Texture Formation in a Polycrystalline Fe–Ni–Co–Al–Ti–B Shape Memory Alloy

Abstract

In polycrystalline Fe-Ni-Co-Al-based shape memory alloys, control of the recrystallization texture is significantly important to improve the ductility by suppressing the brittle precipitates of the β-B2 phase at grain boundaries during aging treatment. In this paper, the texture evolution in the recrystallization process was systematically investigated by means of the electron backscatter diffraction (EBSD) method in an Fe–Ni–Co–Al–Ti–B polycrystalline alloy. The development of a {110}<112> texture was confirmed in the 98.5% cold-rolled sheet specimen. After primary recrystallization annealing, the γ matrix containing the β phase showed the continuous recrystallization remaining in the same orientation with a deformation texture after annealing at 1000°C. Grain growth of the γ phase was interrupted by the β phase. Then, abnormal grain growth of {210}<001> grains began occurring in concurrence with the dissolution of the β phase and the main recrystallization texture changed from {110}<112> to {210}<001> at temperatures higher than 1100°C.

1. Introduction

Most physical and chemical properties of metallic materials are dependent on its crystallographic orientation.¹⁾ The orientation is recognized as one of the most important controlling factors for improving such features as functionality and reliability in industrial fields because of crystalline anisotropy. Moreover, in polycrystalline alloys, the difference in the crystal orientation of adjacent grains causes grain constraint¹⁾ and poses the introduction of dislocations due to the stress concentration near grain boundaries or triple junctions. In the case of shape memory alloys (SMAs), not only mechanical but also shape memory properties, including superelastic (SE) strain and critical stress for stress-induced martensitic transformation, are significantly affected by the direction of applied external fields because of anisotropy;^2,3,4) therefore, a number of studies on texture control have already been reported to maximize the capacities of SMAs in some alloy systems.^{5,6,7,8,9,10,11)}

Recently, Fe-based SMAs, such as Fe–Ni–Co–Al–(Ta, Nb or Ti)–^{12,13,14,15,16,17,18)} and Fe-Mn-Al-Ni-based alloys,^19,20) have been extensively investigated because of their advantages regarding low material cost and good workability, in addition to attractive SE properties. In Fe-Ni-based SMAs that accompany γ(fcc)/α’(bcc or bct) martensitic transformation, it has been calculated that the largest SE strain can be obtained in the [100] direction of the γ phase parallel to the tension and compression direction.^2,12) According to the research of Tanaka et al.,¹²⁾ a strong {530}<001> recrystallization texture has been formed by suitable thermomechanical heat treatment in an Fe–Ni–Co–Al–Ta–B polycrystalline alloy, resulting in a record SE strain of 13.5% and elongation of 20% at room temperature.

The low-temperature aging treatment is an indispensable process to precipitate the coherent γ’-L1₂ phase in the parent phase and to obtain thermoelastic martensitic transformation by the strengthening of the parent phase in Fe-Ni-Co-Al-based SMAs, while the composition of the matrix and the martensitic transformation temperatures change by the aging treatment.²¹⁾ The alloying elements, including Ta, Nb or Ti, are necessary for promotion of γ’ precipitation,²²⁾ among which Ti has a cost advantage. However, the grain boundary of these alloy systems becomes brittle during aging treatment due to the formation of the β-B2 phase along grain boundaries.^12,15,18) Therefore, in the case of polycrystalline alloys, it is necessary to suppress the precipitation of the β phase. From our previous research,¹²⁾ it has been confirmed that those grain boundary precipitates can be reduced by adding 0.05 at.% B, which segregate at grain boundaries and lower the grain boundary energy in Ni-based superalloys.²³⁾ Furthermore, the grain boundary precipitation can also be suppressed by texture control, which results in an increase in the frequency of low energy boundaries, such as low-angle and coincidence site lattice boundaries (CSLBs), coupled with the addition of B.^12,15,18,24) For that reason, the control of texture is critically important to enhance the SE properties and ductility in the Fe-Ni-Co-Al-based system with polycrystalline form.

In our previous paper,²⁵⁾ the effects of the cold-rolling ratio and heating rate on the texture of an Fe–Ni–Co–Al–Ti–B alloy was reported. It was found that the deformation texture of {110}<112> is developed for a reduction ratio higher than 80% and that the primary recrystallization texture at 1000°C is unchanged. It was reported that the secondary recrystallization (SRC) occurs at temperatures higher than 1100°C, in which the texture changes to {210}<001>. Superelasticity has been successfully obtained in a 90% cold-rolling and slowly heated sheet. However, the texture evolution from the primary to the SRC remains an issue for the texture control in this alloy system.

In this paper, texture formation, including the primary and secondary recrystallizations, was systematically investigated using a 98.5% cold-rolled Fe–Ni–Co–Al–Ti–B polycrystalline alloy.

2. Experimental Procedure

A columnar ingot of 20-mm diameter with a nominal alloy composition of Fe-30Ni-15Co-10Al-2.5Ti-0.05B (at.%) was prepared by induction melting under an Ar atmosphere. Hot-rolling was carried out to a thickness of 14 mm at 1200°C in the γ-single phase region, and the obtained sheets were solution-treated at 1200°C for 30 minutes, followed by water-quenching. Then, a sheet specimen with a thickness of 0.2 mm was prepared by cold-rolling with a reduction ratio of 98.5% after removal of the oxidation layer.

Small pieces cut from the cold-rolled sheet were directly heat-treated at different temperatures ranging from 400 to 1200°C under an Ar atmosphere. Some specimens were subjected to the heat treatment with a slow heating rate of 3°C/min from 900 to 1200°C, as shown in the schematic diagram of Fig. 1, to observe the microstructural evolution and change of crystal orientations during the heating process. Deformation and recrystallization texture were measured by means of the electron backscatter diffraction (EBSD) method.

Fig. 1.

Thermomechanical treatment of direct annealing or slow heating.

3. Results

3.1. Deformation Texture

Figure 2 shows the inverse pole figure (IPF) map for rolling direction (RD), transverse direction (TD) and normal direction (ND) taken from the vertical section of (a) an annealed specimen at 1200°C (fcc-γ phase) after hot-rolling and (b) after subsequent cold-rolling (98.5%) at room temperature, the right side of which is the surface of the specimen, and (c) the (100) pole figure of Fig. 2(b).²⁵⁾ The microstructure before cold-rolling has a polygonal grain structure with a grain size of about 390 μm and a random texture. After cold-rolling with a high reduction ratio of 98.5%, this polygonal structure is destroyed and grains elongated along the rolling direction are observed in Fig. 2(b). Note that this specimen has the γ phase even after cold-rolling, meaning that no α’ (bct) martensite was induced by cold-rolling. It was found that the strong {110}<112> brass orientation, which is often observed in fcc metals,^26,27,28,29) and a relatively weaker {110}<001> Goss texture are developed in the 98.5% cold-rolled sheet specimen. Several textures, such as {321}<214> and {211}<011>, were also observed as sub-orientation components. The effect of the reduction ratio on texture formation was reported in our previous paper and it was found that the {110}<112> texture is strongly developed when the reduction ratio is 90% or higher.²⁵⁾

Fig. 2.

IPF maps of the vertical section (a) before and (b) after 98.5% cold-rolling and (c) the (100) pole figure obtained from (b). (Online version in color.)

3.2. Primary Recrystallization

Figure 3 shows the IPF maps in the RD of the fcc and bcc phases of the direct-annealed specimens in the temperature range from 400 to 800°C for 24 hours, and the corresponding (100) pole figures are shown in Fig. 4. Judging from the microstructure in Fig. 3, recrystallization does not occur in the fcc-γ phase annealed at 400°C. The fraction of the fcc-γ phase gradually decreases with increasing temperature up to 700°C. Although it is hardly determined whether recrystallization of the fcc-γ phase has occurred after annealing at 500–700°C due to its low volume fraction, the equiaxial grains are observed in the 800°C specimen, indicating recrystallization. Figure 4(a) shows that the orientation of the fcc-γ phase remained the same as that in the deformation texture (Fig. 2) at all the temperatures.

Fig. 3.

IPF maps in RD of (a) fcc phase and (b) bcc phase after annealing at a temperature range from 400 to 800°C for 24 hours. Volume fraction V_F of the bcc phase is indicated on the top right in (b). (Online version in color.)

Fig. 4.

(100) pole figure of annealed specimens. (a) fcc phase and (b) bcc phase annealed at 400 to 800°C for 24 hours. (Online version in color.)

The bcc phase was also observed at every temperature as shown in Fig. 3, and its volume fraction increases with increasing annealing temperature up to 700°C. However, the texture of the bcc phase quenched from 500°C or 600°C (e.g. {211}<011>) and 700°C or 800°C (e.g. {332}<113> and {100}<011>) are different, as shown in Fig. 4(b).

3.3. Secondary Recrystallization

IPF maps and (100) pole figures of the γ phase directly annealed at temperatures above 1000°C are presented in Fig. 5. A significant change in crystal orientation from the {110}<112> deformation texture was not observed at 1000°C, and all the grains had polygonal shapes, which indicates that (primary) recrystallization has finished. When the sample was heated up to 1100°C (Fig. 5(b)), some γ grains with different texture orientations (e.g. {210}<001> and {211}<011>) have started to abnormally grow. Moreover, the major texture of {210}<001> is developed together with the minor texture of {211}<011> at 1200°C (Fig. 5(c)). Although the information on the fcc-γ phase is shown in Figs. 5(a)–5(c), Fig. 5(d) shows the IPF for both the fcc and bcc phases at 1100°C. It is found that the bcc phase in Fig. 5(d) (and shown as black in Figs. 5(a) and 5(b)) remains up 1100°C. This bcc phase is considered from the phase diagram to be the NiAl-β phase with the B2 structure.^30,31,32) The precipitates are probably related to the same texture between the deformation and recrystallization texture, which will be discussed later. By annealing up to 1200°C, the β phase is perfectly dissolved and the grain size of the γ phase becomes much larger. This result suggests that the SRC at around 1100°C, which is caused by the dissolution of the β precipitates inhibiting the grain growth of the γ phase, is greatly important to develop the {210}<001> texture.

Fig. 5.

IPF maps and the (100) pole figures for fcc phase in cold-rolled specimens after annealing at (a) 1000°C for 3 hours, (b) 1100°C for 3 hours and (c) 1200°C for 3 hours. Black indicates the β precipitates. (d) High magnification images of IPF map in RD and phase map (red: fcc, green: bcc) at 1100°C. (Online version in color.)

The mean grain size of the γ matrix is summarized in Fig. 6(a), and the texture intensities of each texture component against the annealing temperature are shown in Fig. 6(b). The texture intensity of the <001> component remarkably starts to increase at 1120°C and has a noticeable increase at about 1160°C, accompanied with a decrease of intensity in the <112> component and a slight increase in the <110> component. As a consequence, the γ matrix of about 70% is occupied by {210}<001> grains at 1200°C. Thus, the {210}<001> grains are suggested to have the highest growth rate. The change in the major texture component from <112> to <001> in Fig. 6(b) may be related with the drastic grain growth shown in Fig. 6(a).

Fig. 6.

(a) Mean grain size of the γ phase and (b) the texture intensity parallel to RD with tolerance of 10° from the ideal orientation versus annealing temperature. (Online version in color.)

It is concluded that the microstructural evolution by the present thermomechanical treatment is explained as follows; i) development of a deformation texture with strong {110}<112> and weak {110}<001> by cold-rolling after annealing at 1200°C, ii) primary continuous recrystallization and precipitation of the β phase without significant orientation change by heating to about 1000°C, and iii) SRC accompanying dissolution of the β phase and the texture change from {110}<112> to a main {210}<001> and minor {211}<011> by further heating at 1160–1200°C.

4. Discussion

4.1. Primary Recrystallization

Deformation texture changes to a different orientation after recrystallization. For example, a significant orientation change from a {110}<112> brass orientation to a {100}<001> cube orientation occurs in the primary recrystallization stage in Fe–Ni binary alloys,³³⁾ which does not accompany the precipitation of a secondary phase. However, a behavior similar to that in the present alloy has been reported in an Al–Mn–Fe ternary alloy,³⁴⁾ in which the nanoscaled Al₁₂(Mn, Fe) phase precipitates prior to recrystallization. The precipitates disturb the migration of the sub-boundaries and the crystal orientation in recrystallization texture after annealing becomes the same as that in deformation texture, which is called “recrystallization in-situ”. Because recrystallization and precipitation are in diffusion-controlled reactions, both are obviously affected by each other. For the present alloy, the β phase was confirmed by in-situ X-ray diffraction measurement to precipitate prior to the recrystallization of γ matrix, which suggests that the β phase particles may play a role as inhibitors against grain boundary migration. Further investigation is required to clarify the issue in detail.

4.2. Secondary Recrystallization

It was found that the {110}<112> texture changes to the {210}<001> texture through the SRC process in this alloy system. In this section, the microstructural change in this process and the reason for preferential growth of {210}<001> grains than, for example, other {110}<112> grains will be discussed in terms of grain boundary energy, pinning force and grain boundary mobility. The velocity of the grain boundary migration v is generally given by the following form;   

$$v=M\cdot \mathrm{\Delta }{G}_d,$$(1)

where M and ΔG_d are the mobility and driving pressure, respectively. Each factor will be discussed.

4.2.1. Driving Pressure and Pinning Pressure

To investigate the initial stage of the SRC, annealing from 900 to 1110°C with a heating rate of 3°C/min was conducted. An abnormally growing grain with {210}<001> is represented in Fig. 7, where the black particles are the β phase. Precipitate particles, called inhibitors, hinder grain growth and they cause SRC by pinning migration of some grain boundaries, such as in Fe–Si electrical steels³⁵⁾ and also in the present study as discussed above. Here, the driving pressure for SRC and pinning pressure by the β phase are evaluated.

Fig. 7.

Grain boundary characteristics around abnormally grown {210}<001> grains at 1110°C. (Online version in color.)

The pinning pressure by the β phase was estimated by Zener’s equation,³⁶⁾   

$$P_z=\frac{3F_v{\sigma }_h{V}_m}{2R_{\beta }},$$(2)

where F_v and R_β are the volume fraction and the mean grain radius of the β phase, respectively, and σ_h is the grain boundary energy of the γ matrix. The V_m is a molar volume, which is 6.96×10⁻⁶ m³/mol calculated from the lattice parameter a = 0.359 nm determined by the X-ray diffraction technique. Figure 8(a) shows F_v and R_β of the β phase as a function of the annealing temperature of the present alloy measured by the EBSD technique. These experimental values of F_v and R_β and σ_h = 835 mJ/m² from SUS304 stainless steel³⁷⁾ are used for Eq. (2).

Fig. 8.

Diagrams of (a) Volume fraction and mean grain size of the β phase, (b) the estimated pinning pressure by the β phase and driving pressure for abnormal grain growth.

The estimation of the driving pressure for SRC of the γ phase was carried out using the Gibbs-Thomson effect by the following Hillert’s equation³⁸⁾   

$$\mathrm{\Delta }{G}_d=V_m\left(\frac{{\sigma }_l}{{\overline{R}}_l}-\frac{{\sigma }_h}{R_h}\right),$$(3)

where σ_l and σ_h are grain boundary energies of the low-angle boundaries (LABs) and high-angle boundaries (HABs), respectively. R_l is the mean grain radius of the γ matrix and R_h is the grain radius of the abnormal grain. LABs energy σ_l was calculated using the Read-Shockley equation,³⁹⁾ given by,   

$${\sigma }_l={\sigma }_h\frac{{\theta }_{ave}}{{\theta }_h}\left(1-ln\frac{{\theta }_{ave}}{{\theta }_h}\right),$$(4)

The average misorientation angle θ_ave is obtained as 6.3° by analyzing the misorientation of normal {110}<112> grains around the abnormal grains.

The estimated values of the pinning pressure and the driving pressure are plotted against the temperature in Fig. 8(b). Because the pinning pressure of the β phase estimated as 1.74 J/mol at 1000°C is five times larger than the driving pressure for SRC of 0.34 J/mol, grain growth hardly occurs at this temperature. With an increase of temperature, the pinning pressure decreases because of a decrease in the volume fraction of the β phase (and an increase in the mean grain size below 1100°C), and the driving pressure becomes larger than the pinning pressure at temperatures higher than approximately 1096°C. Then, the specific {210}<001> and {211}<011> grains may be able to grow abnormally above this temperature. This estimation is in good agreement with the experimental result of the SRC starting temperature T_s = 1100°C of the <100> grains shown in Fig. 6(b). Once the {210}<001> and {211}<011> grains start to grow, they can grow faster than the γ matrix grains because of their much higher mobility, as will be mentioned later.

To discuss the reason that the {210}<001> becomes the main orientation after secondary recrystallization, the first factor we need to consider is the grain boundary energy (Eq. (3)). However, the difference in the grain boundary energy seems to be low because the misorientation of the {211}<011> to the {110}<112> grains is close to that of the {210}<001> grains (about 5°). The interfacial energy between the γ abnormal grain and the β precipitates, which yields the pinning pressure in Eq. (2), is the second factor to be considered here. The pining pressure is larger for an abnormal grain with higher γ/β interfacial energy than for that with a lower interfacial energy. If the {211}<011> grains have a higher interfacial energy than the {210}<001> grains (ex. {211}<011> and {210}<001> have lower and higher coherency to the β precipitate, respectively), the pinning pressure becomes larger in the {211}<011> than in the {210}<001> grains. However, the texture development of the β phase is not strong and a specific orientation relationship between the γ phase with the {211}<011> and {210}<001> textures and the β phase was not confirmed in this study. Another factor affecting the pinning pressure is the particle size of the β precipitates dispersed on the γ grain boundaries (Eq. (2)). The atomic diffusivity in the grain boundaries around the {210}<001> may be higher than that around the {211}<011> because of the higher mobility, as will be discussed later. Therefore, the β precipitates may grow faster with the Ostwald ripening on the grain boundaries surrounding the {210}<001> abnormal grains, which reduces the pinning pressure. This inference is only one possibility for explaining the preferential growth against the {211}<011> and also the γ matrix grains with a {110}<112> texture, and no direct evidence has been obtained yet.

4.2.2. Mobility of the Grain Boundary

As shown in Fig. 7, the abnormally growing grain is unexceptionally surrounded by HABs (>15°) and the other fine grains that mostly have LABs with a small fraction of HABs and low percentage of CSLBs, consisting of Σ=3 of 1%–2%, 7 of 5%–7%, 13b of 3%–6% and others less than 1%. In the SRC process of fcc alloys, various orientation relationships between the grain in matrix and abnormally growing grains have been reported by many researchers, such as the 30°–50° <111> rotation,⁴⁰⁾ the 22° and 38° <111> rotation,⁴¹⁾ the 19° <100> rotation⁴¹⁾ and the 30° <100> rotation.^42,43) Furthermore, those grain boundaries indicate a higher mobility than others. The rotation angles of the {210}<001> and {211}<011> abnormal grains around <111> to the {110}<112> matrix, which were observed in the present work, are near 40° and 30°, respectively. Furthermore, both the boundaries are in the 30°–50° <111> rotation relation and are considered to have higher motilities than other conventional boundaries.⁴⁴⁾ Moreover, the average misorientation angle among the {110}<112> matrix grains is within about 6.3° and the grain boundary mobility of LABs is expected to be low. Hence, the {210}<001> and {211}<011> grains that have a large and specific misorientation angle with the {110}<112> matrix can grow faster than other grains during the SRC process.

On the difference in mobility between the {110}<112> and the {210}<001> grains, Huang et al. reported the effect of rotation angle in the <111> axis for grain boundary mobility in Al alloys and indicated that the growth rate of grains with a misorientation of 40° is about four times faster than that of grains with a misorientation of 30°.⁴⁵⁾ If this tendency also meets the present case in the Fe alloy with the same fcc structure, the {210}<001> grains with a 40° rotation about the <111> axis should have a much higher mobility than {211}<011> grains with a 30° rotation. This may be the prevailing reason for the {210}<001> texture being the main component in the SRC.

5. Conclusions

In the present study, recrystallization behavior was investigated using 98.5% cold-rolled an Fe-30Ni-15Co-10Al-2.5Ti-0.05B alloy and the following results were obtained:

(1) The microstructure before cold-rolling is oriented randomly with a grain size of 390 μm. After 98.5% cold-rolling, a deformation texture, mainly of {110}<112> and {110}<001>, is developed in the elongated grain microstructure.

(2) A significant orientation change of the γ matrix from the {110}<112> texture is not observed during the primary recrystallization process, i.e. continuous recrystallization occurs. The invariance of crystal orientation is probably related to the presence of the β phase, which may interrupt the migration of the grain boundaries.

(3) At temperatures higher than 1100°C (in the secondary recrystallization process), the pinning pressure from the β precipitates becomes smaller than the driving pressure for the grain growth of the {210}<001> and {211}<011> γ grains. Abnormal grain growth of these grains occurs because of the higher mobility of these grains than that of the textured γ matrix.

(4) The main texture of the γ phase after secondary recrystallization is {210}<001>. The reason for the preferential growth of {210}<001> grains over the {211}<011> grains is thought to be that the {210}<001> grains with misorientation by 40° about <111> axis with respect to the {110}<112> matrix have higher mobility. The relatively lower pinning pressure from the β precipitates might be another possible reason.

Acknowledgements

This work was supported by JSPS KAKENHI Grant Number 15H05766. The authors would like to thank Dr. K. Okuda, Prof. T. Furuhara and Prof. H. Kokawa for valuable discussions.

References

• 1)   U. F.  Kocks,  C. N.  Tomé and  H.-R.  Wenk: Texture and Anisotropy: Preferred Orientations in Polycrystals and Their Effect on Materials Properties, Cambridge University Press, Cambridge, UK, (2000), 282.
• 2)   H.  Sehitoglu,  X. Y.  Zhang,  T.  Kotil,  D.  Canadinc,  Y.  Chumlyakov and  H. J.  Maier: Metall. Mater. Trans. A, 33 (2002), 3661.
• 3)   H.  Horikawa,  S.  Ichinose,  K.  Morii,  S.  Miyazaki and  K.  Otsuka: Metall. Trans. A, 19 (1988), 915.
• 4)   Y. I.  Chumlyakov,  N. S.  Surikova and  A. D.  Korotaev: Fiz. Met. Metalloved., 82 (1996), 148 (in Russian).
• 5)   H.  Inoue,  N.  Miwa and  N.  Inakazu: Acta Mater., 44 (1996), 4825.
• 6)   Y. C.  Shu and  K.  Bhattacharya: Acta Mater., 46 (1998), 5457.
• 7)   K.  Gall,  T. J.  Lim,  D. L.  McDowell,  H.  Sehitoglu and  Y. I.  Chumlyakov: Int. J. Plast., 16 (2000), 1189.
• 8)   Y.  Sutou,  T.  Omori,  R.  Kainuma,  N.  Ono and  K.  Ishida: Metall. Mater. Trans. A, 33 (2002), 2817.
• 9)   H. Y.  Kim,  T.  Sasaki,  K.  Okutsu,  J. I.  Kim,  T.  Inamura,  H.  Hosoda and  S.  Miyazaki: Acta Mater., 54 (2006), 423.
• 10)   R.  Chulist,  M.  Czerny,  A.  Panigrahi,  M.  Zehetbauer,  N.  Schell and  W.  Skrotzki: IOP Conf. Ser. Mater. Sci. Eng., 375 (2018), 012006.
• 11)   C.  Zhang,  C.  Zhu,  S.  Shin and  K.  Vecchio: J. Alloy. Compd., 768 (2018), 277.
• 12)   Y.  Tanaka,  Y.  Himuro,  R.  Kainuma,  Y.  Sutou,  T.  Omori and  K.  Ishida: Science, 327 (2010), 1488.
• 13)   J.  Ma,  B.  Kockar,  A.  Evirgen,  I.  Karaman,  Z. P.  Luo and  Y. I.  Chumlyakov: Acta Mater., 60 (2012), 2186.
• 14)   P.  Krooß,  C.  Somsen,  T.  Niendorf,  M.  Schaper,  I.  Karaman,  Y.  Chumlyakov,  G.  Eggeler and  H. J.  Maier: Acta Mater., 79 (2014), 126.
• 15)   T.  Omori,  S.  Abe,  Y.  Tanaka,  D. Y.  Lee,  K.  Ishida and  R.  Kainuma: Scr. Mater., 69 (2013), 812.
• 16)   H. E.  Karaca,  A. S.  Turabi,  Y. I.  Chumlyakov,  I.  Kireeva,  H.  Tobe and  B.  Basaran: Scr. Mater., 120 (2016), 54.
• 17)   Y. I.  Chumlyakov,  I. V.  Kireeva,  V. V.  Poklonov,  Z. V.  Pobedennaya and  I.  Karaman: Tech. Phys. Lett., 40 (2014), 747.
• 18)   D.  Lee,  T.  Omori and  R.  Kainuma: J. Alloy. Compd., 617 (2014), 120.
• 19)   T.  Omori,  K.  Ando,  M.  Okano,  X.  Xu,  Y.  Tanaka,  I.  Ohnuma,  R.  Kainuma and  K.  Ishida: Science, 333 (2011), 68.
• 20)   T.  Omori and  R.  Kainuma: Shap. Mem. Superelast., 3 (2017), 322.
• 21)   J.  Ma,  B. C.  Hornbuckle,  I.  Karaman,  G. B.  Thompson,  Z. P.  Luo and  Y. I.  Chumlyakov: Acta Mater., 61 (2013), 3445.
• 22)   Y.  Tanaka,  R.  Kainuma,  T.  Omori and  K.  Ishida: Mater. Today Proc., 2 (2015), S485.
• 23)   E. W.  Ross and  C. T.  Sims: Superalloys II, eds. by C. T. Sims, N. S. Stoloff and W. C. Hagel, Wiley-Interscience Publication, New York, (1987), 97.
• 24)   H.  Fu,  W.  Li,  S.  Song,  Y.  Jiang and  J.  Xie: J. Alloy. Compd., 684 (2016), 556.
• 25)   D.  Lee,  T.  Omori,  K.  Han,  Y.  Hayakawa and  R.  Kainuma: Shap. Mem. Superelast., 4 (2018), 102.
• 26)   J.  Hirsch and  K.  Lücke: Acta Metall., 36 (1988), 2863.
• 27)   J.  Hirsch,  K.  Lücke and  M.  Hatherly: Acta Metall., 36 (1988), 2905.
• 28)   T.  Leffers and  R. K.  Ray: Prog. Mater. Sci., 54 (2009), 351.
• 29)   S. R.  Kalidindi: Int. J. Plast., 17 (2001), 837.
• 30)   Y.  Himuro,  Y.  Tanaka,  I.  Ohnuma,  R.  Kainuma and  K.  Ishida: Intermetallics, 13 (2005), 620.
• 31)   L.  Zhang and  Y.  Du: Calphad, 31 (2007), 529.
• 32)   V.  Raghavan: J. Phase Equilib. Diffus., 31 (2010), 455.
• 33)   S.  Zaefferer,  T.  Baudin and  R.  Penelle: Acta Mater., 49 (2001), 1105.
• 34)   N.  Inakazu and  K.  Lücke: J. Jpn. Inst. Met., 49 (1985), 921 (in Japanese).
• 35)   S.  Taguchi and  A.  Sakakura: Acta Metall., 14 (1966), 405.
• 36)   C. S.  Smith: Trans. Am. Inst. Min. Metall. Eng., 175 (1948), 15.
• 37)   L. E.  Murr: Interfacial Phenomena in Metals and Alloys, Addison-Wesley, Reading, (1975), 87.
• 38)   M.  Hillert: Acta Metall., 13 (1965), 227.
• 39)   W. T.  Read and  W.  Shockley: Phys. Rev., 78 (1950), 275.
• 40)   P. A.  Beck,  P. R.  Sperry and  H.  Hu: J. Appl. Phys., 21 (1950), 420.
• 41)   M. L.  Kronberg and  F. H.  Wilson: Trans. Am. Inst. Min. Metall. Eng., 185 (1949), 501.
• 42)   O.  Dahl and  F.  Pawlek: Z. Metallkd., 28 (1936), 266.
• 43)   F. H.  Wilson and  R. M.  Brick: Trans. Am. Inst. Min. Metall. Eng., 161 (1945), 173.
• 44)   F. J.  Humphreys and  M.  Hatherly: Recrystallization and Related Annealing Phenomena, 2nd ed., Elsevier, Oxford, UK, (2004), 121.
• 45)   Y.  Huang and  F. J.  Humphreys: Acta Mater., 47 (1999), 2259.