Availability Evaluation of SiC Anode in Chloride Molten Salt Electrolysis

Abstract

INTRODUCTION Carbon or graphite is commonly used to anode material in molten salt electrolysis because of their stability and reasonable price. However, they are consumed with CO or CO2generation by electrochemical reaction with oxide ion, and the consumption is undesirable in electrolysis in chloride melt. In this study, the anodic behavior of carbide was investigated to develop the inert anode material. EXPERIMENTAL A eutectic mixture of LiCl-KCl (41:59 in mole) with Li2O was used as an electrolytic bath. Anode was a silicon carbide (SiC) rod, and a graphite rod was also used for comparison. Cathode was an Al plate, and reference electrode was a Ag/Ag+ couple, and its potential was calibrated with a Li/Li+couple. The electrodes were inserted into the molten salt at 723 K in an Ar atmosphere. Reactions were studied by cyclic voltammetry and potentiostatic electrolysis. The generated gas by potentiostatic electrolysis was analyzed by gas chromatography. After the electrolysis, the specimen was analyzed by SEM-EDX and XRD. RESULTS and DISCUSSION Fig.1 shows typical cyclic voltammograms at the SiC and graphite anode. At the graphite electrode, oxidation current increased around 2.65V (vs. Li/Li+, the same hereinafter) and 3.5V, and arose suddenly above 3.8V. At the SiC electrode, oxidation current was increased around 2.65V, 3.2V, 3.5V and arose suddenly above 4.0V. Bubble formation was seen around 3.5V and above 3.8V (4.0V) at both electrodes. The generated gas at 3.5V was identified as CO2 by gas chromatograph, and it was considered that the generated gas above 3.8V (4.0V) was Cl2. Gas was not generated around 3.2V at the SiC electrode, and SiO2was detected on by XRD after potentiostatic electrolysis at 3.2V. The mass of the SiC electrode increased by potentiostatic electrolysis at 3.2V for 4h and 6h, while it decreased by the electrolysis for 8h, as shown in Table 1. The mass didn’t change by potentiostatic electrolysis at 3.5V for 4h, but it decreased by the electrolysis for 6h and 8h. The mass of the SiC electrode tended to decreased with the electrolysis time at both potentials. The mass decreased by the electrolysis at 4.5V even for 4h. The O content after the potentiostatic electrolysis at higer potential was smaller than that at lower potential as shown in Table 2. The surface of the SiC electrode after the electrolysis was uneven and a dense SiO2layer was not found at the cross-section. From the result mentioned above, SiO2 formation occurred below 3.2V at SiC, and CO2 was generated around 3.5V. This result indicate that SiO2 formation precedes CO2 generation at SiC, but that the formed SiO2 cannot prevent the consumption of SiC by CO2 generation. The formed SiO2 layer was still uneven and partially peered out during electrolysis. The formation of a dense and coherent SiO2 layer should be necessary to prevent the consumption of SiC by electrolysis in molten chloride. Figure 1