Machine Learning-Assisted Dual-Metal MOF Luminescence Sensor Array for High-Throughput Analysis of EDCs and Pesticides in Water

This study presents a novel "all-in-one" luminescence sensor array based on a dual-metal organic framework Ga-Tb MOF (GMT) that combines dual luminescence centers to produce six distinct emission channels. The sensor displays unique luminescence fingerprint responses to various endocrine-disrupting chemicals (EDCs) and pesticides through multiple mechanisms, including inner filter effect, antenna effect, photoinduced electron transfer, and competitive coordination. By applying machine learning algorithms, the system achieves high-accuracy identification of 21 pollutants with 93.7% recognition accuracy in real lake water and complete discrimination (100%) of binary mixtures. Quantitative detection using random forest regression (RFR) shows excellent performance (R² > 0.99, LODs as low as 0.07   mg/L, recoveries of 98.8-99.6%) with remarkable anti-interference capabilities in complex matrices. This work offers an efficient and practical approach for the simultaneous detection of multiple pollutant types, highlighting significant potential for environmental monitoring applications.