The diffusion of silver and gold tracers in silver-gold in crystals of 0, 8, 17, 35, 50, 66, 83, 94 and 100 at. % gold has been measured. It is shown that the limiting error in such measurements is due to temperature uncertainty rather than to the sectioning process. The activation energies obtained do not vary in proportion to the melting point or heat of fusion, and the deviations cannot be rectified in terms of lattice parameter arguments. The activation energies in the pure metals are better accounted for by the theory of Turnbull and Hoffman than by that of Swalin. The suggestion that the vacancy migration energy should vary as (cu -Cn) is not confirmed. From the change in frequency factor with composition it is deduced that the activation entropy of migration of a vacancy decreases linearly with composition by 1.5R from pure silver to pure gold. The dependence of diffusion coefficient on gold content is compared with the theories of Hoffman, Turnbull, and Hart, of Reiss, of Manning, and of Lidiard, and impurity correlation factors of reasonable magnitudes are obtained. Particularly difficult to treat, however, is the decrease in diffusion coefficients resulting from additions of the rapid diffuser silver to the slow diffuser gold. An experiment to measure the effects of vacancy flux directly is proposed.

No abstract

No abstract

No abstract