed in the Blue-Brittle- 38. J.A. Wright and A.G. Quarrell, Effect 5, May 1978, p 150-153
ness Temperature Range, Trans. ISIJ, of Chemical Composition on the Oc- 52. G.D. Funnell, Observations on Effect
Vol 13, 1973, p 293-302 currence of Intergranular Fracture in of Aluminum Nitride on Hot Ductility
23. A.L. Tsou et al., The Quench-Aging Plain Carbon Steel Castings Contain- of Steel, in Hot Working and Forming
of Iron, J. Iron Steel Inst., Vol 172, ing Aluminum and Nitrogen, J. Iron Processes, Book 264, The Metals So-
Oct 1952, p 163-171 Steel Inst., Vol 200, April 1962, p ciety, 1980, p 104-107
24. T.C. Lindley and C.E. Richards, The 299-307 53. K. Yamanaka et al., Relation Between
Effect of Quench-Aging on the Cleav- 39. L. Barnard and R. Brook, Intergranu- Hot Ductility and Grain-Boundary
age Fracture of a Low-Carbon Steel, lar Fracture of Alloy Steels, J. Iron Embrittlement of Low-Carbon Killed
Met. Sci. J., Vol 4, May 1970, p 81-84 Steel Inst., Voi 205, July 1967, p 756- Steels, Trans. ISIJ, Vol 20, 1980, p
25. A.S. Keh and W.C. Leslie, Recent 762 810-816
Observations on Quench-Aging and 40. N.E. Hannerz, Influence of Cooling 54. S.L. Gertsman and H.P. Tardif, Tin
Strain Aging of Iron and Steel, in Rate and Composition on Intergranu- and Copper in Steel: Both Are Bad,
Materials Science Research, Vol 1, lar Fracture of Cast Steel, Met. Sci. J., Together They're Worse, Iron Age,
Plenum Publishing, 1963, p 208-250 Vol 2, 1968, p 148-152 Vol 169 (No. 7), Feb 14, 1952, p 136-
26. E.R. Morgan and J.F. Enrietto, Aging 41. N.H. Croft, Use of Solubility Data to 140
in Steels, in AISI 1963 Regional Tech- Predict the Effects of Aluminum and 55. P. Bjornson and H. Nathorst, A Spe-
nical Meeting, American Iron and Nitrogen Contents on the Susceptibil- cial Type of Ingot Cracks Caused by
Steel Institute, 1964, p 227-252 ity of Steel Castings to lntergranular Certain Impurities, Jernkontorets
27. E. Stolfe and W. Heller, The State of Embrittlement, Met. Technol., Vol 10, Ann., Vol 139 (No. 6), 1955, p 412-438
Knowledge of the Aging of Steels: I, Aug 1983, p 285-290 56. D.A. Melford, Surface Hot Shortness
Fundamental Principles, Stahl und 42. N.H. Croft et al., Origins of Dendritic in Mild Steel, J. Iron Steel Inst., Vol
Eisen, Vo190 (No. 16), 1970, p 861-868 A1N Precipitates in Aluminum-Killed- 200, April 1962, p 290-299
28. G. Lagerberg and B.S. Lement, Mor- Steel Castings, Met. Technol., Vol 10, 57. A. Nicholson and J.D. Murray, Sur-
phological and Phase Changes During April 1983, p 125-129 face Hot Shortness in Low-Carbon
Quench-Aging of Ferrite Containing 43. N.H. Croft et al., Intergranular Frac- Steel, J. Iron Steel Inst., Vol 203, Oct
Carbon and Nitrogen, Trans. ASM, ture of Steel Castings, in Advances in 1965, p 1007-1018
Vol 50, 1958, p 141-162 the Physical Metallurgy and Applica- 58. W.J.M. Salter, Effects of Alloying El-
29. J.D. Baird, Strain Aging of Steel A tions of Steels, Book 284, The Metals ements on Solubility and Surface En-
Critical Review, Iron Steel, Vol 36, Society, 1982, p 286-295 ergy of Copper in Mild Steel, J. Iron
1963, p 186-192,326-334,368-374,400- 44. S.C. Desai, Longitudinal Panel Crack- Steel Inst., Vol 204, May 1966, p 478-
405, and 450-457 ing in Ingots, J. Iron Steel Inst., Vol 488
30. J.D. Baird, The Effects of Strain-Ag- 191, March 1959, p 250-256 59. W.J.M. Salter, Effect of Chromium on
ing Due to Interstitial Solutes on the 45. E. Colombo and B. Cesari, The Study Solubility of Copper in Mild Steel, J.
Mechanical Properties of Metals, Met. of the Influence of A1 and N on the Iron Steel Inst., Vol 205, Nov 1967, p
Rev., Vol 16, Feb 1971, p 1-18 Susceptibility to Crack Formation of 1156-1160
31. R.D. Butler and D.V. Wilson, The Medium Carbon Steel Ingots, Metall. 60. W.J.M. Salter, Effect of Mutual Addi-
Mechanical Behavior of Temper Ital., Vol 59, 1967, p 71-75 tions of Tin and Nickel on the Solubil-
Rolled Steel Sheets, J. Iron Steel 46. L. Ericson, Cracking in Low Alloy ity and Surface Energy of Copper in
Inst., Vol 201, Jan 1963, p 16-33 Aluminum Grain Refined Steels, Mild Steel, J. Iron Steel Inst., Vo1207,
32. F. Garofalo and G.V. Smith, The Ef- Scand. J. Metall., Vol 6, 1977, p 116- Dec 1969, p 1619-1623
fect of Time and Temperature on Var- 124 61. H.J. Kerr and F. Eberle, Graphitiza-
ious Mechanical Properties During 47. R. Sussman et al., Occurrence and tion of Low-Carbon and Low-Carbon-
Strain Aging of Normalized Low Car- Control of Panel Cracking in Alumi- Molybdenum Steels, Trans. ASME,
bon Steels, Trans. ASM, Vo147, 1955, num Containing Steel Heats, in Me- Vol 67, 1945, p 1-46
p 957-983 chanical Work. and Steel Processing, 62. S.L. Hoyt et al., Summary Report on
33. C.C. Li and W.C. Leslie, Effects of Vol 17, American Institute of Mining, the Joint E.E.I.-A.E.I.C. Investiga-
Dynamic Strain Aging on the Subse- Metallurgical, and Petroleum Engi- tion of Graphitization in Piping, Trans.
quent Mechanical Properties of Car- neers, 1979, p 49-78 ASME, Vol 68, Aug 1946, p 571-580

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 730 / Service Characteristics of Carbon and Low-Alloy Steels

63. R.W. Emerson and M. Morrow, Fur- Embrittlement of 18Ni(350) Maraging 95. B.J. Schulz and C.J. McMahon, Jr.,
ther Observations of Graphitization in Steel, Metall. Trans., Vol 2, Sept Alloy Effects in Temper Embrittle-
Aluminum-Killed Carbon-Molybdenum 1971, p 2665-2672 ment, in Temper Embrittlement of Al-
Steel Steam Piping, Trans. AIME, Vol 79. W.C. Johnson and D.F. Stein, A loy Steels, STP 499, American Society
68, Aug 1946, p 597-607 Study of Grain Boundary Segregants for Testing and Materials, 1972, p 104-
64. J.G. Wilson, Graphitization of Steel in in Thermally Embrittled Maraging 135
Petroleum Refining Equipment, Weld. Steel, Metall. Trans., Vol 5, March 96. C.J. McMahon, Jr. et al., The Influ-
Res. Counc. Bull., No. 32, Jan 1957, p 1974, p 549-554 ence of Mo on P-Induced Temper Em-
1-10 80. E. Nes and G. Thomas, Precipitation brittlement in Ni-Cr Steel, Metall.
65. A.B. Wilder et al., Stability of AISI of TiC in Thermally Embrittled Marag- Trans. A, Vol 8A, July 1977, p 1055-
Alloy Steels, Trans. AIME, Vol 209, ing Steels, Metall. Trans. A, Vol 7A, 1057
Oct 1957, p 1176-1181 July 1976, p 967-975 97. H.L. Marcus and P.W. Palmberg, Au-
66. E.J. Dulis and G.V. Smith, Roles of 81. H.J. Rack and P.H. Holloway, Grain ger Fracture Surface Analysis of a
Aluminum and Nitrogen in Graphitiza- Boundary Precipitation in 18Ni-Ma- Temper Embrittled 3340 Steel, Trans.
tion, Trans. ASM, Vol 46, 1954, p raging Steels, Metall. Trans. A, Vol AIME, Vol 245, July 1969, p 1664-1666
1318-1330 8A, Aug 1977, p 1313-1315 98. D.F. Stein et al., Studies Utilizing
67. A. Preece et al., The Overheating and 82. G.F. Vander Voort, Failures of Tools Auger Electron Emission Spectrosco-
Burning of Steel, J. Iron Steel Inst., and Dies, in Failure Analysis and Pre- py on Temper Embrittlement in Low
Part I, Vol 153, 1946, p 237p-254p; vention, Vol 11,9th ed., Metals Hand- Alloy Steels, Trans. ASM, Vol 62,
and, Part III, Vol 164, 1950, p 37-45 book, American Society for Metals, 1969, p 776-783
68. I.S. Brammar, A New Examination of 1986, p 563-585 99. P.W. Palmberg and H.L. Marcus, An
the Phenomena of Overheating and 83. L.D. Jaffee and J.R. Hollomon, Hard- Auger Spectroscopic Analysis of the
Burning of Steels, J. Iron Steel Inst., enability and Quench Cracking, Trans. Extent of Grain Boundary Segrega-
Vol 201, Sept 1963, p 752-761 AIME, Vol 167, 1946, p 617-626 tion, Trans. ASM, Vol 62, 1969, p
69. R.D.N. Lester, Overheating in Steels, 84. M.C. Udy and M.K. Barnett, A Lab- 1016-1018
SteelTimes, Vol 193 (No. 513), 15 July oratory Study of Quench Cracking in 100. D.F. Stein, Reversible Temper Em-
1966, p 96-102 Cast Alloy Steels, Trans. ASM, Vol brittlement, Anna. Rev. Mater. Sci.
70. G.D. Joy and J. Nutting, Influence of 38, 1947, p 471-487 1977, Vol 7, 1977, p 123-153
Intermetallic Phases and Non-Metallic 85. J.W. Spretnak and C. Wells, An Engi- 101. L.D. Jaffe and D.C. Buffum, Isother-
Inclusions Upon the Ductility and neering Analysis of the Problem of mal Temper Embrittlement, Trans.
Fracture Behavior of Some Alloy Quench Cracking in Steel, Trans. ASM, Vol 42, 1950, p 604-618
Steels, in Effect of Second-Phase Par- ASM, Vol 42, 1950, p 233-269 102. F.L. Carr et al., Isothermal Temper
ticles on the Mechanical Properties of 86. C. Wells, Quench Cracks in Wrought Embrittlement of SAE 3140 Steel,
Steel, Iron and Steel Institute, 1971, p Steel Tubes, Met. Prog., Vol 65, May Trans. AIME, Vol 197, Aug 1953, p
95-100 1954, p 113-121 998
71. T.J. Baker and R. Johnson, Overheat- 87. T. Kunitake and S. Sugisawa, The 103. M.P. Seah, Grain Boundary Segrega-
ing and Fracture Toughness, J. Iron Quench-Cracking Susceptibility of tion and the T-t Dependence of Tem-
Steel Inst., Vol 211, Nov 1973, p 783- Steel, Sumitomo Search, No. 5, May per Brittleness, Acta Metall., Vol 25,
791 1971, p 16-25 1977, p 345-357
72. D.R. Glue et al., Effect of Composi- 88. P. Gordon, The Temper Colors on 104. B.C. Woodfine, Some Aspects of
tion and Thermal Treatment on the Steel, J. Heat Treat., Vol 1, June Temper-Brittleness, J. Iron Steel
Overheating Characteristics of Low- 1979, p 93 Inst., Vol 173, March 1953, p 240-255
Alloy Steels, Met. Technol., Vol 2, 89. G.F. Vander Voort, Visual Examina- 105. J.M. Capus, Austenite Grain Size and
Sept 1975, p 416-421 tion and Light Microscopy, in Frac- Temper Brittleness, J. Iron Steel Inst.,
73. R.C. Andrew et al., Overheating in tography, Vol 12, 9th ed., Metals Vol 200, Nov 1962, p 922-927
Low-Sulphur Steels, J. Australasian Handbook, ASM INTERNATION- 106. R. Viswanathan and A. Joshi, Effect
Inst. Met., Vol 21, June-Sept 1976, p AL, 1987, p 91-165 of Microstructure on the Temper Em-
126-131 90. K. Balajiva et al., Effects of Trace brittlement of Cr-Mo-V Steels, Metall.
74. R.C. Andrew and G.M. Weston, The Elements on Embrittlement of Steels, Trans. A, Vol 6A, Dec 1975, p 2289-
Effect of Overheating on the Tough- Nature, Vol 178, 1956, p 433 2297
ness of Low Sulphur ESR Steels, J. 91. W. Steven and K. Balajiva, The Influ- 107. D.L. Newhouse et al., Temper Em-
Aast. Inst. Met., Vol 22, Sept-Dec ence of Minor Elements on the Iso- brittlement Study of Nickel-Chromi-
1972, p 171-176 thermal Embrittlement of Steels, J. um-Molybdenum-Vanadium Rotor
75. R.C. Andrew and G.M. Weston, The Iron Steel Inst., Vol 193, Oct 1959, p Steels, I: Effects of Residual Ele-
Effect of the Interaction Between 141-147 ments, in Temper Embrittlement of
Overheating and Tempering Tempera- 92. J.R. Low et al., Alloy and Impurity Alloy Steels, STP 499, American Soci-
ture on the Notch Toughness of Two Effects on Temper Brittleness of ety for Testing and Materials, 1972, p
Low Sulphur Steels, J. Aust. Inst. Steel, Trans. AIME, Vol 242, Jan 3-36
Met., Vol 22, Sept-Dec 1972, p 200- 1968, p 14-24 108. J. Watanabe and Y. Murakami, Pre-
204 93. R.H. Greaves and J.A. Jones, Tem- vention of Temper Embrittlement of
76. G.E. Hole and J. Nutting, Overheating per-Brittleness of Nickel-Chromium Chromium-Molybdenum Steel Vessels
of Low-Alloy Steels, Int. Met. Rev., Steels, J. Iron Steel Inst., Vol 102, by Use of Low-Silicon Forged Steels,
Vol 29, 1984, p 273-298 1920, p 171-222 Proc. API Refin. Dept., Vol 60, 1981,
77. G.J. Spaeder et al., The Effect of Hot 94. R.H. Greaves and J.A. Jones, Tem- p 216-224
Rolling Variables on the Fracture per-Brittleness of Steel: Susceptibility 109. S. Takayama et al., The Calculation of
Toughness of 18Ni Maraging Steel, to Temper-Brittleness in Relation to Transition Temperature Changes in
Trans. ASM, Vol 60, 1967, p 418-425 Chemical Composition, J. Iron Steel Steels Due to Temper Embrittlement,
78. D. Kalish and H.J. Rack, Thermal Inst., Vol 111, 1925, p 231-255 Metall. Trans. A, Vol IIA, Sept 1980,

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 731

p 1513-1530 monds, Tempered Martensite Embrit- 27, March 1971, p 119-127

110. R. Viswanathan and S.M. Bruemmer, tlement: Role of Retained Austenite 139. A.B. Kinzel, Chromium Carbide in
In-Service Degradation of Toughness and Cementite, Met. Sci., Vol 13, June Stainless Steel, Trans. AIME, Vol 194,
of Steam Turbine Rotors, J. Eng. 1979, p 325-334 May 1952, p 469-488
Mater. Technol. (Trans. ASME), Vol 125. M. Sarikaya et al., Retained Austenite 140. R. Stickler and A. Vinckier, Electron
107, Oct 1985, p 316-324 and Tempered Martensite Embrittle- Microscope Investigation of the Inter-
111. B.C. Woodfine, Temper-Brittleness: ment in Medium Carbon Steels, Met- granular Corrosion Fracture Surfaces
A Critical Review of the Literature, J. all. Trans. A, Vol 14A, June 1983, p in a Sensitized Austenitic Stainless
Iron Steel Inst., Vol 173, March 1953, 1121-1133 Steel, Corros. Sci., Vol 3, 1963, p 1-8
p 229-240 126. H. Kwon and C.H. Kim, Tempered 141. M.A. Streicher, General and Inter-
112. J.M. Capus, The Mechanism of Tem- Martensite Embrittlement in Fe-Mo-C granular Corrosion of Austenitic
per Brittleness, in Temper Embrittle- and Fe-W-C Steel, Metall. Trans. A, Stainless Steels in Acids, J. Electro-
ment in Steel, STP 407, American So- Vol 14A, July 1983, p 1389-1394 chem. Soc., Vol 106, March 1959, p
ciety for Testing and Materials, 1968, 127. R.M. Horn and R.O. Ritchie, Mecha- 161-180
p 3-19 nisms of Tempered Martensite Em- 142. K.T. Aust et al., Heat Treatment and
113. C.J. McMahon, Jr., Temper brittlement in Low Alloy Steels, Met- Corrosion Resistance of Austenitic
Brittleness--An Interpretive Review, all. Trans. A, Vol 9A, Aug 1978, p Type 304 Stainless Steel, Trans. ASM,
in Temper Embrittlement in Steel, STP 1039-1053 Vol 59, 1966, p 544-556
407, American Society for Testing and 128. J.P. Materkowski and G. Krauss, 143. K.T. Aust et al., Intergranular Corro-
Materials, 1968, p 127-167 Tempered Martensite Embrittlement sion and Electron Microscopic Studies
114. B.L. Eyre et al., Physical Metallurgy in SAE 4340 Steel, Metall. Trans. A, of Austenitic Stainless Steels, Trans.
of Reversible Temper Embrittlement, Vol 10A, Nov 1979, p 1643-1651 ASM, Vol 60, 1967, p 360-372
in Advances in the Physical Metallur- 129. N. Bandyopadhyay and C.J. McMa- 144. K.T. Aust et al., Intergranular Corro-
gy and Application of Steels, Book hon, Jr., The Micro-Mechanisms of sion and Mechanical Properties of
284, The Metals Society, 1982, p 246- Tempered Martensite Embrittlement Austenitic Stainless Steels, Trans.
258 in 4340-Type Steels, Metall. Trans. A, ASM, Vol 61, 1968, p 270-277
115. L.J. Klingler et al., The Embrittle- Vol 14A, July 1983, p 1313-1325 145. K.T. Aust, Intergranular Corrosion of
ment of Alloy Steel at High Strength 130. C.L. Briant, Role of Carbides in Tem- Austenitic Stainless Steels, Trans.
Levels, Trans. ASM, Vol 46, 1954, p pered Martensite Embrittlement, AIME, Vol 245, Oct 1969, p 2117-2126
1557-1598 Mater. Sci. Technol., Vol 5, Feb 1989, 146. J.S. Armijo, Impurity Adsorption and
116. G. Delisle and A. Galibois, Tempered p 138-147 Intergranular Corrosion of Austenitic
Martensite Brittleness in Extra-Low- 131. F. Zia-Ebrahimi and G. Krauss, The Stainless Steel in Boiling HNO3-
Carbon Steels, J. Iron Steel Inst., Vol Evaluation of Tempered Martensite K2CrzO 7 Solutions, Corros. Sci., Vol
207, Dec 1969, p 1628-1634 Embrittlement in 4130 Steel by Instru- 7, 1967, p 143-150
117. M. Gensamer, "Study of the Effects mented Charpy V-Notch Testing, 147. J.S. Armijo, Intergranular Corrosion
of Vacuum Melting on 550°F Temper- Metall. Trans. A, Vol 14A, June 1983, of Nonsensitized Austenitic Stainless
ing Embrittlement," WADC Techni- p 1109-1119 Steels, Corrosion, Vol 24, Jan 1968, p
cal Report 57-85, ASTIA Document 132. J.A. Peters et al., On the Mechanisms 24-30
AD 130850, Wright Air Development of Tempered Martensite Embrittle- 148. T.M. Devine et al., Mechanism of
Center, Sept 1957 ment, Acta Metall., Vol 37, Feb 1989, Intergranular Corrosion of 316L Stain-
118. J.M. Capus and G. Mayer, The Influ- p 675-686 less Steel in Oxidizing Acids, Scr.
ence of Trace Elements on Embrittle- 133. B. Strauss et al., Carbide Precipitation Metall., Vol 14, 1980, p 1175-1179
ment Phenomena in Low-Alloy Steels, in the Heat Treatment of Stainless Non- 149. A. Joshi and D.F. Stein, Chemistry of
Metallurgia, Vol 62, Oct 1960, p 133- Magnetic Chromium-Nickel Steels, Zh. Grain Boundaries and Its Relation to
138 Anorg. Allg. Chem., Vol 188, 1930, p Intergranular Corrosion of Austenitic
119. E.B. Kula and A.A. Anctil, Tempered 309-324 Stainless Steel, Corrosion, Vol 28,
Martensite Embrittlement and Frac- 134. E.C. Bain et al., The Nature and Pre- Sept 1972, p 321-330
ture Toughness in SAE 4340 Steel, J. vention of lntergranular Corrosion in 150. C.L. Briant, The Effects of Sulfur and
Mater., Vol 4, Dec 1969, p 817-841 Austenitic Stainless Steels, Trans. Phosphorus on the Intergranular Cor-
120. S.K. Banerji et al., Intergranular Frac- ASST, Vol 21, June 1933, p 481-509 rosion of 304 Stainless Steel, Corro-
ture in 4340-Type Steels: Effects of 135. K.G. Caroll et al., Chromium Distri- sion, Vol 36, Sept 1980, p 497-509
Impurities and Hydrogen, Metall. bution Around Grain Boundary Car- 151. C.L. Briant, The Effect of Alloying
Trans. A, Vol 9A, Feb 1978, p 237-247 bides Found in Austenitic Stainless Elements on Impurity Induced Inter-
121. C.L. Briant and S.K. Banerji, Phos- Steel, Nature, Vol 184, 1959, p 1479- granular Corrosion, Corrosion, Voi
phorus Induced 350°C Embrittlement 1480 38, April 1982, p 230-232
in an Ultra High Strength Steel, Met- 136. C.W. Weaver, Grain-Boundary Pre- 152. C.L. Briant, Grain Boundary Segrega-
all. Trans. A, Vol 10A, Jan 1979, p cipitation in Nickel-Chromium-Base tion of Phosphorus and Sulfur in
123-126 Alloys, J. Inst. Met., Vol 90, 1961- Types 304L and 316L Stainless Steel
122. C.L. Briant and S.K. Banerji, Tem- 1962, p 404 and Its Effect on Intergranular Corro-
pered Martensite Embrittlement in 137. S. Aim and R. Kiessling, Chromium sion in the Huey Test, Metall. Trans.
Phosphorus Doped Steels, Metall. Depletion Around Grain-Boundary A, Vol 18A, April 1987, p 691-699
Trans. A, Vol 10A, Nov 1979, p 1729- Precipitates in Austenitic Stainless 153. C.S. Pande et al., Direct Evidence of
1737 Steel, J. Inst. Met., Vol 91, 1962-1963, Chromium Depletion Near the Grain
123. G. Thomas, Retained Austenite and p 190 Boundaries in Sensitized Stainless
Tempered Martensite Embrittlement, 138. R.J. Hodges, Intergranular Corrosion Steels, Scr. Metall., Vol 11, 1977, p
Metall. Trans. A, Vol 9A, March 1978, in High Purity Ferritic Stainless 681-684
p 439-450 Steels: Effect of Cooling Rate and 154. P. Rao and E. Lifshin, Microchemical
124. H.K.D.H. Bhadeshia and D.V. Ed- Alloy Composition, Corrosion, Vol Analysis in Sensitized Austenitic

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 732 / Service Characteristics of Carbon and Low-Alloy Steels

Steel, in Proceedings of the 8th Annu- Growth of M23C 6 Carbide on Grain ment of High Chromium Steels Within
al Conference of the Microbeam Anal- Boundaries in an Austenitic Stainless Temperature Range of 570-1100°F,
ysis Society, 1977, p I18A-118F Steel, Trans. AIME, Vol 242, May Arch. Eisenhiittenwes., Vol 15, Oct
155. R.A. Mulford et al., Sensitization of 1968, p 814-819 1941, p 175-182
Austenitic Stainless Steels: II, Com- 169. C. Da Casa et al., M23C6 Precipitation 184. W. Dann6hl, Discussion of Ref 183
mercial Purity Alloys, Corrosion, Vol in Unstabilized Austenitic Stainless (and The Embrittlement of High-Alloy
39, April 1983, p 132-143 Steel, J. Iron Steel Inst., Vol 207, Oct Chrome Steels in the Temperature
156. E.L. Hall and C.L. Briant, Chromium 1969, p 1325-1332 Range About 500°, by G. Bandel and
Depletion in the Vicinity of Carbides 170. W.O. Binder et al., Resistance to Sen- W. Tofaute, Arch. Eisenhiittenwes.,
in Sensitized Austenitic Stainless sitization of Austenitic Chromium- Vol 15, 1942, p 307-320), Arch. Eisen-
Steels, Metall. Trans. A, Vol 15A, Nickel Steels of 0.03% Max. Carbon hiittenwes., Vol 15, 1942, p 319
May 1984, p 793-811 Content, Trans. ASM, Vol 41, 1949, p 185. H.D. Newell, Properties and Charac-
157. C.L. Briant and E.L. Hall, A Compar- 1301-1370 teristics of 27% Chromium Iron, Met.
ison Between Grain Boundary Chro- 171. S. Danyluk et al., Intergranular Frac- Prog., Vol 49, May 1946, p 977-1028
mium Depletion in Austenitic Stain- ture, Corrosion Susceptibility, and Im- 186. J.J. Heger, 885°F Embrittlement of the
less Steel and Corrosion in the purity Segregation in Sensitized Type Ferritic Chromium-Iron Alloys, Met.
Modified Strauss Test, Corrosion, Vol 304 Stainless Steel, J. Mater. Energy Prog., Vol 60, Aug 1951, p 55-61
42, Sept 1986, p 522-531 Syst., Vol 7, June 1985, p 6-15 187. A.J. Lena and M.F. Hawkes, 475°C
158. S.M. Bruemmer and L.A. Chariot, 172. R.A. Lula et al., Intergranular Corro- (885°F) Embrittlement in Stainless
Development of Grain Boundary sion of Ferritic Stainless Steels, Trans. Steels, Trans. AIME, Vol 200, May
Chromium Depletion in Type 304 and ASM, Vol 46, 1954, p 197-230 1954, p 607-6 ! 5
316 Stainless Steels, Scr. Metall., Vol 173. A.P. Bond, Mechanisms of Intergran- 188. C.A. Zapffe, Fractographic Pattern for
20, 1986, p 1019-1024 ular Corrosion in Ferritic Stainless 475°C Embrittlement in Stainless
159. E.P. Butler and M.G. Burke, Chromi- Steels, Trans. AIME, Vol 245, Oct Steel, Trans. AIME, Vol 191, March
um Depletion and Martensite Forma- 1969, p 2127-2134 1951, p 247-248
tion at Grain Boundaries in Sensitized 174. R.J. Hodges, Intergranular Corrosion 189. H. Masumoto et al., The Anomaly of
Austenitic Stainless Steel, Acta Met- in High Purity Ferritic Stainless the Specific Heat at High Tempera-
all., Voi 34, March 1986, p 557-570 Steels: Isothermal Time-Temperature tures in a-Phase Alloys of Iron and
160. C. StawstrOm and M. Hillert, An Im- Sensitization Measurements, Corro- Chromium, Sci. Rep. Res. Inst., T6-
proved Depleted-Zone Theory of In- sion, Vol 27, April 1971, p 164-167 huko Univ. A, Vol 5, 1953, p 203-207
tergranular Corrosion of 18-8 Stainless 175. J.J. Demo, Mechanism of High Tem- 190. S. Takeda and N. Nagai, Experimen-
Steel, J. Iron Steel Inst., Vol 207, Jan perature Embrittlement and Loss of tal Research on Superlattices in Iron-
1967, p 77-85 Corrosion Resistance in AISI Type Chromium System, Mere. Fac. Eng.,
161. C.S. Tedmon, Jr. et al., Intergranular 446 Stainless Steel, Corrosion, Vol 27, Nagoya Univ., Vol 8, 1956, p 1-28
Corrosion of Austenitic Stainless Dec 1971, p 531-544 191. R.M. Fisher et al., Identification of
Steel, J. Electrochem. Sot., Vol 118, 176. J.A. Davis et al., Intergranular Corro- the Precipitate Accompanying 885°F
Feb 1971, p 192-202 sion Resistance of a 26Cr-lMo Ferritic Embrittlement in Chromium Steels,
162. R.L. Fullman, A Thermodynamic Stainless Steel Containing Niobium, Trans. AIME, Vol 197, May 1953, p
Model of the Effects of Composition Corrosion, Vol 36, May 1980, p 215- 690-695
on the Susceptibility of Austenitic 220 192. R.O. Williams and H.W. Paxton, The
Stainless Steels to Intergranular Stress 177. T.M. Devine et al., Influence of Heat Nature of Aging of Binary Iron-Chro-
Corrosion Cracking, Acta Metall., Vol Treatment on the Sensitization of mium Alloys Around 500°C, J. Iron
30, 1982, p 1407-1415 18Cr-2Mo-Ti Stabilized Ferritic Stain- Steel Inst., Vol 185, March 1957, p
163. G.S. Was and R.M. Kruger, A Ther- less Steel, Metall. Trans. A, Vol 12A, 358-374
modynamic and Kinetic Basis for Un- Dec 1981, p 2063-2069 193. G.F. Tisinai and C.H. Samans, Some
derstanding Chromium Depletion in 178. T.M. Devine and A.M. Ritter, Sensi- Observations of 885°F Embrittlement,
Ni-Cr-Fe Alloys, Acta Metall., Vol 33, tization of 12 Wt Pct Chromium, Tita- Trans. AIME, Vol 209, Oct 1957, p
May 1985, p 841-854 nium-Stabilized Ferritic Stainless 1221-1226
164. S.M. Bruemmer, Sensitization Devel- Steel, Metall. Trans. A, Vol 14A, Aug 194. M.J. Blackburn and J. Nutting, Metal-
opment in Austenitic Stainless Steel. 1983, p 1721-1728 lography of an Iron-21% Chromium
Measurement and Prediction of Ther- 179. P. Payson, Prevention of Intergranular Alloy Subjected to 475°C Embrittle-
momechanical History Effects, Corro- Corrosion in Corrosion-Resistant ment, J. Iron Steel Inst., Vol 202, July
sion, Vol 44, July 1988, p 427-434 Chromium-Nickel Steels, Trans. 1964, p 610-613
165. E.M. Mahla and N.A. Nielsen, Car- AIME, Vol 100, 1932, p 306-333 195. M.J. Marcinkowski et al., Effect of
bide Precipitation in Type 304 Stain- 180. T.M. Devine, Mechanism of Inter- 500°C Aging on the Deformation Be-
less Steel--An Electron Microscope granular Corrosion and Pitting Corro- havior of an Iron-Chromium Alloy,
Study, Trans. ASM, Vol 43, 1951, p sion of Austenitic and Duplex 308 Trans. AIME, Vol 230, June 1964, p
290-322 Stainless Steel, J. Electrochem. Soc., 676-689
166. R. Stickler and A. Vinckier, Morphol- Vol 126, March 1979, p 374-385 196. R. Lagneborg, Metallography of the
ogy of Grain-Boundary Carbides and 181. T.M. Devine, Jr., Influence of Carbon 475°C Embrittlement in an Iron-30%
Its Influence on Intergranular Corro- Content and Ferrite Morphology on Chromium Alloy, Trans. ASM, Vol
sion of 304 Stainless Steel, Trans. the Sensitization of Duplex Stainless 60, 1967, p 67-78
ASM, Vol 54, 1961, p 362-380 Steel, Metall. Trans. A, Vol llA, May 197. R. Lagneborg, Deformation in an Iron-
167. F.R. Beckitt and B.R. Clark, The 1980, p 791-800 30% Chromium Alloy Aged at 475°C,
Shape and Mechanism of Formation of 182. F.M. Becket, On the Allotropy of Acta Metall., Vol 15, Nov 1967, p
M23C6 Carbide in Austenite, Acta Stainless Steels, Trans. AIME, Vol 1737-1745
Metall., Vol 15, Jan 1967, p 113-129 131, 1938, p 15-36 198. T. DeNys and P.M. Gielen, Spinodal
168. L.K. Singhal and J.W. Martin, The 183. G. Riedrich and F. Loib, Embrittle- Decomposition in the Fe-Cr System,

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 733

Metall. Trans., Vol 2, May 1971, p cial Report 64, The Iron and Steel ASM, Vol 42, 1950, p 981-1007
1423-1428 Institute, 1959, p 118-124 229. J.H. Jackson, The Occurrence of the
199. P.J. Grobner, The 885°F (475°C) Em- 216. E.O. Hall and S.H. Algie, The Sigma Sigma Phase and Its Effect on Certain
brittlement of Ferritic Stainless Steels, Phase, Metall. Rev., Vol 11, 1966, p Properties of Cast Fe-Ni-Cr Alloys, in
Metall. Trans., Vol 4, Jan 1973, p 61-88 Symposium on the Nature, Occur-
251-260 217. E.R. Jette and F. Foote, The Fe-Cr rence, and Effects of Sigma Phase,
200. R.O. Williams, Further Studies of the Alloy System, Met. Alloys, Vol 7, Aug STP 110, American Society for Test-
Iron-Chromium System, Trans. 1936, p 207-210 ing and Materials, 1951, p 101-127
AIME, Vol 212, Aug 1958, p 497-502 218. A.J. Lena and W.E. Curry, The Effect 230. L.P. Stoter, Thermal Aging Effects in
201. P.J. Grobner and R.F. Steigerwald, of Cold Work and Recrystallization on AISI Type 316 Stainless Steel, J.
Effect of Cold Work on the 885°F the Formation of the Sigma Phase in Mater. Sci., Vol 16, 1981, p 1039-1051
(475°C) Embrittlement of 18Cr-2Mo Highly Stable Austenitic Stainless 231. R.G. Ellis and G. Pollard, The Obser-
Ferritic Stainless Steels, J. Met., Vol Steels, Trans. ASM, Vol 47, 1955, p vation of Sigma Phase After Short
29, July 1977, p 17-23 193-210 Aging Times in a Duplex Steel, J. Iron
202. T.J. Nichol et al., Embrittlement of 219. P. Duhaj et al., Sigma-Phase Precipi- Steel Inst., Vol 208, Aug 1970, p 783-
Ferritic Stainless Steels, Metall. tation in Austenitic Steels, J. Iron 784
Trans. A, Vol 11A, April 1980, p 573- Steel Inst., Vol 206, Dec 1968, p 1245- 232. H.D. Solomon and T.M. Devine, In-
585 1251 fluence of Microstructure on the Me-
203. W. Haoquan et al., Influence of An- 220. B. Weiss and R. Stickler, Phase Insta- chanical Properties and Localized
nealing and Aging Treatments on the bilities During High Temperature Ex- Corrosion of a Duplex Stainless Steel,
Embrittlement of Type 446 Ferritic posure of 316 Austenitic Stainless in MiCon 78: Optimization of Process-
Stainless Steel, J. Mater. Eng., Vol 9, Steel, Metall. Trans., Vol 3, April ing, Properties, and Service Perfor-
1987, p 51-61 1972, p 851-866 mance Through Microstructural Con-
204. W. Trietschke and G. Tammnann, The 221. M.T. Shehata et al., A Quantitative trol, STP 672, American Society for
Alloys of Iron and Chromium, Zh. Metallographic Study of the Ferrite to Testing and Materials, 1979, p 430-461
Anorg. Chem., Vol 55, 1907, p 402-411 Sigma Transformation in Type 316 233. Y. Maehara et al., Effects of Alloying
205. E.C. Bain and W.E. Griffiths, An In- Stainless Steel, in Microstruct. Sci., Elements on ~ Phase Precipitation in
troduction to the Iron-Chromium- Vol 11, Elsevier, 1983, p 89-99 ~--,/ Duplex Phase Stainless Steels,
Nickel Alloys, Trans. AIME, Vol 75, 222. J.K.L. Lai et al., Precipitate Phases in Met. Sci., Vol 17, Nov 1983, p 541-547
1927, p 166-213 Type 316 Austenitic Stainless Steel 234. W.H. Johnson, On Some Remarkable
206. P. Chevenard, Experimental Investi- Resulting From Long-Term High Changes Produced in Iron and Steel by
gations of Iron, Nickel, and Chromium Temperature Service, Mater. Sci. the Action of Hydrogen and Acids,
Alloys, Tray. Mem., Bur. Int. Poids et Eng., Vol 49, 1981, p 19-29 Proc. R. Soc., Vol 23 (No. 158), 1875,
Mesures, Vol 17, 1927, p 90 223. J. Barcik and B. Brzycka, Chemical p 168-179; reprinted in Hydrogen
207. F. Wever and W. Jellinghaus, The Composition of ~r Phase Precipitated Damage, American Society for Met-
Two-Component System: Iron-Chro- in Chromium-Nickel Austenitic als, 1977, p 1-12
mium, Mitt. Kaiser-Wilhelm Inst., Vol Steels, Met. Sci., Vol 17, May 1983, p 235. B. Craig, Hydrogen Damage, in Cor-
13, 1931, p 143-147 256-260 rosion, Vol 13, 9th ed., Metals Hand-
208. D.C. Ludwigson and H.S. Link, Fur- 224. T. Andersson and B. Lundberg, Effect book, ASM INTERNATIONAL,
ther Studies of the Formation of Sigma of Mo on the Lattice Parameters and 1987, p 163-189
in 12 to 16 Per Cent Chromium Steels, on the Chemical Composition of Sig- 236. J.P. Hirth and H.H. Johnson, Hydro-
in Advances in the Technology of ma Phase and M23C 6 Carbide in an gen Problems in Energy Related Tech-
Stainless Steels and Related Alloys, Austenitic 25Cr-20Ni Steel, Metall. nology, Corrosion, Vol 32, Jan 1976, p
STP 369, American Society for Test- Trans. A, Vol 8A, May 1977, p 787-790 3-26
ing and Materials, 1965, p 299-311 225. G.N. Emanuel, Sigma Phase and Oth- 237. R.P. Frohmberg et al., Delayed Fail-
209. J.H.G. Monypenny, The Brittle Phase er Effects of Prolonged Heating at ure and Hydrogen Embrittlement in
in High-Chromium Steels, Metal- Elevated Temperatures on 25 Per Cent Steel, Trans. ASM, Vol 47, 1955, p
lurgia, Vol 21, 1939-1940, p 143-148 Chromium-20 Per Cent Nickel Steel, 892-925
210. F.B. Foley, The Sigma Phase, Alloy in Symposium on the Nature, Occur- 238. A.R. Troiano, The Role of Hydrogen
Cast. Bull., July 1945, p 1-9 rence, and Effects of Sigma Phase, and Other Interstitials in the Mechan-
211. D.A. Oliver, The Sigma Phase in STP 110, American Society for Test- ical Behavior of Metals, Trans. ASM,
Stainless Steels, Met. Prog., Vol 55, ing and Materials, 1951, p 82-99 Vol 52, 1960, p 54-80
May 1949, p 665-667 226. G. Aggen et al., Microstructures Ver- 239. H.H. Johnson et al., Hydrogen, Crack
212. G.V. Smith, Sigma Phase in Stainless: sus Properties of 29-4 Ferritic Stain- Initiation, and Delayed Failure in
What, When and Why, Iron Age, Vol less Steel, in MiCon 78: Optimization Steel, Trans. AIME, Vol 212, Aug
166, 30 Nov 1950, p 63-68; 7 Dec 1950, of Processing, Properties, and Service 1958, p 528-536
p 127-132 Performance Through Microstructural 240. I.M. Bernstein, The Role of Hydrogen
213. A.J. Lena, Sigma Phase--A Review, Control, STP 672, American Society in the Embrittlement of Iron and Steel,
Met. Prog., Vol 66, July 1954, p 86-90; for Testing and Materials, 1979, p 334- Mater. Sci. Eng., Vol 6, 1970, p 1-19
Aug 1954, p 94-99; Sept 1954, p 122- 366 241. M.R. Louthan, Jr. et al., Hydrogen
126, 128 227. J.J. Gilman, Hardening of High-Chro- Embrittlement of Metals, Mater. Sci.
214. F.B. Foley and V.N. Krivobok, Sigma mium Steels by Sigma Phase Forma- Eng., Vol 10, 1972, p 357-368
Formation in Commercial Ni-Cr-Fe tion, Trans. ASM, Vol 43, 1951, p 242. M.R. Louthan, Jr., Effects of Hydro-
Alloys, Met. Prog., Vol 71, May 1957, 161-192 gen on the Mechanical Properties of
p 81-86 228. G.J. Guarnieri et al., The Effect of Low Carbon and Austenitic Steels, in
215. F.B. Pickering, The Formation of Sig- Sigma Phase on the Short-Time High Hydrogen in Metals, American Soci-
ma in Austenitic-Stainless Steels, in Temperature Properties of 25 Chromi- ety for Metals, 1974, p 53-77
Precipitation Processes in Steels, Spe- um-20 Nickel Stainless Steel, Trans. 243. I.M. Bernstein et al., Effect of Dis-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 734 / Service Characteristics of Carbon and Low-Alloy Steels

solved Hydrogen on Mechanical Be- Impurities and Their Effects on Hy- Based on A-286, Metall. Trans. A, Vol
havior of Metals, in Effect of Hydro- drogen-Induced Cracking, Metall. 6A, July 1975, p 1431-1442
gen on Behavior of Materials, The Trans. A, Vol 9A, May 1978, p 625-633 272. G.G. Hancock and H.H. Johnson, Hy-
Metallurgical Society, 1976, p 37-58 258. R. Viswanathan and R.I. Jaffie, drogen, Oxygen, and Subcritical
244. I.M. Bernstein and A.W. Thompson, "Clean Steels" to Control Hydrogen Crack Growth in a High Strength
Effect of Metallurgical Variables on Embrittlement, in Current Solutions to Steel, Trans. AIME, Vol 236, April
Environmental Fracture of Steels, Int. Hydrogen Problems in Steels, Ameri- 1966, p 513-516
Met. Rev., Vol 21, Dec 1976, p 269-287 can Society for Metals, 1982, p 275- 273. D.P. Dautovich and S. Floreen, The
245. A.W. Thompson, Effect of Metallur- 278 Stress Intensities for Slow Crack
gical Variables on Environmental 259. N. Bandyopadhyay et al., Hydrogen- Growth in Steels Containing Hydro-
Fracture of Engineering Materials, in Induced Cracking in 4340-Type Steel: gen, Metall. Trans., Vol 4, Nov 1973,
Environment-Sensitive Fracture of Effects of Composition, Yield p 2627-2630
Engineering Materials, The Metal- Strength, and H2 Pressure, Metall. 274. S.J. Hudak, Jr. and R.P. Wei, Hydro-
lurgical Society, 1979, p 379-410 Trans. A, Vol 14A, May 1983, p 881- gen Enhanced Crack Growth in 18 Ni
246. J.P. Hirth, Effects of Hydrogen on the 888 Maraging Steels, Metall. Trans. A,
Properties of Iron and Steel, Metall. 260. J. Kameda and C.J. McMahon, Jr., Vol 7A, Feb 1976, p 235-241
Trans. A, Vol I IA, June 1980, p 861- Solute Segregation and Hydrogen-In- 275. R.P. Gangloff and R.P. Wei, Gaseous
890 duced Intergranular Fracture in an Al- Hydrogen Embrittlement of High
247. A.W. Thompson and I.M. Bernstein, loy Steel, Metall. Trans. A, Vol 14A, Strength Steels, Metall. Trans. A, Vol
Microstructure and Hydrogen Embrit- May 1983, p 903-911 8A, July 1977, p 1043-1053
tlement, in Hydrogen Effects in Met- 261. H. Asahi et al., Effects of Mn, P, and 276. M. Iino, The Extension of Hydrogen
als, The Metallurgical Society, 1981, p Mo on Sulfide Stress Cracking Resis- Blister-Crack Array in Linepipe
291-308 tance of High Strength Low Alloy Steels, Metall. Trans. A, Vol 9A, Nov
248. C.G. Interrante, Basic Aspects of the Steels, Metall. Trans. A, Vol 19A, 1978, p 1581-1590
Problems of Hydrogen in Steels, in Sept 1988, p 2171-2177 277. M. Iino, Influence of Sulfur Content
Current Solutions to Hydrogen Prob- 262. A.W. Thompson, The Mechanism of on the Hydrogen-Induced Fracture in
lems in Steels, American Society for Hydrogen Participation in Ductile Linepipe Steels, Metall. Trans. A, Vol
Metals, 1982, p 3-17 Fracture, in Effect of Hydrogen on 10A, Nov 1979, p 1691-1698
249. G.M. Pressouyre, Current Solutions Behavior of Materials, American In- 278. B.E. Wilde et al., Some Observations
to Hydrogen Problems in Steel, in stitute of Mining, Metallurgical, and on the Role of Inclusions in the Hy-
Current Solutions to Hydrogen Prob- Petroleum Engineers, 1976, p 467-479 drogen Induced Blister Cracking of
lems in Steels, American Society for 263. T.D. Lee et al., Effect of Hydrogen on Linepipe Steels in Sulfide Environ-
Metals, 1982, p 18-34 Fracture of U-Notched Bend Speci- ments, Corrosion, Vol 36, Nov 1980, p
250. H.G. Nelson, Hydrogen Embrittle- mens of Spheroidized AISI 1095 Steel, 625-632
ment, in Embrittlement of Engineering Metall. Trans. A, Vol 10A, Feb 1979, 279. R.E. Cramer and E.C. Bast, The Pre-
Alloys, Vol 25, Treatise on Materials p 199-208 vention of Flakes by Holding Railroad
Science and Technology, Academic 264. H. Cialone and R.J. Asaro, The Role Rails at Various Constant Tempera-
Press, 1983, p 275-359 of Hydrogen in the Ductile Fracture of tures, Trans. ASM, Vol 27, Dec 1939,
251. H.H. Johnson, Keynote Lecture: Plain Carbon Steels, Metall. Trans. A, p 923-934
Overview on Hydrogen Degradation Vol 10A, March 1979, p 367-375 280. C.A. Zapffe and C.E. Sims, Hydro-
Phenomena, in Hydrogen Embrittle- 265. R. Garber et al., Hydrogen Assisted gen, Flakes and Shatter Cracks, Met.
ment and Stress Corrosion Cracking, Ductile Fracture of Spheroidized Car- Alloys, Vol 11, May 1940, p 145-151;
American Society for Metals, 1984, p bon Steels, Metall. Trans. A, Vol 12A, June 1940, p 177-184; Vol 12, July
3-27 Feb 1981, p 225-234 1940, p 44-51; Aug 1940, p 145-148
252. H.C. Rogers, Hydrogen Embrittle- 266. H. Cialone and R.J. Asaro, Hydrogen 281. F.B. Foley, Flakes and Cooling
ment in Engineering Materials, Mater. Assisted Fracture of Spheroidized Cracks in Forgings--A Problem in
Prot., Vol 1, April 1962, p 26, 28-30, 33 Plain Carbon Steels, Metall. Trans. A, Ordnance, Met. Alloys, Vol 12, Oct
253. E.A. Steigerwald et al., Discontinuous Vol 12A, Aug 1981, p 1373-1387 1940, p 442-445
Crack Growth in Hydrogenated Steel, 267. M.B. Whiteman and A.R. Troiano, 282. C.A. Zapffe, Defects in Cast and
Trans. AIME, Vol 215, Dec 1959, p Hydrogen Embrittlement of Austenitic Wrought Steel Caused by Hydrogen,
1048-1052 Stainless Steel, Corrosion, Vol 21, Met. Prog., Vol 42, Dec 1942, p 1051-
254. J.T. Brown and W.M. Baldwin, Jr., Feb 1965, p 53-56 1056
Hydrogen Embrittlement of Steels, 268. M.L. Holzworth and M.R. Louthan, 283. C.A, Zapffe, Sources of Hydrogen in
Trans. AIME, Vol 200, Feb 1954, p Jr., Hydrogen-Induced Phase Trans- Steel and Means for Its Elimination,
298-303 formations in Type 304L Stainless Met. Prog., Vol 43, March 1943, p
255. T. Toh and W.M. Baldwin, Jr., Duc- Steels, Corrosion, Vol 24, April 1968, 397-401
tility of Steel With Varying Concentra- p 110-124 284. E.R. Johnson et al., Flaking in Alloy
tions of Hydrogen, in Stress Corrosion 269. R.M. Vennett and G.S. Ansell, A Steels, in Proceedings of the National
Cracking and Embrittlement, John Study of Gaseous Hydrogen Damage Open Hearth Conference, Vol 27,
Wiley & Sons, 1956, p 176-186 in Certain FCC Metals, Trans. ASM, American Institute of Mining, Metal-
256. S. Marshall et al., Relationship Be- Vol 62, 1969, p 1007-1013 lurgical, and Petroleum Engineers,
tween Hydrogen Content and Ductili- 270. N. Narita et al., Hydrogen-Related 1944, p 358-377
ty of Steels, in Electric Furnace Steel Phase Transformations in Austenitic 285. A.W. Dana, Jr. et al., Relation of
Conference, Vol 6, American Institute Stainless Steels, Metall. Trans. A, Vol Flake Formation in Steel to Hydrogen,
of Mining, Metallurgical, and Petro- 13A, Aug 1982, p 1355-1365 Microstructure, and Stress, Trans.
leum Engineers, 1949, p 59-73 271. A.W. Thompson and J.A. Brooks, AIME, Vol 203, Aug 1955, p 895-905
257. C.L. Briant et al., Embrittlement of a Hydrogen Performance of Precipita- 286. W.L. Kerlie and J.H. Richards, Origin
5 Pct Nickel High Strength Steel by tion-Strengthening Stainless Steels and Elimination of Hydrogen in Basic

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 Embrittlement of Steels / 735

Open-Hearth Steels, Trans. AIME, Embrittlement, Scr. Metall., Vol 8, J., Voi 50, April 1971, p 174s-182s
Vol 209, Dec 1957, p 1541-1548 1974, p 519-526 318. W.F. Savage et al., Intergranular At-
287. J.M. Hodge et al., Effect of Hydrogen 303. C.H. Desch, The Solidification of tack of Steel by Molten Copper, Weld.
Content on Susceptibility to Flaking, Metals From the Liquid State (2nd J., Vol 57, Jan 1978, p 9s-16s
Trans. AIME, Vol 230, Aug 1964, p report), J. Inst. Met., Vol 22, 1919, p 319. W.F. Savage et al., Copper-Contami-
1182-1193 241-276 nation Cracking in the Weld Heat-
288. J.E. Ryall et al., The Effects of Hy- 304. E. Heyn, Internal Strains in Cold- Affected Zone, Weld. J., Vol 57, May
drogen in Rolled Steel Products, Met. Wrought Metals, and Some Troubles 1978, p 145s-152s
Forum, Vol 2 (No. 3), 1979, p 174-182 Caused Thereby, J. Inst. Met., Vol 12, 320. W.F. Savage et al., Liquid-Metal Em-
289. A.P. Lingras, Hydrogen Control in 1914, p 3-37 brittlement of the Heat-Affected Zone
Steelmaking, in Electric Furnace Steel 305. H.S. Rawdon, The Use of Mercury by Copper Contamination, Weld. J.,
Proceedings, Vol 40, Iron and Steel Solutions for Predicting Season Crack- Vol 57, Aug 1978, p 237s-245s
Society, 1983, p 133-143 ing in Brass, Proc. ASTM, Vol 18 (No. 321. L. Habraken and J. Lecomte-Beckers,
290. A.R. Huntington, discussion of paper 2), 1918, p 189-219 Hot Shortness and Scaling of Copper-
by C.H. Desch, The Solidification of 306. H. Moore and S. Beckinsale, The Re- Containing Steels, in Copper in Iron and
Metals From The Liquid State, J. Inst. moval of Internal Stresses in 70:30 Steel, John Wiley & Sons, 1982, p 45-81
Met., Vol 11, 1914, p 108-109 Brass by Low Temperature Anneal- 322. B. Straub, Microscopic Steel Investi-
291. W. Rostoker et al., Embrittlement by ing, J. Inst. Met., Vol 23, 1920, p gation, Stahl Eisen, Vol 34 (No. 50),
Liquid Metals, Reinhold, 1960 225-245 Dec 1914, p 1814-1820
292. N.S. Stoloff, Liquid Metal Embrittle- 307. J.C. Lynn et al., Solid Metal Induced 323. G.F. Vander Voort, Failures of Loco-
ment, in Surfaces and Interfaces, H, Embrittlement of Steel, Mater. Sci. motive Axles, in Failure Analysis and
Syracuse University Press, 1968, p Eng., Vol 18, March 1975, p 51-62 Prevention, Vol 11, 9th ed., Metals
157-182 308. J.M. Johnson et al., Zinc Embrittle- Handbook, American Society for Met-
293. A.R.C. Westwood et al., Adsorption- ment of Stainless Steel Welds, in Em- als, 1986, p 715-727
Induced Brittle Fracture in Liquid- brittlement by Liquid and Solid Met- 324. A.P. Druschitz and P. Gordon, Solid
Metal Environments, in Fracture, als, American Institute of Mining, Metal-Induced Embrittlement of Met-
Vol 3, Academic Press, 1971, p 589- Metallurgical, and Petroleum Engi- als, in Embrittlement by Liquid and
644 neers, 1984, p 415-434 Solid Metals, American Institute of
294. M.H. Kamdar, Embrittlement by Liq- 309. D.L. Hammon et al., Embrittlement Mining, Metallurgical, and Petroleum
uid Metals, in Progress in Materials of Engineering Materials During High- Engineers, 1984, p 285-316
Science, Vol 15, Part 4, Pergamon Temperature Fatigue in a Liquid-Lith- 325. J.C. Lynn and W.R. Warke, Delayed
Press, 1973, p 289-374 ium Environment, in Embrittlement by Failure of Steel Due to Solid Metal
295. M.H. Kamdar, Mechanism of Embrit- Liquid and Solid Metals, American Induced Embrittlement, in Embrittle-
tlement and Brittle Fracture in Liquid Institute of Mining, Metallurgic~al, and ment by Liquid and Solid Metals,
Metal Environment, in Fracture 1977, Petroleum Engineers, 1984, p 549-562 American Institute of Mining, Metal-
Vol 1, Pergamon Press, 1978, p 387- 310. R. Genders, The Penetration of Mild lurgical, and Petroleum Engineers,
405 Steel by Brazing Solder and Other 1984, p 343-365
296. M.G. Nicholas and C.F. Old, Review: Metals, J. Inst. Met., Vol 37, 1927, p 326. Y. Asayama, Metal-Induced Embrit-
Liquid Metal Embrittlement, J. 215-221 tlement of Steels, in Embrittlement by
Mater. Sci., Vol 14, Jan 1979, p !-18 311. H. Schottky et al., The Red-Shorten- Liquid and Solid Metals, American
297. N.S. Stoloff, Recent Developments in ing of Steels by Metals, Arch. Eisen- Institute of Mining, Metallurgical, and
Liquid-Metal Embrittlement, in Envi- hiittenwes., Vol 4, 1931, p 541-547 Petroleum Engineers, 1984, p 317-331
ronment-Sensitive Fracture of Engi- 312. R.R. Hough and R. Roils, The High- 327. Y. Iwata et al., Delayed Failure of
neering Materials, American Institute Temperature Tensile Creep Behavior Cadmium Plated Steels at Elevated
of Mining, Metallurgical, and Petro- of Notched, Pure Iron Embrittled by Temperatures, Jpn. Inst. Met., Vol 31,
leum Engineers, 1979, p 486-518 Liquid Copper, Scr. Metall., Vol 4, 1967, p 77-83
298. M.H. Kamdar, Liquid Metal Embrit- Jan 1970, p 17-24 328. D.N. Fager and W.F. Spurr, Solid
tlement, in Embrittlement of Engineer- 313. R.R. Hough and R. Rolls, Creep Frac- Cadmium Embrittlement: Steel Al-
ing Alloys, Vol 25, Treatise on Mate- ture Phenomena in Iron Embrittled by loys, Corrosion, Vol 27, Feb 1971, p
rials Science and Technology, Liquid Copper, J. Mater. Sci., Vol 6, 72-76
Academic Press, 1983, p 361-459 1971, p 1493-1498 329. S. Mostovoy and N.N. Breyer, The
299. N.S. Stoloff, Liquid and Solid Metal 314. R.R. Hough and R. Rolls, Copper Dif- Effects of Lead on the Mechanical
Embrittlement, in Atomistics of Frac- fusion in Iron During High-Tempera- Properties of 4145 Steel, Trans. ASM,
ture, Plenum Press, 1983, p 921-953 ture Tensile Creep, Metall. Trans., Vol 61, 1968, p 219-232
300. N.S. Stoloff, Metal Induced Vol 2, Sept 1971, p 2471-2475 330. R.D. Zipp et al., A Comparison of
Embrittlement--A Historical Perspec- 315. R.R. Hough and R. Rolls, Some Fac- Elevated Temperature Tensile Frac-
tive, in Embrittlement by Liquid and tors Influencing the Effects of Liquid ture in Nonleaded and Leaded 4145
Solid Metals, American Institute of Copper on the Creep-Rupture Proper- Steel, in Electron Microfractography,
Mining, Metallurgical, and Petroleum ties of Iron, Scr. Metall., Vol 8, Jan STP 453, American Society for Test-
Engineers, 1984, p 3-26 1974, p 39-44 ing and Materials, 1969, p 111-133
301. M.G. Nicholas, A Survey of Litera- 316. E.A. Asnis and V.M. Prokhorenko, 331. W.R. Warke and N.N. Breyer, Effect
ture on Liquid Metal Embrittlement of Mechanism of Cracking During the of Steel Composition on Lead Embrit-
Metals and Alloys, in Embrittlement Welding or Deposition of Copper Onto tlement, J. Iron Steel Inst., Vol 209,
by Liquid and Solid Metals, American Steel, Weld. Prod. (USSR), Vol 12 Oct 1971, p 779-784
Institute of Mining, Metallurgical, and (No. 11), 1965, p 15-17 332. N.N. Breyer, Some Effects of Certain
Petroleum Engineers, 1984, p 27-50 317. S.J. Matthews and W.F. Savage, Trace Elements on the Properties of
302. F.A. Shunk and W.R. Warke, Speci- Heat-Affected Zone Infiltration by High Strength Steels, in Proceedings
ficity as an Aspect of Liquid Metal Dissimilar Liquid Weld Metal, Weld. of the 31st Electric Furnace Confer-

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user
 736 / Service Characteristics of Carbon and Low-Alloy Steels

ence, Vol 31, American Institute of Cracking of Fe-Cr-Ni Base Alloys, Cor- 359. R.D. Kane, Role of H2S in Behavior of
Mining, Metallurgical, and Petroleum rosion, Vol 26, Nov 1970, p 451-486 Engineering Alloys, Int. Met. Rev.,
Engineers, 1974, p 183-189 346. N.A. Nielsen, Observations and Vol 30 (No. 6), 1985, p 291-301
333. N.N. Breyer and K.L. Johnson, Liq- Thoughts on Stress Corrosion Mecha- 360. R.F. Hehemann, Stress Corrosion
uid Metal Embrittlement of 4145 Steel nisms, J. Mater., Vol 5, Dec 1970, p Cracking of Stainless Steels, Metall.
by Lead-Tin and Lead-Antimony Al- 794-829 Trans. A, Vol 16A, Nov 1985, p 1909-
loys, J. Test. Eval., Vol 2, Nov 1974, p 347. R.N. Parkins, Stress Corrosion Crack- 1923
471-477 ing of Low-Strength Ferritic Steels, in 361. R.H. Jones and R.E. Ricker, Stress-
334. S. Dinda and W.R. Warke, The Effect The Theory of Stress Corrosion Crack- Corrosion Cracking, in Corrosion, Vol
of Grain Boundary Segregation on ing in Alloys, W.S. Maney & Sons, 13, 9th ed., Metals Handbook, ASM
Liquid Metal Induced Embrittlement 1971, p 167-185 INTERNATIONAL, 1987, p 145-163,
of Steel, Mater. Sci. Eng., Vol 24, 348. B.F. Brown, Stress Corrosion Crack- 187-188
1976, p 199-208 ing of High Strength Steels, in The 362. E.E. Denhard, Jr., Effects of Compo-
335. J.J. Heger, Stress-Corrosion of Stain- Theory of Stress Corrosion Cracking sition and Heat Treatment on the
less Steels, Met. Prog., Vol 67, March in Alloys, W.S. Maney & Sons, 1971, Stress Corrosion Cracking of Austen-
1955, p 109-116 p 186-204 itic Stainless Steels, Corrosion, Vol
336. H.H. Uhlig, New Perspectives in the 349. R.W. Staehle, Stress Corrosion 16, July 1960, p 359t-369t
Stress Corrosion Problem, in Physical Cracking of the Fe-Cr-Ni Alloy Sys- 363. H.R. Copson, Effect of Composition
Metallurgy of Stress Corrosion Frac- tem, in The Theory of Stress Corrosion on Stress Corrosion Cracking of Some
tures, Vol 4, AIME Metallurgical So- Cracking in Alloys, W.S. Maney & Alloys Containing Nickel, in Physical
ciety Conferences, Interscience, 1959, Sons, 1971, p 223-288 Metallurgy of Stress Corrosion Frac-
p 1-28 350. A.J. Sedriks, Stress-Corrosion Crack- ture, Vol 4, AIME Metallurgical Soci-
337. N.A. Nielsen, The Role of Corrosion ing of Stainless Steels and Nickel Al- ety Conferences, Interscience, 1959, p
Products in Crack Propagation in Aus- loys, J. Inst. Met., Vol 101, 1973, p 247-272
tenitic Stainless Steel: Electron Micro- 225-232 364. J.G. Hines and E.R.W. Jones, Some
scopic Studies, in Physical Metallurgy 351. R.W. Staehle, Stress Corrosion Effects of Alloy Composition on the
of Stress Corrosion Fractures, Vol 4, Cracking (and Corrosion Fatigue), Stress-Corrosion Behaviour of Austen-
AIME Metallurgical Society Confer- Mater. Sci. Eng., Vol 25, 1976, p itic Cr-Ni Steels, Corros. Sci., Vol 1,
ences Interscience, 1959, p 121-154 207-215 1961, p 88-107
338. J.G. Hines and R.W. Hugili, Metallo- 352. G.J. Theus and R.W. Staehle, Review 365. M.O. Speidel, Stress Corrosion Crack
graphic and Crystallographic Exami- of Stress Corrosion Cracking and Hy- Growth in Austenitic Stainless Steel,
nation of Stress Corrosion Cracks in drogen Embrittlement in the Austen- Corrosion, Vol 33, June 1977, p 199-
Austenitic Cr-Ni Steels, in Physical itic Fe-Cr-Ni Alloys, in Stress Corro- 203
Metallurgy of Stress Corrosion Frac- sion Cracking and Hydrogen 366. K.L. Money and W.W. Kirk, Stress
tures, Vol 4, AIME Metallurgical So- Embrittlement of Iron-Base Alloys, Corrosion Cracking Behavior of
ciety Conferences, Interscience, 1959, National Association of Corrosion En- Wrought Fe-Cr-Ni Alloys in Marine
p 193-226 gineers, 1977, p 845-892 Atmosphere, Mater. Perform., Vol 17,
339. H.L. Logan, Stress Corrosion Crack- 353. G.E. Moiler, Designing With Stainless July 1978, p 28-36
ing in Low Carbon Steel, in Physical Steels for Service in Stress Corrosion 367. C. Edeleanu, Stress Corrosion Crack-
Metallurgy of Stress Corrosion Frac- Environments, Mater. Perform., Vol ing in Austenitic Stainless Steels, in
tures, Vol 4, AIME Metallurgical So- 16, May 1977, p 32-44 Stress Corrosion Cracking and Em-
ciety Conferences, Interscience, 1959, 354. A.J. Sedriks, Corrosion of Stainless brittlement, John Wiley & Sons, 1956,
p 295-310 Steels, John Wiley & Sons, 1979, p p 126-139
340. J.F. Bates and A.W. Loginow, Princi- 139-193 368. T. Suzuki et al., Austenitic-Ferritic
ples of Stress Corrosion Cracking as 355. H.E. Hanninen, Influence of Metallur- Stainless Steels With High Chromium
Related to Steels, Corrosion, Vol 20, gical Variables on Environment-Sensi- Low-Nickel Content, Nippon Kinzoku
June 1964, p 189t-197t tive Cracking of Austenitic Alloys, Gakkaishi, Vol 32 (No. 11), 1968, p
341. R.N. Parkins, Stress-Corrosion Crack- Int. Met. Rev., Vol 24 (No. 3), 1979, p 1171-1177
ing, Metall. Rev., Vol 9 (No. 35), 1964, p 85-135 369. K.C. Thomas et al., Stress Corrosion
201-260 356. B.M. Gordon, The Effect of Chloride of Type 304 Stainless Steel in Chloride
342. P.R. Swann and J.D. Embury, Micro- and Oxygen on the Stress Corrosion Environments, Corrosion, Vol 20,
structural Aspects of Stress-Corrosion Cracking of Stainless Steels: Review March 1964, p 89t-92t
Failure, in High-Strength Materials, of Literature, Mater. Perform., Vol 370. F.K. Bloom, Stress Corrosion Crack-
John Wiley & Sons, 1965, p 327-362 19, April 1980, p 29-38 ing of Hardenable Stainless Steels,
343. H.L. Logan, The Stress Corrosion of 357. J.E. Truman, Stress-Corrosion Crack- Corrosion, Vol I!, Aug 1955, p 351t-
Metals, John Wiley & Sons, 1966 ing of Martensitic and Ferritic Stain- 361t
344. R.M. Latanision and R.W. Staehle, less Steels, Int. Met. Rev., Vol 26 (No. 371. R.R. Gaugh, Sulfide Stress Cracking
Stress Corrosion Cracking of Iron- 6), 1981, p 301-349 of Precipitation Hardening Stainless
Nickel-Chromium Alloys, in Funda- 358. F.P. Ford, Stress Corrosion Cracking Steels, Mater. Perform., Vol 16, Sept
mental Aspects of Stress Corrosion of Iron-Base Alloys in Aqueous Envi- 1977, p 24-29
Cracking, National Association of ronments, in Embrittlement of Engi- 372. V.L. Barnwell et al., Effect of Grain
Corrosion Engineers, 1969, p 214-307 neering Alloys, Vol 25, Treatise on Size on Stress Corrosion of Type 302
345. R.W. Staehle et al., Effect of Alloy Material Science and Technology, Ac- Austenitic Stainless Steel, Corrosion,
Composition on Stress Corrosion ademic Press, 1983, p 235-274 Vol 22, Sept 1966, p 261-264

Downloaded from http://dl.asminternational.org/handbooks/edited-volume/chapter-pdf/445938/a0001039.pdf

by Ben Gurion University of the Negev user