Optoelectronic Coupling Modeling Guided by Machine Learning for Targeted Material Selection in All‐Perovskite Tandem Solar Cells

Perovskite‐based solar cells are widely recognized as one of the most promising next‐generation photovoltaic technologies. However, the selection and optimization of charge transport layer materials remain challenging, particularly for all‐perovskite tandem solar cells (TSCs), which involve more complex multi‐interface coupling and interactions. In this work, we propose a rigorous optoelectronic coupling model that integrates machine learning techniques to identify key transport layer parameters affecting the performance of all‐perovskite TSCs, thereby guiding material selection and device optimization strategies. Numerical simulations reveal that both wide‐bandgap and narrow‐bandgap perovskite subcells achieve high performance and exhibit strong tolerance to interface defects when a well‐matched energy band alignment is realized using wide‐bandgap hole and electron transport layers (HTL/ETL). SHAP analysis further indicates that the valence/conduction band offset is the key efficiency‐determining factor, with an almost aligned valence band at the perovskite/HTL interface being optimal for hole extraction. Strategic doping can also modulate the interfacial electric field to mitigate the negative effects of suboptimal band alignment. By optimizing the optical and electrical properties of ETL and HTL materials and incorporating an interfacial buffer layer, we propose a feasible pathway toward achieving 33% efficiency in both n ‐ i ‐ p and p ‐ i ‐ n structured all‐perovskite TSCs.