Machine Learning-Integrated Electrochemical Sensors for Accurate and Continuous Free Chlorine Monitoring

Accurate and continuous monitoring of free chlorine concentrations is essential for ensuring water safety in applications such as drinking water disinfection and food sanitation. Traditional methods for free chlorine detection, including colorimetry and photometry, often involve complex sample preparation and lack real-time monitoring capabilities. Electrochemical sensors provide a promising alternative; however, their long-term accuracy is affected by pH variations, electrode surface conditions, and impurity accumulation. In this study, we developed a machine learning-integrated electrochemical sensor using a glassy carbon (GC) electrode to measure current-potential relationships for free chlorine detection. An automated measurement system was constructed to acquire large datasets across varying pH values and free chlorine concentrations, thereby enabling robust model training. The effects of electrode surface conditions were mitigated by integrating voltammogram data obtained from a chlorine-free background solution (base solution) alongside the target voltammogram data into the machine learning model. The trained model was cross-validated and further tested on real samples collected from a vegetable washing factory. The free chlorine concentrations, measured by an iodine photometric sensor, were used as reference values. The calibration system significantly enhanced the estimation accuracy across all test conditions. In real-sample evaluations, the machine learning model successfully estimated free chlorine levels, despite variations in the base solution parameters and the presence of impurities. These results demonstrate the feasibility of integrating machine learning with electrochemical sensing for accurate and continuous monitoring of free chlorine.