- Email: [email protected]
Role of processing parameters in the development of tri-modal microstructure during isothermal local loading forming of TA15 titanium alloy
Journal of Materials Processing Technology 239 (2017) 160–171
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Role of processing parameters in the development of tri-modal microstructure during isothermal local loading forming of TA15 titanium alloy P.F. Gao, X.G. Fan, H. Yang ∗ State Key Laboratory of Solidification Processing, School of Materials Science and Engineering, Northwestern Polytechnical University, P.O. Box 542, Northwestern Polytechnical University, Xi’an, 710072, PR China
a r t i c l e
i n f o
Article history: Received 12 January 2016 Received in revised form 1 July 2016 Accepted 11 August 2016 Available online 12 August 2016 Keywords: Isothermal local loading forming TA15 titanium alloy Processing parameters Tri-modal microstructure
a b s t r a c t It is critical to precisely control the tri-modal microstructure during the isothermal local loading forming of titanium alloy to obtain high-performance components. To this end, the effect of local loading processing parameters on the development of tri-modal microstructure were experimentally investigated. The key influence factors and laws are revealed as follows. In the first loading step, the deformation temperature plays a decisive role in the volume fraction of equiaxed ␣, which decreases with the increase of temperature. While, the deformation amount and cooling mode present little effects on the microstructure evolution. In the second loading step, the deformation temperature and amount mainly influence the volume fraction, spatial orientation distribution and globularization of lamellar ␣. The volume fraction of lamellar ␣ increases with the temperature decreasing. The spatial orientation distribution of lamellar ␣ gradually changes from homogeneous distribution to concentrated distribution with the increase of deformation amount. Besides, the dynamic globularization fraction of lamellar ␣ producing in the second loading step increases with the deformation amount, and their relationship can be well fitted by Avrami type equation. Moreover, higher temperature in the second loading step is beneficial to decrease the critical strain for the initiation of dynamic globularization and promote the kinetic rate of dynamic globularization. On the other hand, the deformation amount of the second loading step has after-effects on the static globularization of lamellar ␣ in the annealing treatment. If the deformation amount exceeded the critical strain for the initiation of dynamic globularization, the static globularization of lamellar ␣ would produce in the annealing, and the static globularization fraction increases with the deformation amount of the second loading step. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Titanium alloy large-scale complex components with features of high performance and light weight (such as bulkhead) are muchneeded structural parts for the advanced equipment in aviation and aerospace fields. However, the high deformation resistance of material, complex shape and large projection area of the structure make it very difficult to form such components. To overcome this problem, Yang et al. (2011) proposed the isothermal local loading forming technology that integrates the advantages of isothermal forming and local loading forming. This forming technology was named as ILLF by Zhou et al. (2011). During local loading forming, the load is applied to part of the billet and the component is formed
∗ Corresponding author. E-mail address: [email protected] (H. Yang). http://dx.doi.org/10.1016/j.jmatprotec.2016.08.015 0924-0136/© 2016 Elsevier B.V. All rights reserved.
by changing loading region, as shown in Fig. 1. This forming technology can enhance the plasticity of material, control the flow of material, reduce the forming load and enlarge the size of component to be formed, providing a feasible way to manufacture this kind of components. The large-scale complex components of titanium alloy often serve as key load bearing structures in severe conditions. Thus, not only the high quality of macroscopical forming but also the fine microstructure and mechanical properties are required in the forming process. Zhou et al. (2005) pointed out that the tri-modal microstructure, consisted of equiaxed ␣, lamellar ␣ and  transformed matrix, exhibits a good combination of strength, ductility, fracture toughness and fatigue life. The balanced mechanical properties make it a preferable microstructure morphology in the hot working of titanium alloy. However, besides the microstructure morphology, the microstructure parameters (such as the content, scale and distribution of each constituent phase) also play
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
161
Fig. 1. Illustration of local loading forming (Sun and Yang, 2009).
important roles in the final mechanical properties, as demonstrated by Lütjerin (1998). On the other hand, the isothermal local loading forming is a complicated multi-step hot working process with coupling effects of severe uneven deformation and complex temperature routes. The complex thermomechanical processing history may greatly influence the development of tri-modal microstructure as well as its microstructure parameters and mechanical properties. Consequently, in order to realize the precise control of tri-modal microstructure and properties, it is urgent to study the development mechanisms and rules of trimodal microstructure in the isothermal local loading forming of titanium alloy. Since the tri-modal microstructure of titanium alloy was first proposed by Zhou et al. (1996), some investigations have been conducted on its development during the integral forging and subsequent heat treatments. Lütjerin (1998) obtained a so called bi-lamellar microstructure of Ti-6Al-4V alloy, in which the single phase  lamellae are hardened by fine ␣ plates. This bi-lamellar microstructure is similar to the tri-modal microstructure. Zhou et al. (2005) reported the new near- forging process to produce tri-modal microstructure for titanium alloys, i.e. near- forging followed by rapid water-cooling, then high temperature toughening and low temperature strengthening heat treatments. The new near forging process have been successfully applied to manufacture the TC11 alloy compressor disk with tri-modal microstructure. Sun et al. (2014) quantitatively studied the evolution of volume fraction, average grain size and aspect ratio of equiaxed ␣ in the tri-modal microstructure during the near- forging process of TA15 alloy. Recently, Sun et al. (2016) put forward a new method to obtain the tri-modal microstructure in the hot working of TA15 alloy, i.e., the conventional forging combined with subsequent near- and two-phase field heat treatment. And, they investigated the effect of conventional forging conditions on the parameters of equiaxed and lamellar ␣ phases. As for the development of tri-modal microstructure during the isothermal local loading forming of titanium alloy, Gao et al. (2012) proposed that the tri-modal microstructure can be achieved through near- forging in the first loading step combined with conventional forging in the second loading step. Fan et al. (2012a) further found that the volume fraction and morphology of each constituent phase in the final tri-modal microstructure change with the deformation temperature and degrees in the above forming process. However, there is still a lack of quantitative analysis on the effect of local loading processing parameters on the development of tri-modal microstructure, which is necessary to achieve the precise control of tri-modal microstructure and mechanical properties. In this paper, the effects of processing parameters on the development of tri-modal microstructure during isothermal local loading of titanium alloy were quantitatively studied by staged experiment. It will provide a technological basis for the precise control of tri-modal microstructure in the isothermal local loading forming of titanium alloy large-scale complex components.
Fig. 2. Original microstructure of the billet.
Table 1 Local loading processing parameters. Deformation temperature (◦ C) Holding time (min) Reduction rate (%) Nominal strain rate (s−1 ) Cooling mode
910–980 15, 60 30, 50, 70 0.01 Air cooling (AC)Water quenching (WQ)
2. Experimental procedures The material used in this investigation is TA15 titanium alloy. Its chemical compositions are as follows (wt%), Al: 6.06; Mo: 2.08; V: 1.32; Zr: 1.86; Fe: 0.3 and Ti balance. Its -transus temperature is 990 ◦ C. The microstructure of the as-received material is equiaxed microstructure consisted of about 60% primary ␣ phase and  transformed matrix, as given in Fig. 2. The physical analogue experiment of isothermal local loading (Fig. 3) designed in the authors’ previous study (Fan et al., 2011) was conducted in this study. It can reflect the deformation characteristics of local loading forming. The local loading experiment is accomplished in one loading pass, which consists of two loading steps. In each loading step, the specimen was heated to the deformation temperature, deformed isothermally and then cooled down. Fig. 4 shows the change of specimen shape during experiment. After local loading, the specimens were annealed in the route of 810 ◦ C/1 h/AC. The local loading processing parameters in this work are given in Table 1. In this experiment, the effects of processing parameters on the development of tri-modal microstructure were studied in stages. At any stage, the specimen was quartered along the two symmetric planes and prepared for metallographic observation using standard technique. The detailed processing parameters and metallographic observation locations in each stage would be demonstrated in Section 3. The microstructure parameters under different conditions are quantitatively measured by Image-Pro Plus software as follows. The area fraction of each phase is measured as its volume fraction. The size of equiaxed ␣ is measured by the average length of diameters measured at 2 ◦ intervals and passing through the grain’s centroid, called as mean diameter (dmean ). The aspect ratio of each phase is represented by the ratio between major axis and minor axis of ellipse equivalent to object. And the spatial orientation of lamellar ␣ is evaluated by the angle between the major axis of object and the vertical, i.e. the compression direction. For each microstructure parameter, the average value of large numbers of objects is adopted as the final result. According to the basic stereological procedure, i.e., the area fraction (AA ) and volume fraction (VV ) are equivalent, it is reasonable to represent the volume frac-
162
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 3. The physical analogue experiment of local loading forming (Fan et al., 2011): (a) the first loading step; (b) the second loading step; (c) the processing route.
Fig. 4. The change of specimen shape in the physical analogue experiment of local loading forming (Fan et al., 2011): (a) initial specimen; (b) after the first loading step; (c) after the second loading step.
tion of each phase in three-dimensional space by the area fraction in two-dimensional micrograph. In addition, the mean diameter can provide a direct estimation of the grain size of equiaxed ␣ particle. Indeed, for size of true spheres (Dsphere ), Dsphere = 4/dmean . However, for simplicity and because the true three-dimensional shape is currently unknown, the factor 4/ is omitted in this work. The same treatment methods have been applied in the microstructure analysis of ␣+ titanium alloys by Collins et al. (2009). The aspect ratio and orientation are used to depict the morphology and spatial orientation of particle. There is still a lack of strong evidence for relating them to specific three-dimensional morphology features. However, they have been widely used to evaluate the morphology and orientation of lamellar ␣ in the hot working of titanium alloy with lamellar microstructure, as reported by Park et al. (2014) and Bieler and Semiatin (2002). Thus, the aspect ratio and spatial orientation of lamellar ␣ are also used in this work. To acquire the local effective strains on the specimen for microstructure analysis, the isothermal local loading experiment was simulated using DEFORM-3D in previous study (Gao et al. (2011)). The finite element (FE) model was developed in isothermal condition neglecting any thermal events. The constant shear model is employed and the friction factor is determined to be 0.4. For the element meshing, the tetrahedral element is adopted and the automatic remeshing techniques of DEFORM-3D are applied. The flow behavior of material was input in the form of discrete points based on the results in Shen (2007). The FE model has been validated in the work of Gao et al. (2011), which will be used to acquire the local effective strains for microstructure analysis in this work. The effective strain is a common value to evaluate the deformation degree in the metal forming analysis, which is defined as follows: √ ε¯ =
2 (ε1 − ε2 )2 + (ε2 − ε3 )2 + (ε3 − ε1 )2 3
(1)
where, ε¯ is the effective strain, and ε1 , ε2 and ε3 are principal strains.
Fig. 5. Microstructure observation locations for sample after the first loading step.
3. Results and discussion 3.1. Microstructure evolution in the first loading step In the first loading step, the effects of deformation temperature (T1 ), deformation amount and cooling mode on the microstructure are mainly concerned. The detailed processing parameters are T1 of 910, 930, 950 and 970 ◦ C, 15 min holding, 50% reduction, and the cooling mode of AC and WQ. Microstructures of Region A and B in Fig. 5 are compared to study the effect of deformation amount. FE simulation suggests that the effective strains in Region A and B are 0.98 and 0.05, respectively. Figs. 6 and 7 show the microstructures of Region A and B deformed to 50% reduction at diverse temperatures and air cooled in the first loading step, respectively. The microstructures at diverse temperatures all consist of equiaxed ␣ and  transformed matrix. It can be found that the microstructures of Region A and B present the
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
163
Fig. 6. Microstructures of Region A deformed to 50% reduction at diverse temperatures and air cooled in the first loading step: (a) 910 ◦ C; (b) 930 ◦ C; (c) 950 ◦ C; (d) 970 ◦ C.
Fig. 7. Microstructures of Region B deformed to 50% reduction at diverse temperatures and air cooled in the first loading step: (a) 910 ◦ C; (b) 930 ◦ C; (c) 950 ◦ C; (d) 970 ◦ C.
Table 2 The microstructure parameters of Region A and B at different conditions in the first loading step. Sample
◦
910 C, AC 930 ◦ C, AC 950 ◦ C, AC 970 ◦ C, AC 970 ◦ C, WQ
Volume fraction of ␣p (%)
Grain size of ␣p (m)
Aspect ratio of ␣p
Region A
Region B
Region A
Region B
Region A
Region B
49.17 44.16 35.53 19.06 18.85
51.81 44.90 32.94 20.03 19.37
9.54 9.55 9.08 9.02 8.06
10.14 9.79 8.98 8.28 8.12
3.04 2.84 2.50 2.40 2.37
2.35 2.70 2.25 2.57 2.48
Note: ␣p represents equiaxed ␣.
close morphology at each temperature. Their volume fraction, grain size and aspect ratio of equiaxed ␣ are also very close, as shown in Table 2. We can also found from Table 2 that the volume fraction of equiaxed ␣ decreases notably with the temperature increasing, while the grain size and aspect ratio of equiaxed ␣ change little. To compare the microstructures at different cooling modes, the sample at 970 ◦ C was also quenched in water after deformation. The corresponding microstructure image and parameters are given in Fig. 8 and Table 2, respectively. The microstructure after water quenching is composed of equiaxed ␣ and metastable martensitic, which is distinct from that cooled in air (Figs. 6(d) and 7(d)). While, the volume fraction, grain size and aspect ratio of equiaxed ␣ all change little with the cooling mode, as shown in Table 2. 3.2. Microstructure evolution in the second loading step Previous study (Gao et al., 2012) indicates that the lamellar ␣ generates in the heating and holding stage of the second loading step, which is a key process for the development of tri-modal microstructure. Thus, the second loading step was divided into two stages, i.e., the heating and holding stage and deformation stage. Here, the temperature of the first loading step was fixed at 970 ◦ C. 3.2.1. Heating and holding stage First, we checked whether the deformation amount and cooling mode of the first loading step have after-effects on the microstructure evolution in the second loading step. To this end, two samples deformed to 50% reduction underwent different cooling modes (AC and WQ) in the first loading step were firstly prepared. Then, they were hated to 950 ◦ C, hold 15 min, and water quenched to reserve the high temperature microstructure, as shown in Fig. 9. It can be
found that the microstructures underwent different deformation amounts and cooling modes all produce lamellar ␣ and present the same morphology. Besides, the quantitative analyses (Table 3) show that the parameters of equiaxed ␣ and lamellar ␣ are all very close at different conditions. It can be concluded that the deformation amount and cooling mode of the first loading step have little after-effects on the subsequent microstructure evolution. Fig. 10 shows the waster-quenched microstructures after holding at different temperatures and time, which all consist of equiaxed ␣, lamellar ␣ and  phase. It can be found the microstructures hold 60 min (Fig. 10(e),(f)) and 15 min (Fig. 9(c),(d)) have the same morphology and close parameters of equiaxed ␣ and lamellar ␣ (Tables 3 and 4). This indicates that the microstructure evolution changes little with the holding time in this work. Table 4 lists the microstructure parameters after heating and holding at various temperatures. We can find that the temperature mainly influences the volume fraction of lamellar ␣. Increasing the temperature would decrease the volume fraction of lamellar ␣ significantly. 3.2.2. Deformation stage The effects of deformation amount and temperature are mainly concerned in the deformation stage. To this end, the samples were deformed to various reduction rate (30%, 50% and 70%) at 930 and 950 ◦ C in the second loading step. After deformation, microstructure observation was carried out in three typical regions (Region A–C), as shown in Fig. 11. This because Region A–C could reflect the deformation characteristics of the first-loading region, transitional region, and second-loading region, respectively. Table 5 gives the FE simulated effective strains at three typical regions in two loading steps.
164
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 8. Microstructures of sample deformed to 50% reduction at 970 ◦ C and quenched by water in the first loading step: (a) Region A; (b) Region B.
Fig. 9. Microstructures after heating and holding in the second loading step with different cooling modes in the first loading step: (a) AC, Region A; (b) AC, Region B; (c) WQ, Region A; (d) WQ, Region B. Table 3 The microstructure parameters of microstructures in Fig. 9. Cooling mode
Region
AC
A B A B
WQ
Parameters of ␣p
Parameters of ␣l
Volume fraction (%)
Grain size (m)
Aspect ratio
Volume fraction (%)
Aspect ratio
21.79 19.45 22.79 19.95
8.21 8.80 8.03 8.06
2.04 1.87 2.03 1.92
13.25 14.86 12.71 13.18
5.01 5.50 5.24 5.06
Note: ␣l represents lamellar ␣. Table 4 The microstructure parameters after heating and holding at different temperatures in the second loading step. Temperature (◦ C)
930 940 950
Region
A B A B A B
Parameters of ␣p
Parameters of ␣l
Volume fraction (%)
Grain size (m)
Aspect ratio
Volume fraction (%)
Aspect ratio
21.25 22.67 19.06 20.20 21.79 19.45
8.95 8.41 8.53 8.68 8.21 8.80
1.75 2.27 2.07 2.06 2.04 1.87
28.56 29.14 22.33 21.72 13.25 14.86
5.59 5.53 5.04 5.42 5.01 5.50
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
165
Fig. 10. Water-quenched microstructures of Region A and B after different heating and holding treatments in the second loading step: (a) 930 ◦ C, 15 min, Region A; (b) 930 ◦ C, 15 min, Region B; (c) 940 ◦ C, 15 min, Region A; (d) 940 ◦ C, 15 min, Region B; (e) 950 ◦ C, 60 min, Region A; (f) 950 ◦ C, 60 min, Region B.
Fig. 11. Microstructure observation locations for samples at various reductions: (a) 30%; (b) 50%; (c) 70%.
Table 5 Effective strains at microstructure observation locations in Fig. 11. Reduction rate (%)
Strains in step 1 Region A
Region B
Region C
Region A
Region B
Region C
30 50 70
0.52 0.98 1.53
0.26 0.56 0.66
0.10 0.26 0.22
0.10 0.26 0.40
0.26 0.49 0.84
0.44 0.86 1.29
Strains in step 2
166
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 12. Microstructures of samples deformed to different reductions at 930 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 50%; (g)–(i) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
The microstructures of samples deformed to various reductions at 930 ◦ C are shown in Fig. 12. It can be seen that the morphology and distribution of equiaxed ␣ change little with the loading region and reduction. Moreover, the quantitative analyses shows that the volume fraction, grain size and aspect ratio of equiaxed ␣ all locate in narrow ranges, i.e., 21.89 ± 1.17%, 8.87 ± 0.33 m and 2.02 ± 0.15, respectively. Semiatin et al. (2002) pointed out that the  phase is softer and easier to deform than equiaxed ␣ during the hot working of ␣ +  titanium alloy in two phase region. Considering the morphology and distribution characteristics of lamellar ␣ in  matrix, it can be speculated that the lamellar ␣ deforms more easily than the equiaxed ␣. During deformation, the lamellar ␣ would gradually rotate themselves perpendicular to compression direction, and produce kinking and break-up with the increase of deformation amount. So the lamellar ␣ at different loading regions present various morphology and spatial orientation distribution after undergoing different deformation amounts, as shown in Fig. 12. The quantitative results of spatial orientation distribution of lamellar ␣ under different conditions are given in Fig. 13. The result of “ = 0” represents the microstructure prior to deformation (Fig. 10(a) and (b)). From the results at 30% reduction (Fig. 13(a)), it can be found that Region A presents the similar distribution as that prior to deformation. While, the distributions of Region B and C change a little due to their greater strain. Their fraction of lamellar ␣ with smaller title angle decreases, while the fraction with
greater title angles increases. With the reduction rate increasing to 50% and 70% (Fig. 13(b) and (c)), the strain at each loading region, especially Region B and C, increases greatly, which leads to more notable change in the spatial orientation distribution of lamellar ␣. It’s worth noting that Region B and C at 70% reduction present very close distribution, although there exists a big difference on their strains. From the above analyses, it can be found that increasing the deformation amount would aggravate the rotation of lamellar ␣. This makes the spatial orientation distribution of lamellar ␣ changing from homogeneous distribution to concentrated distribution gradually. In addition, it can be concluded that the distribution change begins at a strain lower than 0.49, and mainly produces at strains between 0.49 and 0.84, while almost stops after strain being greater than 0.84. These phenomena are similar to that reported by Park et al. (2014) in the hot deformation of Ti-6Al-4V alloy with lamellar microstructure. Besides the change of orientation distribution, dynamic globularization is another important phenomenon of lamellar ␣ during deformation, as shown in Fig. 12. The dynamic globularization fraction of lamellar ␣ at different conditions are quantitatively measured and given in Fig. 14. In this work, globularization was taken to be an ␣-phase with aspect ratio less than 2.5, and the globularization fraction was calculated by:
fg =
f2.5after − f2.5pre 1 − f2.5pre
(2)
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
167
Fig. 13. Spatial orientation distribution of lamellar ␣ traces relative to the compression axis for samples deformed to different reductions at 930 ◦ C in the second loading step: (a) 30%; (b) 50%; (c) 70%.
Fig. 14. Dynamic globularization fraction of lamellar ␣ for samples deformed to different reductions at 930 ◦ C in the second loading step.
where, fg denotes the globularization fraction, f2.5pre and f2.5after represent the volume fractions of small equiaxed ␣ (aspect ratio <2.5) before and after deformation. It can be seen from Fig. 14 that the deformation amount has significant effect on the dynamic globularization fraction of lamellar ␣. At smaller deformation amount (Region A and B at 30% reduction), the globularization doesn’t occur. This rule is the same as that in the hot working of two phase titanium alloy with lamellar microstructure reported by Wang et al. (2010). They proposed that the globularization of a lamellae may be rationalized to consist of two processes: the break-up of lamellae and the formation
of globules. The break-up of lamellae firstly needs the nativity of intraphase boundary, which usually occurs at kinks or shearing bands of lamellae. So sufficient strain must be imposed, that is to say, there exists a small critical strain for the initiation of dynamic globularization. Once the globularization beginning, the dynamic globularization fraction increases gradually with the deformation amount (Fig. 14), which can also be clearly observed in the microstructure morphology (Fig. 12). To study the effect of temperature, the microstructures of samples deformed at 950 ◦ C in the second loading step are given in Fig. 15. Comparing Figs. 15 and 12, it can be seen that the volume fraction of lamellar ␣ at 950 ◦ C is much smaller than that at 930 ◦ C, but the morphology and spatial orientation distribution are very similar. This is also confirmed by the comparison on the quantitative results of spatial orientation distribution at 950 ◦ C (Fig. 16) and 930 ◦ C (Fig. 13). Therefore, it can be said that the spatial orientation distribution of lamellar ␣ is little influenced by the temperature in the second loading step. Song et al. (2009) found that in the hot working of titanium alloy with lamellar microstructure, the kinetic of dynamic globularization experiences three stages with the increase of deformation amount: the induction stage, the accelerated globularization stage and the decelerated globularization stage. The induction stage corresponds to the critical strain required for the initiation of dynamic globularization. As a result, the variation of globularization fraction with strain presents an S-type curve starting at the critical strain. Meanwhile, they also found that the Avrami type equation (Eq. (3)) can properly depict the three stages variation process.
fdg = 1 − exp[−k · (ε − εc )n ]
(3)
168
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 15. Microstructures of samples deformed to different reductions at 950 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 50%; (g)–(i) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
Table 6 Fitted parameters. Temperature (◦ C)
εc
k
n
930 950
0.320 0.249
0.590 0.712
1.230 1.624
where, fdg is the dynamic globularization fraction, εc is the critical strain for the initiation of dynamic globularization, k is the temperature-dependent kinetic constant, and n is the Avrami exponent. It not only can reflect the increase trend of globularization fraction with strain, but also can catch the critical strain. Moreover, the effect of temperature on the globularization kinetic can also be reflected by the parameter k. Therefore, the Avrami type equation is widely used to fit the kinetic of dynamic globularization during hot working of titanium alloy with lamellar microstructure, as reported by Wang et al. (2010) and Ma et al. (2012). In this work, although the initial microstructure is different from the fully lamellar microstructure reported above, the dynamic globularization process of lamellar ␣ is the same. It also undergoes the induction and globularization stages, so the Avrami type equation (Eq. (3)) is also trialed in fitting the variation of dynamic globularization fraction of lamellar ␣ with strain at different temperatures (Fig. 17). It can be observed that the fitting results agree well with the experimental results at each temperature. The fitted parameters are listed in Table 6.
From Fig. 17, it can be found that the curve of 950 ◦ C has smaller critical strain and greater dynamic globularization fraction at the same strain than that of 930 ◦ C. As mentioned above, the critical strain is related to the break-up of lamellae ␣ which is mainly achieved by two ways: intense shearing and penetration of  phase along the interfaces. The first way is determined by the deformation, while the second way is a temperature-dependent thermally activated process. At higher temperature, the volume fraction of lamellar ␣ is smaller so that each lamellar ␣ would bear more deformation, which is in favor of the first way. Meanwhile, higher temperature could promote the diffusion process, which is beneficial to the second way. Thus, the higher temperature would promote the break-up of lamellae ␣, leading to smaller critical strain. On the other side, the greater dynamic globularization fraction at higher temperature can be explained as follows. Even the emergence of the interfaces is indispensable, the migration rate of intraphase boundary, as the final step for dynamic globularization, restricts the kinetic rate of dynamic globularization. Higher temperature is beneficial to the diffusion and interface migration, so the dynamic globularization fraction increases with temperature.
3.3. Microstructure evolution in the annealing After local loading, the samples were annealed in the route of 810 ◦ C/1 h/AC. Fig. 18 shows the annealed microstructures of samples deformed at 970 and 930 ◦ C sequentially in two loading steps.
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
169
Fig. 16. Spatial orientation distribution of lamellar ␣ traces relative to the compression axis for samples deformed to different reductions at 950 ◦ C in the second loading step: (a) 30%; (b) 50%; (c) 70%.
Table 7 Comparisons of globularization fraction of lamellar ␣ before and after annealing for samples deformed to different reductions. Reduction rate(%)
State
30
Prior to annealing After annealing Prior to annealing After annealing
70
Fig. 17. Dynamic globularization fraction of lamellar ␣ vs. strain in the second loading step (dot: experimental data, line: fitted results using Eq. (3).
It can be found that the microstructure morphology change little in comparison with those before annealing (Fig. 12(a)–(c) and (g)–(i)). The quantitative analyses of microstructures show that the volume fractions of equiaxed ␣ and lamellar ␣ increase by about 3% and 6.5% in the annealing, respectively. This is caused by the precipitation of ␣ phase and the mergence of fine ␣ phase on the equiaxed ␣ and lamellar ␣. Moreover, the increments of contents of equiaxed ␣ and lamellar ␣ are very close at different loading regions and reductions. Further static globularization of lamellar ␣ in the annealing process was also found when comparing the microstructures before
Globularization fraction Region A
Region B
Region C
0 0 0.021 0.030
0 0 0.229 0.286
0.041 0.048 0.432 0.525
and after annealing (Figs. 12 and 18). Fan et al. (2012b) studied the static globularization behavior of TA15 alloy with lamellar microstructure and found that the globularization kinetics rate increases with pre-deformation. Thus, the effect of deformation amount of the second loading step on the static globularization of lamellar ␣ was also explored in this work. Table 7 gives the comparisons of globularization fraction of lamellar ␣ before and after annealing for samples deformed to different reductions. It can be found that the increments of globularization fraction are very small at regions underwent small deformation (Region A–C at 30% reduction, Region A at 70% reduction), but are greater at regions underwent greater deformation (Region B and C at 70% reduction). These suggest that greater deformation amount in the second loading step could promote the static globularization of lamellar ␣ in the annealing, presenting the same rule as that proposed by Fan et al. (2012b). The annealed microstructures of samples deformed at 950 ◦ C in the second loading step (Fig. 19) were also analyzed to study the after-effect of temperature of the second loading step. The comparative quantitative analyses suggest that the temperature in the second loading step has little after-effect on the rules of content
170
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 18. Microstructures after annealing for samples deformed to different reductions at 930 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
Fig. 19. Microstructures after annealing for samples deformed to different reductions at 950 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
increments of each phase and the static globularization of lamellar ␣ in the annealing. Meanwhile, we found that only if the deformation amount in the second loading step exceeded the critical strain for the initiation of dynamic globularization, obvious static globularization would occur. In other words, the occurring of dynamic globularization in second loading step can be approximately taken as the precondition to the static globularization in annealing. 3.4. Mechanical properties of typical microstructures The tensile stress-strain curves and mechanical properties of typical tri-modal microstructures are shown in Fig. 20. Microstructures A and B have the close equiaxed ␣ content but different
lamellar ␣ content. Comparing their mechanical properties, we can found that the yield strength (R0.2 ), ultimate tensile strength (Rm ) and reduced area (Z) of microstructure A are all greater than microstructure B. While, the elongation (A) of microstructure A is a little smaller than microstructure B. Another hand, microstructure C has the close lamellar ␣ content but greater equiaxed ␣ content in comparison with microstructure B. As for the mechanical properties, microstructure C presents the close R0.2 and smaller Rm compared to microstructure B. Nevertheless, both of A and Z of microstructure C are much greater than microstructure B. It can be concluded that the parameters of tri-modal microstructure play great roles in the mechanical properties. To control the microstructure and mechanical properties, it is very critical to study the effect
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
171
Acknowledgements The authors would like to gratefully acknowledge the support of National Natural Science Foundation of China (No. 51605388, 51575449), 111 Project (B08040), Project supported by the Research Fund of the State Key Laboratory of Solidification Processing (NWPU), China (Grant No. 131-QP-2015), and the Open Research Fund of State Key Laboratory of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology. References
Fig. 20. The tensile stress-strain curves and mechanical properties of typical microstructures.
of processing conditions on the microstructure and the relationship between microstructure parameters and properties. The former research content is right the focus of this work, and the later one is being studied. 4. Conclusions In this work, the development of tri-modal microstructure during the isothermal local loading of TA15 titanium alloy were quantitatively studied through a physical analogue experiment. It was found that the local loading processing parameters play critical role in the development of tri-modal microstructure as follows: 1. The volume fraction of equiaxed ␣ in tri-modal microstructure is mainly determined by the deformation temperature of the first loading step. Increasing the temperature would decrease the volume fraction of equiaxed ␣. While, the volume fraction of lamellar ␣ increases with the decrease of temperature of the second loading step, when the temperature of the first loading step is fixed at a higher temperature. 2. The spatial orientation distribution and globularization of lamellar ␣ are strongly dependent on the deformation amount and temperature of the second loading step. With the increasing of deformation amount, the spatial orientation distribution of lamellar ␣ changes from homogeneous distribution to concentrated distribution gradually, and the dynamic globularization fraction of lamellar ␣ approximately increases in a sigmoid way. The relationship between dynamic globularization and deformation amount can be well fitted by the Avrami type equation. Besides, increasing the temperature of the second loading step would reduce the critical strain for the initiation of dynamic globularization and enhance the kinetics rate of dynamic globularization. 3. The deformation amount of the second loading step has aftereffects on the static globularization of lamellar ␣ in the annealing treatment. If the deformation amount exceeded the critical strain for the initiation of dynamic globularization in the second loading step, lamellar ␣ would produce static globularization in the annealing. And the static globularization fraction increases with the deformation amount of the second loading step.
Bieler, T.R., Semiatin, S.L., 2002. The origins of heterogeneous deformation during primary hot working of Ti-6Al-4V. Int. J Plast. 18 (9), 1165–1189. Collins, P.C., Welk, B., Searles, T., Tiley, J., Russd, J.C., Fraser, H.L., 2009. Development of methods for the quantification of microstructural features in ␣ + -processed ␣/ titanium alloys. Mater. Sci. Eng. A 508, 174–182. Fan, X.G., Gao, P.F., Yang, H., 2011. Microstructure evolution of the transitional region in isothermal local loading of TA15 titanium alloy. Mater. Sci. Eng. A 528 (6), 2694–2703. Fan, X.G., Yang, H., Gao, P.F., Yan, S.L., 2012a. Dependence of microstructure morphology on processing in subtransus isothermal local loading forming of TA15 titanium alloy. Mater. Sci. Eng. A 546, 46–52. Fan, X.G., Yang, H., Yan, S.L., Gao, P.F., Zhou, J.H., 2012b. Mechanism and kinetics of static globularization in TA15 titanium alloy with transformed structure. J. Alloy Compd. 533, 1–8. Gao, P.F., Yang, H., Fan, X.G., 2011. Quantitative analysis of the microstructure of transitional region under multi-heat isothermal local loading forming of TA15 titanium alloy. Mater. Des. 32 (4), 2012–2020. Gao, P.F., Yang, H., Fan, X.G., Yan, S.L., 2012. Microstructural features of TA15 titanium alloy under different temperature routes in isothermal local loading forming. Mater. Sci. Eng. A 540 (1), 245–252. Lütjerin, G., 1998. Influence of processing on microstructure and mechanical properties of (␣ + ) titanium alloy. Mater. Sci. Eng. A 243 (1–2), 32–45. Ma, X., Zeng, W.D., Tian, F., Zhou, Y.G., 2012. The kinetics of dynamic globularization during hot working of a two phase titanium alloy with starting lamellar microstructure. Mater. Sci. Eng. A. 548, 6–11. Park, C.H., Kim, J.H., Hyun, Y.T., Yeom, J.T., Reddy, N.S., 2014. The origins of flow softening during high-temperature deformation of a Ti-6Al-4V alloy with a lamellar microstructure. J. Alloy Compd. 582, 126–129. Semiatin, S.L., Montheillet, F.G., Shen, G., Jonas, J.J., 2002. Self-consistent modeling of the flow behavior of wrought alpha/beta titanium alloys under isothermal and nonisothermal hot-working conditions. Metall. Mater. Trans. A 33, 2719–2727. Shen, C.W., 2007. Research on material constitution models of TA15 and TC11 titanium alloys in hot deformation process. Master Thesis, Northwestern Polytechnical University. Song, H.W., Zhang, S.H., Cheng, M., 2009. Dynamic globularization kinetics during hot working of a two phase titanium alloy with a colony alpha microstructure. J. Alloy Compd. 480 (2), 922–927. Sun, Z.C., Yang, H., 2009. Microstructure and mechanical properties of TA15 titanium alloy under multi-step local loading forming. Mater. Sci. Eng. A 523 (1–2), 184–192. Sun, Z.C., Wang, X.Q., Zhang, J., Yang, H., 2014. Prediction and control of equiaxed ␣ in near- forging of TA15 Ti-alloy based on BP neural network: for purpose of tri-modal microstructure. Mater. Sci. Eng. A 591, 18–25. Sun, Z.C., Han, F.X., Wu, H.L., Yang, H., 2016. Tri-modal microstructure evolution of TA15 Ti-alloy under conventional forging combined with given subsequent heat treatment. J. Mater. Process. Technol. 229, 72–81. Wang, K.X., Zeng, W.D., Zhao, Y.Q., Lai, Y.J., Zhou, Y.G., 2010. Dynamic globularization kinetics during hot working of Ti-17 alloy with initial lamellar microstructure. Mater. Sci. Eng. A 527 (10–11), 2559–2566. Yang, H., Fan, X.G., Sun, Z.C., Guo, L.G., Zhan, M., 2011. Recent developments in plastic forming technology of titanium alloys. Sci. China-Technol. Sci. 54 (2), 490–501. Zhou, Y.G., Zeng, W.D., Yu, H.Q., 1996. A new high-temperature deformation strengthening and toughening process for titanium alloys. Mater. Sci. Eng. A 221, 58–62. Zhou, Y.G., Zeng, W.D., Yu, H.Q., 2005. An investigation of a new near-beta forging process for titanium alloys and its application in aviation components. Mater. Sci. Eng. A 393 (1–2), 204–212. Zhou, L., Chang, H., Wang, X.D., 2011. Recent progress of titanium industry, research and development in China. Proceedings of the 12th World Conference on Titanium. Beijing, China, 56–61 (Plenary Lectures).
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Role of processing parameters in the development of tri-modal microstructure during isothermal local loading forming of TA15 titanium alloy P.F. Gao, X.G. Fan, H. Yang ∗ State Key Laboratory of Solidification Processing, School of Materials Science and Engineering, Northwestern Polytechnical University, P.O. Box 542, Northwestern Polytechnical University, Xi’an, 710072, PR China
a r t i c l e
i n f o
Article history: Received 12 January 2016 Received in revised form 1 July 2016 Accepted 11 August 2016 Available online 12 August 2016 Keywords: Isothermal local loading forming TA15 titanium alloy Processing parameters Tri-modal microstructure
a b s t r a c t It is critical to precisely control the tri-modal microstructure during the isothermal local loading forming of titanium alloy to obtain high-performance components. To this end, the effect of local loading processing parameters on the development of tri-modal microstructure were experimentally investigated. The key influence factors and laws are revealed as follows. In the first loading step, the deformation temperature plays a decisive role in the volume fraction of equiaxed ␣, which decreases with the increase of temperature. While, the deformation amount and cooling mode present little effects on the microstructure evolution. In the second loading step, the deformation temperature and amount mainly influence the volume fraction, spatial orientation distribution and globularization of lamellar ␣. The volume fraction of lamellar ␣ increases with the temperature decreasing. The spatial orientation distribution of lamellar ␣ gradually changes from homogeneous distribution to concentrated distribution with the increase of deformation amount. Besides, the dynamic globularization fraction of lamellar ␣ producing in the second loading step increases with the deformation amount, and their relationship can be well fitted by Avrami type equation. Moreover, higher temperature in the second loading step is beneficial to decrease the critical strain for the initiation of dynamic globularization and promote the kinetic rate of dynamic globularization. On the other hand, the deformation amount of the second loading step has after-effects on the static globularization of lamellar ␣ in the annealing treatment. If the deformation amount exceeded the critical strain for the initiation of dynamic globularization, the static globularization of lamellar ␣ would produce in the annealing, and the static globularization fraction increases with the deformation amount of the second loading step. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Titanium alloy large-scale complex components with features of high performance and light weight (such as bulkhead) are muchneeded structural parts for the advanced equipment in aviation and aerospace fields. However, the high deformation resistance of material, complex shape and large projection area of the structure make it very difficult to form such components. To overcome this problem, Yang et al. (2011) proposed the isothermal local loading forming technology that integrates the advantages of isothermal forming and local loading forming. This forming technology was named as ILLF by Zhou et al. (2011). During local loading forming, the load is applied to part of the billet and the component is formed
∗ Corresponding author. E-mail address: [email protected] (H. Yang). http://dx.doi.org/10.1016/j.jmatprotec.2016.08.015 0924-0136/© 2016 Elsevier B.V. All rights reserved.
by changing loading region, as shown in Fig. 1. This forming technology can enhance the plasticity of material, control the flow of material, reduce the forming load and enlarge the size of component to be formed, providing a feasible way to manufacture this kind of components. The large-scale complex components of titanium alloy often serve as key load bearing structures in severe conditions. Thus, not only the high quality of macroscopical forming but also the fine microstructure and mechanical properties are required in the forming process. Zhou et al. (2005) pointed out that the tri-modal microstructure, consisted of equiaxed ␣, lamellar ␣ and  transformed matrix, exhibits a good combination of strength, ductility, fracture toughness and fatigue life. The balanced mechanical properties make it a preferable microstructure morphology in the hot working of titanium alloy. However, besides the microstructure morphology, the microstructure parameters (such as the content, scale and distribution of each constituent phase) also play
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
161
Fig. 1. Illustration of local loading forming (Sun and Yang, 2009).
important roles in the final mechanical properties, as demonstrated by Lütjerin (1998). On the other hand, the isothermal local loading forming is a complicated multi-step hot working process with coupling effects of severe uneven deformation and complex temperature routes. The complex thermomechanical processing history may greatly influence the development of tri-modal microstructure as well as its microstructure parameters and mechanical properties. Consequently, in order to realize the precise control of tri-modal microstructure and properties, it is urgent to study the development mechanisms and rules of trimodal microstructure in the isothermal local loading forming of titanium alloy. Since the tri-modal microstructure of titanium alloy was first proposed by Zhou et al. (1996), some investigations have been conducted on its development during the integral forging and subsequent heat treatments. Lütjerin (1998) obtained a so called bi-lamellar microstructure of Ti-6Al-4V alloy, in which the single phase  lamellae are hardened by fine ␣ plates. This bi-lamellar microstructure is similar to the tri-modal microstructure. Zhou et al. (2005) reported the new near- forging process to produce tri-modal microstructure for titanium alloys, i.e. near- forging followed by rapid water-cooling, then high temperature toughening and low temperature strengthening heat treatments. The new near forging process have been successfully applied to manufacture the TC11 alloy compressor disk with tri-modal microstructure. Sun et al. (2014) quantitatively studied the evolution of volume fraction, average grain size and aspect ratio of equiaxed ␣ in the tri-modal microstructure during the near- forging process of TA15 alloy. Recently, Sun et al. (2016) put forward a new method to obtain the tri-modal microstructure in the hot working of TA15 alloy, i.e., the conventional forging combined with subsequent near- and two-phase field heat treatment. And, they investigated the effect of conventional forging conditions on the parameters of equiaxed and lamellar ␣ phases. As for the development of tri-modal microstructure during the isothermal local loading forming of titanium alloy, Gao et al. (2012) proposed that the tri-modal microstructure can be achieved through near- forging in the first loading step combined with conventional forging in the second loading step. Fan et al. (2012a) further found that the volume fraction and morphology of each constituent phase in the final tri-modal microstructure change with the deformation temperature and degrees in the above forming process. However, there is still a lack of quantitative analysis on the effect of local loading processing parameters on the development of tri-modal microstructure, which is necessary to achieve the precise control of tri-modal microstructure and mechanical properties. In this paper, the effects of processing parameters on the development of tri-modal microstructure during isothermal local loading of titanium alloy were quantitatively studied by staged experiment. It will provide a technological basis for the precise control of tri-modal microstructure in the isothermal local loading forming of titanium alloy large-scale complex components.
Fig. 2. Original microstructure of the billet.
Table 1 Local loading processing parameters. Deformation temperature (◦ C) Holding time (min) Reduction rate (%) Nominal strain rate (s−1 ) Cooling mode
910–980 15, 60 30, 50, 70 0.01 Air cooling (AC)Water quenching (WQ)
2. Experimental procedures The material used in this investigation is TA15 titanium alloy. Its chemical compositions are as follows (wt%), Al: 6.06; Mo: 2.08; V: 1.32; Zr: 1.86; Fe: 0.3 and Ti balance. Its -transus temperature is 990 ◦ C. The microstructure of the as-received material is equiaxed microstructure consisted of about 60% primary ␣ phase and  transformed matrix, as given in Fig. 2. The physical analogue experiment of isothermal local loading (Fig. 3) designed in the authors’ previous study (Fan et al., 2011) was conducted in this study. It can reflect the deformation characteristics of local loading forming. The local loading experiment is accomplished in one loading pass, which consists of two loading steps. In each loading step, the specimen was heated to the deformation temperature, deformed isothermally and then cooled down. Fig. 4 shows the change of specimen shape during experiment. After local loading, the specimens were annealed in the route of 810 ◦ C/1 h/AC. The local loading processing parameters in this work are given in Table 1. In this experiment, the effects of processing parameters on the development of tri-modal microstructure were studied in stages. At any stage, the specimen was quartered along the two symmetric planes and prepared for metallographic observation using standard technique. The detailed processing parameters and metallographic observation locations in each stage would be demonstrated in Section 3. The microstructure parameters under different conditions are quantitatively measured by Image-Pro Plus software as follows. The area fraction of each phase is measured as its volume fraction. The size of equiaxed ␣ is measured by the average length of diameters measured at 2 ◦ intervals and passing through the grain’s centroid, called as mean diameter (dmean ). The aspect ratio of each phase is represented by the ratio between major axis and minor axis of ellipse equivalent to object. And the spatial orientation of lamellar ␣ is evaluated by the angle between the major axis of object and the vertical, i.e. the compression direction. For each microstructure parameter, the average value of large numbers of objects is adopted as the final result. According to the basic stereological procedure, i.e., the area fraction (AA ) and volume fraction (VV ) are equivalent, it is reasonable to represent the volume frac-
162
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 3. The physical analogue experiment of local loading forming (Fan et al., 2011): (a) the first loading step; (b) the second loading step; (c) the processing route.
Fig. 4. The change of specimen shape in the physical analogue experiment of local loading forming (Fan et al., 2011): (a) initial specimen; (b) after the first loading step; (c) after the second loading step.
tion of each phase in three-dimensional space by the area fraction in two-dimensional micrograph. In addition, the mean diameter can provide a direct estimation of the grain size of equiaxed ␣ particle. Indeed, for size of true spheres (Dsphere ), Dsphere = 4/dmean . However, for simplicity and because the true three-dimensional shape is currently unknown, the factor 4/ is omitted in this work. The same treatment methods have been applied in the microstructure analysis of ␣+ titanium alloys by Collins et al. (2009). The aspect ratio and orientation are used to depict the morphology and spatial orientation of particle. There is still a lack of strong evidence for relating them to specific three-dimensional morphology features. However, they have been widely used to evaluate the morphology and orientation of lamellar ␣ in the hot working of titanium alloy with lamellar microstructure, as reported by Park et al. (2014) and Bieler and Semiatin (2002). Thus, the aspect ratio and spatial orientation of lamellar ␣ are also used in this work. To acquire the local effective strains on the specimen for microstructure analysis, the isothermal local loading experiment was simulated using DEFORM-3D in previous study (Gao et al. (2011)). The finite element (FE) model was developed in isothermal condition neglecting any thermal events. The constant shear model is employed and the friction factor is determined to be 0.4. For the element meshing, the tetrahedral element is adopted and the automatic remeshing techniques of DEFORM-3D are applied. The flow behavior of material was input in the form of discrete points based on the results in Shen (2007). The FE model has been validated in the work of Gao et al. (2011), which will be used to acquire the local effective strains for microstructure analysis in this work. The effective strain is a common value to evaluate the deformation degree in the metal forming analysis, which is defined as follows: √ ε¯ =
2 (ε1 − ε2 )2 + (ε2 − ε3 )2 + (ε3 − ε1 )2 3
(1)
where, ε¯ is the effective strain, and ε1 , ε2 and ε3 are principal strains.
Fig. 5. Microstructure observation locations for sample after the first loading step.
3. Results and discussion 3.1. Microstructure evolution in the first loading step In the first loading step, the effects of deformation temperature (T1 ), deformation amount and cooling mode on the microstructure are mainly concerned. The detailed processing parameters are T1 of 910, 930, 950 and 970 ◦ C, 15 min holding, 50% reduction, and the cooling mode of AC and WQ. Microstructures of Region A and B in Fig. 5 are compared to study the effect of deformation amount. FE simulation suggests that the effective strains in Region A and B are 0.98 and 0.05, respectively. Figs. 6 and 7 show the microstructures of Region A and B deformed to 50% reduction at diverse temperatures and air cooled in the first loading step, respectively. The microstructures at diverse temperatures all consist of equiaxed ␣ and  transformed matrix. It can be found that the microstructures of Region A and B present the
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
163
Fig. 6. Microstructures of Region A deformed to 50% reduction at diverse temperatures and air cooled in the first loading step: (a) 910 ◦ C; (b) 930 ◦ C; (c) 950 ◦ C; (d) 970 ◦ C.
Fig. 7. Microstructures of Region B deformed to 50% reduction at diverse temperatures and air cooled in the first loading step: (a) 910 ◦ C; (b) 930 ◦ C; (c) 950 ◦ C; (d) 970 ◦ C.
Table 2 The microstructure parameters of Region A and B at different conditions in the first loading step. Sample
◦
910 C, AC 930 ◦ C, AC 950 ◦ C, AC 970 ◦ C, AC 970 ◦ C, WQ
Volume fraction of ␣p (%)
Grain size of ␣p (m)
Aspect ratio of ␣p
Region A
Region B
Region A
Region B
Region A
Region B
49.17 44.16 35.53 19.06 18.85
51.81 44.90 32.94 20.03 19.37
9.54 9.55 9.08 9.02 8.06
10.14 9.79 8.98 8.28 8.12
3.04 2.84 2.50 2.40 2.37
2.35 2.70 2.25 2.57 2.48
Note: ␣p represents equiaxed ␣.
close morphology at each temperature. Their volume fraction, grain size and aspect ratio of equiaxed ␣ are also very close, as shown in Table 2. We can also found from Table 2 that the volume fraction of equiaxed ␣ decreases notably with the temperature increasing, while the grain size and aspect ratio of equiaxed ␣ change little. To compare the microstructures at different cooling modes, the sample at 970 ◦ C was also quenched in water after deformation. The corresponding microstructure image and parameters are given in Fig. 8 and Table 2, respectively. The microstructure after water quenching is composed of equiaxed ␣ and metastable martensitic, which is distinct from that cooled in air (Figs. 6(d) and 7(d)). While, the volume fraction, grain size and aspect ratio of equiaxed ␣ all change little with the cooling mode, as shown in Table 2. 3.2. Microstructure evolution in the second loading step Previous study (Gao et al., 2012) indicates that the lamellar ␣ generates in the heating and holding stage of the second loading step, which is a key process for the development of tri-modal microstructure. Thus, the second loading step was divided into two stages, i.e., the heating and holding stage and deformation stage. Here, the temperature of the first loading step was fixed at 970 ◦ C. 3.2.1. Heating and holding stage First, we checked whether the deformation amount and cooling mode of the first loading step have after-effects on the microstructure evolution in the second loading step. To this end, two samples deformed to 50% reduction underwent different cooling modes (AC and WQ) in the first loading step were firstly prepared. Then, they were hated to 950 ◦ C, hold 15 min, and water quenched to reserve the high temperature microstructure, as shown in Fig. 9. It can be
found that the microstructures underwent different deformation amounts and cooling modes all produce lamellar ␣ and present the same morphology. Besides, the quantitative analyses (Table 3) show that the parameters of equiaxed ␣ and lamellar ␣ are all very close at different conditions. It can be concluded that the deformation amount and cooling mode of the first loading step have little after-effects on the subsequent microstructure evolution. Fig. 10 shows the waster-quenched microstructures after holding at different temperatures and time, which all consist of equiaxed ␣, lamellar ␣ and  phase. It can be found the microstructures hold 60 min (Fig. 10(e),(f)) and 15 min (Fig. 9(c),(d)) have the same morphology and close parameters of equiaxed ␣ and lamellar ␣ (Tables 3 and 4). This indicates that the microstructure evolution changes little with the holding time in this work. Table 4 lists the microstructure parameters after heating and holding at various temperatures. We can find that the temperature mainly influences the volume fraction of lamellar ␣. Increasing the temperature would decrease the volume fraction of lamellar ␣ significantly. 3.2.2. Deformation stage The effects of deformation amount and temperature are mainly concerned in the deformation stage. To this end, the samples were deformed to various reduction rate (30%, 50% and 70%) at 930 and 950 ◦ C in the second loading step. After deformation, microstructure observation was carried out in three typical regions (Region A–C), as shown in Fig. 11. This because Region A–C could reflect the deformation characteristics of the first-loading region, transitional region, and second-loading region, respectively. Table 5 gives the FE simulated effective strains at three typical regions in two loading steps.
164
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 8. Microstructures of sample deformed to 50% reduction at 970 ◦ C and quenched by water in the first loading step: (a) Region A; (b) Region B.
Fig. 9. Microstructures after heating and holding in the second loading step with different cooling modes in the first loading step: (a) AC, Region A; (b) AC, Region B; (c) WQ, Region A; (d) WQ, Region B. Table 3 The microstructure parameters of microstructures in Fig. 9. Cooling mode
Region
AC
A B A B
WQ
Parameters of ␣p
Parameters of ␣l
Volume fraction (%)
Grain size (m)
Aspect ratio
Volume fraction (%)
Aspect ratio
21.79 19.45 22.79 19.95
8.21 8.80 8.03 8.06
2.04 1.87 2.03 1.92
13.25 14.86 12.71 13.18
5.01 5.50 5.24 5.06
Note: ␣l represents lamellar ␣. Table 4 The microstructure parameters after heating and holding at different temperatures in the second loading step. Temperature (◦ C)
930 940 950
Region
A B A B A B
Parameters of ␣p
Parameters of ␣l
Volume fraction (%)
Grain size (m)
Aspect ratio
Volume fraction (%)
Aspect ratio
21.25 22.67 19.06 20.20 21.79 19.45
8.95 8.41 8.53 8.68 8.21 8.80
1.75 2.27 2.07 2.06 2.04 1.87
28.56 29.14 22.33 21.72 13.25 14.86
5.59 5.53 5.04 5.42 5.01 5.50
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
165
Fig. 10. Water-quenched microstructures of Region A and B after different heating and holding treatments in the second loading step: (a) 930 ◦ C, 15 min, Region A; (b) 930 ◦ C, 15 min, Region B; (c) 940 ◦ C, 15 min, Region A; (d) 940 ◦ C, 15 min, Region B; (e) 950 ◦ C, 60 min, Region A; (f) 950 ◦ C, 60 min, Region B.
Fig. 11. Microstructure observation locations for samples at various reductions: (a) 30%; (b) 50%; (c) 70%.
Table 5 Effective strains at microstructure observation locations in Fig. 11. Reduction rate (%)
Strains in step 1 Region A
Region B
Region C
Region A
Region B
Region C
30 50 70
0.52 0.98 1.53
0.26 0.56 0.66
0.10 0.26 0.22
0.10 0.26 0.40
0.26 0.49 0.84
0.44 0.86 1.29
Strains in step 2
166
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 12. Microstructures of samples deformed to different reductions at 930 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 50%; (g)–(i) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
The microstructures of samples deformed to various reductions at 930 ◦ C are shown in Fig. 12. It can be seen that the morphology and distribution of equiaxed ␣ change little with the loading region and reduction. Moreover, the quantitative analyses shows that the volume fraction, grain size and aspect ratio of equiaxed ␣ all locate in narrow ranges, i.e., 21.89 ± 1.17%, 8.87 ± 0.33 m and 2.02 ± 0.15, respectively. Semiatin et al. (2002) pointed out that the  phase is softer and easier to deform than equiaxed ␣ during the hot working of ␣ +  titanium alloy in two phase region. Considering the morphology and distribution characteristics of lamellar ␣ in  matrix, it can be speculated that the lamellar ␣ deforms more easily than the equiaxed ␣. During deformation, the lamellar ␣ would gradually rotate themselves perpendicular to compression direction, and produce kinking and break-up with the increase of deformation amount. So the lamellar ␣ at different loading regions present various morphology and spatial orientation distribution after undergoing different deformation amounts, as shown in Fig. 12. The quantitative results of spatial orientation distribution of lamellar ␣ under different conditions are given in Fig. 13. The result of “ = 0” represents the microstructure prior to deformation (Fig. 10(a) and (b)). From the results at 30% reduction (Fig. 13(a)), it can be found that Region A presents the similar distribution as that prior to deformation. While, the distributions of Region B and C change a little due to their greater strain. Their fraction of lamellar ␣ with smaller title angle decreases, while the fraction with
greater title angles increases. With the reduction rate increasing to 50% and 70% (Fig. 13(b) and (c)), the strain at each loading region, especially Region B and C, increases greatly, which leads to more notable change in the spatial orientation distribution of lamellar ␣. It’s worth noting that Region B and C at 70% reduction present very close distribution, although there exists a big difference on their strains. From the above analyses, it can be found that increasing the deformation amount would aggravate the rotation of lamellar ␣. This makes the spatial orientation distribution of lamellar ␣ changing from homogeneous distribution to concentrated distribution gradually. In addition, it can be concluded that the distribution change begins at a strain lower than 0.49, and mainly produces at strains between 0.49 and 0.84, while almost stops after strain being greater than 0.84. These phenomena are similar to that reported by Park et al. (2014) in the hot deformation of Ti-6Al-4V alloy with lamellar microstructure. Besides the change of orientation distribution, dynamic globularization is another important phenomenon of lamellar ␣ during deformation, as shown in Fig. 12. The dynamic globularization fraction of lamellar ␣ at different conditions are quantitatively measured and given in Fig. 14. In this work, globularization was taken to be an ␣-phase with aspect ratio less than 2.5, and the globularization fraction was calculated by:
fg =
f2.5after − f2.5pre 1 − f2.5pre
(2)
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
167
Fig. 13. Spatial orientation distribution of lamellar ␣ traces relative to the compression axis for samples deformed to different reductions at 930 ◦ C in the second loading step: (a) 30%; (b) 50%; (c) 70%.
Fig. 14. Dynamic globularization fraction of lamellar ␣ for samples deformed to different reductions at 930 ◦ C in the second loading step.
where, fg denotes the globularization fraction, f2.5pre and f2.5after represent the volume fractions of small equiaxed ␣ (aspect ratio <2.5) before and after deformation. It can be seen from Fig. 14 that the deformation amount has significant effect on the dynamic globularization fraction of lamellar ␣. At smaller deformation amount (Region A and B at 30% reduction), the globularization doesn’t occur. This rule is the same as that in the hot working of two phase titanium alloy with lamellar microstructure reported by Wang et al. (2010). They proposed that the globularization of a lamellae may be rationalized to consist of two processes: the break-up of lamellae and the formation
of globules. The break-up of lamellae firstly needs the nativity of intraphase boundary, which usually occurs at kinks or shearing bands of lamellae. So sufficient strain must be imposed, that is to say, there exists a small critical strain for the initiation of dynamic globularization. Once the globularization beginning, the dynamic globularization fraction increases gradually with the deformation amount (Fig. 14), which can also be clearly observed in the microstructure morphology (Fig. 12). To study the effect of temperature, the microstructures of samples deformed at 950 ◦ C in the second loading step are given in Fig. 15. Comparing Figs. 15 and 12, it can be seen that the volume fraction of lamellar ␣ at 950 ◦ C is much smaller than that at 930 ◦ C, but the morphology and spatial orientation distribution are very similar. This is also confirmed by the comparison on the quantitative results of spatial orientation distribution at 950 ◦ C (Fig. 16) and 930 ◦ C (Fig. 13). Therefore, it can be said that the spatial orientation distribution of lamellar ␣ is little influenced by the temperature in the second loading step. Song et al. (2009) found that in the hot working of titanium alloy with lamellar microstructure, the kinetic of dynamic globularization experiences three stages with the increase of deformation amount: the induction stage, the accelerated globularization stage and the decelerated globularization stage. The induction stage corresponds to the critical strain required for the initiation of dynamic globularization. As a result, the variation of globularization fraction with strain presents an S-type curve starting at the critical strain. Meanwhile, they also found that the Avrami type equation (Eq. (3)) can properly depict the three stages variation process.
fdg = 1 − exp[−k · (ε − εc )n ]
(3)
168
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 15. Microstructures of samples deformed to different reductions at 950 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 50%; (g)–(i) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
Table 6 Fitted parameters. Temperature (◦ C)
εc
k
n
930 950
0.320 0.249
0.590 0.712
1.230 1.624
where, fdg is the dynamic globularization fraction, εc is the critical strain for the initiation of dynamic globularization, k is the temperature-dependent kinetic constant, and n is the Avrami exponent. It not only can reflect the increase trend of globularization fraction with strain, but also can catch the critical strain. Moreover, the effect of temperature on the globularization kinetic can also be reflected by the parameter k. Therefore, the Avrami type equation is widely used to fit the kinetic of dynamic globularization during hot working of titanium alloy with lamellar microstructure, as reported by Wang et al. (2010) and Ma et al. (2012). In this work, although the initial microstructure is different from the fully lamellar microstructure reported above, the dynamic globularization process of lamellar ␣ is the same. It also undergoes the induction and globularization stages, so the Avrami type equation (Eq. (3)) is also trialed in fitting the variation of dynamic globularization fraction of lamellar ␣ with strain at different temperatures (Fig. 17). It can be observed that the fitting results agree well with the experimental results at each temperature. The fitted parameters are listed in Table 6.
From Fig. 17, it can be found that the curve of 950 ◦ C has smaller critical strain and greater dynamic globularization fraction at the same strain than that of 930 ◦ C. As mentioned above, the critical strain is related to the break-up of lamellae ␣ which is mainly achieved by two ways: intense shearing and penetration of  phase along the interfaces. The first way is determined by the deformation, while the second way is a temperature-dependent thermally activated process. At higher temperature, the volume fraction of lamellar ␣ is smaller so that each lamellar ␣ would bear more deformation, which is in favor of the first way. Meanwhile, higher temperature could promote the diffusion process, which is beneficial to the second way. Thus, the higher temperature would promote the break-up of lamellae ␣, leading to smaller critical strain. On the other side, the greater dynamic globularization fraction at higher temperature can be explained as follows. Even the emergence of the interfaces is indispensable, the migration rate of intraphase boundary, as the final step for dynamic globularization, restricts the kinetic rate of dynamic globularization. Higher temperature is beneficial to the diffusion and interface migration, so the dynamic globularization fraction increases with temperature.
3.3. Microstructure evolution in the annealing After local loading, the samples were annealed in the route of 810 ◦ C/1 h/AC. Fig. 18 shows the annealed microstructures of samples deformed at 970 and 930 ◦ C sequentially in two loading steps.
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
169
Fig. 16. Spatial orientation distribution of lamellar ␣ traces relative to the compression axis for samples deformed to different reductions at 950 ◦ C in the second loading step: (a) 30%; (b) 50%; (c) 70%.
Table 7 Comparisons of globularization fraction of lamellar ␣ before and after annealing for samples deformed to different reductions. Reduction rate(%)
State
30
Prior to annealing After annealing Prior to annealing After annealing
70
Fig. 17. Dynamic globularization fraction of lamellar ␣ vs. strain in the second loading step (dot: experimental data, line: fitted results using Eq. (3).
It can be found that the microstructure morphology change little in comparison with those before annealing (Fig. 12(a)–(c) and (g)–(i)). The quantitative analyses of microstructures show that the volume fractions of equiaxed ␣ and lamellar ␣ increase by about 3% and 6.5% in the annealing, respectively. This is caused by the precipitation of ␣ phase and the mergence of fine ␣ phase on the equiaxed ␣ and lamellar ␣. Moreover, the increments of contents of equiaxed ␣ and lamellar ␣ are very close at different loading regions and reductions. Further static globularization of lamellar ␣ in the annealing process was also found when comparing the microstructures before
Globularization fraction Region A
Region B
Region C
0 0 0.021 0.030
0 0 0.229 0.286
0.041 0.048 0.432 0.525
and after annealing (Figs. 12 and 18). Fan et al. (2012b) studied the static globularization behavior of TA15 alloy with lamellar microstructure and found that the globularization kinetics rate increases with pre-deformation. Thus, the effect of deformation amount of the second loading step on the static globularization of lamellar ␣ was also explored in this work. Table 7 gives the comparisons of globularization fraction of lamellar ␣ before and after annealing for samples deformed to different reductions. It can be found that the increments of globularization fraction are very small at regions underwent small deformation (Region A–C at 30% reduction, Region A at 70% reduction), but are greater at regions underwent greater deformation (Region B and C at 70% reduction). These suggest that greater deformation amount in the second loading step could promote the static globularization of lamellar ␣ in the annealing, presenting the same rule as that proposed by Fan et al. (2012b). The annealed microstructures of samples deformed at 950 ◦ C in the second loading step (Fig. 19) were also analyzed to study the after-effect of temperature of the second loading step. The comparative quantitative analyses suggest that the temperature in the second loading step has little after-effect on the rules of content
170
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
Fig. 18. Microstructures after annealing for samples deformed to different reductions at 930 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
Fig. 19. Microstructures after annealing for samples deformed to different reductions at 950 ◦ C in the second loading step (The compression axis is vertical): (a)–(c) 30%; (d)–(f) 70%. For each sample, the three micrographs correspond to Region A, B and C in sequence.
increments of each phase and the static globularization of lamellar ␣ in the annealing. Meanwhile, we found that only if the deformation amount in the second loading step exceeded the critical strain for the initiation of dynamic globularization, obvious static globularization would occur. In other words, the occurring of dynamic globularization in second loading step can be approximately taken as the precondition to the static globularization in annealing. 3.4. Mechanical properties of typical microstructures The tensile stress-strain curves and mechanical properties of typical tri-modal microstructures are shown in Fig. 20. Microstructures A and B have the close equiaxed ␣ content but different
lamellar ␣ content. Comparing their mechanical properties, we can found that the yield strength (R0.2 ), ultimate tensile strength (Rm ) and reduced area (Z) of microstructure A are all greater than microstructure B. While, the elongation (A) of microstructure A is a little smaller than microstructure B. Another hand, microstructure C has the close lamellar ␣ content but greater equiaxed ␣ content in comparison with microstructure B. As for the mechanical properties, microstructure C presents the close R0.2 and smaller Rm compared to microstructure B. Nevertheless, both of A and Z of microstructure C are much greater than microstructure B. It can be concluded that the parameters of tri-modal microstructure play great roles in the mechanical properties. To control the microstructure and mechanical properties, it is very critical to study the effect
P.F. Gao et al. / Journal of Materials Processing Technology 239 (2017) 160–171
171
Acknowledgements The authors would like to gratefully acknowledge the support of National Natural Science Foundation of China (No. 51605388, 51575449), 111 Project (B08040), Project supported by the Research Fund of the State Key Laboratory of Solidification Processing (NWPU), China (Grant No. 131-QP-2015), and the Open Research Fund of State Key Laboratory of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology. References
Fig. 20. The tensile stress-strain curves and mechanical properties of typical microstructures.
of processing conditions on the microstructure and the relationship between microstructure parameters and properties. The former research content is right the focus of this work, and the later one is being studied. 4. Conclusions In this work, the development of tri-modal microstructure during the isothermal local loading of TA15 titanium alloy were quantitatively studied through a physical analogue experiment. It was found that the local loading processing parameters play critical role in the development of tri-modal microstructure as follows: 1. The volume fraction of equiaxed ␣ in tri-modal microstructure is mainly determined by the deformation temperature of the first loading step. Increasing the temperature would decrease the volume fraction of equiaxed ␣. While, the volume fraction of lamellar ␣ increases with the decrease of temperature of the second loading step, when the temperature of the first loading step is fixed at a higher temperature. 2. The spatial orientation distribution and globularization of lamellar ␣ are strongly dependent on the deformation amount and temperature of the second loading step. With the increasing of deformation amount, the spatial orientation distribution of lamellar ␣ changes from homogeneous distribution to concentrated distribution gradually, and the dynamic globularization fraction of lamellar ␣ approximately increases in a sigmoid way. The relationship between dynamic globularization and deformation amount can be well fitted by the Avrami type equation. Besides, increasing the temperature of the second loading step would reduce the critical strain for the initiation of dynamic globularization and enhance the kinetics rate of dynamic globularization. 3. The deformation amount of the second loading step has aftereffects on the static globularization of lamellar ␣ in the annealing treatment. If the deformation amount exceeded the critical strain for the initiation of dynamic globularization in the second loading step, lamellar ␣ would produce static globularization in the annealing. And the static globularization fraction increases with the deformation amount of the second loading step.
Bieler, T.R., Semiatin, S.L., 2002. The origins of heterogeneous deformation during primary hot working of Ti-6Al-4V. Int. J Plast. 18 (9), 1165–1189. Collins, P.C., Welk, B., Searles, T., Tiley, J., Russd, J.C., Fraser, H.L., 2009. Development of methods for the quantification of microstructural features in ␣ + -processed ␣/ titanium alloys. Mater. Sci. Eng. A 508, 174–182. Fan, X.G., Gao, P.F., Yang, H., 2011. Microstructure evolution of the transitional region in isothermal local loading of TA15 titanium alloy. Mater. Sci. Eng. A 528 (6), 2694–2703. Fan, X.G., Yang, H., Gao, P.F., Yan, S.L., 2012a. Dependence of microstructure morphology on processing in subtransus isothermal local loading forming of TA15 titanium alloy. Mater. Sci. Eng. A 546, 46–52. Fan, X.G., Yang, H., Yan, S.L., Gao, P.F., Zhou, J.H., 2012b. Mechanism and kinetics of static globularization in TA15 titanium alloy with transformed structure. J. Alloy Compd. 533, 1–8. Gao, P.F., Yang, H., Fan, X.G., 2011. Quantitative analysis of the microstructure of transitional region under multi-heat isothermal local loading forming of TA15 titanium alloy. Mater. Des. 32 (4), 2012–2020. Gao, P.F., Yang, H., Fan, X.G., Yan, S.L., 2012. Microstructural features of TA15 titanium alloy under different temperature routes in isothermal local loading forming. Mater. Sci. Eng. A 540 (1), 245–252. Lütjerin, G., 1998. Influence of processing on microstructure and mechanical properties of (␣ + ) titanium alloy. Mater. Sci. Eng. A 243 (1–2), 32–45. Ma, X., Zeng, W.D., Tian, F., Zhou, Y.G., 2012. The kinetics of dynamic globularization during hot working of a two phase titanium alloy with starting lamellar microstructure. Mater. Sci. Eng. A. 548, 6–11. Park, C.H., Kim, J.H., Hyun, Y.T., Yeom, J.T., Reddy, N.S., 2014. The origins of flow softening during high-temperature deformation of a Ti-6Al-4V alloy with a lamellar microstructure. J. Alloy Compd. 582, 126–129. Semiatin, S.L., Montheillet, F.G., Shen, G., Jonas, J.J., 2002. Self-consistent modeling of the flow behavior of wrought alpha/beta titanium alloys under isothermal and nonisothermal hot-working conditions. Metall. Mater. Trans. A 33, 2719–2727. Shen, C.W., 2007. Research on material constitution models of TA15 and TC11 titanium alloys in hot deformation process. Master Thesis, Northwestern Polytechnical University. Song, H.W., Zhang, S.H., Cheng, M., 2009. Dynamic globularization kinetics during hot working of a two phase titanium alloy with a colony alpha microstructure. J. Alloy Compd. 480 (2), 922–927. Sun, Z.C., Yang, H., 2009. Microstructure and mechanical properties of TA15 titanium alloy under multi-step local loading forming. Mater. Sci. Eng. A 523 (1–2), 184–192. Sun, Z.C., Wang, X.Q., Zhang, J., Yang, H., 2014. Prediction and control of equiaxed ␣ in near- forging of TA15 Ti-alloy based on BP neural network: for purpose of tri-modal microstructure. Mater. Sci. Eng. A 591, 18–25. Sun, Z.C., Han, F.X., Wu, H.L., Yang, H., 2016. Tri-modal microstructure evolution of TA15 Ti-alloy under conventional forging combined with given subsequent heat treatment. J. Mater. Process. Technol. 229, 72–81. Wang, K.X., Zeng, W.D., Zhao, Y.Q., Lai, Y.J., Zhou, Y.G., 2010. Dynamic globularization kinetics during hot working of Ti-17 alloy with initial lamellar microstructure. Mater. Sci. Eng. A 527 (10–11), 2559–2566. Yang, H., Fan, X.G., Sun, Z.C., Guo, L.G., Zhan, M., 2011. Recent developments in plastic forming technology of titanium alloys. Sci. China-Technol. Sci. 54 (2), 490–501. Zhou, Y.G., Zeng, W.D., Yu, H.Q., 1996. A new high-temperature deformation strengthening and toughening process for titanium alloys. Mater. Sci. Eng. A 221, 58–62. Zhou, Y.G., Zeng, W.D., Yu, H.Q., 2005. An investigation of a new near-beta forging process for titanium alloys and its application in aviation components. Mater. Sci. Eng. A 393 (1–2), 204–212. Zhou, L., Chang, H., Wang, X.D., 2011. Recent progress of titanium industry, research and development in China. Proceedings of the 12th World Conference on Titanium. Beijing, China, 56–61 (Plenary Lectures).