Coulomb correlation effects in zinc monochalcogenides

TL;DR

This study investigates the electronic structure of zinc monochalcogenides using density functional theory, highlighting the importance of Coulomb correlations and the effectiveness of the LDA+U method in correcting band gap inaccuracies.

Contribution

It demonstrates the significant role of Coulomb correlations in zinc monochalcogenides and shows that LDA+U improves the accuracy of electronic structure calculations over standard LDA.

Findings

LDA underestimates band gaps and misplaces Zn-3d levels.

Coulomb correlations are crucial for accurately modeling ZnO.

LDA+U correction aligns calculations more closely with experimental data.

Abstract

Electronic structure and band characteristics for zinc monochalcogenides with zinc-blende- and wurtzite-type structures are studied by first-principles density-functional-theory calculations with different approximations. It is shown that the local-density approximation underestimates the band gap and energy splitting between the states at the top of the valence band, misplaces the energy levels of the Zn-3d states, and overestimates the crystal-field-splitting energy. Regardless of the structure type considered, the spin-orbit-coupling energy is found to be overestimated for ZnO and underestimated for ZnS with wurtzite-type structure, and more or less correct for ZnSe and ZnTe with zinc-blende-type structure. The order of the states at the top of the valence band is found to be anomalous for ZnO in both zinc-blende- and wurtzite-type structure, but is normal for the other zinc monochalcogenides considered. It is shown that the Zn-3d electrons and their interference with the O-2p electrons are responsible for the anomalous order. The typical errors in the calculated band gaps and related parameters for ZnO originate from strong Coulomb correlations, which are found to be highly significant for this compound. The LDA+U approach is by and large found to correct the strong correlation of the Zn-3d electrons, and thus to improve the agreement with the experimentally established location of the Zn-3d levels compared with that derived from pure LDA calculations.