Machine Learning-Based Optimization and Performance Enhancement of CH3NH3SnBr3 Perovskite Solar Cells with Different Charge Transport Materials Using SCAPS-1D and wxAMPS

Abstract

Recent research focuses on enhancing the sustainability of perovskite solar cells (PSCs) by substituting lead with non-toxic materials, identifying tin-based perovskites such as CH₃NH₃SnBr₃ as a viable alternative. This study examines the efficacy of CH₃NH₃SnBr₃ as the absorber layer in conjunction with V₂O₅ as the hole transport layer (HTL) and several electron transport layers (ETLs), including C₆₀, IGZO, WS₂, and ZnSe. The study employs SCAPS-1D simulations to optimize parameters including doping concentration, thickness, and defect density, aiming to improve photovoltaic efficiency. The optimal configuration (FTO/WS₂/CH₃NH₃SnBr₃/V₂O₅/Au) attained a power conversion efficiency (PCE) of 33.54%, surpassing alternative ETL combinations. The results of the SCAPS-1D simulation are analyzed in comparison to those of the wxAMPS simulation. The machine learning model is developed to predict solar cell performance, achieving an accuracy of 82%. The findings underscore the significance of choosing appropriate ETL to enhance PSC efficiency and sustainability.