Optimizing Inorganic CsCuSbCl/CsTiI Dual-Absorber Solar Cells: SCAPS-1D Simulations and Machine Learning.

Perovskite solar cells (PSCs) have emerged as a promising contender in photovoltaics, owing to their rapidly advancing power conversion efficiencies (PCEs) and compatibility with low-temperature solution processing techniques. Single-junction architectures reveal inherent limitations imposed by the Shockley-Queisser (SQ) limit, motivating adoption of a dual-absorber structure comprising CsCuSbCl (CCSC) and CsTiI (CTI)-lead-free perovskite derivatives valued for environmental benignity and intrinsic stability. Comprehensive theoretical screening of 26 electron/hole transport layer (ETL/HTL) candidates identified SrTiO (STO) and CuSCN as optimal charge transport materials, producing an initial simulated PCE of 16.27%. Subsequent theoretical optimization of key parameters-including bulk and interface defect densities, band gap, layer thickness, and electrode materials-culminated in a simulated PCE of 30.86%. Incorporating quantifiable practical constraints, including radiative recombination, resistance, and FTO reflection, revised simulated efficiency to 26.60%, while qualitative analysis of additional factors follows later. Furthermore, comparing multiple algorithms within this theoretical framework demonstrated eXtreme Gradient Boosting (XGBoost) possesses superior predictive capability, identifying CTI defect density as the dominant impact on PCE-thereby underscoring its critical role in analogous architectures and offering optimization guidance for experimental studies. Collectively, this theoretical research delineates a viable pathway toward developing stable, environmentally sustainable PSCs with high properties.