Machine learning-assisted sensor array based on CeO2 nanozymes for intelligent recognition of multiple synthetic phenolic antioxidants.

BACKGROUND: Synthetic phenolic antioxidants are commonly utilized in foods, rubbers and plastics to slow down the oxidation process, but excessive intake of these compounds may potentially harm human health. RESULTS: Hence, we developed a novel nanozyme sensor array utilizing three CeO2 nanozymes with excellent peroxidase-like activity for rapid discrimination among five synthetic phenolic antioxidants, including propyl gallate (PG), butylated hydroxyanisole (BHA), tertiary butylhydroquinone (TBHQ), butylated hydroxytoluene (BHT), and 3,5-Di-tert-butyl-4-hydroxybenzoic acid (BHT-COOH). Wherein, three CeO2 nanozymes with different morphologies were synthesized, which exhibited excellent peroxidase-like activity. Interestingly, the synthetic phenolic antioxidants inhibited the expression of CeO2 nanozymes activity to varying degrees due to their different chemical structures. Inspired by this, we constructed a CeO2 nanozyme sensor array for precise discrimination of five synthetic phenolic antioxidants through pattern recognitions. Furthermore, the sensor array was integrated with machine learning to develop a dual-stage prediction model featuring classifier-independent concentration discrimination and regression-based quantification. Besides, the identification and concentration discrimination of different synthetic antioxidants in actual samples have been achieved with the support of model, demonstrating good practical applicability. SIGNIFICANCE: Therefore, our findings provide a new method for the identification of synthetic phenolic oxidants, providing better oversight of the overuse of synthetic antioxidants.