ABSTRACT
Chemical sensors generally suffer from the cross-sensitivity problem. For this purpose, the identification of volatile organic compounds (VOCs) using unselective sensor requires a combination of sensors followed by pattern recognition methods. In the current study, four quartz crystal microbalances (QCMs) were coated by plasma-enhanced chemical vapor deposition (PECVD) with the pure vapor of hexamethyldisiloxane (HMDSO) at monomer pressures ranging from 5 to 40 Pa. The coated QCM-based sensors were used for the detection of volatile organic molecules. The sensor responses toward different VOCs showed the possibility to tune the chemical affinity of the sensors by changing only the monomer pressure during the sensitive layer deposition process. Fourier transform infrared (FTIR) spectroscopy and atomic force microscopy (AFM) were used to investigate chemical composition and surface morphology of the coated QCM sensors, respectively. The sensor responses have been used as a database for the identification of VOCs using pattern recognition methods such as principal component analysis (PCA) and artificial neural networks (ANNs). PCA is a linear method that allowed the identification of ethanol only among two other VOCs. Complete identification and accurate quantification of the studied VOCs have been achieved by the combination of PCA for data preprocessing and ANN for pattern recognition.