Computer simulation as a tool for the interpretation of total scattering data from glasses and liquids

Abstract

Ever since the pioneering work by McGreevy and Pusztai [Mol. Simul. 1 (1988), pp. 359–367] in developing the reverse Monte Carlo (RMC) technique, computer simulation has become a widely used tool for the interpretation of total scattering data from liquid, glassy and crystalline materials. A variant on this approach, empirical potential structure refinement (EPSR), was introduced in 1996 for the study of molecular materials. Both approaches perform routine Monte Carlo simulations of the systems under study but derive their structural information partly from the assumed knowledge of the system in question (such as density, limits on atomic overlap, likely forces between atoms and molecular structure) and partly from radiation total scattering data, which contain weighted sums of the relevant site–site radial distribution functions (rdfs) via the corresponding partial structure factors. The latter information is carried in the form of a fit factor (RMC) or empirical potential (EPSR) which is included in the Hamiltonian used to assess whether individual atom or molecule moves should be accepted or rejected. The present account illustrates the use of EPSR to investigate two well-known liquids, water and dimethylsulphoxide. In the former case, it is possible, at least in principle, to recover the set of three site–site partial structure factors directly from the available scattering data. In the latter case, this is not possible, even in principle, because there are insufficient isotope or X-ray contrasts available to recover the full set of 10 partial structure factors needed to define the structure of the liquid. Although the information content of the data-sets for each liquid is different, EPSR is able to distinguish the intermolecular force field between models in both cases, and thus produce a set of site–site rdfs which are both consistent with the scattering data and derived from a non-overlapping 3D distribution of molecules.