Kink Formation and Strengthening Effects in TiNi–V Eutectic Alloys with Mille-Feuille Structure

Abstract

Kink formation by rolling and its strengthening effects on the mechanical properties of TiNi–V alloy were investigated. The as-cast TiNi–V alloy had a eutectic lamellar structure consisting of B2-TiNi and bcc-V phases. When the alloy was rolled, kinks were formed in the layered structure. The Vickers hardness of the alloy increased after it was subjected to 30% rolling reduction, and then decreased with increasing annealing temperature. After the rolled alloy was annealed at 973 K, its hardness recovered to the original value; however, the layer structure and kinks remained in the alloy. Thus, both dislocations and kinks were introduced into the alloy by rolling, but only dislocations disappeared upon annealing. Therefore, a layered TiNi–V alloy with only kinks but no dislocations was obtained. A comparison of the yield stresses of as-cast, 30% rolled, and annealed alloy specimens revealed that the magnitude of strengthening by dislocations and kinks was about 170 and 40 MPa, respectively. Therefore, kink formation and kink strengthening effects similar to those observed in Mg-based long-period stacking ordered (LPSO) alloys were observed.

Fig. 6 Stress-strain curves for (a) as-cast, (b) 30% rolled and (c) 30% rolled and annealed (973 K) B2+bcc two-phase Ti₃₅Ni_39.5V_25.5 alloys (d) B2-TiNi single-phase Ti₄₉Ni₄₇V₄ alloy and (e) bcc-V single-phase Ti₈Ni₆V₈₆ alloy.

1. Introduction

Recently, Kawamura et al. reported high strength Mg–Zn–Y alloys, that have attracted attention for their potential application in Mg-based structural alloys.¹⁾ One of the origins of this remarkable high strength is the presence of a long-period stacking ordered (LPSO) phase. In the LPSO phase, closed-packed Mg layers and Zn, Y condensed layers are stacked periodically along the c-axis.^2,3) Therefore, the LPSO phase has a layer structure that consists of two different layers in its unit cell, which is the “mille-feuille” structure. The LPSO phase is often observed in Mg–Tm–Re (Tm: transition metal, Re: rare-earth metal) systems.^4–7) When an LPSO alloy is deformed, kink formation is observed in its layer structure; and these kinks play an important role in improving the mechanical properties of Mg-based alloys. This “kink strengthening” effect is considered a new strengthening method for metallic alloys similar to the methods of solid solution strengthening, grain boundary strengthening, precipitation strengthening and work hardening by dislocations.

Although a layered structure is formed in the LPSO single-phase alloy, another layered structure consisting of two different phases is often observed in many alloy systems. For example, a lamellar structure is formed spontaneously during solidification if the alloy is solidified through an eutectic reaction. Hagihara et al. reported that a layered structure can be obtained by directional solidification in Mg/Mg₁₇Al₁₂,⁸⁾ Mg/Mg₂Yb,⁹⁾ Mg/Mg₂Ca⁹⁾ and Al/Al₂Cu¹⁰⁾ eutectic alloys. When stress is applied parallel to the lamellar interface, a kink band is induced in these alloys. Ueji et al. also reported that kinking occurs in a pearlitic lamellar structure of ferrite/cementite formed through a eutectoid reaction.¹¹⁾ A morphological layered structure also exists. For example, Cu–Nb¹²⁾ and Al–Cu¹³⁾ nanolaminates can be obtained via the accumulative roll bonding (ARB) method, and kinks are formed during compression. In addition, Zhu et al. reported that a lamellar structure of α/β-Ti can be formed by a rolling and annealing technique, and kinks are formed after deformation.¹⁴⁾ Therefore, a mille-feuille structure can be obtained in various alloys using various methods. Whether kink formation in these alloys enhances their mechanical strength is an interesting topic that warrants further investigation.

Non-Pd hydrogen separation and purification alloys are currently strongly desired to enable the generation of large amounts of hydrogen gas at low cost. Although 5A elements such as V, Nb and Ta are promising candidates for non-Pd hydrogen permeation alloys, they are susceptible to severe hydrogen embrittlement. For compatible hydrogen permeability and resistance to hydrogen embrittlement, a TiNi phase is added to Nb, resulting in the formation of a TiNi+Nb two-phase structure.¹⁵⁾ This structure exhibits a hydrogen permeability equivalent to that of pure Pd while exhibiting efficient resistance to hydrogen embrittlement under a hydrogen atmosphere. In TiNi–Nb alloys, a layer structure consisting of B2-TiNi and bcc-Nb phases is formed during eutectic solidification. On the basis of this concept, hydrogen permeable alloys can be obtained in TiNi–V,¹⁶⁾ TiNi–Ta,¹⁶⁾ TiCo–Nb^17,18) and ZrNi–Nb^19,20) systems, and eutectic layer structure is observed in these alloys.

The hydrogen flux J passing through an alloy membrane is expressed as;   

\begin{equation} J = \varPhi\cdot (P_{\text{u}}^{0.5} - P_{\text{d}}^{0.5})/L \end{equation} (1)

where Φ and L are the hydrogen permeability and thickness of the alloy, and P_u and P_d are the hydrogen pressures in the upstream and downstream sides, respectively. To achieve large J values, thin hydrogen permeation alloys are used under a large hydrogen pressure difference. Moreover, thin alloy membranes are usually prepared by forging and rolling. Therefore, in addition to high hydrogen permeability, good mechanical properties are also important for hydrogen permeation alloys.

The formation of kinks and their contribution to mechanical properties have recently been investigated in TiNi–Nb alloys.²¹⁾ When a TiNi–Nb alloy is rolled at room temperature, kinks are formed in the layer structure. Although both dislocations and kinks are introduced in the alloy during deformation, an alloy having a layer structure with kinks but no dislocations can be obtained when the rolled alloy is annealing at 923 K for 1 h. Subsequent tensile tests demonstrated that kink strengthening enhances the tensile strength of the alloy by about 50 MPa. In the present work, we focus on TiNi–V alloys. They are designed based on the same concept as TiNi–Nb and other alloys, that is, TiNi is added to V for compatible hydrogen permeability and resistance to hydrogen embrittlement. They also show good hydrogen permeability similar to TiNi–Nb alloys, especially superior hydrogen permeability at lower temperatures compared to TiNi–Nb alloys.¹⁶⁾ Furthermore, they have a layer structure consisting of B2-TiNi and bcc-V phases in the as-cast state and show good ductility at room temperature similar to TiNi–Nb alloys. Thus, kink formation and kink strengthening effects in the TiNi–V alloys are expected to be similar to those observed in TiNi–Nb alloys. Here, we investigate the microstructural changes and thermal stability of the layer structure and kinks and estimate the strengthening effects by dislocations and kinks quantitatively.

2. Experimental Procedure

The Ti₃₅Ni_39.5V_25.5 (mol%) ingots were prepared by arc melting under an Ar atmosphere using pure elements of Ti (99.5 mass%), Ni (99.9 mass%) and V (99.7 mass%). The ingots were melted several times for macroscopic homogenization. The sample plates were cut from the ingots using an electric discharge machine, and were rolled at room temperature with a 30% thickness reduction. The samples wrapped by Mo foils were enclosed in a quartz tube filled with pure Ar gas with Zr foil to avoid oxidation and direct contact to quartz tube. The capsules were heated in an electric furnace at temperature range from 673 to 1173 K for 1 h, and quenched into a water. The surface of the sample was polished using abrasive paper and a buff. The sample for microstructural observation was soaked in an etchant consisting of water:nitric acid:hydrofluoric acid = 7:2:1. Microstructural observations were carried out by scanning electron microscopy (SEM). The phase structure of samples was identified by X-ray diffraction (XRD) analysis. The Vickers hardness of the samples was measured using a micro-Vickers hardness testing machine under a 0.5 N load for 10 s. The mechanical properties of the samples were measured using a tensile testing machine with an initial strain rate of 3.3 × 10⁻³ s⁻¹. Sample deformation was recorded by a digital video camera, and the strain generated in each sample was traced using a digital imaging correlation (DIC) method.

3. Results and Discussions

3.1 Kink formation and thermal stability of microstructure

Figures 1 show SEM micrographs of (a) as-cast and (b) 30% rolled Ti₃₅Ni_39.5V_25.5 alloys. This alloy shows a eutectic structure consisting of V (gray) and TiNi (white) phases in (a) the as-cast alloy with a phase pitch of a few hundreds nanometers. Both rod and lamellar types of eutectic structure were clearly observed, although only a lamellar-type eutectic structure was observed in the TiNi–Nb alloys.¹⁵⁾ After a 30% thickness reduction by rolling at room temperature, layer bending occurred in the sample (Fig. 1(b)). In particular, kinks were formed in the layers along normal direction (ND). In addition, this alloy exhibited efficient ductility and no peeling of the phase boundary.

Fig. 1

SEM micrographs of (a) as-cast and (b) 30% rolled Ti₃₅Ni_39.5V_25.5 alloys.

Figure 2 shows XRD patterns for (a) as-cast and (b) 30% rolled Ti₃₅Ni_39.5V_25.5 alloys. The Bragg peaks for (a) the as-cast alloy can be indexed as a TiNi phase having a B2 (ordered bcc) structure. The lattice parameters for TiNi and pure V phases have been reported as 0.3007 nm²²⁾ and 0.3031 nm,²³⁾ respectively. Thus, we considered that the Bragg peaks for the V phase overlapped to those for TiNi phase because these two phases have the same structure and similar lattice parameters. The XRD pattern for (b) the rolled alloy, as shown in Fig. 2(b), indicates that peak broadening occurred as a result of rolling and that a monoclinic TiNi phase was formed because of a stress-induced martensitic transformation.

Fig. 2

XRD patterns for (a) as-cast (b) 30% rolled Ti₃₅Ni_39.5V_25.5 alloys.

Dislocations are well known to also be introduced to the alloys during plastic deformation, which increases the yield strength of the alloy. We speculated that the TiNi–V alloy is strengthened by the introduction of both dislocations and kinks generated during rolling. In order to separate the effects of dislocations and kinks on the mechanical strength and to understand the thermal stability of the microstructure, the 30% rolled alloy was annealed at various temperatures for 1 h. Figure 3 shows the change in Vickers hardness (H_v) as a function of the annealing temperature. The hardness of the as-cast alloy was 348 and increased to 399 after rolling, which suggests that this alloy was strengthened by the introduction of both dislocations and kinks. The hardness of the alloy decreased with increasing annealing temperature. After the alloy was annealed at 973 K, its hardness recovered to approximately the same value for the as-cast alloy. Annealing at higher temperatures led to further reductions of the hardness.

Fig. 3

Change in Vickers hardness for 30% rolled Ti₃₅Ni_39.5V_25.5 alloy after annealing for 1 h at various temperature.

Figures 4 show SEM micrographs of the 30% rolled Ti₃₅Ni_39.5V_25.5 alloy annealed at (a) 973 and (b) 1073 K, respectively. In the sample annealed at 973 K (a), the layer structure and kinks remained in the alloy. However, in the sample annealed at 1073 K, most of layer structure had disappeared and isolated granular TiNi and V phases were formed. The XRD patterns for these two alloys are shown in Figs. 5(a) and (b). These two alloys consisted of both TiNi and V phases after annealing. In addition, neither peak broadening due to strain introduction nor the formation of a monoclinic TiNi phase during rolling was not observed after annealing. These experimental results indicate both dislocations and kinks are introduced in the alloy because of rolling, which causes an increase in its hardness. In the sample annealed at 973 K, only the dislocations disappeared; thus, an alloy having a lamellar structure that includes kinks but no dislocations can be obtained via this rolling and annealing process.

Fig. 4

SEM micrographs of Ti₃₅Ni_39.5V_25.5 alloys annealed at (a) 973 K and (b) 1073 K after the alloy were subjected to 30% rolling reduction.

Fig. 5

XRD patterns for Ti₃₅Ni_39.5V_25.5 alloys annealed at (a) 973 K and (b) 1073 K after the alloy were subjected to 30% rolling reduction.

3.2 Quantitative estimation of the strengthening effects by dislocations and kinks

In order to distinguish between strengthening effects caused by dislocations and kinks, we subjected the prepared alloys to tensile tests. Figure 6 shows the nominal stress-strain (ss) curves of (a) as-cast, (b) 30% rolled and (c) annealed (973 K) after rolling, respectively. Because the yield stress was not observed clearly in the ss curves, the yield stress is defined as the stress at which the plastic strain reaches 0.2%, so-called 0.2% proof stress. The yield stress and fracture strain for the (a) as cast alloy were 512 MPa and 5%, respectively. After 30% rolling reduction, the yield stress increased to 722 MPa, which suggests that this alloy was strengthened by the introduction of both dislocations and kinks. However, the fracture stress was approximately the same value as that for as-cast alloy. After the annealing at 973 K, the yield stress decreased to 556 MPa, and the fracture strain increased to 9% because dislocations disappeared during annealing. After annealing, fracture strain of this alloy increased from 5% for the as-cast alloy. The reason for improvement of ductility by rolling and annealing is not perfectly clear. It is considered that micro cracks, micro voids and segregation of impurities formed during solidification in the as-cast alloy disappeared by rolling and annealing process.

Fig. 6

Stress-strain curves for (a) as-cast, (b) 30% rolled and (c) 30% rolled and annealed (973 K) B2+bcc two-phase Ti₃₅Ni_39.5V_25.5 alloys (d) B2-TiNi single-phase Ti₄₉Ni₄₇V₄ alloy and (e) bcc-V single-phase Ti₈Ni₆V₈₆ alloy.

Because the rolled alloy was strengthened by phase boundaries, dislocations and kinks, the contributions of these two factors must be separated quantitatively. Hughes et al. reported that the improvement in strength for pure Ni as a result of grain boundaries and dislocations can be added to the original strength of defect-free Ni.²⁴⁾ This law is often used for quantitative analysis of the mechanical strength for many materials.^25–28) Assuming that same method can be applied to rolled TiNi–V alloy strengthened by phase boundaries, dislocations and kinks, the yield stress for a rolled alloy ($\sigma_{\text{y}}^{\text{rolled}}$) can also be expressed by the following equation;   

\begin{equation} \sigma_{\text{y}}^{\text{rolled}} = \sigma_{\text{a}} + \Delta \sigma_{\text{pb}} + \Delta \sigma_{\text{d}} + \Delta \sigma_{\text{k}} \end{equation} (2)

where σ_a is the virtual strength for the TiNi–V alloy forming a single-phase structure. The strengthening effects by phase boundaries, dislocations and kinks are defined as Δσ_pb, Δσ_d and Δσ_k, respectively. The parameter σ_a can be estimated by the following equation:   

\begin{equation} \sigma_{\text{a}} = \sigma_{\text{y}}^{\text{TiNi}}\cdot V^{\text{TiNi}} + \sigma_{\text{y}}^{\text{V}}\cdot V^{\text{V}} \end{equation} (3)

where $\sigma_{\text{y}}^{\text{TiNi}}$ and $\sigma_{\text{y}}^{\text{V}}$ are the yield stresses for TiNi and V single-phase alloys, and V^TiNi and V^V are their volume fractions in a mille-feuille structure. The chemical compositions of the TiNi and V phases in the Ti₃₅Ni_39.5V_25.5 alloy are Ti₄₉Ni₄₇V₄ and Ti₈Ni₆V₈₆ (mol%), respectively.²⁹⁾ These two single-phase alloys were prepared by arc melting, and test samples were cut from the alloy ingot. Their mechanical properties were measured by a tensile test under the same conditions for two-phase alloys. The ss curves for (d) TiNi single-phase and (e) V single-phase alloys are shown in Figs. 6. The yield stress and fracture strain for the TiNi single-phase alloy were 162 MPa and 7%, respectively. However, the V single-phase alloy did not exhibit plastic strain; thus, the yield stress for this alloy could not be determined. Therefore, the value of σ_a for the TiNi–V alloy was not obtained in the present research. Because the yield stress for the as-cast alloy ($\sigma_{\text{y}}^{\text{as-cast}}$) can be expressed as σ_a + Δσ_pb, the yield stress for the rolled alloy can be expressed as;   

\begin{equation} \sigma_{\text{y}}^{\text{rolled}} = \sigma_{\text{y}}^{\text{as-cast}} + \Delta \sigma_{\text{d}} + \Delta \sigma_{\text{k}} \end{equation} (4)

The difference in yield stress between the as-cast and rolled alloys is assumed to correspond to Δσ_d + Δσ_k, because only the rolled alloy involved dislocations and kinks. The difference in yield stress between the rolled and annealed alloys is assumed to correspond to Δσ_k, because the annealed alloy does not contain dislocations. A comparison of the yield stress for the as-cast, rolled and annealed alloys leads to Δσ_d and Δσ_k of 166 and 44 MPa, respectively. In conclusion, kinks are formed in TiNi–V alloy as a result of rolling, and they improve the strength of the alloy. The magnitude of kink strengthening was approximately one-fourth that of dislocation strengthening.

Kink formation and strengthening effects have also been observed in TiNi–Nb alloy, which has a mille-feuille structure consisting of B2-TiNi and bcc-Nb phases.²¹⁾ Figure 7 compares each strengthening factors for 30% rolled TiNi–V and TiNi–Nb alloys. For the TiNi–Nb alloy, we obtain σ_a and Δσ_pb as 512 and 16 MPa, respectively; that is, $\sigma_{\text{y}}^{\text{as-cast}}$ is 528 MPa. This result means that the yield stress for virtual TiNi–Nb having a single-phase structure is 512 MPa, and that the alloy is strengthened by 16 MPa because of the introduction of phase boundaries forming a mille-feuille structure. The strengthening effects by dislocations and kinks are 206 and 49 MPa in the TiNi–Nb alloy, respectively, which are approximately the same values observed in the TiNi–V alloys. In these two alloys, the magnitude of the strengthening effects induced by dislocations is about 4 times greater than that of the strengthening effects induced by kinks. These two as-cast TiNi–V and TiNi–Nb alloys show good ductility, exhibiting greater than 5% plastic deformation. We speculated that deformation progresses mainly by the introduction and migration of dislocations, which leads to greater strengthening effects of dislocations compared with that of kinks.

Fig. 7

Comparison of each strengthening factors for 30% rolled Ti₃₅Ni_39.5V_25.5 and Ti₄₀Ni₄₁Nb₁₉ alloys. The mechanical strength of the 30% rolled alloy can be expressed as: basic strength of the as-cast alloy ($\sigma_{\text{y}}^{\text{as-cast}}$) with addition of the strengthening effects of dislocations (Δσ_d) and kinks (Δσ_k).

A similar eutectic microstructure has been reported to form not only in TiNi–V and TiNi–Nb systems but also in other alloy systems such as TiNi–Ta,¹⁶⁾ TiCo–Nb^17,18) and ZrNi–Nb.^19,20) Some have shown a good ductility in rolling reduction and tensile tests; thus, their mechanical properties are important because they are also candidates for use as non-Pd hydrogen separation and purification alloys. In addition, kink strengthening effects similar to those in TiNi–V and TiNi–Nb alloys are also observed in other alloy systems. In particular, we found that large kink strengthening effects can be estimated in TiCo–Nb alloys, which has a mille-feuille structure consisting of B2-TiCo and bcc-Nb phases. We plan to investigate the microstructure and mechanical properties of TiCo–Nb alloys, and characterize the relationship between kink strengthening effects, ductility, crystal orientation of constituting phases and phase boundary structure. We expected that the accumulation of basic experimental data will lead to clarification of the kink formation criteria and the kink strengthening mechanism.

4. Conclusion

1. (1)    A eutectic mille-feuille structure consisting of B2-TiNi and bcc-V phases is formed in the as-cast Ti₃₅Ni_39.5V_25.5 alloy, and kinks are formed by rolling at room temperature.
2. (2)    The hardness of this alloy increases after rolling, and decreases with increasing annealing temperature. After the rolled alloy is annealed at 973 K for 1 h, its hardness recovers to the original value and the mille-feuille structure and kinks are still observed. We could thus obtain a mille-feuille type alloy that includes kinks but no dislocations.
3. (3)    By comparing the yield stress of the as-cast, 30% rolled and annealed alloys, we estimate that the strengthening effects of dislocations and kinks increased the yield stress of the alloy by 170 and 40 MPa, respectively.

Acknowledgement

This study was supported by the grants from the Japan Society for the Promotion of Science (JSPS) KAKENHI for Scientific Research on Innovative Areas “MFS Materials Science, Nos. 19H05123 and 21H00097”.

REFERENCES

• 1)   Y.  Kawamura,  K.  Hayashi,  A.  Inoue and  T.  Masumoto: Mater. Trans. 42 (2001) 1172–1176. doi:10.2320/matertrans.42.1172
• 2)   E.  Abe,  Y.  Kawamura,  K.  Hayashi and  A.  Inoue: Acta Mater. 50 (2002) 3845–3857. doi:10.1016/S1359-6454(02)00191-X
• 3)   A.  Ono,  E.  Abe,  T.  Itoi,  M.  Hirohashi,  M.  Yamasaki and  Y.  Kawamura: Mater. Trans. 49 (2008) 990–994. doi:10.2320/matertrans.MC200763
• 4)   M.  Yamasaki,  T.  Anan,  S.  Yoshimoto and  Y.  Kawamura: Scr. Mater. 53 (2005) 799–803. doi:10.1016/j.scriptamat.2005.06.006
• 5)   Y.  Kawamura,  T.  Kasahara,  S.  Izumi and  M.  Yamasaki: Scr. Mater. 55 (2006) 453–456. doi:10.1016/j.scriptamat.2006.05.011
• 6)   S.  Zhang,  J.  Kim,  F.  Wang,  K.  Kang,  H.  Xue,  X.  Chen and  F.  Pan: Mater. Charact. 191 (2022) 112111. doi:10.1016/j.matchar.2022.112111
• 7)   Y.  Kawamura: Materia Japan 54 (2015) 44–49. doi:10.2320/materia.54.44
• 8)   K.  Hagihara,  K.  Hayakawa and  K.  Miyoshi: Mater. Sci. Eng. A 798 (2020) 140087. doi:10.1016/j.msea.2020.140087
• 9)   K.  Hagihara and  K.  Miyoshi: J. Magnesium Alloys 10 (2022) 492–500. doi:10.1016/j.jma.2021.02.006
• 10)   K.  Hagihara,  T.  Tokunaga,  K.  Nishiura,  S.  Uemichi and  S.  Ohsawa: Mater. Sci. Eng. A 825 (2021) 141849. doi:10.1016/j.msea.2021.141849
• 11)   R.  Ueji,  H.  Somekawa,  T.  Inoue and  T.  Hara: Mater. Sci. Eng. A 857 (2022) 144018. doi:10.1016/j.msea.2022.144018
• 12)   T.J.  Nizolek,  M.R.  Begley,  R.J.  McCabe,  J.T.  Avallone,  N.A.  Mara,  I.J.  Beyerlein and  T.M.  Pollock: Acta Mater. 133 (2017) 303–315. doi:10.1016/j.actamat.2017.04.050
• 13)   M.  Yamazaki,  K.  Ishikawa,  T.  Fujii and  Y.  Miyajima: Mater. Trans. 64 (2023) Published online. doi:10.2320/matertrans.MT-MD2022003
• 14)   J.  Zhu,  F.  Briffod,  T.  Shiraiwa,  M.  Enoki and  S.  Emura: Mater. Sci. Eng. A 849 (2022) 143538. doi:10.1016/j.msea.2022.143538
• 15)   W.  Wang,  K.  Ishikawa and  K.  Aoki: J. Membr. Sci. 351 (2010) 65–68. doi:10.1016/j.memsci.2010.01.029
• 16)   K.  Hashi,  K.  Ishikawa,  T.  Matsuda and  K.  Aoki: J. Alloy. Compd. 404–406 (2005) 273–278. doi:10.1016/j.jallcom.2005.02.085
• 17)   K.  Hashi,  K.  Ishikawa,  T.  Matsuda and  K.  Aoki: J. Alloy. Compd. 425 (2006) 284–290. doi:10.1016/j.jallcom.2006.01.028
• 18)   Y.  Saeki,  Y.  Yamada and  K.  Ishikawa: J. Alloy. Compd. 645 (2015) S32–S35. doi:10.1016/j.jallcom.2015.01.248
• 19)   T.  Takano,  K.  Ishikawa,  T.  Matsuda and  K.  Aoki: Mater. Trans. 45 (2004) 3360–3362. doi:10.2320/matertrans.45.3360
• 20)   K.  Ishikawa,  T.  Takano,  T.  Matsuda and  K.  Aoki: Appl. Phys. Lett. 87 (2005) 081906. doi:10.1063/1.2006985
• 21)   K.  Ohno,  Y.  Miyajima and  K.  Ishikawa: submitted to J. Alloy. Compd.
• 22)   H.  Stüwe and  Y.  Shimomura: Z. Metallk. 51 (1960) 180–181. doi:10.1515/ijmr-1960-510308
• 23)   Y.  Smirnov and  V.  Finkel: J. Exp. Theor. Phys. 22 (1966) 750–753.
• 24)   D.  Hughes and  N.  Hansen: Acta Mater. 48 (2000) 2985–3004. doi:10.1016/S1359-6454(00)00082-3
• 25)   M.  Ma,  Z.  Xiao,  X.  Meng,  Z.  Li,  S.  Gong,  J.  Dai,  H.  Jiang,  Y.  Jiang,  Q.  Lei and  H.  Wei: J. Mater. Sci. Technol. 112 (2022) 11–23. doi:10.1016/j.jmst.2021.08.080
• 26)   W.  Chen,  G.  Gao,  X.  Meng,  X.  Zhao,  Y.  Jiang,  M.  Wang,  Z.  Li and  L.  Xiao: J. Alloy. Compd. 912 (2022) 165170. doi:10.1016/j.jallcom.2022.165170
• 27)   W.  Chrominski,  S.  Wenner,  C.D.  Marioara,  R.  Holmestad and  M.  Lewandowska: Mater. Sci. Eng. A 669 (2016) 447–458. doi:10.1016/j.msea.2016.05.109
• 28)   Y.  Sun,  S.  Xu and  A.  Shan: Mater. Sci. Eng. A 641 (2015) 181–188. doi:10.1016/j.msea.2015.06.043
• 29)  K. Mori and K. Ishikawa: unpublished (private communication).