It was recalled that the tracer diffusion of Co and Ga had formerly been described in terms of a triple-defect mechanism which was based upon a divacancy mechanism, and the existence of triple-defect disorder. By using interaction energies which had been used to describe the partitioning of vacancies and atoms in Co-Ga, Monte Carlo simulations were made of tracer diffusivities in Co-Ga. Use was made of a general isolated vacancy diffusion model, with nearest-neighbour jumps and no explicit or implicit binding between the vacancies. The model was fully consistent with triple-defect disorder in Co-Ga. It was shown that the tracer diffusion data could be described well without any need to invoke the triple-defect mechanism. It was found that Ga performed the equivalent of a next-nearest neighbour jump by jumping via the Co sub-lattice, where it had a very short residence time, whereas the triple-defect mechanism required a formal next-nearest neighbour jump for Ga. It was concluded that tracer diffusion measurements alone might not be able to resolve the difference.

A New Interpretation of Tracer Diffusion in Co-Ga. I.V.Belova, G.E.Murch: Philosophical Magazine A, 2001, 81[1], 95-108