Generating amorphous structures by combining Reverse Monte Carlo and molecular dynamics and analyzing with graph theory algorithm*

We present results from atomistic modeling of amorphous zirconia-doped tantala with an aim to understand the sources of mechanical loss in its structure. This material is a candidate for future LIGO gravitational wave detector mirror coatings, where lower mechanical loss coatings are essential to reduce thermal noise. We use Reverse Monte Carlo (RMC) modeling, which is a common technique used to generate atomic models based on experimental data. Here we use pair distribution functions (PDFs) as the reference experimental data set for the RMC. However, the structural solutions from RMC based on PDFs are not unique, and other constraints are needed to rule out unphysical solutions. In this work, we use molecular dynamics as the additional constraint, and show that allowing the structure to perform ionic and volume relaxation in between RMC runs can generate models that are in good agreement with the PDF data and energetically favorable. We also present an analysis of the intermediate range order from these models. This analysis is achieved by converting amorphous structures into graphs, where atoms are vertices and bonds are edges, and analyzing the structures that are responsible for short and intermediate range segments of the PDF.