Effect of Electrolysis Potential and Molar Ratio of CaO to SiO₂ on Si Metal Deposition in Molten CaCl₂

Abstract

The electrodeposition of Si metal in molten CaCl₂ containing various types of calcium silicate has been studied at 1373∼1673 K. It was shown that the influence of the molar ratio of CaO to SiO₂ on Si electrodeposition was not remarkable in contrast with our previous results that the molar ratio of CaO to TiO₂ strongly affected the Ti metal deposition in the similar system. Silicon metal was obtained on MoSi₂ cathode in the melt by potentio-static electrolysis. The cathodic current efficiency was about 30% regardless of the added types of calcium silicate. The current efficiency tended to be better as the electrolysis potential became negative, but the electrolysis below the border potential lead to lower current efficiency due to Ca co-deposition.

Enough metallic Si to evaluate the current efficiency from its weight was obtained by potentio-static electrolysis in molten CaCl₂ with Ca₂SiO₄ at 1673 K. It was shown that the current efficiency of Si metal deposition depended on electrolysis potential, though the further study to clarify the optimum condition for better electrolysis is necessary.

1. Introduction

Metallic Si is produced by carbon reduction of SiO₂ in an arc furnace above 2000 K. Since the energy consumption and CO₂ exhaustion in the process are very large, an energy-efficient and environment-friendly process is desired. Si electrolysis in molten CaCl₂ has been studied by several groups.^1–4) The authors have also researched the Si electrodeposition in molten CaCl₂ containing CaSiO₃ above 1273 K to obtain metallic silicon in liquid; CaSiO₃ dissolved in molten CaCl₂ at 1573 K, and the cathodic reaction of Si occurred.⁵⁾ A thick MoSi₂ layer was formed on Mo cathode by potentio-static electrolysis and metallic Si was detected in the electrodeposit. These results suggested that the cathodic reduction to metallic Si could be possible in the system. However, the suitable condition of Si metal deposition had not been studied in detail.

In this study, various types of calcium silicate were used as solute, and the dependence of cathodic reaction on the molar ratio of CaO to SiO₂ (R_CaO/SiO2, hereinafter) of calcium silicate was investigated based on our previous study in molten CaCl₂ with calcium titanate.^6,7) The electrochemical behavior of calcium silicate was analyzed by cyclic voltammetry, and potentio-static electrolysis was performed. MoSi₂ electrode was used as the cathode to prevent further MoSi₂ formation, and the dependence of the cathodic current efficiency on the electrolysis potential was discussed.

2. Experimental

Commercial CaSiO₃ (Wako Pure Chemical, CaO: 40.0∼48.0, SiO₂: 51.0∼57.0%) and calcium silicates of various R_CaO/SiO2 values synthesized from CaO (Kishida Chemical, >98%) and SiO₂ (Kojundo Chemical Laboratory, >99.9%) were used. It was confirmed by that the synthesized calcium silicate consisted of Ca₂SiO₄, Ca₃Si₂O₇ and CaO. Their compositions changed with R_CaO/SiO2, which were consistent with the phase diagram of the CaO–SiO₂ binary system.⁸⁾ In this paper, molten CaCl₂ containing the calcium silicate of R_CaO/SiO2 = 1.5 is simply written as “the bath of R_CaO/SiO2 = 1.5”, for example, and the concentration of calcium silicate was 5.0 mol% based on the SiO₂ concentration.

The apparatus for electrochemical measurement and electrolysis was almost the same as in our former studies.⁵⁾ Calcium chloride (Kishida Chemical, >95%) containing the calcium silicate was put into a Mo crucible (3N5, Sanwa Metal), and vacuum-dried at 373 K for a day. The salt in a Mo crucible was set in an air-tight furnace tube in an electric furnace (Motoyama, MS-2821). The system was heated, and the salt was held at 1373∼1673 K under a high-purity Ar atmosphere.

A MoSi₂ rod (Kanthal Super 1700, ϕ3.0 mm) of which active surface area limited with an Al₂O₃ sheath was used as the working electrode for cyclic voltammetry. A MoSi₂ electrode with a BN receiver (TEP, pure grade) as shown in Fig. 1 was also used at 1673 K to collect the electrodeposit efficiently for potentio-static electrolysis. A graphite rod (Nippon Techno-Carbon, MF307, ϕ5.0 mm) was used as a counter electrode. A Mo wire was used as a quasi-reference electrode, and its potential was calibrated versus the Mo dissolution potential measured by cyclic voltammetry.

Fig. 1

MoSi₂ cathode with BN receiver.

The electrochemical behavior of Si was investigated by cyclic voltammetry, and then potentio-static electrolysis was performed. The electrodeposit was analyzed by XRD (Rigaku, RINT-2550V) and SEM (JEOL, JCM-6000) - EDX (JEOL, EX-5441).

3. Results and Discussion

Cyclic voltammograms in the baths of R_CaO/SiO2 = 1.0, 1.5 and 2.0 at 1373 K are shown in Fig. 2. In the bath of R_CaO/SiO2 = 1.0, a reduction current peak was observed at −1.3 V. A current bulge was seen around −1.2 in the bath of R_CaO/SiO2 = 1.5, and around −1.4 V in the bath of R_CaO/SiO2 = 2.0. Considering the results of potentio-static electrolysis mentioned after and the report that one-step reduction of Si⁴⁺ + 4e⁻ → Si occurred in the LiF–NF–KF–SiO₂ system,⁹⁾ it is thought that silicate ion was reduced to Si metal by one step under the condition in this study. The reduction step seemed not to depend on the R_CaO/SiO2 value, but the reduction potential was influenced.

Fig. 2

Typical cyclic voltammograms at 1373 K.

Clear current peaks or bulges could not be observed in the cyclic voltammogram above 1573 K. It is considered that stronger convection in the bath at a higher temperature affected the cyclic voltammogram.

Potentio-static electrolysis was performed based on the results of the cyclic voltammogram, and a deposit piled up on the MoSi₂ cathode in all the baths of R_CaO/SiO2 = 1.0, 1.5 and 2.0 at 1573 K. Metallic Si particle was contained in the deposit, and the bath components were included. The XRD pattern of the typical deposit after rinsing is shown in Fig. 3. The deposit mainly consisted of metallic Si, and silicon carbide was not detected in it. The R_CaO/SiO2 value influenced the deposit; metallic Si seemed to be obtained more easily in the baths of R_CaO/SiO2 = 1.5 and 2.0 than in the bath of R_CaO/SiO2 = 1.0. The electrolysis potential also affects the deposition, and CaSi₂ was detected in the deposit according to the condition.

Fig. 3

XRD pattern of electrodeposit obtained by potentio-static electrolysis at 1573 K.

A thick layer of Si metal was formed by potentio-static electrolysis at 1673 K as shown in Fig. 4. Oxygen was slightly detected due to the surface oxidation during the sample preparation, while Ca was not contained in the deposit. The graphite anode was consumed by electrolysis. Considering the temperature and the contents of cathode deposit, it is thought that the reaction: C + O²⁻ → CO + 2e⁻ occurred at the anode.

Fig. 4

SEM image and elemental distributions of cross section of MoSi₂ electrode with electrodeposit obtained at 1673 K.

Figure 5 shows the current efficiency calculated from the weight of electrodeposit and the electric quantity during electrolysis assuming the reaction: Si⁴⁺ + 4e⁻ → Si. The current efficiency tended to better as the electrolysis potential became negative, but it dropped sharply at −2.0 V. CaSi₂ was detected from the deposit by the electrolysis at −2.0 V, so co-deposition of Ca might hinder metallic Si deposition.

Fig. 5

Dependence of current efficiency for Si electrodeposition on electrolysis potential in molten CaCl₂ with Ca₂SiO₄ at 1673 K.

It was clearly indicated that metallic Si could be obtained by electrolysis in molten CaCl₂ containing calcium silicate. However, the current efficiency was still limited to around 30% at most. Since it was shown that R_CaO/SiO2 and electrolysis potential strongly affected the Si deposition as mentioned above, the detailed study on their influence should be necessary for better electrolysis. Metallic Si was deposited in solid under the condition in this study, so the detachment by strong convection in the bath might cause the low current efficiency. The electrolysis above the melting point of Si should be also attempted.

4. Conclusion

Metallic Si was electrodeposited by one-step reduction; the electrolysis potential and temperature strongly influenced the Si deposition, whereas the dependence on R_CaO/SiO2 value was not remarkable as compared with our former results on Ti deposition in molten CaCl₂ containing calcium titanate. A thick metallic Si layer was formed on MoSi₂ cathode by the potentio-static electrolysis at 1673 K, and the current efficiency was about 30% under the suitable condition. A detailed study to realize more efficient electrolysis should be necessary, and also the research to obtain metallic Si in liquid should be desirable.

Acknowledgments

The study was supported by the Ministry of Education, Culture, Sports, Science and Technology, Japan (Grant-in-Aid for Scientific Research (B), #18H01763).

REFERENCES

• 1)   K.  Yasuda,  T.  Nohira and  Y.  Ito: J. Phys. Chem. Solids 66 (2005) 443–447. doi:10.1016/j.jpcs.2004.06.037
• 2)   K.  Yasuda,  T.  Nohira,  R.  Hagiwara and  Y.-H.  Ogata: Electrochim. Acta 53 (2007) 106–110. doi:10.1016/j.electacta.2007.01.024
• 3)   Z.  Chen,  D.J.  Fray and  T.W.  Farthing: Nature 407 (2000) 361–364. doi:10.1038/35030069
• 4)   X.  Yang,  L.  Ji,  X.  Zou,  T.  Lim,  J.  Zhao,  E.T.  Yu and  A.J.  Bard: Angew. Chem. Int. Ed. 56 (2017) 15078–15082. doi:10.1002/anie.201707635
• 5)   T.  Takenaka,  R.  Shimokawa,  H.  Okada and  T.  Morishige: ECS Trans. 86(14) (2018) 29. doi:10.1149/08614.0029ecst
• 6)   H.  Okada,  K.  Shimokawa,  T.  Morishige and  T.  Takenaka: Mater. Trans. 59 (2018) 690–693. doi:10.2320/matertrans.MBW201702
• 7)   T.  Takenaka,  H.  Okada,  R.  Shimokawa and  T.  Morishige: MATEC 321 (2020) 07010. doi:10.1051/matecconf/202032107010
• 8)   B.  Phillips and  A.  Muan: J. Am. Ceram. Soc. 42 (1959) 413–423. doi:10.1111/j.1151-2916.1959.tb12966.x
• 9)   Y.  Sakanaka and  T.  Goto: Electrochim. Acta 164 (2015) 139–142. doi:10.1016/j.electacta.2014.12.159