Self-assembled, ordered polystyrene microsphere-modified electrodes for simultaneous determination of Cu(II), Zn(II), and Pb(II) based on machine learning algorithms

In the process of industrial wastewater discharge, water pollution remains a pressing challenge, with heavy metal pollution being the most critical issue. Electrochemical sensors offer advantages such as rapid response time, ease of integration with microprocessors for on-site testing. This paper proposes the use of self-assembled polystyrene (PS) microspheres as electrode modification materials, leveraging their ordered assembly structure to reduce signal differences between electrodes from different batches, thereby preparing an electrochemical sensor with good repeatability for the determination of heavy metal ions (HMIs) Cu(II), Zn(II), and Pb(II). The experiment employed square wave anodic stripping voltammetry (SWASV) to collect the stripping currents. The designed sensor exhibited sensitivities of 0.905 μA/μM, 0.408 μA/μM, and 0.264 μA/μM, as well as limits of detection (LODs) of 0.13 μM, 0.25 μM, and 0.29 μM (S/N = 3) for Cu(II), Zn(II), and Pb(II), respectively. To enhance detection accuracy, peak current, full stripping current, and feature current were respectively input combined with machine learning (ML) algorithms to construct prediction models to evaluate the performance of the sensor. The results demonstrated that the constructed RF-Feature model significantly improved detection accuracy. It demonstrated excellent reliability, with R 2 values of 0.9952, 0.9973, and 0.9976, and root mean square error (RMSE) values of 0.6580 μM, 0.5838 μM, and 0.8164 μM, for Cu(II), Zn(II), and Pb(II), respectively. The detection range was 1 μM to 70 μM. Furthermore, the effectiveness of the sensor in practical applications was validated by detection in real wastewater samples from industrial and hospital emissions. These results were validated using an inductively coupled plasma mass spectrometry (ICP-MS) method.