Colorimetric - Fluorescence - Photothermal tri-mode sensor array combining the machine learning method for the selective identification of sulfonylurea pesticides.

Though cholinesterase-based method could detect two types of pesticides (organophosphorus and carbamate), they had weak sensing on sulfonylurea pesticides. In our previous work, the peroxidase-like reaction system of nanozyme - HO - TMB showed selective detection of sulfonylurea pesticides, but the single-signal output sensing platform was easily affected by complex matrix background, cross-contamination and human error. Therefore, this work used colorimetric, photothermal, and fluorescent signals of the nanozyme reaction as sensing units for the detection of pesticides. This is the first time that photothermal signals have been used to construct a sensor array. When the concentration of interfering substances was 25 times that of pesticides, the method was still unaffected and had excellent selectivity and anti-interference performance. Meanwhile, a concentration-independent differentiation mode was established based on the K-nearest neighbor (KNN) algorithm. The pesticides were detected and distinguished with 100% accuracy. This work contributed to the detection of sulfonylurea pesticides in complex environmental/food matrices, bridging the gap of existing pesticide detection methods and providing an effective method for food safety detection.