Electrical conduction in solid-state gas sensors

The gas-sensing properties of metal-oxide semiconductors have been studied extensively over the past twenty years. There has been renewed interest recently in their application in differential gas sensor arrays and the association with cellular automata and neural networking methods. A diffusion-based analytical model is considered that relates the electrical conductance of a porous thick-film semiconductor gas sensor to its physical and geometrical properties. The theoretical responses of sensing elements with several configurations have been derived, and compared with experimental data on tin-oxide sensors exposed to simple alcohols. The results obtained broadly agree with the predictions of the basic model at low gas concentrations (<50 ppm); but, at higher gas concentrations (>50 ppm), the model needs to incorporate a dependence of the gas diffusivity upon concentration.

The effect of changes in sensor design on the response is considered, and the optimum coplanar electrode configuration is obtained when the electrodes lie a distance equal to the electrode separation below the semiconductor surface. Thus, the use of a fully-developed analytical model may well lead to improvements in sensor design and to modifications in the decision criteria currently utilized in pattern recognition techniques.