Rapid detection and prediction model establishment of propachlor residues in food assisted by machine learning

The detection of the herbicide propachlor in environmental and food samples is critical due to its potential toxic effects. In this study, a sensitive and selective electrochemical sensor based on a molecularly imprinted polymer (MIP), polyethyleneimine-treated multi-walled carbon nanotubes (PEI-MWCNTs), and gold nanoparticles (AuNPs) was developed for propachlor detection. The PEI-MWCNTs enhanced the conductivity of the sensor, while AuNPs promoted electron transfer during the detection process, resulting in a lower detection limit (78 nM) and a wider linear range (0.5 µM to 70 µM) compared to traditional sensors. Furthermore, the sensor exhibited high selectivity against other herbicides, with interference suppression percentages above 92% for common contaminants. The performance of the sensor was further improved by integrating an XGBoost machine learning algorithm for data analysis, which improved the predictive accuracy of propachlor concentration. Reproducibility tests on three independently prepared sensors showed a relative standard deviation (RSD) below 5%. This innovative approach presents a significant advancement in the field of propachlor detection, providing an effective tool for environmental monitoring and food safety.