Introduction

Due to their high strength-to-density ratio and excellent corrosion resistance, titanium and titanium alloys are widely used in the aerospace field since 1950s (Ref 1). In recent years, pure titanium as biomaterial has been paid extensive attention owing to its excellent biocompatibility and low modulus of elasticity (Ref 2, 3). It is found that some alloying elements are harmful to human body when they are used in making transplantation materials with titanium element (Ref 4), but there is no known toxicity associated with pure titanium up to now. As a result, the use of commercially pure titanium is desirable and receives extensive attention. However, the coarse-grained pure titanium cannot meet the development needs for medical implants due to its limited strength. The most ideal method to minimize the problem is to prepare the ultrafine-grained or nanostructured pure titanium processed by severe plastic deformation (SPD) (Ref 5). It is well known that high plastic strain can be accumulated through simple shear deformation (Ref 6), and torsion test is a simple and effective means to provide large strain dominated by shear deformation. Compared with tension and compression, it is possible for torsion to achieve large deformation without rupture or strain localization (Ref 7, 8). Therefore, it is necessary to explore the deformation behavior of pure titanium during torsion deformation, for it can provide theoretical basis and experimental support to study SPD-induced grain refinement.

When the material is applied to the shape forming, its practical application requires large plastic deformation (Ref 9), especially at high temperature (Ref 10). As is known, temperature has an important effect on deformation resistance, flow stress and metallurgical transformation kinetics for many metals (Ref 11). The hot deformation also can be designed to control final microstructure and mechanical properties (Ref 12). Therefore, many tests are focusing on studying the mechanical properties and microstructure in hot deformation (Ref 13,14,15,16,17,18). These studies show that recovery and dynamic recrystallization (DRX) can occur in high-temperature deformation, which may play an important role in controlling microstructure and corresponding mechanical behaviors (Ref 16,17,18,19). Besides, the phase transition may appear in the process of deformation when temperature reaches a certain degree for metal materials (Ref 12). The high temperature can resolve the problem of the insufficient deformation as well as change the phase distribution and phase morphology effectively. The deformation strengthening and heat treatment strengthening can be simultaneously proposed to prepare the billets with excellent comprehensive properties (Ref 20, 21). This combined strengthening process is called thermo-mechanical treatment, which is very helpful to enhance the strength and plasticity of metals.

As a common deformation technology, torsion tests are often used to simulate high-temperature processes at the industrial level and assess the hot working properties of some metallic alloys. The special studies of the mechanical properties, microstructure and fracture mechanism in pure titanium processed by torsion at room temperature have been carried out (Ref 7, 22, 23). The previous studies state that dislocation strengthening is the dominated strengthening mechanism for pure titanium deformed at room temperature, and no weakening mechanism appear. In addition, the gradient microstructure with elongated grains is observed. However, the previous study does not take into account the impact of temperature on microstructure and mechanical behavior. It can be speculated that the dislocation behavior will change, and dynamic recovery or DRX may appear under the effect of temperature, which can influence its mechanical behavior significantly. Thus, it is of great importance to explore the effects of combined temperature and applied shear stress on microstructure and mechanical properties. And a comprehensive experimental investigation including mechanical property, microstructure discussion and fracture analysis should be conducted. The study of thermo-mechanical treatment could provide a new idea or reference value for improvement in mechanical properties and controlling grains morphology. Therefore, this paper aims at exploring the mechanical properties, microstructure evolution and fracture mechanisms of pure titanium processed by torsion at several different temperatures based on experimental studies.

Materials and Methods

The pure titanium Ti-GR2 (ASTM) rods were first annealed at 823 K for 1 h in argon atmosphere and followed by cooled in air for removing residual stress. The annealed samples were subjected to the torsion deformation. The detailed dimensions of the sample are shown in Fig. 1(a). The torsion tests were conducted in a XC-10 wire torsion testing machine with the torsion speed 30r/min at room temperature (298 K) and different high temperatures (673 and 1073 K). The rod samples were twisted until the occurrence of fracture failure. The number of total torsion turns before fracture at 298 K is 1.3. Thus, Ti-298 K-1.3 is defined to mean twisting 1.3 turns at 298 K. The number of turns at 673 and 1073 K are 9.2 and 11.5 before fracture, and thus, the deformed samples are marked as Ti-673 K-9.2 and Ti-1073 K-11.5, respectively. In addition, the Ti-673 K-0, Ti-673 K-3, Ti-673 K-6, Ti-673 K-9, Ti-1073 K-0, Ti-1073 K-3, Ti-1073 K-6 and Ti-1073 K-9 samples were prepared in order to investigate the microstructure development during 673 and 1073 K. During high-temperature torsion, the middle part of sample was wrapped by induction coil. The samples were heated by high-frequency induction heating (HD-40KW). And the temperature was controlled by infrared temperature measurement technology (STB-42). The samples were twisted immediately when the temperature was attained. After deformation, the samples were quenched in water immediately in order to retain the deformed microstructure.

Fig. 1
figure 1

(a) Detailed dimensions of torsional specimen, (b) and (c) the positions of the microhardness on TS and LS, (d) the sampling locations of the microstructure on TS and LS

Full size image

Microhardness tests were conducted at the load of 200 g and dwell time of 15 s using HXP-1000TM tester. In order to ensure the reliability of the data, eight times tests were carried out on a circle in transversal section (TS). And ten times tests were conducted on a line in longitudinal section (LS). The positions of the microhardness test on TS and LS are shown in Fig. 1(a), (b) and (c). The microstructure of undeformed and deformed rods was characterized by an optical microscope (OM, OLYMPUS/PMG3), as shown in Fig. 1(d). The surface of TS and LS (about 3.0 mm from the center) is taken for microstructure observation. For the Ti-673 K-0, Ti-673 K-3, Ti-673 K-6, Ti-673 K-9, Ti-1073 K-0, Ti-1073 K-3, Ti-1073 K-6 and Ti-1073 K-9 samples, the accumulated equivalent plastic strains are about 0, 2.176, 4.351, 6.527, 0, 2.176, 4.351 and 6.527 on sample surface using equivalent strain calculation formula (Ref 24, 25). The chemical etchant used on the samples is a solution consisting of 5 mL HNO3 with 10 mL HF and 85 mL H2O. The fracture morphology and precipitate of deformed samples were observed by using the scanning electron microscope (SEM, MIRA3 TESCAN and FEINNS450). Transmission electron microscope (TEM, H800) samples were prepared from the deformed samples. The samples were polished mechanically using silicon carbide papers, and the thickness of polished samples is expected to be about 50 µm. A slice with a diameter of 3 mm was obtained, and then twin-jet polishing tests were conducted on these slices in a solution of 5% perchloric acid, 35% butanol and 60% methanol at an applied potential of 40 V and a temperature of 233 K.

Results and Analysis

Hardness Analysis

The hardness distribution of deformed samples on TS at different temperatures is shown in Fig. 2. The average hardness of undeformed sample is about 190 HV, and it is relatively evenly distributed, as shown in Fig. 2(a). The hardness of sample is improved significantly after twisting 1.3 turns at 298 K. It is observed that the hardness increases gradually from center to surface on TS, which is due to the gradient strain induced by torsion deformation. As a result, the hardness increment on surface position is higher than that of center position owing to its accumulated higher plastic strain. According to the previous studies, the dislocation density increases during deformation, and the improvement in hardness is attributed to dislocation strengthening (Ref 22, 23).

Fig. 2
figure 2

(a-c) Variation trends of microhardness vs. radius on TS during torsion, (d-f) the variation trends of microhardness on LS

Full size image

For Ti-673 K-3 sample, the hardness increases gradually from center to surface on TS, especially increasing dramatically on edge section, as shown in Fig. 2(b). With the increase in plastic deformation, the hardness is improved. It can be seen that the hardness of Ti-673 K-6 and Ti-673 K-9 samples is higher than that of undeformed pure titanium, but their hardness is lower than that of Ti-298 K-1.3 sample (except for outside edge). It is inferred that the hardening effect caused by torsion deformation is weakened owing to the appearance of dynamic recovery at 673 K. The dynamic recovery happens under the effects of imposed shear stress and high temperature. The softening caused by dynamic recovery is actually attributed to the dislocation climb and cross-slip at this moment (Ref 26). Therefore, there is a competitive relationship between dislocation hardening and dynamic recovery softening during torsion deformation at 673 K. The sample rotates to 9 turns, and a higher plastic strain can be accumulated during deformation at 673 K, but hardness increment at this time is less than that of deformed sample at 298 K. It is speculated that softening mechanism dominates firstly and then dislocation strengthening dominates subsequently during torsion deformation at 673 K.

In Fig. 2(c), the hardness is improved obviously after the sample is twisted 3 turns at 1073 K. Then the hardness decreases significantly with the increase in torsion turns at 1073 K. Finally, the hardness increases slightly and reaches steady state. It is well known that the accumulated plastic strain can decrease the phase transition point, and thus, the β phase may appear during deformation at 1073 K. The α phase will precipitate from β matrix when the deformed sample is subjected to water cooling rapidly. Along with the increase in plastic deformation, the hardening effect caused by torsion deformation will be weakened with the appearance of dynamic recrystallization. Compared with the deformed sample at 673 K, it is speculated that dynamic recrystallization happens at this moment owing to accumulated higher plastic strain and higher temperature. The fine recrystallization grains are observed in the following microstructure description of Ti-1073 K-9 sample (Fig. 8). The strengthening mechanism depends on the precipitate under this condition. At first, the hardness increases dramatically because the dislocation movement is hindered by precipitate, but this blockade effect is weakened or eliminated with the continuous high temperature. During this process, lots of substructures with low-angle grain boundaries (LAGBs) gradually evolve into fine grains with high-angle grain boundaries (HAGBs) through dislocation movement (Ref 10). Finally, the hardness decreases and returns to the undeformed level. The detailed variation trends are given in Fig. 2(c). Furthermore, the deformation twins are observed in the following transmission electron microscopy (TEM) observation (Fig. 8b and 9). The appearance of deformation twins could absorb dislocation and change the crystal orientation. As a consequence, the resistance of dislocation movement decreases, and thus, the hardness decreases (Ref 23). Compared with the deformed sample at 673 K, there is more obvious gradient in distribution of hardness on TS, and the hardness increases evidently on edge section for deformed sample at 1073 K.

The hardness distribution of deformed samples on LS at different temperatures is shown in Fig. 2(d), (e) and (f). The microhardness is relatively uniformly distributed, and the average hardness of undeformed sample is about 191 HV in Fig. 2(d). The average hardness of Ti-298 K-1.3 on LS is 209 HV, as shown in Fig. 2(d). It can be seen from Fig. 2(e) that the average microhardness for Ti-673 K-3, Ti-673 K-6 and Ti-673 K-9 samples is about 189 HV, 198 HV and 206 HV, respectively. The average hardness for Ti-1073 K-3, Ti-1073 K-6 and Ti-1073 K-9 samples is about 343 HV, 186 HV and 194 HV, respectively, as shown in Fig. 2(f). The variation trends for deformed samples at 298, 673 and 1073 K on LS have good agreements with the previous results on TS.

For one thing, the hardness can be improved by dislocation strengthening and precipitation strengthening. For another thing, the hardness can be decreased by dynamic recovery and dynamic recrystallization. For deformed samples at 298 K, the hardness increment caused by torsion is about 18 HV, which is highest at all temperatures. (At 673 K, the hardness increment is about 15 HV compared with the undeformed sample. For deformed samples at 1073 K, the hardness is first increased obviously, then decreases evidently and finally increases slightly. Before fracture, the hardness increase is about 3 MPa compared with the undeformed sample, which is the smallest of above three temperatures.) At this condition, the increase in hardness is mainly due to dislocation strengthening and no obvious weakening mechanism exists. For deformed samples at 673 K, the hardness increment caused by torsion is smaller than that at 298 K. And it indicates that dislocation strengthening and dynamic recovery mechanisms coexist, but dislocation strengthening dominates in the process of deformation. For deformed samples at 1073 K, the hardness first increases, then decreases and finally increases again. In general, the hardness of deformed sample at 1073 K is approximately equal to that of undeformed sample. It indicates that precipitation strengthening dominates firstly and then dynamic recrystallization dominates subsequently during torsion deformation at 1073 K.

Microstructure Evolution

The optical microstructure of undeformed pure titanium on TS and LS is shown in Fig. 3(a) and (b). The equiaxed grains are observed, and the average grain size is about 30 µm. The optical microstructure of Ti-298 K-1.3 sample on TS and LS is shown in Fig. 3(c) and (d). Compared with that of undeformed sample, the grains of Ti-298 K-1.3 sample are refined and its average size is about 20 µm. The equiaxial grains are still observed on TS, but the grains on LS are elongated due to the presence of gradient shear strain caused by torsion (Ref 22). The optical microstructure of Ti-673 K-0 and Ti-1073 K-0 samples on TS is presented in Fig. 3(e) and (f). The grains are equiaxed in Fig. 3(e), and the grain size keeps unchanged in contrast to that of undeformed sample. For Ti-1073 K-0 sample, the equiaxed grains are still presented, but its grain size is larger than that of undeformed sample. For sample, after quenching at 673 K (without torsion), the temperature is low and recrystallization does not occur. Therefore, the average grain size keeps unchanged compared with that of undeformed sample. For sample, after quenching at 1073 K (without torsion), the temperature is high and recrystallization appears, but the time is limited. Thus, the average grain size is slightly larger than that of undeformed sample.

Fig. 3
figure 3

(a) and (b) Microstructure of undeformed pure titanium on TS and LS, (c) and (d) the microstructure of Ti-298 K-1.3 sample on TS and LS, (e) and (f) the microstructure of Ti-673 K-0 and Ti-1073 K-0 samples on TS

Full size image

The optical microstructure of deformed pure titanium at 673 K is shown in Fig. 4. For Ti-673 K-3 sample, the original HAGBs break up and become unclear, as shown in Fig. 4(a) and (a′). The α phase with strip shape precipitates non-uniformly on TS. Because pure titanium has bulky α phase stable element and can accumulate high plastic strain at 673 K, it is easy for pure titanium to precipitate α phase at this time. Besides, the crystal lattice defects induced by shear stress increases high distortion energy of deformed material. When the deformed pure titanium is quenched in water rapidly, these defects can provide the positions for nucleation of α phase. But the precipitated α phase is not very clear and its volume percentage is low. With the increase in torsion deformation, the accumulated plastic strain increases, and the sample is in a high-temperature state for a longer time. The dynamic recovery is easy to happen at this time, which can make the grains coarsening. It can be seen in Fig. 4(b) and (b′) that the coarse grains with curved grain boundaries appear on TS and LS of Ti-673 K-6 sample. Due to the accumulated plastic strain, the volume percentage of precipitated α phase increases. What is more, the α phase precipitates along the grain boundaries orderly. The precipitated α phase is more fine and the distance of them decreases compared with that in Fig. 4(a) and (a′). It is inferred that the precipitated α phase with high length–diameter ratio can be formed easily at 673 K. With the increase in deformation, the boundaries of coarse grain become blurred and broken for Ti-673 K-9 sample, as shown in Fig. 4(c) and (c′). Parts of α phases in strip shape change to rod-shaped α phase. As a result, the newly formed α phase with rod shape intertwines with previous α phase with strip shape. It indicates that spheroidizing phenomenon of α phase occurs with the increase in plastic deformation. The volume percentage of α phase with high length–diameter ratio decreases, which can make mechanical properties increase. Therefore, it indicates that the improvement in hardness may also be related to the α phase. When the amount of deformation is relatively small, the strengthening caused by precipitate is not obvious, and the hardness increment caused by dislocation strengthening is eliminated. The dynamic recovery dominates at this time, and the grain coarsening phenomenon happens. As a result, the average hardness of Ti-673 K-3 sample is smaller than that of undeformed pure titanium. However, when the amount of deformation is large, the accumulated plastic strain dominates, and the strengthening effect caused by precipitate increases. The grains are refined and grain boundaries break, but the hardness increment is limited due to the increase in softening effect caused by dynamic recovery. Based on the detailed analysis about microstructure, the variation trends of hardness can be explained thoroughly.

Fig. 4
figure 4

(a) and (a′) Microstructure of Ti-673 K-3 sample on TS and LS, (b) and (b′) the microstructure of Ti-673 K-6 sample on TS and LS, (c) and (c′) the microstructure of Ti-673 K-9 sample on TS and LS. (The red dotted line denotes the grain boundaries; the black arrows refer to lamellar α phase.)

Full size image

The optical microstructure of deformed pure titanium at 1073 K observed from TS and LS is shown in Fig. 5. It can be seen from Fig. 5(a) and (a′) that the average grain size of Ti-1073 K-3 sample increases compared with that of undeformed pure titanium. With the increase in plastic deformation, the grain boundaries of Ti-1073 K-6 sample are fuzzy. The strip-shaped α phase is broken, and the newly formed α phase with rod shape is not clear, as shown in Fig. 5(b) and (b′). For Ti-1073 K-9 sample, the rod-shaped α phase and strip-shaped α phase are mixed together, as illustrated in Fig. 5(c) and (c′). From the microstructure observation of LS, these rod-shaped α phases distribute uniformly, which exhibits the characteristics of basket-weave microstructure.

Fig. 5
figure 5

(a) and (a′) Microstructure of Ti-1073 K-3 sample on TS and LS, (b) and (b′) the microstructure of Ti-1073 K-6 sample on TS and LS, (c) and (c′) the microstructure of Ti-1073 K-9 sample on TS and LS (The red dotted line denotes the grain boundaries; the black arrows refer to lamellar α phase) (Color figure online)

Full size image

Compared with that of Ti-673 K-3 sample, the precipitated α phase of Ti-1073 K-3 sample is denser and the distance of α phase is smaller. The α phase with high length–diameter ratio mainly distributes on TS and the rod-shaped α phase distributes non-uniformly on LS, which are different from the Ti-673 K-3 sample. A sudden and dramatic increase in hardness may be due to the aggregation of these rod-shaped α phases. For Ti-1073 K-9 sample, the volume percentage of precipitated α phase with rod shape increases, which is higher than that of Ti-673 K-9 sample. Also, the length–diameter ratio of these α phases is smaller for Ti-1073 K-9 sample. It is inferred that spheroidizing phenomenon of α phase is more evident at 1073 K. The temperature approaches phase transition point and degree of supercooling is small. As a result, the driving force of phase transition is not enough to form the strip-shaped α phase, and thus, the rod-shaped α phase precipitates. These rod-shaped α phases usually distribute along the grain boundaries. Furthermore, the grain boundaries are not seen clearly with the increase in plastic deformation. It is because the α phase is transformed to β phase under large plastic strain at 1073 K, and then the formed β phase transforms to other precipitates during water cooling, which makes the grain boundaries disappear.

The magnified optical microstructure of Ti-673 K-9 and Ti-1073 K-9 samples on LS is shown in Fig. 6. It can be seen from Fig. 6(a) that recrystallization phenomenon is not obvious. However, lots of very fine dynamic recrystallization grains appear around the precipitated α phase, as shown in Fig. 6(b). At 1073 K, the dynamic recrystallization is easy to appear because of the accumulated plastic strain and high temperature. At this time, the required temperature for dynamic recrystallization decreases, and the dynamic recrystallization grains are prone to appear the positions with high storage energy (Ref 10). Based on the microstructure analysis, the average hardness of Ti-1073 K-3 sample is much higher than that of undeformed pure titanium due to precipitation strengthening during torsion deformation. However, the softening effect caused by dynamic recrystallization increases quickly, which results in the variation of hardness at 1073 K. High-resolution figures obtained by using SEM (FEINNS450) are shown in Fig. 7, which is used to observe the precipitate clearly. It is shown that numbers of α phase distribute on TS and LS for Ti-673 K-9 and Ti-1073 K-9 samples, which has a good agreement with Fig. 4 and 5. Besides, it can be seen that large numbers of fine precipitates distribute uniformly in Ti-1073 K-9 sample, which may lead to much higher hardness at 1073 K.

Fig. 6
figure 6

(a) and (b) Microstructure of Ti-673 K-9 and Ti-1073 K-9 samples on LS

Full size image
Fig. 7
figure 7

(a) and (b) SEM microstructure of Ti-673 K-9 and Ti-1073 K-9 samples on LS

Full size image

TEM bright-field images of deformed pure titanium at different temperatures are shown in Fig. 8. Figure 8(a), (b) and (c) shows TEM images of Ti-298 K-1.3, Ti-673 K-9 and Ti-1073 K-9 samples, respectively. In Fig. 8(a), a large number of shear bands (SBs) form along the grain boundaries (GBs), and the dislocation with high density gathers around the SBs. Compared with the analysis result in Fig. 8(a), a large number of dislocations gather in the form of dislocation walls (DWs), and deformation twins occur around the position with high dislocation density in Fig. 8(b). Figure 8(c) shows TEM image of Ti-1073 K-9 sample. Just like the Ti-673 K-9 sample, the SBs and high-density dislocation appear on selected section.

Fig. 8
figure 8

TEM images of (a) Ti-298 K-1.3, (b) Ti-673 K-9 and (c) Ti-1073 K-9 samples

Full size image

The plastic deformation of pure titanium at room temperature is attributed to the dislocation movement. It is speculated that large numbers of SBs occur in the activated slip systems, and the dislocation density increases with the increase in the plastic deformation. As a result, the dislocation will evolve into dislocation cells or walls, and the dislocation density of the cell wall is higher than that of intracellular. Finally, the dislocation of wall agglomerates and cellular structure collapses. And the dislocation evolves into a large number of substructures or subgrains (SGs), as shown in Fig. 8(a) (Ref 22, 23). When the torsion deformation is carried out at 673 K, the accumulated plastic strain increases evidently. It is inferred that the dislocations evolve into DWs during deformation at 673 K, for the energy required for dislocation climb can be provided. Furthermore, the deformation twin occurs, and it is the {10-12} twin, because it is easy for the {10-12} twin to activate, for its shear strain is smallest (Ref 26). The deformation twins are activated to coordinate the plastic deformation when the dislocation slip is hindered. The appearance of deformation twins could make the grains rotate and thus promote dislocation slip. From the results of TEM images of Ti-1073 K-9 sample, the deformation process of pure titanium at 1073 K suffers a series of changes as stated at 298 and 673 K. The deformation mechanism is mainly the dislocation slip. Besides, lots of lamellar α phase are distributed along grain boundaries and large numbers of dislocation gather around these α phase in Fig. 9. Furthermore, many {10-12} twins are detected at the intersection of lamellar α phase and dislocation. It is inferred that the dislocation slip is hindered by lamellar α phase, which expedites the appearance of deformation twins. The appearance of α phase prevents dislocation from moving, and the dislocation gathers in the grain boundaries. Therefore, the hardness is improved significantly (Ref 27). It also can be seen that a large number of α phases are distributed along the grains and grain boundaries (Fig. 4 and 5).

Fig. 9
figure 9

TEM images of (a) Ti-1073 K-9 samples; (b) and (c) the selected area electron diffusion of {10-12} twin and the lamellar α-Ti phase

Full size image

Fracture Analysis

The fracture appearance of deformed samples can be determined exhaustively by SEM, and detailed analysis of the results could reveal the effects of intrinsic microstructural features on mechanical properties of the chosen material. The ductile fracture is attributed to the micro-voids’ nucleation, growth and expansion, while the brittle fracture is related to the nucleation of microscopic cracks and propagation of macroscopic cracks (Ref 28).

Figure 10 shows the SEM images of fractured pure titanium caused by torsion at different temperatures. It can be seen from Fig. 10(a), (b) and (c) that the fracture morphology is dominated by elongated dimples for Ti-298 K-1.3 sample, which indicates that the fracture process is mainly dominated by ductile fracture. Due to the presence of torsional shear stress in the fracture process, the dimples on fracture surface are elongated as shown in Fig. 10(b) and (c). In addition, evident track of torsion deformation on fracture surface is observed. Figure 10(d), (e) and (f) shows the fracture morphology of Ti-673 K-9.2 sample. It can be seen from these images that a large number of shear facets are oriented in the same direction. These non-directional facets are slender, which are arranged in rows and intersected at grain boundaries, as shown in Fig. 10(e). And a large number of equiaxial dimples distribute inside these small facets. In addition, the river pattern and cleavage steps are presented on the fracture surface, which shows the characteristics of intergranular fracture. In summary, the fracture surface shows the characteristics of shear fracture and cleavage fracture. Figure 10(g), (h) and (i) shows the fracture appearance of Ti-1073 K-11.5 sample. The fracture surface presents the characteristics of typical intergranular fracture. The apparent grain boundaries and lots of blocks in polyhedral shape are shown on fracture surface. It indicates that the deformation of the grains is small in the process of torsion at high temperature, but the deformation of the ductile phases at the grain boundary is large. Combining with the previous analysis of microstructure, the α phase is distributed at the grain boundary. These α phases are brittle so that the grain boundaries are easy to crack in the deformation process, resulting in the appearance of intergranular fracture. The precipitated α phase distributes along the grain boundaries, as shown by the arrows shown in Fig. 10(i).

Fig. 10
figure 10

(a-c) Fracture appearance of Ti-298 K-1.3 sample, (d-f) Ti-673 K-9.2 sample, (g-i) Ti-1073 K-11.5 sample

Full size image

Conclusions

To sum up, the experimental results of hardness measurement, microstructure evolution and fracture appearance of pure titanium processed by torsion at different temperatures are analyzed and compared. Some conclusions are drawn as follows:

  1. (1)

    The hardness of pure titanium is improved significantly when torsion deformation is conducted at 298 K, and the improvement in hardness is attributed to dislocation strengthening. The hardness is also improved when it is deformed at 673 and 1073 K compared with that of undeformed sample, but their increments before fracture are smaller than that at 298 K, which is due to dynamic recovery and dynamic recrystallization, respectively.

  2. (2)

    The grains can be refined obviously when the pure titanium is twisted at 298 K, and the elongated grains on LS are observed. The α phase with strip shape firstly precipitates, and finally parts of strip-shaped α phases transform into α phase with rod shape when torsion deformation is conducted at 673 K. When torsion deformation is carried out at 1073 K, the α phase with rod shape firstly precipitates on LS, and finally lots of rod-shaped α phases and strip-shaped α phases appear on TS and LS simultaneously.

  3. (3)

    Lots of fine grains caused by dynamic recrystallization appear on TS and LS during deformation at 1073 K. These fine and equiaxed grains are distributed around the precipitated α phase.

  4. (4)

    A large number of elongated dimples occur on fracture surface for deformed sample at 298 K. When torsion deformation is conducted at 673 K, lots of shear facets are arranged in rows and intersect at the grain boundaries on fracture surface. When the torsion deformation is carried out at 1073 K, lots of blocks in polyhedral shape are presented on fracture surface, which belong to intergranular fracture.