Designing Efficient Lead-free KSn 0.5 Ge 0.5 I 3 Perovskite Solar Cells Using DFT, SCAPS-1D, and Machine Learning

We thoroughly investigated a new KSn0.5Ge0.5I3-based perovskite solar cell (PCE) using DFT and the SCAPS-1D simulator. Our DFT investigation examined the atomic arrangement, electronic and optoelectronic properties, and the SCAPS-1D analysis evaluated the photovoltaic (PV) outcomes of different layers in the KSn0.5Ge0.5I3-based device. The orthorhombic structure demonstrates thermodynamic stability, supported by a tolerance factor of 0.86 and a formation energy of –1.94 eV/atom, and we calculated the band structure, obtaining a bandgap of about 1.87 eV. We compared electron and hole transport layers with the perovskite layer and determined the photovoltaic outputs of the Al/FTO/ZnO/KSn0.5Ge0.5I3/V2O5/Ni structure. The optimized device delivered high performance, achieving an efficiency of 22.31%, a Voc of 1.48 V, a Jsc of 16.86 mA/cm2, and an FF of 89.40%. We also analyzed the effects of perovskite layer thickness, carrier concentration, defect, and operating temperature on the photovoltaic behavior. Furthermore, we applied machine learning methods to optimize the performance-prediction algorithm, which attained a low MAE and RMSE along with a high R2value of 0.9997, demonstrating superior accuracy. The study explored material-level variations to guide the development of optimized and cost-effective solar cell prototypes.