The G0W0 data set

The data set comprises quasiparticle (QP) energies from 286 G₀W₀ band structures of non-magnetic 2D semiconductors covering 14 different crystal structures and 52 chemical elements. The QP energies have been obtained from plane-wave-based one-shot G₀W₀@PBE calculations with full frequency integration and were produced as a part of the Computational 2D Materials Database (C2DB)^20,21. The data set has been described and analysed in detail in ref. ²⁵.

The QP energies of the data set have been calculated under the standard assumption that the G₀W₀ self-energy can be treated within first-order perturbation theory and linearized around the non-interacting reference energy, ω = ε_nk, leading to the expression

where

is the QP weight and ψ_nk is the PBE wave function with eigenvalues ϵ_nk. In practice, the G₀W₀ correction to the PBE energies, ΔEnkQP=EnkQP−ϵnk, were used as targets for the machine learning model.

To ensure the highest data quality, the original data set was filtered such that only states with QP weights between 0.7 and 1.0 were kept. As shown in ref. ²⁵ the MAE on the QP correction of such states due to the linearization of the QP equation is 0.04 eV.