First example of multi-scale reverse Monte Carlo modeling for small-angle scatteringexperimental data using reverse mapping from coarse-grained particles to atoms

Abstract

We propose a new method for multi-scale reverse Monte Carlo (RMC) modeling ofsmall-angle scattering data using reverse mapping from coarse-grained particles to atoms incases where scale separation cannot be assumed. For efficient RMC analysis for small-anglescattering data, it is important to determine a large scale structure with the lowest possiblecomputing cost. In order to find this large scale structure, a method using coarse-grainedparticles instead of atoms is suitable. As our first example, we examine the structure ofexpanded fluid Hg near the critical point. For this system, small-angle x-ray scattering(SAXS) data and wide-angle x-ray diffraction data (XRD) are observed in the samethermodynamic state. First, RMC analysis using coarse-grained particles for SAXS data isperformed. Second, RMC analysis for SAXS and XRD data is performed withthe replacement of a coarse-grained particle by an ad hoc cluster of several Hgatoms. In the present study, we have determined that the size of one coarse-grainedparticle corresponds to ten Hg atoms. The number density for the coarse-grainedparticles is set to one-tenth the actual number density of atoms and the cutofflength is three times (6.9 Å) that of Hg atoms (2.3 Å). As a result, this approach isfound to be successful and the computing cost of RMC analysis can be reduced.