Point Defects at Grain Boundaries Can Create Structural Instabilities and Persistent Deep Traps in Metal Halide Perovskites

Nanoscale. 2025. Vol. 17. No. 4. P. 2224–2234.

Wu Y., Liu D., Chu W., Wang B., Vasenko A., Prezhdo O.

Metal halide perovskites (MHPs) have attracted strong interest for a variety of applications due to their low cost and excellent performance, attributed largely to favorable defect properties. MHPs exhibit complex dynamics of charges and ions that are coupled in unusual ways. Focusing on a combination of two common MHP defects, i.e., a grain boundary (GB) and a Pb interstitial, we develop a machine learning model of the interaction potential, and study structural and electronic dynamics on a nanosecond timescale. We demonstrate that point defects at MHP GBs can create new chemical species, such as Pb-Pb-Pb trimers, that are less likely to occur with point defects in bulk. The formed species create structural instabilities in the GB and prevent it from healing towards the pristine structure. Pb-Pb-Pb trimers produce deep trap states that can persist for hundreds of picoseconds, having a strong negative influence on charge carrier mobility and lifetime. Such stable chemical defects at MHP GBs can only be broken by chemical means, e.g., introduction of excess halide, highlighting the importance of proper defect passivation strategies. Long-lived GB structures with both deep and shallow trap states are found, rationalizing the contradictory statements in literature regarding MHP GB influence on performance.

Keywords: машинное обучение, дефекты, defects, перовскиты, Perovskites, Grain boundaries, Machine Learning, межзёренные границы

Publication based on the results of: Квантовые эффекты в низкоразмерных гибридных наноструктурах (2025)