As a kind of biodegradable material with broad application prospects, Mg alloys can completely dissolve after tissue repair and have better mechanical properties than biodegradable polymers used in clinic [[1], [2], [3]]. Due to solid solution strengthening and precipitation strengthening, WE43 (Mg-4Y-3RE) alloy has good tensile properties at room and elevated temperature [4,5]. It has been reported that WE43 has high corrosion resistance, good biocompatibility and osteogenesis [1,3,6,7]. Thus, WE43 has received extensive attention in bone repair applications. At present, bone screws made of MgYReZr alloy whose composition is similar to WE43 have been applied in clinical practice [8,9]. However, Mg alloys used for bone screws have strict requirements for mechanical properties to prevent breakage, because the bone screws need to withstand high shear stress during implantation. Torsional properties of bone screws are important mechanical parameters as shown in some standards [[10], [11], [12]]. Stainless steel and titanium alloys may easily meet the requirements for the torsional performance of biomedical materials, while the torsional properties of Mg alloys is of utmost concern due to their poor mechanical properties. Although, there are some studies on the torsional properties of Titanium alloys, CP Titanium and Co-Cr alloys [13,14], only the torsional properties, fracture morphology, or the torsion deformation and failure mechanisms were studied. Besides, the previous studies on the mechanical properties of Mg alloys mainly focused on tensile, compressive and bending properties [15,16]. There is a few studies on torsional property characterization, which are only for testing the hot workability of Mg alloys or improving the torsional properties [17,18], but no systematic research was performed on the effects of microstructure on the torsional properties of Mg alloys. Thus, it is necessary to further study the torsional properties of Mg alloys for biodegradable bone screw applications.

Heat treatment and equal channel angular pressing (ECAP) are generally recognized methods to optimize the microstructure and mechanical properties of Mg alloys. Appropriate heat treatment can improve the tensile properties of Mg alloys, which may simultaneously improve the torsional properties. ECAP provides a viable method for obtaining biomedical implants with high performance. Bone screws made of ECAP processed Mg alloys were used for animal experiments, and the results showed these screws provided good curative effects in oral and maxillofacial reconstruction [19]. Mg alloy wires or tubes with high mechanical properties could be obtained by ECAP and subsequent drawing or extrusion, which showed bright prospects in the surgical suture or cardiovascular stent application [20,21]. ECAP usually results in significant grain refinement, optimizing the distribution of second phases and texture, and thus greatly improve the tensile properties of Mg alloys [22,23]. Many researchers have found that ECAP could improve the tensile strength and ductility of Mg alloys simultaneously [24,25]. Others suggest that ECAP could improve the plasticity of Mg alloys and even obtain superplasticity [26,27], or improve the tensile strength [28,29]. Thus, ECAP may also be an effective way to improve the torsional properties of Mg alloys.

Therefore, in this work, the extruded WE43 was processed by heat treatment and ECAP to obtain different microstructures, then torsional tests were performed to study the effects of grain size, second phases and texture on the torsional properties. Although, previous studies on the tensile properties of ECAP processed WE43 were mainly performed on solution treated samples [28,[30], [31], [32]], with few on extruded WE43. Thus, the tensile properties of extruded WE43 after different heat treatment and ECAP were also carried out to present the different effects of microstructure on torsional and tensile properties.

The WE43 (Mg-3.97Y-2.4Nd-0.55 Zr) ingots were homogenized at 500 °C for 18 h and pre-heated at 460 °C for 2 h before extrusion, the bars of Φ10 mm were obtained with an extrusion ratio of 64. Then the extruded bars were subjected to different heat treatment: T4 (525 °C ×8 h), T5 (200 °C ×24 h), T6 (525 °C ×8 h + 250 °C ×16 h), respectively. The extruded bars with dimension of Φ10 mm × 120 mm were also subjected to ECAP for 1, 2 and 4 passes in route Bc (the billet was rotated to 90 ° after each pass) after preheating at 390 °C for 15 min, named as 1p, 2p and 4p, respectively in this study. The internal angle and the outer arc of curvature between the two channels of the ECAP die were both 90 °, so the cumulative strain of billets were 0.907, 1.814, 3.628, respectively after 1, 2 and 4 passes of ECAP [33,34]. The ECAP was carried out at 390 °C and lubricated with MoS₂.

For optical microstructure (OM) observation, the samples were cut along the cross section, ground by SiC papers and polished. Then the polished area was etched with a mixed acid (6 g picric acid, 10 mL acetic acid, 10 mL water and 70 mL alcohol) and observed by optical microscopy. Scanning electron microscopy (SEM: backscattered electron imaging) and energy dispersive spectrometry (EDS) were used to analyze the distribution and composition of the second phases. X-ray diffraction (XRD) was also used to evaluate the phase composition. For texture evaluation, the samples were cut along the longitudinal section of each bar and polished, then the macro texture was measured by XRD (Bruker D8) and the (0002) pole figures were mapped.

The torsional tests were performed with a torsion testing machine for bone screws (MTT501) made by SUNS Company, Shenzhen, China. The torsion specimens were 10 mm in gauge length and Φ5 mm in cross section, and the torsional tests were performed at a speed of 20°/min at room temperature. For tensile tests, dog-bone shaped samples with the gauge dimension of 4 mm × 2 mm × 20 mm were cut from the billets with the tensile direction parallel to longitudinal axis. The surfaces of samples were ground up to 2000 grit to remove the traces of the line cutting. Instron-5569 universal testing machine was used for tensile tests. The tensile tests were performed at room temperature with a strain rate of 8.33 × 10^-4 s^-1. After torsional and tensile tests, the fracture morphology of each sample was observed by SEM.

Fig. 1 shows the optical microstructure of the WE43 in different conditions, and Fig. 2 shows the average grain size of each sample. The average grain size of extruded WE43 was 9.1 ± 3.0 μm, and decreased to 7.2 ± 2.2 μm after T5 treatment. After T4 and T6 treatment, the average grain size grew greatly to more than 70 μm. The average grain size decreased to 6.0 ± 2.3 μm after one-pass ECAP, but grew up again as ECAP pass increased. After 4 passes, the grain size was uneven with co-existence of fine grains and deformed large grains.

Fig. 3 shows the distribution of second phases in WE43 under different conditions. The second phases almost disappeared after T4 treatment, but increased after T5 treatment. After T6 treatment, some second phases appeared again and more distributed at the grain boundaries. After one-pass ECAP, the second phases increased and many fine particles appeared. The second phases were mainly distributed at the grain boundaries after 2 passes, but evenly distributed after 4 passes. The EDS analysis for the second phases in T5 treated samples and XRD results of each sample are shown in Fig. 4. EDS results show that the second phases contain Mg, Nd and a small amount of Y, XRD results indicate that the second phases are Mg₁₂Nd, Mg₄₁Nd₅ and Mg₂₄Y₅.

The (0002) pole figures of the extruded and heat-treated WE43 are shown in Fig. 5. The extruded WE43 exhibited a weak basal fiber texture, and there was no basal fiber texture in T4 treated samples. The texture intensity had a slight reinforcement after T5 treatment, which also presented a basal fiber texture. Compared with the T4 treated WE43, the basal planes were more inclined parallel to the extrusion direction (ED) for T6 treated samples. The (0002) pole figures of ECAP processed WE43 are presented in Fig. 6. After one-pass ECAP, the texture was obviously strengthened, and majority of the basal planes deviated from the ED at about 45°, but there were still some basal planes parallel to the ED. After 2 passes, the basal planes were mainly parallel to the ED, it presented a typical strong basal fiber texture. After 4 passes, the texture intensity was significantly increased and the basal planes were mainly distributed at 45° from the ED.

The data obtained from the torsion test system was the torque T (N m) - torsion angles θ (rad) curves, and they were converted to sheer stress (τ) - strain (γ) curves by the formulae as follows [14,35]:

where R and L are radius and length of specimens, respectively, θ (rad) is the torsional angle and T (N m) is the torque. The maximum shear stress (τ_m), yield shear stress (τ_0.3) and maximum shear strain (γ_max) were obtained from the curves. The reported torsional properties were the average of three individual tests.

In order to study the difference between the effects of microstructure on the torsional and tensile properties, the torsional and tensile properties of each sample were both measured, and the results were shown in Fig. 7, Fig. 8 and Table 1. As shown in Fig. 7(a), the torsional ductility slightly decreased after T4 treatment, but the strength greatly reduced. After T5 treatment, the torsional strength increased and ductility decreased. However, both the torsional strength and ductility of WE43 significantly decreased after T6 treatment. As shown in Fig. 8a, the tensile strength of the T4 treated WE43 was much lower than that of extruded WE43, but the ductility was similar. There was great improvement in tensile strength after T5 treatment, but significant reduction in ductility. After T6 treatment, both tensile strength and ductility slightly decreased.

Fig. 7(b) is the torsional sheer stress-strain curves of extruded and ECAP processed WE43. There was no increase in torsional strength after ECAP, and the ductility decreased. After ECAP, one-pass processed samples had higher torsional strength and medium ductility, and 2-pass processed samples had the highest torsional ductility but the lowest strength. After 4 passes, the torsional strength was higher but the ductility was the lowest. The change in tensile properties was much different from that of torsional properties. As shown in Fig. 8(b), the tensile properties of the extruded WE43 significantly improved after ECAP. The tensile strength improved greatly after 2 passes, while the ductility was constant. After 4 passes, the tensile strength slightly increased, with significant improvement in ductility. The tensile strength and ductility of one-pass ECAP processed WE43 were in the middle.

The torsional fracture always starts from the outermost region of samples because the shear stress applied on the outermost region is maximum and first reaches the ultimate shear stress, so the fracture morphology of the outer region can reflect the torsional ductility to some extent. The torsional fracture of the outer region of each sample is presented in Fig. 9. The torsional fracture was lamellar-like morphology. Fig. 10 shows the tensile fracture morphology of WE43 in different conditions. The extruded and ECAP processed WE43 had a dimple fracture, indicating their high tensile ductility. There were many cleavage steps in T5 and T6 treated samples, so the ductility was lower. Thus, the torsional fracture was much different from tensile facture.

According to the viewpoint of mechanics of materials, there is a consistent law between the torsional and tensile properties for isotropic homogeneous materials. That is higher torsional strength and ductility is always accompanied with higher tensile strength and ductility. However, after heat treatment and ECAP, the changes of torsional properties were significantly different from that of tensile properties (Fig. 7, Fig. 8 and Table 1). The fracture morphology also showed the significant difference between torsional and tensional deformation (Fig. 9, Fig. 10). This might be because the microstructure of WE43 was not homogeneous due to the existence of grain boundaries, unevenly distributed second phases and texture etc. Besides, the deformation mode is different between tension and torsion. In tensile tests, the strain is generated along three dimensions under normal stress. While uniform plane strain is generated when pure shear stress is applied on samples in torsional tests, but the shear stress is not uniform [36]. Microstructure, including grain size, second phases and texture, has significant effects on the mechanical properties of materials. Hence, the effect of grain size, second phases and crystallographic texture on the torsional properties of biodegradable WE43 alloy were discussed respectively.

The extruded WE43 was obtained by high-ratio extrusion and high density of strain energy might be stored inside it. Thus, the grain refinement after one-pass ECAP or T5 treatment might arise from recrystallization. In the subsequent thermal deformation process, recrystallized grains were difficult to be refined [25,37], but the grain growth occurred (Fig. 1(f)), and the deformed microstructure appeared again (Fig. 1(g)). The amount and distribution of second phases changed greatly after ECAP, which might be due to the fine second phase particles precipitated at high temperature, and the severe plastic deformation caused the breakage and redistribution of second phases [23].

The extruded WE43 was obtained at high extrusion temperature (460 °C), the non-basal slip could be activated at this temperature [38]. Besides, the 112¯1 crystal orientation tend to be parallel to the ED in extruded rare-earth Mg alloys, which is called“rare earth” texture and also lead to a weaker basal fiber texture [39,40]. Thus, the texture was weak for extruded WE43. The texture were slightly enhanced after T5 treatment but became random after T4 and T6 treatment, which was because the recrystallization of extruded WE43 could result in texture evolution during heat treatment [41,42]. The recrystallization was insufficient during T5 treatment due to the low temperature, while fully recrystallization occurred in high-temperature treatment. The texture significantly changed after ECAP because the billets were rotated to 90° after each pass of ECAP and the effect of dynamic recrystallization, the basal orientation changed between parallel and off the extrusion axis during the ECAP process [26,33,43,44].

The solid solubility of yttrium is large in WE43 Mg alloy [45], and the EDS and XRD results also indicated that there was little yttrium in the second phases, thus there were enough dissolved alloying atoms for solid solution strengthening. In contrast, fine-grain strengthening and precipitation strengthening might play a more decisive role for the strength of WE43 because the grain boundaries and second phases could effectively obstruct the motion of dislocations. As shown in Figs. 1-3, 7 and 8, the T5 treated WE43 with fine grains and many evenly distributed second phases had the highest torsional and tensile strength, and the torsional strength was greatly reduced with grain growth, such as T4 and T6 treated samples. The T6 treated samples with more second phases had much higher torsional and tensile strength than T4 treated samples, which was due to precipitation hardening, and the synergy between the grain boundaries and the second phases distributed at grain boundaries might greatly improve the strength of WE43. Thus, grain refinement and appropriately distributed second phases would result in higher torsional strength. Ref. [46] also indicated that finer grains led to higher torsional strength of gold wires. The grain size of WE43 was not significantly refined after ECAP, but the second phases were increased and redistributed, and the tensile strength increased. Especially for 2-pass ECAP processed sample with more second phases distributed at the grain boundaries had much higher tensile strength then the extruded samples. However, the torsional strength did not increase and even decreased after ECAP, which indicated that some factors, such as texture might have different influence on torsional and tensile strength, which would be discussed in the next section.

The grain size and second phases also affect the torsional ductility. As shown in Fig. 1, Fig. 2, Fig. 3 and 7, the WE43 with more second phases had lower torsional ductility when the grain size was similar. Generally, large second phase particles lead to high stress concentration and micro-cracks during deformation, making the ductility lower, especially when the second phases are distributed at the grain boundaries, which usually result in intergranular fracture. Thus, the T6 treated WE43 had much lower torsional ductility than T4 treated samples due to many second phases distributed at the grain boundaries. However, the 2-pass ECAP processed samples with many second phases distributed at the grain boundaries had higher torsional ductility than 1p and 4p samples, which was abnormal and indicated that other factors also had significant effects on the torsional ductility. The extruded WE43 with finer grains and more second phases had higher torsional ductility than T4 treated samples, which might be due to the fact that non-basal slip could be promoted for Mg alloys with finer grains, because the activation of non-basal slip would improve the ductility of Mg alloy [47,48]. Moreover, due to the uniform shear strain generated in cylindrical samples during torsional tests, the deformation can be more evenly distributed in each grain to avoid stress concentration if the grains are fine and homogeneous. Thus, the grain size might have greater influence on torsional ductility than tensile ductility. As a result, the T6 treated samples had much lower torsional ductility than T5 treated samples, but had higher ductility in tensile tests (Fig. 7, Fig. 8). Besides, uneven grain size would significantly reduce the torsional ductility, such as 4-pass processed samples had the lowest torsional ductility among all the ECAP processed samples, which might partly be due to the uneven grain size. Thus, grain refinement, the reduction of large second phase particles and the increase in grain size uniformity would increase the torsional ductility.

Besides the effect of grain size and second phases, the texture also plays a great role on the mechanical properties of Mg alloys. The ECAP processed samples with a stronger texture had higher tensile strength, but the torsional strength did not increase and even decreased. Moreover, the change of strength and ductility showed opposite trend in torsional and tensile tests after 2 and 4 passes of ECAP. Thus, the effect of microstructure on torsional and tensile properties were different. The previous analysis indicted that the grain size and second phases had similar effects on torsional and tensile properties, thus the difference between torsional and tensile properties should be attributed to the effect of texture.

Generally, for samples with a strong basal texture, when the tensile stress is parallel to the basal planes, the tensile strength would increase and the ductility would decrease [26,44,49]. This is because the basal slip is the main slip mode for Mg alloys deformed at room temperature [48], and the Schmid factor is zero when the normal stress is parallel to the slip planes. Then the shear stress applied on slip planes would be zero and the slip system is difficult to be activated. As shown in Fig. 7, the 2-pass ECAP samples with a typical basal texture had the highest tensile strength and lower ductility among all the ECAP samples despite the similarity in grain size, which indicated that the strengthening effect of basal fiber texture also played an important role besides the effect of second phases. Besides the effect of grain size and second phases, the T5 and T6 treated samples had higher tensile strength and lower ductility compared with the extruded and T4 treated samples, which might partly be due to the crystal orientation that the basal planes parallel to ED was stronger for T5 and T6 treated samples. When the slip planes deviate from the tensile direction (especially when the angle between them is 45°), the tensile strength of the sample is lower and the ductility is higher, due to the high values of the Schmid factor for slip system activation [26]. Thus, 4-pass processed samples with the basal planes deviating from the ED had a lower tensile strength and the highest ductility.

However, pure shear stress was vertically applied to the longitudinal direction of samples and along the circumference of cross section in torsional test, which was much different from tensile test that the normal stress was parallel to longitudinal axis. A sketch of the applied stress and slip planes for torsional and tensile test is shown in Fig. 11. There are more basal planes parallel to shear stress when there is a strong basal fiber texture, and the shear stress applied on the basal planes is the highest and the basal slip would be enhanced. As a result, the torsional strength is lower and the ductility is higher, which is contrary to tensile test. When the basal planes deviate from the extrusion axis, which means the basal fiber texture is weak, the shear stress applied on the basal planes is lower. The basal slip is difficult to be activated, so the torsional strength is higher and the ductility becomes lower. Thus, it was found that the effect of texture on the torsional and tensile properties is different. As shown in Figs. 6 and 7, the torsional strength of ECAP processed WE43 with a stronger texture did not increase and even decreased. Especially for the 2-pass ECAP processed samples with a typical basal fiber texture had the lowest torsional strength but higher torsional ductility, which was due to texture softening overcoming the strengthening effects of second phases during torsion. While the 4-pass ECAP processed samples with the basal planes deviating from the ED had higher torsional strength and lower ductility due to texture strengthening. Due to basal fiber texture strengthened after T5 treatment and the different effects of texture on tension and torsion, the increase in strength and reduction in ductility during torsional test was much lower than that of tensile test (Figs. 7 and 8 and Table 1). Thus, the basal fiber texture would improve the torsional ductility but reduce the torsional strength.

The effect of microstructure on the torsional properties of extruded, heat-treated and ECAP processed WE43 were evaluated in term of grain size, second phases and texture. The following conclusions could be drawn:

(1)ECAP resulted in weak grain refinement for the extruded WE43 with fine grains, but led to great changes in the amount and distribution of the second phases. Heat treatment and ECAP greatly changed the position and maximum intensity of orientations in the pole figures.

(2)The changes of strength and ductility showed different trends in torsional and tensile tests after heat treatment and ECAP, indicating that the effects of microstructure on torsional and tensile properties were not exactly the same. The grain refinement and properly distributed second phases would improve the torsional strength, which were similar to tensile tests. The grain refinement and increase in grain size uniformity could improve torsional ductility, while the second phase particles decreased the torsional ductility. Besides, the basal fiber texture would reduce the torsional strength but improve the torsional ductility, which was much different from its effect on tensile properties.