WRC 534-2009

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.
STUDIES OF CLADDING AND OVERLAY FOR

PRESSURE VESSEL SERVICE
Martin Prager, Editor
WRC Bulletin 534
THE WELDING RESEARCH COUNCIL, INC.

PO Box 1942
New York, NY 10156
www.forengineers.org
WRC Bulletin 534

Studies of Cladding and Overlay for Pressure Vessel Service
ISBN No. 1-58145-541-0
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 534

ii Studies of Cladding and Overlay for Pressure Vessel Service
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. The problems and solutions are complex because of the diverse
microstructures 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 Methods for Hydrogen
Induced Cracking ( # 530), Metallurgical Studies of Steels for Sour Service Environments ( # 532),
Studies of Cladding and Overlayfor Pressure Vessel Service ( # 534), and Toughness, Fracture and
Fitness for Hydrogen Service ( # 535).
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.
Martin Prager
Executive Director, WRC
WRC Bulletin 534

Studies of Cladding and Overlay for Pressure Vessel Service iii
WRC Bulletin 534

iv Studies of Cladding and Overlay for Pressure Vessel Service
Table of Contents
HYDROGEN IN THICK-WALLED, HIGH-TEMPERATURE HYDROGEN REACTORS DURING

SERVICE AND DURING SHUTDOWN ................................................................................................................... 1
F. Dupoiron, L. Coudreuse, D. Jobard, and J. Charles
RESISTANCE TO HYDROGEN INDUCED DISBONDING OF HOT ROLLED HEAVY GAUGE

CLAD STEEL FOR APPLICATION IN PETROLEUM INDUSTRY PRESSURE VESSELS ..................7
Hideto Kimura, Minoru Suwa, Sadahiro Yamamoto
CORROSION RESISTANT CLAD PLATES FOR HIGH PRESSURE VESSELS IN HYDROGEN

AND OFF-SHORE ENVIRONMENT ...................................................................................................................... 19
F. Dupoiron, L. Coudreuse, D. Jobard, and J. Charles
EVALUATION OF STAINLESS STEEL EXPLOSION CLAD FOR HIGH TEMPERATURE, HIGH

PRESSURE HYDROGEN SERVICE ....................................................................................................................... 27
J. G. Banker, M. S. Cayard
HYDROGEN INDUCED DISBONDING IN WELD-CLAD PRODUCTS: THE INFLUENCE

OF MICROSTRUCTURE ............................................................................................................................................. 35
C. Casciaro, P. Marinelli, A. Solina, R. Valentini
DISBONDING MECHANISM AND ITS PREVENTION .................................................................................. 41

A. Vignes, R. Palengat, P. Bocquet
WRC Bulletin 534

Studies of Cladding and Overlay for Pressure Vessel Service v
WRC Bulletin 534

vi Studies of Cladding and Overlay for Pressure Vessel Service
HYDROGEN IN THICK-WALLED, HIGH-TEMPERATURE HYDROGEN REACTORS

DURING SERVICE AND DURING SHUTDOWN
Reflection of APl/MPC Work and Direction for Future Research
K. Smit and P. F. Ivens

Shell Internationale Petroleum Maatschappy B. V.
The Hague, The Netherlands
ABSTRACT reactors on which - in the light of research results

acquired from the API/MPC sponsored programme - the
A few aspects of engineering and operating high established rules and guidelines need to be
temperature, high pressure hydrogen reactors are reconsidered. Suggestions are made for future
discussed in the light of research results acquired directions for API/MPC sponsored research.
from the API/MPC programme. Considerations are given
for revisions of API document 941, which specifies
steels that are resistant to hydrogen attack, and API MATERIALS FOR HIGH TEMPERATURE HYDROGEN SERVICE
document 946 on cooldown procedures. The derivation of
allowable design stresses at temperatures close to or The 2\CrlMo and 3Cr1Mo steels have been improved in
in the creep range and near the limit for hydrogen recent years to give better hot tensile and creep
attack needs to be reviewed. Recommendations are made properties. The mechanical properties of 2\CrlMo steel
for future research embracing a) the solubility and can be enhanced either by closer restrictions on heat
diffusivity of hydrogen, b) the effect of hydrogen on treatment alone or in combination with micro-alloying
creep rupture strain and strain rate and its with vanadium and/or niobium. The first method is
significance for creep design and defect assessment, suspected to result in reduced resistance to hydrogen
c) the mechanism of disbanding of weld overlay attack, whereas the method of micro-alloying with V
claddings and d) safe limits for the hydrogen and/or Nb, which are carbide stabilising elements, is
concentration in a reactor wall during shutdown and considered beneficial for the resistance to hydrogen
safe pressure limits for reactors at low temperature. attack.
INTRODUCTION A major improvement of hot tensile and creep properties

has been achieved by modifying the 2\CrlMo and 3Cr1Mo
Heavy wall reactors for high temperature hydrogen steels with the addition of 0.25% of vanadium.
service are designed and operated in accordance with Simultaneously, the resistance to hydrogen attack is
established rules and guidelines. Some of these are: improved to the extent that the 2\CrlMo\V steel and its
a) The design code which is locally in force (e.g. weld metals have a better resistance than the
ASME VIII div 2 or BS 5500) for the mechanical conventional 3Cr1Mo steel (Ref. 3). Therefore, it seems
design. The maximum stresses allowed by these codes justified that the temperature limit for the
are generally related to the hot tensile and/or application of 3Cr1Mo steel in high temperature
creep strength of the steel in air at the design hydrogen service as given by the Nelson diagram in API
temperature. document 941 may also be applied to the vanadium
b) API document 941 for the selection of steels which modified 2\CrlMo\V steel. At high hydrogen partial
are resistant to hydrogen attack and decarburisation pressures this limit is 55 °C above the 454 °C
(Ref. 1). temperature limit for conventional 2\CrlMo steel. Also,
c) API document 946 to keep the concentration of the resistance to hydrogen attack of the vanadium
residual hydrogen in the CrMo steel during shutdown modified 3Cr1Mo\V steel is better than the resistance
below the critical level for hydrogen damage of the straight 3Cr1Mo steel
(blistering / flaking) which is stated to be 3 ppm
(Ref. 2). Recent series of creep tests on CrMo steel and
weldments - including heat affected zones - in a high
Developments in hydroprocessing technology demand pressure hydrogen environment have indicated that the
time to rupture in hydrogen may be nearly an order of
higher maximum operating temperature, total pressure,
magnitude smaller than in air (Ref. 4). The reduction
partial hydrogen pressure and reactor volume. At the
same time there have been developments in steel making of time to rupture may have been caused in part by
and reactor fabrication. These developments make it accelerated hydrogen attack under the mechanical load.
The vanadium modified steels are more resistant to
necessary to reconsider the materials engineering
hydrogen attack. Early results from the API/MPC
practices as laid down in the above design rules and
programme suggest that any reductions of the time to
API guidelines. An extensive effort is made by the
creep rupture which is caused by hydrogen relative to
API/MPC task group on materials for pressure vessel
the time to rupture in air will be less severe than for
service with hydrogen at high temperatures and
the plain 2\CrlMo steels. Clearly, the effect of
pressures to adjust the materials technology for high
hydrogen on creep deformation and the synergism of
temperature hydrogen reactors to the demands made by
hydrogen attack and creep need further modelling and
developments in process technology.
experimental verification.
The aim of this paper is to highlight a few aspects of The allowable design stress as given by the design code
engineering and operating high temperature hydrogen is generally related to either the hot tensile
WRC Bulletin 534

Studies of Cladding and Overlay for Pressure Vessel Service 1
properties or the stress to cause rupture in air in

100,000 hours, whichever yields the lowest design
stress value. The above creep test results indicate Temperature, °F
that the allowable design stress should be carefully 40.5 µmoles H /100 gms
2
selected when steels are applied at a temperature close
to the creep range and in particular when this 1400
temperature is also close to the established limit for
hydrogen attack. 1200
CALCULATED HYDROGEN DISTRIBUTIONS THROUGH THE REACTOR 1000

WALL
Models to calculate the hydrogen concentration profile

during operation and the hydrogen outgassing and 800
diffusion during cooldown for a reactor wall with an
internal stainless steel cladding are well established 700
(e.g. Ref. 2 and 5). The models are based on Fick's
first law of diffusion. In our particular model it is 600
assumed that a discrete boundary exists between the
cladding and the base metal and that the hydrogen
concentrations in the boundary layers on either side of 500
the interface are in equilibrium with each other at any
moment. The hydrogen concentrations which are
calculated with these models are largely dependent on
the hydrogen solubility and diffusivity data used. 400
- Hydrogen Solubility and Diffusivity -

300 2
10
The hydrogen solubility under equilibrium conditions Pressure Hydrogen, psi
and the diffusivity are assumed to be given by the
following relationships: Fig. 1. Solubility of Hydrogen in Steel
c Avp.e·B/T (1) and D = X.e·Y/T (2) (ref. 6)
where
C concentration of hydrogen in the steel in ppm Hydrogen Concentration
p partial hydrogen pressure in the gas phase. in ppm(w)
surrounding the steel in bar.
T
D
temperature in K.
2
diffusion coefficient in cm /sec.
A, B, X and Y are constants
1 5.0
Values for A and B for 2.25Cr-1Mo steel have been

calculated for a number of solubility data sets in the
literature: 6
1) Armbruster (Ref. 6)
A= 39.7 ppm/vbar; B = 3333 K 1.0
The data apply to both carbon and "low chromium"
steels and have been reported in ml H2/100g steel
This solubility has often been presented as in Fig.l 0.5 2.25Cr-1 Mo steel
2) API-publication 946 (Ref. 2)
all data of table 4 for temperatures of 700 °F and
above: A= 53.3 ppm/vbar; B = 3600 K
3) Geller and Sun (Ref. 7)
A= 42.7 ppm/..Jbar; B = 3280 K
4) T. Sakai c.s. (Ref. 5)
A= 23.6 ppm/vbar; B = 3257 K
The solubility lines are given in Fig. 2. The top and

0.1 ref's
and bottom of the range of uncertainty differ by about
a factor of 1.7. Other data on ferritic steels (Ref. 8) 7
suggest that much lower solubilities than given in 0.05 6
Fig.2 may occur. The solubility of hydrogen in 2
present-day steels seems to be lower than in older 700 500 400 30cfC 5
steels. Further research should provide a more accurate
knowledge of the solubility of hydrogen as a function
of steel chemistry and microstructure to enable a 1.0 1.5 2.0
critical assessment of the risk of hydrogen damage. The 1000/T(K) __.,.
solubility data for the stainless steel cladding as
given in Refs. 2, 5 and 6 are also given in Fig. 2. The Fig.2. Solubility of Hydrogen at 1 bar Hz
constants A and B are:
WRC Bulletin 534

2 Studies of Cladding and Overlay for Pressure Vessel Service
A in ppm/.Jbar B in K
8.93 650
Ref. 2 1.0
7.6 630
Ref. 5
Ref. 6 14.8 1099
RDF T = 450°C
The literature data on the diffusivity in the cladding

and in the 2\CrlMo steel are fairly scattered. For the
t \_
><;;. 214cr1 Mo Cladding
calculations in this paper the following data have been ~ A: 23.6 8.93
used:
2
Y in K
/ B: 3257 650
X in cm_is
2\CrlMo steel: 2.4 10_4 2132 0.5 '<
cladding: 7.1 10 4555 r

x
)<
~
x
The Hydrogen Concentration during Operation and the )(
x
)<
)<)<~ )<
)(
)<
Effect of a cladding -
2 14Cr1 Mo Claddin~V "
~
The maximum amount of hydrogen that dissolves during A: 42.7 7.6

.l! )<
operation just below the inner surface of a solid B: 3280 630
2\Cr-lMo steel reactor wall can be calculated with the
solubility data from the previous paragraph. The
0
uncertainty in the solubility data causes the same 0 0.05 0.10
uncertainty in the maximum hydrogen concentration. For Cladding/wall thickness ratio--...
example, at 450 °C and 150 bar hydrogen partial
pressure the maximum hydrogen concentration is Fig.3. Reduction Factor for the Hydrogen Partial Pressure
calculated to be between 5.6 and 3.2 ppm when using the
highest and the lowest solubility of Fig. 2. The
maximum hydrogen concentration may be even lower.
The available hydrogen solubility data and diffusivity
Many high temperature hydrogen reactors have an data show considerable scatter. The consequent range of
internal 18/8 stainless steel cladding to protect the uncertainty for the reduction factor of the hydrogen
2\CrlMo steel from corrosive attack by the reactor partial pressure is given in Fig. 3 as a function of
environment. The presence of the cladding will reduce the cladding over wall thickness ratio at a temperature
the maximum hydrogen concentration in the 2\CrlMo steel of 450 °C. This range only represents the uncertainty
during steady state operation. This reduction is mainly in the hydrogen solubility data in Fig. 2, but does not
caused by the low diffusivity of hydrogen in the account for the uncertainty in the diffusivity data.
stainless steel relative to the diffusivity in the The curve in Fig. 3 is only weakly dependent on the
2\CrlMo steel. temperature.
The effect of the cladding on the maximum hydrogen Residual Hydrogen Concentration during Shutdown -
concentration can be expressed by a reduction factor
which is defined as the ratio between the notional A typical hydrogen profile through a reactor wall
hydrogen partial pressure in equilibrium with the before and after a certain cooldown procedure is given
hydrogen concentration just underneath the cladding and in Fig. 4. Similar profiles are given in ref 2 and 5.
the hydrogen partial pressure in the reactor. The After cooldown, a very high hydrogen concentration
hydrogen partial pressure has been chosen for the occurs in the cladding at the interface with the CrMo
definition of the reduction factor rather than the steel; the hydrogen concentration in the CrMo steel is
hydrogen concentration in the steel because the low near the interface because of the much higher
hydrogen partial pressure is the independent variable solubility in the adjacent cladding material and the
in the Nelson diagram (Ref. 1). It is noted that maximum concentration occurs some distance away from
actually the reduction factor is never applied when the interface.
steels are selected for high temperature hydrogen
service with the Nelson diagram. When the temperature Again, the hydrogen profiles are strongly dependent on
through the wall is nearly constant, the reduction the solubility and diffusivity data used in the
factor (RDF) can be shown to be equal to: calculation. The solubility values used for the 2\i;CrlMo
2 steel correspond to the upper boundary of the range in
RDF I 1 1 Fig. 2.
1 Cbm Dbm tell
1
I + Gel Del tbml In an attempt to reduce the residual hydrogen
where: concentration during shutdown, sometimes a degassing
Gel and Cbm the hydrogen solubility in the cladding period is applied during the cooldown of high
and the base metal, respectively, at the temperature hydrogen reactors. The effect of such a
wall temperature and at the same hydrogen degassing period for a thick reactor wall is shown in
activity. Fig. 5, which gives the reduction of the maximum
Del and Dbm the hydrogen diffusivity in the cladding hydrogen concentration in the CrMo steel during
and the base metal, respectively, at the cooldown for a procedure with and a procedure without a
wall temperature. degassing period of 24 hours at 330°C. The benefit of
tel and tbm the thickness of the cladding and the the degassing period on the maximum hydrogen
2\CrlMo steel wall, respectively. concentration is only small (less than 10%),
particularly in relation to the uncertainty in the
hydrogen solubility and diffusivity data.
WRC Bulletin 534

[H] in cladding
t 200
[H] in CrMo-steel
~~~~~~~~~~~~~~~~-in_.Tppm
in ppm
- - prior to cooldown 5 .o
t - Disbanding of the Cladding -
Disbanding of the cladding is generally considered as a

form of hydrogen damage which occurs during cooldown.
- after cooldown Therefore, disbanding is simulated in experiments where
4.0 cladded samples are exposed for short periods to high
'' temperature hydrogen and subsequently cooled in air
3.0 (Ref. 2 and 5). Disbanding is observed in such tests;
100 in particular when the cooling rates are high. However,
2.0 it is uncertain whether this mechanism of disbanding is
relevant for the disbanding that has been observed
1.0 occasionally in actual high temperature hydrogen
reactors. Another feasible mechanism may be hydrogen
0 +5.,--..-+0--.--.-.-.-1..0-0-,.--,---,--...,-2..,.0-0...,..... 0 attack at high temperature along the fusion line where
during overlay welding carbides have dissolved and
cladding CrMo-steel distance in mm-jjoo- reprecipitated which are less resistant to high
,_. I ...
temperature hydrogen attack than the carbides in the
Fig. 4. Hydrogen Distribution through the Reactor Wall base material.
Hydrogen Solubility: CrMo steel Cladding
A (ppm/{ bar): 45 8 In the case that disbanding really occurs during
B (K): 3300 630 shutdown, the amount of residual hydrogen which remains
near the interface between cladding and CrMo steel
[H] in ppm after cooldown could be an important parameter. As for
-r-:----------,
t 4.0
II
I cooldown without 24 hr degassing
cooldown with 24 hr degassing
hydrogen damage in the CrMo steel, the critical
hydrogen concentrations on either side of the interface
to cause disbanding are unknown. The degree of
supersaturation of the hydrogen near the interface is
much smaller than in the CrMo steel away from the
interface (see Fig. 3). This means that the hydrogen
pressure which can build up inside a defect and
3.0 provides a driving force for growth of the defect is
much smaller as well.
The risk of disbanding is mainly dependent on the steel

composition, the fabrication practices used for making
2.0-1-~-.~-.~~,..-~.,.-~-.~-.~~,..----; the cladding (Ref. 10 and 11) and the operating
0 20 40 60 80 temperature and hydrogen partial pressure. The
__....,.. time in hr experiments in Ref. 11 indicate that the cooling rate
Fig. 5. Peak Hydrogen Concentration in the CrMo steel is also an important parameter for disbanding. Hydrogen
during cooldown. (solubility as in fig.4) distributions (Fig. 5) in the test specimens after
cooldown have been calculated for the disbanding
experiments in Ref. 11. Similar to the profile in a
ASSESSMENT OF THE RISK OF HYDROGEN DAMAGE reactor wall after cooldown (Fig. 3), a high peak
concentration of hydrogen occurs in the cladding at the
- Defect Growth during Operation - interface with the CrMo steel. The effect of a
variation of the cooldown rate on the peak hydrogen
The acceleration of hydrogen attack under mechanical concentration is shown in Fig. 6. The difference in
load and the interaction of hydrogen with the creep peak hydrogen concentration is only small and is
deformation process have been mentioned above. In the unlikely to be the sole explanation for the large
case that defects are present in the reactor wall, the difference in disbanding area observed in the tests.
interactions with hydrogen become even more important Other factors like differential thermal expansion may
in the areas of high stress and deformation at the tips have to be considered.
of the defects.
Further API/MPC research should clarify the true cause
The growth of defects by creep has been modelled with of disbanding and the significance of the experimental
the damage mechanics approach, where the growth rate of evidence of disbanding for high temperature, heavy wall
crack-like defects is predicted as a function of the hydrogen reactors.
crack tip stress field which is characterised by the
creep parameter C* (Ref. 9). C* is a function of - Defect growth in the CrMo steel during shutdown -
applied stresses, defect size, geometry and of the
uniaxial stress/creep strain rate relationship of the API's Document 946 (Ref. 2) states that 3 ppm of
material. Hydrogen is now known to reduce the time to dissolved hydrogen is required at ambient temperature
rupture in a creep test, but little is known about the to cause hydrogen-assisted growth of defects in high
effect of hydrogen on the creep rupture strain and the strength 2.25Cr-1Mo steel. Coincidently, this figure of
creep strain rate. 3 ppm is in line with experience of steel fabricators
who have to limit the dissolved hydrogen in their slabs
Further API/MPC research is needed to establish the in order to avoid flaking during cooldown. Other forms
effect of hydrogen on the creep rupture strain and of hydrogen problems may occur at low temperature at
strain rate in CrMo steels and to validate the use of different hydrogen concentrations; for example hydrogen
C for the assessment of defects in a reactor for high induced cracking (HIC) may occur in carbon steels at
temperature hydrogen service. hydrogen concentrations well below 1 ppm, whereas SMAW
weld metals may contain hydrogen concentrations well
above 3 ppm without any cracking problems. Similarly,
WRC Bulletin 534

consideration that diffusion rates below 400 'F (-200
• •
[H] in ppm [H] in ppm 'C) are low and that an incubation period is needed to
reach the equilibrium pressure inside a defect.
Calculation of the hydrogen concentration in the steel
: 6 which is needed at 200 'C to attain an internal
300
i pressure equal to the yield strength (i.e 4500 bar)
using the solubility data in the appendix gives a value
l
of 2.8 ppm. More realistic assumptions of a lower
hydrogen solubility and of significant diffusion rates
200 4 down to room temperature would result in a critical
hydrogen concentration well below 1 ppm for extension
:r----:;;.--------=
:I
of defects, but this does not seem realistic.
In our view the Geiger's model underestimates the

critical hydrogen concentration in the steel for the
100 2
following reasons:
I
.,.....----.u:u······················································ ..• a) Relationship (1) is only valid in the engineering
range of hydrogen pressures. For very high pressures
it may be replaced by:
0 ..._~--~--...---+-------.-------.------..------0
6 4 2 0 10 20 30 C - A.Jp.e-(B-vp)/T (2)
cladding CrMo-steel distance
..,..._-------+I--------1•.,.. in mm with v - 0.071 K/bar (Ref. 13).
a) prior to cooldown ( 450°C, 150 bar) b) The assumption that above a certain temperature
sufficient hydrogen can diffuse into the defect to
after cooldown (after 50 min) build up the equilibrium pressure is an
after 200 hours over-simplification. Proper diffusion modelling is
required, so that factors like the size of the
defect, the finite amount of hydrogen available and
~[\
O>Q.
~.s 300
the temperature history can be taken into account.
Consideration should also be given to trapping of
part of the available hydrogen during cooldown. The
"O G>
0 (.) trapping ratio and the distribution and the size of
o.g the defects in the steel, the welds and the cladding
Q) Q) are strongly dependent on the type and the quality
200
:S£ of the material. Therefore, the tolerable hydrogen
.s Q)
,_...r::.
concentration during cooldown is expected to be
:I: .... strongly dependent on the material qualities as
...... 100
well.
T(°C) pH 2(bar) dT/dt(°C/hr)
c) The criterion that the internal pressure should
1: 450 150 675
equal the yield strength of the steel is
2: 450 150 150
inappropriate for defect growth. It should be
b) 0
0 50 100 150 200
replaced by a criterion based on fracture mechanics
principles. In a fracture mechanics analysis, the
time in hrs ... crack driving force which results from the internal
pressure inside the defect is compared with the
Fig.6. Degassing of a Disbonding Specimen fracture toughness of the surrounding steel. The
a) Hydrogen Concentration Profiles fracture toughness used in the analysis should
b) Hydrogen Concentration in the Cladding at account for any temper embrittlement and hydrogen
the interface during and after cooldown embrittlement of the steel. The analysis can be
extended to assess the risk of initiation of
Hydro9en solubility: CrMo steel Cladding unstable fracture when the vessel is under pressure
A (ppm/{ bar): 23.6 7.6 and the crack driving force resulting from the
B (K): 3257 630 membrane stresses is superimposed on the above crack
driving force resulting from hydrogen pressure in
the critical hydrogen concentration in CrMo steels for the defect.
hydrogen damage during shutdown is expected to be
largely dependent on the quality of the steel. In the API/MPC task group, a considerable amount of
hydrogen and temper embrittlement data have been
A well-known attempt to model the growth of defects as presented. An effort should be made to model the
a result of hydrogen pressure build-up inside the extension of defects during and after cooldown and to
defect in a steel which is over-saturated with hydrogen apply the generated data to derive guidelines for the
has been made by Geiger (Ref. 12). He postulates that avoidance of hydrogen damage during shutdown in terms
the pressure inside the defect can be calculated using of maximum hydrogen concentration during operation and
relationship (1) of the appendix: cooldown procedures.
2
p - (C/A) .exp(2B/T) (1)
For the symbols see the appendix. Flaking will occur

once the internal pressure exceeds the yield strength
of the steel. Implicit in his paper is the
WRC Bulletin 534

9) Prediction of Creep Crack Growth from Uniaxial Creep

REVIEW OF THE RECOMMENDATIONS IN THIS PAPER FOR FUTURE Data.
RESEARCH Nikbin, K.M., Smith, C.F. and Webster, G.A.
Proc. R. Soc. London A 396, 183-197 (1984)
The following directions for research have been 10) Hydrogen Induced Disbanding of Stainless Steel
suggested:
Overlay Weld.
- Assessment of the solubility and diffusivity of Ohnishi, K., Chiba, R., Watanabe, J., Naito, K., and
hydrogen in CrMo steels as a function of steel Okada, H.
chemistry and microstructure. The kinetics of
hydrogen trapping which reduces the diffusivity of 11) Disbonding Mechanism and its Prevention.
the hydrogen must be established. With lower Palengat, R., Bocquet, P., and Vignes, A.
priority, the solubility and diffusivity of hydrogen to be presented at the International Conference on
in the cladding materials should also be assessed. Interaction of Steels with Hydrogen
- Modelling and experimental verification of the Paris, March 1989.
effect of hydrogen on time to creep rupture and the 12) The relation between shutdown procedures and
synergism of hydrogen attack and creep. possible hydrogen induced flaking in thick-wall
- Assessment of the effect of hydrogen on the creep hydrogenation reactors.
rupture strain and the strain rate in CrMo steels, Angeles, O.F., and Geiger, C.H.
and validation of the use of the parameter "C*" in 35th midyear API meeting, Houston; May 1970.
the assessment of defects in reactors for high
temperature hydrogen service. 13) Von Ellerbrock, G., Vibrans, H. and Stuwe, H.P.
- Establishment of the mechanism of disbanding of Acta Metall. 20, No.l (1972) 53
cladded CrMo steel (incl the vanadium modified CrMo
steels) and assessment of the significance of the
experimental evidence of disbanding for high
temperature hydrogen reactors.
- Modelling of the growth of defects during and after
cooldown as a result of hydrogen pressure build-up
inside and hydrogen embrittlement around the defect.
Derivation of a safe hydrogen concentration limit
during shutdown and safe pressure limits for
reactors at low temperature.
REFERENCES
1) API Publication 941
Steels for Hydrogen Service at Elevated Temperatures
and Pressures in Petroleum Refineries and
Petrochemical Plants.
Third edition, May 1983.
2) API publication 946
The effect of outgassing cycles on the hydrogen
content in petrochemical-reactor-vessel steels
July 1981.
3) MPC/API task group on materials for pressure vessel
pressures.
HPV-43: Minutes of meeting of May 11 and 12, 1988.
4) MPC/API task group on materials for pressure vessel
pressures.
HPV-41: Minutes of meeting of October 7 and 8, 1988.
5) Hydrogen distribution through pressure vessel wall.
Sakai,T. Asami K. and Takada H.
paper presented at the Pressure Vessel Research
Committee Meeting, Scottsdale, Jan 1981
6) The solubility of hydrogen at low pressure in iron,
nickel and certain steels at 400 to 600 °C.
Armbruster, M.H.
J. of Am. Chem. Soc. 65 (1943) 1043-1054
7) Influence of Alloy Additions on Hydrogen Diffusion
in Iron and Contribution to the System Iron
Hydrogen.
Geller, W. and Sun, T.
Arch. Eisenhuttenw., 21, 423-430 (1950)
8) Der Einfluss von Wasserstoff auf die
Gebrauchseigenschaften von unlegierten und
niedriglegierten Stahlen.
Haumann, W. e.a.
Stahl und Eisen 107 (1987), No. 12, 585-594
WRC Bulletin 534

RESISTANCE TO HYDROGEN INDUCED DISBONDING OF HOT ROLLED HEAVY

GAUGE CLAD STEEL FOR APPLICATION IN PETROLEUM INDUSTRY PRESSURE
VESSELS
Hideto KIMURA, Minoru SUW A, Sadahiro YAMAMOTO

Materials and Processing Research Center
NKK Corporaion
1 Kokan-cho, Fukuyama 721
Japan
ABSTRACT Sufficient disbonding resistance was also found after

welding md head plate forming.
The advantage of the use of hot rolled clad steel as the
replacement of conventional overlay welded clad steel in INTRODUCTION
hydrogen reactor vessels was discussed focusing on the
resistance to hydrogen induced disbonding. Petroleum refining and other petroleum industrial plants
Heavy gauge clad plate of y-stainless steel on Cr-Mo include hydrogen reactor vessels which are exposed to
steel was thermally hydrogen charged in autoclave, and the 200-600°C under hydrogen partial pressure of 1-60 MPa.
resistance lo disbonding was evaluated with varying The requirement in vessel materials both for high
conditions of cladding and hydrogen charging, mainly in temperature strength and corrosion resistance has typically
terms of disbonded area fraction at the interface. been satisfied by heat resistant Cr-Mo steel overlaid with
The conventional overlaid clad plate was also prepared, austenitic stainless steel. However, the overlaid plates
tested and compared with the clad steel manufactured by occasionally have experienced disbonding of the stainless
hot rolling. The result indicated a superior disbonding steel caused by accumulation of hydrogen at the interface
resistance of rolled clad steel to that of overlaid clad steel between the cladding metal and the base steel, mainly
both in single and in multiple hydrogen charging tests. induced by shutdowns of the plants for inspection and
The high sensitivity to disbonding in overlaid clad steel other purposes[l][2]. (See Figure 1.) Therefore, several
was attributed to the formation of martensite caused by studies have been performed in order to clarify the relation
dilution of weld metal with base steel, and to carbide between the disbonding behaviour and selection of welding
precipitation on grain boundary of cladding weld metal material, welding conditions, chemical composition of
caused by heat migration of carbon from the base steel. base steel, etc.[3][4].
Both of these were suppressed in rolled clad steel and the Although rolled clad plates are considered to possess
advantages of using solid-phase bonding were described. excellent disbonding resistance and economical advantage
Further improvement in disbonding resistance for rolled for producing, they have not usually been used in these
clad steel was achieved by applying higher rolling application, because the mill capacity to obtain the
reduction and/or using Cr-Mo-V steel as the base plate. nesessary large reduction for producing these heavy
WRC Bulletin 534

APPLICATION AT H2 ENVlRONMEN 1
(hydrogen reactor vessels)
400
I I
HYDROGEN INTRUSION IN CLAD STEEL .r.:
Ci 300
c:
....(!)
SHUTDOWN ti
...
C'O
JIS G3601 (;;.200MPa)
(!) 200
.r.:
HYDROGEN ACCUMULATION v;
ASTM A264 (;;.140MPa)
100~~~~~~~~-'-~~~--'
DISBONDING
Initial Reduction
30mm/pass 40mm/pass
{plate thickness : 100mm)
Figure 1 SCHEMATIC FLOW OF HYDROGEN- Figure 2 EFFECT OF INITIAL REDUCTION ON
-INDUCED DISBONDING SHEAR STRENGTH OF CLAD PLATES
gauge clad plates (80-200mm in thickness) has not been conventional overlaid clad steel and the rolled clad steel,
considered satisfactory. However, both the recent progress and, (2)the improvement of joining performance of rolled
in rolling technology and the use of mills with larger clad plate through proper selection for the base material
capacity attained a marked increase in shear strength at the and the insert metal. The commercial-scale production of
bonded interface, which was reasonably applicable to the heavy gauge y-stainless steel I Cr-Mo steel clad plate was
production of heavier-gauge clad plates. The example of also carried out using 9000 ton rolling mill in Fukuyama
the increase in bonding strength of austenitic stainless Works, NKK Corporation, and the properties were
steel clad plate (lOOmm in thickness) is shown in Figure discussed and compared with the conventional overlaid clad
2. The general feature of rolled clad steels and their plate.
production process will be discussed in the following
section. HOT ROLLED CLAD PLATES
The present study was performed to obtain the basic data
on (l)the difference of disbonding behaviour between the The typical process for producing overlay welded and hot
OVERLAID
~ T " /~~
I
I
overlay welding
\ rolling finishing
I
ROLLED
~ -- ~
slab composing rolling separationg
Figure 3 PRODUCTION OF ROLLED AND OVERLAY WELDED CLAD STEELS
WRC Bulletin 534

Table l COMPARISON OF ROLLED CLAD STEEL WITH OVERLAY WELDED CLAD STEEL
mass bonding variety in met- resistance to dimension

productivity strength al selection H2 at interface enlargement
ROLLED good fairly good fairly good fairly good easy
(improved*) good good good good easy

OVERLAID worse good good worse
I fairly easy
*Clad plate production carried out by :
(l)large-capacity rolling mill (9000 tons max.) with shiftable rolls
(2)slow rolling velocity (100rpm--20rpm)
rolled clad plates was shown in Figure 3. In overlaying Therefore the production mill of 9000 ton-class was used
process, the banded weldment of cladding metal is in the present study. In the following sections the
performed, typically using submerged arc welding on base application of hot rolling process to production of the clad
steel. The surface finishing process usually follows. On plates for hydrogen environment was discussed.
the other hand, in the rolled clad production, the cladding
plates are composed with base steels by welding the EXPERIMENTAL PROCEDURE
circumference and degassing at the interface before hot
rolling. The insert metal such as Ni foil can be used in The chemical composition of the base steels and
oroer to help joint at the interface. Then the composed slab cladding metal used for the laboratory study is shown in
is heated and hot rolled for the interface bonding using the Table 2. The 21I4Cr- l Mo steels with lower C content and
plate mill. The clad plates used in the present study was a small addition of B+V were used as well as the standard
composed as the sandwich-type slab and the separation of 21; 4Cr-lMo steel. Type 347 austenitic stainless steel
two bonded clad plates was performed after rolling. The (18%Cr-10%Ni-Nb) with plate thickness of 20mm was
surface finishing is carried out at the final stage. always used as the cladding metal.
The advantageous feature of the rolled clad plate lies in The cladding conditions for rolling and overlaying were
the mass-productivity and capability in dimension summerized in Table 3. Slabs sized (200+20)mm x
enlargement of products. Moreover, the diffusion at the 240mm x 480mm were composed in advance of hot
interface related to the completion of the interface bonding rolling of the clads[5]. Ni foil of 100 µm in thickness was
is to be promoted by controlling the rolling parameter used as the insert metal.
relatively with ease, especially in the case a heavy-load The slabs were then heated at l 200°C and rolled to
reduction is available at rolling process. 55mm in thickness. The rolled plates were normalized at
Table 2 CHEMICAL COMPOSITION OF BASE STEEL AND CLADDING METAL
c Si Mn p s Ni Cr Mo Nb v Ti B sol.Al total N
Std. 0.15 0.11 0.55 0.006 0.00< 0.03 2.43 0.96 - - - - 0.014 0.0044
Base metal 0.20V 0.08 0.19 0.55 0.004 0.001 0.50 2.42 1.04 - 0.20 0.007 0.0008 0.017 0.0033
0.34V 0.11 0.06 0.54 0.003 0.001 0.51 2.42 0.97 0.02 0.34 0.006 0.0009 0.016 0.0037
Claddinl!. meta tvoe347 0.04E 0.62 1.65 0.025 0.008 10.86 17.49 0.17 0.77 0.03 - - 0.005 0.0021
WRC Bulletin 534

Table 3 CLADDING CONDITION IN LABORATORY
ROLLED hot rolling 1200 °C reheat (200+ 20)-+(50+ 5)mmt
heat treatment 900 'C X 30min AC+ 720 'C X 30min AC
+ 690°C X 24h FC
OVERLAID welding materials 1st pass PFB-1/USB-309L
2nd pass PFB-1 FK/USB-347L
welding condition DC foop + 800A, 26V, l 90mm/min
heat treatment 690 °C X 24h FC
900°C x 30 min., then tempered at 720°C x 30 min. overlaid clad plates. After reduction to 50mm in thickness,
followed by stress-relieving at 690°C x 24 hours. two layers of weld metal, type 309L and 347L as the first
Steel slabs normalized at 900°C x 30 min. and tempered and the second layer respectively, were overlaid using
at 720°C x 30 min. were used as the base steel of the submerged arc welding. Stress relieving heat treatment of
690°C x 24 hours followed the overlaying.
The plate specimen of (46+4)mm x 50mm x lOOmm in
dimension was taken from each clad plate and used for
hydrogen charging test. After verifying the bonding at the
150 interface using ultra-sonic testing(U.T.), they were
hydrogen charged in autoclave at 10-20MPa, 400-450°C
E
a. for 24hours. Then the plates were cooled to room
a.
c 100 temperature at the cooling rate 3.5-4.0°C I min., and kept
Q)
for 1 week before inspecting the disbonding. The scanned
c0
(.) data was picture-processed to measure disbonded area ratio.
c
Q) ti <t2 <h<t4
0
O> 50 RESULT AND DISCUSSION
.....
-0
>.
I
10 Disbonding condition
4 Hydrogen concentration in the clad steel of hydrogen
3 reactor vessels during operation and after shutdown was
2
I calculated evaluating the diffusivity of hydrogen in
05 0 50 100 austenitic stainless steel and Cr-Mo steel. The result was
Clodding Bose metal shown schematically in Figure 4. Hydrogen atoms entered
Distance from interface (rrrn) the clad steel at the inner surface of the vessel wall and
Inner surface Outer surface
diffused through the wan thickness to the outer, showing
the linear profile during the operation, but gathered
Figure 4 SCHEMATIC PROFILE OF HYDROGEN
preferentially at the interface showing the curved
DISTRIBUTION IN THE CLAD STEEL SIMULATING
distribution described in Figure 4 after shutdown [4]. The
THE HIGH PRESSURE REACTOR VESSEL AFTER
high concentration of hydrogen evaluated at the cladding I
SHUTDOWN[4]
base metal interface was due to a large difference in both
WRC Bulletin 534

JO Studies of Cladding and Overlay for Pressure Vessel Service
solubility and diffusion coefficient between Cr-Mo steel

and stainless steel. This accumulation was considered to Overlay weided
clod plate
have caused the disbanding.
The overlay welded clad was hydrogen charged in
autoclave and the condition was varied to reproduce the
disbanding of the clad steel in actual reactor vessels. U.T.
was used to detect the disbanding. Figure 5 shows the § 20
·r;
typical scanned ultra-sonic mapping of disbonded portion
at the clad interface. The specimen was overlaid clad plate
0
tt • Rolled
o clad plate
which was exposed to 20MPa hydrogen at 450°C for 24 0'--~r=-~~'--~~~~-'-~~
hours and cooled to room temperature at 4°C I min. U.T. 200 400 600
Max. hydrogen content at the
was carried out 5 days after completing the hydrogen
interface (ppm, calculated vo lue)
charging. By optical and scanning microscopy the crack
was observed in the cladding approximately 50 µm distant Figure 6 RELATIONSHIP BETWEEN CALCULATED
from the interface (Photo 1). The crack propagation was MAXIMUM HYDROGEN CONTENT AT THE
recognized on grain boundary and the feature of the crack INTERFACE AND MEASURED FRACTION OF
was similar to what was reported in disbonding of clad DISBONDED AREA
plates in high pressure hydrogen reactor plants[6].
According to these observation the hydrogen autoclave was using difference method. The calculation was carried out
verified to reproduce the interface disbonding of clad plates on the assumption that the distribution of hydrogen was
caused by shutdown of the reactors. proportional to the saturated solubility of hydrogen in the
Temperature and pressure dependence of hydrogen caldding and the base metal. Three combinations of
accumulation at the interface was simulated on the basis of temperature and pressure parameters, 450°C-20MPa,
Fick's second law, 450°C-1 OMPa and 400°C- l OMPa were assumed, and in
(1) each case the concentration showed the maximum value at
approximately 100 hours after cooling. Only this value is
Length (mm)
expressed in Figure 6. A linear relationship between the
0 110
O-t-~~~~~~~~~~~~~~~--1 calculated maximum hydrogen concentration at the
interface and fraction of disbonded area measured using
U.T. is shown both for overlaid and rolled clad plate.
E
E
~ <~~JY.. ~
Improvement of disbonding resistance in rolled
dad steel
60 The disbanding behavior of rolled and overlaid clad
Horizontal plates were tested and compared. Two hydrogen charging
conditions, 400°C-lOMPaH2 and 450°C-20MPaH2, were
E applied to specimen from both types of clad pates. The
5..c measured disbonded area fraction is shown in Figure 7. In
a.
Q) the case of lower temperature and lower hydrogen pressure
0
(400°C-10MPaH2), the rolled clad steel did not show
Figure 5 THE MAPPED ULTRA-SONIC RESULT disbanding, when both the standard Cr-Mo steel and
OF DISBONDING AFTER THERMAL HYDROGEN Cr-Mo-0.34% V were used as the base steel. The overlaid
CHARGING clad steel showed larger disbanded area fraction than any of
WRC Bulletin 534

Photo 1 MICROSTRUCTURE AT THE INTERFACE OF THE OVERLAY WELDED

CLAD STEEL AFTER HYDROGEN CHARGING OF 450°C, 20MPaH2.
Photo 2 MICROSTRUCTURE AT THE INTERFACE OF THE ROLLED CLAD STEEL

AFTER HYDROGEN CHARGING OF 450°C, 20MPaH2.
WRC Bulletin 534

50 hydrogen-trapping ability of VC I matrix interface

~
~
4 O'C relieving the concentration of hydrogen during cooling at
20MPa
Ill I
(!)
.... 40 I the interface of cladding and the base metal.
I
co I
I
"O I
(!) I
"O
c:
30 \
\
Comparison of rolled and overlaid clad steel
0 I
.c
r/I
I
I
Microstructure at the interface of the rolled clad steel
I
:0 20 \
I after hydrogen charging was examined. In condition 400°C
0 I
I
\ -lOMPaH2, there was no crack in any of rolled clad steel.
c: Q \
.2 10 4o~·c \
\ In condition 450°C-20MPaH2, the clad standard Cr-Mo
t) lOMPa
~ ' steel had the crack in Ni insert metal along the interface,
LI..
0
Std. Std. 0.2V 0.34V while no crack was observed in the clad steel with V-added
Cr-Mo Cr-Mo
base steel (Photo. 2). According to these observation, the
Overlay Rolled
welded disbanding behavior of rolled clad steel was essencially
different from that of overlaid clad steel; the disbanding
Figure 7 EFFECT OF CHEMICAL COMPOSITION occurred in the insert metal or at the interface in rolled clad
OF BASE METAL ON DISBONDING RESISTANCE steel, indicating the possibility of improvement in
disbonding resistance by completing the interface bonding.
rolled clad steels. On the other hand disbanding in overlaid clad steel
In the case of heavier charging condition( 450°C- occurred as grain boundary fracture in the cladding metal
20MPaH2), the largest disbanded area, 5%, was observed near the interface as was shown in Photo 1.
in the clad plate with the standard Cr-Mo steel . In the Figure 6 compares the relationship between the
overlaid clad steel about 50% of the interface area disbanded area fraction measured using U.T. and the
disbonded. calculated maximum hydrogen concentration for rolled and
From these results the superior disbanding resistance of overlaid clad steel. The disbanded area was less in rolled
rolled clad steels to the weld overlaid clad steel, prepared clad steel than in overlaid clad steel, indicating better
by condition given in Table 3, was shown in the described disbanding resistance of the rolled joining at clad interface.
hydrogen atomosphere, and the use of Cr-Mo-V steel as This result also indicated that a different disbanding
the base metal was also shown to be recommendable to mechanism was dominant in each clad steel, which would
suppress interface disbanding. However, the use of rolled be discussed as follows: In overlaid clad steel, through the
clad was enough to obtain a marked improvement in joining process achieved by solidification of liquid phase,
disbonding resistance, even if the standard Cr-Mo steel was dilution of Cr and Ni in the cladding possibly led to cause
used as the base metal. martensitic phase formation in deposited metal, and
The obtained data indicated that the composition of the diffusion of C from Cr-Mo steel to the stainless steel
base steel affected disbanding resistance. V-containing cladding induced the grain boundary precipitation of carbide
steels exhibited better resistance, and this is probably due in cladding metal. On the other hand, in rolled clad steel,
to the distributed gathering site for hydrogen given by VC the change in the chemical composition mentioned above
precipitation in matrix reducing the concentration of was negligible by the use of solid-phase bonding, thus
hydrogen accumulation at the interface. Sakai et al.[7] the joint interface probably took minimum shear strength
reported, concerning the hydrogen embrittlement of Cr-Mo to disbond. In the case where the joint strength was large
steel, that the trapping amount of hydrogen at VC I matrix compared with the shear stress of the cladding or the base
interface increased with lowering temperature, and the metals, the "weakest" part would be the Ni insert layer
better disbanding resistance of clad steel with V addition in where disbonding took place as the result of stress
the base metal would also be explained by the benefit of concentration which is shown in Photo 2 for the
WRC Bulletin 534

Table 4 COMPARISON OF CRACKING SENSITI- -lONi-Nb) was used for plates B and C. The standard
VITY IN OVERLAID AND ROLLED CLAD STEEL 2l/4Cr-lMo steel was used as the base metal for plates A
FACTORS AFFECTING CRACKING SENSITIVITY and B, while V-added 2 l; 4Cr- l Mo-0.2V steel was for plate
H2 accumulation site Harri phase formation C. The finish thickness of plates A, B, and C was
Overlaid C diffusion from Dilution of Cr+Ni in (143+5)mm, (75+4)mm, (100+4)mm respectively. After
Clad Steel base steel-->grain cladding rolling, dehydrogenation heat treatment was carried out,
boundary carbide -->martensitic phase and normalizing at 930°C, tempering at 730°C, and stress
Rolled Joint completing Solid-state bonding relieving at 690°C followed. Single Ni insert was used to
Clad Steel -->no microvoid -.;..less martensitic help bonding. In order to know the influence of thickness
at interface phase of the insert metal, l 00 µm-thick Ni foil was inserted as
Ni insertion single layer in the top-half of the slab, and two sheets of
-->suppressed C dif- the same foil were inserted in the bottom-half. For plate B
fusion and C, Ni foil insert of single layer was used.
-->less extra carbide The mechanical properties of the base metals of these
in cladding clad plates are shown in Table 6. The properties were
examined on the specimen taken at 500mm from the top
of the plate, at the center of the width and thickness. The
condition 450°C-20MPaH2. The comparison of overlaid
tensile tests were carried out for rolling(L), crossing(C),
and rolled clad steel on factors affecting cracking
and normal(Z) direction for plate A while specimen were
sensitivity is summarized in Table 4.
taken and tested only for crossing(C) direction for plate B
and C. The result contented ASTM A387 Grade 22, Class
PROPERTIES OF LARGE-SCALE
2 specification. The impact test was also performed for the
PRODUCED CLAD STEEL
specimen from the same position of plates, and the
transition temperature for each plate was measured to be
Procedure and mechanical properties
Based on the property evaluation for laboratory- lower than -20°C, as is shown in Table 6. Shear stress at
the cladding I base interface was larger than 280MPa for
-manufactured clad steels, three heavy gauge
both single- and double-insert portion of the clad plate.
disbonding-resistant clad plates were produced in trial in
For plate B and C shear stress was larger than 350MPa.
actual plate mills. The chemical composition of the
cladding and the base metals are shown as plates A, B, and
Effect of V addition in the base steel and
C in Table 5. The cladding metal of type 321(
microvoid at the interface
18Cr-10Ni-Ti) was used for plate A, and type 347(18Cr-
The resistance to disbonding of the trial product A is
Table 5 CHEMICAL COMPOSITION OF THE CLAD STEELS HOT ROLLED IN THE TRIAL MASS
PRODUCTION (MASS%)
c Si Mn p s Ni Cr Mo Nb v Ti B sol.Al total N
A 0.14 0.16 0.57 0.006 0.002 - 2.46 1.00 - - 0.008 - 0.018 0.0036
Base metal B 0.13 0.16 0.57 0.006 0.002 - 2.44 l.02 - - 0.012 - 0.010 0.0028
c 0.08 0.15 0.53 0.006 0.002 0.47 2.34 0.97 - 0.19 0.012 0.001( 0.013 0.0033
Cladding metal A 0.041 0.55 l.61 0.026 0.002 10.06 18.37 0.25 - - 0.427 - 0.055 0.0164
B, C 0.050 0.60 l.59 0.029 0.001 10.96 17.49 0.20 0.74 - - - 0.005 0.023 l
WRC Bulletin 534

Table 6 MECHANICAL PROPERTIES OF THE BASE STEELS IN THE TRIAL MASS PRODUCTION
Location Di rec ti or Tensile test CXt) •> Impact test (Xt)b> Shear strength
YS(MPa) TS(MPa) El(%) vEs(J)c> vTs ('C)" > (MP a)
L 417 580 31.6 294 -50 -
A Top c 417 579 31.7 280 -48 293
z 417 576 25. 5 225 -29 -
Bottom - - - - - - 281
B Top c 446 592 30. 3 398 -77 364
Bottom c 460 572 31. 4 - - 353
c Top c 485 596 61. 2 340 -76 380
Bottom c 460 573 63. 2 325 -70 398
ASH! A387-22-2 ~310 515--690 ~18 - - -
a) A and B : 9~, GL=36111l11 , C : lH, GL=50mm
b) 2mmV-notch charpy
c) vEs : Upper she! f energy
d) vTs : 50% FATT
shown and compared with overlay welded clad plate in value than the laboratory value under the same condition.
Figure 8. The specimen for hydrogen charging was taken This was attributed to the difference in the soundness at
at the center of width in the thickness ( 46+4)mm, and lhe interface as rolled; the mill-produced plate A showed
hydrogen charged at 400°C- l OMPaH2, and at 450°C microvoids of approximately 4% in area fraction while no
-20MPaH2. Although disbonding was not observed at microvoid was observed in the clad plate used in
400°C-lOMPaH 2 for both single- and double-insert laboratory. This indicated that the micro voids also gave
portion, the trial product showed disbanding of 20-30% in the accumulation site for hydrogen and the lower reduction
area fraction at 450°C-20MPaH2, which was a larger applied in producing clad plate A led to cause a small
amount of microvoid resulting in a high sensitivity to
disbonding. However, still the disbonded area fraction in
plate A was approximately a half of that in the overlaid
~
~
50 l/2W 0 400°C, IOMP0H2
clad plate at the condition 450°C-20MPaH2, which was
0
<ll
6. 450°C, 20MP0H2
0 40 shown in Figure 9. The relationship between the fraction
-0 of microvoid at the interface and the fraction of disbonded
<ll Top
-0
c 30 area after hydrogen charging is shown in Figure 10.
0
.0
Vl /:::,.
'6 20 Hydrogen charging was carried out also on plate B and
'+-
0 C, and the detected disbanded area fraction is shown in
c 10
.Q Figure 9. Plate B and C did not show dis bonding after
t5 hydrogen charging at 400°C-lOMPa. At 450°C-20MPaH2,
0
.... 0 0 0 0 0
LL
C D Overlay welded a small area fraction (2%) of disbonding was observed in
A B
clod plate plate B, although no disbonding was observed in plate C
Single insert Double inserts
which was with V-added base steel. V addition to the base
Figure 8 RESISTANCE TO DISBONDING IN THE steel also improved resistance to disbonding in the large
TRIAL PRODUCT (steel A) scale trial production.
WRC Bulletin 534

The difference in disbonding resistance between plate A

~
50
I0 1:::,.
400"C, IOMP0H2
450"C,20MP0H2
I and B was considered to be also due to the difference in the
0
~ 40 amount of microvoid at the interface. The area fraction of
0
'O
Cl)
microvoid is one of the best parameters to evaluate the
'O
c 30 joint achievement, and, as is shown in Figure l 0 there
0
..0
(/)
'5 was a relationship between microvoid fraction and fraction

20
0 of disbanded area after hydrogen charging. The charging
c
0
·5 10 • condition was 450°C-20MPaH 2 , and results on the
it 0 0 tJ Ch Q,. Ch
laboratory and factory-produced clad plates A and B with
Ni insert and V-free base steel were summarized. The
A B c head weld Overlay
disbanded area fraction increased with increasing area
plate welded
clod plate fraction of microvoid at the interface before hydrogen
c
Figure 9 RESISTANCE TO DISBONDING IN THE charging.
TRIAL PRODUCTS In summary the superior resistance to disbanding of the
rolled clad plate to overlaid clad plate was recognized in the
I
trial product as well, and both of the completed interface
~ 450°C, 20MP0H2 joint and V addition in the base steel were shown to
50
0 improve the resistance to disbanding.
~
0 40
A
i------o
'O
Cl)
'O 30
Effect of secondary working and welding
c
In order to obtain data on the disbanding resistance after
/~
0
..0
.!!!
lJ
20 typical secondary working, plate C was hot pressed into
0 10 head plate. The shape of the head plate was half ellipsoid
c
.Q •-Laboratory test
u of 600mm in inner diameter, and (46+4)mm specimen was
0 6.-B
it 0 2 4 6 8
taken from the flange where the material was applied the
largest deformation, for measuring the disbanded fraction
Fraction of microboid at interface(%)
of area using U.T. The effect of welding on disbonding
Figure 10 RELATIONSHIP BETWEEN FRACTION resis.tance was evaluated using simulated weld joint
OF MICROVOID AT THE INTERFACE AND hydrogen charged in autocrave as well. The plate supplied
FRACTION OF DISBONDED AREA AFTER for making the weld joint was also taken from the flange
HYDROGEN CHA~R.GING 4 1.5 portion of t.'1e head plate, and type 309L and 347L were
l::lad met:i-1 ~ 22 • ~ "'' ! j ____ j____ used as welding materials for the first layer and the second
i 'N lf:u __ ------T=1=f ===

I third layer respectively. TIG welding was used with heat
input 53kJ/cm. The welding condition for preparing the
Base metal ! I simulation joint is summarized in Figure 11.
1 i The result is shown in Figure 9. After hydrogen
: : charging at 450°C-20MPa, disbanding was not recognized
t. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .J in the flange of the head plate .before and after the
process: TIG
heat input: 200Ax20V,45mrn/rnin simulated welding, indicating the satisfactory resistance of
welding materials : the head plate to disbonding.
1st pass T-309L
2nd/3rd pass T-347L
Figure 11 WELDING CONDITION OF SIMU- Repeated hydrogen charging
LATION JOINT In order to trace the thermal history of hydrogen reactor
WRC Bulletin 534

CONCLUSION
5.0 o B
~
~ D. C head plate Heavy gauge, y-stainless-steel clad plate of Cr-Mo steel
0
Q)
...... 4.0 • Overlay 1Nelded clod plate by hot rolling was thermally hydrogen charged in an
0
"'O
Q)
"'O
.--..---· autoclave, and the resistance to the interface disbonding
was evaluated and compared with conventional overlaid
c
.8V1
3.0
.---- clad plate mainly in terms of disbonded area fraction at the
/6-06~
-
'6
0
c
2.0 interface. The result indicated superior resistance to
disbonding of rolled clad plate to that of overlaid clad plate
·n
0
0
I .0 both in single and in multiple hydrogen charging. The
'-
lJ._ formation of martensite phase caused by dilution of
0 cladding weld metal in base steel, and the carbide
2 3 4 5 precipitated at the grain boundary in weld metal by
Number of hydrogen charge diffusion of carbon from the base steel, were considered to
be the main factors which led to disbonding in overlaid
Figure 12 RESISTANCE TO DISBONDING UNDER clad steel. Both of these were suppressed in rolled clad
REPEATED HYDROGEN CHARGING steel showing the superior resistance to disbonding.
Further improvement in resistance to disbonding in
vessels and examine the resistance to disbonding properly, rolled clad steel was achieved by applying higher rolling
the evaluation after applying the repeated hydrogen reduction and/or using Cr-Mo-V steel as the base plate.
charging to clad steels is important, since the hydrogen Sufficient resistance was practically verified as received and
accumulation at the interface during repeated shutdowns after welding I head plate forming on the trial production
most possibly accelerates the disbonding. Specimens used of rolled clad plate.
for testing disbonding resistance after repeated hydrogen
charging were plate B, plate Cat the head plate flange, and
the overlaid clad plate. The cycle consisted of 48h heating REFERENCES
at 455°C-8.5MPaH2, cooling at 0.75°C/min down to
350°C, and continuing air cooling to the room [!]Naito et al.: Journal High Pressure Inst. Japan,
temperature. The cycle was repeated 5 times with l week's Vol.18, 1980, p.263.
interval between each, measuring the disbonded area by [2]Kikuta et al.: Quart. Journal Japan Welding Soc.,
U.T. at the end of each cycle. Vol. 2, 1984, p.463.
The result is shown in Figure 12. For the overlaid clad [3]Kikuta et al.: Quart. Journal Japan Welding Soc.,
plate disbonding was recognized after the first cycle, area Vol. 2, 1984, p.470.
fraction was increased to more than 2.5% after the second [4]Imanaka et al.: Kawasaki Steel Giho, Vol.17, 1985,
cycle, and further continuous enlargement of disbonded p.84.
area followed during the repetition. For plate B disbonding [5]Kawanami : Journal Japan Soc. Technol. Plasticity,
of less than 0.2% by area fraction was observed after 4 Vol.32, 1991,p.3.
cycles, and for the head plate of plate C no disbonding was [6]Naito et al.: Journal High Pressure Inst. Japan,
observed after 5 cycles. Vol.18, 1980, p.271.
From these results, the superior resistance to dis- [?]Sakai et al.: Journal Iron and Steel Inst. Japan,
bonding of the rolled clad plates to the overlaid clad plate Vol.72, 1986, p.1375.
was practically verified also in repeated hydrogen charging.
WRC Bulletin 534

WRC Bulletin 534

CORROSION RESISTANT CLAD PLATES FOR

HIGH PRESSURE VESSELS HYDROGEN AND
OFF-SHORE ENVIRONMENT
F. Dupoiron, L. Coudreuse, D. Jobard and Charles

Creusot-Loire lndustrie
Le Creusot, France
or gas and oil treatments.

The purpose of this paper is to present the new
availlable clad plates in terms of properties
SUMMARY and behaviour in hydrogen and off-shore me-
dia .New clad plates have been particularly de-
Stainless steel cladding ares very effective in veloped for new fields and exploitation condi-
reducing the hydrogen activity in ca:bon b~c tions:
king steel due to its physical properue~ and. its - High strength quenched and tempered base
corrosion resistance . Clad plates with high material allowing thickness and weight reduc-
strength base material and high corrosion re-
sistant cladding have been developped for off- tioo; .
- High corrosion resistant claddmg : URB6
shore and petrochemical applications . The (UNS N08904) and 6% Molybdenum. super-
performances of thes~ materials are pre~ented austenitic URB26 (UNS 08926) pamcularly
including their corros10n performances m off- developped for highly aggressive fluids .
shore and hydrogen media :Sevral industi:ial After a brief presentation of the clad plate and
experiences are presented m representative their behaviour in hydrogen environment • the
media following section presents the mechanical and
corrosion properties of these materials and the
INTRODUCTION references uses .
Oil and gas , petrochemical and ~hemical in- CLAD PLATES PROCESSES AND PRO~
dustries used widely corrosion resistant alloys PERTIES
behaviour agressive .
High pressure conditionsand the need to use of A clad plate is a composite plate
high alloyed material many cases to formly and integrally bonding a layer of a cor-
use clad materials . Recently new process rosion resistant metal to a carbon or low
conditions and the desire of weigtht and cost steel backing plate also called base material .
savings has lead producers and. use.rs to deye- years clad plates (even .
lop the production and the fabncauon of high plates) or solid plates have been extensively
strength and/or corrosion resistant clad used. Among them the classical 18-10
steels , 3
WRC Bulletin 534

321) , the ferritic stainless steels • but also Typical results of shear and ram tests are also
more sophisticated alloys, high corrosion re- presnted on Figure 1. For the shear test , the
sistant products including duplex and super minimum ASTM 263-5 recommendation is
duplex stainless steels , super austenitic 140MPa . Both explosively bonded clads and
stainless steels • nickel base or copper alloys or hot rolled bonded clads present generally shear
more noble metal as titanium and zirconium . strength higher than an average value of 250
Other combinations are also availlable for spe- MPa.
cific applications [l] .
THEORIDCAL ASPECTS ABOUT
The two components , base metal and clad me- CLADDING AND HYDROGEN EM-
tal • are usually bonded together by explosion BRITTLEMENT
or hot rolling , or a combination explosion
and hot rolling . Table Il gives lists of hydrogen diffusivity
Mechanical properties of the bond ; Ultrasonic coefficients at 20 and 100°C for low alloyed
inspection is always used to check the bond normalized steels and austenitic stainless
quality of clad plates . In addition , to confirm steels. It is shown that the hydrogen diffusion
o-
coefficient is at least 1 5 times lower for aus-
Clad metal Process Plate th. Bend tests results tenitic stainless steels (fee structures). This
means that even with a high activity of
hydrogen on the inner wall of an austenitic
grade base clad comp. Tens. Side
clad plate reactor , only a little quantity of
hydrogen diffuses through the clad metal
UR 825 expl+roll 88 4 no crack no crack no crack towards the base material.
Moreover the time needed to obtain the steady
Monel400 hot roll 13,5 3 no crack no crack nocrad state (equilibrium conditions) through 1 mm
thick wall is of about 30 years at room
URB6 hot roll 25 3 no crack no crack no crack temperature and 115 days at 100°C.The use of
higher thicknesses of clad metal requires much
UR 625 hot roll 11 2 no crack no crack no crack
longer time to obtain the steady state since e
Titanium expl+rol 65 3 nocrad no crack no crad is proportional to square of the thickness (0 =
Table I : Examples of bend tests results T2/D)*. A two mm thick plate results thus in 4
x greater time of steady state.
* T = Thickness
the bond strength several destructive tests are
carried out . Table I present the typical results
Moreover the hydrogen activities resulting
of compression , tension and side bend tests .
from corrosion reaction on the surface of a
dad plate (stainless steel) or a C-Mn steel are
Figure 1 : Ram and Shear test results
MP a To Low alloyed Stainless
600 oc steel steel (304)
400
Diffusion 20 10- 1< D<lO-o tno- 11
coefficient 6
100 5.10- < D<5.10- 5 # s.10- 10
200 2
unit: cm /s
Steady state time 20 10-j< t<2.10-=> t # 10~ (30 years)
0 for a 1 mm thick 100 2.10-2< t<2.10- 3 t #2.10"/)
13 Cr 316L UR B6 UR625 C22 plate unit : s (115 days)
IHJ Ram test ~ Shear

Table II - Hydrogen parameters D ( diffusion
coefficient ) and t ( steady state time ) for low
alloy steel and austenitic stainless steel
WRC Bulletin 534

different since the corrosion rate of the carbon T bm =thickness of the base metal
manganese steel is much higher. Due to the T clad=thickness of the cladding
very low corrosion rates for stainless steels ,
very low hydrogen activity is expected . Then (for example with Sbm = 0.01 cm3 H /cm2/
2
is there risk of hydrogen embrittlement for the
base material ? First of all the hydrogen atm2 , Sclad = 103 Sbm• Tbm = 50 mm and
activity on the inner wall of the clad metal T clad = 6 mm we obtain at room temperature
already explained will be less than that of a ba-
se metal due to the limitation of the corrosion Ra=0,2).
reactions (cathodic part of the phenomena).
Then the hydrogen activity, as explained in For that reason, clad plates provide a safe
figure 2 is lowered through the thickness of solution in the most critical service conditions
the clad metal. At the interface due to the which can produce hydrogen embrittlement of
differences of diffusivity we observe a large <;-Mn base steels. To investigate the condi-
reduction of the hydrogen activity. This drop t1on.s and behaviour of material in hydrogen
of hydrogen activity can be obtained by the envrronment , C.L.I research center has deve-
following relationship : lopped a computerized code to determine the
hydrogen concentration in the pressure vessels
wall according to the thermal cycles . Such co-
de is very effective to determine the desirable
1 coo~ing conditions ( cooling rate , steps ... )
particularly for the petrochemical installations
1 + Sbm . Dlm:L . I..ct00
S clad D clad T bm
CLAD PLATES MATERIAL
with:
RaH =reduction of hydrogen activity at the
'I?e cladding technology included an optimisa-
bond tion of the chemical compositions and heat
Sbm =hydrogen solubility in the base metal treatment of both base metal and clad metal
Sclad=hydrogen solubility in the cladding for each specific application. Sevral high cor-
material rosion and high strength clad plates have been
designed for oil and gas as well as off-shore
Dbm =hydrogen diffusion coefficient in the
applications [2] .
base metal
Dclad=hydrogen diffusion coefficient in the
cladding
Grade Rp0,2 Rm A% Kv (T 0 C)
Th=50mm MPa MP a J
BS 1501&224 >305 490-610 >18 31 (-40°C)
A516 Gr 70 >260 485-620 >21 20 (-40°C)
STE35S >345 490-630 >22 30(-40°C)
SESOO >500 600-750 >18 75 (-40°C)
- Cladding Base SE700 >690 790-940 >16 40(-40°C)

Figure 2 : Hydrogen activity profile through the Table III : Mechamcal properties of base material
wall of a clad plate
WRC Bulletin 534

HIGHER STRENGTH BASE MATE-

RIALS Grades Composition % PREN
The use of high strength base metal helps to

c Cr Ni Mo N otbe l'Cr+3,3Mo+16N
reduce the weight of vessels which is 316L 0,02 17 12 2 om 24,7

extremely important for offshore applications.
UR45N 0,02 22 5,5 3 0,16 34,5
The first method is the use of high C UR52N+ 0,02 25 6,5 3,5 0,23 40,2
equivalent normalized steels like BS 1501 -
224 - 490 and A 516 Gr 70 . In that case, the UR825 0,02 22 41 3 Ti 31,9
use of very clean steels as earlier presented,
results in both high hydrogen embrittlement URB6(N) 0,02 20 25 4,3 0,14 Cu 36,4
resistance behavior and high toughness proper- URB26 0,02 20 25 6,3 0,2 Cu 44,0
ties at low temperature. A second possibility is
the use of microalloying base metal like STE
355 steel or BS 1501-225 grade with improved Table IV : Cladding grades for off-shore applications
cleanliness, similar to the H-resistantgrades. has its own properties and according to service
A third solution, recently developed is the use conditions and corrosivity the level of alloying
of quenched and tempered base metal in nickel, chromium or molybdenum may be
(Superelso 500 and Superelso 700). The me- optimised.
chanical properties are presented on Table III
.This solution combines very high strength
and toughness. A large reduction of wall thick- Duplex family ;
ness may be achieved. This has been recently The need of new cladding material providing
successfully used for several offshore both high corrosion resistance and cost saving
separators with a 904 LN ( UR B6N) alloy , has led to the developement of several new
cladding (see references ) . corrosion resistant clad plates.Nitrogen addi-
tions are obviously beneficial for duplex alloys
since they simultaneously improve the
CORROSION RESISTANT CLADDING structural stability of the alloy and HAZ
MATERIALS (lower increase of ferrite level), the strength of
the austenitic phase and the corrosion
Improved cladding materials properties of the alloy particularly the pitting
and crevice corrosion resistance.
In order to make the forming easier and to ob- This explains why the nitrogen level of the UR
tain for clad plates, even after hot forming and 45N UNS 31803 slowly increased (0.12 ...
post welding treatments, the same corrosion 0.14 ... 0.17) and the developement of the UR
properties as for fully annealed solid plates se- 45N+ with high nitrogen (0,18% and molybde-
veral specific grades have developed , with num (3,3%) . The most spectacular
chemical composition improved in order to improvements nevertheless concern the
obtain higher thermal stability . This effect is highest alloyed superduplex stainless steels.
obtained by controled molybdenum additions The new 25% Cr superduplex alloy contains
and low residual ( carbon , silicon , phospho- now 3.6 Mo , and nitrogen additions higher
rus ) order to stabilize the austenitic phase than 0,22 %. The pitting index (Cr+ 3.3 Mo+
against intermetallic phase transformations 16 N) is now 40 minimum. This results in an
phase which can reduce the corrosion resis- increase of the corrosion resistance to the level
tance main corrosion resistant grades of superaustenitic alloys particularly for
used for clad plates ares presented on Table seawater applications.
IV. Three families are presented :
-the duplex grades, -the austenitic stainless
steels and -the nickel base alloys. Each family
WRC Bulletin 534

Austenitic and super;mstenitic stainless

Figure 3 :Pitting potential in 30 g/I NaCl solution
at Ci0°C
md
As discussed previously due to their austenitic
800 structure , these grades are well adapted to
sour gas an oil and hydrogen uses.
(JOO For oil or gas production and treatment a wide
variety of clean (low sulfur and oxygen) 304,
400 316, 321and347 steels has been delivered.
Among the superaustenitic grades UR B6
200 (UNS 08904) and UR B26 (UNS 08926) have
been developed in clad plates associated with
0 high strength quenched and temper type (Su-
URB6 URB6N perelso 500) .
f;sJ Quenched Tempered
I i Q-T+PWHT Corrosion resistance of Superaustenitic clad
plates
II Quenched +Hot forming+ Quenched+ Temper
f::::::~ Quenched +Hot forming +Hot forming +PWHT Figures 3 to 7 presents several results of cor-
rosion studies carrried out on UR B6 (08904)
, URB6N , and 6 Mo grade URB26 (08926).
In order to simulate fabrication operations and
°c Figure 4 : Critical pitting T°C according heat treatments of the carbon steel base mate-
100
ASTMG48A rials , several types of heat treatment have
80
been simulated .
Corrosion tests in chloride containing media (
60
Soiid iS..:s:'S1 pitting and crevice tests) and general corrosion
Clad~
tests have been carried out . Such tests are also
able to detect structure or local composition
40 changes.
20 UR B6 and B6N : Typical critical pitting po-
0
Figure 6 : UR B26-Critical pitting T°C according
316L UR 825 UR UR UR UR UR ASTM G48A for different states
45N B6 B6N 52N+ B26 625
100
°c Figure 5: Critical crevice T°C according
50
u
40
ASTM G78 - FeCl3 6%
Eso
0
Solid rsss'!
30
Clad~ 0
Solid Clad Plate CP
20
Quenched Q+Temper
10
tentials results determined in 30 g/l NaCl solu-
tion at 60°C are presented on figure 3 and the
CPT in ferric chloride solution according
316L UR 825 UR UR UR UR UR ASTM G48A and B on the figure 4 and 5 .
45N B6 B6N 52N+ B26 625
UR B6N specificaly designed to improve the
WRC Bulletin 534

Figure 7 :UR B26· test according SEP 1877 n for PROJECT APPLICATION GRADES
different state FIELD base clad
Sleipner gas separator SE500 URB6N
Sleipner gas separator SE500 URB6N
East Brae gas separator SE500 316L
East Brae gas separator SE500 316L
Ekofisk heat exchanger SE500 Ti
Embla gas separator SE500 URB6N
Troll gas gas separator SE500 316L
Solid Clad Plate CP
Quenched Q+Temper Froy gas separator SE500 URB6N
N'Kossa separators SE500 316L
phases stability presents an excellent corrosion
behaviour even after several heat treatment si- Table V : SE 500 clad plates Industrial applications
mulating fabrication steps . As shown in figure
3 , the pitting potential at 60°C appears to be
unaffected by several heat treatments which is
not the case for the nitrogen free UNS N08904 UK and Norway sectors) , and Table V pre-
( URB6) less stable after several heat treat- sents the main applications in off-shore . Seve-
ments . No detrimental effects of heat treat- ral projects using Superelso 500 + URB6N
ment are reported on the pitting potential of clad plates have shown the interest of this so-
UR B6N. This grade appears to be better adap- lution to improve the efficiency and reliability
ted to clad plates than standard N08904 . It is of the installation as well as the saving of
also important to point out that the high locali- weight and cost . Clad plates are used in many
sed corrosion resistance of UR B6N , particu- fields like ALWYN FIELD (TOTAL),
larly when it is compared to AISI 316L and Trenganu (PETRONAS - Malaysia), Lybian
Alloy 825 as show for ASTM G48A test on fi- (AGIP), BRAE B. (MARATHON),
gure 3. GULLFAKS A, B,C, OSEBERG 2,
SLEIPNER (STATOIL), FULMAR (SHELL),
SNORRE (SAGA) and BP MILLER (BP).
UR B26 -UNS 08926 - This alloy is a high Several refinery vessels are also built from
corrosion resistant material , particularly well clad plates like vacuum towers (ESSO -
adapted in high chloride media . Several type SHELL), Flexicokers (ESSO), heat
of tests have been carried out after cladding exchangers (ULTRAMAR - CND), Coke
operations , solution annealing and several ty- drums (CONOCO) or hydrotreater
pes of heat treatments in order to simulate (PETRONOR - Spain). Clad plates may also
temper and post weld heat treatments of the be used for carrying oil and gas from the field
carbon steel . The results presented on Figure to the continent in order to be treated on shore.
4 to 7 show the high corrosion properties of
this alloy . The tests in chloride media and
acidic environments show that the high corro- CONCLUSION
sion properties of UR B26 remain at level
equivalent to the solid plates after the heat This paper has presented :
treatments of the base material .
-Some theoritical considerations about
REFERENCES AND APPLICATIONS: cladding and hydrogen embrittlement. It has
been shown that austenitic stainless steel clad-
Figure 8 give typical applications of clad dings are very effective in reducing the
plates for offshore applications in North Sea ( hydrogen activity in the carbon manganese
WRC Bulletin 534

oil and gas, has demonstrate that clad plate

Icu
.-------------.HEIDRU
OFFSHORE REFERENCES! == present one of the better solutions in terms of
reliability , safety and economy , particularly
c::::>
BRENT 0 SNORRE 0 "' for hydrogen environment .
/'10 0 GULFAKSc:::Jf\.
= TROLL,
c::::> 0ALWYN 0 c=f*~
0 OSEJ?ERG ~ ,
c::::i- '-="-' •
0
NORTH WESTERN fl. FRljlY
EUROPEAN Oc::t*
CONTINENTAL SHELF. BRUl:'.E HEIMDAL
Og;;; 0 c::::i REFERENCES :
[l] J.Charles , D.Jobard , F.Dupoiron and
EAST BRAE c=f ci *fl. D.Catelin- NACE 1988-St LOUIS -USA-Pa-
c::::i
0
o
MILLER c::::i=_,= per N09
SCOTT fl. 0 PIPER 0 SLEIPNER c::::i= *
c::::> 0 fl.0SALTIREALBA
0,11. [2]J.Charles - D.Jobard - D.Catelin - 2nd Int.
ARBROATH Conf. on Surf. Technology, TWI , Stratford
Ofl. upon Avon - 1987
0 GAN~ET J.JOANNE DUNBAR
c::::> 0 c5 0 c::::i=*
FULMAR EKOFISK
Ofl!flii0ci*
~EM_B}A
c::::>
c::::> PRESSURE VESSEL /':., TUBES,FLANGES,VALVES ..
= CLAD VESSEL * BASE METAL: SE 500
D 316 D URSBS D UR825 D UR45N D Ti
D UR B6N Q UR 826 !Ill UR 625 D UR 52N+
Figure 8 : Clad plates and s.s references in North Sea
plates due to their physical properties (H.

diffusivity ... ) and also due to a sharp decrease
of the corrosion reactions. Clad plates are
obviously an alternative to high hydrogen
embrittlement resistant material.
-New deyelopments concerning stainless steels

for cladding . It has been particularly pointed
out the importance of structural stability of the
corrosion resisting alloy during the specific
thermal cycles on the corrosion behaviour .
Specific cladding material have been develop-
ped and investigated , they present high corro-
sion properties even after heat treatment
(PWHT).
-New developments concern high strength
quenched tempered steels used has base metal
which provides weight and cost savings.
-Awide variety of industrial applications in oil
and gas production have been presented . Ex-
perience of clad plates , with more than 100
000 t for petrochemical industries , off-shore ,
WRC Bulletin 534

WRC Bulletin 534

EVALUATION OF STAINLESS STEEL EXPLOSION CLAD

FOR HIGH TEMPERATURE, HIGH PRESSURE HYDROGEN SERVICE
J. G. Banker
Explosive Fabricators, Inc.
Louisville, CO
M. S. Cayard
CLI International, Inc.
Houston, TX
ABSTRACT cladding and base metal may be susceptible to

disbanding under certain high temperature hydrogen
Hydrogen disbanding tests were performed on conditions.
specimens of T347 stainless steel explosion clad to
alloy steel, SA387. One test series used a standard
rectangular autoclave specimen. The other used a Under high temperature, high pressure operating
new cylindrical specimen design which more conditions, hydrogen diffuses outward from the
realistically simulates pressure vessel service. One interior of the vessel, reaching a steady state diffusion
specimen was manufactured with an intentionally condition during service. Upon cool-down the
induced nonbond flaw. The 167°C (300°F)/hour hydrogen solubility decreases. Since the hydrogen
cooling rate of these tests simulated extreme refinery diffusion rate is considerably higher in the Cr-Mo
shut down conditions. Ultrasonic and destructive steel than in the stainless steel, the bulk of the
testing demonstrated that dad disbonding did not hydrogen is dissipated by diffusion outward to the
occur and that the intentional nonbond did not grow. external vessel surface.
Two potential mechanisms can result in clad

INTRODUCTION metal disbanding under cool-down conditions. In the
most common mechanism, molecular hydrogen can
Pressure vessels used in high temperature, high a cc um ulate at laminations or bond defects, eventually
pressure hydrogen service, such as hydrotreating and resulting in internal pressures which exceed the bond
hydrocracking reactors, are typically constructed of strength. The defect may then enlarge to reduce
Cr-Mo steels internally clad with stainless steel. The hydrogen pressure. The second mechanism may occur
Cr-Mo steel provides superior high temperature in hard martensitic regions which can be formed
mechanical properties and resists high temperature during clad manufacture by some weld overlay
hydrogen attack. processes. These regions may be susceptible to
accumulation of hydrogen in the martensitic lattice
The stainless steel cladding provides corrosion structure, significantly reducing ductility and
resistance. When properly selected, each of the toughness. This condition can eventually lead to crack
component materials adequately meets the respective initiation and subsequent crack propagation [1].
performance requirements and effectively resists
hydrogen. However, the interface between the
WRC Bulletin 534

The resistance of the dad metal to disbonding is through the sample is restricted during cool-down by
dependent upon a variety of variables including the OD overlay, hence diffusion of hydrogen is
materials processing methods, clad manufacture tech- unidirectional, more accurately simulating service
niques and service conditions. conditions.
Since disbonding can potentially result from

hydrogen accumulation at bond anomalies, or
nonbonds, an understanding of the relationship
between the presence of nonbonds and hydrogen
service conditions is critical. In this work, two
hydrogen disbonding tests were performed. One
specimen was selected from a typical explosion
bonded clad plate. The other specimen was
manufactured with an intentionally induced nonbond
flaw.
HYDROGEN DISBONDING TEST

METHODS
Figure 1. Disbonding test specimen configuration
Several test methodologies currently exist for commonly used to conduct qualification tests of clad
evaluating the susceptibility to hydrogen disbonding. and overlay for hydrogen service
These methods are extensively discussed in the
referenced literature [1-5].
The tests of this program were performed by high

temperature, high pressure hydrogen charging in an
autoclave. The exposure conditions were as follows: Cled
A
Temperature: 850°F (454°C) J1 J1
0
0
Hydrogen Partial Pressure: 2,150 psi (150 kg/cm 2) I:Q) I:Q)
Bose Mel:..ol >
Exposure time: 48 hours 6 0
Cool-Down Rate: 300°F /hr (167°C/hr)
Number of Cycles: 2
A
Two different specimen configurations were used. Secllan A-A
The first test used the specimen configuration shown
in Figure 1. The specimen was 2" x 3" x 2.188" thick
(50 x 75 x 56 mm) with cladding on one face only. In Figure 2. New specimen configuration consisting of
a specimen of this type, the hydrogen diffuses out stainless steel on the top and side surfaces to better
through five of the six sides during cool-down. stimulate in-service hydrogen diffusion behavior
However, in vessel service the hydrogen is primarily during cool-down
limited to unidirectional diffusion. Consequently, the
laboratory evaluation is less severe than service under
the same conditions. VARIABLES AFFECTING DISBONDING
The second test used a newly proposed specimen Component Metal Variables
design which provides a more realistic simulation of
hydrogen diffusion during cool-down [1]. This The composition, manufacturing method, and
specimen, Figure 2, was 3" OD x 3" thick (75 mm OD heat treat condition of the component metals are
x 75 mm thick) with stainless steel on one flat face critical for hydrogen resistance. These variables are
and the OD. The stainless steel on the face was extensively presented elsewhere and will not be
applied by explosion bonding; the stainless on the addressed in detail here [1-3, 6]. The test specimens
OD was applied by weld overlay. Hydrogen diffusion were produced from Cr-Mo steel, SA-387,
WRC Bulletin 534

manufactured using procedures incorporating both The stainless steel is then deposited onto the interior
vacuum degassing and sulfide shape control surface of the alloy steel components using one of
processes. The cladding metal was austenitic wrought several potential welding processes. In many cases,
plate of the niobium stabilized grade SA-240 Type the overlay metal is deposited in two sequential
347. layers to compensate for dilution in the initial layer
by the unalloyed base metal. The interface
The test specimens of this program were given characteristics are very dependent on the overlay
the heat treatments typically experienced in process and heat input parameters. Incorrect
manufacture and fabrication of refinery vessels. The selection or control of welding parameters can result
base steel was normalized and tempered prior to in formation of hard martensitic regions adjacent to
cladding. The stainless steel was solution heat treated the fusion line in the Cr-Mo steel which are
and water quenched prior to cladding. The clad susceptible to hydrogen cracking.
product was given a simulated Postweld Heat
Treatment (PWHT) at 1,275°F (690°C) for 12 hours Explosion Clad. Explosion cladding uses the
prior to hydrogen testing. energy of an explosive detonation to create metal-
lurgical bonding between the stainless steel and base
Clad Manufacture Variables metal [7]. Prior to bonding, the component metals
are heat treated to achieve optimum performance
Clad for hydrogen service refinery vessels is characteristics. Typically, the stainless steel is solution
primarily manufactured using one of three different heat treated and water quenched; the base steel is
manufacturing technologies: roll bond clad, weld normalized and tempered or quenched and tempered.
overlay, and explosion clad. Each of these processes During the bonding process the faying surfaces are
results in distinctly different interface metallurgical dynamically cleaned by jetting away the surface
characteristics and potentially significant differences layers. The nascent surfaces are then brought
in hydrogen disbonding characteristics. Although not together under pressures of several million psi. There
addressed herein, there are significant variations in is no bulk heating of the metals, subsequently the
the cladding processes of individual manufacturers heat treat conditions are not altered by the bonding
which may result in further variance in the hydrogen operation and no measurable diffusion occurs. The
service performance of the specific products. clad plates are then formed into cylinders and heads
for subsequent vessel fabrication. If hot forming is
Roll Bond Clad. Roll bond clad is produced by required for fabrication some diffusion will occur in
hot rolling a multilayer pack, consisting of cladding the interface region; however, since the combined
metal plates and base metal plates appropriately time and temperature exposure of these operations is
prepared for bonding. Subsequent to the hot rolling typically short, interface carbide concentrations are
operation the composite plates are heat treated to much lower.
achieve the required metallurgical conditions for each
component metal. Typically the clad plate is first The three different processes result in clad
heated to the stainless steel solution annealing metals that differ significantly in cladding metal
temperature followed by air cooling. The plate is then chemistry uniformity and corresponding corrosion
heated to the normalizing range for the Cr-Mo steel resistance, in bond strength, in bond uniformity, and
followed by air cooling or quenching. Lastly, the plate in component metal mechanical properties. The effect
is tempered at the appropriate temperature for the of these variables upon hydrogen disbanding is
base steel; a temperature which is normally above the currently not well known. Service experience of
vessel PWHT. The clad plates are formed into vessels constructed by methods other than weld
cylinders and heads for subsequent vessel fabrication. overlay has not been well documented, consequently
The manufacturing and heat treating sequence may laboratory testing is being used to evaluate clad
induce significant diffusion between the component disbonding characteristics.
metals resulting in a concentration of carbides in the
interface region. Service Operation Parameters
Weld Overlay. Weld overlay is typically produced Exposure Conditions. The in-service exposure
in a very different sequence of operations. The base conditions influence the probability of hydrogen
steel is manufactured to the required condition and disbanding. The potential for disbonding increases
fabricated into and heads prior to overlay. with increasing temperature and/or hydrogen partial
WRC Bulletin 534

pressure. The higher the service temperature and/or was produced in the center of the explosion bonded
hydrogen partial pressure, the higher the hydrogen dad face, Figure 3. Ultrasonic inspection indicated no
concentration available at the base metal/dad metal other nonbonds in the clad face or the overlayed OD.
interface during shutdown conditions. Consequently
the pressures that can be generated by internal
accumulation of molecular hydrogen are much
greater. The pressure and temperature conditions of
this test program were chosen to represent a
significant percentage of refinery vessels as
/
documented elsewhere [8]. I
/
'
Cool-Down Rate. The cool-down rate used during
shutdown has a dramatic effect on the susceptibility
of hydrogen disbonding, with increasing cooling rates
increasing the potential for disbanding. Under higher
cooling rates the diffusion rate of hydrogen in the Cr-
Mo steel decreases dramatically, potentially creating
conditions in which the hydrogen cannot outgas to
the exterior surface. The result is a high
concentration of supersaturated atomic hydrogen in
the steel which can coalesce at bond anomalies
creating disbanding. At sufficiently high cooling rates,
it is expected that any of the stainless steel/Cr-Mo
clad products can be forced to incur hydrogen
cracking of the steel and/ or separation of the dad.
The 300°F (167°C)/hr rate used in this test program
is considered to be more extreme than can possibly
be achieved in realistic refinery operation. Intentional Defect 1~
1'347 ~losive Clad
Base Material
047 Weld Overiay
TEST PROGRAM
Hydrogen disbanding tests were performed on

two explosion dad samples. The clad was manufac- Figure 3. Schematic of Specimen presenting the size
tured by Explosive Fabricators, Inc., using standard of the cladding components and intentional nonbond
production procedures. The stainless steel cladding
was SA-240 Type 347, 0.188-inch (4.8 mm) thick. The The hydrogen exposure sequence performed by
base steel was Cr-Mo per SA-387 Grades 11 and 22. CLI International and inspections performed by DNV
Industry, Houston, Texas, are as follows:
Specimen A was a standard rectangular hydrogen
disband test block, Figure 1, with stainless steel 1. Ultrasonic Inspection of all clad surfaces
explosion bonded onto one face only. The base steel 2. SPWHT at 1275°F (690°C) for 12 hours, heating
was SA387 Grade 11 Class 2, 2-inch (50 mm) thick. and cooling rates above 800° F ( 425° C) per
Ultrasonic inspection prior to testing showed no ASME Code.
nonbond indications. 3. Ultrasonic Inspect
4. Hydrogen charge at 2,150 psi (150 kg/cm 2) in an
Specimen B was manufactured to the new design, autoclave at 850°F (454°C) for 48 hours followed
Figure 2. The base steel was SA-387 Grade 22 Class by cooling to room temperature at 300°F
2, 2.5-inch (67 mm) thick. Stainless steel was (167°C) per hour.
explosion bonded onto one of the flat faces. The OD 5. Hold specimen at room temperature for 7 days
was enclosed by weld overlay. The overlay was 6. Ultrasonic inspect
applied using FCAW, the first pass was E309LTl; the 7. Repeat Steps 4, 5 and 6
second pass was E347LTl. The back surface was not
clad. An intentional nonbond, 0.5-in. (12.5 mm) OD
WRC Bulletin 534

interface region and no measurable increase in

concentration of carbon.
Figure 4. Photomicrograph of Specimen B bond zone

prior to heat treatments and hydrogen disbond
testing. (50x, Nital etch)
The ultrasonic inspections were performed using

straight beam techniques with pitch-catch transducers. Figure 5. Photomicrograph of Specimen B bond zone
The precision was sufficient to detect indications of after hydrogen disbond testing (SOX, Nital)
0.03" (0.75 mm) or greater.
The ultrasonic inspections indicated that no

disbanding occurred in either the explosion clad or
the overlay during or after either of the two ther-
mal/hydrogen cycles. Furthermore, there was no
change in the size of the intentional nonbond. Figure ·-·-·----
300
4 shows the interface microstructure of Specimen B
before testing; Figure 5 shows an area adjacent to the 280
nonbond after the test sequence. 0:-

0
0 260-
:l
g
Specimens A and B exhibit pretest properties ~ 240
which are typical of production clad materials. Table w
1 presents the comparative data of Specimens A and
B with that of 59 clad plates produced during 1992
z
0
~
I 34 7 STAINLESS STEEL ---b-; - - Cr Mo STEEL
and 1993.
Samples of the same metal combination as

1 BO
160 ,---..........,--,.----r--,---T- -
___ - - - ··-- ···· ---------·· - - I
,,_,,_,.
--.------.~-,---,-
0 030 0 025 0 020 0 015 0 010 0 005 0 005 0 010 0 015 0 020 0 025 0 030
Specimen A were examined for hardness and diffu- BONOZONE
sion in the interface region. The specimens were
DISTANCE FROM BONDZONE
given the same heat treatments as Specimen A prior
to examination. Figure 6 presents the microhardness
scan across the bond zone. Figure 7 presents the scan Figure 6. Microhardness scan across explosion
for Carbon, Chromium, and Nickel across the bond bonded interface of Specimen A material after all
zone. The data indicate no hardness increase in the heat treatments of this program
WRC Bulletin 534

CONCLUSIONS variables utilized in this study. Increases in wall

thickness, hydrogen pressure, and cooling rate would
The two disbanding tests indicate that the result in exposures more severe than the test
explosion bonded stainless steel/Cr-Mo steel prod- program. Conversely, decreases in these parameters
ucts are not disbanded by the high temperature, high would result in less severe conditions.
pressure hydrogen exposure and subsequent
shutdown. The specimen sizes and simulated service
conditions are typical of those that are currently used
Specimen B demonstrates a newly proposed within the industry for product qualification.
specimen design which more closely simulates the Additional testing under more severe conditions
behavior of a clad refinery vessel. The hydrogen would be beneficial for determining the range of
diffusion during cool-down is unidirectional as would suitability of explosion clad for refinery vessel
occur in service. This research program demonstrates applications. Similarly, testing of other clad types
that hydrogen disbanding specimens of the design of under similar conditions would provide beneficial
Figure 2 provide a simple, cost effective method for information to those specifying materials of construc-
closely simulating material behavior in clad refinery tion for refinery hydrogen service.
vessels.
REFERENCES
[1] M.S.Cayard, R.D.Kane, C. E. Stevens,

"Evaluation of Hydrogen Disbanding of Stainless
Steel Cladding for High Temperature Hydrogen
j: 17
Service," 1994 Annual Corrosion Conference,
(!J
Baltimore, MD, (Houston, TX, National Association
~ 16
of Corrosion Engineers, 1994)
_J
~ 15
0
z 14 z [2) T. Sakai et al, "Hydrogen Induced Disbanding of
oil
0.12 al Weld Overlay in Pressure Vessels and its Prevention,"
~ 13 .•. ·························I ~
~
2 ° Current Solutions to Hydrogen Problems in Steels, p.
--:::x:-7~S~
01
12
I
() 11- +"" 340, (Metals Park, OH: American Society for Metals.
000
1982).
---,--~,-----,..------,--,------,-----r
10 -·- ...--,--------,---- - . - - 006
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14
t
!101i020NE
DISTANCE FROM BONDZONE (INCHES)
CHROMIUM ·-+- NICKEL ~)!(--CARBON

[3] G.M. Pressouyre, J.M. Chaillet, and G. Valette,
"Parameters Affecting Hydrogen Induced Disbanding
of Austenitic Stainless Cladded Steels," Current
Solutions to Hydrogen Problems In Steels, p. 349,
Figure 7. Microanalysis scan for Carbon, Chromium, (Metals Park, OH: American Society for Metals,
and Nickel in Specimen A stainless steel cladding 1982).
metal after an heat treatments of this program
[4] T. Fujii et al, "A Safety Analysis on Overlay
Specimen B also addresses the concerns of the Disbonding of Pressure Vessels for Hydrogen
effects of high temperature, high pressure hydrogen Service," Current Solutions of Hydrogen Problems in
exposure on explosion clad which contains an area of Steels, p. 361, (Metals Park, OH: American Society
nonbond at the interface. The clad face of the for Metals, 1982).
specimen included an intentionally created 0.5-inch
(12.5 mm) OD nonbond. The simulated service [5] T.P. Groeneveld, "The Effect of Austenitic
conditions did not cause growth of the nonbond as Stainless Steel Weld Overlay or Cladding on the
indicated by post exposure ultrasonic inspection and Hydrogen Content and Hydrogen Attack of the
destructive evaluation. Underlying Steel in Petrochemical Reactor Vessels,"
International Conference on the Interaction of Steels
It is critical to recognize that this research with Hydrogen in Petroleum Industry Pressure Vessel
simulated vessels with wall thicknesses, hydrogen Service, p. 311, (New York, NY: The Materials
partial pressures, and cooling rates similar to the test Properties Council, 1993).
WRC Bulletin 534

[6] R. Blondeau et al, "Contribution to a Solution to

the Disbonding Problem in 2 1/4 Cr-1 Mo Heavy
Wall Reactors," Current Solutions to Hydrogen
Problems in Steels, p. 356, (Metals Park, OH:
American Society for Metals, 1982).
[7] J.G. Banker, E.G. Reineke, "Explosion Welding,"

ASM Handbook: Volume 6, Welding and Brazing
(Materials Park, OH: ASM International, December
1993).
[8] Minutes of API Refining Subcommittee on

Corrosion and Research, Attachment IV, API Mid-
year Refining Meeting, May 14-16, 1984, New
Orleans, LA, 1984.
TABLE I
COMPARISON OF PROPERTIES OF TEST SPECIMENS WITH THOSE OF

PRODUCTION CLAD PLATES
SPECIMEN A SPECIMEN B PRODUCTION

Shear Strength 71,243 69,444 75,469 (average)
lb/in 2 12,250 Std. Dev.
N/mm 2 492 480 521 (average)

78 Std. Dev.
Non-Bond None 0.195 in 2 0.003 % of area
1.22 cm 2
NOTE:Production shear data is taken from 59 clad plates produced in 1992 and 1993 and is exclusively
from clad plates of stainless steels types 321 and 347 clad to Cr-Mo steels SA-387, Grades 11 and 22,
Classes 1 and 2. Production UT data is from all clads produced in a 2 month period of 1993, yielding
18,000 sq-ft (1,700 sq-m).
WRC Bulletin 534

WRC Bulletin 534

HYDROGEN INDUCED DISBONDING IN WELD-CLAD

PRODUCTS: THE INFLUENCE OF MICROSTRUCTURE
C.CASCIARO, P. MARINELLI
Belleli Ricerche Scpa
TARANTO, ITALY.
A.SOLINA, R.V ALENTINI

Dip. di Ing. Chimica, Chim. Ind. e Scienza dei Mat.
UNIVERSITA' DI PISA, PISA, ITALY
ABSTRACT presence of sensitive microstructure and high tensile

stress induced by PWHT [5], can induce disbonding of
Pressure vessels such as hydrocracking reactors are clad overlay from base metal.
fabricated from 2.25 Cr- I Mo steel weld-clad with The microstructural factors responsible for the
austenitic stainless steel. In these reactors, hydrogen disbonding are: the presence of martensitic layer near the
diffuses into the steels and can cause disbonding of the interface, the presence of austenitic coarse grains and the
weld cladding after shut down and cooling to ambient heavy precipitation of carbide at the fusion line after
temperature. PWHT [6,7]. Welding procedures and PWHT may
The aim of this study is to clarify the influence of modify these factors that are in most cases
welding procedure , hydrogen charging and cooling rate interdependent.
on dis bonding susceptibility. In the present research the disbonding phenomenon of
Autoclave tests, metallographic examinations and Submerged Arc Welding (SAW) was investigated for
hydrogen measurements, are carried out to verify various welding procedure specifications. The aim of the
disbonding behaviour of different kinds of cladding and work is to characterize the structural transition zone and
results are correlated with theoretical considerations. to correlate the critical rnicrostructural parameters on
The influence of microstructures of interface zone dis bonding susceptibility.
across fusion line are considered on disbonding
phenomena. Austenitic coarse grain zone, microhardness
and carburized zone have been carefully examined and EXPERIMENTAL PROCEDURE
they have been correlated to welding procedure
specifications. Materials: The chemical compositions of steels
employed are shown in Table 1. The base material is a
quenched and tempered plate (300x500xl00 mm) of
INTRODUCTION SA387 Gr22 Cl2 .
All welds have been executed using a Submerged Arc
The disbonding phenomenon is a typical hydrogen strip cladding with AISI309 and/or AISI347. Welding
embrittlement problem of low alloy steels such as 2.25 Procedure Specifications (WPS) are summarised in Table
Cr - 1 Mo, when heavy sections are clad with an 2. The plate coupons were used to produce cylindrical test
austenitic stainless steel (AISI 300). Very thick pieces with 50 mm diameter and 30 mm height. All
hydrogenetion reactors,built with these materials and specimens have been Post Weld Heat Treated at 690°C
stainless steel clad are of normal use in petroleum for 16h.
industry [ l]. In these vessels the shut down from high Test pieces for hydrogen charging (almost five
temperature and high hydrogen pressure induces a specimens for each procedure) were placed in an autoclave
hydrogen saturation at room temperature at the at 475°C - 150 bar for the time required to obtain 99% of
overlay/base metal interface region [2,3,4]. the saturation concentration[8] (40 ppm).
High hydrogen concentration combined with the
WRC Bulletin 534

Table 1 - Chemical Composition of investigated steels hydrogen measurements and metallographic

examinations[9, 10]. The disbonding was evaluated by
El. Base Metal AISI 309L AISI 309L AISI 347L ultrasonic examination (U.T.) using a reference block
~%2 2.25Cr1Mo ~Nb2 containing 2 mm diameter flat bottom holes. The
c 0.140 0.013 0.018 0.014 disbonding area has been weighted by ratio between crack
Mn 0.55 1.65 2.05 1.70 signals and the reference one. Micrographic examinations
Si 0.17 0.31 0.33 0.15 were carried out to localise the defect and confirm the
p 0.0037 0.011 0.014 0.014 presence of cracks. Microhardness tests and x-ray
s 0.0038 0.001 0.001 0.012 microanalysis measurements (WDX) have been carried
Ni 0.055 13.24 12.44 10.42 out to study the transition zone across the fusion line.
Cr 2.25 23.56 23.70 19.35
Mo 1.01 0.08 0.07 0.10
!'..1b 0.79 0.48 EXPERIMENTAL RESULTS
Every signal higher than 20% of reference one has
been recorded; Dis bonding phenomena, conventionally,
Table 2 - Welding Procedure Specifications (WPS) occur only when signals were higher than 50% of
reference; every indication among 20 and 50 % have been
Procedures 1° layer 2° layer Heat Input C@l recorded as "light disbanding".
(AISI) (AISI) [kJfcm2]* Nieg No disbanding occurs when the cooling rate is 100
GSAW 309L 16.10 1.59 °C/h; when the cooling rate is 600 °C/h every specimen
HSAW 309L 2G.44 1.57 shows a disbanding of about 20% and no appreciable
ISAW 309LNb 15.87 1.68 difference has been noted among WPSs. For the cooling
LSAW 309L Nb 19.51 1.68 rate of 300 °C/h, meaningful differences were found;
MSAW 309L 347L 16.43 1.57 Fig.2 shows the influence of WPS on disbonding
NSAW 309L 347L 19.25 1.60 susceptibility.
OSAW 309L Nb 14.00 1.69
PSAW 309L 347L 14.00 1.57
UT
* The Heat Input for unit area is calculated as~
s w
60
were V is the arc voltage [V], I is the welding current Cooling rate 300 °C/h
[A], sis welding speed [cm/min] and w is the strip width 0.9
[cm]. 0.8
Preliminary tests have been carried out to check the 0.7

('l'2
required cooling rates (100, 300 and 600°C/h), as shown c<j
<l) 0.6
in Fig.1. ....
<t)[) 0.5
i::
475 •c
:ai:: 0.4
0
..0
rJl
0.3
Q
150 bar 0.2
0.1
Typicai cycie of 0 p
G H L M N 0
hydrogenation
Welding Procedure specifications
Fig. 2 Influence of WPS on disbonding

Oh 4h 48h susceptibility
Fig. 1 Typical cycle of hydrogenation Extensive results of hydrogen measurements,

compared to theoretical calculations, have been reported
in a previous work [10]; ratios between maximum values
After cooling some specimens have been machined and of hydrogen concentrations calculated at the fusion line,
stored in liquid nitrogen to obtain test pieces suitable for and initial concentration in in ox stainless steel (calculated
WRC Bulletin 534

and subsequentely verified with experimental tests), have

been shown in Fig.3.
12
.....=
-e
u
u
~
~
Fig.Sa Typical hydrogen disbanding (SEM)

(500x)
100.000 300.000 600.000
Cooling rate [°C/h]
Fig. 3 Effect of cooling rate on ratio Cmax/Cin
Typical hydrogen induced disbanding cracks are shown

in Fig. 4-5.
Fig.Sb Same particular at more magnification

(1500x)
The Fig.6a-c shows a typical profile of microhardness

for an "as welded" specimen and for a PWHTreated one.
In as welded condition a martensitic layer was formed at
the interface inox/base material; after PWHT a massive
carbide precipitation occurred too in this region and
therefore the maximum values of hardness were located at
the fusion line; Fig.7 shows the WPS influence on peak
hardness values.
Typical distributions of Cr and Ni, obtained by
microanalysis measurements, are shown in Fig.8-9.
Fig. 4 Typical Hydrogen disbanding (650x)
WRC Bulletin 534

50 1111 M (AW)
8 G(AW)
500 M (PWHTreated)
G (PWHTreated)
40
400
2001----......
10•,.._--'-~..l..--I...~'---'---'~-'---'----'
-150 -50 50 150 250 100,__.......___,_~..__.___...___.__..~...._~
Distance from interface ( µm) -150 -50 50 150 250

Fig. 6a Microhardness of sample G. Distance from interface ( µm)
Fig.6c Microhardness of sample M.

500 II l(AW)
I (PWHTreated)
500
~ 400
>
::r::
'-'
VJ
VJ
(].)
:::::
] 300
..c
c
I-<
u
......
~
200
1001...--'---'-~....__.___,~_.__._~.._~
-150 -50 50 150 250

Distance from interface ( µm) G H L M N 0 P
Fig.6b Microhardness of sample I. Fig. 7 Effect of WPS on peak value of microhardness

afterPWHT
WRC Bulletin 534

38 Studies of Cladding and Overlay for Pres sure Vessel Service
disbonding behaviour when the cooling rate was 600

°C/h. High cooling rates, probably, induce not only
hydrogen peak concentrations but also other phenomena
that have to be studied in future works.
Hydrogen disbonding behaviour of single layer 308L
S.S. is better than that of 347 , this phenomenon is
related to the chemical analysis of deposited weld metal
[8]. Fig.10 shows the influence of the ratio Creq/Nieq
on disbonding susceptibility at the cooling rate of 300
°C/h for one layer cladding (works to evaluate the
influence of the second layer are in progress; single layer
procedures seem to have a better behaviour of double
layer ones). Technical literature gives us a criterion to
distinguish between two different behaviours[l 1].
1.0
Fig.8 Sample G: Cr profile across fusion line (750x)
0.8
0.6
t\2
ro
~on
.....,
i::: 0.4
"O
i:::
0
,.0
.,...,
VJ
Cl
0.2
0.0..._'-l~""-1~-'-...._....__._...._....__,__,__,__.___.
1.55 1.60 1.65 1.70

Creq/Nieq
Fig.9 Sample G: Ni profile across fusion line (750x)
Fig. I 0 Effect of Creq/Nieq on disbonding susceptibility

DISCUSSION AND CONLUSION
Perhaps, whith the employed cooling rates after
The most important parameter influencing hydrogen welding, if Creq/Nieq is less than 1.6 solidification
disbonding susceptibility is the cooling rate after begins with ferrite followed by austenite formation; when
hydrogenation. High disbonding susceptibility was found Creq/Nieq is greater than 1.6 solidification begins with
at a cooling rate of 600°C/h. This higher susceptibility is austenite followed by ferrite. Impurities segregate to
not related to Welding Procedure Specifications and can grain boundaries of the last solidified phase and therefore
not be clearly explained by hydrogen peak concentration the second solidification mode is more critical because
only (Fig.3). Hydrogen peak concentration is not so hydrogen embrittlement occurs at austenitic grain
much higher than that found for a cooling rate of 300 boundaries [8]. The ratio Creq/Nieq is related to strip
°C/h; some additional tests with a smaller initial chemical composition and to dilution (that is a function
hydrogen charge (28 ppm) and with a double PWHT of welding parameters).
(690°C x 30h + 600 °C x 15 h) have been carried out, Microhardness measurements do not lead to any
but no appreciable difference has been noted on conclusion about disbonding susceptibility of claddings
WRC Bulletin 534

tested in this work: all peak hardness values are of about JO. C. Casciaro, M. Olive, A. Solina, R. Valentini
400 HV after PWHT and no meaningful relation has "Hydrogen diffusion and distribution in weld-clad
been found between peak hardness values and disbonding fabricated products with special reference to dis bonding",
susceptibility. Weld. Int., Vol.7, 1993, p.544.
Experimental tests and theoretical considerations about 11. J.W. Elmer, S.M. Allen, T.W. Eagar
the thermal cycle of the weld, lead to conclude that "Microstructural development during solidification of
different strip widths don't induce different disbonding stainless steel alloys", Metall. Trans.,20A, 1989, 2124
behaviours, when the appropriate sets of welding
parameters are used. The only difference is the number of
overlap zones that seem to be more critical, and therefore
a better behaviour is achievable with 90mm wide strips.
A similar research has been carried out, also, on
Electro Slag (ES) cladding procedure; it will be the
subjiect of a separate report. Results obtained by ES
cladding procedure were in many cases better than those
obtained by SAW procedure.
ACKNOWLEDGEMENTS
Authors thank Mr. A.Bertoni and Mr F.Fusari of

BELLELI Spa for the usefull discussion on metallurgical
and procedural problems of the research.
REFERENCES
1. F. Broschi, G. Camona, G. A. Mondora "Impiego

degli acciai 2.25Cr-1 Mo per recipienti in pressione di
grosso spessore: quadro di riferimento e problematiche
metallurgiche", Roma, 1984
2. K. Ohnishi, R. Chiba, J. Watanabe, K. Naito, H
Okada " Hydrogen induced disbonding of stainless steel
overlay weld", Pressure Vessel Research Committee
Meeting in Scottsdale, 1981, 1-35
3. T. Sakai, K. Asami, H. Takada "Hydrogen distribution
through pressure vessel wall'', ibidem
4. K. Smit, P.F. Ivens "Hydrogen in thick-walled, high
temperature hydrogen reactors during service and during
shutdown", Int. Conference on Interaction of steels with
hydrogen in petroleum industry pressure vessel service,
1989, Paris
5. R. Palengat, P. Bocquet, A. Vignes "Disbanding
mechanism and its prevention", ibidem
6. H.A. Schimoller, J.L. Ruge , Pro. Int. Conf. Residual
Stresses in Welded Construction and their effects, The
Welding Institute, 1977, 251
7. M. De Sanctis, L. Paganini, A. Solina, R. Valentini
"Disbonding of type weld overlays induced by postweld
heat treatments and hydrogen charge", J. Heat Treating,
vol.7, 1989, 35-47
8. G.M. Pressouyre, J.M. Chaillet, G. Valette
"Parameters affetting the hydrogen disbonding of
austenitic stainless cladded steels", Proceedings of the
First International Conference, ASM, Washington DC,
1982, 349
9. K. Ohnishi "Hydrogen in austenitic stainless steel
under high temperature, high pressure hydrogen
atmosphere", WRC Bullettin, 240, 1978
WRC Bulletin 534

DISBONDING MECHANISM AND ITS PREVENTION
A. Vignes and R. Palengat

FRAMATOME
71380 Saint Marcel, France
P. Bocquet
Creusot-Loire lndustrie
71202 Le Creusot, France
ABSTRACT DISBONDING MECHANISM
The study carried out covers the hydrogen induced The cooling down of pressure vessels from high
cold cracking phenomenon, so called disbanding, at the temperature and hitjh hydrogen pressure induces a hydro-
interface of 2.25 Cr - 1 Mo steel plates, cladded with gen saturation at room temperature at the overlay/base
stainless steel. metal interface region [1]. The disbanding phenomenon
The main factors responsible for the disbanding phenome- can immediately occur if this region contains sensitive
non are the presence of austenitic coarse grains and the microstructures such as :
thin martensitic layer near the interface. It seems that - a thin martensitic layer carburized by the postweld
the thin martensitic layer has the more detrimental heat treatment (PWHT),
effect. The solutions using a low heat input to limit - carburized austenitic coarse grain boundaries near
only the amount of austenitic coarse grains are partly this thin martensitic layer.
efficient under the worst conditions. We have thus in- The disbanding phenomenon can propagate along these two
vestigated a very efficient solution consisting in the regions.
performance of a duplex heat treatment, which mainly Any parameter which generates these microstructu-
acts on the thin martensitic layer and prevents any res promotes the sensitivity to the disbanding phenome-
risk of disbanding. non [2] [4]. For instance, a PWHT at a higher tempera-
The Maglay process can be applied, using wide ture or the use of the electroslag welding process
strips, even in the presence of a large amount of aus- increases the risk of disbanding [3].
tenitic coarse grains and a strongly carburized thin
martensitic layer. CURRENT SOLUTIONS
INTRODUCTION A lot of solutions were developped to limit the

susceptibility to disbanding, by the control of the
Pressure vessels such as hydrocracking reactors, stainless steel microstructure near the thin martensi-
used at high temperatures and under high hydrogen pres- tic layer :
sure, are made of 2.25 Cr - 1 Mo steel and the internal - use of submerged arc process with narrower strips or
surface is weld overlaid with austeno-ferritic stainless strips equal to 75 mm, in order to limit the percentage
steel. FRAMATOME has carried out developments to apply of austenitic coarse grains [3].
the Maglay process so as to obtain a high metallurgical - increase of the welding current and the welding speed
quality of the weld overlay and a high deposit rate. to obtain a finer austenitic grain size for SAW process
However, the base metal-to-overlay weld metal [2] and ESW process [2] [3]. This solution cannot be
interface is sensitive to the hydrogen induced cracking, used for wide strips (e.g. 150 mm) because the indus-
so called disbanding phenomenon. Its resistance to the trial application appears to be very difficult.
disbanding phenomenon depends on the welding process - use of special strip chemistry to develop high delta
and the postweld heat treatment (PWHT). ferrite or some quantities of martensite in this area.
We have thus performed disbanding tests in order Tests were recently performed to study the effect
to compare : of a duplex PWHT in order to prevent the disbanding
- various welding processes, including Maglay process, phenomenon, even if the microstructure near the inter-
and welding parameters, face is not adequate [5].
- various PWHT types,
and thus avoid the disbanding phenomenon in one or two
layer overlay.
WRC Bulletin 534

TRIAL PROCEDURE
WELDING CONDITIONS TEMPERATURE ( 0 0 TEMPERATURE ( 0 0
2.25 Cr - 1 Mo plates of 150 mm thick were weld

overlaid under various conditions :
- strip/flux processes : SAW and ESW,
,---·..,_\
690'C - 30 H
.
690'f - 30.,.H
r-"-1
J Tu_o~~- 1s H
- strip dimensions : 60 x 0.5 mm - 75 x 0.4 mm - 150 x I . ... +
0.4 mm, . \
- welding speed and welding current : conventional and I . I \
higher values,
- first layer microstructure : normal delta ferrite
. \ ... \
(approx. 6 %) and high delta ferrite (> 10 %) and low R.PWHT TIME S.PWHT TIME
silicon using a special 309 L strip,
- first layer chemical composition : 309 L or 347 type,
- one or two layers. TEMPERATURE ('Cl TEMPERATURE ( 0 0
The table 1 shows the welding conditions applied
and the chemical composition obtained in the first 690°C - 30 H
layer. 1--------.,
I I 650°C - 30 H
I I :---·--·--·-\
POSTWELD HEAT TREATMENT CONDITIONS (PWHT) I I 600°C . I
I \ {i"nT\ I
!
I
i
I I I I
Four types of PWHT were used, according to Fig. 1: I I I I I
I
I
.
I I I I
- a single heat treatment called regular PWHT (R.PWHT) - I I I I I
. I
I
I 11 I
690°C - 30 hours, I II I ! i
- a special heat treatment (S.PWHT) - 690°C - 30 hours + II I I \
600°C - 15 hours, D.PWHT TIME LT PWHT TIME

- a duplex heat treatment (D.PWHT) - 690°c - 30 hours +
600°C - 15 hours,
- a single heat treatment at low temperature, called LT Fig. 1 Postweld heat treatment
PWHT - 650°C - 30 hours.
TABLE 1. WELDING CONDITIONS - CHEMICAL COMPOSITION AND MICROSTRUCTURE OF THE FIRST LAYER
STRIPS WELDING PARAMETERS

TEST WELDING CHEMICAL COMPOSITION lST LAYER
NAME PROCESS 1st 2nd Dimensions Welding Welding OF THE lST LAYER b
FERRITE
layer layer current speed (%} (SCHAEFFLER)
(mm) (A) (cm/mn)
SAW1 309 L 308 L 60 x 0.5 750 14 c = 0.043 ; Ni = 9.28 12

Si = 0.800 ; Cr = 19.32
- SAW
SAW2 (2 layers) 309 L 347 75 x 0.4 1200 19 c = 0.062 ; Ni = 9.03 6.2
Si = 0.829 ; Cr = 18.09
ESW1 1260 13.5 c = 0.035 ; Ni = 10.14 8

Si = o.497 ; Cr = 19.48
'--- 75 x o.4
ESW2 1800 21 c = 0.043 ; Ni = 8.45 7
Si = 0.735 ; Cr = 17.06
~- normal 308 L
ESW4 ESW 309 L 2580 14 c = 0.030 ; Ni = 10.45 8.7
MAGLAY Si = o.454 ; Cr = 19.96
~- process 150 x 0.4
ESW5 (2 layers) 2750 21 c = 0.041 ; Ni = 9,13 5.4
Si = 0.370 ; Cr = 17.75
'---
ESW3 75 x 0.4 1250 13 c = 0.029 ; Ni = 8.41 12
Si = 0.362 j Cr = 18.98
'---- special 347
ESW6 309 L 150 x 0.4 2500 14 c = 0.027 ; Ni = 8.45 13
Si = 0.387 ; Cr = 18.97
ESW MAGLAY c = 0.026 ; Ni = 9.84
ESW7 process 347 150 x 0.4 2500 14 Si = 0.512 ; Cr = 17.39 4.9
(1 layer) Nb = 0.574
HS ESW 309 L 347 75 x 0.4 2200 30 c = 0.036 ; Ni = 8.30 12
ESW (2 layers) Si 0.40 ; Cr = 18.7
WRC Bulletin 534

The effect of the duplex heat treatment was alrea-

dy studied by the Welding Institute [5]. The PWHT 690°C
30 hours induces an important carburization of the thin
martensitic layer ; the interface is thus very sensitive.
In fact, this holding time is representative of heavy
wall pressure vessels, including all repairs.
DISBONDING TEST PROCEDURE
Disbanding specimens were obtained according to

Fig. 2. Each specimen has undergone the following opera-
tions :
- hydrogen charging in an autoclave at 450°C under a
pressure of 150 kg/mm2. The holding time was defined
according to the overlay thickness. The hydrogen level
at the interface was thus identical for all specimens,
- after charging, the specimens were cooled down without
hydrogen pressure at various rates : 675°C/h - 250°C/h
and 150°C/h within 450 to 225°C, 15 days after hydrogen charging
- the disbanding was evaluated within 15 days after
cooling, by ultrasonic testing (probe : 5 MHz, 0 10 mm, Fig. 3 DSCAN patterns
focussed on the defect ; reference block containing 1 mm
dia. flat bottomed holes ; immersed testing with 100 %
automatic DSCAN). The Fig. 3 shows an example of the EXPERIMENTAL RESULTS AND DISCUSSION
patterns obtained,
- macrographic examinations were systematically carried All results are indicated in the table 2. The
out to localize the defect and thus confirm the presence following effects can be observed :
of disbanding.
EFFECT OF THE SPECIMEN COOLING RATE
This factor has a significant effect on the dis-

banding level (Fig. 4). The disbanding increases with
the cooling rates. For all mixed processes, the disban-
ding level is :
- very small (lower than 5 %) for the slowest cooling
rate (150°C/h),
. . _______.0 - within 0 to 30 % for the intermediate cooling rate

(250°C/h),
- high (20 to 50 %) for the fastest cooling rate
( 675°C/h).
100
DISBONDING (%)
Fig. 2 Disbanding specimen e SAW1 60 X 0.5 MM

oSAW2 75 X 0.11 MM
__ ..,..
50 ~~~~~ ~H 8:~:
l< ES\13 75 X 0.4 MM
-- ----
40
6ES114 150 X 0.4 MM
4'ES115 150 x 0.11 MM
•ES\16 150 x 0.•1 MM
"1ESll7 150 X 0.4 MM ,/'
_,..,.,.,/ -- ,,.-
d.
oA
'O
'VHS ES\I 75 x O.'I MM/
30 ,Lf'.a. 00
I
'O'•
/' 6
20
I
I
I
I
I
-A ...........
,, .... .,..,,,,,. ..... ------')<
10
I
I 0
• ...
,. ,, ...
;.~
,_,.,,,, , ... "
A/
100 200 300 •100 500 600 700 COOLING RATE ·uu
Fig. 4 Effect of the specimen cooling rate

on disbanding
Before hydrogen charging
Other authors [1Il have found that the disbanding
is generated for cooling rates within 100 to 200°C/h.
WRC Bulletin 534

Studies (~f Cladding and Overlay for Pressure Vessel Service 43
These results lead to two important observations : Influence of the Type of Weld Overlay
- for comparing the results, it is necessary to use the
same cooling rates. In the opposite case, the risk of - for the slow cooling rate (150°C/h), no difference is
errors may be important, observed between all the procedures (Fig. 5). The criti-
- a cooling rate capable of discriminating weld overlay cal disbanding threshold was not reached.
procedures can be applied.
TABLE 2. DISBONDING RESULTS
DISBONDING (%) DISBONDING (%)

TEST PWHT TEST PWHT
NAME TYPE COOLING RATE NAME TYPE COOLING RATE
15o·c;h 25o·c;h 675·c;h 15o·c;h 25o·c;h 675·c;h
R.PWHT 4 8 25 R.PWHT 5 24-30 40
SAW1 -------- -------- -------- ------- -------- -------- --------
D.PWHT 3 ESW4 S.PWHT 30 20-28
------- -------- -------- --------
R.PWHT 1 15 30 D.PWHT 0 0-8
SAW2 -------- -------- --------
D.PWHT 10 R.PWHT 0 30 45
ESW5 ------- -------- -------- --------
R.PWHT 4 10 30-40-45 D.PWHT 12
-------- -------- --------
ESW1 S.PWHT o-4 9 R.PWHT 0 0 40
-------- -------- -------- ------- -------- -------- --------
D.PWHT 8 S.PWHT 20
ESW6 ------- -------- -------- --------
R.PWHT 4 14 40 D.PWHT 8
ESW2 -------- -------- -------- ------- -------- -------- --------
D.PWHT 10 LT.PWHT 0
R.PWHT 0 20 R.PWHT 14 40-50
-------- -------- -------- ------- -------- -------- --------
S.PWHT 2 ESW7 S.PWHT 40-35
ESW3 -------- -------- -------- ------- -------- -------- --------
D.PWHT 0 LT PWHT 10
-------- -------- --------
LT.PWHT 7 HS ESW R.PWHT 25-37
EFFECT OF THE WELDING PROCEDURE - for the medium cooling rate (250°C/h), a significant
difference was observed between the welding procedures
In order to understand the effect of the welding (Fig. 6). The use of wide strips (150 mm) increases the
procedure, we have analysed the interface behaviour under disbanding sensitivity.
the worst conditions (R.PWHT 690°C - 30 hours).
D!SBONDING m DISBONDING (%)
50
"t
40 j 40
WllJE TRIPS
30
~
30
NllRR~ STR f PS
i
20
20
W!IJE STRIPS
= = cs:=
10 10
= =
SPECll\l
30~~
0
0 PROCESS SllW1 Sf\112 ES\11 ES\12 ES\14 ES\15 ES\16 ES\17
~_09L+
308L ~giht
3091.+
o~~rf 308t ;:;r
µvoe>
308L •>34]"' 308l 347
60. 5 • 75 • ZS • 150. 150 • J'.>U • 150 •
•Hw n .....
''HH h 4HH tJ 4HH 0.4191 kJ.•1m l,411'1 ~J.4HH
Fig. 5 Effect of the welding procedure Fig. 6 Effect of the welding procedure
(specimen cooling rate = 150°C/h) (specimen cooling rate = 250°C/h)
WRC Bulletin 534

The Welding Research Council, Inc .
. for narrow strips (less than or equal to 75 mm), the In fact, the cooling rate threshold to discriminate
disbanding level is the same in SAW ans ESW processes, the effect of the welding procedure varies with the strip
even using various parameters (SAW2 and ESW2 tests). size. If the cooling rate is slow, any welding procedure
Micrographic examinations show small amounts of marten- cannot be discriminated. The medium cooling rate only
si te in the weld overlay near the carburized thin mar- discriminates the welding procedures using wide strips.
tensitic layer (Fig. 7). The presence of martensite does The high cooling rate only discriminates the welding
not decrease the disbanding level. procedures using narrow strips.
Effect of the Welding Parameters
The welding parameters have a major effect on the

weld overlay interface microstructure, mainly on the austenitic coarse
martensite grain ratio. The coarse grain size ratio is the percen-
tile length of the interface affected by planar grains
(coarsest grains). The Fig. 9 shows the effect of the
heat input on the coarse grain size ratio.
martensi tis,.
thin layer /\USTENITIC COARSE GRAIN RATIO (%)
material
e S/\111 60 x 0.5 Hl1
x 200 o ESl/1 75 x O.Q H11
50 l> ESW•1 150 x O.Q H11
Fig. 7 Interface microstructure of ESW2 test X ES\13 75 x o,q HM
9 ES\16 150 x 0,11HM
V ES\17 150 x 0,Q HM
. for wide strips (150 mm), only a special chemical com- •10
position (to obtain low silicon and high delta ferrite
in the weld metal) significantly reduces the disbanding
phenomenon. In this case, the disbanding level is less 30
than that observed with the SAW process.
20
- for the fast cooling rate (675°C/h), the same general
behaviour is observed (Fig. 8); therefore, for each cate-
gory of strip size, there exist some differences. 10
DISBONDJNG CU
H/t~ROW STRIPS WIDE STRIPS 0
50 100 150 200 250
50
~j
Fig. 9 Effect of the welding heat input on
QO
= the austenitic coarse grain ratio
30
Furthermore, strips with special chemical composi-
tions give a smaller percentage of planar grains because
20
= the high level of delta ferrite in the first layer pre-
vents the generation of austenitic coarse grains near
SPECIAL the thin martensite layer (Fig. 10).
309 L STRIP
10 ESW4 test
(X 200)
Fig. 8 Effect of the welding procedure

(specimen cooling rate = 675°C/h)
. for narrow strips, the ESW process (ESW1 and ESW2

tests) gives higher disbanding. The use of a strip with ESW6 test
a special chemical composition (ESW3 test) significantly (X 200)
reduces the disbanding. The level obtained is less than
that observed with the SAW process. The high speed (HS)
ESW process has only a beneficial effect compared to the overlay
ESW process .
. for wide strips, the disbanding level is the same for
all the we~ding procedures applied. The use of a strip
with a special chemical composition cannot decrease the
sensitivity to disbanding.
Fig. 10 Interface microstructure of ESW4 and ESW6 tests
WRC Bulletin 534

The use of a narrow strip enables to apply a low On these figures, the points representative of
heat input, thus to obtain a better austenitic micro- strips with special chemistry, are outside the scatter
structure. After R.PWHT, the effect on the disbanding band, containing all the other points. In fact, the dis-
level is depicted on Fig. 11 thru 13. banding phenomenon depends on the coarse grain ratio
(Fig. 14).
D!SBOND!NG m
DISBONDING (%)
50
50
110
SPEClf'EN COOLING RATE = 675°C/H
110
30
30
~~~~~ ~g:g:~: SPEW'EN COOLING RATE 250"C/H
20 DESW1 75•0,qMH
llESW 2 75 • 0.4 MM
)( ESW 3 75 • O,q MH 20
6ESW q 150 x 0,q MH
10 AESW 5 150 • O,q MH
~HSll 6 150 • o,q MH
TESll 7 150 • 0.4 MH 10
?HS ESW 75 X 0.4 MH
SPECil'EN COOLING RATE = 150°C/H
50 100 150 200 250 300 HEAT l NPUT
(KJ/CH)
10 20 30 110 50 AUSTEN ITI C COARSE
GRAIN RATIO (%)
Fig. 11 Effect of the welding heat input
(specimen cooling rate = 675°C/h) Fig. 14 Effect of the austenitic coarse grain ratio
on disbanding
D!SBOND!NG (%)
Some examinations were performed with a transmis-
e SAW1 60 • 0.5
0 SA\12 75 • O,q MH
MM sion electron microscope using accurately located thin
o ESW1 75 • O,q MH foils in order to observe the carburized martensitic
50 111 ES\12 75 • 0.11 MH layer. Fig. 15 shows Maglay interfaces using strips with
6 ES\ill 150 • o,q MM
A ESW5 150 • o.q MM standard and special chemistries. No difference can be
® ESW6 150 x 0.11 MH observed. In both cases, the precipitates M23 C6 are
110 "ES\17 150 • o.q Ml1
similar. Thus, the sensitivity of the carburized layer
is probably similar because the same PWHT was performed.
Therefore, it seems normal to only observe the effect of
the austenitic coarse grain microstructure.
Fig. 12 Effect of the welding heat input

(specimen cooling rate = 250 °C/h)
DlSRONDiNG OD
e SAW1 60 • 0.5 M11

o SAW2 75 • 0.11 MM
50
o ESW1 75 • O.'I MM
11 ESW2 75 x 0.4 MH
X ES\13 75 • 0.4 MM
6 ESWll 150 • 0.'I MM
A ES\15 150 x O.'I MM
110 ® ESl/6 150 • Q,q MM
30
20
10
.. 0
Ill 0
50 100 150 200 250 300 HEAT INPUT

(KJ/CH)
Fig. 13 Effect of the welding heat input Fig. 15 Transmission electron microscope examinations
(specimen cooling rate = 150°C/h) (martensitic thin layer microstructure)
WRC Bulletin 534

EFFECT OF THE POSTWELD HEAT TREATMENTS (PWHT) DISBONDING (%)
Generally, the PWHT is carried out at approximate-

ly 690°C, with a holding time depending on the vessel
thickness and the number of PWHT, including those for 50
any future repairs. For heavy wall vessels, the interface
sensitivity to the disbanding phenomenon is thus increa- •10
sed by a more important carburization. The Welding Ins-
titute has demonstrated that the risk of disbanding can
be reduced by the following factors 30
R.PWHT - • .-----------:
.7~~
- a duplex PWHT (D.PWHT). This type of hIBat treatment
consists in performing a first PWHT at normal tempera-
ture (690°C approximately) to obtain the best mechanical 20
properties for the 2.25 Cr - 1 Mo material. Then after
~·_.,?s.PWHT
cooling down to room temperature a second heat treatment
is carried out at a lower temperature to de-sensitize
the interface.
- a single PWHT at a lower temperature (LT PWHT), for
10
0
.
/
/6
/
e
//
("
\
.. ~ /
- _,/
/
./·
/
D.P~
upgraded 2.25 Cr - 1 Mo steels (e.g. ASME Code Case 100. 200 300 HEAT INPUT
(KJ/tH)
1960). Decreasing the temperature leads to an interface 2 LAYER OVERLAY WITH NORMAL 309 l STR IP
microstructure less sensitive to the hydrogen induced
cracking and to less stresses in the interface region.
FRAMATOME has studied the effect of these solu- Fig. 17 Effect of the PWHT type
tions. (specimen cooling rate = 250°C/h)
Effect of a Duplex Postweld Heat Treatment Electron examinations using thin foils were carried
out (Fig. 18). After RJPWHT (690°C - 30 hours), the
As the disbanding level depends on the welding fresh martensite was not revealed, probably due to the
heat input, we have analysed the effect of the PWHT type low amount of the thin martensitic layer which is auste-
for various welding procedures and each cooling rate nitized and to the difficulties met to accurately posi-
applied to disbanding specimens. tion the thin foils.
2 layer-overlay with standard strip. For various
cooling rates applied to disbanding specimens, the fol-
lowing behaviour was observed (Fig. 16 and 17). M23 C6
DISBOND!NG (%)
0. 22 )-Im
Fig. 18 Transmission electron microscope examinations

(martensite thin layer microstructure)
The S.PWHT gives a full efficiency only if the

heat input is lower than 140 or 150 kj/cm. Over this
critical level, the efficiency is less and less impor-
tant. In this case, some of the austenite generated at
2 LAYER OVERLAY \llTll NORMAL 309 L STRIP 690°C will transform into fresh martensite during the
final cooling down to room temperature.
Fig. 16 Effect of the PWHT type
(specimen cooling rate = 675°C/h) Other welding procedures. The S.PWHT and the
D.PWHT performed on the weld overlay either using 2
Only the D.PWHT gives a full efficiency even for layer special strips or 1 layer standard strip, give
high heat input. The interface de-sensitization mecha- similar results to those obtained using a 2 layer overlay
nism could be explained as follows : with standard strips (Fig. 19 and 20). In fact, no
- presence of a thin martensitic layer, immediately further effect of special strips was observed.
after cladding, The respective effects of S.PWHT and D.PWHT are
- carburization of the interface and partial austeniti- clearly depicted on Fig. 21. For the fastest cooling
zation of the thin martensitic layer during the first rate, the full efficiency of the D.PWHT is confirmed up
heat treatment (690°C - 30 hours). Then generation of to 30 % of austenitic coarse grains. On the contrary,
fresh martensite from the austenite by cooling, the efficiency of the S.PWHT decreases as the austenitic
- tempering of the fresh martensite by the second heat coarse grain ratio increases.
treatment (600°c).
WRC Bulletin 534

In fact, we think that the disbanding phenomenon

DJSBONDJNG (%)
occurs in the thin martensitic layer because the fresh
x 2 LAYER OVERLAY WITH SPECIAL 309 L STRIP
01 LAYER OVERLAY W!Tfl NORMAL 347 STRIP
I S.PWHT COND!TION
martensite is the most sensitive. After it propagates
either into the austenitic coarse grain boundaries or
50 in the thin martensitic layer. Consequently, we can
explain the reason why the D.PWHT is more efficient.
Thus, in order to prevent any risk of disbanding, it is
preferable to control the sensitivity of the thin mar-
/ ·~
IJO
. tensitic layer to preclude the disbanding generation.
. 0 \
S.PWllT SCATTER BAND OF / / . For instance, the presence of a large amount of auste-
30 2 LAYER OVERLAY WITll / \ nitic coarse grains and an important carburization res-
NORM/IL 309 L STRIP / ' /./ pectively induced by the weld overlay and the PWHT
20 ./ /)I. (690°C - 30 hours), both necessitate the performance of
./ /' a D.PWHT. Otherwise, the disbanding level will be high.
/ /
10 ( // Effect of the Low Temperature PWHT (LT PWHT)
100
\ /--:/
200 300 HEAT INPUT
Disbanding tests were carried out after a PWHT at
6505C - 30 hours (called LT PWHT) for three types of
(KJ/CH) weld overlay :
- 2 layer ESW process using special 309 L strips ( 75 x
Fig. 19 Effect of the S.PWHT for various 0.4 mm and 150 x 0.4 mm)
welding procedures - 1 layer ESW process using standard 347 strip (150 x
(specimen cooling rate = 675°C/h) 0.4 mm).
The Fig. 22 compares the disbanding level for
each overlay, after cooling at 675°C/h. The efficiency
DISBONDING CXl
of the LT PWHT is obvious, even for the one layer weld
x 2 LAYER OVERLAY WITll SPECIAL 309 L STRIP <D.PWllT CONDITION) overlay which contains a lot of austenitic coarse
grains (ESW7 test). The disbanding level is less than
50 or equal to 10 %.
DISBONDING <Xl
30 50
20 40
D.PWllT SCATTER BAND OF 2 LAYER
10 :------- --- --------- -----

OVERLAY WITll NORM/IL 309 L STRIP
--~ 30
I I
I I
I I
20
100 200 300 HEAT INPUT
(KJ/CH)
10
Fig. 20 Effect of the D.PWHT for various
welding procedures
(specimen cooling rate = 675°C/h) 0
PROCESS SW7 C150•0.4MHl
DISBOND!NG C%l OVERLAY SPECIAL 309 l + SPECIAL 309 l + 347

TYPE 347 347
&9o•c 650°c 690°C 650"C 690°C 650"C
/.
PWltT
30 H 30 H SPWllT DPWflT 30 H 30 H DPWHT 30 H 30 H SPWllT
40 TYPE RPWllT Tf'lffr l«'lfrr TNlT TPlfJT
30
20 /: Fig. 22 Effect of the low temperature PWHT (LT PWHT)

(specimen cooling rate = 675°C/h)
A PWHT at 650°C has a dual effect :
~--
- the amount of austenite at 650°C is less than that
10
obtained at 690°C. Thus the fresh martensite is less
-~----01 e S.PWllT important after cooling down to room temperature. There-
I o D.PllHT fore, the sensitivity to disbanding generation after
:./0 I
I
LT PWHT is probably less important than after R.PWHT at
10 20 30 40 50 AUSTEN ITl C COARSE 690°C but more important than after D.PWHT.
GRAIN RATIO (%)
- the carburization of the interface (thin martensitic
layer and austenitic coarse grain boundaries) is less
Fig. 21 Effect of the austenitic coarse grain ratio important than after R.PWHT, S.PWHT and D.PWHT. Thus
for S.PWHT and D.PWHT types the sensitivity to disbanding propagation is less im-
(specimen cooling rate = 675°C/h) portant in the coarse grain boundaries.
WRC Bulletin 534

As a conclusion, we can assume that the LT PWHT is [3] SAKAI T., ASAMI K., KATSUMATA M., TAKADA H. and
an other solution to significantly decrease the risk of TANAKA 0. : "Disbanding of Weld Overlay in Pressure
disbanding. But the best solution would be the following Vessels and its Prevention". Proceedings of the
duplex PWHT : 650°C with cooling down to room tempera- First International Conference, American Society
ture+ an otherPWHT at 600°C. In this PWHT, the genera- for Metals, WASHINGTON DC, November 1982, pp. 340-
tion and propagation of the disbanding phenomenon would 348.
be very low.
[4] COUDREUSE L., GERARDIN D., CHARLES J., BLONDEAU R.
SOLUTIONS PROPOSED TO PREVENT THE DISBONDING PHENOMENON and PALENGAT R. : "Etude du Phenomene de Disbanding
a l'Interface entre Divers Revetements en Aciers
We want not to apply solutions that would lead to Inoxydables Austenitiques et l'Acier 2.25 Cr - 1 Mo".
obtain any martensite in the first layer of weld overlay Cinquiemes Journees d'Etude sur les Appareils a
or high welding speed, or high delta ferrite in the first Pression, AFIAP, Vol. 3, 1986, p. 179-194.
layer. For one or two layer overlays, using 75 x 0.4 mm
strips, we propose the following solutions to avoid any [ 5] GI TOS MF,, ROBINSON JL. and GOOCH TG. : "Disbanding
risk of disbanding : of Stainless Steel Cladding in High Temperature
- for the pressure vessels as stress relieved at 690°C, Hydrogen Service", internal report.
performance of a duplex heat treatment 690°C + 600°C -
5 hours,
- for the pressure vessels as stress relieved at a low
temperature, performance of the duplex PWHT : a first
PWHT at low temperature, then a second PWHT at 600°C -
5 hours.
CONCLUSIONS
The various tests carried out enabled to emphasize

interesting factors :
- the disbanding significantly depends on the cooling
rate applied to the specimens from the autoclave. As a
result, in order to compare the result reliability, this
experimental parameter has to be taken into account.
- after regular PWHT (690°C - 30 hours), the risk of
disbanding increases with the amount of austenitic
coarse grains at the interface. The amount of austenitic
coarse grains can be limited by acting on the heat input.
But the risk of disbanding can be still important since
the thin martensitic layer is sensitive (presence of
fresh martensite).
- to certainly prevent any risk of disbanding, an appro-
priate PWHT must be performed to act on the thin marten-
sitic layer sensitivity :
. a duplex PWHT (690°C - 30 hours + 600°C - 5 hours) is
fully efficient even if the level of austenitic grain is
important,
. a low temperature PWHT (650°C) is as efficient as the
duplex PWHT.
From these key points, simple rules can be stated
for various parameters (strip width, number of layers),
enabling to apply the Maglay process capable of obtai-
ning :
- a high productivity,
- an excellent metallurgical quality of the weld overlay,
- an excellent resistance to the disbanding phenomenon
by the use of a duplex postweld heat treatment.
REFERENCES
[1] DEVAUX J.C., DUBOIS D. and LEBLOND J.B. : "Computer

Simulation of Hydrogen Diffusion in Stainless Clad
Low-Alloy Steels". Proceedings of the First Inter-
national Conference, American Society for Metals,
WASHINGTON DC, November 1982, pp. 45-52.
[2] HOKADA H., NAITO K., WATANABE J., OHNISHI K. and

CHIBA R. : "Disbanding of Stainless Weld Over lay in
Hydrodesulfurizing Reactor". Proceedings of the
First International Conference, American Society for
Metals, washington de, November 1982, pp. 331-339.
WRC Bulletin 534


WRC 534-2009

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 534-2009

Uploaded by

Copyright:

Available Formats

INTENTIONALLY LEFT BLANK

For more information, see www.forengineers.org

Copyright © 2009 The Welding Research Council.

Welding Research Council

STUDIES OF CLADDING AND OVERLAY FOR

Martin Prager, Editor

WRC Bulletin 534

THE WELDING RESEARCH COUNCIL, INC.

WRC Bulletin 534

ISBN No. 1-58145-541-0

WRC Bulletin 534

Considerable research aimed at understanding hydrogen-steel interactions has been

WRC Bulletin 534

WRC Bulletin 534

HYDROGEN IN THICK-WALLED, HIGH-TEMPERATURE HYDROGEN REACTORS DURING

RESISTANCE TO HYDROGEN INDUCED DISBONDING OF HOT ROLLED HEAVY GAUGE

CORROSION RESISTANT CLAD PLATES FOR HIGH PRESSURE VESSELS IN HYDROGEN

EVALUATION OF STAINLESS STEEL EXPLOSION CLAD FOR HIGH TEMPERATURE, HIGH

HYDROGEN INDUCED DISBONDING IN WELD-CLAD PRODUCTS: THE INFLUENCE

DISBONDING MECHANISM AND ITS PREVENTION .................................................................................. 41

WRC Bulletin 534

WRC Bulletin 534

HYDROGEN IN THICK-WALLED, HIGH-TEMPERATURE HYDROGEN REACTORS

Reflection of APl/MPC Work and Direction for Future Research

K. Smit and P. F. Ivens

ABSTRACT reactors on which - in the light of research results

INTRODUCTION A major improvement of hot tensile and creep properties

WRC Bulletin 534

properties or the stress to cause rupture in air in

CALCULATED HYDROGEN DISTRIBUTIONS THROUGH THE REACTOR 1000

Models to calculate the hydrogen concentration profile

- Hydrogen Solubility and Diffusivity -

Values for A and B for 2.25Cr-1Mo steel have been

The solubility lines are given in Fig. 2. The top and

WRC Bulletin 534

The literature data on the diffusivity in the cladding

cladding: 7.1 10 4555 r

The maximum amount of hydrogen that dissolves during A: 42.7 7.6

WRC Bulletin 534

Disbanding of the cladding is generally considered as a

The risk of disbanding is mainly dependent on the steel

WRC Bulletin 534

consideration that diffusion rates below 400 'F (-200

In our view the Geiger's model underestimates the

For the symbols see the appendix. Flaking will occur

WRC Bulletin 534

9) Prediction of Creep Crack Growth from Uniaxial Creep

WRC Bulletin 534

RESISTANCE TO HYDROGEN INDUCED DISBONDING OF HOT ROLLED HEAVY

Hideto KIMURA, Minoru SUW A, Sadahiro YAMAMOTO

ABSTRACT Sufficient disbonding resistance was also found after

WRC Bulletin 534

Figure 3 PRODUCTION OF ROLLED AND OVERLAY WELDED CLAD STEELS

WRC Bulletin 534

mass bonding variety in met- resistance to dimension

(improved*) good good good good easy

Table 2 CHEMICAL COMPOSITION OF BASE STEEL AND CLADDING METAL

WRC Bulletin 534

Table 3 CLADDING CONDITION IN LABORATORY

ROLLED hot rolling 1200 °C reheat (200+ 20)-+(50+ 5)mmt

heat treatment 900 'C X 30min AC+ 720 'C X 30min AC

OVERLAID welding materials 1st pass PFB-1/USB-309L

2nd pass PFB-1 FK/USB-347L

welding condition DC foop + 800A, 26V, l 90mm/min