Research On The Zirconia-based Electrochemical Gas Sensor Array

The pollution of vehicle emissions to environment has become increasingly serious in recent years, especially by the hydrocarbon, carbon oxygen compounds, and nitrogen oxides, enough attention must be paid to. Air-fuel ratio control and three-way catalyst technology were mainly used to reduce the exhaust emissions, as these are related to the sensor, so the performance of the sensor is very important. And because of simple structure, preparation technology mature, wide range of gas and easy to measure the output electric potential, the gas sensor on the basis of yttrium stabilized zirconia(YSZ) solid electrolyte has become a research focus in the car gas sensor. In addition, considering the automobile exhaust gas sensor must give a high sensitivity and selectivity response to target gas in high temperature, electrochemical sensor based on YSZ has been received widespread attention in current market.Traditional gas sensors can’t give a description in detail to a variety of gases with only a single working electrode, which led to the research of sensor array. Sensor array contains more than one working electrode(while shared with a reference electrode), each electrode is mainly sensitive to one gas. The use of neural network will be effective for signal separation later.The research object of this article is the harmful gas of automobile exhaust, a real-time monitoring system based on sensor array would be set up, so a series of work was carried out, mainly included: 1) The construction of dynamic distribution system. According to the requirements of the study, the LabVIEW software platform was used, also the mass flowmeters and a NI data card of USB-6009 were used to build a real-time and dynamic configuration of the concentration of gas composition. 2) The design of automatic acquisition system. The LabVIEW software platform was combined with an acquisition NI card of USB-6221, then it was connected to a set of experimental test equipment, to realize automatic acquisition of sensor signals. 3) The screening of sensing electrode materials. In order to get high sensitivity and high selectivity electrodes, it was necessary to select the electrode materials. SnO2(sintered at 1000 oC) for C3H6, ZnFe2O4(sintered at 1100 oC) for NO and ZnO(+30wt.% In2O3) for CO were used finally. The best proportion of In2O3 doped to ZnO(30wt.%), the best sintering temperature(sintered at 1200 oC) and the optimum working temperature(500 oC) were also studied. In addition, XRD, SEM and impedance spectrum test for some explanation were also conducted. 4) Preparation and test of sensor array. Good filtering materials were prepared on an YSZ tube with a common Pt reference electrode to form a sensor array, the single and mixed gases were tested respectively. The results showed that the response of the sensor array for single gas was similar to the response of sensor with only a single working electrode, however, gas cross interference characteristics were existed when to the mixed gases. 5) Neural network algorithm. In order to eliminate the cross sensitivities of the mixed gases, the neural network algorithm was presented in this paper. Adaboost integrated BP neural network was used to optimize, the integration of signal was separated effectively. The operation precision of Adaboost integrated BP neural network was compared with the operation precision of single BP neural network, the results showed that the prediction error of Adaboost integrated BP neural network was lower than single BP neural network. 6) Compensated CO sensor. In order to improve the selectivity of CO sensor, ZnO(+30wt.% In2O3) was used as sensing electrode and cylinder Cr2O3 material was used as reference electrode to form a compensated CO sensor. Results showed that it really improved the selectivity of the sensor.In this paper, the results showed that the combine of electrochemical sensor array based on YSZ with neural network algorithm to form a system for detection was really successfully applied to classify the identification and concentration of C3H6, CO and NO in mixed gas, while the prediction error was not higher than 2%. This sensor array can detect gas concentration range of 0-500 ppm, and effectively improve the detection accuracy of the hybrid vehicle exhaust.