Materials Characterization 96 (2014) 213224

Contents lists available at ScienceDirect

Materials Characterization
journal homepage: www.elsevier.com/locate/matchar

Microstructural characterization of weld joints of 9Cr reduced activation

ferritic martensitic steel fabricated by different joining methods
V. Thomas Paul , S. Saroja, S.K. Albert, T. Jayakumar, E. Rajendra Kumar
Materials & Metallurgy Group, Indira Gandhi Centre for Atomic Research, Kalpakkam, Tamil Nadu 603 102, India

a r t i c l e

i n f o

Article history:
Received 2 June 2014
Received in revised form 7 August 2014
Accepted 14 August 2014
Available online 15 August 2014
Keywords:
RAFM steel weld
Transmission Electron Microscopy
INRAFM steel
PWHT of RAFM steel
Microstructural characterization

a b s t r a c t
This paper presents a detailed electron microscopy study on the microstructure of various regions of weldment
fabricated by three welding methods namely tungsten inert gas welding, electron beam welding and laser beam
welding in an indigenously developed 9Cr reduced activation ferritic/martensitic steel. Electron back scatter diffraction studies showed a random micro-texture in all the three welds. Microstructural changes during thermal
exposures were studied and corroborated with hardness and optimized conditions for the post weld heat
treatment have been identied for this steel. HollomonJaffe parameter has been used to estimate the extent
of tempering. The activation energy for the tempering process has been evaluated and found to be corresponding
to interstitial diffusion of carbon in ferrite matrix. The type and microchemistry of secondary phases in different
regions of the weldment have been identied by analytical transmission electron microscopy.
2014 Elsevier Inc. All rights reserved.

1. Introduction
The reduced activation ferritic/martensitic (RAFM) steels, which
were initially developed for special high active waste storage of radioactive structures of fusion reactors, has been considered as the primary
material for construction of rst wall and test blanket module in the fusion reactors [13]. The chemical composition of 9Cr RAFM steel has
been arrived at by modifying the conventional grade 91 steel (9Cr
1Mo0.06Nb0.2 V0.05 N) with the substitution of Mo and Nb, the
high induced radioactive elements by W and Ta [4,5]. Tungsten addition
increases not only the creep rupture strength but also the ductile to brittle transition temperature (DBTT). Tantalum plays an important role in
lowering the DBTT of the RAFM steel by way of rening the prioraustenite grain size. The physical metallurgy of ferritic/martensitic
steels with respect to stability of microstructure on thermal and irradiation exposures, inuence of chemical composition and heat treatments
has been extensively studied by several researchers [6,7]. The 9Cr RAFM
steels possess desirable properties such as good mechanical properties,
adequate creep resistance up to 873 K, good weldability and limited radiological activation in addition to better void swelling resistance
[812]. Ferritic/martensitic steels in general, bear a complex microstructure and the mechanical properties are inuenced by the prior austenite
grain size (PAGS), martensite lath size, dislocation density, chemistry
and size of the precipitates.

Corresponding author.
E-mail address: vtp@igcar.gov.in (E. Rajendra Kumar).

http://dx.doi.org/10.1016/j.matchar.2014.08.013
1044-5803/ 2014 Elsevier Inc. All rights reserved.

Fabrication of structural components involves welding and other

joining procedures involving heating and cooling cycles, during which
the ferritic steels undergo microstructural and microchemical changes
[13]. The microstructure and microchemistry that evolve during
welding and subsequent post weld heat treatments (PWHTs) are sensitive to the alloy composition. Recent works on similar reduced activation steels give insight on detailed microstructure and mechanical
properties of weld and heat affected zone (HAZ), upon varying the
heat input and cooling rates [12,14,15]. In fusion welding, the solidied
metal as well as base metal undergo a number of thermal cycles which
depend on the distance from the heat source and number of passes
employed [16,17]. The weldment would possess a highly heterogeneous microstructure consisting of martensite, tempered martensite
and ferrite. The failures during creep testing and long term service exposures are found to occur in the HAZ region of the weld. This is referred as
type IV cracking [18,19]. Creep test conducted in weld samples showed
that ne grained heat affected zone (FGHAZ) exhibits minimum creep
rate than coarse grained heat affected zone which in fact shows higher
creep strength than base metal. A large number of creep cavities were
also found in FGHAZ region of the weld joint. However tests on simulated FGHAZ showed that fracture occurs by necking with no evidence of
cavities. Rupture life of the joint was found to be improved by decreasing the HAZ width [20]. Advanced welding techniques with low heat
input such as electron beam welding (EBW) and laser beam welding
(LBW) are alternate methods to reduce the HAZ width thereby minimizing the soft zone.
The present paper presents the results of a detailed study on the
microstructural modications that take place in the indigenously
manufactured 9Cr reduced activation ferritic martensitic (RAFM) steel

214

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

Table 1
Chemical composition of the two RAFM steels used in this study [22].
Steel

Cr

Mn

Ta

Ti

Ni

Mo

Nb

1W
IN-RAFM

9.04
9.03

0.08
0. 12

0.55
0.56

0.22
0.24

1.0
1.39

0.06
0.06

0.022
0.029

0.005
0.002

b0.002
b0.002

b0.002
b0.001

0.005
0.005

0.005
0.005

0.001
0.002

0.001
0.001

on minimization of hardness and microstructural variations across the

fusion zone.

Table 2
Welding parameters.
Welding technique

TIG

EB

Laser

Welding speed (m/min)

Heat input (kJ/mm)
Gas ow (l/min)

0.15
0.9
10

0.6
0.36

1
0.21
30

during fusion welding. The microstructural characteristics of the

weldment obtained by conventional tungsten inert gas (TIG) welding
are compared with high power-density EBW and LBW processes. The
modications during subsequent PWHTs have also been analyzed
employing conventional and electron microscopy techniques. The tempering of martensite in various regions of the weldment during PHWT
is discussed in detail to evaluate the mechanism and kinetics of the
tempering process. The PWHT conditions have been optimized based

2. Experimental
9Cr RAFM steels with 1 and 1.4 mass% of tungsten procured from M/s
MIDHANI, Hyderabad in the form of 12 mm and 6 mm thick plates have
been used in the present study. The steel was used in the standard heat
treated condition namely normalizing at 1253 K/1 h followed by tempering at 1033 K/1 h. Chemical composition of the two steels is given in
Table 1. Based on an earlier work on structureproperty correlations the
1.4% W steel has been designated as an Indian RAFM or INRAFM steel
which possesses good creep resistance with adequate ductility for fabrication of the test blanket module [21,22]. Autogenous bead-on-plate welds
were fabricated with both steels using 200 mm 200 mm 12 mm steel
plates. EBW and LBW joints of INRAFM steel were fabricated using 6 mm
plate as per optimized conditions [23] to obtain high quality welds. Details

Fig. 1. Microstructure of the normalized and tempered 9Cr1W RAFM steel. a) SEM micrograph showing tempered martensite structure. b) Thin foil TEM micrograph showing triple point
(arrow marked) of prior austenite grains and martensitic laths. c) Lenticular inter & intralath precipitates. d) Course globular precipitates in a well recovered region.

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

of the welding processes are given in Table 2. Autogenous TIG weld process was specically chosen to avoid compositional changes if any in
the weld metal as well as to simulate conditions similar to EBW and
LBW where ller material is not used.
Classication of different regions of the weldment as a function of
distance from fusion zone has been made for study of microstructural
and microchemical features. Further TIG welded joints were heat treated at three different temperatures in the range of 8731033 K for durations of 15, 30, 60 and 120 min to optimize the PWHT condition for the
indigenous steel. EB and laser welded joints were subjected to the optimum PWHT at 1033 K for 60 min.

215

The microstructure across the cross section of the weldment has

been studied on samples prepared by standard metallography methods.
Optical microscopy, scanning electron microscopy (SEM) and microhardness measurements (applied load of 100 g) were carried out. SEM
was carried out using a Philips XL30 ESEM at an operating voltage of
30 kV. Micro-texture evolution during solidication or during transformation has been studied on longitudinal cross section close to the weld
surface (face side) in the three types of joints using electron back scatter
diffraction (EBSD) technique in the SEM. The scan area is 390 pixels
horizontal 306 pixels vertical with a step size of 0.6 m. EBSD analysis
has been carried out on PWHT samples to obtain high value of indexing.

Fig. 2. TEM micrograph of carbon extraction replica of 1W RAFM steel showing. a) inter and intra lath boundary carbides (diffraction evidence from MX (left) and M23C6(right) in
inset). b & c) EDS spectrum from Cr rich M23C6 carbide and V rich MX carbide respectively. d) Histogram showing the size distribution of carbides in normalized and tempered steel.

216

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

Further, transmission electron microscopy (TEM) specimens from

each region of the weldment were prepared by sectioning parallel to fusion line with a precision cutting device. Thin foils were prepared from
these samples by jet thinning method using perchloric acid (10%) and
methanol (90%) mixture as electrolyte. Carbon extraction replicas

were prepared from each region of the weld and were used extensively
to identify the type and chemistry of carbides. TEM studies were carried
out on thin foils as well as carbon extraction replicas using Philips CM
200 under an operating voltage of 120200 kV, with INCA energy
dispersive spectrometer (EDS). The EDS spectra were quantied by

Fig. 3. Hardness prole and micrographs from various regions of TIG weldment of RAFM steel. a) Hardness plot across weldment of 1 and 1.4W steel, b) weld showing columnar grains and
a typical martensitic structure (inset), c) coarse grained heat affected zone, d) ne grained heat affected zone showing ferrite grains (arrow marked) with lower defect density, e) crystal
orientation map (standard IPF Z coloring) from the weld region showing random texture and f) subset region of a prior austenite grain with coarse martensite substructure.

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

CliffLorimer method using standard KAB values [24]. Activation energy

for tempering process has been calculated by an Arrhenius analysis
of the rate of reduction of hardness with temperature as a kinetic
parameter.
3. Results
Section 3.1 deals with characterization of the steels in as received
condition, while Section 3.2 presents the results on the microstructural
modications during welding and PWHT. Section 3.3 presents the
results on optimization of PWHT and the kinetics of tempering of
martensite.

217

wide distribution of sizes ranging from 10 to 300 nm. It could be clearly

classied as M23C6 based on the distinct microchemistry. In contrast, the
distribution of MX was very sparse and the size was conned to b20 nm.
It was possible to establish the co-existence of the two types of precipitates only at high magnications. A meaningful analysis of the individual size distribution could not be made due to the poor statistics at high
magnications. In view of the above fact the measured size distribution
of 10300 nm (Fig. 2d) which corresponds to M23C6 reects the presence of two size distributions of the M23C6 carbides. The 1.4W steel
also showed similar kind of trend in M23C6 composition except the W
content which varied from 10 to 15%. A minor increase in hardness
(240 VHN) and a renement in PAGS (10 m) are manifestations of increased W content in 1.4W steel.

3.1. Base Metal Characterization

3.2. Microstructural Characterization of Weldment
The SEM micrograph of the as received normalized and tempered
steels is shown in Fig. 1a. It reveals the presence of a number of precipitates decorating the prior austenite grain boundaries as well as within
the grains. The average size of prior austenite grains (PAGS) was evaluated as 14 for 1W steel. The average hardness was measured as 230 VHN,
in agreement with the observed tempered martensitic microstructure
[12,25].
TEM studies on thin foil (Fig. 1b) showed large number of dislocations and retention of lath morphology in the prior austenite grains
(with characteristic boundary triple point-arrow marked) [2,7]. The average lath size was estimated as 700800 nm. The laths were decorated
with carbides, predominantly of lenticular morphology, as seen in
Fig. 1c. However, few regions of ferrite with low dislocation density
are also observed (Fig. 1d). Interestingly, the carbides present in regions
of lower dislocation density, were found to be globular and comparatively coarser in size.
Fig. 2a shows the micrograph of the carbon extraction replica revealing the distribution of precipitates along the martensitic laths. Microdiffraction patterns from two typical precipitates are shown as insets
in Fig. 2a (diffracting planes are marked). On matching with standard
d-spacing values and angles between the planes, the phases were identied as M23C6 (zone axis b 011 N) and MX (zone axis b 001 N) type of
precipitates. Fig. 2b shows the EDS spectrum of M23C6 carbide which
is found to be rich in Cr, while Fig. 2c shows the V and Ta rich nature
of MX. Analysis of a number of carbides showed that the M23C6 contains
5560 wt.% Cr, 2030 wt.% Fe and 1020 wt.% W. Repeated analysis on
several regions indicated the predominance of M23C6 which showed a

The microstructural and micro-chemical changes and micro-texture

in the different zones of the weldments prepared by three welding techniques have been studied, which will be discussed in sequence in this
section.
3.2.1. Characterization of the Autogenous TIG Weld
Fig. 3a shows the hardness prole across the autogenous TIG weld of
two steels, simulating all the regions of the weldments. Hardness value
was found to be high in the weld and in the HAZ close to fusion zone and
beyond that a steady decrease of hardness was observed. The weld region exhibited a complete martensitic structure, as seen in Fig. 3b. It reveals a solidication structure with elongated prior austenite grains.
TEM micrograph (inset) revealed that the weld region comprised of
lath martensite with a high density of dislocations. The lath size was estimated to be in the range of ~400 nm. Average hardness of weld metal
was measured as 465 and 490 VHN for 1 & 1.4W steel respectively, in
agreement with the martensitic microstructure. The extent of HAZ
was ~ 1.6 mm for a weld bead of thickness 2 mm. Fig. 3c shows the
SEM micrograph of CGHAZ region close to fusion line. Microstructure
was found to be martensitic with an average PAGS of ~26 m and possessed the highest hardness (490 & 515 VHN for 1 & 1.4W steel respectively) in the weldment. The average PAGS in the FGHAZ region (Fig. 3d)
near to base metal measured only ~10 m. This region exhibited ferrite
grains of lower defect density in addition to martensitic regions of high
defect density. The hardness in FGHAZ was low ~ 215 VHN. The EBSD
crystal orientation map from the weld region in Fig. 3e shows a random

Fig. 4. Bright eld TEM micrographs of carbon extraction replicas from HAZ showing. a) Dissolution of carbides in the CGHAZ region except a few retained primary TaC precipitates. b) High
number density of carbides in FGHAZ.

218

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

microtexture, while Fig. 3f clearly conrms the large size of the prior
austenite grains in the weld region. The difference in prior austenite
grain size in the different regions of the weldment can be understood in
terms of the peak temperature attained by the region as a function of distance from the heat source. Immediate to FGHAZ a very narrow region

showed a duplex microstructure of tempered martensite and martensite.

TEM extraction replica micrographs of CGHAZ and FGHAZ regions (Fig. 4a
& b) showed no evidence for precipitates other than few primary TaC precipitates in the former but a high number density of carbides in the latter,
which also explains for the low hardness in this region.

Fig. 5. Hardness proles and microstructures of EB weldment of 1.4W RAFM steel. a) Hardness prole across the weldment b) columnar microstructure in the weld. c & d) EBSD crystal
orientation map (standard IPF Z coloring) from weld region and a typical prior austenite grain. e) CGHAZ with typical martensitic structure but PAGS comparable to base metal and
f) FGHAZ with signatures of large number of precipitates.

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

3.2.2. Characterization of the EB and Laser Welds

Hardness prole across the EB and LB weld joints is shown in Figs. 5a
and 7a respectively. Hardness of the weld metal was estimated to 474
and 497 VHN for EB and LB welds. Figs. 5b and 7b show columnar structures of martensite typical of the weld regions of the EB and LB welds.
The liquid metal is expected to attain a much higher temperature than
in the case of TIG weld, although the residence time at the peak temperature is low due to higher cooling rates. The prior austenite grains are
clearly discerned in the EBSD crystal orientation maps in Figs. 5c and
7c. No evidence for preferred orientation is observed. The prior austenite grains in the weld metal are distinctly smaller in the EB and LB welds
(Figs. 5d and 7d) as compared to TIG welds. The CGHAZ (Figs. 5e and 7e)
with a martensitic structure possessed the highest hardness in the
weldment, but the PAGS was comparable to the base metal. The
FGHAZ (Figs. 5f and 7f) exhibited martensitic or duplex (martensite
plus ferrite) structure with ner PAGS depending up on the distance
from the fusion zone. The variation in the number density of precipitates between the two regions is clearly demonstrated in the TEM micrograph of carbon extraction replica shown in Fig. 6a and b and the
precipitate distributions in Fig. 6c and d. The width of the HAZ in laser

219

beam welds was measured as ~ 0.7 mm, which is far lower than the
TIG (1.6 mm) bead on plate welds.
The microstructural parameters are compared in Table 3. The important differences observed in EBW and LBW as compared to TIG welding
are highlighted below:
a) LBW showed low width of the weld resulting in a narrow HAZ. EBW
also exhibited similar characteristics except the entry side of the
beam.
b) Smaller PAGS of the weld metal in EBW and LBW compared to TIG
welds could be due to lower residence time at peak temperature
and high cooling rates associated with the welding technique. The
lower residence time at peak temperature and high cooling rates
are also manifested in ner lath size in EBW and LBW as compared
to TIG weld.
3.3. Optimization of Post Weld Heat Treatment Conditions
It is well known that PWHTs are necessary in ferritic/martensitic
steels to remove the residual stresses, to improve toughness and to

Fig. 6. Extraction replica micrograph taken from heat affected zone of EB weldment of 1.4W RAFM steel showing, a & b) the presence of few ne carbides in CGHAZ and large number of
coarse carbides in FGHAZ. c & d) Size distribution of carbides in CGHAZ and FGHAZ.

220

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

Fig. 7. a) Microhardness prole across laser weld of 1.4W RAFM steel. b) Fully martensite structure in weld. c & d) EBSD crystal orientation map from weld region and a typical prior
austenite grain. e) Martensite with carbides in CGHAZ and f) FGHAZ showing martensite with ne PAGS.

obtain a relatively stable homogeneous microstructure across the

weldment. In order to optimize the same, heat treatments at 873,
973 and 1033 K for durations ranging from 0.25 to 2 h were carried
out on representative TIG welds followed by hardness measurements and
microstructural investigations. Fig. 8a shows the microhardness proles

from the weldments, heat treated at 873, 973 and 1033 K for 1 and 2 h
durations.
Fig. 8bd shows the SEM micrographs of weld region heat treated for
973 K/2 h, 1033 K/1 h and 1033 K/2 h respectively. The lath morphology
of martensite is evident in the specimen treated at lower temperatures,

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

Table 3
Microstructural parameters in three welding methods.

Thickness of weld metal (mm)

Width of HAZ (mm)
Hardness (VHN) weld
CGHAZ
PAGS (um) BM
Weld
CGHAZ
FGHAZ
Weldlath size (nm)

TIG weld
(1.4W)

EB weld
(1.4W)

Laser weld
(1.4W)

6
1.6
490 10
515 10
10
41
26
10
~300500

4.0
1.4
474 10
484 10
18
19
21
10
~150350

1.2
0.7
497 10
520 10
19
20
19
7
~200400

due to a lesser extent of recovery and recrystallisation. It is evident that

the effect of temperature in reducing the hardness is more pronounced
than the duration of treatment at any temperature. It is obvious from
the gure that 873 K is not adequate for tempering. Reduction in hardness after 2 h of heat treatment at 973 K was also not substantial
(339 VHN). However, the average hardness after 1 and 2 h of treatment
at 1033 K reduced to 298 and 272 VHN respectively.
Supporting this data, the weld showed a well tempered martensitic
structure with uniformly distributed precipitates after treatment for 2 h
at 1033 K (Fig. 8d) as compared to 1 h (Fig. 8c). Hence, the PWHT
conditions have been optimized as 2 h at 1033 K for a plate thickness
of 12 mm and above.

221

Fig. 9a represents the hardness of the weld region as a function of

time at each temperature. The observed exponential decrease with
time is represented by the following HollomonJaffe parameter [26]
combining the effect of temperature and time.

P H TC log t

where
P
H
T
t
C

HollomonJaffe parameter
HARDNESS value (VHN)
Temperature (K)
time (h)
a constant which depends on the composition of the steel and
thermal history.

From Eq. (1), it is possible to obtain the same tempering hardness

value by suitable selection of tempering temperature and time. The
value of C has been evaluated as 15.2 by an appropriate combination of
temperature and time of tempering. A plot of hardness vs Hollomon
Jaffe parameter is given in Fig. 9b. The hardness proles of the EBW and
LBW joints subjected to PWHT treatment for 1 h at 1033 K have been
shown in Figs. 5a and 7a respectively. From the hardness value of weld
region (250270 VHN) it is understood that the above said treatment is
sufcient for EBW and LBW joints of 6 mm plate.

Fig. 8. Optimization of PWHT for 1W RAFM steel. a) Microhardness prole before and after various PWHTs. b, c & d) Micrographs of weld region after PWHT of 973 K/2 h, 1033 K/1 h and
1033 K/2 h.

222

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

(Fig. 9c) followed by saturation conrms that major part of tempering

occurs within the initial few minutes and thereafter, recovery is sluggish. Hence, recovery rate is calculated from the slope of the initial portion of the curve for each temperature.
Fig. 9c shows the plot between ln (VHN/t) and 1/T. The linear
behavior with a negative slope suggests that the process follows an
Arrhenius behavior. The activation energy estimated from the slope is
0.65 eV. This value is comparable to the activation energy for interstitial
diffusion of carbon in -ferrite [27] which is ~0.80 eV. This conrms that
the interstitial diffusion of carbon in -ferrite is the rate controlling step
in the tempering process. The lower value of activation energy of carbon
diffusion in the present study is due to the high density of defects in the
martensite aiding the diffusion process.
4. Discussion

Fig. 9. Effect of temperature and duration of tempering on hardness and recovery rate.
a) Variation of hardness of weld zone (RAFM 1W) with respect to time and temperature.
b) Hardness vs HollomonJaffe parameter: T temperature in K & t time in h.
c) Arrhenius plot of the rate of recovery of weld region.

3.3. Kinetics of Tempering of Martensite

The kinetics of tempering in the weld region has been studied by
monitoring the temperature dependence of a kinetic parameter, in
this case the recovery rate. The rapid reduction in initial hardness

The RAFM steel in normalized and tempered condition exhibited a

tempered martensite structure with characteristic inter & intralath/
grain boundary precipitates of Cr rich M23C6 and V/Ta rich MX. The
lower number density of MX is in agreement with the low V/Ta content
in the steel. Although the identity of X has not been established during
this study, the MX precipitates are expected to be carbonitrides rather
than carbides, since dissolution of stable V(Ta)N in ferrite is higher
than that of V(Ta)C at tempering temperatures [28]. Thermocalc data
on the equilibrium phases in similar steel also reports that undissolved
precipitates at 1323 K are MX carbonitrides [29]. The microstructural
parameters of RAFM steel have been compared with that of modied
9Cr1Mo steel in Table 4. The role of W and Ta in pinning the migrating
austenite grain boundaries is clearly evident in the ne prior austenite
grain size, although there is a contribution from the lower solutionizing
temperature for RAFM steel. The decrease in average PAGS from 14 to
10 m with increase in W content from 1 to 1.4 wt.% is in agreement.
The weld (TIG) metal microstructure of RAFM steel consisted of
martensite devoid of precipitates. It is obvious that all carbides have
gone into solution in the fusion zone and the high cooling rate during
solidication and subsequent solid state transformation produces ne
lath martensite, with super saturation of carbon. Formation of martensite in the weld and HAZ is understood from the continuous cooling
transformation (CCT) diagram reported in literature for mod. 9Cr
1Mo steel [6]. For this class of steel, a shallow cooling from austenite
can result in martensitic structure at room temperature. Hardening in
the coarse grained HAZ (CGHAZ) is attributed to the high carbon supersaturation of the martensite due to dissolution of carbides on exposure
to high temperature (far above AC3), while softening in the FGHAZ is the
consequence of fresh precipitation or coarsening of the carbides due to
exposure to lower temperature (above ~873 K), around the tempering
temperature.
The observed variation in PAGS in the HAZ is understood based on
the peak temperature experienced by each region as a function of distance from the heat source. In addition to peak temperature the stay
time of each region at that temperature also controls the stability of carbide or in other words the carbon content in the transformed austenite
phase. The observed low hardness of the martensite in FGHAZ can be
understood against this background [30]. The presence of few ferrite
grains with lower defect density amidst lath martensite indicates a
mixed microstructure suggesting a correspondence to the intercritical
HAZ. This is attributed to the microscopic region witnessing a temperature regime corresponding to the intercritical temperature of Ac3 N
T b Ac1 for this steel.
The high energy density of the laser/electron beam allows the surface of the material to attain the liquidus temperature rapidly, allowing
for a short interaction time of the beam as compared to TIG welding or
other traditional welding methods. With far less time for the energy to
dissipate into the interior of the work piece, the high power density
welding methods result in small heat-affected zones which ensure
good integrity of the weld. The lower PAGS (Table 3) in the weld as

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

223

Table 4
Comparison of microstructural/microchemical parameters of two ferritic steels [27].
Steel

Heat treatment

PAGS
(average)
(m)

Lath size
(nm)

Hardness
(VHN)

M23C6

MX

9Cr1WVTa

1253
1033
1333
1033

14

700800

230 5

1030 nm

40

5001000

205 5

30300 nm
Cr, W rich
40400 nm
Cr, Mo rich

9Cr1MoVNb

K/1
K/1
K/1
K/1

h (n)
h (t)
h (n)
h (t)

well as in CGHAZ exhibited by high energy density welding methods of

electron beam and laser beam is also understood against this background. The variation in hardness in weld metal is not signicant for
the three welding methods despite a ner substructure in EB and LB
welds as compared to TIG welds. This is understood in terms of lower
solute content of martensite resulting from incomplete dissolution of
carbides due to lower residence time in the austenite phase eld
which compensates the effect due to ner substructure.
Heat treatment of the martensite structure resulted in a variety of
microstructures and marked change in the microchemistry of the
phases. Analysis on carbon extraction replicas showed that the number
density of carbides did not change signicantly after a short duration of
heat treatment. It is evident that the observed reduction in hardness is
due to coarsening of the precipitates. It is clear from Fig. 8d, that 2 h of
PWHT at 1033 K results in a well tempered structure with uniformly
distributed precipitates as compared to lower temperature or durations.
Comparison of average hardness (272 VHN) and microstructure
(Fig. 8d) of the PWHT weld with the normalized and tempered steel
(230 VHN) (Fig. 1a), treatment of 2 h duration at 1033 K is considered
to be the optimum PWHT for TIG welds of RAFM steel.

5. Summary & Conclusion

The microstructure of various regions of a weldment fabricated by
three welding methods namely TIG, EBW and LBW of indigenously
developed RAFM steel was studied by detailed electron microscopy.
The important results of the study are summarized below:
Prior austenite grain size in weld region of TIG weld was far higher
than the EBW and LBW due to the higher residence time of the steel
at temperatures above Ac3.
Random micro-texture was observed in all the three welds.
Prior austenite grain size in HAZ is controlled by the peak temperature
and cooling rate in the region as a function of distance from the heat
source.
Variation of hardness across the HAZ is understood based on the
decreasing peak temperature and thereby cooling rate resulting
in decreasing dislocation density and supersaturated carbon in
the martensite.
Negligible amount of carbides in CGHAZ in TIG weld is attributed to
the complete dissolution of carbides at high temperatures prevailing
in the CGHAZ. The higher carbon super saturation in the martensite
and the high dislocation density due to high cooling rate explains
for the high hardness in this region, despite the large grain size.
Weld metal in all the cases showed comparable hardness despite ner
lath size in EBW and LBW due to incomplete dissolution of carbide at
short residence time in the high temperature regime.
Optimum condition for PWHT is identied as 2 h at 1033 K for TIG
welds of RAFM steel, while for EBW and LBW joints of 6 mm plate, a
PWHT treatment of 1033 K/1 h is sufcient.
The activation energy of the tempering process has been evaluated
and found to be matching with the activation energy for migration
of carbon in ferrite matrix.

4080 nm

Acknowledgments
The authors would like to express their thanks to Dr. P.R. Vasudeva
Rao, Director, Indira Gandhi Centre for Atomic Research, for his constant
support.
References
[1] M. Tamura, H. Hayakawa, M. Tanimura, A. Hishinuma, T. Kondo, Development of potential low activation ferritic and austenitic steels, J. Nucl. Mater. 141143 (1986)
10671073.
[2] R.L. Klueh, A.T. Nelson, Ferritic/martensitic steels for next-generation reactors, J.
Nucl. Mater. 371 (2007) 3752.
[3] J.-F. Salavy, L.V. Boccaccini, P. Chaudhuri, S. Cho, M. Enoeda, L.M. Giancarli, R.J. Kurtz,
T.Y. Luo, K. Bhanu Sankara Rao, C.P.C. Wong, Must we use ferritic steel in TBM?
Fusion Eng. Des. 85 (2010) 18961902.
[4] A. Lindau, M. Schirra, First results on the characterization of the reduced-activationferriticmartensitic steel EUROFER, Fusion Eng. Des. 5859 (2001) 781785.
[5] R.L. Klueh, E.E. Bloom, The development of ferritic steels for fast inducedradioactivity decay for fusion reactor applications, Nucl. Eng. Des. 2 (1985)
383389.
[6] M. Tamura, H. Hayakawa, A. Yoshitake, A. Hishinuma, T. Kondo, Phase stability of
reduced activation ferritic steel: 8%Cr2%W0.2%V0.04%TAFe, J. Nucl. Mater.
155157 (1988) 620625.
[7] R. Schaublin, P. Spatig, M. Victoria, Microstructure assessment of the low activation
ferritic/martensitic steel F82H, J. Nucl. Mater. 258263 (1998) 11781182.
[8] H. Hayakawa, A. Yoshitake, M. Tamura, S. Natsume, A. Gotoh, A. Hishinuma,
Mechanical properties of welded joints of the reduced-activation ferritic steel:
8%Cr2%V0.04%TaFe, J. Nucl. Mater. 371 (1991) 179181.
[9] Baldev Raj, K. Bhanu Sankara Rao, A.K. Bhaduri, Progress in the development of
reduced activation ferriticmartensitic steels and fabrication technologies in India,
Fusion Eng. Des. 85 (2010) 14601468.
[10] A. Alamo, A. Castaing, A. Fontes, P. Wident, Effects of thermal aging behavior of F82H
weldments, J. Nucl. Mater. 283287 (2000) 11921195.
[11] P. Aubert, F. Tavassoli, M. Rieth, E. Diegele, Y. Poitevin, Review of candidate welding
processes of RAFM steels for ITER test blanket modules and DEMO, J. Nucl. Mater.
417 (2011) 4350.
[12] Shuhui Zheng, Wu. Qingsheng, Qunying Haung, Shaojun Liu, Yangyang Han,
Inuence of different cooling rates on the microstructure of the HAZ and welding
CCT diagram of CLAM steel, Fusion Eng. Des. 86 (2011) 26162619.
[13] T. Hirose, K. Shiba, T. Sawai, S. Jitsukawa, M. Akiba, Effect of heat treatment process
for blanket fabrication on mechanical properties of F82H, J. Nucl. Mater. 329333
(2004) 324327.
[14] Zhizhong Jiang, Litian Ren, Jihua Huang, Xin Ju, Huibin Wu, Qunying Huang, Yican
Wu, Microstructure and mechanical properties of the TIG welded joints of fusion
CLAM steel, Fusion Eng. Des. 85 (2010) 19031908.
[15] B. Arivazhagan, G. Srinivasan, S.K. Albert, A.K. Bhaduri, A study on inuence of heat
input variation on microstructure of reduced activation ferritic martensitic steel
weld metal produced by GTAW process, Fusion Eng. Des. 86 (2011) 192197.
[16] M. Vijayalakshmi, S. Saroja, V. Thomas Paul, R. Mythili, V.S. Raghunathan, Microstructural zones in the primary solidication structure of weldment of 9Cr1Mo
steel, Metall. Trans. 30A (1999) 161174.
[17] V. Thomas Paul, S. Saroja, P. Hariharan, A. Rajadurai, M. Vijayalakshmi, Identication
of microstructural zones and thermal cycles in a weldment of modied 9Cr1Mo
steel, J. Mater. Sci. 42 (2007) 57005713.
[18] D.J. Allen, S.J. Brett, Failure of a modied 9Cr header endplate, Proceedings of International Symposium on Case Histories on Integrity and Failures in Industry, Held in
Milan, Italy, Engineering Materials Advisory Services Ltd, London, UK, 1999, pp.
133143.
[19] J.A. Francis, W. Mazur, H.K.D.H. Bhadeshia, Type IV cracking in ferritic power plant
steels, Mater. Sci Technol. 22 (12) (2006) 13871395.
[20] K. Shinozaki, D.-J. Li, H. Kuroki, H. Harada, K. Ohishi, Analysis of degradation of creep
strength in heat affected zone of weldment of high Cr heat resisting steels based on
void observation, ISIJ Int. 42 (2002) 15781584.
[21] K. Laha, S. Saroja, A. Moitra, R. Sandhya, M.D. Mathew, T. Jayakumar, E. Rajendra
Kumar, Development of India-specic RAFM steel through optimization of tungsten
and tantalum contents for better combination of impact, tensile, low cycle fatigue
and creep properties, J. Nucl. Mater. 439 (2013) 4150.

224

V. Thomas Paul et al. / Materials Characterization 96 (2014) 213224

[22] T. Jayakumar, M.D. Mathew, K. Laha, S.K. Albert, S. Saroja, E. Rajendrakumar, C.V.S.
Murthy, G. Padmanabham, G. Appa Rao, S. Narahari Prasad, Reduced activation
ferritic martensitic steel and fabrication technologies for the Indian test blanket
module in ITER, Fusion Sci. Technol. 65 (2014) 171185.
[23] K.A. Gopal, S. Murugan, S. Venugopal, K.V. Kasiviswanathan, Laser welding of precision
engineering components, Indian Weld. J. 44 (2011) 5459.
[24] G. Cliff, G.W. Lorimer, Quantitative Microanalysis with High Spatial Resolution, The
Metal Society, London, 1981.
[25] H. Tanigawa, T. Hirose, K. Shiba, R. Kasada, E. Wakai, H. Serizawa, Y. Kawahito, S.
Jitsukawa, A. Kimura, Y. Kohno, A. Kohyama, H. Mori, K. Nishimoto, R.L. Klueh,
M.A. Sokolov, R.E. Stoller, S.J. Zinkle, Technical issues of reduced activation ferritic/
martensitic steels for fabrication of ITER test blanket modules, Fusion Eng.Des. 83
(2008) 14711476.

[26] J.M. Vitek, R.L. Klueh, Precipitation reactions during heat treatment of ferritic steels,
Metall. Trans. 14A (1983) 10471055.
[27] V. Raghavan, Material Science and Engineering, Prentice-Hall, New Delhi, 1984. 168.
[28] V. Foldyna, Z. Kubon, A. Jakobava, V. Vodarek, in: A. Strang, D.J. Gooch (Eds.), Microstructural Developments and Stability in High Chromium Ferritic Power Plant Steels,
Book, 667, The Inst. of Materials, London, 1997, pp. 7392.
[29] J. Hald, Metallurgy and creep properties of new 912%Cr steels, Steel Res. 67 (9)
(1996) 369374.
[30] Smallman, A.W.H. Ngan, Modern Physical Metallurgy and Advanced Materials, 4th
ed. Butterworth-Heinemann, 2007. 421.