Enhancement of sulfide-based absorber and charge transport layer solar cell performance using machine learning and the SCAPS-1D simulator

Abstract

Lead(II) sulfide (PbS)-integrated solar devices are attracting attention from scientists because of their extraordinary semiconducting attributes. However, their full potential has not been achieved due to issues such as incompatibility of the band topologies at the absorber/ETL and HTL/absorber interfaces and recombination of carriers at the front and rear metal contacts. In addition to examining the impacts of the SnS HTL and SnS₂ ETL layers on the performance parameters, the primary focus of this work is to optimize the layer properties of the recently suggested Al/FTO/SnS₂/PbS/SnS/Ni photocell. The SCAPS simulation program was used to conduct this study. Higher performance was achieved by analysing the performance characteristics, including changes in the defect concentration of every stratum, thickness, doping concentration, capacitance (C)-voltage (V), interfacial defects, operating temperature, resistance, and front and back metals. At a thin (900 nm) PbS layer thickness, this device works very well at a lower acceptor density (1 × 10¹⁷ cm^-3). The Al/FTO/SnS₂/PbS/Ni reference cell's PCE, V_OC, J_SC, and FF values were calculated to be 22.96%, 0.99 V, 26.99 mA cm^-2, and 84.08%, respectively. Furthermore, the optimal Al/FTO/SnS₂/PbS/SnS/Ni structure, which introduces SnS between the PbS and Ni, has PCE, V_OC, J_SC, and FF values of 31.43%, 1.12 V, 31.46 mA cm^-2, and 89.10%, respectively. We subsequently developed a machine learning (ML) model to predict the output parameters of the photo devices. Using ML, the performance matrix of the photocells under study was predicted with an accuracy rate of 83.75%. The study sheds light on this important field and provides a workable method for constructing cost-effective PbS-based photovoltaic cells.