Abstracts Category : Other

Modeling and Simulation of Amorphous Materials

by Anup Pandey

The general and practical inversion of diffractiondata - producing a computer modelcorrectly representing thematerial explored -is an important unsolved problem fordisorderedmaterials. Such modeling should proceed by using our full knowledgebase,both from experiment and theory. In this dissertation, weintroduce a robust method,Force-Enhanced Atomic Refinement (FEAR),which jointly exploits the power of ab initioatomistic simulationalong with the information carried by diffraction data. Asapreliminary trial, the method has been implemented using empiricalpotentials foramorphous silicon (a-Si) and silica ( SiO_2 ). Themodels obtained are comparable to theones prepared by theconventional approaches as well as the experiments. Using abinitiointeractions, the method is applied to two very differentsystems: amorphous silicon (a-Si)and two compositions of a solidelectrolyte memory material silver-doped GeSe_3 . It isshown thatthe method works well for both the materials. Besides that, thetechnique iseasy to implement, is faster and yields results muchimproved over conventionalsimulation methods for the materialsexplored. It offers a means to add a prioriinformation in firstprinciples modeling of materials, and represents a significantsteptoward the computational design of non-crystalline materialsusing accurate interatomicinteractions and experimentalinformation. Moreover, the method has also been used tocreate acomputer model of a-Si, using highly precise X-ray diffractiondata. The modelpredicts properties that are close to the continuousrandom network models but with no apriori assumptions.In addition,using the ab initio molecular dynamics simulations (AIMD) weexploredthe doping and transport in hydrogenated amorphous silicona-Si:H with the most popular4impurities: boron and phosphorous. Weinvestigated doping for these impurities and therole of H in thedoping process. We revealed the network motion and H hoppinginducedby the thermal fluctuations significantly impacts conductionin this material. In the lastsection of the dissertation, weemployed AIMD to model the structure of amorphous zincoxide (a-ZnO)and trivalent elements (Al, Ga and In) doped a-ZnO. We studiedthestructure and electronic structure of these models as well asthe effect of trivalent dopantsin both the structure and electronicstructure of a-ZnO.Advisors/Committee Members: Drabold, David A. (Advisor).