摘要
Simultaneous and accurate analysis of multiple heavy metals (HMs) in environment by electrochemical method still faces huge challenges, such as complex experimental process, low detection efficiency and the interference arising from electrochemical interactions between different HMs. To address these issues, we propose a sensing platform combining S-TiO 2 @Ti 3 C 2 MXene with machine learning strategies for simultaneous electrochemical detection of Cu 2+ and Hg 2+ in water samples. Specifically, surface morphology and crystal structure of S-TiO 2 @Ti 3 C 2 were characterized via SEM, XRD, and XPS. Electrochemical behaviors were systematically investigated using cyclic voltammetry (CV) and differential pulse voltammetry (DPV). Orthogonal test-based multi-objective optimization method integrated with comprehensive balance approach were employed to determine the optimal sensing parameters, which were identified pH of 4.5, deposition time of 210 s, and deposition potential of −1.5 V. The fabricated sensor exhibited high sensitivities of 44.122 and 143.795 μA·L·mg −1 , low limits of detection of 0.392 and 0.093 μg·L −1 , and broad linear ranges of 0.001–1000 μg·L −1 for Cu 2+ and Hg 2+ , respectively. Furthermore, machine learning strategy was introduced to resolve the poor accuracy issue associated with traditional calibration curve analysis for the simultaneous detection of Cu 2+ and Hg 2+ across the entire linear range. The results indicated that minimum Redundancy Maximum Relevance (mRMR) algorithm effectively extracted valuable and critical information of Cu 2+ and Hg 2+ from the full voltammetric curves, and the constructed support vector machine (SVM) model remarkably improved detection accuracy. The reliability of the proposed SVM model was validated through simultaneous detection of HMs in actual environmental water samples, yielding favorable recovery rates.
危险品化学品经营许可证（带存储）