Materials Science & Engineering A 796 (2020) 140006

Contents lists available at ScienceDirect

Materials Science & Engineering A

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

Study on microstructure and tensile properties of 316L stainless steel

fabricated by CMT wire and arc additive manufacturing
C. Wang a, T.G. Liu a, P. Zhu b, Y.H. Lu a, *, T. Shoji a
a
National Center for Materials Service Safety, University of Science and Technology Beijing, Beijing, 100083, China
b
Suzhou Nuclear Power Research Institute, Suzhou, 215004, China

A R T I C L E I N F O A B S T R A C T

Keywords: An austenitic stainless steel 316L part was fabricated by cold metal transfer wire and arc additive manufacturing
Wire and arc additive manufacturing (WAAM) (CMT-WAAM), and its microstructure, microhardness and tensile properties were investigated. Results showed
316L that the as-built 316L part exhibited a multilayered structure along the building direction. In the transverse
Microstructure
direction (perpendicular to scanning direction) of each layer, there was also a multilayered structure of alter­
Mechanical properties
nating overlapping zone (OA) and re-melting zone (RA). Compared with the OA, the RA had higher ferrite
content, smaller austenite dendrite size, more dispersed orientation and lower residual stress. The overall
multilayered structure and the intra-layer non-equilibrium microstructure exhibit a great influence on the me­
chanical properties of as-built 316L part. Along the building and transverse direction, the microhardness dis­
tribution in the OA was uniform, while the RA showed a trend of lower hardness in the middle and higher
hardness on both sides of the RA layer. The effect of multilayered structure on tensile properties was stronger in
the transverse direction than that in the building direction. The deformation feature was obviously inconsistent
between the OA and RA. Local necking and fracture always occurred in the OA. Microvoids trended to initiate at
silicate impurity particles, grow into large cracks, and finally lead to material failure during tension.

1. Introduction that periodic martensite laths within the block-shaped ferrite matrix
were observed in the inner layers. Zhang et al. [10] found two different
Wire and arc additive manufacturing (WAAM) includes such benefits areas, inner layer area (ILA) and partial melting area (PMA), in the high
as high deposition efficiency, high wire utilization, short period of nitrogen austenite stainless steel parts by CMT-WAAM, which are
manufacturing, low cost, less size limitation, easy repair of parts and less formed under the two solidification models, ferrite-austenite (FA) and
environmental pollution [1–3]. It can be used for in-situ composite austenite-ferrite (AF), respectively. Chen et al. [6] studied the micro­
manufacturing and large-size parts forming. The WAAM technique is structure of 316L stainless steel produced by gas metal arc additive
widely used in aerospace [2,4], marine engineering [5] and other fields. manufacturing (GMA-AM). Under the subsequent thermal cycles, the δ
Previous studies on the WAAM fabricated materials mainly focused on phases redissolved into austenite, meanwhile the intermetallic σ phases
microstructure analysis and mechanical properties measurement under formed at γ/δ interfaces. Most δ and σ phases had strip morphologies in
different manufacturing procedures. heat affect zones (HAZ), but they were fine vermicular morphologies in
Non-equilibrium microstructure can be widely observed in the the other zones due to the thermal cycles. Though it is known that the
WAAM materials [3,6]. Wang et al. [7] studied the 316L stainless steel non-equilibrium microstructure of WAAM materials is correlated with
manufactured under different arc modes, and found that most grain the deposition process, a precision comparison analysis between the
orientations are along the building direction. Haden et al. [8] found an microstructure characteristics and the deposition process is lack, such as
obvious change of the microstructure near fusion line in 304 samples the correlations between the directional growth and deposition process.
produced by arc additive manufacturing, which is attributed to the Anisotropy of mechanical properties is another remarkable feature of
transformation of solidification mode. Ge et al. [9] gave a detail inves­ the WAAM materials [11]. Rodriguez et al. [12] studied the tensile
tigation on the microstructure of 2Cr13 specimens printed by cold metal properties of 316L samples produced by CMT-WAAM. It was found that
transfer wire and arc additive manufacturing (CMT-WAAM), showing the yield strength along the building direction was lower than that along

* Corresponding author.
E-mail address: [email protected] (Y.H. Lu).

https://doi.org/10.1016/j.msea.2020.140006
Received 4 June 2020; Received in revised form 24 July 2020; Accepted 27 July 2020
Available online 31 July 2020
0921-5093/© 2020 Elsevier B.V. All rights reserved.
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

the scanning direction, which may be related to the preferential orien­ 2.2. Microstructure characterization
tation of austenite grains due to epitaxial grain growth. However, Wu
et al. [13] found that the yield and tensile strength along the building The metallographic sample locations are shown in Fig. 1(b), where
direction was better than that along the scanning direction for 316L “A”, “B” and “C” represent the metallographic sample locations in the
samples produced by metal inert gas additive manufacturing (MIG-AM). BOT, SOT and BOS planes, respectively. The BOT plane is perpendicular
IvánTabernero et al. [1] studied the CMT 316L stainless steel, and they to the scanning direction; the SOT plane is perpendicular to the building
found that the tensile strengths were similar in different directions, but direction; the BOS plane is perpendicular to the transverse direction. The
the yield strength and elongation along the building direction were microstructure demonstrated multilayered structures in both the
lower than that along the scanning direction. The anisotropic mechan­ building direction and the transverse direction. The multilayered
ical properties of WAAM materials must be correlated with the structures showed a periodic structure in the building direction, and an
non-equilibrium microstructure, but the concrete correlation is unclear alternating structure, covering re-melting area (RA) and overlapping
[14]. area (OA), in the transverse direction.
In addition, all these published results were obtained on the basis of Metallographic samples with dimensions of 10 × 10 × 5 mm3 were
thin-walled structure samples, and there was a lack of research on the machined from the as-built 316L part using a wire-electrode cutting
block samples manufactured by WAAM. In the deposition process, the machine. According to the layer gap size and scanning step size, four
repeatedly complex thermal history, including rapidly heating and layers of weld beads were covered in metallographic samples along the
cooling, directional heat extraction, high temperature gradients, and building direction, and two pairs of RAs and OAs were covered along the
even partial re-melting, affects the microstructure and mechanical transverse direction. Each metallographic sample was cleaned, ground
properties [9]. The effect of the non-equilibrium microstructure on the and polished to a mirror finish. Subsequently, they were electrochemical
mechanical properties of 316L was unknown. In this study, based on a etched in 10% oxalic acid solution at 10 V voltage. The microstructure
316L block sample produced by CMT-WAAM, the relationships between was observed by an Olympus BX53 M optical microscope (OM) and a
the microstructure and mechanical properties would be analyzed. Zeiss Auriga scanning electron microscope (SEM). The sample located in
“A” zone was polished again with a suspension (colloidal silica polishing
2. Experiment suspension: deionized water mixed at 1:1) for 20 min for electron
backscattered diffraction (EBSD) mapping in a SIGMA 300 SEM equip­
2.1. Material fabrication ped with EDAX EBSD probe with a scanning step of 6 μm. The total
mapping area is more than 3.7 × 1.7 mm2.
316L stainless steel wire with a diameter of 1.0 mm was used to
deposit, and the substrate was stainless steel plate with dimensions of
2.3. Mechanical tests
200 × 200 × 8 mm3. Before deposition, the substrate surface was pol­
ished and cleaned. CMT advanced 4000 R was used as the deposition
Microhardness measurements were recorded at 0.2 mm intervals
power source, and CMT mode was adopted. The dwelling time between
under a 100 g load with a dwell time of 10 s in the BOT plane along the
layers was 20 s which is enough to reduce the interlayer temperature.
building direction and the SOT plane along the transverse direction.
The optimal deposition parameters were obtained from previous ex­
As shown in Fig. 1(b), vertical tensile sample (VT) and horizontal
periments: arc current 150 A, arc voltage 14.1 V, wire feeding rate 11 m/
tensile sample (HT) were cut from each plane of the as-built 316L part by
min, scanning speed 0.7 m/min, and overlapping rate 40%. The Ar-2%
a wire-electrode cutting machine, which is listed in Table 2. The tensile
O2 mixed shielding gas was used at a flow rate of 20 L/min. Each layer
samples, containing different multilayered structures, were defined with
was scanned in the same way with oscillation deposition strategy. Fig. 1
their plane, such as BOT-VT. Schematic diagram of these tensile samples
(a) shows the schematic diagram of scanning path, where the bold gray
is shown in Fig. 2. The tensile samples in the BOT plane contained an
lines indicate the fusion lines, B direction is the building direction, S
alternating structure (covering two RAs and three OAs) and a periodic
direction is the scanning direction, and T direction is the transverse
structure (covering five layers of weld beads); the tensile samples in the
direction. The scanning directions of adjacent weld beads are opposite in
the layer, but they are the same between layers. An as-built block sample
Table 1
with a geometric size of 100 × 100 × 100 mm3 (Fig. 1(b)) was fabri­
Chemical composition (wt%) of the as-built 316L part.
cated, using the above deposition parameters, and its measured chem­
ical composition (wt%) was listed in Table 1. C Cr Ni Mo Mn Si S P Fe

0.013 19.24 12.92 2.64 1.68 0.30 0.0021 0.020 Bal.

Fig. 1. (a) Schematic diagram of deposition process; (b) Sampling locations of the metallographic samples (“A”, “B” and “C”), small tensile samples (“D”) and
tensile samples.

2
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Table 2 grains due to different corrosion degrees could reflect the differences in
Tensile samples and small tensile samples in different planes. grain orientations to some extent. Fig. 4(a) and (b) show that the areas
BOT plane SOT plane BOS plane with dispersed and concentrated grain orientations appear alternatingly
along the transverse direction in the BOT and SOT planes. In Fig. 4(a),
Tensile samples BOT-VT BOT-HT SOT-VT SOT- BOS-VT BOS-
HT HT the areas with dispersed grain orientation are located at the semicircle
Small tensile RA-VT OA-VT RA-HT fusion line, where materials re-melted in deposition process; the areas
samples OA-HT with concentrated grain orientations are located at the overlapping area
of adjacent weld beads. Thus, we define the two areas as the re-melting
area (RA) and the overlapping area (OA). As shown in Fig. 4(a) and (c),
fusion lines divide the microstructure into many layers along the
building direction in the BOT and BOS planes, and the grain orientations
are similar between layers. Because the microstructure in the BOT plane
shows multilayered structures in both the building and transverse di­
rections, we focus on the microstructure in the BOT plane in the
following study.
Fig. 5 demonstrates the microstructures of two layers of the OA and
RA in the BOT plane. The black ferrite is distributed within the white
austenite matrix. As shown in Fig. 5(a), in the OA, austenite dendrites
grow along the building direction in each layer. The ferrite exhibits
skeleton morphology in the bottom part of the OA layer in Fig. 5(b); the
skeleton ferrite develops into lath ferrite in the upper part of the OA
layer in Fig. 5(d); the ferrite at the fusion lines exhibits strip morphology
with almost no secondary dendrite arm in Fig. 5(c). As shown in Fig. 5
(e), in the RA, austenite dendrites grow perpendicular to the semicircle
fusion line in each layer. The ferrite exhibits narrow granular or acicular
morphology in the bottom part of the RA layer in Fig. 5(f); the ferrite
develops into coarse lath ferrite in the upper part of the RA layer in Fig. 5
(h); the ferrite with acicular and lath morphologies is on both sides of the
Fig. 2. Schematic diagram of the tensile samples in (a) the BOT plane (defined fusion line, respectively, in Fig. 5(g). Compared with the microstructures
as BOT-VT and -HT), (b) the SOT plane (defined as SOT-VT and -HT), (c) the
in the OA and RA, both of them present periodic change in the intra-
BOS plane (defined as BOS-VT and -HT) and (d) tensile sample dimensions.
layer along the building direction. The size and morphology of ferrite
in the upper part of layer are similar, but they much differ in the bottom
SOT plane only contained the alternating structure; the tensile samples part of layer.
in the BOS plane only contained the periodic structure. All the tensile Fig. 6 shows the EBSD results of the BOT plane for as-built 316L part.
samples were tensile failure at tensile rate of 0.0025/s. As shown in Fig. 6(a), below the fusion lines of the RA, austenite
Several small tensile samples were cut from “D” zone in Fig. 1(b), columnar grains tend to grow perpendicular to the fusion lines and to­
which is listed in Table 2. Each small tensile sample only contained RA wards the center of the molten pool. The epitaxial growth of grains can
or OA in parallel section, and is defined with its area, such as RA-VT. be observed at the fusion lines, and the growth direction changes to the
Schematic diagram of the small tensile samples is shown in Fig. 3. transverse direction. Then, some small grains nucleate on epitaxial
Before tensile tests, these samples would undergo grinding and polish­ grains and continue to grow towards the center of the molten pool. In the
ing. All the tensile samples were tensile failure at a tensile rate of OA, austenite columnar grains grow along the building direction, and
0.0025/s. Zeiss Auriga SEM was used to observe the fractographic fea­ grow epitaxially at the fusion line. As shown in Fig. 6(a), the OA has
tures and the corresponding surface morphologies of the tensile samples. <001> fiber texture in the building direction, which accounts for 99.7%
of the OA. The texture contains 83.8% Goss {110} <001> component
3. Results (Fig. 6(b)) and 10.5% cubic {100} <001> component by texture
component analysis. The crystal orientation of the {110} <001> texture
3.1. Microstructure is shown in Fig. 6(d). As shown in Fig. 6(a) and (c), the RA has relatively
dispersed grain orientations. In Fig. 6(a), the white grain boundaries
Fig. 4 shows the macrostructures in the three planes of the as-built present low angle grain boundaries with a misorientation angle of
sample. The fusion lines exhibit semicircle morphology in the middle 3–15◦ , and the black grain boundaries present high angle grain ones
of the weld bead and straight-line morphology in the overlapping area in with a misorientation angle of more than 15◦ . It is found that there are
the BOT plane (Fig. 4(a)). The fusion lines exhibit corrugated more sub-grains in the OA than that of the RA. As shown in Fig. 7, most
morphology in the SOT and BOS planes (Fig. 4(b) and (c)). The grain grain boundaries are distributed in the range of 1–15◦ with low angle
orientations affect corrosion resistance, thus the diverse brightness of

Fig. 3. Schematic diagram of the (a) RA-VT; (b) OA-VT; (c) RA-HT; (d) OA-HT small tensile samples and (e) their dimensions.

3
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 4. Macrostructure of as-built 316L part in the (a) BOT, (b) SOT and (c) BOS planes.

Fig. 5. Microstructure of the OA (a–d) and RA (e–h) in the BOT plane: (a, e) overall micrographs of two layers; (b-d, f-h) high magnification micrographs of the
bottom part in the N layer, fusion line and the upper part in the N-1 layer, respectively.

grain boundaries features. The frequencies of the high angle grain direction in the BOT plane and along the transverse direction in the SOT
boundaries in the RA are higher than that in the OA. Wang et al. [15] plane. Because of the multilayered structure of microstructure, the dis­
pointed out that the recrystallized fraction of grains relates to the fre­ tribution of microhardness in the intra-layer is focused. As shown in
quency of high angle grain boundaries, and the microstructure with Fig. 10(a), along the building direction, the microhardness value of the
more high angle grain boundaries will result in enhanced strength and OA is uniform in the layer, but that of the RA shows the lower hardness
toughness. It indicates that the RA may have higher strength and in the middle and higher hardness on both sides of the RA layer. As
toughness. shown in Fig. 10(b), the microhardness value of the OA is much lower
Fig. 8 shows the kernel average misorientation (KAM) map in the than that of the RA, which forms an alternating distribution of high-and
BOT plane for the as-built 316L part. KAM map indicates the degree of low-microhardness layers along the transverse direction. The micro­
plastic deformation in the material, which is based on the nearest hardness value of the OA is uniform in the layer, but that of the RA also
neighbors of each point and with an upper limit of 5◦ . As shown in Fig. 8, shows lower hardness in the middle and higher hardness on both sides of
the relatively high KAM value in the OA and the upper part of the RA the RA layer, which is similar to the microhardness trend in the RA layer
layer seems to be related to the shrinkage of weld metal after deposition. along the building direction.
Fig. 9 shows the EBSD phase diagram and SEM micrographs of as- Fig. 11 shows the stress-strain curves of VTs and HTs in three planes,
built 316L part. EBSD phase diagram (Fig. 9(a)) indicates that the and the corresponding yield strength, tensile strength and elongation
microstructure consists of austenite (γ) and delta-ferrite (δ). The low values are summarized in Table 4. The tensile properties of the BOT and
carbon concentration in Table 1 leads to no carbide for as-built 316L BOS planes show obvious anisotropy. The tensile strength differences of
part. As shown in Fig. 9(b), many spherical particles with a diameter of VT and HT are 9.1% and 9.6% in the BOT and BOS planes, respectively,
about 1 μm can be observed in the austenite matrix. The elements of but that in the SOT plane is only 1.44%. It is related to the slender
these particles are listed in Table 3, which were detected by energy austenite grains morphologies due to the directional growth in the
dispersive spectroscopy (EDS). The particles contain higher O, Si and Mn building direction. In the BOS plane, though the microstructure has a
elements, and should be silicate impurities. periodic structure along the building direction, which may not affect the
tensile properties. And the BOS-VT and BOS-HT have the highest elon­
gation. In the BOT and SOT planes, the microstructure has an alternating
3.2. Mechanical properties structure along the transverse direction, which leads to the great
decrease of elongation of all tensile samples in these planes. In the BOT
Fig. 10 shows the microhardness distribution along the building

4
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 6. (a) EBSD inverse pole figure with <uvw> directions plotted || BD, pole figure of the (b) OA and (c) RA, as well as (d) crystal orientation in pentagram location
of the OA.

plane, there is also a periodic structure along the building direction, surface cracks appear in the OA, which indicates the bad elongation of
which leads to the further decrease of elongation in the building the OA. Both the BOT-HT (Fig. 12(a)) and SOT-HT (Fig. 12(b)) show the
direction. inconsistent deformation of alternating structure, and obvious local
Fig. 12 shows the overall deformation of tensile samples in three necking can be observed in the OA. But the periodic structure may not
planes. The BOT-VT shows the inconsistent deformation of both alter­ affect the samples in the BOS plane in Fig. 12(c), and the deformation is
nating and periodic structures in Fig. 12(a), and folds appear periodi­ uniform. Therefore, the effect of the alternating structure on the tensile
cally in the RA along the building direction. The SOT-VT shows the property is stronger than that of the periodic structure, which results in
inconsistent deformation of alternating structures in Fig. 12(b), and the obvious local necking or fracture of the OA, and the decrease of

5
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 7. Misorientation angle distribution of the (a) OA and (b) RA in the BOT plane.

Fig. 8. Kernel average misorientation map in the BOT plane for as-built 316L part.

Fig. 9. (a) EBSD phase diagram and (b) SEM micrographs of as-built 316L part.

elongation in the BOT and SOT planes. The tensile properties of the OA
Table 3
and RA in the BOT plane will be discussed in the following study, and the
EDS results of the particle elements in Fig. 9(b).
effect of the multilayered structure on different tensile properties will be
Element/wt% O Si Mo Cr Mn Fe Ni analyzed.
Point 1 5.61 3.31 1.10 16.52 6.14 56.14 11.20 Fig. 13 shows the stress-strain curves of the small tensile samples in
Point 2 6.75 4.12 1.02 16.44 6.97 54.40 10.31 the BOT plane. Compared with the tensile samples in the BOT plane,
Point 3 5.09 3.16 1.54 17.25 5.95 56.05 10.96 small tensile samples show totally different tensile properties. On the
Point 4 3.88 1.94 1.35 17.37 4.38 59.10 11.98
whole, the RA has higher elongation than the OA, and the HTs have
higher elongation than the VTs. In the OA, the OA-VT has higher tensile
strength and lower elongation than the OA-HT. But in the RA, the RA-HT
has higher tensile strength and higher elongation than the RA-VT.

6
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 10. Microhardness distribution of the RA and OA in the intra-layer along the (a) building and (b) transverse directions.

direction. The slip bands appear preferentially in the austenite den­

drites, and they are perpendicular to the gain growth direction. Some
slip bands connect through the dendrite boundaries, and most slip bands
are inhibited. In the left grain of Fig. 15(d), the inhibition of the ferrite or
dendrite boundaries to the slip bands are not observed. In HTs, multiple
slip systems are activated and deform the samples, but in VTs, the ferrite
and dendrite boundaries inhibit the extension of slip bands and reduce
the deformation of samples, which is consistent with the results in
Fig. 13. In Fig. 15(e), higher magnification micrograph shows that
microvoids initiate at the silicate impurity particle with a diameter of
about 1 μm.
Fig. 16 shows the corresponding fracture sections of small tensile
samples in the BOT plane. There are a lot of dimples at the fracture in all
samples, indicating that the fracture modes belong to plastic fracture.
The size of dimples in the RA (Fig. 16(a) and (c)) are smaller than that in
the OA (Fig. 16(b) and (d)), which indicates that the OA may have
higher tensile strength than that of the RA. As shown in Fig. 16(a) and
(c), the fracture section of the RA is relatively flat and the dimples are
Fig. 11. Stress-strain curves of VTs and HTs in three (BOT, SOT and elongated. The dimples in the RA may be tear dimples caused by the
BOS) planes. cracks on the surface.

4. Discussion
Table 4
Tensile properties of VTs and HTs in three (BOT, SOT and BOS) planes.
4.1. Microstructure
Yield strength/MPa Tensile strength/MPa Elongation/%

BOT-HT 400.6 586.9 28.45 The multilayered structure of microstructure was the most important
BOT-VT 346.5 545.3 31.06 feature of the CMT-WAAM 316L samples, which was caused by the
SOT-HT 384.8 563.1 28.42
complex thermal histories [16]. The multilayered structure included
SOT-VT 382.8 555.1 37.25
BOS-HT 394.1 577.5 41.87 change of ferrite, austenite, and residual stress. In this study, it was
BOS-VT 361.3 526.9 46.24 focused on the multilayered microstructures in the BOT plane, the pe­
riodic structure along the building direction and the alternating struc­
ture along the transverse direction.
Fig. 14 shows the macro fracture surface morphologies of the small
As shown in Fig. 5, the ferrite showed great differences in the OA and
tensile samples in BOT plane. As shown in Fig. 14(a) and (b), the RA-VT
RA, and periodically changed in the layers along the building direction.
has obvious local necking at the fracture location, but the OA-VT has
It was reported that, in the austenitic stainless steel weld metals, the
little necking, which leads to the bad elongation. It is consistent with the
ferrite content and morphology were affected by the peak temperature,
results in Fig. 13. In Fig. 14(c) and (d), both the RA-HT and OA-VT have
cooling rate, and thermal cycles during nonequilibrium solidification at
large deformation in the parallel section, resulting in higher elongation.
high cooling rates [17,18]. The behaviors of residual ferrite with cooling
Fig. 15 shows corresponding high magnification micrographs of
rate were related to the primary solidification mode [19]. According to
micro fracture surface morphologies of small tensile samples in Fig. 14.
the chemical composition of the material, there were different solidifi­
In all micrographs, it is found that folds form nearly perpendicular to the
cation modes in the liquid phase [20,21]:
tensile direction, and impurity particles separate from matrix to form
microvoids at the white arrows. As shown in Fig. 15(a) and (b), the
1) Austenite mode: L→L + γ→γ Creq/Nieq<1.25
ferrite protrusions, growing along the tensile direction, inhibits the
2) Austenite-Ferrite mode: L→L + γ→L + δ + γ→γ + δ 1.25<Creq/
extension of the slip bands. In Fig. 15(c), cracks form from the micro­
Nieq<1.48
voids and extend along the fold. In the right grain of Fig. 15(d), the
3) Ferrite-Austenite mode: L→L + δ→L + δ + γ→γ + δ 1.48<Creq/
growth direction of the austenite dendrites is close to the tensile
Nieq<1.95

7
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 12. Overall deformation of the (a) BOT-VT and -HT, (b) SOT-VT and -HT, and (c) BOS-VT and -HT, respectively.

direction (Fig. 5(e)). The result was agreement with the study of Haden
et al. [8], which reveals a repeat microstructure shift near the interface
and links the abrupt texture changes with the local thermal histories.
Elmer et al. [19] also observed the change of solidification mode in
re-melting area, and the areas with austenite-ferrite solidification mode
were larger with a higher scan speed. The OA had a lower temperature
gradient and cooling rate than that of the RA, which led to larger
dendrite spacing and smaller areas with austenite-ferrite solidification in
the bottom part of the OA layer. But, compared with the upper part of
the OA layer, there was higher cooling rate in the bottom part of the OA
layer, which caused lower primary dendritic spacing and more incon­
spicuous secondary dendritic arm [24]. As a result, the ferrite showed
strip morphology without secondary dendritic arm near the fusion lines
(Fig. 5(c)), and gradually changed from the skeleton morphology in the
bottom part of the OA layer (Fig. 5(b)) into the lath morphology in upper
part of the OA layer (Fig. 5(d)) along the building direction. The results
were consistent with the study of Wang et al. [7], which reveals the
microstructure change along the building from fine columnar structures
Fig. 13. Stress-strain curves of small tensile samples in the BOT plane. without secondary dendrites to coarse columnar structures with sec­
ondary dendrites.
4) Ferrite mode: L→L + δ→δ→δ + γ Creq/Nieq>1.95 Grains usually tended to grow in a competitive grain growth process.
The easy growth direction of FCC crystal was <001> [27], because
Where Creq/Nieq value is calculated by the following equations: Creq grains would grow faster along the direction containing fewer atoms
= %Cr + %Mo + 1.5%Si + 0.5%Nb; Nieq = %Ni + 30%C + 0.5%Mn [28]. Most grains grew along the easy growth direction, that corre­
[22]. According to the results in Table 1, the Creq/Nieq value in this study sponded with the thermal gradient of the solidifying liquid [3,16,29].
was 1.58 (Creq/Nieq = 22.33/14.15), so the solidification mode of Therefore, the grains grew perpendicular to the fusion lines in the RA, as
as-built 316L part was ferrite-austenite mode. But higher cooling rate in shown in Fig. 6. Fig. 17(a) shows that the semicircle fusion line of the RA
the middle of the weld bead might lead the liquid metal to undercool to led to the dispersed orientation, which caused more high angle grain
below the austenite solvus temperature, and the solidification mode boundaries in the RA (Fig. 7). In the OA, strong texture was observed,
switched from primary ferrite to primary austenite solidification mode which was caused by the directional temperature gradient and heat flow
[23]. [11,30]. The grains of the OA were affected by the adjacent weld beads,
In the solidification process of liquid metal, the morphology and size which resulted in that the heat flow was along the building direction,
of grains were determined by temperature gradient and cooling rate [24, and the grain orientation was <001> || the building direction (Fig. 17
25]. In the deposition process, the temperature of weld bead presented (b)). But in the cross section of the building direction, grain orientation
Gauss distribution curve in the cross-section was perpendicular to was related to the scanning speed, which was between the scanning
scanning direction [10], so the highest temperature gradient was located direction and the transverse direction [31]. In the OA, the most grain
in the RA. Higher temperature gradient and cooling rate of the RA led to orientation was close to 45◦ from the scanning direction, forming the
the austenite-ferrite solidification mode. As shown in Fig. 5(f), granular most {110} <001> texture. The growth direction had a few deviations
ferrite or acicular ferrite formed in the bottom of the RA layer. As the from the building direction, which was attributed to deviation of the
liquid front advanced and the melt pool shrunk, the cooling rate maximum temperature gradient from the building direction [16]. The
decreased rapidly [26], and austenite-ferrite solidification mode trans­ texture in the OA caused more low angle grain boundaries, as shown in
formed into ferrite-austenite solidification mode. As shown in Fig. 5 (h), Fig. 7.
lath ferrite formed in the upper part of the RA layer, which resulted in
periodic change of ferrite morphology in the RA layer along the building

8
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 14. Macro fracture surface morphologies of the (a) RA-VT, (b) OA-VT, (c) RA-HT, and (d) OA-HT.

4.2. Mechanical properties results were consistent with tensile properties of the small tensile sam­
ples in Fig. 13, which was caused by the different microstructures in the
As shown in Fig. 10, along the building direction and transverse di­ OA and RA. On the one hand, the OA had more low angle grain
rection, the microhardness distribution in the OA was uniform, and the boundaries than RA (Fig. 7), which decreased the dislocation activity
RA showed a trend of lower hardness in the middle and higher hardness [35], and thus restrained more slip systems [36,37], as a result, the
on both sides of the RA layer. The ferrite morphology and size of the RA elongation of the OA decreased. On the other hand, the ferrite [38] and
and the OA (Fig. 5) in the upper part of layer were similar, leading to the dendrite boundaries [39] perpendicular to the tensile direction arrested
similar microhardness value of the RA and the OA near the fusion lines. slip bands expansion and crack growth (Fig. 15). Because the main grain
As shown in Fig. 5(f), granular or acicular ferrite in the bottom part of growth direction in the OA was along the building direction, and the
the RA layer with a higher ferrite content increased the microhardness grains growth direction in the RA was perpendicular to the fusion lines,
value. As shown in Fig. 8, the shrinkage of weld metal after deposition the effect of the ferrite and dendrite boundaries was stronger in the OA
caused the higher residual compressive stress in the upper part of the RA than that in the RA. As a result, the OA-VT had the lower elongation and
layer [32], which increased the microhardness. Therefore, under the higher tensile strength than that of the RA-VT.
effects of ferrite and residual stress, the microhardness in the RA showed Both the BOS-HT and SOT-HT showed local necking in the OA after
lower hardness in the middle and higher hardness on both sides of the fracture, which indicates the lower strength of the OA than that of the
RA layer in the building direction. As shown in Figs. 5 and 6, in the RA. As shown in Fig. 15, the ferrite and dendrite boundaries in OA-HT
transverse direction, the RA had a higher ferrite content and a smaller and RA-HT didn’t arrest the slip bands. Therefore, compared with the
austenite grain size than that of the OA in most areas, which increased OA-VT, the tensile strength decreased and the elongation increased in
the microhardness value of the RA. And in Fig. 8, a higher residual the OA-HT. However, the tensile strength of the RA-HT was higher than
tensile stress appeared in the OAs between two adjacent weld beads, that of the RA-VT, which was caused by the high angle grain boundary in
which was also observed by previous studies [33]. It reduced the the RA. The average grain diameter of the RA along the building di­
microhardness value of the OA. Therefore, the microhardness value of rection was larger than that along the transverse direction due to the
the RA was much higher than that of the OA in the transverse direction. slender grain morphology. The smaller effective grain size in the
As shown in Fig. 6, the grain orientation in the middle position of the RA transverse direction increased the stress concentration area caused by
layer in transverse direction was closed to the <100> direction. It was dislocation pile up [38]. Therefore, the tensile strength of the RA-HT
believed that the microhardness of (100) crystal surface was relatively was higher than that of the RA-VT. Meanwhile, the high angle grain
lower because (100) crystal surface was a non-close-packed plane with boundaries arrested the crack propagation and made the material bear
large bond length and relatively weaker atomic binding force [34]. larger plastic deformation before fracture [24,40], which caused that
Therefore, under the effects of ferrite and grain orientation, the micro­ the higher elongation of the RA-HT. As a result, compared with the
hardness in the RA also showed lower hardness in the middle and higher RA-VT, the tensile strength and the elongation increased in the RA-HT,
hardness on both sides of the RA layer in the transverse direction. which also caused the tensile strength of RA-HT was higher than that of
Multilayered structures played a great role in the tensile properties of the OA-HT, and local necking occurred in the OA, as shown in Fig. 12(a)
the CMT-WAAM 316L samples. As shown in Fig. 12(a) and (b), there was and (b).
obviously inconsistent deformation between the OA and RA in the BOT The periodic structure didn’t affect the samples in the BOS plane in
and SOT planes. The BOS-VT showed relief morphologies in the RA, and Fig. 12(c), and the deformation was uniform. Therefore, the BOS-VT and
the SOT-VT showed surface cracks in the OA after fracture, which in­ -HT had the highest elongation.
dicates that the elongation of the RA is higher than that of the OA. The Oxidizing impurities commonly acted as a brittleness phase and

9
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 15. Corresponding high magnification micrographs of micro fracture surface morphologies of the (a) RA-VT, (b) OA-VT, (c) RA-HT, and (d) OA-HT; (e) higher
magnification micrograph of the RA-HT.

ultimately became the initiation sites of cracks [7]. As shown in Fig. 9, area (OA). The RA, in which the grains grew perpendicular to the
many silicate impurity particles existed in as-built 316L part. Because of fusion lines, had finer columnar grains and higher fraction of high
the low plasticity and high strength of impurity particles, the large angle grain boundaries. The OA, in which the grains grew along the
plastic difference between inclusions and austenite resulted in the sep­ building direction, had large-sized directional growth grains with a
aration of inclusions/austenite interface, and formation of many strong texture.
microvoids (Fig. 15(e)). These microvoids could form cracks and expand 2. The as-built 316L part performed non-equilibrium mechanical
along the fold perpendicular to the tensile direction. The ferrite and properties corresponding to the microstructure distribution. The RA
dendrite boundaries could inhibit the slip band extension and crack had higher yield and tensile strength in transverse direction, which
growth, which improved the tensile strength. was opposed to the OA. The RA had higher elongation than the OA,
and both of them showed higher elongation in the transverse direc­
5. Conclusions tion than that in the building direction. Additionally, along the
building direction and transverse direction, the microhardness dis­
In this study, the microstructure and mechanical properties of an as- tribution in the OA was uniform, while the RA showed a trend of
built 316L part manufactured by cold metal transfer wire and arc ad­ lower hardness in the middle and higher hardness on both sides of
ditive manufacturing (CMT-WAAM) were discussed in detail. The con­ the RA layer.
clusions can be summarized as follows: 3. Many silicate impurity particles were scattered in the austenite ma­
trix, which would become the crack initiation sites according to
1. The as-built 316L showed multilayered microstructures both in the surface observation.
building direction and transverse direction. In the building direction, 4. The effect of multilayered structure on the tensile property was
epitaxial growth of grains was observed, and ferrite morphology stronger in the transverse direction than that in the building direc­
showed periodic change. In the transverse direction, there were two tion, and obviously inconsistent deformation between RA and OA
alternating microstructures, re-melting area (RA) and overlapping

10
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

Fig. 16. Fracture sections of the (a) RA-VT, (b) OA-VT, (c) RA-HT, and (d) OA-HT.

Fig. 17. Schematic diagram of grains growing in the (a) RA and (b) OA.

was observed in fractured tensile samples. Local necking or fracture References

always occurred in the OA.
[1] A. Paskual IvánTabernero, P. Álvarez, A. Suárez, Study on arc welding processes for
high deposition rate additive manufacturing, Procedia CIRP 68 (2018) 358–362.
CRediT authorship contribution statement [2] S.W. Williams, F. Martina, A.C. Addison, J. Ding, G. Pardal, P. Colegrove, Wire +
arc additive manufacturing, Mater. Sci. Technol. 32 (7) (2016) 641–647.
C. Wang: Conceptualization, Methodology, Data curation, Writing - [3] C.R. Cunningham, J.M. Flynn, A. Shokrani, V. Dhokia, S.T. Newman, Invited
review article: strategies and processes for high quality wire arc additive
original draft. T.G. Liu: Data curation, Writing - review & editing. P. manufacturing, Addit. Manuf. 22 (2018) 672–686.
Zhu: Resources, Writing - review & editing. Y.H. Lu: Writing - review & [4] F. Martina, J. Mehnen, S.W. Williams, P. Colegrove, F. Wang, Investigation of the
editing. T. Shoji: Writing - review & editing. benefits of plasma deposition for the additive layer manufacture of Ti–6Al–4V,
J. Mater. Process. Technol. 212 (6) (2012) 1377–1386.
[5] A. Queguineur, G. Rückert, F. Cortial, J.Y. Hascoët, Evaluation of wire arc additive
Declaration of competing interest manufacturing for large-sized components in naval applications, Weld. World 62
(2) (2017) 259–266.
[6] X. Chen, J. Li, X. Cheng, B. He, H. Wang, Z. Huang, Microstructure and mechanical
The authors declare that they have no known competing financial
properties of the austenitic stainless steel 316L fabricated by gas metal arc additive
interests or personal relationships that could have appeared to influence manufacturing, Mater. Sci. Eng. A 703 (2017) 567–577.
the work reported in this paper. [7] L. Wang, J. Xue, Q. Wang, Correlation between arc mode, microstructure, and
mechanical properties during wire arc additive manufacturing of 316L stainless
steel, Mater. Sci. Eng. A 751 (2019) 183–190.
Acknowledgements [8] C.V. Haden, G. Zeng, F.M. Carter, C. Ruhl, B.A. Krick, D.G. Harlow, Wire and arc
additive manufactured steel: tensile and wear properties, Addit. Manuf. 16 (2017)
This work was supported by the Beijing Municipal Science & Tech­ 115–123.
[9] J. Ge, J. Lin, H. Fu, Y. Lei, R. Xiao, A spatial periodicity of microstructural
nology Commission (Z181100005218005), National Key Research and evolution and anti-indentation properties of wire-arc additive manufacturing
Development Program of China (2019YFB1900905) and National Nat­ 2Cr13 thin-wall part, Mater. Des. 160 (2018) 218–228.
ural Science Foundation of China (51701017). [10] X. Zhang, Q. Zhou, K. Wang, Y. Peng, J. Ding, J. Kong, S. Williams, Study on
microstructure and tensile properties of high nitrogen Cr-Mn steel processed by
CMT wire and arc additive manufacturing, Mater. Des. 166 (2019).

11
C. Wang et al. Materials Science & Engineering A 796 (2020) 140006

[11] F. Khodabakhshi, M.H. Farshidianfar, A.P. Gerlich, M. Nosko, V. Trembošová, parameters on tailoring the grain morphology of IN718 in electron beam additive
A. Khajepour, Effects of laser additive manufacturing on microstructure and manufacturing, Acta Mater. 112 (2016) 303–314.
crystallographic texture of austenitic and martensitic stainless steels, Addit. Manuf. [27] S.A. David, J.M. Vitek, M. Rappaz, L.A. Boatner, Microstructure of stainless steel
31 (2020). single-crystal electron beam welds, Metall. Trans. A 21 (6) (1990) 1753–1766.
[12] N. Rodriguez, L. Vázquez, I. Huarte, E. Arruti, I. Tabernero, P. Alvarez, Wire and [28] O. Andreau, I. Koutiri, P. Peyre, J.-D. Penot, N. Saintier, E. Pessard, T. De Terris,
arc additive manufacturing: a comparison between CMT and TopTIG processes C. Dupuy, T. Baudin, Texture control of 316L parts by modulation of the melt pool
applied to stainless steel, Weld. World 62 (5) (2018) 1083–1096. morphology in selective laser melting, J. Mater. Process. Technol. 264 (2019)
[13] W. Wu, J. Xue, L. Wang, Z. Zhang, Y. Hu, C. Dong, Forming process, 21–31.
microstructure, and mechanical properties of thin-walled 316L stainless steel using [29] X. Xu, S. Ganguly, J. Ding, S. Guo, S. Williams, F. Martina, Microstructural
speed-cold-welding additive manufacturing, Metals 9 (1) (2019). evolution and mechanical properties of maraging steel produced by wire + arc
[14] J.L. Prado-Cerqueira, A.M. Camacho, J.L. Dieguez, A. Rodriguez-Prieto, A. additive manufacture process, Mater. Char. 143 (2018) 152–162.
M. Aragon, C. Lorenzo-Martin, A. Yanguas-Gil, Analysis of favorable process [30] S. Bahl, S. Mishra, Y. K U, I. Kola, K. Chatterjee, S. Suwas, Non-equilibrium
conditions for the manufacturing of thin-wall pieces of mild steel obtained by wire microstructure, crystallographic texture and morphological texture synergistically
and arc additive manufacturing (WAAM), Materials (Basel) 11 (8) (2018). result in unusual mechanical properties of 3D printed 316L stainless steel, Addit.
[15] Q. Wang, S. Zhang, C. Zhang, C. Wu, J. Wang, J. Chen, Z. Sun, Microstructure Manuf. 28 (2019).
evolution and EBSD analysis of a graded steel fabricated by laser additive [31] S. Chen, G. Guillemot, C.-A. Gandin, Three-dimensional cellular automaton-finite
manufacturing, Vacuum 141 (2017) 68–81. element modeling of solidification grain structures for arc-welding processes, Acta
[16] P. Collins, D. Brice, P. Samimi, I. Ghamarian, H.L. Fraser, Microstructural control of Mater. 115 (2016) 448–467.
additively manufactured metallic materials, Annu. Rev. Mater. Res. 46 (2016). [32] P. Rangaswamy, M.L. Griffith, M.B. Prime, T.M. Holden, R.B. Rogge, J.M. Edwards,
[17] S. Katayama, T. Fujimoto, A. Matsunawa, Correlation Among Solidification R.J. Sebring, Residual stresses in LENS® components using neutron diffraction and
Process, Microstructure, Microsegregation and Solidification Cracking contour method, Mater. Sci. Eng. A 399 (1–2) (2005) 72–83.
Susceptibility in Stainless Steel Weld Metals(Materials, Metallurgy & Weldability), [33] F. Liu, X. Lin, G. Yang, M. Song, J. Chen, W. Huang, Microstructure and residual
1985, p. 14. stress of laser rapid formed Inconel 718 nickel-base superalloy, Optic Laser.
[18] M.H. Chen, C.P. Chou, Effect of thermal cycles on ferrite content of austenitic Technol. 43 (1) (2011) 208–213.
stainless steel, Sci. Technol. Weld. Join. 4 (1) (1999) 58–62. [34] Z. Guo, X. Wang, X. Yang, D. Jiang, X. Ma, H. Song, Relationships between Young’s
[19] J.W. Elmer, S.M. Allen, T.W. Eagar, Microstructural development during modulus, hardness and orientation of grain in polycrystalline copper, Acta Metall.
solidification of stainless steel alloys, Metall. Trans. A 20 (10) (1989) 2117–2131. Sin. 44 (2008) 901–904. Jinshu Xuebao.
[20] A. Schino, M. Mecozzi, M. Barteri, J. Kenny, Solidification mode and residual [35] J. Shi, Predictive Microstructural Modeling of Grain-Boundary Interactions and
ferrite in low-Ni austenitic stainless steels, J. Mater. Sci. 35 (2000) 375–380. Their Effects on Overall Crystalline Behavior, 2009.
[21] J.W. Fu, Y.S. Yang, J.J. Guo, W.H. Tong, Effect of cooling rate on solidification [36] J.B. Jordon, J.B. Gibson, M.F. Horstemeyer, H.E. Kadiri, J.C. Baird, A.A. Luo, Effect
microstructures in AISI 304 stainless steel, Mater. Sci. Technol. 24 (8) (2008) of twinning, slip, and inclusions on the fatigue anisotropy of extrusion-textured
941–944. AZ61 magnesium alloy, Mater. Sci. Eng. A 528 (22–23) (2011) 6860–6871.
[22] Delong, Ferrite in austenitic stainless steel weld metal, Welding Res. Suppl. (1974) [37] A. Yadollahi, N. Shamsaei, S.M. Thompson, D.W. Seely, Effects of process time
53. interval and heat treatment on the mechanical and microstructural properties of
[23] S. Kou, Y. Le, The effect of quenching on the solidification structure and direct laser deposited 316L stainless steel, Mater. Sci. Eng. A 644 (2015) 171–183.
transformation behavior of stainless steel welds, Metall. Trans. A 13 (7) (1982) [38] R.G. Thomas, Effect of delta -ferrite on the creep rupture properties of austenitic,
1141–1152. Weld Metals 57 (1978) 81s–86s.
[24] K. Zhang, S. Wang, W. Liu, X. Shang, Characterization of stainless steel parts by [39] A.M. Mullis, D.J. Walker, S.E. Battersby, R.F. Cochrane, Deformation of dendrites
laser metal deposition shaping, Mater. Des. 55 (2014) 104–119. by fluid flow during rapid solidification, Mater. Sci. Eng. A 304–306 (2001)
[25] W. Ou, T. Mukherjee, G.L. Knapp, Y. Wei, T. DebRoy, Fusion zone geometries, 245–249.
cooling rates and solidification parameters during wire arc additive manufacturing, [40] N. Shamsaei, A. Yadollahi, L. Bian, S.M. Thompson, An overview of Direct Laser
Int. J. Heat Mass Tran. 127 (2018) 1084–1094. Deposition for additive manufacturing; Part II: mechanical behavior, process
[26] N. Raghavan, R. Dehoff, S. Pannala, S. Simunovic, M. Kirka, J. Turner, N. Carlson, parameter optimization and control, Addit. Manuf. 8 (2015) 12–35.
S.S. Babu, Numerical modeling of heat-transfer and the influence of process

12