Ligand Microenvironment-Regulated Nanozymes Enabled Machine Learning-Assisted Sensor Array for Simultaneous Identification of Phenolic Pollutants

Phenolic pollutants pose a great threat to human health due to high toxicity, whereas existing methods are difficult to achieve the rapid recognition of multiple phenolic pollutants. In this study, we developed a novel machine learning-assisted sensor array based on ligand microenvironment-regulated Pt nanozymes for the simultaneous differentiation of five phenolic pollutants (phenol, 2,4-DCP, p -chlorophenol, o -chlorophenol, and m -chlorophenol), wherein four cellulose ligands (carboxymethylcellulose, CMC; methylcellulose, MC; hydroxyethyl cellulose, HC; and hydroxypropyl methyl cellulose, HPMC)-regulated Pt nanozymes (Pt@CMC, Pt@MC, Pt@HC, and Pt@HPMC) with considerable laccase-mimicking activity were designed, and the Pt@CMC nanozyme exhibited the highest catalytic activity, which was about 7.5-folds than that of natural laccase. The calculation of density functional theory revealed that Pt@CMC had a stronger ability for capturing 2,4-DCP molecules, showing higher laccase-like activity. More importantly, the different cellulose ligands endowed four Pt nanozymes with laccase-like activity diverse recognition capability to phenolic compounds; thus, a nanozyme sensor array was developed for the differentiation of five phenolic pollutants. Moreover, the integration of a machine learning algorithm and the nanozyme sensor array successfully achieved accurate identification and prediction of the five phenolic pollutants in real water samples. Therefore, this study provided an emerging sensing strategy for the simultaneous identification of phenolic pollutants, carving a promising path for the application of sensor arrays and machine learning algorithms in environmental monitoring.