Developing Microstructure and Enhancing Strength of Ti–6Al–7Nb Alloy with Heat Treatment Processed by High-Pressure Torsion

Abstract

Ti–6Al–7Nb alloy with and without heat treatment was processed by high-pressure torsion (HPT), and subsequently its microstructures and mechanical properties were investigated. Herein, the development of the microstructure by HPT processing significantly depended on the initial microstructures even though their minimum grain size was approximately equal and less than 100 nm. The needle-like structure was fragmented into several grains by HPT processing with a small number of revolutions, and grain refinement was more easily achieved than the equiaxed structure. The Ti–6Al–7Nb alloy with a duplex microstructure comprising equiaxed and needle-like structures was obtained with an adequate balance of tensile strength (1280 MPa) and elongation to fracture (22%). The results show that the combination of the initial duplex microstructure of equiaxed α grain, needle-like structure of the α′ phase, and the resultant inhomogeneous microstructure obtained by HPT processing through a moderate number of revolutions is effective in achieving a better balance of mechanical properties in the Ti–6Al–7Nb alloy.

1. Introduction

The Ti–6Al–7Nb alloy possesses superior characteristics of lightweightness, low toxicity, and high corrosion resistance;¹⁾ it is widely employed in the medical field for fabricating dental implants, artificial hip joints, and spinal fixators. This alloy was developed for substituting a more widely employed Ti–6Al–4V alloy by replacing V in the Ti–6Al–4V alloy with non-toxic Nb; this was done because V is cytotoxic, and if V ions are released from the Ti–6Al–4V alloy in the human body, it leads to several complications.²⁾ Therefore, the V-free Ti–6Al–7Nb alloy is safer than the Ti–6Al–4V alloy. Moreover, the Ti–6Al–7Nb alloy possesses a microstructure and mechanical properties similar to those of Ti–6Al–4V alloy.³⁾

As previously mentioned, this alloy is widely used in fabricating dental implants. The narrow dental implants are used when the amount of available human bone is limited. However, the narrow dental implants have a higher risk of fracture because of their thin diameter.⁴⁾ Therefore, a good balance between high strength and large elongation is necessary for materials with narrow dental implants for reliable use.

The microstructure of the Ti–6Al–7Nb alloy consists of two phases, α and β, which maintain a good balance of mechanical properties for medical applications. Nevertheless, sufficient strengthening of the alloy cannot be achieved by conventional quenching and aging treatments. To meet this requirement, this study uses a process of high-pressure torsion (HPT). It is now a well-known processing technique to strengthen metallic materials through significant grain refinement.^5,6) It was reported that grain refinement can be achieved in the Ti–6Al–7Nb alloy by HPT processing;^7–11) it was investigated whether the grain refinement of the Ti–6Al–7Nb alloy by severe plastic deformation is effective in realizing superplasticity.^7,12) In addition, the Ti–6Al–7Nb alloy with ultrafine grains showed good cytocompatibility with the alloy before HPT processing.⁸⁾ Furthermore, superplasticity of the Ti–6Al–7Nb alloy with heat treatment can be achieved by HPT processing.¹³⁾ However, it is not clear concerning the effects of HPT processing on the tensile properties at room temperature (about 293 to 303 K) and the differences in the development of the resultant microstructure after HPT processing of Ti–6Al–7Nb alloy with different initial microstructures by heat treatment. It has been reported that the hardness of the HPT-processed Ti–6Al–4V alloy was improved by changing the initial microstructure via heat treatment.^14,15)

This study discusses the possibilities of enhancing the tensile properties via grain refinement by applying heat treatment and HPT processing to the Ti–6Al–7Nb alloy. This is the first study on the effects of HPT processing on the tensile properties at room temperature and differences in the development of the resultant microstructure after HPT processing of the Ti–6Al–7Nb alloy with different initial microstructures upon heat treatment.

2. Materials and Methods

2.1 Materials and heat treatments

A Ti–6Al–7Nb alloy (ASTM F1295, Daido Steel) in a form of rod with 10 mm diameter was cut to lengths of 70 mm. The following three types of heat treatments were applied to the rods.

(1) Solution treatment and aging (STA): The rods were subjected to solution treatment at 1258 K for 1 h, air-cooled to room temperature, and then subjected to aging at 973 K for 4 h followed by air cooling.

(2) Solution treatment and quenching at a temperature lower than the β-transus temperature, T_β = 1278 K¹¹⁾ (STQ < T_β): The rods were subjected to solution treatment at 1258 K for 1 h, followed by quenching in iced water.

(3) STQ > T_β: The rods were subjected to solution treatment at 1308 K (higher than T_β) for 1 h, followed by quenching in iced water.

2.2 HPT processing

Disks with 0.8 mm thickness were cut from the heat-treated rods. For comparison, the disks were cut from the as-received rods without any heat treatment. The disks were processed by HPT at room temperature with a rotation speed of 1 rpm under a pressure of 6 GPa. The number of revolutions (N) imposed on the disks were 1, 5 and 20.

2.3 Structural characterization

The microstructures were characterized by X-ray diffraction (XRD) method, scanning electron microscopy with electron backscatter diffraction (SEM-EBSD) and transmission electron microscopy (TEM). For XRD, the disks were mechanically ground with waterproof SiC paper (up to P800) and polished to mirror-like surface with buff using a monocrystalline diamond suspension (9 µm) and a colloidal silica suspension (0.04 µm). The XRD measurement was performed with the X-ray source of CuKα, the accelerating voltage of 40 kV, the current of 40 mA using a Bruker D8 ADVANCE X-ray diffractometer. For SEM and TEM observations, the disks of both surfaces were mechanically ground with waterproof SiC paper (up to P1000) to a thickness of 0.2 mm. The disks were punched out to small disks with a diameter of 3 mm. They were polished with buff using a diamond suspension of 9 µm and colloidal silica suspension of 0.04 µm to a thickness of 0.1 mm. For further thinning, the electrolytic polishing was performed with a twin-jet electrochemical polisher (Tenu-pol5, Struers) using a solution of HClO₄:CH₄(CH₂)₃OH:CH₃OH = 5:35:60 (vol%) at 243 to 253 K under an applied voltage of 20 V. SEM was carried out at 15 kV using JEOL FE-SEM 7800F. The volume fraction of α phase of the alloy before HPT was calculated from pixel information of a binary image by SEM. For EBSD, ion milling was performed on small disks with 3 mm diameter and 0.1 mm thickness using GATAN PIPS 681 at an accelerating voltage of 4 kV and an incident angle of 4°. The crystal orientation was analyzed using EBSD with a software provided by TSL Solutions. TEM was operated at 100 or 200 kV for microstructural observation and recording the selected-area electron diffraction (SAED) patterns using a HITACHI H-7100 or a JEOL JEM-2100. The SAED pattern was taken from entire bright field image. The grain size was measured from TEM images by taking an arithmetic mean value of the major and minor diameters of the grains.

2.4 Mechanical properties evaluation

The mechanical properties were evaluated by Vickers microhardness measurement and tensile tests. For the hardness measurement, the HPT-processed disks were polished to mirror-like surfaces using the same procedures as for the XRD measurement. The hardness measurement was performed with an applied load of 2.942 N for 15 s using SHIMADZU HMV-1 along the 12 radial directions from the center to the edge of each disk in an incremental step of 0.5 mm (see Fig. 1). The average value of hardness was taken from 12 measurements at equal distances from the disk center. For the tensile tests, the HPT-processed disks were first ground to thicknesses of 0.6 mm, and tensile specimens were cut from the disks at a position 2 mm away from the disk center using an electrical discharge machine (ROBOCUT α-C400iA, FANUC). Figure 1 shows the tensile specimen with gauge length of 1.5 mm and the gauge section of 0.7 mm width and 0.6 mm thickness. The specimens were carefully polished with a SiC waterproof paper of P2000. Each tensile specimen was mounted horizontally on the grips and pulled to fracture at room temperature with an initial strain rate of 2 × 10⁻³ s⁻¹ using a tensile testing machine (DT-05, Sagawa Manufacturing) (n = 1).

Fig. 1

Schematic of sample size and evaluated area.

3. Results and Discussions

3.1 Initial microstructures before HPT processing

Figures 2 and 3 shows SEM images together with invers pole figure (IPF) maps and XRD profiles before HPT processing, respectively. From these results, the as-received specimen consists of α and β phases with equiaxed structures, as shown in Fig. 2(a), where the grain size and volume fraction of the α phase were ∼5 µm and 94%, respectively. The microstructure of the STA specimen is duplex as shown in Fig. 2(b), which consists of an equiaxed α phase and lamellar α + β phases with a grain size of α grains of ∼10 µm and a volume fraction of α grains of ∼75%. For the STQ < T_β specimen shown in Fig. 2(c), the microstructure is duplex, consisting of equiaxed α-phase with a grain size of ∼5 µm and a needle-like martensitic structure. The STQ > T_β specimen in Fig. 2(d) shows only a needle-like structure. TEM observation confirmed that the needle-like martensitic structure in both STQ < T_β and STQ > T_β was the α′ phase because $\{ 10\bar{1}1\} $ twins formed within the needle-like structure as shown in Fig. 4.¹⁶⁾

Fig. 2

Scanning electron microscopy (SEM) images and inverse pole figure (IPF) maps of initial microstructures of (a) as-received, (b) STA, (c) STQ < T_β, and (d) STQ > T_β of Ti–6Al–7Nb alloy.

Fig. 3

X-ray diffraction (XRD) patterns of initial microstructures of as-received, STA, STQ < T_β, and STQ > T_β of Ti–6Al–7Nb alloy before high-pressure torsion (HPT).

Fig. 4

Transmission electron microscopy (TEM) image of STQ > T_β of Ti–6Al–7Nb alloy before high-pressure torsion (HPT).

3.2 Mechanical properties before and after HPT processing

Figure 5 shows the Vickers microhardness plotted against the distance from the disk center for the (a) as-received, (b) STA, (c) STQ < T_β, and (d) STQ > T_β specimens before and after HPT processing. The average hardness values before the HPT processing were 310 HV, 340 HV, and 355 HV in the STA, STQ < T_β, STQ > T_β, respectively. The hardness of the specimens with STQ treatment before the HPT processing was larger than 325 HV of the specimens in the as-received condition due to the formation of the martensitic α′ phase. The hardness values increased by the HPT processing with increasing the number of revolutions for all the initial microstructures. The maximum values were 395 HV, 400 HV, 395 HV, and 395 HV in the as-received, STA, STQ < T_β, and STQ > T_β specimens, respectively, and all of them were achieved after 20 revolutions. The Vickers microhardness of all specimens increased in early stages and reached saturation with increasing the equivalent strain.

Fig. 5

Vickers microhardness versus distance from disk center of (a) as-received, (b) STA, (c) STQ < T_β, (d) STQ > T_β of Ti–6Al–7Nb alloy before and after high-pressure torsion (HPT).

Figure 6 shows the stress-strain curves before and after the HPT processing. The ultimate tensile strength (UTS) and the elongation to fracture before and after the HPT processing are summarized in Table 1 and Fig. 7. UTS increased by the HPT processing regardless of all the initial microstructures. The elongation to fracture decreased after the HPT processing with an increasing number of revolutions. In particular, the elongation to fracture (total elongation) significantly decreased after more than 5 revolutions, thereby indicating that brittle fracture occurred. However, the total elongation of the as-received, STA, and STQ < T_β specimens after 1 revolution were almost the same as those before the HPT processing. A good combination of UTS with 1280 MPa and total elongation with 22% was achieved in the STQ < T_β specimen after one revolution.

Fig. 6

Stress-strain curves of (a) as-received, (b) STA, (c) STQ < T_β, (d) STQ > T_β of Ti–6Al–7Nb alloy before and after high-pressure torsion (HPT).

Table 1 Tensile properties of ultimate tensile strength (σ_UTS) and elongation to fracture (l) for non-heat treated (as-received) and heat-treated (STA, STQ < T_β, STQ > T_β) Ti–6Al–7Nb alloy before and after high-pressure torsion (HPT).

Fig. 7

Tensile properties of as-received, STA, STQ < T_β, STQ > T_β of Ti–6Al–7Nb alloy before and after high-pressure torsion (HPT).

3.3 Microstructure after HPT processing

3.3.1 TEM observations

Figures 8–11 show TEM micrographs of the as-received, STA, STQ < T_β, and STQ > T_β specimens after HPT processing for (a) 5 and (b) 20 revolutions, respectively. The dark-field images were taken with the diffraction indicated by the arrow in the SAED pattern. After HPT processing, ultrafine grains with sizes of several hundred nanometers were visible in some areas of the disks, as shown in Figs. 8–11. While coarse grains with a few micrometers in size cover the majority of the as-received, STA, and STQ < T_β samples until 5 revolutions, as shown in Fig. 12. The SAED pattern from the region of ultrafine grains showed a ring-type pattern in the samples after HPT processing. This indicates that the grain size became smaller with random orientations. However, the SAED pattern of the coarse grains region showed a spot-like diffraction pattern in some areas. This suggests that the low-angle grain boundaries formed in the coarse grains. Thus, the microstructure was inhomogeneous until 5 revolutions, consisting of ultrafine grains and coarse grains. However, the fraction of the coarse grains decreased with an increasing number of revolutions. Then, the microstructure was homogeneous after 20 revolutions, consisting only of ultrafine grains.

Fig. 8

Transmission electron microscopy (TEM) images of as-received Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for (a) 5 revolutions⁹⁾ and (b) 20 revolutions.

Fig. 9

Transmission electron microscopy (TEM) images of STA-treated Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for (a) 5 revolutions⁹⁾ and (b) 20 revolutions.

Fig. 10

Transmission electron microscopy (TEM) images of STQ < T_β-treated Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for (a) 5 revolutions and (b) 20 revolutions.

Fig. 11

Transmission electron microscopy (TEM) images of STQ > T_β-treated Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for (a) 5 revolutions and (b) 20 revolutions.

Fig. 12

Transmission electron microscopy (TEM) images of STA-treated Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for 5 revolutions.

The minimum grain size was reached after 20 revolutions and was ∼90, ∼70 and for the α phase in the as-received and STA specimens, respectively, and ∼70 nm, ∼80 nm for the α′ phase in the STQ < T_β and STQ > T_β specimen. Furthermore, the grain sizes for the β phase in the as-received and STA specimens were ∼65 nm and ∼40 nm, respectively, which were smaller than those of the α grains in both the as-received and STA specimens.

3.3.2 SEM observations

Figure 13 shows IPF maps from EBSD analysis for the as-received and STA specimens after HPT processing for 1 and 5 revolutions. It was difficult to observe the microstructures by EBSD analysis for the STQ < T_β, and STQ > T_β specimens after 20 revolutions because of intense internal strain introduced by severe deformation through the HPT processing. The dark regions in the IPF maps correspond to areas where imaging was difficult because the grain size is too small to be detected and/or the area is too strained to be analyzed. The IPF maps in both specimens indicate that the grain size decreases with increasing number of HPT revolutions. This is consistent with the TEM observations shown in Fig. 8–9. Close inspection reveals that the color of IPF maps changes within the same grains, indicating that low-angle grain boundaries formed within the grains.

Fig. 13

Inverse pole figure (IPF) maps of non-heat treated (as-received) and heat-treated (STA) Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for 1 and 5 revolutions.

From the IPF map of the STQ > T_β specimen, the needle-like structure was fragmented into several grains, as indicated by white arrows in Fig. 14 (STQ > T_β, N = 1). There was almost no difference in the crystal orientation of several fragmented grains.

Fig. 14

Inverse pole figure (IPF) maps of STQ > T_β-treated Ti–6Al–7Nb alloy after high-pressure torsion (HPT) for 1 revolution.

3.4 Developing the resultant microstructure and mechanical properties

Despite the fact that the hardness before HPT processing was different in each initial microstructure, the maximum values of hardness were similar to each other after the HPT processing. The reason for this is considered because the grain sizes after 20 revolutions were approximately the same in all conditions.

As a result of EBSD analysis and TEM observations after processing for a small number of revolutions (i.e., small equivalent strain), the size of coarse grains was 1 µm or more, resulting in an inhomogeneous microstructure. However, as the equivalent strain increases, the grain size becomes finer to 100 nm or less, resulting in a homogeneous microstructure comprising only ultrafine grains. Furthermore, from the results of crystal orientation analysis by EBSD, it was found that the evolution of grain refinement differs between the equiaxed structure, lamellar structure, and needle-like structure even if the strain amount by HPT processing is the same. It is considered that the ultrafine grains observed by TEM are mainly grains refined from the lamellar structures for the STA specimen and the needle-like structures for the STQ < T_β and STQ > T_β specimens. Microstructural development is more likely to be inhomogeneous when the number of revolutions is less in materials having an HCP structure such as pure Ti, Ti alloy,^17,18) Zr¹⁹⁾ and Mg alloy with a duplex structure.²⁰⁾ Three factors may contribute to the formation of such inhomogeneous microstructures. First, the HCP structure has a limit for the slip system. Second, α-Ti has a large critical shear stress at room temperature. Third, because the HPT processing involves shear deformation in only one direction, the amount of strain introduced differs depending on the initial crystal orientation, even if the crystal structure is the same. In addition, the plastic deformation of the HCP structure generally occurs due to dislocation slip and twin formation; however, it has been reported that dislocation slip is more likely to occur during the HPT processing than twin formation.²¹⁾

According to the IPF map for the STQ specimen as shown in Fig. 14, the needle-like structure was fragmented when subjected to HPT processing. The needle-like structure has a high residual stress before the HPT processing and consists of a high volume fraction of initial defects, which lead to fast grain fragmentation with formation of subgrains.^22,23) The grain size becomes small with increasing the volume fraction of lamellae structure because the boundaries act as nucleation sites for grain fragmentation.²⁴⁾ Therefore, the tensile strength after HPT becomes high when the initial microstructure has a lamellae structure or a needle-like structure.

From the TEM images of Fig. 12, the coarse grains from the equiaxed structure are existed after HPT processing with a small number of revolutions. The decrease of elongation to fracture was small when HPT was processed for 1 revolution if the initial microstructure has the equiaxed structure, as shown in Figs. 6(a)–(c). The inhomogeneous structure is effective for a good balance of strength and ductility because soft regions from large grains deform plastically more than hard regions from small grains²⁵⁾ and play a big role in keeping the elongation. Therefore, it is considered that a better balance of tensile properties was achieved in STQ < T_β with a duplex structure comprising equiaxed and needle-like structures after HPT processing for 1 revolution.

As shown in Fig. 7, the elongation could be maintained at a low number of revolution, but brittle fracture was shown at 5 rotations or more. The cross-section of the specimens after HPT processing was observed to investigate the cause of these brittle fractures. The results are shown in Fig. 15. The cracks, as shown by the arrows, were observed in the cross section. These cracks were confirmed in all specimens in which the elongation to fracture was significantly decreased due to low cold workability of Ti alloy. These cracks were not observed in the specimens before the HPT processing. Therefore, it is considered that the cracks occurred during HPT processing, and brittle fracture occurred because of the propagation of this crack.

Fig. 15

SEM images of cross section of Ti–6Al–7Nb alloy (STA) after high-pressure torsion (HPT) for (a) 5 revolutions and (b) 20 revolutions.

4. Conclusions

In this study, a Ti–6Al–7Nb alloy was subjected to three types of heat treatment and was processed by HPT including the alloy without heat treatment. The resultant microstructures and mechanical properties of the alloys were investigated. The conclusions are drawn as follows:

1. (1)    When the HPT processing was applied to the Ti–6Al–7Nb alloy with various initial microstructures, the grain size was refined to ultrafine grains with sizes less than 100 nm, which was almost the same regardless of the initial microstructures.
2. (2)    By HPT processing, the tensile strength increased while maintaining higher elongations to fracture. The duplex microstructure comprising equiaxed and needle-like structures processed by the HPT for 1 revolution exhibited a better combination of the ultimate tensile strength and the elongation to fracture, which were 1280 MPa and 22%, respectively.
3. (3)    The grain refinement by fragmentation of the needle-like structure and inhomogeneous microstructure formed after HPT processing with a small number of revolutions in a duplex structure comprising equiaxed and needle-like structures is effective for enhancing the tensile strength while maintaining elongation to fracture.
4. 　  

Therefore, a duplex microstructure with an equiaxed structure and a needle-like structure and an inhomogeneous microstructure after HPT processing with heat treatment is useful for increasing the balance of mechanical properties for application in medical devices.

Acknowledgments

The authors are grateful to Prof. Zenji Horita (Kyushu Institute of Technology, Japan) for providing the samples and helpful discussions during the experiment. This work was supported by the Japan Science Promotion Society (Grant-in-Aid for Early-Career Scientists, Grant Number JP21K17007). A part of this work was carried out under the GIMRT Program of the Institute for Materials Research, Tohoku University. This work was also supported by Light Metal Educational Foundation, Inc. Part of this research is based on the Cooperative Research Project of the Research Center for Biomedical Engineering.

REFERENCES

• 1)   E.  Kobayashi,  T.J.  Wang,  H.  Doi,  T.  Yoneyama and  H.  Hamanaka: J. Mater. Sci. Mater. Med. 9 (1998) 567–574. doi:10.1023/A:1008909408948
• 2)   M.  Niinomi: Mater. Sci. Eng. A 243 (1998) 231–236. doi:10.1016/S0921-5093(97)00806-X
• 3)   M.F.  Semlitsch,  H.  Weber,  R.M.  Streicher and  R.  Schön: Biomaterials 13 (1992) 781–788. doi:10.1016/0142-9612(92)90018-J
• 4)   B.R.  Chrcanovic,  J.  Kisch,  T.  Albrektsson and  A.  Wennerberg: Clin. Implant Dent. Relat. Res. 20 (2018) 58–67. doi:10.1111/cid.12572
• 5)   R.Z.  Valiev,  Y.  Estrin,  Z.  Horita,  T.G.  Langdon,  M.J.  Zechetbauer and  Y.T.  Zhu: JOM 58 (2006) 33–39. doi:10.1007/s11837-006-0213-7
• 6)   K.  Edalati and  Z.  Horita: Mater. Sci. Eng. A 652 (2016) 325–352. doi:10.1016/j.msea.2015.11.074
• 7)   M.  Ashida,  P.  Chen,  H.  Doi,  Y.  Tsutsumi,  T.  Hanawa and  Z.  Horita: Mater. Sci. Eng. A 640 (2015) 449–453. doi:10.1016/j.msea.2015.06.020
• 8)   P.  Chen,  M.  Ashida,  H.  Doi,  Y.  Tsutsumi,  Z.  Horita and  T.  Hanawa: Mater. Trans. 57 (2016) 2020–2025. doi:10.2320/matertrans.MI201515
• 9)  M. Ashida, T. Hanawa, P. Chen, Y. Tsutsumi and Z. Horita: The 14th World Conference on Titanium, MATEC Web Conf. 321 (2020) 12002.
• 10)   T.S.  Pinheiro,  J.  Gallego,  C.  Bolfarini,  C.S.  Kiminami,  A.M.  Jorge, Jr. and  W.J.  Botta: Int. J. Mater. Res. 15 (2012) 792–795. doi:10.1590/S1516-14392012005000102
• 11)   M.  Janeček,  J.  Stráský,  J.  Čížek,  P.  Harcuba,  K.  Václavová,  V.V.  Polyakova and  I.P.  Semenova: Metall. Mater. Trans. A 45 (2014) 7–15. doi:10.1007/s11661-013-1763-2
• 12)   K.  Watanabe,  M.  Ashida,  T.  Masuda,  P.  Kral,  Y.  Takizawa,  M.  Yumoto,  Y.  Otagiri,  V.  Sklenicka,  T.  Hanawa and  Z.  Horita: Mater. Trans. 60 (2019) 1785–1791. doi:10.2320/matertrans.ME201924
• 13)   J.M.  Cubero-Sesin,  J.E.  Gonzalez-Hernandez,  E.  Ulate-Kolitsky,  K.  Edalati and  Z.  Horita: Mater. Sci. Eng. 194 (2017) 012041. doi:10.1088/1757-899X/194/1/012041
• 14)   J.  Fu,  H.  Ding,  Y.  Huang,  W.  Zhang and  T.G.  Langdon: J. Mater. Res. Technol.-JMRT 4 (2015) 2–7. doi:10.1016/j.jmrt.2014.10.006
• 15)   Y.C.  Wang and  T.G.  Langdon: Mater. Sci. Eng. A 559 (2013) 861–867. doi:10.1016/j.msea.2012.09.034
• 16)   H.  Matsumoto,  H.  Yoned,  K.  Sato,  S.  Kurosu,  E.  Maire,  D.  Fabregue,  T.J.  Konno and  A.  Chiba: Mater. Sci. Eng. A 528 (2011) 1512–1520. doi:10.1016/j.msea.2010.10.070
• 17)   K.  Edalati,  E.  Matsubara and  Z.  Horita: Metall. Mater. Trans. A 40 (2009) 2079–2086. doi:10.1007/s11661-009-9890-5
• 18)   Y.  Todaka,  M.  Umemoto,  A.  Yamazaki,  J.  Sasaki and  K.  Tsuchiya: Mater. Trans. 49 (2008) 47–53. doi:10.2320/matertrans.ME200714
• 19)   Y.B.  Wang,  M.  Louie,  Y.  Cao,  X.Z.  Liao,  H.J.  Li,  S.P.  Ringer and  Y.T.  Zhu: Scr. Mater. 62 (2010) 214–217. doi:10.1016/j.scriptamat.2009.10.034
• 20)   J.  Čížek,  I.  Procházka,  B.  Smola,  I.  Stulíková,  R.  Kužel,  Z.  Matěj,  V.  Cherkaska,  R.K.  Islamgaliev and  O.  Kulyasova: Mater. Sci. Eng. A 462 (2007) 121–126. doi:10.1016/j.msea.2006.01.177
• 21)   Y.J.  Chen,  Y.J.  Li,  J.C.  Walmsley,  N.  Gao,  H.J.  Roven,  M.J.  Starink and  T.G.  Langdon: J. Mater. Sci. 47 (2012) 4838–4844. doi:10.1007/s10853-012-6343-0
• 22)   H.  Shahmir and  T.G.  Langdon: Mater. Sci. Eng. A 704 (2017) 212–217. doi:10.1016/j.msea.2017.07.099
• 23)   Q.  Chao,  P.D.  Hodgson and  H.  Beladi: Metall. Mater. Trans. A 45 (2014) 2659–2671. doi:10.1007/s11661-014-2205-5
• 24)   P.  Castany,  F.  Pettinari-Sturmel,  J.  Douin and  A.  Coujou: Mater. Sci. Eng. A 483–484 (2008) 719–722. doi:10.1016/j.msea.2006.10.183
• 25)   E.  Ma and  T.  Zhu: Mater. Today 20 (2017) 323–331. doi:10.1016/j.mattod.2017.02.003