Study of the effect of relative humidity on the classification of volatile organic compounds for a quartz crystal microbalance sensors array

Gas sensors are a crucial part of systems called electronic noses. One type of sensor frequently used is the one based on quartz crystal microbalance (QCM). However, relative humidity (RH) is a parameter that strongly affects the response of these sensors. Therefore, a study of the effects of RH on the classification of different volatile organic compounds is necessary. In the present work, an array of gas sensors with ethyl-cellulose, polymethyl methacrylate, Apiezon T and Apiezon L sensing films, was constructed using the drop-casting method. A dynamic system was used to obtain the measurements of the sensor responses. The characterization of the QCM sensors response, as a function of humidity and concentration of ethanol (EtOH), heptane (hp), and ethyl acetate (EtOAc) was performed. Furthermore, the classification of the three compounds was implemented using principal component analysis (PCA), Mahalanobis distance and canonical quadratic discriminant analysis, obtaining 100% accuracy. In addition, a principal component regression (PCR) with cross-validation study was performed to predict the concentration and obtain the limit of detection values for each compound, which were as low as 33.33 ppm to EtOH, 12.5 ppm to hp and 25.0 ppm to EtOAc. • Study of the humidity effect on the classification of three VOCs using QCM sensors • A QCM sensors array with four sensing films (EC, PMMA, ApT, ApL) was used • 95-100% classification accuracy was achieved using PCA, MD and QDA • LODs of 33.3 ppm (EtOH), 12.5 ppm (heptane), 25 ppm (EtOAc) were obtained by PCR-CV • PCR models were obtained with R 2 >0.90 for predicting VOC concentrations