- Email: [email protected]
Microstructural evolution and its effect on the mechanical behavior of Ti-5Al-5Mo-5V-3Cr alloy during aging
Materials Science & Engineering A 731 (2018) 239–248
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Microstructural evolution and its effect on the mechanical behavior of Ti5Al-5Mo-5V-3Cr alloy during aging ⁎
Baozhen Jianga,b, , Satoshi Emuraa, Koichi Tsuchiyaa,b, a b
T
⁎
Research Center for Structural Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan Graduate School of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan
A R T I C LE I N FO
A B S T R A C T
Keywords: Ti-5553 alloy High-pressure torsion α phase precipitates Mechanical properties Strengthening mechanism
Development of different microstructures with α phase precipitates by a series of aging of Ti-5Al-5Mo-5V-3Cr (mass%, Ti-5553) alloy and their influences on the mechanical behavior were studied. Coarse acicular α phase formed in the solution treated (ST) samples, while ultrafine equiaxed α phase formed in the high-pressure torsion deformed (HPTed) samples. The strengthening by the precipitation of α phase could be explained by the Hall-Petch relationship between the yield strength and the α-to-α interphase spacing. The Hall-Petch coefficient of the HPTed samples was larger than that of the ST samples, because the equiaxed α phase was harder than the acicular α phase. The ductility of the HPTed samples was enhanced by the homogeneous distribution of equiaxed α phase.
1. Introduction Ti alloys have been used in many industrial fields thanks to their attractive combination of mechanical properties. Among them, metastable β-Ti alloys exhibit unique combination of high strength-todensity ratio, excellent hardenability and good corrosion resistance [1–4]. Ti-5Al-5Mo-5V-3Cr (mass%, Ti-5553) alloy is a recently developed metastable β-Ti alloy with promising mechanical properties for aerospace structural applications [1,5–7]. Ti-5553 alloy exhibits a higher strength and a better high-cycle fatigue property over Ti-6Al-4V (Ti-64) alloy which is the most widely used Ti alloy in aerospace industry [8]. The wider processing window and the better deep-hardenability of Ti-5553 alloy make it an ideal candidate for the thick forging applications in the landing gear components to replace Ti-10V2Fe-3Al (Ti-1023) alloy [5–7]. Ti-5553 alloy can be heat treated in a section size up to 152 mm accompanied with only a slight drop in properties at the thicker section size by air cooling, whereas Ti-1023 alloy requires a water quenching and the section size is limited to 76 mm. This is mostly due to the more sluggish precipitation of α phase in Ti-5553 alloy because of the addition of Mo and Cr. There have been several studies on Ti-5553 alloy, showing that the microstructure can vary substantially as a function of thermomechanical processing and heat treatment and the mechanical properties of Ti-5553 alloy are critically dependent on its microstructure [9–14]. Quantitative microstructural analysis were carried out to clarify the influence of microstructure, especially the α phase precipitates, on
⁎
the flow behavior [9], the fracture toughness [11] and the tensile properties [12–14] of Ti-5553 alloy. The volume fraction and size of α phase precipitates were considered as important microstructural features that could affect the mechanical properties. From the viewpoint of precipitation/dispersion hardening mechanism, the influence of α phase precipitates on the mechanical properties is mostly due to the creation of a large number of α/β interfaces which act as effective dislocation barriers. Recently, it has been shown that severe plastic deformation, such as high-pressure torsion (HPT) and equal channel angular processing (ECAP), can bring about a change in the morphology of α phase from acicular to equiaxed [14–19]. Zafari et al. [17] investigated the mechanism for the change in the morphology of α phase in HPT deformed Ti-5553 alloy by examining the α precipitation at the very early aging stages. The present authors carried out a series of aging on HPT deformed Ti-5553 alloy to investigate the formation and growth process of equiaxed α phase [19]. Ultrafine equiaxed α phase distributed homogeneously in Ti-5553 alloy subjected to HPT deformation of 10 revolutions. However, in all the cases mentioned above, the influence of the precipitation of equiaxed α phase on the mechanical properties was not included. Further investigations are required to develop a deeper fundamental understanding of the strengthening mechanism by the precipitation of equiaxed α phase. The present study shows the results of a systematic study done on the HPT deformed Ti-5553 alloy. The evolution of microstructure and mechanical properties as a function of aging is presented. As a comparison and to establish the microstructure-
Corresponding authors at: Research Center for Structural Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan. E-mail addresses: [email protected] (B. Jiang), [email protected] (K. Tsuchiya).
https://doi.org/10.1016/j.msea.2018.06.064 Received 7 March 2018; Received in revised form 14 June 2018; Accepted 15 June 2018 Available online 18 June 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 1. BSE micrographs of the ST samples after aging at 823 K for (a) 0.3 ks, (b) 1.2 ks, (c) 14.4 ks and (d) 360 ks; aging at 873 K for (e) 0.3 ks, (f) 1.2 ks, (g) 14.4 ks and (h) 360 ks; aging at 923 K for (i) 0.3 ks, (j) 1.2 ks, (k) 14.4 ks and (l) 360 ks; and aging at 1023 K for (m) 0.3 ks, (n) 1.2 ks, (o) 14.4 ks and (p) 360 ks.
diameter of 10 mm. Disk samples with a thickness of 0.9 mm were sliced from the bar and were solution treated above the β transus temperature (1129 K) at 1273 K for 3.6 ks. After grinding the surface to a thickness of 0.85 mm, deformation by high-pressure torsion (HPT) was applied on the disk samples at room temperature under a pressure of 5 GPa and a rotation speed of 0.2 rpm for 10 revolutions. Hereafter, the solution treated samples are referred as ST samples and the HPT deformed samples are referred as HPTed samples. The microstructures of Ti-5553 alloy after ST and HPT deformation have been reported by the present authors [19]. A series of aging were done on the ST samples and HPTed samples. The aging temperatures were 823 K, 873 K, 923 K and 1023 K, and the aging time ranged from 0.3 ks to 360 ks. The
mechanical properties correlation, the evolution of microstructure and mechanical properties as a function of aging in the Ti-5553 alloy without HPT deformation is also presented. The influences of the equiaxed α phase and the acicular α phase on the mechanical behavior will be discussed in detail.
2. Experimental An ingot of Ti-5553 alloy was prepared by cold crucible levitation melting (CCLM) with a weight of around 1.2 kg. The analyzed chemical composition is Ti-5.04Al-5.14Mo-4.91V-3.04Cr (mass%). After hot forging and rolling at 1473 K, a bar of the alloy was cold swaged to a 240
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
823 K and 873 K for 0.3 ks. The bright-field (BF) images in Fig. 3(a) and (d) show an ultrafine duplex microstructure which is composed of α and β phases. The dark-field (DF) images in Fig. 3(b) and (e) show the ultrafine equiaxed α phase precipitates. The average size of equiaxed α phase is around 50 nm and 80 nm after aging at 823 K and 873 K for 0.3 ks, respectively. Although a short aging appears to have started the β recrystallization and α precipitation, both the α and β phases are not well-defined and cannot exhibit clear grain boundaries. As demonstrated by the corresponding selected area electron diffraction (SAED) patterns in Fig. 3(c) and (f), the ring patterns show the nanoscale grains with high-angle misorientations. The TEM micrographs of the HPTed samples aged at 823 K and 923 K for 14.4 ks are shown in Fig. 4. The well-defined α and β phases can be seen. The interfaces between the α phase and β phase are sharp with no significant sign of stress between the two adjacent phases. As shown by the SAED pattern in Fig. 4(f), the reflection of [0001]α is superimposed on [11 0]β diffraction pattern, indicating the equiaxed α phase obeys the Burgers orientation relationship with the β phase, which is described by {0001}α//{011}β and < 112 0 > α// < 111 > β [23]. The variation in the size of acicular α phase and equiaxed α phase is shown in Fig. 5(a) and (b), respectively. Both the acicular α phase and the equiaxed α phase grow on increasing the aging time or aging temperature. The size of equiaxed α phase is much smaller than the length of acicular α phase, but is a litter larger than the width of acicular α phase. The large error bars in Fig. 5 suggest that there is a large variation in the size of α phase, which can be seen clearly from Fig. 1 and Fig. 2. Fig. 6(a) and (b) show the volume fraction of acicular α phase and equiaxed α phase, respectively. The volume fraction of acicular α phase increases significantly during aging from 0.3 ks to 14.4 ks, and keeps almost no change by further aging. A high volume fraction of equiaxed α phase is obtained after aging for 0.3 ks. Further aging to long time leads to a very slow increase in the volume fraction of equiaxed α phase. These results also show that the volume fraction of acicular α phase and equiaxed α phase is quite similar when the aging time is ranging from 3.6 ks to 360 ks. Fig. 7(a) and (b) show the α-to-α interphase spacing in the ST samples and the HPTed samples, respectively. The change in the α-to-α interphase spacing of acicular α phase over time is irregular, while there is an increase in the α-to-α interphase spacing of equiaxed α phase over time. For both the acicular and equiaxed α phases, higher aging temperature leads to a larger α-to-α interphase spacing. The α-to-α interphase spacing of equiaxed α phase is smaller than that of the acicular α phase. Basically, a higher volume fraction of α phase or a smaller size of α phase at a certain volume fraction leads to the smaller α-to-α interphase spacing, as indicated by Eq. (1). Fig. 7 also shows large error bars, which implies a random distribution of α phase. The nucleation sites for α phase include the prior β grain boundaries, secondary phase such as ω phase [24] and dislocations [25]. In the present study, the main nucleation sites for α phase are the metastable ω phase in the ST samples and the HPT-induced dislocations in the HPTed samples. Once nucleated, the growth of α nucleus is governed by the migration of α/β interfaces which is diffusion-controlled. In general, α nucleus is bounded by a combination of coherent or semicoherent facets and incoherent interfaces so as to minimize the total interfacial free energy [26,27]. The incoherent interfaces migrate much faster than the coherent or semicoherent interfaces, leading to the formation of the acicular shape. As reported by our recent study [19], the equiaxed α phase obeys Burgers orientation relationship with one of the surrounding β phase. It can be considered that only one interface is coherent or semicoherent and the other interfaces are incoherent which can migrate much faster, leading to the formation of the equiaxed morphology. The growth of α phase requires the partitioning of alloying elements between α phase and β phase. The growth of acicular and equiaxed α phases over aging time is accompanied by the enrichment of Al element
solution treatment and all the aging were performed in an Ar atmosphere followed by water quenching. Backscattered electron (BSE) observations were carried out on the cross-section of the samples mounted in resin by using scanning electron microscopy (SEM) on JEOL JSM-7001F operated at 20 kV. Based on the BSE images, quantitative microstructural analysis on the volume fraction and size of α phase as well as the α-to-α interphase spacing were measured by software package ImageJ. At least 5 BSE images were selected for the analysis of the volume fraction of α phase. The size of α phase was measured manually and more than 200 α-phase particles were selected for the analysis. In most cases, the α and β phases are easily identified by their relative contrast - a result of their respective average atomic mass. In the case of unclear α/β interfaces, it is essential to threshold (binary contrast enhancement) the micrographs to precisely delineate the two phases. Detailed information on the ImageJ basics can be found in the handout by Larry Reinking [20]. The α-to-α interphase spacing, d, was estimated by the linear intercept method. The α-to-α interphase spacing is defined as the mean edge-to-edge distance along random straight lines between all possible pairs of particles, and can be calculated by the following equation [21]:
d = (1 − Vα )/ NL
(1)
where Vα is the volume fraction of α phase and NL is the number of particle interceptions per unit length of test line. This equation is valid regardless of size, shape or distribution of the particles. At least 30 test lines which are randomly oriented with respect to the structure were used. Microstructural characterization was also carried out using transmission electron microscopy (TEM) on JEOL JEM-2100F operated at 200 kV. 3 mm discs for TEM analysis were cut from the median plane and perforated by twin-jet electropolishing at a temperature of 228 K in an electrolyte of 6 vol% perchloric acid, 30 vol% butanol and 64 vol% methanol. Vickers microhardness was measured with an applied load of 0.98 N for 15 s on the cross-section of mounted samples. 57 indentations were conducted on each sample. The nanohardness of α and β phases was measured by nanoindentation using a Hysitron Triboindenter TI950 in a load control mode with a Berkovich indenter. Both the α and β phases were indented in at least 30 positions with a peak load of 500 μN at a loading/unloading rate of 50 μN/s. The indenter was held at peak force for 5 s before unloading. The nanohardness was calculated by the Oliver-Pharr method [22]. Miniature tensile specimens with a gauge section of 4 mm × 1 mm × 0.7 mm were cut by electric discharge machine (EDM) from the disk samples. The tensile specimens were electrochemically polished to remove the damage layer from EDM cutting and to obtain a smooth surface. The tensile tests were carried out at room temperature with a strain rate of 2 × 10−3 s−1 and the strain was monitored by a video-extensometer with a 3 µm resolution. At least 4 tensile specimens were prepared for each condition. 3. Results and discussion 3.1. Microstructural evolution After aging at different temperatures and for different time, α phase precipitates with different size and morphology were obtained in the Ti5553 alloy. Some selected BSE images on the microstructures in the ST samples and in the HPTed samples are shown in Figs. 1 and 2, respectively. The α phase precipitates exhibit a darker contrast in the BSE images because of their higher Al content and lower Mo content. There can be seen two different morphologies of α phase precipitated in the Ti-5553 alloy, namely, acicular α phase in the ST samples and equiaxed α phase in the HPTed samples. The precipitation of equiaxed α phase is much more rapid than the precipitation of acicular α phase, as shown in Fig. 2 that a large amount of equiaxed α precipitates form after aging for 0.3 ks. Fig. 3 shows the TEM micrographs of the HPTed samples aged at 241
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 2. BSE micrographs of the HPTed samples after aging at 823 K for (a) 0.3 ks, (b) 1.2 ks, (c) 14.4 ks and (d) 360 ks; aging at 873 K for (e) 0.3 ks, (f) 1.2 ks, (g) 14.4 ks and (h) 360 ks; aging at 923 K for (i) 0.3 ks, (j) 1.2 ks, (k) 14.4 ks and (l) 360 ks; and aging at 1023 K for (m) 0.3 ks, (n) 1.2 ks, (o) 14.4 ks and (p) 360 ks.
in α phase and rejection of Mo, V and Cr elements from α phase, as investigated by the present authors [19]. Al element is known to be an α stabilizer, while Mo, V and Cr elements are β stabilizers. The higher diffusion rate of alloying elements at higher aging temperature leads to the more rapid growth of α phase to a coarser size. But aging at higher temperature can inhibit the formation of metastable ω phase in the ST samples and can promote the recovery of deformed microstructure in the HPTed samples, reducing the nucleation site for α phase. That is the reason why a lower volume fraction of α phase is obtained after aging at a higher temperature. The HPT-induced high density of dislocations can act as preferential nucleation sites for α phase and can provide enhanced diffusion of
alloying elements by many orders of magnitude [28–30], resulting in the more rapid formation of α phase in the HPTed samples after aging for 0.3 ks. However, there is a pile-up of β stabilizers near the equiaxed α/β interface in the β grain at the early aging stage, which is caused by the overlap of diffusion fields [19]. The pile-up of β stabilizers ahead of growing α phase leads to a reduction in the α/β interface energy, reducing the driving force for the growth of equiaxed α phase and contributing to the ultrafine size of equiaxed α phase.
3.2. Nanohardness of β and α phases The nanohardness of β and α phases was measured by 242
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 3. TEM microstructure in the HPTed sample aged at 823 K for 0.3 ks: (a) BF image, (b) DF image and (c) SAED pattern; and TEM microstructure in the HPTed sample aged at 873 K for 0.3 ks: (d) BF image, (e) DF image and (f) SAED pattern.
[32]. Dong et al. [33] studied the micromechanical behavior of α phase with different morphologies in Ti-6Al-4V alloy. The equiaxed α phase was harder than the acicular α phase. This was explained by dislocation hardening effect that dislocations accumulated easily in the equiaxed α phase due to its hexagonal or spherical shape as well as its small size. However, further investigation needs to be carried out to clarify the difference in the nanohardness of equiaxed α phase and acicular α phase. In addition, a wide variation of nanohardness in the α phase is achieved. The nanohardness of α phase not only depends on its chemical composition, but also depends on its crystallographic orientation [34–37].
nanoindentation in the ST sample and the HPTed sample aged at 1023 K for 360 ks. The typical load-displacement curves are shown in Fig. 8(a). The penetration depth of β phase is deeper than that of α phase in both the ST sample and the HPTed sample, indicating that the β phase is softer than the α phase. The penetration depth of acicular α phase in the ST sample is deeper than that of equiaxed α phase in the HPTed sample, indicating the acicular α phase is softer than the equiaxed α phase. Fig. 8(b) shows the nanohardness of α phase and β phase calculated from the unloading part of the load-displacement curves. The inserts show the corresponding in-situ scanning probe microscopy (SPM) images after nanoindentation tests in the ST sample and the HPTed sample. The white arrows show the indentations in the α phase and the black arrows show the indentations in the β phase. In the ST sample, the nanohardness of β phase is ranging from 3.82 GPa to 4.28 GPa with an average value of 4.03 ± 0.12 GPa and the nanohardness of acicular α phase is ranging from 4.50 GPa to 5.20 GPa with an average value of 4.83 ± 0.22 GPa. In the HPTed sample, the nanohardness of β phase is ranging from 4.10 GPa to 4.43 GPa with an average value of 4.22 ± 0.11 GPa and the nanohardness of equiaxed α phase is ranging from 5.32 GPa to 6.10 GPa with an average value of 5.77 ± 0.22 GPa. Min et al. [31] investigated the nanohardness of β and α phases in Ti15Mo (mass%) alloy and Ti-15Mo-3Al (mass%) alloy. The average nanohardness of β phase and α phase in Ti-15Mo alloy was 3.43 ± 0.36 GPa and 2.48 ± 0.41 GPa, respectively. The average nanohardness of β phase and α phase in Ti-15Mo-3Al alloy was 3.83 ± 0.27 GPa and 4.34 ± 0.36 GPa, respectively. It was considered that the enrichment of Al in α phase could significantly harden the α phase. The enrichment of Al element which has fairly large atomic size difference to Ti element and also large solid solubility in the α phase can cause significant substitutional solid solution hardening. It was reported that the enrichment of Al in the α phase can increase the c/a ratio by decreasing the a axis without causing much change in the c axis
3.3. Mechanical properties Fig. 9(a) shows the average hardness of the aged ST samples. The hardness of the ST samples before aging is 288.1 ± 5.3 Hv. The precipitation of α phase by aging usually results in an increase in the hardness, such as aging at 823–923 K. But the precipitation of α phase at 1023 K has little contributions to the hardness. Fig. 9(b) shows the average hardness of the aged HPTed samples. The original hardness of the HPTed sample before aging is 380.7 ± 20.7 Hv, which is due to the ultrafine-grained structure and high density of lattice defects. After aging at 823 K and 873 K, the formation of ultrafine α phase increases the hardness. Peak hardness is attained quickly by aging for 0.3 ks. However, the hardness becomes lower than the original hardness after aging at 923 K and 1023 K, although there are numerous α phases formed. This is mainly due to the rapid recovery of the deformed structure during aging. The hardness is higher in the HPTed samples than in the ST samples when the aging time is ranging from 0.3 ks to 3.6 ks, but becomes almost the same by ranging the aging time from 14.4 ks to 360 ks. The tensile properties after aging at different temperatures for 14.4 243
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 4. TEM microstructure in the HPTed sample aged at 823 K for 14.4 ks: (a) BF image, (b) DF image and (c) SAED pattern; and TEM microstructure in the HPTed sample aged at 923 K for 14.4 ks: (d) BF image, (e) DF image and (f) SAED pattern from the region indicated by circle in (d).
3.4. Strengthening mechanism by α phase
ks were also investigated. Fig. 10(a) and (b) shows the tensile stressstrain curves of the ST samples and the HPTed samples, respectively. The tendency is that there is a decrease in the ultimate tensile strength (UTS) accompanied by an increase in the elongation to failure (EL) with increasing the aging temperature. Similar UTS and EL are obtained in the ST samples and the HPTed samples.
This study showed that acicular α phase formed in the ST samples and equiaxed α phase formed in the HPTed samples by aging. For the Ti-5553 alloy with a variety of microstructures, a wide range of mechanical properties could be achieved. These observations showed the effects of the α phase precipitates on the mechanical properties and allowed us to discuss the possible strengthening mechanism. As mentioned in the results part, the α phase precipitates are harder
Fig. 5. Average size of α phase in (a) the ST samples and (b) the HPTed samples after a series of aging. Length (L, solid lines) and width (W, broken lines) of the acicular α phase are shown for the ST samples. 244
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 6. Volume fraction of α phase in (a) the ST samples and (b) the HPTed samples after a series of aging.
than the β matrix. Therefore, plastic deformation should preferentially occur in the β matrix and the α/β interfaces can act as barriers to dislocation movement. This is the widely accepted strengthening mechanism of the α phase precipitates in metastable β-Ti alloys. The relationship between the yield strength (σy) and the α-to-α interphase spacing (d) was studied. Fig. 11 shows the relationship between σy and d−1/2. Here, the yield strength was calculated by the empirical relationship of σy = Hv/3 [38]. The diagram in Fig. 11 shows a very reasonable linear relationship between σy and d−1/2. The linear regression fit gives correlation factors of R2 = 0.96 and R2 = 0.95 for the ST samples and the HPTed samples, respectively. The good linear relationship between σy and d−1/2 suggests that the strengthening mechanism by α phase precipitates follows the Hall-Petch relation which is expressed as follows [39,40]:
σy = σ0 + kD−1/2
that yielding takes place when the stress concentration is large enough to cause the slip to propagate from one grain to the neighboring grain. The same interpretation can be considered to be valid in the two-phase alloys that the phase interfaces play the same role as the grain boundaries. The dislocations in the β phase matrix pile up in the vicinity of α/β interfaces, consequently causing the strengthening of β phase matrix. Then the slip can pass through the α/β interfaces and the α phase undergoes plastic deformation. Similar discussions on the deformation mechanism of Ti alloys with β phase and α phase have been reported [14,41,42]. Here it should be noted that the yield strength appears to keep almost no change when the α-to-α interphase spacing is below 0.15 µm, implying the Hall-Petch relation is no longer valid. The possible reasons for this phenomenon are that the array of dislocations may be limited and the dislocation sources can hardly exist in this kind of very small α-to-α interphase spacing. As shown in Fig. 11, the Hall-Petch relation for the ST samples is given as below:
(2)
where σ0 is the friction stress for starting dislocation movement in the matrix, k is the Hall-Petch coefficient individual for each material and representing the relative hardening contribution of the grain boundaries and D is the grain size. It is well known that the Hall-Petch relation is initially developed for single-phase polycrystalline materials. The Hall-Petch relation can be explained as that the stress concentration occurs due to the pile-up of dislocations at the grain boundaries and
σy (MPa) = 887 + 198d−1/2 (μm−1/2)
(3)
And the Hall-Petch relation for the HPTed samples is given as following:
σy (MPa) = 760 + 259d−1/2 (μm−1/2)
Fig. 7. Average α-to-α interphase spacing in (a) the ST samples and (b) the HPTed samples after a series of aging. 245
(4)
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 8. (a) Typical load-depth curves for the β and α phases in the ST sample and HPTed sample aged at 1023 K for 360 ks; and (b) nanohardness of the β and α phases. Inserts in (b) show the corresponding in-situ SPM images. Black and white arrows show the indents in β phase and α phase, respectively.
Fig. 9. Average Vickers microhardness in (a) the ST samples and (b) the HPTed samples after a series of aging.
Fig. 10. Tensile stress-strain curves of (a) the ST samples and (b) the HPTed samples after aging at different temperatures for 14.4 ks.
246
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
[46–49]. This kind of net-work like structure can deform in a more homogeneous manner thanks to the accommodation of plastic deformation by the softer β phase, promoting the ductility. 4. Conclusions In the present study, the influence of aging on a solution treated (ST) and a high-pressure torsion (HPT) deformed Ti-5Al-5Mo-5V-3Cr alloy was reported in terms of the microstructural characterization and the mechanical properties. The conclusions can be drawn as follows: 1. Acicular α phase formed in the ST samples while equiaxed α phase formed in the HPTed samples. The ST samples and the HPTed samples had a similar volume fraction of α phase after long time aging. The particle size of equiaxed α phase was much smaller than the length of acicular α phase, but was a litter larger than the width of acicular α phase. The α-to-α interphase spacing of equiaxed α phase was smaller than that of the acicular α phase. 2. The α phase was harder than the β phase in Ti-5553 alloy and the equiaxed α phase was harder than the acicular α phase. 3. The ST samples and the HPTed samples had similar mechanical properties after long time aging. 4. The strengthening by the precipitation of α phase could be explained by the Hall-Petch relationship between the yield strength and the α-to-α interphase spacing. The Hall-Petch coefficient of the HPTed samples was larger than that of the ST samples because of the higher strength of equiaxed α phase. The ductility could be enhanced by the homogeneous formation of equiaxed α phase.
Fig. 11. Relationship between σy and d−1/2 for the ST samples and the HPTed samples.
Acknowledgements This work was supported in part by a Grant-in-Aid for Scientific research on Innovative Area, “Bulk Nanostructured Metals,” through MEXT, Japan (Contract No. 22102004). References [1] [2] [3] [4] [5] [6] [7]
Fig. 12. Relationship between the elongation to failure and the α-to-α interphase spacing for the ST samples and the HPTed samples.
The friction stress for starting dislocation motion (σ0) and the HallPetch coefficient (k) are different for the ST samples and the HPTed samples. It is puzzling that the HPTed samples exhibit a lower friction stress than the ST samples, inconsistent with the nanoindentation results that the strength of β phase in the HPTed samples is the same as or even a little higher than that in the ST samples. For the Hall-Petch coefficient, it was reported that the grain boundary hardening could result in a substantial increase in the Hall-Petch coefficient [43–45]. Therefore, in the present case, it is considered that the higher HallPetch coefficient in the HPTed samples is mainly due to the higher strength of equiaxed α phase. The influence of α phase precipitates on the ductility is another part that should be considered. Generally, the smaller α-to-α interphase spacing indicates the shorter mean free path for dislocation movement in the β phase, which eventually reduces the ductility. However, as shown by the relationship between the elongation to failure and the αto-α interphase spacing in Fig. 12, the HPTed samples exhibit a better ductility than the ST samples when they have the same α-to-α interphase spacing. The ductility of the HPTed samples is enhanced by the homogeneous distribution of equiaxed α phase, as discussed in a previous study by the present authors [14]. The softer equiaxed β phase is surrounded by the harder equiaxed α phase, forming net-work like structure which is quite similar to the so-called “harmonic structure”
[8] [9] [10] [11] [12] [13]
[14]
[15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
247
D. Banerjee, J.C. Williams, Acta Mater. 61 (2013) 844–879. R.R. Boyer, Mater. Sci. Eng. A 213 (1996) 103–114. I. Weiss, S.L. Semiatin, Mater. Sci. Eng. A 243 (1998) 46–65. S. Ankem, C.A. Greene, Mater. Sci. Eng. A 263 (1999) 127–131. R.R. Boyer, R.D. Briggs, J. Mater. Eng. Perform. 22 (2013) 2916–2920. S.L. Nyakana, J.C. Fanning, R.R. Boyer, J. Mater. Eng. Perform. 14 (2005) 799–811. J.D. Cotton, R.D. Briggs, R.R. Boyer, S. Tamirisakandala, P. Russo, N. Shchetnikov, J.C. Fanning, JOM 67 (2015) 1281–1303. S. Veeck, D. Lee, R. Boyer, R. Briggs, Adv. Mater. Process. 162 (2004) 47–49. N.G. Jones, R.J. Dashwood, D. Dye, M. Jackson, Metall. Mater. Trans. A 40A (2009) 1944–1954. S.K. Kar, A. Ghosh, N. Fulzele, A. Bhattacharjee, Mater. Charact. 81 (2013) 37–48. A. Ghosh, S. Sivaprasad, A. Bhattacharjee, S.K. Kar, Mater. Sci. Eng. A 568 (2013) 61–67. S. Shekhar, R. Sarkar, S.K. Kar, A. Bhattacharjee, Mater. Des. 66 (2015) 596–610. B.A. Welk, Microstructural and property relationships in β-titanium alloy Ti-5553 (Master's thesis), The Ohio State University, 201 West 19th Avenue, Columbus, OH 43210, 2010. B.Z. Jiang, S. Emura, K. Tsuchiya, Improvement of ductility in Ti-5Al-5Mo-5V-3Cr alloy by network-like precipitation of blocky α phase, Mater. Sci. Eng. A 722 (2018) 129–135. W. Xu, D.P. Edwards, X. Wu, M. Stoica, M. Calin, U. Kühn, J. Eckert, K. Xia, Scr. Mater. 68 (2013) 67–70. W. Xu, X. Wu, M. Stoica, M. Calin, U. Kühn, J. Eckert, K. Xia, Acta Mater. 68 (2012) 5067–5078. A. Zafari, K. Xia, Scr. Mater. 124 (2016) 151–154. B.Z. Jiang, K. Tsuchiya, S. Emura, X.H. Min, Mater. Trans. 55 (2014) 877–884. B.Z. Jiang, S. Emura, K. Tsuchiya, J. Alloy. Compd. 738 (2018) 283–291. L. Reinking, ImageJ basics, 2007. 〈https://imagej.nih.gov/ij/docs/pdfs/ImageJ. pdf〉. E. Underwood, Quantitative stereology for microstructural analysis, in: J. McCall, W. Mueller (Eds.), Microstructural Analysis, Springer, US, 1973, p. 47. W.C. Oliver, G.M. Pharr, Mater. Res. 7 (1992) 1564–1583. W.G. Burgers, Physica 1 (1934) 561–586. S. Nag, R. Banerjee, R. Srinivasan, J.Y. Hwang, M. Harper, H.L. Fraser, Acta Mater. 57 (2009) 2136–2147. T. Makino, R. Chikaizumi, T. Nagaoka, T. Furuhara, Mater. Sci. Eng. A 213 (1996)
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
51–60. [26] T. Furuhara, S. Takagi, H. Watanabe, T. Maki, Metall. Mater. Trans. A 27 (1996) 1635–1646. [27] S. Zherebtsov, G. Salishchev, S.L. Semiatin, Philos. Mag. Lett. 90 (2010) 903–914. [28] X. Sauvage, F. Wetscher, P. Pareige, Acta Mater. 53 (2005) 2127–2135. [29] D. Setman, E. Schafler, E. Korznikova, M.J. Zehetbauer, Mater. Sci. Eng. A 493 (2008) 116–122. [30] S.V. Divinski, G. Reglitz, H. Rosner, Y. Estrin, G. Wilde, Acta Mater. 59 (2011) 1974–1985. [31] X.H. Min, L. Zhang, K. Sekido, T. Ohmura, S. Emura, K. Tsuchiya, K. Tsuzaki, Mater. Sci. Tech. 28 (2012) 342–347. [32] D.E. Gordon, J.W. Hagemeyer, J. Mater. Sci. 10 (1975) 1725–1731. [33] J.Z. Dong, F.G. Li, C.P. Wang, Mater. Sci. Eng. A 580 (2013) 105–113. [34] T.B. Britton, H. Liang, F.P.E. Dunne, A.J. Wilkinson, Proc. R. Soc. A (2010) 695–719. [35] C. Zambaldi, Y.Y. Yang, T.R. Bieler, D. Raabe, J. Mater. Res. 27 (2012) 356–367. [36] J. Kwon, M.C. Brandes, P. Sudharshan Phani, A.P. Pilchak, Y.F. Gao, E.P. George, G.M. Pharr, M.J. Mills, Acta Mater. 61 (2013) 4743–4756. [37] E. Lee, Microstructure evolution and microstructure/mechanical properties
[38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
248
relationship in α+β titanium alloys (Ph.D. thesis), The Ohio State University, 201 West 19th Avenue, Columbus, OH 43210, 2004. P. Zhang, S.X. Li, Z.F. Zhang, Mater. Sci. Eng. A 529 (2011) 62–73. E.O. Hall, Proc. Phys. Soc. Lond. B64 (1951) 747. N.J. Petch, J. Iron Steel Inst. 174 (1953) 25. N.G. Jones, M. Jackson, Mater. Sci. Tech. 27 (2011) 1025–1032. A.G. Paradkar, S.V. Kamat, A.K. Gogia, B.P. Kashyap, Mater. Sci. Eng. A 520 (2009) 168–173. K. Takeda, N. Nakada, T. Tsuchiyama, S. Takaki, ISIJ Int. 48 (2008) 1122–1125. S. Takaki, D. Akama, N. Nakada, T. Tsuchiyama, Mater. Trans. 55 (2014) 28–34. C.H. Caceres, Gemma E. Mann, J.R. Griffiths, Metall. Mater. Trans. A 42 (2011) 1950–1959. S.K. Vajpai, K. Amayama, M. Ota, T. Watanabe, R. Maeda, T. Sekiguchi, G. Dirras, A. Tingaud, IOP Conf. Ser.: Mater. Sci. Eng. 63 (2014) 012030. H. Fujiwara, R. Akada, A. Noro, Y. Yoshita, K. Ameyama, Mater. Trans. 49 (2008) 90–96. T. Sekiguchi, K. Ono, H. Fujiwara, K. Ameyama, Mater. Trans. 51 (2010) 39–45. D. Orlov, H. Fujiwara, K. Ameyama, Mater. Trans. 54 (2013) 1549–1553.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Microstructural evolution and its effect on the mechanical behavior of Ti5Al-5Mo-5V-3Cr alloy during aging ⁎
Baozhen Jianga,b, , Satoshi Emuraa, Koichi Tsuchiyaa,b, a b
T
⁎
Research Center for Structural Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan Graduate School of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan
A R T I C LE I N FO
A B S T R A C T
Keywords: Ti-5553 alloy High-pressure torsion α phase precipitates Mechanical properties Strengthening mechanism
Development of different microstructures with α phase precipitates by a series of aging of Ti-5Al-5Mo-5V-3Cr (mass%, Ti-5553) alloy and their influences on the mechanical behavior were studied. Coarse acicular α phase formed in the solution treated (ST) samples, while ultrafine equiaxed α phase formed in the high-pressure torsion deformed (HPTed) samples. The strengthening by the precipitation of α phase could be explained by the Hall-Petch relationship between the yield strength and the α-to-α interphase spacing. The Hall-Petch coefficient of the HPTed samples was larger than that of the ST samples, because the equiaxed α phase was harder than the acicular α phase. The ductility of the HPTed samples was enhanced by the homogeneous distribution of equiaxed α phase.
1. Introduction Ti alloys have been used in many industrial fields thanks to their attractive combination of mechanical properties. Among them, metastable β-Ti alloys exhibit unique combination of high strength-todensity ratio, excellent hardenability and good corrosion resistance [1–4]. Ti-5Al-5Mo-5V-3Cr (mass%, Ti-5553) alloy is a recently developed metastable β-Ti alloy with promising mechanical properties for aerospace structural applications [1,5–7]. Ti-5553 alloy exhibits a higher strength and a better high-cycle fatigue property over Ti-6Al-4V (Ti-64) alloy which is the most widely used Ti alloy in aerospace industry [8]. The wider processing window and the better deep-hardenability of Ti-5553 alloy make it an ideal candidate for the thick forging applications in the landing gear components to replace Ti-10V2Fe-3Al (Ti-1023) alloy [5–7]. Ti-5553 alloy can be heat treated in a section size up to 152 mm accompanied with only a slight drop in properties at the thicker section size by air cooling, whereas Ti-1023 alloy requires a water quenching and the section size is limited to 76 mm. This is mostly due to the more sluggish precipitation of α phase in Ti-5553 alloy because of the addition of Mo and Cr. There have been several studies on Ti-5553 alloy, showing that the microstructure can vary substantially as a function of thermomechanical processing and heat treatment and the mechanical properties of Ti-5553 alloy are critically dependent on its microstructure [9–14]. Quantitative microstructural analysis were carried out to clarify the influence of microstructure, especially the α phase precipitates, on
⁎
the flow behavior [9], the fracture toughness [11] and the tensile properties [12–14] of Ti-5553 alloy. The volume fraction and size of α phase precipitates were considered as important microstructural features that could affect the mechanical properties. From the viewpoint of precipitation/dispersion hardening mechanism, the influence of α phase precipitates on the mechanical properties is mostly due to the creation of a large number of α/β interfaces which act as effective dislocation barriers. Recently, it has been shown that severe plastic deformation, such as high-pressure torsion (HPT) and equal channel angular processing (ECAP), can bring about a change in the morphology of α phase from acicular to equiaxed [14–19]. Zafari et al. [17] investigated the mechanism for the change in the morphology of α phase in HPT deformed Ti-5553 alloy by examining the α precipitation at the very early aging stages. The present authors carried out a series of aging on HPT deformed Ti-5553 alloy to investigate the formation and growth process of equiaxed α phase [19]. Ultrafine equiaxed α phase distributed homogeneously in Ti-5553 alloy subjected to HPT deformation of 10 revolutions. However, in all the cases mentioned above, the influence of the precipitation of equiaxed α phase on the mechanical properties was not included. Further investigations are required to develop a deeper fundamental understanding of the strengthening mechanism by the precipitation of equiaxed α phase. The present study shows the results of a systematic study done on the HPT deformed Ti-5553 alloy. The evolution of microstructure and mechanical properties as a function of aging is presented. As a comparison and to establish the microstructure-
Corresponding authors at: Research Center for Structural Materials, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan. E-mail addresses: [email protected] (B. Jiang), [email protected] (K. Tsuchiya).
https://doi.org/10.1016/j.msea.2018.06.064 Received 7 March 2018; Received in revised form 14 June 2018; Accepted 15 June 2018 Available online 18 June 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 1. BSE micrographs of the ST samples after aging at 823 K for (a) 0.3 ks, (b) 1.2 ks, (c) 14.4 ks and (d) 360 ks; aging at 873 K for (e) 0.3 ks, (f) 1.2 ks, (g) 14.4 ks and (h) 360 ks; aging at 923 K for (i) 0.3 ks, (j) 1.2 ks, (k) 14.4 ks and (l) 360 ks; and aging at 1023 K for (m) 0.3 ks, (n) 1.2 ks, (o) 14.4 ks and (p) 360 ks.
diameter of 10 mm. Disk samples with a thickness of 0.9 mm were sliced from the bar and were solution treated above the β transus temperature (1129 K) at 1273 K for 3.6 ks. After grinding the surface to a thickness of 0.85 mm, deformation by high-pressure torsion (HPT) was applied on the disk samples at room temperature under a pressure of 5 GPa and a rotation speed of 0.2 rpm for 10 revolutions. Hereafter, the solution treated samples are referred as ST samples and the HPT deformed samples are referred as HPTed samples. The microstructures of Ti-5553 alloy after ST and HPT deformation have been reported by the present authors [19]. A series of aging were done on the ST samples and HPTed samples. The aging temperatures were 823 K, 873 K, 923 K and 1023 K, and the aging time ranged from 0.3 ks to 360 ks. The
mechanical properties correlation, the evolution of microstructure and mechanical properties as a function of aging in the Ti-5553 alloy without HPT deformation is also presented. The influences of the equiaxed α phase and the acicular α phase on the mechanical behavior will be discussed in detail.
2. Experimental An ingot of Ti-5553 alloy was prepared by cold crucible levitation melting (CCLM) with a weight of around 1.2 kg. The analyzed chemical composition is Ti-5.04Al-5.14Mo-4.91V-3.04Cr (mass%). After hot forging and rolling at 1473 K, a bar of the alloy was cold swaged to a 240
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
823 K and 873 K for 0.3 ks. The bright-field (BF) images in Fig. 3(a) and (d) show an ultrafine duplex microstructure which is composed of α and β phases. The dark-field (DF) images in Fig. 3(b) and (e) show the ultrafine equiaxed α phase precipitates. The average size of equiaxed α phase is around 50 nm and 80 nm after aging at 823 K and 873 K for 0.3 ks, respectively. Although a short aging appears to have started the β recrystallization and α precipitation, both the α and β phases are not well-defined and cannot exhibit clear grain boundaries. As demonstrated by the corresponding selected area electron diffraction (SAED) patterns in Fig. 3(c) and (f), the ring patterns show the nanoscale grains with high-angle misorientations. The TEM micrographs of the HPTed samples aged at 823 K and 923 K for 14.4 ks are shown in Fig. 4. The well-defined α and β phases can be seen. The interfaces between the α phase and β phase are sharp with no significant sign of stress between the two adjacent phases. As shown by the SAED pattern in Fig. 4(f), the reflection of [0001]α is superimposed on [11 0]β diffraction pattern, indicating the equiaxed α phase obeys the Burgers orientation relationship with the β phase, which is described by {0001}α//{011}β and < 112 0 > α// < 111 > β [23]. The variation in the size of acicular α phase and equiaxed α phase is shown in Fig. 5(a) and (b), respectively. Both the acicular α phase and the equiaxed α phase grow on increasing the aging time or aging temperature. The size of equiaxed α phase is much smaller than the length of acicular α phase, but is a litter larger than the width of acicular α phase. The large error bars in Fig. 5 suggest that there is a large variation in the size of α phase, which can be seen clearly from Fig. 1 and Fig. 2. Fig. 6(a) and (b) show the volume fraction of acicular α phase and equiaxed α phase, respectively. The volume fraction of acicular α phase increases significantly during aging from 0.3 ks to 14.4 ks, and keeps almost no change by further aging. A high volume fraction of equiaxed α phase is obtained after aging for 0.3 ks. Further aging to long time leads to a very slow increase in the volume fraction of equiaxed α phase. These results also show that the volume fraction of acicular α phase and equiaxed α phase is quite similar when the aging time is ranging from 3.6 ks to 360 ks. Fig. 7(a) and (b) show the α-to-α interphase spacing in the ST samples and the HPTed samples, respectively. The change in the α-to-α interphase spacing of acicular α phase over time is irregular, while there is an increase in the α-to-α interphase spacing of equiaxed α phase over time. For both the acicular and equiaxed α phases, higher aging temperature leads to a larger α-to-α interphase spacing. The α-to-α interphase spacing of equiaxed α phase is smaller than that of the acicular α phase. Basically, a higher volume fraction of α phase or a smaller size of α phase at a certain volume fraction leads to the smaller α-to-α interphase spacing, as indicated by Eq. (1). Fig. 7 also shows large error bars, which implies a random distribution of α phase. The nucleation sites for α phase include the prior β grain boundaries, secondary phase such as ω phase [24] and dislocations [25]. In the present study, the main nucleation sites for α phase are the metastable ω phase in the ST samples and the HPT-induced dislocations in the HPTed samples. Once nucleated, the growth of α nucleus is governed by the migration of α/β interfaces which is diffusion-controlled. In general, α nucleus is bounded by a combination of coherent or semicoherent facets and incoherent interfaces so as to minimize the total interfacial free energy [26,27]. The incoherent interfaces migrate much faster than the coherent or semicoherent interfaces, leading to the formation of the acicular shape. As reported by our recent study [19], the equiaxed α phase obeys Burgers orientation relationship with one of the surrounding β phase. It can be considered that only one interface is coherent or semicoherent and the other interfaces are incoherent which can migrate much faster, leading to the formation of the equiaxed morphology. The growth of α phase requires the partitioning of alloying elements between α phase and β phase. The growth of acicular and equiaxed α phases over aging time is accompanied by the enrichment of Al element
solution treatment and all the aging were performed in an Ar atmosphere followed by water quenching. Backscattered electron (BSE) observations were carried out on the cross-section of the samples mounted in resin by using scanning electron microscopy (SEM) on JEOL JSM-7001F operated at 20 kV. Based on the BSE images, quantitative microstructural analysis on the volume fraction and size of α phase as well as the α-to-α interphase spacing were measured by software package ImageJ. At least 5 BSE images were selected for the analysis of the volume fraction of α phase. The size of α phase was measured manually and more than 200 α-phase particles were selected for the analysis. In most cases, the α and β phases are easily identified by their relative contrast - a result of their respective average atomic mass. In the case of unclear α/β interfaces, it is essential to threshold (binary contrast enhancement) the micrographs to precisely delineate the two phases. Detailed information on the ImageJ basics can be found in the handout by Larry Reinking [20]. The α-to-α interphase spacing, d, was estimated by the linear intercept method. The α-to-α interphase spacing is defined as the mean edge-to-edge distance along random straight lines between all possible pairs of particles, and can be calculated by the following equation [21]:
d = (1 − Vα )/ NL
(1)
where Vα is the volume fraction of α phase and NL is the number of particle interceptions per unit length of test line. This equation is valid regardless of size, shape or distribution of the particles. At least 30 test lines which are randomly oriented with respect to the structure were used. Microstructural characterization was also carried out using transmission electron microscopy (TEM) on JEOL JEM-2100F operated at 200 kV. 3 mm discs for TEM analysis were cut from the median plane and perforated by twin-jet electropolishing at a temperature of 228 K in an electrolyte of 6 vol% perchloric acid, 30 vol% butanol and 64 vol% methanol. Vickers microhardness was measured with an applied load of 0.98 N for 15 s on the cross-section of mounted samples. 57 indentations were conducted on each sample. The nanohardness of α and β phases was measured by nanoindentation using a Hysitron Triboindenter TI950 in a load control mode with a Berkovich indenter. Both the α and β phases were indented in at least 30 positions with a peak load of 500 μN at a loading/unloading rate of 50 μN/s. The indenter was held at peak force for 5 s before unloading. The nanohardness was calculated by the Oliver-Pharr method [22]. Miniature tensile specimens with a gauge section of 4 mm × 1 mm × 0.7 mm were cut by electric discharge machine (EDM) from the disk samples. The tensile specimens were electrochemically polished to remove the damage layer from EDM cutting and to obtain a smooth surface. The tensile tests were carried out at room temperature with a strain rate of 2 × 10−3 s−1 and the strain was monitored by a video-extensometer with a 3 µm resolution. At least 4 tensile specimens were prepared for each condition. 3. Results and discussion 3.1. Microstructural evolution After aging at different temperatures and for different time, α phase precipitates with different size and morphology were obtained in the Ti5553 alloy. Some selected BSE images on the microstructures in the ST samples and in the HPTed samples are shown in Figs. 1 and 2, respectively. The α phase precipitates exhibit a darker contrast in the BSE images because of their higher Al content and lower Mo content. There can be seen two different morphologies of α phase precipitated in the Ti-5553 alloy, namely, acicular α phase in the ST samples and equiaxed α phase in the HPTed samples. The precipitation of equiaxed α phase is much more rapid than the precipitation of acicular α phase, as shown in Fig. 2 that a large amount of equiaxed α precipitates form after aging for 0.3 ks. Fig. 3 shows the TEM micrographs of the HPTed samples aged at 241
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 2. BSE micrographs of the HPTed samples after aging at 823 K for (a) 0.3 ks, (b) 1.2 ks, (c) 14.4 ks and (d) 360 ks; aging at 873 K for (e) 0.3 ks, (f) 1.2 ks, (g) 14.4 ks and (h) 360 ks; aging at 923 K for (i) 0.3 ks, (j) 1.2 ks, (k) 14.4 ks and (l) 360 ks; and aging at 1023 K for (m) 0.3 ks, (n) 1.2 ks, (o) 14.4 ks and (p) 360 ks.
in α phase and rejection of Mo, V and Cr elements from α phase, as investigated by the present authors [19]. Al element is known to be an α stabilizer, while Mo, V and Cr elements are β stabilizers. The higher diffusion rate of alloying elements at higher aging temperature leads to the more rapid growth of α phase to a coarser size. But aging at higher temperature can inhibit the formation of metastable ω phase in the ST samples and can promote the recovery of deformed microstructure in the HPTed samples, reducing the nucleation site for α phase. That is the reason why a lower volume fraction of α phase is obtained after aging at a higher temperature. The HPT-induced high density of dislocations can act as preferential nucleation sites for α phase and can provide enhanced diffusion of
alloying elements by many orders of magnitude [28–30], resulting in the more rapid formation of α phase in the HPTed samples after aging for 0.3 ks. However, there is a pile-up of β stabilizers near the equiaxed α/β interface in the β grain at the early aging stage, which is caused by the overlap of diffusion fields [19]. The pile-up of β stabilizers ahead of growing α phase leads to a reduction in the α/β interface energy, reducing the driving force for the growth of equiaxed α phase and contributing to the ultrafine size of equiaxed α phase.
3.2. Nanohardness of β and α phases The nanohardness of β and α phases was measured by 242
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 3. TEM microstructure in the HPTed sample aged at 823 K for 0.3 ks: (a) BF image, (b) DF image and (c) SAED pattern; and TEM microstructure in the HPTed sample aged at 873 K for 0.3 ks: (d) BF image, (e) DF image and (f) SAED pattern.
[32]. Dong et al. [33] studied the micromechanical behavior of α phase with different morphologies in Ti-6Al-4V alloy. The equiaxed α phase was harder than the acicular α phase. This was explained by dislocation hardening effect that dislocations accumulated easily in the equiaxed α phase due to its hexagonal or spherical shape as well as its small size. However, further investigation needs to be carried out to clarify the difference in the nanohardness of equiaxed α phase and acicular α phase. In addition, a wide variation of nanohardness in the α phase is achieved. The nanohardness of α phase not only depends on its chemical composition, but also depends on its crystallographic orientation [34–37].
nanoindentation in the ST sample and the HPTed sample aged at 1023 K for 360 ks. The typical load-displacement curves are shown in Fig. 8(a). The penetration depth of β phase is deeper than that of α phase in both the ST sample and the HPTed sample, indicating that the β phase is softer than the α phase. The penetration depth of acicular α phase in the ST sample is deeper than that of equiaxed α phase in the HPTed sample, indicating the acicular α phase is softer than the equiaxed α phase. Fig. 8(b) shows the nanohardness of α phase and β phase calculated from the unloading part of the load-displacement curves. The inserts show the corresponding in-situ scanning probe microscopy (SPM) images after nanoindentation tests in the ST sample and the HPTed sample. The white arrows show the indentations in the α phase and the black arrows show the indentations in the β phase. In the ST sample, the nanohardness of β phase is ranging from 3.82 GPa to 4.28 GPa with an average value of 4.03 ± 0.12 GPa and the nanohardness of acicular α phase is ranging from 4.50 GPa to 5.20 GPa with an average value of 4.83 ± 0.22 GPa. In the HPTed sample, the nanohardness of β phase is ranging from 4.10 GPa to 4.43 GPa with an average value of 4.22 ± 0.11 GPa and the nanohardness of equiaxed α phase is ranging from 5.32 GPa to 6.10 GPa with an average value of 5.77 ± 0.22 GPa. Min et al. [31] investigated the nanohardness of β and α phases in Ti15Mo (mass%) alloy and Ti-15Mo-3Al (mass%) alloy. The average nanohardness of β phase and α phase in Ti-15Mo alloy was 3.43 ± 0.36 GPa and 2.48 ± 0.41 GPa, respectively. The average nanohardness of β phase and α phase in Ti-15Mo-3Al alloy was 3.83 ± 0.27 GPa and 4.34 ± 0.36 GPa, respectively. It was considered that the enrichment of Al in α phase could significantly harden the α phase. The enrichment of Al element which has fairly large atomic size difference to Ti element and also large solid solubility in the α phase can cause significant substitutional solid solution hardening. It was reported that the enrichment of Al in the α phase can increase the c/a ratio by decreasing the a axis without causing much change in the c axis
3.3. Mechanical properties Fig. 9(a) shows the average hardness of the aged ST samples. The hardness of the ST samples before aging is 288.1 ± 5.3 Hv. The precipitation of α phase by aging usually results in an increase in the hardness, such as aging at 823–923 K. But the precipitation of α phase at 1023 K has little contributions to the hardness. Fig. 9(b) shows the average hardness of the aged HPTed samples. The original hardness of the HPTed sample before aging is 380.7 ± 20.7 Hv, which is due to the ultrafine-grained structure and high density of lattice defects. After aging at 823 K and 873 K, the formation of ultrafine α phase increases the hardness. Peak hardness is attained quickly by aging for 0.3 ks. However, the hardness becomes lower than the original hardness after aging at 923 K and 1023 K, although there are numerous α phases formed. This is mainly due to the rapid recovery of the deformed structure during aging. The hardness is higher in the HPTed samples than in the ST samples when the aging time is ranging from 0.3 ks to 3.6 ks, but becomes almost the same by ranging the aging time from 14.4 ks to 360 ks. The tensile properties after aging at different temperatures for 14.4 243
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 4. TEM microstructure in the HPTed sample aged at 823 K for 14.4 ks: (a) BF image, (b) DF image and (c) SAED pattern; and TEM microstructure in the HPTed sample aged at 923 K for 14.4 ks: (d) BF image, (e) DF image and (f) SAED pattern from the region indicated by circle in (d).
3.4. Strengthening mechanism by α phase
ks were also investigated. Fig. 10(a) and (b) shows the tensile stressstrain curves of the ST samples and the HPTed samples, respectively. The tendency is that there is a decrease in the ultimate tensile strength (UTS) accompanied by an increase in the elongation to failure (EL) with increasing the aging temperature. Similar UTS and EL are obtained in the ST samples and the HPTed samples.
This study showed that acicular α phase formed in the ST samples and equiaxed α phase formed in the HPTed samples by aging. For the Ti-5553 alloy with a variety of microstructures, a wide range of mechanical properties could be achieved. These observations showed the effects of the α phase precipitates on the mechanical properties and allowed us to discuss the possible strengthening mechanism. As mentioned in the results part, the α phase precipitates are harder
Fig. 5. Average size of α phase in (a) the ST samples and (b) the HPTed samples after a series of aging. Length (L, solid lines) and width (W, broken lines) of the acicular α phase are shown for the ST samples. 244
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 6. Volume fraction of α phase in (a) the ST samples and (b) the HPTed samples after a series of aging.
than the β matrix. Therefore, plastic deformation should preferentially occur in the β matrix and the α/β interfaces can act as barriers to dislocation movement. This is the widely accepted strengthening mechanism of the α phase precipitates in metastable β-Ti alloys. The relationship between the yield strength (σy) and the α-to-α interphase spacing (d) was studied. Fig. 11 shows the relationship between σy and d−1/2. Here, the yield strength was calculated by the empirical relationship of σy = Hv/3 [38]. The diagram in Fig. 11 shows a very reasonable linear relationship between σy and d−1/2. The linear regression fit gives correlation factors of R2 = 0.96 and R2 = 0.95 for the ST samples and the HPTed samples, respectively. The good linear relationship between σy and d−1/2 suggests that the strengthening mechanism by α phase precipitates follows the Hall-Petch relation which is expressed as follows [39,40]:
σy = σ0 + kD−1/2
that yielding takes place when the stress concentration is large enough to cause the slip to propagate from one grain to the neighboring grain. The same interpretation can be considered to be valid in the two-phase alloys that the phase interfaces play the same role as the grain boundaries. The dislocations in the β phase matrix pile up in the vicinity of α/β interfaces, consequently causing the strengthening of β phase matrix. Then the slip can pass through the α/β interfaces and the α phase undergoes plastic deformation. Similar discussions on the deformation mechanism of Ti alloys with β phase and α phase have been reported [14,41,42]. Here it should be noted that the yield strength appears to keep almost no change when the α-to-α interphase spacing is below 0.15 µm, implying the Hall-Petch relation is no longer valid. The possible reasons for this phenomenon are that the array of dislocations may be limited and the dislocation sources can hardly exist in this kind of very small α-to-α interphase spacing. As shown in Fig. 11, the Hall-Petch relation for the ST samples is given as below:
(2)
where σ0 is the friction stress for starting dislocation movement in the matrix, k is the Hall-Petch coefficient individual for each material and representing the relative hardening contribution of the grain boundaries and D is the grain size. It is well known that the Hall-Petch relation is initially developed for single-phase polycrystalline materials. The Hall-Petch relation can be explained as that the stress concentration occurs due to the pile-up of dislocations at the grain boundaries and
σy (MPa) = 887 + 198d−1/2 (μm−1/2)
(3)
And the Hall-Petch relation for the HPTed samples is given as following:
σy (MPa) = 760 + 259d−1/2 (μm−1/2)
Fig. 7. Average α-to-α interphase spacing in (a) the ST samples and (b) the HPTed samples after a series of aging. 245
(4)
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
Fig. 8. (a) Typical load-depth curves for the β and α phases in the ST sample and HPTed sample aged at 1023 K for 360 ks; and (b) nanohardness of the β and α phases. Inserts in (b) show the corresponding in-situ SPM images. Black and white arrows show the indents in β phase and α phase, respectively.
Fig. 9. Average Vickers microhardness in (a) the ST samples and (b) the HPTed samples after a series of aging.
Fig. 10. Tensile stress-strain curves of (a) the ST samples and (b) the HPTed samples after aging at different temperatures for 14.4 ks.
246
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
[46–49]. This kind of net-work like structure can deform in a more homogeneous manner thanks to the accommodation of plastic deformation by the softer β phase, promoting the ductility. 4. Conclusions In the present study, the influence of aging on a solution treated (ST) and a high-pressure torsion (HPT) deformed Ti-5Al-5Mo-5V-3Cr alloy was reported in terms of the microstructural characterization and the mechanical properties. The conclusions can be drawn as follows: 1. Acicular α phase formed in the ST samples while equiaxed α phase formed in the HPTed samples. The ST samples and the HPTed samples had a similar volume fraction of α phase after long time aging. The particle size of equiaxed α phase was much smaller than the length of acicular α phase, but was a litter larger than the width of acicular α phase. The α-to-α interphase spacing of equiaxed α phase was smaller than that of the acicular α phase. 2. The α phase was harder than the β phase in Ti-5553 alloy and the equiaxed α phase was harder than the acicular α phase. 3. The ST samples and the HPTed samples had similar mechanical properties after long time aging. 4. The strengthening by the precipitation of α phase could be explained by the Hall-Petch relationship between the yield strength and the α-to-α interphase spacing. The Hall-Petch coefficient of the HPTed samples was larger than that of the ST samples because of the higher strength of equiaxed α phase. The ductility could be enhanced by the homogeneous formation of equiaxed α phase.
Fig. 11. Relationship between σy and d−1/2 for the ST samples and the HPTed samples.
Acknowledgements This work was supported in part by a Grant-in-Aid for Scientific research on Innovative Area, “Bulk Nanostructured Metals,” through MEXT, Japan (Contract No. 22102004). References [1] [2] [3] [4] [5] [6] [7]
Fig. 12. Relationship between the elongation to failure and the α-to-α interphase spacing for the ST samples and the HPTed samples.
The friction stress for starting dislocation motion (σ0) and the HallPetch coefficient (k) are different for the ST samples and the HPTed samples. It is puzzling that the HPTed samples exhibit a lower friction stress than the ST samples, inconsistent with the nanoindentation results that the strength of β phase in the HPTed samples is the same as or even a little higher than that in the ST samples. For the Hall-Petch coefficient, it was reported that the grain boundary hardening could result in a substantial increase in the Hall-Petch coefficient [43–45]. Therefore, in the present case, it is considered that the higher HallPetch coefficient in the HPTed samples is mainly due to the higher strength of equiaxed α phase. The influence of α phase precipitates on the ductility is another part that should be considered. Generally, the smaller α-to-α interphase spacing indicates the shorter mean free path for dislocation movement in the β phase, which eventually reduces the ductility. However, as shown by the relationship between the elongation to failure and the αto-α interphase spacing in Fig. 12, the HPTed samples exhibit a better ductility than the ST samples when they have the same α-to-α interphase spacing. The ductility of the HPTed samples is enhanced by the homogeneous distribution of equiaxed α phase, as discussed in a previous study by the present authors [14]. The softer equiaxed β phase is surrounded by the harder equiaxed α phase, forming net-work like structure which is quite similar to the so-called “harmonic structure”
[8] [9] [10] [11] [12] [13]
[14]
[15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
247
D. Banerjee, J.C. Williams, Acta Mater. 61 (2013) 844–879. R.R. Boyer, Mater. Sci. Eng. A 213 (1996) 103–114. I. Weiss, S.L. Semiatin, Mater. Sci. Eng. A 243 (1998) 46–65. S. Ankem, C.A. Greene, Mater. Sci. Eng. A 263 (1999) 127–131. R.R. Boyer, R.D. Briggs, J. Mater. Eng. Perform. 22 (2013) 2916–2920. S.L. Nyakana, J.C. Fanning, R.R. Boyer, J. Mater. Eng. Perform. 14 (2005) 799–811. J.D. Cotton, R.D. Briggs, R.R. Boyer, S. Tamirisakandala, P. Russo, N. Shchetnikov, J.C. Fanning, JOM 67 (2015) 1281–1303. S. Veeck, D. Lee, R. Boyer, R. Briggs, Adv. Mater. Process. 162 (2004) 47–49. N.G. Jones, R.J. Dashwood, D. Dye, M. Jackson, Metall. Mater. Trans. A 40A (2009) 1944–1954. S.K. Kar, A. Ghosh, N. Fulzele, A. Bhattacharjee, Mater. Charact. 81 (2013) 37–48. A. Ghosh, S. Sivaprasad, A. Bhattacharjee, S.K. Kar, Mater. Sci. Eng. A 568 (2013) 61–67. S. Shekhar, R. Sarkar, S.K. Kar, A. Bhattacharjee, Mater. Des. 66 (2015) 596–610. B.A. Welk, Microstructural and property relationships in β-titanium alloy Ti-5553 (Master's thesis), The Ohio State University, 201 West 19th Avenue, Columbus, OH 43210, 2010. B.Z. Jiang, S. Emura, K. Tsuchiya, Improvement of ductility in Ti-5Al-5Mo-5V-3Cr alloy by network-like precipitation of blocky α phase, Mater. Sci. Eng. A 722 (2018) 129–135. W. Xu, D.P. Edwards, X. Wu, M. Stoica, M. Calin, U. Kühn, J. Eckert, K. Xia, Scr. Mater. 68 (2013) 67–70. W. Xu, X. Wu, M. Stoica, M. Calin, U. Kühn, J. Eckert, K. Xia, Acta Mater. 68 (2012) 5067–5078. A. Zafari, K. Xia, Scr. Mater. 124 (2016) 151–154. B.Z. Jiang, K. Tsuchiya, S. Emura, X.H. Min, Mater. Trans. 55 (2014) 877–884. B.Z. Jiang, S. Emura, K. Tsuchiya, J. Alloy. Compd. 738 (2018) 283–291. L. Reinking, ImageJ basics, 2007. 〈https://imagej.nih.gov/ij/docs/pdfs/ImageJ. pdf〉. E. Underwood, Quantitative stereology for microstructural analysis, in: J. McCall, W. Mueller (Eds.), Microstructural Analysis, Springer, US, 1973, p. 47. W.C. Oliver, G.M. Pharr, Mater. Res. 7 (1992) 1564–1583. W.G. Burgers, Physica 1 (1934) 561–586. S. Nag, R. Banerjee, R. Srinivasan, J.Y. Hwang, M. Harper, H.L. Fraser, Acta Mater. 57 (2009) 2136–2147. T. Makino, R. Chikaizumi, T. Nagaoka, T. Furuhara, Mater. Sci. Eng. A 213 (1996)
Materials Science & Engineering A 731 (2018) 239–248
B. Jiang et al.
51–60. [26] T. Furuhara, S. Takagi, H. Watanabe, T. Maki, Metall. Mater. Trans. A 27 (1996) 1635–1646. [27] S. Zherebtsov, G. Salishchev, S.L. Semiatin, Philos. Mag. Lett. 90 (2010) 903–914. [28] X. Sauvage, F. Wetscher, P. Pareige, Acta Mater. 53 (2005) 2127–2135. [29] D. Setman, E. Schafler, E. Korznikova, M.J. Zehetbauer, Mater. Sci. Eng. A 493 (2008) 116–122. [30] S.V. Divinski, G. Reglitz, H. Rosner, Y. Estrin, G. Wilde, Acta Mater. 59 (2011) 1974–1985. [31] X.H. Min, L. Zhang, K. Sekido, T. Ohmura, S. Emura, K. Tsuchiya, K. Tsuzaki, Mater. Sci. Tech. 28 (2012) 342–347. [32] D.E. Gordon, J.W. Hagemeyer, J. Mater. Sci. 10 (1975) 1725–1731. [33] J.Z. Dong, F.G. Li, C.P. Wang, Mater. Sci. Eng. A 580 (2013) 105–113. [34] T.B. Britton, H. Liang, F.P.E. Dunne, A.J. Wilkinson, Proc. R. Soc. A (2010) 695–719. [35] C. Zambaldi, Y.Y. Yang, T.R. Bieler, D. Raabe, J. Mater. Res. 27 (2012) 356–367. [36] J. Kwon, M.C. Brandes, P. Sudharshan Phani, A.P. Pilchak, Y.F. Gao, E.P. George, G.M. Pharr, M.J. Mills, Acta Mater. 61 (2013) 4743–4756. [37] E. Lee, Microstructure evolution and microstructure/mechanical properties
[38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49]
248
relationship in α+β titanium alloys (Ph.D. thesis), The Ohio State University, 201 West 19th Avenue, Columbus, OH 43210, 2004. P. Zhang, S.X. Li, Z.F. Zhang, Mater. Sci. Eng. A 529 (2011) 62–73. E.O. Hall, Proc. Phys. Soc. Lond. B64 (1951) 747. N.J. Petch, J. Iron Steel Inst. 174 (1953) 25. N.G. Jones, M. Jackson, Mater. Sci. Tech. 27 (2011) 1025–1032. A.G. Paradkar, S.V. Kamat, A.K. Gogia, B.P. Kashyap, Mater. Sci. Eng. A 520 (2009) 168–173. K. Takeda, N. Nakada, T. Tsuchiyama, S. Takaki, ISIJ Int. 48 (2008) 1122–1125. S. Takaki, D. Akama, N. Nakada, T. Tsuchiyama, Mater. Trans. 55 (2014) 28–34. C.H. Caceres, Gemma E. Mann, J.R. Griffiths, Metall. Mater. Trans. A 42 (2011) 1950–1959. S.K. Vajpai, K. Amayama, M. Ota, T. Watanabe, R. Maeda, T. Sekiguchi, G. Dirras, A. Tingaud, IOP Conf. Ser.: Mater. Sci. Eng. 63 (2014) 012030. H. Fujiwara, R. Akada, A. Noro, Y. Yoshita, K. Ameyama, Mater. Trans. 49 (2008) 90–96. T. Sekiguchi, K. Ono, H. Fujiwara, K. Ameyama, Mater. Trans. 51 (2010) 39–45. D. Orlov, H. Fujiwara, K. Ameyama, Mater. Trans. 54 (2013) 1549–1553.