WRC 532-2009

INTENTIONALLY LEFT BLANK
WRC - The Welding Research Council brings together science and engineering specialists in
developing the solutions to problems in welding and pressure vessel technology. They exchange
knowledge, share perspectives, and execute R and D activities. As needed, the Council organizes
and manages cooperative programs.
MPC – A Council of the WRC, the Materials Properties Council is dedicated to providing
industry with the best technology and the best data that can be obtained on the properties of
materials to help meet today’s most advanced concepts in design and service, life assessment,
fitness-for-service, and reliability and safety.
PVRC – A Council of the WRC, the goal of the Pressure Vessel Research Council is to
encourage, promote and conduct research in the field of pressure vessels and related pressure
equipment technologies, including evaluation of materials, design, fabrication, inspection and
testing.
For more information, see www.forengineers.org
WRC Bulletins contain final reports from projects sponsored by the Welding Research Council, important
papers presented before engineering societies and other reports of current interest.
No warranty of any kind expressed or implied, respecting of data, analyses, graphs or any other
information provided in this publication is made by the Welding Research Council, and the use of any
such information is at the user’s sole risk.
All rights are reserved and no part of this publication may be reproduced, downloaded, disseminated, or
otherwise transferred in any form or by any means, including photocopying, without the express written
consent of WRC.
Copyright © 2009 The Welding Research Council.

All rights, including translations, are reserved by WRC.
Printed in the United States of America.
ISSN 0043-2326
Library of Congress Catalog Number: 85-647116
Welding Research Council

20600 Chagrin Blvd.
Suite 1200
Shaker Heights, OH 44122
www.forengineers.org
The Welding Research Council, Inc.
ETALLURGICAL STUDIES OF STEELS FOR

SOUR SERVICE ENVIR NMENTS
Martin Prager, Editor
WRC Bulletin 532
WELDING RESEARCH COUNCIL, INC.

Box 1942
New York, NY 10156
www.forengineers.org
WRC Bulletin 532

Metallurgical Studies for Steels for Sour Service Environments
ISBN No. 1-58145-539-9
Library of Congress
Catalog Card Number: 85-647116
Copyright © 2009 by
Welding Research Council, Inc.
All Rights Reserved
Printed in U. S. A
WRC Bulletin 532

ii Metallurgical Studies for Steels for Sour Service Environments
FOREWORD
In recent years there has been increasing need to acquire technology for the safe handling
of hydrogen for petroleum and other energy related applications. The pace ofresearch into
metal/hydrogen interactions accelerated at the start of the space age due to the hydrogen's
efficiency per unit weight as a rocket fuel. Practical applications called for hydrogen to be
transported and used at high pressures thereby increasing its hazard potential. Exotic and
expensive alloys could be used to put astronauts into space because cost was of little concern
where safety was an issue. For most conventional applications steel is still the practical choice.
It has been long recognized that at ordinary and elevated temperatures hydrogen can be
destructive to steels, not just to high-strength steels, but also to steels of ordinary strength levels.
Hydrogen's capability to enter and then diffuse through the metallic lattice, accelerated by stress
gradients and seeking out points of weakness where it can concentrate or react, renders it capable
of destroying pressure retaining metals from the inside where damage defies detection and until
it is too late and the pressure containment as failed or is no longer safe.
Considerable research aimed at understanding hydrogen-steel interactions has been

conducted over the last half century. An important objective has been to permit the reliable use
of steels of higher strength levels in aggressive hydrogen environments. Tied to this objective is
the necessity of fabricating the higher strength steels without cold cracking due to hydrogen
introduced during welding. problems and solutions are complex because of the diverse
rnicrostructures and compositions that have been developed to achieve the performance goals set
for steels.
This Welding Research Council (WRC) Bulletin is part of a series that captures the
essential studies of the interaction of steel-hydrogen interactions in recent years. Topics include
Modern Vanadium Steels.for High Temperature Petroleum Reactors ( # 524), Fabrication and Repair of
Low Alloy Steel Pressure Equipment ( # 525), Performance of Steels in Hydrogen Charging
Environments ( # 526), Practical Aspects of Hydrogen Attack ( # 527), Test Method~ for Hydrogen
Induced Cracking ( # 530), Metallurgical Studies of Steels for Sour Service Environments ( # 532),
Studies of Cladding and Overlay for Pressure Vessel Service ( # 534), and Toughness, Fracture and
Fitness.for Hydrogen Service ( #
The papers included have been presented at international conferences sponsored by WRC' s sister
organization the Materials Properties Council Inc. (MPC). The technology reported in this series provides
a comprehensive view of practical solutions to engineering problems and advances in the knowledge
about hydrogen and steel interactions.
WRC Bulletin 532

Metallurgical Studies for Steels for Sour Service Environments iii
WRC Bulletin 532

iv Metallurgical Studies for Steels for Sour Service Environments
Table of Contents
REFINERY EXPERIENCE WITH CRACKING OF PROCESS VESSELS EXPOSED TO WET
H 2 S ENVIRONMENTS .................................................................................................................................................... 1
M. L. Bullen, M. J. Humphries, R. D. Merrick
DEVELOPMENT OF PRESSURE VESSEL STEEL SERIES FOR WET H 2 S REFINERY SERVICE .. .11
A. Streisselberger, V. Schwinn
QUENCHED AND TEMPERED STEELS FOR H 2 S CONTAINING PRESSURE VESSELS .............. 27

L. Coudreuse, C. Zrnudzinski, P. Bocquet, A. Cheviet
INFLUENCE OF ACCELERATED COOLING ON THE HYDROGEN INDUCED CRACKING

AND SULPHIDE STRESS CRACKING BEHAVIOR OF LINEPIPE STEELS WITH A LOW
CARBON CONTENT .................................................................................................................................................... 35
H. Margot-Marette, B. Charnont, E. Arnoris, and C. Perdrix
EFFECT OF SOME METALLURGICAL FACTORS ON THE SUSCEPTIBILITY TO SSCC OF

X60 TO X70 PIPELINE STEELS ............................................................................................................................... 43
F. Terreur, G. Monfort, and J. Defourny
INFLUENCE OF THERMAL TREATMENT ON THE HYDROGEN EMBRITTLEMENT OF

A-516 GR. 60 STEEL ..................................................................................................................................................... 51
G. Anteri, G. Merlone, and J. Ovejero Garcia
CARBON MANGANESE STEELS FOR SOUR SERVICE - IMPROVEMENT OF HIC AND SSC
RESISTANCE .................................................................................................................................................................. 59
L. Coudreuse, C. Zrnudzinski, J. Charles, and L. Cadiou
ON THE EFFECT OF NON-METALLIC INCLUSIONS ON HYDROGEN DAMAGE IN

AL-KILLED STEEL ..................................................................................................................................................... 69
Estela R. de Schiapparelli, Sixto Prado
STUDY THE SULFIDE STRESS CRACKING OF LOW ALLOY AND C-Mn STEELS ................73
H. Margot-Marette, J.C. Charbonnier, L. Coudreuse, and J. Charles
INFLUENCE OF MICROSTRUCTURE ON DIFFUSION AND TRAPPING HYDROGEN

IN LOW CARBON STEEL HEAT AFFECTED ZONES ..................................................................................... u i
E. Bollinger, C. and P. Manolatos
THE INFLUENCE OF SEGREGATION ON HIC IN PRESSURE VESSEL STEELS ............................. 89

W. B. Morrison and K. Abernethy
WRC Bulletin 532

Metallurgical Studies for Steels.for Sour Service Environments v
WRC Bulletin 532

vi Metallurgical Studies for Steels for Sour Service Environments
REFINERY EXPERIENCE WITH CRACKING OF

PROCESS VESSELS EXPOSED TO WET H2S ENVIRONMENTS
M. L Bullen
Exxon Research & Engineering Co.
Florham Park, New Jersey, U.S.A.
M. J. Humphries
Essa Production Malaysia
Kemasek, Malaysia
R. D. Merrick
Exxon Research & Engineering Co.
Florham Park, New Jersey, U.S.A.
ABSTRACT and promotes their absorption into the steel. The

hydrogen atoms, because of their small size, diffuse
Cracking of equipment exposed to wet H2S has through the steel. The generation and absorption of
recently become of drastically increased concern to the atomic hydrogen is enhanced by the presence of cyanides
Petroleum Refining Industry. Although the oil produc- in waters. For this reason the effects of hydrogen
tion industry (particularly pipelines) has experienced charging are most common in cyanide producing processes,
such cracking, it was believed that control of hardness such as fluid catalytic cracking and coking units.
would eliminate cracking in refining environments. The However, it has been found wherever a separate water
1984 catastrophic failure of a pressure vessel at an oil phase containing hydrogen sulfide (wet H2S) is present;
refinery in Chicago led to a complete change in atti- even in the vapor space of crude storage tanks. Hydro-
tude. Two previously, little-recognized factors were gen charging does not require hydrogen in the process
identified. First, control of hardness is insufficient stream.
to control through-wall cracking in wet H2S environ-
ments. And, secondly, hydrogen can cause significant Hydrogen Cracking
embrittlement of soft pressure vessel carbon steels. The effect of hydrogen absorbed by carbon steels is
well known and extensively reported in the literature
Subsequent to the failure, extensive inspection (3,4,5,6,7). The primary effects are:
programs in the oil refining industry have revealed a • Cracking in high strength or high hardness steels
substantial incidence of cracking in wet H2S services (Sulfide Stress Cracking).
with vessels built to controlled hardness standards • Blister formation originating at non-metallic
(NACE RP-04-72). This paper presents Exxon's strategy inclusions (Hydrogen Blistering).
for equipment inspection, the findings of the inspection • Hydrogen induced cracking producing step-wise cracks
program, metallurgical analyses of the cracks and the to the metal surface (HIC).
results of a limited laboratory study on the effect of • Loss of ductility under a slow tensile loading
stress on one wet H2S cracking mechanism, namely SOHIC. (Hydrogen Embrittlement).
SOHIC, discussed in more detail, can lead to through-
wall cracks by link-up of blisters. Of these, sulfide stress cracking and hydrogen embrit-
tlement have the potential to produce thru thickness
Much of the information in this paper was previ- cracking in pressurized equipment. With the relatively
ously published by the National Association of Corrosion low strength carbon and carbon-manganese steels used in
Engineers (1). A second study at Exxon investigated the pressure vessel construction, the occurrence of sulfide
fracture toughness on hydrogen charged low strength stress cracking is controlled by limiting the hardness
carbon steels. The results of this second program are to 200 BHN per NACE RP-04-72 (5). By imposing the hard-
discussed in a separate paper (2). ness limits the cracking experienced on pressure welds
(6) has been arrested. However, many of the older ves-
BACKGROUND sels were constructed before this ~riteria was estab-
lished. Hydrogen embrittlement has not resulted in
Hydrogen Charging cracking of equipment during the normal operating cycle.
Before discussing the basis for the inspection However, if repairs are made to the equipment the
recommendations a brief review will be made of how stresses that result from weld shrinkaqe can be suffi-
hydrogen is charged into the steel and its effects. cient to cause cracking unless proper procedures are
Hydrogen damage is not a new finding; it has been followed. These procedures generally include baking at
recognized since the late 1940's when it became trouble- about 4500F (2300C) to drive out the dissolved hydrogen
some in the fluid catalytic cracking unit light ends before welding. Since hydrogen cracking could be pres-
equipment. When steel is in contact with water contain- ent in a number of refinery services, the inspection
ing hydrogen sulfide, it will corrode producing hydrogen program addressed pressure vessels in these services.
as will happen in many acids. However, the HzS inhibits
the hydrogen atoms from combining to form hydrogen gas
WRC Bulletin 532

Metallurgical Studies for Steels for Sour Service Environments 1
Inspection Program Since the inspection was directed at vessels con-

The inspection program that was developed is pre- taining wet HzS, this environment was common to all
sented in Tables 1 and 2. This program was for pressure cracked and uncracked vessels. The quantity of H2S,
vessels (including shell and tube heat exchangers) fab- when known, varied from 85 wppm to 14 wt% in the vessels
ricated of ferritic steels. It was developed consider- that cracked. However, for over 70% of the cases the
ing the probability of cracking based upon the process quantity is unknown. Therefore, no true lower limit can
environment and equipment history. The priority for be established. Cyanide was present in about 60% of the
inspection is given by the order of listing and wet H2S cases. The use of polysulfide, water wash or inhibitors
plus cyanides takes priority over wet H2S for each cate- to control corrosion and hydrogen activity apparently
gory. had no significant effect on the incidence of cracking.
The reasoning that formed the basis for the need Visual Inspection Findings
and priority of inspection follows: The nature of the cracking can be described by the
visual appearance as recorded by the inspectors. The
Efforts should concentrate on equipment with a prior following are typical observations:
history of hydrogen induced problems.
• Cracking of weld repairs to the pressure boundary • Cracking on bottom side of tray support weld in HAZ.
after the vessel has been in service and possibly • Nozzle welds with cracking both along the weld and
charged with hydrogen is most probable. perpendicular to the weld in the shell plate.
• Cracking of non-stress relieved equipment in wet H2S • Downcomer to shell welds. Intermittent longitudinal
and cyanide service was found at one refinery. cracks all in the welds. Cracks as deep as the
welds.
Inspection Technique • Numerous new cracks found after welding and stress
The inspection of pressure vessels in wet hydrogen relief.
sulfide environments was conducted with the sensitive • Cracks both parallel and across pressure weld seams.
wet fluorescent magnetic particle technique (WFMPT)
because this technique had proven to be the most sensi- As indicated by these observations and the results of
tive to tight surface breaking cracks. WFMT inspection, the cracking has been limited to the
areas adjacent to weldments. However, in one case
In an investigation of amine cracking, cracks that cracking was found between two welds in an area of high
were approximately 0.5 inch (13 mm) deep were not stress.
detected visually or with dye penetrant or dry powder
magnetic particle techniques. The cracking was, how- Metallographic Study
ever, clearly evident when the wet fluorescent magnetic To gain a better understanding of the cracking,
particle inspection (WFMT) method was used. That these samples were taken from a number of vessels. This is a
cracks went undetected prompted a change in inspection on-going program to further define the cracking. How-
practices for equipment in other services. Wet fluores- ever, the samples examined to date have provided signif
cent magnetic particle inspection was therefore applied icant insight into the causes of cracking.
to equipment in wet H2S service since it could contain
similar tight, surface-breaking cracks. Sulfide Stress Cracking. Cracking found associated
with the repair of a main seam weld is shown in Figure
RESULTS 1. This photomicrograph shows two separate cracks; one
surface breaking which was detected by WFMT and a sub-
Inspection Findings surface crack discovered during the metallographic exam-
ination. This sample was from an Absorber/Deethanizer
Inspections in accordance with this program have in a fluid catalytic cracking light ends system. The
been conducted during the scheduled downtime of units in surface breaking crack is shown in detail in Figure 2.
a number of refineries worldwide. A summary of the The microhardness readings (converted to Brinell values)
inspection findings for 588 pressure vessels (including are shown. This crack, which is transgranular, initiat-
shell and tube heat exchangers) is presented in Table 3. ing in a hard HAZ and stopping in soft base metal, is
These represent refineries in the U.S.A., Canada, Europe typical of sulfide stress cracking. The subsurface
and Japan. The amount and severity of cracking found crack, shown in Figure 3, is generally intergranular and
was a surprise. Cracks that were deeper than the corro- is not limited to a hard microstructure. With the
sion allowance were found on 29% of the vessels exception of lower hardness, this could be considered
inspected. 80% of these cracks were more than 25% of cold cracking from hydrogen absorbed in the steel before
the vessel wall thickness The longest crack was about the repair weld was made. However, as shown in Figure
10 ft. The cracks were primarily in pressure and nozzle 4, this crack has the characteristics of hydrogen
welds and were parallel to the weld in the heat affected induced cracking (HIC). Of interest is the end of the
zone (HAZ} or adjacent base metal. Since the location sulfide stress crack shown in Figure 5. The tip of the
of the cracks was, for the most part, reported based crack is blunt and forked, indicating that it has not
upon visual examination, the data only indicates that actively lengthened in the immediate past. Therefore
the cracks are near but not in weldments. this crack may have existed in the vessel for an
extended period of time.
It should be noted that none of these vessels were
inspected by WFMT after original fabrication, so some of Sulfide stress cracking was found in another vessel
the smaller defects may date back to that time. as transverse cracks in the weld metal of a main spam
weld. A typical crack with converted microhardness
Contrary to expectations only a few cracks were readings is shown in Figure 6. This crack occurred in
reported where field weld repairs had been made. Like- the surface weld pass which had a hardness greater than
wise, cracking was found in about 20% of the vessels 250 BHN equivalent. This was a manual weld. Of inter-
that had been stress relieved. est was that a standard laboratory Brinell test gave a
hardness of 205 BHN. A Rockwell B test gave the same
WRC Bulletin 532

2 Metallurgical Studies for Steels for Sour Service Environments
equivalent hardness of 205 BHN. These data suggest that operation and discontinuity stresses at nozzles and
cracking can be expected even though the macrohardness attachments in a PWHT'd vessel. The performance of the
criteria of NACE RP-04-72 are met. steels at the 50% 'fS is felt to be the key to assessing
whether the PWHT can prevent SOHIC. The 903 yield
Intergranular Cracking. Inspection of the shell of stress level is typical of a non-PWHT weldment and was
a heat exchanger cooling the overhead stream from the selected to demonstrate the worse case cracking.
fractionator on a fluid catalytic cracking unit found
numerous cracks near main seam welds and between the It should be noted that the tests were designed to
welds attaching a slip-on flange. This cracking was produce SOHIC in a short time period (30 days) and so
unlike the previous samples examined. The cracks were employed severe test conditions. First, the NACE
in base metal with an "as-rolled" microstructure, and TM-01-77 environment has a high hydrogen activity.
were intergranular and branched as shown in Figures 7 Secondly, the test specimens were completely immersed in
and 8. The deposits in the cracks were primarily oxides the test solution. In a pressure vessel only one side
with only a minor amount of sulfur present. In addition of the vessel wall is exposed to the wet H2S environ-
to the large crack there were a number of shallow cracks ment. Lastly, the test specimens are very thin (0.25
with similar morphology. The microhardness reading inches [6 mm]). In contrast, pressure vessel walls are
(converted to BHN) around the crack was less than 150 typically 0.5 inch (12 mm) or thicker.
BHN.
Test Results
This cracking is not typical of sulfide stress The tests were evaluated by sectioning the specimen
cracking but does have the characteristics of an anodic longitudinally and examining the microstructure to
mechanism such as caustic, amine and carbonate cracking. determine quantity and the degree of stacking of inter-
Of these the only possibility is carbonate cracking from nal blister cracks. The stacking of the internal
the C02 produced in the cracking process. However, this blister cracks has been observed as a precursor to SOHIC
needs to be confirmed by similar experiences in other (13). After the stacking, link-up of the blisters and,
refineries. ultimately, failure occurs. For this reason conditions
which cause stacking of the internal blister cracks are
Hydrogen Induced Cracking. A second sample from considered to cause SOHIC failures as well.
the Absorber Oeethanizer showed that crack extension by
HIC was also occurring. This is shown in Figures 9, 10 Reviewing the photomicrographs of the cross-sec-
and 11. The crack shown in Figure 9 initiated at a very tions (Figure 12) for the two solutions, it is apparent
hard zone (> 300 BHN) associated with a repair weld the NACE TM-01-77 solution is the more severe environ-
adjacent to an original weld. This was a sulfide stress ment. More internal blistering is observed with the
~rack which changed from transgranular to HIC when it NACE TM-01-77 solution and stacking of the blisters is
~ntered the low hardness base metal. The extension of observed at lower (303 of the yield strength) for the
;he cracking by HIC is clearly shown in Figure 10 and ASTM A285 grade steel. However, for both steels and in
1. This type of crack extension has been described in both solutions, stacking of the internal blisters was
Ghe literature (9,10,11). As illustrated by this observed at stress levels of 503 of the yield strength
sample, small "hydrogen blisters" tend to stack in the and above.
through thickness direction at areas of stress concen-
tration. In this case the stress concentration is DISCUSSION
provided by the sulfide stress crack. In the investiga-
tion of a pipeline failure by this mechanism the phenom- The findings from the inspection program have shown
ena was aptly called SOHIC, stress oriented hydrogen that significant cracks exist in pressure vessels in wet
induced cracking (ll). SOHIC has been found frequently hydrogen sulfide service which have not previously been
in the common API grades of pipeline steels but has only detected by the commonly-used inspection methods. The
recently been reported associated with a pressure vessel application of WFMT inspection will detect these cracks.
failure (12). Since there is no technique to establish the age of the
cracks found, the blunt ends of the cracks, Figures 2
Laboratory Testing and 5, would suggest that they have been present for
A limited laboratory test program was run to eval- some time. It can be concluded that the cracks formed
uate the threshold stress required to cause SOHIC. early in the life of the equipment and had ceased to
Specifically testing was aimed at determining whether propagate. The cracking is only now being detected by
postweld heat treatment (PWHT) would prevent SOHIC. Two applying an improved inspection tool. The question
steels (an ASTM A-285 and A-515 grade steel), from remains, 'how detrimental is the cracking?'.
scrapped pressure vessels, were exposed to stressed
conditions per the procedures in NACE TM-01-77, Sulfide Stress Cracking
of Metals for Resistance to Sulfide Stress Cracking It has been the belief of the petroleum refining
Ambient Temperatures. In addition to the acidified, industry that controlling the hardness of a steel to 200
HzS-saturated salt solution specified in NACE TM-01-77 a BHN would prevent sulfide stress cracking in process
synthetic refinery solution was also used. This was an equipment. The philosophy is incorporated into both
alkaline solution (pH between 7.8 and 8.2) containing NACE Standard RP-04-72 and API Publication 942 with the
ammonium sulfide ([NH4]2S), potassium cyanide (KCN) and testing conducted by a standard Brinell test using a 10
saturated with H2S (the solution contained 1.0 wt% sul- mm ball. The sulfide stress cracks found in this studv
fide and 0.68 wt% cyanide). Proof Rings, supplied by were in weldments that would pass the Brinell test. The
CorTest, Inc., were used to stress the steels. hard zones were only detected by a microhardness test.
These hard zones also were not detected by a Rockwell B
Three different stress levels, 303, 503, and 903 of hardness test. These documents also imply that the
the actual yield stress (YS), were tested. These stress carbon and carbon-manganese steels will achieve these
levels were uniform across the test specimen. The 303 hardnesses with normal welding procedures using shield
yield stress was chosen to simulate the residual stress metal arc welding. The samples reported here showed
level in a PWHT'd vessel. The 503 yield stress was cho- hardness in both the HAZ 2 and 9) and in
sen to simulate the combined stress level from residual,
WRC Bulletin 532

the weld metal itself (Figure 6). In these hard areas The Role of PWHT in Eliminating Cracking
sulfide stress cracks initiated. They did, however, The observation that the narrow hard areas associ-
stop when entering soft metal. ated with weldments are susceptible to sulfide stress
cracking has led us to the conclusion that vessels
Intergranular Cracking exposed to wet H2S service should be PWHT'd. PWHT will
The intergranular cracking was not expected in the minimize the probability of cracking from sec by reduc-
catalytic cracking fractionator overhead exchangers. ing both the hardness and residual stresses, both pre-
The cracks look like those found in amine or caustic requisites for SCC. (We have defined wet H2S as a
service. However, neither of these could have been process environment containing water with more than 50
present. The other possibilities are cyanide or carbon- wppm of H2S.)
ate cracking since the fractionator overhead contains
both cyanide and carbon dioxide. Cracking when cyanides Similarly our experience and literature (15) have
are present is generally reported to be transgranular demonstrated that the "carbonate" intergranular stress
(14). Parkins (15) reported intergranular cracking with corrosion cracking mechanism is similar to other types
cyanides, hydrogen sulfide and carbon dioxide present. of alkaline cracking and can be controlled through PWHT.
However in recent discussions Parkins now believes the Furthermore, in the refining industry, equipment con-
cracking was primarily the result of the carbon dioxide taining carbon dioxide (i.e., Fluidized Catalytic
with the hydrogen sulfide and cyanides providing the Cracking Unit and Coker Fractionation and Light Ends
proper environmental conditions. An analysis of the equipment) will also contain more than 50 wppm of H2S.
scale in the cracks showed oxide as the bulk product As a result, equipment in carbonate equipment will be
with only traces of sulfides. The scale in the trans- PWHT'd as protection against sulfide stress cracking and
granular cracks in the hard microstructure was all sul- will not require a unique protection system against
fide. It has been concluded that the cracking is from intergranular ("carbonate" cracking).
the carbon dioxide in the process stream although this
has not been reported previously. In contrast, our laboratory testing has shown that
PWHT is not sufficient to make a vessel immune to SOHIC.
Hydrogen Induced Cracking However, by preventing sulfide stress cracking and
Extension of the sulfide stress cracks by HIC is intergranular (carbonate) cracking (by PWHT), it does
most significant. As mentioned previously this mode of reduce the likelihood of introducing cracks which act as
crack extension has only been recently reported (12) in stress concentrators, a contributor to SOHIC. For
a pressure vessel. The presence of SOHIC in several of maximum confidence alternative approaches are required
the samples examined indicates it is not "one-of-a-kind" to prevent SOHIC. For new projects we have approached
related to unusual conditions. All the conditions it in one of two ways. One approach has been to use a
necessary for SOHIC as presented by Ikeda (13) and Iino clad Type 304 stainless steel construction in wet H2S
(9) were found to exist. These were: services. The Type 304 stainless steel lining prevents
corrosion of the carbon or ferritic steel by the wet H2S
• Soft microstructure environment and thereby eliminates the concern of SOHIC.
• Stress concentrator This approach provides a high degree of confidence that
• Residual welding stress the base metal will be protected, but is also very
• High hydrogen activity costly. Another approach we are exploring is the use of
HIC resistant steels. The philosophy is that these
The crack extension has occurred in the soft base metal "clean" steels do not provide any traps for hydrogen to
where the sulfide stress crack has terminated. This collect and form internal blisters. As indicated, we
sharp crack in non-stress-relieved weldments will con- are still experimenting to define our requirements for a
centrate both the residual welding stresses as well as HIC resistant steel in refinery services. The HIC
the applied process stress. These vessels had a high resistant steel will also be PWHTd as discussed above.
hydrogen activity as shown by the hydrogen blistering
found. The sulfide stress cracking in the shallow, hard CONCLUSIONS
heat affected zones becomes more significant when their
potential to extend by SOHIC is recognized. • The cracking found by WFMT does not represent a new
deterioration mode. It is the detection of cracks
In the limited laboratory test program the presence not detectable with the inspection methods typically
of stacked internal blisters at stress levels in the 50% used in refinery pressure vessels previously.
of yield stress range demonstrates that postweld heat
treatment (PWHT) of equipment will not eliminate SOHIC. • Significant cracks can initiate from very small hard
Equipment which has been PWHT'd will have lower residual zones associated with weldments. The initially
stresses overall. However, at nozzles and other attach- small cracks can extend by the link-up of hydrogen
ments, a stress level above 50% of the yield stress is blisters to create a through-thickness crack.
likely. Consequently, a vessel with PWHT will be sus-
ceptible to SOHIC. • Stress oriented hydrogen induced cracking (SOHIC)
which is common in pipeline steels can occur in the
We acknowledge that the laboratory tests required common grades of pressure vessel steels.
severe testing conditions, most notably high hydrogen
charging rates, to obtain stacking of the internal • Monitoring the hardness of weld metal and heat
blisters at the low stress conditions. However, we affected zone with the Brinell test is not adequate
continue to believe that PWHT is inadequate to prevent to prevent sulfide stress cracking. The laboratory
SOHIC. It is our philosophy that at lower hydrogen and field Brinell test did not detect the small hard
charging (low stress conditions) and longer times that zones that initiated cracks.
stacking of the internal blister cracks may also occur.
This has not yet been confirmed by our inspections. •Thermal stress relief (postweld heat treatment)
appears to be an effective method to reduce sulfide
stress cracking and intergranular cracking.
WRC Bulletin 532

• Postweld heat treatment (PWHT) of a vessel is not 15. R.N. Parkins, R. Usher, JOURNAL OF APPLIED CHEMIS-
sufficient to make it immune to SOHIC. Even with TRY, No. 9, September 1959, p. 445.
PWHT there are areas in the vessei with stresses
above the 50% yield stress level where the initia-
tion of cracking has been repeatedly observed in the
accelerated laboratory tests. An alternative
approach is required to prevent SOHIC.
REFERENCES
l. R.D. Merrick, MATERIALS PROTECTION, January 1988,
p. 30.
2. M.J. Humphries, J.E. Mclaughlin, L.J. Pargeter,

"Toughness Characteristics of Hydrogen Charged
Pressure Vessel Steels", International Conference
on Interaction of Steels with Hydrogen in Petroleum
Industry Pressure Vessel Service, March 28, 1989,
MPC, Paris, France.
3. R.T. Effinger, M.L. Renquist, A. Wachter and J.G.
Wilson, "Hydrogen Attack of Steel in Refinery
Equipment", 16th Mid-Year Meeting of American
Petroleum Institute's Division of Refining, May 1,
1951, API Washington.
4. G.A. Nelson, Welding Research Supplement, January
1955, p. 12-s. Fig. 1 Cracking found in an Absorber/Deethanizer in
FCCU Light Ends Unit. Surface breaking crack
5. C.M. Hudgins, R.K. McGlasson, P. Mehdizadeh and shown in detail in Figure 2; the subsurface
W.M. Rosborough, CORROSION, Vol. 22, August 1966, crack in Figure 3. Cracks in base metal adja-
p. 238. cent to a repair weld. Nital etch. 6X
6. H.C. Rogers, MATERIALS PROTECTION, April 1962, p.

26.
7. NACE Standard RP-04-72, NACE,Houston, TX.
8. E.L. Hildebrand, "Aqueous Phase H2S Cracking of
Hard Carbon Steel Weldments - A Case History", API
Meeting, May 1970, API, Washington, D.C.
9. Makio Iino, METALLURGICAL TRANSACTIONS, Vol. 9A,
November 1978, p. 1581.
10. I. Class, International Conference on Metallic
Corrosion, 1963, p. 342.
11. W. Bruckoff, G. Schmitt, D. Steinmetz, "Rupture of
A Sour Gas Line Due to Stress Oriented Hydrogen
Induced Cracking Failure Analysis, Experimental
Results and Corrosion Prevention", CORROSION/85,
Paper No. 389, National Association of Corrosion
Engineers, Houston, TX, 1985.
12. H.J. McHenry, T.R. Shives, D.T. Read, J.D.
McColskey, C.H. Brady, P.T. Portscher, "Examination
of A Pressure Vessel That Ruptured at the Chicago
Refinery of the Union Oil Company on July 23,
1984", Rept. No. NBSIR 86-3049, March 1986,
National Bureau of Standards, Boulder, CO 80303.
13. A. Ikeda, T. Kaneko, T. Hashimoto, M. Takeyama, Y.
Sumitomo, T. Yamura, "Development of Hydrogen
Induced Cracking (HIC) Resistant Steels and HIC
Test Methods for Hydrogen Sulfide Service", CIM
Conference, Paper No. 6, Session 11, August 1983, Fig 2. Surface breaking crack from Figure I. Micro-
Edmonton, Canada. hardness values shown in Brinell units next to
Knoop 500 gm indentations. The crack is a
14. Von H. Buchholtz, R. Pusch, Stahlund Eisen, Vol. typical sulfide stress crack terminating in soft
62, No. 2, January 8, 1942, p. 21. base metal. Nital etch. 30X
WRC Bulletin 532

l with microhard-
Crack appears to
be intergranular with transgranular features.
Not typical of sulfide stress cracking. Figure
4 shows detail of cracking. Nital etch. 20X
Fig. 6 Transverse cracking in the weld metal of a main

seam weld of a pressure vessel. Microhardness
values are shown in Brine1l units. This sulfide
stress crack terminated in the soft previous
passes. Nital etch. 50X
End of crack shown in Figure 3.

hydrogen expanded laminations is
is typical of hydrogen induced cracking.
etch.
Fig. 7 Cracking found in she n of heat exc:hange·r

FCCU fractionator overhead system.
indicates ing is
past. 30X
WRC Bulletin 532

Fig. 10.
Fig. 8 Cracks near that 7 showing the

intergranu1 ar nrrrn;ln;;t. oxide in cracks.
Nita l etch.
Fig. 11. 10 but

12X
WRC Bulletin 532

JVli?ta,rtwrr<z,cal ~'''"''•" v,•vv•U •v• Sour Service Environments
•v• 7
Fig. 12. SOHIC tests on Proof Rings.
WRC Bulletin 532

INSPECTION PROGRAM FOR EQUIPMENT EXPOSED

TO WET H?S > 50 WPPM AND CYANIDES > 20 WPPM
Equipment Ins ection
Existing Non-PWHT Vessels
History of cracking, blistering or WFMT, 100% of shell and head welds, internal attachment
HIC welds and weld repairs/alterations at next scheduled
turnaround.
Previous welded repairs/alterations WFMT, 100% of weld repairs and alterations and spot
(with/without PWHT) examination of other welds at the next scheduled turn-
around.
No prior welded repairs/alterations WFMT on a spot basis of shell and head welds in selected
vessels at next turnaround.
Existing PWHT Vessels
History of cracking WFMT, 100% of shell and head welds, internal attachment
welds and weld repairs/alterations at next scheduled
turnaround.
History of blistering or HIC WFMT, 100% of weld repairs and alterations and spot
examination of other welds at next scheduled turnaround.
Previous welded repairs/alterations WFMT, 100% of weld repairs and alterations at next
(with/without PWHT) scheduled turnaround.
No prior welded repairs/alterations No special inspection required.
TABLE 2
INSPECTION PROGRAM FOR EQUIPMENT
EXPOSED TO WET H?S > 50 WPPM
Equipment Ins ection
Existing Non-PWHT Vessels
History of cracking, blistering or WFMT, 100% of shell and head welds, internal attachment
HIC welds and weld repairs/alterations at next scheduled
turnaround.
Previous welded repairs/alternations WFMT, 100% of weld repairs and alterations of selected
(with/without PWHT) vessels at next scheduled turnaround.
No prior welded repairs/alterations No special inspectin required.
Existing PWHT Vessels
History of cracking WFMT, 100% of shell and head welds, internal attachment
welds, and weld repairs/alterations at next scheduled
turnaround.
History of blistering or HIC WFMT, 100% of weld repairs and alterations and spot
examination of other welds at next scheduled turnaround.
Previous welded repairs/alterations WFMT, 100% of weld repairs and alterations of selected
(with/without PWHT) vessels at next scheduled turnaround.
No prior welded repairs/alterations No special inspection required.
WRC Bulletin 532

INSPECTION FINDINGS
Vessels Inspected 588

Cracks Found 244
Cracks Deeper than Corrosion Allowance 170
Crack Depth (% wall thickness)
< 25 62
25-49 53
50-74 43
75-100 41
Crack Length > 2 in (50 mm) 136

Longest: 10 ft (3. 05 m)
Location of Cracks:
In pressure and/or nozzle welds 261
In attachment welds 61
Vessels Stress Relieved and Cracked 52
Stream Composition for Cracked Vessels:
HzS quantity from 85 wppm to 14 wt%
Cyanide reported to be present in 63%
of cracked vessels
WRC Bulletin 532

JO Metallurgical Studies for Steels for Sour Service Environments
DEVELOPMENT OF PRESSURE VESSEL STEEL SERIES FOR

WET H2 S REFINERY SERVICE
A. Streisselberger, V. Schwinn
Research & Development
Dillinger Hutte GTS
66748 Dillingen/Saar
Germany
ABSTRACT on temperature and time for PWHT in order to avoid

impairment of mechanical properties. The different
Cracking in refinery service due to uptake of effects of PWHT are pointed out.
hydrogen has led to specific requirements for pressure
vessel steels used for wet H2 S refinery service. The
generally qualification test to cover this aspect is the
HIC test as defined by the NACE Standard TM0284. INTRODUCTION
The usual restrictions in steel chemistry such as Cracking due to uptake of hydrogen in H2 S-bearing
limitation of sulfur content to extreme low values media plays an important role in performance of refinery
(:::; 1O ppm) only help to improve but do not optimize the industry materials, since this phenomenon impairs safety
resistance against H2 S related cracking. Therefore a and efficient operation. The importance of these aspects
special production route that includes each by itself as is demonstrated by many service failures [1 - 4].
well as a combination of one subsequent to the other Especially the catastrophic failure in 1984 at the Union
optimized steps of steel making and plate manufactur- Oil Chicago Refinery where a pressure vessel ruptured
ing was developed. The main production steps are and caused an explosion and fire that killed 17 people
explained. [4] has started a lot of inspection programs in the oil
refining industry. Examples are collected in the literature
As a result of this optimization a new family of [5 - 8]. These inspection programs revealed high portions
unalloyed and low alloyed steels is presented. These of H2 S related cracking in several cases.
steels meet the HIC test demands of modern specifica-
tions, i.e. they can be ordered at 3 different levels of As a result of these investigations the demand for
HIC test acceptance criteria. On principle these steels pressure vessel steeis with additional spE~c1tic rP!rn111m-
are available according to all main pressure vessel steel ments has sharply increased. In case of the replace-
standards. Above all the steel grades according to ment of old and the construction of new vessels so called
ASTM/ASME A/SA516 Gr60, Gr65 and Gr70 and HIC resistant i.e. HIC tested steels are more and more
P275 ... and P355 ... according to EN10028 part 3 are used. For instance the number of customer inquiries
delivered as sour service steels even in case of low for HIC tested pressure vessel steels at Dillinger HOtte
tonnage per ordered lot. has increased ninefold during the last three years.
As inspection HiC Test as defined by NACE Standard
During fabrication of vessels PWHT is generally TM0284 [9] is mostly used. In order to get a more severe
recognized and used to improve resistance against H2 S condition the test solution defined in NACE Standard
related cracking. But special attention must be focussed TM0284 (pH 5) is frequently replaced by the test solution
according to NACE Standard TM0177 [1
WRC Bulletin 532

Metallurgical Studies for Steels for Sour Service Environments ll
In the first part of this paper the current possibilities vessel steels for use in wet H2 S bearing media and fol-
and philosophy regarding the production of pressure lowing theoretical considerations which were described
vessel steels with optimized resistance against HIC are elsewhere [13, 14] optimized series of production steps
presented. It is important to mention that all descrip- have been developed. Typical prescriptions such as re-
tions refer to the production of unalloyed and low alloyed strictions in composition (e.g. sulfur content) or perform-
steels as for instance defined in ASTM/ASME SA/A ing a vacuum and Ca treatment only help to improve
A516 [11] or EN10028 part 3 [12]. the resistance against HIC. The desired optimization of
HIC resistance can only be obtained by an optimization
PWHT is frequently used to reduce cracking ten- of the total processing route during steel making plus
dency in H2 S refinery service. The second part of this plate production. The philosophy of the production of
paper critically points out that PWHT has on the one pressure vessel steels for sour gas applications by Dil-
hand beneficial effects regarding H2 S related cracking linger HOtte makes use of this assessment (Figure1).
resistance, on the other hand, PWHT can have detri-
mental effects on mechanical properties. Steelmaking and Casting
The production steps in the steel plant which were

described in more detail in another paper [15] are
PHILOSOPHY OF PRODUCTION ROUTE schematically shown in Figure 2.
To ensure the success of each individual metallurgi-
As a consequence from the experience in the pro- cal process stage a process-time-model and a process-
duction of numerous orders of line pipe and pressure temperature-model have been implemented.
high
non sour gas steels
>-
:!:::
-c
(].)
E
:.0 0.. low sulfur-content
:;::; 0
0.. Ci3
(].)
(.) >
(].) low phosphorus-content
(/)
:;;:, "O
(/)
ctS high cleanness
(.)
I
previous u
·a,
-
I
©
>
0
sour gas steels .....
..2
ctS
......
Q)
::?!
vacuum treatment
Ca-treatment
().)
_J
...
llllllllll---------
new sour gas steels
! -------
optimization· of the total
production route
0
FIGURE 1: Philosophy of production of the new sour service steels
WRC Bulletin 532

Hot Metal Desulfurization. The production of steel such a way that the probability for crack nucleation is
with high cleanliness is started with the hot metal desul- negligible. As it will be described later a small window
furization. of Ca/S-ratio which is derived from older steel types with
higher sulfur contents looses its relevancy for the very
Converter Treatment The aimed carbon range and low sulfur contents, less than 1O ppm.
a phosphorous content less or equal than 100 ppm are
achieved in a 185 t BOF converter with bottom stirring. Casting. The casting is achieved in a sealed-off
In combination with balanced addition of carbon and system including shrouding of the pouring stream.
manganese these reduced P contents are one precon- Casting can be performed by continuous or ingot casting.
dition for avoiding local hard areas which can be However continuous casting is preferentially used due
favourable sites for crack propagation. to its more homogeneous behaviour over thickness and
length of the cast product. As it will be described later
Vacuum Tank Degassing. The application of a special positioning of specimens should be required in
vacuum tank degassing allows simultaneously a strong case of ingot casting in order to test the worst position.
desulfurization and degassing, that means nitrogen and The main special advantages of the Dillinger vertical
hydrogen removal. In combination with the above bending type CC-machine compared to curved casters
mentioned hot metal desulfurization a maximum sulfur are [16]:
content of 1O ppm is obtained on the final cast. • enhanced refinement of non-metallic inclusions
• no locally enriched areas of non-metallic inclusions
Cleanness Stirring. After vacuum degassing a • exact adjustment of roll alignment for various
cleanness stirring with optimized inertgas ratio is slab thicknesses to minimize center segregation.
performed. As a result, the total oxygen content is re- Furthermore bulging is prevented by an intensive
duced to values of max .. 002 %. cooling that gives low strand surface temperatures. This
practice stiffens the shell of the strand. Additionally it
Shape Control. For inclusion shape control a well should be mentioned that for the continuous casting
defined amount of CaSi cored wire is added. On the machine of Dillinger HC!tte bending of the strand takes
one hand the very low sulfur contents, the cleaness place only after complete solidification.
stirring and the shape control reduce the density of Figure 3 demonstrates that only a vertical caster
potential sites for crack initiation to a minimum level. with bending after complete solidification can guaran-
On the other hand remaining inclusions are modified in tee highest internal steel cleanness and its homogeneity.
or
0
• 0 000
Production Hot metal Converter Vacuum tank Casting (CC or ingot)

step desulfurization process degassing unit
-------+----·---- ------'.---- -·----- -;--------------·-·+- ----·-----·-·----·-----·-·--
Metallurgical
mechanism Desulfurization Dephos- Desulfurization, Prevention of resulfurization,
phorization, Nitrogen removal, Minimization of segregation,
Decar- Hydrogen removal, Assurance of cleanness level
burization Cleanness,
Shape control
in the steel for steels with HIC resistance
WRC Bulletin 532

0 curved strand machine Rolling Shape Factor -

10
°
2 .jF!fio:tJ
+ t1
1
= m
lJ'tij
100
• vertical, bending with
liquid core "'" R: Rafi Radius
0
i 1•t stand (5.Sm-4-high stand)
~~ ~~-
-
vertical, bending after 0 Work roll : 11 BOmm
_e> complete solidification ~- - Max. torque: 2 x 4500kNm
....
... 80 0 0
AC synchr. motor
.8 Steel grade : API - X65 (2 x 10900kW)
E Reduction = ..!2..::...'.L (x 100%) Max. force 10BOOOkN
::I
c: Ve 1.0 m/min lo
G' 60 Superheat 25 K
al•t = 1.71 (250->146mm) I 12 passes b) •t = 1.71 (250-146mm) I 3 passes
E
:::i
-
8
/\I
Ill
c:
40
0
"iii
::I
(j <F-
.5 20
oo aot--t----=;:::;::~;;:?::~7,5:::7,5:::1')~
GI ~
e ts
~
~ .c
gi eo+-+---77WW77':~~~~
0 e High shape
~ factor rolling
«I
~ ~o;---t----7'~~~)""
loose ~t Fix o
side 4 side §
·~
Slab thickness position u 201+--J-----L-+------+-4
"
a:
FIGURE 3: Inclusion distribution for different ca-
ster types [ 21 ]
Ratio of reduction of thickness
(Slab thickness/Plate thickness)
FIGURE 5: Influence of high shape factor rolling
({/) u::::dllllllllllllllllllllllllllllllllllWij!h 11dll111111111mm111111111111111~1111111th

or
Production Reheating Rolling Warm pilling Normalizing I Tempering or

step l simulated Stress
1 relief treatment
I (on specimen)
I I J
----------+----------~-------~-----------r----------
Metallurgical 1 I I I
mechanism Homogenisation I High Shape I Adapted I Improvement of I Optimization of
I Factor rolling I H-Effusion I microstructure I microstructure
FIGURE 4: Optimized production steps in the heavy plate mill for pressure vessels steels with high HIC
resistance
WRC Bulletin 532

Plate-making mers insist on not performing PWHT a tempering treat-

ment at a typical stress relief temperature should be
The production steps in the heavy plate mill are sket- performed directly on plate. However post weld heat
ched in Figure 4. treatment of the full vessel is recommended. The reason
for that will be described in more detail in the last part of
Reheating. The reheating of the slabs respectively this paper.
ingots is performed in pusher or bogie hearth furnaces.
Rolling. High shape factor rolling that means high

reduction per pass is feasible by using the very power- HIC TEST PROCEDURE, RESULTS AND
ful 5.5 m-4 high-stand. High shape factor rolling impro- QUALITY ASSURANCE
ves properties in through thickness tensile test respec-
tively toughness properties in mid thickness (Figure 5). Test results
Warm Pilling. If necessary an adapted H-effusion As a result of the just explained application of the
treatment is applied immediately after rolling based on optimized production route it is possible to produce
model calculations. steels for pressure vessels with a high level of HIC
resistance, that means low sensitivity for cracking.
Normalizing. The normalizing is performed at well Figure 6 shows the frequency curve of CLR of plates
defined and reproducible conditions. After normalisation produced with narrow tolerances of the target values
the microstructure is more homogeneous and tough, for all production parameters.
that means more capable of resisting crack propagation.
It is important to mention, that in the case where the
Stress Relief respectively Tempering. After above outlined specific production route is applied there
normalisation a simulated stress relief treatment re- is no demand for an extreme lowering of the phosphorus
spectively post weld heat treatment (PWHT) is applied content (Figure 7) which would increase production
on coupons taken from the plate. In case where custo- costs unnecessarily.
100
number of values: 271
HIC test solution: NACE TM0177
#
Cl)
(/)
CCl
0 60 _.__ _
d::.
_J
()
40 ......__ _
20
0
0 2 3 4 5 6 7
CLR (average 9 sections), %
FIGURE 6: Frequency of CLR from plates produced with narrow target ranges for the production parame-
ters steel grades: A516 Gr60, Gr65 and Gr70 with HIC test level 5; plate thickness: 6 + 80mm
WRC Bulletin 532

100 I
I HIC test solution: NACE TM0177 I

<ft I
cf. I
C')
/\ 80
I .•
I
I
I
q
Ci) limit for I •• I

c::: HIC resistant -1 I
0 I
uQ)
60
steels I
I I
I
(f)
Q)
~
a:
I
I
I
I )
I
....! 40 I
/
(.)
.... I
0
'+- I
~ I
c:::
Q)
20 .
I
/
::J I
.1 /
O"' 41
Q) 0 I.
...... 0
u..
••
0 T
> 40 50 60 70 80 90 100 110 120
:o; 50 60 70 80 90 100 110 120
Phosphorus content ranges, ppm

FIGURE 7: Influence of phosphorus content on CLR steel grades: A516 Gr60, Gr65 and Gr70;
plate thickness: 6 + 80mm
100
'#-
cf.
(")
I HIC test solution: NACE TM0177 I
/\ 80
Ci)
c:::
0
uQ)
(f) 60
CJ)
~ I
a:
....! 40
(.)
._
.e
~
c::
Q)
20
::J
O"
....
CD
u.. --' _ _ _ _ _ _4
~-----· ~-----·
______ .., ~-----...I ~---
~
0
1.5 2 3 4 5 6
2 3 4 5 6 7
Ca/S-ranges (heat analysis)

FIGURE 8: Influence of Ca/Son CLR for plates with sulfur contents ::;;10 ppm;steel grades: A516 Gr60, Gr65
and Gr70; plate thickness: 6 + 80mm
WRC Bulletin 532

Furthermore the importance of a small window of location of HIC Specimens. As already mentio-
Ca/S-ratios which is derived from older steel types with ned plates rolled from ingots are in contrast to plates
higher sulfur contents looses its relevancy for the very rolled from continuous cast slabs handicapped by their
low sulfur contents,::.;; 1Oppm (Figure 8). Therefore it is pronounced dependance of HIC properties on the
recommended to avoid specifying severe restrictions in location of specimens. Figure 10 gives a schematic
Ca/S ratios for very low sulfur contents::.;; 10 ppm, but to illustration of the influence of location of HIC specimens
require a Ca-treatment in any case. on the test results.
For that reason it is recommended to test half width

position in case of plates rolled from ingots. Furthermore,
Supplementary Test Conditions
because it is not in each case assured that the top of
plate is identical to the top of ingot, top and bottom of
HIC Test on Plate Thickness greater than 30 mm.
plate should be tested.
The NACE TM0284 test method was originally prepa-
red to provide a standard set of test conditions for con-
sistent evaluation of pipeline steels and test results from Assurance of Designated HIC Test level. The
different laboratories. The test is only applicable to steels assurance of a high level of HIC resistance over the full
with wall thicknesses from 5 to 30 mm. Since the stan- plate respectively heat requires a good knowledge and
dard specimen thickness is equal to the wall thickness control of all equipments. Further it demands a good
the maximum specimen thickness will be 30 mm. For knowledge of the influence of certain scatter in the
higher material thickness the specimen should be limited production (deviations or incidents) on the HIC resist-
to this maximum of 30 mm. In order to test the whole ance. Consequently for all essential production
thickness for plate thickness higher than 30 mm, sub- parameters target values are defined during the design
specimens should be taken. In the meantime the NACE stage. Because certain scatter can not be avoided a
working group T-1 F-20 that is responsible for the stan- special adapted quality control system for HIC resistant
dard NACE TM0284 has perceived the need for defining steels must be installed that is able to state a deviation
specimen dimensions also for thicknesses greaterthan of actual values and to decide if such plates must be
30 mm in NACE TM0284. prohibited from release and delivery or foreseen for
To demonstrate this necessity, the following test has additional testing to prove conformity with specification.
been carried out: Two plates with only limited resistance
to stepwise cracking and in a thickness equal to 50 mm That means a high level of HIC resistance over the
were chosen and two types of specimen were taken. whole vessel is only assured if the steel producer knows
One type has had a specimen thickness equal to the this purpose of the steel. Only in that case the steel
plate thickness, consequently 50 mm. The other type producer can take all necessary measures of process
has consisted of subspecimens from the lower and control during production and apply such a special
upper surface and from the mid-thickness position. All adapted quality control system.
specimens were immersed into NACE TM0177 solu-
tion. After this test, the specimens with thickness equal It does also mean that there is a big difference
to 50 mm were also cut into subspecimens and their between intentionally as HIC resistant produced plus
results were compared to those from the other type. HIC tested steels and only HIC tested steels. Figure 11
shows schematically that in such a case where basic
The results of the evaluation of all specimens are requirements as for instance low sulfur and low phos-
shown in Figure 9a. The figure demonstrates the consi- phorous contents are met for steels from normal
derable difference between the damaging of the two production it is possible to find also good HIC test results
types of specimens for the mid-thickness position. For for such plates. Nevertheless it exists a lot of plates
the specimens with thickness equal to the plate thick- with unsatisfying HIC test results because they are
ness, resistance is pretended for the mid-thickness area, outside the tolerances of other production parameters
because of their lower and slower hydrogen charging. for HIC resistant steels. Furthermore, as already
As a consequence of these test results, to assure mentioned above, even the HIC tested plate from the
resistance even for severe attack, the position on HIC normal production can have unsatisfactory behaviour
subspecimens as shown in Figure 9b are used to test at other test positions because the homogeneity is not
the steels with plate thickness greater than 30 mm. It assured by the special adapted quality control measures
should be noted that this procedure to define HIC applied for production of HIC resistant steels.
specimens is similar to that foreseen in the new edition
of NACE TM0284.
WRC Bulletin 532

dimensions of specimen, mm: t = 50; w = 20; I = 100

plate A plate B Pos. of specimen
0D
50 @ @ surface
t = plate thickness
• 0 mid thickness
cut before
HIC-test w
upper surface D t/3
D
mid thickness
lower surface 9 t/3
t/3
10 20 30
CTR, %(specimen cut before HIC-test)
evaluation
t
FIGURE 9a: Influence of specimen thickness on CTR from two HIC susceptible steels in NACE TM0177
solution (pH3) Ingot width (W)
·5/12W
·1/3 w
Top
0
Bottom
"'11 1--17"7'.h":"h '
~/"*
I- HIC Suscepibility
B medium (about 30% of 'D')

W = 20 ± 0.5 mm L = 100 ± 1 mm (L =test piece length)
t = full material thickness orientation: longitudinal C E'.'.Z'.a Medium (about 60% of 'D')
T = subspecimen thickness, to be reduced max. 1 mm on each
surface, maximum thickness: 30 mm
D - High
(all dimensions in mm) FIGURE 10: Schematic illustration of the influence

FIGURE 9b: Procedure for defining position of HIC of specimen location on the SWC
subspecimen for sour service steels susceptibility of hot-rolled product
with thickness >30 mm from a size [ 20 ]
WRC Bulletin 532

HIC resistant
nhr" .. = non resistant

plates from production
Cl) hr""" = HIC resistant
Q) of non-HIC resistant
«$ steels within produc- pi =target values of
-0..
0
c
tion parameter of HIC
resistant steels
parameters for each
production step
·-e00 ti =tolerance of each
0.. target value
FIGURE 11: Target values and tolerances of HIC resistant and non-resistant steels
Test solution CLR acceptance criteria for

test solution per NACE TM0177
12,5%
37,0%
NACE TM0177
63,0%
63,0%
FIGURE 12: Statistics on customer's re<lUirerne1ntsfor HIC test
WRC Bulletin 532

CUSTOMER ORIENTED STEEL DESIGN INFLUENCE OF PWHT

(DICREST STEELS)
In the meantime PWHT is mostly recognized to lower
Figure 12 gives an overview of customer require- hardness in welds and heat affected zone (HAZ) and to
ments for steels with defined HIC resistance in standard reduce residual stresses and thereby to minimize the
lab test. The highest level is a CLR-value less or equal risk of SSCC. It will be demonstrated that PWHT has
than 5 % in NACE TM0177 solution. As demonstrated other beneficial effects but can also have detrimental
above this highest level of HIC resistance can be influence. In order to minimize detrimental effects the
reached by using the optimized production route with conditions of PWHT, that means mainly temperature
narrow tolerances of the target values for all production and time, should be carefully selected. It should be kept
steps. For other HIC test levels, that means higher in mind that the following description is only valid for
allowable CLR values respectively higher pH of the test steels similar or according to ASTM A516 or EN 10028
solution the allowable tolerances of the target values part 3 (P275 ... or P355 ... ).
can be increased.
In order to take into consideration these different

customer requirements three levels which differ in Influence on HIC
maximum CLR values (Table 1) as well as in production
costs are offered. HIC Test on Plate. As already mentioned in the first
part of this paper a PWHT and especially its temper
effect improves HIC resistance.
acceptance criteria
(HIC test per NACE TM0284)
steel designation
test solution
(HIC test level)
CLR (%) CTR(%) CSR(%)
::;; ::;; ::;;
NACE TM0177
Level5 5 1,5 0,5
(pH 3)
NACE TM0177
Level 10 10 3 1
(pH 3)
NACE TM0177
15 5 2
(pH 3)
Level 15*
NACE TM0284
0,5** 0, 1** 0,05**
(pH 5)
* test solution to be agreed

** values of "O" are routinely achieved
Table 1: Acceptance criteria for sour service steels with test level 5, 1O and 15
WRC Bulletin 532

Figure 13 demonstrates the influence of the temper- there has been found no further improvement. It is
ature of PWHT on CLR. In order to get a better statisti- important to mention that PWHT improves HIC resist-
cal significance 27 specimens per test condition have ance especially for plates with lower thickness. This is
been investigated from three plates from different heats a result of the temper effect on the less tough micro-
each in 10 mm thickness. Starting from the normalized structural constituents which are more pronounced in
testing condition CLR is the lower the higher the temper- the case of higher cooling rate, that means lower plate
ature of PWHT. For temperatures higher than 550°C thickness.
material condition
~
- - - -- -- - - - ---
_________ J_________________ --------------
~o~;i~d- - , - - - - - - - - - - - - ~;:;~;;;,:ii;;d+-PWHT-
cf 100 -+--4:3--------+--~--------~--------------~-1
LO
/\
en
c O plate A
0 D plate B
t5(!) A plate C
({)
CJ)
~
er:
...J
0
20
400 450 500 550 600 650
without PWHT Temperature, °C (time: 60min)
FIGURE 13: Influence of PWHT on CLR (test solution: NACE TM0177); steel grade: A516 Gr70;
plate thickness: 1O mm
35 14
-:!(
0
en
30
* plate A
~
0
en
c::
12
10 *
i * Atplate
D plate B
c:: 25 + - - - - 1 0
1
0 D plate B t5CD 8
·~
20 (/)
CD
(/) O'l 6
O'l 15 D
> ~ 4
~ 10 a::
a: ....J
(.)
2 ---
....J
0 5
0
l*1 ~
-
~ m
"""
-....
0 I 500 550 600 625 650
0 2.5 5 10 I
without I
PWHT temperature, °C
Extent of cold working, % PWHTj
(compression)
FIGURE 14b: Influence of temperature of PWHT
(holding time 30 min) of cold worked
FIGURE 14a: Influence of cold working on CLR in (5%) HIC specimens on CLR in
NACE TM0177 solution (steel grade: NACE TM0177 solution (steel grade:
A516Gr70) A516 Gr70)
WRC Bulletin 532

Cold Deformation. The obtained level of HIC The higher the Hollomon-Parameter the higher the
resistance on plate can be impaired by subsequent cold change of tensile and thoughness properties compared
processing. Figure 14a demonstrates the detrimental to the initial, i.e. normalized condition. Figure 15 and
effect of a too high extent of cold working in terms of 16 give examples for the loss of tensile strength and
CLR. The cold deformation has been performed by Charpy-V-impact energy due to PWHT with different
compression of the HIC specimens before HIC test. The Hollomon-Parameter that means different temperatures,
impairment can be compensated by PWHT (Figure times and/or cooling respectively heating rates.
14b). It is shown, that already for a temperature of
500 °C the influence of this simulated type of cold In order to meet the required properties the steel
deformation has almost disappeared. analysis has to be designed in dependance on the
conditions of PWHT. For example the tensile strength
Influence on Tensile and Thoughness Properties in the normalized condition must be higher in case of
high PWHT temperature than in case of low PWHT
The influence of the temperature, holding time and temperature. That means that the carbon equivalent has
cooling respectively heating rate of a PWHT can be to be increased in case of high PWHT temperature
expressed by the Hollomon-Parameter [17]: respectively high value of Hollomon-Parameter.
(1) HP= T (20 +log t0 q) · 10-3 Influence on Hardness
with H = Hollomon-Parameter As already mentioned PWHT is mostly used to

T = Temperature in K lower hardness in welds and heat affected zone and
t eq = total equivalent time in h = thereby to minimize the risk of SSCC. In order to inves-
sum of equivalent heating, tigate the influence of PWHT temperature hardness
holding and cooling time [18] surveys have been performed on welded joints [19].
Temperature, °C 575 600 625 650 675

Holding time, h 2 2 2 2 2
Heating/cooling rate, K/h 50 50 50 50 50
40 -r~~~.,-~~~~......,.~~~~~...,....~~~~--r--~~~~-..~~~~~~
"'E
E
--z 20
..c
uQ)
Ci .!::::! 0
c (ii
~ E
t)
0c
..!l2
·wc
w
I-
-20
2
.!;;; I=' -40
(!.) I
(.)
c s:CL
~
Q) + -60
:i::: "O
0 Q)
.!::::!
(ii
E -80
0
c
(Ji
I- -100
17 17,5 18,5 19 19,5 20
HP= T x (20 +log t 0 q) x 10-3
FIGURE 15: Influence of Hollomon-Parameter on tensile strength of steel grade A516 Gr70
WRC Bulletin 532

Temperature, °C 550 625 575 650 600

Holding time, h 2 2 2 2 2
Heating/Cooling rate, K/h 50 50 50 50 50
40 -...~~~~~~~,.......~~~-.-~~~~,.......~~~.......~~~~...-~~~~~~~~~....
...,
.!:
0 'OQ)
20
if, .!::::!
'"? (ij
cti
E
,_ 0
>- g
0) -
(D ~
55 0 -20
...... '
u
cti -l-
a. I
-~ $: -40 -------
.!: 0..
Q) +
u "O
c: Q)
~ .!::::!
& (ij
o E 0 -80
c:
>
..c:
0
16 16,5 17 17,5 18 18,5 19 19,5 20

HP= T x (20 +log teq) x 10-3
FIGURE 16: influence of Hoiiomon-Parameter on impact properties of steel grade A516 Gr70
300
*as welded
275 Li PWHT 1 (570°C) I - - ' · · · - - - - - - + - - - - - - - - · · - - + - - - - - - - - - - + - - -..·· _____, position of indentations:
upper surface -2 mm
o PWHT2 (600°C)
250 o PWHT (620°C)
0
I
>
I 225 *
(/)
(/)
~
Q)
c
200 * 0 *
Li
..c:
cti
ch
* Li
D
I
D
0
(D
..:.:::
175
*
* ~
>
u D
D
I * * *
150
0 ®
I
125
100
6 4 2 0 weld 0 2 4 6
Distance from fusion line, mm

FIGURE 17: Influence of PWHT temperature on hardness of welded joint; steel grade: A516
plate thickness: 68 mm; CMn steel, CE 11 w: .45%
WRC Bulletin 532

Figure 17 shows for example a hardness survey for [6] C.F. Miller and E.J. Sperling, ,,Crack inspection
the position 2 mm below the upper surface. It is obvious of sour and amine service vessels in hydrocarbon pro-
that a PWHT decreases hardness after welding, but duction operations", Corrosion 92, Nashville, April 1992
there has been found no great difference in hardness
for the three investigated temperatures. [7] Panel Discussion, ,,Cracking of carbon and low
alloy steels in wet sour environments - pipeline, piping
and pressure vessels", Corrosion 92, Nashville, April
1992
CONCLUSIONS
[8] NACE Task group meeting T-8-16, ,,Cracking in
Safety as well as efficient operation in the petroleum wet H2S environment", Corrosion 92, Nashville, April
refinery industry has been impaired by H2 S related 1992
cracking. Therefore HIC testing with customer de-
pendent requirements is more and more applied. [9] NACE Standard TM0284-87, ,,Evaluation of
Pipeline Steels for Resistance to Stepwise Cracking"
Improved steels embody a specific philosophy of pro- (Houston, TX: NACE, 1987)
duction that includes
• alternatives of customer oriented HIC test [10] NACE Standard TM0177-86, ,,Testing of Metals
levels for Resistance to Sulfide Stress Cracking at Ambient
• the performance of a special optimized Temperatures" (Houston, TX: NACE, 1986)
production route and
• the application of a special adapted quality [11] ASTM A516, ,,Standard Specification for
control system. Pressure Vessel Plates, Carbon Steel for Moderate and
Lower-Temperature Service (Philadelphia, 1990)
PWHT is generally recognized and used to improve
resistance against H2 S related cracking. But the para- [12] EN 10028-3, ,,Flat products made of steels for
meters of PWHT, that means temperature, holding time pressure purposes - Part 3: Weldable fine grain steels,
and heating respectively cooling rate should be carefully normalized" (Brussel, 1992)
selected by balancing its beneficial influence and the
possible impairment of mechanical properties. [13] B. Bergmann, H.A. Jungblut, A. Streisselberger
and V. Schwinn, ,,Modellization of HIC under service and
test conditions", Int. Conference on Pipeline Reliability,
REFERENCES Calgary, June 1992
[1] R.S. Treseder, ,,Historical review of sulfide stress [14] B. Bergmann, A. Streisselberger, N. Bannen-
cracking of alloy steels", 2nd Int. Conference on ,,Sour berg and H.A. Jungblut, ,,Development of line pipe steels
service in the oil, gas and petrochemical industries", based on model calculations of HIC", Int. Conference
London, May 1987 on Pipeline Reliability, Calgary, June i 992
[2] M. Reuter, ,,Schaden und Vorschriften bei Ania- [15] B. Bergmann and N. Bannenberg, ,,Erschmel-
gen zur Forderung, Reinigung und zum Transport von zen und VergieBen van Sauergasstahlen bei der AG
Sauergas", Technische Oberwachung, 15, 3 (1974): p. der Dillinger Huttenwerke", Stahl und Eisen 112, 7
65 [3] A.J.R. Reed, ,,Problems with pressure vessels in (1992): p. 83
sour gas service", Mat. Perform., 6 (1977): p. 29
[16] B. Bergmann, N. Bannenberg, A. Streisselber-
[4] H.I. Mc. Henry, P.T. Purtscher and T.R. Shives, ger and J. Bauer, ,,Advanced Casting Practices and
,,Observations of hydrogen damage in a failed pressure TMCP Technology Applied to Offshore Plate and
vessel", Carros. Sci., 27 (1987): p. 1041 Linepipe", Houston Materials Conference, Houston, April
1993
[5] M.l. Bullen, M.J. Humphries and R.D. Merrick,
,,Refinery experience with cracking of proces vessels [17] NF A 36-200, ,,Influence des traitements ther-
exposed to wet H 2 S environments", Int. Conference on miques apres soudage sur les proprietes mecaniques
,,Interaction of steels with hydrogen in petroleum indu- des aciers pour chaudieres et appareils a pression",
stry", Paris, March 1989 AFNOR 1982
WRC Bulletin 532

[18] T.F. Gulrin, B. Sc. D. Scott, D.M. Haddrill and J. [20] A. Ikeda, F. Terasaki, M. Takeyama, I. Takeuchi
Glen, ,,The Influence of Stress Relief on the Properties and Y. Nara, ,,Hydrogen Induced Cracking Suscebtibity
of C and CMn Pressure-Vessel Plate Steels", Proceed- of Various Steel Line Pipes in the Wet H2 S Environment",
ings of the Cont. on the Effect of Modern Fabrication NACE National Conference, Houston, 1978,
Techniques on the Properties of Steels, paper no. 621, unpublished paper No. 43
Glasgow, May 1972
[19] V. Schwinn and F. Hanus, ,,Corrosion and [21] M. Kodama, I. Nagay, T. Koshigawa, Y. Habu,
hardness tests on welded joint of DICREST steels", 0. Haida and J. Matsuno, ,,Cont. Casting", Proc. Fourth
DICREST report no. 3 and 4 of Dillinger HCJtte, 1993 Int. Iron and Steel Congr., London, May 1982, paper
respectively 1994 No.23
WRC Bulletin 532

WRC Bulletin 532

QUENCHED AND TEMPERED STEELS

FOR H2S CONTAINING PRESSURE VESSELS
L. COUDREUSE - C. ZMUDZINSKI - P. BOCQUET - A CHEVIET

Creusot-Loire Industrie - BP 56 - 71202 Le Creusot Cedex, FRANCE
ABSTRACT environments, conditions can be found for which large

quantities of hydrogen can be introduced in the material (1).
Up to now the use of normalized steels has been the rule
Several types of degradation have been described in the case
for the fabrication of pressure vessels. Normalized steels
of low alloy steels.
with improved HIC resistance are proposed for sour service
(very low sulfur steels, low phosphorus), however there is - Hydrogen Induced Cracking (HIC). HIC manifests by
still some concerns on the SOHIC behaviour of such steel. the initiation and propagation of internal cracks in the steel
Tests have been performed in order to compare the without external applied stresses.
behaviour of normalized and quenched and tempered steels.
It is shown that the tempered bainite microstructure of - Sulfide Stress Cracking (SSC). In the case of pressure
quenched and tempered steels gives an improved HIC and vessel steels, SSC manifests mainly in hard zones such as in
SSC resistance when compared with the pearlite/ferrite welded areas.
microstructure. Quenched and tempered steels can achieve - Stress Oriented Hydrogen Induced Cracking (SOHIC).
the same levels of mechanical characteristics as the In fact SOHIC can be considered as a combination of HIC
normalized steels, with a lower Ceq value. This will result in and SSC.
lower hardness values in HAZ. Tests performed on
simulated HAZ microstructures indicate that this is The two first phenomena have been extensively studied,
favourable for the SSC behaviour. and ways of improvment of steel behaviour have been
proposed.
SOHIC phenomenon has not been so much studied, and
INTRODUCTION in fact only recently test methods have been to
In oil and gas production equipment or in petroleum investigate this phenomenon (2).
refineries, carbon steels and low alloy steels can be severely HIC and SOHIC manifest mainly on carbon and low
embrittled when they are used in presence of wet H2S. The alloy steels used for the fabrication of line-pipe and pressure
embrittlement is the consequence of the absorption in the vessels. In the following we will focuse on the case of
steel of hydrogen resulting from the corrosion reaction at the pressure vessel steels. Although the problems are the same
surface of the steel. In oil and gas production, risks of for p!peline steels and pressure vessel steels, the answer of
hydrogen embrittlement increase when the amount of H2S the steel producer to the problem is not exactly the same.
increases and when the of the environment decreases. In In the case of pipeline steels, there is a demand for a high
refinery equipments very low pH is seldom encountered, but productivity. High productivity weld processes have to be
it has been shown that even in alkaline wet H2S used, without post weld heat treatments. In order to decrease
WRC Bulletin 532

the weight of the pipes, there is also a demand for higher segregation areas with bainitic and martens1tlc
grades. As a consequence, in the last twenty years there was microstructures have to be avoided. In segregated areas a
an evolution from low strength grades normalized steels to correlation between the phosphorous content and the
higher strength grades Thermomechanical Control Process occurence of cracking has been found. Then a decrease of
Steels (TMCP steels). TMCP steels are generaly very low the average amount of phosphorous increases the HIC
carbon steels with microalloying element such as Nb, Ti, V. resistance. The presence of segregation can also be reduced
The mechanical characteristics are achieved through the use by an optimized cropping of the top of the ingots. It is also
of controlled rolling and accelerated cooling (3). important to adapt the size and the shape of the ingot to the
final size of the product. A tempering of the plates, such as
In the case of pressure vessel steels, the evolution has not
the tempering performed by PWHT, has also a positive
been the same and the use of normalized steels is still the
effect on HIC resistance of segregated areas.
rule in most of applications. There are several limitations for
the use of TMCP in the case of pressure vessels. For heads, Another way to decrease the sensitivity of segregated
hot forming is not possible since it is not possible to perform areas is to decrease the amount of alloying elements such as
heat treatment after forming. There are also some questions C and Mn. But in the case of normalized steels these
about the characteristics of welded areas. Due to the very elements cannot be reduced under a certain level, in order to
low carbon content, there is a soft area just near the HAZ, guarantie the mechanical characteristics. The addition of
and cracking has already been observed in this area (4). microalloying elements, such as Nb, V, Ti, can allow a
reduction of Ceq content, but as we will see later, has other
In the following we will consider the ways to improve
consequences on the hardness of heat affected zones.
the H 2S behaviour of steels for pressure vessels. We will
consider the case of normalized steels and also quenched In Table 1 we have listed the normalized steels which are
and tempered steels which appear very attractive. generally supplied with improved HIC resistance. It can be
seen that steels with guaranteed yield strength up to 355
MPa can be supplied.
1 - NORMALIZED STEELS WITH IMPROVED H~S. In Figure 1 we have reported histogram of HIC test
BEHAVIOUR results obtained on normalized steels with improve HIC
resistance. CLR values given is the average value obtained
HIC resistance of normalized steels depends mainly on from 3 specimens taken over the thickness of the plate
two parameters : cleanliness and microstructure. It is now (maximum specimen thickness : 30 mm).
well admitted that improved HIC resistance is achieved by
the use of very clean steels. The reduction of the size and Some recent studies indicate that modern normalized
number of inclusions reduces the number of crack initiation steels with improved HIC resistance can develop SOHIC
sites. On an other hand, globular inclusions are less cracks in presence of external stresses. SOHIC cracks are
dangerous than elongated inclusions. As a consequence, favoured by stress concentrations, which can be present in
steels with improved HIC resistance are very clean steels welded areas or in notches. The occurence of SOHIC cracks
with: has been evidenced by four point bending tests (2). It seems
- very low sulfur content (S ~ 0.001 %)

TABLE 1 : NORMALIZED STEELS USED FOR THE FABRI-
- very low oxygen content (0 ~ 0.001 %). CATION OF H2S CONTAINING PRESSURE VESSELS AC-
Calcium treatment is often proposed as a way to obtain CORDING TO DIFFERENT STANDARDS
globular inclusions. However we have demonstrated that in
the case of very low sulfur content, calcium treatment is not steel grades ASTM NF A 36205 BS 1501-224 EN HHl28
necessary since sulfide inclusions are already globular. In NF A 36201 BS 1501-225
this case, if calcium treatment is not well controlled a

decrease of the cleanliness can be obtained through the Grade 60 A516gr60 A 42 gr400 P265
UTS> 400MPa (415/550) (410/490) (400/520)
formation of calcium oxides (5).
However if a very good cleanliness is a necessary Grade 65 A516gr65 A 48 gr460 1'295
UTS>450MPa (450/585) (470/560) (460/580)
condition to improve RIC behaviour, it is not a sufficient
one.
Grade 70 A516 gr 70 A 52 gr 490 P355
UTS >485 MPa (485/620) (510/610) (490/610)
Once a crack has initiated, its propagation depends on the A537 ell E 355
microstructure around the crack tip. Then it is necessary to (485/620) (490/610)
improve the microstructure of the steel too. Particularly
WRC Bulletin 532

that ferrite/pearlite banded microstructures are more characterization of plates.

favourable for the development of HIC and SOHIC cracks.
These results have still to be confirmed but it is agreed,
2.1 - Laboratory simulation of quenching and
and this has been shown by the development of TMCP
tempering
steels for line pipe, that the normalized structure is not the
most resistant to hydrogen embrittlement. For pressure Low grades normalized industrial heats have been
vessel steels, the use of quenched and tempered steels seems selected with the chemical compositions given in Table 2.
to be an interesting alternative to the normalized steels.
TABLE 2: CHEMICAL COMPOSITION OF STEELS USED

40
FOR LABORATORY SIMULATION OF QUENCHING AND
TEMPERING
35
30 Grade th c s p Si Mn Ni Cr Mo Cu Al Ccq
(I) (mm) (2)
;if! 25
A42 105 0.145 0.001 0.005 0,236 1,09 0,157 0,085 0,048 0,118 0,023 0,37
;:I
"'~ 20
..... A516 gr60 8-32 0.124 0,001 0,005 0,285 0,932 0,08 0,059 0,022 0,07 O,Q3 0,305
0
If!. 15 A516 gr60 25 0,156 0,001 0,005 0,256 0,89 0,076 0,129 0,034 0,052 0,022 0,345
10
(1) : Grade corresponding to the normali7.cd stale
(2): Ccq =C+Mn/6+(Cr+Mo+V)/5+(Ni+Cu)/15
5
2.1.1 - Tests on a 105 mm thick plate

0
0
II \) 'O ...v ;:;
i:i.: i:i.: i:i.: /\
i:i.:
..l ..l ..l ..l i:i.: 20, 50 and 105 mm thick coupons have been sampled
u u uv uv ..l
u
::. N
"'
from the plate as indicated in Figure 2. The coupons have
been austenitized at 925°C for 1 h. The austenitization has
CLRVALUES
been followed by water quenching.
FIGURE 1 : HISTOGRAM OF HIC TEST RESULTS ON NOR-
MALIZED STEELS (135 TESTS IN NACE SOLUTION) Tempering at 600 and 650°C have then been performed.
In Figure 3, the microstructure at the normalized state and
2 - BEHAVIOUR OF QUENCHED AND TEMPERED after quenching and tempering are compared. The
STEELS H~S ENVIRONMENTS normalized state corresponds to a ferrite/pearlite
microstructure, whereas the quenched and tempered
A research programm has been performed in our research coupons give ferrite/bainite microstructures. The amount of
center in order to study the behaviour of quenched and bainite is higher for the low thickness coupons than for
tempered steels in H 2S environment. those of greater thickness.
The objective of the research was to study quenched and

tempered steels with LIJ.e mechanical characteristics of the
highest grades of normalized steels given in table 1.
YS::?: 355 MPa
490 :s:; UTS < 620 MPa.
The approach used was the following.
- Selection of industrial heats of lower grades of
normalized steels (A42, A516 gd 60 or 65)
I• HIC test sample
- Reaustenitization of test coupons and then realization of
the quenching and tempering treatment.
- Characterization of the quenched and tempered coupons FIGURE 2 : SAMPLING OF COUPONS FOR QUENCHING IN
(Mechanical characteristics, HIC and SSC tests). THE 105mm THICK PLATE, AND LOCATION OF HIC TEST
- Tests on industrial plates : Quenching, tempering and SAMPLES
WRC Bulletin 532

HIC test results are given in Table 3. Whereas for the

normalized test a CLR value of about 6 % is found at 1/2
thickness, maximum CLR values of about 3 % are found on
the quenched and tempered coupons. It is observed that for
quenched and tempered coupons the amount of cracking
increases when the thickness of the coupon increases. This
indicates that the most quenched microstructures give the
highest HIC resistance.
TABLE 3 : MECHANICAL CHARACTERISTICS AND HIC

TESTS RESULTS AFTER LABORATORY OF Q&T
Thickness Sampling Tempering YS UTS El% NACE TEST

(mm) (MPa) (MPa) CLR% CTR%
600°C-lh 394 529 31,6 0 0

surface 650°C-lh 376 509 32,2 0 0
20
1/2 thick. 600°C-lh 412 569 26,2 0 0
650°C-lh 396 540 27,5 " 0
Normalized state (1/2 thickness) Q&T state ;20mm thick coupon
temp 600°C- l h
50 600°C-lh 368 526 33,4 0 0
(I) 650°C-lh 355 505 34,4 0 0
105 600°C-lh 349 516 33,4 2,5 0,3

(1) 650°C-lh 334 491 36,2 3,1 0,5
105 1/2 thick. normalized 290 450 37 5,9 1,1
(1) properties al 1/2 thickness of Q&T coupons
, The ferrite/pearlite specimen gives the highest amount of

cracking. It must be observed L.11at for the normalized state
the results are not representative of the actual behaviour of
the plate since generally the HIC tests are performed on
specimens taken through the whole thickness of the plate
(average between the surfaces and mid-thickness).
Q&T state ;50 thick coupon Q&T state ;105 thick coupon
l/2 thickness ;temp 600°C-lh 112 thickness ;temp 600°C-lh
SSC tests have been performed according to the NACE
TM 01-77 90 - Method A - on tensile specimens.
Specimens have been taken from the 20 and 105 mm
thick coupons after quenching and tempering, and from the
FIGURE 3 :MICROSTRUCTURES OBSERVED ON THE
105MM THICK PLATE :NORMALIZED AND Q&T STATES plate at the normalized state. The results are shown on
(xlOO) figure 4. It is clear from this figure that quenched and
tempered microstructures give an improved SSC resistance.
Again the best behaviour is obtained with the 20 mm thick
quenched and tempered coupon with quite a fully bainitic
microstructure.
Mechanical characteristics measured after quenching and
tempering are given in Table 3. For the 20 mm and 50 mm
thick coupons, the aimed mechanical characteristics have
2.1.2 - Tests on lower thickness plates
been achieved. For the 105 mm thick coupon lower
mechanical characteristics are obtained. Quenching and tempering have been performed on
coupons of full thickness of 3 plates between 8 and 32 mm
On these different coupons, HIC tests have been
thick (see Table 2).
performed according to NACE TM 02-84 standard (with the
NACE TM 01-77 test solution). In each case HIC tests have Two plates : (8 and 32 mm thick) have a very low Ceq
been performed on 20 mm thick specimens (specimens sizes (0.31 and 0.30 respectively).
20 x 20 x 100 mm) corresponding to the mid-thickness of
A 25 mm thick plate has a Ceq value of 0.35.
the plate (which is the most sensitive area of the plate). The
WRC Bulletin 532

5 % for normalized steels ; CLR<l % for Quenched and tem-

pered steels).It is obseved that for about the same Ceq va-
lue,the HIC behaviour of the8mm thick plate is better than
the behaviour of the 32 mm thick plate.It is important to no-
te that for low thickness plates, very good HIC behaviour are
obtained on quenched and tempered material.That is not al-
;: 250 - 1 - - - - - - - + - - 1 ' way the case for normalized plates.
:>.
~ 200
r-
~ TABLE 4: TESTS RESULTS ON LOW THICK.NESS Q&T
;=, 150
z PLATES (LABORATORY SIMULATION ON A516 gde 70)
100
Thickness State Tempering YS UTS El% NACE test
(mm) (MPa) (MPa) CLR %CTR%
50
Normalized 312 435 33 0,8 0,25

N·1 Q&T ;105mm; Q&T ; 20mm ; 1/2 25 Q&T 600°C-lh 360 510 32 0 0
105rnm; 1/2 tl1ickness I/2thickness thickness 488 33 0 0
Q&T 650°C-Jh 345
Normalized 322 431 30 4,2 1,25

8 Q&T 600°C-Jh 452 578 28 0 0
0.8
Normalized 300 438 34 2,5 0,35
v;
>- 32 Q&T 600°C-lh 325 473 35 0,5 0,04
0 0.6
N
IC-
E o.4
-~
0.2
0 1~CLR ~CTR I
N; Q&T ;IOSmm; Q&T; 20mm; 112
105mm; 1/2thickness 1/2thickness thickness
4,5
-:,.
4 i
FIGURE 4 :SSC TEST RESULTS, COMPARISON BE1WEEN
3,5 i
NORMALIZED(N) AND QUENCHED AND TEMPERED
(Q&T) STATES
g:""' 2,5 It
The quenching and tempering treatments have been

performed on test coupons (200 mm x 200 mm x whole
u
i:.:
...:!
u
2 k ~
N
thickness). 1,5 k ~
~ ~
~
The obtained mechanical characteristics are given in It'-; ~
~
Table 4. ~
0.5
~
For the 25 mm thick plate, the mechanical characteristics
0
~~
~
' '· /.
~
' ,, ~i00
after tempering at 600°C are in agreement with the aimed ;5 ;5 ;5 ;5 ;5
] ] u
'tl
-~
;5
values. For the 8 and 32 mm thick wit..'1 low Ceq -;; ~ ~ ~ ~ ~
values, the mechanical characteristics are reached only for E 0 0
~ 0 &:,
'z"E 0 Q
"""'f-
0 l() 0 l() 0
z
c 'oC
f- "" zc "" "' 0
""
the lowest thickness plate.For these two plates,tempering «!
Cl
'"'
«!
Cl
'"' «!'Cl"'
«!
Cl
'"'
«!
Cl
«!
Cl
has only be performed at 600°C. Plate: 25 mm thick 8mm thick 32 mm thick
HIC tests have been perfonned according to the NACE

TM 02-84 standard with the NACE TM 01-77 test solu-
tion.Test results obtained on these 3 plates are summarized
in Table 4 and in Figure 5. FIGURE 5 : HIC TEST RESULTS ON LOW THICK.NESS PLA-
TES, COMPARISON WITH NORMALIZED STATE
HIC test results confirm that lower CLR values are
obtained on the quenched and tempered coupons. For all the
tests performed low CLR values have been obtained (CLR <
WRC Bulletin 532

2.2 - Tests on industrial plates 3 - BEHAVIOUR OF WELDED AREAS

Quenched an tempered plates have been produced
industrially. The chemical composition of the plates are
3.1 - Correlation between hardness and SSC
given in Table 5. The Ceq contents were between 0.38 and
behaviour
0.39. In order to obtain the aimed mechanical
characteristics, the treatment conditions were the following : Welded areas and particularly Heat Affected Zones
(HAZ) are known to be sensitive to Sulfide Stress Cracking
Austenitization: 925°C - 15'
phenomenon. In order to study the SSC behaviour of HAZ,
Water quenching: tests have been performed according to NACE standard TM
01-77 (tensile tests) on specimens with simulated HAZ.
Tempering : 600°C - 30' for 24 mm thick
Thermal simulations have been performed on Gleeble
2 h for 47 and 63 mm thick plates. machine.
The microstructures of the plates are illustrated on Figure The specimens used are shown on Figure 7. The gage
6. They are very similar to the microstructures obtained on length of the tensile SSC specimens has been reduced in
laboratory coupons (homogeneous bainite-ferrite order to obtain quite an homogeneous microstructure on the
microstructure). whole gage length. Tests have been performed on A516 Gr.
70 and TSTE 355 steels with the chemical compositions and
Test results are summarized in Table 5 :
mechanical characteristics given in table 6. The TSTE 355
- Mechanical characteristics are in accordance with the steel was a Nb modified steel so that the Ceq value was
aimed value. lower than for the A516 gde 70 steel. The simulation heat
treatments performed and the resulting hardness are given in
- No crack was detected on the plates after HIC test (in
table 7.
NACE TM 01-77 test solution).
TABLE 5: CHEMICAL ANALYSIS(Wt%) OF INDUSTRIAL TABLE 6: CHEMICAL ANALYSIS OF STEELS USED FOR
Q&T PLATES AND HIC TEST RESULTS THERMAL SIMULATION
th c s p Si Mn Ni Cr Mo Ceq YS UTS NACE test Grade th c s p Si Mn Ni Cr Mo Cu Nb Ceq

(mm) (MPa) MPa) CLR% CTR% (I) (mm)
24 0,161 0.001 0,006 0,31 0.972 0.224 0,151 0,055 0,39 390 545 0,23 0,02 A516 gr70 50 0,188 0,001 0,008 0,267 1,045 0,200 0,189 0,055 0,113 I 0,432
47 0,165 0,000< 0,005 0,29 0,95 0,223 0,131 0,065 0,38 387 558 0 0 E355 35 0,114 0,0005 0,008 0,29 1,340 0,195 0,096 0,056 0,118 0,026 0,389
63 0,17 0,000 0,00( 0,28 0,92 0,220 0,13( 0,068 0,38 373 546 0 0
JOO mm
SamplC:.'i for thermal simulation
simulated
HAZ
Sampling of SSC specimen
24 mm thick ; Ceq;0,39
FIGURE 6: EXEMPLE OF MICROSTRUCTURES OBSERVED

ON INDUSTRIAL QUENCHED AND TEMPERED PLATES FIGURE 7 : THERMAL SIMULATION SPECIMENS, SAM-
PLING OF SSC SPECIMENS
(xlOO)
WRC Bulletin 532

In Figure 8 the threshold for cracking (a720 h) is given as the higher heat input, hardness below 250 Hv is obtained
a function of hardness in the heat affected zone. It is clear even in the as welded condition.
from figure 8 that for hardness higher than 250 Hv there is
The SSC behaviour of the welds has been studied
an increase of SSC sensitivity. For the A516 Gr. 70 steel,
through a four point bending test as illustrated in figure 9.
PWHT is effective in reducing the hardness to low level,
Specimens have been exposed for one month in the NACE
and there is an improvment of the SSC behaviour. For the
TM 01-77 solution under a stress level of 90 % of the
Nb containing steel, it is also observed that for hardness
guaranted Y.S. The occurence of cracking has been analysed
below 250 Hv there is an improvement of the HAZ
through metallographic examination. Results are
behaviour. However in this case for coarse grain areas,
summarized in Table 9. Cracks have been observed only for
PWHT is not sufficient to reduce the hardness below 250
A516 Gr. 70 and TTSTE 355 steels in the as welded
Hv, and in this case the improvment of SSC behaviour is not
condition. For these two steels a PWHT is effective in
so great.
reducing the SSC sensitivity of the HAZ (even for the Nb
modified steel for which the hardness after PWHT is higher
than 250 Hv). For the quenched and tempered steel no crack
TABLE 7: SIMULATION HEAT TREATMENT PERFORMED
AND RESULTING HARDNESSES
was observed even in the as welded condition. This has to
be correlated with the low Ceq content.
Weld area Simulation PWHT Hardness (HV 30)
simulated treatment IA516 gde70 E 355
Coarse grain 1250°C-ls No 340/360 355/360 TABLE 8: COMPARISON OF HAZ HARDNESS UNDER SIN-
15kJ/cm At800-500 =8s GLE BEAD DEPOSIT (HV30)
Coarse grain 1250°C-ls No 290/320 310/330
25kJ/cm At800-500 =8s 600°C-lh 220/230 270/280 Heat Input PWHT A516 gd70 E3SS Q&T
(kJ/cm) (Ceq=0,43) (Ceq=0,39) (Cea=0.36)
Reaffected area l 250°C- I s+900°C- ls No 220/240 230/235
25kJ/cm At800-500 =8s 600°C-lh 190/200 205/215 15 No 315 320 265
(preheat 100°C 600°C-lh 235 285 225
2mm from fusion 1250°C-ls No 230/240 205/210
line 25kJ/cm AtS00-500 =8s 600°C-lh 175/195 175/185 25 No 285 310 240
(preheat I 00°C 600°C-lh 230 280 215
TABLE 9 :RESULTS OF FOUR POINTS BENDING TESTS
Heat Input PWHT AS16 gd70 E355 Q&T

(kJ/cm) (Cea=0,43) (Cea=0,39) (Cea=0,36)
15 No Cracks Cracks No crack

preheat 100°C 600°C-lh No crack No crack No crack
a E 355
. 25 No Cracks Cracks Na crack
200 (preheat 100°C 600°C-lh No crack No crack No crack
• A516 gde 70
100 L--~--.L--~---'---"'--t---'
"
100 200 300 400 4 - CONCLUSIONS - ADVANTAGES OF 0 AND T
HV 10
STEELS
FIGURE 8: RESULTS OF SSC TESTS ON SIMULATED HAZ The use of normalized steels is the rule for the fabrication
of pressure vessels working in H2S environment. Some
recent studies indicate that the ferrite/pearlite microstructure
3.2 - Welding of Quenched and Tempered steels found in normalized steels is not the best one. Normalized
Due to their lower Ceq content obtained without steels with improved HIC resistance can be produced with
microalloying addition, the hardness in HAZ of quenched yield strength up to 355 MPa, however there are still some
and tempered steels is expected to be lower than in classical concerns on the SOHIC phenomenon in these steels. In the
normalized steel. case of pipe line steels, TMCP steels with very good HIC
resistance are now produced, but there are some limitation
In Table 8 the hardness under single bead deposit have for their use for pressure vessel fabrication (hot for-
been compared for different steels in the as welded ming,welding). The use of quenched and tempered steels
conditions and after PWHT at 600°C for 1 h. It is observed seems to be a very interesting way to improve the HIC
that the Q and T steel gives the lowest hardness level. For resistance.
WRC Bulletin 532

REFERENCES
H=150mm
h=50 mm
t=6mm (1) R.D. KANE, S.M. WILHEM and J.W. OLDFIELD,
do=l2 mm "Review of Hydrogen Induced Cracking of Steels in
Wet H 2 S Refinery Service". Interaction of Steels with
Hydrogen in Petroleum Industry Pressure Vessel
CT H 2 (1-h/H) (1+2h/H)
d= 3Et Service, The Materials Properties Council Inc., New
0 = Contrainte maximale
York, 1993.
maximum tensile stress(MPa)
(2) R.D. KANE, "Final Report of Findings and
E=module d'Y oung Recommandations of the MPC Sponsored Program on
(modulus of elasticity=210 000)
the Susceptibility of Process Equipment and Steels to
Cracking in Aqueous Hydrogen Charging
Environments". The Materials Properties Council Inc.,
New York, Feb. 1993.
(3) H. T AMEHIRO and M. CHINO, "The Progress in
Pipeline Material Properties". Publ. Nippon Steel
Corporation, Apr. 1991.
Ue !f:jl •I
(4) "Full Scale and Laboratory Tests Approach to H 2 S
~. tfll •I
Induced Environmental Fracture of Pipeline Steels and
Welds", Final Report, ECSC Research 7210-KE/l 16.
N.A.C.E. TM OI-77 solution
NaCl 50 g/l + CH3COOH ->pH = 3
saturee en H 2S I saturated with HzS
(5) L. COUDREUSE, C. ZMUDZINSKI, J. CHARLES
and L. CADIOU, "Carbon Manganese Steels for Sour
Service : Improvement of HIC and SSC resistance",
Ibl, p. 193.
FIGURE 9 :PRINCIPLE OF FOUR POINT BENDING TESTS
When compared with normalized steels for a same level of

mechanical characteristics, it is shown that quenched and
tempered steels give improved HIC and SSC resistance. The
tests performed confirm the higher hydrogen embrittlement
resistance of homogeneous tempered bainite microstructures
when compared with normalized microstructures.
A desired level of mechanical properties can be achieved
with a lower Ceq content in the case of quenched and
tempered steel. This presents several advantages.
A decrease of Ceq content improves the homogeneity of
the steel and decreases the risk of finding hard zones in the
plates. A decrease of Ceq is also favourable for the welding
of view. A better weldability is low hardness
levels can be obtained in HAZ even in the as welded
conditions. Then in some cases PWHT treatment can be
avoided after welding.
In the case of quenched and tempered steels it can be
expected that grades could be used in sour service
when compared with normalized steels.
WRC Bulletin 532

INFLUENCE OF ACCELERATED COOLING ON THE

HYDROGEN INDUCED CRACKING AND SULPHIDE STRESS CRACKING
BEHAVIOR OF UNEPIPE STEELS WITH A LOW CARBON CONTENT
H. Margot-Marette and B. Chamont

IRSID-UNIREC
Saint-Germain-en-Laye, France
E. Amoris and C. Perdrix

SOLLAC, Usine de Dunkerque
Dunkerque, France
ABSTRACT mechanical treatments (2 - 8). In particular, the

segregation problem remains incompletely resolved.
This study was carried out on low carbon linepipe Meanwhile, this parameter plays a very important
steels in order to improve our knowledge of the role on the hydrogen cracking resistance (9 - 11).
accelerated cooling process after controlled rolling Accelerated cooling af+-er controlled rolling is
and its consequentia 1 effects on mechanica 1 proi:erties one of the means to reduce segregat~on effects (1,
and hydrogen ernbrittlement resistance in hydrogen 12, 13). This process also presents the advantage to
sulphide environments. provide good mechanic;al properties without ad::litions
The stee 1 plates were manufactured on an experi- of expensive elements to the chemical composition
mental rolling mill with a cooling unit. ( 1) •
The main conclusions are : The aim of the vresent work is :
Accelerated cooling improves the mechanical pro- - to show the influence of accelerated cooling
perties with respect to air cooling but in our conditions on the hydr0gen induced cracking
experimental conditions the variation of cooling para- resistance of linepipe steels with low C c0ntent,
meters have a weak influence on the mechanical proper- - tu point out the effect of an apJ?lied stress
ties. In contrast, cooling parameters play a role on in H2S aquecus environment in relation to the
the hydrogen induced cracking (HIC) resistance, espe- rolling direction, and
cially in presence of a moderate level of segregation. - lastly, to contribute to the comprehension of
The lower the finishing cooling temperature (FCT), the the hydrogen embrittlement mechanism,
higher is the HIC resistance, while a too low FCT is The study was concerned with low carbon micro-
to be avoided. a lloyed steels for linepii:e. The controlled rolling
An optimal cooling rate of 10 to 15°C/s allows and accelerated cooling were conducted on an
the steelmaker to obtain improved HIC rcosistance experimental rolling mill with a cooling unit, The
throughout a large range of finishing cooling influence of cooling rate and finishing temperature
temperature .. of cooling on mechanica 1 properties and hydrogen
Lastly, it is possible to improve sulphide stress induced cracking (HIC) resistance was examined. For
cracking resistance by accelerated cooling, mainly in a few cases, the effect of an applied stress in H2S
the transverse direction, but the best conditions corrosion environment in relation +-.o the rolling
remain to be defined. direction was alsu considered.,
Key words Accelerated cooling - Hydrogen inducal MATERIALS

cracking - Sulphide stress cracking - Linepii:e steels.
The experi.mental v;ork was performed on fou:c
INTRODUCTION industria 1 steels manufactured by continuous cas-
ting, with SiCa treatment to control the inclusion
The use of linepii:e steels in more and more moq:hology and to obtain a low sulphur content
severe conditions has caused to be developed, these (('. 0.002 per cent)
recent years, fabrication processes to simultaneously Table 1 shows the chemical composition of the
obtain good mechanical pr0perties and a good studied steels.
resistance to hydrogen induced cracking in hydrogen The carbon content varied throughout the range
sulphide (H2S) environments (1). 0.030 - 0.050 per cent, while tl»e rnang.3ne«e ccntmt
The improvement of "in-use" properties can be was between 1.15 to 1.58 per cent. The average
obtained by decreasing impurities (S, P, O, H), oxygen content was 0.0015 per cent.
inclusion morphology control, diminution of segre- Starting from pieces of rough ;::>lates measuring
gation, optimal steel composition and lastly thermo- 70x70x200rnm, thermomechan;cal treatment:s were
WRC Bulletin 532

STEEL c Mn p s Si Al Ni Cr Cu Nb v Ti N
0.040 1. 162 0.013 0.002 0.221 0.022 0.028 0.021 0.262 0.049 0.046 0.0049
0.029 1. 581 o. 01 7 o. 002 o. 161 0.038 0.026 0.024 0.313 <l. 041 0.061 0. 012 0.0059
0.042 1.426 0.010 0.002 0.212 o. 032 0.028 0.021 0.265 0.040 0.010 o. 0085
4 0.050 1.165 0.013 0.002 o. 312 0.041 0.033 0.017 0.274 0.045 0.066 0.012 0. 0059
Table - Chemical composition in weight %. Average oxygen content : 0,0015 %
STEEL CONTROLLED ROLLING CONDITIONS COOLING CONDITIONS
Reheating Reduction Finishing 't: Cooling rate Finishing cooling

Temperature 2d stage Rolling Temperature
oc % Temperature 0 c °C/s oc
30 600, 540, 490, 450, 410

115 0 67 815
10 5 75, 530, 500, 450, 365
30 600, 560, 500, 450, 410

2 115 0 67 785
10 590, 550, 490, 4 50, 385
30 605, 560, 520, 450, 390

115 0 67 785
10 620, 580, 510, 475, 420
4 1150 75 830 15 590, 550, 520, 475
Table 2 - Thermomechanical and cooling conditions.
conducted on an experimental rolling mill. The details CAR were determined.

of the experimental conditions are presented in table SSC tests were performed with the constant load
2. A heating temperature of about 1150°C was realised method, according to the NACE standard TM 01-77 1 the
so that the niobium and vanadium were almost comple- non failure threshold being determined after 720
tely in solution. The finishing temperature of rolling hours. Testing machines were CORTEST (TM) proof rings.
was, in all cases, except me, slightly higher than The specimens were machined in the long and transverse
the Ar transformation point. The reduction in second directions.
stage iaried between 67 and 75 % and the fina 1 thick-
ness was 12 mm. RESULTS AND DISCUSSION
For the steels 1 to 3, two austenite cooling
rates were studied : 30°C/s and 10°C/s. For the stee 1 Mechanica 1 proi:erties
4, this rate was 15°C/s. The usual air cooling
corresponds to 1.5 °C/s. For all the steels, the The mechanica 1 properties were studied in the
variation of finishing cooling temperature (FCT) was transverse direction. Figures to 4 show the
between about 600°C and 400°c. evolution of tensile properties (YS and UTS) and
figures 5 to 8 those of toughness (trans it ion
EXPERIMENTAL PROCEDURE temperature) in relation to the finishing cooling
temperature (FCT) for each cooling rate.
Mechanical properties in the transverse We can see a very weak variation of the tensile
direction were tested on tensile specimens (5 mm properties with the FCT except for the extreme tempe-
gauge diameter and 25 mm gauge length) and full-size ratures. Low FCT gives slightly increased UTS but
charpy v-notch impact test specimens taken from the lower YS. An increasing cooling rate from 10 to 30°C/s
mid-thickness of the plates. enhances slightly the yield strength and ultimate
HIC tests were conducted according to the NACE tensile strength (about 20 MPa). But there is a strong
standard TM 02-84 in a chloride solution acidi:'ied effect en tensile strength ('I.I 50 MPa) by accelerated
with acetic acid (pH3) and saturated with H2S (TM cooling rather than air cooling.
01-77 standard). FCT seems to have no great influence on the tran-
The cracking susceptibility was evaluated by an sition temperature, moreover, it is noticed that the
ultrasonic method described formerly (14) and based higher the cooling rate, the lower the transition tem-
on an echographic method with focalised transducer. perature. The clear beneficial effect of accelerated
B and C scans (parallel and perpendicular cooling on the tensile properties and toughness compa-
respectively to rolling plane) were established and red to air cooling is mainly due to the refinement of
allowed to localise the cracking. Crack area ratios both the structure and the carbides precipitation, con-
WRC Bulletin 532

600
TRANSVERSE DIRECTION 600 TRANSVERSE DIRECTION
~ ~
~
590
560 / /
560 /
7 7 ~
/ / ?
/ / /
~ 540
/
/
/ / / / / /
/ ~ 540
/
/
/
/
/
/ /
/
13 520 Ill 520
/
v
/
~ v v v
I: ~ t: "
~ r, v v
~
~ " " " "v / / /

500
~
v
v
v
t: "
~ ~ ~
~
~ v /
~ 48a v
v v v v
v t:
~ 460
~
~ " /
w
:::! 460 v
v
/
v
v
v "
/ v
~
~ v
v
v v "" e 440 " ~
v "
r~
441)
420
v " " 420
"
/
/
,, ~
t:
t:
/
/
v ~ ~
v v /
41)0 " " 400
AIR 1600 540 490 450 41011375 530 500 450 3651 F.C. T. <°C> AIR 1600 560 500 450 41011590 :550 490 450 38!5! F.C, T, (OC)
:l1l°C/s 10°C/s COOLING RATE :l1l 0 C/s 10°C/s COOLING RATE
Fig. 1. Effect of accelerated cooling on tensile Fig. 2. Effect of accelerated cooling on tensile
properties (YS, UTS) - Steel 1. propert.ies (YS, UTS) - Steel 2.
600 TRANSVERSE DIRECTION 600 TRANSVERSE DIRECTION
560 560
~ 540 ~ 540
Ill
I:
520
~ 500
~ 480
w
~ 460 :::! 460
~ 440 m 440
420 420
41)0 400
All'! J 605 560 520 450 39011 620 590 510 475 420 I F. c. T. <°C > F.C.T. (°C)
30°C/s 10°C/s COOLING RA TE COOLING RATE
Fig. 3. Effect of accelerated cooling on tensile Fig. 4. Effect of accelerated cooling on tensile
properties (YS, UTS) - Steel 3. properties (YS 1 UTS) - Steel 4.
TRANSVERSE
- DIRECTION
TRANSVERSE DIRECTION - -
-120 - -
n
n~
-120
ID FATT 50%1
1 ID FATT 50ltl -
~
I I
-110 -110
-
-100 -100
u ~ -90
..
~
1il
-90
-80 ~ -ea
,.: ,.:
,.: -70 ,.: -70
< <
..: ..: -60
-60
-50
-50
-40
AIR j600 541) 490 450 41011575 530 500 450 365) F.C.T. !°Cl
:l1l 0 C/s 10°C/s COOLING RATE 30°C/s
Fig. 5. Effect ofaccelerated cooling on transition Fig. 6. Effect of accelerated cooling on transition
temperature - Steel 1. temperature - Steel 2.
WRC Bulletin 532

-120 }:FAN~VERSE QIR~TION ~ TRANSV~SE_QIRECTION
-110
- [D FATT 50%1
-120
I CJ FATT :ioxj
-110
-100 -100
~ -90 u
..
~ -80
..
2. -90
Iii -so
,.: ,.:
,.: -70 ,.: -70
< <
u'. u'.
-60 -60
-50 -50
-40
AIR 1600 560 520 450 390 II 620 580 510 475 4201 F.C. T. 1•C) AIR 1600 540 490 450 1 F.C. T. 1°Cl
30°C/s 1o•c1s COOLING RA TE 150C/s COOLING RATE
Fig. 7. Effect of accelerated cooling on transition Fig. 8. Effect of accelerated cooling on transition
temperature - Steel 3. tem[12rature - Steel 4.
10 100
9
I CJ C.A.R. 90
ICJ C.A.R.
8
80
7
E E 70
E
...
~
...g 60
5 ~ 50
~
< 4 ~ 46
<
tl tl
~l
5 3
30
20
10
AIR j 600 540 490 450 41011 575 530 500 450 3651 F.C. T. (•Cl AIR 1600 560 500 450 41011590 550 490 450 3851 F.C.T. <°C)
30°C/s 1o•c1s COOLING RA TE 30°C/s 1o•c1s COOLING RATE
Fig. 9. Effect of accelerated cooling on HIC Fig. 10. Effect of accelerated cooling on HIC
resistance - Steel 1. resistance - Steel 2.
10
I CJ C.A.R. I CJ C.A.R.
~n
~
E
I-
6
~
<
~ 4
tl
3
AIR 1605 560 520 450 39011620 580 510 475 4201 F.C.T. c•ci AIR 1 600 540 490 450, F.C.T. 1°C!
30°C/s 10°c1s COOLING RATE 15oc1s COOL !NG RATE
Fig. 11. Effect of accelerated cooling on HIC Fig. 12. Effect of accelerated cooling on HIC
resistance - Steel 3. resistance - Steel 4.
WRC Bulletin 532

38 Metallurgical Studies.for Steels.for Sour Service Environments
secutive to the increasing cooling rate during the aus- figure 13). Figures 14 to 17 illustrate the effect of
tenite transformation ( 15). FCT for a given cooling rate (30°C/s) on this
The small influence of finishing cooling tempe- segregation zone. For the higher temperature (600°C),
rature on the tensile properties between 450°C and a thin martensitic islets band ( 20 µm, HV :;dOO) is
600°C is linked to very slight and subtle changes in observed. As the FCT decreases, this band becomes
the final ferritic structure of the products as will larger and softer. For 450°C, practically m
be seen later. segregated band ai:peared, whereas for 410°C, a 200 pm
Lastly, the hardening eff.ect of microalloyed thickness bainitic band is revealed.
elements is linked to their quantities in solution These observations were confirmed on the other
before transformation. While the good solubility of steels.
vanadium guarantees that the vanadium content, soluble The HIC is directly Hnked to the presen.ce of
at the heating temperature, is totally free at the these segregated bands, and is localised in
transformation, sucn is not the cas2. for niobium mid-thickness region of the plates. The improvement
where 20 to 30 per cent of the total content is able of the HIC behaviour by accelerated cooling is mainly
to precipitate during rolling as strain induced due to a smoothing of segregation areas 1 the coo ling
precipitation ( 16) • The differences in mechanica 1 operation in a limited area probably preventLrig the
properties between the steels studied is mainly diffusion of carbon to the segregation zone as
expressed by their content in Mn, Nb, V and Ti. described by TAMEHIRO ( 12) •
HIC resistance
Figures 9 to 12 show the HIC results in relation

to the cooling conditions (temperature and rate). We
observe an inconsistant effect of the accelerated
cooling on the HIC resistance. Steel 1 is completely
resistant to HIC in all conditions. Ste"'l 2 has > 50%
CAP. when air cooled, minimum CAR (generally< 10%) when
cooled at 10°C/s, and poor CAR (20-50%) when cooled at
the fastest rate. Stee 1 3 has a similar pattern to
Steel 2 1 but is much more resistant to HIC in all
conditions. Stee 1 4 has improved CAR when coo led at
15°C/s rather than air cooled. Therefore, it ai:pears
that a cooling rate of 10-15°C/s provides the best HIC
resistance.
Finishing cooling temperature effect. We

generally observe that the lower the FCT 1 the higher
is the resistance to HIC. Nevertheless a lower limit
is observed in a few case (by example see stee 1 2 for
30°C/s cooling rate and FCT 450°C) under which the HIC
susceptibility again increases. This has already been
mentionned by other workers (12 1 17). Fig. 13. Controlled rolling - Mid-thickness -
Steel 2. x 200.
Cooling rate effect. The higher the cooling rate, MARTENSITIC ISLETS BAND
the narrower is the finishing cooliHg tamperature
range where HIC disappears. We can notice, i;articu-
larly 1 that for a 10 to 15 °C/s coo ling rate, a large
temperature range allows to attain an excellent resis-
tance to HIC. This fact is very interesting from the
industrial point of view, because it is easier to
obtain a reproducible HIC quality guarantee.
Steel influence. Steel 1 which presents a very

high HIC resistance with controlled rolling, keeps its
good resistance regardless of the coo ling conditions.
It is easier to enhance the HIC resistance with steels
3 and 4 than with steel 2. This behaviou.t: can probably
be explained by a large difference in the segregation
levels of the steels.
Micrographic examinations, were conducted in an

attempt to exJ?lain the accelerated coolL1g effect on
HIC behaviour. Observations were made at 1/4
thickness and at mid-thickness.
Generally an homogeneous structure is observed
at 1/4 thickness. Controlled rolling is characterized Fig. 14. FCT 610°C. Cooling rate 30°C/s.
by a regular distribution of pearlite islets in a Mid-thickness - Steel 2. x 500.
ferritic matrix, whereas accelerated cooling is
characterized by a fine acicular ferritic structure
with a few bainitic is lets.
At mid-thickness, a segregation area is
r~vealed. For controlled rolling, we observe the pre-
sence of a ferritic perlitic banded structure (cf
WRC Bulletin 532

Fig. 15. FCT 560°C. Cooling rate 30°C/s. Fig. 16. FCT 4S0°C. Coolillg rate 30°C/s.
Mid-thickness - Steel 2. x 200. Mid-thic)~ness- Steel 2. x 350. no segregation.
SSC resistance
SSC tests were ccnd11cted on steel 4, in order to

point out the difference in behaviour between
:industrial controlled rolling in the a+'/!' field
(plate 4A, Finishing Rolling Temperature = 680°C) and
expe~imental accelerated cooling (plate 4B 1 FRT
835 °C) •
The results obtained in the long and transverse
directions are shown respectively on figures 18 and
19 and in table 3. It appears that the SSC resistance
can be improved by accelerated cooling. This
improvement is noticeable ma inly in the transverse
direction. We can observe an increasing SSC
transverse threshold from o. 65 YS (for a + ¥ CR
plate) to 0.8 YS (for AC plate). The best SSC
resistance is due to a more homogeneous structure
leading to isotropic "in-use" properties.
Nevertheless / work must be pursued in order to define
more accurately the influence of thermomechanica l and
cooli'1g conditions on SSC resistance.
Fig. 17. FCT 410°C. Cooling rate 30°C/s.
Mid-thickness - Si:ee 1 2. x 100.
LONG DIRECTION TRANSVERSE DIRECTION

PLATE
YS UTS aNF aNF/YS YS UTS aNF aNF/YS
CONTROLLED
4A ROLLING 465 535 410 0.85 500 5 55 325 o. 65
a+ tr
FRT = 680°C
ACCELERATED
4B COOLING 460 550 440 0.95 460 540 370 0.80
FRT = 835°C
Table 3 - Comparison of SSC resistance between (a+'() controlled rolling and accelerated cooling plates of
steel 4.
WRC Bulletin 532

The Welding Research Council, Tnc.
STEEL 4 CJ 4A
Long dtrect ton
110 ....
C9 48
100 ....
Long dtrect ton
Ill
>- 90 -- -
1--~~~~~~~~~.+-~~~~~~~.::.....==-..+-~~~~~~_,,,~ ......~
~ -(:+Y}cR-~
(f) 80 ....
(f) ""'"""
w
0£'.
I- 70 ....
(f)
0
w 60
_J
Q_
Q_ 50 ....
<
40 ....
I
10 100 1000
TIME TO FAILURE (Hours)
Fig. 18. SSC resistance. Comp'3.risoo between controlled Rolling aDd Accelerated
cooling. Long direction - Steel 4.
STEEL 4 CJ 4A
110 Trans. di.rect ton
100 C9 48
Trans .. dt~ctton
Ill
>- 90
~
(f) 80 A.C.
(f)
w
0£'.
I- 70
(f)
{<»P)C.R.
0
w 60
_J
Q_
0... 50
<
40
10 100 1000
TIME TO FAILURE (Hours)
Fig. 19. SSC resistance. Comp'3.rison between controlled rolling and accelerated
cooling. Transverse direction - Steel 4.
WRC Bulletin 532

CONCLUSIONS 7. Nakai, Y., Kurahashi 1 M., Emi, T., Haida, o.,

"Development of Steels resistant to HIC in Wet H2S
This work was conducted with low sulphur, low Environment" Transactions ISIJ, Vol. 19, 1979 1 pp.
carbcn linepipe steels to improve our know ledge on 401-410.
accelerated cooling process after controlled r0lling 8. Kimura, M,, Totsuka 1 N., Kurisu, T., Amano,
and its consequential effects on the mechanical K., Matsuyama, J., Nakai, Y. 1 "Sulfide Stress
proi:erties and hydrogen embrittlement resistance in Corrosion Cracking of Linepipe, Corrosion 86 - NACE -
hydrogen sulphide en vircnment. Houston, Pai;:er 160.
With respect too controlled rolling, acceler,,ted 9. Taira, T. 1 Kobayashi, Y., Matsumoto, K., Mat-
cooling at rates greater than 10°C/s largely improved sumoto, s . , Terunima, T., Arikata, K .. , "Resistance of
the tensile properties (about 50 MPa higher). Faster Pipeline Steels to Wet Sour Gas, Proceedings of the
cooling rates in our exi:erimental conditions, do not Ist Intern. Conf, "Current Solutions to Hydrogen
seem to play a determining role on the mechanical Problems in Steels" ASM. Washington, Nov. 1982 1 pp.
properties. 173-180.
In contrast, accelerated cooling rates seem to 10. Haida, o,, Kitaoka, H,, Habu 1 Y, 1 Kakihara,
play an important role in determining the HIC resis- s., Bada 1 H., Shiraishi, s., "Macro and Semi macro
tance. Accelerated cooling between 10 to 15°C/s is Features of the Centerline Segregation in c.c. slabs
very beneficial but its effect also &!pends on the and their effect on Product Quality" ~'lctions
finishing temperature and the segregation level of the ~Vol. 24 1 1984, pp. 891-898.
steel. For the steels studied, a cooling rate of 10 to 11. Matsumoto, K. 1 Kobayashi, K, 1 Ume, K.,
15°C/s allows the steels to obtain an excellent HIC Murakami, K., Taira, K., Arikata, K., "Hydrogen
resistance (crack area ratio lcwer than 2 per cent) Induced Cracking Susceptibility of high Strength
throughout a large range of finishing temperature of Linepipe Steels", Corrosion-NACE, Vol. 42, N° 6, June
cooling. This point is very important :tro:;; the 1986, w. 337-345.
industrial point of view because it easily allows a 12. Tameiro, H., Takeda, T., Matsuda, s.,
mean to improve HIC resistance. Yamamoto, K., Okumura, N., "Effect of Accelerated
Finally, it is f'OSsible to improve SSC resistance Cooling after Controlled Rolling on the Hydrogen
by accelerated cooling mainly in the transverse Induced Cracking Resistance of Linepipe Steel"
direction. Nevertheless, work must be pursued to have Transactions ISIJ, Vol. 25, 1985, r:p. 982-988.
better knowledge of the thermomechanical and cooling 13. Ohtani, M., Hashimoto, T., Komizo, Y.,
parameters influencing the SSC behaviour. Murayama, J., Sawamura, T., Bessyo, K., Kyogoku, T.,
"development of low PCM High Grade Linepipe for
BIBLIOGRAPHIC REFERENCES Artie Service and Sour Enviror1ment" 1 Proceedings of
the Intern. Conf. on Technology and Applicaticns of
1. International Conference on Pipe Technology, High Strength Low Alloy Steels A.S.M.
ROMA, Nov. 1987. Philadelphia, 3-6 oct. 1983, Paper 8306-057.
2. Dewsnap, R,F., Jones, C.L. 1 Lessells, J., Mor- 14. Margot-Marette, H., Bardou, G., Albert,
rison, Rudd, W.J., Walker, E.F., Wilkins, R. "A Review J.C., Cassier, 0. 1 "Application of an echographic
of Information on Hydrogen Induced Cracking and ultrasonic method to the quantification of Hydrogen
sulphire Stress Corrosion Cracking in Linepii:e steels. Induced Cracking of Linepii;:e Steels". This
Offshore technology report OTH 86 256 1 Lendon Her Conference, Pai;:er n° 18.
Majesty's Stationery Office, 1987. 15. Ouchi, c., Okita, T., Yamamoto, s.,
3. Margot-Marette, H., Charbonni0r 1 J.C., "Effects of Interrupted Accelerated Cooling after
Bridoux, D. 1 Perdrix, C, 1 "Influence of Metallurgical Controlled Rolling on the Mechanical properties of
Parameters on HIC of Linepii:e Steels", Re-,ue de low alloy Steels". Transactions ISIJ, Vo]. 22, 1982 1
Metallurgie - CIT, Dec. 1986 1 pp. 891-899. w. 608-616.
4. Ikeda, A., Nakanishi, M., Kaneko, T., 16. Perdrix, c., Chamont, B., Amoris, E.,
Sumitomo, Y. 1 Takeyama, M., "Improvement of Hydro- Biausser, H., "A model for Mechanical Proi;:erties of
gen Sulfide Cracking Susceptibility in Linepipe Steels Thermomechanically tretated Plates". Proceedings of
for Sour Gas Services" 1 Sumitomo Search, Nov. 1981, N° Thermec 88 Conf. ISIJ, Tokyo, 6 - 10 June 1988 1 w.
26, pp. 91-104. 807-814.
5 .. Taira; T .. 1 Tsukada, K.,, Kobayashi, Y.,, 11 HIC 17 .. Tamehiro, H.. , Murata, M., Takeda, T .. , Habu,
and SSC Resistance of Sour gas Linepipe Steels" Nippon R., Nagumo, M., "Application of Accelerated Cooling
Kokan Technical Report, Civerseas 1 ' ' 0 31, 1981. after Controlled Rolling to Linepii;:e Steel". Procee-
6. Hersleb, G., Poepperling, R.K. Schwenk, W. 1 dings of the 3d. Intern. Conf, Steel Rolling~
"Occurence and Prevention of HIC and SSC of Lew alloy Pipe Tube ApPlications. ISIJ, Tokyo, 22-23 sept,
Pipeline Steels" Corrosion-NACE, vol., 36, N° 5, 1981v 1985.
w. 247-255.
WRC Bulletin 532

EFFECT OF SOME METALLURGICAL FACTORS ON THE

SUSCEPTIBILITY TO SSCC OF X60 TO X70 PIPELINE STEELS
F. Terreur, G. Monfort and J. Defoumy

Centre de Recherches Metaliurgiques
Abbaye du Val Benoit, Liege, Belgium
ABSTRACT - hydrogen induced cracking ( HIC). on the one hand,

which develops in direction parallel to the
The aim of this paper is to precise the surface of the product and often becomes apparent
influence of some alloying elements and of micro- in the form of blisters ;
structure on the behaviour to corrosion by wet H S sulphide stress corrosion cracking (SSCC), on the
2
of grade X60 to X70 pipe line steels produced 5y other hand, which generally occurs along the
controlled rolling and accelerated cooling after through thickness direction, only when the compo-
rollin9. nent or the specimen is submitted to a mechanical
The trials are performed according to the NACE stress or a local embrittlement of the steel.
standard TM-01 77 with the samples stressed to 80 1.
of the yield strength. In bath cases, cracks result from the diffusion
As regards the microstructure of the steel and of atomic hydrogen generated at the steel surface,
its effect, acicular ferrite microstructure appears which recombines under molecular hydrogen in speci-
beneficial to decrease sulphide stress corrosion fic locations in the metallic structure and causes a
cracking. local increase of the internal pressure or a local
Concerning the chemical composition, the detri- embrittlement of the steel.
mental roles of sulphur and phosphorus have first Steels for drilling as well as pipelines for
been precised, the latter element mainly governing gas and oil or LPG tanks may then be sensitive to
the phenomenon of cracking through its very local one or bath of these corrosion phenomena. It is
concentrations. Carbon is to be limited firstly to therefore interesting to document the most
favour the obtention of acicular ferrite type micro- favourable types of steels in such an environment as
structure and secondly to lower the influence of well as the optimal processing conditions of these.
manganese. Neither copper nor nickel were found to This is, however, a difficult task not only because
have any influence in the NACE test conditions. In of the great number of factors influencing the
the BP test conditions for which a higher pH value sensitivity of a steel to hydrogen embrittlement but
prevails, copper is, however, beneficial. also owing to the important role exerted by
The influence exerted by microprecipitates has interactions between two or several metallurgical
been investigated and depends on the size and parameters. It is thus hard, in many circumstances,
distribution of these particles in the steel matrix. to clearly define the role played by one of these
Finally, the presence of a welded zone was parameters and to draw for instance firm conclusions
found to decrease strongly the behaviour of steels regarding the effect of a given allaying element.
even when the base steel passes the test. Taking that fact into account, this paper tries ta
identify, from the invest1gations carried out at
1. INTRODUCTION CRM, a set of conditions whose fulfilment allows to
produce steels resisting to hydrogen sulphide, a
Steels used in the petrochemical industry are special emphasis being put on stress corros1on
more and more submitted to the action of corrosive cracking and on metallurgical products processed by
agents particularly aggressive under certain condi- accelerated cooling after hot rolling.
tions. One of these, which is found in the whole
chain of the oil industry, is the hydrogen sulphide 2. TESTS CONDITIONS
which is well known as a promoter of hydrogen
embrittlement when present together with water. The A certain number of procedures are applied ln
action of hydrogen induces the apparition of two pratice ta appraise the sensitivity of steels to
types of cracks : hydrogen sulphide corrosion.
WRC Bulletin 532

Metallurgical Studies.for Steels for Sour Service Environments 43
They differ either by the nature of the corrosive TABLE I

agent IBP solution with a pH of 4.5, NACE solution
at pH 3.5), or by the presence or not of a stress Range of the mechanical properties
applied to the specimen, or again by the way this
possible stress is developed (under flexion or Re IMPal Rm I MP a) A 17.)
tension). The test described in the NACE recommen-
dation TM-01-77 (1) sets the steel under extremely 268-564 436-735 17 - 4 1
hard conditions: -3
- on the one side, the corrosive agent is a solution Range of the composition ( 10 7.)
of water containing 5 Z NaCl acidified by the
addition of 0.5 7. acetic acid, which fixes the c Mn Si Al
initial pH around 3. This solution is continuously
saturated in H s and the pH during the test does 40-160 440-1620 • 300 • 50 10-40 2-32
2
not exceed 4.5 ;
on the other side, the test temperature is speci- Nb Mo Ni Cu
fied to 24 ~ 2.a·c. a range corresponding to the
most dangerous area. 0-60 0- 2 50 • 50-250 • 50-235
Such conditions recognized in the literature as

minimizing the cracking resistance 121 are made 5. INFLUENCE OF THE MICROSTRUCTURE
again more critical by the application of external
load. To pass the test, the specimen of the studied For the population of the investigated steels,
steel must resist 720 h. cracks or failures were systematically recorded in
As a rule, a complete characterization accor- specimens showing heterogeneous microstructures such
ding to this recommendation requests to appraise the as mixed ferrite and pearlite. ferrite and bainite
duration to rupture as a function of the stress or ferrite dotted with pearlite and bainite. As it
level. Nevertheless, an accelerated procedure cor- is examplified in Figure 1, cracks were associated
responding to an often encountered specification with the harder constituents such as bainite or
consists in testing the specimen under a stress pearlite.
equal to 80 7. of the steel yield strength.
3. SSCC TESTING UNIT
The CRH testing system presents the following

features (a I 15 cells (one per sample I have been
adapted on creep testing machines which allow a
constant load all over the test period : these
machines have a 2 tons load capacity which is enough
for testing X60-X70 steels ; lb) the gas circuits
are controlled and switched automatically by
computer le) data logging is used extensively to
monitor and record the temperature in the cells, the
elongation of the sample and other parameters ; Id)
the concentration in the room is continuously
monitored (e) a battery power supply is used for
the data logging system; If) the H S is neutralized
by a three-step treatment after bub5ling in the test
solution. A detailed description of parts of this
unit can be found in a previous paper (3).
4. INVESTIGATED STEELS
The research was mainly oriented to pipeline

steels of grades X60 to X70 produced under control-
led rolling followed by on line accelerated cooling
according to a new technique developed by the
Rolling Department of CRH (4). Other types of steels
were also investigated either for comparison pur-
poses or to enlarge the population of the tested
products.
40 different continuous cast steels were
examined whose range of mechanical properties and
chemical compositions is summarized in Table !.
In complement to the tests on material produced
u<1der industrial conditions, trials were also per-
formed on some of these steels re-hotrolled in the
laboratory so as to modify their microstructure,
which appeared as an important factor and whose
effect will first be approached in this paper.
and
FIG. 1 SENSITIVE MICROSTRUCTURES
WRC Bulletin 532

It is particularly important to observe, on the Conclusions in the same direction are round in
contrary, that a quite reproducible success to the the literature 15,6,7,8.9) but to support this
SSCC test could be met by a low carbon steel descri- finding, different thermal heat treatments were
bed in Table II and whose microstructure after applied to the hereabove mentioned steel so as to
accelerated cooling displays an homogeneous 1c1cular modify the microstructure from the favourable one
ferrite with only some bainitic islands, shown in initially found :
Fi9ure 2. two normalization heat treatments at 820'C and
950'C let carbide bands to appear ;
li\~_t,_CJ!. - a reheating at 950'C followed by a cold water
~t_g~_:!,_Q3!.U.!.l'.t9 the NACE test quenching and post-tempering at &oo'c resulted in
an increased bain1tic amount.
0.06 7. Mn 1 7. Si : 0.3 The so achieved mechanical properties as well
P:0.0247. 0 ' 0 0 4 7. as the results to the NACE test are listed in Table
Al: 0.070 7. Nb 0' 0 50 7. III. For comparison purposes, the original data
Phosphorus segregation ratio lower than concerning this steel are also reported. None of
Steel processed by Accelerated cooling those modified microstructures resisted more than
Microstructure of acicular ferrite one day under the NACE test conditions. It is thus
Thickness : 11 mm clearer from this, that homogeneous structures free
Re : 540 MPa Rm : 600 MPa A : 36 from hard phases are necessary, in particular those
Cha2py V at - 60'C : of the acicular ferrite types.
306 J/cm (Rolling direction) Thermal treatments were applied to other steels
2 at temperature of 1200'( so as to redissolve the
217 J/cm (Transverse direction)
carbides and improve their initial microstructure.
Then a rolling sequence was applied followed by an
accelerated cooling. Such tests which will not be
detailed here, allowed to reach the target of an
acicular ferrite but 1n spite of this, failures
induced by SSCC occurred. Such disappointing results
show that the microstructure does not govern alone
the phenomenon. A strong influence is indeed exerted
by the steel chemistry. as it is related hereunder
to clarify the understanding of the phenomenon.
6. INFLUENCE OF THE STEEL CHEMISTRY
6. 1. Sulphur.
The influence of the HnS inclusions on the

hydrogen cracking is well-known : such inclusions
are very dangerous when they are elongated since
they constitute rupture starters by the notch effect
they induce and the high internal pressure they
develop owing to the recombination of atomic hydro-
gen in molecules in the interface inclusion-steel.
Our results support these findings and indicate
that a sulphur level lower or close to 0.005 7. and a
globularization treatment with Si-Ca bring a solu-
tion to this problem.
6.2. Phosphorus
Phosphorus is an element with a well-known

effect for promoting several segregations and the
building up of hard banded area of bain1te and
martensitel7). A greater attention was paid to this
element during this research after that the thermal
treatments related above and aimed at homogeneiz1ng
the microstructure did not lead to the expected
improvements.
Phosphorus homogenization in steels is rather
hard. Consequently, this element is found essential-
ly in the central segregation, with segregation
coefficients which may reach a value equal to 8.
This segregation may be revealed in metallography
either by the Stead's or by the Oberhoffer's
reactive. The choice was finally decided in favour
of the Stead's reactive, which gave clearer results
in the case examined at CRM. This reactive deposits
copper on the areas with low phosphorus content,
which makes them appear darker.
FIG. 2 AC!CULAR FERRITE HICROSTRUCTURE

(STEEL RESISTING TO SSCC)
WRC Bulletin 532

_,T-"A'-'B'-'L,_,E~,_r....
r.,.r__,__,,_Behaviour of modified microst.ructure~
State of the steel Micro structure Mechanical properties Time to

from table II failure
Re(MPa) Rm(MPa) A( 'l.) at NACE
I test
Normalized 820 Ferrite-pearlite 41 0 1,55 28.2 10 H

Normalized 950 Ferrite-pearlite 337 442 30.2 20 H
Reheat and quenched Ferrite--bainite I
450 512 27.6 14 H
Original Acicular ferrite 5 "° I
600 36 >720 H
-· ---------
Figure presents two examples of the results More elaborate examinations were carried out
obtained. The first steel in the upper part shows with a scanning electron microscope coupled with a
important segregations. It will be noted that the wave dispersion spectrometer. As in all cases, the
hydrogen induced cracks are located preferably in here investigated steels come from continuous cas-
the white zones rich in phosphorus. For comparison ting and present certain central segregation,
purpose the image of phosphorus distribution is also which appears as a preferential location for the
shown for the steel of Table II which passes the occurrence of cracks. The analysis by the microprobe
NACE test without cracking : a so marked segregation was concentrated in this area. Two distribution
of phosphorus does not here appear. profiles of P (and also Mn) are presented in Figure
4 and are specially significant. They correspond
indeed to the steel of Table II which passes the
NACE test and to another one having a similar chemi-
cal composition as well as a same acicular ferrite
microstructure, obtained after the laboratory rehea-
ting and hot-rolling treatments. The main signifi·
cant difference between both steels is the local
maximum phosphorus content : close to 0.050 'l. for
the steel which passes the test and to 0.100 z for
the steel which fails the test. Such local contents
correspond to respective segregation ratios of 2 and
3.
One observes thus in those conditions that a
local phosphorus amount of about 0.050 z in a steel
with content in manganese close to 1 'l. and in
carbon near 0.06 is still acceptable for an
acicular ferrite microstructure.
Taking into account that the segregation ratio
of phosphorus may range between 2 and B (7,15)
depending on certain conditions, the recommendable
phosphorus maximum global content lies between 0.005
and 0.020 'l.. The mean acceptable phosphorus level is
thus to be closely associated to the steel chemistry
and the conditions of casting, reheating. hot-rol-
ling and cooling which are the main parameters
influencing in total the central segregation on the
end product and the local phosphorous contents.
These two elements strongly interact for some

important phenomena :
- they influence the y-a transformation parameter,
and therefore the final microstructure ;
possible apparition of banded structure is
strongly linked to those elements ;
the size of the Nb carbo-nitride microprecipitates
tends to increase when carbon content is high and
manganese content is low ;
b. Weak phosphorus segregation the segregation coefficient of carbon depends on
the manganese content which, consequently, could
influence the microstructure.
FIG 3 EXAMPLES OF PHOSPHORUS SEGREGATIONS

SHOWN BY STEAD'S REACTIVE.
WRC Bulletin 532

TAJl.U__lY
n(%) Calculation of ~anganese diffusion rates
t'I~
Formula used :
66000]
DMn, T ( 0. 4 86 + 2 - 53 c) exp [ -
0' 11 Mn) ( 1 +
1 . 99T
Mn(%) 2
DMn,T diffusion rate of Mn at temperature T I cm Is)
1.4 T temperature I K)
1.2 Mn,C contents ( z)
1
0.8 1. Steel with c : 0.06 z Mn : 1 z
Composition (%) ---------------------------------
C:0.061 Mn:0.992 P:0.024 0
2
3,2 1 0 - 12 cm /s
Mn,100o'c
2
4,6 10 - 10 cm /s
Steel passing the test DMn,1300'C
2. Steel with c 0,02 z Mn 2 z
·---·--------- ---··---------
(%) 2
3.5 1 0 - 12 cm /s
0.11 DMn,lOOO'C
DMn,1300'C
2
5. 0 1 0 - 10 cm /s
o.os rL Mn(%)
Manganese being well-known as increasing
o ~11..Mn A l J2 efficiently the tensile strength while not affecting
too much the toughness, it is thus beneficial for
~· v·~~J1.5
pipe steels to limit the carbon so as to keep admis·
1 sible Mn contents high enough. A decreased carbon
Composition {%) value goes also in the right direction to promote
the formation of acicular ferrite during accelerated
C:0.081 Mn: 1,244 P:0.034 cooling. Favourable carbon contents are therefore
lower than or close to 0.05 Z.
Steel failing the test 6.4. Copper and nickel
FIG. EXAMPLES OF CENTERLINE SEGREGATION SCANS. The extremely positive influence of copper is
well-known under SSCC when the testing solution pH
All of these quoted phenomena can influence the stays above 4.5. (8,11,15). For sourer solutions,
SSCC behaviour of steels. It is thus logical to none effect is, however, recorded (16) _
consider carbon and manganese together as concerns Nickel, for its part, is reported as without
this problem. effect or slightly unfavourable because of an acce-
Our experiments showed that manganese is a leration of the pitting corrosion in steel (16).
segregating element as phosphorus but to a lower The influence of copper and nickel has been
extent a segregation ratio between 1.1. and 1 .5. appraised on four casts whose compositions (Table V)
was observed (Figure 4). Examinations with the and microstructures are comparable except as regards
scanning electron microscope indicated that at a both the above elements. Results of our tests are
local manganese content of 1.3 7. in a steel contai- listed also in Table V. They show that the steels
ning 0.06 z carbon, cracks could be avoided. For fail the test with equivalent durations to failure.
steels with such a medium C content 12 0.05 Zl a Those four steels were then tested under other
nominal value of manganese near 1 7. should then be conditions, namely those of the BP test, i.e. for a
accepted, as often recommended in literature pH close to 5. Results in Table V show that the
(7,10,11,12,13). As the segregation coefficient of steel with high copper content then succeeds the
manganese is related to the carbon content of the test contrarily to the three others which display
steel, the acceptable maximum content of manganese durations to failure a bit superior to those
depends also on the carbon level. Calculations corresponding to the NACE test, but comparable
carried out at CRM and based on the diffusion rate between each other. Nickel has therefore neither an
are reported in Table IV. They show that same manga- influence in those conditions.
nese diffusion rates and thereby same segregation
ratios would be observed in steels containing
respectively 0,06 7. C - 1 Z Mn and 0,02 Z C - 2 Z
Mn. This ts confirmed by other findings (14) accor-
ding to which for carbon contents lower than 0.04 Z,
no detrimental effect of manganese appears up to 1 .7
while for higher C contents, a pronounced effect
is seen.
WRC Bulletin 532

3
STEEL Composition ( 10 - ) Mechanical properties Time to failure
c Mn p s Cu Ni Re(MPa) Rm(MPa) AI/.) NACE BP
-- ---
TF 1 0 95 10 77 22 8 38 41 434 524 23.2 42 H
TF20 79 992 22 7 .f1..i 53 430 51 7 2 6. 7 ?7 H >720 H
TF30 85 1067 20 6 89 1.li 441 525 28. 7 15 H 42 H
TF40 76 115 2 23 4 42 48 432 527 2 6. 6 16 H 68 H
6.5. Microprecipitates induced during hot rolling On the contrary, coherent precipitates of fine size
and cooling and uniformly distributed in the matrix do not seem
to induce any trapping and let hydrogen diffuse
Shimogori (17) reported beneficial effect without damage in the crystalline network. Other
played by a fine coherent precipitation, in the works are presently carried out at CRM to verify
scope of an investigation of the influence of carbo- whether such an assumption is correct and agrees
ni tr ides an HIC. with other findings (20).
To approach the influence of microprecipitates, TABLE VII
Nb-Ti HSLA steel (Table VI) was processed at CRM Microprecipitates in th~ Nb-Ti Steels
according to two thermal processing routes in an
experimental rolling mill, using as start materials, Nb content 0. 031
samples taken on concast industrial plates outside Ti content 0. 11 8
the central segregation.
Incoherent precipitates (che~~cal determination by
l.Aflb.LV_.l electrolytic dissolution, 1 O /.).
Influence of microprecipitates
Steel Nb Ti
3
Base material (10- /.)
HT 1 15 67
C : 91 Mn 803 p : 12 13 Nb 31 Ti 11 8 HT 2 22 98
3
Roll mg sequences HT1 HT2 (including 40 10- insoluble titanium as TiN, TiS)
Initial thickness 24 mm 24 mm
12so·c 3
Reheat temperature 1250'( Coherent precipitates (computed, 10- /.)
Rolling in three stages to 9 mm 9 mm
end thickness
Finish rolling temperature 1ooo'c 900°( Steel Nb Ti Total : Nb + Ti
Accelerated cooling s'c/s 8. Cl s
Stop cooling temperature 5oo'c 5oo·c HT 1 16 51 67 ( 61 /.) •
HT 2 9 20 29 I 2 6 /.)
Properties
* ratio coherent/(total - insoluble)
Re (MP a) Rm (MP a) NACE test
This also shows the great influence played by
HT1 483 630 Passes ) 720 hours the thermal history that a steel undergoes ; and
HT2 427 534 Fails 10 hours specially the influence of finishing temperature as
well as the cooling rate on the run-out table which
exert a determining effect on the microprecipitation
The operating conditions were set according to state on the end product. It is quite clear that
the CRM experience in this field (i8,19l and are this conclusion related to Nb-Ti steel hot-rolled in
described in table VI. It resulted in two varieties such conditions as to develop coherent precipitates
whose properties and performances to the NACE test in the metallic matrix is in good agreement with the
are also listed : one succeeds the test, the other above reported beneficial effect of homogeneous
not. acicular ferritic microstructure - in Nb steels -.
Metallographic examinations have shown that the Accelerated cooling results indeed in acicular
behaviour to SSCC was in close correlation with the ferrite structure (depending on the plate thickness)
distribution of the microprecipitates. These are as well as in a reduced formation of incoherent
very finely distributed and coherent with the precipitation formed at higher temperature. This
metallic matrix in the variety succeeding the test. effect, already detected in Nb steels, is particu-
The other product which fails, shows rougher and larly observed in Nb-Ti steels.
incoherent precipitates (Table VII). Those incohe-
rent precipitates play certainly the role of traps
for hydrogen, saturate in that element which is
present in a quasi illimited amount in this test and
initiate a premature failure.
WRC Bulletin 532

7. INFLUENCE OF A WELDED AREA 4. Heterogeneous structures containing hard or

carbon enriched constituents were found the most
An investigation of sensitivity to SSCC in pipe sensitive to hydrogen embrittlement. In our experi-
steels would not be complete without considering the ments, cracks nave been associated with the presence
effect of a welded area. No line is indeed free from of bands or local area of pearlite, bainite or mar-
welds on the one hand while the important micro- tensite. Segregated areas, in particular at the cen-
structural modifications induced by a welding opera terline, were often prone to SSCC.
tion can be detrimental even in the case of a parent 5. The susceptibility to SSCC of the segregated
steel resisting to cracking. areas has been related to local higher concentra-
Tests were therefore carried out on welded tions of phosphorus. Metallographic observations
zones and specially as concerns the steel listed in have been confirmed by local analysis carried out
Table II, whose base metal passes the NACE test with a scanning electron microscope coupled with a
(Table VIII). Cracks and failures were here systema- wave dispersion spectrometer. Local phosphorus con-
tically recorded and found to propagate either in centrations lower than 0.05 Z allowed to avoid the
the heat affected zone or in the fused metal. occurrence of cracks. Taking into account that the
segregation ratio of phosphorus may range between 2
and 8 depending on certain conditions, the accep-
TABLE VI II table phosphorus global content is between 0.005 and
Influence of a welded area 0.020 'l.. The mean acceptable phosphorus level is to
on the steel of Table Il be associated to the steel chemistry and the condi-
tions of casting, reheating and hot rolling which
METAL NACE TEST are the main parameters influencing the central se-
gregation. The policy of the steelmaker as regards
Base metal Passes > 720 hours phosphorus content in terms of SSCC is therefore
Welded metal Fails 36 hours strictly bound to the control of the harmful segre-
gations and the confidence which can be given to
These results are therefore a complementary this control on a statistical basis.
evidence of the difficulty to satisfy the NACE test 6. As regards manganese, no detrimental effect
as regards not only the steel chemistry but also the was observed for contents near 1 Z in steel contai-
metal microstructure. This latter parameter is real- ning around 0,05 Z C. This is consistent with other
ly of determinant influence and raises the question statements from literature in this regard which
of a structure integrity made in a steel passing the consider that limitations in manganese are to be
NACE test in the as received condition but contai- linked with the steel C-content : higher manganese
ning welds with a lot of different local structures. levels being admissible for lower C contents,
Considering the poor results obtained at the NACE account being taken, however, of the different
test, the question also arises whether this proce- phenomena induced by combined action of carbon and
dure is not excessively severe, as compared to the manganese as regards the 1-Q. transformation tempe-
real environmetal and stressing conditions encoun- rature, the propensity to form banded structures,
tered in practice. There are indeed not so many the effect on the size of the microprecipitates.
relations of SSCC problems occurring in pipe lines. 7. With the experimented steels, neither copper
Although this was not the scope of this investiga- nor nickel were found to exert a significant influ-
tion, it seems that it would be useful to reconsider ence in the NACE test conditions while it was con-
in the future the problem of tt s cracking under firmed that in the BP test, copper was very benefi-
2
actual environmental conditions so as to define a cial.
more realistic laboratory testing procedure. 8. It was shown that the type of microprecipi-
tates, for the part of them which exceeds the stoe-
8. CONCLUSIONS chiometric contents bound to sulfur, may influence
the behaviour of steels ; finely distributed (and
Our research on SSCC on low carbon steels has probably coherent with the metallic matrix) preci-
allowed to draw the following conclusions pitates being preferable to large and incoherent
1 _ Sensitivity of the steels to SSCC was mainly precipitates formed at higher temperature ; such
appraised according to the NACE TM-01-77 procedure difference may be probably related to hydrogen
at the usually requested stress level of 80 Z of the trapping mechanisms.
actual yield stress. Passing the test requires 9. From data reported elsewhere I 18), it
avoidance of fracture after a 30 days exposure_ It appears that hot rolling conditions, accelerated
appeared that the stress corrosion testing unit cooling as well as steel chemistry (C-Mn-N ) influ-
built by C.R.M. with a fully automated gas circuit ence to a large extent the balance betwee~ incohe-
allowed to perform the tests in a reproducible rent and coherent precipitates. In particular, high
manner, and in safe conditions. rolling temperature and accelerated cooling decrease
2. The NACE criterion 1s extremely severe the incoherent fraction. At the same time, such
regarding the steel chemistry and microstructure. accelerated cooling is prone to induce acicular fer-
However, the following statements are based on this ri tic microstructure. Our results have clearly shown
criterion except when the contrary is mentionned. the favourable behaviour of homogeneous acicular
3. Elongated MnS inclusions have to be prohibi- ferrite structure produced by accelerated cooling
ted as they play the role of crack starters even in for Nb steels with a carbon of 0.06 z. As regards
the H.!.C. test. Low levels of sulfur plus globula- the incidence of welding, it was evidenced that the
rization treatment are necessary. The negative low sensitivity of a base metal may be impaired in
influence of sulfur disappears for a level content the heat affected zone, due to a modification of
up to 0.005 Z. microstructure.
WRC Bulletin 532

Now, the main conclusion to derive is the fact (12) F. Sauvage, M. Poupon, M. Bridoux,
that sensitivity to SSCC is the result of a wide set B. Lefevbre , Y. Provou, Realisation de tubes soudes
of interacting parameters, the role exerted by a resistant a la corrosion par H s, Revue de Metal-
2
given one being strongly balanced by the level of lurgie, CIT, February 1985, 129-135.
the others. As already said, conditions to fulfil (13) P. Bufalini, M. Ghersi, M. Pontremoli,
are very constraining, owing to the high, if not A. De Vito, A. Aprile, C. Jannone, G. De Leonardis,
excessive, severity of the NACE procedure applied at Accelerated cooling after control rolling for line-
80 7. yield stress. Nevertheless, this investigation pipe, BTF, Special issue 1985, 86-89.
showed that a steel processed by accelerated cooling ( 14) W. Haumann, F. 0. Koch, Development of
such as the one hereunder mentioned passes the test: Linepipe with improved resistance against Hydrogen
C : 0,06 7. Mn : 1 7. Si : 0,3 7. P : 0,024 7. induced cracking (HIC) and Sulfide stress Corrosion
S : 0,004 1. Al : 0,07 1. Nb : 0,05 7. Cracking (SSCC), Presented at the International
Phosphorus segregation ratio lower than 2. Conference on Pipe Technology, Rome, November 17-19,
Microstructure of acicular ferrite. 1987.
(15) A. Ikeda, M. Nakanishi, T. Kanedo,
ACKNOWLEDGMENTS Y. Sumitomo, M. Takeyama, Improvement of Hydrogen
Sulfide Cracking Susceptibility in Lines Pipes for
This research was carried out under the joint Sour Gas Services, The Sumitomo Search, n· 26,
sponsorship of IRSIA (!nstitut belge pour l'encou- November 1981, 91-104.
ragement de la Recherche Scientifique dans l' lndus- (16) L. Coudreuse, J. Charles, C. Zmudzinski,
trie et !'Agriculture) and ECSC (European Coal and H. Margot-Marette, J.C. Charbonnier, F. Moussy,
Steel Community). Etude de la rupture differee en milieu H s (SSC)
2
d'aciers faiblement allies et de base C-Mn, ECSC
REFERENCES Contract 7210-KB/306, Final Report.
(17) K. Shimogori, Y. Torii, K. Kitahata,
(1) Testing of Metals for Resistance to Effects of Carbo-nitrides on Hydrogen Induced
Sulphide Stress Cracking at Ambient Temperature - Cracking (H!C) in Steels, R.D. Kobe Steel Enginee-
NACE Standard TM-01-77. ring Reports, Vol. 34, n· 2, April 1984, 20-24.
(2) R.D. Kane, Roles of H S in behaviour of (18) J.C. Herman, V. Leroy, Incoherent and
engineering alloys, International Metals Review, coherent Precipitations Induced by Hot Deformation
Vol. 30, n' 6, 1985, 291--301. and accelerated Cooling in HSLA steels, Internatio-
( 3) G. Monfort, J. Halleux, A. Braga rd, The nal Conference on Physical Metallurgy of Thermome-
Automation and Instrumentation of Testing.H s for chanical Processing of steels and other metals
2
cracking, Corrosion Science, Vol. 27, n T0-11, (THERMEC-88), June 6-10, 1988 - Tokyo Japan.
1987, 1145-1152. ( 19) J. c. Herman, v. Leroy, Precipitation
(4) H. Galant, S. Wilmotte, Das Mulpic induite par la deformation a chaud dans les trains
Abkuhlverfahren - Ein Jahr Erfahrung mit der Blech- finisseurs, ECSC Contract 7210-EA/211.
stra sse bei den Forges de Clabecq, Stahl und Eisen, (20) G.M. Pressouyre, J. Dollet, B. Vieillard-
105 (1985 - n' 22 - pp 1191-1194). Baron, Evolution des connaissances concernant la
(5) H. famehiro, T. Takeda, S. Matsuda, fragilisation des aciers par l'hydrogene, Memoires
K. Yamamoto, N. Okumura, Effect of Accelerated et Etudes Scientifiques Revue de Metallurgie, April
Cooling after Controlled Rolling on the Hydrogen, 1982, 161-176.
Induced Cracking Resistance of Line Pipe Steel.
(6) T. Taira, Y. Kobayashi, Development of line
pipe for sour gas service, Presented at the confe-
rence "Steels for line pipe and pipeline fittings",
Grosvenor House, London, October 21-23, 1981.
(7) T. Taira, Y. Kobayashi, K. Matsumoto,
K. fsukada, Resistance of line pipe steels to wet
sour gas, Corrosion - NACE, Vol. 40, n· 9, September
1984, 478-486.
(8) K. Yamada, Y. Satoh, N. Tanaka, H.Murayama,
Z. Chana, K. Itoh, Influence of metallurgical
factors on HIC of high strength ERW line pipe for
sour gas service, Proceedings of the international
conference on technology and applications of HSLA
steels, Philadelphia, October 3-6, 1983. 835-842.
(9) M. Pontremoli, G. Buzzichelli,A. De Vito,
M. Celant, Design, test methodologies and production
of pipeline steels for sour service, BTF
- special issue, 1986, 108-118.
(10) W.B. Morrison, Factors leading to the
development of steels with improved hydrogen induced
cracking susceptibility, ECSC Contract 7210-KE/811
(F5.2/84) - Final Report.
(11) C.L. Jones, P. Rodgerson, A. Brown,
Mechanism of hydrogen induced cracking in pipeline
steels, Proceedings of the International Conference
on Technology and Applications of HSLA Steels.
WRC Bulletin 532

INFLUENCE OF THERMAL TREATMENT ON THE

HYDROGEN EMBRITTLEMENT OF A-516 GR. 60 STEEL
G. Anteri, G. Merlene and J. Ovejero Garcia

Direcci6n Proyectos Agua Pesada
Comisi6n Nacional de Energia At6rnica
1429 Buenos Aires, Argentina
ABSTRACT
Hydrogen embrittlement (HE) tests were It is important to note that most line-
carried out on a commercial A-516 Gr 60 steel. pipe steels and some pressure vessel steels
It was shown that a correlation exists between are used in the normalized state (ferritic-
the microstructural state of the steel and the pearlitic structure) .
HE susceptibility. The martensitic structure
is particularly brittle while the structures In addition, such microstructure very
quenched and tempered at 500°C, spheroidized often presents segregation bands frequently
and normalized, are least sensitive to HE. The rich in manganese, phosphorus and sulphur.
presence of precipitates (Fe 3C, carbides), Therefore, in service, hydrogen cracking tends
which reduce hydrogen penetration as well as to follow these bands.
its transport through dislocations reduces the
HE. These results indicate that is always On the other hand, a catastrophic rupture
necessary to pay attention to the possible by hydrogen damage in the A-516 steel due to
structural modifications that a process of the high hardness (450 HB) of some zones in the
fabrication such as welding can induce and also microstructure of the heat affected zone was
to the selection of the proper thermal reported elsewhere[3].
treatment.
Due to these reasons, the selection of
the proper HE resistant steel and its thermal
treatment deserves a special consideration.
Thus, in this article, the influence of

thermal treatments like quenching, quenching
INTRODUCTION and tempering at 180 and 500°c, normalizing
and spheroidizing, on embrittlement
Hydrogen embri ttlement of a steel (ASTM 60) to be used
which frequently in vessels for the heavy water in-
dustries ammonia synthesis was analyzed.
refineries, heavy water plants,
of heavy economic losses. The role of hydrogen
characteristics (ultimate
Susceptibility to HE is strongly de- striation) of the specimens
pendent on the microstructure as well as other cathodic and un1hj•dico•~ena.teid
metallurgical variables [1,2). examined test.
surfaces were analyzed by scanning
It is well known that martensitic micro- microscopy (SEM).
structures are embrittled by hydrogen, their
susceptibility increasing with increasing their EXPERIMENTAL PROCEDURE
strength level. However, different micro-
structures, characterized by the same strength The present work was performed on a com-
level, could present different resistance to mercial A-516 silicon killed, ferritic carbon-
HE. manganese steel with a low sulfur content.
WRC Bulletin 532

Table 1: Composition of A-516 Gr 60
c Si Mn p s cu Ni Cr Mo v Sn N
wt% wt% wt% wt% wt% wt% wt% wt% wt% wt% wt% wt%
0.12 0.26 1.06 .014 .005 0.01 0.01 0.01 .001 .003 .001 .002
tests were performed in transverse sections of

The composition of the steel is summarized the tension specimen heads.
in Table 1. Rectangular blanks were cut from
a 12.7 mm thick plate. These blanks were heat RESULTS
treated in order to obtain a variety of
mechanical properties and microstructures: Results are summarized in Table 2. The
figures of hardness, ultimate tensile strength,
Normalized (N), as received, 900°C, 58 elongation and reduction of area presented
min, air cooling. there are the average values of each series of
specimens.
Quenched (Q) ; austenitized in a lean
endothermic atmosphere for 30 min. at In Figure N° 1 are given all the RA values
900°C, then quenched in an ice-brine bath. obtained for each state as well as the RA loss
value.
Quenched and tempered at low temperature
(QTL), quenched as above and then tempered It may be seen that the property most
for 6 hours at 180°C. affected by hydrogen is ductility. The hard-
Quenched and tempered at high temperature
(QTH) , quenched as above and tempered for
1 hour at 5oo·c.
80 00 80
Spheroidized (S), normalized plus 80 0
annealing at 700°C for 6 hours. 6go
70
M
6 g 70
ASTM round tensile specimens with a gauge 60 60

diameter of almost 6 mm were machined after
heat treatment (ASTM A-370). Six specimens for ~
.'l 50 8 50
each state of the material were tested; three ;i § 0
of them were charged with hydrogen. O RA uncharged specimens
Hydrogen was introduced into s~ecimens.at

•
"a:
I.() A RA charged specimens
x RA loss
6 40
room temperature by means of cathodic charging 30 6x 30
in a solution of lN H2 SQ4 with a cathodic 6 20

20
current density of 5 mA/cm2 • The charge time M
was 48 hr. 10
10
Tensile specimens both charged and un-

charged were coated with a thin layer of Cu by State QTH 5 OTL Q
dipping in a Cu2 SQ4 to prevent hydrogen loss HSN (227) ( 135l ( 1351 ( 3701 ( 3801
until testing [4).

Fig. 1. RA and RA loss of the specimen tested
Tensile tests of the specimens were
carried out at a strain rate of 2, 5 io·3 s·' on
an Instron testing machine at room temper- ness and the ultimate tensile strength values
ature. The reduction of area was calculated remain almost unchanged.
from measurements on the broken specimens.
From these values the embrittlement index
RA,oss for each state of the material was
calculated as follows: As could be expected from previous works
[ 5] , the steel A-516 Gr 60 shows in the
normalized state a low hydrogen susceptibility
(RA'°"' = 8%). In this state, the material
possesses a microstructure composed of
where RAo and ~ represent the reduction of equiaxial ferrite and fine pearlite. The
area of uncharged and charged specimens, re- inclusions content according to the ASTM E45
spectively. standard was evaluated. Every type of
inclusion were found to correspond to the thin
The fracture surf ace of most of the series with the average number of 2 for Type
specimens, both hydrogenated and unhydro- A inclusions (MnS) and 1 for the other types:
genated, were examined by SEM and fractographs B (aluminates), C (silicates) and D (oxides).
were taken. The microstructure was examined The microstructure is "banded" as a result of
by optical microscopy. Additional hardness chemical inhomogeneities in solidification
which persists through processing. The
WRC Bulletin 532

Table 2: Results
CHARGED HARDNESS UlS ELONGATION RA RA LOSS

STA TE •1.
WITH H HBN MPa. in 25,4mm % •1.
NO 475 37 74
NORMALIZED 135 8
..
YES 471 35 68
NO 1225 13 50
OUENCHED 380 66
YES 1250 7 17
0 T AT NO 1235 12 47
370 27
180 •c YES 1225 9 34
---- -~-- ·--
0 l AT NO 738 22 75
0 227 0
soo c YES 717 24 75
NO 441 43 79
SPHEROIDIZ ED 135 10
YES 431 38 71
pearlitic bands are frequently rich in

manganese. The ferritic grain size is ASTM N'
8 [ 5 J.
The fractograph (Fig. 2) reveals that for

both charged and normal specimens the fracture
me~hanism was of nucleation and coalescence of
voids.
Fig. 3. Ductile fracture in the quenched and

uncharged specimen
When tempering is performed at low

temperature, hydrogen susceptibility decreases
with respect to the quenched material.
However, the index (27%) is near the known
limit of 30% [6].
Fig. 2. Typical fracture surface of
normalized and charge specimen As could be predicted, the possible
microstructure changes caused by the treatment
The quenched specimens showed a high could not be metallographically detected.
susceptibility of hydrogen embrittlement
(~ = 66%) as could be predicted for a In charged as well as in normal spec-
structure mainly composed of lath martensite. imens, mixed fracture is produced with pre-
valence of voids. However, there exists a
Fractographs clearly reveal the dif- certain trend to intergranular in hydrogenated
eren~es between hydrogenated and normal specimens (Fig. 5).
specimens. . In the latter ones, fracture is
mostlr du~tile, characterized by voids (Fig. As could be expected, when steel was
3) while in the hydrogenated specimens a mixed tempered up to 5 oo • c, a tempered martens i te
fr~cture appears prevailingly quasi-cleavage structure was obtained; precipitates of
(Fig. 4). globular cementite in the martensitic laths.
WRC Bulletin 532

Although a complete study of quantitative

metallography was not carried out, the sizes
of a number of these carbides were measured
by means of SEM observations, giving a mean
value of 0.15 µm (Fig. 6).
In this case the index decreased to 0%

indicating the insensitivity to hydrogen in
this state. Corroborating this result, the
fractographic inspection revealed the sim-
ilarities between the fracture surfaces of
hydrogenated and normal specimens, which
resulted ductile in the whole (Fig. 7).
Finally, the spheroidized state exhibited

an intermediate index between QTL and N states.
Metallographic inspection in SEM revealed that
the mean size of the carbide particles is 0.7
µm. The "banded" structure was still present
(Fig. 8).
Fig. 4. Quenched and charged specimen: pre-

dominant cleavage with some plastic
deformation and secondary cracks
Fig. 7. Fracture aspect of charged QTH

specimen
Fig. 5 Fracture surface of QTL and charged

specimen
Fig. 8. Microstructure of spheroidized state
Longitudinal sections of tensile charged

as well as normal specimens were also observed
SEM micrograph of the QTH state by SEM, concluding that the number of voids is
Fig. 6. significantly higher in the latter ones (Figs.
9 a and b).
WRC Bulletin 532

Fractographic observation revealed that

in hydrogenated and normal samples the.frac-
ture was ductile (Fig. 10). However, in the
case of charged specimens, small quasi-cleavage
zones uniformly distributed on the fracture
surface were present (Figs. 11 and 12).
Fig. 10. Fracture surface of spheroidized and

uncharged specimen
Fig. 11. Fracture surface of spheroidized and

charged specimen
Fig. 9. SEM micrographs of longitudinal

section near the fracture for the
spheroidized specimens: a) un-
charged, b) charged
DISCUSSION
As a result of microstructural changes

due to thermal treatments of steels, vari-
ations are produced in:
the diffusion coefficient and solubility

of hydrogen in steel [l, 2].
the sensitivity of steel to hydrogen
embrittlement (1, 2].
Fig. 12. Spheroidized and charged specimen:
This is to say that there exists a cleavage zone
relationship between the microstructure and
the behavior of hydrogen in steels.
WRC Bulletin 532

The substructure of lath martensite

consists of a high density of dislocations ferrite interface represents the moderate traps
arranged in cells. Typically, dislocations of hydrogen [10, 11]. As the interfacial area
density is estimated at between 0.3 and 0.9 x decreases, the number of traps also decreases
10 12 cm/cm3 [7, 8]. The stress field of dis- and, co~sequently, the amount of hydrogen
locations and cell wall in lath martensite trapped in these sites also diminishes. This
provide lower energy sites for hydrogen than means that the hydrogen which is not trapped
the normal interstitial lattice positions. on in the lattice must segregate at the chemical
the other hand, at room temperature, the major inhomogeneities of the "banded" zones
part of the hydrogen introduced during 48 hours (gen7rally rich in manganese, sulphur, etc.)
of cathodic charge will be located in those lead7ng to Hydrogen Induced Cracking. This is
sites. Consequently, both the inter-lath and confirmed by the appearance of brittle zones
intergranular fracture modes (Fig. 4) observed in the fracture surface (Fig. 12)[12].
in this work are the results of hydrogen
segregation in these sites. The structure quenched and tempered at
500 'C presents very small particles of Fe c
The high dislocations density (reversible (0.15 MID) uniformly distributed in the matri~
hydrogen traps) and, therefore, the high level w~ile spher?idized pearlite presents much
of residual stresses present in the fresh bigger particles of Fe 3 C (0.7 MID) inhomo-
martensite is the origin of the high sen- geneously distributed in the ("banded") matrix.
sitivity to hydrogen embrittlement. This means Consequently, the structure quenched and
that dislocations, with high density, act as tempered at 500'C presents a higher number of
hydrogen sources giving origin to a high traps (interfaces Fe 3 C-ferrite) than the
sensitivity of steel to HE (RA,"""= 66%). spheroidized structure. This gives origin to
a higher resistance to HE (Table 2).
Tempering at 180 • C produces carbon
segregation in dislocations and cell walls and
possibly carbide precipitation ( 8 J • A decrease
in the level of residual stresses was also CONCLUSIONS
produced. This means that occupation of the
low energy sites by carbon and carbide and the The results obtained on the hydrogen
consequent decrease in the number of hydrogen embrittlement susceptibility of the A-516 Gr
reversible traps {hydrogen sources) were pro- 60 steel show that:
duced. The decrease of the number of hydrogen
sources and of the residual stresses increases A correlation exists between the micro-
the resistance of steels to HE [9, 10]. This structural state of the steel and the HE
argument is in agreement with the ~results susceptibility, but not with its strength
obtained here. level.
When the steel is tempered at 500'C, Susceptibility to HE is:

cementite (Fe 3C orthorombic) is formed. The very high in fresh martensite,
nucleation sites are frequently martensite lath high in the quenched and tempered at
boundaries [ 7, 8] . On the other hand, recovery 180'C state,
of the dislocations contained between the lath low in the spheroidized and
boundaries also occurs. The final result of normalized state,
the process is a low dislocations density (low least sensitive in the quenched and
"lev~!l" of residual stress) and spheroidal Fe 3 C
tempered at 500'C state.
particles.
The high dislocations density and the high
The small number of dislocations yields level of residual stresses associated with
a low number of reversible hydrogen traps them, ~r~ . the origin of the high
(hydrogen sources). On the other hand, the susceptibility to HE of fresh martensite.
presence of spheroidal Fe 3 C particles means
that irreversible hydrogen traps (hydrogen The presence of small particles of Fe c
sinks) are present. Therefore, both effects uniformly distributed in the matrix a~e
contribute to increase the resistance of steel responsible for the high resistance to HE
to HE. This argument is also in agreement with of QTH (500'C).
our results (Table 2) [10].
Fractography of HE spheroidized micro-
Normalized steel presents a fine lamellar structure showed small brittle zones
pearlite structure (moderate hydrogen traps) (quasi-cleavage in a ductile matrix). The
and a low level of internal stresses. More- "ban~e~" zone constitutes a region
over, its chemical composition shows a low sensitive to hydrogen induced brittle
carbon (0.12%) and sulphur (0.005%) content. fracture.
Due to these characteristics, this steel
exhibits a low sensitivity to HE. This fact
is confirmed by the results obtained in the
present work (Table 2 and Fig. 2). ACKNOWLEDGEMENTS
When the lamellar pearlite transforms into Thanks are due to Department of Materials
spheroidal pearlite, the amount of interfacial C.N.E.A. for the provision of facilities 1 also
area (Fe3 C-ferrite) decreases. The Fe 3 C to Mr. J. Nieto and Mrs. Norma Picco for their
technical assistance.
WRC Bulletin 532

REFERENCES
( 1) "Effect of metallurgical variables on

environmental fracture of steels," I. M.
Bernstein and A. W. Thompson, Inter-
national Metals Reviews, Vol. 21 (1976),
pp 267-286
(2) Proc. 2nd J.I.M. Intl. Syrop. on Hydrogen

in Metals Trans., Japan Inst. of Metals,
21 (1980)
( 3) Weld failure in pressure vessel generates.

Revealing report, Welding Journal, April
1986, pp 57
(4) A. Million, c. Million, "L'hydrogene dans

les aciers et dans les joints soudes,"
Dunond, Paris, 1971, pp 181
(5) Internal report 1981. Hydrogen Embrittle-

ment Group, C.N.E.A.
(6) c. Zmudzinski, L. Breitn and M. Toitot -

Proc. Conf. on Environmental Degradation
of Engineering Materials, Virginia, USA,
pp. 575/83 (1977)
(7) "Tempering of low carbon martensite," G.R.

Sperch, Trans. TMS AIME, 245 (1969), 2553
(8) "Tempering of steel," G. R. sperch and w.

c. Leslie, Met. Trans. 3, (1972), 1043
(9) "The influence of heat treatment and cold

rolling on hydrogen embrittlement of a low
carbon steel," G. F. Merlene, A. J. Funes
and J. overjero Garcia, Mem. Etud. Scien.
Rev. Met. 82 (1985), 337
(10) Hydrogen distribution and hydrogen induced

cracking in A-516 Gr 60 steel, M. I. Luppo
and J. Ovejero-Garcia. 4th International
Conference: Hydrogen and Materials,
Beijing, 9-13 May, 1988
(11) Hydrogen in iron and steel: interactions,

traps and crack paths," R. Gibala and D.
s. de Migilio, Hydrogen in Metals, ed. by
I. M. Bernstein
(12) Hydrogen effects in the fracture of

spheroidized plain carbon steels, H.
Cialone and R. J. Asaro. Proc. 3rd
International conference: Hydrogen
Effects in Metals, 1980, pp 767-776, Ed.
by I. M. Bernstein and A. w. Thompson
WRC Bulletin 532

WRC Bulletin 532

CARBON MANGANESE STEELS FOR SOUR SERVICE -

IMPROVEMENT OF HIC AND SSC RESISTANCE
L Coudreuse, C. Zmudzinski and J. Charles

Creusot-Loire lndustrie, Material Research Center
71202 Le Creusot Cedex, France
L Cadiou
Creusot-Loire lndustrie, Plate Division
ABSTRACT
SSC occurs in all kind of steels, even stainless ones. It is
Metallurgical factors influencing the HIC and SSC behavior generally recognized tliat, the higher the yield strength, the
of low alloy C-Mn steel plates for pressure vessels are poorer the resistance of the steel. However this rule is not
investigated. Numerous HIC and SSC tests have been alway valid. For example it has been possible to develop
performed on a great number of plat.es. From the results high yield strength steel with good SSC resistance by
obtained it is shown that the behavior of steels can be addition of microalloying elements such as V, Ti, Nb, wich
improved by : delay the tempering softening and act as hydrogen traps
1) Use of very clean steels : The sulfur and (5,6). On the other hand low alloy C-Mn steels with low
phosphorus content must be kept as low as possible. ~t is mechanical characteristics can be very sensitive to S.S.C
shown that for very low sulfur content (< 0,001 %), calcmm (7). In fact it is the micro structure that governs the
treatment is not necessary since sulfides are already i:;lobu!ar. hydrogen embrittlement behavior of ferritic steels. And for a
Oxygen has also to be kept low (< 0,002 %) to av01d oxide given class of steels the following classification can be
inclusions. established (from the poorer to the better) : quenched
structures, normalized structures, quenched and tempered
2) Avoiding segregation : It is sho~n that structures. For low alloy C-Mn steels with normalized
cracking occurs in hard segragated bands. HIC cr~ckmg has structures (ferrite/perlite) SSC is a complex phenomena
been found to be related to phosphorus segregation. C and since internal HlC cracks can be present irI tlie metal.
Mn content must be kept as low as po?sible, improv~ment of
mechanical properties can be obtained by alloying and During these last years, according to the quantity of steel
microalloying additions. concerned, most of tlie research has focused on the problem
of line pipe steels. These steels are now produced by
continuous casting and controlled rolling. Steels with
1 - INTRODUCTION improved HlC and SSC resistance have been developped.
These are low sulfur, ultra low carbon (< 0,05)
When they are used in hydrogen sulfide environnments low microalloyed steels with an acicular ferrite homogeneous
alloy steels can be severely embrittled. Hydrogen Induce microstructure obtained by an accelerated cooling after
Cracking (HIC) and Sulfide Stress Cracking (SSC) are the rolling. In such conditions X 80 pipeline steels can be
two ways by which hydrogen degradation occurs. produced with a good HIC and SSC behavior (8, 9).
However these processes are effective up to 30/40 mm
HIC manifests itself in "soft" steels in the form of surface thicknesses (which are enough for line pipe). But H2S
blisters and/or internal cracks in the absence of an applied
stress. Hydrogen absorbed from the solution is trapped on problems can also be found in pressure vessels (1 O) where
microstructural defects, such as non metallic inclusions or the thicknesses concerned can be greater than 40 mm. In
carbides. High hydrogen pressures can be developed and such applications the use of normalized steels is still the rule
cracks can initiate and propagate when locally the stress and the ways of improving the HIC and SSC resistance
intensity factor Kr becomes greater than KrH (threshold for cannot be the same than for line pipe steels.
crack propagation in presence of hydrogen). So cracking
depends on many parameters : sizes of the defect, hydrogen
In this paper experimental results obtained on normalized
concentration, microstructure, internal stresses (1,3). For a C-Mn low alloy steels are presented. The influence of some
given microstructure and a given type of defect one can metallurgical parameters is analysed and discussed. Finally
consider that there is a critical concentration of hydrogen for ways of improvement of the H2S resistance of normalized
cracking (4).
WRC Bulletin 532

3
Table 1 : Range of chemical compositions of steels tested (lo wt %)
c s p Si Mn Ni Cr Mo Cu v Nb
Mini 80 0.3 5 210 800 100 100 40 70 0 0
maxi 210 14 16 400 1550 500 285 95 250 100 40
Table 2 : Range of mechanical characteristics 3 - EXPERIMENT AL RESULTS · INFLUENCE

OF METALLURGICAL PARAMETERS
3.1 - Influence of cleanliness

Y.S. U.T.S Steels
(MPa) (MPa) grades It is well known that H.I.C cracks can initiate on defects of
the structure where high hydrogen pressure can be
developed. Among these defects non metallic inclusions are
the more detrimental. Using the fracture mechanic concepts,
280 420 A 42 FP (NFA 36-205) it can be shown that little, globular and randomly dispersed
BS 1501 224 gd 400 inclusions are less susceptible than big, elongated and
aligned inclusions. Among these the elongated sulfide
A 516 gde 60 inclusions have been found to be the worst.
A 48 FP (NFA 36-205) We have shown by QTM inclusion analysis that the

BS 1501 224 gde 460 reduction of the sulfur content leads to a decrease in the
number and the size of sulfide inclusions and that they
A 516 gde 70 become more globular. This is shown in figure 1, where the
A52 sulfide shape parameter has been reported versus the sulfur
content.
E 355 (NFA 36-201)TTSTE 355
BS 1501 224 gde 490 Another way to achieve globularisation of the sulfides is to
BS 1501 225 gde 490 perform a calcium treatment. Influence of calcium treatment
on the sulfide shape parameter is also given on figure 1. It
can be seen that calcium treatment is effective when the S
E 420 FP (NFA 38-201) content is greater than 0,002 %. On the other hand for very
••
,, TTSTE420 low sulfur contents, sulfides are already globular and a Ca
treatment is not necessary. Even it can be worst, since it can
be responsible of the formation of big oxide inclusions .
580 770 E 460 (NFA 36-201 The correlation between the HIC behavior and the sulfur
content is shown in figure 2. It is clear from this figure that
a decrease in S content is effective to improve the HIC
behavior. For S content > 0.002 %, the calcium treatment is
effective in reducing the HIC sensitivity but this is not true
plates are presented. Shape p&rameter
(LIT)
2 - MATERIALS AND EXPERIMENTAL
PROCEDJJRES
4 4 _,..
Tests have been performed on a great number of heats with
variable chemical analysis. The range of chemical
composition of steels tested is given in table 1, range of
mechanical characteristics are given in table 2, with the
corresponding grades. Plates with thicknesses from
15 mm to 150 mm have been tested.
Tests have been performed either at the normalized state, or

after stress relieving treatments.
0 '-...__..._....__,_-'-_,__..__.__.__._~~__,___,.__.
0 4 10 12 14
Steels are produced by the way of ingot casting. 3
SULFUR CONTENT ( t 0 - %)
On these different steels HIC tests and SSC tests have been
performed according to the NACE standards TM 02-84 and
TM-01-77. For the HIC tests both the TM 02-84 solution
and the NACE solution have been used. SSC tests have Figure 1: Influence of sulfur content on the shape
been performed using proof rings with the NACE solution. parameter (Length/Thickness).
WRC Bulletin 532

The influence of sulfur content on the SSC behavior is

shown on figure 4 where the parameter cr72o'Re have been
C.L.R (%)
100------------,-------. reported versus the sulfur content, for steels with a high
yield strength (Y S > 400 MPa). Results are shown for
IH-1.C TESTS IN 'BP SOLUTION! SSC specimens taken in three orientations : long direction,
transverse direction, and short transverse direction. It
appears that the correlation, between sulfur content and
SSC behavior is very marked for the short transverse
direction. In fact for this orientation, if there are HIC
I 0 ··- ...................... ·······~---···
cracks, they will propagate easily under the effect of
stresses. The correlation between SSC behavior and sulfur
content is not so good for the other orientations since the
scatter is large. However the better results are obtained with
the low sulfur steels.
SIGMA 720 I Y. S.
0,9
0,8
0,7
IO 100 0,6
3 0,5
(SULFUR CONTENT (10 . % ))
0,4
SSC TESTING NACE
0,3 SOLUTION
0,2
Figure 2 : Influence of sulfur content
on the maximum CLR value. 0,1
10 15
3
SULFUR CONTENT (10. %)
for low sulfur content. In fact in low sulfur steels, in the
absence of segregation, the HIC behavior is very good Figure 4 : Influence of sulfur content on the
and only very little cracks (1 < 0,1 mm) can be found SSC behavior.
initiating on oxide inclusions. An example of such a crack is
given in figure 3. The worst results found in low sulfur
steels are associated with a high oxide content. Thus 3.2 - Influence of segregation
calcium treatment for very low sulfur steels can become
harmful, because of the great tendancy of calcium to form H.I.C behavior depends not only on the cleanliness of the
oxides. steel but also on the microstructure. Cracking can initiate
and propagate more easily in less ductile microstructures.
Thus segregated zones with bainitic or martensitic structures
are harmfull. Segregation depends on the chemical
composition, and also on the casting process. It is well
known that for plates made from continuous casting the
segregation is located at 1/2 thickness, whereas for ingot
casting segregation is found at the top of the ingot. For
ingot casting the rate of segregation depends on the ingot
weight and shape. In order to avoid top segregation a
minimum cropping allowance is imposed in the mother plate
(part corresponding to the top of the ingot). In figure 5 the
hardness profile through the thickness of a plate has been
reported for different percentages of cutting from the top of
the plate. As the percentage of cropping increases, the
homogeneity through the thickness of the plate becomes
better. It is clear that a minimum percentage of the top of the
plate (corresponding to the top of the ingot) must be
removed in order to avoid inhomogeneities in the plate. The
cutting level depends on the casting parameters.
Quantitative electron microanalysis has been performed

through segregated areas of different plates in order to
correlate the occurence of cracking to the rate of
segregation. Results are given in table 3. Examples of
x200 profiles are given on figure 6. From table 3 it can be seen
that the main difference between cracked areas and
uncracked one's is the segregation of phosphorous. Rate of
segregation between 235 and 670 % are found in cracked
areas whereas segregation is between 100 and 230 % in
Figure 3 : Little crack initiating on an oxyde inclusion. uncracked one's.
WRC Bulletin 532

IN 10 N
~
a.
0,02
0 ,01
lD 15 20 25
HV 10 N
%
1,5"
250 I ,90
~1,70
200 2
2
I, 50
150 Ot .~ ............
1'30
10 15 20 25
HV 10 N
No crack No crack
I, IO
µm
2~
a/
·~tv~~f!WJ 0,03
150
~
a.
0,02
Figure 5 : Hardness profile through the thickness of a
25 mm thick plate influence of headcutting.
I.a.!ili:....J. 0,01
Measurement of segregation rates in

cracked and uncracked areas for different heats
c p Si Mn
%
Analysis (10" 3 Wt %) 114 7,7 290 1340 1,90
Cracked 350 14 21
Seg. (%)
Uncracked 100 13 23,5
Analysis(lO-~ - ' Wt%) 142 11.4 365 1360
Cracked 488 20 29
Seg. (%)
Uncracked 233 23 23
1,30
Analysis (10" 3 Wt%) 165 8,8 268 1340
No crack
Cracked 670 51 53 1, 10
Seg. (%)
Uncracked 115 37 37
3
Analysis (10 Wt %) 80 8 320 1450
Cracked
(2 zones)
235 17 18 b/
Seg. (%) 402 18 21
Uncracked 135 16 16
Figure 6 : Segregation profiles through a crack area (a)
Seg.% Ayerage % :teg . % ana!ysjs x 100 and an uncracked one (b).
% analysis
WRC Bulletin 532

For the other alloying elements, the difference. is 1?"ot so

large and in a few cases, the r~tes of segregat10n. m the
uncracked areas are higher than m cracked areas. It 1s clear
that HIC can be associated with phosphorus segregation.
Some authors have correlated the hardness in segregated
zones to the occurence of cracking (11) - Examples of
continuous hardness measurements performed through
cracked and uncracked areas are given on figure 7. Because 100
of the crack it is not alway possible to measure the hardness Cl B.P. Norm.
of the cracked area, however it has been observed, that Ill B.P. Temp.
hardness is generally higher in cracked zones. But in fa?t it !;fl. 80
i::i:: A NACE Norm.
is not possible to give a critical hardness value for cracking,
~ Ji. NACE Temp.
because hardness value depends strongly on both the size of u 60
the indentor and the width of the segregated area.
There is not a macroscopic parameter that can be related to 40

the intensity of segregation. But since in se¥Tegated areas
there is an increase in hardness (due to the increase of C,
20
Mn and P content), one can try to establish a relation ____________ J._J._____________________________ _
between C content, Mn Content or Ceq with the HIC and 10
8
SSC behavior. On figure 8 the relation between the CLR 6
value and the Mn or the Ceq (C + Mn/6) contents is shown 4
2
for a large number of tests. There is not a perfect correlation Cl
O{
between the CLR value and the Mn content or the Ceq,
nevertheless the risks of cracking increase when Mn content 0,8 1,0 1,2 1,4 1,6
increases or when (C + Mn/6) increases.
Mn (wt%)
100
Cl RP.Norm.
Ill B.P.Temp.
80
*
i::i::
~ 60
A
I"'
NACE Norm.
NACE Temp.
Cl
u
100
40
Ill
50
Crack
20
Hl
-------------~-~--------
o.o 1.0 2.0 3.0
8
6 Ill Ill
• Cl
d (mm) 4 Cl
a) cracked area 2 Ill
Hv 10 N
IH
250
Mark 250 300 350 400 450
C+Mn/6 wt%)
200
Figure 8: Influence of Mn content (a) and Ceq (C + Mn/6)

content (b) on the CLR value low sulfur steels
150
(S < 0,002 %).
100
From the two plots in figure 8 it can be proposed that a
good HIC behavior is obtained when C + Mn/6
< 0,35 % and when Mn< 1,3 %. However good results
50 can be obtained with higher carbon equivalent or with
higher Mn content. This indicates that chemical composition
1.0
is not sufficient to describe segregation (other parameters
2.0 3 .o
such as, plate thickness, ingot weight and shape act on the
b) uncracked area d (mm)
phenomena). The influence of the thickness of the plate is
important since the risk of transformation of segregated
areas into martensite or bainite increases when the thickness
decreases. Then for the lower thicknesses more precautions
Figure 7 : Hardness profiles through cracked areas (a) must be taken. A very good cleanliness is necessary and
and uncracked one's (b).
ingot shape and weight must be adapted to the final product.
Chemical composition has also to be balanced in
consideration of the final product.
WRC Bulletin 532

Concerning the influence of segregation on the SSC , it has

not been possible to establish a correlation between SSC
behavior and a metallurgical parameter such as Ceq. In fact Acid Solution
for SSC tests performed on tensile specimens in the long or (IN H2S0 4+AS20,)
transverse direction, there is no correlation between the
occurence of HIC cracks and the threshold stress for Specimen
cracking. This point is discussed in another paper
presented at this conference. (12)
The effect of tempering treatment on the HIC behavior of

0
selected steels with hard bands is shown in figure 9. A large Hydroeen H!Jalysis
Hydrwen chargjng
decrease in the CLR value can be obtained with the
(Vacuum extraction at 650 °C)
tempering treatment. This can be associated with the
softening of the hard bands (confirmed by microhardness
measurement), and to a decrease of internal stresses in
segregated areas.
Ii Normalized
~ Temp. 580 °c 2h " Cracks
~ ll No crack
CJ 60
40
No crack
Examination for cracks on pnlj5hM qps.s sertinn5

20
Figure 10 : Principle of the CK calculation.
0
Ref. A !:l c D E F CLR=l0% CLR=29% CLR =0% CLR=0%
101===t=~=t===t===1
Figure 9 : Examples showing the influence of tempering
s0.. • i
treatment on the CLR values 0..
'-'
(303)* : maximum microhardness values (Hv 50 g)

measured on bainitic areas.
The concept of critical concentration for cracking is

useful to describe the HIC resistance (13). Then it can be
attractive to class the steels versus the CK value. The
technique used for the determination is described in
figure 10. Little specimens (0 8 m) are nyc1ro1gen
charged by charging in an acid solution
S04) at different current densities (in order to have different
Norm. Temp. Norm.
hydrogen concentrations). After charging, the specimens
are stored in liquid nitrogen, before hydrogen analysis. C : 0,125 ; Mn : 1,340 C: 0,113; Mn :1,350
Hydrogen analysis are performed by vacuum extraction at S : 7ppm ; Nb:0,030 S : 6ppm ; Nb :0,027
650°C. The specimens are then cut and metallographically
examined to look at the cracks. Thickness : 12mm Thickness: 32mm
Examples of such a determination on two plates (at the

normalized and tempered are given in figure 11. The
CLR values measured during tests are also given. The
first plate is a low thickness one (12 mm) with a
ferrite/parlite/bainite rnicrostructure homogeneously
distributed. The second one is a 32 mm thick plate with an Figure 11 : CK determination on two plates at the normalized
homc•ge:necms ferrite/parlite microstructure. state and at the normalized and tempered state.
WRC Bulletin 532

The CK values are comprised between 0,8 and 2 ppm.

There is no correlation between the CLR value and the CK.
The plate with the ferrite + bainite structure has a CK value,
of about 1,5 ppm at the normalized state (CLR = 25 %)
whereas the CK value is about 0,8 ppm at the tempered
state (CLR = 0). For this plate the higher the CK value the
higher the CLR is. On the other hand for the plate with the
ferrite/parlite structure there is a decrease of the CK value
with tempering. I
I
I
These results show that the CK value is not sufficient to I e I
4 I I
estimate the H.I. C. sensitivity of a steel. In fact for a given I
. .
I
external hydrogen activity, the hydrogen concentration in 3 I
the steel depends on its trapping capacities (ie, I
microstructure, inclusions, precipitates ... ), its 2 / .. I

,
.. ,
I
microstructure and on the hydrogen absorption I I
I
mechanisms. I I
I
I I
I "' •
UA.~_,,.-....,.._D-.111-_,.541411<-
"'"'
~~--.-~~~~......:.~Pfl'Yi
1
I
The CK value must be compared to the quantity of 1.5 2 2.5

hydrogen (CH) that can be absorbed in the steel during l0- 3 wt 'JI. c s p Sl Mn Ni Cr Mo Cu
exposure to a given environment. ""
Cr 3'55
120 l s 3'0 l3D 2"" 100 40 140 32
( TSTE 355)
HEAT TEAT Nonma l i zed 900" C + 58Q•C - air

On figure 11 we have reported the quantity of hydrogen
introduced in little cylindrical specimens during exposure to
the NACE solution (squares on the figure). Cracks have
only been found in the steel with the CLR value of 28 % ; Figure 12 : Correlation between the CLR value and the
2,1 ppm Hz have been introduced during exposure to the amount of hydrogen collected by degassing
at 45°C for 72 h in a glycerin bath.
NACE solution, that is higher than the CK value. For the
steels where cracking has not been observed during HIC HEATED LADLE REFINING
test (CLR = 0), the quantity of hydrogen introduced during
exposure to the NACE solution is below CK, so it is not
surprising that these steels are not sensitive to HIC during
the NACE test.
In figure 12 are given results obtained during HIC tests on a
TT STE 355 steel. The CLR values measured during HIC
tests in NACE solution are reported versus the amount of
hydrogen collected by degassing at 45°C for 72 h in a .·
glycerin bath. It has been substantiated that for such a steel
more than 90 % of the total amount of hydrogen is degassed
at this temperature. So the values measured can be
compared to the values obtained for CK determination. For l~NQJ'
......"-........,
FINAL AOJUSTEMEN
01' CH!MICAL
ANAL V$1S
1
--C~A-MBER
LOC<
DEGASSING
the heats tested(> 30), it is shown that below 1,2 ppm there I
is never cracking whereas above 1,7 ppm there is alway

cracking. The higher the HIC value, the higher the ::.
hydrogen content, in fact this is not surprising since part of
the analised hydrogen comes from the cracks. L..r--=-1
From these results it is shown that it is difficult to estimate Figure 13 : Principle of the Heating Laddle Refining process
the risks of cracking from the analysis of hydrogen content.
However for low alloy C-Mn steels with a normalized or It has been shown in our studies that the HlC resistance of
normalized + tempered microstructure, it can be proposed very low sulfur steels (S <0,001 %) can be very good even
that below l ppm risks of cracking are very low and that without calcium treatment. Calcium which is effective to
above 2 ppm cracking can be expected. globularise sulfides when S > 0,002 %, can decrease the
cleanliness low sulfur steels by formation of oxides.
In very low sulfur steels little cracks (l < 0,1 mm) can
4 - DISCUSSION • IMPROYEMENT OF H E. initiate on oxides. Consequently the oxygen level has to be
RESISTANCE kept low. The new steel making processes available allow
the manufacture of very clean steels. An example of such a
From the above results it is clear that to effect an melting process is shown on figure 13 : It is called the
improvement of the HIC and SSC resistance one needs to Heating laddie refining process, where dephosphorization
act both on the cleanliness of the steel and on the and desulfuration are performed in different !addles under
segregation problems. vacuum. With such a process, sulfur contents as low as
0,001 %, phosphorus contents less than 0,005 % and
oxygen contents less than 0,002 % can be obtained.
WRC Bulletin 532

TABLE 4
Results of SSC tests on

tensile specimens taken through welded joints 30
T TSTE 355 STEEL
00
.µ NACE Solution
Temp. HV max. CJ (MPa) % YS Result .-1
Treat. ::I
!I) Thickness 10nomm
610°C - 2 h 293 250 70% 0 failure in 480 h w
(3 specimens) ~
20
570°C - 2 h 270 250 70% 0 failure in 480 h 4-1
(3 specimens) 0
" 285 80% O failure in 480 h
(3 specimens)
~z
Limiting the occurence of segregation is quite a more 10
difficult problem. It has been shown that in hard bands HIC
can be correlated to the phosphorus segregation : The rate of
segregation for phosphorus is higher in cracked areas than
in uncracked one's. Thus a decrease in the amount of
phosphorous can improve the HIC resistance of the hard
bands. Another way to limit the occurence of hard bands is
to decrease the Ceq. However the decrease of Ceq is not 0
compatible with an increase of mechanical properties 0 0/1 1/2 2/3 3/4 4/5 5/6 6/7 7/8
characteristics. Decrease in C and Mn content must then be CLR (%)
balanced by hardening elements such as Ni, Cr, Mo or by
microalloying elements (V, Nb, Ti). On figure 14 are
presented HIC results obtained on a T STE 355 steel with a Figure 14 : Histogramme of HIC results obtained on steel
Nb addition. Tests have been performed in the NACE TI STE 355 in NACE solution.
solution. Results are quite good (most of the CLR value are
below 1 %). But there are other problems with these kinds
of steel. Because of Nb addition the hardness value in the
H.A.Z, of welds even after tempering treatments can be
quite high (until! 290 HV). This hardness level can be f'QNl'U T~IQN~
associated to the Niobium carbide precipitation. In most of
the specifications hardness values below 248 HV are The influence of some metallurgical parameters has been
required, an then such a steel is not acceptable. Some SSC investigated from numerous HIC and SSC tests performed
tests have been performed in order to check the H2S on C-Mn steels with different chemical compositions. It is
behavior of hard H.A.Z. SSC specimens have been taken shown that.
through welded joints with high hardness values (up to
290 HV). The H.A.Z. has been localized in the middle of - Sulfur content must be decreased as low as possible, to
the specimen. Results are given in table 4. No failure have have little and globular sulfides. It is shown that for these
been observed in 480 h at 70 % and 80 % of the base conditions (S < 0,001 % ) calcium treatment is not
material yield strength value. On the other hand failures necessary, and can even be detrimental due to the formation
have been shown to occur in soft areas of H.A.Z. in oflarge oxide inclusions.
linepipes, when low carbon steels are used. It is clear that
for new generation of steels (low C, microalloying - Oxygen must also be kept so low as possible in order to
additions) hardness value alone is not sufficient to estimate increase the cleanless of the steel and so as to reduce the
the HzS behavior. More work is needed on the hydrogen potential sites of crack nucleation.
embrittlement resistance of H.A.Z. microstructures (in
relation with the precipitation mechanisms). - Phosphorus content must be reduced particularly for high
strength C-Mn normalized plates. This is explained by the
In fact when Ceq increases, and thicknesses decrease more high segregation effect of phosphorus which, with
attention must be focused on the cleanliness of the steel and simultaneous increase of carbon and manganese, promote
on the phosphorus content. Phosphorus can have two the transformation of austenite to martensite or baini.te.
negative effects. Firstly it can favor the martensite and
bainite transformation during cooling in segregated areas. - Segregation must be kept as low as possible in order to
Secondly it can have an embrittling effect on the grain avoid the formation of hard zones sensitive to HIC. The
boundaries. best ways are the reduction of carbon and manganese
additions particularly for small thicknesses plates. However
In spite of all the precautions taken it seems difficult to one must point out that good HIC results can be obtained
produce resistant plates in the normalized state with a for higher alloyed steel but, in that cases, special cares must
yield strength level higher than 400/450 MPa. The only be taken on the cleanliness of the steel and on the casting
way, that has to be investigated, is to use quenched and parameters.
tempered steels. In this case the C content can be reduced,
and the low carbon tempered bainite microstructure is - Stress relieved treatments can help to improve HIC results
expected to be H.I.C. resistant. by softening hard bands and by reducing internal stresses in
segregated areas.
WRC Bulletin 532

- 1;ficro. additions can. help in decreasing the C. equivalent

rat10 without decreasmg the mechanical properties. Such APPENDIX
steels can provide for the base metal and H.A.Z. good HIC
and SSC results. Nevertheless the hardness in HAZ can be CONTINUOUS HARDNESS MEASUREMENT
higher than the required 248 HV maximum values and one
must be careful in welding parameters in order to avoid
excess of intergranular precipitation in H.A.Z. More work An apparatus has been developed in order to perform
has to be done in this field. dynamical continuous hardness measurements. A scratch is
performed on the surface of the specimen without cutting
Finally we have shown some comparison between CK nor tearing. The depth of the groove performed is linked to
values and HIC results. It is observed that the behavior of the hardness of the metal.
the different steels in hydrogen charging capacity during
HIC tests are very dispersed. This explains partially the The specimen is fixed in moving supports, the vertical
differences observed in HIC results between several steels. motion of the indenter, made of diamond or carbide, is
measured with an accuracy of 0, 1 µ.m. A first record of the
roughness is made with the unloaded indenter, the second
record gives the surface profile when the indenter is loaded.
The difference between the two measurements gives the
REFERENCES depth of the groove.
1 G.M. PRESSOUYRE, proc. lfil: Int. Conf. on "Current The experiment has been automated : Both the displacement
Solutions to Hydrogen Problems in steels" Ed. by C.G of the specimen and the vertical motion of the indenter are
INTERRANTE and G.M. PRESSOUYRE, ASM, Publ., sent to a micro computer. The hardness values are given in
1982, p 18. Vickers hardness scale. The measuring step depends on the
length of the scratch : For example for a 50 mm length
2 M. IINO, Met. Trans., 9A, 1978, p. 1581-1590. scratch the measuring step is 16 µm.
3 T. TAIRA et al, Corrosion, 37, (1981), 5. This technique is well adapted to measure hardness
heterogeneity in narrow areas (segregation, weld
4 J. CHARLES, L. COUDREUSE, G.M. structures ... ) An example of hardness profile trough a
PRESSOUYRE and L. CADIOU welded structure is given on figure Al.
Corrosion 87, San Francisco, March 9-13, 1987, Paper
number 88. HV SN
5 J.C. CHARBONNIER, H. MARGO-MARETTE, !vt 300
AUCOUTURIER and A.M. BRASS
Mem. Sc. Rev. Met, 5, 1984, p 225-238.
250
6 A. CHA VANE, M. HABASH!, G.M. PRESSOUYRE
and J. GALLAND
Corrosion, 42, N° 1, Jan. 1986, pp 54-61.
7 J.C. TURN JR., B.E. WILDE, and C.A. TRO ANOS

Corrosion, 39, N° 9, 1983, pp 364-370.
8 H. NAKA TE, T. KOBAYASHI, Y. YAMAGUCHI

Tetsu to Hogane, 70, May 1984, p 548.
HA.Z. B.M.
9 P. BUFALINI, M. DE LISI, A. DE VITO,
M. PETRUCCI
Int. Conf. on Pipe Technoiogv Roma 17/19 Nov, 1987.
10 H.I. Mc HENRY, D.T. READ and T.R. SHIVES ~ : Hardness profil through a welded structure
Mat. Performance, August 1987, pp 18-24.
11 T. KANERO A. IKEDA - T. HA SHIMOTO

Tetsu to Hogane, 88, Dec. 1982, p 1286.
12 H. MARGOT-MARETTE, J.C. CHARBONNIER

L. COUDREUSE, J. CHARLES
This Conference.
13 L. COUDREUSE, and J. CHARLES

Corrosion science, 27, N° 10/11, 1987 pp 1169-1183
14 Y. KOBAYASHI, K. UME, T. HYODO and

T. TAIRA
lb 13, pp 1117-1136.
WRC Bulletin 532

WRC Bulletin 532

ON THE EFFECT OF NON-METALLIC INCLUSIONS ON

HYDROGEN DAMAGE IN Al-KILLED STEEL
Estela R. de Schiapparem
Departamento Materiales
Comisi6n Nacional de Energia At6mica
i429 Buenos Aires, Argentina
Sixto Prado
Facultad de Ciencias Exactas
Universidad Nacional de Trujillo, Peru
ABSTRACT the type of inclusions which produces HPC.

Among the most important points of view,
The failure analysis in a pressure vessel Ciszewski [5] mentions the difference in te~
Al- Killed steel of petroleum industry was selated stresses, generated during quenching
made. After 40 years of service, this reactor around some inclusions, mainly oxides, in
presented hydrogen pressure cracking (HPC). tempered steels, caused by low thermal coef-
The effect of metallurgical variables and in- ficient compared with that of the matrix.
clusions was considered. Smialowska and Lunarska (6] analyse the type
Accurate determinations of chemical compo- of the matrix-inclusion bonding in low car-
sition of the different inclusions compounds bon steel. Silicates will be the inclusions
were performed by energy dispersive analysis which may initiate internal blistering since
(EDAX) of replica extracted particles.Clusters they have the weakest type of matrix bonding.
of alumina inclusions appear to be preferred Wilde et al [7], Pressouyre et al [8],
sites for the occurrence of blisters. Complex Yoichi et al [9] among others researchers
inclusions with the phases: Ca-aluminates, mu~ have observed that the most advantageous
lite, alumina and manganese oxide frequently sites for hydrogen damage initiation are
presented blisters. Blisters appeared to be elongated manganese sulphide inclusions. The
nucleated in voids produced during working of difference in opinion may be partly due to
the steel. Blisters were never found in manga- the difficulties in characterization of the
nese sulphide inclusions. inclusions in steel matrix, Schiapparelli et
The relations between the stress field al [10, 11].
around some inclusions and inclusions type
present on susceptibility to HPC was taken EXPERIMENTAL AND RESULTS
into account.
The failure analysis in a specimen
INTRODUCTION extracted from a catalytic hydrogenation fuel
oil pressure vessel was made. The reactor
Hydrogen pressure cracking is most preva- performed with a hydrogen gaseous and hydro-
lent in low and medium strength alloys like carbon mixture at 540°C temperature and
carbon steel used in the petroleum industry, 25xl05Pa pressure, was in service for 40
and it is observed when the material has been years. The analysed specimen dimensions are
exposed, for example, to hydrogen charging, 140x28xl0 mm and the chemical composition
gaseous charging at high temperatures, cathod- (W%) of the specimen is the follO'\tling: C:
ic charging or corrosion reactions. In the 0.14; S: 0.006; Si: 0.25; Al: 0.05; Mn: 1.11;
surface the hydrogen is absorbed and adsorbed P: 0 . O1; Mo: 0. 0 2; Cr: 0. 0 2.
by some mechanism that is not clear yet and
afterwards put into the steel as atomic hydro- Inclusion Content
gen and diffuses inward where i t can precipit-
ate as molecular hydrogen at internal voids, Size, distribution, morphology, shape
inclusion-matrix interfaces and i t can build and number of the inclusions were studied by
hydrogen gas pressures great enough to force conventional light microscopy, quantitative
their expansion by plastic deformation or microscope, and scanning electron microscope
cleavage, leading to the coalescence of micro- (SEM), in the later, etching with Br/CH30H
cracks or microvoids [1,4]. In the bulk the solution. In order to study the chemical co~
voids or microcracks above mentioned can be position of the phases that comprises the
associated with hard inclusions, hence the inclusions, a non-conventional technique was
susceptibility of the steels to H~C could used [11]. Inclusions were removed from the
depend on the inclusion content and other fac- matrix by epoxy replicas,etching with Br/
tors. There are several opinions, concerning CH30H solution, and analysed with EDAX. The
WRC Bulletin 532

Br/CH30H solution was chosen, in order to presented HPC, which consist of microvoids in
avoid the disolution of the inclusions consid- the interface matrix hard inclusions.
ering that the inclusions phases are nobler The same inclusions were deformed in the
than the matrix. The same formalism of thin sense of becoming larger in the direction of
foil analysis was used in our work. Stoichiom- rolling. All the observations were made on
etric calculation was possible due to the abs- the surface of the sample in planes contain-
ence of the matrix effects. ing the rolling direction, Fig. 3, giving a
The type and number of the inclusions pr~ standard metallographic preparation and exam-
sent in this material are given in Table I. ined for internal crack by SEM. Table 1
The microstructure of this material is shows the microcracks associated with dif-
shown in Fig. 1. ferent types of inclusions.
Rolling
Direction
Fig. 3. Extraction of test specimen///

Surfaces observed by SEM.
Magnification: 100 Fig. 4 and 5 show microstructures obtain

Fig. 1. Photomicrograph of the microstructure ed by SEM where we can see the microcracks
observed by optical microscopy associated to the inclusions and in Fig. 6
MnS-Al203 inclusion is shown (sulphide type
II)according to Sims clasification which is
not associated with HPC.
a) 110 mm length b) 45 mm length

Fig. 2.- Photomicrograph of oblate macrovoid
observed in the bulk of the material, 110 mm
length and of a fissure on the exterior sur-
face of 45 mm length.
Fig. 4.4-9 1 µm. SEM micrograph of Al203
This macrovoid produced a fissure in inclusion which presents HPC.
the exterior surface of approximately 45 mm
length. The surface of the above mentioned
Ultrasonic Inspection
macrovoid was observed by SEM and EDAX, many
macrosegregations of alumina particles of
An ultrasonic inspection of the specimen
about 1 mm length each one were found. Taking
of 240xl40x28 mm was made. A Kraunt-Krames
into account the working temperature of the
Mod. USE 1 with a detector SE of 4 MHz and
reactor i t was thought the possibility of the
¢ = 10 mm was used. This detector was used
formation of CH4 together with H2, but decar-
because it was necessary to evaluate the
bonization was not observed in the matrix
piece through one face only. The specimen
around the macrovoid. was divided into quadrille rolling of 20x20
Observations by SEM far from the region
mm and 50x50 mm in order th make easier the
of the macrovoid showed that all the material
information obtained. Four measurements were
WRC Bulletin 532

70 Metallurgical Studies.for Steels for Sour Service Environments
made for an area of 20x20 mm to ensure a com- to HPC. While Nakai et al, when studying the
plete inspection of the piece. The specimen effect of rolling conditions on hydrogen da-
presented reflections all over the area in dif- mage (HD) susceptibility, showed that a
ferent distances beneath the surface. The mor- greater deformation of MnS particles increas-
phology seems to correspond to a rolling flaw, es the HD susceptibility [12]. Pressouyre et
because of the flat reflections;however we al, observed that MnS inclusion concentrations
cannot obtain enough information about HPC increase HD susceptibility. In this work, the
from this study~ size and shape of the inclusions are consider-
ed to depend on the sulphur concentration in
steel. Nevertheless, the authors do not consi-
der that the shape, size or distribution of
the inclusions depend mostly on the conditions
of solidification, rolling and thermal treat-
ments. It is also believed that the differen-
ce in opinion could result from mistakes in
the technique chosen for characterization of
the inclusions.
The steel susceptibility to HPC in worked
steel depends on the differences in the ther-
mal contractions of the inclusions and matrix,
the elastic constants of the inclusions, and
rolling conditions, while in non worked steel
i t depends only on the differences in thermal
contractions of the inclusion and the matrix
during heat treatments. This is understood
as follows: the stress localization occurs
around inclusions for two reasons: (1) the
Fig. 5.~10 µm. SEM micrograph of clusters
difference in thermal contraction during
of Al20 3 which present HPC.
cooling and (2) the concentration of remote
applied stresses resulting from the dif feren-
ce in elastic constants between matrix and
inclusions.
Analytical and empirical expressions are
suggested in several works to show the rela-
tions of stress field around an inclusion
with the size; shape and chemical composition
of the same [13, 14]. Nevertheless, according
to literature [15] the relative contribution
of tesselated stresses versus applied stress
magnification is still somewhat unclear.
The stresses generated around large and
hard particles during cooling and work proce~
sing could exceed the yield strength. There-
fore, greater deformation is likely to occur
in the matrix around hard inclusions. As
Bernard et al reported that hydrogen solubili
ty in iron increasses with the strain level
[16], i t can be inferred that around hard in-
clusions there is a higher hydrogen concentr~
tion in solid solution. This concentration
could be in equilibrium with H2 (gaseous) in-
side microvoids in matrix-inclusions inter-
1 µm. SEM micrograph of MnS-Al203
faces or interfaces of different hardness of
which is not associated with HPC.
the same inclusion. The aforementioned micro-
voids appeared during heat treatments and or
steel working processes. If H2(gaseous) pres-
DISCUSSION AND CONCLUSIONS
sure in the microvoids exceeds the yield
strength the blister will grow.
The HPC consists of a macrovoid that is
Hydrogen concentration in solid solution
observed in Fig. 2 and studying the specimen
in the steel matrix will depend on the extent
far from the macrovoid i t was observed inter-
of deformation and on the hydrogen charging
nal voids in the interface matrix-hard in- conditions. We can thus understand the growth
clusions such as alumina in all the specimen; of blisters around hard inclusions in worked
while the soft inclusions which deform plasti steels and the non existence of blisters in
caly during the rolling of the steel such as-
those interfaces between matrix-soft MnS
MnS and silicates, are not associated with
type inclusions.
HPC, hence the susceptibility to HPC does not
depend on the number, size or shape of the
MnS inclusions.
Great differences were found however
in the literature with respect to experimen-
tal results on the effect of MnS inclusions
WRC Bulletin 532

Sour Sen:ice Environments 71
[ 16] S. Bernard, I. Talbot; Compt. Rend.

Table 1.- Type of inclusions and their Vol. 224, p. 1193, 1957.
behaviour to HPC
Type of Diameter Number per Shape Occurrence of

Inclusions range(µm) metric ton blister in
interface AKNOWLEDGEMENTS
matrix-inclu-
sion (%) The authors thank the Materials Depart-
ment, the colleagues involved in this work,
Alumina 5-10 s.10 15 Angular 60 and specially the Metallography Division, as
and complex and sphe- well as Thermal Treatments, for permission
inclusions ric al to publish this work.
Clusters 10-500 s.10 14 Clusters 100

of alumina 1.103 i.1014 Clusters 100
particles
MnS-MnO 15-20 9.1015 Elongate 0
30 3.109 Elongate 0
Complex
Inclusion 10-20 6.1015 Elongate 0
Oxide-Sul-
phide
REFERENCES
[l] De Kazinski, F.J. 1954, J. Iron Steel

Ins. 177: 85-92.
[2] Garofalo, F.; Chou, Y.T.: Ambegaokar, R.;
1070. Acta Metall. 8: 504-12.
[3] Bilby, B.A.; Hewitt, J.; 1962. Acta
Metall. 10:587-600.
[4] Tetelman, A.S.; Robertson, W.D.; 1963.
Acta Metall. 1: 415-26.
[ 5] Ciszewski, R. Proceedings Stress Corros-
ion Cracking and Hydrogen Embrittlement
of Iron Base Alloys (NACE), June 1983.
[6] Szklarska, Z.; Smialowska and Lunarska,
E.; Korrosion 32, 1981, 478-485.
[ 7] Wilde, B.; Kim, C. and Phelps, E.; Cor-
rosion 36 (1980) 625.
[8] Pressouyre, G.;Blondeau, R.; Primon,G.
and Gandiou, L.; American Society for
Metals, Metals Park OH 44073 (1982) 212-
219.
[9] Yoichi, N~; Kurahashi, H~; Pmi, T,; Hida,
O.; Transactions ISIJ 19(1979) 401-410.
[ 10 l Schiapparelli, E.; Hey, A.; Memoires et
Et Scient. Revue de Metallurgie Mai
(1984) 273-278.
[ 11] Schiapparelli, E.R. Non Destructive Tes-
ting Communications 3 (1987)pp. 39-46.
[ 12] Y. Nakai, H. Kurahashi, T. Emi, 0. Haida;
Trans. ISIJ Vol. 19, p. 401, 1971.
[ 13] D. Broosklank, K.W. Andrews, Production
and Application of Clean Steel; The
Iron and Steel Institute, London, U.K.,
p. 186, 1972.
[ 14] R. Kiesling; I. Nordberg; Applied Mech.
Vol. 1, p. 179,1933.
[15] W. Philbrook; Int. Metal Review, p. 223,
September 1977.
WRC Bulletin 532

STUDY OF THE SULFIDE STRESS CRACKING OF

LOW ALLOY AND C-Mn STEELS
H. Margot-Marette and J. C. Charbonnier

IRSID-UNIREC
78105 Saint-Germain-en-Laye, France
L Coudreuse and J. Charles

Creusot-Loire lndustrie
Materials Research Center
ABSTRACT The use of steels in such environrr:ents poses two

major problems :
A study on the metallurgical factors influence en in the absence of stresses, the hydrogen
SSC behaviour was conducted throughout a large scale induced cracking HIC : blistering, step-wise cracking
of' low alloy and C-Mn steels used for the operation or straigth-line cracking,
and conveyance of petroleum products. Stee 1 plates - in the presence of externa 1 stresses, induced
were manufactured either by ingot casting, hot rolling delayed fracture or sulphide stress cracking SSC,
and normaliz:ing or by continuous casting and The first phenomenon has already been the
controlled rolling .. In this study, SSC resistance is subjec.:t of many in.vestigati.ons and the mai_r1
compare::1 with HIC resistance and analysed. metallurgica 1 factors involved have been determined :
It is shown that : inclusions, segregations of elements as c, Mn, P
- non metallic inclusions cleanliness is of prime leading to the formation of bainitic-martens itic
importance. Particularly, the noxious effect of bands during cooling, imp.1rties (S, P), rolling
elongated inclusions (MnS) or alignments (aluminates) conditions and other factors (1-18).
is evidenced on SSC as well as on HIC resistance, The second phenomenon, by contrast, is more
- the segregations are m:>re detrimental on HIC complex and the metallurgical factors affecting the
than for SSC when externa 1 stresses are ai;p lied in the SSC behaviour are poorly known. Similarly, the
long or transverse directions, relationship between the HIC and SSC phenomena i f
- Phosphorus acts mainly on the surface reactions any, have not been clearly identified yet,
enhancing the introduction of hydrogen in the metal The aim of the present work, which was supported
and its permeation rate. Consequently the specimen is by ECCE (19) was to clarify the metallurgical parame-
susceptible to more rapidly attain a critical hydrogen ters and the mechanisms involved in the delayed
concentration, fracture of steels used for the operation and the
- work hardening contribute simultaneously to a conveyance of petroleum products : i.e. low alloy and
large hydrogen trai;ping, promoting HIC sensitivity, C-Mn steels.
and to a local strengthening reducing the SSC sensiti- The study was concerned with industrial and
vity. Thu.s a compromise must be found in order to experimental heats.
obtain a good resistance to both HIC and SSC. The iadustria 1 heats were manufactured by two
different processes :
Key words Sulphide stress cracking - Hydrogen in - :L1got casting followed by hot rolling and nor-
duced cracking Petroleum products Hydrogen malizing (low alloy and C-Mn steels) ,
sulphide environment. continuous casting followed by controlled
rolling (C-Mn linepipe steels).
INTRODUCTION The influence of the following main
metallurgical parameters inclusions (sulphides or
Hydrogen embri tt lemen t as a res ult of hydrogen oxides) 1 segregations 1 phosphorus, rolling
sulphide (H2S) hydrolysis and corrosion reactions conditions, specimen sampling direction in relation
leads to many industria 1 problems. In the petroleum to rolling direction, yield strength level, were
industry, in i:articular 1 developments in oil well examined.
drilling and operating techniques are such that the
materia 1 used are being subjected to increasingly MATERIALS
severe operating conditions : higher acidity, chloride
medium charged with H2S. Table 1 shows the chemical compositions of the
steels studied in relation to the three manufacture
classes :
WRC Bulletin 532

HEATS CLASS PROCESS c Mn Si s p Ni cu Cr Mo v Nb Ti Ceq
130 1230 255 0.3 6 160 70 100 40 0 440

INGOT I I I I I I I I I I
:;;] CASTING 190 155 0 370 14 16 500 250 285 95 100 550
H
~
E-<
UJ
::0
Q
z 50 835 225 4 0 0 0 0 0 25 0 310
H
2 CONTINUOUS I I I I I I I I I I I I I
CASTING 120 1525 465 2 19 320 360 35 195 65 50 30 360
...,
<i:
E-< 50 kg Ingot 3
z
µ:j
:;:
H VACUUM 100 1500 295 2 I 60 25 360
0::
µ:j
p., MELTING 33
>:
µ:j
Table 1 - Chemical composition average 10

-3
15 in weight Ceq c +- +
Mn Cr + Mo + v
+
cu + Ni
6 5 15
First class industrial heats - ingot products experimental heats where the P content variat:iDn was in
Second class industrial heats - continuous cas- a wider range (3 to 33 thousandth per cent in weight).
ting products Lastly, the influence of rolling conditions was
Third class experl.mental heats - 50 kg ingot conducted on linepipe steels obtained by controlled
Vacuum melting products. rolling, the finishing temperature of rolling being
The influence of sulphur content was studied only either in the two-phased field (a+o) or in the auste-
with the first class steels, whereas the second class nitic field ( o).
steels were low sulphur linepipe steels treated by Table 2 collects the variation of mechanical
silicocalcium~ properties of the different kinds of steels.
The influence of segregations was studied with rt will be noted that the thickness of the first
specimens machined in the core or in the skin of the class steel plates varied between 20 and 100 mm, whe-
heavy plates or the slabs. reas those of the linepipe steel plates was 20-24 mm.
The influence of phosphorus content was examined Microstructures were usually ferritic perlitic
for all of the steels and more particularly with banded structures.
Long Direction Transverse Direction

HEATS CLASS PROCESS YS UTS YS UTS Plate
MP a MP a MP a MP a Thickness
HOT 400 550 400 550 20

ROLLING I I I I I
...,
<i: NORMALIZING 590 825 580 825 100
H
0::
E-<
UJ
::0
Q
z CONTROLLED 400 490 410 510 20
H
2 ROLLING I I I I I
550 650 600 670 24
CONTROLLED
ROLLING 460 565 470 570 14 mm
Table 2 - Mechanical Properties average.
WRC Bulletin 532

EXPERIMENTAL TECHNIQUES
SSC behaviour was systernaticaly investigated by

the constant load test in accordance with the NACE
standard TM 01-77 1 the non failure threshold stress
being evaluated after 720 hours. Testing machines were
CORTEST (TM) proof rings,
In addition, for some cases, the slow strain rate
technique was also used in the same enviion'llent (20).
Quantitative estimation of inclusions coHcentra- (.f)
- - fang and transverse direction
tion were carried out. The inclusions were idenfied
according to the norm NF A 04-106 : type A alignments
>
"0.8 ----short direction
U1
(sulphidesj 1 type B alignments (aluminateo) type [; U1
isolated particles (globular oxides). The total ~ 0.6
projecte:i lengths of the A and B alignments were U1
measured by Texture Analysis Surface Microscopy. "O
0 0.4
.c
RESULTS AND DISCUSSION U1
OJ
L. 0.2
Influence of inclusions alignments .c
....,
0
Sulphides. It has been demonstrated by many wor- 0 5
kers (21 to 24) that sulphide inclusions morphology
10
plays an important role in relation to the in sernice
properties and particularly in HIC behaviour / the ali- Fig. 1. Influence of the sulphur content on the
gnments being the more detrimental feature (1 1 2 1 4 1 SSC behaviour of normalized low a.Lloy steel plates.
7, 8, 12, 13).
Figure 1 illustrates the noxious effect of the
sulphur content and Figure 2 those of the sulphide
alignments on the SSC resistance, with respect to the
sampling direction. The most detrimental effect is
obviously obtained for the short transverse direction. (f) --long and transverse direction
Oxides. The decrease of the sulphur content down
to 20 pp;;;7nd the shape control of inclusions by Si-Ca
> 0.8
'-... ----short direction
U1
additions 1 prevents the presence of MnS elongated in- U1
clusions. Nevertheless, inclusions are still to be e o.6
taken into account, mainly inclusions of type B ti
(aluminates) which ai;:pear in alignments. :g 0.4
Their detrimental effect on HIC resistance has 0
.c
been clearly demonstrated in a previous work ( 18).
Figure 3 shows that 1 the lower the non rupture thrcs- e 0.2
U1
no ld 1 the higher the projected length of alignments. :5

This effect is ma inly seen for the transverse O'--~~.=...;;....<;;......;.,_.;_;~-.L~~~~...i._~~~_J
direction. Obviously it will be even more detrimental 0 50 100 150 200
in the short transverse direction,
projected length of sulphides ( JJ m/m m 2 )
It will be noticed however, that the detrimental
effect of inclusions is more severe with MnS elongated
inclusions than with aluminates inclusions alignments. Fig. 2. Influence of the projected length of sul-
It is then clear that the size and thickness of phides on the SSC behaviour of low alloy steels
MnS inclusions as well as their distribution in the plates.
matrix play an important role in the loss of cohesion
at the interface defect/matrix and cm the hydrogen
trapping mechanism ( 11). If only the presence of
oxides is taken into account, others metallurgical
parameters can drastically influence the SSC behaviour ui
and masked their detrimental effect.
>
"0.8
U1
Ill
Segregations ~ 0.6
U1
Many studies have already described the influence
of primary segregation of elements like P and Mn ( 2, ~ 0.4 ' '•
7 1 14 1 16 1 25 - 27) and secondary like C (14, lS, 2'7)
.c transverse direction
U1
on the formation of hard bands and on the HIC crack ~ 0.2
propagation, :5
Although the high susceptibility of the hard
bands to hydrogen embrittlement was clearly demonstra-
ted ( 18) / the constant load method did not allowed to
distinguish the behaviour of long or transverse ten- projected length of type 8 alignments
sile specimens machined in the core (segregated area)
or in the skin (no segregation) of normalized steel
plates or of controlled rolled steel plates coming Fig. 3. Influence of the projected length of
from either the core or the skin of the slabs. (Cf. tyoe B alignments on the SSC behaviour of line-
figures 4 and 5). pipe steels plates.
WRC Bulletin 532

A large discrepancy is only observed in the

fractc<re times and in the thresholds. Obviously it is
not the case in the short direction as disC'..:ss2d later
on. 1.0 . - - - - - - - - - - - - - - - - - - - .
ui
>' 0.8-
'--..
lll
00
0 0
.
00
0 00 • •
lll
~ 0.6-
0 •
..
• 0 0 0 0
lll
0 • o.
J, 0 i
ui ~ 0.4 -
!1j .t::
~ 0.8 - ~ long direction
lll
~ 0.2-
.t::
o long direction
...,~ 0.6- ..... • transverse direction
lll ~ 0 ~~~~~-~l~~~~-'--'~~~~-"'--'~~~~...J
~ 0.4 - transverse direction 0 5 10 15 P(10- 3 % )
.t::
lll
~ 0.2 - Fig. 6. Evolution of SSC threshold VS P content for low
...,
.t:: sulphur steel (S< 2-10-3%). Normalized low alloy steel
plates.
' - r - - - - 1 '__,,___.;
skin v2 thickness
Fig. 4. Influence of the specimen position on ;:;sc
ui 1.0 .-----<>----------------.
threshold. Normalized low-alloy steel pli'ites. >'
';;;- 0.8~
lll
steels Q)
ti 0.6 ~
:g
0.9 _g 0.4 -
lll
ui o.s1--~--~~~-'\',/~/,,,'./'.:"6~~-1-~~~~~~-.j Q)
.... 0.2 - o long direction
>
......... 0.7
.t::
..... • transverse direction
Ill
~"' 0.6
0 I
0 5 10
Fig. 7. Evolution of SSC threshold VS P content. Lows

linepipe steels.
0.3
10 100 1000 et~~·)I t. !experimental heats

time to failure (hours)
gla,-:n I
--1-- -~: -
I I
Fig. 5. Influence of segregation on SSC resistance.
Linepipe steels. Long direction.
Phosphorus
2.0 ~
Phosphorus is an element which highly segregates
during solidification and to which an important effect
I I'{ I
was attributed in the formi'ition of hard bands 1 because

of its hardenability. ,, 1 ··· r~
The interference of metallurgical parameters
-1---- t------
r
generally masks the influence of P content in the low
concentration range (50 to 170 i;:pm) as we can see in 1.0~· I
Figures 6 and 7. It was then not i;ossible co clearly
·~
demonstrate its effect 01 industrial heats.
Nevertheless / the importance of reducing as low
as possible the residua 1 con tent in P was demonstrated
with experimenta 1 heats on HIC ( 18) as we 11 as on SSC
0.Sr----------+-I I
(Figure 8) testing with the slow strain rate
technique.
00'.----'--------'l_ ___J_~
Furthermore, hydrogen permeation measurements
conducted in acidic medium (28) showed that fhosphorus
plays a main role on the surface reactions, promoting 10 20 30
3
the hydrogen pick up and increasing its permeation ra- 10- % P content
te through the metal. It is then possible 1 in the ma-
teria 1 to attain more rapidly a critical hydrogen con-
centration leading to cracking in brittle lo"ations. Fig. 8. Influence of P content on SSC susceptibility.
Slow strain rate technique. L direction.
WRC Bulletin 532

Rolling conditions
It has bPen displayed in the above- mentioned

former work ( 18) that HIC resistance was lower when
controlled rolling ended in the two phases field a +~
than when it finished in the ii field.
On the contrary 1 we observe a higher resistance
to SSC for the specimen machined in plate rolled in 1.0 r-----------,,---------.
the a +1' field than in 6 field (Figure 9).
When the finishing temperature of rolling decrea- ui
ses, larger and larger volumic fractions of ferrite > O.B 1-
are work hardened, causing an increase in the disloca- ~

tions density. This favours the hydrogen trapping in ~ 0.6-
1... 0 0
the metal and therefore, critical hydrogen concentra- ......
Ul
tion is more easily reached. "O 0.4 l-
In the presence of external stress, once the mi- o.r::.
crodefect initiated, the propagatiun depends on the 0 long direction
surrounding state of stress and on the material ~ 0.2 t-
plasticity. £ • transverse direction
As observed by Iino (29), the increasing of work o~~~~~_._~~~~-"'--1~~~~-'--'~~~--l
hardening hinders the SSC propagation mechanism. 400 500 600 Y.S.(MPa)
Moreover, the decreasing grain size with the "hard"
controlled rolling certainly promotes the SSC
resistance. Fig. 10. Correlation between SSC threshold and YS
for linepipe steels.
0,9
0,8
0, 7
0,6
l Ul
> OB1-
'-,_
Ul
.
~ 0.6~
1.0,.-----------------
.o
••
,0 ••
..
0
oeO
ti Oo. 00
IJNF 0,5
vs 0,4
ill8F>AR3
li!J6F<AR3
:g 0.4-
0
.r::.
~~·i;oo
..
~ 0.2~
0,3
long direction
0
0,2
£ • transverse direction
0, 1 0 ....__ _ ___._1_ _ _ _...l,__ _ _ _L __ _ __J
400 500 600 Y.S.(MPa)
Fig. 9. Influence of Finishing Temperature of Rolling

on SSC behaviour. Fig. 11. Correlation between SSC threshold and YS
For normalized plates.
Yield strength level
The above-mentioned mechanisms are in good H.1.C• a

.........
accordance with the observations made on linepipe
steels : in the transverse dirPction 1 the highe!: the
yield strength, the higher the resistance to SSC
throughout the range 400-600 MPa (Figure 10).
On the contrary, for the normalized plates, no
correlation appears between the SSC thresholds and the
yj.eld strengths (Figure 11).
This fact shows the great influence of the work
hardening on the SSC resistance mechanism.
HIC and SSC relationship
The relation between HIC and SSC is complex.

Generally / no relationship exists between HIC and SSC
resistance when the external stresses are ag::ilied in
the long or transverse direction.
It is possible to observe a high SSC threshold
stress close to the yield strength of the material
even if the HIC susceptibility is high (Figure 12). In
the same way 1 a great HIC resistance does not lead
systematically to a great SSC resistance (Figure 13). Fig. 12. HIC sensitive sample, non fractured in
720 hours.
WRC Bulletin 532

Segregations must be reduced at the maximal

a
........
extent to obtain a large HIC resistance, although they
do not seem to have a determining role when the
external stresses are awlied in the long or transverse
direction.
- The detrimental effect of phosphorus 1 though
well present, is generally masked by others metallurgi-
ca 1 parameters in industrial heats. Nevertheless, it
seems to act on the surface reactions and to promote
the pick up of hydrogen in metals. It is therefore
advised to limit its content.
- Lastly, the controlled rolling conditions must
be adjusted in order to obtain a good HIC resistance as
well as a good SSC resistance.
BIBLIOGRAPHIC REFERENCES
1. Moore, E.M., Warga, J.J., "Factors influencing

the Hydrogen Cracking Sensitivity of Pipe line Steels"
Materials Performance June 1976, pp. 17-23.
2. Miyoshi, E., Tanaka, T,, Terasaki, F., Ikeda,
A., "Hydrogen Induced Cracking of Steels under wet H2S
Environment" 1 Trans .• ASME, Journal of Engineering for
Fig. 13. HIC resistant sample, fractured by SSC in Industry, Vol. 98, Nov. 1976, pp, 1221-1230.
25 hours. 3. Ikeda, A., Morita, Y,, Terasaki, F., Takeyama. 1
M. 1 "On the Hydrogen Induced Cracking of linepipe
In the short transverse direction, a straight Steels under wet H2S Environment" Proceedinqs of the
relation may appear between the two phenomena for the Intern. Congress of Hydrogen in Metals, PARIS, June
HIC sensitive steel, because in this case, the 1977, Paper 4 A 7.
extema 1 stress aw lied contribute to a lamelling 4. Nishimura, T., Inagaki, H., Tanimura, M., "Hy-
tea ring mechanism. drogen Cracking in Sour Gas Pipeline Steels" See ref.
It is not suprising, from this fact, that the 3, Paper 3 E 9.
metallurgical parameters seem not always to affect in 5. Iino, M., "The Extension of Hydrogen Blister-
the same way the HIC or SSC behaviour of the mate- Crack Array in Line pipe Steels". Metallurgica 1 Trans.
rials. It depends on the initiation sites and on the !ii_ Vol 9A, Nov. 1978, i;p. 1581-1590.
surrounding stress field, as it has been developed w "Influence of Sulphur Content on the Hydrogen Induced
prior works ( 30 1 31) 1 by a mechanica 1 analysis of HIC Fracture in Linepipe Steels" Metallurgical Trans. A,
and SSC cracking whose conclusions may be summarized Vol. 10A 1 Nov. 1979, i;p. 1691-1698.
as follows : 6. Iino, M., Nomura, N, 1 Takezawa 1 H,, Gondoh, H.,
- Alignments of inclusions (sulphides or oxides) "Linepipe Steels Resistant to Hydrogen Blistering and
act as well on HIC as SSC susceptibili<:y because they Cracking". Revue de Meta llurgie, Vo 1. 76, N° 8-9, 1980,
are the most efficient initiation sites for fractures pp. 592-609.
and impose the direction of the initial defect. 7. Nakai, Y., Kurahashi, H., Emi, T,, Haida, o.,
Segregations leading to the hard bands "Development of Steels Resistant to Hydrogen Induced
formation act differently because when they are Cracking in Wet H2S Environment", Transactions ISIJ,
initiation sites, it is the surrounding stress field Vol. 19, 1979, pp. 401-410.
which governs the crack propagation. 8. Hersleb, G., Poepperling, R.K., Schwenck, W.,
- Phosphorus ma inly acting on surface reactions "Occurence and Prevention of Hydrogen Induced Stepwise
for these steels, its deliterious effect is probably Cracking and Stress Cerros ion Cracking of low-alloy
the same on HIC or SSC. Pipeline Steels". Corrosion-NACE, Vol. 36, N° 5 1 May
1981, PP. 247-256,
CONCLUSIONS 9. Taira, T., Tsukada, M., Kobayashi, Y., Inagaki,
H., Watanabe, T., "Sulfide Corrosion Cracking of Line-
This study was conducted on a large scale of pipe for Sour Gas Service", Corrosion-NACE, Vol., 37 1 N°
industrial low alloy and C-Mn steels 1 manufactured by 1, Jan. 1981 1 pp. 5-15.
two different processes : 10. Ikeda, A,, Nakanishi, M., Kaneko, T., Sumito-
'1,mprvvement uf ·Bych.~·gen Sulf i-
11
- i.rigot casting followed by hot rolling and nor- mo / Y .. 1 Takeyama 1 M .. '
malizing, de Cracking Susceptibility in Linepipes for Soor Gas
- continuous casting followed by controlled ro 1- Services". The Sumitomo Search, N° 26, Nov. 1981 1 w
ling. 91-103.
It was aimed to clarify ..:he influence of metal- 11. Pressouyre, G.M., "Application of the Traps
lurgical parameters on SSC and to contribute to a best Threory of Hydrogen Embrittlement to Hydrogen Induced
understanding of the HIC and SSC mechanisms. Results Cracking in H2S Environment", Proceedings of the 3d.
allow to point out the main parameters in botr, cases. Intern. Congress on Hydrogen and Materials, PARIS, June
They suggest wha.t could be done to enhance the 1982, Paper E01.
resistance of materials used for the operation. and the 12. Pressouyre, G,M., Blondeau, R,, Primon, G.,
conveyance of petro le urn products. Cadiou, L,, "Very Low Inclusion and Impurity Content
- In particular, of prime importance is the non Steels as a Solution to resist Sour Environments" 1
metallic inclusions cleanliness. Not only must be Proceedings of the Ist. Intern. Conf, "Current
considered the elongated MnS but also the oxides Solutions to Hydrogen Problems in Steels", ASM,
alignments. Washington, Nov. 1982, i;p. 212-218.
WRC Bulletin 532

13. Brown,A,, Jones, L., "Hydrogen Induced Cra- 23. Bernard, G,, Hersant 1 T. 1 Moliexe, F., Moussy 1
cking in Pii;:e line Steels" Cerros ion-NAr:E, Vol. 40 1 N° F., "Mechanisms and Quantitative Description of Ductile
7, July 1984 1 r:p. 330-336. Fracture in Relation to the Elongated Sulphid!ls", Mem.
14. Taira, T., Kobayashi, Y., Matsumoto, K., Mat- Sci. Revue de Metallurgie, Vol. 76 1 N° 11, Nov. 1979,
sumoto, s., Terunuma, T., Arikata, K, 1 "Resistance of i;:p. 667-687.
Pipeline Steels to Wet Sour Gas", See ref. 1.2, .i;p. 24. Moussy, F., Roesch, L. / "Infuence of Inclu-
173-180. sions Distribution and Moq:hology on Plastic Deforma-
15. Ohtani, M. Mashimoto, T., Komizo, Y. 1 tion and Fracture Behaviour of Steels", Fina 1 Report N°
Murayama, J, / Sawamura, T., Bessyo, K., Kyogoku, T. 1 7210 MA/ 307 / Commission of the European Communities
"Developnent of low PCM High Grade Linepii;:e for Artie 1981.
Service and Sour Environment" Proceedings of the 25. Kohn / A. / "Dendritic Segregation and Banded
Intern. Conf. on Technology and Awlications of High Structure" 1 Revue de Metallurgie. Jan. 1970 1 pp. 23-26.
Strength Low Alloy Steels I ASM-PHILADELPHIA, 3-6 oct. 26. Taira, T., Kobayashi, Y., Matsumoto, K., Tsu-
1983 1 Paper 8306-057. kada, K,, "Resistance of Linepipe Steels to Wet Sour
16. Matsumoto, K., Kobayashi, Y., Ume, K., Muraka- Gas", Corrosion-NACE, Vol. 40 1 N° 9, Sept. 1984 1 pp.
mi, K., Taira, T. 1 Arikata, K., "Hydrogen Induced Cra- 478-486.
cking Susceptibility of High Strength Linepii;:e Steels" 27. Haida 1 o., Kitaoka 1 N,, Habu, H. 1 Kakihara,
Corrosion-NACE, Vol. 42 1 N° 6 1 June 1986, pp. 337-345. s., Bada 1 H. 1 Shiraishi, s., "Macro and Semi-macrosco-
17. Tamehiro, H., Takeda, T., Matsuda, s., pic Features of the Centerline Segregation in CC Slabs
Yamamoto, K., Okumura 1 N. 1 "Effect of Accelerated and their Effect on Product Quality" 1 Transactions
Cooling after Controlled Rolling on the Hydrogen f§l.-I. 1 Vol. 24, 1984 1 pp. 891-898,
Induced Cracking Resistance of Linepipe Steel" 28. Margot-Marette, H, 1 Jessie, T,, Caplet, J.L.,
Transactions ISIJ 1 Vo 1. 2 5 1 1985 1 fP. 98 2-988. Charbonnier, J.C., Brass, A.M. 1 "Application of Electro
18. Margot-Marette, H,, Charbonnier, J.C. Bridoux, chemical Permeation Technique to Hydrogen Diffusion and
D. 1 Perdrix, C. 1 "Influence of Metallurgical Parameters Trapping in Steels, Correlation With Hydrogen Induced
on HIC of Linepipe Steels" Revue de Metallurgie - CIT, Cracking Resistance" 1 Procredings of the Fourth Intern.
Dec. 1986, pp. 891-899. Congress Hydrogen and Materials 1 Beijing, CHINA, 9-13 1
19. Final Research N° 7210 KB/306. Commission of May 1988, Paper N° 37.
the European Communities. Jan. 1987. 29. Iino, M. / Nomura, N. / "Influence of Tensile
20. Margot-Marette, H., Bardou, G., Charbonnier, Properties and Metallurgical Structure on HIC Resistan-
J.C., "1'he Application of the Slow Strain Rate Test ce of Linepipe under Stress" / Nov. 85 1 Private Commu-
Method for the Development of Linepii;:e Steels Res is- nication.
tant to Sulphide Stress Cracking, Corrosion Science, 30. Charbonnier, J. c. / Margot-Marette, H. / Moussy,
Vol. 27 1 N° i0-11 1 1987 1 pp. 1009-1026, F., "Influence of Metallurgical Parameters on HIC and
21. Sims, C.E. 1 "The non Metallic Constituent of SSC Behaviour of Linepipe Steels", Corrosion 87 NACE,
Steel", Transactions of the Metall. Soc. AIME, Vol. SAN FRANCISCO 9-13 March 1987, Paper N° 293.
215 1 June 1959, pp. 367-393. 31. Moussy, F, 1 Charbonnier, J.C., Margot-Marette,
22. Baker, T.J. Charles, J.A., "Type II Manganese H. / Determination of the Stress TE11sor Component Gover-
Sulphides / their deformation and effect on Steel ning Fracture in Hydrogen Charged Pipe Steels" / See
Fracture", Journal of the Iron and Steel Institute, Ref. 28, Paper N° 83.
March 1973, pp. 187-192.
WRC Bulletin 532

WRC Bulletin 532

INFLUENCE OF MICROSTRUCTURE ON
DIFFUSION AND TRAPPING OF HYDROGEN
IN LOW CARBON STEEL HEAT AFFECTED ZONES
E. Bollinger
CRDM-SOLLAC
5938i Dunkerque, France
C. Duret-Thual and P. Manolatos

Centre de Recherche Commun de la Siderurgie Groupe B
42702 Firminy, France
ABSTRACT In an attempt to clarify the role played by the different types of

microstructures present in the HAZ, the following approach has been
The influence of metallurgical factors such as microstructure, considered :
residual stresses, ... , on the behaviour of materials in environments - realization of large simulated HAZ. A special care has been taken to
leading to hydrogen embrittlement is well known. -pie presence of produce specimens showing very homogeneous structures,. in
graded microstructures such as those encount~red m hea~ ~f~ected order to study only microstructure variations. A Gleeble machme
zones (HAZ) may then modify the material suscept1b1hty to has been used for this purpose
hydrogen induced cracking. In an attempt to clarify the role played - comparison of hydrogen diffusion, hydrogen induced degradation
by the different microstructures present in HAZ, large simulated of the different structures. Electrochemical permeation tests have
HAZ have been produced using a Gleeble machine to be studied by then been carried out.
the hydrogen electrochemical permeation technique. The perm~ation
and outgassing rates are clearly related to the volume fraction of 2 - OBTENTION OF SPECIMENS SIMULATING HAZ
martensite. In ferrite-pearlite structures, the influence of small AlN
and NbC-type precipitates on permeation has also been shown. Welding cycles are mainly characterized by two parameters:
- the maximum temperature, which is reached after a heat impulse of
three seconds, arbitrarily settled at l 350°C for usual test pieces.
1 - INTRODUCTION This allows to get the coarse grain structure which is situated just
under the fusion line of real welding
The effects of hydrogen on materials, especially low alloy carbon - the cooling down rate, defined by the necessary time for
steels used in hydrogen sulphide containing media (exploitation, temperature falling from 973 to 573 K, according to a law about
storage and transport of gas, ... ) are well known and have been flowing out of heat, such as that proposed by Rosenthal (6).
clearly identified ( 1).
For example, in the case of plates used for manufacturing large
The main degradations observed are hydrogen induced crac~ng welded tubes, one can find that the cooling down parameters at
(cracks and blisher formation in the absence of external apphed 973/573 respectively equal to 100 and 30 seconds duly represent the
stress) and sulphide stress cracking (rupture with external applied processes of longitudinal (SAW) and end to end welding (SMAW).
stress lower than the yield strength).
2.1 Tested material
Various test methods have been developped to evaluate roughly if To be sure to isolate the part played by the microstructure, it was
a material is prone to hydrogen embrittlement in H2S-containing necessary to use a quite homogeneous product (chemical
media. Among them, the most commonly used in gas and petroleum composition and inclusions distribution). As the actual qualities of
industry are NACE TM-02-84 and TM 01-77 standard~ related steels for pipes with low Ceq are very little quench hardenable
respectively to hydrogen induced cracking (HIC) and sulphide stress (fig. 1), in order to get a wide range of structures and to make sure
cracking (SSC). The tests as specified in these standards allow to of the sensitivity of the method, we used a steel A 516 Gr. 70 rolled
compare the behaviour of different materials and to eliminate in some laminated into a 80 mm thick plate. The chemical analysis is given
cases those not convenient for a use in these environments. below:
The number of tests carried out on low alloy carbon steels has Element C S P Si Mn Ni Cr Cu N Al Nb
made it possible to show the infl::ience of metallurgical p~ameters 10-3 % 146 1.5 15 372 1130 211 23 20 7 38 28
such as composition, heterogeneity of structure (segregations, ... ),
level and nature of inclusions (2-5).
The structure is made of ferrite and well-distributed thin bands of
However it does not permit to make predictions, in particular pearlite. The amount of inclusions has been evaluated elsewhere (7) :
when graded microstructures are present as it is the case in heat it is rather low (table 1). Finally, the behaviour in HIC TM 02-84
affected zones (HAZ). and SSC TM 01-77 tests, is comparatively good (table 2).
WRC Bulletin 532

2.2 Sampling and specimen design Two samples treated in a different manner show clearly distin-
The cell used for permeation tests implies the production of disk- guishable behaviours : the chosen conditions for permeation are
shaped test pieces of 20 mm diameter and approximately 1 mm thick. appropriate to make difference within the metallurgical structures
In order to set in action a sufficient volume of material having a studied.
homogeneous structure and apt to be quickly cooled down, it has
been necessary to find out a specially designed test-piece (fig. 2). A slight scattering in the permeation currents (maximum current
0.8 to 0.9 µNcm2) can be noticed. It is usually attributed to small
The evolution of structure and hardness thus obtained is shown variations at the hydrogen entry side (12). Clearly the interpretation
on figure 3 as a function of the cooling parameter 6.t 973/573. of the results cannot be affected by these small changes.
The different investigated conditions are given in table 3. 3.2. Influence of cooling rate
It comes out from figure 7, that the higher is the quenching rate,
Micro graphic examinations and micro hardness profiles (fig. 4) the larger is the permeation breakthrough-time (figure 7a) and the
have proved the homogeneity of the useful part of the specimens. larger is the amount of hydrogen desorbed after interruption of
Disks intended to permeation tests (0 20 mm) have been cut by charging (figure 7b).
electroerosion in the center of this zone.
For the highest cooling rate which lead to mixed bainitic-
The evolution of hardness and structure with the cooling martensitic structures, it seems that the permeation kinetics and the
parameter is shown on fig. 3 : graded micr~structures have ?een amount of hydrogen desorbed during outgassing are related to the
obtained ranging from almost 100 % martenslte to 100 % femte + fraction of martensite. The trapping effect associated to martensite
pearlite. thus appears to be much more important than that of iron carbides
particles which come with the formation of bainite. It is also shown
3 - PERMEATION TESTS that the trapping capacity of the ferrito-perlitic structure (6.t 973/573
= 500 sec) is much lower.
The permeation cell used in this study has been developped o.n the
basis of the method proposed by Devanathan and Stachurski (8). 3.3. Influence of a heat treatment
The experimental condi~ions are ~ummarized !n tal;>le 4 ; ~hey are In figure 8, the behaviours of three ferrite-pearlite structures are
typical of many permeat10n expenments descnbed m the literature compared: as produced (6.t 973/573 = 500 sec), heat treated 30 min
(9). and 4 hours at 873 K. It shows that :
- the thermal treatment for 30 minutes at 873 K does not alter in a
detectable manner the permeation characteristics : the breakthrough
A preliminary study of the method (10) has shown us that, in time and the transient for permeation or outgassing are very similar
these experimental conditions (see table 4), the passive layer grown - holding for 4 hours at 873 K induces a noticeable reduction of
on the detection side plays the role of barrier to hydrogen permeation permeation rate. Moreover, the quantity of hydrogen liable to be
i.e. it interferes with the exit of hydrogen from the metallic sample. trapped by the structure is more important than that of the previous
Thus the hydrogen content at the exit side cannot be considered to be specimens.
zero as it is supposed in many diffusion models presented in the
literature (11). Moreover the evolution of the passive layer with time Using electron transmission microscopy it has been possible to
may induce variation of this hydrogen content which may affect attribute this effect to the precipitation of very small particles (a few
considerably the permeation curve (for example, breakthrough time). nanometers in diameter) containing mainly Al and Nb (probably AlN
In this case, the exploitation of the curves according to the diffusion and NbC type precipitates) (figure 9).
models usually presented in the literature to obtain quantitative data
(diffusion coefficient, amount of trapping, ... ) cannot be used to The additional trapping provided by these particles is more
produce reliable data. important than that corresponding to the decrease in the amount of
dislocations accompanying the heat treatment.
' Furthermore, the comparison of materials which form different
passive layers should not be done by this method without assessing 4 - EVALUATION OF RESISTANCE TO HYDROGEN EMBRIT-
that these passive layers show identical behaviours to hydrogen 1LEMENT
permeation.
In order to complete these results and roughly estimate the
However, in this study, using optimized experimental conditions resistance of the different specimens to hydrogen induced cracking,
and by working on samples of the same composition but with further tests have been carried out with altering the charging
va."ious microstrnctures, it has been possible to obtain from the conditions (charging solution H2S04 0,1 N + 15 mg/1 NaAs02,
permeation curves informations on hydrogen diffusion truly related cathodic current density 5 mA/cm2 applied during 15 hours). These
to the specimen microstructure. conditions lead to significant absorption of hydrogen and the injected
quantities may cause metal cracking which appears externally by
We have compared, for different microstructures, the permeation blistering. The experimental conditions at the detection side remain
and outgassing curves after interrupting the cathodic charging. The the same as those cited before.
passivation time before charging on one hand and the permeation
duration before outgassing on the other hand have been strictly In these conditions, it has been observed that an increased in the
controlled and are the same for all the tests. At least two permeation fraction of martensite induced an increased degradation of the metal
tests have been carried out in each case (see table 3). on the charging face. On the other hand, it has not been possible to
differentiate the behaviour of specimens which do not contain
3 .1. Results martensite as they did not exhibit any cracks in these conditions.
At first, we have checked the sensitivity of the permeation test to
detect different behaviours from different microstructures ; we have
also controlled that the scattering introduced by the way of producing The data produced by these last tests are only qualitative.
the specimens was limited. Figure 6 shows permeation (a) and out Although the same type of experimental conditions have been used to
gassing (b) curves of three samples, two of them having received the damage specimens and calculate from permeation results critical
same thermal cycle on the Gleeble machine. Samples treated in the concentrations of hydrogen (13) (i.e. the concentration required to
same way show close characteristics with regard to permeation : the crack initiation), we have seen before that the hypothesis of zero H
possibility of repeating the test (simulation of structure, dispersion concentration at the exit face does not apply in these conditions. In
connected with electrochemical process) appears satisfactory. the particular case of high hydrogen fluxes at the entry face, at least
two problems are encountered, taking place at the exit face. The first
WRC Bulletin 532

one (presence and evolution with time o~ the passive layer) has h<?en REFERENCES
identified and its incidence on permeation results has been st~d1e;d
especially in the case of low fluxes (10) ..Th~ second.one which is (1) La fragilisation des aciers par l'hydrogene - S. Talbot-Besnard
specific to high fluxes is due to the recombmatton reaction (H + H -> Mem. Sc. Rev. Met. - Mai 1979 - p. 293-307
H2) at the exit face : gas bubbles have been observed to evolve from
the detection side. In this case the permeation current does not (2) A. Brown, C.L. Jones
provide a measure directly related to the hydrogen flux going
Corrosion 83 - Paper n° 155
through the metal since an important part of hydrogen does not
participate to the oxidation reaction (H -> H+ + e-). For ~~ese (3) J.C. Charbonnier, H. Margot-Marette, F. Moussy, D. Bridoux,
reasons the curves obtained cannot be used to evaluate cntical C. Perdrix
hydrogen concentrations Corrosion 87 - Paper n° 293
5 - DISCUSSION AND CONCLUSION (4) J. Charles, V. Lemoine, G.M. Pressouyre, L. Cadiou
The different tests carried out have made obvious some quite
different permeation behaviours relat~d to ~e app~ied thermal cycles (5) B.E. Wilde, C.D. Kim, E.H. Phelps
and to the rnicrostructures of th.e specrmens mvesttgated. Corrosion NACE - Vol. 36, n° 11 - 1980 - p. 625-632
For example, the effect of martensite both on the quantity of (6) D. Rosenthal
hydrogen trapped and on the sensitivity to hydrogen induced Weld. J. - Vol. 20, n° 5 - 1941 - p. 220-234
cracking has been shown.
(7) D. Zmudzinski
The influence of the precipitation of AlN and NbC type particles Creusot Loire Industrie - RT n° 87066 C - Juin 1987
on the amount of trapped hydrogen has also been clearly seen.
However it has not been possible to detect how this precipitation (8) M. Devanathan, Z. Stachurski
could affect the resistance to hydrogen cracking in our experimental Proc. Roy. Soc. London - A 270 - 1962 - p. 90
conditions.
(9) A. Chavane - M. Abashi - G.M. Pressouyre - J. Galland
With regards to the behaviour of HAZ in situations inv?lving Corrosion NACE - Vol. 42, n° 1 Jan. 1986 - p. 54-61
hydrogen embrittlement, it can be deduced from these expenments
that: (10) P. Manolatos, J. Le Coze, C. Duret-Thual
- low grade energies of welding, by helping the formation. of C.R. Acad. Sci. Paris, t. 306, Serie II - 1988 - p. 775-780
structures rich in martensite are potentially dangerous : that is a
well known phenomenon in cold-crakcing (14) (11) N. Boes, H. Ziichner
- when a post-weld heat treatment has to be applied, it is necessary to J. of the Less Com. Met. - n° 49 - 1976- p. 223-240
study carefully the evolution of trapping involved by secondary
precipiiations. (12) A.M. Brass
These de Docteur es Sciences Physiques - Universite de Paris
Sud - Centre d'Orsay - 1983
(13) J. Charles - L. Coudreuse - G.M. Pressouyre - L. Cadiou

(14) K. Easterling
Introduction to the Physical Metallurgy of Welding - p. 180 -
Butterworths - 1983
Table 1
Inclusion content according to ASTM and SEP standards
Heat Method A MethodD

reference A B C D A B C D
46 823 TC 0.0 0.5 0.0 ' "

l.V 0.0 " '
v.1 0.0 15.9
46 823 TP 0.0 0.5 0.0 1.0 0.0 0.6 0.0 13.2
SEP 1570m
Heat MethodK MethodM

reference Ko Kl K2 K3 K4 SS OA OS 00
OA 00 OA 00 OA 00 OA 00 OA 00
46 823 TC 2 40 0 25 0 10 0 0 0 0 0.5 2.0

46 823 TP 7 55 0 32 0 4 0 0 0 0 0.5 2.0
WRC Bulletin 532

Table2
Results of HIC and SSC tests on the base metal
(normalisation at 1183 K +annealing for 2 h at 853 K)
Test solution (NACE TM 01-77 solution):

- 0.5 % acetic acid
- 50 g/l NaCl H2S saturated
- pH = 3 (:£ 4.5 at the end of the test)
HIC test (average of 3 specimens) : CLR = 4.8 %

CTR=2%
CSR =0.24 %
SSC test (4 specimens) : a 720 h = 100 % Y.S. = 320 MPa
Table 3
Gleeble samples for permeation tests
Ref. Trnax ti t973 Post Structure*

573 (sec) treatment
1573 K 10 martensite
2 1573 K 20 60 % martensite
40 % bainite
3 1573 K 50 40 % martensite
60 % bainite
4 1573 K 500 ferrite + pearlite
5 1573 K 500 30 min at 873 K ferrite + pearlite
6 1573 K 500 4 hat 873 K ferrite + pearlite
* semi-quantitative evaluation from metallographical examinations
Table4
Experimental conditions for permeation tests
- Preparation of specimens : mechanical polishing under water with SiC grit paper to 1200 grade -
degreasing with acetone
- Cathodic and anodic solutions : NaOH 0.1 N deaerated with stirred argon 24 hours before testing and all
along the test
- Cathodic charging current density : l mNcm2
Imposed potential at the detection side : + 390 m V/N.H.E.
Temperature: 24.7 ± 0.2°C (297.7 ± 0.2 K)
- Passivation time before cathodic charging: 16 ± 0.5 h
- Permeation time before interruption of cathodic charging : 7 .5 ± 0.5 h
WRC Bulletin 532

- HV5
350
10-3 % c Mn p s Si Al Cu Nb v Ti N
30
250
••• •• A 41 1238 14 4 250 33 273 40 48 7 8
200
• ~
......, Ii c
B
B
c
37
52
1397 10
1152 14 2
330 44
340 39
277 48
271 43
69
63 11
7
6
"A
150
100
1 10 100 1000
M 973/573 (sec.)
Figure 1 Hardness of the HAZ in relation with the thermal cycle
Lqi25 mm
$ 20 mm
() (]) ()
Figure 2 : Gleeble specimen designed for permeation tests
Sampling conditions
WRC Bulletin 532

a b
HVS
450
400 ••• a
350
••
~ 1.....--..l
b 20 l'-'11
20 pm 300
250
• ••
c
200
150
• ••
d
L\ t 973/573 (sec)
100
1 10 100 1000
c d
Figure 3: change in microstructure and hardness related to the cooling rate
6mm
HV0,5
---"-L____J--
c••• 'llo-
E E
E E
400 ~ .. ~ a
"c::<. "Gl.
~
300
··'? 0
.6
...
200
•a
a
100
t 973/573 20 sec
10 5 0 5 10 d(mm)
Figure 4 Microhardness profiles in the center of the Gleeble specimens
WRC Bulletin 532

SKETCH OF THE PERMEATION CELL PERMEATION TEST
charging compartment detection compartment

current at the
etectlon side
ELECTROLYIB ELECTROLYIB
PERMEATION
C.E
Imposed
Imposed E
H
2
C.E
H + +
\ +
H - H + e
/ =0
breakthrough
lime
I time
cathodic current cut

/ passivation time
start of cathodic
charging
Figure 5
(b)
(a)
~ 2
Figure 6 Hydrogen permeation (a) and outgassing (b)

curves showing the sensitivity of the method
2 - 2' : At 973/573 = 20 sec
4 : ~t 973/573 = 500 sec
WRC Bulletin 532

(a) (b)
Figure 7 Hydrogen permeation (a) and outgassing (b) curves - Influence of cooling rate
1 : ~t 973/573 = 10 sec 2 : 20 sec 3 : 50 sec 4 : 500 sec
(a)
(b)
2.0 2.5 l.O
Figure 8 Hydrogen permeation (a) and outgassing (b) curves - Influence of post-treatment
4 : b. t 973/573 = 500 sec 5 : 4 + 873 K - 30 min 6 : 4 + 873 K - 4 h
0,1 pm 0,05 pm
a b
Figure 9 Micrographs obtained by E.T.M. showing the precipitation occurring during heat
treating at 873 K
a : /J. t 973/573 = 500 sec b : a + 873 K - 4 h
WRC Bulletin 532

88 Metallurgical Studies.for Steels for Sour Service Environments
THE INFLUENCE OF SEGREGATION ON HIC

IN PRESSURE VESSEL STEELS
W. B. Morrison and K Abernethy

British Steel Technical, Swinden laboratories
Moorgate, Rotherham, South Yorkshire, England
ABSTRACT the segregated region divided by the bulk or average element

concentration. The technique used to determine Mn segregation
The influence of the composition, microstructure, hardness has been described in detail (2). A range of steels was examined
and geometry of segregation bands on HIC was studied using mainly of the linepipe and structural types and the results
composite specimens of a low carbon hot rolled steel containing a compared with those from other studies (3,4,5). The maximum Mn
segregation zone in the form of an artificial segregation band. The segregation ratio was around 2.0(2,5) which implies a peak Mn
artificial segregation band (ASB) compositions were chosen to content in the segregated zone of 1.6 to 3.0% for a bulk Mn content
study the influence of segregation intensity at different base of 0.8 to 1.5%. Co-segregation of other elements with Mn has been
manganese and phosphorus levels. noted in some studies (2,3). Butler (2) has found a relationship
Testing was done in NACE solution. Extensive cracking between the Mn segregation ratio and the segregation ratios of Nb,
occurred within the ASB when the manganese content reached a P, Cu, Si and Ni for a wide variety of steel compositions as
level of about 2.0% irrespective of the level of the other elements. illustrated in Fig. 1. It is important to note the extent of P and Nb
Two base levels of phosphorus were examined, 0.015 and 0.005%.
Element ratio
At a lower manganese level of 1.4% the lower phosphorus content
caused a large improvement in HIC performance. 11.0
The most important factor influencing HIC was the
microstructure (and related hardness) of the segregation band 10.0
which in hot rolled or normalised steels is influenced mainly by the
manganese segregation. Thus the probability ofHIC failure can be 9.0
reduced by reducing the manganese content where possible to the
order of 0.8% maximum. It is also advantageous to reduce the 8.0
phosphorus content to a low level of around 0.005%.
Other effective procedures for minimising the influence of
segregation on HIC include tempering after normalising and
quenching and tempering.
::: [
INTRODUCTION
Hydrogen induced cracking (HIC) of steel plates is

5.J
manifested by cracks which are parallel to the plate surface, may be 4.0
straight or stepwise and are associated with inclusions and with
hard segregation bands of martensite, bainite or pearlite. 3.0
Considerable steelmaking development has occurred in order to
minimise the detrimental influence of inclusions. This has 2.0
included reducing the S level to very low values of <0.002% and
introducing elements to control sulphide shape. Likewise it has
been noted that to reduce HIC problems related to segregation
bands it is necessary to control chemical composition, casting and
heat treatment (1). Such segregation bands can occur in plates Manganese ratio
produced via both ingot and continuous slab casting.
The intensity and distribution of segregated regions can be
assessed accurately using electron probe microanalysis since this is SEGREGATION RATIOS FOR OTHER ELEMENTS FIG. 1
the only technique with suitable spatial resolution. It is common to COMPARED WITH THAT FOR MANGANESE
define the intensity of segregation as the element concentration in
WRC Bulletin 532

segregation which can be expressed in terms of a power Jaw. The

expression for Pis as follows:-
All dimensions in mm
where Kp and KMn are P and Mn segregation ratios respectively.
Thus at a Mn segregation ratio of2 the P segregation ratio would be 126
around 10. The exact form of this relationship would depend upon
factors such as size of the segregated regions (5) and processing RD
history of the steel since P, due to its higher diffusion rate, would
redistribute itself more rapidly than Mn. /
There are comparatively few data related to the C content
of the segregated regions (4,5). The C diffuses to the segregation
bands during transformation and the amount segregated would Insert
depend upon factors such as bulk C content, cooling rate and the 0
I- 1 0 x 180
alloy level of the segregation band. In thick bands the C content is
likely to be non-uniformly distributed with the edges of the bands 10 r=. -------"""
ior:=-:.__~--1'--~~~--,.
having the highest C level (6). This should be around 0.5% for a
band containing 2.5% Mn (7) and would be higher for lower Mn
contents.
Although segregation bands are recognised as sites for
hydrogen cracks with microstructure and hardness being of major
importance the role of the composition and geometry of the Weld all round
segregation bands is difficult to quantify due to the variable nature
of such bands in commercial steel. One way of controlling the
composition and geometry is to manufacture composite specimens DESIGN OF COMPOSITE FIG. 2
containing artificial segregation bands. Matsumoto et aJB made use PREPARED FOR WELDING AND ROLLING
of such a technique in conjunction with a hot working simulator to
study the influence of C, Mn and P on HIC. The present paper
describes a project which uses a similar technique. However, in this designed to provide samples with three thicknesses of artificial
instance samples were prepared by rolling plates in a laboratory segregation band (ASB), namely 240, 100 and 30 um.
mill to commercial schedules. Band compositions were based upon
the relationships illustrated in Fig. 1 and the influence of band HIC Test Procedure
thickness was an integral part of the study. Although the steel used
in the major part of the investigation was a low C linepipe grade the Test piece sampling and preparation were carried out
results obtained apply also to pressure vessel steels. according to established procedure. Four test pieces were prepared
from each plate with their major axis parallel to the rolling
EXPERIMENTAL PROCEDURE direction. All tests were carried out in H2S saturated 5% sodium
chloride 0.5% acetic acid solution (pH -3.5) formulated according to
Composite Manufacture NACE standard TM-01-77. Cracks were measured at a
magnification of X30 on three metallographically prepared sections
Composite samples were made up using commercial slab and from each test piece using the standard sectioning procedure. The
inserts of laboratory steel representing segregation band graphical illustration of crack systems is given in Fig. 3 together
compositions. The chemical composition of these steels is given in with a numerical definition. The system employed for defining
Table 1. The laboratory ingots (50 kg vacuum melts) were hot crack parameters is as follows:-
rolled to 5 mm plate and then ground to 2 mm thick to provide All cracks, except isolated cracks with length less than 0.1
insert material. The insert was placed between 2 pieces of slab as mm, are included in the CLR % calculation. Cracks with an extent
shown in Fig. 2. A TIG root run followed by manual metal arc less than 0.1 mm are not included in the CSR % calculation. An
welding was used to seal the insert inside the composite block. The isolated crack or crack system is one having a separation value
blocks were then soaked at 1200°C and controlled rolled in a greater than 0.25 mm.
laboratory mill to a plate 15 mm thick using 2 holding periods to
finish at a temperature of 750°C. Composite manufacture was
TABLE 1
BASE LOW CARBON STEEL AND ASB COMPOSITION
c Si Mn p s Cu Ni Also! N Nb Ca 0
Base 0.05 0.28 0.78 0.012 0.001 0.33 0.20 0.036 0.0090 0.033 0.0050 0.0022
Steel
PL710 0.08 0.34 0.86 0.012 0.004 0.36 0.22 0.030 0.0063 0.048 0.0081 0 0031
PL705 0.29 0.36 0.84 0.018 0.003 0.19 0.21 0.040 0.0060 0.032 0.0037 0.0025
PL712 0.36 0.56 1.53 0.105 0.002 0.51 0.35 0.045 0.0052 0.20 0.0048 0.0014
PL713 0.34 0.78 2.14 0.345 0.003 0.84 0.48 0.053 0.0066 0.70 0.0065 0.0024
PL715 0.34 0.57 1.53 0.037 0.001 0.52 0.35 0.043 0.0058 0.20 0.0039 0.0025
PL716 0.35 0.80 2.17 0.111 0.002 0.85 0.49 0.046 0.0063 0.70 0.0057 0.0020
PL718 0.35 0.46 2.13 0.049 0.002 0.40 0.30 0.038 0.0054 0.11 0.0051 0.0020
WRC Bulletin 532

- - - - - - - - - - - w ________..
E
TC+.
•--+:;!_- ASB
..
S - Crack separation W Width
ELC Cr~ck lencth T Thickness
ETC - Crack extent TYPICAL INCLUSION LEVEL IN FIG. 4

ARTIFICIAL SEGREGATION BAND
E
Crack length ratio (CLR) l: ~ xlOO %
w only in the 100 um ASB (Table 2) the other 2 composites being free
ELC x ETC of cracks thus demonstrating the high quality of the interfaces.
Crack sensitivity ratio (CSR) x 100 % Metallographic examination of the composites did not reveal any
l: Wx T
interfacial inclusions although the general inclusion level of the
DEFINITION OF CRACK SYSTEMS FIG. 3 laboratory steel inserts was significantly higher than that of the
AND RATIO CALCULATIONS commercial parent steel (Fig. 4). Also an examination of testpieces
after HIC testing showed that the great majority of cracks occurred
RESULTS AND DISCUSSION within the ASB as shown in Fig. 5.
The distribution of Mn in the vicinity of the 240 um ASB
A problem commonly encountered with the composite shown in Fig. 4 is given in Fig. 6. As expected the Mn content is
samples used in the present project is that if great care is not taken relatively uniform across the ASB. This can be compared with the
in the manufacturing process the interfaces between the insert
(ASB) and the parent material become contaminated causing fine
oxide inclusions to form. These can then nucleate cracks during
HIC testing giving spurious results. Composite quality was tested
by using insert PL 710 which closely matched the composition of the
parent steel and then subjecting the resultant composite sample to
HIC testing. A small amount of cracking (CLR = 1.7%) occurred
TABLE2
HIC RESULTS FROM LOW CARBON STEEL
ASB CLR CSR

1~_,,_--~c~--..;;:;;::-~--~~~~--::::'~'----""""'~IE
Insert
PL710
Thickness
µm
%
(Av)
%
(Av) 1
240 0 0
100 1.7 0
30 0 0
PL605 240 7.6 0.14
100 0.1 0
30 2.3 0.02
PL712 240 37 0.68
100 30 0.41
30 27 0.17
PL713 240 70 0.96
100 71 0.40
30 55 0.02
PL715 240 8.6 0.30
100 12 0.19
30 0 0
PL716 240 61 0.98
100 60 0.18
30 56 0
PL718 240 51 0.69
100 47 0.32 CRACKS IN ARTIFICIAL FIG. 5
30 38 0 SEGREGATION BAND (PL718l
WRC Bulletin 532

Manganese ferrite. Using the composite technique it is possible to simulate the

wt. % presence of individual segregation bands but it would be very
difficult to simulate a segregation region composed of a group of
2.6
closely spaced bands.
Table 2, which contains the HIC results, shows that when the
2.4 Mn content of the ASB reached 2% (PL713, PL 716 and PL718) i.e.
when the Mn segregation ratio was 2.5 for a base level Mn of 0.8%
or when the ratio was 1.4 for a base level of 1.4% severe cracking
2.2 occurred. In the latter case (PL 718) the other elements (Si, P etc.)
were at lower levels than those in the other high Mn ASBs and this
gave a lower hardness and appeared to promote less cracking. At a
2.0
Mn content of around 2.0% the microstructure of the ASBs
contained martensite and therefore were of high hardness. Due to
1.8 the uneven distribution of C in the thicker bands there was a
tendency for the average hardness of the thicker bands to be lower
than those of the thin bands. However, maximum hardness was
1.6 relatively independent of band thickness. Figure 8 shows how the
hardness of the 30 um ASB increases as Mn content increases. The
scatter in hardness results can be largely explained by the fact that
1.4
even at similar Mn contents there was a variation in the content of
the other elements. Nevertheless, the relationship is clearly
defined and shows that Mn has the main influence on the hardness
1.2
of the ASB, assuming that sufficient C is available to harden the
structure.
1.0
Microhardness
HV25
0.8 800
0.6 ......~-'-~~"'-~--'~~-'-~~'--~-I-~--'
0 50 100 150 200 250 300 350 700
Distance, µm
600
MANGANESE DISTRIBUTION ACROSS INSERT FIG. 6
MADE FROM STEEL PL716
Mn distribution across the centreline of a commercial plate which 500 •

shows a series of Mn peaks which correspond to individual
segregation bands as illustrated in Fig. 7. It is commonly observed
that centreline segregation in commercial steels is composed of 400
individual bands of high C constituent separated by bands of
Manganese 300 0 o CLR < 10%
wt. %
MnS G CLR > 10 < 30%
2.8
~
00
200 0 9 CLR > 30%
2.6
100'--~~~~~.._~~~~~-'-~~~~~~
2.4 1.0 2.0 3.0

Mn, 3
2.2
2.0
I RELATIONSHIP BETWEEN MANGANESE LEVEL
AND HARDNESS OF THE 30 µm ARTIFICIAL_
SEGREGATION BANDS AND THE HIC PERFORWiANCE
FIG. 8
~ ~
Above a hardness of around 300 H v in the 30 um ASB, HIC
becomes significant (> 10% CLR). In the literature it has been
1. 8
widely observed that HIC tends to occur in bands of low
temperature transformation structure (3,9-12) and it has also been
1.6 noted that above 300 Hv HIC becomes particularly severe (13) as
observed in the present work. Nevertheless despite the observation
that hard bands are the most likely to suffer from hydrogen
1.4 cracking, cracking can also occur in pearlite bands of hardness
around 200 Hv. A comparison of the HIC results from inserts
PL 710 and PL 705 reveals that although cracking levels were low in
1.2 0 50 100 150 200 250 300 350 both cases PL705 had a much higher CLR value than PL710. The
only compositional difference between the two inserts is that the C
Distance, µm content was initially higher in PL 705 prior to manufacture of the
composite, the composition otherwise of both inserts representing
MANGANESE DISTRIBUTION ACROSS FIG. 7 that of the commercial steel i.e. nil segregation. Figure 9 shows
CENTRELINE SEGREGATION IN X60 LINEPIPE cracking in pear lite bands in the 240 um ASB of insert steel PL 705.
WRC Bulletin 532

Although during manufacture of the composite the C tends to There are various possible solutions which could be adopted
diffuse away from the insert material into the parent steel, enough to avoid HIC problems related to segregation bands in commercial
remains in the ASB of the final composite to give some marked steels. They are to reduce alloy levels, particularly Mn, P and C, to
pear lite banding. Cracking only occurred in the insert steel and not use rapid cooling (accelerated cooling) after rolling, to temper after
in the pearlite adjacent to the ASB. The explanation probably lies normalising, and to employ quenching and tempering as a process
in the fact that the morphology of the pearlite in the 2 regions is route.
quite different. In the insert the pearlite exists mainly as long It is advisable to reduce the Mn content to as low a value as
continuous bands while in the parent steel it takes the form of much possible, of the order of0.8%. With a maximum segregation ratio of
shorter bands or colonies. A possible contributory factor is that the 2.0 this will ensure a maximum Mn content in the segregation
inclusion level, as already pointed out, is fairly high in the insert bands of 1.6%. It is also important to keep P levels to a minimum,
steels and could thus act as initiation sites for hydrogen cracks for example less than 0.005%. Other elements such as Cu and Ni
which then propagate along the pearlite bands. Further limited are useful for replacing Mn in order to maintain mechanical
trials (14) with insert steel of even higher C (0.6%) but of otherwise properties. Of particular benefit is the fact that they have a similar
similar composition gave much more severe HIC indicating the segregation ratio to that of Mn but a smaller influence on
major influence of pear lite banding per se. hardenability.
It is assumed that during processing C will diffuse to the
segregation bands during the austenite to ferrite transformation.
However, if this can be prevented, the segregation bands, even if
high in Mn, will develop a microstructure much less susceptible to
HIC. There are two ways by which the C level of the segregation
bands can be minimised; (a) reduce the C content in the parent steel
to very low levels ( < 0.02%) in conjunction with increased alloy
levels to maintain properties (3,18) and (b) employ accelerated
cooling after hot rolling (19) or quenching after austenitising. The
effectiveness of accelerated cooling in reducing C diffusion to the
segregation bands during transformation is reduced as plate
thickness increases and so accelerated cooling after rolling could
only usefully be applied below a certain plate thickness, the value
of which would depend upon compositional, structural and cooling
rate factors.
Microhardness
HV25 g
CRACKING IN PEARLITE BANDS (PL705) FIG. 9 500+~

( Normalised
+ Edge
0 Centre
The influence of reducing the P content can be observed in
Table 2 by comparing inserts PL715 and 716 with inserts PL712
and 713 since the former are low P versions (base level 0.005%) of
400 ',,~ l:l. Parent
the latter (base level 0.015%). For a high Mn segregation ratio of

2.5, giving 2.0% Mn for a base content of 0.8%, the reduction in P 300 10
1-
had a small beneficial but not significant effect on HIC
susceptibility (PL716 and 713). However, at a lower Mn
--o- - - - - - -::t;
segregation ratio of 1.75, giving 1.4% Mn for a base content of 0.8%,
the lower P content caused a large improvement in HIC 200 A-------"------,t:;.------t:.
performance reducing CLR from around 30% to around 10% (PL 715
and 712). The main difference in the two ASB microstructures was
the presence of some martensite in the ASB containing the high P 0 0 550 600 650
level which contributed, together with solid solution hardening, Tempering temperature, °C
towards an increase in hardness of around 60 Hv. This effect of Pon
hardenability although not widely recognised has been observed by
other workers (15,16). The reason for the very much smaller effect
of P in the higher Mn inserts lies in the fact that the high Mn
500
content in these inserts was sufficient to cause martensite r Quenched
:::~- ----~
formation without any contribution from the P content. In the
steels studied the element with the most important influence on the
microstructure of the ASB was Mn. This is not unexpected since
Mn is the major alloying element and is known to have a
pronounced effect on hardenability.
The results in Table 2 show that in the majority of casts the
thinnest ASB (30 um) cracked least but at greater thicknesses there
was no systematic effect of band thickness i.e. the 100 um and the
240 um bands cracked to a similar extent. The observation that the 200 - - A
thinnest band cracked least despite being generally harder than the
thicker bands indicates that there is a true thickness effect and
agrees with a study (17) of cracked commercial steel samples which 0 .._~,-----'-------'--------'
suggested that the thickness of a segregation band affects the 0 550 600 650
probability of it cracking. Thick bands are more likely to contain
Tempering temperature, °C
suitable nuclei for hydrogen cracking, such as inclusions. Band
thickness in the final plate will be influenced by the initial as-cast
structure and by the amount of rolling reduction. Although thick EFFECT OF TEMPERING ON HARDNESS OF FIG. 10
plates would tend to have the thickest segregation bands this NOP.MALISED AND QUENCHED STRUCTURAL
adverse effect on HIC would be counteracted to some extent by the STEEL CONTAINING ARTIFICIAL SEGREGATION
slower cooling rate which would tend to decrease hardness. BAND OF Mn SEGREGATION RATIO 2.0
WRC Bulletin 532

Metallurgical Studies.for Steels.for Sour Service Environments 93
TABLE3 5. To optimise HIC performance the parent plate manganese

and phosphorus contents should be reduced with targets of
BASE STRUCTURAL STEEL AND ASB COMPOSITIONS
approximately 0.8% and 0.005% respectively.
6. The influence of segregation on HIC in pressure vessel steels
c Si Mn p s Cu Ni Al sol N Nb can be minimised tempering after normalising and by employing
quenching and tempering as a process route.
Base 0.09 0.38 1.45 0.020 0.003 0.02 0.29 0.035 0.004 0.020
Steel ACKNOWLEDGMENTS
ASB 0.30 1.03 3.10 0.24 0.013 0.05 0.52 0.040 0.0077 0.25 The results reported here are partly from an ECSC sponsored
Tempering after rolling, normalising or quenching is an project and acknowledgment is made of the financial assistance
effective method of reducing the hardness of segregation bands and provided.
therefore making them less susceptible to HIC. This is a The authors wish to thank Dr. R. Baker, Director of Research
particularly advantageous process step in the manufacture of and Development, British Steel Technical, for permission to publish
pressure vessel plates. Several workers (10,12,20,21) have this work.
compared quenched and tempered steels with steels in the
controlled rolled condition and have attributed the improvement at REFERENCES
least partly to a more uniform microstructure. Avoidance of
elongated microstructural constituents, particularly if hard, is 1. Dewsnap, R.F., Jones, C.L., Lessells, J., Morrison, W.B.,
essential if a high resistance to HIC is required. Figure 10 shows Rudd, W.J., Walker, E.F., and Wilkins, R., 'A Review of
the effect of tempering on the hardness of ASBs of Mn segregation Information on Hydrogen Induced Cracking and Sulphide
ratio 2.0 in a structural steel in the normalised and quenched Stress Corrosion Cracking in Linepipe Steels', Offshore
conditions. The compositions of the steel and the ASB are given in Technology Report OTH 86 256, Department of Energy,
Table 3. In both the normalised and quenched conditions 1987.
tempering has reduced the hardness of the ASB to a similar level. 2. Butler, S.A., 'Segregation Studies in Low Alloy Steels by
Note, however, that there remains a large hardness difference Energy Dispersive X-ray Spectrometry', Scanning, Vol. 9,
between the ASB and the parent plate in the case of the normalised 1987, pp 204-214.
and tempered steel whereas the difference is quite small in the 3. Taira, T., Kobayashi, Y., Matsumoto, K., Matsumoto,
quenched and tempered steel. When subjected to HIC testing in S.,Terunuma, T., and Arikata, K., 'Resistance of Pipeline
NACE solution (Table 4) the Q + T steel shows a considerable Steels to Wet Sour Gas', Hydrogen Problems in Steel,
improvement over the N + T steel. This is considered to be a result ASM, Metals Park, Ohio, 1983, pp 173-180.
of the greater uniformity of microstructure within the Q + T steel. 4. Jones, C.L., Rodgerson, P., and Brown, A., 'Mechanisms of
Cracking is nevertheless quite substantial in the ASB of the Q + T Hydrogen Induced Cracking in Pipeline Steels',
steel due to the presence of unmodified sulphides and illustrates the International Conference on Techniques and Application of
point made earlier that both high cleanness and structural HSLA Steels, Philadelphia, P.A., 3-6 October 1983, pp 809-
uniformity are necessary for HIC resistance. 825.
5. Saeki, T., Imura, H., Oonishi, Y., Niimi, H., Miwa,
TABLE4 E.,Yoshida, T., Igari, S., and Kitamine, S., 'Effect of Bulging
HIC RESULTS FROM STRUCTURAL STEEL WITH ASB and Solidification Structure on Segregation in Continuously
Cast Slab', Trans ISIJ, Vol. 24, No. 11, 1984, pp 907-916.
6. Tamehiro, H., Yamada, N., and Matsuda, H., 'Effect ofthe
CLR CSR Thermo-Mechanical Control Process on the Propertiesof
Condition % % High Strength Low Alloy Steel', Trans ISIJ, Vol. 25,No. 1,
(Av) (Av)
1985, pp 54-61.
7. Hillert, M., 'Prediction of Iron Base Phase Diagrams',
Normalised 58 0.32 Hardenability concepts with applications to steel, (Proc.
Conf.) 24-26 October 1977, Met. Soc. AIME. Warrendale,
Normalised + 33 0.13 P.A., 1978, pp 5-27.
Tempered 8. Matsumoto, K., Kobayashi, Y., Ume, K., Murakami,
K.,Taira, K., and Arikata, K., 'Hydrogen Induced Cracking
Quenched+ 28 0.12 Susceptibility of High Strength Line Pipe Steels', Corrosion,
Tempered Vol. 42, No. 6, June 1986, pp 337-345.
CONCLUSIONS 9. Miyoshi, E., Tanaka, T., Terasaki, F., and Ikeda,
A.,'Hydrogen-Induced Cracking of Steels Under Wet H 2 S
1. Composite specimens were prepared which successfully Environment', ASME Publication Paper No. 75-Pet-2,
simulated steel containing centreline segregation bands. The (1975).
compositions of the artificial segregation bands (ASBs) were based 10. Taira, T., Kobayashi, Y., Inagaki, H., andVlatanabe T., 1
on the microanalysis of segregation bands in a variety of 'Sulphide Corrosion Cracking of Linepipe For Sour Gas
commercial steels. Service', Corrosion 79, Paper 171.
11. Steinberg, B.G., and Kane, R.D., 'Stepwise Cracking of
2. Extensive HIC occurred when samples containing ASBs with Linepipe Steels in Simulated Oilfield Environments' 14th
around 2.0% manganese were tested in NACE solution. This level OTC (4430), Houston, Texas, 3-6 May 1982, pp 369-377'. - -
of manganese represents a high segregation ratio of 2.5 for a base 12. Taira, T., Tsukada, K., Kobayashi, Y., Tanimura,
manganese content of 0.8% or a ratio of 1.4 for a base content of M.,lnagaki, H., and Seki, N., 'HIC and SSC Resistance of
1.4%. Linepipe for Sour Gas Service', NKK Tech Report Overseas,
3. At a high manganese level in the segregation band of 2.0% a No. 31, 1981, pp 1-13.
reduction of parent plate phosphorus content from 0.015 to 0.005% 13. Moore, E.M., 'Hydrocarbon Induced Damage in Sour
did not significantly influence HIC. However, at a manganese level WetCrude Pipelines', Middle East Oil Technical Conference
of 1.4% in the band the reduction in phosphorus caused a large ofSPE, Morama, Bahrein, 14-17 March 1983, pp 587-594.
improvement in HIC performance. 14. Morrison, W.B., Unpublished work, British Steel pie, 1988.
4. The influence of manganese and phosphorus content on HIC 15. Grange, R.A., 'Estimating the Hardenability of Carbon
could be linked to the microstructure and related hardness of the Steels', Met. Trans. 4, 1973, p 2231-2244.
segregation band.
WRC Bulletin 532

16. Clarke, B.D., Mcivor, I., 'Effect of Phosphorus on

theMicrostructure and Strength of High Carbon Steel Rod',
to be published.
17. Butler, S.A., Unpublished work, British Steel plc, 1988.
18. Ohtani, H., Hashimoto, T., Sawamura, T., Bessyo, K. and
Kyogoku, T., 'Influence of Metallurgical Factors on HIC of
High Strength ERW Linepipe for Sour Gas Service',Int. Conf.
HSLA Steels Technology and Applications,Philadelphia,
P.A., USA, 3-6 October 1983, American Society for Metals,
Metals Park, Ohio 44073 USA, 1984, pp 843-854.
19. Tamehiro, H., Takeda, T., Matsuda, S., Yamamoto, K.,and
Okumura, N., 'Effect of Accelerated Cooling AfterControlled
Rolling on the Hydrogen Induced CrackingResistance of
Linepipe Steel', Trans ISIJ, Vol. 25,1985, pp 982-988.
20. Taira, T., Tsukada, K., Kobayashi, Y., Inagaki, H.
andWatanabe, T., 'Sulfide Corrosion Cracking of Linepipe for
Sour Gas Service', Corrosion, Vol. 37, No. i, 1981, pp 5-16.
21. Nishimura, T., Inagaki, H., and Tanimura, M.,
'HydrogenCracking in Sour Gas Pipeline Steels', Second
International Conference on Hydrogen in Metals, Paris, 6-11
June 1977, Paper No. 3E9.
WRC Bulletin 532


WRC 532-2009

Uploaded by

Copyright:

Available Formats

WRC 532-2009

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WRC 532-2009

Uploaded by

Copyright:

Available Formats

INTENTIONALLY LEFT BLANK

For more information, see www.forengineers.org

Copyright © 2009 The Welding Research Council.

Welding Research Council

ETALLURGICAL STUDIES OF STEELS FOR

Martin Prager, Editor

WRC Bulletin 532

WELDING RESEARCH COUNCIL, INC.

WRC Bulletin 532

ISBN No. 1-58145-539-9

WRC Bulletin 532

Considerable research aimed at understanding hydrogen-steel interactions has been

WRC Bulletin 532

WRC Bulletin 532

QUENCHED AND TEMPERED STEELS FOR H 2 S CONTAINING PRESSURE VESSELS .............. 27

INFLUENCE OF ACCELERATED COOLING ON THE HYDROGEN INDUCED CRACKING

EFFECT OF SOME METALLURGICAL FACTORS ON THE SUSCEPTIBILITY TO SSCC OF

INFLUENCE OF THERMAL TREATMENT ON THE HYDROGEN EMBRITTLEMENT OF

ON THE EFFECT OF NON-METALLIC INCLUSIONS ON HYDROGEN DAMAGE IN

INFLUENCE OF MICROSTRUCTURE ON DIFFUSION AND TRAPPING HYDROGEN

THE INFLUENCE OF SEGREGATION ON HIC IN PRESSURE VESSEL STEELS ............................. 89

WRC Bulletin 532

WRC Bulletin 532

REFINERY EXPERIENCE WITH CRACKING OF

ABSTRACT and promotes their absorption into the steel. The

WRC Bulletin 532

Inspection Program Since the inspection was directed at vessels con-

WRC Bulletin 532

WRC Bulletin 532

WRC Bulletin 532

2. M.J. Humphries, J.E. Mclaughlin, L.J. Pargeter,

6. H.C. Rogers, MATERIALS PROTECTION, April 1962, p.

WRC Bulletin 532

Fig. 6 Transverse cracking in the weld metal of a main

End of crack shown in Figure 3.

Fig. 7 Cracking found in she n of heat exc:hange·r

WRC Bulletin 532

Fig. 8 Cracks near that 7 showing the

Fig. 11. 10 but

WRC Bulletin 532

Fig. 12. SOHIC tests on Proof Rings.

WRC Bulletin 532

INSPECTION PROGRAM FOR EQUIPMENT EXPOSED

WRC Bulletin 532

Vessels Inspected 588

Crack Length > 2 in (50 mm) 136

WRC Bulletin 532

DEVELOPMENT OF PRESSURE VESSEL STEEL SERIES FOR

ABSTRACT on temperature and time for PWHT in order to avoid

WRC Bulletin 532

The production steps in the steel plant which were

non sour gas steels

WRC Bulletin 532

Production Hot metal Converter Vacuum tank Casting (CC or ingot)

in the steel for steels with HIC resistance

WRC Bulletin 532

0 curved strand machine Rolling Shape Factor -

FIGURE 5: Influence of high shape factor rolling

({/) u::::dllllllllllllllllllllllllllllllllllWij!h 11dll111111111mm111111111111111~1111111th

Production Reheating Rolling Warm pilling Normalizing I Tempering or