Machine Learning-Assisted Laccase-like Pt@GA-Cu Nanozyme-Based Colorimetric Method for Discrimination and Detection of Phenolic Compounds in Water

Abstract

Phenolic compounds represent a class of hazardous environmental pollutants characterized by high toxicity, low biodegradability, and a strong tendency to bioaccumulate, leading to considerable risks for ecological safety and human health. In this work, a novel bimetallic nanozyme Pt@GA-Cu was synthesized by loading platinum nanoparticles (Pt NPs) onto a copperbased laccaselike material. The synergistic interaction between Pt and Cu significantly enhanced the electron transfer rate and catalytic efficiency, resulting in better laccase‐like performance than that of the individual materials. By this enhanced performance, a timedependent threechannel colorimetric sensor array was constructed. With the integration of machine learning (ML), specifically linear discriminant analysis (LDA) and hierarchical cluster analysis (HCA), the as‐developed sensor successfully not only distinguished four distinct phenolic compounds (2,4-DP, 4cP, CAT, and 2,6DMP) with high accuracy, but also exhibited exceptional discriminative capabilities even when deployed within intricate multi-component phenolic solutions. Furthermore, to evaluate the functionality of the sensor in practical condition, real water samples were tested using standard addition methods, and satisfactory recovery rates (93.15-112.5%) and low relative standard deviations (<2.2%) were obtained. This study provides an efficient and reliable and strategy to rapidly detect and differentiate among various phenolic pollutants, offering considerable potential for environmental monitoring applications.

Keywords: Laccase-like nanozymes, Phenolic compounds, colorimetric sensor array, machine learning, Pt@GA-Cu