Tracer Diffusion in an Ordered Alloy: Application of the Path Probability and Monte Carlo Methods

摘要

Tracer diffusion technique has been extensively utilized to investigate diffusion phenomena and has contributed a great deal to the understanding of the phenomena. However, except for self diffusion and impurity diffusion, the meaning of tracer diffusion is not yet satisfactorily understood. Here we try to extend the understanding to concentrated alloys. Our major interest here is directed towards understanding the physical factors which control diffusion through the comparison of results obtained by the Path Probability Method (PPM) and those by the Monte Carlo simulation method (MCSM). Both the PPM and the MCSM are basically in the same category of statistical mechanical approaches applicable to random processes. The advantage of the Path Probability method in dealing with phenomena which occur in crystalline systems has been well established. However, the approximations which are inevitably introduced to make the analytical treatment tractable, although their meaning may be well-established in equilibrium statistical mechanics, sometimes introduce unwarranted consequences the origin of which is often hard to trace. On the other hand, the MCSM which can be carried out in a parallel fashion to the PPM provides, with care, numerically exact results. Thus a side-by-side comparison can give insight into the effect of approximations in the PPM. It was found that in the pair approximation of the CVM, the distribution in the completely random state is regarded as homogeneous (without fluctuations), and hence, the fluctuation in distribution is not well represented in the PPM. These examples thus show clearly how the comparison of analytical results with carefully carried out calculations by the MCSM guides the progress of theoretical treatments and gives insights into the mechanism of diffusion. (ERA citation 10:011333)