Recognition of volatile organic compounds using SnO2 sensor array and pattern recognition analysis

Abstract A sensor array with 10 sensors integrated on a substrate was developed to recognize various kinds and quantities of volatile organic compounds (VOCs), such as benzene, toluene, ethyl alcohol, methyl alcohol, and acetone. The sensor array consists of gas-sensing materials using SnO 2 as the base material, plus a heating element based on a meandered platinum layer, all deposited on the substrate. The sensors on the sensor array are designed to produce a uniform thermal distribution and show a high and broad sensitivity and reproductivity to low concentrations through the usage of nano-sized sensing materials with high surface areas and different additives. By utilizing the sensing signals of the array with an artificial neural network, a recognition system can then be implemented for the classification and quantification of VOCs. The characteristics of the multi-dimensional sensor signals obtained from 10 sensors are analyzed using the principal component analysis (PCA) technique, and a gas pattern recognizer is implemented using a multi-layer neural network with an error-back-propagation learning algorithm. Simulation and experimental results demonstrated that the proposed gas recognition system is effective in identifying VOCs. For real-time processing, a DSP board can be used to implement the proposed VOC recognition system in conjunction with a neural network.