Intelligent nanozyme-based colorimetric sensor array for identifying six major teas and quantifying adulteration.

The rapid discrimination of tea varieties and detection of adulteration are critical for food safety and quality control. In this study, a three-channel colorimetric sensor array was constructed using MIL100(Fe), Au nanoclusters (Au NCs), and MIL100(Fe)/Au NCs to distinguish tea polyphenols. The sensing mechanism is based on the differential inhibition of peroxidase-like activity arising from the interaction between structurally diverse tea polyphenols and the active sites, which enables the establishment of specific characteristic fingerprint patterns for each analyte. The experimental results demonstrated that, in the current dataset, the colorimetric sensor array reached a maximum internal cross-validation accuracy of 100% for five tea polyphenols and exhibited excellent clustering across a dynamic range of 10 nM to 10 mM. Through a systematic evaluation of 8 machine learning algorithms and the optimization of the K-Nearest Neighbors (KNN) model by combining data fusion strategies, the highest internal cross-validation accuracy for classifying six major teas reached 100%. Furthermore, a linear quantitative identification model for Longjing tea adulteration (ranging from 0% to 100%) was established, along with nanomolar-level ultrasensitive detection of specific tea polyphenols. This study confirms that machine learning-assisted nanozyme sensor arrays has significant application potential in the rapid assessment of tea quality and market supervision.