Structural and electronic properties of monolayers: Enhancing computational accuracy

Abstract

Matching theoretical and experimental lattice constants is essential for accurately simulating real-life materials. DFT optimizations often yield lattice constants that deviate from experimental values, especially in 2D materials. For instance, databases like C2DB rely on these theoretical values, which might not always reflect experimental data. Furthermore, PBE-based calculations can misclassify the band gap type while also underestimating its magnitude. A notable example is hexagonal boron nitride, where PBE predicts a direct band gap, although experimentally, it is indirect. This work aims to identify functionals that produce lattice constants and electronic properties more aligned with experimental results for monolayers, thereby enhancing the accuracy of machine learning models in materials discovery.