Machine learning-assisted selective sensing of ornidazole in milk and juice by molecularly imprinted CeFeO3/CHIT nanocomposites

Abstract

Ornidazole (ODZ), as a third-generation nitroimidazole antibiotic, has the potential to cause neurotoxicity and the spread of drug resistance when abused, making the development of highly sensitive and selective detection methods essential. This study develops a molecularly imprinted CeFeO₃/chitosan (CHIT) modified electrode (MIP-CeFeO₃/CHIT/GCE) and utilizes machine learning to predict its electrochemical performance. Through electrochemical detection, the modified electrode shows a good linear response within the range of 0.05–140 nM, with a detection limit as low as 0.0143 nM, significantly outperforming existing methods. Additionally, it exhibits high selectivity, good reproducibility, and stability for ornidazole. In practical electrochemical detection of milk and orange juice samples, the recovery rates range from 98.80 % to 102.83 % and 99.30 %–103.88 %, respectively. Furthermore, by integrating a machine learning model, the modified electrode achieves intelligent and precise electrochemical detection of ornidazole. This work not only provides a high-performance electrochemical sensor for trace detection of ornidazole but also, more importantly, combines machine learning with molecular imprinting technology, offering a new design approach for intelligent modified electrodes in food safety detection. It provides an efficient solution for the rapid monitoring of antibiotic residues in food samples.