Machine learning-assisted construction of a lignin carbon dots sensor array for detecting food colorants

Food safety monitoring is crucial due to the widespread use and potential toxicity of synthetic food colorants. High-sensitivity techniques such as chromatography are routinely employed but require costly equipment and skilled operators. Here we show that a three-channel fluorescence sensor array based on amino acid-derived, nitrogen-doped lignin carbon dots (CDs) enables accurate and interference-resistant detection of food colorants. To elucidate the role of specific functional groups in fluorescence response, a stepwise chemical modification strategy was employed to systematically investigate the contribution of surface chemistry to food colorants recognition. This array distinguished four common colorants (100-300 μM) with nearly 100 % accuracy in blind tests using hierarchical cluster analysis (HCA), linear discriminant analysis (LDA), and principal component analysis (PCA), surpassing conventional single-channel systems. These results demonstrate the promise of lignin-based CDs as green, cost-effective sensing elements and introduce a robust multi-channel fluorescence strategy for practical on-site monitoring of food colorants.