Screening and Many-Body Effects in Two-Dimensional Crystals: Monolayer MoS$_{2}$

TL;DR

This study systematically investigates how computational choices affect the electronic and optical property predictions of monolayer MoS₂ using GW and GW-BSE methods, highlighting challenges in modeling 2D materials.

Contribution

It clarifies the impact of Coulomb interaction treatment and convergence parameters on GW-BSE results for 2D semiconductors, providing guidelines for accurate calculations.

Findings

Variations in GW-BSE results stem from treatment of Coulomb interactions and convergence issues.

Fast spatial screening variations in 2D systems pose computational challenges.

Common divergence removal methods can lead to false convergence in exciton calculations.

Abstract

We present a systematic study of the variables affecting the electronic and optical properties of two-dimensional(2D) crystals within \textit{ab initio} GW and GW plus Bethe Salpeter Equation (GW-BSE) calculations. As a prototypical 2D transition metal dichalcogenide material, we focus our study on monolayer MoS${}_2$. We find that the reported variations in GW-BSE results in the literature for monolayer MoS${}_2$ and related systems arise from different treatments of the long-range Coulomb interaction in supercell calculations and convergence of k-grid sampling and cutoffs for various quantities such as the dielectric screening. In particular, the quasi-2D nature of the system gives rise to fast spatial variations in the screening environment, which are computationally challenging to resolve. We also show that common numerical treatments to remove the divergence in the Coulomb interaction can shift the exciton continuum leading to false convergence with respect to k-point sampling. Our findings apply to GW-BSE calculations on any low-dimensional semiconductors.