The use of advanced Monte Carlo techniques can greatly improve sampling of
the spatial distribution and partitioning of impurities in
multi-component, multi-phase systems that occur on time scales
inaccessible to conventional molecular dynamics simulations. We will carry
out constant pressure simulations of the Gibbs ensemble or expanded Gibbs
ensemble followed by Gibbs-Duhem integration along coexistence lines to
cover a wider range of state points. The Gibbs ensemble allows one to
study directly phase equilibria of systems with arbitrary number of
components and phases by using multiple simulation boxes that are
thermodynamically connected but do not share a physical interface.
Configurational-bias Monte Carlo techniques will be used to enhance phase
transfers of the alloying elements and, in particular, identity exchanges
between two different alloying elements, e.g. oxygen with sulfur.
Aggregation-volume-bias techniques will be exploited to direct the
transfers to specific sites, e.g. interstitial or vacancy sites. These
techniques have been successfully used to investigate solid-liquid-vapor
phase boundaries, to determine precise partition constants of multiple
analytes in chromatographic systems, and to calculate the vapor-liquid
coexistence curve of water using first-principles methods (Siepmann,
Gillan, and Price).
