ORIGINAL RESEARCH ARTICLE

Austenite Morphology and Distribution Dependence

of Impact Toughness in S32101 Duplex Stainless
Steel Laser Welds
BAOJUN WANG, YONGPENG ZHANG, XINJUN SHEN, PENGCHENG HUAN,
QINGYU ZHANG, and XIAONAN WANG

In this study, S32101 duplex stainless steel (DSS) was welded by laser and then subjected to
post-weld heat treatment (PWHT) at 1100 C for various times to tailor the weld microstruc-
ture. The impact toughness of the welds was tested at  40 C, and the crack propagation
behavior was characterized in detail by electron back-scattered diffraction (EBSD). After
PWHT for 4.5 and 30 minutes, the austenite volume fraction increases from 17.8 pct (HT0
weld) to 35.4 pct (HT4.5 weld) and 37.7 pct (HT30 weld), and the impact absorbed energies are
4.0, 5.1, and 20.7 J, respectively. Therefore, the austenite volume fraction is not the key factor in
determining weld toughness. The increase in austenite in the HT4.5 weld is due to the coarsening
of the original austenite, while it is mainly the formation of intragranular austenite (IGA) in the
HT30 weld, which contributes to the uniform distribution of austenite. The EBSD analysis
shows that the IGA has higher crack propagation resistance than widmanstätten austenite (WA)
and grain boundary austenite (GBA) because the crack can deflect at the IGA/ferrite phase
boundaries. Coherent or semicoherent orientation relationships (ORs) also increase the fracture
critical stress. Therefore, ductile fraction occurs in the HT30 weld, and its toughness improves
significantly. In addition, the local plastic deformation caused by IGA agglomeration in the
macroscopic cleavage fracture region also enhances the HT30 weld toughness. Thus, the
austenite morphology and distribution play a key role in the impact toughness of the welds.

https://doi.org/10.1007/s11661-024-07314-x
 The Minerals, Metals & Materials Society and ASM International 2024

I. INTRODUCTION Welding is an essential connection method in engi-

neering construction. However, when the medium-thick-
DUPLEX stainless steel (DSS) has outstanding ness DSS plates are connected by traditional arc
mechanical properties and corrosion resistance due to welding, opening grooves, filling wires, and even mul-
its balanced microstructure, including ferrite and tiple passes welding are necessary due to their low
austenite.[1,2] Compared with the second-generation energy density,[4,6] which increases costs and reduces
standard DSS and super duplex stainless steel (SDSS), efficiency. Laser welding with high energy density can
the second-generation lean S32101 DSS uses cheaper connect medium-thickness DSS plates only by one pass
manganese and nitrogen instead of nickel, which is without opening grooves and filling wires.[7] Moreover,
conducive to saving production costs.[3,4] Therefore, laser welding has faster welding speed, higher precision
S32101 DSS has been widely used in petrochemicals, and higher automation, which has been widely used in
pressure pipelines of natural gas transportation and engineering manufacturing.[8,9] However, excessive fer-
structural parts.[2,5] rite phase and precipitates caused by rapid cooling of
laser welding harm mechanical properties and corrosion
properties.[10–14] Impact loads often exist in the service
environment, so it is necessary to tailor the weld
microstructure and improve toughness.
Optimizing welding process parameters, adding nitro-
BAOJUN WANG, YONGPENG ZHANG, XINJUN SHEN, gen gas during welding, and post-welding heat treatment
QINGYU ZHANG, and XIAONAN WANG are with the School of (PWHT) are often used to tailor the weld microstruc-
Iron and Steel, Soochow University, Suzhou 215100, P.R. China. ture.[10,12,15–17] However, optimizing process parameters
Contact e-mails: [email protected], [email protected] can only weaken non-equilibrium solidification due to
PENGCHENG HUAN is with the The State Key Laboratory of
Rolling and Automation of Northeastern University, Shenyang low heat input.[17] It is undoubtedly effective in balanc-
110819, P.R. China. ing ferrite/austenite proportion by adding nitrogen
Manuscript submitted July 12, 2023; accepted January 7, 2024.

METALLURGICAL AND MATERIALS TRANSACTIONS A

during welding, as nitrogen is the austenite-forming heat treatment), HT4.5 (heat treatment for 4.5 minutes),
element. However, the production of nitrides is also and HT30 (heat treatment for 30 minutes), respectively.
inevitable.[12,16] Many studies have proved that PWHT
can balance the phase proportion and eliminate precip-
B. Microstructure Characterization and Impact
itation simultaneously.[10,15] In addition, many research-
Toughness Testing
ers have found that the weld impact toughness is not
only related to phase equilibrium[18–20] but also to After heat treatment, microstructure characterization,
microstructure morphology.[10,21] Therefore, obtaining Charpy impact testing and observation of impact
the austenite with appropriate morphology through fracture behavior were performed. The specimens were
PWHT is also essential. Haghdadi et al.[21] and Wu polished and etched with Beraha’s II solution (100 mL
et al.[22] found that elongated austenite has higher H2O + 20 mL HCl + 1 g K2S2O5). Then, the
cryogenic toughness than equiaxed and widmanstätten microstructures were observed using an RX50M optical
austenite (WA) in the rolled DSS. However, the austen- microscope (OM) and a SU-5000 field emission scanning
ite morphology, distribution and size in the microstruc- electron microscope (FE-SEM). The EBSD analysis was
ture of rolled steel plates differ significantly from the conducted by a ZEISS G300 FE-SEM with a step size of
weld microstructure. Furthermore, Zhang et al.[10] sug- 0.4 lm, and a HKL Channel 5 and OIM Analysis
gested that the non-coherent relationship between software were used for postprocessing. According to
spheroidized intragranular austenite (IGA) and ferrite GB/T 229-2020, subsized Charpy impact specimens of
in electron beam welds contributes to toughness. How- 55 9 10 9 5 mm with V-notch were machined [Fig-
ever, studies on the austenite morphology (including ures 1(b) and (c)]. A PIT452D-2 impact testing machine
grain boundary austenite (GBA), WA, and IGA) and was used to achieve the impact absorbed energy of the
distribution in the welds are lacking. specimens at  40 C, and the impact experiment was
In the present study, S32101 DSS was welded by laser, repeated three times. To analyze the fracture mecha-
and then the morphology and distribution of austenite nism, a TESCAN VEGA3 SEM was used to observe the
in the welds were tailored by PWHT at 1100 C for macro- and micromorphologies of the impact fractures,
various times. Electron back-scattered diffraction and the microstructures near impact cracks were char-
(EBSD) analysis was used to characterize the acterized by EBSD analysis. Nickel plating was applied
microstructure and crack propagation behavior in to the fracture surfaces to preserve the microstructures
detail. The influence of PWHT on the austenite mor- near impact cracks and obtain good characterization
phology and distribution in the welds was discussed, and results. The nitrogen content in the welds was measured
their influence mechanism on impact toughness was by a TCH600 ONH analyzer to quantify nitrogen loss in
illustrated based on the crack propagation behavior. the welds.

II. EXPERIMENTAL PROCEDURES III. RESULTS

A. Welding and Heat Treatment Procedures A. Microstructure of the Welds
The chemical composition of the used S32101 DSS is Figures 2(a) through (d) shows the welds cross-sec-
listed in Table I. Before welding, the samples were tion, optical microstructures, and the corresponding
machined to 160 9 75 9 6.5 mm by wire cutting and statistical data. The microstructures were observed at
then polished and cleaned. A welding platform with a the center position of the welds [Figure 2(a)]. All the
CWX-3000 fiber laser, shielding gas system and custom welds contain ferrite and austenite phases [Figures 2(b)
fixture was used, as shown in Figure 1(a). Our previous through (d)]. Three kinds of austenite exist in the welds,
tests proved that the welding parameters listed in i.e., GBA, WA, and IGA, marked by the yellow arrow.
Table II could avoid the formation of weld pores. Thus, It is evident that the GBA and WA are long strips, while
these parameters were adopted, and the relationship the IGA has a blocky shape, and the uniformity of the
between the microstructure and impact toughness of the austenite distribution in the welds is mainly affected by
welds can be truly reflected. After welding, the speci- IGA. In the HT0 weld, the austenite is mainly GBA
mens were reheated to 1100 C and held for 4.5 and [Figure 2(b)], the original austenite coarsens obviously
30 minutes, respectively, followed by quenching in water in the HT4.5 weld [Figure 2(c)], the number of IGA
to obtain austenite with different morphologies and increases significantly, and the austenite grains are
distributions and to avoid the formation of a precipi- distributed uniformly in the HT30 weld [Figure 2(d)].
tated phase. The specimens were named HT0 (without Figure 2(e) shows the volume fractions of austenite
and IGA in the welds. After PWHT, the volume fraction
of austenite in the welds increases from 17.8 to 35.4 and
Table I. Chemical Composition of Used S32101 DSS 37.7 pct, and the IGA increases from 5.1 to 12.8 and
(Weight Percent) 18.6 pct, respectively. In addition, the size distribution
and number density of IGA in the welds were quanti-
C Cr Ni N P Si Mo Mn Cu Fe tatively counted [Figure 2(f)], and the size is represented
0.02 21.64 1.52 0.19 0.02 0.52 0.16 5.21 0.13 bal. by the equivalent circular diameter. The proportion of
IGA larger than 4 lm increases from 36.9 to 59.3 and

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 1—Schematic diagrams of the laser welding system and specimen preparation process: (a) Laser welding system; (b) Microstructure
observation specimens and Charpy impact specimens; (c) Position and size of Charpy impact sample. TD is the transverse direction, RD is the
roll direction, and ND is the normal direction.

Table II. Welding Parameters for S32101 DSS high cooling rate was adopted after PWHT, so the
formation of carbide is inhibited in the welds.
Parameter Value
Laser Power (kW) 3 B. Impact Toughness of the Welds
Welding Speed (mm/minutes) 500
Heat Input (kJ/mm) 0.36 Figure 4 shows the weld impact energy and ductile
Defocusing Amount (mm) 2 fracture ratio after PWHT. After PWHT, the impact
Flow of Pure Argon (L/minutes) 30 energy increases from 4.0 to 5.1 J and 20.7 J, respec-
tively, consistent with the trend that the ductile fracture
ratio rises from 4.3 to 12.5 and 42.1 pct. Therefore, the
significant improvement in weld toughness is closely
related to the increase in the ductile fracture ratio.
66.5 pct, respectively. The number density of IGA Figure 5 shows the macro- and micromorphologies of
increases from 3280 to 3505/mm2 and 4220/mm2. Thus, the fracture surfaces after PWHT. The HT0 and HT4.5
the high proportion of IGA in the HT30 weld is mainly welds have no obvious plastic deformation on the
attributed to the significant nucleation and growth of macroscale [Figures 5(a) and (b)], and the micromor-
IGA after long-time PWHT. phology is cleavage facets, so almost an entire brittle
Due to the non-equilibrium transformation of the fracture occurs. The HT30 weld exhibits obvious plastic
weld microstructure, secondary precipitates are easy to deformation [Figure 5(c)], and dimples and cleavage
form in the weld. Figure 3 shows SEM images of the facets exist on the fracture surface. It is worth noting
welds. There is a large amount of secondary precipitates that the size of cleavage facets in the HT0 weld is large
in the HT0 weld, while there are no precipitates in the and smooth, and tear ridges caused by GBA at ferrite
HT4.5 and HT30 welds. SEM-EDS point and mapping grain boundaries indicate that cleavage cracks propa-
show that the secondary precipitates are rich in N, while gate steadily between and within ferrite grains [Fig-
Cr was not obviously enriched or even poor [Figures 3(b) ures 5(a1) and (a2)]. In the HT4.5 weld, the surface of the
through (d)], which are considered as nitride (chromium cleavage facets begins to be rough, but the size is still
nitride).[23] In addition, the nitride precipitates are large, and the cleavage crack propagates straight [Fig-
distributed in the interior of the coarse ferrite grains. ure 5(b1)]. For the HT30 weld, the height difference of
This precipitation behavior commonly occurs in duplex adjacent cleavage facets appears frequently, and the size
stainless steel welds.[12,20] As for the carbide, Kose of cleavage facets decreases significantly [Figure 5(c2)],
et al.[24–26] reported that Cr23C6 is often formed in the indicating that cleavage cracks propagate more tortu-
DSS weld when the cooling rate is slow, such as air ously in ferrite grains. Therefore, the cleavage fracture
cooling. In our present study, water quenching with a behavior in the three welds indicates that the HT30 weld

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 2—Cross-section, optical microstructures and the corresponding statistical data of the welds: (a) Cross-section; (b) HT0 weld; (c) HT4.5
weld; (d) HT30 weld; (e) Volume fractions of austenite and IGA; (f) Size distribution and number density of IGA (Color figure online).

has higher crack propagation resistance than the HT0 high cooling rate also inhibits the growth of austenite at
and HT4.5 welds. the ferrite grain boundaries (i.e., GBA) and the nucle-
ation inside the ferrite (i.e., IGA).[28,29] In addition,
about 0.05 wt pct nitrogen loss under the action of high
IV. DISCUSSION energy laser also brings difficulty for austenite formation
in the weld. Therefore, the HT0 weld without PWHT
A. Effect of PWHT on Weld Microstructure contains less austenite [Figure 2(b)]. A low fraction of
Figure 6 shows the equilibrium phase diagram of the austenite has limited storage capacity for nitrogen
S32101 DSS. With the decrease in temperature, ferrite element, resulting in supersaturation of nitrogen in the
forms first, and then austenite and nitrides generate in ferrite with low nitrogen solubility, so the nitrides are
sequence. In the laser welded welds, the rapid growth of formed in the interior of ferrite grains [Figures 3(a)
ferrite along the direction of heat flow leads to the through (d)].
formation of coarse ferrite columnar crystals and ferrite As shown in Figure 6, the theoretical austenite
grain boundaries with high energy provide a favourable volume fraction considering nitrogen element loss is
condition for austenite formation.[27,28] However, the 36.9 pct at 1100 C (Figure 6), much higher than that in
the HT0 weld (17.8 pct). Therefore, the PWHT will

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3—SEM images and elemental analysis of the welds: (a) HT0 weld; (b) and (c) the observation and analysis of nitrides; (d) SEM-EDS
mapping of nitride; (e) HT4.5 weld; (f) HT30 weld. The N element was not detected in the ferrite by spot scanning due to its low content since
Cr2N was precipitated.

significantly increase the austenite content in the welds.

After PWHT, the phase fractions in the HT4.5 and
HT30 welds are 35.4 and 37.7 pct, respectively, which
reached an equilibrium state.
In this study, the PWHT temperature exceeds the
solution temperature of nitrides (903 C), resulting in
the disappearance of nitride precipitates in the HT4.5
and HT30 welds [Figures 3(e) and (f)]. The dissolution
of nitrides increases the nitrogen content of the ferrite,
creating conditions for the formation of austenite.
However, Xie et al. pointed out that the formation of
IGA requires more active energy to overcome the
constraints of the surrounding ferrite lattice, so the
formation of IGA is slow.[28] Therefore, a small
amount of IGA is generated during the short-time
PWHT and the growth of the original austenite
domains. As a result, austenite in the HT4.5 weld is
Fig. 4—Impact energy and ductile fracture ratio of the welds after mainly composed of coarsened GBA and WA [Fig-
PWHT.

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 5—Macro- and micro morphologies of impact fracture surfaces of the various welds: (a) HT0 weld; (b) HT4.5 weld; (c) HT30 weld. (a1),
(a2) The top and the bottom micromorphologies of HT0 fracture surface. (b1), (b2) The cleavage facets and the dimples of HT4.5 fracture
surface. (c1), (c2) The cleavage facets and the dimples of HT30 fracture surface.

Fig. 7—IPF of austenite in the HT30 weld (Color figure online).

GBA and WA increase the proportion of IGA, as they

have no essential difference compared with IGA.

B. Fracture Mechanism for Austenite with Different

Fig. 6—The equilibrium phase diagram of S32101 DSS calculated by Morphology and Distribution
JMatPro.
In ferrite-austenite duplex steels, austenite has better
plastic deformation ability than ferrite due to more slip
systems. Therefore, the typical way to improve the weld
ure 2(c)]. When the PWHT time is extended to toughness is to increase the austenite content.[18–20,30]
30 minutes, sufficient time is provided for the nucle- However, in this study, the impact absorbed energy of
ation and growth of IGA, the size and the number the HT4.5 weld with 35.4 pct austenite is only 5.1 J,
density of IGA increase significantly [Figures 2(d) which is close to that of the HT0 weld with 17.8 pct
through (f)]. In addition, to follow the principle of austenite (4.0 J). The impact absorbed energy of the
minimizing interface energy, the long strips of WA and HT30 weld with 37.7 pct austenite reaches 20.7 J.
GBA gradually tend to be equiaxial [marked by black Obviously, the impact toughness of the welds is deter-
dashed rectangles in Figure 7). The decompositions of mined by other factors.

METALLURGICAL AND MATERIALS TRANSACTIONS A

 Previous studies have shown that microscopic inter-
faces with high misorientation, including grain bound-
aries and phase boundaries, play an important role in
crack propagation.[31–33] In our present study, the effect
of grain boundaries can be ignored because of their low
density [Figures 9(a) through (f)]. After PWHT, the
high-angle phase boundary density increases from
0.191 to 0.281 lm1 and 0.266 lm1, respectively. The
misorientation angle is mainly concentrated at ~ 44 deg,
which suggests that K–S (~ 42.85 deg) or N–W ORs
(~ 45.98 deg) exist between the austenite and ferrite
[Figure 9(g)]. Because the exact coherent ORs almost do
not exist, misorientation angles between 41.85 and
46.98 deg are regarded as coherent or semicoherent
ORs, which are represented by the red line in
Figures 9(d) through (f). Obviously, WA, IGA and
one side of GBA exhibit coherent or semicoherent ORs
with adjacent ferrite grains, consistent with the previous
reports in DSS weld.[25,34,35] Meanwhile, the density of
coherent or semicoherent interfaces located in the
interior of ferrite grains in the HT30 weld is higher
than that in the HT0 and HT4.5 welds [Figure 9(h)].
Although the types of WA/ferrite and IGA/ferrite
phase interfaces are the same, they show different roles
in crack propagation. In detail, cracks spread directly
through WA and are usually deflected at IGA. Accord-
Fig. 8—Inverse pole figures (IPFs) and local kernel average ing to fracture theory, the degree of stress concentration
misorientation (KAM) images of impact fracture surfaces: (a) HT0 determines the formation of cracks.[36] Thus, when the
weld; (b) HT4.5 weld; (c) HT30 weld; (a) through (c) IPFs; (a1), (a2) crack propagates to WA/ferrite phase boundaries, the
and (c1) are KAM images in (a) and (c) located by dashed stress release is relatively difficult due to the additional
rectangles, respectively; (d) Position of EBSD analysis (Color
figure online). constraints of coherent or semicoherent ORs,[10] and
then the crack spreads directly. However, IGA/ferrite
To clarify the relationship between the microstructure has a shorter phase boundary than WA/ferrite, resulting
and crack propagation in the welds, the crack propaga- in a relatively weak additional constraint and conducive
tion behaviors of the HT0, HT4.5, and HT30 welds were to crack deflection. In addition, Interface sliding is
investigated through EBSD (Figure 8). The crack inevitable when crack deflection at phase boundaries.
propagation path in the HT0 weld is straight, and the Compared with general phase interfaces, sliding along
direction of main crack propagation deviates from the coherent or semicoherent interfaces increases the surface
impact load direction, resulting in deflections of 36 deg energy (c) and then critical fracture stress (rc ) increases
(22 fi  14 deg), 125 deg ( 105 fi 20 deg) and 26 deg according to the Griffith fracture theory
(26 fi 0 deg) inside the ferrite. The accumulation of h i1=2
pEc [37]
IGA leads to 91 deg abnormal deflection and higher (rc ¼ ð1v2 Þd0 ), which also contributes to increas-
kernel average misorientation (KAM) values [green ing crack propagation resistance.
arrow in Figures 8(a) and (a1)], meaning local plastic Wang et al.[36] pointed out that the variation of crack
deformation occurs.[26] However, due to the small size propagation path length caused by crack deflection can
and low density of the IGA in the HT0 weld, the be quantitatively described by the roughness parameter
occurrence of crack deflection and local plastic defor- RL (the ratio of the true propagation length Lt to its
mation is limited [marked by red arrows in Figures 8(a) projected length L¢). The RL values of the three welds
through (a2)]. The crack goes straight through WA are 1.23, 1.15, and 1.84, respectively, as shown in
without deflection [marked by white dashed rectangles Figure 10. It is obvious that the frequent crack deflec-
in Figure 8(b)], so the fracture profile is flat in the HT4.5 tion caused by the high density and uniform distribution
weld, which has high percentage of GBA and WA of IGA increases the crack propagation path. So, the
[Figure 8(b)]. In the HT30 weld, effective and frequent energy consumption will be increased with the elonga-
crack deflection and the fracture profile becomes signif- tion of the crack propagation path.
icantly rough due that high density of IGA [red arrow in According to the above discussion, the toughening
Figure 8(c)]. The area of plastic deformation induced by mechanism of the HT30 weld can be summarized in
IGA accumulation increases [Figure 8(c1)]. Meanwhile, Figure 11. Because the IGA has higher crack propa-
the coarsened GBA in the HT4.5 weld does not improve gation resistance than WA and GBA, the high-density
its toughness. Hence, the IGA is more critical in and larger IGA in the HT30 weld increase resistance
inhibiting crack propagation than WA and GBA. for crack propagation by frequent crack deflection. As

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 9—(a) through (c) Distributions of grain boundaries (Blue lines indicate high-angle grain boundaries with misorientation h ‡ 15 deg, and
green lines indicate low-angle grain boundaries with misorientation 2 deg £ h < 15 deg); (d) through (f) Distributions of misorientation angle
(red line represents 41.85 to 46.98 deg); (g) Phase boundary densities; (h) Point-point and point-origin misorientation angle along the yellow
dashed lines in (a) through (c). (a), (d) HT0 weld; (b), (e) HT4.5 weld, and (c), (f) HT30 weld (Color figure online).

Fig. 10—Fracture surface profile and roughness calculation of the Fig. 11—Schematic diagram of the fracture mechanism for the HT30
three fracture surfaces. weld.

METALLURGICAL AND MATERIALS TRANSACTIONS A

a result, more and larger plastic deformation regions DATA AVAILABILITY
are induced, and the weld toughness improves
significantly. Data will be made available on request.

CONFLICT OF INTEREST
V. CONCLUSIONS
The authors declare that they have no conflicts of
In this study, lean S32101 DSS was welded by laser, and interest.
the PWHT process was used to tailor the microstructure
in the welds. The influence of the austenite morphology
and distribution on the impact toughness was revealed by
observing the crack propagation path using EBSD. The REFERENCES
main conclusions are as follows: 1. Z. Feng, J. Li, and Y. Wang: Steel Res. Int., 2017, vol. 88, p.
1700177.
1. Nonequilibrium solidification and nitrogen loss in 2. T. Dandekar, A. Kumar, R. Khatirkar, D. Mahadule, and G.
the laser welding pool result in a low austenite volume Ayyappan: J. Mater. Eng. Perform., 2021, vol. 30, pp. 2916–29.
fraction (17.8 pct) and nitride precipitates in the HT0 3. M. Lilias, P. Johansson, H. Liu, and C. Olsson: Steel Res. Int.,
2008, vol. 79, pp. 466–73.
weld. After PWHT, the nitride precipitates disap- 4. J. Pilhagen and R. Sandstrom: Mater. Sci. Eng. A, 2014, vol. 602,
pear, and the austenite volume fractions of the HT4.5 pp. 49–57.
and HT30 welds increase to 35.4 and 37.7 pct, 5. Y. Fang, Z. Gao, W. Bai, and P. Sun: Int. J. Mater. Res., 2018,
respectively, close to the theoretical equilibrium vol. 109, pp. 485–93.
6. N. Ouali, K. Khenfer, B. Belkessa, J. Fajoui, B. Cheniti, B. Idir,
value of 36.9 pct. The austenite in the HT30 weld and S. Branchu: J. Mater. Eng. Perform., 2019, vol. 28, pp.
distributes uniformly owing to the highest propor- 4252–64.
tion of IGA (18.6 pct) formed through nucleation 7. L. Chen, B. Wang, X. Qi, X. Shen, Q. Sun, and X. Wang: Steel
and growth of IGA and the decomposition of WA. Res. Int., 2022, vol. 93, p. 2200349.
8. S. Hu, B. Yang, B. Zeng, D. Zheng, and Z. Yang: Mater. Sci.
2. WA, one side of GBA, and IGA exhibit coherent or
Forum, 2016, vol. 861, pp. 141–46.
semicoherent ORs with the adjacent ferrite grains. 9. A. Ghosh, D. Misra, and S. Acharyya: Lasers Manuf. Mater.,
However, the additional stress generated by the 2019, vol. 6, pp. 228–46.
coherent or semi-coherent ORs at the IGA/ferrite 10. Z. Zhang, H. Jing, L. Xu, Y. Han, L. Zhao, X. Lv, and J. Zhang:
interface is relatively weak, making the cracks prone J. Manuf. Process., 2018, vol. 31, pp. 568–82.
11. F. Cao, G. Huang, W. Hou, R. Ni, T. Sun, J. Hu, Y. Shen, and A.
to deflection at the IGA/ferrite interface. Therefore, Gerlich: J. Mater. Process. Technol., 2022, vol. 307, p. 117660.
IGA has higher crack propagation resistance than 12. Z. Zhang, Y. Han, X. Lu, T. Zhang, Y. Bai, and Q. Ma: Mater.
WA and GBA. Sci. Eng. A, 2023, vol. 872, p. 144936.
3. The HT0 weld has poor impact toughness (4.0 J) due 13. M. Keskitalo, K. Mäntyjärvi, J. Sundqvist, J. Powell, and A.
Kaplan: J. Mater. Process. Technol., 2014, vol. 216, pp. 381–84.
to low austenite volume fraction and nitride precip- 14. K. Ceyhun: J. Mater. Sci., 2020, vol. 55, pp. 17232–54.
itates. The high proportions of GBA and WA in the 15. A. Magalhaes, C. Magalhaes, M. De Lima, J. Cruz, L. Godefroid,
HT4.5 weld do not improve the weld toughness R. Bertazoli, and G. De Faria: Weld. J., 2020, vol. 99, pp.
(5.1 J). The IGA with high density and uniform 185–202.
distribution in the HT30 weld causes frequent crack 16. R. Lai, Y. Cai, Y. Wu, F. Li, and X. Hua: J. Mater. Process.
Technol., 2016, vol. 231, pp. 397–405.
deflection and larger local plastic deformation, hin- 17. F. KoleniČ, L. Kovac, and D. Drimal: Weld. World, 2011, vol. 55,
dering crack propagation inside the coarse ferrite and pp. 19–25.
improving toughness significantly. 18. K. Qi, G. Wang, Y. Jin, J. Gu, Z. Zhang, and Y. Tao: Int. J. Mod.
Phys. B, 2020, vol. 34, p. 2040061.
19. T. Toth, S. Krasnorutskyi, J. Hensel, and K. Dilger: Int. J. Press.
Vessel. Pip., 2021, vol. 191, p. 104354.
20. N. Haghdadi, C. Ledermueller, H. Chen, Z. Chen, Q. Liu, X. Li,
G. Rohrer, X. Liao, S. Ringer, and S. Primig: Mater. Sci. Eng. A,
ACKNOWLEDGMENTS 2022, vol. 835, p. 142695.
21. N. Haghdadi, P. Cizek, P. Hodgson, and H. Beladi: Mater. Sci.
This work was financially supported by the Aca- Eng. A, 2018, vol. 745, pp. 369–78.
demic Start-up Foundation of Soochow University 22. X. Wu, Z. Song, J. He, H. Feng, B. Wang, and M. Wu: Mater. Sci.
Eng. A, 2022, vol. 851, p. 143654.
(NH14400323). 23. A. Mohamed, A. Mohamed, H. Abdel, A. Farahat, and A.
EINikhaily: Sci. Rep., 2023, vol. 13, p. 4592.
24. K. Ceyhun and T. Ceyhun: Mater. Sci. Eng. A, 2023, vol. 862, p.
144476.
AUTHOR CONTRIBUTIONS 25. K. Ceyhun and T. Ceyhun: Mater. Chem. Phys., 2022, vol. 289, p.
126490.
BW: investigation, methodology, analysing and dis- 26. K. Ceyhun and T. Ceyhun: J. Manuf. Process., 2022, vol. 73, pp.
cussing data, writing–original draft. YZ: investigation, 861–94.
methodology. XS: supervision, funding acquisition, 27. S. Saravanan, K. Raghukandan, and N. Sivagurumanikandan: J.
conceptualization, analysing and discussing data, Manuf. Process., 2017, vol. 25, pp. 284–89.
28. X. Xie, J. Li, W. Jiang, Z. Dong, S. Tu, X. Zhai, and X. Zhao:
review & editing. PH: investigation, methodology. QZ: Mater. Sci. Eng. A, 2020, vol. 786, p. 139426.
analyzing and discussing data, supervision. XW: ana- 29. J. Singh and A. Shahi: J. Mater. Process. Technol., 2018, vol. 272,
lysing and discussing data. pp. 137–48.

METALLURGICAL AND MATERIALS TRANSACTIONS A

30. S. Cui, Y. Shi, Y. Cui, and T. Zhu: Eng. Fail. Anal., 2018, vol. 94, 36. H. Wang, Y. Tian, Q. Ye, R. Misra, Z. Wang, and G. Wang:
pp. 226–31. Mater. Sci. Eng. A, 2019, vol. 761, p. 138009.
31. I. Alvarez-Armas, U. Krupp, M. Balbi, S. Herenu, M.C. Marinelli, 37. W. Brace: J. Geophys. Res., 1960, vol. 65, pp. 3477–80.
and H. Knobbe: Int. J. Fatig., 2012, vol. 41, pp. 95–100.
32. B. Donges, A. Giertler, U. Krupp, C. Fritzen, and H. Christ: Publisher’s Note Springer Nature remains neutral with regard to
Mater. Sci. Eng. A, 2014, vol. 589, pp. 146–52. jurisdictional claims in published maps and institutional affiliations.
33. R. Strubbia, S. Herenu, A. Giertler, I. Alvarez-Armas, and U.
Krupp: Int. J. Fatig., 2014, vol. 65, pp. 58–63.
34. X. Zhang, K. Wang, Q. Zhou, J. Kong, Y. Peng, J. Ding, C. Diao, Springer Nature or its licensor (e.g. a society or other partner) holds
D. Yang, Y. Huang, T. Zhang, and S. Williams: Mater. Sci. Eng. exclusive rights to this article under a publishing agreement with the
A, 2020, vol. 773, p. 138856. author(s) or other rightsholder(s); author self-archiving of the accepted
35. L. Karlsson and J. Börjesson: Sci. Technol. Weld. Join., 2014, vol. manuscript version of this article is solely governed by the terms of
19, pp. 318–23. such publishing agreement and applicable law.

METALLURGICAL AND MATERIALS TRANSACTIONS A