Recent advancement in selective gas sensors and role of machine learning

Research on selective gas sensors got significant attention in recent years due to industrialization and rising levels of greenhouse gases/pollutants in the atmosphere. Apart from the detection of gases and volatile organic compounds (VOCs) special emphasis on hydrogen (H₂) detectors is given because H₂ is a good source of renewable energy. Nowadays, gas sensors are installed in indoor and outdoor locations. Moreover, high-performance selective gas sensors are also used in space exploration and to monitor human health. All these applications require enhancement of gas sensing properties, especially gas selectivity of sensing devices. This review discusses the conventional selectivity and the new Machine Learning (ML) selectivity approaches. In the conventional scheme, some of the most common synthesis methods of sensing material and their effect on morphology and device selectivity have been discussed. The incorporation of noble metals and various composites has been found to make sensors extremely sensitive and selective due to the synergistic effect of composite configuration. In modern approach, the selectivity problem of a sensor while facing a mixture of gases has been resolved by analysing the output signals of the sensor using ML techniques. An attempt has been made to compare/evaluate novel advancements in various ML techniques and provide a summary of the prediction accuracy and the classification of gases.