Detecting Chemical Contaminants in Water Using AI

Water is the most important resource on Earth, and among other essential resources, its existence is intimately correlated with life on Earth. Even though water covers about 70% of Earth’s surface, around 97% of its water resources are the saltwater and unfit for human use. About 2% of glaciers and ice caps are inaccessible. Only around 1% of all water resources remain available for domestic, commercial, residential, and agricultural uses. A number of policies were passed to regulate water pollution since it has significant adverse effects on both ecosystems and people. Water contamination detection is the initial stage in pollution control. For testing the quality of water, several approaches have been put forth, including spectroscopy, chromatography, and electrochemical technologies. Conventional testing techniques require large laboratory apparatus, incompatible for on-site testing in real-time zone. Water quality examination may now be done effectively and conveniently with microfluidic devices, which can overcome the drawbacks of conventional testing equipment. In addition, artificial intelligence (AI) is a perfect tool for identifying, categorizing, and forecasting data from microfluidic systems. Significant work is being done on developing AI- and machine learning (ML)-based microfluidic devices for the next stage of water quality monitoring devices. ML, deep learning (DL), robotics, expert systems, fuzzy logic, and natural language processing are the six broad subsets of AI. In water toxicity prediction, fuzzy logic, ML, and DL are the commonly used subset. The rapid advancement of both AI and molecular sciences, and more significantly the confluence of these two domains, has caused an unforeseen shift in water process engineering from conventional to AI-assisted processes. Internet of Things (IoT) and AI-driven sensor systems can efficiently detect water contaminants and perform environmental monitoring. This chapter will cover the use of AI-assisted removal and sensing of some major traditional and emerging water contaminants, viz., quaternary ammonium compounds, heavy metals, pesticides, micro- and nanoplastics, microalgae disinfection by-products, endocrine-disrupting chemical pharmaceutical drugs, and per- and polyfluoroalkyl substances. Lastly, the challenges confronted and future prospects for detection techniques based on microfluidic chips and industrial intelligence are also explored.