Abstract
This paper presents a highly sensitive thermoelectric sensor for catalytic combustible gas detection. The sensor contains two low-stress (+176 MPa) membranes of a combination of stoichiometric and silicon-rich silicon nitride that makes them chemically and thermally stable. The complete fabrication process with details, especially the challenges and their solutions, is discussed elaborately. In addition, a comprehensive evaluation of design criteria and a comparative analysis of different sensor designs are performed with respect to the homogeneity of the temperature field on the membrane, power consumption, and thermal sensitivity. Evaluating the respective tradeoffs, the best design is selected. The selected sensor has a linear thermal characteristic with a sensitivity of 6.54 mV/K. Additionally, the temperature profile on the membrane is quite homogeneous (20% root mean standard deviation), which is important for the stability of the catalytic layer. Most importantly, the sensor with a ligand (p-Phenylenediamine (PDA))-linked platinum nanoparticles catalyst shows exceptionally high response to hydrogen gas, i.e., 752 mV at 2% concentration.
Highlights
The new technological advancement of micro fabrication has opened a platform for research and development to improve the performance and the characteristics of micro sensors for different applications such as combustible catalytic gas sensing [1,2,3]
The resistance temperature detector (RTD) catalytic gas sensor presented by Trochimczyk et al consumes only 2 mW power for a pulsating supply [6]
The key aim was to perform a comparative analysis based on the heat loss, power consumption, the uniformity of the temperature field on the membrane, and the thermal sensitivity of the sensor
Summary
The new technological advancement of micro fabrication has opened a platform for research and development to improve the performance and the characteristics of micro sensors for different applications such as combustible catalytic gas sensing [1,2,3]. The membranes were created by the RIE from the backside with a 10-μm-thick photoresist mask Both the 600 nm SiN and the 500 nm SiO2 were removed by the CF4 flow at 60 sccm with 1800 W coil power and 20 W RF power, whereas the deep etching of the Si wafer was done with SF6 and C4F8 gases at 350 sccm and 130 sccm, respectively, at 1800 W coil power. The RF power was continuously switched between 25 W and 60 W during the deep reactive ion etching of the Si wafer (etch rate 5.85 μm/min) with a chamber pressure of 4.5 Pa. The deep reactive ion etching was stopped on the etch stop layer, SiO2, beneath the membrane that was removed afterwards wet-chemically by oxide etch modified 7:1 of Honeywell, Germany because the incongruity of the internal stress due the presence of silicon oxide layer can hamper the stability of the membrane [4,16]. The process described by Buchner et al [41] inspired the fabrication process
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
