The current research investigates a novel double perovskite halide absorber K2NiCl6 regarding its structural stability as well as electronic and optical properties using Density Functional Theory (DFT) calculations. According to the band structure, the direct band gap of K2NiCl6 is 1.179 eV. Using SCAPS-1D simulator, the emphasized configuration, ITO/IGZO/K2NiCl6/MoTe2/Pt, is also investigated. However, a simulated efficiency of up to 30.11% can be attained by using the hole transport layer (HTL) and electron transport layer (ETL) combination of MoTe2 and IGZO, respectively, and optimized device parameters. To further accelerate device optimization, machine learning models such as Support Vector Machine, Neural Network, CatBoost, GradientBoost, KNN, Random Forest, and XGBoost were trained on 2304 SCAPS-1D simulation results, were obtained by altering the thickness of absorber, doping, defect density, resistivity, and temperature for accurately predict solar cell performance. The Cat Boost model outperforms all other approaches, having a brilliant Coefficient of Determination (R2) of 99.99% and Mean Squared Error (MSE) of 0.0054. To enhance interpretability, SHAP (Shapley Additive exPlanations) analysis were utilized to examine the influence of key parameters on device efficiency. This comprehensive framework, integrating first-principles calculations, numerical simulations, and machine learning, provides valuable insights into the development of stable, high-efficiency, lead-free perovskite solar cells. The outcomes of this contribution may open up useful research directions for making K2NiCl6 as an environmentally friendly absorber material, advancing its prospects for next-generation optoelectronic applications in photovoltaic cells.
Imran et al. (Fri,) studied this question.