silicon carbide micro-devices for combustion gas sensing under harsh conditions (1.16mb)英文精品课

3.1.2 Sensor Activation Activation of SiC gas sensors takes place by switching between oxidizing and reducing gases, to obtain a fast and stable sensor response. All the devices of a n-type 6H-SiC sensor chip or 4H-SiC sensor chip were activated simultaneously by flooding the entire chip with gas at 620 °C. In the activation process, alternately, 1.0% O2 and 10% H2 are sprayed onto the entire face of the SiC chip, which is heated by the micro heaters. The signal from one of the sensors – typically a 1mm diameter device – is monitored during the activation process. The sensor measurements were made by monitoring the gate voltage, while holding the capacitance fixed at midgap bias. 3.1.3 Sensor for UHV Studies We prepared a new sample, SiC7+I+I, for use in UHV studies of gate surface chemistry and sensor performance. This sample had 52 gates with nominal diameters of 200, 300, 500, and 1000 μm and was prepared as described in section 3.1.1. A micrograph of the wire-bonded top surface of the sample is shown in Fig. 3.1.3.1. Prior to installation in UHV the sample was activated at MSU as described in section 3.1.2. Micrographs clearly showed roughening of the gate surface, as discussed in section 4.1. The sample was extensively characterized under atmospheric pressure conditions at MSU, to establish a baseline for comparison with the UHV results. Fig. 3.1.3.2 shows C-V curves in 10% hydrogen and 1% oxygen, at 778 K (505oC) before and after activation, showing the increase in sensitivity due to activation. Fig. 3.1.3.3 shows the sensor response as a function of operating temperature. Fig. 3.1.3.1. Micrograph of the sample used in the UHV studies, SiC7+I+I. Eleven gates were wire-bonded by gold wires for the sensor measurements. Marked gates 43, 76, and 81 were initially connected to the outside. The rest of the wire-bonded gates were grounded. All the UHV sensor measurements included in this report were carried out on gate 43. 11 120010008006004002000-1Gate 4310% H21% O2C (pF)0123Vgate (V)Fig. 3.1.3.2. 1 MHz C-V characteristics in oxygen and hydrogen of gate 43 before and after activation, measured under atmospheric conditions. Left two curves were measured in 10% hydrogen, and right two curves were measured in 1% oxygen. Dashed/solid curves were measured before/after the activation. Measurements were done at 778 K. 0.5Delta-V (V)0.40.30.20.1Gate 810400y = 0.0007x - 0.20622R= 0.9975006007008009001000T (K)Fig. 3.1.3.3. Temperature dependence of the sensor response (10% H2 vs. 1% O2) measured under atmospheric conditions. The signal response increased with temperature at a 0.7 mV/K rate. 12 3.1.4 Sensor Mounting for UHV Studies In Phase I we reported a pattern of device failure during UHV measurements, and speculated that charged particles – used for ion sputter cleaning, Auger electron spectroscopy, and sample heating – might have been causing damage to the oxide layer. This hypothesis, if correct, would have seriously limited our ability to use standard surface analysis techniques on working devices. Testing of the “failed” samples, however, revealed that most of the problems were in fact due to failures of the electrical contacts between the device and its mounting header. When reattached, some of the failed samples exhibited C-V characteristics similar to their original curves. We have now developed improved mounting techniques, including the use of back metallization on the SiC chip, better silver paint (GC Electronics, Silver Print II), and a more careful mounting procedure. Mounted in this way, the new sample (see section 3.1.3) has been repeatedly cycled to temperatures as high as 900 K (630oC) and exposed to ion and electron beams without degradation. 3.2 Design and Construction of “Nano-Reactor” We have designed and built an operational sensor test system to simulate the gas temperature and flow rates encountered in the NETL micro-reactor. The new system ‘nano-reactor” is compatible with the module developed in phase I for sensor testing under industrial conditions. As shown in Fig. 3.2.3.1. the nano-reactor contains: 1) An automated gas switching valve to switch between the oxidizer (1%O2) and a reducer (52 ppm to 10% H2 are currently available, 50% H2 is on order). 2) Two thermometers to observe and control the temperature in the Gas chamber and the sensor surrounding area. 3) High Temperature gas chamber, able to pre-heat the flowing gas up to 1000 °C. 4) A stainless-steel union coupled to the gas-heating chamber and capable of accepting the high temperature test module (used in phase I at NETL). 5) A heater strip to heat and keep the pre-heated gas at a desired temperature. 6) This system also contains two valves, a flow-meter, and a pressure gauge for high-pressure operation. 7) Two one-way valves and a flow meter capable to regulate both pressure and the flow rate (0-1000 sccm). A photograph of the main body of the nano-reactor is shown in Fig. 3.2.3.2. 13 Gas ExistUnionGas Heating Chamber RT ≤T ≤1000 °CThermocoupleThermocoupleSensorHeater StripRT ≤T ≤800 °CSensorElectronic BoxGasAuto-controllerOxidizerTo Sensor Heater Controller and Boonton CapacitameterFig. 3.2.3.1 Blueprint of the Nano-reactor” assembly for sensor measurements at MSU. The thermocouples for monitoring the gas temperature are inserted from both ends: One from the left into the union cross and the second one from the right into the gas heating oven. The gas temperature will vary from 85 to 700 °C. The sensor chip will be maintained at ~ 620 °C for fast response. The coaxial connectors are rated for continuous operation at 200 °C. Fig. 3.2.3.2 Photograph of the “Nano-reactor’ assembly with the electrical breakout box and other accessories. 14ReducersLow / High pressureGas entrance 3.3 Software development We have developed a new LabView program for measuring capacitance-voltage (C-V) curves in UHV, using the Boonton capacitance meter. Fig. 3.3.1 shows the front panel of the program. The left panel of Fig. 3.3.2 shows C-V curves measured in UHV with this program during hydrogen and oxygen exposures. For comparison, C-V curves measured on an earlier sample, which had not been activated at atmospheric pressure prior to the UHV measurements, are shown in the right panel. The much larger sensor response of the fully activated sample is apparent. The sensor response is discussed more fully in section 4.4. In addition to being a valuable tool in its own right, this C-V program represents the first step toward developing an integrated program that performs all the necessary tasks needed for sensor operation in UHV: C-V curve scan, sensor signal measurement, and temperature control. The next step will be to integrate active temperature regulation into the system, using a newly acquired analog interface to control our existing heating system. UHV sensor measurements reported in Phase I were carried out under constant-bias-voltage conditions. For those measurements, on devices that had not been fully activated, constant-voltage measurements were adequate, because the C-V shifts were so small. For the larger shifts encountered on fully activated sample, a feedback system is needed to operate under constant-capacitance conditions. We have now identified optimal settings for the feedback program, and are using it successfully for constant-capacitance measurements. All UHV sensor response measurements in this report were made under constant capacitance conditions, with the device biased near mid-gap. Fig. 3.3.1. Front panel of the C-V scan program developed at Tufts. 15