A thin-film SnO 2 sensor system for simultaneous detection of CO and NO 2 with neural signal evaluation

Abstract

Simultaneous CO and NO 2 measurements are of importance for the ventilation control of automobiles and other applications. For this purpose semiconducting SnO 2 sensors are often used. A well known disadvantage of SnO 2 sensors is the concurrent reaction of the oxidizing NO 2 and the reducing CO on the sensor surface, which causes a near zero sensor signal in the presence of both gases in a certain range of mixtures. A second disadvantage of SnO 2 sensors are the long rise and decay times of the sensor signal. The combination of different SnO 2 sensors, operated at different temperatures and combined with a signal evaluation system based on a specially trained neural forward network (artificial neural net (ANN)) solves this problem. The runtime version of the neural net is a small program, compatible with customary micro controllers. These signal evaluation techniques are applicable to similar problems using sensor arrays or single sensors in a non-stationary operating mode.