IoT integrated and deep learning assisted electrochemical sensor for multiplexed heavy metal sensing in water samples
Sreerama Amrutha Lahari, Nikhil Kumawat, Khairunnisa Amreen, R. N. Ponnalagu, Sanket Goel
Abstract
Abstract Heavy metal measurement is vital for ecological risk assessment and regulatory compliance. This study reports a sensor using gold nanoparticle-modified carbon thread electrodes for the simultaneous detection of Cd²⁺, Pb²⁺, Cu²⁺, and Hg²⁺ in water samples. Differential pulse voltammetry (DPV) was employed, achieving detection limits of 0.99 µM, 0.62 µM, 1.38 µM, and 0.72 µM, respectively, with a linear span of 1–100 µM. The sensor operated effectively in acidic conditions, with excellent selectivity, repeatability, and reproducibility. Real water samples from various lakes in Hyderabad, India, were analyzed to validate their practical application. To extract the sensing features a convolutional neural network (CNN) model was used to process DPV signals, enhancing heavy metal ion classification with high accuracy. Performance metrics such as precision, recall, and F1 score were evaluated. Integration with IoT technology has improved the user experience, advanced heavy metal quantification capabilities, and further enabled remote monitoring.