Researchers from India and Algeria have examined the optimization of perovskite solar cells based on the lead-free double perovskite K2CuCrCl6 by integrating machine learning (ML) models with the SCAPS-1D simulator.
A comprehensive dataset was generated to assess the influence of key parameters, including variations in the electron transport layer (ETL), hole transport layer (HTL), absorber thickness, and the effects of defects, doping, and impurities in the K2CuCrCl6 absorber. An artificial neural network (ANN) model was developed, achieving high predictive accuracy, with statistical analysis revealing a strong correlation between predicted and actual values.
The model’s reliability was confirmed by a Pearson correlation coefficient of 0.950 for power conversion efficiency (PCE) predictions.
SCAPS-1D simulations, guided by ANN-optimized inputs, identified an optimal device configuration consisting of a 100 nm WS2 as ETL, an 800 nm K2CuCrCl6 absorber, and a 500 nm CBTS as HTL, achieving a PCE of 27.99%.
These findings demonstrate the potential of ML-driven approaches to enhance PSC design and optimization, offering a promising pathway toward scalable, high-efficiency photovoltaic technologies with experimental validation.
