Gaussian-Based Quasiparticle Self-Consistent $GW$ for Periodic Systems

TL;DR

This paper introduces a Gaussian-based quasiparticle self-consistent GW method for periodic systems, demonstrating improved consistency in spectral property predictions across various solids, despite systematic band gap overestimations.

Contribution

The paper develops a full-frequency analytic continuation QSGW implementation for periodic systems using Gaussian basis sets and density fitting, advancing computational methods for solid-state physics.

Findings

QSGW overestimates band gaps but reduces dependence on density functionals.

Provides more consistent spectral predictions than G0W0.

Enables application of QSGW in ab initio quantum embedding for solids.

Abstract

We present a quasiparticle self-consistent $GW$ (QSGW) implementation for periodic systems based on crystalline Gaussian basis sets. Our QSGW approach is based on a full-frequency analytic continuation GW scheme with Brillouin zone sampling and employs the Gaussian density fitting technique. We benchmark our QSGW implementation on a set of weakly-correlated semiconductors and insulators as well as strongly correlated transition metal oxides including MnO, FeO, CoO, and NiO. Band gap, band structure, and density of states are evaluated using finite size corrected QSGW. We find that although QSGW systematically overestimates band gaps of tested semiconductors and transition metal oxides, it completely removes the dependence on the choice of density functionals and provides more consistent prediction of spectral properties than $G_0W_0$ across a wide range of solids. This work paves the way for utilizing QSGW in ab initio quantum embedding for solids.