In-Line Electrochemical Impedance Spectroscopy for Real-Time Circuit Identification and Interpretation in Metal Halide Perovskite Single Crystals

Electrochemical impedance spectroscopy (EIS) has played a crucial role in understanding electronic and ionic responses in MHPs through frequency and time domain analyses. This study introduces a novel approach by integrating supervised machine learning to develop in-line automation for classifying and fitting the EIS spectra for metal halide perovskite single crystals (MHPSCs). By employing widely used EIS circuits for MHPSCs, we establish a chronological link and offer a comprehensive physical interpretation. Our model utilizes simulated data to accurately classify experimental EIS spectra and provides precise fitting values of electronic components, enhancing the spectroscopic analysis of MHPSCs. The Decision Tree classifier demonstrates superior accuracy in classifying EIS spectra compared to Random Forest and XGBoost models. Incorporating 3-fold cross-validation confirms its reliability with notable versatility in capturing spectral changes under varying bias conditions. Our study establishes a chronological link between the widely used EIS circuits for MHPSCs and offers comprehensive physical interpretations.