Implementation of an all-electron GW Approximation using the Projector Augmented Wave method: II. Application to the optical properties of semiconductors

TL;DR

This paper applies an all-electron GW approximation with the PAW method to study optical properties of various semiconductors, analyzing local-field and quasiparticle corrections and their impact on dielectric functions and spectra.

Contribution

It demonstrates the application of an all-electron GW method to compute optical properties of semiconductors, highlighting the effects of local-field and quasiparticle corrections on spectra and dielectric constants.

Findings

Quasiparticle corrections improve spectral alignment with experiments.

Local-field effects reduce peak intensities in optical spectra.

Static dielectric constants are significantly improved by local-field corrections.

Abstract

We used our previously implemented GW approximation (GWA) based on the all-electron full-potential projector augmented wave (PAW) method to study the optical properties of small, medium and large-band-gap semiconductors: Si, GaAs, AlAs, InP, Mg$_2$Si, C, and LiCl. The aim being to study the size of both local-field (LF) and the quasi-particle (QP) corrections to the calculated dielectric function obtained using the local density approximation (LDA). We found that while the QP corrections tend to align the calculated structures in the optical spectra with their experimental counterparts, the LF effects don't change these peak positions but systematically reduce the intensities of the so called $E_1$ and $E_2$ structures in all the optical spectra. The reduction of the intensity of the $E_1$ peak worsen the agreement with experiment while that of $E_2$ improves it. We then show that the local-field correction improves considerably the calculated static dielectric constants of all studied semiconductors. Because the static dielectric constant is a ground state property, the remaining discrepancy with experiment should be attributed to the the LDA itself. On the other hand, as expected, the calculation of the static dielectric constant using the GW quasiparticle energies and including the LF effects is underestimated for all the semiconductors. The excitonic effects should then correct for this discrepancy with experiment.