Inversion of diffraction data for amorphous materials

Abstract

The general and practical inversion of diffraction data–producing a computer model correctly representing the material explored–is an important unsolved problem for disordered materials. Such modeling should proceed by using our full knowledge base, both from experiment and theory. In this paper, we describe a robust method to jointly exploit the power of ab initio atomistic simulation along with the information carried by diffraction data. The method is applied to two very different systems: amorphous silicon and two compositions of a solid electrolyte memory material silver-doped GeSe3. The technique is easy to implement, is faster and yields results much improved over conventional simulation methods for the materials explored. By direct calculation, we show that the method works for both poor and excellent glass forming materials. It offers a means to add a priori information in first-principles modeling of materials, and represents a significant step toward the computational design of non-crystalline materials using accurate interatomic interactions and experimental information.

Realistic inversion of diffraction data for an amorphous solid: The case of amorphous silicon

“…In this paper, we focus on the classic and persistently vexing problem of amorphous silicon. The details of the methods can be found elsewhere [8,9]. The method is fast enough to make it easy to implement with ab initio interactions (Siesta here) and plane-wave DFT (Vasp) as we used in ternary chalcogenide materials in Ref.…”

“…The problem was that this scheme often failed to converge. We therefore amended Ecmr and introduced ab initio force-enhanced atomic refinement (ab initio FEAR) [9]. In effect we alternate between partially fitting the RDF (or structure factor) using RMC and carrying out partial relaxations using ab initio interactions, as we explain in detail in References 8,9.…”

“…We therefore amended Ecmr and introduced ab initio force-enhanced atomic refinement (ab initio FEAR) [9]. In effect we alternate between partially fitting the RDF (or structure factor) using RMC and carrying out partial relaxations using ab initio interactions, as we explain in detail in References 8,9. By carrying out the iteration in "bite-sized" bits rather than iterated full relaxations as in the original Ecmr, we find that the method is robust, working for silver-doped chalcogenides with plane-wave DFT and for WWW a-Si with Siesta and also for forms of amorphous carbon [28].…”

“…We should clarify that in our previous work on a-Si [9] we used the WWW RDF as input "experimental data", whereas in this work we have used high energy X-ray diffraction data from Laaziri et al [14]. WWW models are a fixture of the modeling community (a continuous random network of ideal four-fold coordination and involving up to 100,000 atoms [13,29], and represent an important benchmark that a new method must handle.…”

“…In our applications of FEAR, we have always started with a random model, and even for a complex ternary [9] the method converges with satisfactory and chemically sensible results. In effect, chemical information is provided through the partial CG relaxations, and the method explores the configuration space rather well, thanks to the excellent RMCProfile code [32].…”