Due to the complicated gas-sensing process on the surface of metal oxide semiconductors (MOS), it is hard to accurately characterize the transient changes of MOS gas sensor. This paper introduces a modified Hill equation to quantitatively analyze the instantaneous resistance of MOS gas sensors operating in temperature-pulsed operation mode, to extract the gas-sensing dynamic features and key parameters for the selective identification of various volatile organic compounds (VOCs) gases (such as ethanol, formaldehyde, toluene and acetone) by machine learning algorithms. The sensors were tested with the pulse-mode at different temperatures for multiple gases. The characteristic features of the sensing process during the pulse-on period were investigated. Key parameters were extracted via the modified Hill equation, which have shown obvious dependence on the gas type and concentration. Excellent selectivity for the VOC gases was attained at low concentrations of 500 ppb thanks to the equation's further integration with the principal component analysis (PCA) approach. The incorporation of this modified Hill equation in temperature-pulsed operation mode provides the vital features for the possible applications of MOS gas sensors with the emerging machine learning algorithm.
Read full abstract