Introduction

Metal Ti and Ti alloys with low density, advantageous high special strength (σ/ρ, σ is the strength and ρ is the density), and favorable erosion resistance in numerous aggressive environments are widely applied in aviation, navigation, medical, chemical engineering, and other fields (Ref 1, 2). Given that the performance requirement in high-tech fields continues to increase, the development of new Ti alloys with high and ultrahigh properties is important (Ref 3, 4). Several TiZrAlV series alloys (abbreviated to TZAV alloys) with ultrahigh strength have been reported (Ref 5, 6). Two representative alloys, namely, TZAV-20 (Ref 7) and TZAV-30 (Ref 8), with ultrahigh special strength exceeding 340 MPa/(g/cm3) present tremendous application potential in aviation, automotive, navigation, and other fields.

However, the high cost of Ti alloys from difficulties in processing severely limits their practical applications (Ref 9). Thus, processability improvement has always been the focus for Ti alloys. Plastic forming processes are deeply dependent on strain-hardening behavior (Ref 10). Thus, research on strain-hardening behavior of Ti alloys is important for improving processability and promoting practical applications. Mechanical properties largely depend on the phase constitution and microstructure. Wang (Ref 11) and Venkatesh (Ref 12) showed that the strain-hardening exponent (n) of Ti alloys decreases with the increase in strain rate and/or with the decrease in temperature. Zuo et al. (Ref 13) showed that the n of the ferrite-martensite dual-phase steels increases with the increasing martensite fraction. Ertorer et al. (Ref 14) showed that the n of pure Ti decreases with the decrease in average grain size. However, investigation on strain-hardening behavior of new TZAV series alloys is rare.

As mentioned, strain-hardening behavior of Ti alloys largely depends on phase constitution and microstructure, which can be controlled and adjusted effectively by plastic working and heat treatment. In the current paper, the phase constitution and microstructure of the TZAV-30 alloy were adjusted via heat treatment. Strain-hardening behavior of the TZAV-30 alloy with various phase constitution and microstructures was investigated. The influence mechanism of phase constitution and microstructure on strain-hardening behavior was also discussed. The current findings are beneficial to improve processability and promote the practical application of Ti alloys.

Experimental Procedure and Methods

Sponge Ti (99.7 wt.%), HZr-1 level sponge Zr (Zr + Hf ≥ 99.5 wt.%, Hf ≤ 3 wt.%), industrially pure Al (99.5 wt.%), and V (99.9 wt.%) were used to prepare alloy with the following composition in weight percentage: Ti-30Zr-5Al-3V (TZAV-30). The alloy was melted by ZHT-001 type vacuum-consumable electro-arc furnace, followed by cogging at 1050 °C and heat forging at 950 °C successively. The detailed preparation procedure is shown in Ref 15. Specimens with size of 10 mm × 10 mm × 70 mm (width × height × length) were sectioned as heat-treated specimens. Heat treatments were performed in a SK-G06143 tubular vacuum heat treatment furnace in a protective argon atmosphere. The detailed heat treatments are shown in Table 1. Bone-shaped plate specimens with a 21 mm gauge length, 3 mm gauge width, and 2 mm gauge thickness were prepared for tensile tests.

Table 1 Heat treatments performed on TZAV-30 alloy
Full size table

The phase constitution of the specimen with size of 10 mm × 10 mm × 5 mm was examined by x-ray diffraction (XRD) that uses Cu Kα radiation [ASTM: D5380-93(2014)]. S-4800 type scanning electron microscopy (SEM) and JEM2010 type transmission electron microscopy (TEM) were used to test and/or analyze the microstructure. SEM specimens were prepared by using a typical quantitative metallographic method. The etching solution is a mixed solution (5 mL HF, 10 mL HNO3, and 85 mL H2O). TEM specimens were prepared by twin-jet electrochemical polishing in a solution of 10% perchloric acid and 90% methanol at 13 V and 238 K. Tensile tests were conducted on an Instron mechanical testing system at room temperature, and the strain rate was 5 × 10−4/s in the range from 10−3 to 10−4/s according to the ISO 6892-1:2009 (Ref 16). Strain throughout the entire testing process was monitored with an extensometer with 12.5 mm gauge length.

Statistical thickness distribution of α plate in various specimens is obtained from SEM or TEM micrographs by using standard test methods for determining average grain size (ASTM: E112-12). The volume fraction of the retained β phase was measured by point-counting method. The SEM/TEM micrograph was covered by a suitable grid. The percentage of the number of grid points located at the β phase in the total number of grid points was regarded as volume fraction of the retained β phase. The true stress-strain (S-e) curve was calculated from engineering stress-strain (σ-ε) curve using formulas S = σ(1 + ε) and e = ln (1 + ɛ). Uniform plastic deformation is the deformation stage at true strain from e e to e b in S-e curve, where e e and e b are the true strain at elastic limit and ultimate tensile strength, respectively.

To avoid occasional results, all tests on each specimen were performed for a minimum of three times. The data points signed in the figures are the average values. The error bars in the figures show the relative deviations between the three repetitive results and average value.

Results

Microstructure of Specimens with Various Treatments

Figure 1 shows that the XRD patterns of TZAV-30 alloy underwent various heat treatments. Analysis reveals that FC, AC, and WA specimens are composed of a two-phase equilibrium of α + β, whereas WQ specimen consists of major α″ martensite and several retained β phases. Previous reports (Ref 8, 17) also showed similar results, in which the phase constitution of annealed and aged TZAV-30 specimens is an equilibrium of two phases, α + β.

Fig. 1
figure 1

The XRD patterns of the examined alloy with various heat treatments

Full size image

Figure 2 shows the micrographs of TZAV-30 specimens with various heat treatments. The micrographs of Fig. 2(a) to (c) show that FC, AC, and WA specimens possess similar basket weave microstructures. However, the content of the retained β phase and grain size of α plate in the aforesaid three specimens are different. The microstructural feature of WQ specimen is characterized by major acicular α″ martensite and retained β phase, as shown in Fig. 2(d).

Fig. 2
figure 2

Micrographs of specimens with various heat treatments (a) SEM of FC, (b) SEM of AC, (c) TEM of WA, and (d) SEM of WQ

Full size image

Strain-Hardening Behavior of Specimens with Various Treatments

Figure 3 shows the tensile results of TZAV-30 alloy with various heat treatments. True stress-strain (S-e) curves of various specimens are shown in Fig. 3(a). To investigate strain-hardening behavior, the uniform plastic deformation stage has been marked on each S-e curve. Holloman (Ref 18) showed that the relationship between S and e during uniform plastic deformation stage in the S-e curve of metal, and alloys can be expressed as S = Ke n., where K is a constant and n is the strain-hardening exponent. According to the Holloman equation, the ln S − ln e during uniform plastic deformation stage should be a straight line, and the n value is the slope of the straight line of ln S − ln e. The ln S − ln e fitted straight line during the uniform plastic deformation stage of each examined specimen is obtained from the corresponding S-e curve, and all ln S − ln e fitted straight lines are shown in Fig. 3(b). Thus, the n of various specimens can be obtained from Fig. 3(b). The average value is 0.068 for FC specimen, 0.121 for AC, 0.412 for WQ, and 0.061 for WA. The particular strain-hardening behavior of WQ specimen should result from the metastable α″ martensite. Previous results (Ref 19, 20) showed that the metastable α″ martensite in Ti alloys possesses low strength, low elastic modulus, and high strain-hardening exponent.

Fig. 3
figure 3

Tensile results of TZAV-30 alloy with various heat treatments (a) S-e curves, the stage between two triangles marked on each curve is the uniform plastic deformation stage, (b) lnS − lne curves of uniform plastic deformation stage and corresponding fitted straight lines

Full size image

Discussion

Effects of Phase Constitution on Strain-Hardening Exponent

As mentioned above, the strain-hardening exponent n of WQ specimen with α″ martensite and retained β phase is considerably higher than that of other specimens with two-phase α + β equilibrium. To investigate the effect of phase constitution on strain-hardening behavior, we obtained the strain-hardening rate versus true strain (dS/de − e) curve of each specimen from the corresponding S-e curve; and all dS/de − e curves are shown in Fig. 4(a). The FC, AC, and WA specimens with the same phase constitution present similar dS/de − e curves, which can be characterized by three stages, as reported by Ahmad (Ref 21). The initial plastic deformation stage with high dS/de is defined as Stage I; the rapid decrease stage in dS/de − e curve is Stage II; and in Stage III, the dS/de is almost a constant near zero. However, Salem et al. (Ref 22) showed a second hardening behavior after the rapid decrease (Stage II) in the dS/de − e curve of pure α-Ti, as shown in Fig. 4(b); the second hardening behavior results from the formation of twins during plastic deformation. However, no twins can be found in the examined alloy during plastic deformation (Ref 23). Previous reports (Ref 24-26) also showed that the second hardening behavior in dS/de − e curve of alloys with hexagonal close-packed (hcp) crystal structure is intimately linked with deformation twins. Won et al. (Ref 27) also proved that the strain hardening along rolling direction with twining deformation is more evident than that along normal and transverse directions with dislocation slip in high purity Ti. The dS/de − e curve of WQ specimen is also shown in Fig. 4(a). The dS/de − e curve of WQ specimen is markedly different from that of FC, AC, and WA specimens but is similar with the curves in Fig. 4(b). However, the mechanism of second hardening behavior in the dS/de − e curve of WQ specimen should not result in the formation of twins. Apart from the formation of twins, martensitic transition is another key factor of second hardening behavior in dS/de − e curve. The previous result showed that martensitic transition induced by tensile stress in WQ specimen will occur (Ref 28). Neelakantan and Hanada et al. (Ref 29, 30) also revealed that martensitic transition induced by stress in Ti alloys with non-equilibrium phases can result in second hardening phenomenon (Ref 31, 32).

Fig. 4
figure 4

The dS/de − e curves of (a) the examined alloy with various treatments, (b) α-Ti with various grain sizes of 30 and 130 μm

Full size image

Effects of Phase Content on Strain-Hardening Exponent

According to XRD theory (Ref 33), the relative content of each phase in the multi-phase alloy presents a similar trend with the relative peak intensity of corresponding phase in the same XRD pattern. Thus, the trend of peak intensity ratio I β /I α in various treated specimens can indicate the trend of relative phase content. Here, I β is a sign of (200) peak intensity of β phase near 2θ = 55.2°, and I α is a sign of (102) peak intensity of α phase near 2θ = 52.3°. Based on the results of XRD, the I β /I α value is approximately 0.105 for FC specimen, 0.496 for AC and 0.086 for WA. Thus, the β phase content order from high to low is in the order of AC, WA, and FC specimens. Variations of n with I β /I α , as well as volume fraction of β phase (f β ), are shown in Fig. 5 and Table 2. The n value increases from 0.068 for FC specimen to 0.121 for AC with increasing I β /I α ratio from 0.105 to 0.496, and variation of n with f β is similar to that with I β /I α ratio. A previous result (Ref 23) showed that the plastic deformation mechanism of TZAV-30 alloy with α + β phases involves only dislocation glide. Thus, strain hardening of the examined alloy originates mainly from the accumulation of the forest of dislocations and their interaction (Ref 34, 35). In fact, the dislocation motion and multiplication during plastic deformation in β phase with bcc structure are easier and faster than that in the α phase with hcp structure. Therefore, the n value in the examined alloy with α + β phases is largely dependent on the content of the β phase. Kang et al. (Ref 36) also showed that deformation of β phase in Ti-4%Cr alloy caused significant strain hardening.

Fig. 5
figure 5

Variations of n with relative phase content in specimens with α + β two phases

Full size image
Table 2 Microstructure and strain-hardening exponent of specimens with various treatments
Full size table

Effects of Grain Size on Strain-Hardening Exponent

Based on the micrographs in Fig. 2, variation of thickness (h) of α plate in WA, AC, and FC specimens with α + β phases and its effect on n are shown in Fig. 6. Statistical h distributions of α plate in Fig. 6(a) to (c) show that the h values are primarily around 0.35 μm for WA specimen, 0.85 μm for AC, and 4.2 μm for FC. Previous results (Ref 37, 38) showed that the strain-hardening behavior of metal and alloys depends on the microstructure parameters, such as grain size. Variation in n with average h of α plate in Fig. 6(d) and Table 2 reveals that the trend line of n versus h is a parabola, that is, n increases first with the increase in h but then decreases with further increase in h. This trend is inconsistent with previous results (Ref 39, 40), which showed that n decreases gradually with increase in grain size. The key factor of this discrepancy is the relative content of β phase. According to section 4.2, strain-hardening behavior and the n value of the examined alloy depend greatly on the relative content of β phase. Thus, the influence of h on strain-hardening behavior is obscured by the relative content of β phase. Although the WA specimen possesses smaller h, the relative content of β phase in that specimen is also lower. Therefore, n of WA specimen is lower than that of AC instead. Previous result (Ref 23) also revealed that the effect of grain size on the mechanical properties of the Ti alloy is obscured by the relative content of β phase and aspect ratio of the α plate.

Fig. 6
figure 6

Statistical h distributions of specimens (a) WA, (b) AC, (c) FC, and (d) the trend line of n vs. h of α plate

Full size image

Conclusions

The strain-hardening behavior of TZAV-30 alloy with various heat treatments is investigated in this work. The conclusions are summarized as follows:

  1. (1)

    The strain-hardening behavior of the examined alloy depends considerably on heat treatment. Strain-hardening exponent is approximately 0.061 for WA specimen, 0.068 for FC, 0.121 for AC, and 0.412 for WQ.

  2. (2)

    Metastable α″ martensite and phase transition induced by tensile stress can increase the strain-hardening exponent of TZAV-30 alloy significantly.

  3. (3)

    The relative content of β phase is a major factor of strain-hardening exponent of TZAV-30 alloy with α + β equilibrium phases instead of the thickness of the α plate.