Effect of Different Chromium Additions on the Microstructure

and Mechanical Properties of Multipass Weld Joint of Duplex

Stainless Steel
DONG HOON KANG and HAE WOO LEE

The correlation between the mechanical properties and ferrite volume fraction (approximately
40, 50, and 60 Ferrite Number [FN]) in duplex stainless steel weld metals were investigated by
changing the Cr content in filler wires with a flux-cored arc-welding (FCAW) process. The
interpass temperature was thoroughly maintained under a maximum of 423 K (150 °C), and
the heat input was also sustained at a level under 15 KJ/cm in order to minimize defects. The
microstructure examination demonstrated that the d-ferrite volume fraction in the deposited
metals increased as the Cr/Ni equivalent ratio increased, and consequently, chromium nitride
(Cr2N) precipitation was prone to occur in the ferrite domains due to low solubility of nitrogen
in this phase. Thus, more dislocations are pinned by the precipitates, thereby lowering the
mobility of the dislocations. Not only can this lead to the strength improvement, but also it can
accentuate embrittlement of the weld metal at subzero temperature. Additionally, the solid-
solution strengthening by an increase of Cr and Mo content in austenite phase depending on the
reduction of austenite proportion also made an impact on the increase of the tensile and yield
strength. On the other hand, the impact test (at 293 K, 223 K, and 173 K [20 °C, –50 °C, and
–100 °C]) showed that the specimen containing about 40 to 50 FN had the best result. The
absorbed energy of about 40 to 50 J sufficiently satisfied the requirements for industrial
applications at 223 K (–50 °C), while the ductile-to-brittle transition behavior exhibited in
weldment containing 60 FN. As the test temperature decreased under 223 K (–50 °C), a narrow
and deep dimple was transformed into a wide and shallow dimple, and a significant portion of
the fracture surface was occupied by a flat cleavage facet with river patterns.

DOI: 10.1007/s11661-012-1310-6
Ó The Minerals, Metals & Materials Society and ASM International 2012

I. INTRODUCTION significance in the industry. Large industrial applica-

tions of duplex stainless steels utilize welding as a
DUPLEX stainless steels (DSS) are widely used in manufacturing process. Therefore, it is also very impor-
many environments and operating conditions, such as in tant to maintain an optimum phase balance in weld
the marine, chemical, petrochemical, nuclear, fertilizer, metal after welding. However, welding of DSS results in
and food industries, because of their superior perfor- a disturbed phase balance, and this problem is more
mance in comparison to traditional austenitic stainless significant in flux-cored arc welding (FCAW) processes
steel. DSSs have higher strengths than austenitic stain- because of their low heat input and very fast cooling
less steels, higher toughness levels than ferritic stainless rates.[9] The resultant weld metals contain higher ferrite
steels, good weldability, and high resistance to stress levels and have been reported to have inferior proper-
corrosion cracking. These good properties depend on the ties.[10] In order to restore the toughness of welded
two-phase microstructure that consists of approximately connections, weld filler materials are usually overalloyed
equal volume fractions of c-austenite and d-ferrite with 2 to 4 pct more Ni than in the base material.[11]
phases.[1–4] Since the development of first-generation During recent years, the authors studied the mechanical
DSSs in the 1930s, continuing research effort have been behaviors of DSS welds with varying alloying elements
exerted to improve both mechanical and corrosion and reported that the mechanical properties were
properties, particularly by controlling alloying elements strongly influenced by the change in the shape and
of Cr, Mo, N, W, etc.[5–8] Currently, however, most of volume fraction of austenitic phase with different
the research on DSS has been performed on the wrought amounts of additions. For example, Muthupandi
products, and only a limited number of studies has been et al.[12] examined the effect of nickel and nitrogen
reported on the weld products of DSSs, despite their addition on the microstructure and mechanical proper-
ties of DSS weld metals, and they found that not only
DONG HOON KANG, Master Degree, and HAE WOO LEE, does the introduction of Ni or N not influence the
Professor, are with the Department of Materials Science and hardness of resultant weld metals, but also these
Engineering, Dong-A University, Saha-gu, Busan 604-714, Republic elements improve the impact toughness substantially at
of Korea. Contact e-mails: hwlee@dau.ac.kr; cartoon83@nate.com
Manuscript submitted December 18, 2011. subzero temperature. Park and Lee[13] studied the effect
Article published online August 1, 2012 of nitrogen and heat treatment on the microstructure

4678—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

and tensile properties of 25Cr-7Ni-1.5Mo-3W-xN DSS distortion, and this was performed in the plat position by
castings, and consequently, it was revealed that the a FCAW process using three newly designed types of
increase of ferrite volume fraction decreases the tensile 1.2-mm-diameter filler wires, which were fabricated based
strength and the elongation, and it increases the yield on the AWS 2209TO (1)-1/4 electrode by modifying the
strength of the casting linearly. Some authors[14] have Cr contents in the flux to obtain three different volume
reported that manganese additions increased the tough- fractions. Each groove and backing material was buttered
ness and the ultimate tensile strength of stainless steel, up to 4 mm with the same filler wires in this study to
but it had little effect on the yield strength. Cr is a strong minimize the dilution between the AISI304L austenitic SS
ferrite-stabilizing element and is known to improve the and DSS weld metal, so the buttering layer was not
resistance to corrosion by the formation of a stable considered in the investigation. A schematic diagram of
passive film on the structure of the stainless steel.[15,16] the weldment and the welding parameters is shown in
Despite the beneficial effect of Cr in stainless steel, the Figure 1 and Table I, respectively. The welded specimens
effect of Cr on the mechanical behaviors of DSS welds were sectioned as shown in Figure 2 indicating the
has not been well established. location of the different test specimens, and then the
The aim of this current study was, therefore, to correlation of weld metal microstructure with mechanical
examine the effect of the addition Cr on the microstruc- properties was investigated.
ture and mechanical properties of the weld fusion zone.
Specifically, the investigation focused on a comparison
B. Equivalent Formula
of the low-temperature impact behavior of the DSS weld
with different d/c ratios, based on the SEM micro- The Cr/Ni equivalent ratio was computed to the
graphic and fractographic observations. WRC-1992 formula.[17]
Creq ¼ Cr þ Mo þ 0:7Nb
II. EXPERIMENTAL DETAILS
Nieq ¼ Ni þ 35C þ 20N þ 0:25Cu
A. Test Panel/Filler Wire Preparation
and Welding Process
The test panels, which were 300 mm long 9 100 mm
wide 9 12 mm thick, were butt welded in four layers with
a root gap opening of 3 mm and a total V-groove angle of
34 deg under a strong restraint condition to avoid thermal

Fig. 1—Schematic diagram of the weldment. Fig. 2—Location of test specimens.

Table I. Welding Parameters

No. of Interpass Travel Heat Input

Test No. Layers Current (A) Voltage (V) Temperature (°C) Speed (cm/min) (KJ/cm)
No. 1 1 140 28 19 26 9.0
2 170 29 77 28 10.2
3 180 30 106 34 9.4
4 180 30 127 21 15.3
No. 2 1 140 28 22 30 7.6
2 170 29 72 29 10.0
3 180 30 92 32 9.9
4 180 30 120 21 15.5
No. 3 1 140 28 15 27 8.6
2 170 29 45 28 10.3
3 180 30 77 30 10.8
4 180 30 123 21 15.0
Shielding gas/flow rate: CO2 (100 pct), 20 mL/min, polarity: DCRP(+) electrode extension (mm): 15 to 20.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4679

 Table II. Chemical Composition of the Weld Metal (Weight Percent)

As-Weld Metals C N Si Mn P S Cr Ni Mo Cu Nb Creq Nieq Creq/Nieq

No. 1 0.032 0.112 0.81 0.62 0.024 0.008 21.30 8.71 3.41 0.07 0.003 24.72 12.08 2.02
No. 2 0.029 0.114 0.82 0.59 0.025 0.008 22.03 8.73 3.28 0.07 0.003 25.31 12.03 2.10
No. 3 0.030 0.115 0.9 0.61 0.024 0.007 22.82 8.70 3.37 0.07 0.004 26.19 12.06 2.18

Table III. FN of Weld Metals

Ferrite Scope

Specimen Face Center Root Average WRC-1992

No. 1 50 40 47 45 42
No. 2 60 50 56 55 51
No. 3 75 60 70 68 62

7.5 9 10 9 55 mm) as shown in Figure 4; they were

tested at room temperature, 223 K and 173 K (–50 °C
and –100 °C) according to ASTM E23-05. The data
were then converted to a standard size (10 by 10 mm)
Fig. 3—FNs of specimens on the WRC-1992 diagram. according to the A 370-05 acceptance criteria table. The
strength was measured using a universal testing machine
The chemical analysis for the three welds, which were (AG-25TG; Shimadzu, Tokyo, Japan). Round-bar-type
named as No. 1, No. 2, and No. 3, respectively, was tensile test specimens (ASTM E8-04) were prepared
performed using a hybrid optic spectrometer (Metal- from the fully deposited metal in a direction parallel to
LAB75/80J, GNR, srl, Rome, Italy), and the results are the welding direction. The hardness tests were per-
shown in Table II. formed on a Vickers hardness tester FM700 applying a
maximum load of 1 kg for dwell times of at least
10 seconds. More detailed studies of the hardness
C. Ferrite Number (FN) Prediction alternation were performed using Vickers HV0.05
Ferrite measurements were also used to produce method. Thirty measurements of each phase in weld
quantitative relationships between alloy composition metal at a depth of 7 mm from the surface, and the
and ferrite volume fraction, and the results were then results were averaged for subsequent analysis.
plotted on the WRC-1992 weld constitution diagram as
shown in Figure 3. Good agreement was shown between
the solidification mode of primary ferrite alloys and III. RESULTS AND DISCUSSION
those predicted by the WRC-1992 diagram. A ferrite-
scope (MP30E-S, Fischer Scientific, Schwerte, Ger- A. Macrostructure/Microstructure
many) was used in the nondestructive evaluation to Figure 5 shows a good external FCAW macrostruc-
observe the ferrite content on the weldments in terms of ture (12 mm, four pass). The geometry of the weld joint
the FN, and the details are listed in Table III. was fully penetrated and deposited. Additionally, no
visible porosity or defect was observed on the weldment.
D. Microstructure/Fracture Analysis As shown in Figure 6, the transformation sequence for
duplex stainless steel is as follows[2]:
The specimens were ground with 2000-grit emery paper
and polished successively with 3-lm diamond pastes,
rinsed with water, subjected to ultrasonic cleaning, and L ! L þ F ! F ! F þ A ½F mode
dried in air. Then they were immersed in boiling Muraka-
mi’s reagent (10 g K3Fe(CN)6, 10 g KOH, and 100 mL The nature of the ferrite-to-austenite transformation in
H2O heated to 373 K (100 °C) for 1 hour) in order to dependent on both the chemical composition and its total
measure its volume fraction.[18] For observation of the thermal history.[10,19] Welds, however, and their heat-
microstructure and fractograph, scanning electron micro- affected zones (HAZs) are rapidly cooled from temper-
scopy with an energy dispersive spectroscopy (SEM-EDS) atures near the ferrite solvus, so there is a tendency for
(JSM-6700f; JEOL Ltd., Tokyo, Japan) detector was used. appreciably more ferrite in the weld metal and HAZ of a
duplex stainless steel than there is in the base metal. It is
considered that, therefore, the chemical composition can
E. Impact/Tensile/Hardness Test be even more crucial factor to maintain the ferrite-
The specimens for the Charpy-V impact test were austenite balance in the weld metal. For this reason, each
machined (subsize type A – V notch, dimensions: of the specimens was manufactured by controlling the

4680—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 4—Dimensions of tensile and impact specimens. (a) Charpy (simple-beam) subsize (Type A, V-notch) impact test specimens. (b) Round
tension test specimen (threaded end).

Fig. 5—Macrography of the four-pass weld joint.

Cr/Ni equivalent ratio while maintaining the heat input

constantly, for the purpose of changing the ferrite-
austenite volume fraction of the weldment.[20,21]
In the weld metal, ferrite solidification involves
epitaxial growth from the parent material at the fusion
boundary. The initial dendrite growth is oriented in
relation to the thermal gradient, and this produces a Fig. 6—Relationship of solidification type to the pseudobinary phase
columnar ferritic structure. The initial nucleation and diagram.
growth process of the austenite phase occurs intergra-
nularly; then, it completes the coverage of the ferrite optical microstructure of the weld metal, the dark part
grain boundaries. Additional austenite may form as represents ferrite while the light part represents austen-
Widmannstätten side plates off the grain boundary ite. The microstructure of Figure 7(a) is composed of
austenite or it may form intragranularly within the about 40 FN, Figure 7(b) is composed of about 50 FN,
ferrite grains.[10] As can be seen in Figure 7, in the and Figure 7(c) is composed of about 60 FN by

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4681

 Fig. 8—X-ray diffraction results of specimen No. 3 (as-weld).

not mean they do not exist. They may be present in

concentrations of less than 3 pct, which is the lowest
detection limit, or there are only atom clusters in the
initial stages of formation of a precipitate coherent with
the matrix. According to Ramirez et al.,[22] these
precipitates are believed to be Cr2N and secondary
austenite (c2). The formation of these Cr2N within the
ferritized microstructure is aided by this rapid cooling
from temperatures above 1373 K (1100 °C), which
creates a supersaturated solid solution, primarily of Cr
and N, in the interior of the ferrite grains, resulting in a
competition between chromium nitride and austenite
precipitation. These chromium nitrides precipitate from
the ferrite by nucleation and growth, following a kinetic
‘‘C’’ curve. Nucleation takes place at dislocation,
inclusions, and grain boundaries.[23] This chromium
nitride precipitation has been shown to severely impair
the corrosion resistance and the toughness of the
DSSs.[23–25] When the material in this metastable con-
dition is reheated, as in multipass welding, the most
apparent changes in the microstructure are the dissolu-
tion of the intragranular nitrides and the precipitation of
secondary austenite (c2). According to these stud-
ies,[26,27] the nitrogen liberated by the nitride dissolution,
in the temperature range of 1273 K to 1473 K (1000 °C
to 1200 °C), facilitates the c2 nucleation. It is also
well known that the c2 precipitation improves the
toughness of the material.[26] Figure 9 presents an
SEM micrograph of specimen No. 3 (as-weld). Needle-
like Cr2N precipitates (arrow marks) were found in the
Fig. 7—Optical micrograph of the weld metal: dark etched regions ferrite grains or at the a/a subgrain boundaries. These
(ferrite), light etched region (austenite). (a) No. 1, 42 FN; (b) No. 2,
51 FN; and (c) No. 3, 62 FN.
particles had an average thickness of 0.05 to 0.2 lm and
a length of 0.3 to 1 lm.

reference of WRC-1992 formula. In addition to the

changes in the austenite morphologies present in the B. Hardness Test
fusion zone, some precipitates could also be observed The hardness values were recorded on a transverse
within the ferrite grains. section of the deposited weld metal at a depth of 7 mm
Figure 8 indicates the X-ray diffraction results of from the surface using a Vickers hardness testing
specimen No. 3. These figures show only peaks corre- machine, and the results are presented in graphical
sponding to the reflections of ferrite and austenite form as shown in Figure 10(a). Comparing the three
crystalline planes; no secondary phase was identified by specimens, No. 3 had the highest hardness value,
this technique. The fact that the X-ray test did not detect followed in order by specimens No. 2 and No. 1.
other precipitates in the material’s microstructure does Figure 10(b) shows the variation of hardness measured

4682—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 9—Nitride colony in the interior of a ferrite grain in specimen
No. 3: (a) secondary electron SEM image, (b) TEM micrographs of

Cr2N precipitates: bright-field image and selected-area diffraction
pattern z ¼ ½111a ==½0110Cr2 N .

for ferrite and austenite phase in each specimens. The

hardness in both phases increased proportionally within
similar value with the increase of Cr content in the filler
wire. For example, the indentation size on both the
ferrite and austenite phases in the specimen No. 2 are
nearly the same as the hardness of the both phases are in
the same range, i.e., 270 to 285 HV0.05 (Figure 10(c)).
This means that differences in the ferrite–austenite ratio
have little effect on the hardness of the material.
In general, it has been reported that the ferrite and
austenite phases do not differ much in composition Fig. 10—Hardness testing results. (a) Hardness traverses along 7-mm
because the substitutional elements do not have time to distance from top surface in FCAW process. (b) Variation of hard-
partition significantly during DSS welding and with ness measured for ferrite and austenite phase in each specimens. (c)
interstitial nitrogen; only the austenite amount varies Vickers microhardness indentation on ferrite and austenite phase of
but not its hardness.[10] But the findings show a regular specimen No. 2.
difference between the chemical composition of d and
c phase in spite of the fusion zone. Table IV shows the SEM-EDS. The d-ferrite phase contains a few more Cr
average chemical composition of major alloying and Mo, while the Ni and N are partitioned preferen-
elements in the ferrite and austenite phase analyzed by tially to the austenite phase. Although the ferrite volume

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4683

 Table IV. Average Major Alloying Element Content of Ferrite and Austenite Phases

Phase Volume Cr Ni Mo N P Cu
No. 1 Ferrite 35 pct 23.49 8.24 3.99 0.08 0.11 0.05
Austenite 65 pct 20.11 9.81 2.81 0.27 0.03 0.38
No. 2 Ferrite 43 pct 23.70 7.91 3.68 0.06 0.12 0.04
Austenite 57 pct 21.53 10.02 2.98 0.29 0.04 0.42
No. 3 Ferrite 51 pct 23.81 7.72 3.42 0.06 0.13 0.05
Austenite 49 pct 21.87 10.31 3.07 0.30 0.03 0.45

fraction is high in specimen No. 3, it should be noted

that Cr content slightly increased in d-ferrite due to an
increase of Cr content in the filler wire, whereas Cr and
Mo are incremented sequentially for each c-austenite
depending on the decrease of the austenite volume
fraction. Also, there is no observable change in nitrogen.
It is, therefore, also estimated that substitutional effect
by Cr and Mo is more influential than interstitial solid-
solution strengthening by nitrogen in austenite phase.
On the other hand, the hardness of ferrite phase has
risen slightly despite the reduction in Mo content. The
possibility of chromium nitride formation during welding
has been indicated in several studies.[28,29] If the ferrite
content is high, such as in the weld metal and HAZ under
Fig. 11—The result of mechanical strength test.
rapid cooling conditions, then an intense nitride precip-
itation reaction occurs upon cooling since the solubility
limit of the ferrite is exceeded and the nitrogen has variation of phase ratio by Cr content leads to significant
insufficient time to partition to the austenite. The change in the tensile properties of DSS. Li et al.[8]
presence of chromium nitride particles within a ferrite noted that the increase in c-phase as a result of increase in
grain can be clearly seen as shown in Figure 9. Since Mn and N enhances the ultimate tensile strength and
sigma (r) and chi (v) phases were not observed in the weld ductility. Park and Lee[13] also reported that an increase
metal, the higher hardness value of ferrite phase can be of ferrite volume fraction decreases the tensile strength,
attributed to Cr2N precipitation. In a word, it can be and the elongation and increases the yield strength of the
deduced that the presence of Cr2N play an important role DSS castings almost linearly with decreasing N content.
as a pinning site limiting the movement of mobile It is well known that in DSS, most of the nitrogen is
dislocation. The frequently observed accumulations and dissolved in austenite, which makes austenite the stron-
alignment of chromium nitrides at ferrite grain bound- ger phase due to interstitial solid-solution strengthening
aries correspond to the high dislocation density of those by nitrogen.[31] In this study, on the other hand, it is
boundaries and can be explained through the diffusion interesting to note that both the YS and UTS were
impediment of nitrogen at these dislocations during rapid increased along with increasing the volume fraction of
cooling. The partial annealing zone (PAZ) adjacent to the ferrite. As mentioned above, this result was caused by
base metal and the base metal ruled out profile analysis three crucial factors: (1) solid-solution strengthening by
due to the buttering layer diluted with the AISI 304L. an increase of Cr and Mo content in austenite phase,
(2) stacking of mobile dislocation by the precipitates in
ferrite phase, (3) the difference of the Peierls stress
C. Tensile Behavior between the body-centered cubic (bcc; ferrite) and face-
The tensile properties obtained after the tensile tests centered cubic (fcc; austenite) materials.[32] This is further
for each specimen are shown in Figure 11 and compared discussed in Section III–D.
with those of SAF 2207. Each data point represents the Figure 12 illustrates the damage process: cleavage
average of at least three test results. The results show that crack nucleation in the ferrite and subsequent growth in
as the chromium in AWS 2209 filler wire increases, there the austenite that creates cavities. Final rupture is
is increase in both yield strength (YS) and ultimate tensile produced by the coalescence of these cavities. Although
strength (UTS) as compared with those of SAF 2207. rupture is initiated by cleavage, the overall fracture
The YS has increased by 31 pct and 43 pct and UTS has mechanism remains ductile.
also increased by 9 pct and 14 pct. The elongation of the
weld metal normally has a lower value than that of the
base metal.[30] However, the elongation of No. 1 and No. D. Impact Toughness
2 can reach a level that is about 25 pct of the value of the The impact toughness values estimated by Charpy
base metal due to an increased 9 wt pct Ni content in impact testing both at 293 K, 223 K, and 173 K (20 °C,
filler wire, whereas that of No. 3 reduced by increasing –50 °C, and –100 °C) for the FCAW metals are shown
the ferrite volume fraction. It is obvious that the in Figure 13. As a result, the absorbed energy of

4684—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

 (38 FN), apparently due to a change in solidification
mode causing segregation and precipitation of interme-
tallic phase.[33] But these results (Figure 13) mean that
weld metals must be maintained below 61 FN (50 pct)
in order to restore toughness, although the filler wire
with 2 to 4 pct more Ni than in the base material is used.
The ductile-to-brittle transition temperature (DBTT)
was normally determined at half the value of the total
impact energy against the test temperature. Then, the
values of DBTT are probably expected to be existed just
below 223 K (–50 °C).
At subzero temperature, welds typically exhibit higher
strength and lower toughness than their base metal
counterparts. Chan[34] reported that the inferior weld
metals toughness is associated with high nonmetallic
inclusion and delta ferrite content, as well as higher
strength level, and the d-ferrite phase undergoes a ductile-
to-brittle transition at or below 223 K (–50 °C). On the
other hand, according to Lee et al.,[35] the austenite
stainless steel weld metal fracture surface displayed
Fig. 12—Surface crack nucleation in the ferrite and subsequent void ductile dimple rupture features at room temperature
growth during a tensile test. and even at 77 K (–196 °C). It can then be concluded that
the austenite matrix has restricted the cleavage fracture
mechanism that could have been formed around the
deformed delta ferrite phase. In conclusion, the observed
ductile-to-brittle transition behavior of the duplex stain-
less steel weld metals can be associated with the high
ferrite proportion in their microstructure.
Since ferrite is of a bcc structure, its yield strength
(which is a function of temperature) increases as the
temperature is lowered due to an increased lattice friction
stress and pinning of mobile dislocations with interstitial
atoms. On the other hand, the cleavage fracture stress of
ferrite is not a function of temperature and is only varied
by microstructural parameters as grain size and disloca-
tion density.[36] Additionally, it also may be explained
with the Peierls stress, which was studied with regard to
movement of the dislocation by temperature variation.
The critical feature of the Peierls stress is that the yield
strength is closely connected with the temperature; that is,
the Peierls stress dramatically increases from the slow-
Fig. 13—Results of Charpy V-notch impact tests for weld metal in down of the dislocation as the temperature decreased. The
FCAW process. yield strength of the material increases as a result. This
means a decrease of the absorbed energy at a lower
specimen No. 1 was the highest, followed in order by temperature due to an inverse relationship between the
specimens No. 2 and No. 3. The toughness values of yield strength and the toughness in the material.[37–39]
No. 1 (42 FN) and No. 2 (51 FN) show that although According to Kacar,[40] at room temperature, the cleav-
differences exist in the ferrite content among the weld age fracture stress of ferrite is much higher than its yield
metals, they all exhibit nearly the same toughness until strength, and consequently, plastic deformation prevails
223 K (–50 °C) impact tests, and these weld metals and ductile behavior is verified. As temperature decreases,
sufficiently meet the subzero impact toughness require- and at a certain low temperatures, the yield strength of
ments for industrial applications, whereas No. 3 (61FN) ferrite becomes higher than its cleavage fracture stress. At
showed a significant drop. On the other hand, the results this stage, a transition from ductile fracture through
obtained from the test conducted at 173 K (–100 °C) plastic deformation to brittle fracture by cleavage takes
showed that the ferrite–austenite ratio has a significant place. Since the austenite phase of the duplex stainless
role to play in declining the low-temperature toughness. steel is of an fcc structure, neither of its yield nor fracture
It is well known that the influence of the ferrite stress would be a function of temperature.[41,42] A possible
content on the absorbed energy was reported to be additional reason is that the ferrite matrix was highly
negligible up to about 50 pct to 60 pct (60-85 FN) modulated, and all the dislocations were heavily jogged
ferrite,[10] whereas there is a clear negative influence at such that no straight dislocations can be observed.
higher ferrite levels. Furthermore, there may be a Besides, there were Cr2N precipitates situated in the
negative effect due to ferrite contents below 35 pct matrix or dislocation lines. This evidence suggested that

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4685

as the ferrite phase increases, more dislocations are
pinned by the precipitates on the dislocation and the
modulated matrix, thereby lowering the mobility of the
dislocations. This pinning behavior is expected to lead to
the embrittlement of the weld metals.

E. Fractographic Observation
Figure 14 shows detailed fracture morphologies as
obtained under an SEM of the fracture surfaces from
Charpy impact specimens varied with ferrite number. In
the as-welded condition, the three specimens showed the
similar tendency for fracture morphologies at three
different temperatures (293 K, 223 K, and 173 K
[20 °C, –50 °C, and –100 °C]), respectively. Figure 14(a)
exhibits a ductile fracture mechanism characterized by a
linked series of the structure that was composed of deep,
narrow, and equiaxed dimples. A mixed mode of ductile
and brittle fracture was observed after a 223 K (–50 °C)
impact test (Figure 14(b)). The fracture appearance
reveals dimples along with the occasional area of
quasi-cleavage. As for the rest, the continuous dimple
link snapped due to the existence of sparse quasi-
cleavage areas. On the other hand, a significant portion
of the fracture surface was occupied by a flat cleavage
facet that had river patterns (as shown in Figure 14(c)),
suggesting that the brittle fracture mechanism has
become dominant. The whole appearance of the dimple
assumed the form of a shallow, stretched, and quantal
shape.[43] Notably, river pattern areas coexist with small
shallow dimples, which could be caused by the fracture
of secondary austenite, and the tearing edges existed
amidst the dimple rupture, which resulted in the break
off of a stream of dimples such that the ligaments of
austenite are sheared between the cleavage planes.[44]
Using an energy-dispersive X-ray analyzer equipped
in the scanning electron microscope, a quantitative
determination of chemistry was carried out on both
cleavage of quasi-cleavage facets and dimple zones,
respectively. The result of over 20 repeated tests were
averaged to achieve 23.9 wt pct Cr and 7.7 wt pct Ni
contents on the cleavage or quasi-cleavage facets,
whereas the composition of average 20.3 wt pct Cr
and 10.02 wt pct Ni was obtained on the dimple zone,
for the No. 2 specimen. These results revealed that
cleavage or quasi-cleavage occurred in the ferrite phase,
whereas a ductile fracture with microvoid coalescence
occurred in the austenitic phase. Viewed with respect to
the overall phenomenon, such as above, the parameter
that could contribute to the deformation mechanism of
duplex stainless steel weldments, particularly at a
subzero temperatures, is predicted to result from the
Fig. 14—SEM micrograph of the fracture surfaces of No. 3 speci-
amount of d-ferrite phase contained in the microstruc- men: (a) 293 K (20 °C), (b) 223 K (–50 °C), and (c) 173 K (–100 °C).
tures. And it is also apparent that a drastic drop in
toughness is connected with the embrittlement and
cleavage of d-ferrite.[45] and mechanical properties of DSS weld metals were
investigated. The study results can be summarized as
follows:
IV. CONCLUSIONS
1. The delta ferrite increased as the addition of Cr in
The effect of the ferrite volume fraction by changing the deposited metals increased, and consequently,
chromium content of filler wire on the microstructure chromium nitride (Cr2N) precipitation was prone to

4686—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

 occur in the ferrite domains due to low solubility of 10. R.N Gunn: Duplex Stainless Steel, Woodhead Publishing Ltd.,
nitrogen in this phase. Thus, more dislocations are Abington, VA, 1997, pp. 110–19.
11. J.R. Davis: Stainless Steels (ASM Specialty Handbook), ASM
pinned by the precipitates, thereby lowering the International, Materials Park, OH, 1994.
mobility of the dislocations. This leads not only to 12. V. Muthupandi, P. Bala Srinivasan, V. Shankar, S.K. Seshadri,
the strength improvement but also to embrittlement and S. Sundaresan: Mater. Lett., 2005, vol. 59, pp. 2305–09.
of the weld metal at subzero temperatures. 13. Y.H. Park and Z.H. Lee: Mater. Sci. Eng. A, 2001, vol. 297,
pp. 78–84.
2. The frequently observed accumulations and align- 14. M. Kemp, A. Van Bennekom, and F.P.A. Robinson: Mater. Sci.
ment of chromium nitrides at ferrite subgrain Eng. A, 1995, vol. 199, pp. 183–94.
boundaries correspond to the high dislocation den- 15. J. Sakai, I. Matsushima, Y. Kamemura, M. Tanimura, and T.
sity of those boundaries and can be explained Osuka: Duplex Stainless Steels, R.A. Lula, ed., ASM Interna-
through the diffusion impediment of nitrogen at tional, Materials Park, OH, 1983, pp. 211–31.
16. R.A. Perren, T. Suter, C. Solenthaler, G. Gullo, P.J. Uggowitzer,
these dislocations during rapid cooling. H. Böhni, and M.O. Speidel: Corros. Sci., 2001, vol. 43, pp. 727–
3. The hardness of austenite phase was largely in- 45.
creased by substitutional effect by Cr and Mo 17. D.J. Kotecki and T.A. Sievert: Weld. J., 1992, vol. 71, pp. 171–78.
rather than interstitial solid-solution strengthening 18. D.E. Nelson and W.A. Baeslack, III: Mater. Charact., 1997,
vol. 39, pp. 467–77.
by nitrogen, as the volume fraction of austenite de- 19. R. Badji, M. Bouabdallah, B. Bacroix, C. Kahloun, B. Belkessa,
creased. That of the ferrite phase has risen linearly and H. Maza: Mater. Charact., 2008, vol. 59, pp. 447–53.
despite the reduction in Mo content, on the other 20. D.J. Lee, J.C. Byun, J.H. Sung, and H.W. Lee: Mater. Sci. Eng. A,
hand, due to the precipitation hardening of Cr2N. 2009, vol. 513, p. 154.
This phenomenon is also affecting the strength 21. Y.H. Kim, D.J. Lee, J.C. Byun, K.H. Jung, J.I. Kim, H.J. Lee,
Y.T. Shin, S.H. Kim, and H.W. Lee: Mater. Des., 2011, vol. 32,
properties of DSS weld metals. pp. 330–36.
4. The absorbed energy of about 40 to 50 FN suffi- 22. A.J. Ramirez, J.C. Lippold, and S.D. Brandi: Metall. Mater.
ciently satisfied the requirements (34 J at 253 K Trans. A, 2003, vol. 34A, pp. 1575–97.
[–20 °C]) for industrial applications at 223 K 23. S.D. Brandi and A.J. Ramirez: Duplex Stainless Steels ’97, Proc.
Conf., KCI, Maastricht, the Netherlands, 1997, pp. 405–10.
(–50 °C), while the ductile-to-brittle transition 24. V.S. Moura, L.D. Limda, J.M. Pardal, A.Y. Kina, R.R.A.
behavior exhibited in weldment containing 60 FN. Corte, and S.S.M. Tavares: Mater. Charact., 2008, vol. 59,
As the test temperature decreased less than 223 K pp. 1127–32.
(–50 °C), a narrow and deep dimple was trans- 25. B. Deng, Z. Wang, Y. Jiang, T. Sun, J. Xu, and J. Li: Corros. Sci.,
formed into a wide and shallow dimple. A signifi- 2009, vol. 51, pp. 2969–75.
26. J.C. Lippold and A.M. Al-Rumaih: Duplex Stainless Steels ’97,
cant portion of the fracture surface was occupied Proc. Conf., KCI, Maastricht, the Netherlands, 1997, pp. 405–
by a flat cleavage facet with river patterns. 10.
27. S.D. Brandi, J.C. Lippold, and W. Lin: Duplex Stainless Steels ’97,
Proc. Conf., KCI, Maastricht, the Netherlands, 1997, pp. 411–18.
ACKNOWLEDGMENT 28. T. Ogawa and T. Koseki: Weld. J., 1989, vol. 68, pp. 181–91.
29. T.G. Gooch: ASM Conference Proc. on Duplex Stainless Steels,
This research was supported by Basic Science Research 1982, pp. 573–602.
30. J.D. Kordatos, G. Fourlaris, and G. Papadimitiou: Scripta Ma-
Program through the National Research Foundation of ter., 2001, vol. 44, pp. 401–08.
Korea (NRF) funded by the Ministry of Education, Sci- 31. J.W. Simmons: Mater. Sci. Eng. A, 1996, vol. 207, p. 159.
ence and Technology (Grant 20100021533). 32. F.R.N. Nabarro: Mater. Sci. Eng. A, 1989, vol. 113, pp. 315–26.
33. S. Pak and L. Karlsson: Scand. J. Metall., 1990, vol. 19, pp. 19–23.
34. J.W. Chan: Adv. Cryog. Eng., 1992, vol. 55, pp. 34–38.
35. D.J. Lee, J.C. Byun, J.H. Sung, and H.W. Lee: Mater. Sci. Eng. A,
REFERENCES 2009, vols. 513–514, pp. 154–59.
1. I. Varol and W.A. Baeslack, III: Mater. Charact., 1997, vol. 39, 36. G. Han, S. Fukuyama, and K. Yokogawa: Mater. Sci. Technol.,
pp. 553–73. 1999, vol. 15, pp. 905–09.
2. J.C. Lippold and D.J. Kotecki: Welding Metallurgy and Welda- 37. R. Peierls: Proc. Phys. Soc., 1940, vol. 52, pp. 34–37.
bility of Stainless Steels, 1st ed., Hoboken, NJ, 2005, pp. 230–34. 38. G. Schoeck: Mater. Sci. Eng. A, 2005, vol. 400, pp. 7–17.
3. H. Ericksson and S. Bernhardsson: Corros. Sci., 1991, vol. 47, 39. B. Deng, Z. Wang, Y. Jiang, T. Sun, J. Xu, and J. Li: Corros. Sci.,
pp. 719–25. 2009, vol. 51, pp. 2969–75.
4. J. Oredsson and S. Bernhardsson: Mater. Perf., 1983, vol. 22, 40. R. Kacar: Mater. Des., 2004, vol. 25, pp. 1–9.
pp. 35–42. 41. R. Kacar and M. Acarer: Mater. Sci. Eng. A, 2003, vol. 363,
5. J.I. Son, S.S. Kim, J.H. Lee, and B.H. Choi: Metall. Mater. Trans. p. 290.
A, 2003, vol. 34A, pp. 1617–25. 42. H. Sieurin and R. Sandstom: Eng. Fract. Mech., 2006, vol. 73,
6. Y.H. Jang, S.S. Kim, and J.H. Lee: Mater. Sci. Eng. A, 2005, p. 377.
vol. 396, pp. 302–10. 43. Ceramics Metals and Information Center: SEM/TEM Fractogra-
7. S.B. Kim, K.W. Paik, and Y.K. Kim: Mater. Sci. Eng. A, 1998, phy Handbook, Metals and Ceramics Information Center, Hun-
vol. 247, pp. 67–74. tington, CA, 1975.
8. J. Li, Y. Xu, X. Xiao, J. Zhao, L. Jiang, and J. Hu: Mater. Sci. 44. S. Bonnet, J. Bourgoin, J. Champredonde, D. Guttmann, and M.
Eng. A, 2009, vol. 527, pp. 245–51. Guttmann: Mater. Sci. Technol., 1990, vol. 6, pp. 221–29.
9. S. Kou: Welding Metallurgy, 2nd ed., Wiley, Hoboken, NJ, 2003, 45. K. Chandra, R. Singhal, V. Kain, and V.S. Raja: Mater. Sci. Eng.
pp. 3–29. A, 2010, vol. 527, pp. 3904–12.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4687