Publication date: 15th December 2025
Interpreting the impedance response of perovskite solar cells (PSCs) is challenging due to coupled ionic and electronic motion, which creates complex, overlapping signals. While drift-diffusion (DD) modelling can extract meaningful physical parameters, its complexity makes direct parameter estimation from experimental data impractical. This work overcomes this limitation by using DD simulations to generate a large synthetic dataset of impedance spectra, which is used to train machine learning (ML) models. For a standard TiO₂/MAPI/spiro cell, a Gradient Boosting Regressor was the most effective at predicting key recombination and ionic parameters. Interpretative analysis revealed that open-circuit measurements best probe recombination losses, while short-circuit conditions are optimal for extracting ionic properties like concentration and mobility. The trained models successfully analyzed experimental data, predicting ion concentrations of (1.3-3.3) × 10¹⁷ cm⁻³ and mobilities of (5-7) × 10⁻¹¹ cm²V⁻¹s⁻¹. This approach demonstrates a viable pathway to accurately derive efficiency-determining physical parameters from impedance measurements in PSCs.
