Planar Microwave Sensor suitable for Artificial-Intelligence (AI) based detection of Volatile Organic Compounds

Abstract

Development of a rapid sensors for detecting volatile organic compounds (VOCs) is a need of the hour to effectively mitigate the adverse effect of VOCs on environmental pollution. In this line, the current paper presents the design and development of a non-invasive split-ring resonator (SRR)-based microwave sensor for detecting liquid VOCs, specifically isopropyl alcohol (IPA), acetone, ethanol, and methanol. Artificial intelligence (AI) based algorithms are gaining popularity in developing a highly-selective sensor circuit. In the proposed sensor, the SRR circuit is optimized for better detection sensitivity and the multi resonant behavior of the circuit offers adequate selectivity. The designed sensor offers better re-usability and thereby supporting AI-based algorithms for continuous monitoring of VOCs in real-time. Transmission coefficient ($S_{21}$) of the sensor is measured over the frequency range of 0.8–6 GHz for different VOCs with varying concentrations. Analysis of variance (ANOVA) and post hoc Tukey tests are employed to discern significant variations in the measured data. Principle component analysis (PCA) and discriminant analysis are performed over the measured data to classify the VOCs. These analytical results show that the proposed sensor can be used for generating huge data set to support AI based algorithms in detecting VOCs in real-time.