Journal of Nuclear Materials 355 (2006) 68–82

[Link]/locate/jnucmat

Sensitization behaviour of modified 316N and 316L

stainless steel weld metals after complex annealing
and stress relieving cycles
N. Parvathavarthini a, R.K. Dayal a,*, H.S. Khatak a,
V. Shankar b, V. Shanmugam b
a
Indira Gandhi Centre for Atomic Research, Corrosion Science and Technology Division, Materials Characterisation Group,
Kalpakkam, Tamil Nadu 603 102, India
b
Materials Technology Division, Materials Development Group, Indira Gandhi Centre for Atomic Research, Kalpakkam – 603 102
Tamil Nadu, India

Received 15 September 2005; accepted 20 April 2006

Abstract

Sensitization behaviour of austenitic stainless steel weld metals prepared using indigenously developed modified 316N
(C = 0.05%; N = 0.12%) and 316L (C = 0.02%; N = 0.07%) electrodes was studied. Detailed optical and scanning electron
microscopic examination was carried out to understand the microstructural changes occurring in the weld metal during
isothermal exposure at various temperatures ranging from 500 C to 850 C (773–1123 K). Based on these studies the
mechanism of sensitization in the austenite–ferrite weld metal has been explained. Time–temperature–sensitization
(TTS) diagrams were established using ASTM A262 Practice E test. From the TTS diagrams, critical cooling rate
(CCR) above which there is no risk of sensitization was calculated for both materials. The heating/cooling rates to be
followed for avoiding sensitization during heat treatment cycles consisting of solution-annealing and stress-relieving in
fabrication of welded components of AISI 316LN stainless steel (SS) were estimated taking into account the soaking time
and the number of times the component undergoes thermal excursions in the sensitization regime. The results were vali-
dated by performing controlled heating and cooling heat treatment trials on welded specimens.
 2006 Elsevier B.V. All rights reserved.

1. Introduction Breeder Reactor (PFBR) being built at Kalpakkam.

Welding is extensively employed in the fabrication
AISI 316LN SS (C = 0.024–0.03% and N = of PFBR components. Sensitization in the base
0.06–0.08%) has been chosen as the primary struc- metal is not a problem at the heat-affected zone of
tural material for the 500 MWe Prototype Fast 316LN SS weldments because carbon content of
the base material is 60.03%. Normally, the weld
*
Corresponding author. Tel.: +91 44 27480121; fax: +91 44
metal does not get sensitized because of the duplex
27480081. austenite–ferrite structure in which chromium
E-mail address: rkd@[Link] (R.K. Dayal). depletion due to carbide precipitation does not

0022-3115/$ - see front matter  2006 Elsevier B.V. All rights reserved.
doi:10.1016/[Link].2006.04.006
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 69

readily occur. Even if carbide precipitation takes facing because of the high stresses built-up at
place, rapid diffusion of chromium from the chro- the coating-substrate interface. Without this
mium-rich delta ferrite phase results in quicker heat treatment cracking of the coating can
homogenisation. However, this immunity from take place during inservice thermal cycling
sensitization and subsequent intergranular corro- because of the mismatch between the expan-
sion (IGC) can be severely compromised especially sion coefficients of the hard face coating and
on heat treatment in the temperature range of trans- stainless steel substrate.
formation. The severity of this problem depends on (iii) Dimensional stabilization treatment is required
the carbon content as well as the extent to which the for relieving peak residual stresses and is
duplex structure is lost through the dissolution of performed before final machining to prevent
ferrite. For PFBR components welding of the AISI distortion during final machining and assem-
316LN SS is to be carried out using modified 316N bly. It is performed at a temperature 50 C
(as per AISI/AWS SFA-5.4) electrodes developed above the expected peak transient temperature
indigenously. Weld metal cracking is controlled by (the highest temperature the component is
optimizing the chemical composition of the welding likely to experience for short duration of time
consumables. Carbon in the range of 0.045–0.055% during service).
and nitrogen in the range of 0.06–0.1% are specified
to provide weld joints with improved creep strength According to ASM hand book [2], heat treatment
and freedom from sensitization in the as-welded periods of 3–4 h per 25 mm are used for dimensional
state. In addition the ferrite content in the weld stabilization treatment and 1 h per 25 mm is used
metal is specified to be between 3 and 7 FN (mea- for solution-annealing treatment. Heating and
sured magnetically) to promote ferrite solidification cooling rates are to be controlled because very fast
mode. cooling rates will result in distortion and reintroduc-
Following three types of heat treatments are tion of residual stress, while too slow a cooling will
required for austenitic stainless steel components lead to sensitization. Hence optimum heating/cool-
[1]: ing rates are to be established for welded
components.
(i) Solution-annealing at 1050 C (1323 K) or A large volume of literature exists with reference
above for full stress-relieving, restoration of to carbide precipitation and sensitization behaviour
mechanical properties and corrosion resis- of austenitic stainless steel base metals [3–7], where
tance particularly when the maximum allow- nucleation of the carbide phase requires a certain
able level of cold work is exceeded. incubation period. Sensitization kinetics in the
Following this treatment, slow cooling is molybdenum-containing grades such as 316N and
necessary to avoid reintroduction of residual 316L are governed by the carbon and nitrogen con-
stress. When the welded components are solu- tents [5]. Duplex stainless steels represent a different
tion-annealed, delta ferrite dissolves and the situation where the large area of austenite/ferrite
weld metal behaviour is similar to that of interphase interfaces serves to reduce the barrier
austenitic base metal of higher carbon content to nucleation and the precipitation kinetics are
(0.05% as against 0.03% max. in base metal) faster [8]. In duplex stainless steels, precipitation
for which CCR to avoid sensitization is much of chromium carbides and nitrides leads to pit-
higher than that of the base metal. Therefore, ting corrosion [9], while sigma phase formation
it becomes necessary to optimize the heating increases susceptibility to stress corrosion cracking
and cooling cycles to avoid sensitization as [10]. Austenitic weld metal with only a few percent
well as to preclude reintroduction of residual of retained ferrite phase represents an intermediate
stresses. situation with reference to both these cases. While
(ii) Some components of PFBR encounter wear the predominantly austenitic structure renders the
(adhesive or abrasive) due to sliding move- material susceptible to sensitization by chromium
ment and erosion due to high velocity sodium. depletion, some amount of protection from these
Nickel base, cobalt-free hard facing is used for phenomena is afforded by the ferrite present in
mating parts in most of the nuclear steam welds [11] and castings [12]. The amount of fer-
system components. Stress-relieving at 750– rite present and therefore the austenite/ferrite
850 C (1023–1123 K) is required after hard interfacial area can decrease significantly with heat
70 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

treatment through transformation of the ferrite to Table 1b

austenite, carbide and sigma phase. Under such Welding parameters for modified 316L SS weld metal
conditions, austenitic stainless steel weld metal Pad Arc voltage Arc current Speed Heat input
and welded components normally immune to sensi- no. (V) (A) (mm/min) (J/mm)
tization, can be rendered susceptible to corrosion 1 20 80 180–200 480–533
attack [13]. Sensitization data on austenitic stainless 2 20–24 90–100 180–210 540–735
3 20 80–90 170–210 457–568
steel weld metal is scarce. In the present case, the 4 24–26 100–110 170–220 720–918
complex manufacturing cycle necessitated establish-
ment of time–temperature–sensitization (TTS)
diagrams, calculation of CCR from TTS diagrams
and validation of the estimated CCR. Validation Table 2a
was performed taking into account the number of Chemical composition for modified 316N SS weld metal
times the component undergoes thermal excursions Element Wt.%
in sensitization regime with various soaking times Carbon 0.050
by performing controlled heating/cooling heat Nitrogen 0.120
treatments as necessary. To begin with, systematic Chromium 18.5
Nickel 11.1
investigations were carried out to understand the Molybdenum 1.9
sensitization behaviour by establishing the TTS Manganese 1.4
diagram for the weld metal prepared using modified Silicon 0.46
316N and 316L electrodes. Detailed microstructural Sulphur 0.006
characterisation of the various heat treated speci- Phosphorus 0.025
Tantalum <0.010
mens has been carried out using optical and Titanium 23 ppm
scanning electron microscopy to explain the mecha- Niobium <0.07
nism of sensitization in modified 316N and 316L Vanadium 0.075
stainless steel weld metals. Boron <10 ppm
Copper 0.210
Cobalt 0.060
2. Experimental
Delta ferrite
As per WRC 92 1.8 FN
Weld pads were prepared by Shielded Metal Arc
By Ferritescope 3.5, [4.1, 3.9] FN
Welding (SMAW) process using modified 316N
electrodes (3.15 mm diameter) and 316L (3.15 mm
diameter) of M/s. Mailam India Ltd., and the weld-
Table 2b
ing parameters used are presented in Tables 1a and
Chemical composition for modified 316L SS weld metal
1b. The chemical composition of the weld metals
Element Wt.%
was analysed using optical emission spectroscopy
and the results are given in Tables 2a and 2b. The Carbon 0.020
Nitrogen 0.070
equipment used was Jobin Yvon make model JY-
Chromium 18.5
132 F. Liquid penetrant and radiographic examina- Nickel 11.5
tions were performed to ensure that the weld pads Molybdenum 2.3
were free from porosity and defects. From the weld Manganese 1.6
pads, specimens of dimensions 80 · 10 · 3 mm were Silicon 0.28
Sulphur 0.009
cut for sensitization studies as per ASTM Standard
Phosphorus 0.045
Titanium <0.080
Niobium <0.070
Table 1a Vanadium 0.075
Welding parameters for modified 316N SS weld metal Copper 0.210
Cobalt 0.060
No. of Arc voltage Arc current Speed Heat input
passes (V) (A) (mm/min) (J/mm) Delta ferrite 8–11 FN
1 22–23 115–120 160 991
2 22–23 115–120 186 852
3–16 21–22 120–130 160 1008 A262 Practice E test. Delta ferrite measurements
17–26 21–22 130–140 290 600
were carried out using Ferritescope. Specimens
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 71

Table 3 3. Results and discussion

Delta ferrite content of modified 316L SS weld metal after
various heat treatments
3.1. Development of TTS diagram
Temp./time Ferrite number
24 h 40 h 50 h 100 h 200 h The results obtained for the various heat treated
625 C (898 K) – – 5.27 2.67 1.74 specimens in the ASTM standard A262 practice E
650 C (923 K) 3.86 2.94 2.68 1.18 – for modified 316N SS weld metal and modified
675 C (948 K) – – 1.10 0.36 0.04 316L SS weld metal are collectively presented in a
TTS diagram in Fig. 1. Sensitized specimens with
cracks after exposure and bending are indicated by
made from weld pads prepared using 316N elec- closed circles. Specimens with no cracking in the
trodes were heat treated at various temperatures bent portion were adjudged free from sensitization
ranging from 500 C to 850 C (773–1123 K) for (points marked X). From these data points, sensi-
various durations ranging from 20 min to 20 h. tized and non-sensitized area are separated by draw-
Specimens made from weld pads prepared using ing a smooth curve such that the curve represents
316L electrodes were heat treated at 625 C the beginning of sensitization process. Efforts were
(898 K), 650 C (923 K) and 675 C (948 K) for made to keep all the non-sensitized data points out-
durations ranging from 24 h to 200 h. Changes in side the curve except one data point at 550 C
delta ferrite content after ageing treatment of 316L (823 K) which has fallen on the curve. This has
SS weld metal are presented in Table 3. happened in order to maintain the smoothness of
The heat treated specimens were electrolytically the curve. By drawing the curve in this way it is
etched in 10% oxalic acid for a few seconds at ensured that there is no data point outside the curve
2 V in order to reveal the microstructure. Optical which will show sensitization.
and scanning electron microscopic examination From the diagram, it is clear that the modified
was carried out for characterising the microstruc- 316N SS weld metal gets sensitized in the tempera-
ture. The specimens were analysed for the possible ture range of 625–725 C (898–998 K). The mini-
formation of secondary phases. Bulk extraction of mum time required for sensitization at the nose
the precipitates was carried out for modified 316N temperature (tmin) is about 25 min. From the data,
SS weld specimens by keeping the specimens at the CCR above which there is no risk of sensitiza-
1.5 V with respect to platinum cathode in 10% tion, was calculated using the method developed
Hydrochloric acid – 90% methanol solution for a by Dayal and Gnanamoorthy [16]. From these
duration of 24 h. This results in selective dissolu- calculations, the CCR for weld metal was deter-
tion of austenite phase [14], while carbides and mined as 160 C/h. This critical cooling rate is appli-
other secondary phases remain undissolved and cable only when cooling of the weld metal starts
were collected by centrifuging. The extracted from 725 C (998 K).
powder samples were washed in methanol, dried It can be seen that the 316L SS weld metal gets
and analysed by X-ray diffraction technique using sensitized in the temperature range of 625–675 C
Cu Ka radiation. The modified 316L SS weld spec-
imens were etched in modified Murakami reagent
for 15 s to reveal secondary intermetallic phases No attack Partial attack Attack

such as sigma and chi. 900 Modified 316N 316L 1173

Critical cooling rate - 160 °C / h Critical cooling rate - 1 °C / h
All the heat treated specimens were tested to 850 1123
TEMPERATURE, K
TEMPERATURE,˚C

determine occurrence of sensitization as per ASTM 800 1073

standard A262 practice E [15]. According to this 750 1023

700 973
standard, the specimen to be tested was embedded
650 316L 923
in copper turnings and exposed to boiling copper Modified 316N
600 873
sulphate–sulphuric acid solution for 24 h. Then the 550 823
specimens were bent slowly through 90 angle 500 773
around a mandrel of 6 mm diameter. Those speci- 0.1 1.0 10 100 1000
mens in which cracks were observed in the bent TIME, h

portion at a magnification of 20· were categorised Fig. 1. Time–temperature–sensitization diagrams for modified
as sensitized. 316N SS weld metal and 316L SS weld metal.
72 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

(898–948 K). The minimum time required for sensi-

tization at the nose temperature (tmin) is about 40 h.
From this diagram, CCR above which there is no
risk of sensitization, was calculated as 1 C/h. This
value is applicable only when cooling of the weld
metal starts from 675 C (948 K).

3.2. Microstructural changes during heat treatment

3.2.1. Modified 316N stainless steel weld metal

The initial microstructure of modified 316N SS
weld metal specimen is presented in Fig. 2. Duplex
austeno-ferritic structure with delta ferrite of
vermicular morphology can be clearly seen in the Fig. 3(a). Optical micrograph of modified 316N SS weld metal
microstructure. Delta ferrite was estimated to be in aged at 525 C (798 K) for 10 h showing delta ferrite and carbides
the range of 3.5–4.1 FN when measured using at delta/austenite boundaries.
Ferritescope on the top bead. However, in some
regions of the multipass weld metal, considerable
transformation of the ferrite had taken place. In these
regions, nucleation of M23C6 carbide is expected in
the as-welded condition itself by exposure to the
temperature range of 450–750 C (723–1023 K) by
thermal cycling during multipass welding. The heat
treatments given subsequently to the test specimens
served to further transform the ferrite to various
extents depending upon the temperature and time
of exposure.
When heat treated at 525 C (798 K) for 10–15 h,
the delta ferrite has transformed to carbides and
austenite and growth of M23C6 carbides along delta
ferrite/austenite (d/c) boundaries has taken place.
This is presented in Figs. 3(a) and 3(b), respectively,
Fig. 3(b). Optical micrograph of modified 316N SS weld metal
for 10 h and 15 h exposure. Initially, the carbide film aged at 525 C (798 K) for 15 h showing carbides at intercellular
regions.

is continuous (Fig. 3(a)), which is shown by the

darkening of the d/c boundary. The delta ferrite
grains are also considerably intact, while in
Fig. 3(b), most of the delta ferrite phase has trans-
formed and the remaining phases in the intercellular
regions are carbide particles, as shown by discrete
particles with dark appearance on etching with
10% oxalic acid. Precipitation of intermetallic
phases such as sigma or chi phase is not expected
at this temperature [17,18]. The microstructure in
Fig. 3(a) corresponds to the unsensitized region
and 3(b) to the sensitized region.
The microstructural changes with heat treatment
Fig. 2. Optical micrograph showing the initial microstructure of at 600 C (873 K) are shown in Figs. 4(a)–(c), for
modified 316N SS weld metal (delta ferrite with vermicular treatment times of 1 h, 2 h and 3 h respectively.
morphology). Here Fig. 4(a) shows delta ferrite lined with carbide
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 73

film in some areas (dark boundaries) and the delta

ferrite shows considerable signs of transformation.
After 2 h (Fig. 4(b)) the breakup of delta ferrite is
more advanced and carbide films have broken up
into discrete particles that have coarsened and are
present throughout the microstructure. When age-
ing is further continued for 3 h, the carbide particles
have coarsened appreciably. The series of three
micrographs corresponds to the progression from
unsensitized, to nearly sensitized to sensitized
microstructure. The association of delta ferrite with
absence of sensitization and its disappearance with
sensitization is therefore clear. X-ray diffraction
analysis of extracted precipitates of a specimen aged
Fig. 4(a). Optical micrograph of modified 316N SS weld metal
aged at 600 C (873 K) for 1 h showing carbides (dark regions at 600 C (873 K) for 10 h confirmed that M23C6
marked C) and untransformed delta ferrite. carbide was the most predominant phase present
and delta ferrite was absent in the sensitized
condition.
An extreme case where delta ferrite was totally
absent in the sensitized microstructure is shown in
Fig. 5, which shows a specimen aged at 650 C
(923 K) for 10 h. Here, the austenite grain bound-
aries have been decorated extensively with coarse
carbide precipitates. (The round particles in the
microstructure are slag inclusions from the welding
flux, as confirmed by EDAX analysis.) This type of
grain boundary precipitation is typically found in
single-phase austenitic base metal and is associated
with sensitization.
At higher temperatures of ageing such as 850 C
(1123 K), delta ferrite almost completely transforms
to sigma phase, as illustrated in Fig. 6. Any carbide
Fig. 4(b). Optical micrograph of modified 316N SS weld metal precipitation that is observed is likely to have
aged at 600 C (873 K) for 2 h showing discrete carbides and already formed during multipass welding, as it does
break up of delta ferrite.

Fig. 5. Optical micrograph of modified 316N SS weld metal aged

Fig. 4(c). Optical micrograph of modified 316N SS weld metal at 650 C (923 K) for 10 h showing austenite grain boundaries
aged at 600 C (873 K) for 3 h showing coarsening of carbides. being decorated with coarse carbides.
74 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

3.2.2. Modified 316L SS weld metal

The initial microstructure of modified 316L SS
weld metal is presented in Fig. 8. The as-deposited
weld metal showed a discontinuous network of delta
ferrite of vermicular morphology. The delta ferrite
was estimated to be in the range of 8–11 FN. When
heat treated at 625 C (898 K) for 50 h, delta ferrite
content decreased to 5.27 FN and Fig. 9(a) shows
microstructure obtained after etching in oxalic acid
indicating delta ferrite lined with carbide film.
Fig. 10(a) is the corresponding microstructure after
etching in modified Murakami reagent which shows
sigma phase precipitated within delta ferrite
through out the specimen. In this etchant, matrix
Fig. 6. Optical micrograph of modified 316N SS weld metal aged
at 850 C (1123 K) for 1 h showing sigma phase.
and ferrite phases are stained light and medium
tan colour respectively and sigma is stained reddish
brown colour. Transformation of delta ferrite and
not redissolve at this temperature. The extracted
appearance of M23C6 and sigma can be clearly seen
precipitate for a specimen aged at 850 C (1123 K)
in Fig. 10(a). Since the carbide network and the
for 1 h showed predominantly sigma phase,
associated chromium depletion are discontinuous,
although some M23C6 as remnant of the unaged
the specimen did not fail in ASTM A262 Practice
multipass weld microstructure is possible.
E test. Further ageing at the same temperature for
The fractograph of the specimen aged at 600 C
100 h reduces delta ferrite content to 2.7 FN (Table
(873 K) for 10 h and subjected to ASTM A262 Prac-
3). Delta ferrite grains are still present and nearly
tice E test is presented in Fig. 7. As illustrated in this
continuous network of carbides are seen in
figure, fracture along interdendritic regions has been
Fig. 9(b). The corresponding microstructure of the
observed in sensitized weld metal all along the
specimens in Fig. 10(b) showing elongated sigma
edges. At the centre of the specimen ductile fracture
phase. This specimen also did not fail in ASTM
was observed. This is in contrast to the intergranu-
A262 Practice E test for the same reason explained
lar fracture along the austenite grain boundaries
for the 50 h aged specimen. Extended ageing
usually observed in the HAZ of austenitic stainless
resulted in IGC attack. Fig. 9(c) shows continuous
steel base metals. The association of the chromium
carbide network and Fig. 10(c) shows more of elon-
depletion with the dendritic structure is because of
gated sigma particles as network. The failure in U
the carbide precipitation at these boundaries, where
delta ferrite was also present.

Fig. 7. Fractograph of modified 316N SS weld specimen aged at Fig. 8. Optical micrograph showing the initial microstructure of
600 C (873 K) for 100 h and subjected to ASTM A262 Practice E modified 316L SS weld metal (delta ferrite with vermicular
test showing fracture along interdendritic features. morphology).
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 75

Fig. 9. Microstructure obtained for modified 316L SS weld metal Fig. 10. Microstructure obtained for modified 316L SS weld
(a) 625 C (898 K) – 50 h; (b) 625 C (898 K) – 100 h; (c) 625 C metal (a) 625 C (898 K) – 50 h; (b) 625 C (898 K) – 100 h; (c)
(898 K) – 200 h (Etchant – 10% oxalic acid). 625 C (898 K) – 200 h; (Etchant – modified Murakami solution).

bend test can be attributed to sensitization and IGC progressively transforms to carbide and sigma phase
due to Cr depletion and inherent loss in ductility as is evident from the FN (Table 3) and the micro-
due to sigma formation. structures given in Fig. 11(a–d). Here 40 h itself
When aged at 650 C (923 K) for various dura- leads to IGC. When aged at 675 C (948 K) for
tions viz., 24 h, 40 h, 50 h and 100 h delta ferrite 50 h, 100 h and 200 h , the delta ferrite content
76 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

Fig. 11. Microstructure obtained for modified 316L SS weld metal (a) 650 C (923 K) – 24 h; (b) 650 C (923 K) – 40 h; (c) 650 C (923 K)
– 50 h (d) 650 C (923 K) – 100 h; (Etchant – 10% oxalic acid).

changes from 8–11FN to 1.11, 0.36 and 0.04. The M23C6 which proceeds till the carbon potential of
microstructures obtained are given in Fig. 12(a)–(c). the surrounding austenite decreases sufficiently.
Of all the 316L SS weld specimens studied, the spec- After that if sufficient chromium is available sigma
imen heat treated at 675 C (948 K) for 200 h forms. The amount of sigma phase increases with
showed the maximum reduction in ductility. After increase in ageing temperature due to increased dif-
exposure to Strauss test solution, the specimen fusivity of sigma forming elements. Initially sigma
failed even before doing the bend test. Fig. 13 shows was present as elongated particles in a discontinu-
the fracture surface and presence of bulky particles ous network. With increase in ageing time and
are seen in the fractograph which are identified as temperature, the network is broken and large sphe-
sigma by EDAX. Similar to modified 316N SS weld roidised particles are formed. ASTM A262 Practice
metal, fracture along interdentritic regions has been E test detects susceptibility to IGC associated with
observed. Intergranular failure along austenite grain M23C6 precipitation and is not intended to detect
boundary was not observed. that due to sigma phase embrittlement. However,
From all the above microstructures it can be in these cases, because of long term high tempera-
summarized that thermal ageing of modified 316L ture ageing, sigma phase contributes to reduction
SS weld metal results in dissolution of delta ferrite in ductility even before exposure to the corrodant
and precipitation of M23C6 along delta ferrite/aus- as per ASTM A262 Practice E test. Under such con-
tenite interfaces as well as sigma phase precipitation ditions failure in ASTM A262 Practice E test has
within delta ferrite. Carbides are formed initially taken place predominantly due to the inherent loss
while isolated sigma particles could be observed in ductility due to sigma phase embrittlement, while
towards the end of transformation. Carbon pro- IGC due to chromium depletion played a minor
motes the removal of chromium in the form of role.
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 77

Fig. 13. Fractograph of modified 316L SS weld specimen aged at

675 C (948 K) for 200 h and subjected to ASTM A262 Practice E
test showing fracture along interdendritic features.

tion behaviour of single phase austenitic stainless

steel and duplex austenite–ferrite stainless steel weld
metals. In the case of single phase austenitic stain-
less steel certain time–temperature combinations
will be sufficient to precipitate chromium-rich
M23C6 carbide but insufficient to rediffuse chro-
mium back to the depleted zone in the austenite
near carbides. This will result in the formation of
envelopes of chromium-depleted zones around
M23C6 carbides. Since the carbides precipitate along
austenite boundaries, the linking of the chromium-
depleted zones provides a continuous path of lower
corrosion resistance along the grain boundaries for
the propagation of IGC.
In the case of duplex weld metal containing aus-
tenite and delta ferrite also, thermal exposure in the
sensitization range results in M23C6 carbide precip-
itation. Since chromium diffusion is much faster in
ferrite than in austenite, the carbides grow faster
in ferrite than in austenite. Consequently, a wide
but shallow chromium depleted zone develops on
the ferrite side and a very deep but narrow chro-
mium depleted zone develops on the austenite side.
Delta ferrite starts transforming to secondary
austenite. As long as the delta ferrite present is in
Fig. 12. Microstructure obtained for modified 316L SS weld
metal (a) 675 C (948 K) – 50 h; (b) 675 C (948 K) – 100 h; (c)
contact with the growing carbide, quicker homoge-
675 C (948 K) – 200 h; (Etchant – 10% oxalic acid). nisation of chromium depletion takes place and
hence the material will be free from sensitization.
However, when delta ferrite transforms more or less
3.3. Microstructural evolution and sensitization completely to M23C6 and secondary austenite, the
behaviour carbide starts drawing chromium from the sur-
rounding austenitic matrix. This leads to formation
Based on the above microstructural studies, the of narrow and steep chromium concentration gradi-
following differences can be inferred in the sensitiza- ents along the austenite grain boundaries and the
78 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

material gets sensitized. When the sensitized mate- Table 4

rial is exposed to corrosive medium corrosion takes Details of the controlled heating–soaking–cooling cycles carried
out for solution-annealing of modified 316N SS weld metal and
place along the interdentritic boundaries leading to results obtained in ASTM A262 Practice E test
failure. This mechanism is clearly illustrated by the
No. Heating-holding-cooling cycles Result obtained in
microstructures shown in Figs. 2–6 and fractograph ASTM Practice
shown in Fig. 7 and is valid in the temperature E test
range in which carbide precipitation is the predom- 1 Room temperature to 1050 C Attack
inant mechanism of transformation of delta ferrite. (1323 K) in 2 h
Time to sensitization in the peak temperature range Hold at 1050 C (1323 K) for 1 h
in the present case is influenced greatly by the nitro- 1050 C (1323 K) to
gen content of the weld metal and the times reported Room temperature @50 C/h
here are much longer than for 316 (base or weld 2 Room temperature to 1050 C Attack
metal compositions) containing less nitrogen. These (1323 K) in 2 h
Hold at 1050 C (1323 K) for 1 h
findings are consistent with the observations of
1050 C (1323 K) to
Nakao et al. [13] for AISI type 308 stainless steel, Room temperature @65 C/h
although the kinetics of transformation are much
3 Room temperature to 1050 C No attack
faster in their case because of very low nitrogen
(1323 K) in 2 h
levels (<0.01%). Hold at 1050 C (1323 K) for 1 h
It is interesting that in some cases, the micro- 1050 C (1323 K) to
structures corresponding to sensitized specimens Room temperature @75 C/h
aged at 650 C (923 K) showed austenite grain
boundaries decorated with carbides. These grain
boundaries are not normally visible in duplex stain- welded components are cooled from 1050 C
less steel weld metal because of the strong attack by (1323 K). The heating rate is not a concern because
the etchant on the interphase boundaries in prefer- the carbides formed during heating part of the heat
ence to these austenite grain boundaries. In the total treatment will be solutionised at 1050 C (1323 K).
absence of delta ferrite, the material essentially Cooling at the rate of 75 C/h can be safely followed
behaves as high carbon base material and chromium without the risk of sensitization. It is well known
carbide precipitation at the grain boundaries has that solution-annealing at 1050 C (1323 K) results
therefore resulted in sensitization. in complete dissolution of delta ferrite present in
For temperatures of 725 C (998 K) and above, the weld metal and subsequent transformation to
transformation is predominantly to sigma phase. austenite. In other words, after solution-annealing
Sigma phase formation takes place by consumption the weld metal behaves as base metal (austenitic
of chromium present in delta ferrite, with very little stainless steel) of higher carbon content. In our ear-
or no contribution of chromium from the surround- lier investigation, sensitization behaviour of austen-
ing austenitic matrix. That is why sigma phase for- itic stainless steel Type 316 having different carbon
mation alone does not result in sensitization of the and nitrogen contents were compared [19]. It was
microstructure. found that for stainless steel containing C-0.054%
and N-0.053% [alloy 1] the CCR was determined
4. Validation of CCR of modified 316N and as 365 C/h, for C-0.043% and N-0.075% [alloy 2]
316L SS weld metals by controlled the CCR was 17 C/h, and for C-0.03% and N-
heating–soaking–cooling trials 0.086% the CCR was as low as 0.43 C/h. These
data clearly indicates that lowering the carbon con-
4.1. Solution-annealing of modified 316N SS weld tent and increasing the nitrogen content results in
metals significant reduction in the kinetics of sensitization.
Since the composition of the weld metal prepared
For modified 316N SS weld metal subjected to using welding electrode containing 0.045–0.055%C
solution-annealing heat treatment (Table 4 – No. and 0.06–0.1%N is expected to lie in between that
1, 2, 3), heating and cooling up to 65 C/h results of alloy 1 and 2, the CCR for solution annealed
in sensitization whereas 75 C/h does not lead to weld metal is also expected to lie between 365 C/h
sensitization. From the above results, it can be seen and 17 C/h and our results are in agreement with
that CCR lies between 65 C/h and 75 C/h when this prediction.
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 79

Table 5
Details of the controlled heating–soaking–cooling cycles carried out for stress-relieving of modified 316N SS weld metal and results
obtained in ASTM A262 Practice E test
No. Heat treatment Heating-holding-cooling cycles Result obtained in ASTM
Practice E test
1 Stress-relieving Room temperature to 850 C (1123 K) @90 C/h No attack
Hold at 850 C (1123 K) for 4 h
850 C (1123 K) to Room temperature @90 C/h
2 Solution-annealing Room temperature to 1050 C (1323 K) in 2 h Attack
Hold at 1050 C (1323 K) for 1 h
1050 C (1323 K) to Room temperature @90 C/h,
@110 C/h, @140 C/h and @200 C/h
Stress-relieving Room temperature to 850 C (1123 K) @90 C/h,
@110 C/h, @140 C/h and @200 C/h
Hold at 850 C (1123 K) for 4 h
850 C (1123 K) to Room temperature @90 C/h,
@110 C/h, @140 C/h and @200 C/h
3 Solution-annealing Room temperature to 1050 C (1323 K) in 2 h Attack
Hold at 1050 C (1323 K) for 1 h
1050 C (1323 K) to Room temperature @90 C/h
Stress-relieving Room temperature to 850 C (1123 K) @90 C/h
Hold at 850 C (1123 K) for 2 h
850 C (1123 K) to Room temperature @90 C/h
4 Solution-annealing Room temperature to 1050 C (1323 K) in 2 h Attack
Hold at 1050 C (1323 K) for 1 h
1050 C (1323 K) to Room temperature @90 C/h
Stress-relieving Room temperature to 750 C(1023 K) @90 C/h
Hold at 750 C(1023 K) for 4 h
750 C(1023 K) to Room temperature @90 C/h

Table 6
Details of controlled heating–soaking–cooling cycles carried out for modified 316L SS weld metal and results obtained in ASTM A262
Practice E test
No. Heat treatment Heating–soaking–cooling cycles Result obtained in ASTM
A262 Practice E test
1 Solution-annealing [Step 1] Room temperature to 1050 C (1323 K) in 2 h No attack
Hold at 1050 C (1323 K) for 1 h
1050 C (1323 K) to Room temperature @120 C/h,
@50 C/h, @30 C/h, @20 C/h and @10 C/h
2 Pre-heating [Step 2] Room temperature to 500 C (773 K) @120 C/h,
@50 C/h, @30 C/h, @20 C/h and @10 C/h
Hold at 500 C (773 K) for 1 h
3 Stress-relieving [Step 3] 500 C (773 K) to 850 C (1123 K) @120 C/h,
@50 C/h, @30 C/h, @20 C/h and @10 C/h
Hold at 850 C (1123 K) for 4 h
850 C (1123 K) to Room temperature @120 C/h,
@50 C/h, @30 C/h, @20 C/h and @10 C/h
4 Pre-heating after Room temperature to 500 C (773 K) @120 C/h,
repair [Step 4] @50 C/h, @30 C/h, @20 C/h and @10 C/h
Hold at 500 C (773 K) for 1 h
5 Stress-relieving after 500 C (773 K) to 850 C (1123 K) @120 C/h,
repair [Step 5] @50 C/h, @30 C/h, @20 C/h and @10 C/h
Hold at 850 C (1123 K) for 4 h
850 C (1123 K) to Room temperature @120 C/h,
@50 C/h, @30 C/h, @ 20 C/h and @10 C/h
80 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

4.2. Stress-relieving 4.2.2. Two step heat treatment for hard-faced

components containing modified 316N SS welds
4.2.1. Single step stress-relieving for modified 316N For the hard-faced components containing mod-
SS weld metals ified 316N SS welds which may be subjected to two
For weld metal subjected to stress-relieving heat step heat treatment involving solution-annealing
treatment (Table 5 – No. 1) heating and cooling at and stress-relieving (Table 5 – No. 2), heating and
the rate of 90 C/h does not result in sensitization, cooling at various rates of viz., 90 C/h, 110 C/h,
since this cooling rate is higher than the CCR 140 C/h and 200 C/h also resulted in severe sensi-
(75 C/h). tization. This is expected because CCR for this

Fig. 14. Microstructure obtained for modified 316L SS weld metal after solution-annealing, stress-relieving and cooling at the rate of (a)
120 C/h; (b) 50 C/h; (c) 30 C/h; (d) 20 C/h; (e) 10 C/h (Etchant – 10% oxalic acid).
 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82 81

composition is 75 C/h (for single phase austenite). The above five steps were simulated as indicated
As explained in Section 4.1, the heating rate to reach in Table 6 with five different heating and cooling
solution-annealing temperature is not a concern rates, viz., 120 C/h, 50 C/h, 30 C/h, 20 C/h and
because soaking dissolves all the carbides which 10 C/h. It has been found that the weld metal does
are nucleated due to the thermal exposure in the not get sensitized even at the rate of 10 C/h. The
sensitization range. However, during cooling, microstructure obtained after the above heat treat-
M23C6 nucleates and grows when passing through ment is presented in Fig. 14(a–e) The fastest cooling
the sensitization range. rate also viz., (120 C/h) has resulted in isolated car-
Subsequent stress-relieving exposes the material bide precipitation (Fig. 14(a)) and isolated sigma
twice to the sensitization range (heating and cool- phase. With progressive decrease in heating and
ing). Therefore totally three times the welded com- cooling rates, the duration of exposure to sensitiza-
ponent is exposed to the sensitization range and tion regime increases nucleation of carbides along
hence three times faster cooling (225 C/h) only will grain boundaries and sigma phase throughout the
avoid sensitization. specimen. Since continuous carbide network and
Since the maximum permissible heating/cooling associated chromium depleted zones are not seen
rate which will retain the desired mechanical prop- in the microstructure it can be concluded that there
erty of the hard face coating is 90 C/h, some more is no risk of sensitization during the stress relieving
attempts were made to modify the heat treatments treatment for the rates presented in Table 6.
either by soaking at 850 C (1123 K) for a lesser
duration viz., 2 h (Table 5 – No. 3) or lower soaking
temperature, viz., 750 C (1023 K) for a longer 5. Conclusions
duration (Table 5 – No. 4) and cooling at 90 C/h
were attempted but all the heat treatment resulted The sensitization behaviour of modified 316N SS
in severe sensitization. and 316L SS weld metal prepared using indige-
nously developed, modified 316N SS electrodes
4.3. Stress-relieving of hard faced components and 316L SS electrodes respectively was established
containing modified 316L SS welds by performing both isothermal and continuous
heating and cooling heat treatment cycles. Suscepti-
From Section 4.2.2 it can be concluded that for bility to IGC was assessed by ASTM A262 Practice
hard faced 316LN SS components containing E test. Microstructural changes taking place during
316L SS welds, indigenously developed modified various thermal exposures were characterized using
316N SS welding electrodes cannot be used because optical and scanning electron microscopy studies
the CCR is very much higher and will certainly lead and electrochemical phase separation and X-ray dif-
to reintroduction of residual stress and cracking of fraction techniques. The mechanism of sensitization
hard face deposit. Hence 316L electrodes with lower was explained based on these studies:
carbon only will enable stress-relieving heat treat-
ment to be performed at a permissible rate so that 1. TTS diagrams were developed for modified 316N
sensitization as well as reintroduction of residual SS weld metal and 316L SS weld metal. Modified
stress can be avoided simultaneously. As explained 316N SS weld metal exhibited sensitization in the
in Section 3.1, CCR for 316L SS weld metal will temperature 525–725 C (798–998 K). The mini-
be about 1 C/h. This prediction is validated by mum time required for sensitization was 25
performing the actual stress-relieving heat treatment minutes. The critical cooling rate above which
envisaged for bottom plate of the grid plate assem- there is no risk of sensitization is 160 C/h (when
bly for PFBR as follows: cooling starts from 725 C (998 K)). The 316L SS
weld metal gets sensitized in the temperature
(1) Solution-annealing at 1050 C (1323 K). range of 625 C (898 K) to 675 C (948 K). The
(2) Pre-heating before hard facing. minimum time required for sensitization at the
(3) Stress-relieving after hard facing. nose temperature (tmin) is about 40 h. From this
(4) Pre-heating before repair (assuming that diagram, CCR above which there is no risk of
repair is required). sensitization, was calculated as 1 C/h. This value
(5) Stress-relieving after repairing the hard face is applicable only when cooling of the weld metal
deposit. starts from 675 C (948 K).
82 N. Parvathavarthini et al. / Journal of Nuclear Materials 355 (2006) 68–82

2. In both the materials, delta ferrite transforms to Group, Dr V.S. Raghunathan, Associate Director,
M23C6 carbides and austenite. In low carbon Materials Characterisation Group, for their keen
weld metal, in addition to carbides, sigma phase interest and support during the course of the inves-
is also observed. In modified 316N SS weld metal tigation. The authors gratefully acknowledge the
failure in ASTM A 262 Practice E is due to assistance provided by Mr M. Sangiah in carrying
corrosion of chromium depleted zone formed out heat treatments and Mrs K. Parimala in carry-
by carbide precipitation whereas in 316L SS weld ing out corrosion tests. The authors thank Mr
metal, the loss in ductility due to sigma phase [Link] for his help in preparing the line
also contributes to the failure in U bend test. drawing.
3. For short ageing times prior to onset of sensitiza-
tion in the critical temperature range, untrans-
formed delta ferrite was present along with References
M23C6 carbides. In the presence of delta ferrite
[1] S.K. Ray, V. Shankar, V. Balasubramaniam, V.K. Sethi,
in close contact, chromium supply for carbide Heat treatment of Austenitic Stainless steel components, in:
precipitation comes from this phase, which pro- Proceedings of the Seminar Materials R&D for PFBR held
tects the surrounding austenite from chromium on 1–2 January 2003, p. 215.
depletion. [2] ASM committee on Heat Treating, Metals Hand Book, 9th
Ed., ASM, Metals Park, OH, 1981, p. 647.
4. Intergranular attack was not observed in speci-
[3] V. Cihal, Intergranular Corrosion of steel and alloysMaterial
mens aged above the critical temperature range Science Monograph, vol. 18, Elsevier, New York, NY, 1984.
(for carbide precipitation), in which transforma- [4] A.J. Sedricks, Corrosion of Stainless Steel, 2nd Ed., John
tion was predominantly to sigma phase. Wiley, NY, 1996.
5. For modified 316N SS weld metal, when cooling [5] E. Folkhard, Welding Metallurgy of Stainless Steel,
Springer-Verlag, Vienna, 1988.
starts from 1050 C (1323 K) (solution-annealing
[6] H.E. Hannien, Int. Met. Rev 3 (1979) 85.
temperature), the critical cooling rate is 75 C/h. [7] E.A. Trillo, L.E. Murr, Acta Mater. 47 (1) (1999) 235.
The difference in the CCR has been attributed to [8] H.D. Solomon, T.M. Devine Jr., in: Proceedings of the
the transformation characteristics of delta ferrite. Conference on Duplex Stainless Steels, American Society for
6. For the stress-relieving heat treatment of hard Metals, Metals Park, OH, 1983, p. 693.
[9] E. Angelini, B. Benedetti, G. Maizza, F. Rossalbino,
faced components containing weld metal fabri-
Corrosion 55 (6) (1999) 606.
cated with modified 316N indigenous electrode, [10] E. Erauzkin, A.M. Irisarri, in: Applications of Stainless
heating/cooling at the rate of 200 C/h also Steels ’92, vol. 2, Stockholm, Sweden, 9–11 June 1992, p.
results in sensitization. Since solution-annealing 870.
followed by stress-relieving at a heating/cooling [11] H. Menendez, T.M. Devine, Corrosion 46 (5) (1990) 410.
[12] C.D. Lundin, Maintenance and Repair Welding in Power
rate above CCR will certainly lead to reintroduc-
Plants, Orlando, FL, USA, 9–11 December 1991, p. 46.
tion of the residual stress and cracking of hard [13] Y. Nakao, K. Nishimoto, M. Ishizaki, Welding Int. 6 (7)
face deposit, welding electrode of such high car- (1992) 523.
bon should not be used wherever hard facing [14] T.P.S. Gill, R.K. Dayal, J.B. Gnanamoorthy, Welding J. 9
and double heat treatment are required. Use of (1979) 3755.
electrodes of lower carbon is expected to solve [15] Recommended Practices for Detecting Susceptibility to
Intergranular corrosion in Stainless Steels, A-262-02; ASTM
the problem. Annual Book, vol. 3.02, ASTM Publications, Philadelphia,
2003.
[16] R.K. Dayal, J.B. Gnanamoorthy, Corrosion 36 (1980) 104.
Acknowledgements [17] T.P.S. Gill, M. Vijayalakshmi, V. Shankar, P. Rodriguez,
Scr. Metall. Mater. 27 (1992) 313.
[18] T.P.S. Gill, V. Shankar, M.G. Pujar, P. Rodriguez, Scr.
The authors are grateful to Dr Baldev Raj, Direc- Metall. Mater. 32 (1995) 1595.
tor, Indira Gandhi Centre for Atomic Research, Dr [19] N. Parvathavarthini, R.K. Dayal, J. Nucl. Mater. 305 (2002)
S.L. Mannan, Director, Metallurgy and Materials 209.