Enhanced non-invasive and non-enzymatic detection of Saliva glucose using CuO nanoparticles modified CuO nanorods: An experimental and machine learning modelling of electrochemical sensing

The conventional monitoring of blood glucose is invasive, painful, and often inhibits patient compliance, driving the urgent need for reliable, non-invasive alternatives. This research work investigated the fabrication of a CuO nanoparticles modified CuO nanorods (CuONP/CuONR) based high-performance, non-enzymatic electrochemical glucose sensor electrode for the direct and non-invasive glucose detection in human saliva. The microstructural, optical and crystallinity structure of CuONP/CuONR were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM), UV-visible optical spectroscopy (UV-Vis), Raman-spectroscopy and X-ray diffraction (XRD). Electrochemical glucose sensing performance of CuONP/CuONR were analysed using cyclic voltammetry (CV), square wave voltammetry (SWV), amperometric i-t and electrochemical impedance spectroscopy (EIS). The CuONP/CuONR electrode has demonstrated a fast glucose sensing response time, high linear glucose detection range (0 to 325 µM), high sensitivity (5272 µAmM −1 cm − 2 ), very good repeatability, reproductivity, stability and selectivity. Additionally, various machine learning model were also implemented to enhance the detection limit and increase the accuracy of glucose detection. The artificial neural network (ANN) model appeared as a most suitable ML model to interpret the current density vs. glucose concentration relationships, revealing the CuONP/CuONR electrode working capability in real-world application for high precision/intelligent healthcare system. The practical applicability of CuONP/CuONR was validated by successfully detecting glucose in human saliva and also correlated with blood glucose concentration, highlighting its potential as a scalable, cost-effective, non-enzyme and non-invasive platform for glucose sensing application.