In work reported previously, measurements of silver and gold tracer activities were made in silver-gold-alloy Kirkendall diffusion couples in which the tracers were originally placed in the interface. Semilogarithmic plots of these data were straight lines, but the coefficients measured were not self-diffusion coefficients because of the effects of the composition gradient. These straight-line graphs are now explained, and an approximate equation for the diffusion coefficients is derived from the theory reported earlier. Some of the measured values disagree with the calculation and the disagreement is outside of the estimated errors.
Read full abstract