Numerical Modeling and Machine Learning-Assisted Analysis of Ultra-Thin CuSbS2 Solar Cells Incorporating SnS2 ETL and V2O5 BSF Layers

Abstract

The clean and environmentally friendly characteristics of photovoltaic solar panel research have long been a source of fascination. A promising option for solar absorber material for ultrathin film solar cells is the triple chalcostibite CuSbS2 (Copper Antimony Sulfide) system, It has earth-abundant components, affordable prices, vacuum-free production techniques, and an exceptionally high light absorption coefficient. However, the efficiency of a typical CuSbS2/CdS heterojunction solar panel is extremely low because of the Schottky barrier that forms at the back contact and the significant recombination of carriers at the CdS/CuSbS2 interface. In this article, SnS2 (tin disulfide) is suggested as a substitute for the CdS layer in CuSbS2-based TFSCs (Thin Film Solar Cells). SnS2, CuSbS2, and V2O5 have been used as ETL (Electron Transport Layer), absorber layer, and BSF (Back Surface Field) layer, respectively. A new n-p-p+heterojunction solar cell based on Al/FTO/SnS2/CuSbS2/Ni has been designed and simulated using the SCAPS-1D photovoltaic cell simulator. It has been investigated how integrating the V2O5 BSF layer affects the photovoltaic performances of the CuSbS2- based heterojunction solar cell, about the back-contact recombination of carriers, and the built-in potential. Furthermore, a systematic study has examined the effects of several device characteristics, including operating temperature, shunt and series resistances, back-contact metalwork function, carrier concentration, and layer thickness. The outcomes are examined in connection with the device’s photovoltaic characteristics to maximize the suggested solar cell’s efficiency. At high temperatures, the improved CuSbS2-based solar cell exhibits good performance stability and a maximum efficiency of 31.09%, VOC = 1.39 V, JSC = 25.09 mA/cm2 , FF=88.54%. Additionally, a Random Forest Machine Learning algorithm predicts the optimum PCE (Power Conversion Efficiency) using semiconductor parameters. The model quantifies each parameter’s importance using SHAP (Shapley Additive Explanations) values, revealing their contributions. The model predicts performance accurately and provides precise results with a mean correlation coefficient (R2 ) of 0.84. This study highlights CuSbS2-based potential as a viable material for sustainable solar cells.