Impact of Morphology on the High Cycle Fatigue Behavior of Ti-6Al-4V for Aerospace

Abstract

The mechanical properties of Ti-6Al-4V alloy are affected by its microstructures. However, the effects of these microstructures on the high cycle fatigue behavior of Ti-6Al-4V alloy with a mixed structure (α + β phases) remain unknown. In this study, three alloy specimens were prepared using different hot-deformation methods, and their microstructures were investigated by optical microscopy and scanning electron microscopy. Fatigue tests were then performed to determine their high cycle fatigue and fatigue crack propagation behavior. All specimens showed a bimodal structure, but the morphology of each phase (e.g., diameter, shape, and volume fraction) showed notable differences. Among the samples prepared, the forged sample (FS) showed the lowest fatigue strength in all cycles. The fatigue strength of the homogeneously rolled sample (HS) was slightly higher than that of the rolled sample (RS) below 10⁶ cycles but lower above 10⁶ cycles. Compared with those of RS and HS, the secondary α (α_s) grain width of FS was twofold larger. The interconnected primary α (α_p) phase clusters in HS appeared to promote microcrack propagation.

1. Introduction

Ti-6Al-4V alloy is widely used in aerospace applications, such as turbine engines, airframe applications, and space shuttles, owing to its excellent corrosion resistance and mechanical properties. Particularly, Ti-6Al-4V alloy is compatible with carbon-fiber-reinforced plastics (CFRP) that are also finding wide uses in aerospace applications [1]. As the use of composite materials in aircraft increases, the applications of Ti-6Al-4V, which demonstrates small differences in galvanic corrosion, thermal expansion coefficient, and high strength-to-weight ratios compared with other composite materials, can be expected to increase [2,3].

The desired mechanical properties of Ti alloys can be effectively obtained by controlling their microstructure. Annealed Ti-6Al-4V alloy, which is mainly used in industrial fields such as aviation, contains large amounts of the hcp-α phase, along with a small amount of the bcc-β phase. The distribution, shape, and volume fraction of these phases influence the mechanical properties of the alloy, especially its fatigue properties. Research has confirmed that the different microstructural properties of alloy materials contribute to their mechanical properties.

The size and fraction of the primary α (α_p) phase, the width of the secondary α (α_s) phase, and the size and fraction of the lamellar colonies in Ti-6Al-4V alloys with a mixed structure change according to the speed and degree of cold working; thus, the fatigue characteristics of these alloys can also be expected to change, depending on these characteristics [4,5]. Boyer and Wallem [6] studied the complex relationship between the properties and microstructures of Ti-6Al-4V alloy obtained via the hot-deformation method and suggested that the mechanical properties of this alloy are closely related to its microstructural properties. Lütjering [7] optimized the mechanical properties of Ti-6Al-4V by controlling its microstructure and found that the latter fundamentally affects properties such as strength, plasticity, fatigue, and creep. However, the effects of various microstructural features on the fatigue properties of Ti-6Al-4V alloy with a mixed structure, are complicated; thus, further research on this topic is necessary. High fatigue strength is an important property of Ti-6Al-4V alloys used as aerospace materials.

In this study, the high cycle fatigue behaviors of three Ti-6Al-4V alloys with different microstructures obtained using different hot-deformation methods were studied. Optical microscopy (OM) and scanning electron microscopy (SEM) were used to observe the microstructures of the alloys, as well as the size and distribution of each phase. Based on these observations, the effects of microstructural features on the high cycle fatigue and fatigue crack propagation behavior of the Ti-6Al-4V alloys were assessed.

2. Materials and Methods

2.1. Materials

The effect of microstructure on the fatigue strength of Ti-6Al-4V alloys was investigated. Samples were prepared and denoted forged sample (FS), rolled sample (RS), and homogeneously rolled sample (HS) according to their thermomechanical processing method. Each sample was ground and etched with Kroll’s solution. The surface structure of the samples was then observed under an optical microscope (Olympus BX51, Tokyo, Japan).

Ti-6Al-4V alloy is most widely used in the annealed state. Alloys with an α + β mixed structure are obtained by annealing below the beta transformation (T_β) temperature [5]. In this study, FS was hot-forged at 1223 K, i.e., below the T_β temperature (1263 K), and then annealed at 978 K for 7.2 h. Both RS and HS were hot-rolled at 1203 K and annealed at 978 K. HS was manufactured using an ingot of a smaller volume compared with those used to prepare FS and RS; therefore, this sample is considered to have a higher cooling rate. In addition, although the cooling rates of the internal and external portions of an ingot tend to differ, the cooling rate of HS is considered to be relatively homogeneous because of the small size of its ingot.

During deformation, the α + β phase is deformed plastically. Moreover, different textures that can influence the mechanical properties of the alloy material can be formed by tailoring the deformation mode [8].

2.2. Microstructural Observation

Prior to surface observation, all samples were polished sequentially using waterproof emery papers with grit sizes of #120, #320, #600, #1000, and #2400 and then buffed to a mirror finish using Al₂O₃ powder with average particle diameters of 3 and 1 μm. The samples were etched using 5% HNO₃ + 2% HF solution, and their microstructures were observed under an optical microscope. The volume fraction and particle diameter of the α_p phase and width of the α_s phase were determined using ImageJ software (v1.53, National Institutes of Health, Bethesda, MD, USA).

In general, the size of the equiaxed α phase and presence (or absence) of a mixed structure in a metal alloy are determined by the processing conditions, such as the degree of the process before annealing, the annealing temperature, and the cooling rate in the temperature range in which the α and β phases coexist [9]. Figure 1 illustrates how the OM images were processed using ImageJ. ImageJ is image processing software developed by the National Institutes of Health (USA). This software can be used to analyze geometric characteristics, such as length, area, major axis, minor axis, perimeter, and Feret’s diameter [8]. In Figure 1b, the α + β colonies in the OM image were marked with a black pen, and the image was converted into a binary image using the Make Binary tool. Geometric characteristics were then obtained from the outlined image shown in Figure 1c using the Analyze Particles function in the Analyze tool bar.

The grains of Ti-6Al-4V are irregular in shape; therefore, performing morphological analysis of these grains is challenging. In this case, the Feret diameter or equivalent circle diameter (ECD) is used to measure the diameter of grains. The Feret diameter of a grain refers to the longest dimension of a grain independent of its orientation and is defined as the distance between two parallel planes [9]. These two parallel planes must restrict the grain between them and be perpendicular to the specified direction [10]. The ECD of a particle represents the diameter of a sphere that occupies the same two-dimensional surface area as the particle [11]. In this study, the average grain diameter of α_p grains was measured via the ECD method from the ImageJ results.

2.3. Fatigue Tests

The geometry of the tensile and fatigue test specimens cut from the samples used in this study is shown in Figure 2. The center of each specimen was buffed to a mirror finish to remove surface defects. Tensile tests were performed according to ASTM E8 in air at room temperature using an Instron-type machine. The load was measured using the load cell installed in the machine, and the strain was measured using a strain gauge attached to the specimen.

The high cycle fatigue test was performed according to ASTM E466 using a uniaxial load-type electro-servo-hydraulic machine with a sine wave signal under a load ratio of 0.1 and a frequency of 10 Hz. In this study, 1 × 10⁷ cycles were set as the run-out condition. The fatigue properties of the alloys were evaluated by obtaining the maximum cyclic stress (S) and number of cycles to failure (N) and plotting the S–N curve for each sample. Coefficients of determination (R²) were also calculated using Excel 2019 (Microsoft Corp., Redmond, WA, USA) to obtain the trendline of S-N curve.

3. Results and Discussion

3.1. Materials

Figure 1 shows a photograph of the specimens obtained after heat treatment. The white and round areas in the image indicate the precipitated α_p phase, and the areas showing stacked long and thin grains indicate a lamellar structure in which the acicular metamorphic β and α_s phases are finely layered. The specimens show bimodal structures composed of α and β phases, but the morphology of each phase, such as its diameter, shape, and volume fraction, shows differences depending on the sample. Compared with RS and HS, FS shows an α_s phase which is close to equiaxed. Compared with FS, RS exhibits a narrower needle-like α_s phase. Finally, HS shows a long and thin α_s phase in the rolling direction and a finer α_s phase.

3.2. Mechanical Properties

Stress-controlled fatigue tests were performed on the three samples to investigate the effect of microstructural properties on the high cycle fatigue behavior of Ti-6Al-4V alloy. Tensile tests were then performed with a crosshead speed of 8.33 × 10⁻⁶ m·s⁻¹ to analyze the mechanical properties of the Ti-6Al-4V specimens; the results are shown in

Table 1. Tensile properties of the Ti64 samples in this study.

Specimen	Yield Strength (MPa)	Tensile Strength (MPa)	Tensile Elongation (%)
FS	871	956	21
RS	876	974	18
HS	885	985	20

. No significant difference in elongation and yield strength was observed among the specimens. The high cycle fatigue resistance of metals tends to be proportional to their tensile strength in many cases [12]. According to Table 1, the tensile strength of HS was approximately 30 MPa higher than that of FS, but no significant difference between the strengths of these specimens was noted. In addition, the tensile strength of the samples increased in the order of FS < RS < HS.

Figure 3 shows the maximum stress amplitudes observed over 10⁴–10⁷ fatigue failure cycles. In this study, a power trendline was fitted to the data using the Excel “Insert Trend Line” function, to obtain a trendline close to the perfect fit (R² = 1). R² values of FS, RS and HS are 0.75, 0.73 and 0.95, respectively. The fatigue limit was defined as the highest stress at which the specimen did not fail even after 10⁷ cycles, in this study. The fatigue limits of RS, HS and FS were 500, 450 and 350 MPa, respectively. FS clearly showed the lowest fatigue strength in all cycles. Moreover, the fatigue strength of HS was slightly higher than that of RS in the region below 10⁶ but lower in the region above 10⁶. HS showed smaller variations in fatigue behavior compared with RS and FS (Figure 3).

Figure 4 shows SEM images of the fatigue-fractured surface of the specimens. The high cycle fatigue resistance of Ti-6Al-4V alloy is more influenced by microstructural factors, such as the α_p and α_s phases and α + β colonies, rather than its tensile strength. As shown in Figure 4, the fracture surface of all three specimens exhibited cleavage facets with cracks growing in the grain while crossing the α phase, in addition to a river pattern.

The most important microstructural parameters affecting the fatigue strength of specimens with a bimodal structure are the volume fraction and size of the α_p phase and the width of the α_s phase. Figure 5 and Figure 6 show the results of the geometric analyses of the longitudinal cross-sections of the fatigue specimens. Figure 5, in particular, shows the distribution of the diameters of the α_p grains in all specimens, as measured by the ECD method.

Fatigue crack propagation resistance is generally determined by the slip reversibility of an alloy material [13]. The metallurgical factors with the greatest influence on slip reversibility are grain size and stacking fault energy. The lower the stacking fault energy of a material, the higher its slip reversibility and the better its fatigue crack propagation resistance [14]. In this study, because the three specimens had the same chemical composition, differences in their stacking fault energy can be considered to be negligible.

When the fatigue crack propagation mechanism is cleavage failure, smaller grain sizes lead to lower slip reversibility and higher fatigue crack propagation rates [14]. The average diameters of the α_p grains of FS, RS, and HS were 15, 15, and 13 μm, respectively, and the diameter of the α_p grains of HS showed a more homogeneous distribution compared with those of FS and RS. A difference in grain diameter of approximately 10 μm does not significantly affect the fatigue crack propagation behavior of a metal. Therefore, in this study, the effect of α_p grain diameter on the fatigue behavior of Ti alloy is considered to be negligible.

The shape of the cleavage facet in Figure 4a–c is similar to that of the α_p phase. In particular, the diameter of the cleavage plane is not significantly different from the measured average diameter of the α_p phase. This finding indicates that the α_p phase boundary plays a major role in changing the crystal plane in which cleavage failure occurs. Therefore, fatigue crack propagation occurs through the α + β colonies, and β phase factors do not affect the fatigue crack propagation or fatigue behavior of the alloys.

The width, rather than the diameter, of the α_s phase should be preferentially considered when assessing the strength of bimodal Ti-6Al-4V alloys composed of a lamellar microstructure (α_s + β phase) [15]. When the width of α_s grains is narrow, good crack propagation resistance occurs because of the high grain boundary density [8]. Figure 6 shows the widths of over 150 α_s grains in each sample. The average widths of the α_s grains of FS, RS, and HS are 2.8, 1.1, and 1.2 μm, respectively. Similar to the results of their distribution and mean values, RS and HS did not show significant differences with respect to their α_s grain width. The α_s grain width of FS was over twice those of RS and HS. Thus, among the samples, FS may be expected to have the lowest grain boundary density, which could explain its low fatigue strength.

In addition to the width of the α_s phase, the volume fraction of this phase is an important microstructural parameter that affects the fatigue strength of Ti alloys [15]. In this study, the volume fractions of the α_p phase of FS, RS, and HS were measured using ImageJ and found to be 44%, 43%, and 56%, respectively. Compared with the volume fractions of the other samples, the volume fraction of HS is clearly higher; no sharp difference between the volume fractions of the α_p phase of FS and RS was noted. In the high cycle fatigue region (above 10⁶ cycles), the fatigue strength of HS was lower than that of RS, but the specimens have nearly the same properties (i.e., average grain diameter of the α_p phase and average width of the α_s phase). The lower fatigue strength of HS may be related to the interconnections formed by most of its α_p grains, as shown in Figure 1c.

As indicated in Figure 4, all samples showed brittle fractures, such as cleavage fracture surface and river patterns. Fatigue cracks are considered to propagate within the α_p phase, rather than the grain boundaries. Crack propagation is generally less likely to occur than crack formation under low cycle (high load) fatigue, crack propagation is the dominant domain, determining fatigue life. In contrast, under high cycle (low load) fatigue, crack formation is less likely to occur than crack propagation, and the number of cycles required for crack formation to occur over the entire fracture cycle is high. As mentioned earlier, slip reversibility and the slip system are important factors influencing fatigue crack generation in Ti-6Al-4V alloys without inclusions. α_p phase with hcp structure has limited slip system, which can act as a notch for the crack initiation. Consequently, microcracks are easy to be nucleated and propagated along the slip bands within the α_p phase. Furthermore, interconnected α_p clusters can serve as large, single grains through which microcracks easily propagate [8]. In the case of HS, which is characterized with low crack propagation resistance and a high proportion of interconnected α_p grains, cracks within the α_p phase propagate toward adjacent α_p phases without significant resistance in the high fatigue cycle region (above 10⁶ cycles), possibly leading to its slightly lower fatigue strength compared with RS.

4. Conclusions

In this study, the effects of microstructural features on the high cycle fatigue and fatigue crack propagation behavior of three Ti-6Al-4V samples, each having different microstructures obtained under different thermomechanical treatment conditions, were investigated. The primary findings of this study are as follows.

• Among the samples prepared, FS showed the lowest fatigue strength in all cycles. In addition, particularly in the region above 10⁶ cycles, the fatigue strength of HS was lower than that of RS.
• The α_s phase width of FS was two times greater than those of RS and HS. This large width could lead to a lower grain boundary density, which may explain the low fatigue strength of FS.
• Interconnected α_p phase clusters in HS can serve as large, single α_p grains through which microcracks easily form and propagate.