International Journal of Fatigue 176 (2023) 107839

Contents lists available at ScienceDirect

International Journal of Fatigue

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

Investigation of fatigue behavior of laser powder bed fusion Ti-6Al-4V:

Roles of heat treatment and microstructure
Jianwen Liu a, b, c, Kai Zhang a, b, c, *, Jie Liu a, Hao Wang a, Yi Yang a, Liangming Yan d,
Xinni Tian b, c, e, Yuman Zhu b, c, Aijun Huang b, c
a
School of Materials and Chemistry, University of Shanghai for Science and Technology, Shanghai 200093, China
b
Monash Center for Additive Manufacturing (MCAM), Monash University, Notting Hill, VIC 3168, Australia
c
Department of Materials Science and Engineering, Monash University, Clayton, VIC 3800, Australia
d
College of Materials Science and Engineering, Inner Mongolia University of Technology, Hohhot, Inner Mongolia 010051, China
e
Suzhou Industrial Park Monash Research Institute of Science and Technology, Suzhou, China

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

Keywords: High cycle fatigue (HCF) performance of additive manufactured (AM) titanium products is of great significance
Laser powder bed fusion for its structural and functional applications. However, tying fatigue performance to the complex microstructures
Ti-6Al-4V in AM titanium alloys is challenging. The work here carried out a thorough investigation into the influence of
Grain boundary α-phase
microstructures (particularly the grain boundary α-phase (GB-α) with varied morphologies) on the fatigue per­
Fatigue performance
formance of laser powder bed fusion (LPBF) Ti–6Al–4V (Ti-64). Results showed that the improvement in the
high-cycle fatigue life of LPBF Ti-64 could be achieved by the formation of low aspect ratio α lath and discon­
tinuous GB-α via optimized post-fabrication heat treatment. Discontinuous GB-α could fully accommodate the
deformation, improving the fatigue crack propagation resistance. Moreover, α lath with a low aspect ratio could
lead to less strain accumulation on the interface between adjacent α lath, and thereby inhibit the crack initiation
at these interfaces. This study enhances the understanding of how LPBF-induced complex microstructures in­
fluence fatigue behavior, and provides a pathway for the improvement of fatigue performance of additive
manufactured titanium alloy.

1. Introduction along the build direction [10,11]. High residual stresses in conjunction
with the presence of acicular α’ martensite causes high strength, limited
Titanium alloy Ti-6Al-4V (Ti-64) is an advanced lightweight material ductility, and low fracture resistance of as-built LPBF Ti-64 [12].
that can fulfill the demand for structural applications with high specific Post-fabrication heat treatment could transform the acicular α’
strength, high toughness, and high fracture resistance [1–3]. Laser martensite into an equilibrium α+β phase and reduce thermal stress,
powder bed fusion (LPBF), as a typical metal additive manufacturing which leads to ductility improvement [13,14]. However, GB-α, which is
(AM) technology, could directly fabricate complex-shaped 3D metal hardly observed in an as-built state under mesoscale characterization
parts layer-by-layer, which provides higher design freedom and low (like SEM), could form at the early heat treatment stage [15,16]. With
material waste [4–6]. Titanium alloys fabricated by LPBF have been the further process of heat treatment, GB-α could get coarsened
successfully applied in applications in multiple industries, including following the classical Lifshitz, Slyozov, and Wagner (LSW) theory [15].
aerospace, nuclear, and automotive applications [7,8]. The intrinsic In LPBF Ti-64 alloy, continuous GB-α embedded at columnar β grain
thermal history of the LPBF process could introduce the unique micro­ boundary further result in anisotropic tensile properties [17,18]. Spe­
structure into titanium alloy. In LPBF Ti-64 alloy, for instance, the cifically, in the sample with tension applied along the transverse di­
extremely high cooling rates (up to 106 ◦ C/s) could lead to the formation rection, the tensile load is perpendicular to the grain boundaries and acts
of undesired acicular α′ martensite microstructure and high residual to separate the adjacent two columnar β grains [17,19,20]. This leads to
stresses, while preventing the grain boundary α-phase (GB-α) formation reduced ductility compared with the sample loading along the longitu­
[9]. The strong directional solidification could result in columnar grains dinal direction [17,21]. With the heat treatment temperature close to

* Corresponding author at: School of Materials and Chemistry, University of Shanghai for Science and Technology, Shanghai 200093, China.
E-mail address: [email protected] (K. Zhang).

https://doi.org/10.1016/j.ijfatigue.2023.107839
Received 11 May 2023; Received in revised form 14 July 2023; Accepted 16 July 2023
Available online 19 July 2023
0142-1123/© 2023 Elsevier Ltd. All rights reserved.
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 1. (a) Scanning electron microscopy images showing the Ti-64 powders used in this study, (b) The powder size distribution of the Ti-64 powders, (c) Schematic
illustration showing the transverse fatigue sample cutting from the printed part, and (d) Geometry size of fatigue specimens.

the β-transus, GB-α tends to become discontinuous [19,20]. Discontin­ treatment temperatures combined with shorter dwell time could result
uous GB-α could accommodate uniform deformation with adjacent prior in thinner α lath and thereby could lead to low fracture toughness and
β-grain under loading in both the transverse or longitudinal directions, fatigue performance [30,31]. This is attributed to the fact that fine α lath
which reduces the tensile property anisotropy (even though columnar thickness with smaller effective slip length could reduce both the fatigue
grains morphology still exists) [22,23]. crack initiation and fatigue crack growth resistance of LPBF Ti-64. In
Although the influence of the GB-α morphology on the tensile comparison, high heat treatment temperatures combined with long
properties in AM titanium alloy has been successfully revealed, its ef­ dwell time could enable α lath to break-up and become more equiaxed,
fects on the fatigue behavior which includes the high-cycle fatigue life, which enhances the fatigue performance of additive manufactured Ti-
fatigue crack initiation, and propagation behavior remain unknown. 64. This is since that α lath with more equiaxed morphology could
Only a few studies in conventional manufactured titanium alloys have accommodate deformations and reduce the stress concentration at the
demonstrated that continuous GB-α could provide the preferred lath interface, which could increase the fatigue crack initiation resis­
pathway for fatigue crack propagation and thus degrade fatigue per­ tance [32].
formance [24,25]. However, due to the high heating and cooling rates of While there have been some published studies reported on the in­
the LPBF process, the microstructures in LPBF Ti-64 are remarkably fluence of heat treatment on the tensile and fatigue performances of
different from that in conventional manufactured Ti-64, which causes LPBF Ti-64 [33,34], it still remains unclear how LPBF-induced micro­
the GB-α related fatigue behavior knowledge in conventionally manu­ structures (especially regarding the GB-α with varied morphologies,
factured titanium alloys not transferrable to LPBF Ti-64 [26]. Previous coupled with α laths) affect the fatigue behavior of LPBF Ti-64. In this
studies have shown that GB-α in conventional manufactured Ti-64 could work, a thorough investigation into the influence of microstructure on
form in the as-fabricated state accompanied by forming widmanstätten α the fatigue performance of LPBF Ti-64 was conducted. The improvement
colonies with similar crystallographic orientations [27]. On the con­ in the high-cycle fatigue life of LPBF Ti-64 could be achieved by forming
trary, GB-α in LPBF Ti-64 is normally not identifiable in the as-built discontinuous GB-α and α lath with a low aspect ratio via optimal post-
state, while rapidly forming during the post subtransus annealing heat fabrication heat treatment. Further characterization revealed that α lath
treatment. This GB-α formation is not accompanied by α colony for­ with a low aspect ratio could lead to less strain accumulation on the α
mation and thus has different crystallographic orientation from the lath interface and reduce the tendency for crack initiation and propa­
adjacent α lath within prior-β grains [13,15,28]. This could lead to more gation at these interfaces. In the meanwhile, discontinuous GB-α with
strain incompatibility between GB-α with the surrounding grains during various crystallographic orientations could accommodate the deforma­
the deformation process, resulting in a different fatigue behavior tion homogeneously and thus increase the crack initiation and propa­
compared to that of GB-α in conventional manufactured titanium alloys. gation resistance. This study provides a pathway for the improvement of
Furthermore, the morphology of α lath in additive manufactured Ti- fatigue performance of additive manufactured titanium alloy.
64, as another influencing factor in determining the fracture-related
properties of titanium alloys, is also mainly controlled by the post-
fabrication heat treatment parameters [14,29]. Relatively low heat

2
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 2. Optical microscopy images of (a) as-built sample, (b) HT1 sample, and (c) HT2 sample, along the build direction. The statistical analysis of the prior-β grain
sizes of (d) as-built sample, (e) HT1 sample, and (f) HT2 sample.

Fig. 3. Representative backscattered electron micrographs of the (a) as-built sample, (b) HT1 sample, and (c) HT2 sample. The corresponding EBSD inverse pole
figure (IPF) orientation maps of (d) as-built sample, (e) HT1 sample, and (f) HT2 sample. GB-α was highlighted by black arrows in (e-f). GB-α was identified according
to their corresponding locations in the prior-β phase reconstructed IPF maps, Supplementary Fig. 1.

2. Experimental methods laser power 265 W, scan speed 1100 mm/s, hatch distance 0.15 mm, and
layer thickness 0.04 mm. The as-built samples with a density of 99.9 %
Gas-atomized spherical Ti-64 powder with a particle size range of were obtained. As-built sample (100 × 10 × 10 mm3) was heat-treated
18–58 μm was used in this work with the chemical compositions (in wt. at 850 ◦ C and 950 ◦ C for 2 h followed by water quenching (WQ) to
%) of Ti-6.20Al-4.04V-0.15O-0.010C-0.028N-0.0043H-0.20Fe. Scan­ achieve the different GB-α morphologies. 850◦ C/2h/WQ (HT1) was
ning electron microscopy (SEM) examination confirmed that the powder used to obtain continuous GB-α, while 950◦ C/2h/WQ (HT2) was used to
particles have spherical morphologies, as shown in Fig. 1 (a-b). Laser obtain discontinuous GB-α [19].
powder bed fusion (LPBF) fabrication was in the EOSINT M290 with Fatigue samples were extracted by using electrical discharge

3
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

mode (stress ratio R = 0.1) and at a frequency of 110 Hz under room

temperature. The tensile properties of the heat-treated samples were
obtained in our previous study [19]: the ultimate tensile stress (UTS)
and yield stress (YS) of LPBF Ti-64 after the HT1 process were 1041.8 ±
10.5 MPa and 942.4 ± 10.0 MPa, respectively; the UTS and YS of the
HT2 sample were 1010.8 ± 18.1 MPa and 890.1 ± 8.5 MPa,
respectively.
The samples for microstructure characterization were first ground
from 220# to 5000# SiC sandpaper and then polished with a mixture of
90 vol% silica suspension OP-S and 10 vol% H2O2. A ZEISS Gemini SEM
300 field emission gun scanning electron microscope (FEG-SEM) was
used to characterize microstructure features in backscattered electron
(BSE) mode. The FEG-SEM was operated at 20 kV acceleration voltage
and 10 mm work distance. Furthermore, the morphologies and crystal­
Fig. 4. High cycle fatigue (HCF) S-N curves for LPBF Ti-64 samples after HT1 lographic orientations of the microstructure were also evaluated via
and HT2 process. Horizontal arrows indicate the sample run-out after electron backscatter diffraction (EBSD). EBSD measurements were ac­
107 cycles. quired using an Oxford Symmetry detector apparatus at a step size of
0.25 µm and processed by AztecCrystal 2.1 software.
machining along the transverse direction, Fig. 1 (c). This ensured that
most GB-α would have the maximum influence on the fatigue perfor­ 3. Results
mance, given that columnar β-grain tends to be parallel to the building
direction. High-cycle fatigue (HCF) samples were machined in a Nadcap- 3.1. Microstructure characterization
certified lab and were prepared by using the longitudinal polishing
process to fully reduce the surface roughness and remove the surface The microstructure characterization shows that the as-built samples
stress layer introduced by the machining process. The HCF samples were consisted of columnar grain morphology along the longitudinal direc­
threaded cylindrical dogbones with a gauge diameter of 4 mm and a tion, with a size of 138.2 ± 39.3 μm and an aspect ratio of ~ 6.2, Fig. 2
length of 10 mm, which follows ASTM E466 (Fig. 1 (d)) [35]. RUMUL (a,d). This was a typical microstructure in AM-processed titanium alloys
100 kN fatigue machine was used to conduct the tests in tension–tension due to the small melt pool and strong thermal gradients [36].

Fig. 5. (a1-a2) and (b1-b2) are the respective characterized fracture surfaces of LPBF Ti-64 samples after the HT1 and HT2 processes: (a-1) HT1 sample failed at 600
MPa with 82,800 cycles, (a-2) failed at 600 MPa with 55,000 cycles, (b-1) HT2 sample failed at 600 MPa with 440,400 cycles, and (b-2) failed at 600 MPa with
118,100 cycles. Red dashed lines indicate the positions of longitudinal cross-sectional characterization in Section 3.3. (For interpretation of the references to color in
this figure legend, the reader is referred to the web version of this article.)

4
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 6. Secondary electron micrographs of the sample fracture surfaces, (a-c) HT1 sample failed at 600 MPa with 82,800 cycles, and (d-f) HT2 sample failed at 600
MPa with 440,400 cycles. Lip-like smooth areas were shown by blue arrows in (c). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)

Fig. 7. Optical morphologies of longitudinal cross-sectional characterization of HCF testing failed samples: (a) HT1 samples failed at 600 MPa with 82,800 cycles,
and failed at 600 MPa with 55,000 cycles, and (b) HT2 samples failed at 600 MPa with 440,400 cycles, and failed at 600 MPa with 118,100 cycles.

Meanwhile, high-magnification BSE and EBSD images show the acicular and HT2. It was found that prior-β grains retained the columnar
α’ martensite within columnar prior-β grains, which arise due to the morphology, which confirmed that LPBF Ti-64 heat treated below β
rapid cooling rate involved in the LPBF process, Fig. 3 (a, d). The average transus temperature inherits the prior-β grain morphology of the as-built
thickness of the α’ martensite was 0.53 ± 0.09 μm. Furthermore, the state. The average prior-β grain aspect ratio was measured as 5.3 in the
rapid cooling of the LPBF process could inhibit GB-α precipitation, and HT1 sample, and 4.9 in the HT2 sample. And the average prior-β grain
thus GB-α was not found on the prior-β grain boundaries in the as-built width was 145.4 ± 71.3 μm in the HT1 sample, and 158.3 ± 56.5 μm in
sample. Fig. 2 (b-e) shows the morphologies of prior-β grains after HT1 the HT2 sample.

5
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 8. (a) The EBSD IPF orientation maps of the HT1 sample failed at 600 MPa with 82,800 cycles. (b) The corresponding BSE image along the fatigue crack profile
of the HT1 sample. The prior-β grain boundaries are highlighted by yellow-color lines in the images. (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this article.)

Further detailed microstructure characterization of the HT1 sample This indicated that the formation of flat facets was strongly associated
shows that acicular α’ martensite was transformed to the equilibrium α with the columnar prior-β grains. In the meanwhile, the lip-like smooth
and β phases during the HT1 process (Fig. 3 (b, e)). The thickness of areas without obvious dimples were observed in the HT1 sample (as
transformed α lath was 1.53 ± 0.25 μm and the aspect ratio was 7.1. shown by the blue arrows in Fig. 6 (c)). This could be attributed to the
Moreover, continuous GB-α could be identified on the columnar prior-β presence of continuous GB-α, which was also found in [12,38,39]. In the
grain boundary, with the thickness of 1.66 ± 0.11 μm. In the meanwhile, case of the HT1 sample, the columnar prior-β grains boundaries were
the crystallographic orientations within the same GB-α were almost decorated with continuous GB-α which was commonly considered a
consistent (Fig. 3 (e)). After the HT2 process, α laths were present within preferred pathway for fatigue crack propagation [24,40]. In compari­
prior-β grains, with a thickness of 2.01 ± 0.64 μm and an aspect ratio of son, the fracture surface of the HT2 sample shows dimples indicative of
5.3, Fig. 3 (d, f). In the meantime, the presence of discontinuous GB-α ductile failure, Fig. 6 (d-f). The size of the dimples was measured as 3.72
was identified, with an average thickness of 2.15 ± 0.13 μm. Such ± 0.93 μm. Furthermore, detailed fractography characterization showed
discontinuous GB-α has various crystallographic orientations, which are the fatigue striations in both HT1 and HT2 samples, Fig. 6 (b, e). The
indicated by arrows in Fig. 3 (f). The formation of discontinuous GB-α width of fatigue striations was measured as 0.89 ± 0.06 µm in the HT1
was attributed to the factor that HT2 treating temperature was close to sample and 0.57 ± 0.04 µm in the HT2 sample. The finer fatigue stria­
β-transus temperature and led to a high β-phase volume fraction, tion spacing in the HT2 sample indicated a slower crack propagation
resulting in more GB-α nucleation sites [37]. rate.

3.2. Fatigue performance 3.3. Fatigue crack growth behavior

High-cycle fatigue (HCF) performance of heat treated samples is Longitudinal cross-sectional characterization was conducted to
exhibited in Fig. 4, which shows that the HT2 samples have superior investigate the fatigue crack growth behaviors of LPBF Ti-64 samples
fatigue properties than the HT1 samples at every investigated stress. The after heat treatment. The cross-sectional profiles of the HT1 failed
failed fatigue cycles of HT1 samples were 104 cycles at 600 MPa, while samples show smooth fatigue crack propagation geometries (Fig. 7 (a)),
those of HT2 samples at the same stress level reached 105 cycles. indicating lower fatigue crack propagation resistance. In comparison,
Furthermore, the HT2 sample could run out to 107 cycles under 500 MPa the cross-sectional profiles of the HT2 failed samples show more
maximum stress, while the HT1 samples just reached 105 cycles in the tortuous geometries (indicated by white arrows in Fig. 7 (b)), which
stress level studied, which further showed that the fatigue performance strongly suggested the higher crack propagation resistance.
of the HT1 samples was obviously worse than the HT2 samples. Microstructural features, particularly the prior-β grains, and the GB-α
Examination of the fracture surfaces of the failed samples shows that were further analyzed along the cross-sectional profiles, as shown in
the fatigue failure process of the HT1 sample and the HT2 sample both Fig. 8 and Fig. 9. For the HT1 sample, the cross-sectional profiles ex­
consist of three distinctive stages: crack initiation (from inevitable amination found that some crack propagation was along continuous GB-
defect introduced by the LPBF process), crack propagation, and final α, which was highlighted by the black circle in Fig. 8 (a). Although the
fracture regime (Fig. 5). The crack initiation sites were all along the EBSD indexing near the cracks was very poor due to the severe defor­
sample sub-surfaces (as indicated by the red circled areas in Fig. 5). mation within the plastic zone, the corresponding BSE image (Fig. 8 (b))
Furthermore, the fracture surfaces of the failed HT1 samples consisted of confirmed that the microstructure with the same crystallographic
large flat facets without obvious dimples, which were shown in the orientation was the GB-α. In the meanwhile, the detailed BSE exami­
yellow arrows in Fig. 6 (a). These regions with large facets indicated nation showed crack propagation along the α lath interface (highlighted
rapid crack propagation. The sizes of such flat facets exceeded 100 μm, by the yellow arrows in Fig. 9 (a-c). These smooth crack profiles also
which was consistent with the dimensions of columnar prior-β grains. show low fracture resistance. In comparison, a high degree of crack

6
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 9. High-magnification backscattered electron micrographs of longitudinal cross-sectional characterization of HCF testing failed samples: (a-b) HT1 samples
failed at 600 MPa with 82,800 cycles, and (c) failed at 600 MPa with 55,000 cycles. (d-e) HT2 samples failed at 600 MPa with 440,400 cycles, and (f) failed at 600
MPa with 118,100 cycles. The regions in (a) and (d) were further characterized by EBSD, as shown in Fig. 12.

deflection could be clearly identified on the crack cross-sectional profile 4. Discussion

of the HT2 sample, particularly on the α lath interface, which indicates
that the presence of coarse α lath in the HT2 sample could resist the 4.1. Effect of α laths on the fatigue behavior
crack propagation (Fig. 9 (d-f)). Furthermore, in the case of the HT2
sample, fatigue crack propagation along the discontinuous GB-α was not As shown by the longitudinal cross-sectional characterization, a
observed (Fig. 10 (a)). The full optical microscope microstructure smooth fracture surface was found in the HT1 sample, which indicated
characterization of the cross-sectional crack profile of the HT2 sample the relatively faster crack propagation in the HT1 sample. The
further confirmed this viewpoint, Supplementary Fig. 2. This result in­ morphology of α laths is one of the key factors that determine the fatigue
dicates that discontinuous GB-α with different crystallographic orien­ crack initiation and growth behavior. On the one hand, the width of α
tations was not the preferential crack path (Fig. 10 (b)). lath is directly related to the effective slip lengths which is the primary
crack propagation determining factor in titanium alloys [31,41]. It has

7
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 10. (a) The BSE image along the fatigue crack profile of the HT2 sample failed at 600 MPa with 440,400 cycles. (b) The corresponding EBSD IPF map of the same
area. Discontinuous GB-α was highlighted by blue arrows in (a) and black cycle in (b). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)

Fig. 11. EBSD IPF of α phase in (a) HT1 sample and (b) HT2 sample with contour lines of high Schmid factor value (>0.47) for basal and prismatic slip system (pr:
prismatic 〈a〉 slip, ba: basal 〈a〉 slip).

been demonstrated that the decreased α lath thickness could result in behavior, especially in titanium alloys due to the limited deformation
reduced fatigue performance due to the smaller effective slip length modes of α phase [47]. In this work, IPF of α phase with contour lines of
[42,43]. Since α laths within the same parent prior-β grain have similar high Schmid factor value (>0.47) was introduced to illustrate the in­
crystallographic orientations, slip can be easily transferred across the fluence of α texture on the deformation modes [48–50]. It should be
neighboring α laths. As a result, α lath with the smaller thickness could noted that only basal 〈a〉 and prismatic 〈a〉 slip systems were under
easily induce the occurrence of intensive slip bands, which directly lead consideration in this work due to the low maximum stress applied in the
to low fatigue crack propagation resistance. fatigue testing (<80% of yield stress)1. Results show that the α phase in
Moreover, the aspect ratio of α lath was correlated with the effect of α the HT1 sample is favorable both for basal 〈a〉 slip systems and prismatic
lath interface on the fatigue crack initiation threshold, and thus can 〈a〉 slip systems, while the α phase in the HT2 sample is solely favorable
affect the fatigue performance of AM titanium alloys [23,44]. According for basal 〈a〉 slip systems, Fig. 11. This is attributed to the fact that 〈2 11
to crystal plasticity finite element model analysis in the literature [45], it 0〉 crystallographic orientation of α phase that favored prismatic 〈a〉 slip
was found that with the aspect ratio of grain decreases to 1 (equiaxial systems is caused by the transformation of strong 〈0 0 1〉β texture into 12
morphology), the effect of grain boundaries on fracture resistance di­ α-variants [50,51]. However, most α phases with 〈2 11 0〉 orientation is
minishes, i.e. the number of dislocations accumulated on the grain difficult to retain during the heat treatment process since variant se­
boundaries during deformation decreases. In contrast, α lath with small lection happened. α phases in the as-built state undergone the trans­
thickness and large aspect ratio could lead to more strain accumulation formation from α to high-temperature metastable β and back to α during
around the interface between the adjacent α laths. This high strain the heat treatment process, which could lead to 72-fold multiplication of
accumulation could lead to fatigue crack initiation from the interface, α orientation since a single α grain could theoretically produce 6 β ori­
which also results in poor high-cycle fatigue properties. entations and a single β grain can transform into 12 α orientations (as­
In this study, although the HT1 sample and HT2 sample have similar sume that all variants could be selected with equal probability) [27].
α lath thickness, there are differences in their aspect ratios. α lath in the Compared to the HT1 samples, the HT2 samples with more high-
HT2 sample has a smaller aspect ratio (the average aspect ratio of α lath
was 7.1 in the HT1 sample and 5.3 in the HT2 sample). The difference in
the aspect ratio is attributed to the fact that higher heat treatment 1
The different deformation modes of the α phase of titanium alloy could be
temperatures in HT2 (850 ◦ C for HT1 and 950 ◦ C for HT2) could provide
assessed with respect to the Schmid factor (SF) analysis, which was basal,
sufficient thermal energy and promote element diffusion to split α lath, prismatic, and pyramid slip modes [36]. In the present study, it was noted that
and thus result in a low aspect ratio in HT2 sample [23,46]. α lath with a pyramid slip mode was difficult to active since the maximum stresses applied in
low aspect ratio could result in less strain accumulation and stress the fatigue test were lower than the yield strength of the alloy. Both basal and
concentration on the α lath interface during the cyclic loading, which prismatic slip system has significantly lower critical resolved shear stresses
could reduce the tendency for crack initiation and slightly enhance the (CRSS) than other slip systems like pyramid slip systems. Thus, only basal and
crack propagation resistance as well. prismatic slip systems were considered in this study.
Apart from the morphology effect, the crystallographic orientations
of α lath could be another factor that affects the fatigue crack growth

8
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 12. The in-depth characterization of the fatigue crack propagation in both HT1 and HT2 samples, with (a-c) HT1 sample failed at 600 MPa with 82,800 cycles
and (d-f) HT2 sample failed at 600 MPa with 440,400 cycles. (a) EBSD IPF map of α phase in HT1 sample. The Schmid factor distribution maps for (c) basal 〈a〉 slip
systems and (d) prismatic 〈a〉 slip systems in the HT1 sample. (d) EBSD IPF map of α phase in HT2 sample. The Schmid factor distribution maps for (e) basal 〈a〉 slip
systems and (f) prismatic 〈a〉 slip systems in the HT2 sample. The crack profiles were extracted from BSE images, and outlined with red lines. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

9
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 13. Slip trace analysis for basal 〈a〉 and prismatic 〈a〉 slip systems in (a) HT1 sample and (b) HT2 sample.

Fig. 14. (a) EBSD IPF map of HT1 sample. The misorientation profiles along the scanning lines from (b) point A to B and (c) point C to D. The misorientation profile
was set to ‘point-to-point’.

temperature β phases (with more α phases undergoing the α → β → α This shows that α laths in the HT1 sample have ‘soft’ crystallographic
process) could therefore lead to fewer α phases with 〈2 11 0〉 orientation, orientations. Due to the similarly aligned close-packed planes, easy slip
and was not favorable for prismatic 〈a〉 slip systems2. transmission can occur between soft grains, which are prone to fatigue
Further Schmid factor distributions analysis was carried out to reveal damage and lead to minimal fatigue crack growth resistance. Thus, the
the slip activity of grains. In the examined area the of HT1 sample, it is fatigue crack was easy to grow along these grains, resulting in a smooth-
found that most grains were favorable both for basal 〈a〉 slip systems and appearing fracture pathway and poor high-cycle fatigue properties. In
prismatic 〈a〉 slip systems (Fig. 12 (b-c)). The SF values of both slip comparison, most α laths in the examined area of HT2 failed sample are
systems were larger than 0.4, which is consistent with the IPFs in Fig. 11. solely favorable for basal 〈a〉 slip systems with SF values larger than 0.4,
while relatively unfavorable for prismatic 〈a〉 slip systems with SF values
less than 0.3 (Fig. 12 (e-f)). The fewer favorable slip systems can lead to
less slip transmission across grains, and thereby smaller effective slip
2
The higher heat treatment temperature of HT2 could cause more ‘initial’ α length and higher fatigue crack propagation resistance in the HT2
phases introduced by LPBF to be transformed into metastable β phases during sample. The correlation between the slip traces and retracted crack
the HT process (the equilibrium β-fraction was 27 % at 850 ℃ used in HT1 and
profiles shows that the crack paths in most grains of the HT1 sample are
77 % at 950 ℃ used in HT2 [37]). The high-temperature metastable β phases
on either basal or prismatic planes (as highlighted by black arrows in
were re-transformed to α phases via Burgers orientation relationships
Fig. 13 (a)). On the contrary, slip trace analysis shows that only a few
((0001)α//{1 1 0}β and <11 2 0>α//<1 1 1>β) during subsequent water
quenching (WQ) of the heat treatment process. Furthermore, as the cooling rate
grains in the HT2 sample have their basal planes as the favored fatigue
of WQ was much less than that of the LPBF process, the WQ process could crack propagation pathways, which is highlighted by black arrows in
produce different variant selection tendencies from the LPBF process [55]. Fig. 13 (b). Due to the minimal resistance, close-packed planes

10
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

Fig. 15. (a) EBSD Image quality (IQ) of HT2 sample with the GB-α orientations displayed. (b-g) (0001) pole figures showing the orientations of selected discon­
tinuous GB-α and the corresponding grain rotation angles between the adjacent GB-α.

perpendicular to the loading axis are favorable for crack advancing and is undisputed that the continuous GB-α could cause fatigue crack initi­
transmission between grains. Most grains in the HT2 sample only have ation. This is since that the variations in the crystallographic orientation
their basal planes as the favored fatigue crack propagation pathway, and geometry of neighboring columnar prior β-grains could cause plastic
which increases the difficulties of slip transfer between adjacent α lath deformation incompatibility, which result in the stress concentration at
(due to the limited favorable slip plane) and results in abrupt crack path the GB-α and made it vulnerable sites for materials failure [17,53]. In the
changes and more tortuous crack profiles in the HT2 sample as found in meanwhile, due to the GB-α layer perpendicular to the loading axis, the
this study. entire length of the continuous GB-α is exposed to opening tension
during the cyclic loading. In our previous study, it has confirmed that the
4.2. Effect of GB-α morphology on the fatigue behavior deformation and dislocations were mostly accommodated within the
continuous GB-α during the opening tension process, which can result in
Grain boundary α-phase (GB-α) is the transformed α phase that exists strain localization within continuous GB-α and render it the precursor to
at the prior-β grain boundary, whose morphology is normally divided crack initiation [54]. Moreover, in the HT1 failed sample, continuous
into continuous and discontinuous [52]. In this work, the GB-α mor­ GB-α could be identified on the fatigue crack profile during the crack
phologies are continuous in the HT1 sample and discontinuous in the propagation stage (as shown in Fig. 8), which shows that continuous GB-
HT2 sample, which could lead to the different fatigue behaviour of the α could serve as a pathway for crack growth. Microstructure charac­
HT1 and HT2 samples. This morphology difference is attributed to the terization shows continuous GB-α has a similar crystallographic orien­
fact that the HT2 temperature was close to β-transus, leading to a high β tation, Fig. 3 (e). The misorientation profiles within continuous GB-α
volume fraction (reported as approximately 87% in [37]) and more GB-α show the maximum misorientation angle measured as ~1.0◦ , Fig. 14.
nucleation sites. Thus, continuous GB-α with similar crystallographic orientation could
Despite of the observation that intrinsic defects were the primary be considered as a preferential crack pathway during the fatigue loading
factor leading to the fatigue crack initiation in the HT1 sample (Fig. 5), it process, and has minimal fracture resistance, as shown by the flat

11
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

fracture surface in Fig. 8. In the meanwhile, due to the tensile loading Acknowledgment
axis perpendicular with GB-α direction, the full crack opening effect
could accelerate crack propagation along the GB-α layer between two This research work was sponsored by Sailing Program
neighboring prior-β grains, which further results in a reduction in fa­ (20YF1431600), Natural Science Foundation of Inner Mongolia (Grant
tigue crack growth resistance in HT1 sample. No. 2020MS05061), and internal funding from Univeristy of Shanghai
Since columnar grains still existed in the HT2 sample, which also for Science and Technology. The authors acknowledge the use of the
could lead to strain incompatibility between the adjacent columnar instruments and scientific and technical assistance at the Monash Centre
grains, the prior-β grain boundary is potentially the vulnerable site for for Electron Microscopy, a Node of Microscopy Australian. The authors
crack initiation and propagation during the fatigue process. However, in also appreciate the access to the facilities provided by Advanced Mate­
the HT2 sample, GB-α is not found along the crack profile, Fig. 10. This is rials Research Institute, Yangtze Delta Analytical Characterization
because the presence of discontinuous GB-α could uniformly accom­ Platform.
modate the deformation together with the surrounding matrix during
the mechanical loading process, which avoids the crack initiation at Appendix A. Supplementary data
discontinuous GB-α [22,23,54]. On the other hand, different GB-α at the
same prior β-grain boundary has different crystallographic orientations, Supplementary data to this article can be found online at https://doi.
Fig. 3 (f) and Fig. 10. The calculated misorientations between the org/10.1016/j.ijfatigue.2023.107839.
adjacent GB-α were measured to be 52◦ , 35◦ , 3◦ , 2◦ , 88◦ , and 77◦ , which
show the relatively large crystallographic orientation difference be­ References
tween the neighboring discontinuous GB-α in HT2 sample and thereby
can inhibit the fatigue crack growth along the weak prior-β grain [1] Noronha J, Rogers J, Leary M, Kyriakou E, Inverarity SB, Das R, et al. Ti-6Al-4V
hollow-strut lattice materials by laser powder bed fusion. Addit Manuf 2023;72:
boundaries (Fig. 15). Thus, the prior-β grain boundary with discontin­ 103637.
uous GB-α is considered not detrimental to the fatigue performance in [2] Zhang T, Huang Z, Yang T, Kong H, Luan J, Wang A, et al. In situ design of
HT2 sample. advanced titanium alloy with concentration modulations by additive
manufacturing. Science (80-) 2021;374(6566):478–82.
[3] Zhang D, Qiu D, Gibson MA, Zheng Y, Fraser HL, StJohn DH, et al. Additive
5. Conclusions manufacturing of ultrafine-grained high-strength titanium alloys. Nature 2019;576
(7785):91–5.
[4] Rogers J, Qian Ma, Elambasseril J, Burvill C, Brice C, Wallbrink C, et al. Fatigue
In summary, this study presents a thorough investigation into the test data applicability for additive manufacture: A method for quantifying the
influence of the microstructure on the fatigue performance and fatigue uncertainty of AM fatigue data. Mater Des 2023;231:111978.
crack growth behavior of laser bed powder fusion Ti-64. The improve­ [5] Gu D, Shi X, Poprawe R, Bourell DL, Setchi R, Zhu J. Material-structure-
performance integrated laser-metal additive manufacturing. Science 2021;372
ment in the high-cycle fatigue life of LPBF Ti-64 could be achieved by
(6545).
discontinuous GB-α and low aspect ratio α lath which was introduced by [6] Hu YN, Wu SC, Withers PJ, Zhang J, Bao HYX, Fu YN, et al. The effect of
post-fabrication heat treatment. Discontinuous GB-α has different crys­ manufacturing defects on the fatigue life of selective laser melted Ti-6Al-4V
tallographic orientations, and could accommodate the deformation and structures. Mater Des 2020;192. https://doi.org/10.1016/j.matdes.2020.108708.
[7] Tekdir H, Yetim AF. Additive manufacturing of multiple layered materials
improve the fatigue crack propagation resistance. α lath with a low (Ti6Al4V/316L) and improving their tribological properties with glow discharge
aspect ratio could lead to less strain accumulation on the α lath interface surface modification. Vacuum 2021;184:109893. https://doi.org/10.1016/j.
and thus reduce the tendency for crack initiation and propagation at vacuum.2020.109893.
[8] du Plessis A, Yadroitsava I, Yadroitsev I. Effects of defects on mechanical properties
such interface. In comparison, continuous GB-α, which has a similar in metal additive manufacturing: A review focusing on X-ray tomography insights.
crystallographic orientation within, could act as a preferential crack Mater Des 2020;187:108385. https://doi.org/10.1016/j.matdes.2019.108385.
path and thus lead to poor fatigue performance of LPBF Ti-64. In the [9] Yang J, Yu H, Yin J, Gao M, Wang Z, Zeng X. Formation and control of martensite
in Ti-6Al-4V alloy produced by selective laser melting. Mater Des 2016;108:
meantime, α lath formed alongside continuous GB-α has soft crystallo­ 308–18. https://doi.org/10.1016/j.matdes.2016.06.117.
graphic orientations and higher aspect ratio, which was prone to fatigue [10] Chen X, Zhang J, Chen X, Cheng X, Huang Z. Electron beam welding of laser
damage and could lead to low fracture resistance as well. additive manufacturing Ti–6.5Al–3.5Mo–1.5Zr–0.3Si titanium alloy thick plate.
Vacuum 2018;151:116–21. https://doi.org/10.1016/j.vacuum.2018.02.011.
[11] Simonelli M, Tse YY, Tuck C. On the texture formation of selective laser melted Ti-
CRediT authorship contribution statement 6Al-4V. Metall Mater Trans A Phys Metall Mater Sci 2014;45:2863–72. https://doi.
org/10.1007/s11661-014-2218-0.
[12] Liu CM, Wang HM, Tian XJ, Tang HB, Liu D. Microstructure and tensile properties
Jianwen Liu: Conceptualization, Formal analysis, Data curation,
of laser melting deposited Ti-5Al-5Mo-5V-1Cr-1Fe near β titanium alloy. Mater Sci
Writing – original draft. Kai Zhang: Conceptualization, Visualization, Eng A 2013;586:323–9. https://doi.org/10.1016/j.msea.2013.08.032.
Funding acquisition, Supervision, Writing – review & editing. Jie Liu: [13] Vrancken B, Thijs L, Kruth JP, Van Humbeeck J. Heat treatment of Ti6Al4V
Data curation, Investigation, Writing – original draft. Hao Wang: produced by Selective Laser Melting: Microstructure and mechanical properties.
J Alloys Compd 2012;541:177–85. https://doi.org/10.1016/j.
Investigation, Resources. Yi Yang: Data curation, Formal analysis. jallcom.2012.07.022.
Liangming Yan: Funding acquisition, Investigation. Xinni Tian: [14] Cao S, Chu R, Zhou X, Yang K, Jia Q, Lim CVS, et al. Role of martensite
Methodology, Resources. Yuman Zhu: Resources, Conceptualization, decomposition in tensile properties of selective laser melted Ti-6Al-4V. J Alloys
Compd 2018;744:357–63.
Writing – review & editing. Aijun Huang: Supervision, Resources. [15] Liu J, Zhang K, Yang Y, Wang H, Zhu Y, Huang A. Grain boundary α -phase
precipitation and coarsening : Comparing laser powder bed fusion with as-cast Ti-
Declaration of Competing Interest 6Al-4V. Scr Mater 2022;207:114261. https://doi.org/10.1016/j.
scriptamat.2021.114261.
[16] Chang K, Wang X, Liang E, Zhang R. On the texture and mechanical property
The authors declare that they have no known competing financial anisotropy of Ti6Al4V alloy fabricated by powder-bed based laser additive
interests or personal relationships that could have appeared to influence manufacturing. Vacuum 2020;181:109732. https://doi.org/10.1016/j.
vacuum.2020.109732.
the work reported in this paper. [17] Carroll BE, Palmer TA, Beese AM. Anisotropic tensile behavior of Ti-6Al-4V
components fabricated with directed energy deposition additive manufacturing.
Data availability Acta Mater 2015;87:309–20. https://doi.org/10.1016/j.actamat.2014.12.054.
[18] Simonelli M, Tse YY, Tuck C. Effect of the build orientation on the mechanical
properties and fracture modes of SLM Ti-6Al-4V. Mater Sci Eng A 2014;616:1–11.
Data will be made available on request. https://doi.org/10.1016/j.msea.2014.07.086.
[19] Li F, Qi B, Zhang Y, Guo W, Peng P, Zhang H, et al. Effects of heat treatments on
microstructures and mechanical properties of ti6al4v alloy produced by laser solid
forming. Metals (Basel) 2021;11(2):346.

12
J. Liu et al. International Journal of Fatigue 176 (2023) 107839

[20] Xie Y, Gong M, Zhang R, Gao M, Zeng X, Wang F. Grain boundary discontinuity and [38] Sun W, Ma Y, Huang W, Zhang W, Qian X. Effects of build direction on tensile and
performance improvement mechanism of wire arc additive manufactured fatigue performance of selective laser melting Ti6Al4V titanium alloy. Int J Fatigue
Ti–6Al–4V. J Alloys Compd 2021;869:159287. https://doi.org/10.1016/j. 2020:130. https://doi.org/10.1016/j.ijfatigue.2019.105260.
jallcom.2021.159287. [39] Xu ZW, Liu A, Wang XS. The influence of building direction on the fatigue crack
[21] Zhao R, Chen C, Shuai S, Hu T, Fautrelle Y, Liao H, et al. Enhanced mechanical propagation behavior of Ti6Al4V alloy produced by selective laser melting. Mater
properties of Ti6Al4V alloy fabricated by laser additive manufacturing under static Sci Eng A 2019:767. https://doi.org/10.1016/j.msea.2019.138409.
magnetic field. Mater Res Lett 2022;10(8):530–8. [40] Kumar KN, Muneshwar P, Singh SK, Jha AK, Pant B, George KM. Effect of Grain
[22] Liu CM, Wang HM, Tian XJ, Liu D. Development of a pre-heat treatment for Boundary Alpha on Mechanical Properties of Ti5.4Al3Mo1V Alloy. JOM 2015;67:
obtaining discontinuous grain boundary α in laser melting deposited Ti-5Al-5Mo- 1265–72. https://doi.org/10.1007/s11837-015-1443-3.
5V-1Cr-1Fe alloy. Mater Sci Eng A 2014;604:176–82. https://doi.org/10.1016/j. [41] Lavogiez C, Dureau C, Nadot Y, Villechaise P, Hémery S. Crack initiation
msea.2014.03.028. mechanisms in Ti-6Al-4V subjected to cold dwell-fatigue, low-cycle fatigue and
[23] Wang J, Lin X, Wang M, Li J, Wang C, Huang W. Effects of subtransus heat high-cycle fatigue loadings. Acta Mater 2023:244. https://doi.org/10.1016/j.
treatments on microstructure features and mechanical properties of wire and arc actamat.2022.118560.
additive manufactured Ti–6Al–4V alloy. Mater Sci Eng A 2020;776:139020. [42] Qiu J, Feng X, Ma Y, Lei J, Liu Y, Huang A, et al. Fatigue crack growth behavior of
https://doi.org/10.1016/j.msea.2020.139020. beta-annealed Ti-6Al-2Sn-4Zr-xMo (x = 2, 4 and 6) alloys: Influence of
[24] Foltz JW, Welk B, Collins PC, Fraser HL, Williams JC. Formation of grain boundary microstructure and stress ratio. Int J Fatigue 2016;83:150–60.
α in β Ti alloys: Its role in deformation and fracture behavior of these alloys. Metall [43] Liu YX, Chen W, Li ZQ, Tang B, Han XQ, Yao G. The HCF behavior and life
Mater Trans A Phys Metall Mater Sci 2011;42:645–50. https://doi.org/10.1007/ variability of a Ti-6Al-4V alloy with transverse texture. Int J Fatigue 2017;97:
s11661-010-0322-3. 79–87. https://doi.org/10.1016/j.ijfatigue.2016.12.030.
[25] Sauer C, Lütjering G. Influence of α layers at β grain boundaries on mechanical [44] Wilson-Heid AE, Wang Z, McCornac B, Beese AM. Quantitative relationship
properties of Ti-alloys. Mater Sci Eng A 2001;319–321:393–7. https://doi.org/ between anisotropic strain to failure and grain morphology in additively
10.1016/S0921-5093(01)01018-8. manufactured Ti-6Al-4V. Mater Sci Eng A 2017;706:287–94. https://doi.org/
[26] Xia M, Liu A, Lin Y, Li N, Ding H, Zhong C. Densification behavior, microstructure 10.1016/j.msea.2017.09.017.
evolution and fretting wear performance of in-situ hybrid strengthened Ti-based [45] Jiang M, Devincre B, Monnet G. Effects of the grain size and shape on the flow
composite by laser powder-bed fusion. Vacuum 2019;160:146–53. https://doi.org/ stress: A dislocation dynamics study. Int J Plast 2019;113:111–24. https://doi.org/
10.1016/j.vacuum.2018.11.023. 10.1016/j.ijplas.2018.09.008.
[27] Obasi GC, Birosca S, Quinta Da Fonseca J, Preuss M. Effect of β grain growth on [46] Sabban R, Bahl S, Chatterjee K, Suwas S. Globularization using heat treatment in
variant selection and texture memory effect during α → β → α phase transformation additively manufactured Ti-6Al-4V for high strength and toughness. Acta Mater
in Ti-6 Al-4V. Acta Mater 2012;60:1048–58. https://doi.org/10.1016/j. 2019;162:239–54. https://doi.org/10.1016/j.actamat.2018.09.064.
actamat.2011.10.038. [47] Hémery S, Villechaise P. In situ EBSD investigation of deformation processes and
[28] Wu SQ, Lu YJ, Gan YL, Huang TT, Zhao CQ, Lin JJ, et al. Microstructural evolution strain partitioning in bi-modal Ti-6Al-4V using lattice rotations. Acta Mater 2019;
and microhardness of a selective-laser-melted Ti-6Al-4V alloy after post heat 171:261–74. https://doi.org/10.1016/j.actamat.2019.04.033.
treatments. J Alloys Compd 2016;672:643–52. [48] Bridier F, Villechaise P, Mendez J. Analysis of the different slip systems activated
[29] Cao S, Hu Q, Huang A, Chen Z, Sun M, Zhang J, et al. Static coarsening behaviour by tension in a α/β titanium alloy in relation with local crystallographic
of lamellar microstructure in selective laser melted Ti–6Al–4V. J Mater Sci Technol orientation. Acta Mater 2005;53:555–67. https://doi.org/10.1016/j.
2019;35(8):1578–86. actamat.2004.09.040.
[30] Xie Y, Gong M, Luo Z, Li Q, Gao M, Wang F, et al. Effect of microstructure on short [49] Ter Haar GM, Becker TH. The influence of microstructural texture and prior beta
fatigue crack growth of wire arc additive manufactured Ti-6Al-4V. Mater Charact grain recrystallisation on the deformation behaviour of laser powder bed fusion
2021;177:111183. produced Ti–6Al–4V. Mater Sci Eng A 2021;814:141185. https://doi.org/10.1016/
[31] Zhang K, Liu Y, Tian X, Yang Yi, Zhu Y, Bermingham M, et al. Fatigue behaviour of j.msea.2021.141185.
L-DED processed Ti-6Al-4V with microstructures refined by trace boron addition. [50] Li R, Wang H, He B, Li Z, Zhu Y, Zheng D, et al. Effect of α texture on the anisotropy
Int J Fatigue 2023;168:107454. of yield strength in Ti–6Al–2Zr–1Mo–1V alloy fabricated by laser directed energy
[32] Jha JS, Toppo SP, Singh R, Tewari A, Mishra SK. Deformation behavior of Ti-6Al- deposition technique. Mater Sci Eng A 2021;824:141771.
4V microstructures under uniaxial loading: Equiaxed Vs. transformed-β [51] Cepeda-Jiménez CM, Potenza F, Magalini E, Luchin V, Molinari A, Pérez-Prado MT.
microstructures. Mater Charact 2021:171. https://doi.org/10.1016/j. Effect of energy density on the microstructure and texture evolution of Ti-6Al-4V
matchar.2020.110780. manufactured by laser powder bed fusion. Mater Charact 2020;163:110238.
[33] Kahlin M, Ansell H, Moverare J. Fatigue crack growth for through and part-through https://doi.org/10.1016/j.matchar.2020.110238.
cracks in additively manufactured Ti6Al4V. Int J Fatigue 2022;155:106608. [52] Lütjering G, Williams JC. Titanium. 2nd ed. Vol. 53. Springer Berlin Heidelberg;
https://doi.org/10.1016/j.ijfatigue.2021.106608. 2007. doi: 10.1017/CBO9781107415324.004.
[34] Tarik Hasib M, Ostergaard HE, Li X, Kruzic JJ. Fatigue crack growth behavior of [53] Kumar P, Ramamurty U. Microstructural optimization through heat treatment for
laser powder bed fusion additive manufactured Ti-6Al-4V: Roles of post heat enhancing the fracture toughness and fatigue crack growth resistance of selective
treatment and build orientation. Int J Fatigue 2021;142:105955. https://doi.org/ laser melted Ti–6Al–4V alloy. Acta Mater 2019;169:45–59. https://doi.org/
10.1016/j.ijfatigue.2020.105955. 10.1016/j.actamat.2019.03.003.
[35] ASTM. E466 Standard Practice for Conducting Force Controlled Constant [54] Liu J, Zhang K, Gao X, Wang H, Wu S, Yang Yi, et al. Effects of the morphology of
Amplitude Axial Fatigue Tests of Metallic Materials. Test 2002;03:4–8. grain boundary α-phase on the anisotropic deformation behaviors of additive
[36] Zhu Y, Zhang K, Meng Z, Zhang K, Hodgson P, Birbilis N, et al. Ultrastrong manufactured Ti–6Al–4V. Mater Des 2022;223:111150.
nanotwinned titanium alloys through additive manufacturing. Nat Mater 2022;21 [55] Zhao Z, Chen J, Lu X, Tan H, Lin X, Huang W. Formation mechanism of the α
(11):1258–62. variant and its influence on the tensile properties of laser solid formed Ti-6Al-4V
[37] Pederson R, Babushkin O, Skystedt F, Warren R. Use of high temperature X-ray titanium alloy. Mater Sci Eng A 2017;691:16–24. https://doi.org/10.1016/j.
diffractometry to study phase transitions and thermal expansion properties in Ti- msea.2017.03.035.
6Al-4V. Mater Sci Technol 2003;19:1533–8. https://doi.org/10.1179/
026708303225008013.

13