Abstract
The system NaF-AlF3 constitutes the main part of the electrolyte used in primary aluminum production. While it is generally accepted that molten NaF-AlF3 contains Na+, F−, and a number of fluoro-aluminate anion complexes, the detailed structure of the melt is still disputed. In the current study, five structure models, which all contained five anionic entities and Na+ as the only cation, were examined. The activities of NaF(l) and AlF3(s) as computed from ideal Temkin models were compared with the formal activities. The structure models were also used to study the conditions at the cathode. The current is carried by Na+; hence, there is a net diffusion flux of aluminum-containing species toward the cathode. The individual ion fraction gradients and fluxes were computed, based on the Stefan–Maxwell equations for transport in a multicomponent system and the prerequisite that the ionic species should be at equilibrium at any point. The five structure models tested showed similar behavior. The change in the NaF/AlF3 ratio across the boundary layer was almost equal for the four models, while the cathodic concentration overvoltage varied from 81 to 112 mV. The results give no reasons for preferring one structure model above the other.
