Introduction High temperature gas sensors play an important role in monitoring or controlling energy conversion processes. Sensors containing functional materials often need to be heated to a certain operating temperature to achieve functionality. This can be realized using planar thick-film heaters, which are screen-printed on the sensor substrate. When sensors are installed in the exhaust gas flow in order to achieve a fast response behavior, high volume flow has cooling effects on the sensor and leads to an inhomogeneous temperature distribution. This might cause a changed sensor function and cross-sensitivities [1]. For this reason, protective caps are used to reduce the flow influence on the temperature distribution with sufficient response time [2]. Resulting temperature distribution can be simulated using FEM models but cannot be verified directly inside protective caps [3, 4]. The present work solves this uncertainty in two ways to directly measure the temperature homogeneity within a cap. Methods and Setup Sensors are built up on alumina substrates (Figure 1a). On the reverse side, a meander-shaped platinum structure acts as heating element. Two methods are used to determine the temperature distribution on the front side.Firstly, the temperature distribution is recorded by an IR camera (Figure 1c). For optically access, an IR-transparent glass replaces one part of the cap. The requirements for this glass are the transmission of the IR radiation (here attention must be paid to the appropriate wavelength of the camera) and a certain temperature resistance, since the cap and the glass are also heated by the heat radiation of the sensor.Secondly, the temperature distribution is directly measured by means of thermocouples (Figure 1b). Instead of a sensor structure, a matrix of planar screen-printed thermocouples (Figure 1d) is applied onto the front side of the transducer. For this purpose, Pt and Au feedlines are arranged so that five different measuring points are realized at the sensor tip, i.e., that area where a gas sensitive layer would be located. Temperature values are derived from the measured thermvoltage on basis of Seebeck coefficients for Pt/Au given in the literature. Results and Discussion In order to verify the function of the printed thermocouples, the sensor temperature was increased to 600 °C by means of the heater on the reverse side. The heat distribution was recorded by the thermocouples without the protective cap. Thermal images were taken simultaneously with the IR camera. The measured temperature of the thermocouples corresponds well to the thermal images.Afterwards the sensor was placed inside a protective cap containing the described IR-permeable glass window. After increasing the sensor temperature to 600 °C, thermal images of the sensor tip could be taken again. The transmission coefficient of the glass is required for this purpose. To check the accuracy of the transmission coefficient, the temperatures of the thermocouples were recorded simultaneously. The temperature measured by the IR camera again corresponds to the thermocouples values. Thus, the application of this method was also checked and confirmed.Both methods were used to investigate the influence of a protective cap on the temperature distribution on the front side of the sensor at 600 °C operating temperature. Figure 1e) and f) show the temperature distribution of a sensor in a horizontal mounting position recorded by an IR camera (temperature gradients in general come from the individual heater structure and several heat losses, especially heat flow along the substrate material). Figure 1e) shows the measurement without a protection cap. The thermocouple measurement shows a temperature gradient of 6.78 °C/mm between point 1 and 2. The temperature gradient derived from the thermal image is 5.88 °C/mm. Both data (thermocouple and thermal image) agree very well. Then the same measurement was carried out with a protective cap (Figure 1f). The temperature gradient, measured with the thermocouples, is 4.97 °C/mm. The thermal image shows a value of 4.43 °C/mm. The protective cap thus has a positive influence on the temperature homogeneity of a sensor even without a flow. Outlook In the next steps, the temperature distribution in an exhaust flow is investigated by means of the thermocouples. Variations in sensor orientation or mounting positions is possible. Here as well it is possible to investigate the influence of a protection cap. A high temperature setup might be realized by use of screen-printed Pt/PtRh thermopiles [5].
Read full abstract