First-principles study of excitonic effects in Raman intensities

TL;DR

This study uses first-principles calculations to analyze excitonic effects on Raman intensities in silicon, revealing significant enhancements and improved agreement with experiments when these effects are included.

Contribution

It introduces a finite-difference scheme combined with Bethe-Salpeter equation solutions to accurately include excitonic effects in Raman intensity predictions.

Findings

Excitonic effects cause a tenfold increase in Raman intensity in silicon.

The proposed double-grid technique reduces computational effort significantly.

Results align better with experimental data than previous models.

Abstract

The ab initio prediction of Raman intensities for bulk solids usually relies on the hypothesis that the frequency of the incident laser light is much smaller than the band gap. However, when the photon frequency is a sizeable fraction of the energy gap, or higher, resonance effects appear. In the case of silicon, when excitonic effects are neglected, the response of the solid to light increases by nearly three orders of magnitude in the range of frequencies between the static limit and the gap. When excitonic effects are taken into account, an additional tenfold increase in the intensity is observed. We include these effects using a finite-difference scheme applied on the dielectric function obtained by solving the Bethe-Salpeter equation. Our results for the Raman susceptibility of silicon show stronger agreement with experimental data compared with previous theoretical studies. For the sampling of the Brillouin zone, a double-grid technique is proposed, resulting in a significant reduction in computational effort.