Single resistive sensor for selective detection of multiple VOCs employing SnO2 hollowspheres and machine learning algorithm: A proof of concept

Selective detection of harmful gasses and volatile organic compounds (VOCs) in the ambient has become a major challenge. Primarily, semiconducting metal-oxide based gas sensors sense various gases simultaneously, hence their selectivity is poor. This paper presents a single chemiresistive metal-oxide gas sensor for identification of multiple VOCs accurately by employing highly sensitive microstructure and machine learning tools. Tin oxide (SnO2) hollowspheres were taken as sensing material that were prepared through optimized hydrothermal route. Different characterizations were carried out to confirm the formation of desired morphology and structural features. The sensor device was fabricated by controlled drop cast technique over gold based interdigitated electrodes. The sensor showed remarkable response towards the target VOCs with high sensitivity and fast recovery time. Incorporation of machine learning algorithm on the obtained sensor data provided accurate identification of all the VOCs (best performance shown by random forest). In addition, the quantitative prediction of gas concentration was performed for each target gas using regression model. In comparison to e-noses (having array of sensors with different sensing material), a single chemiresistive metal-oxide sensor with proper machine learning tool is simple, economic, compact and easy to fabricate.