Abstract
In the last few years reliable low-temperature diffusion coefficients have been measured for a number of impurity diffusion systems with the aid of microsectioning techniques. Using this data together with the macrosectioning data above 70% of the melting temperature, a curved Arrhenius plot can be obtained. These slightly curved Arrhenius plots have been analysed with respect to monovacancy and divacancy contributions by applying the Morrison routine (1975). Very accurate monovacancy diffusion parameters have been obtained for Ni, Co and In in copper and for In in nickel, while the divacancy parameters show large uncertainties. Analyses reveal that there are only small discrepancies between ΔQ1V, the difference between the monovacancy diffusion energies of impurity and self-diffusion, and ΔQ, the difference between the effective diffusion energies measured in the high-temperature range (> 70% of the melting temperature). While for impurity diffusion in copper marked divacancy contributions are observed, there are indications that the divacancy contributions in silver are distinctly smaller, at least for those impurities that diffuse faster than silver.