Study on high cycle fatigue performance at elevated temperature for a selective laser melted Ti6Al4V alloy

Selective laser melted Ti6Al4V alloy has broad application prospects in aeroengine field. In this study, high cycle fatigue tests were carried out systematically using hourglass shaped specimen (K t=1) at 400 °C. The S-N curves were acquired and compared with data of the casting and the forging. The anisotropy and the defect effects of the fatigue performance were analysed. The LOF defects on the surface and subsurface preferentially act as the fatigue crack initiation site for almost all the tested HCF specimens. The position and irregularity of defect have greater influence on elevated temperature fatigue life than the defect size. The fatigue strength of horizontal orientation is lower than that of vertical orientation, because of the characteristics of larger size, more irregular shape and higher density for the crack source defects of horizontal specimens. The fatigue performance of the SLM Ti6Al4V alloy in this study is better than that of the casting and even the forging, but the dispersion of fatigue data of the SLM alloy is much greater than that of the both traditional process alloys.


Introduction
Ti6Al4V alloy is an α+β type two-phase titanium alloy which has a medium strength, and long-term service below 400℃.Because of the excellent comprehensive mechanical properties, good process plasticity, corrosion resistance and low specific weight [1,2] , it is widely used in aeroengine fan disc and blade, compressor disc and blade and other parts.In recent years, with the rise of additive manufacturing (AM) technology, the AM process especially of the selective laser melting (SLM) process for the Ti6Al4V alloy has received the focus of research in the word, and there are many reports on the related process research and mechanical behavior research [3~9] , which makes the AM process of the Ti6Al4V alloy greatly developed, and obtained some applications in aerospace and other fields.Currently, the main reports for the research on AM Ti6Al4V alloy include process optimization, posttreatment process, relationship between microstructures and properties, fatigue performance and life prediction, etc. [10~16] .For aeroengine components, especially load-bearing components, research on fatigue behavior of AM materials is one of the most important directions.It is not only necessary to ascertain the fatigue characteristics of different AM process materials and accumulate performance data, but also to reveal the deformation and damage mechanism of AM materials, and establish property characterization and life prediction methods, thus provide technical support for AM process optimization and component life assessment.There are many reports on the room temperature fatigue behavior for AM Ti6Al4V alloy in recent years.For example, Fu et al. [17] studied the high cycle fatigue (HCF) behavior of SLM Ti6Al4V alloy under different stress ratios and discussed the influence of stress ratio in detail.Gao et al. [18] studied the HCF behavior of the annealed laser powder bed fusion (LPBF) Ti6Al4V alloy, and revealed that the surface defects geometrical parameters and anisotropy severity governed the significant fatigue anisotropy and scatter.Bhandari et al. [19] studied different post-processing methods to improve fatigue properties of LPBF Ti6Al4V alloy.Qu et al. [20] studied the fatigue properties and microstructure of LPBF Ti6Al4V alloy, and developed a fatigue life prediction model based on the Murakami theory.Tang et al. [21] investigated the effects of build orientation on the anisotropic fatigue performance of directed energy deposition (DED) Ti6Al4V alloy, specifically the effect of microstructures and defects on fatigue performance anisotropy.Kumar et al. [22] studied the effect of process parameters on pore shape, size and distribution in SLM Ti6Al4V specimens, and investigated the influence of pore characteristics on the HCF life under rotating bending fatigue conditions.Jiao et al. [23] studied the HCF behavior of a SLM Ti6Al4V alloy, and systematically revealed the effect of shape, size and location of defects on the fatigue performance, and discussed the life prediction method based on fracture mechanics theory.Gillham et al. [24] presented a critical distance theory to assess the fatigue life of notched SLM Ti6Al4V alloy.Benedetti et al. [25] presented a novel method for building the Kitagawa-Takahashi diagram and checked by the SLM Ti6Al4V alloy.In summary, researches on fatigue behavior of AM Ti6Al4V alloy at room temperature have achieved a wealth of research results, but researches on the elevated temperature testing conditions have not been widely carried out, and few results about this aspect are reported.This study aims at the HCF performance of different orientations of the SLM Ti6Al4V alloy at 400℃.The anisotropy of fatigue properties was systematically studied, and the fractography was analyzed, and the influence of defects on fatigue anisotropy and life was revealed.

Materials and processing condition
The cylindrical rods (Φ13×70 mm) were manufactured using the SLM process, with both vertical and horizontal orientations.The arrangement of the rods on the build platform is shown in Fig. 1.A shell and core concept were employed for the scan strategy, rotating each layer at an angle of 67°.A laser with a power of 380W melted each final layer to a thickness of 60 μm, scanning at a rate of 1250mm/s with a spacing of 0.12mm.The entire manufacturing process took place in an argon atmosphere.To minimize residual stress in the products, post-built annealing heat treatment was conducted at 800℃ for 2 hours under vacuum conditions, followed by cooling in an argon atmosphere.The particle size distribution of the powder is shown in Fig. 2, ranging from 16.7μm to 59.7μm diameter particles.Table 1 provides the chemical composition of the alloy used.Metallographic images revealing the distinct textures between vertical (ZX and ZY) and horizontal (XY) planes is shown in Fig. 3a.Elongated columnar grains were observed on vertical planes while horizontal plane exhibited the equiaxial characteristics with an average grain size around 140μm.Fig. 3b and Fig. 3c demonstrate that columnar crystals consisted approximately 40% fine acicular α basket-weave structures and about 60% α+β basket-weave structures surrounded by β grain boundaries respectively, and some lack-of-fusion (LOF) defects were also present.

Experimental procedure
The round rod blanks, as shown in Fig. 1, were machined into hourglass-shaped specimens with a total length of 70mm and a minimum section diameter of 5mm in the working region.To eliminate any machining imperfections, the gauge surfaces of the specimens were manually polished to achieve a final roughness Ra of 0.4μm along the longitudinal axis.HCF tests were conducted at 400℃ using the QBG-50 high-frequency testing machine, following ASTM E466 standard guidelines until specimen failure or attainment of 1×10 7 cycles, which was considered as "run-out".The testing frequency was approximately 100 Hz.Three stress ratios (R=-1, 0.1 and 0.5) were employed for testing purposes.About 30 samples were tested under each stress ratio, and the results were plotted on S-N curves for analysis.Post-fatigue tests involved examination of fracture surface morphologies using FEI nano Field emission SEM and statistically analyze characteristics related to crack initiation defects.

S-N curves of the SLM Ti6Al4V alloy
The HCF test results at 400℃ and three stress ratios are depicted in Fig. 5.The numerical values following the black arrows indicate the count of specimens that experienced "run-out."The S-N curves were obtained using the three-parameter exponential function as described below.
The fatigue performance of the SLM Ti6Al4V alloy exhibits significant anisotropy for different stress ratios at 400℃, as depicted in Fig. 5.The fatigue strength was found to be lower in the horizontal orientation compared to the vertical orientation.Additionally, the data at lower stress levels show higher dispersion, particularly near the fatigue limit region where most specimens have life gaps of millions of cycles.Consequently, a majority of data clusters were observed around 1×10 5 cycles and the fatigue limit life 1×10 7 cycles.This implies that the alloy is more prone to early fracture occurring around 10 5 cycles even when subjected to the fatigue limit stress.

Comparison of S-N curves between SLM and traditional processed Ti6Al4V alloys
Fig. 6 shows the S-N curves comparison of Ti6Al4V alloy prepared by SLM, casting and forging processes under testing conditions of 400℃ and stress ratios of R=-1, 0.1 and 0.5.Table 3 lists the comparison results, where ">" indicates "better than" and "=" indicates "equal".It can be seen that when R=-1, the vertical orientation of the SLM Ti6Al4V alloy had the highest fatigue performance, and that of the horizontal orientation was second only for the 2# forging.When R=0.1, the fatigue performance of the SLM Ti6Al4V alloy was inferior to forgings in the higher stress region, but superior to the casting, and second only for the 2# forging in the lower stress region.When R=0.5, the fatigue performance of both horizontal and vertical orientations of the SLM Ti6Al4V alloy was superior to the forgings and castings.According to the above results, the HCF performance of the SLM Ti6Al4V alloy in this study is better than that of the castings and even the forgings.However, it also shows that the dispersion of fatigue data of the SLM alloy is much greater than that of the both traditional process alloys.

Fractography of the SLM Ti6Al4V alloy
A total of 74 fractured specimens were produced in this study.The fracture morphology of all fractured specimens was observed by SEM.It is showed that the crack sources of 71 out of 74 specimens were surface or subsurface LOF defects, and the other three specimens were surface slip.Among the 71 specimens of defect initiation, 61 specimens were initiated from single defect and 10 specimens were initiated from multiple defects.Among the specimens of multi-source initiation, one was from vertical orientation, and 9 were from horizontal orientation.It indicates that the horizontal specimen has a higher defect density than the vertical specimen, and also confirms that the fatigue performance of the horizontal specimen is lower than that of the vertical specimen.By analyzing the typical characteristics of the crack source region, the expansion region and the final rupture region of the fractured specimens, it is found that the fractography characteristics under different conditions are basically the same.Take the fractured specimen with information of horizontal, R=0.1, σmax=480MPa and Nf=6.5×10 4 cycles as an example (Fig. 7).The crack source region (region Ⅰ) was characterized by single source, and cracks initiated from an irregular LOF on the surface.Radiative edges and obvious fatigue striations with strip size of about 0.8μm can be seen in the expansion region (region Ⅱ).The final rupture region (region Ⅲ) was rough and uneven, with lots of dimples, and exhibits ductile fracture.Fig. 8 shows the multi-source initiation fractography of the specimen with information of horizontal, R=0.1, σmax =500MPa and Nf =5.1×10 4 cycles.It can be seen that the cracks were mainly initiated from three subsurface LOFs in different locations.Two of them were irregular and the last one was nearly elliptical.3,27,28] .In this study, the surface and subsurface defects were equivalent to semi-ellipses, and the size of the defect was represented by the √areaeff of the semi-ellipse, as shown in Fig. 9.The effect factors such as defect shape, size, relative position to the specimen surface of the fatigue life was studied.Fig. 10 shows the typical defect morphologies of three pairs of specimens with the same testing conditions.It can be seen that for the first group of the specimens (1# and 2#), the 1# specimen, which has the characteristics of smaller defect size, closer distance from the surface, and irregular shape, shows a shorter life, indicating that the defect position and shape have a greater impact on fatigue life than the defect size.For the second group of specimens (3# and 4#), both of them have nearly elliptical LOFs.The 3# specimen which has a smaller defect size and closer to the surface, exhibits a shorter fatigue life, indicating that while the defect shape is the same, the position has a greater influence on fatigue life than the defect size.For the third group of specimens (5# and 6#), the 5# specimen has two crack source regions, and both of them are irregular LOFs with larger size.The 6# specimen is single source initiation, and the defect size is smaller, so exhibits lower fatigue life.In conclusion, the relative position of the defect with the surface and the irregularity of the defect have the greater influence on elevated temperature fatigue life than the defect size.the subsurface defect [27,28]  of both directions, and the defect size √areaeff values were obtained.Fig. 11 shows the lognormal distribution of √areaeff values of defects for specimens of both directions.It can be seen that the defect sizes and the sizes distribution span of the horizontal specimens were larger than that of the vertical specimens.The size range of the vertical specimens was 20μm~70μm, and the median size was 42μm, and the number of defects in the range of 40μm~50μm was the largest.The size range of the horizontal specimens was 30μm~340μm, and the median size was 55μm, and the number of defects in the range of 50μm~60μm was the largest, what's more, there were a number of more than 100μm "super" defects.Through SEM observation of the defects in the crack source region, it was also found that the defects of the vertical specimens were mainly elliptical shapes, and those of the horizontal specimens were mainly irregular shapes.In summary, the defects of the horizontal specimens have the characteristics of larger size, more irregular shape and higher defect density, which are the main reason why its elevated temperature fatigue performance is lower than that of the vertical specimens.

Conclusions
The HCF performance of the SLM Ti6Al4V alloy tested at 400℃ has been studied.The conclusions are as follows: (1) The fatigue cracks of the SLM Ti6Al4V specimens (Kt=1) at 400℃ are mainly initiated from the LOF defects on the surface and subsurface.The relative position of the defect with the surface and the irregularity of the defect have the greater influence on elevated temperature fatigue life than the defect size.(2) The HCF performance of the SLM Ti6Al4V alloy exhibits significant anisotropy on different stress ratios at 400℃, because of the difference of shape, size and density of LOF defects in crack source region between both direction specimens.Compared with vertical specimens, the horizontal specimens with defect characteristics of larger size, more irregular shape and higher defect density exhibit lower fatigue lives.(3) The fatigue performance of the SLM Ti6Al4V alloy is better than that of the casting and even the forging, but the dispersion of fatigue data of the SLM alloy is much greater than that of the both traditional process alloys.

Figure 1 .
Figure 1.Layout of the rods for HCF specimens on the build platform

Figure 2 .
Figure 2. Morphologies and particle size distribution of Ti6Al4V powders

Figure 4 .
Figure 4. Geometry and dimension of the HCF specimen

Figure 10 .Figure 11 .
Figure 10.Defect morphology in crack source region of the SLM Ti6Al4V alloy

Table 1 .
Chemical composition of the SLM Ti6Al4V alloy

Table 2 .
Formula parameter values of the SLM Ti6Al4V alloy S-N curves under different conditions

Table 3 .
Comparison of fatigue performance of Ti6Al4V alloys with different processes