Effect of Halide Ions on Electrodeposition Behavior and Morphology of Electrolytic Copper Powder

Abstract

To investigate the effect of halide ions on the electrodeposition behavior and morphology of copper powder, polarization curves were measured and constant current electrolysis of 300 and 500 A·m⁻² was conducted in an electrolytic solution containing 0.079 mol·dm⁻³ Cu²⁺ and 0.5 mol·dm⁻³ free H₂SO₄ at 293 and 303 K without stirring. Cl⁻ promoted the deposition of copper powder, while Br⁻ and I⁻ suppressed deposition. The current efficiency for copper deposition increased with the addition of Cl⁻ and decreased with the addition of Br⁻. The addition of Cl⁻ reduced the average particle size of the copper powder and caused the dendrite-shaped branches and trunks to grow thinner and longer, resulting in a lower tap density. In contrast, the addition of Br⁻ caused the average particle size, average crystallite size, and tap density of the copper powder to decrease. With increasing Cl⁻ concentration, the current efficiency for copper deposition increased, that is, copper deposition was promoted. This even occurred in the region in which Cu²⁺ ion diffusion was the rate-determining process, indicating that the deposition of copper powder was affected by the charge-transfer process. The change in the morphology of the copper powder with the addition of halide ions is attributed to the change in the charge-transfer process. The deposition of copper powder appears to proceed under a mixed rate-determining process involving the diffusion of Cu²⁺ ions and charge transfer.

 

This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 85 (2021) 207–212. Captions of all figures and tables are slightly modified.

Fig. 7 SEM images of copper powders deposited from electrolytes with various Cl⁻ concentrations: (a) 0 mg·dm⁻³, (b) 10 mg·dm⁻³, (c) 100 mg·dm⁻³, and (d) 200 mg·dm⁻³.

1. Introduction

Copper powder is widely used in powder metallurgy and electronic applications owing to its excellent electrical and thermal conductivities. There are four well-known commercial manufacturing methods for producing copper powder: a chemical reduction method, an atomization method, a dropping method in liquid, and an electrolysis method.^1–4) The advantages of the electrolysis method over the other methods are low manufacturing cost, low energy consumption, and high purity.

Electrolytic copper powder is generally shaped like a tree branch called a dendrite. Because of this shape, electrodeposition at one branch interferes with that at others, making it difficult to fill the interdendritic region. Consequently, electrolytic copper powder has a low density. In electronic applications, this low density is important. For example, for conductive pastes prepared by mixing resin and copper powder, the lower density of electrolytic copper powder relative to that of spherical copper powder allows the resin to be imparted with conductivity by adding a small amount of copper powder, and the properties of the resin, such as softness and flexibility, can be sufficiently maintained. With the miniaturization of electronic devices, electronic materials that use copper powder are becoming increasingly smaller and thinner. Therefore, the electrolytic copper powder used in electronics must also be fine. Many studies have reported methods for controlling the shape of electrolytic copper powder.^5–8)

This study focused on the correlation between the particle size and crystallite size of electrolytic copper powder. The particle size is the average particle size of the electrolytic copper powder, while the crystallite size is the average size of the smallest unit that can be regarded as a single crystal among the crystal grains constituting the individual electrolytic copper powder particles. The overpotential during electrolysis has a substantial influence on crystallite size. As the copper reduction reaction is more suppressed and the overpotential increases, the nucleation rate becomes faster than growth rate of the nuclei, resulting in decrease in crystallite size. In the electrolytic refining of copper, it has been reported that the presence of halide ions in the electrolytic solution suppresses the copper reduction reaction and reduces the size of the electrodeposited crystal grains.⁹⁾ For electrolytic copper powder, the presence of halide ions in the electrolytic solution may reduce the crystallite and particle sizes. However, the mechanism differs greatly between electrolytic refining, which reduces copper ions into plates, and electrodeposition of copper powder, which reduces copper ions into powder. In general, electrolytic refining of copper is performed under charge-transfer rate-determining conditions, in which electron transfer on the electrode surface is the rate-determining process. In contrast, electrodeposition of copper powder is performed under diffusion rate-determining conditions, in which the supply of copper ions is rate-determining factor. In this study, we used electrochemical measurements to investigate the effect of halide ions on the electrodeposition behavior of copper powder, even under the diffusion rate-determining conditions in which copper powder is electrodeposited. In addition, we electrodeposited copper powder, investigated the effect of halide ions on the morphology of copper powder, and determined whether copper powder with a small particle size could be electrodeposited.

2. Experimental

For the electrochemical measurements, a four-necked flask (0.3 dm³) was used as the electrolytic cell. An electrolytic solution was prepared using copper sulfate (Mitsui Mining & Smelting Co., Ltd.) and a commercially available special-grade reagent. They were dissolved in pure water to achieve a Cu²⁺ concentration of 0.079 mol·dm⁻³ and a free H₂SO₄ concentration of 0.5 mol·dm⁻³. Cl⁻, Br⁻, and I⁻ were used as the halide ions. Cl⁻, Br⁻, and I⁻ were added in the form of hydrogen chloride, hydrogen bromide, and hydrogen iodide, respectively, to achieve concentrations of 10 to 1000 mg·dm⁻³.

The measurement of polarization curve and constant current electrolysis were performed by the three-electrode method. A copper plate (0.2 cm²) with one side masked with insulating tape was used as the working electrode, a platinum wire was used as the counter electrode, and an Ag/AgCl (saturated KCl, 0.199 V vs. NHE, 298 K) electrode was used as the reference electrode. The bath temperature was 293 K, and it was operated under non-stirring conditions. The polarization curves were measured at a potential scanning speed of 60 mV·min⁻¹ in the potential range from the immersion potential to −1.0 V. In electrolysis at constant current, referring to the measurement results of the polarization curve, electrolysis was performed for 5 min at 300 A·m⁻², where the copper electrodeposition reaction was limited by the diffusion of Cu²⁺ ions.

An electrolytic cell with a capacity of 80 dm³ and a buffer tank with a capacity of 100 dm³ were used for the copper powder electrolysis test. Copper sulfate (Mitsui Mining & Smelting Co., Ltd.) and industrial sulfuric acid were used for the electrolytic solution; they were dissolved in pure water to achieve a Cu²⁺ concentration of 0.079 mol·dm⁻³ and a free H₂SO₄ concentration of 0.5 mol·dm⁻³. For the halide ions, Cl⁻ and Br⁻ were added in the form of hydrogen chloride and hydrogen bromide, respectively, to achieve concentrations of 10 to 200 mg·dm⁻³. Because I⁻ generates a toxic gas with a strong distinct odor during electrolysis, its use in the copper powder electrolysis test was postponed. The bath temperature was 303 K, and the amount of electrolytic solution circulating between the electrolytic cell and buffer tank was set to 3 dm³·min⁻¹. An oxygen-evolving electrode (5 × 7 cm²) with a titanium substrate coated with iridium oxide was used as the anode, and a titanium plate (5 × 7 cm²) was used as the cathode. The current density was 500 A·m⁻², and the electrolysis time was 120 min. The electrodeposited copper powder was washed and dried, and the powder characteristics were measured. The average particle size was measured by the volume cumulative calculation method using a laser diffraction/scattering particle size distribution analyzer. The crystallite size was calculated from the half width of 200 peaks in the X-ray diffraction pattern using the Scherrer equation.¹⁰⁾ The tap density was measured according to JIS Z 2512: 2006 using a tapping device. The particle shape was observed using scanning electron microscopy (SEM).

3. Results

3.1 Deposition behavior of copper powder

Figure 1 shows the cathode polarization curves during copper deposition from electrolytic solutions containing Cl⁻. Regardless of the Cl⁻ concentration, copper deposition started at approximately 0 V, the current density became constant at approximately −0.15 V, and the current density began increasing when the potential became less noble than −0.6 V. In the charge-transfer rate-determining region of copper deposition (0 to −0.15 V), the current density increased as the Cl⁻ concentration increased. In the region between −0.2 and −0.6 V, the current density became constant, exhibiting the diffusion-limited current density of Cu²⁺ ions. In this region, the Cl⁻ concentration had no observable effect. The increase in current density in the region less noble than −0.6 V was due to hydrogen evolution. In the hydrogen evolution region, the potential at which the current density began to increase became more noble when the Cl⁻ concentration was 1000 mg·dm⁻³, and the current density was large because the deposition of copper powder increased the effective reaction area.

Fig. 1

Polarization curves for copper deposition in electrolytes with various Cl⁻ concentrations. (Cl⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

Figure 2 shows the cathode polarization curves during copper deposition from electrolytic solutions containing Br⁻. When Br⁻ was added, the current density decreased in the charge-transfer rate-determining region of copper deposition (0 to −0.2 V). The reduction in the current density was greatest when the Br⁻ concentration was 100 mg·dm⁻³. The addition of Br⁻ did not noticeably affect the diffusion-limited current density of Cu²⁺ ions and the current density in the potential region less noble than −0.8 V.

Fig. 2

Polarization curves for copper deposition in electrolytes with various Br⁻ concentrations. (Br⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

Figure 3 shows the cathode polarization curves during copper deposition from electrolytic solutions containing I⁻. When I⁻ was added, the current density decreased in both the charge-transfer rate-determining region (0 to −0.3 V) and the hydrogen evolution region (less noble than −0.7 V). In both cases, there was a large decrease in current density when the I⁻ concentration was 10 mg·dm⁻³. When the I⁻ concentration was 1000 mg·dm⁻³, turbidity was observed in the electrolytic solution. It is possible that CuI, which has very low solubility, was produced.

Fig. 3

Polarization curves for copper deposition in electrolytes with various I⁻ concentrations. (I⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

Based on the polarization curves, halide ions affect the charge-transfer process in the copper deposition reaction but do not affect the diffusion-limited current density of Cu²⁺ ions. Thus, halide ions are unlikely to affect the diffusion process of Cu²⁺ ions.

Figure 4 shows the evolution of the cathode potential over time during constant-current electrolysis at 300 A·m⁻² in electrolytic solutions containing Cl⁻. From the polarization curve shown in Fig. 1, the diffusion of Cu²⁺ ions is the rate-determining process at a current density of 300 A·m⁻². As shown in Fig. 4, a noble potential was temporarily observed at the start of electrolysis because charge transfer is the rate-determining process immediately after the start of electrolysis. As electrolysis progressed further, the polarization increased considerably, then the potential gradually became more noble because the effective reaction area increased with the deposition of copper powder. The potential became more noble as the Cl⁻ concentration increased, and the Cl⁻ had a depolarization effect on copper deposition.

Fig. 4

Evolution of the cathode potential over time during copper deposition at 300 A·m⁻² in electrolytes with various Cl⁻ concentrations. (Cl⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

Figure 5 shows the evolution of the cathode potential over time during constant-current electrolysis at 300 A·m⁻² in electrolytic solutions containing Br⁻. The cathode potential gradually became more noble with increasing electrolysis time, similar to that with the electrolytes containing Cl⁻. The potential became less noble as the Br⁻ concentration increased; that is, Br⁻ had a polarization effect on copper deposition.

Fig. 5

Evolution of the cathode potential over time during copper deposition at 300 A·m⁻² in electrolytes with various Br⁻ concentrations. (Br⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

Figure 6 shows the evolution of the cathode potential over time during constant-current electrolysis at 300 A·m⁻² in electrolytic solutions containing I⁻. The cathode potential gradually became more noble with increasing electrolysis time, similar to that with the electrolytes containing Cl⁻ and Br⁻. The potential became less noble as the I⁻ concentration increased; that is, I⁻ had a polarization effect on copper deposition. The addition of I⁻ had a larger polarization effect on copper deposition than that with the addition of Br⁻.

Fig. 6

Evolution of the cathode potential over time during copper deposition at 300 A·m⁻² in electrolytes with various I⁻ concentrations. (I⁻ concentration, ◇ 0 mg·dm⁻³, ● 10 mg·dm⁻³, △ 100 mg·dm⁻³, □ 200 mg·dm⁻³, ○ 1000 mg·dm⁻³)

3.2 Characteristics of electrolytic copper powder

Table 1 shows the current efficiencies and powder characteristics of the copper powders deposited from electrolytic solutions containing Cl⁻. The current efficiency of copper deposition increased as the Cl⁻ concentration increased, while the average particle size and tap density of the copper powder decreased. The decrease in tap density may indicate the growth of the dendrite branches. However, the addition of Cl⁻ did not change the crystallite size. Based on the electrochemical measurements shown in Figs. 1 and 4, it was unclear whether Cl⁻ had a polarization effect on copper deposition. Therefore, it is concluded that the addition of Cl⁻ does not affect the crystallites.

Table 1 Properties of copper powders electrodeposited from electrolytes with various Cl⁻ concentrations.

Figure 7 shows surface SEM images of the copper powders deposited from electrolytic solutions containing Cl⁻. From these SEM images, it was confirmed that the grain size of the copper powder became smaller, and the branches and trunks of the dendrites grew thinner and longer with increasing Cl⁻ concentration. It was thought that finer crystallites were required to make finer copper powder. However, when Cl⁻ was added to the electrolyte, even though the crystallite size did not change significantly, the average particle size decreased.

Fig. 7

SEM images of copper powders deposited from electrolytes with various Cl⁻ concentrations: (a) 0 mg·dm⁻³, (b) 10 mg·dm⁻³, (c) 100 mg·dm⁻³, and (d) 200 mg·dm⁻³.

Table 2 shows the current efficiencies and characteristics of the copper powders deposited from electrolytic solutions containing Br⁻. As the Br⁻ concentration increased, the current efficiency of copper deposition decreased, and the average particle size of the copper powder decreased. Furthermore, both the tap density and crystallite size of the copper powder were reduced by the addition of Br⁻. The polarization effect of Br⁻ on copper deposition was confirmed in the electrochemical measurements shown in Figs. 2 and 5. It is thought that the crystallite size of the copper powder became finer owing to the influence of the polarization effect. The concentrations of Cl⁻ and Br⁻ ions incorporated into the copper powder were measured, and Br⁻ was found to be incorporated at a concentration approximately 10 times higher than that of Cl⁻.

Table 2 Properties of copper powders electrodeposited from electrolytes with various Br⁻ concentrations.

Figure 8 shows SEM images of the surface of the copper powders deposited from electrolytic solutions containing Br⁻. The particle size of the copper powder decreased with increasing Br⁻ concentration.

Fig. 8

SEM images of copper powders deposited from electrolytes with various Br⁻ concentrations: (a) 0 mg·dm⁻³, (b) 10 mg·dm⁻³, (c) 100 mg·dm⁻³, and (d) 200 mg·dm⁻³.

4. Discussion

The addition of Cl⁻ and Br⁻ had different effects on the deposition behavior of the copper powder. As a result, the current efficiency and crystallite size of the copper powder also differed with the addition of Cl⁻ and Br⁻. The current efficiency of the copper powder increased with the addition of Cl⁻, while it decreased with the addition of Br⁻. In addition, the crystallite size of the copper powder was not noticeably different with the addition of Cl⁻, but it became finer with the addition of Br⁻.

From the potential–pCl diagram⁹⁾ of the Cu–Cl–H₂O solution, when Cl⁻ coexists in the solution above a certain concentration, Cu²⁺ ions are reduced to metallic copper via CuCl(s) or CuCl₂⁻, which are intermediates for Cu⁺. For copper deposition from a 0.7 mol·dm⁻³ Cu²⁺ solution, an experiment using a rotating ring disk electrode found that CuCl₂⁻ was produced in the presence of 350 mg·dm⁻³ of Cl⁻ in the electrolytic solution.⁹⁾

In this study, the Cu²⁺ concentration was as low as 0.079 mol·dm⁻³. Consequently, the Cl⁻/Cu²⁺ ratio was high, and it is expected that CuCl₂⁻ would be more likely to form as a reaction intermediate. Therefore, it is inferred that the deposition of copper powder from a solution containing Cl⁻ proceeds according to eqs. (1) and (2).   

\begin{equation} \text{Cu$^{2+}$} + \text{2Cl$^{-}$} + \text{e$^{-}$} \to \text{CuCl$_{2}{}^{-}$} \end{equation} (1)

  

\begin{equation} \text{CuCl$_{2}{}^{-}$} + \text{e$^{-}$} \to \text{Cu} + \text{2Cl$^{-}$} \end{equation} (2)

The addition of Cl⁻ improved the current efficiency of copper deposition (Table 1), which is thought to be due to the depolarization effect of Cl⁻ on copper deposition (Figs. 1 and 4). For the deposition of iron group metals, the presence of Cl⁻ promotes the deposition reaction via the reaction intermediates MCl⁺ and MCl_ad (M: Fe, Ni, Co).¹¹⁾ It is thought that the same effect occurred during copper deposition in this study, which promoted the reaction. Generally, when the deposition overpotential (difference between the equilibrium potential and deposition potential) becomes small, the nucleation rate of the deposited material becomes slower than its growth rate. As a result, the crystal grain size of the deposited copper increases. However, in this study, although the overpotential for copper deposition decreased owing to the coexistence of Cl⁻, the copper particle size decreased, which cannot be explained by the deposition overpotential theory. It is thought that the addition of Cl⁻ reduces the particle size of the copper powder because it causes the deposition reaction in eq. (2) to occur preferentially at the tips of highly reactive dendrites. If CuCl₂⁻ preferentially adsorbs at the tip of the dendrite, it will cause the dendrite to grow thin and long (Fig. 7), and as a result, the particle size of the copper powder will become finer.

In contrast, when Br⁻ coexists in the solution, Cu²⁺ ions are reduced to copper via CuBr_ad, which is an adsorption intermediate for Cu⁺. The solubility products of CuCl(s) and CuBr(s) are 1.9 × 10⁻⁷ and 5.3 × 10⁻⁹, respectively; thus, CuBr is more stable. For copper deposition from a 0.7 mol·dm⁻³ Cu²⁺ solution, an experiment using a rotating ring disk electrode found that CuBr was produced in the presence of 8 to 800 mg·dm⁻³ of Br⁻ in the bath.⁹⁾ Therefore, it is inferred that the deposition of copper powder from a solution containing Br⁻ proceeds according to eqs. (3) and (4).   

\begin{equation} \text{Cu$^{2+}$} + \text{Br$^{-}$} + \text{e$^{-}$} \to \text{CuBr$_{\text{ad}}$} \end{equation} (3)

  

\begin{equation} \text{CuBr$_{\text{ad}}$} + \text{e$^{-}$} \to \text{Cu} + \text{Br$^{-}$} \end{equation} (4)

The addition of Br⁻ reduced the current efficiency for copper deposition (Table 2), which is thought to be due to the polarization effect of Br⁻ on copper deposition (Figs. 2 and 5). Because CuBr has a small solubility product and is more stable, it is expected that the reduction reaction from CuBr_ad to metallic Cu (eq. (4)) will be slow. This is thought to be the reason that copper deposition was polarized when Br⁻ was added. Furthermore, it is thought that the addition of Br⁻ reduced the particle size of the copper powder more than the addition of Cl⁻ because it increased the deposition overpotential. Br⁻ was incorporated into the powder at a concentration nearly 10 times that of Cl⁻ (Tables 1 and 2) because the reduction reaction from CuBr_ad to metallic copper (eq. (4)) was slow; therefore, unreduced CuBr_ad was incorporated into the copper powder. Alternatively, it is possible that Br⁻ specifically adsorbed on the cathode surface was directly incorporated into the copper powder. It has been reported that the specific adsorption ability of Br⁻ is greater than that of Cl⁻.¹²⁾

When I⁻ was added to the solution, the polarization effect on copper deposition was even greater than that when Br⁻ was added (Figs. 3 and 6). The solubility product of CuI(s) is 1.4 × 10⁻¹², which is even smaller than that of CuBr, indicating CuI is more stable. Therefore, it is inferred that addition of I⁻ further suppressed the reduction reaction from CuI_ad to metallic copper.

Based on the polarization curves shown in Figs. 1–3, halide ions affect the charge-transfer process but not the Cu²⁺ diffusion process in the copper deposition reaction. In the potential region less noble than −0.8 V in Fig. 1, the current density increased with increasing Cl⁻ concentration. Because the diffusion of Cu²⁺ ions is the rate-determining process for ordinary smooth plating, this increase in current density is thought to be due to an increase in hydrogen evolution, and the current efficiency for copper deposition is expected to decrease. However, in the copper powder deposition test, the current efficiency increased with increasing Cl⁻ concentration (Table 1), demonstrating that copper deposition is promoted by Cl⁻ ions, even in the potential region less noble than −0.8 V. This suggests that copper deposition is also affected by the charge-transfer process, even in this potential range. It is thought that the change in the morphology of the copper powder due to the addition of halide ions was also affected by the charge-transfer process. In previous studies, it has been thought that the deposition of copper powder proceeds under a diffusion rate-determining process, in which the supply of Cu²⁺ ions is the rate-determining factor. However, based on the results of this study, it is highly probable that the deposition of copper powder proceeds under a mixed rate-determining process that includes the Cu²⁺ ion diffusion process and the charge-transfer process.

It was found that copper powder with a small particle size can be electrolyzed by adding halide ions to the electrolytic solution. Among them, Cl⁻ is considered to be superior for the production of copper powder because it reduces the particle size and improves the current efficiency.

5. Conclusion

Based on this investigation of the effect of halide ions on the electrodeposition behavior and morphology of copper powder, the following was found. For the deposition of copper powder, Cl⁻ had a depolarization effect, Br⁻ and I⁻ had a polarization effect, and the current efficiency increased when Cl⁻ was added and decreased when Br⁻ was added. When Cl⁻ was added, the average particle size of the copper powder decreased, and the branches and trunks of the dendrites grew thinner and longer, resulting in a smaller tap density. In contrast, when Br⁻ was added, the crystallite size and average particle size of the copper powder decreased owing to the polarization effect, and the tap density also decreased. For the deposition of copper powder, the current efficiency increased as the Cl⁻ concentration increased, and copper deposition was promoted, even in the region in which Cu²⁺ diffusion was the rate-determining process. This suggests that the deposition of copper powder is affected by the charge-transfer process. It is thought that the change in the morphology of the copper powder due to the addition of halide ions was also affected by the charge-transfer process. That is, the deposition of copper powder proceeds under a mixed rate-determining process that includes the Cu²⁺ ion diffusion process and the charge-transfer process.

REFERENCES

• 1)   Y.  Watanabe: Resources Processing 37 (1990) 130–137. doi:10.4144/rpsj1986.37.130
• 2)   K.  Tamura and  T.  Takeda: J. Jpn. Soc. Powder Powder Metallurgy 10 (1963) 153–159. doi:10.2497/jjspm.10.153
• 3)   T.  Abe and  Z.  Hara: J. Jpn. Soc. Powder Powder Metallurgy 16 (1969) 279–283. doi:10.2497/jjspm.16.279
• 4)   T.  Shirayanagi and  Y.  Nanya: J. Jpn. Soc. Powder Powder Metallurgy 11 (1964) 23–28. doi:10.2497/jjspm.11.23
• 5)   G.  Orhan and  G.  Hapçı: Powder Technol. 201 (2010) 57–63. doi:10.1016/j.powtec.2010.03.003
• 6)   N.D.  Nikolić,  L.  Avramović,  E.R.  Ivanović,  V.M.  Maksimović,  Z.  Baščarević and  N.  Ignjatović: Trans. Nonferrous Met. Soc. China 29 (2019) 1275–1284. doi:10.1016/S1003-6326(19)65034-X
• 7)   N.D.  Nikolić,  P.M.  Živković,  M.G.  Pavlović and  Z.  Baščarević: J. Serb. Chem. Soc. 84 (2019) 1209–1220. doi:10.2298/JSC190522066N
• 8)   K.I.  Popov,  M.G.  Pavlović,  M.D.  Maksimović and  S.S.  Krstajić: J. Appl. Electrochem. 8 (1978) 503–514. doi:10.1007/BF00610795
• 9)   H.  Nakano,  S.  Oue,  K.  Aoya,  H.  Kuboyama,  H.  Fukushima and  S.  Kobayashi: J. MMIJ 123 (2007) 165–170. doi:10.2473/journalofmmij.123.165
• 10)   T.  Watanabe: Hyomen Gijutsu 40 (1989) 280–286. doi:10.4139/sfj.40.280
• 11)   M.  Yasuda,  I.  Ohno and  S.  Haruyama: Hyomen Gijutsu 41 (1990) 312–317. doi:10.4139/sfj.41.312
• 12)  H.H. Bauer, translated by R. Tamamushi and G. Satoh: Electrodics, (Tokyo Kagaku Dojin, Tokyo, 1982) p. 73.