Multiscale schemes for the predictive description and virtual engineering of materials.

von Lilienfeld-Toal, Otto Anatole

Sandia National Laboratories

关键词: Hydrocarbons;    Chemical Composition;    Engineered Materials-Properties;    Forecasting;    Molten Salts;   

DOI  :  10.2172/991008
RP-ID  :  SAND2010-6251
RP-ID  :  AC04-94AL85000
RP-ID  :  991008

美国|英语

来源: UNT Digital Library

【 摘 要 】

This report documents research carried out by the author throughout his 3-years Truman fellowship. The overarching goal consisted of developing multiscale schemes which permit not only the predictive description but also the computational design of improved materials. Identifying new materials through changes in atomic composition and configuration requires the use of versatile first principles methods, such as density functional theory (DFT). Using DFT, its predictive reliability has been investigated with respect to pseudopotential construction, band-gap, van-der-Waals forces, and nuclear quantum effects. Continuous variation of chemical composition and derivation of accurate energy gradients in compound space has been developed within a DFT framework for free energies of solvation, reaction energetics, and frontier orbital eigenvalues. Similar variations have been leveraged within classical molecular dynamics in order to address thermal properties of molten salt candidates for heat transfer fluids used in solar thermal power facilities. Finally, a combination of DFT and statistical methods has been used to devise quantitative structure property relationships for the rapid prediction of charge mobilities in polyaromatic hydrocarbons.