The optimized effective potential and the self-interaction correction in density functional theory: Application to molecules

Citation
J. Garza et al., The optimized effective potential and the self-interaction correction in density functional theory: Application to molecules, J CHEM PHYS, 112(18), 2000, pp. 7880-7890
Citations number
78
Categorie Soggetti
Physical Chemistry/Chemical Physics
Journal title
JOURNAL OF CHEMICAL PHYSICS
ISSN journal
00219606 → ACNP
Volume
112
Issue
18
Year of publication
2000
Pages
7880 - 7890
Database
ISI
SICI code
0021-9606(20000508)112:18<7880:TOEPAT>2.0.ZU;2-U
Abstract
The Krieger, Li, and Iafrate approximation to the optimized effective poten tial including the self-interaction correction for density functional theor y has been implemented in a molecular code, NWChem, that uses Gaussian func tions to represent the Kohn and Sham spin-orbitals. The differences between the implementation of the self-interaction correction in codes where plane waves are used with an optimized effective potential are discussed. The imp ortance of the localization of the spin-orbitals to maximize the exchange-c orrelation of the self-interaction correction is discussed. We carried out exchange-only calculations to compare the results obtained with these appro ximations, and those obtained with the local spin density approximation, th e generalized gradient approximation and Hartree-Fock theory. Interesting r esults for the energy difference (GAP) between the highest occupied molecul ar orbital, HOMO, and the lowest unoccupied molecular orbital, LUMO, (spin- orbital energies of closed shell atoms and molecules) using the optimized e ffective potential and the self-interaction correction have been obtained. The effect of the diffuse character of the basis set on the HOMO and LUMO e igenvalues at the various levels is discussed. Total energies obtained with the optimized effective potential and the self-interaction correction show that the exchange energy with these approximations is overestimated and th is will be an important topic for future work. (C) 2000 American Institute of Physics. [S0021-9606(00)30114-3].