Preliminary Study for the Implementation of a Software Method for Humidity Compensation in E-noses for Outdoor Applications

Abstract

The competitive adsorption of water over Volatile Organic Compound (VOC) in Metal Oxide Sensor (MOS) is well known in literature and is one of the main disturbing factors that contributes to worsen electronic nose (e-nose) performance in classifying and quantifying odours. This constitutes one of the most important limitations for the large-scale diffusion of e-noses since it hinders their use for several applications. In this paper, we investigate the possibility to implement the Orthogonal Signal Correction (OSC) method for compensating the relative humidity effect on MOS sensors. Two different e-noses have been used for this study, each one equipped with a different sensor array comprising only MOS sensors. In order to investigate the relative humidity (RH) effect on the sensors response for different compounds, four calibrants (acetone, butanol dimethyl-disulphide and toluene) have been analyzed at differente levels of RH (20%, 50% and 80%, for a fixed temperature of 20°C) for training the instruments to compensate its intereference. The concentration of the sample analyzed has been set equal to 2.5 ppm. The results achieved proved that the OSC implementation can be a suitable method for the mitigation of the RH interference on MOS sensors, since its application significantly improved the performances of classification, increasing the global accuracy above 70% for both e-noses considered in this study.