SpecSwap-RMC: a novel reverse Monte Carlo approach using a discrete set of local configurations and pre-computed properties

We present a novel approach to reverse Monte Carlo (RMC) modeling, SpecSwap-RMC, specifically applicable to structure modeling based on properties that require significant computer time to evaluate. In this approach pre-computed property data from a discrete set of local configurations are used and the configuration space is expressed in this basis. Atomistic moves are replaced with swap moves of contributions to a sample set representing the state of the simulated system. We demonstrate the approach by fitting jointly and separately the EXAFS signal and x-ray absorption spectrum (XAS) of ice Ih using a SpecSwap sample set of 80 configurations from a library of 1382 local structures with associated pre-computed spectra. As an additional demonstration we compare SpecSwap and FEFFIT fits of EXAFS data on crystalline copper, finding excellent agreement. SpecSwap-RMC thus extends RMC structure modeling to any property that can be computed from a structure irrespective of computational expense, but at the cost of a reduced configuration space. The method is general enough that it can be applied to any sets of computed properties, not necessarily limited to structure determination.