Quantitative Decoding of Complex Gas Mixtures Using Mixed-Potential Sensor Arrays

An array of mixed potential electrochemical sensors is used to determine the individual concentrations of NOx, NH3 and C3H8 gases in a mixture containing oxygen. The sensors comprise a dense La0.8Sr0.2CrO3 or AuPd sensing electrode and dense Pt electrode with porous YSZ electrolyte. Bayesian model training is used to analyze the data and to find the composition of the gas mixture. The model is successfully demonstrated for predicting binary gas mixtures with less than 10% error. Ternary and quaternary gas mixture analysis is underway to test the reliability of this model to predict the gas composition with accuracy. This work aims to provide robust and accurate NOx sensing for vehicle on board applications.