Machine learning-assisted dual-channel FeZnCu@CNP nanozyme sensor array for qualitative and quantitative analysis of multiple pesticides.

Abstract

To fulfil rapid identification and quantification of various pesticide residues, we first synthesized a FeZnCu@CNP nanozyme with dual laccase- and peroxidase-like activities, enabling differential responses to eight pesticides (Gly, Car, Met, TBZ, CBZ, PCNB, ACE, and MCPA) via coordination and hydrogen-bonding interactions. The dual-channel nanozyme sensor array based on machine learning was developed, successfully discriminating these pesticides at 0.1-100 μg/mL for both qualitative and quantitative detection with an LOD of 0.013 μg/mL. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) confirmed effective differentiation of eight individual pesticides and their mixtures at varied concentrations and ratios. The concentration-independent discrimination model based on the K-nearest neighbors (KNN) algorithm demonstrated outstanding accuracy in pesticide detection and differentiation. Importantly, this model demonstrated strong anti-interference and high feasibility for fruit/vegetable and blind sample identification. This sensor array enables concentration-independent qualitative identification and accurate quantitation of eight residues, offering a supplementary rapid-screening strategy for multi-pesticide detection.