Abstract
A statistical theory of the effect of solutes on solvent self-diffusion in dilute-alloys is given. It is based on the vacancy model used previously by Hoffman, Turnbull and Hart and includes a qualitative consideration of correlation effects. Two limiting cases are considered: (i) in which the vacancy jump frequencies are not greatly changed by the solute atoms (‘weak perturbation’ limit), (ii) in which the vacancies attach themselves strongly to solute atoms and where the jump frequencies may be greatly different (‘tight-binding’ limit). In both cases it is shown how experimental data on solute enhanced diffusion of solvent atoms combined with knowledge of solute tracer diffusion rates may be used to deduce ‘correlation factors’ for solute diffusion. Analysis of existing data on dilute Ag alloys yields similar correlation factors whichever of the two limiting equations we use. The results are not consistent with the absence of isotope effects in Cd and In diffusion in Ag: this may indicate that the model is inadequate for the discussion of solute enhanced diffusion. Further experimental work, particularly on isotope effects, is called for.