A lattice kinetic Monte Carlo code for the description of vacancy diffusion and self-organization in Si

We have investigated the mechanism driving void evolution in crystalline Si by means of On lattice' kinetic Monte Carlo (MC) simulations. The implementation of an efficient algorithm allowed us to compare, on a macroscopic time scale, the predictions of simulation codes using three different binding models. We demonstrate that the use of a modified Ising model, taking also into account second neighbour interaction, results in a MC code embodying the main physical features on vacancy (V) interaction, derived by Tight Binding Molecular Dynamics (TBMD) simulations. In particular the dependence of cluster stability and reactivity on size and geometrical configuration can be efficiently taken into account. We have found that, when this model is used, the ripening process is also driven by the migration of small V clusters, and not only by free Vs as predicted by less refined models. This feature strongly modifies the kinetics of growth and the basic mechanism of void ripening.