Ozone detection in the ppb-range with improved stability and reduced cross sensitivity

In the following paper, a novel method for determining the concentration of ozone using commercial microstructured semiconductor gas sensor elements is presented. The “virtual multisensor” principle was used with a quasi continuous sensor temperature variation (temperature cycle) to improve stability and reduce cross sensitivity to typical interfering gases compared to constant temperature operation. Characteristic features extracted from the sensor resistance response allow accurate prediction of the ozone concentration in the relevant range from 0 to 100 ppb. The feature extraction and concentration prediction algorithm considerably improves reproducibility, yielding valid results also for aged sensors and even when applied to different sensors of the same type reducing the necessity of individual sensor calibration. The combination of low power consumption and simple data processing allows cost-efficient realization of handheld devices or ozone detection in home air cleaners.