Machine Learning-Assisted Pulse Electrodeposition of Copper for Enhanced Nitrate Sensing

Abstract

Overexposure to nitrate, the most stable and prevalent form of dissolved inorganic nitrogen, harms the environment, causing soil acidification, eutrophication, and water contamination. Among various methods for nitrate detection, electrochemical sensors have attracted considerable attention due to their inherent simplicity, high sensitivity, and low cost. However, several challenges remain, including the overpotential for nitrate reduction reaction, which leads to poor selectivity, repeatability and stability. In this work, copper modified electrodes fabricated by pulse electrodeposition method were developed for the selective detection of nitrate. The electrode modification process that determines the sensing performance was investigated by machine learning approaches to understand the relationship between the sensors’ output and the copper deposition parameters. The developed networks successfully predicted the peak current, peak potential, and current stability for electrochemical reduction of nitrate based on the pulse electrodeposition parameters. Furthermore, the most important parameter that influenced the nitrate reduction peak current was revealed by the sensitivity analysis of the designed networks. The experimental results indicate that the proposed sensor achieved a sensitivity of 9.928 μA/mM and a linear range of 0.1 to 20 mM, along with satisfactory recoveries in real sample analysis.