Predicting facet properties of Sn-based perovskites using machine learning interatomic potentials and active learning schemes

The surface properties of halide perovskites play a crucial role in determining the performance of perovskite-based solar cells, as they can influence the stability, defect formation, and charge carrier transport properties. In this study, we predict the structural and energetic properties of higher-index facets of Sn-based perovskites. These properties are calculated by combining density functional theory calculations with machine learning interatomic potentials, MLIPs, which are trained on lower-index surfaces. We also implemented active learning schemes to enhance accuracy for out-of-distribution data and improve convergence and surface energy, and diagram predictions.