Strong renormalization of the electronic band gap due to lattice polarization in the GW formalism

TL;DR

This paper introduces a generalized GW approach that includes lattice polarization effects, significantly improving the accuracy of band gap predictions in polar semiconductors and insulators.

Contribution

It derives a new GW formalism incorporating lattice polarization, addressing overestimations in band gaps caused by neglecting lattice contributions.

Findings

Lattice polarization reduces GW band gaps by over 15% in polar materials.

Including lattice effects decreases the band gap error by a factor of three.

The model accurately predicts band structures for various binary semiconductors and insulators.

Abstract

The self-consistent GW band gaps are known to be significantly overestimated. We show that this overestimation is, to a large extent, due to the neglect of the contribution of the lattice polarization to the screening of the electron-electron interaction. To solve this problem, we derive within the GW formalism a generalized plasmon-pole model that accounts for lattice polarization. The resulting GW self-energy is used to calculate the band structures of a set of binary semiconductors and insulators. The lattice contribution always decreases the band gap. The shrinkage increases with the size of the longitudinal-transverse optical splitting and it can represent more than 15% of the band gap in highly polar compounds, reducing the band-gap percentage error by a factor of three.