To RMC or not to RMC?: The use of reverse Monte Carlo modelling

The 'classical' approach to structural studies of liquids and glasses, developed over the last 40 years, has been through derivation of the partial pair distribution functions and interpretation of these in terms of peak positions, coordination numbers etc. These parameters are then used to produce a simple picture of the average local structure. An alternative approach, strongly developed in the last decade, is full atomistic modelling of the structure, either by direct modelling of experimental data using techniques such as reverse Monte Carlo modelling (RMC) or empirical potential structure refinement (EPSR), or indirectly through simulation techniques such as Monte Carlo (MC) or molecular dynamics (MD), or more recently ab-initio MD. In this comment we raise some questions that should be considered before a decision is made to use any particular method. We then illustrate, with a particular example, why fully atomistic structural modelling, in this case RMC, can be so powerful even though the apparently limited information available from a diffraction pattern does not seem sufficient to obtain such detailed results. (C) 2003 Elsevier Science Ltd. All rights reserved.