Enhancement of fatigue resistance of additively manufactured 304L SS by unique heterogeneous microstructure

Abstract

Selective laser melting (SLM), as a revolutionary technology for metal manufacturing, attracts tremendous attention because it can produce complex components to benefit the customised production. Here we report that additively manufactured 304L austenitic stainless steel (SS) with low stacking fault energy (SFE) show superior fatigue resistance than its conventional counterparts due to the unique heterogeneous microstructure despite containing relatively high porosity. A series of detailed microstructural characterisations were applied to systematically disclose the fatigue enhancement mechanism of additively manufactured parts. Direct evidence is offered to show the obvious progressive work hardening and strain rate hardening caused by the heterogeneous microstructure during cyclic deformation, thus enhancing the fatigue crack initiation resistance. The microstructural results reveal that the cellular substructure plays a decisive role in regulating the dislocation motion during cyclic deformation, resulting in the intergranular fatigue cracking along HAGBs rather than twin boundaries.

Full text

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=nvpp20
Virtual and Physical Prototyping
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/nvpp20
Enhancement of fatigue resistance of additively
manufactured 304L SS by unique heterogeneous
microstructure
Hongzhuang Zhang , Mengtao Xu , Punit Kumar , Changyou Li , Weibing Dai ,
Zhendong Liu , Zhenyuan Li & Yimin Zhang
To cite this article: Hongzhuang Zhang , Mengtao Xu , Punit Kumar , Changyou Li , Weibing Dai ,
Zhendong Liu , Zhenyuan Li & Yimin Zhang (2021): Enhancement of fatigue resistance of additively
manufactured 304L SS by unique heterogeneous microstructure, Virtual and Physical Prototyping,
DOI: 10.1080/17452759.2021.1881869
To link to this article: https://doi.org/10.1080/17452759.2021.1881869
Published online: 12 Feb 2021.
Submit your article to this journal
Article views: 84
View related articles
View Crossmark data

Enhancement of fatigue resistance of additively manufactured 304L SS by unique
heterogeneous microstructure
Hongzhuang Zhang
a
, Mengtao Xu
a
, Punit Kumar
b
, Changyou Li
a
, Weibing Dai
a
, Zhendong Liu
a
, Zhenyuan Li
a
and Yimin Zhang
c
a
School of Mechanical Engineering and Automation, Northeastern University, Shenyang, People’s Republic of China;
b
School of Mechanical
and Aerospace Engineering, Nanyang Technological University, Singapore, Republic of Singapore;
c
College of Mechanical and Automotive
Engineering, Zhaoqing University, Zhaoqing, People’s Republic of China
ABSTRACT
Selective laser melting (SLM), as a revolutionary technology for metal manufacturing, attracts
tremendous attention because it can produce complex components to benefit the customised
production. Here we report that additively manufactured 304L austenitic stainless steel (SS) with
low stacking fault energy (SFE) show superior fatigue resistance than its conventional
counterparts due to the unique heterogeneous microstructure despite containing relatively high
porosity. A series of detailed microstructural characterisations were applied to systematically
disclose the fatigue enhancement mechanism of additively manufactured parts. Direct evidence
is offered to show the obvious progressive work hardening and strain rate hardening caused by
the heterogeneous microstructure during cyclic deformation, thus enhancing the fatigue crack
initiation resistance. The microstructural results reveal that the cellular substructure plays a
decisive role in regulating the dislocation motion during cyclic deformation, resulting in the
intergranular fatigue cracking along HAGBs rather than twin boundaries.
ARTICLE HISTORY
Received 1 December 2020
Accepted 24 January 2021
KEYWORDS
Additive manufacturing;
fatigue resistance;
microstructure; working
hardening mechanisms
1. Introduction
Austenitic stainless steels (SS), such as 316L and 304L,
are widely applied in kitchen tools, medical implants,
oil drilling rigs, nuclear power plants, and other fields
due to their excellent corrosion resistance (Shen et al.
2015; Koyama et al. 2017; Ferreri et al. 2020). Additive
manufacturing (AM), as a disruptive technology in
modern industries, can produce metal parts with custo-
mised geometries that are constrained by the design of
traditional processes, thereby attracting extensive atten-
tion and researches (MacDonald and Wicker 2016;
Martin et al. 2017; Khairallah et al. 2020). In recent
studies, austenitic stainless steels via a laser powder-
bed-fusion (PBF) technique have shown both excellent
ductility and tensile strength, which overcame the
strength-ductility trade-offthat commonly existed in
pure metals and alloys (Liu et al. 2018; Wang et al.
2018). However, due to the existence of intrinsic AM
defects, the fatigue properties of additively manufac-
tured austenitic steels are seriously affected and need
to be deeply understood (Sanaei and Fatemi 2020).
Preventing dislocation motion through adjusting
microstructures is widely applied in traditional materials
to improve the strength of metal, but it sacrifices the
useful ductility, which is called strength–ductility trade-
off(Kumar, Van Swygenhoven, and Suresh 2003;Wei
et al. 2014; Li et al. 2016; Liu et al. 2018). Surprisingly,
the process of selective laser melting (SLM) of PBF
© 2021 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Changyou Li chyli@mail.neu.edu.cn School of Mechanical Engineering and Automation, Northeastern University, Shenyang 110819,
People’s Republic of China; Yimin Zhang zhangyimin@zqu.edu.cn College of Mechanical and Automotive Engineering, Zhaoqing University, Zhaoqing
526061, People’s Republic of China
VIRTUAL AND PHYSICAL PROTOTYPING
https://doi.org/10.1080/17452759.2021.1881869

technique can produce special microstructure due to its
local rapid cooling rate and repeated thermal cycle (Liu
et al. 2018), which makes the SLMed parts have extre-
mely superior static tensile properties far exceeding tra-
ditional metallurgical processing (Herzog et al. 2016).
However, surface roughness (Maleki et al. 2020), gas por-
osity (Svensson, Ackelid, and Ab 2010), keyhole porosity
(Cunningham et al. 2019), lack-of-fusion (LoF) defects
(Darvish, Chen, and Pasang 2016), and residual stress
(Bartlett and Li 2019) induced during AM process can
significantly reduce the fatigue performance of addi-
tively manufactured parts. According to (Zhang et al.
2020a), internal defects and surface defects have been
proved to have more critical impact on fatigue life of
additively manufactured 304L SS than residual stress.
The irregular LoF pores induced by un-optimised
process parameters significantly facilitate fatigue crack
initiation and a lesser extent fatigue crack propagation
(Smith et al. 2019), while the small spherical defects
(<100 μm in diameter) have less effect on fatigue prop-
erties than LoF defects. Therefore, the occurrence of the
LoF pores should be minimised by optimisation of
process parameters to reduce the overall porosity, thus
enhancing the fatigue strength of additively manufac-
tured stainless steels (Kumar et al. 2020). Furthermore,
the position and size of internal pores can also affect
the high cycle fatigue resistance of additively manufac-
tured 316L SS (Andreau et al. 2021). Surface pores are
far more deleterious than internal pores because
fatigue cracks usually originate from the surface. There-
fore, abundant surface treatment methods were applied
to eliminate/minimize the surface defects, thereby
improving fatigue performance of additively manufac-
tured austenitic stainless steels (Maleki et al. 2020). In
addition, some studies have adopted post-heat treat-
ment to adjust the microstructure in an attempt to
improve the fatigue properties. The as-built and stress-
relieved annealed SLMed 316L SS samples can exhibit
better fatigue performance than traditionally manufac-
tured counterparts, which was presumed to be caused
by the strong resistance of cellular substructure to dislo-
cation flow and crack initiation (Elangeswaran et al.
2020). The solute annealing or hot isostatic pressed
(HIP) can enhance the ductility and high cycle fatigue
performance due to the coarse-grained microstructures
and residual stress relaxation (Leuders et al. 2013).
According to the above researches (Cao et al. 2020; Elan-
geswaran et al. 2020; Kumar et al. 2020; Pegues, Roach,
and Shamsaei 2020), we found that the as-built AMed
austenitic stainless steel can show superior fatigue per-
formance than solute annealing samples and its tra-
ditionally fabricated counterparts despite containing
relatively high porosity. The superior fatigue properties
may be related to the unique microstructure caused by
the special manufacturing process of PBF technique.
However, the mechanism of unique microstructure on
fatigue resistance of additively manufactured austenitic
stainless steel remains elusive.
In the last five years, many researchers were com-
mitted to obtaining a comprehensive understanding of
the microstructure-property relationship of additively
manufactured austenitic stainless steels (Liu et al. 2018;
Wang et al. 2018; Zhao et al. 2020; Zhu et al. 2020).
The twining-induced plasticity (TWIP) behaviour is a
key factor for the excellent ductility of the additively
manufactured austenitic stainless steel (Pham, Dovgyy,
and Hooper 2017a; Zhu et al. 2020). The transformation
induced plasticity (TRIP) effect of austenitic stainless
steel with low stacking fault energy (SFE) (calculated
by Eq. (1)) leads to the secondary work hardening and
maintains plastic deformation stability (Shen et al.
2015; Polatidis et al. 2018; Li et al. 2019b; Polatidis
et al. 2020; Zhu et al. 2020).
SFE =−53 +6.2Ni +0.7Cr +3.2Mn +9.3Mo (1)
where the elements are in weight percentage (wt.%). In
addition, the solidification-enabled cellular structures,
low-angle grain boundaries (LAGBs) and pre-existing dis-
location network can enhance the strength, while the
hierarchically heterogeneous microstructure ensures
high uniform elongation (Wang et al. 2018). The pre-
existing dislocation network can slow down but not
entirely block the dislocation motion, and can also
promote the formation of a high density of nano-twins
during plastic deformation (Liu et al. 2018). These mech-
anisms contribute to the combination of high strength
and high ductility of additively manufactured austenitic
stainless steels. However, at present, almost all the
researches focus on the microstructure-property
relationship about static strength and ductility, and
little on microstructure-property about fatigue strength,
which is critical to the application reliability and devel-
opment of AM technique. The mechanism of dislocation
flow and crack formation in the fatigue process is
especially different from that in the static tensile
process due to cyclic deformation, high strain rate, and
long period. Although the intrinsic defects of AM parts
play an important role in reducing the fatigue properties
(Sanaei and Fatemi 2020), the fatigue properties exceed-
ing its wrought counterparts (Kumar et al. 2020; Pegues,
Roach, and Shamsaei 2020) may be potentially related to
the unique heterogeneous microstructure effect, which
is still lacking effective researches.
Therefore, this article systematically investigates the
original heterogeneous microstructure of 304L austeni-
tic stainless steel by selective laser melting (SLM) and
2H. ZHANG ET AL.

microstructural evolution after fatigue loading through
numerous electron microscopy characterisations. The
strengthening mechanism of unique heterogeneous
microstructure on the fatigue resistance of additively
manufactured 304L SS was fully discussed base on
obtained microstructural results. This research aims to
fully understand the effect of the microstructure of addi-
tively manufactured parts on the fatigue strength and to
provide new data support and valuable suggestions for
reliability improvement of metal additive manufacturing
techniques.
2. Experimental procedures
2.1. Sample manufacturing process
The gas-atomized spherical 304L austenitic stainless
steel (SS) powers with the granular sizes of 15–53 μm
were applied for selective laser melting (SLM) processes.
A BHGD-LM150 laser melting 3D printer equipped with
an IPG fibre laser with a maximum 200 W power
output was used to manufacture the parts. The pre-
heated temperature of 100°C was applied to the
Cr12MoV carbon-based die steel substrate for reducing
the residual stress during SLM process (Liu et al. 2018).
And the chamber was continuously purged with pure
argon to keep oxygen level below 1000 ppm during
SLM processes (Wang, Palmer, and Beese 2016). The
optimised processing parameters of laser beam diam-
eter of 80 μm, laser power of 140 W, scanning speed of
1100 mm/s, hatch spacing of 50 μm, layer thickness of
30 μm, and interlayer rotation angle of 67° were
applied to achieve the porosity less than 0.2%, which
was measured by image analysis method given in
(Thijs et al. 2010). The calculated volumetric energy
density (VED) according to (Ghayoor et al. 2020)is
133.33 J/mm
3
. Island scanning strategy with a random
order and the size of 3 × 3 mm
2
was applied to reduce
the residual stress during SLM process (Qiu, Adkins,
and Attallah 2013; Lu et al. 2015). The rotation angle
between neighbouring islands in each layer was 90°,
and the coutour scan was performed to improve
surface quality (Lu et al. 2015; Pham, Dovgyy, and
Hooper 2017a). The wire electrical discharge machining
(EDM) was used to obtain the test samples for fatigue
testing after SLM. The narrow side surfaces of the rec-
tangular samples were polished to Ra = 0.4 for eliminat-
ing the detrimental effects of the EDM process on the
fatigue tests. The thickness and gauge width of speci-
mens are 2 and 15 mm, respectively. The chemical com-
position of 304L SS power was given by the
manufacturer, and the chemical composition of as-
built samples was measured by the chemical
composition testing centre (Table 1). Based on the
chemical composition of as-built samples, the stacking
fault energy (SFE) of as-built 304L SS is calculated to
be ∼27.34 mJ/m
2
, which is relatively low SFE (Schramm
and Reed 1975). The low SFE favours deformation-
induced twinning (Obrtlík, Kruml, and Polák 1994), thus
increasing the strength and ductility (Sun et al. 2018).
The SFE value in this study is very close to the SFE calcu-
lated by the given chemical composition in (Wang,
Palmer, and Beese 2016) and the SFE of SLMed 316L
SS (Sun et al. 2018).
2.2. Tensile and fatigue tests under the
supervision of AE and TI techniques
Tensile tests and fatigue tests were performed through a
200 kN electro-hydraulic fatigue testing machine. The
loading rate of 0.05 mm/min was used for tensile
testing according to ASTM E8. The stress amplitude of
500 MPa, stress ratio (R) of 0.1, and loading frequency
of 10 Hz were used for fatigue testing. Two replicates
were prepared for static tensile tests, while six replicates
were prepared for fatigue tests. The obtained values of
tensile strength are 690 and 694 MPa, respectively. The
average fatigue life is 51,204 cycles. The calculated stan-
dard deviation of fatigue life is 5425 cycles, which indi-
cates the dispersion of fatigue data is relatively low. In
addition, acoustic emission (AE) technique and thermal
imager (TI) were used to monitor the process of tensile
and fatigue testing. The AE144S piezoelectric sensors
with the resonance frequency of 150 kHz and the fre-
quency range of 100–500 kHz, the pre-amplifiers of
40 dB and the sampling frequency of 1 MHz were
selected for AE testing, which is the same as our previous
research (Zhang et al. 2020d).
2.3. Microstructure characterisation
The Ultra Plus field emission scanning electron
microscopy (FESEM) and Crossbeam 550 focused ion
beam (FIB) scanning electron microscopy (FIB-FESEM)
equipped with electron backscattered diffraction
(EBSD) detector and STEM detector were used to
measure fracture surface, grain orientation, and
texture. The step sizes selected for EBSD observation
were 0.5 μm and the collected data were processed by
Channel 5 software. The calibration rate of EBSD
Kikuchi pattern has been researched by more than
92%. Before SEM and EBSD examinations, the extracted
samples were electrochemically etched at 15 V for 15 s
in 10% oxalic acid solutions. The Transmission electron
microscopy (TEM, G20) and transmission Kikuchi diffrac-
tion (TKD) equipped in the FIB-FESEM were used to
VIRTUAL AND PHYSICAL PROTOTYPING 3

obtain the distribution of dislocation and geometric dis-
location density (GND). The step sizes for TKD obser-
vation were 5 nm. The calibration rate of TKD Kikuchi
pattern has been researched by more than 98%. The
samples for TEM observation and TKD observation
were prepared by ion milling after double jet
electropolishing.
3. Results and discussion
3.1. Microstructure
3.1.1. Grain characteristics
There are different microstructures of SLMed 304L SS
on the cross-sections between parallel and perpendicu-
lar to building direction, which are defined as the ver-
tical plane and horizontal plane, respectively. The SEM
images of both horizontal plane (Figure 1(a)) and verti-
cal plane (Figure 1(f)) show there are few gas pores
(marked as yellow dotted circle) and no lack-of-fusion
(LoF) defects, which indicates that the volume fraction
of voids is far less than 0.2% by measured method
given in (Thijs et al. 2010). From IPF image of horizon-
tal plane (Figure 1(b)), scanning tracks with different
scanning direction, hatch spacing, and interlayer
rotation can be identified and measured. The measured
values are consistent with the setting of SLM exper-
iment. For vertical plane, the molten pool boundaries
recognised from SEM image (Figure 1(f)) are depicted
within the IPF image (Figure 1(g)). The direction of
the columnar gains growth is closely aligned with the
<001> building direction, which is the maximum heat
flow direction during SLM process (Figure 1(g)). The
small deviation of grain growth direction from building
direction is mainly due to the adopted island scanning
strategy and rotation angles between layers (Figure 1
(j)), which can also be found in (Zhou et al. 2015;
Kong et al. 2019).
In addition, Figure 1(k) shows the schematic
diagram for the formation of the ‘nested’grain struc-
ture shown in Figure 1(l), which is the typical grain
structure extracted from Figure 1(b). Such ‘nested’
structure has been observed and reported (Pham,
Dovgyy, and Hooper 2017a; Ferreri et al. 2020;
Pegues, Roach, and Shamsaei 2020). Grain growth
during SLM process is affected by scanning strategy,
the repetition of heating and cooling, and directional
heat conduction acting on the shape of molten pool
(Pham, Dovgyy, and Hooper 2017a; DebRoy et al.
2018). The thermal conductivity through the metal
powder is extremely low compared to the thermal
transfer along with solidified material and substrate
(Pham, Dovgyy, and Hooper 2017a). Therefore, the
heat in the new molten pool is mainly dissipated
along <001> crystallographic direction through under-
lying solidified material, resulting in the directional soli-
dification of the grains in the new molten pool during
the cooling process (Figure 1(k1, k2)). If a new molten
pool is deposited precisely at the same position of the
molten pool of underlying solidified material, a coaxial
‘nested’structure can be formed (Polatidis et al. 2018).
The nucleation and growth of new grains in this study
are heterogeneous and anisotropic due to the
influence of island scanning and rotation angles
between layers, resulting in non-coaxial and irregular
‘nested’structures (Figure 1(k2, k3, k4)). A typical
‘nested’structure extracted from Figure 1(b) is shown
in Figure 1(l) and the corresponding pole image is
shown in Figure 1(m).
The SLMed 304L SS showed a high density of low
angle grain boundaries (LAGBs, <15°, ∼86.9% and
∼88.7% of total GBs for horizontal plane and vertical
plane, respectively), a low density of high angle grain
boundaries (HAGBs, ≥15°, ∼13.1% and ∼11.3% of total
GBs for horizontal plane and vertical plane, respectively)
and no twin boundaries, as shown in the GB maps
(Figure 1(c, h)) and the misorientation histograms
(Figure 1(d, i)). The kernel average misorientation
(KAM) map (Figure 1(e)) showed the local misorienta-
tions across grains and the KAM value is concentrated
in the range of 0–3°. The distribution of local misorienta-
tions show a strong relationship with the HAGBs. The
high local misorientations are concentrated around the
HAGBs (Wang et al. 2018)(Figure 1(c)). The horizontal
plane and vertical plane are epitaxial grains with an
average grain size of 6.1 μm and columnar grains with
an average grain size of 7.7 μm according to Figure 1
(b, g), respectively.
3.1.2. Cellular substructures
The higher-magnification SEM images (Figure 2(a-c))
show substructures formed by solute segregation and
residual stress owing to the high cooling rate, which
can be frequently found in SLMed austenitic stainless
Table 1. Chemical compositions of 304L SS powder and as-built samples (wt%).
Element C N Cr Ni Mn Si Mo O Fe
AISI Max 0.03 –18.0–20.0 8.0–12.0 Max 2.0 Max 1.0 –Bal.
Power 0.0086 –18.810 10.02 0.420 0.510 –0.0430 Bal.
As-built 0.0110 0.0150 19.100 10.30 0.811 0.942 0.045 –Bal.
4H. ZHANG ET AL.

steels (Saeidi et al. 2015; Pham, Dovgyy, and Hooper
2017a; Liu et al. 2018; Wang et al. 2018; Ghayoor et al.
2020; Zhu et al. 2020). The SEM images show a large
area of equiaxed cellular substructures (sub-grains) and
a small area of elongated columnar substructures in
horizontal plane (Figure 2(a, b)). The different shapes
of substructures inside each large grain in horizontal
planes may be attributed to the different grain orien-
tations in the Zdirection of horizontal plane. The high
residual stress and plastic deformation caused by extre-
mely high cooling rate and thermal gradient during SLM
process lead to the deformation of equiaxed grains
(Saeidi et al. 2015; Liu et al. 2018), thus dividing defor-
mation zone and non-deformation zone in Figure 2(a,
b). The residual stress on the top surface of the SLMed
304L SS part measured by μ-X360 is −67.6 MPa. The
residual stress in SLM process is much higher than the
measured value owing to the imbalance of the thermal
expansion in melting and cooling shrinkage in solidifica-
tion. Additionally, the interlayer rotation of 67° applied
may cause varying directions of thermal gradients,
thus leading the cellular substructures to deform in mul-
tiple directions. A region (marked by purple dotted
circle) similar to the ‘slip’zone is found between the
deformation zones and non-deformation zones in
Figure 2(a). It should be noted that the boundary lines
shown in Figure 2(a, b) are the molten pool boundaries
rather than GBs, and the columnar substructures can
cross the molten pool boundaries (Figure 2(b)). The
primary cellular arm spacing (PCAS) can be used to
descript the size of sub-grain structures (Ma, Wang,
and Zeng 2017) and is measured to be 0.886 μm
(Figure 2(c)). And PCAS can be used to calculate the
cooling rate according to Eq. (2) (Ghayoor et al. 2020),
˙
T=0.33 80
l
 (2)
where λis the PCAS and
˙
Tis the cooling rate. The calcu-
lated cooling rate is ∼8.4 × 10
6
K/s. Some areas are
higher due to the existence of finer sub-grain structures.
In addition, the schematic diagrams of Figure 2(d, e)
was drawn for clearly showing the cellular walls deco-
rated with a high-density pre-existing dislocation
network (Figure 2(e)), which has been proved to
promote the formation of a high density of nano-twins
during plastic deformation (Liu et al. 2018). The size
and wall thickness of cellular structures are known to
be related to solidification conditions, including
thermal gradient, cooling rate and solidification front
velocity (Prashanth and Eckert 2017; Wang et al. 2018).
The TEM images shown in (Figure 2(g-i)) fully indicate
the cellular sub-grains are decorated by pre-existing
Figure 1. The microstructures of a selective laser melting (SLM) produced 304L SS. (a) The SEM image, (b) inverse-pole figure (IPF)
map with drawing scanned tracks, (c) grain boundaries (GBs) map, (d) the histograms of GB misorientations and (e) the kernel average
misorientation (KAM) map for a cross-section perpendicular to the building direction, named as horizontal plane. (f) The SEM image,
(g) IPF map with drawing molten pool boundaries, (h) GB map, (i) the histograms of GB misorientations and (j) growth direction of
columnar grain for the cross-section along to the building direction, named as vertical plane. (k) The schematic diagram for the for-
mation of ‘nested’grain structure. (l) The IPF map of cross-section extracted from (b). (m) The corresponding pole figure of (l).
VIRTUAL AND PHYSICAL PROTOTYPING 5

dislocation network, and a small number of dislocations
are distributed at the cell interiors.
3.1.3. Phase analysis
The XRD diffraction patterns (Figure 3(a)) and EBSD
observation (Figure 3(b) and Figure 8(d)) of fatigue
specimens further show that the deformed SLMed
304L SS contains only austenite, which is similar to the
studies on SLMed austenitic SS including 304L (Wang,
Palmer, and Beese 2016) and 316L (Pham, Dovgyy, and
Hooper 2017a; Wang et al. 2018), but is different from
the studies on SLMed 304L SS (Polatidis et al. 2020;
Zhu et al. 2020).
It can be considered that the difference of chemical
composition leads to the change of SFE and M
d30
of
additively manufactured 304L SS, thus affecting the
Figure 2. The images of cellular sub-grain structure of SLMed 304L SS. (a-c) SEM images; (d) The schematic diagram of cellular sub-
grain microstructure; (e) pre-existing dislocation network formed during SLM process; (f) The pre-existing dislocation network; (g-i)
The TEM images showing the cellular sub-grain structure decorated with pre-existing dislocation network.
Figure 3. Phase analysis. (a) XRD patterns of SLMed 304L SS; (b) Phase map from EBSD.
6H. ZHANG ET AL.

transformation induced plastic (TRIP) effect. The M
d30
temperature is usually the temperature at which 50
vol. % of the austenite transforms to martensite under
an applied true strain of 30%, and can be empirically cal-
culated by Eq. (3).
Md30(◦C)=413 −462(C+N)−9.2Si −8.1Mn
−13.7Cr −9.5Ni −18.5Mo (3)
where the elements are in weight percentage (wt.%). The
SFE of ∼27.34 mJ/m
2
and M
d30
of ∼25.40°C in this study
are close to that of (Wang, Palmer, and Beese 2016;
Pham, Dovgyy, and Hooper 2017a; Wang et al. 2018)
and are different from that of (Polatidis et al. 2020; Zhu
et al. 2020). The value of M
d30
can be widely used to
verify the occurrence of the martensitic phase transform-
ation (Angel 1954; Wang, Palmer, and Beese 2016). The
M
d30
temperature (∼25.40°C) below the room tempera-
ture led to the absence of martensitic transformation in
this study. In addition, the temperature of the parts
increases further with the cyclic loading (Figure 8(i)),
resulting in the temperature much higher than M
d30
,
thus further inhibiting TRIP effect. The Cr
eq
/Ni
eq
can be
used to infer the solidification process according to
(Ghayoor et al. 2020), and can be calculated as,
Creq =Cr +Mo +1.5Si +0.5Nb
Nieq =Ni +30C +0.5Mn (4)
where the elements are in weight percentage (wt.%). The
calculated Cr
eq
/Ni
eq
ratio is ∼1.47 by Eq. (4), which falls
into the austenitic-ferritic mode (L→L+γ→L+δ+γ→γ
+δ→γ). The high cooling rate (∼8.4 × 10
6
K/s) during
SLM process leads all ferritic phase to transform into aus-
tenitic phase, which explains the reason for all austenitic
phase in as-built parts.
3.2. Mechanical testing
3.2.1. Tensile testing
The 304L SS produced by the SLM process parameters
used in this study shows excellent strength (∼692 MPa)
and ductility (∼70%) (Figure 4(a)). Interestingly, its
strength in this study is higher than similar studies
on SLMed 304L SS (Ghayoor et al. 2020; Hou et al.
2020; Zhao et al. 2020), and is close to the strength
of SLMed 304L SS in (Zhu et al. 2020) and the strength
of SLMed 316L SS in (Liu et al. 2018). The superior
tensile properties shown in this study may be caused
by the unique scanning strategy, including island scan-
ning strategy, interlayer rotation of 67°, and inter-island
rotation of 90°, which can enhance the homogeneity of
the microstructure of parts (Figure 1(a and f)). The
island scanning strategy can also reduce the vector
length, resulting in significant reductions in residual
stress, thus minimising the adverse effects of residual
stress (Bartlett and Li 2019). In addition, the optimised
process parameters make the densification of parts
reach more than 99.8% and obtain cellular substruc-
tures with very small size (∼0.886 μm), which signifi-
cantly contribute to the superior tensile properties.
According to the existing literature (Liu et al. 2018;
Wang et al. 2018), additively manufactured austenitic
stainless steels can break the strength-ductility trade-
off, and exhibit the combination of high strength and
high ductility compared to the conventional austenitic
steels. The important reason is that the mechanical
properties can be improved by the collective effects
of hierarchically heterogeneous microstructures, includ-
ing solidification cellular structure, LAGBs, and dislo-
cations (Wang et al. 2018). In addition, the unique
cellular structure decorated by the high-density pre-
existing dislocation network can adjust and stabilise
dislocation flow (Figure 2), promoting the working
hardening during tensile testing (Liu et al. 2018). In
the elastic deformation (low strain) stage, the pre-exist-
ing dislocation network structures around cell walls can
prevent the dislocation motion (Liu et al. 2018), thus
increasing the dislocation storage ability, resulting in
high yield strength. The strain hardening rate
decreases rapidly in the stage of elastic deformation
and decreases slowly and steadily in the plastic defor-
mation till failure (Figure 4(a)). These differences may
be caused by different chemical compositions, which
can fully indicate the occurrence of second strain hard-
ening behaviour (Zhu et al. 2020). The strain hardening
can delay the onset of necking to enhance the duct-
ility. There is an explanation for strain hardening of
SLMed 304L SS from the microstructure. The pre-exist-
ing dislocation network of SLMed 304L SS with low
stacking fault density (∼27.34 mJ/m
2
) can ensure
stable and continuous dislocation motion and nano-
twins formation, thus ensuring the stable plastic flow
during the whole plastic deformation (Liu et al. 2018).
Therefore, the SLMed 304L SS can maintain high
strength and high ductility even if a small number of
pores and oxide inclusions reduce the strength (Zhao
et al. 2020; Zhu et al. 2020). The AE characteristic par-
ameters presented in Figure 4(d) further indicate that
there is almost no microstructural damage before
failure. The plastic deformation was produced steadily
under the regulation of the sub-grain structures deco-
rated with pre-existing dislocations. The SLMed 304L SS
lost internal structural stability instantaneously with the
release of dislocation energy storage caused by strain
hardening, resulting in the cracks and then rapid
deterioration.
VIRTUAL AND PHYSICAL PROTOTYPING 7

3.2.2. Fatigue testing
The fatigue test shows that the hysteresis loops are
very narrow, which was caused by the lower stress
level (500 MPa) than the yield strength (∼551.76 MPa)
and the high elastic modulus (∼17,810 MPa). The
average fatigue life of SLMed 304L SS is 51,204 ±
5425 cycles, which is beyond the fatigue life of other
studies on PBFed austenitic stainless steels (Lee,
Pegues, and Shamsaei 2020; Pegues, Roach, and Sham-
saei 2020), and is close to fatigue life of (Parvez et al.
2019). This study selected the part with the fatigue
life of 56,629 to draw the hysteresis loops and strain
amplitudes in fatigue history (Figure 4(b, c)). The
strain amplitude and the position of the hysteresis
loop almost remain nearly unchanged in the previous
55,000 cycles, while the strain amplitude gradually
increases in the last 600 cycles. This means that the
last 600 cycles belong to fatigue crack initiation and
crack propagation process. Therefore, the 55,000 pre-
vious cycles before fatigue crack initiation dominate
the fatigue performance of parts (Pegues, Roach, and
Shamsaei 2020). Combined with the conclusions of
tensile tests, each cycle can be assumed to represent
a static tensile process with an extremely high strain
rate. But this tensile process does not enter the
plastic deformation stage and is only in the range in
the elastic deformation stage of Figure 4(a).
The AE characteristic parameters in fatigue tests show
excellent real-time monitoring ability of internal damage
and can reveal additional information (Zhang et al.
2020a,2020c)(Figure 4(f)). With the increasing fatigue
cycles, it can be found that the AE amplitude, the
count curve, and the slope of the cumulative count
curve slowly increase, while the RMS and cumulative
energy almost remain unchanged. These indicate that
the dislocation motion and the interaction between
grains inside the SLMed 304L SS may be gradually inten-
sified during the long-term fatigue loading. This further
demonstrates that the accumulation of fatigue cycles
leads to the increase of dislocation storage, making dis-
locations flow across the cell walls, thus eventually
leading to the initiation of fatigue cracks. However, the
amplitude of AE signals of fatigue testing is generally
higher than that of tensile testing, which may be due
to the higher velocity of dislocation motion by extremely
high strain rate. The regular vibration of the crosshead
during fatigue loading may also have some influence
on the increase of AE amplitude.
3.3. Texture and twining during tensile testing
The tensile sample selected two regions for EBSD obser-
vation, one of which was close to the fracture and the
other was slightly away from the fracture (Figure 5(a)).
Figure 4. Tensile and fatigue tests under the monitoring of acoustic emission (AE) technique. (a) True stress-strain curveand strain hard-
ening rate curve for tensile tests. (b) The cyclic stress-strain hysteresis loops for fatigue tests. (c) The strain response under stress-con-
trolled conditions. (d) The AE characteristic parameters of amplitude, count, cumulative count, RMS and cumulative energy during tensile
test. (f) The AE characteristic parameters of amplitude, count, cumulative count, RMS and cumulative energy during fatigue test.
8H. ZHANG ET AL.

The corresponding EBSD maps (Figure 5) can be used to
analyze the changes of crystallographic orientation by
comparing the un-deformed EBSD maps in Figure 1(a-
e). It can be found that strong deformation twinning
was formed in both areas T1 and area T2 (Figure 5(c, d,
j)), which indicates that twin induced plasticity (TWIP)
occurred during the tensile tests (Rahman, Vorontsov,
and Dye 2015). The area T2 is closer to the fracture
than that of area T1, which made the twinning in area
T2 more than that in area T1 owing to higher plastic
deformation by necking (Figure 5(a)). In addition, the
excessive twinning and slip caused by excessive defor-
mation reduced the texture strength and made the crys-
tallographic orientation more random (Figure 5(i, l)). By
observing the typical grains with twins in Figure 5(e, f),
it can be found that the Schmidt factor of twins is
higher than that of matrix grains, which means that
the twins are easier to activate the slip system than
the matrix grains. The Schmidt factor was calculated
through the X load direction and (110) <111> slip
system. According to Figure 5(g, h), twins nucleated
from pre-existing dislocations around HAGBs, and twin-
ning is controlled by the glide of Shockley partials with
Burgers vector 1/6 <112> on (111) planes (Gutierrez-
Urrutia, Zaefferer, and Raabe 2010; Rahman, Vorontsov,
and Dye 2015; Pham, Dovgyy, and Hooper 2017a). In
addition, a model has been proposed, that is, the twin-
ning stress of twin nucleation is strongly proportional
to the SFE of the metal (Romanov, Vladimirov, and
Nabarro 1992),
t
twin =
g
SF
Kbs
(5)
where τ
twin
is the critical twining stress for creating a
twin, γ
SF
is SFE (∼27.34 mJ/m2), and Kis the constant 2
determined by (Romanov, Vladimirov, and Nabarro
1992) and b
s
is the Burgers vector for a Shockley
partial dislocation, which is 2.58 × 10
−10
m for austenite
phase (Pokharel et al. 2019). The SLMed 304L SS with low
SFE results in very low twinning stress of ∼52.98 MPa,
which explains the strong TWIP behaviour of SLMed
304L SS during tensile testing (Gutierrez-Urrutia,
Zaefferer, and Raabe 2010). Additionally, compared
with the KAM value of non-deformation (Figure 1(e)),
the KAM values of both area T1 and area T2 increased
significantly (Figure 5(m-p)). The KAM value of area T1
and area T2 is concentrated in the range of 0–4° and
1–9°, respectively.
3.4. Texture change during fatigue testing
Due to the complexity of the microstructure changes of
fatigue sample, six areas for EBSD observation were
selected to reveal the differences and named them
F1-F6 (Figure 6(S1, S2)), of which F6 contains a little
crack for later discussion in Section 3.5. The macro-
cracks and voids caused by tensile stress can be seen
from Figure 6(S1, S2). The deformation of F1 and F2 is
extremely large compared to that of F3, F4 and F5
due to closer to the fatigue fracture. The SEM images
of fatigue fracture (Figure 6(S3)) show that the fatigue
crack initiated from the top surface, which has not
been machined or polished. The well-known typical
crack propagation region and typical instantaneous
fracture region can be recognised in Figure 6(S3).
Therefore, the fatigue resistance can be considered to
be influenced by surface quality under the premise of
ensuring low porosity in the parts (Kasperovich and
Hausmann 2015). Additionally, necking can be found
around the instantaneous fracture region (boundaries
were marked with yellow dashed line in Figure 6(S3)),
which is similar to the fracture surface of tensile
samples. A large number of large pores and a small
number of small pores are distributed on the surface
of instantaneous fracture region (Figure 6(S3)). The
large irregular pores are mainly the lack-of-fusion
(LoF) defects caused by the inadequate penetration of
the molten pool (Darvish, Chen, and Pasang 2016),
while the small spherical pores are gas pores formed
due to the entrapment of the shielding gas or alloy
vapours inside the molten pool (Yu et al. 2019). A
similar fracture surface can be found in (Kuo et al.
2020; Zhang et al. 2021). These pores can become the
sites of stress concentration, thus affecting the yield
stress and ductility (Kuo et al. 2020). Therefore, the key-
holes of top surface usually become the sites of fatigue
crack initiation (Figure 6(S3)). The post-treatment
method of eliminating keyholes on the surface of
parts, such as laser re-melting (Yu et al. 2019), may
further improve the fatigue performance of additively
manufactured 304L SS. In addition, the high thermal
gradient and high cooling rate in print process can
lead to high residual tensile stress, shifting the mean
stress amplitude, thereby affecting the fatigue crack
initiation. However, residual stress has a limited effect
on the fatigue performance of additively manufactured
parts compared to surface roughness and internal por-
osity (Zhang et al. 2020b).
By observing Figure 6(a1-e5), the grain shapes, grain
orientation, twins, GBs and KAM values change
obviously with the increase of the distance from
fatigue fracture surface of SLMed 304L SS. The equiaxed
grains around the fracture surface of fatigue samples are
elongated and deformed under the action of tensile
strain (Figure 6(a2, b2)). According to a previous study
on TWIP steel (Gutierrez-Urrutia, Zaefferer, and Raabe
VIRTUAL AND PHYSICAL PROTOTYPING 9

2010), the twin stress is usually less than slip stress
(KH−P
tw ≤KH−P
slip ), and the grain refinement in the
micrometer range cannot inhibit the twinning. There-
fore, deformation twins are generated with the acti-
vation of multiple slips at a small plastic strain close to
yielding (Figure 6(a2, b2, c2)). For the selected regions
of F4 and F5 (Figure 6(d2, e2)) far away from the fracture
surface, the strain produced by fatigue tests is too small
to form twins. Therefore, the fractions of twin bound-
aries decreased with the increase of distance between
scanning area and fracture surface (Figure 6(a4, b4, c4,
d4, f4)). In addition, this study reveals that twining
during fatigue testing is affected by necking. From the
hysteresis loops and strain amplitude response on
Figure 4(b, c), the strain amplitude of the previous
55,000 cycles is nearly unchanged and is ∼2.67%, while
Figure 5. The EBSD maps of SLMed 304L SS for tensile tests. (a) The tensile sample for EBSD testing area T1 (b-i, m-n) and area T2 (j-l,
o-p). (b) SEM images of area T1 with of enlarged image. (c) Euler + band contrast map with the pole images in (i). (d) Band contrast +
grain boundaries (GB) map. (e) The IPF map of typical grains containing twins extracted from (c, d). (f) The 3D crystal orientation with
the Schmid factor of twins and substrate, respectively. (g) The GBs maps of (c). (h) The distribution of GBs misorientation of (g). (l) The
IPF image of area T2. (k) The extracted subsets from (j). (l) The pole images of (j). (m) The kernel average misorientation (KAM) map of
area T1. (n) The distribution of KAM values of area T1. (o) The KAM map of area T2. (p) The distribution of KAM values for area T2.
10 H. ZHANG ET AL.

that of the last 600 cycles rises significantly to ∼8%.
Combined with the AE parameters plotted in Figure 4
(d), the increasing strain amplitude leads to the sharp
increase of the AE parameters, which means the increase
of the damage rate of SLMed 304L SS. Consequently, the
range of 55,000–55,600 cycles can be identified as the
crack propagation and the instantaneous fracture
shown in Figure 6(S3), which makes the strain amplitude
increase and forms necking. It can be demonstrated that
the necking near the fracture is caused by the large
strain amplitude, resulting in more twins of F1 and F2
than that F3, F4 and F5. The twin boundaries (TB) frac-
tion in this study is much lower than that in (Gao et al.
2020), which is due to the grain boundary engineering
(GBE) process used in their parts. From the previous
analysis (Section 3.1), the pre-existing dislocation
network around the cell walls can hinder the dislocation
motion to enhance the ductility of SLMed 304L SS.
The KAM values in Figure 6(a5, b5, c5, d5, e5) are
plotted together in Figure 7(a), it can be found that
the KAM value increases with the decrease of the dis-
tance between EBSD scanning area and fracture
surface (shown in Figure 6(S1)). The texture of F1 and
F2 are random (Figure 7(b)), which is similar to that of
T1 of tensile sample (Figure 6(g)). The <110> texture
strength of F3, F4 and F5 became stronger due to the
further distance from the fracture surface. Although
the contingency and randomness of selective EBSD
area and the reduction of Kikuchi patterns calibration
rate by excessive deformation may cause the error of
Figure 6. The EBSD maps of SLMed 304L SS for fatigue tests. (S1) Fatigue sample for EBSD testing including six areas, named as F1-F6,
respectively. (S2) SEM image of the enlarged area in S1. (S3) The SEM images for fatigue fracture. (a1-a5) SEM image, IPF map, GB map,
GB histogram and KAM image for F-1, respectively. (b1-b5) Corresponding EBSD maps for F-2. (c1-c5) Corresponding EBSD maps for F-
3. (d1-d5) Corresponding EBSD maps for F-4. (e1-e5) Corresponding EBSD maps for F-5.
VIRTUAL AND PHYSICAL PROTOTYPING 11

this conclusion, this conclusion is reasonable owing to
the obvious monotonic trend of KAM values and
<110> texture strength.
3.5. TEM and TKD analysis for showing the key
role of cellular structures
The size and shape of cellular substructures were
changed by the external force of tensile tests and
fatigue tests compared with as-built cellular structures
(Figure 2(g)), as shown in Figure 8(a). Under the stain
before fatigue crack nucleation (∼2.67%), the long-
range motion of mobile edge dislocations of cell interior
can be obstructed by the pre-existing dislocations
around the cellular boundaries, which have high geo-
metric dislocation density (GND). In the research about
stress relaxation and vanish of cellular structures of addi-
tively manufactured parts (Li et al. 2020), it has been
pointed out that the mobile dislocations in cell interior
with low GND density cannot penetrate the cell walls,
and will be absorbed and eliminated preferentially by
the pre-existing dislocation network. This was estimated
by measured physical activation volume (V*). Therefore,
the active dislocation of the cell interior can be impeded
and absorbed by the pre-existing dislocations walls, thus
increasing the density of GNDs of cell walls and signifi-
cantly improving the yield strength, consequently con-
tributing to the excellent fatigue properties by
enhancing the resistance of fatigue crack nucleation.
However, due to the extremely high strain rate of
fatigue testing, the true activity of mobile dislocations
is much higher than that of static tensile testing. The
continuously increasing energy of active dislocation
flow caused by the high strain rate and cumulative
cyclic loading may make the mobile dislocations pene-
trate the cellular boundaries (Figure 8(b, c)). In this
meantime, the pre-existing dislocations allow the
obstructed dislocation flow across the cellular walls
and play a key role in maintaining the stable motion of
mobile dislocation during fatigue testing. This regulation
process which is beneficial to the fatigue properties
causes the cellular boundaries migration, leading to
the drastic changes of position, shape, size and wall
width of cellular structures (Figure 8(b, c)). Finally, the
typical dislocation walls/lines (marked by the white
dotted line) with a higher density of GNDs were
formed, which is also found in the Fe-Ni-Cr alloy at 700
°C with additional strain (Li et al. 2019a). In addition,
for the mechanism of penetrating the dislocation walls
of mobile dislocations, Kong et al. (Kong et al. 2020)
have revealed that the Burger vector direction of the
GNDs of two sides of cell boundaries of additively man-
ufactured 316L SS is opposite. The dislocations of cell
interior can penetrate across the cell boundaries under
appropriate strain by the external force, then leading
to the consequent annihilation of dislocations and the
decrease of GND density. In addition, the dislocation
lines originate from the HAGBs, which indicates that
the dislocation lines with a high density of GNDs can
provide the driving force for twinning, contributing to
the forming of twins (Figure 8(d)).
The interaction of formed twins and remained pre-
existing dislocation walls can result in the effects of dis-
location-twins and dislocation-cellular structures effect,
which can induce a progressive hardening mechanism
and promote the enhancement of fatigue properties. It
is widely known that the twinning is the main mechan-
ism of traditional metals under the action of external
force. Twinning can improve the plasticity of materials,
known as twinning induced plasticity effect (TWIP). For
additively manufactured parts with heterogeneous cel-
lular structures, the massive deformation-induced
twins under high deformation can penetrate and dis-
torted the cellular boundaries. The plasticity of additively
manufactured parts is affected by the combination of
deformation twins and cellular structures. From the
post-mortem TEM observation, the twins originate
from the HAGBs and penetrate through the cellular
boundaries and LAGBs (Figure 8(f)). The deformation
twins are greatly weakened by HAGBs and a small
number of deformation twins can be found to penetrate
slightly the HAGBs and propagate in another grain
(Figure 8(g)), which is due to the fact that the much
higher misorientation around the HAGBs than that
around the cellular boundaries. But in fact, the unique
cellular structures and crystallographic orientation of
grains together regulate the motion of mobile dislo-
cations to induce the dislocation lines with a high
density of GNDs, dominating the formation of twins. In
addition, the interlacing sets of deformation twins are
found under higher strain (Figure 8(h)), which can be
seen in the additively manufactured parts, but rarely in
the traditionally manufactured counterparts. The direc-
tion of the second set of twins intersects that of the
first set of twins, making the first set of twins buckle
(Figure 8(k, l)). The additional second set of deformation
twins may be formed by the regulation effect of retained
pre-existing dislocations and the blockage of the first set
of deformation twins. The interlaced deformation twin
and the pre-existing network may contribute to the
work hardening plasticity and the resistance of fatigue
cracking by impeding the dislocation motion.
To further explore the role of cellular structures on
the working hardening behaviour, we carried out mul-
tiple transmission Kikuchi diffraction (TKD) with the
nanoscale resolution for fatigue samples. The as-built
12 H. ZHANG ET AL.

TKD results (Figure 9(a-d)) reveals that the strong dislo-
cation trapping and retention at the cellular walls were
demonstrated by the larger local misorientation at/
near the cellular boundaries in KAM map, especially at
the triple junctions (Figure 9(c)). This KAM map is
especially similar to (Wang et al. 2018). Therefore, the
cellular structures can significantly impact the dislo-
cation accumulation (GND density), leading to the
work hardening behaviour of additively manufactured
parts. When giving cyclic deformation, the dislocation
walls with large local misorientation (GND density)
were formed due to the deformation and migration of
the cellular boundaries by the dislocation penetration
and annihilation (Figure 9(g, k)). In addition, the strain
rate hardening can help the sustain of homogenous
deformation and then delays the strain localisation and
necking (Li et al. 2020). The combination of work harden-
ing and strain rate hardening plays a critical role in
fatigue life by affecting most of fatigue history. The dis-
location lines with extremely high local misorientation
trigger the formation of LAGBs accompanied by higher
GND density (Figure 9(f, j)). When high strain, the KAM
maps in Figure 9(o, s) show the deformation twins
with higher local misorientation penetrate and subdi-
vide the cellular structures, making the twins, dislocation
lines and cellular structures capture and retain the active
dislocation, thus contributing to the a progressive work
hardening mechanism of fatigue testing. The twins and
HAGBs are shown to have higher blockage capability of
dislocation motion by the KAM maps in Figure 9(o, s).
But the work hardening mechanism of the combination
of dislocation-twins, dislocation-dislocation lines and
dislocation-cell walls contributes to the last stage of
fatigue history, and may have relatively limited effect
on the fatigue life because the fatigue micro-crack may
have initiated.
The fatigue life-enhancing mechanisms of the unique
heterogeneous cellular substructures of additively man-
ufactured parts can be drawn in Figure 10 according to
the above in-depth detailed analysis. The cellular struc-
ture decorated by the pre-existing dislocations
accompanied by the relatively high GND and stored
energy can impede the dislocation motion of the cell
interior to improve the yield strength, thus delaying
the nucleation of fatigue crack (Figure 10(a)). But this
working enhancing mechanism only works for a short
time (nearly 100 cycles) at the beginning of fatigue
testing. As the fatigue loading, the stress-controlled
fatigue testing causes the high strain rate and two
sides of cellular structures have opposite Burgers
vector directions of the GNDs, which allow the mobile
dislocation to penetrate the cell boundaries (Figure 10
(b)). These mobile dislocations will be finally absorbed
and annihilated with the migration of cellular bound-
aries and the decrease of GNDs (Figure 10(c)). The
pinning effect of pre-existing dislocations of cell bound-
aries in this period (100∼55,000 cycles) can effectively
maintain the dislocation motion stable to significantly
improving the resistance of fatigue crack nucleation.
However, it is difficult for active dislocations to break
through the HAGBs because of the extremely high
GNDs. Since the duration of the combination mechan-
ism of work hardening mechanism and strain rate hard-
ening of cellular structures accounts for most of the
fatigue history, the regulation effect of cell structure
plays a key role in enhancing fatigue performance.
Then, with increased strain, the dislocation lines were
formed by the deformation and migration during the
regulation progress of cellular structures (Figure 10(d)),
which can offer the GNDs density for effectively trigger-
ing twinning (Figure 10(e)). The formation of dislocation
lines also is controlled by the slip system of austenitic
grains under external force except for the regulation
effect of cellular structures. With the increase of strain
caused by the sustained fatigue loading, fatigue crack
start to nucleate, the twins form along the previously
formed dislocation lines at the expense of stored
energy of the dislocation lines (Figure 10(e)). Moreover,
Figure 7. The KAM values and pole images for F1-F5.
VIRTUAL AND PHYSICAL PROTOTYPING 13

the second twins can make the first twins buckle, which
is due to the interaction between the regulation effect of
cellular substructure and the blockage effect of the first
twins. The mechanisms of the interaction between inter-
laced twins and cellular structures can impede and
adjust the dislocation motion, thus improving the resist-
ance of fatigue cracking to slightly increase the fatigue
life (the last 600 cycles) (Figure 10(f)). Therefore, the cel-
lular structures can lead to beneficial work hardening
mechanisms at every stage of fatigue tests by tuning
the dislocation motion, which results in the excellent
fatigue properties of additively manufactured metal
parts. Among these hardening mechanisms, the regu-
lation effect of cellular structures for leading the dislo-
cations to penetrate the cell boundaries and
consequently annihilate gives the key contribution for
improving fatigue life (100∼50,000 cycles). This regu-
lation effect significantly relays the fatigue crack nuclea-
tion, thus prolonging the fatigue history stage before
forming the fatigue crack.
As shown in Figure 1(b and c), especially dense
HAGBs are distributed near/in the molten pool bound-
aries, which can also be found in (Pham, Dovgyy, and
Hooper 2017b; Wang et al. 2018; Kong et al. 2020).
This may be caused by the epitaxial growth of grains
due to the thermal gradient of the molten pool during
SLM process (Basak and Das 2016) and the re-melting
effect by neighbouring scanning tracks (Zhu et al.
2020). The high density of HAGBs contributes to the
tensile strength and fatigue strength of polycrystals
Figure 8. The TEM images show the dislocation flow process and the forming process of twins under external force. (a-c) The dislo-
cation flow of cell interior and the change of size and shape of cells by the strain caused by the external force of fatigue tests (N=
20,000, ε≈2.67%). (d-f) The deformation twins originated or blocked by HAGBs and propagated through cellular walls associated with
the disappearance of pre-existing dislocations (N= 50,000, ε≈3.40%). (i-l) The intersections of twins labelled as Set 1 and Set 2 (after
fracture, ε≥8.07%).
14 H. ZHANG ET AL.

(Zhang and Wang 2008) due to the strong ability of
HAGBs to hinder dislocation motion. In addition, the
molten pool boundary is the interface between the
molten pool and the underlying solidified materials
(solid–liquid interface), leading to high strain incompat-
ibility and high interfacial energy, resulting in high
residual strain around molten pool boundary. The high
residual strain may be at the microscopic individual
grain scale (intergranular strain), atomic scale
(vacancies), and even macroscopic plastic deformation
(microcracks) (Bartlett and Li 2019). The adopted
process parameters in some literature resulted in
obvious microcracks (defects) near/at the molten pool
boundaries (Zhou et al. 2015; Li et al. 2018; Nadammal
et al. 2021), which is actually caused by the excessive
residual strain. In Section 3.6, It can be found that the
molten pool boundary shows a high possibility of
fatigue crack initiation. The area around molten pool
Figure 9. The Post-mortem TKD analysis for showing the regulation effect and the GND density of cellular structures during fatigue
testing. (a-d) The TEM image, band contrast (BC) map, KAM map and IPF map of as-built cellular structures (N=0,ε= 0%). (e-h) The
TEM image, BC, KAM and IPF maps of deformed cellular structures (N= 20,000, ε≈2.67%). (i-l) The TEM image, BC, KAM and IPF maps
of formed dislocation lines (N= 20,000, ε≈2.67%). (m-p) The TEM image, BC image, KAM image, and IPF image with 3D crystal orien-
tation and SF of deformed twins (N= 50,000, ε≈3.40%). (q-t) The TEM image, BC, KAM and IPF maps with 3D crystal orientation and SF
of various sets of deformed twins (after fracture, ε≥8.07%).
VIRTUAL AND PHYSICAL PROTOTYPING 15

boundaries has extremely high residual strain induced
by the mismatched expansion/shrinkage during SLM
process, which provides high interface energy and
high residual GB dislocation modulus at HAGBs,
making HAGBs around molten pool boundary become
the preferential sites of fatigue crack initiation. Inap-
propriate process parameters may lead to unacceptable
residual strain, thereby greatly reducing the fatigue life
of additively manufactured parts.
In addition, for twins generated during cyclic defor-
mation, intergranular fatigue cracks have the probability
of nucleation from twin boundaries (Li et al. 2013).
Fatigue crack nucleation along twin boundaries or
HAGBs is a competitive process (Zhang and Wang
2008). For additively manufactured 304L SS, the
pinning effect of twins on active dislocations and the
regulating behaviour of cellular substructures can effec-
tively improve fatigue crack initiation resistance (Liu
et al. 2018). During fatigue testing, twins are generated
under the high cyclic strain, penetrate the cellular sub-
structures, and are eventually surrounded by the cellular
substructure decorated by a high density of pre-existing
dislocations (Suryawanshi, Prashanth, and Ramamurty
2017). It can be inferred that the dislocation motion
around the twins is tuned by the cellular substructure,
which reduces the accumulation of dislocations and
improves the fatigue crack initiation resistance at the
twin boundaries. However, the active dislocations can
penetrate the cell walls but cannot transfer through
GBs, they will pill-up at HAGBs, thus consequently
leading to the intergranular fatigue cracks at HAGBs
with the increase of cyclic deformation. It cannot be
ignored that the higher interface energy and higher
residual dislocation at the HAGBs near the molten pool
boundary induced by mismatched expansion/shrinkage
further contribute to making HAGBs become the prefer-
ential fatigue cracking sites (Li et al. 2013). Combined
with the previous discussions in TEM and TKD character-
isations, the impediment and stabilisation of active dislo-
cation flow by cellular substructure can delay the
accumulation of dislocations at HAGBs, thus delaying
the fatigue crack initiation and improving fatigue per-
formance of SLMed 304L SS. In addition, we think that
the complex intergranular fatigue cracking mechanism
Figure 10. Schematic illustration of dislocation flow of cell wall and cell interior under external force. (a) As-built. (b) The dislocation
annihilation of cell interior by the opposite Burgers vector direction of GNDs of two sides of cell walls. (c) The formation of dislocation
lines by the dislocation emission. (d) The formation and propagation of dislocation lines associated with the decrease of GNDs around
the cell walls. (e) The twins nucleate from the HAGBs and propagate along the dislocation lines. (f) The second set of twins makes the
first set of twins buckle.
16 H. ZHANG ET AL.

of cellular boundaries, twin boundaries and HAGBs is still
not clear enough and needs further study.
3.6. Fatigue crack initiation
Figure 11(a) and (e) show SEM images of the fatigue
crack extracted from Figure 6(S2). The effect of micro-
structures on the process of the fatigue crack initiation
and crack propagation was explored, which is of great
significance to the fatigue resistance of additively man-
ufactured parts. The fatigue cracking initiates at the
molten pool boundaries, as shown in the enlarged
image of Figure 11(a). In addition, the fatigue cracks
nucleate exactly from the interface between the
HAGBs around molten pool boundaries and propagate
along HAGBs according to Figure 11(b, c). Finally, these
fatigue cracks are blocked by other HAGBs. Therefore,
the molten pool boundary may become the stress-sensi-
tive region during fatigue loading due to higher defect
possibilities. Once the HAGBs of grains coincides with
the molten pool boundaries, the overlapping position
may become the initiation point of fatigue cracking.
Improving the lap quality between adjacent molten
pools may improve the fatigue strength by reducing
the probability of fatigue crack nucleation. From the
equiaxed grains marked as red curves in Figure 11(c)
and the GB map in Figure 11(b), the molten pools first
form an equiaxed grain from the centre of the molten
pool, while other grains expand and grow towards the
boundaries of the molten pool with this equiaxed
grains as the centre. The larger KAM map shows that
the areas around the crack have higher plastic defor-
mation, as shown in Figure 11(d). In addition, the GBs
boundaries map, IPF map and KAM values for F2 show
these phenomena again (Figure 11(f, g, h)).
The typical grains (Figure 11(i, k)) containing twins
extracted from Figure 11(i) show that the Schmidt
factor of twins is higher than that of the substrate,
which means twins are easier to activate the slip
system than the matrix grains. This result is similar to
that of tensile testing in Figure 5. Additionally, we
draw the GOS map of F4 in Figure 11(l) to show the
plastic deformation of grains, which reveals that the
grains with the largest deformation are often distributed
on the outer grains of the ‘nested’structure in Figure 1
(k). Therefore, the growth along building direction
(<001> crystallographic orientation) of grains of the
new layer affects the shape and orientation of the orig-
inal grains of the underlying layers. The external strains
make the outer elongated grains easier to deform
owing to the lower deformation resistance, thus
decreasing the fatigue resistance. Combined with dis-
cussions in Section 3.5, the HAGBs of the outer grains
of the ‘nested’structure have higher interface energy
and high residual GB modulus, resulting in higher
residual stress. The high residual stress increases the
true stress amplitude during cyclic loading, leading to
the decrease of deformation resistance of outer grains
of ‘nested’structure, contributing to fatigue crack
initiation, thus reducing the fatigue life of additively
manufactured austenitic steels.
4. Conclusions
The mechanism of microstructure on the fatigue proper-
ties of SLMed 304L SS was deeply studied by numerous
microstructural characterisations. Based on the exper-
imental results and analysis, the key conclusions can
be drawn as follows.
(1) The unique heterogeneous microstructures of addi-
tively 304L SS with low SFE can significantly improve
the fatigue resistance by regulating mobile dislo-
cation motion during cyclic deformation. During
the whole fatigue history, the heterogeneous micro-
structure shows different working enhancement
behaviours in each cyclic deformation stage. In the
stage of residual dislocation accumulation for indu-
cing fatigue crack initiation, the mobile dislocations
caused by cyclic deformation are constrained by cel-
lular boundaries and can slightly penetrate the cellu-
lar boundary due to the high strain rate. These
behaviours result in the migration of cellular bound-
aries, the formation of dislocation lines, and provide
the driving force for the formation of subsequent
interlaced twins. After twin formation, the combi-
nation of cellular boundary, dislocation line, twin
boundary, and HAGBs tunes active dislocation
motion during cyclic deformation, further enhance
the resistance of fatigue crack initiation.
(2) The intergranular fatigue cracking appears predomi-
nantly at HAGBs around molten pool boundaries
rather than twins induced by cyclic deformation.
The mobile dislocations induced by cyclic defor-
mation cannot pill-up at twin boundaries under
the coordination of cellular dislocation structure
and dislocation lines. These mobile dislocations
cannot be transferred through HAGBs, and even-
tually pill-up at HAGBs under cyclic loading. In
addition, the high mismatched expansion/contrac-
tion at the molten pool boundary leads to the
high interface energy and high residual GB
modulus at HAGBs, making HAGBs around the
molten pool boundary become the preferential
sites of intergranular fatigue cracking.
VIRTUAL AND PHYSICAL PROTOTYPING 17

(3) Cyclic deformation accumulation can lead to the
lattice rotation and reduce the preferential <110>
texture strength of as-built parts, resulting in more
random texture.
(4) The outer grains of ‘nested’structures induced by
the print process have lower deformation resistance,
which may reduce the fatigue property of additively
manufactured parts.
This work has systematically determined the mechan-
ism of heterogeneous microstructure to improve the
fatigue resistance, thoroughly investigated the fatigue
failure process of additively manufactured austenitic
stainless steels, and deepened the understanding of
excellent additive manufacturing technology. The work
suggests that optimising process parameters of SLM
process and adding additional post-processing can be
used to tune the unique microstructure to obtain the
desired fatigue property.
Acknowledgments
The work was supported by National Natural Science Foun-
dation of China (grant number 52075087), the Fundamental
Research Funds for the Central Universities (grant number
N2003006), National Natural Science Foundation of China
(grant number U1708254). Special thanks are due to the
instrumental from Analytical and Testing Center, Northeastern
University.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
The work was supported by National Natural Science Foun-
dation of China (grant number 52075087), the Fundamental
Research Funds for the Central Universities (grant number
N2003006), National Natural Science Foundation of China
(grant number U1708254).
Notes on contributors
Hongzhuang Zhang is currently a PhD candidate of Northeast-
ern University.
Mengtao Xu is currently a PhD candidate of Northeastern
University.
Punit Kumar is currently Research Fellow of Nanyang Techno-
logical University.
Changyou Li is a professor of Northeastern University.
Weibing Dai is currently a PhD candidate of Northeastern
University.
Zhendong Liu is currently a PhD candidate of Northeastern
University.
Figure 11. The EBSD maps containing fatigue crack for exploring the fatigue crack initiation. (a-d) SEM image with an enlarged image
of crack, SEM image with GBs overlay, SEM image with IPF map overlay and KAM map for F6, respectively. (e-h) SEM image, SEM image
with GBs overlay, SEM image with IPF overlay and KAM map for F2, respectively. (i) IPF map of F3. (j-k) typical grains containing twins
extracted from (i). (l) GOS map for showing the deformation of grains of F3.
18 H. ZHANG ET AL.

Zhenyuan Li is currently a PhD candidate of Northeastern
University.
Yimin Zhang is a professor of Shenyang University of Chemical
Technology.
References
Andreau, Olivier, Etienne Pessard, Imade Koutiri, Patrice Peyre,
and Nicolas Saintier. 2021.“Influence of the Position and
Size of Various Deterministic Defects on the High Cycle
Fatigue Resistance of a 316L Steel Manufactured by Laser
Powder Bed Fusion.”International Journal of Fatigue 143:
105930. doi:10.1016/j.ijfatigue.2020.105930.
Angel, Tryggve. 1954.“Formation of Martensite in Austenitic
Stainless Steels.”Journal of Iron and Steel International 177:
165–174.
Bartlett, Jamison L., and Xiaodong Li. 2019.“An Overview of
Residual Stresses in Metal Powder Bed Fusion.”Additive
Manufacturing 27: 131–149. doi:10.1016/j.addma.2019.02.
020.
Basak, Amrita, and Suman Das. 2016.“Epitaxy and
Microstructure Evolution in Metal Additive Manufacturing.”
Annual Review of Materials Research 46 (1): 125–149.
doi:10.1146/annurev-matsci-070115-031728.
Cao, Yinfeng, Ziad Moumni, Jihong Zhu, Yahui Zhang, Yajun
You, and Weihong Zhang. 2020.“Comparative
Investigation of the Fatigue Limit of Additive-
Manufactured and Rolled 316 Steel Based on Self-Heating
Approach.”Engineering Fracture Mechanics 223: 106746.
doi:10.1016/j.engfracmech.2019.106746.
Cunningham, Ross, Cang Zhao, Niranjan Parab, Christopher
Kantzos, Joseph Pauza, Kamel Fezzaa, Tao Sun, and
Anthony D. Rollett. 2019.“Keyhole Threshold and
Morphology in Laser Melting Revealed by Ultrahigh-Speed
X-ray Imaging.”Science 363 (6429): 849–852. doi:10.1126/
science.aav4687.
Darvish, K., Z. W. Chen, and T. Pasang. 2016.“Reducing Lack of
Fusion During Selective Laser Melting of CoCrMo Alloy:
Effect of Laser Power on Geometrical Features of Tracks.”
Materials & Design 112: 357–366. doi:10.1016/j.matdes.
2016.09.086.
DebRoy, T., H. L. Wei, J. S. Zuback, T. Mukherjee, J. W. Elmer, J. O.
Milewski, A. M. Beese, A. Wilson-Heid, A. De, and W. Zhang.
2018.“Additive Manufacturing of Metallic Components –
Process, Structure and Properties.”Progress in
Materials Science 92: 112–224. doi:10.1016/j.pmatsci.2017.
10.001.
Elangeswaran, Chola, Antonio Cutolo, Gokula Krishna
Muralidharan, Kim Vanmeensel, and Brecht Van
Hooreweder. 2020.“Microstructural Analysis and Fatigue
Crack Initiation Modelling of Additively Manufactured 316L
after Different Heat Treatments.”Materials & Design 194:
108962. doi:10.1016/j.matdes.2020.108962.
Ferreri, Nicholas C, Reeju Pokharel, Veronica Livescu, Donald W
Brown, Marko Knezevic, Jun-Sang Park, Michael A Torrez,
and George T Gray III. 2020.“Effects of Heat Treatment
and Build Orientation on the Evolution of ɛand α′
Martensite and Strength During Compressive Loading of
Additively Manufactured 304L Stainless Steel.”Acta
Materialia 195: 59–70.
Gao, Shubo, Zhiheng Hu, Martial Duchamp, PS Sankara Rama
Krishnan, Sravya Tekumalla, Xu Song, and Matteo Seita.
2020.“Recrystallization-Based Grain Boundary Engineering
of 316L Stainless Steel Produced via Selective Laser
Melting.”Acta Materialia 200: 366–377.
Ghayoor, Milad, Kijoon Lee, Yujuan He, Chih-hung Chang, Brian
K Paul, and Somayeh Pasebani. 2020.“Selective Laser
Melting of 304L Stainless Steel: Role of Volumetric Energy
Density on the Microstructure, Texture and Mechanical
Properties.”Additive Manufacturing 32: 101011.
Gutierrez-Urrutia, I., S. Zaefferer, and D. Raabe. 2010.“The Effect
of Grain Size and Grain Orientation on Deformation
Twinning in a Fe–22wt.% Mn–0.6wt.% C TWIP Steel.”
Materials Science and Engineering: A 527 (15): 3552–3560.
doi:10.1016/j.msea.2010.02.041.
Herzog, Dirk, Vanessa Seyda, Eric Wycisk, and Claus
Emmelmann. 2016.“Additive Manufacturing of Metals.”
Acta Materialia 117: 371–392.
Hou, Juan, Wei Chen, Zhuoer Chen, Kai Zhang, and Aijun
Huang. 2020.“Microstructure, Tensile Properties and
Mechanical Anisotropy of Selective Laser Melted 304L
Stainless Steel.”Journal of Materials Science & Technology
48: 63–71. doi:10.1016/j.jmst.2020.01.011.
Kasperovich, Galina, and Joachim Hausmann. 2015.
“Improvement of Fatigue Resistance and Ductility of
TiAl6V4 Processed by Selective Laser Melting.”Journal of
Materials Processing Technology 220: 202–214. doi:10.1016/
j.jmatprotec.2015.01.025.
Khairallah, Saad A., Aiden A. Martin, Jonathan R. I. Lee, Gabe
Guss, Nicholas P. Calta, Joshua A. Hammons, Michael H.
Nielsen, et al. 2020.“Controlling Interdependent Meso-
Nanosecond Dynamics and Defect Generation in Metal 3D
Printing.”Science 368 (6491): 660–665. doi:10.1126/science.
aay7830.
Kong, Decheng, Chaofang Dong, Xiaoqing Ni, Zhang Liang,
Cheng Man, and Xiaogang Li. 2020.“Hetero-Deformation-
induced Stress in Additively Manufactured 316L Stainless
Steel.”Materials Research Letters 8 (10): 390–397. doi:10.
1080/21663831.2020.1775149.
Kong, Decheng, Chaofang Dong, Xiaoqing Ni, Liang Zhang,
Jizheng Yao, Cheng Man, Xuequn Cheng, Kui Xiao, and
Xiaogang Li. 2019.“Mechanical Properties and Corrosion
Behavior of Selective Laser Melted 316L Stainless Steel
after Different Heat Treatment Processes.”Journal of
Materials Science & Technology 35 (7): 1499–1507. doi:10.
1016/j.jmst.2019.03.003.
Koyama, Motomichi, Zhao Zhang, Meimei Wang, Dirk Ponge,
Dierk Raabe, Kaneaki Tsuzaki, Hiroshi Noguchi, and Cemal
Cem Tasan. 2017.“Bone-Like Crack Resistance in
Hierarchical Metastable Nanolaminate Steels.”Science 355
(6329). doi:10.1126/science.aal2766.
Kumar, Punit, R. Jayaraj, J. Suryawanshi, U. R. Satwik, J.
McKinnell, and U. Ramamurty. 2020.“Fatigue Strength of
Additively Manufactured 316L Austenitic Stainless Steel.”
Acta Materialia 199: 225–239.
Kumar, K. S., H. Van Swygenhoven, and S. Suresh. 2003.
“Mechanical Behavior of Nanocrystalline Metals and
Alloys11 the Golden Jubilee Issue—Selected Topics in
Materials Science and Engineering: Past, Present and Future,
Edited by S. Suresh.”Acta Materialia 51 (19): 5743–5774.
Kuo, C. N., C. K. Chua, P. C. Peng, Y. W. Chen, S. L. Sing, S. Huang,
and Y. L. Su. 2020.“Microstructure Evolution and Mechanical
VIRTUAL AND PHYSICAL PROTOTYPING 19

Property Response via 3D Printing Parameter Development
of Al–Sc Alloy.”Virtual and Physical Prototyping 15 (1): 120–
129. doi:10.1080/17452759.2019.1698967.
Lee, Seungjong, Jonathan W. Pegues, and Nima Shamsaei.
2020.“Fatigue Behavior and Modeling for Additive
Manufactured 304L Stainless Steel: The Effect of Surface
Roughness.”International Journal of Fatigue 141: 105856.
Leuders, S., M. Thöne, A. Riemer, T. Niendorf, T. Tröster, H. A.
Richard, and H. J. Maier. 2013.“On the Mechanical
Behaviour of Titanium Alloy TiAl6V4 Manufactured by
Selective Laser Melting: Fatigue Resistance and Crack
Growth Performance.”International Journal of Fatigue 48:
300–307. doi:10.1016/j.ijfatigue.2012.11.011.
Li, Haizhou, Hongyang Jing, Lianyong Xu, Yongdian Han, Lei
Zhao, Zhengxin Tang, Bo Xiao, and Yu Zhang. 2019a.
“Fatigue Behavior, Microstructural Evolution, and Fatigue
Life Model Based on Dislocation Annihilation of an Fe-Ni-
Cr Alloy at 700°C.”International Journal of Plasticity 118:
105–129. doi:10.1016/j.ijplas.2019.02.006.
Li, Zan, Zhiqiang Li, Zhanqiu Tan, Ding-Bang Xiong, and Qiang
Guo. 2020.“Stress Relaxation and the Cellular Structure-
Dependence of Plastic Deformation in Additively
Manufactured AlSi10Mg Alloys.”International Journal of
Plasticity 127: 102640. doi:10.1016/j.ijplas.2019.12.003.
Li, Zhiming, Konda Gokuldoss Pradeep, Yun Deng, Dierk Raabe,
and Cemal Cem Tasan. 2016.“Metastable High-Entropy
Dual-Phase Alloys Overcome the Strength-Ductility Trade-
Off.”Nature 534 (7606): 227–230. doi:10.1038/nature17981.
Li, Yajing, Dunji Yu, Bingbing Li, and Xu Chen. 2019b.
“Martensitic Transformation of an Austenitic Stainless Steel
Under Non-proportional Cyclic Loading.”International
Journal of Fatigue 124: 338–347. doi:10.1016/j.ijfatigue.
2019.03.020.
Li, Linlin, Zhenjun Zhang, Peng Zhang, Cuihong Li, and
Zhefeng Zhang. 2018.“Dislocation Arrangements Within
Slip Bands During Fatigue Cracking.”Materials
Characterization 145: 96–100. doi:10.1016/j.matchar.2018.
08.039.
Li, L. L., P. Zhang, Z. J. Zhang, and Z. F. Zhang. 2013.“Effect of
Crystallographic Orientation and Grain Boundary Character
on Fatigue Cracking Behaviors of Coaxial Copper
Bicrystals.”Acta Materialia 61 (2): 425–438. doi:10.1016/j.
actamat.2012.08.045.
Liu, Leifeng, Qingqing Ding, Yuan Zhong, Ji Zou, Jing Wu, Yu-
Lung Chiu, Jixue Li, Ze Zhang, Qian Yu, and Zhijian Shen.
2018.“Dislocation Network in Additive Manufactured Steel
Breaks Strength-Ductility Trade-off.”Materials Today 21 (4):
354–361. doi:10.1016/j.mattod.2017.11.004.
Lu, Yanjin, Songquan Wu, Yiliang Gan, Tingting Huang,
Chuanguang Yang, Junjie Lin, and Jinxin Lin. 2015.“Study
on the Microstructure, Mechanical Property and Residual
Stress of SLM Inconel-718 Alloy Manufactured by Differing
Island Scanning Strategy.”Optics and Laser Technology 75:
197–206. doi:10.1016/j.optlastec.2015.07.009.
Ma, Mingming, Zemin Wang, and Xiaoyan Zeng. 2017.“A
Comparison on Metallurgical Behaviors of 316L Stainless
Steel by Selective Laser Melting and Laser Cladding
Deposition.”Materials Science and Engineering: A 685: 265–
273. doi:10.1016/j.msea.2016.12.112.
MacDonald, E., and R. Wicker. 2016.“Multiprocess 3D Printing
for Increasing Component Functionality.”Science 353
(6307): aaf2093. doi:10.1126/science.aaf2093.
Maleki, Erfan, Sara Bagherifard, Michele Bandini, and Mario
Guagliano. 2020.“Surface Post-Treatments for Metal
Additive Manufacturing: Progress, Challenges, and
Opportunities.”Additive Manufacturing 101619. doi:10.
1016/j.addma.2020.101619.
Martin, John H., Brennan D. Yahata, Jacob M. Hundley, Justin A.
Mayer, Tobias A. Schaedler, and Tresa M. Pollock. 2017.“3D
Printing of High-Strength Aluminium Alloys.”Nature 549
(7672): 365–369. doi:10.1038/nature23894.
Nadammal, Naresh, Tatiana Mishurova, Tobias Fritsch, Itziar
Serrano-Munoz, Arne Kromm, Christoph Haberland, Pedro
Dolabella Portella, and Giovanni Bruno. 2021.“Critical Role
of Scan Strategies on the Development of Microstructure,
Texture, and Residual Stresses During Laser Powder Bed
Fusion Additive Manufacturing.”Additive Manufacturing
38: 101792. doi:10.1016/j.addma.2020.101792.
Obrtlík, Karel, TomásšKruml, and Jaroslav Polák. 1994.
“Dislocation Structures in 316L Stainless Steel Cycled with
Plastic Strain Amplitudes Over a Wide Interval.”Materials
Science and Engineering: A 187 (1): 1–9.
Parvez, M., Y. Chen, J. Newkirk, and F. F. Liou. 2019.
“Comparison of Fatigue Performance Between Additively
Manufactured and Wrought 304L Stainless Steel Using a
Novel Fatigue Test Setup.”Proceedings of the Solid
Freeform Fabrication, 353–363.
Pegues, Jonathan W., Michael D. Roach, and Nima Shamsaei.
2020.“Additive Manufacturing of Fatigue Resistant
Austenitic Stainless Steels by Understanding Process-
Structure–Property Relationships.”Materials Research
Letters 8 (1): 8–15. doi:10.1080/21663831.2019.1678202.
Pham, M. S., B. Dovgyy, and P. A. Hooper. 2017a.“Twinning
Induced Plasticity in Austenitic Stainless Steel 316L Made by
Additive Manufacturing.”Materials Science and Engineering:
A704: 102–111. doi:10.1016/j.msea.2017.07.082.
Pham, M. S., B. Dovgyy, and P. A. Hooper. 2017b.“Twinning
Induced Plasticity in Austenitic Stainless Steel 316L Made
by Additive Manufacturing.”Materials Science and
Engineering: A 704: 102–111. doi:10.1016/j.msea.2017.07.
082.
Pokharel, Reeju, Anirban Patra, Donald W. Brown, Bjorn
Clausen, Sven C. Vogel, and George T. Gray III. 2019.“An
Analysis of Phase Stresses in Additively Manufactured 304L
Stainless Steel Using Neutron Diffraction Measurements
and Crystal Plasticity Finite Element Simulations.”
International Journal of Plasticity 121: 201–217. doi:10.
1016/j.ijplas.2019.06.005.
Polatidis, E., J. Capek, A. Arabi-Hashemi, C. Leinenbach, and M.
Strobl. 2020.“High Ductility and Transformation-Induced-
Plasticity in Metastable Stainless Steel Processed by
Selective Laser Melting with Low Power.”Scripta Materialia
176: 53–57. doi:10.1016/j.scriptamat.2019.09.035.
Polatidis, E., W. N. Hsu, M. Smid, T. Panzner, S. Chakrabarty, P.
Pant, and H. Van Swygenhoven. 2018.“Suppressed
Martensitic Transformation Under Biaxial Loading in Low
Stacking Fault Energy Metastable Austenitic Steels.”Scripta
Materialia 147: 27–32. doi:10.1016/j.scriptamat.2017.12.026.
Prashanth, K. G., and J. Eckert. 2017.“Formation of Metastable
Cellular Microstructures in Selective Laser Melted Alloys.”
Journal of Alloys and Compounds 707: 27–34. doi:10.1016/j.
jallcom.2016.12.209.
Qiu, Chunlei, Nicholas J. E. Adkins, and Moataz M. Attallah.
2013.“Microstructure and Tensile Properties of Selectively
20 H. ZHANG ET AL.

Laser-Melted and of HIPed Laser-Melted Ti-6Al-4V.”Materials
Science and Engineering: A 578: 230–239. doi:10.1016/j.msea.
2013.04.099.
Rahman, K. M., V. A. Vorontsov, and D. Dye. 2015.“The Effect of
Grain Size on the Twin Initiation Stress in a TWIP Steel.”Acta
Materialia 89: 247–257. doi:10.1016/j.actamat.2015.02.008.
Romanov, A., V. I. Vladimirov, and F. R. N. Nabarro. 1992.
“Disclinations in Crystalline Solids.”In Dislocations in Solids,
edited by F. R. N. Nabarro, 135–136. North Holland:
Elsevier Science Publishers, Amsterdam etc.
Saeidi, K., X. Gao, Y. Zhong, and Z. J. Shen. 2015.“Hardened
Austenite Steel with Columnar Sub-Grain Structure
Formed by Laser Melting.”Materials Science and
Engineering: A 625: 221–229. doi:10.1016/j.msea.2014.12.
018.
Sanaei, Niloofar, and Ali Fatemi. 2020.“Defects in Additive
Manufactured Metals and Their Effect on Fatigue
Performance: A State-of-the-Art Review.”Progress in
Materials Science, 100724. doi:10.1016/j.pmatsci.2020.
100724.
Schramm, R. E., and R. P. Reed. 1975.“Stacking Fault Energies of
Seven Commercial Austenitic Stainless Steels.”Metallurgical
Transactions A 6 (7): 1345–1351.
Shen, Y. F., N. Jia, Y. D. Wang, X. Sun, L. Zuo, and D. Raabe. 2015.
“Suppression of Twinning and Phase Transformation in an
Ultrafine Grained 2GPa Strong Metastable Austenitic Steel:
Experiment and Simulation.”Acta Materialia 97: 305–315.
doi:10.1016/j.actamat.2015.06.053.
Smith, Thale R., Joshua D. Sugar, Julie M. Schoenung, and Chris
San Marchi. 2019.“Relationship Between Manufacturing
Defects and Fatigue Properties of Additive Manufactured
Austenitic Stainless Steel.”Materials Science and
Engineering: A 765: 138268. doi:10.1016/j.msea.2019.138268.
Sun, Zhongji, Xipeng Tan, Shu Beng Tor, and Chee Kai Chua.
2018.“Simultaneously Enhanced Strength and Ductility for
3D-Printed Stainless Steel 316L by Selective Laser Melting.”
NPG Asia Materials 10: 127–136. doi:10.1038/s41427-018-
0018-5.
Suryawanshi, Jyoti, K. G. Prashanth, and U. Ramamurty. 2017.
“Mechanical Behavior of Selective Laser Melted 316L
Stainless Steel.”Materials Science and Engineering: A 696:
113–121. doi:10.1016/j.msea.2017.04.058.
Svensson, M., U. Ackelid, and A. Ab. 2010.“Titanium Alloys
Manufactured with Electron Beam Melting Mechanical and
Chemical Properties.”Paper Presented at the Proceedings
of the materials and Processes for Medical Devices
Conference.
Thijs, Lore, Frederik Verhaeghe, Tom Craeghs, Jan Van
Humbeeck, and Jean-Pierre Kruth. 2010.“A Study of the
Microstructural Evolution During Selective Laser Melting of
Ti–6Al–4V.”Acta Materialia 58 (9): 3303–3312. doi:10.1016/
j.actamat.2010.02.004.
Wang, Zhuqing, Todd A. Palmer, and Allison M. Beese. 2016.
“Effect of Processing Parameters on Microstructure and
Tensile Properties of Austenitic Stainless Steel 304L Made by
Directed Energy Deposition Additive Manufacturing.”Acta
Materialia 110: 226–235. doi:10.1016/j.actamat.2016.03.019.
Wang, Y. Morris, Thomas Voisin, Joseph T. McKeown, Jianchao
Ye, Nicholas P. Calta, Zan Li, Zhi Zeng, et al. 2018.“Additively
Manufactured Hierarchical Stainless Steels with High
Strength and Ductility.”Nature Materials 17 (1): 63–71.
doi:10.1038/nmat5021.
Wei, Yujie, Yongqiang Li, Lianchun Zhu, Yao Liu, Xianqi Lei,
Gang Wang, Yanxin Wu, et al. 2014.“Evading the
Strength- Ductility Trade-offDilemma in Steel Through
Gradient Hierarchical Nanotwins.”Nature Communications
5. doi:10.1038/ncomms4580.
Yu, Wenhui, Swee Leong Sing, Chee Kai Chua, and Xuelei Tian.
2019.“Influence of re-Melting on Surface Roughness and
Porosity of AlSi10Mg Parts Fabricated by Selective Laser
Melting.”Journal of Alloys and Compounds 792: 574–581.
doi:10.1016/j.jallcom.2019.04.017.
Zhang, Jinliang, Jianbao Gao, Bo Song, Lijun Zhang, Changjun
Han, Chao Cai, Kun Zhou, and Yusheng Shi. 2021.“A Novel
Crack-Free Ti-Modified Al-Cu-Mg Alloy Designed for
Selective Laser Melting.”Additive Manufacturing 38:
101829. doi:10.1016/j.addma.2020.101829.
Zhang, Hongzhuang, Changyou Li, Weibing Dai, Yuzhuo Liu,
Sen Tian, Wenchao Huang, Dawei Jia, David He, and Yimin
Zhang. 2020a.“Static Compression Testing CFRP Single-lap
Composited Joints Using X-ray μCT.”Composite Structures
234: 111667. doi:10.1016/j.compstruct.2019.111667.
Zhang, Hongzhuang, Changyou Li, Mengtao Xu, Weibing Dai,
Punit Kumar, Zhendong Liu, Zhenyuan Li, and Yimin
Zhang. 2020b.“The Fatigue Performance Evaluation of
Additively Manufactured 304L Austenitic Stainless Steels.”
Materials Science and Engineering: A, 140640. doi:10.1016/j.
msea.2020.140640.
Zhang, Hongzhuang, Changyou Li, Mengtao Xu, Zhendong Liu,
Dawei Jia, and Yimin Zhang. 2020c.“A Novel Method for
Damage Analysis of CFRP Single-Lap Bolted, Bonded and
Hybrid Joints Under Compression.”Composite Structures
251: 112636. doi:10.1016/j.compstruct.2020.112636.
Zhang, Hongzhuang, Changyou Li, Mengtao Xu, Zhendong Liu,
Dawei Jia, and Yimin Zhang. 2020d.“A Novel Method for
Damage Analysis of CFRP Single-Lap Bolted, Bonded and
Hybrid Joints Under Compression.”Composite Structures
251: 112636.
Zhang, Z. F., and Z. G. Wang. 2008.“Grain Boundary Effects on
Cyclic Deformation and Fatigue Damage.”Progress in
Materials Science 53 (7): 1025–1099. doi:10.1016/j.pmatsci.
2008.06.001.
Zhao, Jianguang, Juan Hou, Liang Chen, Binbin Dai, Xiaojing
Xiong, Lei Tan, Kai Zhang, and Aijun Huang. 2020.
“Evaluating Impact Performance of a Selective Laser
Melted 304L Stainless Steel with Weak Texture.”Materials
Today Communications 25: 101299.
Zhou, Xin, Kailun Li, Dandan Zhang, Xihe Liu, Jing Ma, Wei Liu,
and Zhijian Shen. 2015.“Textures Formed in a CoCrMo
Alloy by Selective Laser Melting.”Journal of Alloys and
Compounds 631: 153–164. doi:10.1016/j.jallcom.2015.01.
096.
Zhu, Zhiguang, Weilin Li, Quy Bau Nguyen, Xianghai An,
Wenjun Lu, Zhiming Li, Fern Lan Ng, Sharon Mui Ling Nai,
and Jun Wei. 2020.“Enhanced Strength–Ductility Synergy
and Transformation-Induced Plasticity of the Selective
Laser Melting Fabricated 304 L Stainless Steel.”Additive
Manufacturing 35: 101300.
VIRTUAL AND PHYSICAL PROTOTYPING 21