Real space information from Fluctuation electron microscopy: Applications to amorphous silicon

TL;DR

This paper introduces a method combining experimental data and first-principles calculations to model medium-range order in amorphous silicon, revealing that inhomogeneities like voids and ordered grains explain FEM observations.

Contribution

The paper develops Experimentally Constrained Molecular Relaxation (ECMR) to integrate all available information for modeling complex materials, applied here to amorphous silicon.

Findings

FEM data can be explained by a continuous random network with inhomogeneities.

A new void-based model for medium-range order in amorphous silicon is proposed.

ECMR effectively combines experimental and theoretical data for material modeling.

Abstract

Ideal models of complex materials must satisfy all available information about the system. Generally, this information consists of experimental data, information implicit to sophisticated interatomic interactions and potentially other {\it a priori} information. By jointly imposing first-principles or tight-binding information in conjunction with experimental data, we have developed a method: Experimentally Constrained Molecular Relaxation (ECMR) that uses {\it all} of the information available. We apply the method to model medium range order in amorphous silicon using Fluctuation Electron microscopy (FEM) data as experimental information. The paracrystalline model of medium range order is examined, and a new model based on voids in amorphous silicon is proposed. Our work suggests that films of amorphous silicon showing medium range order (in FEM experiments) can be accurately represented by a continuous random network model with inhomogeneities consisting of ordered grains and voids dispersed in the network.