Extraction of Qualitative Photoluminescence and Series Resistance Images from a Single Electroluminescence Measurement

Electroluminescence (EL) imaging is a commonly used characterization tool in solar cell production lines. EL images provide information on performance losses caused by increased recombination and series resistance ( R s ). However, separating recombination and R s ‐related defects presents significant challenges. This study proposes a deep learning approach to extract qualitative photoluminescence (PL, related to recombination) and R s images from a single EL measurement. The developed model has demonstrated high accuracy on unseen simulations. While further fine‐tuning and additional work are needed to ensure robustness for practical use, this approach holds significant potential to streamline solar cell inspection processes in production lines, eliminating the need for expensive additional equipment.