The use of a tungstic oxide semiconductor as a sensor for ozone at concentration levels relevant to atmospheric monitoring applications is an important advance in attempts to produce cheap, lightweight and reliable instruments. Problems of stability are a possible obstacle to this application. A model that describes the response of these sensors to ozone is proposed here and using it an explanation for the drift of resistance with time at constant concentrations of ozone is given. Consideration of this drift model enables a measurement routine to be employed that compensates for the drift observed experimentally, thus producing a reliable calibration of the sensor.

The behaviour of gas-sensitive resistors based on WO3 towards small concentrations of ozone in air can be understood with a simple model involving the reaction of ozone with surface oxygen vacancies. This model has been validated by comparison with experimental results for the effects of varying oxygen partial pressure on the ozone response. A complete description of the behaviour of devices constructed by printing WO3 as porous layers onto an impermeable substrate requires consideration of the effects of the microstructure of such a device upon its response. A very simple series-parallel equivalent circuit model captures the effects and allows a simple interpretation of the sensor behaviour, including the quadratic limiting steady state resistance response to ozone and the effects of variation of device thickness. An important fact that allows WO3 to be used at rather high temperatures as an effective ozone sensor is that ozone does not decompose at any discernible rate on the oxide surface. Saturation of the oxide surface at ambient temperature with water vapour inhibits the ozone response when the sensor is subsequently heated. The effect can be removed by heating at sufficiently high temperature. Water vapour also gives a high-temperature sensor response, but appears to act at sites different to those that mediate the response to ozone.