Interfacial Assembly of Metal-Phenolic Network-Confined AIEgen Nanotags: Enhancing Emission Performance for Machine Learning-Assisted Hue-Decodable Ratiometric Fluorescent Immunochromatographic Detection

Abstract Monitoring carcinogenic pentachlorophenol (PCP) is vital but limited by the trade-off between analytical accuracy and operational simplicity. Herein, we based on theoretical calculations to develop a green-emitting carrier (ETTPA@YTP) by functionalizing aggregation-induced emission luminogens (AIEgens, ETTPA) with a metal-phenolic network (tannic acid/Y3+) through interfacial assembly. This innovative approach resolves the common issues of colloidal instability that limit the applications of AIEgens in the bioanalytical field. Our results are both in theoretical and experimental evidence, indicating that MPN-mediated dense packing significantly enhances emission efficiency and photostability. We further integrated the ETTPA@YTP-antibodies (Abs) probe with orange-red quantum dot microspheres to create a traffic-light-like hue-decodable ratiometric fluorescent immunochromatographic assay (RFICA). This assay displayed a concentration-dependent green-to-orange-red transition, providing clear visual indicators of PCP presence. Benefiting from RGB-based self-calibration, the RFICA offered 1.54-fold and 4-fold improvements in detection limit and cut-off value, respectively, compared to conventional intensity-based ETTPA@YTP-FICA. To reduce visual errors, we introduced a machine learning platform for objective classification and quantification. This method demonstrated excellent recoveries (80.39%−123.04%) and high versatility for diverse analytes. This work successfully bridges the gap between qualitative visual screening and precise laboratory-grade quantification, offering a robust solution for environmental and food safety monitoring.