Bayesian decoding of the ammonia response of a zirconia-based mixed-potential sensor in the presence of hydrocarbon interference

Zirconia-based mixed-potential sensors are a promising technology for monitoring levels of nitrogen oxides and ammonia in diesel engine exhaust. However, in addition to the target gases these sensors react to unburned hydrocarbons present in the gas mixture. The observed cross-interference between target and non-target gases cannot be fully mitigated by applying different bias currents to the sensor. On the other hand, sensor sensitivity and selectivity toward various components of the mixture depend on the bias current setting, allowing us to effectively create an array of sensors by applying different bias currents to the same device. Here we show how such an array can be used to predict absolute concentrations of ammonia in the presence of propylene. Our Bayesian framework can be easily generalized to other types of sensors and to more complex chemical mixtures. It consists of two steps: the calibration step, in which the parameters of the model are determined a priori in the laboratory setting, and the prediction step, which mimics the deployment of the device in real-world conditions. We investigate a linear model, in which response of the sensor to each gas is assumed to be additive, and a nonlinear model, which takes interference between gases into account. We find that the nonlinear model, although more complex, yields more accurate predictions. We also find that relatively few sensor readings and bias current settings are required to make reliable predictions of gas concentrations in the mixture, making our approach feasible in a variety of automotive and other technological settings.