Machine Learning-Guided Cobalt@Copper Dual-Metal Electrochemical Sensor for Urinary Creatinine Detection

Abstract

By utilizing the synergistic effects of a dual-metal cobalt@copper electrode and advanced machine learning algorithms, we have developed a reliable and cost-effective electrochemical sensor for creatinine monitoring. The sensor’s active surface was fabricated through the sequential electrodeposition of copper and cobalt nanoparticles, with their complexation with creatinine confirmed via cyclic voltammetry and spectroelectrochemical analyses. The combined contributions of both transition metals significantly enhanced the sensor’s sensitivity and selectivity, yielding a linear detection range of 0.00-4.00 mM, a sensitivity of 6.06 ± 0.65 μA mM-1, and a limit of detection of 0.13 mM. The sensor demonstrated excellent selectivity against common interferences, including urea, lactate, ascorbic acid, uric acid, dopamine, and glucose. Its practical application was demonstrated in urine samples, with results showing strong agreement with the standard creatinine assay. Machine learning models, such as Random Forest, Extra Trees, and XGBoost, were employed to optimize data analysis, delivering high predictive accuracy and uncovering key electrochemical features critical to the sensor’s performance.