Development of Monolithic Micro-Led Gas Sensor Based E-Nose System for Real-Time, Selective Gas Prediction

As interest in environmental pollution, indoor air quality, and industrial safety increases, the demand for gas sensors is rapidly increasing. Until now, only high sensitivity and fast response speed were considered important, so there was a problem of high-power consumption and bulky size, and they were installed only in limited places. For higher utilization of gas sensors in the future, low power and mass production are required. In this study, a micro-sized, ultra-violet micro-LED gas sensor (mu LG) that can operate in an ultra-low-power was fabricated, and using a convolutional neural network algorithm, an e-nose system was developed. As a result, it was possible to distinguish four different gas species and predict the concentration of each gas in real time.