A theoretical model for the mixed-alkali effect was proposed which was applicable to quasi 2-dimensional ionic conductors. It was based upon a consideration of the long-range Coulomb fluctuations that were produced by random fields, of the other alkali, and affected the barrier heights for the migration of cations. At low concentrations of the other alkali, the activation energy for conductivity was an almost linear function of either alkali concentration; in the 3-dimensional case. However, in the case of a quasi 2-dimensional cationic conductor such as Na/Ag -alumina, the theoretical value of the activation energy differed from the 3-dimensional value by a logarithmic multiplier that depended upon the composition and temperature. This model furnished predictions which were in good quantitative agreement with experimental data.

V.N.Bondarev: Solid State Ionics, 1996, 89[1-2], 93-8