Comments on "Efficient Band Gap Prediction for Solids" [Phys. Rev. Lett. 105, 196403 (2010)]

TL;DR

This paper critiques prior claims about LDA's failure in predicting semiconductor band gaps, emphasizing that when solved self-consistently using the BZW method, LDA can accurately describe these gaps.

Contribution

It clarifies that LDA's predictive success depends on self-consistent solutions, challenging previous assertions of its failure in band gap calculations.

Findings

LDA can accurately predict band gaps when solved self-consistently.

Previous failures attributed to LDA are due to incomplete solution methods.

The BZW method improves the predictive capability of LDA for semiconductors.

Abstract

An oversight of several previous local density approximation (LDA) results appears to have led to an incomplete picture of the actual capability of density functional theory (DFT), with emphasis on LDA, to describe and to predict the band gaps of semiconductors [Phys. Rev. Lett. 105, 196403 (2010)]. LDA is portrayed as failing to describe the band gap of semiconductors. In light of the content of the literature, this characterization is misleading. These comments are intended to note some of these previous results and to provide an assessment of LDA capability that is drastically different from that of failure to describe or to predict the band gaps of several semiconductors. This true capability is apparent when the required system of equations of DFT (or LDA) is solved self-consistently as done in the Bagayoko, Zhao, and Williams (BZW) method