SCAPS-1D modeling and machine learning prediction of performance enhancement of Ba₂Br₂ [Cs₂Sn₃Br₁₀] solar cells

This study systematically investigated key device fabrication parameters through numerical simulations using SCAPS-1D and evaluated various combinations of the electron transport layer (ETL) and hole transport layer (HTL) to optimize the performance of Aurivillius type perovskite solar cells (PSCs) based on Ba₂Br₂[Cs₂Sn₃Br₁₀] (BCSBr). The results show a significant improvement in the power conversion efficiency (PCE) of the PSCs, highlighting the crucial role of optimizing device fabrication parameters (including absorber layer (AL) thickness, ETL, and HTL etc.) in promoting efficient charge transport and minimizing recombination losses. The PCE further improves when the AL thickness exceeds 0.5 µm, likely due to enhanced light absorption and carrier generation rates. Furthermore, impedance analysis under different illumination intensities reveals the charge transfer resistance ( R ct ) and recombination losses, while capacitance-frequency (C-f) analysis at different temperatures demonstrates the influence of defect states and carrier dynamics. Under 1.6 sun irradiance, the optimized PCE of the PSC reached to 26.71%, with a fill factor (FF) of 86.57% and an open-circuit voltage ( V oc ) of 1.0065 V. Furthermore, we employed eight different machine learning (ML) models to predict the PCE of the device. The XGBoost model predicted a PCE of 27.95%, higher than previously reported PCEs for the same material. These results demonstrate the great potential of BCSBr-based PSCs as a highly efficient and promising high-performance photovoltaic technology for future applications in the solar energy industry.