Vol.4 No.2 2011

Research paper : Thermoelectric hydrogen gas sensor (W. Shin et al.)−101−Synthesiology - English edition Vol.4 No.2 (2011) by controlling the composition and thickness of the catalyst[5]. In the case of new born sensors, it is generally difficult to be accepted as a reliable technology for practical use. In order to demonstrate the reliability of the sensor, our thermoelectric sensors have been installed and tested for a year at Ariake hydrogen station in Tokyo. This field tests has validated the stable sensitive system-level safety operation, showing response to 100 ppm hydrogen in air [6].3.2 The other elemental technology:technology of evaluationAssembling each elemental technology, a sensor device is created. Then the method analyzing or checking the "workmanship" of the final product of the device also becomes an important issue. The device performance may satisfy the manufacturer, but it should be evaluated or tested for characteristics which are determined by social demands, customer needs or specifications. In this study, the following three testing methods were utilized as shown in Fig. 3:・ new test method to observe the surface temperature changes on the sensor device in the sensor operation using IR camera,・ standard test methods for the response speed (Appendix ISO CD26142), ・ system technology to test the gas sensing performance for serial products.As noted above, the thermoelectric gas sensor consists of two components of the ceramic catalyst and the thermoelectric device. The first test method is to evaluate the two different elemental technologies of the sensor separately to get a clear balance between the individual components. The second test method is a simple test method not using any special or complicated equipment so that anyone may evaluate the performance of sensor response in the same way. It is as simple as just filling the gas to be detected into a chamber with a capacity of 30 liters. This method is adopted as an appendix in the international standards issued in the year 2010 (ISO) as an evaluation method for the testing of sensor response time.The third one is the test method for mass production. How much is the manufacturing cost of every single gas sensor element? Fabrication of small sensor elements costs little but testing of sensors costs much. The testing equipment for the materials properties is commercially available but the evaluation device for a “product” such as sensors is not commercially available. Sometimes the testing machine of the products is the most important secret or know-how for the manufacturers. So we started from an early stage of research and development efforts to develop a high-efficiency testing system as a top priority of mass production technology of the device.3.3 The first step is to set a scenarioThe integration of the elemental technologies mentioned so far is part of how the research progressed, aiming at solving various technical problems in order to meet the social demands. Figure 4 shows a chronological scenario of the thermoelectric hydrogen sensor development. We received a three-year research fund for the development from fiscal year (FY) 2000, and the achievement was the “knowledge” of fusing catalytic combustion and thermoelectric conversion as an elemental technology. The idea was very fragile, and as soon as we started, we found that even if the sensor was based on an excellent principle, to bring about satisfactory Fig. 2 Correlation between the sensor performances and elemental technologiesThree performance requirements, selectivity, sensitivity, and stability are of trade-off relations. Novel working principle of thermoelectric conversion of local temperature differences in sensor device induced by catalytic combustion can bring about the selective, stable, and sensitive detection.BAStabilitySensitivitySelectivityH2OH2O2CatalystThermoelectricElectrodeWaferThermoelectric filmPt catalystΔT ⇒ seebeck effect ΔVH2+O/catalyst ⇒ H2O+heatTemperature difference, ΔVA-B⇒ SignalCatalystThermoelectric hydrogen sensorMicrofabricationOperating principle


page 30