Machine Learning-Driven Process Optimization for CdTe PV using Large Dataset for Repeatable PCE Over 18%

This study leverages machine learning (ML) model, such as Random Forest (RF) to optimize the fabrication process for cadmium telluride (CdTe) thin-film photovoltaics. Using historical data from over 8000 devices, RF model established relationships between processing conditions, such as deposition time and temperature, and device performance metrics, including open-circuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and efficiency. This model identified optimal fabrication conditions, achieving repeatable power conversion efficiencies exceeding 18%, with a peak of 20.05% on several hundred cells. This approach highlights the potential of AI-driven methodologies in academic research, paving the way for enhanced efficiency, scalability, and a deeper understanding of device-material interactions in CdTe PV technology.