- © The Mineralogical Society Of America
Ionic or atomic diffusion controls chemical exchange between and within different crystalline, melt and fluid phases. It can control the kinetics of phase transitions, the rate at which minerals grow, the degree of compositional zoning in minerals, and other important geochemical processes. Diffusion also plays a central role in the rheological properties of minerals and melts. In diffusion creep stress is accommodated through ions migrating from regions of high stress to regions of low stress. This is achieved either through bulk diffusion or grain boundary diffusion. In dislocation creep the rate-limiting step is often dislocation climb—the process where a dislocation has to migrate out of the dislocation plane to avoid an obstacle—and this requires ionic diffusion.
There are two main reasons for studying diffusion theoretically. The first is to determine the atomistic mechanism of diffusion. This can provide understanding of the underlying mechanisms, and allow one to extrapolate results to other systems. When different sets of experimental results disagree, theory can often help to decide which is correct or explain the differences. But the most valuable reason for using theory is to predict diffusion properties for systems or conditions where no data exist. As will be shown, theoretical calculations can be used to predict absolute diffusion rates very accurately—perhaps as accurately as is obtainable in high-pressure and high-temperature experiments. However, let it be understood that we are not advocating abandoning experimentation for theory.
In this paper we will use our recent ab initio calculations on the absolute diffusion rates of periclase and perovskite as an example of what can be done and how to do it.
Ab initio vs. empirical potentials
The basis for the techniques described here are atomistic; the system is described as set of interacting atoms or ions, and it is these interactions that govern its behavior. The interactions can either be …