IONIC TRANSPORT-PROPERTIES IN OXIDE GLASSES DERIVED FROM ATOMIC-STRUCTURE

The atomic environments of alkalis are affected by the presence of other alkalis, suggesting that they are juxtaposed in the glass structure. Indeed there is mounting evidence from local structure spectroscopies, e.g., X-ray absorption fine structure (XAFS) from alkalis and oxygen magic angle spinning nuclear magnetic resonance (MAS-NMR), pointing to alkalis being coordinated primarily to non-bridging oxygens and microsegregated in silicate glasses. Reconfiguration of oxygens must therefore accompany the nearest neighbour hopping of alkalis. Simple expressions for the conformational energy associated with this and for the electrostatic energy contributions to the total microscopic energy barrier, E(a), facing the migrating alkali are presented. Cooperative effects manifest in the ac electrical conductivity, sigma(ac), are considered and it is argued that these come into play for high concentrations of alkali in the de electrical conductivity, sigma(dc), and in the self diffusion, D, and that the macroscopic activation energy, W,is related to E(a) by W=E(a)/beta, where beta is the Kohlrausch exponent which can be obtained from sigma(ac). It is proposed that since at low concentrations of alkali (i.e., below the percolation threshold) beta approaches unity, alkalis become decoupled and hopping to more distant sites will become more frequent in consequence, with E(a) approximating to the binding energy of the alkali. In mixed alkali glasses, similar considerations apply to each of the alkalis. The majority alkali will mainly engage in nearest neighbour hopping, largely unaffected by the presence other alkali, while the minority alkali, because it is largely surrounded by foreign alkalis, will be constrained to hop to more distant sites. By parameterising these expressions for E(a) for both single and mixed alkali glasses from XAFS and MAS-NMR results, one is able to predict W for sigma(dc) and D, obtaining good agreement with literature values for sodium and potassium silicates. Simple hopping expressions are used to predict the pre-exponents for these glasses in terms of the alkali concentration, the hopping distance and the hopping attempt frequency, all which can be obtained from the glass composition, alkali XAFS and far IR absorption, respectively. Agreement with transport values is encouraging. From the pre-exponents and activation energies, one can therefore quantitatively predict sigma(dc) and D at any temperature, including the pronounced minimum in the sigma(dc) isotherms across a sodium-potassium silicate glass series.