Screening High-Performance Hybrid Halides Scintillators: A Comprehensive Analysis and Prediction Model

Machine learning models were applied to predict the scintillation performances of organic–inorganic hybrid metal halides (OIMHs), focusing on their photoluminescent quantum yield (PLQY). Random Forest and Decision Tree algorithms identified the most critical structural parameter of organic molecules influencing the M···M distance between metal ions and correlated PLQY value, with an optimal distance of approximately 8 Å correlating with enhanced luminescence efficiency. This prediction was experimentally validated through the synthesis of several OIMH compounds, demonstrating strong agreement between predicted and measured PLQY values. The machine learning approach not only enabled the screening of efficient compounds but also deepened the understanding of how structural factors, such as the structure of organic molecules, govern scintillation properties. These findings underscore the potential of machine learning in accelerating the development of next-generation luminescent materials with improved performance, offering a powerful tool for future material design and optimization.