Accepted Manuscript

Some studies on P91 steel and their weldments

C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini

PII: S0925-8388(18)30121-X
DOI: 10.1016/j.jallcom.2018.01.120
Reference: JALCOM 44584

To appear in: Journal of Alloys and Compounds

Received Date: 8 October 2017

Revised Date: 4 December 2017
Accepted Date: 9 January 2018

Please cite this article as: C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Some studies on P91 steel
and their weldments, Journal of Alloys and Compounds (2018), doi: 10.1016/j.jallcom.2018.01.120.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Some studies on P91 steel and their weldments

C. Pandey1, M. M. Mahapatra2, P. Kumar1, N. Saini1

1
Department of Mechanical and Industrial Engineering
Indian Institute of Technology Roorkee, Uttrakhand-247667, India.

PT
2
School of Mechanical Sciences, Indian Institute of Technology
Bhubaneswar, Odisha 751013, India

RI
*Corresponding author: [email protected], [email protected]
Telephone: +91-1332285518 Fax No. +91-1332285665, Country/Area code: 247667

SC
1 Introduction
2 Key alloying element
2.1 Chromium (Cr)
2.2 Molybdenum (Mo) and tungsten (W)

U
2.3 Vanadium (V) and Niobium (Nb)
2.4 Nickel (Ni) manganese (Mn) and cobalt (Co)
AN
2.5 Carbon (C) and nitrogen (N)
3 Microstructural feature and precipitates morphology
3.1 M23C6
M

3.2 MX
3.3 Laves phase (Fe, Cr)2 (W, Mo)
D

3.4 Z-phase
TE

4 Heat treatment characteristic of P91 steel

5 Weldability of P91 steel
5.1 Weld metal (Tp˃Tm)
EP

5.2 Grain growth zone or CGHAZ (Tm˃Tp˃˃Ac3)

5.3 FGHAZ or grain refined zone (Tϒδ ˃Tp˃Ac3)
C

5.4 ICHAZ or partially transformed zone (Ac3 ˃TP ˃Ac1)

5.5 Over-tempered base metal (TP < Ac1)
AC

5.6 Summary
6 Influence of welding process parameters and heat treatment on characteristic of P91
welds
7 Diffusible hydrogen in P91 welds
7.1 HIC mechanism in weld
7.2 Diffusible hydrogen (HD) measurement

1
 ACCEPTED MANUSCRIPT

7.2.1 Glycerin method

7.2.2 Mercury method
7.3 Effect of diffusible hydrogen on weldability
8 Shrinkage and residual stress in weldments
9 Fundamental of creep

PT
10 Creep characteristic and microstructure evolution in P91 base metal
10.1 Precipitate coarsening

RI
10.2 Martensite lath widening and recovery
10.3 PAGBs disappearance

SC
10.4 Emergence of subgrain boundaries
10.5 Precipitation of inter-metallic phase (Laves and Z-phase)

U
11 Creep characteristic of P91 weldments
12 Future work
AN
13 Conclusion

Abstract:
M

In present research article, microstructure evolution in P91 steel and their weldments are
D

reviewed in as-virgin and heat treatment and creep exposure condition. The thermal stability of
P91 steel is derived from solid solution strengthening, sub-grain hardening and precipitation
TE

hardening. The initial microstructure plays an important role in deciding the mechanical
properties of P91 steel and their weldment in long-term ageing and creep exposure condition.
EP

Effects of various alloying elements present in P91 steel and their related phase have also been
discussed in details. The role of grain coarsening, Cr/Fe ratio, lath widening and dislocation
density on creep rupture life of base metal and weldments are discussed. The combined effects of
C

lath martensitic microstructure, residual stress and diffusible hydrogen content on performance
AC

of P91 steel material are also discussed.

Keywords: P91; Normalizing; Tempering; Residual stress; Diffusible hydrogen; Weldments.

1 Creep strength ferritic steels: Introduction

In recent years, there has been renewed efforts to reduce the environmental pollution from the
thermal power plant by significantly decreasing the CO2 emission which has led to endeavors to

2
 ACCEPTED MANUSCRIPT

improve the thermal efficiency of power plants. By increasing the operating temperature and
pressure of thermal power plants, improved efficiency and hence reduction in the CO2 emission
are achieved. Hence, any improvement in the efficiency of thermal power plant can thus have a
beneficial effect on environment conditions [1–3]. The main problem associated with increasing
in operating temperature and pressure of power plants is related to the suitability of material as

PT
components operating at such high temperature undergo microstructural changes with time and
possible degradation such as creep, fatigue, hydrogen induced cracking, embrittlement, corrosion

RI
and oxidation. The high operating temperature and pressure requirement of such modern day
power unit leads to the development of high temperature creep strength enhanced ferritic (CSEF)

SC
steels. For nuclear and thermal power plants, creep strength enhanced ferritic (CSEF) steels has
been considered as superior material over austenitic stainless steel due to its low coefficient of

U
thermal expansion, good thermal conductivity and high creep rupture strength [4–6]. Apart from
high creep strength, CSEF steels have some other good properties such as ductility, toughness,
AN
high resistance to stress corrosion cracking and oxidation. Good weldability and microstructural
stability at elevated temperature for a long time is also an attractive property of CSEF steels [7].
M

Therefore, CSEF steels are used for pipelines, power unit component being the parts of boiler
and steam generators that subjected to high operating temperature. CSEF steels are commonly
D

known as Cr-Mo alloys steel which contain 9-12 % Cr. The development of CSEF steels started
from last few decades, starting from P9, P91, and P92.
TE

In mid 1960s, 12%CrMoV steels were developed for the fabrication of thin and thick walled
power plants components. The operating temperature for such components was 565 oC. The
EP

creep strength of 12% steels was derived from the solid solution strengthening and precipitation
strengthening. The most of 12% Cr steels were found to be failed in long-term creep exposure
C

due the unexpected precipitation of Z-phase (Cr, V, Nb) N [8,9]. Recent investigators have
AC

demonstrated that Z-phase precipitation is strongly accelerated by the high Cr content [10,11]. In
order to further enhance the creep strength of 12% Cr steels, the Cr percentage brought down to
9 %. The 9% Cr CSEF steel was developed first at Oak Ridge National Laboratory (ORNL)
under the designation P9 (9Cr-1Mo) [12].Over the years, improvements were incorporated in
CSEF 9Cr-1Mo steel by the addition of strong carbide and nitride forming elements such as V
and Nb with controlled amount of N, and the modified steel was designated as Grade

3
 ACCEPTED MANUSCRIPT

91(X10CrMoVNb 9-1) steel. Grade 91 steel, also known as modified 9Cr-1Mo-V, is designated
as P91 for pipe, T91 for tube, and ASTM A335 for plate.
2 Key alloying element
The optimum alloy design for the P91 steel is necessary to obtain the 100 % martensitic
microstructure and free from δ-ferrite. Swindeman et al. [13] reported a method, on Chromium–

PT
Nickel Balance (CNB), to balance the composition of P91 steel and given by modified
Newhouse’s equation (1);

RI
CNB = Cr + 6Si + 4Mo + 1.5W + 5 Nb + 9Ti + 11V + 12 Al − 40C − 30 N − 2Mn − 4 Ni − 1Cu (1)

SC
To ensure the absence of delta ferrite, the CNB value should be less than 10. In the range of 10–
12, the δ-ferrite patches might be observed in the microstructure. The δ-ferrite patches will be

U
observed in microstructure for CNB above 12. Chemical composition of P91 steel is depicted in
Table 1. The elements present in the composition play an important role in microstructure
AN
stabilization during creep exposure.

2.1 Chromium (Cr)

M

Cr is principle alloying element in P91 steel and acts as the ferrite stabilizer. Cr provides the
solid solution strengthening with addition to iron (Fe). Cr attributed the creep strength to P91
D

steel by forming the Cr-rich M23C6 type carbides precipitates. M23C6 enhances the creep strength
TE

of P91 steel by pinning the movement of sub-grain boundaries and dislocations [14]. At high
concentration, Cr enhances the wear resistance, corrosion resistance and hardenability of steel.
EP

Higher Cr content have reported a negative effect on impact properties [15]. Due to ferrite
stabilizer, higher Cr content above 13% leads the formation of δ-ferrite.
C

2.2 Molybdenum (Mo) and tungsten (W)

AC

Mo and W are added to enhance the solid solution strengthening of P91 steel and act as ferrite
stabilizer. Depends on other ferrite and austenite stabilizers, amount of W and Mo must be in
limit to avoid the formation of δ-ferrite. W and Mo impede the motion of free dislocations by
introducing the local strain with in matrix. The W and Mo enhanced the solid solution
strengthening in terms of increase in hardness and creep strength value [16]. For short-term
creep, Mo addition extend the primary creep stage and enhance the creep strength by retarding
the coarsening rate of M23C6 [17]. For long term creep, W and Mo addition lead the formation of

4
 ACCEPTED MANUSCRIPT

laves phase (Fe2W, Fe2Mo). The precipitation of laves phase reduces the wt % of W and Mo in
solid solution which resulted in reduction of solid solution strengthening [18]. The formation of
laves phase in short-term creep exposure enhances the creep strength due to precipitation
hardening and it also provide the pinning effect to the movement of sub-grain boundaries
[19,20]. In long-term creep exposure, precipitation of laves phase reduces the creep rupture

PT
because of higher coarsening rate [21]. Mo and W also helped to reduce the martensitic start (Ms)
temperature and increase the lower critical temperature (Ac1).

RI
2.3 Vanadium (V) and niobium (Nb)

SC
V and Nb in P91 steel act as strong carbide, nitride and carbo-nitride former of type MX (M: V,
Nb; X: C, N) or [M(C, N)]. The formation of fine MX type precipitates enhances the
precipitation strengthening. MX precipitates help to impede the dislocation movement and also

U
slowed down the recovery rate of dislocation structure [22,23]. Sometimes addition of V results
AN
in poor strength and fracture toughness due to formation of δ-ferrite [24]. Both V and Nb
addition promote the formation of carbide precipitates that reduces the carbon content in matrix
M

and enhances the corrosion strength of steel.

2.4 Nickel (Ni) manganese (Mn) and cobalt (Co)

D

Ni, Mn, and Co are added to ensure the 100 % austenitic transformation during the austenitizing
TE

process which further transformed into 100 % martensite on cooling. Ni and Mn have found
same effect on microstructure evolution but Mn acts as poor austenite stabilizer. Addition of Ni
and Co ensures the good toughness. Addition of Ni and Mn in P91 steel promote the formation
EP

of unwanted M6 (C, N) phase that shows high coarsening rate during the creep exposure and also
destabilize the M23C6 [25]. Addition of Co shows the poor solid solution strengthening in iron
C

(Fe) comparison to Mn and Ni. Gustafson and Ågren [26] had reported that Co enhanced the
AC

creep strength by reducing the coarsening rate of M23C6 and it also reduced the curie point. Co is
also added to suppress the δ-ferrite formation [25]. Ni and Mn content have observed a strong
influence on the lower critical temperature (Ac1) [27].

2.5 Carbon (C) and nitrogen (N)

5
 ACCEPTED MANUSCRIPT

C and N are essential for the carbide and carbo-nitride formation that provide the precipitation
hardening and enhance the creep strength. Both C and N show higher solubility in austenite
comparison to ferrite.

3 Microstructural feature and precipitates morphology

PT
P91 steel is subjected to normalizing and tempering treatment to achieve the stable martensitic
microstructure. For normalizing treatment, as-received P91 steel is heated in austenitizing

RI
temperature range that results in formation of carbon saturated solution of α-iron. Normalizing
treatment leads to formation of untempered lath martensitic microstructure with negligible

SC
precipitates. The precipitates that stabilize the grain boundary are found to be dissolved at
normalizing temperature [28]. Tempering of P91 steel is performed in temperature range of 760-
788 oC for tempering time of range 1-3 h to make homogenous distribution of carbides inside

U
fully tempered martensitic structure [29]. At initial stage of tempering Cr-rich M3C particles are
AN
evolved at the lath boundaries along with fine MX precipitates. As tempering proceeds, M3C get
replaced by more stable trigonal M7C3 (M: Fe, Cr, Mn) which further converted to stable cubic
M

structure of M23C6 (M: Fe, Cr, Mo) type. The sequence of precipitation are given below [30];

Solid .solution → M 3C + MC → M 3C + MC + M 23C 6 → M 3C + MC + M 23C 6 + M 7 C 3 →

(2)
D

MC + M 23C 6 + M 7 C 3
TE

During the tempering treatment, composition of M23C6 is found to be changed with continuous
increase in Cr/Fe ratio up to 1. Increase in Cr/Fe ratio confirms the great stability of Cr-rich
M23C6 phase [28]. The increase in Cr/Fe ratio attributed to the evolution of metastable M23C6
EP

phase due to thermal activation driven diffusion of Cr into the carbides and M23C6 progresses
towards achieving the thermally stable Cr23C6 configuration [31]. For higher ductility, high
C

tempering temperature is preferred while for large components having required high strength,
AC

low tempering temperature is referred. Swindeman et al. [13] reported the transformation
temperatures for P91 steel which is depicted in Table 2. The normalizing of P91 steel above Ac1
temperature resulted in transformation of austenite (FCC) into lath martensite (BCT). 100 %
transformation of austenite into martensitic is necessary for optimum strength. Incomplete
transformation of austenite resulted in the formation of retained austenite that led to poor
strength before tempering. A schematic evolution of precipitates during the normalizing and
tempering treatment is shown in Fig. 1. In our previous analysis, detailed study of P91 steel in
6
 ACCEPTED MANUSCRIPT

normalized condition and tempered condition have been represented [32–34]. Secondary electron
micrograph of P91 steel, in normalized condition is shown in Fig. 2. In normalized condition,
spatial orientation of columnar laths inside the PAGBS are clearly observed, as shown in Fig.
2(a). At higher magnification, undissolved fine precipitates in order of nanometer are shown in
Fig. 2(b). The optical micrograph clearly shows the columnar laths formation inside the PAGBs.

PT
The micrograph of P91 steel in normalized and tempered (N&T) condition is shown in Fig. 3. In
N&T condition microstructure shows the several elements such as prior austenite grain

RI
boundaries (PAGBs), blocks, laths, sub-grains, packets, carbide precipitates, and other
precipitates. The prior austenite grain boundaries (PAGBs) and lath boundaries are decorated

SC
with fine M23C6 in different morphologies and sizes ranging from 116 nm to 285 nm [34]. The
matrix region confirmed the fine MX type [M: V, Nb and X: C, N] precipitates having size

U
varying from 75 nm to 185 nm. The average size of precipitates presents both at the grain
boundaries and grain interior was measured to be 145 ± 40 nm. The PAGBs are clearly indicated
AN
by the black dotted line in SEM micrograph as shown in Fig. 3(a-b). M23C6 carbide precipitates
and MX type fine precipitates decorated along PAGBs and in matrix region are clearly indicated
M

by red circle in SEM micrograph at lower magnification as well as higher magnification as

shown in Fig. 3(a-b). Fig. 3(c-d) indicates the EDS spectra of white particles present at the
D

PAGBs and within the matrix. The EDS spectra of white particles present at the PAGBs
confirmed that white particles were enriched with Cr, Fe, and Mo, and Mn, formed the carbide
TE

precipitates of M23C6 types as also reported by Arivazhagan et al. [35]. The EDS spectra of EDS-
spot 2, present within matrix region indicate the formation of fine M23C6 precipitates, lean in Cr,
EP

Mo, Mn, and Nb compared to M23C6 precipitates at the grain boundaries, as shown in Fig. 3(d).
The average grain size was measured about 35±12.54 µm using the SEM micrograph and good
C

agreement was found with grain size measured from the optical micrograph of N&T P91 steel
metal which was about 41±10.45 µm. The typical optical micrograph of P91 steel is shown in
AC

Fig. 3(e).

The morphology of precipitates is stated below:

3.1 M23C6

Cr-rich M23C6 type carbide precipitates with face centered cubic structure (lattice
parameter=1.057-1.068 nm) appears at early stage of the tempering process. Ni, Mo Mn and Fe

7
 ACCEPTED MANUSCRIPT

also encountered in M23C6 composition during the early stage of tempering but Cr/Fe ratio is
found to be increased with tempering duration. M23C6 precipitates are generally nucleated along
the prior austenite grain boundaries (PAGBs), lath boundaries and sub-grain boundaries [36].
The M23C6 precipitates provide pinning effect to subgrain boundaries and retards the movement
of boundaries under loading condition [25]. This result in large driving force required for the

PT
martensite recovery and ultimately improved creep strength is achieved. The sizes of M23C6
phase in the initial stage of tempering play an important role during the service exposure. Coarse

RI
M23C6 precipitates are formed along the PAGBs, and lath boundaries while fine M23C6 are inside
the intra-lath region.

SC
3.2 MX

Formation of fine MX precipitates occurs in the presence of strong carbide and nitride former.

U
MX precipitates have face centered cubic structure with size varying from 25-50 nm. MX
AN
precipitates inside the intra-lath region enhance the creep strength of P91 steel by pinning free
dislocation movement. In P91 steel, MX phase forms the V-rich VN or Nb-rich Nb (C, N).
M

Different morphology of MX in P91 steel is reported by Maruyama et al. [25]. Type I coarse
NbX phase shows higher thermal stability compared to other MX phase and remain undissolved
D

after normalizing treatment. Fine platelet VX and spherical NbX type precipitates are formed
during the tempering treatment and shows higher stability during creep. The VX ‘wings’ shows
TE

the morphology of MX precipitates, nucleate on the NbX particles (Type I) during the tempering
treatment and creep test. The fine MX precipitates are distributed uniformly inside the intra-lath
EP

structure. The VN and NbC precipitates are observed during the initial stage of tempering
process and fraction area is found to be increased during the creep exposure [25,37]. Jones et al.
[28] stated the MX particles are partially dissolved during the normalizing treatment and remain
C

in the microstructure in the temperature range of 600-1240 oC. Maruyama et al. [25] reported
AC

that the precipitate remain undissolved are mainly Type I NbX. The undissolved precipitates
limit the austenitic grain growth during normalizing treatment. The fraction area of MX
precipitates is found to be decreased with increase in normalizing temperature while size of MX
precipitates increased [38]. The schematic evolution of M23C6 and fine MX precipitates along the
prior austenite boundaries, lath boundaries, packet boundaries and inside intra-lath region is
shown in Fig. 4 [39].

8
 ACCEPTED MANUSCRIPT

3.3 Laves phase

In P91 steel, laves phase is an intermetallic phase of W and Mo which may recognized as (Fe,
Cr)2 (W, Mo) [40,41]. Laves phase form a hexagonal structure with lattice parameter 0.473 (a)
and 9.772 (c) nm. In normal tempering range, laves phase has never been seen like MX and

PT
M23C6. At initial stage of tempering process, laves phase remain absent but start forming during
the creep exposure. At early stage of precipitation, it shows short bar-like shape. The main
constituent of laves phase (W and Mo) are generally segregate along the PAGBs and lath

RI
boundaries during tempering treatment and provide the preferential sites for laves phase
nucleation during creep exposure. The laves phase precipitation swallows the M23C6 carbides by

SC
nucleating and growing on it and form a cluster around the PAGBs after certain time [18]. This
resulting, the PAGBs have free from the pinning effect of M23C6 and other precipitates. Hence,

U
laves phase formation limits the life of component under high-temperature service condition.
AN
Laves phase of different size and shape are distributed heterogeneously inside the matrix. In
9CrW steel, nucleation and growth of laves phase was observed for first 10000 h, at creep
exposure of 600 oC and 650 oC and after that slow coarsening rate was reported for both the
M

temperature [20]. The slow coarsening rate of laves phase was also reported by Hatterstand and
Andren [42]. The growth rate and coarsening of laves phase is positively affected by alloying
D

element as W, Co and Mo. Precipitation of laves phase lead the reduction in wt. % of W and Mo
TE

in solid solution that resulting in decrease in solid solution strengthening and creep strength of
steel. Panait et al. [43] have reported that the bigger size laves phase acts as cavity trigger. The
laves phase of average size more than 130 nm shifted the ductile mode of fracture to brittle. This
EP

resulted in sudden loss of creep strength. Laves phase precipitation reduced the Charpy
toughness of P91 steel [44]. For short term creep exposure, the precipitation of laves phase
C

enhance the creep strength by retarding the recovery of sub-grain structure and also provide the
AC

pinning effect to boundaries [45]. For long-term creep exposure, positive effect of laves phase
lost early because of high coarsening rate. Laves phase formation in early stage of creep
exposure decrease the minimum creep rate but high coarsening rate promote the creep rate
acceleration.

3.4 Z-phase

9
 ACCEPTED MANUSCRIPT

In P91 steel, Z-phase precipitation occurs during the long-term service exposure in temperature
range of 650-800 oC. Z phase (CrNbN) was first detected in Nb alloyed creep resistant steel and
proposed a body central tetragonal structure. Modified Z-phases was introduced in 1985 in 18Cr-
12Ni-VNbN steel. The modified Z-phase is formed by replacing the half of Nb form Z-phase.
The modified Z-phase having unit cell formula Cr(V,Nb)N shows tetragonal structure with

PT
lattice parameter a=0.286 nm and, c=0.739 nm. The original Z-phase shows tetragonal structure
with lattice parameter of 0.304 and 0.739 nm. The CrVN shows the identical structure as

RI
modified Z-phase [46]. Sawada et al. [47] have observed that the composition of Z-phase vary
with creep exposure temperature and time. In Z-phase composition, Cr and V content were

SC
observed to be increased with increase in creep exposure time, while Nb content decreased.

The modified Z-phase comes out after long term exposure at temperature about 600-700 oC [10].

U
The dissolution temperature of modified Z phase was reported about 800 oC which is much
AN
lower than the solution temperature of original Z-phase (1200-1250 oC) [10]. The heating of
creep exposure steel above 800 oC resulted in formation of MX nitride from the Z-phase [48].
The Z-phase was observed as more stable nitride in P91 steel compared to other phase like
M

modified Z-phase and CrVN.

D

Cipolla et al. [9] have studied the evolution of Z-phase for varying creep exposure temperature in
range of 600-700 oC. The fastest precipitation of Z-phase was observed about 650 oC. The
TE

crystallographic transformation of cubic crystal structure to tetragonal lead to formation of Cr(V,

Nb)N from (V, Nb)N. The two possible mechanism has been reported for the nucleation of Z-
EP

phase [48], as shown in Fig. 5. First one is nucleation of Z-phase on existing V and Nb-rich
nitride along the PAGBs or packet boundaries. Other mechanism deal with the thermally driven
diffusion of Cr from matrix to MX precipitates. In P91 steel, diffusion mechanism of Cr is
C

responsible for the Z-phase nucleation. The transformation of MX to Z-phase takes lot of time
AC

due to slow diffusion of Cr into MX at 650 oC. After the nucleation, Z-phase grows very fast and
dissolves all the surrounding MX precipitates. The diffusion of Cr to MX is rate controlling
processes and mainly controlled by the Cr level in the matrix. The long-term creep exposed
specimen reveals the presence of both coarse and fine Z-phase particles that indicate the slow
and continuous nucleation of Z-phase.

10
 ACCEPTED MANUSCRIPT

PAGBs and packet boundaries provides the preferential location for Z-phase precipitation by
consumption of V and Nb-rich nitrides [9]. Danielsen and Hald [10] developed a thermodynamic
model for calculation of driving force for Z-phase precipitation and found that Cr has strong
influence on Z-phase precipitation compared to V and Nb. Higher Cr content produced large
driving force for Z-phase precipitation. Hence, precipitation of Z-phase in 12%Cr steels has

PT
been reported too fast compared to 9% Cr (P9, P91) steels. N is other important element that
accelerate the formation of Z-phase [49]. In P91 steel, C is not involve directly in the formation

RI
of Z-phase but affect the driving force required for Z-phase precipitation. In P91 steel, C forms
the Cr-rich M23C6 type carbide precipitates and thus lowers the Cr content in the matrix and

SC
ultimately reduces the growth of Z-phase. Hence, formation of Z phase in P91 steel is
responsible for immature creep failure since it is associated with the consumption of V and Nb-

U
rich fine MX particle which make an important contribution to creep strength by pinning the
movement of dislocations.
AN
The episilon (Fe2.4C), cementite (Fe3C), chi (Fe2C), M2X (M: Cr), M6(C, N), M7C3 and
Cr3C are other phases which observed in P91 steel microstructure in different stage of heat
M

treatment.
D

4 Heat treatment characteristic of P91 steel

P91 steel was developed by adding the strong carbide and carbonitride former element such as
TE

0.2 wt.% V, 0.08 wt.% Nb and 0.05 wt.% N [24] in P9 steel. Normalizing of P91 steel is
generally carried out in austenitizing temperature range of 1040-1060 oC for 20-40 min, followed
EP

by air cooling [33]. The subsequent tempering is performed in temperature range of 730-780 oC
for 1-2 h, followed by air cooling. The strength of P91 steel is primarily derived from their stable
C

microstructure. The stability of microstructure is governed by the tempered martensitic lath

structure, sub-grain size, prior-austenite grain boundaries (PAGBs), lath width, lath boundaries,
AC

precipitates size and their distribution inside the structure, dislocation density, precipitate
morphology and their amount [50,51]. Presence of high dislocation density, pinning precipitates
like Cr-rich M23C6 [M: Cr, Fe, Mn, Mo], V and Nb-rich MX and solute atom result a large
driving force required for recovery that produce a stable structure as required for high creep
strength. The mechanical properties of P91 steel are governed by the solid solution hardening,
sub-grain size hardening, and precipitation hardening. The microstructure and mechanical

11
 ACCEPTED MANUSCRIPT

properties of P91 steel was found to be more sensitive to normalizing and tempering temperature
and time. The recent development and finding related to the effect of normalizing and tempering
treatment on microstructure evolution and mechanical properties of P91 steel are discussed
below.
Shrestha et al. [52] studied the effect of varying normalizing temperature (1020-1100 oC),

PT
normalizing time (2-8 h), tempering temperature (690-790 oC), tempering time (2-20 h) on
microstructure evolution and hardness of P91 steel. Martensitic lath size and PAGs size was

RI
observed to be increased with increase in normalizing temperature (fixed normalizing time) and
time (fixed normalizing temperature). The hardness of steel was found to be decreased with

SC
increase in normalizing temperature and time. Hardness of P91 steel was reported to be
decreased with increase in tempering temperature upto lower critical temperature (Ac1). The
reaction associated with thermal phase transformation for Ac1, Ac3, 1270 oC temperature are
given below;
U
AN
α + M23C6 + (V , Nb)(C, N) →α + M23C6 + (V , Nb)(C, N) + γ ...............820( Ac1)o C (3)

α + γ + M23C6 + (V , Nb)(C, N) → γ + (V , Nb)(C, N)...............................870( Ac3 )o C (4)

M

γ + (V, Nb)(C, N) →γ +δ + Nb(C, N)......................................................1270o C (5)

D

The dissolution of M23C6 precipitates started above Ac1 temperature and found to be completed
at Ac3 temperature. V-rich MX precipitates was also dissolved at temperature of 1270 oC. The
TE

dissolution of precipitates results in the higher availability of C and N in matrix which lead to
increase in solid solution strengthening. Effect of varying normalizing temperature (1010-1100
EP

o
C) and tempering temperature (730-780 oC) on microstructure evolution and mechanical
properties were investigated by Golañski [53] . Strength properties were reported to be optimum
C

in normalizing temperature range of 1010-1040 oC while optimum plastic properties like Charpy
toughness and ductility were obtained in temperature range of 1070-1100 oC. Increase in
AC

tempering temperature from 730 oC to 780 oC led to reduction in strength and increase in plastic
properties. Golañski and Słania [54] have reported similar results for cast and forged P91 steel.
Jones et al. [28] have studied the effect of cooling medium and varying tempering temperature
(650-760 oC) on precipitate evolution. Annealing of tempered sample was also performed for
varying annealing temperature (550-650 oC) and time of 1000 h. In air-cooled condition, coarse
lath structure was obtained clearly due to slow cooling rate. In water quenched condition,

12
 ACCEPTED MANUSCRIPT

microstructure was found to be free from precipitates except residue precipitates of NbC and
(Fe,Cr)3C that remain undissolved during normalizing. For low tempering temperature of 550 oC,
hardness remain unaffected with tempering time while at higher tempering temperature, hardness
decreased continuously with increase in tempering time and stabilize after 1 h. V-rich V-Cr-Nb
precipitates was observed just after 16 min of tempering at 760 oC. Cr-rich Cr-V-Fe precipitates

PT
was also observed after 16 min of tempering and converted to more stable Cr-rich (Cr, Fe,
Mo)23C6 after 30 min. A continuous increase in Cr/Fe ratio was also reported with increase in

RI
tempering time at tempering temperature of 760 oC. Lower tempering temperature (650 oC and
700 oC) has not found a significant effect on dislocation structure and lath morphology of

SC
normalized sample. The precipitates evolution at low tempering temperature was also less as
compared to tempering at 760 oC for 1 h of tempering time. Normalizing temperature have also

U
found a great effect on precipitation morphology of fine MX and M3X precipitates [38]. The
precipitates present in microstructure of P91 steel was dissolved during normalizing treatment
AN
except MX and M3C. The size of M3C was reported to be independent with normalizing
temperature while MX size increased with increase in normalizing temperature. The fraction area
M

of both precipitates was observed to be decreased with increase normalizing temperature.

Change in chemical composition of MX precipitates was also observed with increase in
D

normalizing temperature. Yoshino et al. [55] also reported that spherical MX and platelet of M3C
was distributed heterogeneously in normalized condition of P91 steel. The formation of M3C
TE

might be happened during the cooling from normalizing temperature in lean region of V and Nb
in matrix. Noreña et al. [56] had performed the study related to the evolution of minor phases
EP

under different normalizing and tempering condition. In tempering temperature range of 300-400
o
C, microstructure of P91 steel was mainly consisted of Nb-rich spherical shape MX particles
C

which were not dissolved during normalizing. The evolution of V-rich MX precipitates was
observed for tempering above 600 oC. The size of V-rich MX precipitates was less compared to
AC

Nb-rich MX precipitates. The Fe-rich M3C precipitates were not dissolved completely during
normalizing temperature up to 1250 oC. M3C particles were observed at tempering temperature
of 500 oC or below than 500 oC. Noreña and Bruzzoni [57] have studied the effect of different
heat treatment condition on microstructure evolution and their effect on hydrogen diffusion and
trapping in P91 steel. Higher diffusion coefficient was measured for air-cooled condition having
untempered lath structure. Microstructure in furnace cooled condition having ferrite and

13
 ACCEPTED MANUSCRIPT

precipitates showed the lower value of diffusion coefficient compared to air-cooled condition.
The minimum value of diffusion coefficient was measured for sample tempered at 500 oC. The
ductile-brittle transaction temperature (DBTT) of P91 steel was studied for varying reheat
temperature [58]. Strength of P91 steel was observed to be increased with increase in reheat
temperature due to dissolution of existing precipitates in temperature range of 1025-1100 oC. The

PT
DBTT was observed to be strongly dependent on grain size and higher DBTT was reported for
coarse grain structure. Optimum strength and Charpy toughness combination was observed at

RI
normalizing temperature of 1025 oC. Selvi et al. [59] had performed a comparative study on
effect of single and double normalizing treatment on tensile and creep behaviour of P91 steel.

SC
Double normalizing treatment led to formation of fine PAGs and M23C6 precipitates. The creep
rupture life was observed to be decreased for double normalizing treatment. Improved ductility

U
with minor strength loss was observed for double normalizing treatment. Fine PAGs formation
led to higher tensile ductility. In double normalizing treatment, water quenching was performed
AN
to increase the number of dislocation density in microstructure. Karthikeyan et al. [60] have also
studied the effect of double normalizing treatment on microstructure evolution, DBTT, creep
M

strength and tensile properties of P91 steel. The double austenitizing based normalizing and
tempering treatment (DNT) was observed to be superior in terms of grain size, tensile properties,
D

creep strength and DBTT characteristic compared to conventional normalizing and tempering
treatment (CNT). Long term ageing was also reported at 550 oC for 5000 h. The M2X formation
TE

was observed more frequently during long term exposure for CNT. Laves phase formation was
also reported for long term treatment. Karthikeyan et al. [61] have also reported the similar
EP

results for DNT treatment. The DNT treatment was resulted in formation of fine PAGs with
homogeneous tempered microstructure while CNT produced the heterogeneous microstructure of
C

‘ferrite + martensite’ with low hardness. Pandey et al. [32–34,51] have studied the effect of
varying normalizing temperature, tempering temperature and various cooling medium on
AC

microstructure evolution and mechanical properties of P91 steel. M23C6, M7C3, Cr2N, M3C, M2X
were the main precipitates observed in normalized and tempered condition of P91 steel.
Maximum normalizing temperature was recommended about 1100 oC while tempering beyond
760 oC was not preferable. Maximum tempering time was recommended about 6 h.
Effect of long-term ageing on precipitate evolution behavior and room temperature tensile
properties of P91 steel were also performed by many of researchers. Paul et al. [31] had studied

14
 ACCEPTED MANUSCRIPT

the effect of long term ageing on microstructure evolution of P91 steel in service temperature
range of 500-600 oC. The lath structure was retained even after ageing up to 10000 h. A
continuous coarsening was observed for M23C6 type carbide with increase in exposure time and
temperature while negligible coarsening was reported for V and Nb-rich carbide. A continuous
increase in Cr/Fe ratio was also observed with increase in ageing time. Pandey et al. [62] have

PT
also reported a continuous increase in Cr/Fe ratio with increase in ageing time. Laves phase
formation was reported after the ageing up to 5000 h that led to increase in DBTT. Homolova´ et

RI
al. [63] studied the thermal deformation history on microstructure evolution and stability of P91
steel in temperature range of 580-650 oC. For low temperature range (580-620 oC), laves phase

SC
formed after the long duration while for sample without deformation produced laves phase after
1000 h annealing at 650 oC. For all temperature, Cr/Fe ratio in M23C6 increased with ageing time.

U
Kafexihu et al. [64] performed ageing of P91 and X20CrMoV121 steel at 650 and 760 °C for
varying ageing time (2 h, 4320 h, 8760 h and 1752 h). Study was performed related to room
AN
temperature tensile properties and hardness after long term ageing. It was observed that
tempering at 650 °C have relatively small effect on mechanical properties as compared to
M

tempering at 760 °C. Senior et al. [65] studied the effect of ageing on plastic behaviour of 9Cr-
1Mo steel. A continuous reduction in ductility was observed with increase in ageing time. The
D

ductility failure was mainly led by inclusion and precipitate nucleated voids. The reduction in
ductility after 1000 h was attributed to nucleation of voids on carbide particles. After 5000 h,
TE

laves phase formation was observed that provided void nucleation sites and led to reduction in
ductility. After laves phase formation, void nucleation rate at carbides was decreased.
EP

Sathyanarayanan et al. [66] studied the effect of ageing temperature and time on DBTT of P91
steel by using the reference temperature approach under dynamic loading condition (Tody). At
650 oC, a drastic increase in Tody was observed while at 600 oC, a moderate increase was
C

observed. Ageing in temperature range of 600-650 oC for 10000 h was resulted in formation of
AC

laves phase that led to embrittling effect and ultimately increase in Tody. Golañski and Kepa [67]
performed the ageing of cast P91 steel and studied their effect on microstructure evolution and
mechanical properties. The ageing was performed in temperature range of 550-600 oC for
varying ageing time (2000 h and 6000 h). Ageing up to 6000 h have found a negligible effect on
the tensile properties, hardness and lath morphology of P91 steel. Ageing resulted in a drastic

15
 ACCEPTED MANUSCRIPT

decrease in Charpy toughness value. The decrease in Charpy toughness value was attributed to
formation of laves phase after long-term ageing of 6000 h.
5 Weldability of P91 steel
In present years, the weldability of P91 steel is a serious issue. It has been reported that,
compared to that of weld, a higher creep rupture life of P91 base metal is observed regardless of

PT
the arc welding process and heat treatment used [68]. It occurs due to presence of non-
equilibrium structure in heat affected zone (HAZ) [69,70]. The non-equilibrium structure formed

RI
due to high heating and cooling rate in the weld thermal cycle. P91 steel can be welded by any of
the commonly used arc welding processes such as manual metal arc welding (MMA), gas

SC
tungsten arc welding (GTAW), shielded metal arc welding (SMAW), electron beam welding
(EBW), and submerged arc welding (SAW) [29,71,72]. A lot of efforts have been made to

U
develop P91 electrode that provide optimum strength and toughness to weldments. The initial
problem faced during welding of P91 steel is selection of matching filler wire/rod that provide
AN
optimum composition in weld fusion zone. For each welding process, different composition of
filler wire is used. The typical composition of filler metals are given in Table 3.
M

Composition of Ni and Mn in filler wire required more attention due to their strong
influence on temperature Ac1, Mf, and Ms. In P91 base metal, maximum 1%(Ni+Mn) is allowed.
D

For 1.4% (Ni+Mn), Mf temperature decreased lower than 200 oC (preheat temperature) that led
to formation of retained austenite in filler metal. Santella et al. [27] reported a numerical
TE

equation to calculate the Ac1 temperature based on Mn and Ni contents. The equation is given
below;
EP

Ac1 (o C) = 854.5 ± 0.6 − 43.9 ±1× (Mn+ Ni) − 9 ± 0.4(Mn+ Ni)2 (6)
Increase in Ni and Mn content in filler metal lower down the Ac1 for weld metal. The V and Nb
C

are the other elements in filler wire that affects the mechanical properties of weld fusion zone.
AC

Arivazhaga et al. [73] studied the effect of V+Nb content in weld fusion zone on Charpy
toughness of weld fusion zone. Higher Charpy toughness was observed for the low V+Nb
content in weld zone. Fusion welding process introduces the heterogeneity in microstructure by
its nature. Single phase tempered martensitic structure in P91 weld zone provides optimum
strength and toughness compared to two phase of tempered martensite and δ-ferrite. In P91
weldments, formation of δ-ferrite in weld fusion zone leads to reduction in mechanical properties
of weldments. The welding of P91 steel required more attention than austenitic steel because it

16
 ACCEPTED MANUSCRIPT

required preheating before welding to avoid cracking, Post weld heating before post weld heat
treatment (PWHT) to remove the diffusible hydrogen, and finally PWHT to temper the brittle
martensitic structure. The welding thermal cycle results the formation of narrow non-equilibrium
heat-affected zone. The temperature experienced by the HAZ lies in the range of just above Ac1
to melting temperature (Tm). The HAZ formed in P91 weldment is difficult to distinguish

PT
because of continuous change in microstructure as move away from fusion line. On the basis of
temperature experienced and prior austenite grain size, the subzones form in P91 weldments is

RI
classified into weld zone (WZ), coarse grained HAZ (CGHAZ), fine grained HAZ (FGHAZ) and
inter-critical HAZ (ICHAZ), and over-tempered base metal (OTBM) [29]. A schematic evolution

SC
of sub-zone in multi-pass welded P91 joint is shown in Fig. 6.
The unique characteristic of each zone present in P91 weldments are given below;

U
5.1 Weld metal (Tp˃Tm)
The microstructure of weld fusion zone shows a complex structure due to nature of multi-pass.
AN
The complex structure of weld zone is attributed to the effect of each pass on previous pass. The
temperature gradient between weld and base metal leads to rapid solidification that resulted in
M

formation of columnar austenite grains. The subsequent passes is resulted the normalization of
previous deposit metal. For multi-pass welding, as move away from the first pass, the subsequent
D

passes might be the cause of auto-tempering of previously deposit metal and resulted in
formation of fine former austenite grain structures in previously deposited metal. The schematic
TE

diagram showing the microstructure evolution in different state of weld fusion zone is depicted
in Fig. 7 [74]. Sireesha et al. [75] have already reported the significance of filler metal
EP

composition on mechanical properties of P91 weldments. In P91 steel, weld fusion zone is
characterized with martensitic lath packets or groups with spatial orientation in one direction.
C

The untempered lath martensitic microstructure makes the weld fusion zone to more susceptible
towards cracking. The untempered lath martensitic microstructure is characterized with high
AC

strength and poor toughness. Untempered lath martensitic microstructure formation in weld zone
cause a mismatch in mechanical properties of weld and parent metal that resulted in poor creep
rupture life of cross-weld specimen. The retention of δ-ferrite during the welding also a serious
issue that lead to poor mechanical properties of weld fusion zone. δ-ferrite (BCC) phase
nucleates from liquid during the solidification of weld pool and transform to austenite (FCC) on
cooling which further convert to martensite (BCT) on cooling below Ms temperature. In P91

17
 ACCEPTED MANUSCRIPT

steel, high content of ferries stabilizers (mainly Cr) enhance the formation of δ-ferrite. Fast
cooling rate during the welding process does not allow the complete transformation of austenite
to martensite. Zhou et al. [76] reported that high cooling rate about 200 oC/s led to formation of
granular or block morphology δ-ferrite and simultaneously the amount of martensite in
microstructure decreased. The formation of δ-ferrite was observed to be closely related to the

PT
composition of weld metal and its formation could be controlled about 3% by balancing the
ferrite and austenite former in filler and base metal [29]. To calculate the formation of ferrite

RI
retention in weld fusion zone, the following empirical relations have been suggested [75] ;
Creq= % Cr + 6% Si +4% Mo +1.5% W+11% V+5% Nb+12% Al +8% Ti-40% C-2% Mn-4%

SC
Ni-2% Co-30% N- % Cu (7)

Kaltenhauser ferrite factor (FF)= %Cr + 6% Si +4% Mo +8% Ti+2% Al + 4% Nb -2% Mn- 4%

U
Ni-40% C-40% N (8)
AN
For fully martensitic structure free from δ-ferrite, Creq value lower than 10 and FF value
lower than 8 are preferred. Although, in some studies having Creq and FF lower than proposed
M

limit, showed the formation of δ-ferrite [77,78]. The other way to control the formation of δ-
ferrite is to control the heat input that directly affects the cooling rate. The effect of heat input on
D

δ-ferrite formation was studied by Arivazhagan et al. [79]. It was observed that high heat input
during the multi-pass welding process led the slow cooling rate that resulted in formation of δ-
TE

ferrite in welds. Weld without preheat and with moderate cooling rate was resulted in complete
absence of δ-ferrite. In autogenous TIG (A-TIG) welding process, fast cooling rate due to single
EP

pass led the formation of retained δ-ferrite phase. Presence of high amount of ferrite stabilizer
also lead the formation of δ-ferrite [80,81]. Arivazhagan et al. [80] had reported that in multi-
pass welding, slow cooling rate and narrow solidification range resulted in less formation of
C

retained δ-ferrite. Multi-pass welding also provided sufficient time for diffusion to compensate
AC

segregation that led the homogeneous transformation of ferrite to austenite.

In previous study, δ-ferrite patches were clearly observed in weld fusion zone of autogenous
tungsten arc welded joint of P91 and P92 steel, as shown in Fig. 8(a) [81]. The morphology of δ-
ferrite is shown at higher magnification of Fig. 8(b). It was observed that δ-ferrite did not
respond to PWHT and it remained in the microstructure, as shown in Fig. 8(c). At higher
magnification, precipitates are clearly seen at the δ-ferrite grains. A negligible amount of

18
 ACCEPTED MANUSCRIPT

precipitates are observed in δ-ferrite region. Wang et al. [82] confirmed that precipitate located
along the δ-ferrite grains were mainly ferrite stabilizer like Mo, V, Nb and Cr. Distribution of Cr
was reported to be homogeneous compared to other ferrite stabilizer elements. In multi-pass
GTAW weldments, δ-ferrite presence was not observed in as-welded as well as PWHT
condition, as shown in Fig. 8(e-f).

PT
The δ-ferrite formation in weld fusion zone have a noticeable effect on mechanical
properties. δ-ferrite formation in P91 steel weldments are generally reported at highest peak

RI
temperature [83,84]. The region exposed at temperature more than 1200 oC leads the formation
fine PAGs because of δ-ferrite nucleation on preferential sites of PAGBs. The δ-ferrite

SC
nucleation limits the growth of austenite and resulting in fine PAGs formation. Chandravathi et
al. [84] reported that formation of soft δ-ferrite in microstructure led to reduction in hardness.

U
The hardness of martensite and δ-ferrite patches (present in weld fusion zone) are shown in Fig.
AN
9. Schafer [85] studied the effect of δ-ferrite formation on tensile properties and observed that δ-
ferrite upto 8% have no any significant effect on the strength. However, a different opinion was
reported related to effect of δ-ferrite on toughness and ductility. Some author(s) [83,84] have
M

reported positive effect of retained δ-ferrite on toughness while others [86] emphasize the
negative influence. Schafer [85] reported that combined effect of δ-ferrite and brittle ‘dendritic’
D

carbides of M23C6 types (M≈65%Cr+30%Fe+rest) reduced the toughness by encapsulate the δ-

TE

ferrite. However, pure and soft δ-ferrite increased the toughness and ductility. Anderko et al. [87]
have also reported the negative influence of δ-ferrite on impact toughness of fusion zone.
Presence of δ-ferrite in base metal, weld zone or HAZ have found a negative influence on creep
EP

strength of P91 steel [15,88]. Presence of δ-ferrite accelerated the recovery rate of martensite and
growth rate of laves phase and ultimately reduced the creep and fatigue strength of steel. Effect
C

of welding process on metallurgy of HAZ in P91 weldments is shown in Fig. 10.

AC

The different microstructure of HAZ is obtained for different welding process. The
microstructure evolution in HAZ is mainly influenced by the heating and cooling rate during
weld thermal cycle, peak temperature experienced (Tp), dwell time, number of welding passes
and finally subsequent post weld heating (PWH) and post weld heat treatment (PWHT). High
heating rate during the weld thermal cycle resulted the shifting of transformation temperature,
recrystallizing temperature and solution temperature on higher side. As shown in Fig. 11, at

19
 ACCEPTED MANUSCRIPT

higher peak temperature, austenite transformation to δ-ferrite is started that led to reduction in
grain size because of δ-ferrite nucleation on austenite grains. The C and other austenite stabilizer
has poor solubility in ferrite that resulted in escape of austenite stabilizer from the ferrite and
make the ferrite region enriched form austenite stabilizer. Hence, segregated region differ from
the original composition and cooling cycle resulted in incomplete transformation of austenite to

PT
martensite and lead to formation of retained austenite or δ-ferrite. In multilayer welding, a
homogeneous and complete transformation of δ-ferrite to austenite has been observed. This

RI
occurred due to large time for diffusion to compensate the segregational process.
5.2 Grain growth zone or CGHAZ (Tm˃Tp˃˃Ac3)

SC
The region adjacent to the fusion line called coarse grained heat affected zone (CGHAZ) and
experienced the temperature much above than the transformation temperature Ac3. On the basis

U
of temperature experienced, grain growth zone can be divided into two zones. One zone having
AN
Tp is more than the temperature required for austenite to δ-ferrite transformation (Tϒδ). The
region consisted of untempered lath martensite with some patches of δ-ferrite. The δ-ferrite
generally nucleates along the PAGBs in band of 0.3 to 0.5 mm wide adjacent to fusion boundary
M

[29]. Zone having Tp is less than the Tϒδ leads the complete formation of austenite to martensite.
High peak temperature in CGHAZ leads the formation of coarse austenite grains with complete
D

dissolution of precipitates (M23C6, M7C3) [89]. The dissolved precipitates make the boundaries
TE

free from pinning force and it grows quickly. The CGHAZ region is characterized with coarse
austenite grains, and martensitic microstructure with negligible precipitates of M3C and NbX that
remain undissolved upto 1200 oC [55]. The CGHAZ zone represents the highest hardness and
EP

low toughness compared to other HAZs because of complete dissolutions of precipitates and
resulting increase in presence of C and N in the solution (solid solution strengthening). In
C

CGHAZ, Wang et al. had reported the high length/width martensitic blocks inside the coarse
AC

PAGs [36].

5.3 FGHAZ or grain refined zone (Tϒδ ˃Tp˃Ac3)

The peak temperature experienced by the fine grained heat affected zone (FGHAZ) is just above
the Ac3, about 1100 oC which make difficult the complete dissolution of existing precipitates
especially Cr-rich M23C6 [36,90]. The undissolved precipitates limit the growth of austenite by
pinning the austenite grain boundaries and lead the formation of fine grains of austenite. In

20
 ACCEPTED MANUSCRIPT

FGHAZ, the martensite with low L/w ratio is formed inside the fine PAGs. The low temperature
may lead transformation of tempered martensite into austenite with partial dissolution of the
precipitates. It also leads the coarsening of undissolved existing precipitates. Martensitic
transformation from the fine PAGs results in local strain distribution in FGHAZ [82]. The fine
equiaxed austenite grains, coarse existing precipitates of Cr-rich M23C6 and fine MX and M3C

PT
precipitates are the main characteristic feature of FGHAZ. The FGHAZ is characterized with low
hardness and moderate toughness because of weak solid solution strengthening compared to

RI
CGHAZ [91]. The most common premature type IV cracking generally occurs in the FGHAZ
during the long-term service condition [92]. In long-term exposure, undissolved coarse

SC
precipitates of M23C6 type acts as cavity nucleation sites [93]. A schematic evolution of
microstructure in FGHAZ for different heat treatment state is shown in Fig. 7. Schematic of

U
FGHAZ during welding reveals that the fine M23C6 gets completely dissolved while coarse
M23C6 dissolved partially. The fine MX precipitates remain unchanged during the welding.
AN
5.4 IC-HAZ or partially transformed zone (Ac3 ˃TP ˃Ac1)
The narrow width of IC-HAZ about 200 µm [94] make it difficult to identify the precise location.
M

The temperatures experienced by the IC-HAZ lying in the range of Ac1 and Ac3 and lead the
partial transformation of α to ϒ. The partial tempering of tempered martensitic microstructure
D

results in the nucleation of austenite grains at the favored position like PAGBs or martensitic
TE

boundaries. In IC-HAZ, partial dissolution of precipitates and coarsening of precipitates both are
observed simultaneously [29,35]. After the cooling, IC-HAZ shows a complex structure of
untempered lath martensite, retained over-tempered martensite, newly-formed PAGs, coarse
EP

tempered PAGs with newly formed fine precipitates and existing coarse precipitates along the
new PAGs. In IC-HAZ, higher density of precipitates is observed along the newly formed PAGs.
C

Among all HAZs, IC-HAZ showed maximum fraction and largest size of Cr-rich M23C6. Less
AC

fraction of MX (VX) was observed in IC-HAZ and FGHAZ compared to CGHAZ because of
higher stability of MX [36]. In IC-HAZ, high angle grain boundaries (HAGBs) with higher
mobility and energy contributes the higher potential to undergo recovery process that lead the
poor creep strength of material [82]. The IC-HAZ is characterized with low hardness and great
toughness among all HAZs due to lower solubility of carbon in solid solution [36,68]. The soft
zone formation in P91 weldments make it more susceptible to cracking during the creep
exposure. The IC-HAZ hardness derive from two competing effect- (1) hardness increase due to

21
 ACCEPTED MANUSCRIPT

untempered lath formation, (2) hardness decrease due to over-tempering of tempered martensite
from the base metal. Wang et al. [94] performed the detailed study of IC-HAZ formed in P91
weldments. The newly formed PAGs was revealed clearly by Nital etching while former PAGs
were recognized by tracking the retained Cr-rich M23C6 precipitates along the boundaries. The
schematic diagram that reveals the formation of IC-HAZ from base metal is shown in Fig. 12.

PT
5.5 Over-tempered base metal (TP < Ac1)
The temperature experienced by the over-tempered base metal is very less and does not

RI
allow any type of phase transformation. The weld thermal cycle might be cause of local
heating/tempering of original base metal. The hardness measured in the OTB region is generally

SC
less than the as-received base metal. The local over heating causes coarsening of existing
precipitates and grain.

U
5.6 Summary
The temperature experienced by the HAZs and their corresponding transformation reaction with
AN
microstructure characteristic is summarized in Table 4.
In P91 weldments, soft FGHAZ/IC-HAZ shows higher susceptibility to Type IV cracking that
M

make a challenging issue for P91 steel applications [89,95]. The precipitation strengthening has
been derived from the particle size, shape, volume and their distribution in matrix. Hence, the
D

distribution of precipitates inside the matrix is also play an important role during creep. In P91
steel, welding cycle results in dissolution of M23C6 carbides in some region of HAZs while a
TE

minute effect is observed on V and Nb-rich MX precipitates [92,93]. The dissolution of M23C6
precipitates affects the creep strength of P91 weldments to a great extent. The formation of δ-
EP

ferrite patches in some local region of HAZ acts as preferential sites for creep cavities and results
in harmful reduction in creep strength [88,96]. Wang et al. [82] have reported the lower
C

concentration of Cr and Mo in δ-ferrite grains formed in IC-HAZ and CGHAZ compared to δ-

ferrite grains formed in base metal. The ferrite stabilizers were found to be distributed along the
AC

boundaries between δ-ferrite grains and PAGs in base metal and HAZ. The δ-ferrite grains
present in base metal was observed to be undissolved in HAZ.
6. Influence of welding process parameters and heat treatment on characteristic of P91
welds

As discussed in previous section, welding results in the formation of heterogeneous

microstructure across the P91 weldments with some δ-ferrite patches in localized area of weld

22
 ACCEPTED MANUSCRIPT

fusion zone and HAZ. The δ-ferrite patches and heterogeneous microstructure have found a
negative influence on performance of P91 weldments. Before put in service, P91 weldments
have recommended heat treatment to overcome the heterogeneity across the weldments and
improve the mechanical properties. There are lot of works have been published related to
different type heat treatment of P91 weldments to improve the mechanical properties of weld

PT
fusion zone and HAZ.

According to the ASME Boiler and Pressure Vessel Code SA-335 [97], PWHT for P91

RI
weldments recommended in temperature range of temperature of 730-770 oC for 1 h, but
American Welding Society (AWS) suggests that PWHT should be carried out at 730-760 oC for

SC
1 h [98]. In some cases, normalizing and tempering (N&T) treatment is also recommended in
place of subsequent PWHT. As per ASME Boiler and Pressure Vessel Code SA-335, the

U
normalizing is performed in temperature range of 1040-1080 oC for 2 h, followed by air cooling
AN
up to room temperature then tempering in temperature range of 730-800 oC for 3 h, followed by
air cooling. To achieve the appropriate weld toughness, a Larson-Miller parameter ‘P’ ≥ 21 is
desirable for the heat treatment [99]. The Larson-Miller parameter P is given by Equation (9);
M

−3
P = K (log t + C ) × 10 (9)
D

where ‘k’ is the soak temperature in Kelvin, ‘t’ is the holding time in hours, and ‘C’ is the
TE

material specific constant. For the value of C = 21, the PWHT temperature is calculated about
755 oC for 2-3 h. For industrial purpose, both AWS and ASME recommended the post weld
duration about 1 h. PWHT at760 oC for 2 h has been recommended by so many researchers to
EP

overcome the non-equilibrium structure gradient and quenching stresses [68,99]. Generally, post-
weld heating of P91weldments by 250-300 oC for 30 to 60 min and then air cooling up to 100 oC
C

was recommended before PWHT to remove the diffusible hydrogen. The importance of this
cooling up to 100 oC is to ensure the formation of fully martensitic microstructure just before
AC

PWHT, as residual austenite does not respond to the PWHT and leads to the formation of
harmful un-tempered martensite [100].

Toughness of P91 steel weld is an essential property of weldment evaluated during hydro testing.
To avoid brittle fracture during hydro testing of vessels, higher impact toughness of weldment is
desired. Pressure vessels are subjected to hydro testing for strength and leaks. There are many

23
 ACCEPTED MANUSCRIPT

specifications that stress on the requirement of toughness of P91 steel weld. For example,
according to IGCAR-India specification for the prototype fast breeder reactor, the impact energy
of 45 J (minimum average of 38 J) is required at room temperature for P91 welds. However,
toughness of P91 weld is influenced by many variables such as notch location, welding
processes, filler wire composition, post-weld temperature and tempering time [73,101,102].

PT
Silwal et al. [68] studied the effect of PWHT on microstructure evolution and toughness of sub-
zones of P91 weldments. The PWHT was performed for 2 h at different tempering temperature.

RI
In as-welded condition, the lowest toughness was reported in weld zone and CGHAZ while
highest in IC-AHZ. Similar observation of toughness was also reported by Moitra et al. [102].

SC
The optimum tempering temperature was reported at 760 oC for 2 h that provide maximum
cross-weld toughness value for HAZs. Tempering below 730 oC for 2 h has found no significant

U
effect on toughness of CGAHZ. Higher toughness and lower hardness was observed at tempering
below about 20 oC from Ac1 temperature for 2 h. PWHT above Ac1 was resulted in drastic
AN
decrease in toughness of HAZs due to formation of fresh martensite. A minimum PWHT
temperature of 730 oC for 2 h was recommended to achieve the adequate HAZ toughness.
M

Sireesha et al. [77] studied the effect of filler metal (P9, P91, Inconel filler) composition and
subsequent PWHT on microstructure evolution and mechanical properties of fusion zone of P91
D

weldments. P9 and P91 filler exhibited the poor toughness less than the minimum required value
of 41 J. PHWT at 760 oC has produced a drastic increase in toughness of weld fusion zone for
TE

both P9 and P91 filler. Tempering at 730 oC has found poor effect on toughness and ductility due
to partial tempering. Inconel filler produced the minimum strength and maximum toughness of
EP

fusion zone (110 J˃ min. required) in as-welded condition and remain unaltered during the
PWHT. For P91 filler, acceptable fusion zone properties were observed for tempering at 760 oC
C

for 6 h. Gutiérrez et al. [103] had developed the P91 pipe welds using the GTAW process and
studied the hardness variation across the weldments for varying PWHT. In as-welded condition,
AC

a great hardness gradient was observed across the P91 weldments. Hardness gradient was found
to be reduced continuously with increase in tempering time from 1 to 8 h at 760 oC. PWHT at
760 oC for 2 h was reported to be enough for optimum strength reestablishment after the
welding. Effect of varying PWHT time on mechanical properties of A-TIG P91 weldments was
reported by Maduraimuthu et al. [104]. A continuous increment was observed in toughness value
with increase in tempering time. Higher hardness and poor toughness was observed for 2 h of

24
 ACCEPTED MANUSCRIPT

tempering that occurred due to partial tempering and higher solid solution strengthening.
Tempering for 8 h produced the maximum toughness of 200 J with poor hardness due to over-
tempering and extensive coarsening of M23C6. Optimum toughness and hardness combination
was reported for 6 h of tempering at 760 oC. Arivazhagan et al. [105] had also observed a
continuous increase in toughness with increase in tempering time in range of 2 to 10 h. PWHT

PT
temperature about 30-50 oC below the Ac1 was recommended that led to adequate diffusion
without austenite phase formation [106]. Vimalan et al. [107] have studied the effect of N&T

RI
treatment by surpassing the subsequent PWHT by using the Ac1 of material. For weld metal, Ac1
was observed to be less form the base metal due to higher Ni+Mn content in weld metal. Hence,

SC
for P91 weldments, safe PWHT temperature was reported to be 30 oC less than the Ac1 of weld
metal. Ac1 of weld metal is also influenced by the heating rate and reported to be 792 oC for 28
o
C/hr and 812 oC for 220 oC/h. Chen and Yamashita [108] have also reported a continuous

U
increment in Ac1 value with increase in heating rate. It was observed that PWHT above Ac1 of
AN
base or weld metal, degraded the mechanical properties of weldments due to formation of
untempered martensite. For weld metal with high Mn+Ni content, lower Ac1 was observed and
M

for such weld metal PWHT above but close to Ac1 led the reverse transformation of ϒ-phase.
The ϒ-phase was observed to be less to show any type of reverse tendency in mechanical
D

properties. In such welds, the formation of fresh martensite was accelerated at PWHT
temperature about 30 oC higher than Ac1. For weld metal with low Mn+Ni content, no any
TE

reverse transformation tendency was observed during the PWHT above the Ac1 due to formation
of soft ferrite. The weld with high Mn+Ni content showed the different creep tendency for low
EP

and high stress. Under low stress, creep rupture time was reported to be decreased with increase
in PWHT temperature but showed a reverse tendency at PWHT temperature above than Ac1. For
C

high level of stress, creep rupture time decreased continuously. For weld having low Mn+Ni
content, the creep rupture time was observed to be decreased with PWHT temperature (from
AC

lower to more than Ac1) for both low and high level of stress. For weld with high Mn+Ni content,
Chen and Yamashita [108] had recommended that upper limit of PWHT temperature should be
more than Ac1 but for weld with low Mn+Ni, the upper limit of PWHT should be lower than Ac1.
Taniguchi and Yamashita [109] also studied the role of Mn+Ni content and Ac1 point in deciding
the PWHT temperature for welds. For weld having high Mn+Ni content or lower Ac1, PWHT
temperature (760 oC), higher than Ac1 (730 oC) was recommended. For weld with low Mn+Ni

25
 ACCEPTED MANUSCRIPT

content or higher Ac1, PWHT temperature (780 oC) lower than Ac1 (785 oC) was recommended.
The detailed analysis of PWHT on creep rupture time was also performed by many of
researchers. Albert et al. [89] studied the effect of PWHT and their duration on creep behaviour
of P122 steel welds. Variation in PWHT duration was not found a noticeable effect on creep
rupture time [89,110]. The PWHT and creep test were resulted in significant reduction in

PT
hardness of CGHAZ and FGHAZ but no any significant variation was noticed in IC-HAZ.
PWHT of weld led the microstructural change and damage that occurred in HAZ while no any

RI
significant change was observed on creep rupture life and Type IV fracture mechanism. It was
reported that heterogeneity present in microstructure resulted in multi-axial stress state that led

SC
the final fracture of welds during creep. The effect of multi-axial stress on creep rupture life was
observed to be more dominant compared to microstructural change. Li et al. [110] had performed

U
study related to the creep behaviour of flux-core arc welded P91 weld joint for varying PWHT
temperature (600-840 oC) and time (2-8 h). Longer creep rupture life was observed for
AN
specimen’s heat treated at low temperature about 600 oC (partial tempering) for 2 h duration but
the required toughness of HAZs was achieved only after tempering at 720 oC for 2 h. It was
M

observed that tempering above the Ac1 led the higher strain rate and ductility. It was reported that
tempering above Ac1 with low tempering time of 2 h, led the fracture in IC-HAZ zone (Type IV)
D

but increase in tempering time from 2 to 8 h resulted in shifting of fracture location from IC-
HAZ to weld zone (Type I). Li et al. [110] concluded that upper limit of PWHT should not
TE

exceed the Ac1 of filler, weld or base metal while lower limit should be sufficient to restore the
toughness in weld and HAZs to limit the cold cracking.
EP

From the literature related to effect of PWHT, it is clear that PWHT reduces the strength and
heterogeneity across the weldments but not completely which affects the creep behaviour of
C

weldments during the long-term creep exposure. Renormalization of weldments after the welding
AC

might be useful in effective reduction of strength gradient compared to subsequent PWHT. It

also leads the homogeneous microstructure formation across the weldments and alleviates the
Type IV cracking of joint. Renormalizing treatment of bigger size pipe components is
impractical and uneconomical. Local renormalizing of welding is another way of microstructure
homogenization but lead the formation of other soft zone. For small size components, the
renormalizing is treated as the effective way of heat-treatment of weldments. In recent years, lot
of works has been published related to N&T treatment of weldments instead of subsequent

26
 ACCEPTED MANUSCRIPT

PWHT. Initially, Abd El-Salam et al. [111] had incorporated the effect of normalizing/tempering
(N&T) treatment on microstructure homogenization of shielded metal arc welded P91 joint and
compared it with subsequent PWHT. For N&T treatment, optimum degree of hominization was
observed along the weldment that was attributed due to the recrystallization effect during the
normalizing. Retained δ-ferrites were also observed to be less in weld fusion zone and CGHAZ

PT
for N&T treatment compared to subsequent PWHT. For N&T treatment, microstructure gradient
was reduced to a great extent [74]. A schematic of normalizing & tempering treatment process

RI
and their effect on microstructure evolution in sub-zone of P91 weldments are shown in Fig. 13
[74]. Abd El-Salam et al. [112] further studied the effect of N&T treatment on mechanical

SC
properties of P91 weld joint and successfully compared it with subsequent PWHT. In as-welded
condition, a great hardness gradient was observed along the weldments. Subsequent PWHT was

U
resulted in considerable lowering in hardness value of sub-zones of P91 weldments due to
tempering reaction but it was not sufficient to remove the hardness gradient across the
AN
weldments, completely. The N&T treatment was observed to be more effective way to reduce the
hardness gradient across the weldments by considerable lowering in microstructure gradient of
M

sub-zones. Compared to subsequent PWHT, N&T treatment produced the high-degree of

microstructure homogeneity across the sub-zone of P91 weldments. The impact toughness was
D

also reported to be higher in case of N&T treatment compared to PWHT. The N&T treatment
resulted in drastic increment in creep rupture life~90% more than the subsequent PWHT.
TE

Manugula et al. [72] had performed the N&T and subsequent PWHT for electron beam welded
reduced activated ferritic martensitic steel. The evolution of microstructure and their role on
EP

mechanical properties was studied for both the heat treatment process. The weld was produced
with poor toughness and high hardness due to presence of δ-ferrite in martensitic microstructure.
C

Subsequent PWHT resulted in significant reduction in hardness of weld fusion zone but the
heterogeneous distribution of hardness across the weldments was still present. Toughness
AC

measured for weld fusion zone was still much lower than the base metal. The N&T treatment
produced the homogeneous microstructure across the weldments with absence of δ-ferrite in
fusion zone that led the uniform hardness variation across the weldment and higher impact
toughness than the base metal. The hardness measured in N&T condition was almost similar to
base metal hardness. The improved tensile properties were also observed in N&T condition that
was comparable to that of base metal. Hence, it can be stated that for small components, N&T

27
 ACCEPTED MANUSCRIPT

treatment is most effective compared to subsequent PWHT for improving the toughness,
hardness, creep strength and tensile properties of weldments. Pandey et al. [71,113,114] had also
studied the effect of N&T treatment on microstructure evolution and mechanical behaviour of
P91 GTAW and SMAW weldments and compared with the subsequent PWHT. For N&T
condition, a homogeneous microstructure distribution (in terms of grain size, precipitate size and

PT
volume fraction of precipitates) with negligible hardness gradient was observed along the P91
weldments. The improved tensile properties and Charpy toughness were also observed for N&T

RI
treatment. The microstructure of sub-zone of P91 weldments in PWHT and N&T condition is
shown in Fig. 14 [69]. In PWHT condition, different microstructures were observed for different

SC
zones, as shown in Fig. 14. In N&T condition homogeneous microstructure with coarse
precipitates were observed in each zone. The coarse precipitates present in N&T condition might

U
be cause of creep strength reduction of P91 steel weldments. Effect of long-term ageing on
microstructure evolution and mechanical properties of P91 weldments have also reported by
AN
Pandey et al.[70].

PWHT have found a significant effect on residual stress distribution. Few studies have been
M

performed to investigate the effect of PWHT on residual stress distribution in P91 weldments.
Paddea et al. [99] studied the effect of PWHT distribution on residual stress distribution in P91
D

welded pipe. Highest tensile residual stress was reported in IC-HAZ which is also so called
TE

region for Type IV cracking during the creep exposure. PWHT resulted in drastic reduction of
tensile residual stress value and it was measured about 24% of yield strength of material near the
vicinity of HAZ. The region of tensile hydrostatic stress was also observed near the vicinity of
EP

HAZ. The most important finding was that the region of highest tensile stress and hydrostatic
stress coincided with the ICHAZ. In weld fusion zone, a compressive residual stress was
C

observed near the final welding pass. Venkata et al. [115] had modeled the effect of PWHT time
AC

and temperature on residual stress distribution in P91 welds. Tempering time up to 2 h produced
a significant relaxation in residual stress beyond that no effect was observed. PWHT temperature
played an important role in stress relaxation. Relaxation of residual stress was observed to be
increased with increase in tempering temperature. Maximum tempering temperature was
recommended below the Ac1. To study the effect of PWHT on residual stress in P911 weldments,
a finite element model was developed by Yaghi et al. [116]. In as-welded condition, residual
stress was measured by Deep Hole Drilling (DHD) technique and successfully compared with

28
 ACCEPTED MANUSCRIPT

numerically obtained values. The numerical model was successfully implemented to study the
effect of PWHT on residual stress distribution. PWHT of weldments was resulted in drastic
decrease in peak of residual stress.

The other parameters that affect the mechanical properties of P91 weldments are

PT
shielding gas composition, preheating and inter-pass temperature during welding and filler
composition. Arivazhagan et al. [105] studied the effect of shielding gas composition on
toughness of P91 steel welds. Pure argon shielding introduced the inclusion content in weld.

RI
Hence, to overcome the problem, mixture of argon (Ar) and carbon dioxide (CO2) is preferred. A
continues decrement in toughness was observed with increase the CO2 content in Ar. For pure

SC
CO2, poor toughness was observed. Optimum toughness and sound weld was produced for the
95% Ar+5% CO2 mixture. Arivazhagan and Vasudevan [80] performed a comparative study of

U
activated flux tungsten inert gas (A-TIG) and GTAW and studied their effect on microstructure
AN
evolution and mechanical properties of weld fusion zone. In A-TIG weld, fast cooling rate led
the formation of retained δ-ferrite while GTAW process provided more time for diffusion and
led the complete homogeneous transformation of δ-ferrite to austenite. In A-TIG process, high
M

peak temperature and cooling rate resulted in higher C percentage in solution that led to higher
hardness in A-TIG weld. In, GTAW poor hardness was observed due to less C percentage in
D

solution. In A-TIG weld, higher content of V+Nb and δ-ferrite might be the cause of poor
TE

toughness of fusion zone. For GTAW, PWHT at 760 oC for 2 h was reported to be sufficient to
obtain the minimum desired toughness value while in A-TIG weld, the duration of PWHT
extended. In A-TIG, the minimum required toughness value (47 J) was obtained after tempering
EP

at 760 oC for 3 h. The effect of preheat temperature on mechanical behaviour of P91 weldment
had been studied by Li et al [29] . Preheating before welding is preferred to avoid the cold
C

cracking in weld joint. Proper selection of preheat temperature before the welding lead the
AC

reduction in δ-ferrite by reducing the peak temperature and cooling rate [117]. Preheat
temperature control the martensitic formation that makes the weld susceptible towards cold
hydrogen induced cracking. Preheat temperature mainly depend on the Ms temperature which
have a great influence of material composition. Difference in Ms temperature in order of 50-60
o
C have been observed for upper and lower contents limit. Too high preheat temperature results
in formation crack during cooling while too low temperature leads the crack during welding
cycle due to excessive martensitic formation. Preheat temperature more than Ms leads the

29
 ACCEPTED MANUSCRIPT

formation of excessive martensite during cooling that results in cracking due to high residual
stress. Interpass temperature also played the similar role during the welding. Preheat and
interpass temperature is recommended in temperature range of 200-300 oC [77]. The preheat and
inter-pass temperature during the welding of P91 steel is calculated by using the following
equations;

PT
Tp = (M s − 90) ± 10o C (10)

RI
Tp = (M s −190) ± 10o C (11)

SC
The proper selection of preheat temperature have found a pronounce effect on tensile residual
stress and distortion and it reduced the values of tensile residual stress and distortion [117].

U
Preheat temperature have played an important role in deciding the joint susceptibility towards
hydrogen induced cold cracking which are discussed in latter section.
AN
Preheat temperature, inter-pass temperature, post weld heating, post weld heat treatment,
normalizing /tempering treatment, filler composition, and shielding gas composition are the main
M

deciding factors that affects the microstructure evolution and performance of P91 joints.

7 Diffusible hydrogen in P91 welds

D

The residual stresses, weld microstructure and diffusible hydrogen present in deposited metal are
TE

the main parameters which affect the weldability of P91 steel. The residual stress and
microstructure of P91 steel weldment can be optimized by using the proper PWHT as suggested
EP

by the researchers [29,99] but the hydrogen induced cracking (HIC) in P91 steel weld is a serious
problem. The researchers have recommended the use of low hydrogen electrode, preheating and
C

post weld heating to avoid the HIC in P91 weldments. Diffusible hydrogen combined with
residual stress and crack susceptible microstructure results the HIC in weldments that lead the
AC

catastrophic failure of weld component in service at ambient temperature. A combined effect of

sensitive microstructure, high residual stress and sufficient level of hydrogen on HIC is shown in
Fig. 15.

In arc welding process, the arc temperature was reported about 10000 oC that lead to arc plasma
by dissociation of atoms to ions [118]. The flux with several ingredients in various arc welding
process such as SMAW, FCAW and SAW used to prevent the molten metal from oxidation

30
 ACCEPTED MANUSCRIPT

[119]. In ingredient processes the chemically bonded water in structure that dissociate into
hydrogen and introduces in weld pool. The SMAW process shows higher hydrogen content
compared to SAW and FCAW. The solubility of hydrogen increases with temperature and results
in more hydrogen content in weld pool while solubility decreases during the cooling of weld
pool. Rapid cooling after the welding does not provide sufficient time to escape the dissolved

PT
hydrogen content. Hence, rapid cooling results in formation of supersaturated weld metal with
hydrogen. The trapped hydrogen in weld pool shows higher propensity to diffuse within or out of

RI
weld.

7.1 HIC mechanism in weld

SC
The sequence of action that lead to HIC in welds are given below [120];

U

Hydrogen atom, ions or molecules produced due to dissociation of moisture present in arc
medium. The chemically bonded water in electrode flux (SMAW, FCAW), hydrogen in
AN
shielding gas (GTAW, GMAW) and water in powder flux (SAW) are the major source of
hydrogen [121,122]. The water present in electrode and powder flux shows higher stability at
M

temperature upto 1000 oC and cannot be removed by baking at temperature about 300 oC.
Sometimes, moisture absorbed by the flux from surrounding, hydrocarbon contamination
D

(oil, grease), oxide layer on base metal, bulk hydrogen in base metal and surrounding of
welding atmosphere are also act as the minor source of hydrogen [123], which can be
TE

removed easily by baking of electrode, preheating and cleaning of plate.

Hydrogen atoms, ions get dissolved in weld pool during the heating cycle and molten pool
EP

become supersaturated with hydrogen after the cooling.

In P91 steel, austenite normally transforms completely to martensite during the air cooling.
C

The transformation temperature of austenite to ferrite (TF) is higher than the austenite to
martensite (TB). The hydrogen atom present in weld metal ferrite migrates to the HAZ
AC

austenite. The hydrogen diffusion coefficient for the austenite phase is lower than the ferrite
phase but the solubility of a hydrogen atom in austenite phase is more than the ferrite phase.
Hence, during the welding process, the austenite phase present in HAZ absorbs the hydrogen
which cannot migrate to fusion boundary during the cooling cycle. After further cooling
austenite phase present in the HAZ and weld zone transforms to martensite and hydrogen

31
 ACCEPTED MANUSCRIPT

atoms get trapped in the martensite in a state of high energy. The schematic of hydrogen
diffusion is shown in Fig. 16.

The high energy hydrogen diffusion creates the defects and discontinuities in weld and HAZ.

The residual tensile stress developed in weldments due to the volumetric phase
transformation combine with trapped hydrogen and leads the growth of cracks and

PT
discontinuities.

All hydrogen enter the weld zone and HAZ is not free to diffuse in weld zone or outside. The

RI
one part of absorbed hydrogen remains trapped in weldments and shows no contribution to
delay cracking. The so called absorbed hydrogen is called residual hydrogen and varies with

SC
trap density present in weldments. The other part of hydrogen that is free to diffuse into
weldments or outside is called diffusible hydrogen (HD) and lead the HIC due to migration

U
ability. In P91 steel, the weld fusion zone has higher hardenability compared to the base metal
AN
due to this the driving force of H to migrate to the base steel may not exist.
In weld metal with poor hardness, HAZ acts as preferential site for HIC because of high stress
level. In case of weld with high hardness like P91 and P92 steel, the HIC might be initiate in
M

fusion zone because of high critical stress. Hence, it can be stated that to reduce the
susceptibility of weld fusion zone and HAZ to HIC, the sensitive microstructure, HD and
D

residual stress should be minimized or absent. It is difficult to control of residual stress and
TE

hardenability in weldments but the control of diffusible hydrogen in weldments is possible.

The diffusible hydrogen content can be controlled up to some extent by the baking the
electrode. However, hygroscopic coating of electrode pickups the moisture from ambient at very
EP

fast rate. The cellulose coating also acts as the major source of hydrogen. To reduce the HD
content in weldments upto a large extent, it is necessary to control the cooling rate during the
C

welding cycle. Slow cooling cycle provide more time for diffusible hydrogen to escape out
AC

which is possible by increasing the heat input but high heat input shows a negative influence
on mechanical properties. The preheat of the base metal is another way to control the cooling
rate and reduce the risk of HIC [124].
7.2 Diffusible hydrogen (HD) measurement
The hydrogen atoms are trapped in the weld pool from welding consumable and environment.
The hydrogen trapped in the weld pool is measured based on various standards. The methods
recommended by the standards are JIS, DIN, AS/NZ, BS, AWS, GOST, and BIS IS [125–127].

32
 ACCEPTED MANUSCRIPT

The four commonly used methods for determination of diffusible hydrogen in weld metal are: (i)
glycerin method, (ii) mercury method in which measurement is carried out at room temperature,
(iii) gas chromatography method and (iv) vacuum hot extraction method. The IIW used the linear
scale increment for hydrogen level measurement by units of 5 (5-10-15 ml/100 g), the
logarithmic scale (4-8-16 ml/100 g) used by the AWS, based on the correlation of lower critical

PT
stress, lower preheat temperature with diffusible hydrogen levels to avoid hydrogen induced
cracking [97,128]. Except mercury method, all the mentioned methods for diffusible hydrogen

RI
measurement are out-of-date and not used in present days. Fydrych and Łabanowski [129] have
also reported that among all methods the mercury method is more accurate and provide reliable

SC
and repeatable results. Various researchers have used the mercury method successfully to
measure the diffusible hydrogen in deposited weld metal due to its reliability, repeatability and

U
lower cost [129–131]. AN
The BIS IS [125–127] standard procedure for glycerin and mercury method to measure
the diffusible hydrogen in the weld metal is given below.
M

7.2.1 Glycerin method

It is one of the oldest method in which utilized the volumetric displacement method. The
D

diffusible hydrogen is allowed to collect over the glycerin in a bath maintained at 45 oC. In
glycerin method, the test assembly heated at 650 oC for 1 h to remove any bulk hydrogen and
TE

after cooling clean with the emery paper. Four test specimens of cross-section 12×25 mm and
length of 125 mm are used. A weld bead pass of 100 mm is deposited on 25 mm surface for each
EP

test specimens. For the bead deposition, a 4 mm diameter baked electrode is preferred. After 30 s
of the bead deposition, slag is removed and sample placed in iced water at maximum temperature
C

about 5 oC for 30 s. After the quenching, sample is cleaned and placed in glycerin bath within 10
s. The glycerin bath is maintained at 45 oC. The hydrogen comes out from all the four samples
AC

collected in different apparatus for 48 hours. The average diffusible hydrogen measured is
corrected at 20 oC and 760 mm Hg. The expression to calculate the diffusible hydrogen (HG) per
100 g of weld metal deposited is given below:

293 B 100
HG = Vg (12)
273+ T 760 M 2 − M1

33
 ACCEPTED MANUSCRIPT

where HG is the diffusible hydrogen measured by the glycerin method (ml/100 g of weld metal
deposited) at NTP (0 oC and 760 mm Hg), Vg is the volume of gas in burette after 48 h, B the
barometric pressure in mm Hg, T the room temperature (oC) when Vg is measured, M1 the mass

PT
of sample in g before deposition of weld and M2 the mass of sample in g after removal from
apparatus.

RI
7.2.2 Mercury method

Sample preparation for the mercury method remains same as used in glycerin method.

SC
The schematic of weld assembly used for diffusible hydrogen measurement in mercury method
is shown in Fig. 17 [132]. Test assembly consisted of specimen piece of dimension 30×15×10

U
mm and run on and run off piece of dimension 45×15×10 mm. During the welding the test
specimen was placed between run on piece and run off piece and welding is carried out by
AN
positioning the test assembly on a copper jig fixture. Immediately after the welding, the test
assembly was quenched in iced water. The run on piece and run off piece are broken with the
M

help of vice and hammer. The center piece placed at -50 oC in cooling chamber for 3 to 5 min
after 10 s of welding. Subsequently, the test pieces were removed from the cooling chamber and
D

washed with alcohol, pure ether and then placed in the glass capsule for drying in the blast of dry
nitrogen. After that the test specimen is admitted through open limb in to the Y-tube of mercury
TE

apparatus and limb is closed. The specimen inside the Y-tube is allowed to evolve the hydrogen
for 72 h and it was collected over the mercury at room temperature. The released hydrogen was
EP

measured by volumetric method by using diffusible hydrogen measuring meter. With the help of
mass of deposited metal, ambient pressure and temperature, collected hydrogen in hydrogen
C

measuring meter are converted into volume of diffusible hydrogen per 100 g of the deposited
metal. The expression to calculate the diffusible hydrogen (HM) per 100 g of weld metal
AC

deposited is given below:

273 B − H 100
H M = Vg (13)
273 + T 760 M 2 − M1

34
 ACCEPTED MANUSCRIPT

where HM is the diffusible hydrogen measured by the mercury method (ml/100 g of weld metal
deposited) at NTP(0 oC and 760 mm Hg), Vg is the volume of gas in Y-tube burette after 72 h, B
the barometric pressure in mm Hg, T the room temperature when Vg is measured, H is the head
of mercury in mm at which Vg is measured, M1 the mass of sample in g before deposition of
weld and M2 the mass of sample in g after removal from apparatus.

PT
The gas chromatography (GC) and hot extraction (HE) have been also successfully implemented
by so many researchers to measure the HD in deposited metal [133–135]. Polymer electrolyte

RI
based hydrogen sensor, hot extraction proton-exchange membrane hydrogen sensor (HE-
PEMHS), vacuum-extraction method are some other alternate ways used for the diffusible

SC
o
hydrogen measurement [136–138]. At 150 C, diffusible hydrogen measured by gas
chromatography method (HGC) shows a good relationship with hydrogen measured by mercury

U
method (HM). The expression is given below[139];
AN
HGC =1.05HM (14)

A linear expression between mercury (HM) and glycerin (HG) method has also been developed by
M

the Kotecki for three different analysis which are given below [97];
D

HG = 0.67HM − 0.80 (15)

TE

HG = 0.64HM − 0.93 (16)

HG = 0.79HM −1.73
EP

(17)

Fydrych and Łabanowski [129] have also reported a linear numerical relation between diffusible
C

hydrogen content in deposited metal measured by the mercury and glycerin method. Relation is
valid for the hydrogen content up to 35 ml/ 100 g of deposited metal. It was also concluded that
AC

for low diffusible hydrogen content, mercury method is more appropriate because of its high
accuracy. The linear expression is given below;

HG = 0.81HM −1.71 (18)

The diffusible hydrogen content measured by mercury method is reported to be higher than
glycerin method and it is also clear from the above numerical expression.

35
 ACCEPTED MANUSCRIPT

HE, HE-PEMHS and Gas chromatography (HE-GC) are used to measure the diffusible hydrogen
at temperature about 400 oC and at same temperature shows a good relation with mercury
method, as given below[133,140];

H HE = 0 . 94 H M − 0 . 37 (19)

PT
H PEHMS = 0.998H M + 0.03 (20)

RI
H HE −GC = 1.005H M + 0.22 (21)

SC
The presence of diffusible hydrogen in deposited metal has been affected by many of
parameters like welding process, working environment, preheating, post weld heating and
welding process variables. Padhy et al. [141] studied the effect of preheating and post weld

U
heating on diffusible hydrogen content in deposited metal for P91 steel and mild steel. For P91
AN
steel, weld was prepared using the SMAW process with E9015-B9 electrode of 4 mm diameter.
The preheating and post heating was conducted in temperature range of 100-250 oC. The
diffusible hydrogen content was observed to be decreased with increase in preheat temperature
M

and maximum reduction for a given preheat condition was reported for the weld in which
measurement carried out after the cooling up to 100 oC. For such condition and preheat
D

temperature of 250 oC, minimum HD content in weld was only about 27% of actual HD. The
TE

combined effect of preheat and post heating showed a drastic effect on HD and it observed below
the measurable limit. The hydrogen content comes out during post heating was increased with
post heating temperature. In post heating temperature range of 200-250 oC, hydrogen diffused
EP

out was very close to HD content measured as per ISO 3690 (without any heat treatment). It was
recommended to perform both preheating and post heating at lower temperature instead of
C

performing preheating at higher temperature. In under water welding, welding process parameter
AC

have reported a significant effect on the HD content [142]. It was reported that diffusible
hydrogen content in deposited metal increased with the increase in welding current, salinity of
water and reverse polarity (-DC to +DC). The minimum diffusible hydrogen level obtained for
the negative polarity and low value of welding current. The increase in welding time, electrode
angle also resulted in the increase in HD content in deposited metal. The increase in welding rate,
time duration between weld and measurement and weld with used electrode resulted in lower HD
content [143]. The HD content measured in air environment was observed to be much lower than
36
 ACCEPTED MANUSCRIPT

the HD content in under water welding condition [144]. Moistening of electrode was resulted a
drastic increase in HD content [129]. The number of bead pass have observed a positive effect on
HD content and it decreased with increase in bead pass [145]. In multi-pass, the subsequent bead
pass results in heating of previous deposited bead that lead to diffusion of enhanced hydrogen.
Pandey et al. [146] had studied the effect of diffusible hydrogen content on tensile and flexural

PT
strength of P91 weld joint. A continuous decrease in tensile and flexural strength was observed
with increase in diffusible content in deposited metal.

RI
Hydrogen embrittlement and residual stress are the primary cause of delayed cracking in
precipitation hardened steels. The susceptibility to HAC is amplified in P91 steel weld because

SC
of high hardenability and lath martensitic microstructure. To avoid the HAC in weld metal, low
hydrogen electrode, preheating and post weld heating are recommended. In some critical cases

U
where it is difficult to use preheating above 150 oC, the only alternative is to use low hydrogen
AN
electrode.

7.3 Effect of diffusible hydrogen on weldability

M

To compare the relative susceptibility of material to HIC and to prevent the cold cracking,
various weldability test such as self-restraining test (Lehigh test, Controlled thermal severity
D

(CTS) test, and Tekken test), externally loaded test (Implant test, tensile restraint and rigid
restrain test), and constant load rupture simulated test are employed. As discussed in previous
TE

section, microstructure, HD content and residual stress are the main governing variables for the
HIC. The results obtained from the weldability test are proposed in terms of these variables. The
EP

self-restraining test results are govern in terms of welding condition, heat input and preheating of
the plate. In externally loaded tests, the stress is used to express the test results. No correlation is
C

reported among the results obtained from the various tests. To measure the susceptibility of HAZ
to HIC, mainly self-restraining or externally loaded test are used.
AC

Leigh test was used for the quantitative analysis of weldability and developed by Stout et
al. [147]. The weldability was measured in term of degree of constraint that required for
weldment cracking. Controlled thermal severity (CTS) test is used to determine the weldability
of steel with high yield strength (300-600 MPa). The testing was performed to control the
hardness of HAZ by modifying the cooling rate and phase transformation by chemical
composition. The fillet weld test of CTS provided less sever condition for the HAZ cracking. In

37
 ACCEPTED MANUSCRIPT

welded structure, high restraining stress results the underestimation of CTS test for weldability
test of steel. Tekken test, a self-restraint method, is more stringent compared to other test for
HIC. Fydrych et al. [148] had used this technique to improve the weldability of steels in
underwater wet welding condition. The technique successfully implemented on S355J2G3 steel
and S500M steel.

PT
In internally loaded or direct test, rating of material was performed on the basis of cracking
amount present in microstructure. The amount of hydrogen was controlled by moisture content

RI
of flux coating or by shielding gas. Internally loaded test involves the complete testing of
weldments including both weld zone and HAZ. However, it is difficult to study the separate

SC
effect of welding on microstructure evolution and restraint level. The externally loaded test or
indirect test involves the constant stress rupture test in which constant load is applied on pre-

U
notched specimen charged with hydrogen. Indirect test allows the independent control of each
variable like stress, microstructure and HD content. A series of constant load rupture test is
AN
conducted to evaluate the cracking susceptibility of material. The stress below which no failure
is observed within specified limit, denoted as lower critical stress (LCS). A simple embrittlement
M

index (EI) has been given to rate the material on the basis of crack susceptibility towards HIC
[149];
D

NTS − LCS
EI = (22)
NTS
TE

Where NTS= notch tensile strength

The susceptibility of material towards HIC increases with increase in embrittlement index (EI).
EP

In externally loaded test, most commonly used method for weldability tests is the
Granjon implant test. The implant test was developed in Institute de Soudure (French Welding
Institute) by Henri Granjon to determine the HAZ susceptibility towards HIC [150]. It is more
C

versatile due to its simplicity and small size specimen and it also able to classify the material in
AC

order of their susceptibility to HIC. Moreover, various factors such as residual stress, preheat
temperature, post weld heating temperature, restraint, hydrogen level, welding condition and
microstructure can be quantified separately by using the Implant test [149].
In implant test, a small cylindrical specimen is fixed inside the hole of base metal. A
circumferential notch is made at the end of cylindrical specimen. A single pass weld bead is
deposited on the top surface of specimen so that a part of notch threaded section is positioned in

38
 ACCEPTED MANUSCRIPT

the HAZ. Along with fuse the top end of cylindrical specimen, weld bead also introduces the
hydrogen and form HAZ. The assembly is shown in Fig. 18.
The assembly is subjected to variable constant load and time of failure for each test is
recorded. The major problem faced in Implant test was the proper allocation of notch in same
HAZ microstructure in each test. In modified Implant test, to overcome the notch problem,

PT
circumferential notch was replaced with helical notch and entire HAZ is traversed by the helical
notch [149]. The weld implant test is successfully implemented by many of researchers to

RI
classify the material’s susceptibility[151–154]. The Implant test is mainly implemented in
grading HSLA steel.

SC
The working of tensile restrain cracking (TRC) test is quite similar to implant test. In
TRC test, simple butt welded steel plate subjected to constant tensile load for long time as

U
similar to implant tests. Constant tensile load lead the crack initiation after certain time and crack
growth increases with time. The low level stress results the longer time for crack nucleation and
AN
growth. No cracking was observed below a certain stress called as critical stress. The low HD
content and higher preheat temperature was resulted in higher critical stress. In rigid restraint
M

cracking (RRC) test, an automatic control mechanism is used to keep the certain restraining
length constant during the weld cooling cycle. By varying the restraining length, restrain
D

intensity and critical stress is evaluated for the particular material, groove geometry and welding
condition. It deals with the transient stress that developed on the joint during the cooling cycle.
TE

Among all the methods, the most popular and commonly used method for weldability test is
Implant test. Manimozhi et al. [152] had reported the effect of diffusible hydrogen level and
EP

preheating on HAZ HIC of Cr-0.5Mo-1.7W steels by using the Granjon implant test. It was
observed that residual hydrogen plays a crucial role in delayed hydrogen cracking compared to
C

initial diffusible hydrogen present in the weld metal. A continuous increase in lower critical
stress (LCS) value was observed with increase in preheating temperature up to 250 oC. The
AC

maximum LCS was measured for low HD content and for high preheating temperature (250 oC).
Albert et al. [151] had studied the effect of preheating on HAZ HIC on 9Cr-1Mo steel by using
the implant tests. A continuous increase in LCS value was observed with increase in the preheat
temperature but still LCS value was less than the yield strength of material at preheat
temperature of 250 oC. Both weld metal and HAZ of 9Cr-1Mo steel was observed to be highly
susceptible to HIC. Magudeeswaran et al. [131] had performed the implant tests on high strength

39
 ACCEPTED MANUSCRIPT

quenched and tempered steel and studied the effect of diffusible hydrogen level on LCS of steels.
Welding consumable have observed a significant effect on the hydrogen induced cracking of
Q&T steel welds. Yue [153] and Yue and Lippold [155] also performed the weld implant tests to
classify the different high strength naval steel (BA-160, HY-100 and HSLA-100) in their
susceptibility order towards HIC. The CGHAZ present in the weldment showed highly

PT
susceptible to HIC compared to other zone. It was also concluded that with use of high heat input
and applying preheat, the cracking tendency of metal can be reduced.

RI
Albert et al. [151] reported that at high-stress level crack propagation occurred through the weld
fusion zone while at low-stress level through the HAZ. The phenomena of crack propagation

SC
explained on the basis of diffusible hydrogen level in the weld fusion zone and concentration of
hydrogen at the crack tip to initiate the cracking for given stress level. Just after welding, the

U
diffusible hydrogen concentration in weld fusion zone would be higher comparison to HAZ. At
high stress level, critical hydrogen concentration required for crack initiation would be less. The
AN
higher concentration of diffusible hydrogen in weld fusion zone makes it more susceptible to
HIC. However, at low-stress level, the concentration of diffusible hydrogen required to initiate
M

the cracking would be higher and it takes more time to reach the critical level at the notch tip.
Figure 19 shows the schematic representation of crack initiation and propagation at high and
D

low-stress level.
At low-stress level, a large amount of hydrogen atom diffuses into HAZ from weld fusion zone
TE

to make the level of hydrogen uniform both in weld fusion zone and HAZ. This resulted, the
diffusible hydrogen concentration increases in HAZ and simultaneously in weld fusion zone
EP

decreases. The HAZ HIC is a more critical issue and if not controlled timely, it can occur
unexpectedly and lead to great loss by catastrophic failure [156,157]. The number of theories has
C

been proposed by the researchers to describe the HIC mechanism such as surface absorption
theory by Petch [158], slip softening model by Beachem [159] and the decohesion theory by
AC

Troiano [160]. Even though a unified theory of HIC is still being pursued.

8 Shrinkage and residual stress in weldments

Joining of structural component, piping and pressure vessels used in nuclear, thermal
fertilizer and chemical power industries are generally carried out by the welding process that
results in residual stress, shrinkage stress and distortion at remarkable level. In welding process,

40
 ACCEPTED MANUSCRIPT

the localized heating and cooling leads the formation of complex residual stress and distortion
that results in catastrophic failure of welded joint. Localized heating and cooling of base,
solidification shrinkage of weld, internal constraint, external constraint like tacking, and phase
transformation result in formation of residual stresses and distortion [161–163]. The differential
contraction of weld and adjacent base metal causes the thermal straining that lead to distortion

PT
and shrinkage stress in welded structure. Thermal straining along the welding direction results
in formation of longitudinal shrinkage while strain perpendicular to direction of welding cause

RI
the transverse shrinkage [164]. The welding distortion occurs due to the internal forces
developed in structure form the resulting combined stresses. Distortion is treated as the pattern

SC
of permanent strains that remains structure after the completion of welding process.
In arc welding process, material is subjected to temperature more than melting

U
temperature of metal. The welding cycle is elevated to rapid heating and cooling. Compared to
the weld structure, size of weld pool is very small and expansion and contraction of the weld is
AN
prevented by the adjacent base metal. During the heating cycle, expansion of heated zone leads
the formation of compressive residual stress and cooling cycle results in shrinkage which is
M

prevented by the base metal. After the cooling, shrinkage resistance cause the formation of
tensile residual stress in weld zone which is balanced with compressive residual stress of base
D

metal [163]. Distortion and residual stress cannot be removed completely but it can be
minimized. Excessive constraint during the welding process leads the formation of high
TE

residual stress while low residual stress formed during the free expansion and contraction.
During welding process, parent metal will either resist the distortion or shrinkage resulting in
EP

residual stress, or it follow the shrinking movement of weld pool resulting in distortion and
shrinkage [165,166]. The residual stress and distortion are influenced by many of parameters
C

like welding process, welding parameters, number of process, groove geometry, material
properties and phase transformation. Weld induced tensile residual stress in welded components
AC

leads the HIC, buckling deformation, stress corrosion cracking (SCC), reduction in fatigue life
and brittle fracture [128,167,168].
The solid phase transformation and welding boundary condition play crucial role in final
stress distribution. Solid phase transformation occurs in two different manners. First one is
formation of secondary phase particles during the heating process and second one is the
transformation of matrix from one atomic packing to other during the heating and cooling

41
 ACCEPTED MANUSCRIPT

process. The transformation occurs during the heating and cooling cycles which influence the
formation of residual stress level. The character of the material is mainly responsible for the
phase transformation. In P91 steel, the martensitic transformation occurs during the welding
cycle whose effect is more complex and acts to influence the level of stress in transformed
zone. Cottam et al. [162] studied the effect of both type of phase transformation on stress

PT
formation. They had reported that the martensitic transformation is more complex and help to
reduce the magnitude of residual stress in transformed zone. The point at which martensitic

RI
transformation starts resulted in an increase in stress level in adjacent non-deformed zone. In
P91 steel weldment, cooling cycle leads the transformation of austenite to martensite. During

SC
the transformation form austenite, stress level was observed to be decreased up to Ms
temperature and after that increased upto final martensitic transformation. The austenite to

U
martensite transformation was resulted in formation of compressive stresses. The tensile
residual stress was reduced with increase in volume fraction of martensite by controlling the
AN
heat input and cooling rate. Austenitizing of steel results in structure transformation from BCC
(ferrite) to FCC (austenite) that leads to reduction in volume. During cooling process,
M

transformation of austenite to martensite results in BCT structure formation that leads to

increase in volume, as shown in Fig. 20 [169]. Yaghi et al. [169] studied the residual stress
D

distribution in P91 welded pipe by incorporating the solid phase transformation. It was
observed that the solid state phase transformation (SSPT) in P91 welding have a great influence
TE

on residual stress distributions and their magnitude. Zubairuddin et al. [117] also reported the
martensitic phase transformation effect on residual stress distribution and observed almost
EP

similar results. SSPT during welding was resulted in considerable lowering in magnitude of
axial and hoop stress on the outer surface of welded pipe (HAZ) and half of the weld fusion
C

zone. Den and Hidekazu [100] had reported that martensitic transformation not only reduced
the magnitude of residual stress but also altered the sign of residual stress. The FE simulation
AC

showed a good agreement with experimental results when martensitic transformation was taken
into account.
Further, Murkawa et al. [170] continued the same research and study the effect of Ms
temperature on magnitude of tensile residual stress developed in weld zone. A low Ms
temperature helped to develop compressive stress component that reduced the magnitude of net
tensile residual stress within the fusion zone. The magnitude of compressive stress was

42
 ACCEPTED MANUSCRIPT

observed to be increased with the decrease in the Ms-temperature and became saturate at
temperature about 200 oC. The tensile nature of longitudinal stress developed in the HAZ was
observed to be un-affected by the Ms-temperature. Preheating has also observed a great
influence on the magnitude of residual stress and their distribution. Preheating before the
welding caused a considerable lowering in peak temperature and cooling rate. Zubairuddin et

PT
al. [117] reported that preheating was resulted in reduction in peak tensile residual stress and
distortion.

RI
A lot of work have been performed related to study of effect of welding process, groove
geometry, welding parameters, and number of welding passes on shrinkage and shrinkage stress

SC
in pipe and plate weldments [164,171,172]. Ghosh et al. [164] performed an analytical study on
shrinkage stress mode, magnitude and distribution in different quadrants of GMAW and pulse

U
GMAW welded pipe for different weld groove designs. The mode and magnitude of shrinkage
stress in different quadrant was observed to be non-uniform and varied as a function of welding
AN
process, parameters and groove geometry. Pulse GMAW process was resulted in lower
magnitude and uniform distribution of transverse shrinkage stress compared to GMAW. Higher
M

the heat input during the welding process was resulted in higher magnitude of transverse
shrinkage stress. For constant heat input, narrow groove design produced the lower shrinkage
D

stress compared to conventional groove design. The multi-pass welding, the weld metal is
subjected to localized solidification shrinkage [171]. The repetitive influence of thermal cycle
TE

from subsequent weld passes affects the development of stress in weld groove upto certain
extent and finally it causes a continuous change in groove design and groove area with every
EP

weld passes [173]. The change in groove geometry with subsequent pass results in groove angle
variation and it will not be uniform at all location in each quadrant of pipe. The change in
C

groove geometry was observed to be more in the case of V-groove weld design than narrow-
groove [22] and this occurred due to less weld metal deposition in narrow groove. Ghosh et al.
AC

[174] studied the effect of Pulse GMAW and GMAW process on transverse shrinkage stress
and distortion of thick butt welded plate. It was observed that use of pulse GMAW produced
low shrinkage stress, bending stress, and distortion compared to GMAW process. For given
heat input, narrow groove weld design produced the lower stresses and distortion compared to
V groove design but too much narrow groove was resulted in higher bending stress. For narrow
groove weld design with pulse GMAW, about 35-45% reduction in transverse shrinkage was

43
 ACCEPTED MANUSCRIPT

reported compared to conventional V groove pulse GMAW and SMAW [175]. Kumar et al.
[176] reported that GTAW process minimized the transverse shrinkage about 30% compared to
GMAW process due to lower heat input used in GTAW. Basavaraju performed a simplified
elastic-plastic approach to measure the strains formation due to radical shrinkage of pipe butt
weld [165]. Number of works has also reported related to 3D finite element (FE) simulation of

PT
residual stress and distortion. Yaghi et al. [177] performed a comparative study between
numerically and experimentally evaluated residual stress value for P91 pipe weldments and

RI
showed a good agreements between the results. The residual stresses at outer surface were
measured using X-ray diffraction while inner surface residual stress measurement performed

SC
using deep-hole drilling technique. Paddea et al. [99] performed an experimental study to
measure the residual stress distribution in P91 girth pipe weldment and also study the effect of

U
PWHT on their magnitude and distribution. The boundary of HAZ and base metal that is IC-
HAZ showed the maximum tensile residual stress value about 600 MPa. The maximum tensile
AN
hydrostatic stress (400 MPa) was also observed near the vicinity of HAZ. The PWHT was
resulted in considerable lowering in residual tensile stress (about 24% of yield strength of base
M

metal) and hydrostatic stress (50 MPa). The compressive residual stress was reported in weld
fusion zone near the final weld pass. Venkata et al. [115] studied the effect of PWHT on
D

residual distribution in electron beam welded P91 plate. The residual stress was measured using
neutron diffraction technique and compared with numerically predicted values. PWHT
TE

temperature and time was observed a significant effect on residual stress distribution. To
minimize the residual stress, maximum PWHT duration about 2 hours and temperature about
EP

770 oC below than Ac1 temperature was reported. Chuvas et al.[178] also reported that PWHT
at 760 oC for 2 h was resulted a drastic decrease in residual stress value both in root and cap of
C

the weld. Kulkarni et al. [179] had performed a comparative analysis of SMAW, GMAW, and
pulse GMAW process and their effect on residual stress formation during welding of thick
AC

stainless steel pipe. Kim et al. [180] predicted the residual stress distribution in P91 welded
structure by using the neutron diffraction technique and to enhance the quality of results, the
experimental results fairly matched with the numerical results obtained from FE simulation.
Welding sequence has observed a great influence on residual deformation. Sattari-Far and
Javadi [181] studied the effect of welding sequence on distortion in pipe-pipe butt joints. The
experimental results were validated with numerically obtained value. Welding leads the

44
 ACCEPTED MANUSCRIPT

diametrical variation in pipes that mainly depend on the welding sequence. The diametrical
variation in welding section was observed to be decreased (negative) and became zero at some
distance away from the welding section and afterwards increased.
Deng et al. [171] performed the residual stress estimation in multi-pass butt-welded
thick pipe by strain gauge method and the experimentally observed residual stress value

PT
matched with numerically obtained residual stress value from 2-D axi-symmetric FE model.
The influence of weld metal yield strength on residual stress was also observed. The stresses

RI
induced in material during the strengthening and heat treatment process have observed a great
effect on residual stresses. The strengthening of material resulted in significant increase in

SC
material yield strength and this might be the cause of higher residual stress after welding. Dong
et al. [173] studied the effect of pipe wall thickness on residual stress distribution. Pipe wall

U
thickness showed a strong influence on residual stress specially the axial component of residual
stress. Axial residual stress showed compressive nature at outer surface while tensile at inner
AN
surface near the vicinity of weld. Axi-symmetric model was also developed for numerical
prediction of residual stress. In start and stop point of welding, presence of both high
M

magnitudes of tensile and compressive were resulted in stress gradient. Zhao et al. [182] also
reported the high magnitude of residual stress near the start/stop position for dissimilar multi-
D

pass welding of 12Cr1MoV/P91 pipe. However, this effect was observed to be relieved by
homogeneous distribution of start/stop position.
TE

The most commonly used method for the residual stress estimation in welded joint is blind hole
drilling method. The magnitude and orientation of principle residual stress can be estimated
EP

using the Eq. 23– 25, as shown below [183].

 


 = +  −   +  +  − 2  (23)
̅ 
C

 


 = −   −   +  +  − 2 
AC

̅
(24)


 !
tan 2 = (25)



Where ɛ1, ɛ2 and ɛ3 are the strains readings of strain rosette elements 1, 2 and 3, respectively. The
angle between element 1 and minimum principle stress are presented by α.

45
 ACCEPTED MANUSCRIPT

In Eq. 23 and 24, calibration coefficient "̅ and # stated are calculated using Eq. 26 and 27 [183].
Where, ɛ %&' and ɛ %&' shows the difference of strain reading in principle direction of 1 and 3,
respectively. For P91 steel, the variations of coefficients "̅ and # with respect to the ratio of
applied stress to yield strength of work materials is depicted in Fig. 21(a). Eq. 28 & 29 are used
to calculate the calibration coefficients a and b used in residual stress measurement [183].

PT
"̅ = ɛ %&' + ɛ %&' /2σ&** (26)

RI
# = ɛ %&' − ɛ %&' /2σ&** (27)

 +
"̅ = − (28)

SC
,

-
# = − , (29)

U
The error variation with respect to ratio of applied stress and yield strength of work materials
AN
(σapp /σy) is presented in Fig. 21(b). The percentage error of stresses in longitudinal and
transverse directions are calculated using Eq. 30 and 31, respectively.

Error % = 100 ∗ σ5&678 − σ5&** /σ5&**

M

(30)

Error % = 100 ∗ σ9&678 − σ9&** /σ9&** (31)

D

For biaxial loading, Figure 22 shows the percentage errors for longitudinal and transverse
TE

directions. The ratio of the ASTM calculated stresses to yield strength (
:;< /
= and
>;< /

= ) were also utilized to develop the relationship between ASTM calculated and corrected
EP

residual stress value. Equations 32 and 33 shows the relations. The error in residual stress
calculation depends on so many factors like biaxial stress ration, yield strength of material, and
applied stress.
C

> = −32.7 − 241
>;< /
= − 0.006
:;< + 1.2
>;< (32)
AC

: = −6.21 + 404.58
:;< /
= − 0.26
:;< − 0.109
>;< (33)

Hempal et al. [184] performed the residual stress measurement in multi-pass welded
ferritic-pearlitic pipes, using the X-ray diffraction technique. The phase transformation in weld
and HAZ played an important role in residual stress formation. In weld zone, phase
transformations have observed a minor influence but in HAZ a great influence was observed

46
 ACCEPTED MANUSCRIPT

due to formation of high strength martensitic phase. At inner surface of weld toe, maximum
tensile axial residual stress was observed which is considered as preferential sites for the fatigue
crack initiation.
9 Fundamental of creep
Creep is a thermally activated process. It is defined as slow time-dependent deformation of

PT
material under the influence of constant load. Creep occurs due to long-term exposure of
material under constant applied stress below than yield strength of material. Material properties,

RI
applied constant load, exposure time, and exposure temperature are the main factors that govern
the rate of deformation. The temperature dependent mechanisms like bulk diffusion, slip and

SC
climb dislocations are response for creep phenomena in crystalline material. The thermally
activated creep process follows the Arrhenius law as given below;
− Ea

U
k = Ae (34)
RT
AN
where k is the strain rate, A is pre-exponential factor, Ea is the activation energy for creep, R is
the universal gas constant, and T is temperature (K).
M

The creep phenomenon of material is associated with three stages: primary creep or transient
stage, secondary or steady state creep, tertiary or unstable creep, as shown in Fig. 23. Primary
creep with relatively short life deals with rapid strain rate initially and after that reduce strain
D

rate. Primary stage is characterized with high dislocation density, subgrain formation, evolution
TE

of precipitates, and strengthening of material. In steady stage, rate of strain change with time
remains almost constant. In P91 steel, steady stage creep evolves the diffusion control structural
EP

change like coarsening of secondary phase specially M23C6, low dislocation density, wide lath
martensite and formation of laves and Z-phase. This stage leads the weakening of material.
Final stage have steep rise in strain rate. In P91 steel, tertiary stage deals with further
C

coarsening of precipitates, void nucleation along the PAGBs and at precipitates, cavitations and
AC

necking.
In each stage, a creep phenomenon is controlled by number of mechanisms that strongly
depend on metallurgy of material and creep exposure temperature and applied stress. Diffusion
flow and dislocation creep are the main phenomena that controlled the mechanism. Stress
controlled and time-dependent movement of vacancies from the grain boundary under tension
to grain boundary under compression is characterized by diffusional creep and the movement is

47
 ACCEPTED MANUSCRIPT

termed as Coble creep while the movement of vacancies inside the grain region is termed as
Nabarro-Herring (NH) creep. Coble creep results in grain deformation that introduces the
macroscopic strain in material, as shown in Fig. 24. Low stress is generally referred as ideal
condition for Coble creep while low stress with high temperature is for NH creep.
Dislocation creep deals with time dependent movement of dislocations through obstacles as well

PT
as grain boundary sliding. Dislocation creep behaves in opposite nature of diffusional creep and
it acts as high stress and low temperature. Under high stress, dislocations move freely from one

RI
lattice plane to other and encountered the secondary phase particles. This type of mechanism is
termed as climb. The continuous movement forces the lattice to rearrange themselves and

SC
ultimately involvement of diffusion. The dislocation creep remains independent to grain size and
grain boundary involvement. Here strain rate controlled by diffusion and climb dislocation.

10
U
Creep characteristic and microstructure evolution in P91 base metal
AN
A lot of researchers have worked on microstructure evolution of P91 base metal under creep
exposure condition. Creep exposure of P91 steel leads the various changes in microstructure
such as lath widening, martensite recovery, precipitate coarsening, grain coarsening, evolution
M

of new phases and emergence of subgrains. The microstructure changes that involved during
creep of P91 steel are discussed below.
D

10.1 Precipitate coarsening

TE

Role of precipitates on microstructure stability have been already discussed in previous section.
In P91 steel, M23C6 and MX are main precipitates that play crucial role during creep exposure.
EP

Coarsening of M23C6 precipitates has been reported by many of researchers [185,186]. The
coarsening rate is governed by the 'Ostwald ripening' mechanism that is characterized with
increase in precipitate size, mean particle-to-particle spacing, reduction in average number of
C

particles for constant volume fraction of precipitates. The long-term creep exposure results in
AC

growth of large particles at the expense of fine particles due to bulk diffusion of atoms from
small particles to large particles. The coarsening process depends on the time as per Ostwald
ripening law;

d3 = do3 + kt (35)
where, d is mean particle diameter after creep, do is initial particle diameter (t=0), and k is
particle growth rate.

48
 ACCEPTED MANUSCRIPT

The reduction in number of particle and increase in particle spacing play a key role in creep
strength reduction by undermines the retardation force of precipitates on boundary pinning, lath
recovery and dislocation. Tamura et al. [187] had reported that excessive precipitates might be
the cause of creep strength reduction in P91 steel. The grain boundaries with excessive
precipitates were believed to be weaker compared to matrix region. The boundaries with

PT
excessive precipitates led the higher rupture life in short term creep with high stress. However,
in case of long-term creep with low stress, a reverse trend was reported.

RI
The high coarsening rate of M23C6 precipitates causes a low influence on creep rupture strength
with time. Panait et al. [185] had studied the growth rate for 100,000 h of creep exposure at 600

SC
o
C and observed that the average precipitate diameter increased from 150-180 nm to 300 nm.
Higher coarsening rate of precipitates make the boundaries free form the pinning effect and

U
resulted in sudden drop of creep strength. The other important change was reported in the
composition of M23C6 during the long term service exposure [185,186]. Aghajani et al. [186]
AN
reported a continuous change in chemical composition of M23C6 with time and about 50000 h a
new equilibrium was established. Panait et al. [185] reported an increase in Cr content in M23C6
M

from 60 at.% to 70 at.% for 100000 h of creep exposure at 600 oC. During the creep exposure,
fine MX particle shows higher stability than M23C6. Aghagini et al. [186] reported that volume
D

fraction, size and chemical composition of MX precipitates remain constant during creep
exposure. Vodarek and Strang [188] had also reported higher chemical stability of MX phase
TE

during creep exposure at 550 oC and 600 oC. Sawada et al. [189] had also reported the very slow
coarsening rate of MX compared to M23C6. Coarsening of MX was observed to be sensitive and
EP

accelerated by strain [190].

10.2 Martensite lath widening and recovery

C

The other additional change that occurs during the creep exposure is martensitic lath
AC

morphology. For 10% Cr steels, Hu et al.[191] studied the lath widening during creep exposure
at 600 oC. The lath width was observed to be increased with rupture life and on further expansion
it evolved into sub-grains. The lath widening and sub-grain formation during creep was resulted
continuous movement and annihilation of dislocations that led to reduction in creep strength.
Sawada et al. [192] reported that the local parts of lath boundary subjected to repetitive bulging
and migration that led to final migration of whole lath boundary at tempering temperature about

49
 ACCEPTED MANUSCRIPT

667 oC. The small laths were observed to be shrunk and get disappeared while the large one got
wider. Sawada et al. [193] reported that lath structure showed higher thermal stability in
tempering condition and lath width remained constant during thermal ageing. The growth rate of
lath width was observed to be faster at higher operating temperature and stress. Hence, the
driving force for lath widening is derived from the strain accumulated by creep deformation. The

PT
lath boundary movement can be stopped by pinning the boundaries from precipitates. As per
Zener relationship, higher resistance to grain coarsening is applied by the volume fraction of

RI
precipitates and it is given by;

4r

SC
D= (36)
3f

where, D is limiting grain size, r is radius of precipitates and f is volume fraction of precipitates.

U
During the bulging out, the lath boundary move towards the adjacent lath and motion is retarded
AN
by the small fine MX precipitates decorated along the lath boundaries. The fine MX particles are
easily crossed by the lath boundary. Hence, growth of lath boundary will continue the bulging
M

out until the next lath or grain boundary reaches. The morphology of lath also deals with its
shape. The needle-shape lath structure provided the more equiaxed lath morphology with higher
D

stability.
TE

10.3 PAGBs disappearance

The long-term creep exposure results in disappearance of PAGBs because of high coarsening
EP

rate. The main precipitates that decorate the PAGBs are Cr, Fe and Mo-rich M23C6 and fine MX.
The PAGBs acts as dislocation barrier during the creep exposure. M23C6 can effectively pin the
C

migration of PAGBs but evolution of unfavorable laves phase on PAGBs during long-term creep
exposure results in swallowing of M23C6 precipitates. Laves phase form a cluster around the
AC

M23C6 by nucleating and growing on it [18]. The clustering behaviour cause the swallowing of
M23C6 particles and boundaries become free from the pinning effect of M23C6 and other
precipitates. Hence, PAGBs becomes disappear during the long-term exposure by evolution of
laves phase and dislocation movements. Maruyama et al. [40] had reported that M23C6 particles
provided more effective pinning to sub-grain boundaries compared to MX precipitates because
the small size MX precipitates are easily crossed by the boundaries.

50
 ACCEPTED MANUSCRIPT

10.4 Emergence of subgrain boundaries

Solid solution strengthening and carbide-stabilized substructure hardening are the main factors
that govern the strength of 9-12% Cr steels [22,194]. The subgrain boundary acts as obstacles
against gliding of mobile dislocations and consider as other important creep strengthening

PT
mechanism [40]. Subgrains are generally formed polygonal shape with size about 1 micron. Yan
et al. [195] reported that creep deformation and enhance in operating temperature without any
stress could promote the formation of subgrains. The mobility of dislocations plays a crucial role

RI
in subgrain formation. The mobility of dislocations could be improved by enhancing the thermal
ageing while number of mobile dislocations would be increased by creep deformation. Yan et al.

SC
[195] reported that both creep deformation and enhanced ageing temperature can accelerate the
subgrain formation. The difference in mobility of dislocations was observed mainly due to

U
difference of dislocation density and subgrain morphology. It was reported that during creep
AN
deformation, quick and interactive movement of dislocation led the formation of subgrain
boundaries under the flowing strain. In creep deformation, the mobility of dislocations could not
be pinned by fine MX precipitates. It was difficult to achieve such high mobility dislocation by
M

enhancing the ageing temperature without stress as flowing strain and many of dislocations were
pinned by fine MX precipitates. Hence, in both the cases different morphology of subgrains were
D

formed. Interaction of flowing strain driven dislocations at high temperature lead to

TE

polygonization and results in formation of subgrain boundaries. Keller et al. [196] had explained
the movements of dislocations. Subgrain size was observed to be increased during long-term
creep exposure while less affected by long-term ageing [185,186]. In short term creep, stability
EP

of subgrains is controlled by flowing strain while in long-term creep precipitates govern the
stability of subgrains. Aghajani et al. [186] reported that the precipitate in contact with subgrains
C

grow faster as compared to precipitates lost their contact with subgrains due to growth. During
AC

the creep deformation, precipitate coarsening results in loss of pinning effect and ultimately
coarsening of subgrains. In long term creep, strains are reduced upto a great extent that did not
affect the subgrain coarsening. Hence, the coarsening of precipitates becomes more critical in
terms of subgrain stability. The schematic formation of subgrains from tempered martensitic
laths is shown in Fig. 25.
In P91 steel, carbide stabilized tempered microstructure and subgrain strengthening are the main
hardening mechanisms in short term and long term creep exposure, respectively. The migration

51
 ACCEPTED MANUSCRIPT

of subgrain boundaries is mainly controlled by pinning force of precipitates. In P91 steel, highest
volume fraction of MX and M23C6 lead the thermal stability of subgrains boundaries [25].
Subgrain boundaries are actually wall of dislocations and easily crossed the fine MX precipitates
during coarsening. However, it is difficult to cross the M23C6 precipitates because of similar size
of M23C6 and subgrain boundaries thickness. Hence, M23C6 can be considered as provide most

PT
effective pinning compared to fine MX. Aghajani et al. [197] and also observed that M23C6
provide most effective pinning to subgrain boundaries compared to MX. Aramaki et al. [40] also

RI
reported that due to similarity in spacing of M23C6 precipitate and subgrain size, M23C6 provided
more effective pinning than MX. Kulesh et al. [198] developed a relation between subgrain

SC
coarsening and dislocation density and observed that dislocation density decreased with subgrain
coarsening. The reduction in dislocation density resulted in decrease in hardness. The decrease in

U
hardness caused a poor mechanical resistance to high temperature strength and promotes the
subgrain and precipitate coarsening.
AN
10.5 Precipitation of inter-metallic phase (Laves and Z-phase)
M

Laves and Z-phase precipitation and their characteristic have already been explained in the
previous section. The formation of inter-metallic phase in pre-service or in virgin state has not
D

been reported yet. The research indicates the formation of intermetallic phase occur in thermal
exposure condition. Laves phase generally nucleates near the vicinity of coarse precipitates, at
TE

PAGBs, or at intergranular sites. In thermal ageing condition, stage during which laves phase
nucleate is still unclear. In P91 steel, Hald [199] had reported the possible nucleation of laves
EP

phase occurred in temperature range of 400-500 °C. For P91 steel, in stress-free ageing laves
phase precipitation was observed about 3000 h at 600 oC and 10000 h at 650 oC [200]. Beyond
the temperature, the other parameters that affects laves phase formation is material composition.
C

For 10 % Cr steel, Cui et al. [21] had observed the precipitation of laves phase about 400 h at
AC

600 oC and 204 h at 650 oC. At higher creep exposure temperature (650 oC) diameter of laves
phase was observed to be high as compared to lower exposure temperature (600 oC) [18].
Specimens at higher creep exposure temperature showed the lower density of laves phase. In the
initial phase of 9-12% Cr steel development, Mo equivalent (Mo + 0.5 W, wt%) was used as the
optimization criteria. 9-12% Cr steel with molybdenum equivalent greater than 1 % lead the
formation of laves phase. The Z-phase precipitation is occurred after a long-term creep exposure.

52
 ACCEPTED MANUSCRIPT

In P91 steel, the precipitation of Z-phase has been reported too slowly compared to 12% Cr
steels. Sawada et al.[201] had reported the possible precipitation of Z-phase about 30000 h at
600 oC under applied stress of 70 MPa. As discussed earlier, the precipitation of Z-phases results
in reduction of creep strength due to consumption of fine MX precipitates. In 11-12% Cr steel,
premature failure of steel during short-term creep about 15000 h is attributed to formation of Z-

PT
phase [2]. Z-phase precipitation does not show any inflection point in creep curve [47]. At 550
o
C and upto 50000 h, Z-phase formation is not considered to be crucial for the creep strength.

RI
11 Creep characteristic of P91 weldments
The study related to the creep characteristic of P91 weldments mainly focused on the type IV

SC
cracking. A schematic evolution of cracks formed in P91 weldments is shown in Fig. 26.
Cracks are defined on the basis of their location in weldments. The welding of P91 steel leads

U
the formation of heterogeneous microstructure across the weldments. The different zone present
AN
in weldments shows different susceptibility to creep damage. Type I and type II crack mainly
develop in deposited weld metal, propagate either through base metal (type I) or cross over in
HAZ or base metal (type II). Type III cracking occurs in CGHAZ and it can be minimized by
M

refining the grain size. In P91 welded joint, localized creep cavitations in soft zone led to
intergranular fracture during the long-term creep exposure at low stress. This is termed as Type
D

IV cracking.
TE

From study, it is clear that the type IV cracking mainly nucleate from the soft-zone present in
weldments but it is difficult to clearly distinguish the location of soft zone. The similar optical
appearance of OTBM, FGHAZ and IC-HAZ make it difficult to distinguish the soft-zone in
EP

weldments. In different study, OTBM, FGHAZ and IC-HAZ has been considered as the soft
zone of weldments [202–204]. The carbide coarsening and partial dissolution of precipitates in
C

FGHAZ and IC-HAZ make it relatively soft compared to CGHAZ and weld zone. The soft
AC

zone formation leads to premature type IV cracking in FGHAZ/IC-HAZ. The type IV fracture
is generally characterized as low ductility fracture because of low strain value about 10%.
Nucleation and growth of creep cavities by diffusion mechanism in FGHAZ and IC-HAZ are
mainly responsible for the type IV cracking [205].
The presence of secondary phase particles at PAGBs or at the intersection of grain boundaries
acts as the preferential sites for void nucleation. The coarsening of retained M23C6 particles was
observed to be accelerated under the creep condition. Hence, retained M23C6 particles in

53
 ACCEPTED MANUSCRIPT

FGHAZ/IC-HAZ acts as preferential sites for crack nucleation. Abd El-Azim et al. [206] had
reported that softer martensitic matrix formation in FGHAZ/IC-HAZ, during the welding
process is responsible for the type IV failure. The coarsening of precipitates and formation of
laves and Z-phase are observed to be faster in IC-HAZ compared to other zone of weldments
[207,208].

PT
A lot of works have been published on type IV cracking of P91 weldments. A limited work
have been reported related to cross-weld creep test of 9-12 % Cr steels [207,209,210]. Type IV

RI
cracking in cross-weld specimen was observed to be more pronounced for stress less than 100
MPa [211]. A lot of researchers have confirmed the poor creep strength of weld joint than the

SC
base metal due to premature type IV cracking [93,206,208,212]. The poor strength of weld
joint is attributed to type IV failure. Watanabe et al. [208] have observed the poor creep

U
strength of weld metal than the base metal for P91 steel weldments. Corresponding to low
temperature and high stress (160 Mpa), Type I fracture was observed while decreased in stress
AN
from 160 MPa to 80/40 MPa at higher temperature (660/650 oC), led the type IV fracture in P91
weldments. During creep, a remarkable growth of laves and M23C6 phase was also reported in
M

FGHAZ. Abd El-Azim et al. [213] have performed a comparative study on creep behaviour of
P91 steel joint and base metal. The creep strength of base metal was observed to be superior to
D

weld joint at higher creep exposure temperature about 650 oC while at lower temperature about
600 oC, it approached to creep strength of base metal. FGHAZ was observed to be softest zone
TE

in weld joints. Fracture location was observed in base metal near to HAZ for short term creep
exposure (426 h) at 600 oC and high applied stress of 150 MPa. The lower applied stress for
EP

same applied temperature was resulted in shifting of fracture location and type IV fracture
occurred in FGHAZ of weld joint. Laha et al. [93] reported the soft zone formation in IC-HAZ
C

of P91 joint and type IV fracture was also associated with IC-HAZ.
P91 steel joint exhibited the lower creep rupture strength than base at higher exposure
AC

temperature and lower applied stress while at low temperature and high applied stress reverse
result was obtained. The fracture behaviour of P91 joint was observed to be similar as reported
by Abd El-Azim [213]. For low temperature and high applied stress, fracture location was
observed in base metal near to HAZ. The fracture location was observed to be shifted from base
metal to IC-HAZ for lower applied stress at higher exposure temperature. Albert et al. [89] have
also studied the effect of PWHT duration on type IV cracking nature in P122 weld joint. The

54
 ACCEPTED MANUSCRIPT

creep test was performed at 650 oC for 70 MPa. PWHT duration was varied from 15 min to 4 h
but no any significant change was observed in creep rupture time and fracture behaviour. Sawada
et al. [95] have studied the cross-weld long-term creep behaviour of E911 joint at 600 oC.
Fracture was noticed in soft FGHAZ. The growth of M23C6 in FGHAZ was observed faster than
the base metal. Z-phase formation was observed both in FGHAZ and base metal FGHAZ

PT
exhibited higher number density of Z-phase than the base metal. For long-term creep exposure at
low level of stress, creep life was observed to be minimum in FGHAZ, which have not minimum

RI
hardness [211]. Hence, Hardness along the weldments in transverse direction cannot be used as
location of type IV cracking. Hence, type IV cracking in P91 weld joint result in reduction in

SC
cross-weld creep strength during the long term creep exposure. Researchers have proposed
various technique such as normalizing/tempering treatment, adjusting welding procedure, and

U
modifying tempering treatment to enhance the cross-weld creep strength of P91 weld joint
[93,205,211].
AN
As discussed above, heterogeneity across the weldments are mainly responsible for the type IV
cracking. PWHT resulted in recrystallization that enhance the non-uniform distribution of
M

precipitates and also reduces the heterogeneity presented across the weldments. To enhance the
cross-weld creep strength of P91 joint, PWHT was not observed to be more effective. To
D

remove the heterogeneity across the sub-zone of P91 welded joint, the entire weldments need to
TE

renormalize. Renormalizing of welded component lead the formation of homogeneous

microstructure and alleviates the premature type IV cracking. Mitsubishi Heavy Industries in
Japan have successfully implemented the N&T of P91 weldments and observed little or no
EP

difference in creep behaviour of cross weld P91 weld joint and base metal for creep exposure
time up to 30000 h and temperature of 700 oC [214]. However, in actual practice, N&T
C

treatment of long welded components is not possible. Local normalizing might be the cause of
AC

secondary soft zone formation.

The half-tempering technique might be also used to enhance the cross-weld creep rupture
strength of weld joint. A half-tempered weld joint have reported a superior creep strength than
conventional weld joint for short-term creep test [215]. Half-tempering technique utilized the
partial tempering of base metal before the welding and completes the tempering as PWHT. The
effect of heat input and preheat temperature on type IV cracking behaviour have been reported
by Francis et al. [216]. A narrow HAZ formation is also expected to enhance the creep rupture

55
 ACCEPTED MANUSCRIPT

life of cross-weld specimen [209]. A joint with narrow HAZ results in lower creep cavitation in
soft zone and minimize the tendency of type IV cracking. The addition of element like B and W
in parent metal composition might be useful to enhance the creep rupture strength of welded
joint [217].
12 Future work

PT

Optimum combination of normalizing and tempering temperature for P91 steel and
weldments seems to have not been reported in the literature.

RI

Effect of N&T and subsequent PWHT on microstructure evolution and mechanical
properties of GTAW and SMAW P91 joints have not been reported in literature.

SC

Effect of groove geometry on mechanical properties and high temperature Charpy
toughness of P91 welds have not reported yet.

U

Effect of groove geometry on transverse shrinkage stress and residual stress of P91pipe
AN
weld joints have not observed in literature.

Effect of varying diffusible hydrogen level on weldability of P91 steel has not reported
in literature.
M

Effect of varying diffusible hydrogen level on tensile properties, residual stress and
flexural strength of single pass and multi-pass shielded metal arc welded P91 joint seems
D

to have not been reported in literature.

TE

Comparative study of PWHT and N&T on microstructure evolution in sub-zones of P91

weld joint and creep rupture life have not reported yet.
EP

13 Conclusions

In present article, microstructure evolution in P91 steel base and weld metal in different heat
C

exposure condition and their effect on performance of material has been reviewed. The virgin
state (N&T) microstructure consisted of tempered martensitic microstructure which derived their
AC

stability from precipitation hardening, solid solution hardening and sub-grain hardening. In
initial state microstructure that governed the mechanical properties of P91 steel mainly consisted
of Cr and Mo-rich M23C6, M7C3, M3C, M2X, Cr2N and MX precipitates while during long-term
creep exposure intermetallic laves and Z-phase governed the stability of microstructure. The
normalizing and tempering temperature were found a pounced effect on the microstructure
evolution and mechanical properties of P91 steel. The elements present in composition of P91

56
 ACCEPTED MANUSCRIPT

steel play an important role in deciding the creep rupture behaviour of P91 steel specially V, Nb,
Cr, Mo and W. Sometimes unbalance in composition led to formation of unwanted δ-ferrite
phase that significantly affected the mechanical properties of P91 steel. The welding of steel has
observed a serious issue due to heterogeneous microstructure formation across the weldments.
To remove the heterogeneity, subsequent PWHT and N&T treatment was recommended by the

PT
researchers but for small components N&T treatment have found to be more effective. The other
factors that affect the performance of weldments are residual stress and diffusible hydrogen

RI
content. Residual stresses can also be controlled by the heat treatment process but the presence of
diffusible hydrogen in weldments led to catastrophic failure of weldments after the welding or

SC
after few years of welding. The performance of weldments was also affected during the long-
term creep exposure due to formation of intermetallic-phases, lath widening and coarsening of

U
subgrains. To overcome the diffusible hydrogen induced cracking, preheating of plate, baking of
electrode, use of low hydrogen electrode was recommended. For long-term creep performance of
AN
cross-weld specimen, addition of B and W, half-tempering, and normalizing/tempering was
suggested.
M

References:
D

[1] M.Ã. Yoshizawa, M. Igarashi, Long-term creep deformation characteristics of advanced

ferritic steels for USC power plants, 84 (2007) 37–43. doi:10.1016/j.ijpvp.2006.09.005.
TE

[2] J. Hald, Microstructure and long-term creep properties of 9 – 12 % Cr steels, Int. J. Press.
Vessel. Pip. 85 (2008) 30–37. doi:10.1016/j.ijpvp.2007.06.010.
EP

[3] C. Pandey, N. Saini, M.M. Mahapatra, P. Kumar, Study of the fracture surface
morphology of impact and tensile tested cast and forged (C&F) Grade 91 steel at room
C

temperature for different heat treatment regimes, Eng. Fail. Anal. 71 (2016) 131–147.
AC

doi:10.1016/j.engfailanal.2016.06.012.

[4] I. Carvalho, H. Schut, D.A. Perminov, A.P. Druzhkov, V.L. Arbuzov, Microstructural
Characterization of Ferritic / Martensitic Steels by Positron Annihilation Spectroscopy,
(n.d.). doi:10.1088/1742-6596/443/1/012010.

[5] R. De Rhode, High Temperature Properties of Improved 9 % Cr, 45 (1991) 163–178.

57
 ACCEPTED MANUSCRIPT

[6] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Effect of creep phenomena on room-
temperature tensile properties of cast and forged P91 steel, Eng. Fail. Anal. 79 (2017)
385–396. doi:10.1016/j.engfailanal.2017.05.025.

[7] P.J. Ennis, A. Czyrska-Filemonowicz, Recent advances in creep-resistant steels for power

PT
plant applications, Sadhana. 28 (2003) 709–730. doi:10.1007/BF02706455.

[8] H. Zheng-fei, Y. Zhen-guo, Identification of the Precipitates by TEM and EDS in

RI
X20CrMoV12 . 1 After Long-Term Service at Elevated Temperature, 12 (2003) 106–111.

[9] L. Cipolla, H.K. Danielsen, D. Venditti, P. Emilio, D. Nunzio, J. Hald, M.A.J. Somers,

SC
Conversion of MX nitrides to Z -phase in a martensitic 12 % Cr steel, Acta Mater. 58
(2010) 669–679. doi:10.1016/j.actamat.2009.09.045.

U
[10] H.K. Danielsen, J. Hald, A thermodynamic model of the Z-phase Cr(V, Nb)N, Calphad
AN
Comput. Coupling Phase Diagrams Thermochem. 31 (2007) 505–514.
doi:10.1016/j.calphad.2007.04.001.
M

[11] A. Golpayegani, H. Andr, H. Danielsen, J. Hald, A study on Z-phase nucleation in

martensitic chromium steels, 489 (2008) 310–318. doi:10.1016/j.msea.2007.12.022.
D

[12] M. Marietta, E. Systems, Ornl/tm—9045 de85 012618, (1984).

TE

[13] R.W. Swindeman, M.L. Santella, P.J. Maziasz, B.W. Roberts, K. Coleman, Issues in
replacing Cr – Mo steels and stainless steels with 9Cr – 1Mo – V steel, Int. J. Press.
EP

Vessel. Pip. 81 (2004) 507–512. doi:10.1016/j.ijpvp.2003.12.009.

[14] Y. Wang, K. Mayer, A. Scholz, C. Berger, H. Chilukuru, K. Durst, W. Blum,

C

Development of new 11 % Cr heat resistant ferritic steels with enhanced creep resistance
for steam power plants with operating steam temperatures up to 650 ◦ C, 511 (2009) 180–
AC

184. doi:10.1016/j.msea.2008.04.116.

[15] X.Y. Liu, T. Fujita, Chromium Chromium, ISIJ Int. 29 (1989) 680–686.

[16] T. Publications, No Title, (2000) 419–426. doi:10.4028/www.scientific.net/KEM.171-

174.419.

[17] T. Publications, No Title, (2000) 499–504. doi:10.4028/www.scientific.net/KEM.171-

58
 ACCEPTED MANUSCRIPT

174.499.

[18] G. Dimmler, P. Weinert, E. Kozeschnik, H. Cerjak, Quantification of the Laves phase in

advanced 9-12% Cr steels using a standard SEM, Mater. Charact. 51 (2003) 341–352.
doi:10.1016/j.matchar.2004.02.003.

PT
[19] J. Hald, L. Korcakova, Precipitate Stability in Creep Resistant Ferritic Steels –
Experimental Investigations and Modelling, ISIJ Int. 43 (2003) 420–427.

RI
[20] L. Korcakova, J. Hald, M.A.J. Somers, Quantification of Laves phase particle size in
9CrW steel, Mater. Charact. 47 (2001) 111–117. doi:10.1016/S1044-5803(01)00159-0.

SC
[21] H. Cui, F. Sun, K. Chen, L. Zhang, R. Wan, A. Shan, J. Wu, Precipitation behavior of
Laves phase in 10%Cr steel X12CrMoWVNbN10-1-1 during short-term creep exposure,

U
Mater. Sci. Eng. A. 527 (2010) 7505–7509. doi:10.1016/j.msea.2010.08.013.
AN
[22] N.N. G. EGGELER, J. C. EARTHMAN, B. ILSCHNER, MICROSTRUCTURAL
STUDY OF CREEP RUPTURE IN A 12 % CHROMIUM FERRITIC STEEL, Acta
M

Metall. 37 (1989) 49–60.

[23] P. Hofer, H. Cerjak, S. Bernhard, The Influence of M As pects on the Service Behaviour
D

of Advanced Power Plant Steels, ISIJ Int. 39 (1999) 874–888.

TE

[24] R.L. Klueh, Elevated temperature ferritic and martensitic steels and their application to
future nuclear reactors, Int. Mater. Rev. 50 (2005) 287–310.
EP

doi:10.1179/174328005X41140.

[25] K. Maruyama, K. Sawada, J. Koike, Strengthening mechanisms of creep resistant

C

tempered martensitic steel, ISIJ Int. 41 (2001) 641–653.

doi:10.2355/isijinternational.41.641.
AC

[26] Å. Gustafson, J. Ågren, Possible Effect of Co on Coarsening of M 23 C 6 Carbide and

Orowan Stress in a 9 % Cr Steel, 41 (2001) 356–360.

[27] M.L. Santella, R.W. Swindeman, R.W. Reed, J.M. Tanzosh, Martensite formation in 9 Cr-
1 Mo steel weld metal and its effect on creep behavior, (n.d.).

[28] W.B. Jones, C.R. Hills, D.H. Polonis, Microstructural Evolution of Modified 9Cr-lMo

59
 ACCEPTED MANUSCRIPT

Steel, Metall. Trans. A. 22 (1991) 1049–1058.

[29] X. Li, M.T. Cabrillat, Y. Lejeail, Study of modified 9Cr-lMo welds, 2006.

[30] A. BALTUŠNIKAS, R. Levinskas, I.L.Ū.T. Ė, Kinetics of Carbide Formation During

Ageing of Pearlitic 12X1M Ф Steel, Mater. Sci. 13 (2007) 286–292.

PT
[31] V. Thomas Paul, S. Saroja, M. Vijayalakshmi, Microstructural stability of modified 9Cr-
1Mo steel during long term exposures at elevated temperatures, J. Nucl. Mater. 378 (2008)

RI
273–281. doi:10.1016/j.jnucmat.2008.06.033.

SC
[32] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Characterization of Cast and Forged
(C&F) Gr. 91 Steel in Different Heat Treatment Condition, Trans. Indian Inst. Met.
(2017). doi:10.1007/s12666-017-1144-4.

[33]
U
C. Pandey, A. Giri, M.M. Mahapatra, Effect of normalizing temperature on
AN
microstructural stability and mechanical properties of creep strength enhanced ferritic P91
steel, Mater. Sci. Eng. A. 657 (2016) 173–184. doi:10.1016/j.msea.2016.01.066.
M

[34] C. Pandey, A. Giri, M.M. Mahapatra, Evolution of phases in P91 steel in various heat
treatment conditions and their effect on microstructure stability and mechanical properties,
D

Mater. Sci. Eng. A. 664 (2016) 58–74. doi:10.1016/j.msea.2016.03.132.

TE

[35] B. Arivazhagan, M. Vasudevan, M. Kamaraj, Influence of Low Nickel ( 0 . 09 wt %)

Content on Microstructure and Toughness of P91 Steel Welds, Met. Mater. Int. 21 (2015)
EP

538–542. doi:10.1007/s12540-015-4351-8.

[36] Y. Wang, R. Kannan, L. Li, Characterization of as-welded microstructure of heat-affected

C

zone in modified 9Cr-1Mo-V-Nb steel weldment, Mater. Charact. 118 (2016) 225–234.
doi:10.1016/j.matchar.2016.05.024.
AC

[37] P. Polcik, T. Sailer, W. Blum, S. Straub, A. Orlova, On the microstructural development

of the tempered martensitic Cr-steel P 91 during long-term creep — a comparison of data,
260 (1999) 252–259.

[38] M. Yoshino, Y. Mishima, Y. Toda, H. Kushima, Influence of Normalizing Heat Treatment

on Precipitation Behavior in Modified 9Cr-1Mo Steel, (2005) 5–10.

60
 ACCEPTED MANUSCRIPT

[39] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, J.G. Thakre, Nano-size Particle
Evolution During Heat Treatment of P91 Steel and Their Effect on Micro Hardness,
Trans. Indian Inst. Met. (2017). doi:10.1007/s12666-017-1215-6.

[40] H.G. Armaki, R. 6. pdfipin. Chen, K. Maruyama, M. Igarashi, Creep behavior and

PT
degradation of subgrain structures pinned by nanoscale precipitates in strength-enhanced 5
to 12 Pct Cr ferritic steels, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 42 (2011)
3084–3094. doi:10.1007/s11661-011-0726-8.

RI
[41] C. Pandey, M.M. Mahapatra, P. Kumar, R.S. Vidyrathy, A. Srivastava, Microstructure-

SC
based assessment of creep rupture behaviour of cast-forged, Mater. Sci. Eng. A. 695
(2017) 291–301. doi:10.1016/j.msea.2017.04.037.

U
[42] H.-O.A. M. Ha¨ttestrand, Microstructural development during ageing of an 11 %
chromium steel alloyed with copper Mats Ha, 318 (2001) 94–101.
AN
[43] C. Panait, W. Bendick, A. Fuchsmann, J. Besson, C. Panait, W. Bendick, A. Fuchsmann,
Study of the microstructure of the grade 91 steel after more than 100 . 000h of creep
M

exposure at 600C To cite this version :, (2010) 0–13. doi:10.1016/j.ijpvp.2010.03.017.

D

[44] S. Kunimitsu, Y. You, N. Kasuya, Y. Sasaki, Y. Hosoi, Effect of thermo-mechanical

treatment on toughness of 9Cr-W ferritic-martensitic steels during aging, 181 (1991) 689–
TE

692.

[45] K. Swada, M. Takeda, K. Maruyama, R. Ishii, M. Yamada, Y. Nagae, R. Komine, Effect

EP

of W on recovery of lath structure during creep of high chromium martensitic steels, 267
(1999) 19–25.
C

[46] H.K. Danielsen, J. Hald, F.B. Grumsen, M. Somers, On the crystal structure of Z -phase
AC

Cr ( V , Nb ) N On the Crystal Structure of Z- Phase Cr ( V , Nb ) N, (2006).

doi:10.1007/BF02586098.

[47] K. Sawada, H. Kushima, K. Kimura, M. Tabuchi, TTP Diagrams of Z Phase in 9 – 12 %

Cr Heat-Resistant Steels, 47 (2007) 733–739.

[48] H.K. Danielsen, J. Hald, On the nucleation and dissolution process of Z-phase Cr ( V , Nb
) N in martensitic 12 % Cr steels, Mater. Sci. Eng. A. 505 (2009) 169–177.

61
 ACCEPTED MANUSCRIPT

doi:10.1016/j.msea.2008.11.019.

[49] F. Yin, W. Jung, S. Chung, Microstructure and creep rupture characteristics of an ultra-
low carbon ferritic / martensitic heat-resistant steel, 57 (2007) 469–472.
doi:10.1016/j.scriptamat.2007.05.034.

PT
[50] S.L. Mannan, S.C. Chetal, B. Raj, S.B. Bhoje, Selection of materials for prototype fast
breeder reactor, Trans. Indian Inst. Met. (2003) 1–35.

RI
http://www.igcar.gov.in/igc2004/PFBR.pdf%5Cnpapers3://publication/uuid/8B5AF2B8-
2294-410B-AEAC-C121CCD8FC4B.

SC
[51] C. Pandey, M. Mahapatra, Evolution of phases during tempering of P91 steel at 760 for
varying tempering time and their effect on microstructure and mechanical properties,

U
Proc. Inst. Mech. Eng. Part E J. Process Mech. Eng. 664 (2016) 58–74.
doi:10.1177/0954408916656678.
AN
[52] T. Shrestha, S.F. Alsagabi, I. Charit, G.P. Potirniche, M. V Glazoff, H.F. Lopez, Effect of
Heat Treatment on Microstructure and Hardness of Grade 91 Steel, Metals (Basel). 5
M

(2015) 131–149. doi:10.3390/met5010131.

D

[53] G. Golañski, Effect of the heat treatment on the structure and properties of
GX12CrMoVNbN9-1 cast steel, Arch. Mater. Sci. Eng. 46 (2010) 88–97.
TE

[54] G. Golañski, J. Słania, EFFECT OF DIFFERENT HEAT TREATMENTS ON

MICROSTRUCTURE AND MECHANICAL PROPERTIES OF THE MARTENSITIC
EP

GX12CrMoVNbN9-1 CAST STEEL, Arch. Metall. Mater. 58 (2013) 25–30.

doi:10.2478/v10172-012-0145-x.
C

[55] M. Yoshino, Y. Mishima, Y. Toda, H. Kushima, K. Sawada, Phase Equilibrium between

AC

Austenite and MX Carbonitride in a 9Cr – 1Mo – V – Nb Steel, ISIJ Int. 45 (2005) 107–
115.

[56] C. Hurtado-noreña, C.A. Danón, M.I. Luppo, P. Bruzzoni, Evolution of Minor Phases in a
P91 Steel Normalized and Tempered at Different Temperatures, Procedia Mater. Sci. 8
(2015) 1089–1098. doi:10.1016/j.mspro.2015.04.172.

[57] C. Hurtado Noreña, P. Bruzzoni, Effect of microstructure on hydrogen diffusion and

62
 ACCEPTED MANUSCRIPT

trapping in a modified 9 % Cr – 1 % Mo steel, Mater. Sci. Eng. A. 527 (2010) 410–416.

doi:10.1016/j.msea.2009.08.025.

[58] A. Chatterjee, D. Chakrabarti, A. Moitra, R. Mitra, A.K. Bhaduri, Materials Science &
Engineering A Effect of normalization temperatures on ductile – brittle transition

PT
temperature of a modi fi ed 9Cr – 1Mo steel, Mater. Sci. Eng. A. 618 (2014) 219–231.
doi:10.1016/j.msea.2014.09.021.

RI
[59] S.P. Selvi, T.S.P. Parameswaran, Effect of Normalization Heat Treatment on Creep and
Tensile Properties of Modified 9Cr – 1Mo Steel, Trans. Indian Inst. Met. 69 (2016) 261–

SC
269. doi:10.1007/s12666-015-0816-1.

[60] T. Karthikeyan, M.K. Dash, R. Mythili, S.P. Selvi, A. Moitra, S. Saroja, SC, J. Nucl.

U
Mater. 494 (2017) 260–277. doi:10.1016/j.jnucmat.2017.07.019.
AN
[61] T. Karthikeyan, V. Thomas Paul, S. Saroja, A. Moitra, G. Sasikala, M. Vijayalakshmi,
Grain refinement to improve impact toughness in 9Cr-1Mo steel through a double
austenitization treatment, J. Nucl. Mater. 419 (2011) 256–262.
M

doi:10.1016/j.jnucmat.2011.08.010.
D

[62] C. Pandey, M.M. Mahapatra, Effect of Long-term Ageing on the Microstructure and
Mechanical Properties of Creep Strength Enhanced Ferritic P91 Steel, Trans. Indian Inst.
TE

Met. (2016) 1–17. doi:10.1007/s12666-015-0826-z.

[63] V. Homolova´, J. Janovec, P. Za´humensky, A. Vy´rostkova, Influence of thermal-

EP

deformation history on e v olution of secondary phases in P91 steel, Mater. Sci. Eng. A.
349 (2003) 306–312.
C

[64] F. Kafexhiu, F. Vodopivec, J.V. Turna, Effect of tempering on the room-temperature

AC

mechanical properties of X20CrMoV121 and P91 steels, Mater. Tehnol. 46 (2012) 459–
464.

[65] B.A. Senior, B.L. Eyres, K. Avenue, C. House, The effect of ageing on the ductility of
9Cr-l Mo steel, Acta Mater. 36 (1988) 1855–1862.

[66] S. Sathyanarayanan, J. Basu, A. Moitra, G. Sasikala, Effect of Thermal Aging on Ductile-

Brittle Transition Temperature of Modified 9Cr-1Mo Steel Evaluated with Reference

63
 ACCEPTED MANUSCRIPT

Temperature Approach Under Dynamic Loading Condition, Metall. Mater. Trans. A. 44

(2013) 2141–2155. doi:10.1007/s11661-012-1510-0.

[67] G. Golañski, J. Kepa, The effect of ageing temperatures on microstructure and mechanical
properties of GX12CrMoVNbN9-1 (GP91) cast steel, Arch. Metall. Mater. 57 (2012)

PT
575–582. doi:10.2478/v10172-012-0061-0.

[68] B. Silwal, L. Li, A. Deceuster, B. Griffiths, Effect of Postweld Heat Treatment on the

RI
Toughness of Heat-Affected Zone for Grade 91 Steel, Weld. Res. 92 (2013) 80s–87s.

[69] C. Pandey, M.M. Mahapatra, P. Kumar, A. Giri, Microstructure Characterization and

SC
Charpy Toughness of P91 Weldment for As-Welded , Post-Weld Heat Treatment and
Normalizing & Tempering Heat Treatment, Met. Mater. Int. 23 (2017) 900–914.

U
doi:10.1007/s12540-017-6850-2.
AN
[70] C. Pandey, A. Giri, M.M. Mahapatra, P. Kumar, Characterization of microstructure of
HAZs in as-welded and service condition of P91 pipe weldments, Met. Mater. Int. 23
(2017) 148–162. doi:10.1007/s12540-017-6394-5.
M

[71] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Effect of normalization and tempering
D

on microstructure and mechanical properties of V-groove and narrow-groove P91 pipe

weldments, Mater. Sci. Eng. A. 685 (2017) 39–49. doi:10.1016/j.msea.2016.12.079.
TE

[72] Vijaya L. Manugula, Koteswararao V. Rajulapati, G. Madhusudhan Reddy, K. Bhanu,

Sankara Rao, Role of evolving microstructure on the mechanical properties of electron
EP

beam welded ferritic-martensitic steel in the as-welded and post weld heat- treated states,
Mater. Sci. Eng. A. 698 (2017) 36–45. doi:10.1016/j.msea.2017.05.036.
C

[73] B. Arivazhagan, R. Prabhu, S.K. Albert, M. Kamaraj, S. Sundaresan, Microstructure and

AC

mechanical properties of 9Cr-1Mo steel weld fusion zones as a function of weld metal
composition, J. Mater. Eng. Perform. 18 (2009) 999–1004. doi:10.1007/s11665-008-9349-
7.

[74] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Homogenization of P91 weldments

using varying normalizing and tempering treatment, Mater. Sci. Eng. A. 710 (2018) 86–
101. doi:10.1016/j.msea.2017.10.086.

64
 ACCEPTED MANUSCRIPT

[75] M. Sireesha, S.K. Albert, S. Sundaresan, M. Sireesha, S.K. Albert, S. Sundaresan,

Importance of filler material chemistry for optimising weld metal mechanical properties in
modified 9Cr – 1Mo steel Importance of filler material chemistry for optimising weld
metal mechanical properties in modified 9Cr – 1Mo steel, Sci. Technol. Weld. Join. 1718
(2016) 247–254. doi:10.1179/136217101101538848.

PT
[76] X. Zhou, Y. Liu, Z. Qiao, Q. Guo, C. Liu, L. Yu, H. Li, Effects of cooling rates on δ-
ferrite/γ-austenite formation and martensitic transformation in modified ferritic heat

RI
resistant steel, Fusion Eng. Des. (2017). doi:10.1016/j.fusengdes.2017.05.095.

SC
[77] M. Sireesha, S.K. Albert, S. Sundaresan, Microstructure and Mechanical Properties of
Weld Fusion Zones in Modified 9Cr-1Mo Steel, J. Mater. Eng. Perform. 10 (2001) 320–
330.

[78]
U
C. Pandey, M.M. Mahapatra, Effect of Heat Treatment on Microstructure and Hot Impact
AN
Toughness of Various Zones of P91 Welded Pipes, J. Mater. Eng. Perform. 25 (2016)
2195–2210. doi:10.1007/s11665-016-2064-x.
M

[79] B. Arivazhagan, G. Srinivasan, S.K. Albert, A.K. Bhaduri, A study on influence of heat
input variation on microstructure of reduced activation ferritic martensitic steel weld metal
D

produced by GTAW process, Fusion Eng. Des. 86 (2011) 192–197.

TE

doi:10.1016/j.fusengdes.2010.12.035.

[80] B. Arivazhagan, M. Vasudevan, A comparative study on the effect of GTAW processes

EP

on the microstructure and mechanical properties of P91 steel weld joints, J. Manuf.
Process. 16 (2014) 305–311. doi:10.1016/j.jmapro.2014.01.003.
C

[81] C. Pandey, M. Mohan, P. Kumar, N. Saini, Dissimilar joining of CSEF steels using
AC

autogenous tungsten-inert gas welding and gas tungsten arc welding and their effect on ?-
ferrite evolution and mechanical properties, J. Manuf. Process. 31 (2018) 247–259.
doi:10.1016/j.jmapro.2017.11.020.

[82] Y. Wang, R. Kannan, L. Zhang, L. Li, Microstructural Analysis of the As-Welded Heat-
Affected Zone of a Grade 91 Steel Heavy Section Weldment Microstructural Analysis of
the As - Welded Heat - Affected Zone of a Grade 91 Steel Heavy Section Weldment,

65
 ACCEPTED MANUSCRIPT

Weld. J. 96 (2017) 203–219.

[83] R.G. Faulkner, J.A. Williams, E.G. Sanchez, A.W. Marshall, In uence of Co , Cu and W
on microstructure of 9 % Cr steel weld metals, 19 (2003).
doi:10.1179/026708303225009652.

PT
[84] K.S. Chandravathi, K. Laha, K.B.S. Rao, S.L. Mannan, Microstructure and tensile
properties of modified 9Cr-1Mo steel ( grade 91 ), Mater. Sci. Technol. 17 (2001) 559–

RI
565.

[85] L. Sch afer, Influence of delta ferrite and dendritic carbides on the impact and tensile

SC
properties of a martensitic chromium steel, J. Nucl. Mater. 263 (1998) 1336–1339.

[86] P.J. Grobenr, W.C. Hagel, The Effect of Molybdenum on High- Temperature Properties of

U
9 Pct Cr Steels, Metall. Transction A. 11A (1980) 633–642.
AN
[87] K. Anderko, L. Sch, A.K. Ewaiom, Effect of the a-ferrite phase on the impact properties
chromium steels, J. Nucl. Mater. 179–181 (1991) 492–495.
M

[88] E.D. Specht, S.M. Allen, Communication Cr Steel Investigated by In-Situ X-Ray
Radiation, (2010) 2–5. doi:10.1007/s11661-010-0371-7.
D

[89] S.K. Albert, M. Matsui, T. Watanabe, H. Hongo, K. Kubo, M. Tabuchi, Variation in the
TE

type IV cracking behaviour of a high Cr steel weld with post weld heat treatment, Int. J.
Press. Vessel. Pip. 80 (2003) 405–413. doi:10.1016/S0308-0161(03)00072-3.
EP

[90] C. Pandey, M.M. Mahapatra, Effect of Groove Design and Post-Weld Heat Treatment on
Microstructure and Mechanical Properties of P91 Steel Weld, J. Mater. Eng. Perform. 25
C

(2016) 2761–2775. doi:10.1007/s11665-016-2127-z.

AC

[91] P. Mayr, H. Cerjak, Transactions of The Indian Institute of Metals The impact of welding
on the creep properties of advanced 9-12 % Cr steels, 9 (2010) 5–10.

[92] J.A. Francis, W. Mazur, H.K.D.H. Bhadeshia, Type IV cracking in ferritic power plant
steels, 22 (2006) 1387–1395. doi:10.1179/174328406X148778.

[93] K. Laha, K.S. Chandravathi, P. Parameswaran, K.B.S. Rao, S.L. Mannan,

Characterization of microstructures across the heat-affected zone of the modified 9Cr-

66
 ACCEPTED MANUSCRIPT

1Mo weld joint to understand its role in promoting type IV cracking, Metall. Mater. Trans.
A. 38 (2007) 58–68. doi:10.1007/s11661-006-9050-0.

[94] Y. Wang, R. Kannan, L. Li, Identification and Characterization of Intercritical Heat-

Affected Zone in As-Welded Grade 91 Weldment, Metall. Mater. Trans. A Phys. Metall.

PT
Mater. Sci. 47 (2016) 5680–5684. doi:10.1007/s11661-016-3736-8.

[95] K. Sawada, M. Bauer, F. Kauffmann, P. Mayr, A. Klenk, Microstructural change of 9%

RI
Cr-welded joints after long-term creep, Mater. Sci. Eng. A. 527 (2010) 1417–1426.
doi:10.1016/j.msea.2009.10.044.

SC
[96] S. Kobayashi, K. Sawada, T. Hara, H. Kushima, K. Kimura, Materials Science &
Engineering A The formation and dissolution of residual δ ferrite in ASME Grade 91 steel

U
plates, Mater. Sci. Eng. A. 592 (2014) 241–248. doi:10.1016/j.msea.2013.10.058.
AN
[97] D.J. Kotecki, Hydrogen reconsidered, Weld. J. 71 (1992) 35–43.

[98] A. American, N. Standard, Practices for Local Heating of Welds in Piping Recommended
M

Practices for, 1999 (1999).

[99] S. Paddea, J.A. Francis, A.M. Paradowska, P.J. Bouchard, I.A. Shibli, Residual stress
D

distributions in a P91 steel-pipe girth weld before and after post weld heat treatment,
TE

Mater. Sci. Eng. A. 534 (2012) 663–672. doi:10.1016/j.msea.2011.12.024.

[100] D. Dean, M. Hidekazu, Prediction of welding residual stress in multi-pass butt-welded

EP

modified 9Cr–1Mo steel pipe considering phase transformation effect, Comput. Mater.
Sci. 37 (2006) 209–219. doi:10.1016/j.commatsci.2005.06.010.
C

[101] Y.C. Jang, J.K. Hong, J.H. Park, D.W. Kim, Y. Lee, Effects of notch position of the
Charpy impact specimen on the failure behavior in heat affected zone, 1 (2007) 419–424.
AC

doi:10.1016/j.jmatprotec.2007.11.272.

[102] A. Moitra, P. Parameswaran, P.R. Sreenivasan, S.L. Mannan, A toughness study of the
weld heat-affected zone of a, Mater. Charact. 48 (2002) 55–61.

[103] N.Z. Gutiérrez, J. V. Alvarado, H. de Cicco, A. Danón, Microstructural study of welded

joints in a high temperature, Procedia Mater. Sci. 8 (2015) 1140–1149.

67
 ACCEPTED MANUSCRIPT

doi:10.1016/j.mspro.2015.04.178.

[104] V. Maduraimuthu, M. Vasudevan, P. Parameswaran, Studies on Improvement of

Toughness in Modified 9Cr-1Mo Steel A-TIG Weld Joint, Trans. Indian Inst. Met. 68
(2015) 181–189. doi:10.1007/s12666-014-0456-x.

PT
[105] B. Arivazhagan, S. Sundaresan, M. Kamaraj, A study on influence of shielding gas
composition on toughness of flux-cored arc weld of modified 9Cr-1Mo (P91) steel, J.

RI
Mater. Process. Technol. 209 (2009) 5245–5253. doi:10.1016/j.jmatprotec.2009.02.006.

[106] T. VUHERER, M. DUNĐER, L.J. MILOVIĆ, M. ZRILIĆ, I. SAMARDŽIĆ,

SC
MICROSTRUCTURAL INVESTIGATION OF THE HEAT-AFFECTED ZONE OF
SIMULATED WELDED JOINT OF P91 STEEL, Metalurgija. 52 (2013) 317–320.

U
[107] G.V.G.R. V Muthupandi, Phase Transformation Behaviour in P91 During Post Weld Heat
AN
Treatment : A Gleeble Study, Trans. Indian Inst. Met. 70 (2017) 875–885.
doi:10.1007/s12666-017-1075-0.
M

[108] L. Chen, K. Yamashita, EFFECTS OF PWHT TEMPERATURE ON MECHANICAL

PROPERTIES OF HIGH-Cr FERRITIC HEAT-RESISTANT, Weld. World. 56 (2012)
D

81–91.
TE

[109] G. Taniguchi, K. Yamashita, Effects of Post Weld Heat Treatment ( PWHT ) Temperature
on Mechanical Properties of Weld Metals for High-Cr, Kobelco Technol. Rev. 32 (2013)
33–39.
EP

[110] L. Li, A. Deceuster, B. Griffiths, J. Housekeeper, B. Silwal, J. WAlker, S. Zhang, Effect

of Post-Weld Heat Treatment on Creep Rupture Properties of Grade 91 Steel Heavy
C

Section Welds, Ph.D. Thesis. (2012).

AC

[111] M. Abd El-Rahman Abd El-Salam, I. El-Mahallawi, M.R. El-Koussy, International Heat
Treatment and Surface Engineering weld joints Part 1 – Microstructure Influence of heat
input and post-weld heat treatment on boiler steel P91 ( 9Cr – 1Mo – V – Nb ) weld joints,
Int. Heat Treat. Surf. Eng. 7 (2013) 23–31. doi:10.1179/1749514813Z.00000000050.

[112] M. Abd El-Rahman Abd El-Salam, I. El-Mahallawi, M.R. El-Koussy, Influence of heat
input and post-weld heat treatment on boiler steel P91 ( 9Cr – 1Mo – V – Nb ) weld joints

68
 ACCEPTED MANUSCRIPT

Part 2 – Mechanical properties, Int. Heat Treat. Surf. Eng. 7 (2013) 32–37.

[113] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, A. Srivastava, Microstructure and

mechanical property relationship for different heat treatment and hydrogen level in multi-
pass welded P91 steel joint, J. Manuf. Process. 28 (2017) 220–234.

PT
doi:10.1016/j.jmapro.2017.06.009.

[114] C. Pandey, M.M. Mahapatra, P. Kumar, A. Giri, Microstructure characterization and

RI
Charpy toughness of P91 weldment for as-welded, PWHT and N&T heat treatment, Met.
Mater. Int. (2017). doi:10.1007/s12540-017-6850-2.

SC
[115] K.A. Venkata, S. Kumar, H.C. Dey, D.J. Smith, P.J. Bouchard, Study on the Effect of Post
Weld Heat Treatment Parameters on the Relaxation of Welding Residual Stresses in

U
Electron Beam Welded P91 Steel Plates, Procedia Eng. 86 (2014) 223–233.
doi:10.1016/j.proeng.2014.11.032.
AN
[116] A.H. Yaghi, T.H. Hyde, A.A. Becker, W. Sun, Finite element simulation of welded P91
steel pipe undergoing post-weld heat treatment, Sci. Technol. Weld. Join. 16 (2011) 232–
M

239. doi:10.1179/1362171810Y.0000000018.
D

[117] M. Zubairuddin, S.K. Albert, M. Vasudevan, S. Mahadevan, V. Chaudhri, V.K. Suri,

Materials and Manufacturing Processes Thermo-Mechanical Analysis of Preheat Effect on
TE

Grade P91 Steel during GTA Welding, Mater. Manuf. Process. 31 (2016) 366–371.
doi:10.1080/10426914.2015.1025964.
EP

[118] S.S. Glickstein, Temperature Measurements in a Free Burning Arc, Weld. J. 55 (1976)
222–229.
C

[119] R.K. Dayal, N. Parvathavarthini, Hydrogen embrittlement in power plant steels, Sadhana.
AC

28 (2003) 431–451. doi:10.1007/BF02706442.

[120] I. Maroef, D.L. Olson, M. Eberhart, G.R. Edwards, Hydrogen trapping in ferritic steel
weld metal, Int. Mater. Rev. 47 (2002) 191–223.

[121] D.E. Bunnell, S. LIu, D.L. Olson, Low Hydrogen S M A W Electrode Moisture Content
Determination Using Electrical Capacitance, J. Mater. Eng. 12 (1990) 159–165.

69
 ACCEPTED MANUSCRIPT

[122] H. Li, T.H. North, Hydrogen absorption and hydrogen cracking in high strength weld
metal, Key Eng. Mater. 69 (1992) 95–112. doi:10.4028/www.scientific.net/KEM.69-
70.95.

[123] J.E. Ramirez, M. Johnson, Effect of Welding Parameters and Electrode Condition on

PT
Alloying Enrichment of Weld Metal Deposited with Coated Cellulosic Electrodes, Weld.
J. 89 (2010) 232–242.

RI
[124] R.S. Funderburk, Key Concepts in Welding Engineering, Weld. Innov. 15 (1998) 1–2.

[125] ANSI/AWS A4.3: 1993 (R2006), Standard methods for determination of the diffusible

SC
hydrogen content of martensitic , bainitic , and ferritic steel weld, (n.d.).

[126] AS/NZS 3752:2006, Welding and allied processes — Determination of hydrogen content

U
in ferritic steel arc weld metal, (2006).
AN
[127] BIS:IS 11802: 1986 (R2003), METHODS FOR DETERMINATION OF DIFFUSIBLE
HYDROGEN CONTENT OF DEPOSITED WELD METAL FROM COVERED
M

ELECTRODES IN WELDING MILD AND LOW ALLOY STEELS, (1986).

[128] P.G. Kumar, K. Yu-ichi, Diffusible Hydrogen In Steel Weldments, Trans. JWRI. 42
D

(2013) 39–62.
TE

http://www.researchgate.net/publication/256499083_Diffusible_hydrogen_in_steel_welds
_-_A_status_review/file/3deec52d8dbf163485.pdf.
EP

[129] D. Fydrych, J. Łabanowski, An experimental study of high-hydrogen welding processes,

Rev. Metal. 51 (2015) 1–6. doi:10.3989/revmetalm.055.
C

[130] D. Fydrych, J. Łabanowski, Determining diffusible hydrogen amounts using the mercury
method, Weld. Int. 26 (2017) 697–702. doi:10.1080/09507116.2011.592682.
AC

[131] G. Magudeeswaran, V. Balasubramanian, G. Madhusudhan Reddy, Hydrogen induced

cold cracking studies on armour grade high strength, quenched and tempered steel
weldments, Int. J. Hydrogen Energy. 33 (2008) 1897–1908.
doi:10.1016/j.ijhydene.2008.01.035.

[132] N. Saini, C. Pandey, M. Mohan, ScienceDirect Effect of diffusible hydrogen content on

70
 ACCEPTED MANUSCRIPT

embrittlement of P92 steel, Int. J. Hydrogen Energy. 42 (2017) 17328–17338.

doi:10.1016/j.ijhydene.2017.05.214.

[133] G.K. Padhy, V. Ramasubbu, S.K. Albert, A. Murugesan, G. Ramesh, Hot extraction of
Diffusible Hydrogen and its Measurement Using a Hydrogen Senor, Weld. World. 56

PT
(2012) 18–25.

[134] G.K. Padhy, V. Ramasubbu, S.K. Albert, Rapid determination of diffusible hydrogen in

RI
steel welds using a modified gas chromatography facility, J. Test. Eval. 43 (2015).
doi:10.1520/JTE20130077.

SC
[135] D. Nolan, M. Pitrun, DIFFUSIBLE HYDROGEN TESTING IN AUSTRALIA
DIFFUSIBLE HYDROGEN TESTING IN AUSTRALIA, Diffus. Hydrog. Test. Austrilia.

U
48 (2004) 14–20.
AN
[136] S.K. Albert, C. Remash, N. Murugesan, T.P.S. Gill, G. Periaswami, S.D. Kulkarni, A New
Method to Measure the Diffusible Hydrogen Content in Steel Weldments Using a Polymer
Electrolyte-Based Hydrogen Sensor, Weld. Res. Suppl. (1997) 251–255.
M

[137] G.K. Padhy, V. Ramasubbu, N. Murugesan, C. Ramesh, N. Parvathavarthini, S.K. Albert,

D

Determination of apparent diffusivity of hydrogen in 9Cr-1MoVNbN steel using hot

extraction-PEMHS technique, Int. J. Hydrogen Energy. 38 (2013) 10683–10693.
TE

doi:10.1016/j.ijhydene.2013.06.077.

[138] S.K. Albert, G.K. Padhy, A brief review on methods for diffusible hydrogen measurement
EP

in welds, Indian Weld. J. 47 (2014) 101–113.

[139] L.C. De Abreu, P.J. Modenesi, P.V. Marques, Comparative study of methods for
C

determining the diffusible hydrogen content in welds, Weld. Int. 9 (1995) 26–31.
AC

[140] T. Kannengiesser, N. Tiersch, COMPARATIVE STUDY BETWEEN HOT

EXTRACTION METHODS AND MERCURY METHOD – A NATIONAL ROUND
ROBIN TEST, Weld. World. 54 (2010) 108–114.

[141] G.K. Padhy, V. Ramasubbu, N. Murugesan, C. Remash, S.K. Albert, G.K. Padhy, V.
Ramasubbu, N. Murugesan, C. Remash, S.K. Albert, Effect of preheat and post-heating on
diffusible hydrogen content of welds, Sci. Technol. Weld. Join. ISSN. 17 (2017) 408–413.

71
 ACCEPTED MANUSCRIPT

doi:10.1179/1362171812Y.0000000023.

[142] D. Fydrych, G. Rogalski, Effect of underwater local cavity welding method conditions on
diffusible hydrogen content in deposited metal, Weld. Int. ISSN. 27 (2013) 196–202.
doi:10.1080/09507116.2011.600033.

PT
[143] A. Swierczynska, D. Fydrych, J. Łabanowski, The Effect of Welding Conditions on
Diffusible Hydrogen Content in Deposited Metal, Solid State Phenom. (2011) 1–8.

RI
doi:10.4028/www.scientific.net/SSP.183.193.

[144] D. Fydrych, G. Rogalski, Effect of shielded-electrode wet welding conditions on diffusion

SC
hydrogen content in deposited metal, Weld. Int. 25 (2011) 166–171.
doi:10.1080/09507116.2010.540828.

U
[145] D. Fydrych, G. Rogalski, Diffusible hydrogen content in the deposited metal of multilayer
AN
welded joints Zawartość wodoru dyfundującego, Metall. Foundry Eng. 40 (2014) 221–
229.
M

[146] C. Pandey, M.M. Mahapatra, P. Kumar, N. Saini, Diffusible Hydrogen Level in Deposited
Metal and Their Effect on Tensile Properties and Flexural Strength of P91 steel, J. Eng.
D

Mater. Technol. 139 (2017) 1–11. doi:10.1115/1.4035764.

TE

[147] R.D. Stout, S.S. Toss, L.J. McGeady, J.E. Doan, Quantitative Measure- ment of the
Cracking Tendency in Welds, Weld. J. Res. Suppl. 25 (1946) 522–531.
EP

[148] D. Fydrych, J. Łabanowski, G. Rogalski, Weldability of high strength steels in wet

welding conditions, Polish Marit. Res. 20 (2013) 67–73. doi:10.2478/pomr-2013-0018.
C

[149] J.M. Sawhill, A.W. Dix, W.F. Savage, Modified implant test for studying delayed
cracking, Weld. Res. Suppl. (1974) 554–560.
AC

[150] H. Granjon, The implants method for studying the weldability of high strength steel, Met.
Constr. Br. Weld. J. 11 (1969) 509–515. doi:10.1515/adms-2016-0018.

[151] S.K. Albert, V. Ramasubbu, S.I.S. Raj, A.K. Bhaduri, HYDROGEN-ASSISTED

CRACKING SUSCEPTIBILITY of modifi ed 9Cr-1Mo steel and its weld metal, Weld.
World. 55 (2011) 66–74.

72
 ACCEPTED MANUSCRIPT

[152] S. Manimozhi, S. Suresh, V. Muthupandi, HAZ hydrogen cracking of 9Cr-0.5 Mo-1.7W

steels, Int. J. Adv. Manuf. Technol. 51 (2010) 217–223. doi:10.1007/s00170-010-2611-9.

[153] X. Yue, Investigation on heat-affected zone hydrogen-induced cracking of high-strength

naval steels using the Granjon implant test: HENRY GRANJON PRIZE 2014 Winner

PT
Category B: Materials Behaviour and Weldability, Weld. World. 59 (2014) 77–89.
doi:10.1007/s40194-014-0181-4.

RI
[154] C. Pandey, N. Saini, M.M. Mahapatra, P. Kumar, Hydrogen induced cold cracking of
creep resistant ferritic P91 steel for different diffusible hydrogen levels in deposited metal,

SC
Int. J. Hydrogen Energy. 41 (2016) 17695–17712. doi:10.1016/j.ijhydene.2016.07.202.

[155] X. Yue, J.C. Lippold, Evaluation of Heat-Affected Zone Hydrogen-Induced Cracking in

U
Navy Steels, Weld. J. 92 (2013) 20–28.
AN
[156] M.D. Rowe, T.W. Nelson, J.C. Lippold, Hydrogen-Induced Cracking along the Fusion
Boundary of Dissimilar Metal Welds, Weld. Res. Suppl. 78 (1999) 31–37.
M

[157] B. Alexandrov, K. Theis, M. Streitenberger, H. Herold, I. Martinek, Cold cracking in

weldments of steel S 690 QT, Weld. World. 49 (2005) 64–73.
D

[158] N.J. Petch, P. Stables, © 1952 Nature Publishing Group, Nature. 169 (1952) 842–843.
TE

[159] C.D. Beachem, A New Model for Hydrogen-Assisted Cracking ( Hydrogen “

Embrittlement ”), Metall. Trans. 3 (1972) 437–451.
EP

[160] A.R. Troiano, The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of
Metals, Metallogr. Microstruct. Anal. 5 (2016) 557–569. doi:10.1007/s13632-016-0319-4.
C

[161] M.M. Mahapatra, G.L. Datta, B. Pradhan, N.R. Mandal, Modelling the effects of
AC

constraints and single axis welding process parameters on angular distortions in one-sided
fillet welds, Proc. IMechE Part B J. Eng. Manuf. 221 (2006) 397–407.
doi:10.1243/09544054JEM617.

[162] R. Cottam, V. Luzin, K. Thorogood, Y.C. Wong, M. Brandt, The role of metallurgical
solid state phase transformations on the formation of residual stress in laser cladding and
heating, Mater. Sci. Forum. 777 (2014) 19–24.

73
 ACCEPTED MANUSCRIPT

doi:10.4028/www.scientific.net/MSF.777.19.

[163] A. De, T. Debroy, A perspective on residual stresses in welding, Sci. Technol. Weld. Join.
16 (2017) 204–208. doi:10.1179/136217111X12978476537783.

[164] P.K. Ghosh, R.R. Kumar, K. Devakumaran, Analytical studies on shrinkage stress

PT
distribution in gma and pulse current gma welds of thick wall stainless steel pipe having
narrow and v-groove design, Indian Weld. J. (2010) 14–23.

RI
[165] C. Basavaraju, Simplified analysis of shrinkage in pipe to pipe butt welds, Nucl. Eng. Des.
197 (2000) 239–247.

SC
[166] C. Pandey, A. Giri, M.M. Mahapatra, On the prediction of effect of direction of welding
on bead geometry and residual deformation of double-sided fillet welds, Int. J. Steel

U
Struct. 16 (2016) 333–345. doi:10.1007/s13296-016-6007-z.
AN
[167] G.A. Webster, A.N. Ezeilo, Residual stress distributions and their influence on fatigue
lifetimes, Int. J. Fatigue. 23 (2001) 375–383.
M

[168] P. Dong, P. Dong, Residual stresses and distortions in welded structures : a perspective for
engineering applications Residual stresses and distortions in welded structures : a
D

perspective for engineering applications, Sci. Technol. Weld. Join. 10 (2004) 389–398.
TE

doi:10.1179/174329305X29465.

[169] A.H. Yaghi, T.H. Hyde, A.A. Becker, W. Sun, Finite element simulation of welding and
EP

residual stresses in a P91 steel pipe incorporating solid-state phase transformation and
post-weld heat treatment, IMechE J. Strain Anal. 43 (2008) 275–293.
doi:10.1243/03093247JSA372.
C

[170] H. Murakawa, B. Miloslav, V. Adan, R. Sherif, D. Cartin M., D. David, N. Kamran M.,
AC

Effect of Phase Transformation onset Temperature on Residual Stress in Welded Thin

Steel Plates, Trans. JWRI. 37 (2008) 75–80.

[171] D. Deng, H. Murakawa, W. Liang, Numerical and experimental investigations on welding

residual stress in multi-pass butt-welded austenitic stainless steel pipe, Comput. Mater.
Sci. 42 (2008) 234–244. doi:10.1016/j.commatsci.2007.07.009.

74
 ACCEPTED MANUSCRIPT

[172] C. Pandey, H.K. Narang, N. Saini, M.M. Mahapatra, P. Kumar, Microstructure and
Transverse Shrinkage Stress Analysis in GTA Welds of P91 Steel Pipe, Int. J. Steel Struct.
17 (2017) 763–774. doi:10.1007/s13296-017-6030-8.

[173] P. Dong, Residual Stress Analyses of a Multi-Pass Girth Weld : 3-D Special Shell Versus

PT
Axisymmetric Models, J. Press. Vessel Technol. 123 (2001) 207–213.
doi:10.1115/1.1359527.

RI
[174] P.K. GHOSH, R.R. Kumar, A.K. Pramanick, Analytical studies on shrinkage stress
distribution in GMA and pulse current GMA welds of thick wall stainless steel pipe

SC
having narrow and V-groove design, Indian Weld. JournalWelding J. (2010) 14–24.

[175] R. Anant, P.K. Ghosh, Experimental Investigation on Transverse Shrinkage Stress and

U
Distortion of Extra Narrow and Conventional Gap Dissimilar Butt Joint of Austenitic
Stainless Steel to Low Alloy Steel, Proc. Int. Conf. Mining, Mater. Metall. Eng. (2014) 1–
AN
5.

[176] S. Arunkumar, P. Rangarajan, K. Devakumaran, P. Sathiya, ScienceDirect Comparative

M

study on transverse shrinkage , mechanical and metallurgical properties of AA2219

aluminium weld joints prepared by gas tungsten arc and gas metal arc welding processes,
D

Def. Technol. 11 (2015) 262–268. doi:10.1016/j.dt.2015.05.008.

TE

[177] A.H. Yaghi, T.H. Hyde, A.A. Becker, W. Sun, G. Hilson, S. Simandjuntak, P.E.J. Flewitt,
D.J. Smith, A Comparison Between Measured and Modeled Residual Stresses in a
EP

Circumferentially Butt-Welded P91 Steel Pipe, J. Press. Vessel Technol. 132 (2010) 1–10.
doi:10.1115/1.4000347.
C

[178] T.C. Chuvas, P.S.P. Garcia, J.M. Pardal, M. da P.C. Fonseca, Influence of Heat Treatment
AC

in Residual Stresses Generated in P91 Steel-pipe Weld, Mater. Res. 18 (2015) 614–621.
doi:10.1590/1516-1439.006315.

[179] S. Kulkarni, P.K. Ghosh, S. Ray, Improvement of Weld Characteristics by Variation in

Welding Processes and Parameters in Joining of Thick Wall 304LN Stainless Steel Pipe,
ISIJ Int. 48 (2008) 1560–1569.

[180] S. Kim, J. Kim, W. Lee, Numerical prediction and neutron diffraction measurement of the

75
 ACCEPTED MANUSCRIPT

residual stresses for a modified 9Cr – 1Mo steel weld, J. Mater. Process. Technol. 209
(2009) 3905–3913. doi:10.1016/j.jmatprotec.2008.09.012.

[181] Y. Sattari-Far, IJavadi, Influence of welding sequence on welding distortions in pipes, Int.
J. Press. Vessel. Pip. 85 (2008) 265–274. doi:10.1016/j.ijpvp.2007.07.003.

PT
[182] Y. Zhao, J. Gong, Y. Wang, G. Wei, Effect of Start / stop Position Distribution on
Residual Stresses in the Multi-pass Welded 12Cr1MoV / P91 Dissimilar pipe, Int. J. Steel

RI
Struct. 14 (2014) 539–546. doi:10.1007/s13296-014-3010-0.

[183] Measurement of residual stresses by the hole drilling strain gauage method -Tech note

SC
TN-503, Vishay Precis. Gr. (2010) 19–33.

[184] N. Hempel, T. Nitschke-pagel, K. Dilger, Residual stresses in multi-pass butt-welded

U
ferritic-pearlitic steel pipes, Weld World. 59 (2015) 555–563. doi:10.1007/s40194-015-
AN
0230-7.

[185] C.G. Panait, W. Bendick, A. Fuchsmann, A.F. Gourgues-Lorenzon, J. Besson, Study of

M

the microstructure of the Grade 91 steel after more than 100,000 h of creep exposure at
600 °C, Int. J. Press. Vessel. Pip. 87 (2010) 326–335. doi:10.1016/j.ijpvp.2010.03.017.
D

[186] A. Aghajani, C. Somsen, G. Eggeler, On the effect of long-term creep on the

TE

microstructure of a 12% chromium tempered martensite ferritic steel, Acta Mater. 57

(2009) 5093–5106. doi:10.1016/j.actamat.2009.07.010.
EP

[187] M. Tamura, H. Sakasegawa, A. Kohyama, H. Esaka, K. Shinozuka, Effect of MX type

particles on creep strength of ferritic steel, J. Nucl. Mater. 321. 321 (2003) 288–293.
doi:10.1016/S0022-3115(03)00281-2.
C

[188] V. Vodarek, A. Strang, Minor-phase composition changes in 12CrMoVNb steels during

AC

long-term exposure, Mater. Chem. Phys. 81 (2003) 480–482. doi:10.1016/S0254-

0584(03)00061-0.

[189] K. Sawada, K. Kubo, F. Abe, Contribution of coarsening of MX carbonitrides to creep

strength degradation in high chromium ferritic steel, Mater. Sci. Technol. 19 (2003) 732–
738. doi:10.1179/026708303225010687.

76
 ACCEPTED MANUSCRIPT

[190] M. Taneike, M. Kondo, T. Morimoto, Accelerated Coarsening of during Creep

Deformation, ISIJ Int. 41 (2001) 111–115.

[191] P. Hu, W. Yan, W. Sha, W. Wang, Y. Shan, K. Yang, Microstructure Evolution of a 10Cr
Heat-Resistant Steel during High Temperature Creep, J. Mater. Sci. Technol. 27 (2011)

PT
344–351. doi:10.1016/S1005-0302(11)60072-8.

[192] K. Sawada, M. Taneike, K. Kimura, F. Abe, In situ observation of recovery of lath

RI
structure in 9 % chromium creep resistant steel, Mater. Sci. Technol. 19 (2003) 739–742.
doi:10.1179/026708303225010696.

SC
[193] K. Swada, K. Maruyama, Y. Hasegawa, T. Muraki, No Title, Key Eng. Mater. 174 (2000)
109–114. doi:10.4028/www.scientific.net/KEM.171-174.109.

U
[194] K. Tak, U. Schulz, G. Eggeler, On the effect of micrograin crystallography on creep of
AN
FeCr alloys, Mater. Sci. Eng. A. 511 (2009) 121–129. doi:10.1016/j.msea.2008.11.070.

[195] W. Yan, W. Wang, Y.Y. Shan, K. Yang, Microstructural stability of 9-12%Cr

M

ferrite/martensite heat-resistant steels, Front. Mater. Sci. 7 (2013) 1–27.

doi:10.1007/s11706-013-0189-5.
D

[196] C. Keller, M.M. Margulies, Z. Hadjem-hamouche, I. Guillot, Influence of the temperature

TE

on the tensile behaviour of a modified, Mater. Sci. Eng. A. 527 (2010) 6758–6764.
doi:10.1016/j.msea.2010.07.021.
EP

[197] A. Aghajani, F. Richter, C. Somsen, S.G. Fries, I. Steinbach, G. Eggeler, On the formation
and growth of Mo-rich Laves phase particles during long-term creep of a 12% chromium
tempered martensite ferritic steel, Scr. Mater. 61 (2009) 1068–1071.
C

doi:10.1016/j.scriptamat.2009.08.031.
AC

[198] R.L. Klueh, D.J. Alexander, E.A. Kenik, Development of low-chromium , chromium-
tungsten steels for fusion 1, J. Nucl. Mater. 227 (1995) 11–23.

[199] J. HAld, Metallurgy and creep properties of new 9-12 % Cr steels, Steel Res. 67 (1996)
369–374.

[200] H. Okamura, R. Ohtani, K. Saito, K. Kimura, R. Ishii, K. Fujiyama, S. Hongo, T. Iseki, H.

77
 ACCEPTED MANUSCRIPT

Uchida, Basic investigation for life assessment technology of modified 9Cr – 1Mo steel,
Nucl. Eng. Des. 193. 193 (1999) 243–254.

[201] K. Sawada, H. Kushima, M. Tabuchi, K. Kimura, Microstructural degradation of Gr.91

steel during creep under low stress, Mater. Sci. Eng. A. 528 (2011) 5511–5518.

PT
doi:10.1016/j.msea.2011.03.073.

[202] R.B. Brucker, W.M. Elger, M.J. Sorek, M.S. Group, M. Engineer-, R.B. Brucker, W.M.

RI
Elger, Microstructure-Thermal History Correlations for HY-130 Thick Section
Weldments, Weld. J. 63 (1984) 254–262.

SC
[203] K. Laha, K.S. Chandravathi, P. Parameswaran, K.B. Sankara, S.L. Mannan, T. Minerals,
M. Society, Characterization of Microstructures across the Heat-Affected Zone of the

U
Modified 9Cr-1Mo Weld Joint to Understand Its Role in Promoting Type IV Cracking,
Metall. Mater. Trans. A. 38 (2007) 58–68. doi:10.1007/s11661-006-9050-0.
AN
[204] Y. Tsuchida, K. Okamoto, Y. Tokunaga, Study of creep rupture strength in heat affected
zone of 9Cr ‐ 1Mo ‐ V ‐ Nb ‐ N steel by welding thermal cycle simulation, Weld. Int. 10
M

(1996) 454–460. doi:10.1080/09507119609549030.

D

[205] D.J. Abson, J.S. Rothwell, Review of type IV cracking of weldments in 9 – 12 % Cr creep
strength enhanced ferritic steels Review of type IV cracking of weldments in 9 – 12 % Cr
TE

creep strength enhanced ferritic steels, Int. Mater. Rev. 58 (2013) 437–473.
doi:10.1179/1743280412Y.0000000016.
EP

[206] M.E.A. El-azim, A.M. Nasreldin, G. Zies, A. Klenk, Microstructural instability of a

welded joint in P91 steel during creep at 600 u C, Mater. Sci. Technol. 21 (2005) 779–
C

791. doi:10.1179/174328405X43216.
AC

[207] S.K. Albert, M. Matsui, T. Watanabe, H. Hongo, K. Kubo, M. Tabuchi, Microstructural

Investigations on Type IV Cracking in a High Cr Steel, ISIJ Int. 42 (2002) 1497–1504.

[208] T. Watanabe, M. Tabuchi, M. Yamazaki, H. Hongo, T. Tanabe, Creep damage evaluation

of 9Cr-1Mo-V-Nb steel welded joints showing Type IV fracture, Int. J. Press. Vessel. Pip.
83 (2006) 63–71. doi:10.1016/j.jelectrocard.2005.07.009.

[209] S.K. Albert, M. Tabuchi, H. Hongo, T. Watanabe, K. Kubo, M. Matsui, M. Tabuchi, H.

78
 ACCEPTED MANUSCRIPT

Hongo, T. Watanabe, K. Kubo, M. Matsui, S.K. Albert, M. Tabuchi, H. Hongo, T.

Watanabe, K. Kubo, M. Matsui, Effect of welding process and groove angle on type IV
cracking behaviour of weld joints of a ferritic steel Effect of welding process and groove
angle on type IV cracking behaviour of weld joints of a ferritic steel, Sci. Technol. Weld.
Join. 10 (2005) 149–157. doi:10.1179/174329305X36034.

PT
[210] F. Abe, M. Tabuchi, F. Abe, M. Tabuchi, Microstructure and creep strength of welds in
advanced ferritic power plant steels Microstructure and creep strength of welds in

RI
advanced ferritic power plant steels, Sci. Technol. Weld. Join. 9 (2004) 22–30.
doi:10.1179/136217104225017107.

SC
[211] J.A. Francis, W. Mazur, H.K.D.H. Bhadeshia, Estimation of Type IV Cracking Tendency
in Power Plant Steels, ISIJ Int. 44 (2004) 1966–1968.

U
[212] K. Shinozaki, D. Li, H. Kuroki, H. Harada, K. Ohishi, T. Sato, Observation of type IV
AN
cracking in welded joints of high chromium ferritic heat resistant steels, Sci. Technol.
Weld. Join. 8 (2003) 289–295. doi:10.1179/136217103225005444.
M

[213] M.E. Abd El-Azim, O.E. El-Desoky, H. Ruoff, F. Kauffmann, E. Roos, Creep fracture
mechanism in welded joints of P91 steel, Mater. Sci. Technol. 29 (2013) 1027–1033.
D

doi:10.1179/1743284713Y.0000000233.
TE

[214] T. Sato, K. Tamura, K. Mitsuhata, R. Ikura, “Improvement of creep rupture strength of

9Cr1MoNbV welded joints by post weld normalising and tempering”, Proc. 5th Int. EPRI
EP

Conf. on “Advances in material technology for fossil power plants”, Marco Island, FL,
USA, October 2007, EPRI, 874., (n.d.).
C

[215] M. Dewitte, C. Coussement, Creep properties of 12% Cr and improved 9% Cr weldments,

AC

Mater. High Temp. 9 (1991) 178–184. doi:10.1080/09603409.1991.11689658.

[216] J.A. Francis, G.M.D. Cantin, W. Mazur, H.K.D.H. Bhadeshia, G.M.D. Cantin, W. Mazur,
H.K.D.H. Bhadeshia, J.A. Francis, G.M.D. Cantin, W. Mazur, H.K.D.H. Bhadeshia,
Effects of weld preheat temperature and heat input on type IV failure, Sci. Technol. Weld.
Join. 14 (2009) 436–442. doi:10.1179/136217109X415884.

[217] M. Kondo, M. Tabuchi, S. Tsukamoto, F. Yin, F. Abe, M. Kondo, M. Tabuchi, S.

79
 ACCEPTED MANUSCRIPT

Tsukamoto, F. Yin, F. Abe, Suppressing type IV failure via modification of heat affected
zone microstructures using high boron content in 9Cr heat resistant steel welded joints,
Sci. Technol. Weld. Join. ISSN. 11 (2006) 216–223. doi:10.1179/174329306X89260.

PT
RI
U SC
AN
M
D
TE
C EP
AC

80
 ACCEPTED MANUSCRIPT

Figure caption(s)

Fig. 1 Schematic evolution of MX and M23C6 precipitates during normalizing and tempering

Fig. 2 (a) Secondary electron micrograph of P91 steel in normalized condition, (a) at lower
magnification, (b) at higher magnification, (c) typical optical micrograph in normalized

PT
condition [33]

Fig. 3 SEM micrograph of P91 steel in normalized and tempered condition at different

RI
magnification indicating carbide precipitates and grain boundaries: (a) at 10000x, (b) at 50000x;
(c) EDS spectra of white particles presents at prior austenite grain boundary (EDS-spot 1) and,

SC
(d) EDS spectra of white particles present within matrix (EDS-spot 2), (e) typical optical
micrograph in normalized and tempered condition. Black dotted lines: equiaxed prior austenite

U
grain boundaries, Circle: Carbides (M23C6)
AN
Fig. 4 Schematic evolution of M23C6 and MX precipitates

Fig. 5 Possible mechanism of Z-phase nucleation [48]

M

Fig. 6 Schematic evolution of sub-zone in P91 weldments

Fig. 7 Schematic of microstructure evolution in weld fusion zone and FGHAZ in as-welded and
D

PWHT condition
TE

Fig. 8 (a) Weld fusion zone in as-welded condition at lower and higher magnification (a), (b),
respectively, weld fusion zone in PWHT condition at lower and higher magnification (c) and (d),
EP

respectively; weld fusion zone for multi-pass GTAW (e) as-welded, (f) PWHT

Fig. 9 Variation in hardness of soft δ-ferrite zone and untempered lath martensite
C

Fig. 10 Schematic of the sub-zones of the HAZ and relation to the calculated equilibrium phase
AC

diagram of P91 weldments

Fig. 11 Schaeffler diagram
Fig. 12 Schematic microstructure (a) for base metal, (b) for IC-HAZ
Fig. 13 Schematic of microstructure evolution in sub-zone of P91 weldments during N&T
treatment

1
 ACCEPTED MANUSCRIPT

Fig. 14 SEM micrograph of sub-zone of P91 pipe weldment in PWHT and N&T condition, (a)
and (b) weld fusion zone, (c) and (d) CGHAZ, (e) and (f) FGHAZ, (g) and (h) IC-HAZ [110]

Fig. 15 Schematic evolution of essential condition that cause HIC

Fig. 16 Schematic representation of diffusion of hydrogen to heat affected zone (HAZ) through

PT
weld fusion zone [121]
Fig. 17 Diffusible hydrogen measurement set up

RI
Fig. 18 Implant test assembly
Fig. 19 Schematic of crack propagation in (a) weld zone, (b) in HAZ

SC
Fig. 20 Schematic diagram of volume change during heating and cooling cycle of P91
weldments

U
Fig. 21(a) Variation of calibration coefficient ͞A and B, (b) Percentage error vs. applied stress
ratio
AN
Fig. 22 Percentage error; (a) longitudinal & (b) transverse direction for 2 mm hole

Fig. 23 Typical Schematic diagram of creep curve

M

Fig. 24 Schematic representation of diffusional creep processes, showing the stress-controlled

flow of vacancies from regions of tension to compression
D

Fig. 25 Schematic evolution of martensite into subgrains (GB-grain boundary; LB-lath

TE

boundary)

Fig. 26 Schematic evolution of cracks in P91 weldments

C EP
AC

2
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

Table caption(s)

Table 1 Chemical composition of Grade 91 steel (Wt- %, Fe-Bal.)

Materials C Cr Mo Mn Si Ni Al V Nb S P N
AWS A.5.28 0.07- 8.0- 0.80- 1.25a 0.15- 1.0a 0.04 0.15- 0.02- 0.01 0.01 0.03-

PT
ER90S-B9 0.13 9.5 1.10 0.30 0.25 0.10 0.07
ASTM 182 P91 0.08- 8.0- 0.85- 0.30- 0.20- 0.40 0.02 0.18- 0.06- 0.01 0.01 0.03-

RI
0.12 9.5 1.05 0.60 0.50 0.25 0.10 0.07
Grade 91 metal 0.086 8.890 0.87 0.39 0.33 0.30 0.003 0.0220 0.096 0.001 0.02 0.034

SC
(Measured) [14]
a
Mn+Ni=1.5% maximum

U
Table 2 Typical temperature for P91 steel
AN
Phase transformation Temperature
Lower critical temperature (Ac1) 810-825 oC
912-930 oC
M

Upper critical temperature (Ac3)

Martensitic formation start temperature 390-400 oC
D

(Ms)
Martensitic formation finish temperature 200-230 oC.
TE

(Mf)
EP

Table 3 Composition of filler metals (wt- %, Fe-Bal.)

Materials C Cr Mo Mn Si Ni Al V Nb S P N
C

GTAW filler 0.097 8.83 0.928 0.560 0.250 0.307 0.002 0.197 0.064 0.004 0.06 0.030
AC

(ER90S-B9)
FCAW filler 0.10 8.83 0.88 0.79 0.28 0.55 0.001 0.20 0.03 0.008 0.02 0.05
(E91T1-B9)
MMA filler 0.08- 8.0- 0.85- 0.40- 0.20- 0.4- 0.002 0.15- 0.04- 0.004 - 0.03-
(E9015-B9) 0.13 10.0 1.10 1.0 0.50 1.0 0.30 0.08 0.07
SAW filler 0.1 8.70 1.0 0.50 0.25 0.60 0.001 0.20 0.05 0.006 0.005 0.05

1
 ACCEPTED MANUSCRIPT

(EB9)

Table 4 HAZs, temperature experienced and their corresponding transformation reaction

Sub-zones Temperature Transformation Microstructure Mechanical

PT
experienced reactions characteristics properties
CGHAZ Tm˃Tp˃Tϒδ γ +δ → M +δ Untempered Poor toughness and

RI
martensite, high strength
undissolved NbX,

SC
dissolved M23C6
Tϒδ˃Tp˃Ac3 Coarse ..γ → M Untempered Poor toughness and
martensite, high strength

U undissolved NbX,
AN
dissolved M23C6
FGHAZ Tϒδ˃Tp˃Ac3 Fine ..γ → M Untempered Adequate toughness
M

martensite, and strength

dissolved fine
D

M23C6, undissolved
MX and coarse
TE

M23C6

γ → M + OM
EP

IC-HAZ Ac3 ˃TP ˃Ac1 Untempered High toughness and

martensite, tempered poor strength
martensite,
C

dissolved very fine

AC

M23C6, undissolved
MX and coarse
M23C6

2
 ACCEPTED MANUSCRIPT

Over-tempered base TP < Ac1 OM Tempered High toughness and

zone martensite, poor strength
Undissolved MX
and coarse M23C6

PT
where, Tp=temperature experienced during welding; Tm=melting point of steel; Tϒδ=temperature

RI
at which ϒ to δ transformation completed during heating; Ac3=upper critical temperature at
which α to ϒ transformation completed on heating; Ac1=lower critical temperature at which α to

SC
ϒ transformation started on heating; ϒ=austenite; δ=ferrite; M= martensite; OM=overtempered
martensite

U
AN
M
D
TE
C EP
AC