Exploring the optoelectronic properties and the machine learning driven impact of the hole transport layer on lead-free Ca3PCl3 perovskite solar cells

Abstract

This study assesses the optoelectronic characteristics of cubic perovskite Ca₃PCl₃ for photovoltaic (PV) applications using first-principles density functional theory (DFT) calculations due to the increasing demand for lead (Pb)-free perovskites in renewable energy. Single-junction PV cells utilizing cubic Ca₃PCl₃ were examined in multiple configurations, employing WS₂ as the electron transport layer (ETL) and integrating three distinct hole transport layers (HTLs): Cu₂O, PEDOT:PSS, and P3HT. Critical determinants affecting PV performance and layer thickness, including band alignment, doping concentrations, defect densities, carrier concentration, interface defects, and recombination-generation processes, were thoroughly examined via SCAPS-1D simulations. The configuration utilizing the P3HT HTL attained a peak efficiency of 24.89%, accompanied by a V_OC of 1.49 V, a J_SC of 18.52 mA cm⁻², and a FF of 89.64%. The Cu₂O, PEDOT:PSS and HTL lacking configuration produced PCEs of 23.83%, 23.79%, and 18.31%, respectively, with V_OC values of 1.43 V, 1.42 V, and 1.35 V; FF values of 88.65%, 88.57%, and 88.17%; and J_SC values of 18.46 mA cm⁻², 18.32 mA cm⁻², and 15.7 mA cm⁻². Additionally, a random forest machine learning algorithm model is employed to predict the optimal PCE, considering various semiconductor properties. By using SHAP (Shapley Additive Explanations) values, the model accurately assesses the significance of each parameter, yielding critical insights into their contributions. The model has a remarkable mean correlation coefficient (R²) of roughly 0.84, indicating outstanding accuracy in forecasting performances and providing dependable and precise outcomes. The findings underscore the promise of Ca₃PCl₃ as an effective absorptive material for eco-friendly PV cells.