Machine-Learning Force Fields Reveal Shallow Electronic States on Dynamic Halide Perovskite Surfaces

Abstract

Previous studies indicated that defects in halide perovskites can generate shallow electronic states, which are crucial for their performance in devices. However, how shallow states persist amid pronounced atomic dynamics on halide perovskite surfaces remains unknown. We reveal that electronic states at surfaces of prototypical CsPbBr₃ are energetically distributed at room temperature, akin to well-passivated inorganic semiconductors, despite the presence of undercoordinated atoms and cleaved bonds. Notably, approximately 70% of surface-state energies appear within 0.2 eV of the valence-band edge. Although deep states can still form, they are rarely energetically isolated and are less likely to act as traps. Accelerating first-principles calculations via machine learning, we show that the unique atomic dynamics in halide perovskites render the formation of deep electronic states at their surfaces unlikely. These findings reveal the microscopic mechanism behind the low density of deep states at dynamic halide perovskite surfaces, which is key to their device performance.