Materials Science & Engineering A 602 (2014) 49–57

Contents lists available at ScienceDirect

Materials Science & Engineering A

journal homepage: www.elsevier.com/locate/msea

Influence of nickel on the toughness of lean duplex stainless

steel welds
Johan Pilhagen n, Rolf Sandström
Department of Materials Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden

art ic l e i nf o a b s t r a c t

Article history: Three weldments with the nickel contents 1.3, 4.9 and 6.0 wt% were made from 30 mm LDX 2101s
Received 21 October 2013 plates. The weldments were subjected to tensile, impact and fracture toughness testing. The aim was to
Received in revised form evaluate the susceptibility for brittle failure in the weld metal at sub-zero temperatures (1C). The amount
29 January 2014
of ferrite was higher for the 1.3 wt% nickel weldment compared to the other two which had similar phase
Accepted 30 January 2014
Available online 6 February 2014
composition and mean free ferrite distance. The result from the tensile testing showed that for the
weldment with the highest nickel content the ductility remained unchanged with decreasing tempera-
Keywords: ture while the other two weldments became less ductile with decreasing temperature. J-integral based
Duplex stainless steel fracture toughness testing showed a significant difference in the susceptibility for brittle failure with
Weld metal
higher values for the weldment with 6 wt% nickel than for the others with lower nickel content.
Impact toughness
& 2014 Elsevier B.V. All rights reserved.
Fracture toughness
Nickel

1. Introduction are therefore usually overalloyed with nickel [8] to promote

Wrought duplex stainless steels (DSS) consist of ferrite and
austenite, usually in equal proportions. The main alloying ele-
ments are chromium, molybdenum, manganese, nickel and nitro-
gen. The first two elements are ferrite stabilizing while the three
later are austenite stabilizing [1]. For the DSS to have high
mechanical strength, good toughness, corrosion resistance and
weldability a fine balancing between the different alloying ele-
ments is needed.
Traditionally these alloys have been used in the offshore
industry, the pulp and paper industry and for pressure vessels
[1]. Nowadays there is an increasing use of duplex stainless steel in
other structural applications due to the maintenance cost savings
that come with the corrosion resistance [2]. In particular, lean
duplex grades, with lower nickel and molybdenum content, give a
competitive cost and are anticipated to be increasingly used for
structural applications [3].
Due to the ferrite content the duplex stainless steels exhibit
ductile to brittle transitions at sub-zero temperatures (1C). To
lower the ductile to brittle transition temperature (DBTT) either
the austenite content can be increased or the toughness of the
ferrite itself can be raised. Both of these parameters are affected by
the nickel content [4–7] which makes the nickel content impor-
tant for duplex stainless steels toughness. Welding consumables
n
Corresponding author. Tel.: þ 46 8 7906252.
E-mail address: pilhagen@kth.se (J. Pilhagen).
austenite formation in the weldments. Fracture toughness mea-
surements between
110 and
40 1C on the weld metal of the
lean duplex LDX 2101s resulted in satisfactory toughness [9,10].
The nickel content in these weldments was in the 8–9 wt% range
with  55% austenite.
The purpose of the present work was to evaluate the influence
of nickel on the fracture toughness of welded LDX 2101s duplex
stainless steel. Three types of weldments were produced with the
aim of having similar microstructure and phase composition but
with different nickel contents. The weldments were subjected to
tensile, impact and fracture toughness testing between room
temperature and
60 1C.
2. Material and welding
The material used in this work was commercially produced
duplex stainless steel LDX 2101s (EN 1.4162, UNS S32101) deliv-
ered by Outokumpu Stainless AB. The material was hot-rolled to
the desired plate thickness of 30 mm followed by solution treating
at 1100 1C and water quenching. The chemical composition for the
plate and filler metals can be found in Table 1.
The weldments were produced in X-joint configuration where
the weldment was parallel to the rolling direction of the parent
plates (T–L orientation). A total of 18 beads of filler metal were
used and the temperature of the plate was held below 150 1C at all
times. The welding parameters can be found in Table 2. The target
for the nickel content in the weldments was 1.5, 5 and 7 wt%.

http://dx.doi.org/10.1016/j.msea.2014.01.093
0921-5093 & 2014 Elsevier B.V. All rights reserved.
50 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57

Table 1
Chemical composition (wt%) of the plate and filler metals. Measured with X-ray fluorescence and combustion analysis.

C Si Mn P S Cr Ni Mo N

LDX 2101s 0.028 0.65 4.95 0.024 0.001 21.5 1.59 0.26 0.22
Autogenous experimental filler, Ø 2.4 mm 0.021 0.74 4.95 0.017 0.001 20.62 1.28 0.10 0.24
Standard commercial filler 23 7 NL, Ø 1.2 mm 0.020 0.50 0.71 0.018 0 23.13 7.30 0.25 0.12

Table 2
Welding parameters for the three weldments.

Root gap, mm Land, mm Groove angle, deg Heat input, kJ/mm Current, A Voltage, V Welding speed, cm/min

1.3 Ni 0 3.5 901 1.1–1.3 245–260 30.9 35–42

5 Ni 0 3.5 901 (1.4) 1.3–1.5 (350) 230–255 (31.9) 32.9 (50) 35–40
6 Ni 2 3 901 (0.85–1.2) 1.4 (200–236) 350 (25.3–29.9) 32 (27–36.5) 50

n
Numbers in parenthesis are for the root when different filler metals or welding methods were used.

Table 3
EDS analysis of the weldments showing the average value. Min–max values in parentheses. Nitrogen and oxygen content measured by combustion analysis.

Fe, wt% Cr, wt% Mn, wt% Ni, wt% N, wt% O, ppm

1.3 Ni 69.56 (67.84–71.94) 22.01 (21.33–23.04) 4.01 (3.53–4.55) 1.34 (1.01–1.71) 0.184 371
5 Ni 69.20 (68.27–70.17) 22.06 (21.27–22.84) 3.04 (2.00–3.73) 4.93 (3.87–5.92) 0.178 348
6 Ni 66.89 (64.78–69.35) 22.92 (21.76–24.14) 1.47 (0.78–2.31) 6.02 (5.30–6.63) 0.137 294

Table 4
Phase composition of ferrite in the weldments. Automatic image analysis based on
ASTM E1245 [11].
α-content, vol% STDEV 95% Confidence interval
1.3 Ni 79.6 3 1.2
5 Ni 57.4 7.9 3.3
6 Ni 59.3 7.7 3.6
The welding procedures for these weldments were:
 Submerged arc welding (SAW) with autogenous experimental
filler. The low nickel content of this autogenous experimental
filler will enable studies of brittle fracture.
 SAW with standard commercial filler for the two root beads
where the dilution is high. The remaining 16 beads were
submerged arc welded with the autogenous experimental filler
and nickel powder additions of 2.5–4.9 g/bead. This configura-
tion enables comparison with the previous low nickel weld.
 Metal inert gas (MIG) welding of the first 7 beads and SAW for
the remaining 11 beads. The standard commercial filler metal
23 7 NL was used for both methods. This configuration serves
as a reference weld.
For the SAW welding the flux material used had the following
chemical composition (wt%): 7 SiO2, 50 CaF2, 36 Al2O3, 3 Cr.
Result from energy-dispersive X-ray spectroscopy (EDS) across
the weldments (through-thickness) is shown in Table 3. Small
samples were also cut out from the weld metal of the tested
specimens for combustion analysis of the nitrogen and oxygen
content. The main differences in chemical composition were in the
manganese, nickel and nitrogen content. The weldments are in
this paper named 1.3 Ni, 5 Ni and 6 Ni.
For the 6 Ni weldment the manganese content and nitrogen
content was likely increased due to dilution with the base metal.
For the 1.3 Ni and 5 Ni weldment the nitrogen was however lower
in the weldment than in the base metal which can possibly be
associated to outgassing of nitrogen during welding.
The phase composition of the weldments can be found in
Table 4. The 5 Ni and 6 Ni weldments had similar phase composi-
tion but the 1.3 Ni had clearly higher ferrite content. The ferrite
content at the root was higher,  62% and  63% for the 5 Ni and
6 Ni respectively, compared to the average values in Table 4. For
the 1.3 Ni weldment the ferrite content was lower, 72% at the
root weld compared to the average values in Table 4. The likely
explanation was that for the 5 and 6 Ni weldments the nickel
content was reduced at the root due to dilution with the base
metal. The lower energy input in the MIG welding also contributed
to the higher ferrite content in the root for the 6 Ni weldment. For
the 1.3 Ni weldment the nickel content was instead increased due
to dilution (lower nickel content in the filler metal compared to
the base metal).
In Fig. 1 representative light optical microscope (LOM) photos
of the microstructure for the three different weldments are shown.
The microstructure consists of intergranular (grain boundary),
intragranular and Widmanstätten austenite in a ferrite matrix.
By observing the microstructure in LOM, the 1.3 Ni weldment
seemed to have less degree of Widmanstätten austenite compared
to the 5 Ni and 6 Ni weldments. The 5 Ni and 6 Ni weldments were
indistinguishable from each other. No intermetallic phases were
observed.
For duplex stainless base metals with its highly elongated
microstructure the austenite lamellar spacing have been found
to influence the impact toughness [12] and used for regression
analysis of yield and tensile strength [13,14]. For the weld metal
the microstructure was more complex so the mean free ferrite
distance was used to characterize the weld metal. Line intercept
counting with LOM at  100 magnification on polished and etched
samples of the weldments revealed that the 5 and 6 Ni weldments
had similar mean free ferrite distance, see Table 5. The much larger
mean free ferrite distance for the 1.3 Ni weldments was explained
by the higher ferrite phase content.
 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57
Fig. 1. Typical microstructure of the three weldments (100  magnification). The
scale bar is 100 μm. Etched with Beraha II reagents. (a) 1.3 Ni, (b) 5 Ni, and (c) 6 Ni.
3. Testing procedure
3.1. Tensile testing
The tensile testing was conducted with an electromechanical
tensile machine equipped with a 50 kN load cell. The initial strain
51
Table 5
LOM measurement of the mean free ferrite distance.
Average Average No. of Standard
ferrite ferrite intercepts deviation,
mean free mean free μm
distance at distance, μm
weld root, μm
1.3 Ni 23.7 31.6 370/538 2.4/5.6
5 Ni 20.5 11.1 1257/785 4.0/1.2
6 Ni 15.9 11.6 806/889 2.4/1.7
(ε o5%) was measured by an extensometer. The tests were
performed with round specimens 5 mm in diameter and 60 mm
length including threads (36 mm gauge length). The specimens
were extracted along the center of the weld metal 2 mm below the
surface.
The specimens were submerged in ethanol and liquid nitrogen
was used to cool the ethanol down to the desired testing
temperature. With a thermometer and a stirrer the temperature
was controlled within 72 1C during the entire tensile testing.
The temperature was held constant at 10 min prior to testing.
The strain rate was 0.00028 s
1.
3.2. Impact and fracture toughness testing
The impact testing was conducted according to the ISO 148-1
standard. The impact specimens were extracted from the weld at
t/4 depth (t¼plate thickness). The notch orientation was T–L.
For the fracture toughness measurement, standard single-edge
notched bend bars specimens, SE(B) were used. The specimen
dimension was 30  64  400 mm3 (thickness  width  length).
The specimen orientation was T–L and the notch-tip was located
in the weld metal center line [15]. Side-grooves were used on all
fracture toughness specimens. The crack length divided by the
specimen width was 0.5 including fatigue pre-crack. The fracture
mechanic testing was done with a 100 kN hydraulic testing
machine and a clip-gauge to measure the crack mouth opening
displacement (CMOD) of the specimen. The specimens were
submerged during testing in ethanol. Liquid nitrogen was used
to cool the ethanol down to the testing temperature. The tem-
perature was held constant at 36 min prior to testing.
The specimens were loaded with a constant displacement of
0.025 mm/s until failure occurred or the maximum force plateau
was exceeded. For the latter case the specimen was subsequently
cooled down until cleavage fracture was reassured to happen and
the specimen was broken up for marking the stable crack growth
that had occurred.
Fracture toughness testing was based on BS7448 Part 2 [15] and
the evaluation of the test results was based on ASTM E 1820-06
[16] and ASTM E 1921-05 [17].
3.3. Local compression
The original intent was to test the welded specimens in a
pristine condition after the welding. This can cause problem with
fatigue precracking because of the residual stresses from the
welding. The residual stresses change the stress state along the
crack front and generate an irregular crack front. Methods to avoid
this irregular fatigue crack growth include reversed bending,
stepwise high R-ratio fatigue and local compression [15,18,19].
The recommended method is local compression [18] and this was
also the method used in this work.
Both sides of the specimens containing the notch-tip and
remaining ligament were compressed until the desired plastic
deformation was reached. The local compression was applied prior
52 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57

Table 6
Straightness of the pre-crack fronts.

T, 1C Straightness of pre-crack front

1.3 Ni þ20 –
1.3 Ni
1 –
1.3 Ni
21 –
1.3 Ni
41 –
5 Ni þ20 Accepted, 9(2–8) pointa
5 Ni
1 Accepted, 9(2–8) point
5 Ni
20 Accepted, 9 pointb
5 Ni
40 Accepted, 9 point
6 Ni þ20 Accepted, 9 point
6 Ni
1 Accepted, 9(2–8) point
6 Ni
21 Accepted, 9 point
6 Ni
40 No local compressionc
6 Ni
50 Accepted, 9 point
6 Ni
60 Accepted, 9 point

a
Pre-crack front measured as stated in ASTM E 1820 points No. 1 and 9 were
ignored.
b
Pre-crack front measured as stated in ASTM E 1820.
c
Pre-crack front as in Fig. 2a.

the pre-crack was thereby not accepted according to the standard

Fig. 2. Photographs showing the crack front after fatigue precracking. (a) as-
received from welding, and (b) after local compression.
[ASTM 1820]. Ignoring the two near surface points resulted in an
accepted pre-crack front. The 6 Ni specimen tested at
1 1C had a
similar thumbnail shape as for the two 5 Ni specimens. The other
specimens had an accepted pre-crack front. No correlation
between total plastic deformation and the degree of thumbnail
shape was found.
The crack length used for all fracture toughness evaluation was
the nine point average measurement except for the 1.3 Ni speci-
mens where the notch length was used.

to fatigue precracking and side grooving. The procedure followed
the one described in BS7448 [15].
The result from the local compression can be seen in Fig. 2. For
the as-received specimen (without local compression) most of the
fatigue crack growth occurred at the sides and hardly at all in the
middle of the specimen. For the local compressed specimens the
fatigue crack growth took place more evenly across the cross
section.
The target for the total plastic deformation was 0.8%. The
resulting value was between 0.63% and 1%. As can be seen in
Fig. 2b the fatigue crack front has grown more in the middle of the
specimen, giving a slightly thumbnail shape which might indicate
that the plastic deformation was a bit too high.
The measurement of the length of the fatigue pre-crack was
measured from nine equally spaced points along the pre-crack
front [ASTM 1820]. The point Nos. 1 and 9 were measured 0.005
times the specimen width from the side-groove edge. The average
of the two values at the edges was averaged with the average of
the remaining values. The maximum deviation from this average
pre-crack length is 1.5 mm for a SE(B) specimen with 30 mm
thickness [ASTM 1820].
In Table 6 the straightness of the pre-crack front for the
specimens is shown. For the 1.3 Ni specimens the fatigue pre-
cracking led to an irregular pre-crack front. Some regions were less
affected by the fatigue precracking than other and some regions
inside the pre-crack region were unaffected by the fatigue pre-
cracking. No meaningful measurement of the pre-crack length
could therefore be obtained. For the 5 Ni specimens at þ20 and at

1 1C the fatigue crack growth was less at the edges of the
specimens. The result was a thumbnail shaped pre-crack front.
The maximum difference between the two near surface points and
the average crack length exceeded 1.5 mm and the straightness of
4. Results
4.1. Tensile tests
The yield and tensile strength of the weldments can be found in
Fig. 3. It was evident that the yield and tensile strength increases
with decreasing temperature. The 1.3 Ni weldment had the lowest
mechanical strength and the 5 Ni weldment the highest.
The fracture behavior of the three weldments at room tem-
perature was ductile failure after necking with some local regions
(circular with a diameter of  1 mm) of initiated transcrystalline
cleavage fractures. In the center of each local cleavage region a gas
pore was found. The failure was likely to occur when cleavage
fracture was initiated from this gas pore. This reduced the bearing
capacity of the specimen causing specimen failure. At 0 1C the
1.3 Ni weldment failed by complete cleavage fracture prior to
necking. With further reduced temperature the elongation to
fracture decreased as shown in Table 7. The 5 and 6 Ni specimens
had the same type of failure event at lower temperatures as at
room temperature. For the 5 Ni specimens the elongation to
fracture decreased with reducing temperature.
4.2. Impact toughness
The Charpy V impact toughness of the weldments is shown in
Fig. 4. The 1.3 Ni weldments had the lowest impact toughness
while the 6 Ni weldments had the highest one. The 1.3 Ni was
brittle at all test temperatures with clearly visible transcrystalline
cleavage fracture facets.
For the 5 Ni specimen at room temperature with the highest
impact toughness, the fracture surface was dominated by shear
lips and ductile failure. Small local regions of arrested cleavage
fractures were also visible. The other two specimens at room
 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57 53

1000 Table 8
1.3 Ni Results from the fracture toughness testing.
950 5 Ni
900 6 Ni T, 1C Jc, kN/m JIc, kN/m Test event
Rp0.2
850 Rm 1.3 Ni þ 20 6.3 Na Unstable failure
1.3 Ni
1 8.3 Na Unstable failure
Stress [MPa]

800 1.3 Ni
21 7.0 Na Unstable failure

1.3 Ni
41 7.7 Na Unstable failure
750
5 Ni þ 20 Na 389.6 Stopped (stable)
700 5 Ni
1 66.6 Na Unstable failure
5 Ni
20 54.4 Na Unstable failure
650 5 Ni
40 23.2 Na Unstable failure
6 Ni þ 20 Na 622.9n Stopped (stable)
600
6 Ni
1 Na 402.6 Unstable failure
550 6 Ni
21 258.1a Na Unstable failure
6 Ni
40 78.3 Na Unstable failure
500 6 Ni
50 35.6 Na Unstable failure
−40 −30 −20 −10 0 10 20 6 Ni
60 34.7 Na Unstable failure
Temperature [°C]
a
The master curve method [17].
n
Fig. 3. Yield and tensile strength of the weldments with different nickel contents The slope of the regression line exceeded (Rp0.2 þRm)/2.
as a function of temperature.

Table 7 1.3 Ni specimens. This was due to local transcrystalline cleavage

Elongation (%) to fracture. fracture as will be discussed in the fractography section below.
These pop-ins occurred at a stress intensity factor range of 0.5–

40 1C 20 1C 0 1C þ20 1C
1.7 kN/m.
1.3 Ni 3.6 5.5 7.1 17.6
The 5 Ni weldment behaved in a ductile way at room tempera-
5 Ni 16.5 20.2 24.9 27.0 ture. At 0 1C a large reduction in fracture toughness in comparison
6 Ni 22.9 23.5 24.2 22.2 to room temperature occurred when the specimen failed in an
unstable way.
The 6 Ni weldment was fully ductile at room temperature and the
200
test was stopped after extensive CMOD. However, no size indepen-
1.3 Ni
180 5 Ni dent initiation fracture toughness value could be obtained because
6 Ni the slope of the regression line exceeded the constraint set by the
160
standard [16]. At 0 1C the weldment had extensive stable crack
Impact toughness [J]

140 growth (on average 2.84 mm) prior to the unstable failure so valid
120 initiation fracture toughness could be obtained. At
20 1C the stable
crack growth decreased to about 100 μm prior to the unstable failure.
100 The result from the fracture toughness testing follows the same
80 ranking as for the impact toughness result, see Table 8 and Fig. 5.
The fracture toughness value for the point of fracture instability
60
was defined by the Jc criteria in ASTM 1820-06 [16]. The ductile
40 initiation fracture toughness, JIc, was evaluated according to the
normalization data reduction method in ASTM 1820-06 [16].
20

0 4.4. Fractography of the fracture toughness specimens

−100 −80 −60 −40 −20
Temperature [°C]
Fig. 4. Charpy V impact toughness of the weldments with different nickel contents
as a function of temperature.
temperature have increased degree of cleavage fracture, 1/3 of the
fracture area at 82J and cleavage dominating fracture surface at
50J. At 0 1C cleavage fracture was dominating.
The 6 Ni specimens were fully ductile at room temperature.
At 0 1C shear lips and deformed notch-tip were present. For the
146J specimen at that temperature, a small local arrested cleavage
region was present. The other two specimens had cleavage
dominating interior. The 6 Ni specimens showed full cleavage
fracture with no apparent notch-tip deformation at
60 1C.
The temperature at 40J impact toughness, the T40J value, was found
to be
13 1C for the 5 Ni weldment and
43 1C for the 6 Ni weldment.
4.3. Fracture toughness
The 1.3 Ni weldment failed in an unstable way at all tested
temperatures. The fracture toughness was very low, 6.3–8.3 kN/m.
During the fatigue precracking pop-in events occurred for the
0 20
Fig. 6a shows an example of a local cleavage fracture that
occurred during the fatigue precracking of the 1.3 Ni specimens.
The stress intensity factor for this pop-in was 1.7 kN/m.
The fracture modes found in the fracture toughness specimens
were microvoid coalescence (MVC), Fig. 6b, and transcrystalline
cleavage fracture, Fig. 6a, c and d. For the 5 Ni and 6 Ni specimens
tested at room temperature the fracture mechanism was crack-tip
blunting with subsequent MVC until the test was stopped (no failure).
For all other specimens the fracture process was crack-tip blunting
with some subsequent MVC until critical cleavage fracture occurred.
Some locally arrested cleavage fractures inside the MVC region could
also be observed. The degree of crack-tip blunting and the amount of
MVC prior to the critical cleavage fracture initiation decreased with
decreasing temperature for all specimens, compare Fig. 6c and d.
The failure of the tensile specimens occurred as stated in Section
4.1 from cleavage fracture initiated from gas pores. Due to the test
setup no area measurements could be made during the testing so no
Bridgman correction of the stress could be made. Therefore it was
not known at which stress the cleavage fracture initiation occurred.
After examination of the impact and fracture toughness speci-
mens no relation between gas pores and fracture initiation was
54 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57

found for these specimens. Only for one fracture toughness speci-
men a gas pore was observed close to the crack-tip, see Fig. 7a.
As was evident from the SEM photo, cleavage fracture occurred
prior to the gas pore and ductile fracture was observed almost
around the entire pore. One can conclude that the cleavage fracture
initiation in toughness specimens was not caused by gas pores.
5. Discussion
5.1. Tensile properties
As shown in Table 4, the phase composition was similar for the
5 Ni and 6 Ni weldments while the 1.3 Ni weldment had much
higher ferrite content. As a consequence the mean free ferrite path
was larger for the 1.3 Ni weldments than for the other weldments,
recall Table 5.
The chemical composition of the weldments mainly differs for
three elements, manganese, nickel and nitrogen, recall Table 3.
Previous regression analysis of the yield and tensile strength of
base material of duplex and austenitic stainless steels includes a
contribution from the ferrite phase, the lamellar spacing and the
manganese, nickel and nitrogen content [13,14]. If one assumes
that the lamellar spacing reflects the value of the ferrite mean free
path, the results from the tensile testing in Fig. 3, can be under-
stood: the 1.3 Ni specimens have higher ferrite content and
a higher mean free ferrite path, which give a reduction in yield
strength compared to the other two types of weldments. This
reduction was somewhat offset by the higher manganese and

700
6 Ni
5 Ni
600
1.3 Ni

500

400
J [kN/m]

300

200

100

0
−60 −40 −20 0 20
Temperature [°C]
Fig. 5. Fracture toughness of the weldments. The data points are from Table 8. Fig. 7. Gas pore close to the crack-tip in the 5 Ni specimen tested at
40 1C.

Fig. 6. SEM examination of various fracture surfaces from the fracture toughness tested specimens. (a) Cleavage fracture initiation causing pop-in during precracking, 1.3 Ni
specimens at þ 20 1C, (b) microvoid coalescence fracture, 6 Ni specimen at þ20 1C, (c) fracture initiation at
21 1C for the 6 Ni specimen and (d) fracture initiation at
40 1C
for the 5 Ni specimen.
 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57
1050
1000
950
True tensile stress [MPa]
900
850
800
750
700
1.3 Ni
650
5 Ni
6 Ni
600
−40 −30 −20 −10 0 10
Temperature [°C]
Fig. 8. True tensile stress at necking or failure before necking of the tensile
specimens.
nitrogen content in the 1.3 Ni weldment. With the same reasoning
the 5 Ni weldment have higher yield strength than the 6 Ni
weldment due to the higher manganese and nitrogen content. The
tensile strength was further increased compared to the 6 Ni
weldment due to the lower nickel content in agreement with
the observations. A similar argument for tensile specimens can be
found in [20] for some experimental duplex stainless alloys with
different phase composition, grain sizes, manganese and nickel
contents. The lower tensile strength for the 1.3 Ni weldment can
also partially be explained by the fact that this weldment only
reached necking at room temperature. One should note that the
published regression analyses were made from specimens tested
at room temperature so any temperature dependence of the yield
and tensile strength was not incorporated.
5.2. The influence of nickel
The results of the tensile tests, Fig. 3, show a similar engineer-
ing stress evolution (yield and tensile strength) with decreasing
temperature for the three different weldments. If one instead
evaluates the tensile testing by using true stress and true strain
under the assumption of constant volume, the three weldments
behavior with decreasing temperature were different. Fig. 8 shows
the true stress at necking for the 5 Ni and 6 Ni tensile specimens
and the true stress at necking at room temperature and the true
stress at failure at the other testing temperatures for the 1.3 Ni
weldment. The 1.3 Ni and 5 Ni specimens have a fairly tempera-
ture independent true stress to necking/failure while the 6 Ni
specimens have a true stress to necking that increases with
decreasing temperature. This difference in behavior with decreas-
ing temperature was also clearly seen for the true strain to
necking, see Fig. 9. The 1.3 Ni and 5 Ni weldments loose ductility
with decreasing temperature while for the 6 Ni weldment the
ductility was approximately unchanged for the tested temperature
interval.
From impact toughness testing of low carbon alloys with
varying nickel content it has been shown that nickel increases
the cleavage stress and shifts the ductile-to brittle transition
temperature to lower temperatures [6,7]. One explanation given
is that nickel increases the ability for dislocations to cross-slip in
the ferritic phase [7] which would decrease the local stress from
dislocation pile-ups around cleavage fracture initiation sites. This
increased ability for cross-slip due to nickel has been observed by
Jolley [6]. The high cleavage fracture susceptibility for the 1.3 Ni
weldment can be explained by the high ferrite content with large
mean free ferrite path and low nickel content. An explanation for
55
0.2
0.18
0.16
0.14
εtrue,necking [m/m]
0.12
0.1
0.08
0.06
0.04
1.3 Ni
0.02 5 Ni
6 Ni
0
−50 −40 −30 −20 −10 0 10 20 30
20
Temperature [°C]
Fig. 9. True strain to necking at necking or failure before necking for the tensile
specimens.
the reduced ductility for the 5 Ni weldment with decreasing
temperature would then be that the necking (point of instability)
occurs at lower strains due to change in dislocation behavior. For
the 6 Ni weldment this change in dislocation behavior occurs at
lower temperatures than tested and therefore the strain to necking
was unchanged.
The difference in fracture behavior during fracture toughness
testing between the 5 Ni and 6 Ni weldments with decreasing
temperature was that the 5 Ni specimens failed catastrophically at
a lower force. Both weldments exceeded 90 kN at room tempera-
ture and with reduced temperature the 5 Ni failed at 63 kN at 0 1C,
57 kN at
20 1C and 38 kN at
40 1C. The 6 Ni weldment
remained at  90 kN until
40 1C where it failed at 71 kN. The
crack-tip blunting and the region of ductile fracture prior to
cleavage fracture decreased greatly with decreasing temperature
for the 5 Ni weldment.
For impact toughness and fracture toughness specimens, the
material volume that was affected by the loading, the process
zone, was much smaller compared to the entire volume of the
tensile specimen. The results were therefore largely dependent on
the notch/crack-tip location in relation to cleavage fracture initia-
tion sites (inclusion, carbides, nitrides, intermetallic phases, etc.)
in the material. This gives high scatter for impact and fracture
toughness when tested in the ductile to brittle transition region.
This can be seen from the impact toughness testing where three
specimens at each temperature were tested, Fig. 4. For the fracture
toughness testing only one specimen per temperature was tested
so the scatter was unknown. The results from the impact and
tensile testing indicate, however, that the difference in fracture
toughness between the 5 Ni and 6 Ni reflects the increased
resistance to cleavage fracture initiation in the 6 Ni weldment.
The two weldments 5 Ni and 6 Ni had similar phase composi-
tion and microstructure characteristics, recall Tables 4 and 5. The
chemical composition differs mainly in the nitrogen and manga-
nese content. However, for duplex stainless steel weld metals the
element partitioning is less expressed than for the base metal.
Normally during welding the cooling of the weld metal is too rapid
for the diffusion of the substitutional solute elements (Cr, Mo, Mn
and Ni) to play a significant role in the austenite formation [21,22].
Instead the austenite formation is governed by the diffusion of the
interstitial elements carbon and in particular nitrogen [21,22]. The
nitrogen content in the ferrite was thereby likely to be similar
between the two weldments. To the authors' knowledge, manga-
nese has no inherent detrimental effects on the ferrites suscept-
ibility for cleavage fracture except for the increase in yield
strength. Reduced impact toughness with increased manganese
56 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57

Table 9
Microstructure properties and nickel content for various duplex stainless steel weld metals. Standard deviation in parenthesis.

2205 WM [24] 2304 WM [10] LDX 2101s WM [9] 5 Ni 6 Ni

α Phase contenta, % 63.9 ( 7 2.7) 44.0 ( 79) 45 ( 76) 57.4 ( 7 7.9) 59.3 ( 7 7.7)
γ Phase contenta, % 36.1 (7 2.7) 56.0 ( 7 9) 55 ( 76) 42.6 ( 7 7.9) 40.7 ( 7 7.7)
Average mean free distance, weld roota, μm – – 16.6 ( 7 3.9) 20.5 ( 7 4.0) 15.9 ( 7 2.4)
Average mean free distance, welda, μm 22.9 ( 7 3.4) – – 11.1 ( 7 1.2) 11.6 ( 7 1.7)
Average Ni content in weldb, wt% 7.7 ( 7 0.2) – 8.2 (7 0.6) 5 6.1 ( 7 0.4)
Average Ni content in αc, wt% 7.4 ( 7 0.3) – 8.3 (7 0.4) – 6.2 ( 7 0.2)
Average Ni content in γc, wt% 7.9 ( 7 0.4) – 8.2 (7 0.5) – 6.1 ( 7 0.2)

a
Measured from etched samples by LOM.
b
Area analyses on polished samples by EDS.
c
Average spot analyze, 7 measured spots for each phase at random locations.

200
2205 WM, K−joint the filler metal usually, the nickel content in the weld metal can
180 2304 WM, X−joint decrease with decreasing fracture toughness as a result [25]. The
2101 WM, X−joint
160 lean duplex stainless steel with its low nickel content is more
6 Ni, X−joint
140 5 Ni, X−joint susceptible to this than DSS grades with higher nickel content.

120
JC [kN/m]

100 7−9 wt% Ni 6. Conclusions

80 From plates of LDX 2101s duplex stainless steels three types of

60
6 wt% Ni
40 5 wt% Ni
20
0
−120 −100 −80 −60 −40 −20
Temperature [°C]
Fig. 10. Sub-zero temperatures fracture toughness of DSS weld metals (SAW). The
references for the 2205, 2304 and LDX 2101s WM can be found in [24,10,9]
respectively.
content can be found for high strength steel, however this
toughness reduction was attributed to the change in the micro-
structure [23]. The likely explanation for the increased toughness
for the 6 Ni weldment was therefore the higher nickel content.
In Table 9 the phase content, average mean free distance in the
ferrite and the nickel content for various DSS steel weld metals are
shown. Due to the low partitioning of the substitutional elements
these elements were more evenly distributed between the two
phases in the weld metal compared to the base metal. This can be
seen for nickel in Table 9. The other elements chromium, molyb-
denum and manganese had similar evenly distribution between
the phases.
Fig. 10 shows the measured fracture toughness at low tem-
peratures for these weld metals. The 2205, 2304 and LDX 2101
specimens in References [24,10,9] respectively were not subjected
to local compression and had therefore pre-crack fronts similar to
Fig. 2a. This type of pre-crack front has been shown to give
increased KIc fracture toughness in the transition temperature
region [10]. However, the unstable fracture surfaces for these weld
metals consisted to a large degree of MVC fracture (but not enough
for avoiding unstable failure of the specimens). This was markedly
different from the fracture surfaces of the 5 Ni and 6 Ni specimens
in this report which showed in principle full cleavage fracture at
and below
40 1C (recall Fig. 6d). This indicates that the 2205,
2304 and LDX 2101 weld metals from the references had inherent
higher toughness than the weld metals tested in this paper. When
considering the phase composition, microstructure and chemical
content of the weldments the increase in fracture toughness
correlates well with the increase in nickel content.
One can also draw the conclusion from Fig. 10 that if dilution
with the base metal occurs, which has lower nickel content than
weldments with different nickel content yet similar microstruc-
ture were made for the purpose of investigating the susceptibility
for brittle failure at sub-zero temperatures. Tensile, impact tough-
ness and fracture toughness testing were conducted.
 Using filler metal with low nickel content (1.3 wt%) resulted in
a weld metal with high ferrite content and low toughness at
room temperature and sub-zero temperatures.
 With the addition of nickel powder the nickel content was
increased to 5 wt% which resulted in a weldment with appro-
priate ferrite content. The ductility and toughness were sig-
nificantly increased.
 By using a filler metal with the higher nickel content of 7.3 wt%,
the ductility and toughness were further increased compared
to the 5 wt% nickel weldment.
 Published fracture toughness data on other duplex stainless
steel weldments together with the results of the present
investigation show that the fracture toughness at sub-zero
temperatures increases with increasing nickel content in the
range from 1 to 9 wt% nickel.
Acknowledgments
The authors would like to express their gratitude to the VINN
Excellence Center Hero-M and Outokumpu Stainless AB for finan-
cing this study. Outokumpu Stainless AB is also gratefully acknowl-
edged for delivering the material and for conducting the impact
toughness testing and combustion analysis. Valuable support from
Mikael Johansson, Ravi Vishnu and Jan Y Jonsson at Outokumpu
Avesta Research Centre is acknowledged.
References
[1] R.N. Gunn, Duplex Stainless Steels – Microstructure, Properties and Applica-
tions, Woodhead Publishing Ltd., Abington Hall (2003) 175–184.
[2] J. Charles, Rev. Métall. 105 (2008) 155–171, http://dx.doi.org/10.1051/
metal:2008028.
[3] G. Gedge, J. Constr. Steel Res. 64 (11) (2008) 1194–1198.
[4] R.N. Gunn, Duplex Stainless Steels – Microstructure, Properties and Applica-
tions, Woodhead Publishing Ltd., Abington Hall (2003) 16–17.
[5] V. Muthupandi, P. Bala Srinivasan, V. Shankar, S.K. Seshadri, S. Sundaresan,
Mater. Lett. 59 (18) (2005) 2305–2309.
 J. Pilhagen, R. Sandström / Materials Science & Engineering A 602 (2014) 49–57
[6] W. Jolley, Trans. Metall. Soc. AIME 242 (1968) 306–314.
[7] S. Floreen, H.W. Hayden, T.M. Devine, Metall. Trans. 2 (1971) 1403–1406.
[8] R.N. Gunn, Duplex Stainless Steels – Microstructure, Properties and Applica-
tions, Woodhead Publishing Ltd., Abington Hall, 2003(pp. 122).
[9] H. Sieurin, R. Sandström, E.M. Westin, Metall. Mater. Trans. A 37A (2006)
2975–2981.
[10] H. Sieurin, E.M. Westin, M. Liljas, R. Sandström, Weld. World 53 (3–4) (2009),
R24–R33.
[11] ASTM E 1245: Standard Practice for Determining the Inclusion or Second-
Phase Constituent Content of Metals by Automatic Image Analysis, vol. 03.01.
[12] E.M. Westin, B. Brolund, S. Hertzman, Steel Res. Int. 6 (2008) 473–481.
[13] H. Sieurin, J. Zander, R. Sandström, Mater. Sci. Eng. A 415 (2006) 66–71.
[14] R.N. Gunn, Duplex Stainless Steels – Microstructure, Properties and Applica-
tions, Woodhead Publishing Ltd., Abington Hall, 2003(pp. 58).
[15] BS7448 Part 2: Method for Determination of KIc Critical CTOD and Critical J
Values of Welds in Metallic Materials, seventh draft (revised), 1996.
[16] ASTM E 1820-06: Standard Test Method for Measurement of Fracture Tough-
ness, vol. 03.01.
[17] ASTM E 1921-05: Standard Test Method for Determination of Reference
Temperature, T0, for Ferritic Steels in the Transition Range, vol. 03.01.
57
[18] O.L. Towers, M.G. Dawes, Welding institute research on the fatigue pre-cracking
of fracture toughness specimens, Elastic–Plastic Fracture Test Methods: The
Users Experience, ASTM STP 856, April 1985, Ann Arbor.
[19] M.G. Dawes, H.G. Pisarski, S.J. Squirrell, Fracture Mechanics Tests on Welded
Joints, in: J.D. Landes, A. Saxena, J.G. Merkle (Eds.), Nonlinear Fracture
Mechanics: Vol. II–Elastic – Plastic Fracture, ASTM STP 995, American Society
for Testing and Materials, Philadelphia, 1989, pp. 191–213.
[20] S.M. Wessman, S. Hertzman, R. Pettersson, R. Lagneborg, M. Liljas, Mater. Sci.
Technol. 24 (3) (2008) 348–355.
[21] R.N. Gunn, Duplex Stainless Steels – Microstructure, Properties and Applica-
tions, Woodhead Publishing Ltd., Abington Hall, 2003(pp. 120).
[22] E.M. Westin, S. Hertzman, Element distribution in lean duplex stainless steel
welds, In: E.M. Westin, Microstructure and Properties of Welds in the Lean
Duplex Stainless Steel LDX 2101, Doctoral Thesis in Material Science, 2010,
KTH Royal Institute of Technology, Stockholm, Sweden.
[23] E. Keehan, L. Karlsson, H.-O. Andrén, L.-E. Svensson, Weld. J. 85 (10) (2006)
218–224.
[24] H. Sieurin, R. Sandström, Eng. Fract. Mech. 73 (2006) 377–390.
[25] L. Karlsson, H. Arcini, E.-L. Bergquist, J. Weidow, J. Börjesson, Weld. World 54
(11–12) (2010) 350–359.