Search

Standard density functional approximations often give questionable results for odd-electron radical complexes, with the error typically attributed to self-interaction. In density corrected density functional theory (DC-DFT), certain classes of density functional theory calculations are significantly improved by using densities more accurate than the self-consistent densities. We discuss how to identify such cases, and how DC-DFT applies more generally. To illustrate, we calculate potential energy surfaces of HO· Cl− and HO·H2O complexes using various common approximate functionals, with and without this density correction. Commonly used approximations yield wrongly shaped surfaces and/or incorrect minima when calculated self consistently, while yielding almost identical shapes and minima when density corrected. This improvement is retained even in the presence of implicit solvent. [ABSTRACT FROM AUTHOR]

We introduce a simple semiempirical anharmonic Kirkwood–Keating potential to model AxB1-xC-type semiconductors. The potential consists of the Morse strain energy and Coulomb interaction terms. The optical constants of pure components, AB and BC, were employed to fit the potential parameters such as bond-stretching and -bending force constants, dimensionless anharmonicity parameter, and charges. We applied the potential to finite temperature molecular-dynamics simulations on AlxGa1-xAs for which there is no lattice mismatch. The results were compared with experimental data and those of harmonic Kirkwood–Keating model and of equation-of-motion molecular-dynamics technique. Since the Morse strain potential effectively describes finite temperature damping, we have been able to numerically reproduce experimentally obtained optical properties such as dielectric functions and reflectance. This potential model can be readily generalized for strained alloys. [ABSTRACT FROM AUTHOR]