Ole Edvard Kongstein

Purification of metallurgical grade silicon by electrorefining in molten salts

Abstract Electrochemical studies on silicon deposition were performed in molten salt electrolytes. Purification of metallurgical grade silicon by electrorefining was carried out in molten Si–chloride salts at temperatures from 973 K to 1223 K. It was found that the use of a liquid alloy anode of silicon and copper was beneficial in molten CaCl 2 with NaCl, CaO and dissolved Si. ICP-AES analysis results showed efficient removal of metal impurities, such as titanium, aluminum and iron, which are present in significant quantities in the feedstock. The contents of boron and phosphorus in the silicon after electrorefining were reduced from 36×10 −6 and 25×10 −6 to 4.6×10 −6 and 2.8 ×10 −6 , respectively. The energy consumption of electrorefining was estimated to be about 9.3 kW·h/kg.

Anodic Deposition of Cobalt Oxide during Electrowinning of Cobalt in a Chloride Electrolyte

During electrowinning of cobalt in a chloride electrolyte, in addition to anodic chlorine and oxygen evolutions, precipitation of a black substance was observed on the dimensionally stable anode. The black substance was investigated by means of X-ray diffraction and found to be trivalent cobalt oxy-hydroxide (α-CoOOH). The deposition of cobalt oxy-hydroxide and possible ways to prevent this deposition were studied. Experiments showed that the deposited cobalt oxy-hydroxide could be dissolved cathodically. The amount of charge for cathodic dissolution was used to estimate the deposited amount of cobalt oxy-hydroxide. The partial current density for the cobalt oxy-hydroxide formation increased with increasing anodic potential, indicating an electrochemical deposition reaction [Co 2+ (aq) + 2H 2 O → 3H + (aq) + CoOOH + e - ]. In agreement with thermodynamics, the deposition of cobalt oxy-hydroxide decreased with decreasing pH. The deposition was no longer observed at low pH (pH = 1.0), in good agreement with the thermodynamic calculations. The results showed that the cathodic pulses on the anode during electrowinning can be used to prevent accumulation of cobalt oxy-hydroxide on the anode. It was also found that the addition of small amounts of hydrogen peroxide suppressed the deposition of cobalt oxy-hydroxide.

Proton Transport During Electrodeposition of Cobalt in Chloride Electrolytes

In acid cobalt chloride electrolytes the electrodeposition of cobalt and evolution of hydrogen take place simultaneously. On a rotating platinum disc electrode the current efficiency for cobalt deposition was calculated based on the anodic charge needed for stripping of the cobalt deposits divided by the total cathodic charge. The results showed a decrease in current efficiency with increasing rotation rate. From the partial current density for hydrogen evolution the diffusion coefficient for protons was calculated. The derived diffusion coefficients varied with overpotential and pH, but were always one order of magnitude lower compared to what was expected from literature values. This was attributed to water drag when transporting cobalt ions towards the electrode. The variations in the calculated diffusion coefficients were explained by change in the transfer number when the proton concentration was changed.

Current Efficiency in Hall-Héroult Cells: The Role of Mass Transfer at the Cathode

The current efficiency (CE) in aluminium cells is governed by transport of dissolved metal (mainly sodium) across the boundary layer at the cathode. The transport takes place by ordinary mass transfer, but since solutions of alkali metals and their salts show electronic conductivity, the CE is also influenced by loss of electrons. The dependence of convection is not necessarily the same for the two loss mechanisms. A laboratory experiment was designed, where the mass transfer coefficient in the so-called Sterten-Solli laboratory cell for measuring CE was varied in a controlled manner by means of a mechanical stirrer. The effect of stirring on the mass transfer coefficient (k) was first surveyed by recording the limiting current density for potassium ferri- and ferrocyanide in an aqueous electrolyte as a function of the stirring rate, followed by measuring the CE in cryolitic melts at different stirring rates. It turned out that plots of the CE versus the mass transfer coefficient produced straight lines that extrapolated back to 99% CE at k = 0. This means that predictions of the CE can be made by using equations for ordinary mass transfer.