Activities in the system cryolite-alumina

Reasons are outlined for thinking that accepted values for the activity of Al(2)O(3) in solution in cryolite are in error, and a new analysis is made of literature cryoscopic data. A general treatment is given for the effect of incongruent melting caused by the nonstoichiometry of solid cryolite. The effect is substantial, and accurate calculations require a knowledge of the effect of the solute on the activity of NaF in the solution, since that controls the composition of the solid separating. The results in the case of alumina as solute show that the slope of the log (activity) vs log (concentration) plot is about 3 in dilute solution and 1.5 in concentrated solution, corresponding to species containing 1 and 2 atoms of oxygen, respectively. Any calculations employing the earlier relation a(Al2O3) proportional to (wt pct Al(2)O(3))(2.77), which was based on electrochemical measurements of activity, need to be redone.