Inhibition of Electropulsing Nanocrystallization in Amorphous ZrCu under Helium Atmosphere

Abstract

Nanocrystallization of amorphous ZrCu was induced by applying a pulse current (electropulsing), which exponentially decayed from a certain initial current density (i_d0) with the time constant of 3 ms. For electropulsing, the ribbon specimen was sandwiched between high-thermal-conductive AlN (thermal conductivity κ = 170 W/m/K at 273 K) plates to remove the Joule heat. Nanocrystallization was found when electropulsing was conducted in a vacuum below 10⁻² Pa at the pulse current of i_d0 ≥ 0.6 GA/m². In contrast, electropulsing conducted under 10⁵ Pa He (κ = 0.144 W/m/K at 273 K) required i_d0 above 1.1 GA/m². Thermal conductivity and heat capacity of 10⁵ Pa He were negligibly smaller than those of the AlN plates. The increase in i_d0 for nanocrystallization with increasing He pressure indicates that He penetrated the amorphous structure to inhibit the activation of cooperative motions of many atoms and larger i_d0 is required for nanocrystallization at higher He pressure.

Fig. 5 Resistivity changes of a-Zr₅₀Cu₅₀ by electropulsing at various He pressures. The resistivity normalized by the value in as-spun state (ρ₀) is shown.

1. Introduction

Some of the metal oxides, semiconductors, and organic materials are easily consolidated in the amorphous state by cooling them from the melt. On the other hand, a rapid cooling is often required to prepare amorphous alloys to prevent crystallization. Rapid cooling shapes the amorphous alloys into thin tapes or tiny pieces. Most of the amorphous alloys are thermally unstable and their crystallization is induced by thermal treatments below the glass transition temperature. Others, however, exhibit glass transition state when warming-up and such thermally stable amorphous alloys are classified as metallic glasses. A recent development in technology enables the preparation of amorphous alloys at the cooling rate of below 0.1 K/s. Bulk-sized metallic glasses can therefore be formed.¹⁾

Crystallization of amorphous alloys is also induced by heavy plastic deformations through ball-milling and nanoindentation.^2–5) However, under ultrasonic irradiation, crystallization is enhanced by the elastic deformation, not heavy plastic deformations.^6,7) Also, electromagnetic treatments induce crystallization of amorphous alloys. By applying considerable high currents, Joule heating also induced crystallization after heating the specimens to the glass transition temperatures.^8,9) High current treatment was applied for a shaping or sintering process of amorphous alloys and hard materials.^10–12)

Crystallization of amorphous alloys, like the intermetallic compound formation, precipitation, recrystallization, and grain growth in metals, occurs by passing the current in mild conditions.¹³⁾ Mizubayashi et al. observed crystallization of amorphous (a-) Ti₅₀Cu₅₀, Pd₈₀Si₂₀,¹⁴⁾ Zr₆₀Cu₃₀Al₁₀,¹⁵⁾ and Zr₅₀Cu₅₀¹⁶⁾ alloys by also passing a pulse current, which exponentially decayed with the time constant of τ = 1–10 ms from the initial current density of i_d0 > ∼1.4 GA/m². In their experiments, a ribbon-shaped specimen was sandwiched between two plates of thermal-conductive ceramics to remove Joule heat, and electropulsing was conducted in air. When crystallization was induced by electropulsing, the maximum temperature rise of about of 200 K was observed at around 10 ms.^14–16) Mizubayashi et al. also performed electropulsing experiments in liquid nitrogen and observed nanocrystallization at a similar initial current density.¹⁷⁾ In a-Zr₅₀Cu₅₀, high-temperature phases of cubic (c-) ZrCu or monoclinic (m-) ZrCu intermetallic compounds, which are thermally equilibrium above 984 K,¹⁸⁾ were formed by electropulsing. Furthermore, the dynamic Young’s modulus of amorphous alloys showed a local minimum at the measuring frequency of f ∼ 100 Hz¹⁹⁾ and the corresponding relaxation time of 1/2πf ∼ 1.7 ms was comparable to the time constant of the electropulsing experiments. These observations suggested that the cooperative local motions of atoms were resonantly activated in the amorphous alloys by electropulsing; thus, nanocrystallization without diffusional atom motions was effectively induced like the martensitic transformation.^14,17) However, the detailed process of nanocrystallization remains sketchy.

Recently, we found that nanocrystallization was induced at the initial current density of 0.6 GA/m² in a vacuum. This initial current density was extremely smaller than >∼1.4 GA/m² in previous electropulsing experiments conducted in air. To account for the nanocrystallization process by electropulsing and the effect of the gas atmosphere, we investigated the nanocrystallization behavior of a-Zr₅₀Cu₅₀ in various He gas pressures.

2. Experimental Procedures

Thin tapes of a-Zr₅₀Cu₅₀ alloys were prepared by melt spinning in a high-purity Ar atmosphere. An ingot of Zr₅₀Cu₅₀ was prepared by arc melting and was used as the base material of amorphous alloys. The spun thin tape was cut into ribbons of about 80 mm long each. To make the cross-section uniform, the surface parallel to the Cu roll and both sides of the ribbon were polished mechanically with the lapping film up to #15000. The mechanical polishing was conducted in water flow to cool it. The thickness and width of specimens after polishing were about 30 µm and 1.7 mm, respectively. The ribbons, whose uniformity of the cross-section was within 10%, were used for the electropulsing experiment. Finally, the polished ribbon was cut into 60 mm in length.

A discharge current from a capacitor bank was applied to the ribbon specimen in electropulsing.¹⁴⁾ When a capacitor (capacitance C) charged at voltage V is discharged by connecting a resistor (resistance R, the combined resistance of the ribbon amorphous specimen and the electric circuit in the present case) in series, the discharge current decreases as,   

\begin{equation*} I(t) = I_{0} \exp (-t/\tau)\ \text{or}\ i_{\text{d}}(t) = i_{\text{d0}} \exp(-t/\tau) \end{equation*}

where t is the time after the discharge, τ = RC the time constant and I₀ = V/R the initial current. Further, i_d is the current density in the specimen, that is I(t) = Ai_d(t) and I₀ = Ai_d0, where A is the cross-section of the ribbon specimen. In this experiment, the time constant was fixed at τ = 3 ms because nanocrystallization of a-Zr₅₀Cu₅₀ by electropulsing was most effectively induced at τ = 3 ms in our previous experiments.¹⁶⁾ The combined resistance including the circuit resistance is ∼1.8 Ω. The capacitor of C = ∼1700 µF was used. Nanocrystallization of a-Zr₅₀Cu₅₀ was observed by electropulsing at i_d0 ≥ 0.6 GA/m² in a vacuum. For i_d0 = 0.6 GA/m², the capacitor was charged to the voltage V = ∼55 V or the electrostatic energy of ∼2.6 J. Electropulsing was repeated by increasing i_d0, which actually increased V. When the resistance of the specimen was decreased by electropulsing, the capacitance changed, time constant being 3 ms. The AlN-plate-sandwiched specimen (Fig. 1) was set into a vacuum chamber and electropulsing was carried out under the He pressure between <10⁻² and 10⁵ Pa.

Fig. 1

Schematic drawing for electropulsing experiment setup. The a-Zr₅₀Cu₅₀ ribbon specimen is sandwiched between high-thermal-conductive AlN plates to remove Joule heat. Cu wires shaped into a semicircular cross section are embedded in one of the AlN plates for electrical conductivity.

In this study, high-thermal-conductive AlN plates (thermal conductivity κ = 170 W/K/m and 100 × 10 × 5 mm³ in size) were used to remove Joule heat (Fig. 1). The heat capacities of the AlN plate and a-Zr₅₀Cu₅₀ ribbon specimen are 3.5 J/K and 0.038 J/K, respectively. In our previous electropulsing experiments, Joule heat was removed by sandwiching the specimen between two aluminum nitride (AlN)/boron nitride (BN)-composite ceramic plates (κ = 90 W/K/m and 100 × 10 × 5 mm³). Using a thin thermocouple, the AlN/BN-sandwiched specimen temperature measurement during electropulsing indicated that the maximum temperature rise of about 200 K was attained at t = ∼8 ms where i_d0 = 1.8 GA/m² and τ = 4.3 ms.¹⁶⁾ The resistivity of the specimen after electropulsing was measured by a four-probe method without removing the ribbon specimen from the AlN plates. The X-ray diffraction (XRD) analyses were carried out in the θ-2θ scan mode with Cu-Kα radiation (Panalytical X’pert).

3. Results and Discussion

3.1 Resistivity change by electropulsing in a vacuum and under 10⁵ Pa He

Figure 2 shows the resistivity change of a-Zr₅₀Cu₅₀ by electropulsing in a vacuum (<10⁻² Pa) and under 10⁵ Pa He. In the experiment, i_d0 was increased by 0.01 or 0.1 GA/m² thus keeping τ = 3 ms, and the resistivity was measured after electropulsing. In a vacuum, the resistivity of a-Zr₅₀Cu₅₀ showed a step-like decrease after electropulsing at i_d0 = 0.6–0.7 GA/m². When electropulsing was performed under 10⁵ He Pa, the resistivity showed a gradual decrease with increasing i_d0 above 1.0 GA/m². Figure 3 shows the X-ray diffraction patterns of the specimen where the resistivity was reduced by electropulsing. Both in a vacuum and under 10⁵ Pa He, the formation of c-ZrCu and m-ZrCu intermetallic compounds (equilibrium phases above 984 K) were found. But, no Zr₂Cu and Cu₁₀Zr₇ intermetallic compounds (equilibrium phases below 971 K) were found.¹⁸⁾ In our previous investigations, nanocrystallites of ∼10 nm in size were observed in the electron microscopy of the specimens of which resistivity was decreased by electropulsing.¹⁷⁾ It was reported that the icosahedral local order in the liquid state prevented the formation of the B2 structure in the ZrCu system and it was attributed to the high glass-forming ability of the ZrCu system.²⁴⁾ Liquid alloy near the temperature of the liquidus line should, as expected, contain locally ordered units of crystals formed below the liquidus line.^16,25,26) These locally ordered units play an important role in the cooling formation of the amorphous state. These results indicate that nanocrystallization by electropulsing is initiated from the locally ordered units in as-spun a-Zr₅₀Cu₅₀ while the nanocrystallites of high-temperature intermetallic phases are formed by electropulsing. In other words, the local structures in the amorphous state can be revealed by electropulsing. Furthermore, the progress or initiation of nanocrystallization appears to be inhibited under 10⁵ Pa He because the decrease in the resistivity and fraction of the crystallized amount in the XRD pattern were smaller than in the vacuum.

Fig. 2

Resistivity changes of a-Zr₅₀Cu₅₀ by electropulsing in vacuum (<10⁻² Pa, blue symbols) and in 10⁵ Pa He (orange symbols), where electropulsing was repeated with increasing the initial current density (i_d0). Both in a vacuum and under 10⁵ Pa He, specimens with the initial resistivity (as-spun state) below 2.2 µΩm (filled symbols) showed the decrease in resistivity at i_d0 larger by 0.1–0.2 GA/m² than those above 2.2 µΩm (open symbols).

Fig. 3

XRD patterns for a-Zr₅₀Cu₅₀ of which resistivity was decreased by electropulsing in vacuum (<10⁻² Pa) and under 10⁵ Pa He. The pattern for as-spun specimen and peak positions reported for cubic c-ZrCu,²⁰⁾ monoclinic m-ZrCu,²¹⁾ Cu₁₀Zr₇²²⁾ and Zr₂Cu²³⁾ intermetallic compounds are also shown.

In a vacuum, as shown in Fig. 2, the specimens with the initial resistivity below 2.2 µΩm (filled blue symbols) showed a decrease in resistivity after electropulsing at i_d0 > 0.7 GA/m² whereas those above 2.2 µΩm (open blue symbols) showed a decrease at 0.6 < i_d0 < 0.7 GA/m². Under 10⁵ Pa He, the specimens with the initial resistivity below 2.2 µΩm (filled orange symbols) showed the decrease in resistivity at i_d0 larger by 0.1–0.2 GA/m² than those above 2.2 µΩm (open orange symbols). The lower resistivity might reflect the local ordering or structural relaxation; the larger driving force is required to initiate nanocrystallization in the specimens with the lower resistivity.²⁷⁾ The experimental results of a-Zr₅₀Cu₅₀ with the resistivity of 1.9–2.2 µΩm are discussed below to highlight the effect of He on nanocrystallization by electropulsing.

3.2 Electropulsing by changing the atmosphere from vacuum to 10⁵ Pa He and vice versa

As already mentioned in Fig. 2 and 3, the formation of non-equilibrium crystalline phases was common for electropulsing in a vacuum and under 10⁵ Pa He. However, nanocrystallization in 10⁵ Pa He required much larger i_d0 than in the vacuum. Under the electropulsing process, the pulse current is expected to act as a driving force for the crystallization. The larger i_d0 for nanocrystallization in 10⁵ Pa He reflects the inhibition of nanocrystallization by He penetrated the specimen. When He is chemically trapped at certain sites in the specimen, the inhibition of nanocrystallization is expected even after the atmosphere is changed from 10⁵ Pa He to vacuum. The change in the resistivity by electropulsing was investigated in vacuum subsequently after nanocrystallization by electropulsing in 10⁵ Pa He. The result is shown in Fig. 4(a). In 10⁵ Pa He, the resistivity decreased by about 17% after electropulsing at i_d0 = 1.0 GA/m². Electropulsing was therefore conducted in vacuum (<10⁻² Pa). The step-like decrease in the resistivity in vacuum was observed at 0.7 GA/m², same as i_d0,c in vacuum, as shown in Fig. 2. The result for the atmosphere change from vacuum to 10⁵ Pa He is shown in Fig. 4(b). The resistivity had showed the step-like decrease by electropulsing with i_d0 = 0.6 GA/m² in a vacuum. The atmosphere was then changed to 10⁵ Pa He and the electropulsing was conducted with increasing i_d0 from 0.5 GA/m². In 10⁵ Pa He, a further decrease in the resistivity was observed by electropulsing above i_d0 = 0.9 GA/m². These observations suggest that nanocrystallization by electropulsing is inhibited by He and the effect of He on the nanocrystallization by electropulsing is reversible for the He pressure, for instance, He is not chemically trapped in or on the specimen.

Fig. 4

(a) Resistivity change of a-Zr₅₀Cu₅₀ by electropulsing. The resistivity normalized by the value in as-spun state (ρ₀) is shown. The atmosphere was changed from 10⁵ Pa He to vacuum (<10⁻² Pa) and electropulsing was repeated subsequently after the resistivity decrease by electropulsing in 10⁵ Pa He with i_d0 = 1.0 GA/m². (b) Similar to (a), but the atmosphere was changed from vacuum (<10⁻² Pa) to 10⁵ Pa He. Electropulsing was repeated in 10⁵ Pa He subsequently after the resistivity decrease by electropulsing in vacuum with i_d0 = 0.6 GA/m².

3.3 Helium pressure dependence on critical current density for electropulsing nanocrystallization

Helium pressure dependence on the change in the resistivity by electropulsing was investigated to examine its role on crystallization by electropulsing. Figure 5 depicts the change in the resistivity after electropulsing with increasing i_d0 under various He pressures. Compared with the result in a vacuum (<10⁻² Pa), the current density required for nanocrystallization by electropulsing showed a gradual increase with increasing He pressure from 100 to 10⁵ Pa. Figure 6 shows the He pressure (P_He) dependence on the critical current density (i_d0,c, the decrease in the resistivity beyond 1% was observed after electropulsing). Notably, the concentration of hydrogen in metal is proportional to the square root of the hydrogen pressure because the thermal equilibrium is achieved between the hydrogen gas and hydrogen atom in the metals (Sievelt’s low). Helium is a monoatomic molecule and the thermal equilibrium concentration is expected to be in proportional to P_He; however, the He pressure dependence of i_d0,c in Fig. 6 seems to be parallel to (P_He)^0.5 rather than P_He.

Fig. 5

Resistivity changes of a-Zr₅₀Cu₅₀ by electropulsing at various He pressures. The resistivity normalized by the value in as-spun state (ρ₀) is shown.

Fig. 6

He pressure (P_He) dependence of the critical initial current density (i_d0,c) for nanocrystallization by electropulsing.

Helium is the lightest noble gas and its thermal conductivity (κ) is the highest (0.144 W/K/m at 10⁵ Pa and 273 K). Electropulsing in He atmosphere removes Joule heat more than electropulsing in a vacuum. In this study, however, the specimen was sandwiched between high-thermal-conductive AlN plates. The thermal conductivity of AlN is 170 W/K/m, which is about one thousand times higher than that of 10⁵ Pa He gas. Further, the thermal conductivity of gases is almost constant at the pressure range where the mean free path is smaller than the characteristic length such as the distance between the heat and cold sources or the dimensions of the container.²⁸⁾ The AlN plates are considered as the cold source, and the thickness of the specimen (about 30 µm) corresponds to the characteristic length of He. The mean free path of He at 10⁵ Pa and 300 K is about 200 nm. The thermal conductivity of He in the present case between 670 and 10⁵ Pa is independent. Helium pressure dependence in Fig. 5 or 6 indicates that i_d0,c is a monotonic increasing function of He pressure for the range of 10²–10⁵ Pa. The temperature increase estimated in our previous experiments remained below 200 K where AlN/BN (κ = 90 W/K/m) plates were used under electropulsing with i_d0 = 1.8 GA/m² and τ = 4.3 ms.¹⁶⁾ These observations suggest that nanocrystallization by electropulsing is independent of the thermal effect induced by Joule heat.

Different from crystallization by thermal annealing, the non-equilibrium high-temperature intermetallic phases were formed in a-Zr₅₀Cu₅₀ by electropulsing as mentioned above. It was pointed out that the cooperative motions of many atoms became active by electropulsing.^14,17) Furthermore, the high-temperature crystalline phases in non-equilibrium were formed by electropulsing through the cooperative motions of localized crystalline-like structures embedded in the amorphous state. From these facts, we assume that He is penetrated in the a-Zr₅₀Cu₅₀ specimen to suppress the cooperative motions of many atoms hence electropulsing crystallization is inhibited under the He atmosphere. An inhomogeneous microstructure, in which strongly bonded regions (SBRs) are connected with weakly bonded regions (WBRs), was reported to explain the acceleration of crystallization in amorphous alloys under ultrasonic perturbation around the glass transition temperature.^6,7) In the microstructure, the atomic mobility in WBRs is much higher than that in SBRs. Another type of microstructure, a “core–shell” model which consists of a free-volume zone and its surrounding elastic envelope, was proposed to account for the inelastic deformation of a Zr-based bulk metallic glass.²⁹⁾ The atoms in free-volume zones undergo the local structural rearrangement and the Newtonian-flow behavior of the free-volume zones is attained. Both microstructural models suggest that atomic motions become more active in loosely-packed soft regions than in densely-packed hard regions under external fields. On the other hand, it was reported that inert-gas atoms are absorbed at the low density regions in the amorphous alloys. The high-angle annular dark-field scanning transmission electron microscope observation indicated that small particles of solid argon were formed in the free-volume zones of a Pd₄₀Ni₂₀P₂₀ metallic glass after argon ion-milling process.³⁰⁾ These reports suggest that He atoms penetrate into the low density regions in a-Zr₅₀Cu₅₀ and suppress the excitation of collective motions around the low density regions, hence the nanocrystallization by electropulsing is inhibited under He atmosphere.

The addition of hydrogen enhances the strength of amorphous alloys. The hydrogens are trapped at the free volumes and they impede the development of shear bands at the plastic deformation.³¹⁾ On the other hand, it was reported that hydrogens in the amorphous alloys induced a large internal friction peak below room temperature.³²⁾ The effect of He on nanocrystallization is reversible under the He pressure as shown in Fig. 4. The behavior of He atoms in a-Zr₅₀Cu₅₀ might be different from that of hydrogen trapped in the free volumes. It was pointed out that the local response at the atomic level under the external stress field is very heterogeneous in the amorphous alloys.³³⁾ The evaluation of He concentration and the state in the amorphous structure should be required to account for the detailed role of He and the mechanism of the electropulsing nanocrystallization at the atomic level.

The enhancement in plastic deformation by applying high power current (electromagnetic pulsing deformation) is attracting more attention as a powerful technique for hard and brittle metallic materials.^34–37) The glass transition-like behavior of the grain boundaries was suggested from the experimental³⁸⁾ and simulation^39,40) studies on nanocrystalline metals. The results of this study also contribute to further enhancement of these electromagnetic pulsing deformation techniques for amorphous alloys and provide a deep insight into the characteristics of the grain boundaries in nanocrystalline metals.

4. Conclusion

Through electropulsing, nanocrystallization of a-Zr₅₀Cu₅₀ was induced by applying a pulse current, which was exponentially decayed with the time constant of 3 ms. The specimen was sandwiched between two high-thermal-conductive AlN plates (thermal conductivity κ = 170 W/m/K) to remove Joule heat by the electropulsing. When electropulsing was conducted in a vacuum below 10⁻² Pa, nanocrystallization was observed with the initial current density above 0.6 GA/m². The initial current density to induce nanocrystallization was gradually increased to ≥1.0 GA/m² with increasing He pressure to 10⁵ Pa. Thermal conductivity of He gas was almost constant (κ = 0.144 W/m/K) at room temperature in the pressure range between 670 and 10⁵ Pa in this study. Furthermore, the specimen was sandwiched between the high-thermal-conductive AlN plates, of which the thermal conductivity is about one thousand times larger than that of 10⁵ Pa He at room temperature. Our previous electropulsing experiments showed that the maximum temperature rise during electropulsing was about 150 K when the current density of 1.5 GA/m² was applied to the specimen sandwiched between two AlN/BN plates (κ = 90 W/m/K) in air. In addition, an excitation of cooperative motions of many atoms was suggested from the dynamic elasticity measurement by applying the alternating external force or the electropulsing with various time constants. These features indicate that electropulsing excites the cooperative motions of many atoms in a-Zr₅₀Cu₅₀ thus leading to nanocrystallization. Helium suppresses the cooperative motions by penetrating into the low density regions of the amorphous structure and as a result inhibits nanocrystallization by electropulsing.

Acknowledgements

The authors thank Prof. H. Mizubayashi for valuable discussions. The study was partially supported by the KAKENHI Grant 19H05166 from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

REFERENCES

• 1)   A.  Inoue and  A.  Takeuchi: Acta Mater. 59 (2011) 2243–2267.
• 2)   J.J.  Kim,  Y.  Choi,  S.  Suresh and  A.S.  Argon: Science 295 (2002) 654–657.
• 3)   W.H.  Jiang,  F.E.  Pinkerton and  N.  Atzmon: J. Appl. Phys. 93 (2003) 9287–9290.
• 4)   A.L.  Greer,  Y.Q.  Cheng and  E.  Ma: Mater. Sci. Eng. Rep. 74 (2013) 71–132.
• 5)   Y.  Mao,  J.  Li,  Y.-C.  Lo,  X.  Qian and  E.  Ma: Phys. Rev. B 91 (2015) 214103.
• 6)   T.  Ichitsubo,  E.  Matsubara,  H.S.  Chen,  J.  Saida and  T.  Yamamoto: J. Chem. Phys. 125 (2006) 154502.
• 7)   T.  Ichitsubo and  E.  Matsubara: Mater. Sci. Eng. A 521–522 (2009) 236–242.
• 8)   S.-B.  Qiu and  K.-F.  Yao: J. Non-Cryst. Solids 354 (2008) 3520–3524.
• 9)   Y.-j.  Huang,  X.  Cheng,  H.-b.  Fan,  S.-s.  Guan,  Z.-l.  Ning and  J.-f.  Sun: J. Iron Steel Res. Int. 23 (2016) 69–73.
• 10)   O.A.  Graeve,  R.  Kanakala,  L.  Kaufman,  K.  Sinha,  E.  Wang,  B.  Pearson,  G.  Rojas-George and  J.C.  Farmer: Mater. Lett. 62 (2008) 2988–2991.
• 11)   Y.  Zhang,  P.  Sharma and  A.  Makino: Mater. Trans. 52 (2011) 2254–2257.
• 12)   O.A.  Graeve,  M.S.  Saterlie,  R.  Kanakala,  S.D.  De La Torre and  J.C.  Farmer: Scr. Mater. 69 (2013) 143–148.
• 13)   H.  Conrad: Mater. Sci. Eng. A 287 (2000) 227–237.
• 14)   H.  Mizubayashi,  N.  Kameyama,  T.  Hao and  H.  Tanimoto: Phys. Rev. 64 (2001) 054201.
• 15)   H.  Mizubayashi,  T.  Hao and  H.  Tanimoto: J. Non-Cryst. Solids 312–314 (2002) 581–584.
• 16)   H.  Mizubayashi,  T.  Takahashi,  K.  Nakamoto and  H.  Tanimoto: Mater. Trans. 48 (2007) 1665–1670.
• 17)   T.  Hao,  H.  Tanimoto and  H.  Mizubayashi: Mater. Trans. 46 (2005) 2898–2907.
• 18)   E.F.  Kneller,  Y.  Khan and  U.  Gorres: Z. Metallk. 77 (1986) 43–48.
• 19)   H.  Mizubayashi,  T.  Usui and  H.  Tanimoto: Mater. Sci. Eng. A 370 (2004) 260–263.
• 20)   E.M.  Carvalho and  I.R.  Harris: J. Mater. Sci. 15 (1980) 1224–1230.
• 21)   A.  Zhalko-Titarenko,  M.  Yevlashina,  V.  Antonov,  B.  Yavorskii,  Yu.  Koval and  G.  Firstov: Phys. Status Solidi B 184 (1994) 121–127.
• 22)   J.  Glimois,  P.  Forey,  J.  Feron and  C.  Becle: J. Less-Common Met. 78 (1981) 45–50.
• 23)   M.  Nevitt and  J.  Downey: Trans. Metall. Soc. AIME 224 (1962) 195–196.
• 24)   V.  Wessels,  A.K.  Gangopadhyay,  K.K.  Sahu,  R.W.  Hyers,  S.M.  Canepari,  J.R.  Rogers,  M.J.  Kramer,  A.I.  Goldman,  D.  Robinson,  J.W.  Lee,  J.R.  Morris and  K.F.  Kelton: Phys. Rev. B 83 (2011) 094116.
• 25)   Y.  Yokoyama,  K.  Fukaura and  A.  Inoue: Mater. Sci. Eng. A 375–377 (2004) 427–431.
• 26)   P.  Yu,  H.Y.  Bai,  M.B.  Tang and  W.L.  Wang: J. Non-Cryst. Solids 351 (2005) 1328–1332.
• 27)   F.  Puosi,  N.  Jakse and  A.  Pasturel: J. Phys. Condens. Matter 30 (2018) 145701.
• 28)  J.F. O’Hanlon: A User’s Guide to Vacuum Technology, (John Wiley & Sons, Inc., New Jersey, 2003).
• 29)   J.C.  Ye,  J.  Lu,  C.T.  Liu,  Q.  Wang and  Y.  Yang: Nat. Mater. 9 (2010) 619.
• 30)   T.  Miyauchi,  H.  Kato and  E.  Abe: Mater. Res. Lett. 2 (2014) 94–99.
• 31)   L.S.  Luo,  B.B.  Wang,  F.Y.  Dong,  Y.Q.  Su,  E.Y.  Guo,  Y.J.  Xu,  M.Y.  Wang,  L.  Wang,  J.X.  Yu,  R.O.  Ritchie,  J.J.  Guo and  H.Z.  Fu: Acta Mater. 171 (2019) 216–230.
• 32)   H.  Mizubayashi,  T.  Takahashi,  K.  Nakamoto and  H.  Tanimoto: Mater. Trans. 48 (2007) 1822–1827.
• 33)   T.  Egami,  T.  Iwashita and  W.  Dmowski: Metals 3 (2013) 77–113.
• 34)   G.  Kaltenboeck,  M.D.  Demetriou,  S.  Roberts and  W.L.  Johnson: Nat. Commun. 7 (2016) 10576.
• 35)   P.  Yiu,  C.H.  Hsueh and  C.H.  Shek: J. Alloy. Compd. 658 (2016) 795–799.
• 36)   C.  Ma,  S.  Suslov,  C.  Ye and  Y.  Dong: Mater. Des. 165 (2019) 107581.
• 37)   Y.  Huang,  H.  Fan,  S.  Guan,  Z.  Ning,  F.  Cao,  D.  Daisenberg and  J.  Sun: J. Alloy. Compd. 789 (2019) 704–711.
• 38)   T.  Xi,  T.  Sato,  R.  Suzuki and  H.  Tanimoto: Mater. Trans. 59 (2018) 47–52.
• 39)   P.  Keblinski,  S.R.  Phillpot,  D.  Wolf and  H.  Gleiter: Phys. Rev. Lett. 77 (1996) 2965–2968.
• 40)   D.  Wolf: Curr. Opin. Solid State Mater. Sci. 5 (2001) 435–443.