Prediction and Fine Screening of Small Molecular Passivation Materials for High‐Efficiency Perovskite Solar Cells Via an Enhanced Machine Learning Workflow

Abstract Interface engineering is pivotal for enhancing the efficiency and stability of perovskite solar cells (PSCs), yet traditional experimental approaches for identifying effective passivation materials are labor‐intensive and time‐consuming. Leveraging the power of machine learning (ML), a robust and interpretable workflow is presented for the fine screening of small molecular passivation materials. By integrating multi‐level feature engineering, including structural, physical, and electronic properties, employing advanced ML models, benzodithiophene terthiophene rhodanine derivative (BTR‐Cl) is identified as a highly effective passivator for the perovskite/hole transport layer (HTL) interface. The fine‐screening capability of the ML workflow enables precise prediction of BTR‐Cl's superior performance. Experimental validation shows that BTR‐Cl optimizes energy level alignment, reduces surface defects, and significantly suppresses non‐radiative recombination, leading to a champion power conversion efficiency (PCE) of 25.36% with an open‐circuit voltage (V OC ) of 1.186 V. Furthermore, BTR‐Cl effectively inhibits halogen ion migration, enhancing device stability. This study highlights the transformative potential of ML in fine screening and accelerating the discovery of advanced materials for high‐performance PSCs.