Machine Learning Assisted Metal Oxide‐Bismuth Oxy Halide Nanocomposite for Electrochemical Sensing of Heavy Metals in Aqueous Media

Vijayalakshmi Kailasam,Radha Sankararajan,Muthumeenakshi Kailasam,Sreeja Balakrishnapillai Suseela
DOI: https://doi.org/10.1002/crat.202300173
2024-01-20
Crystal Research and Technology
Abstract:In this article, Silver Oxide Bismuth OxyBromide nanocomposite is identified for the electrochemical detection of Nickel and Copper ions in aqueous solution. The CV and DPV analysis is carried out for with different electrolytes. Linear response is obtained with a correlation coefficient of 98%. ML models are built for various nickel concentrations classification and an accuracy of 93.2% is achieved. Heavy metal in excess quantity is one of the major inorganic pollutants in water. It causes several hazards to human life and ecosystem. It exists in traces in most of the commonly available drinking water sources from lakes, ponds, wells, etc., However, their presence in treated water is relatively significant. As the treated water is primarily used for agricultural purposes, it is necessary to monitor and measure their concentration. This requires sensing of metals in aqueous medium with good sensitivity and stability. Recently, nanosensors coupled with electrochemical transducer is preferred for analyzing heavy metal in aqueous solutions. In this work, Silver oxide‐bismuth oxy bromide coated with nafion is proposed as an electrochemical sensor for detection of heavy metal ions in aqueous solution. Cyclic voltammetry (CV) behavior of the proposed electrode is observed in different electrolytes. Further, Differential Pulse Voltammetry (DPV) study shows that current increases with trace nickel and copper metal ions of different concentration. Further, machine learning (ML) algorithms such as Naïve Bayes, ANN, SVM and decision trees are employed for nickel ions to train the cyclic voltammetry data and evaluate its performance. Naïve Bayes algorithm provides the best accuracy of 93.2% among all the models.
crystallography
What problem does this paper attempt to address?