Physical chemistry in halide metallurgy

which immediately shows (4) that the transfer of condensed aluminum into the vapour state via the aluminum monochloride is significantly less endothermic than the direct evaporation and can, therefore, be expected to occur at lower temperatures. (The data used (4) at that time in kcal were hI = -89.5, h3 = -71.2 and H(ev) = 74.6; for a quantitative evaluation see the thermodynamics of Reaction 3 below using more modern data.) At about the time the author came to these conclusions, Klemm and Voss (5) quite independently investigated experimentally the reaction between liquid aluminum and solid aluminum fluoride by the method which Zintl (6) had applied for the identification (and preparation) of volatile unstable lower oxides and attributed Willmore's discovery to the formation of a volatile aluminum monofluoride. However, the use of the vapour of aluminum cWoride instead of Willmore's solid fluorides would have the great advantage that the resulting monohalide vapour should, on cooling, disproportionate into solid or liquid aluminum and aluminum trichloride vapour which, because of its low sublimation temperature (182°C), must remain in the gas phase. The aluminum should therefore condense free from any admixture of the halide. The expectations about the transport of aluminum in aluminum chloride (and also bromide) vapour were confirmed in experiments (7) in which aluminum chloride vapour under reduced pressure was passed over aluminum and aluminum scrap (7.5 per cent metallic impurities) at temperatures of about 1250K. A condensate of pure aluminum, free from any cWoride, was obtained in far greater amounts than could have been expected by evaporation alone. At 1250K and at pressures below 1/100 atm the yield of distilled aluminum corresponded to nearly total (80 to 90 per cent) conversion of aluminum trichloride into monochloride; it fell to lower values at lower II AI(Q) = AI(g) + H(ev)