1889
Metal-Coated Fiber for Concentration Detection in Gas Mixtures Using the 3-Omega Excitation Method

Tuesday, 30 May 2017: 10:30
Grand Salon A - Section 4 (Hilton New Orleans Riverside)
S. Kommandur, A. Mahdavifar, P. Hesketh, and S. K. Yee (Georgia Institute of Technology)
Gas sensors have been widely used to determine the composition of gas mixtures, and to detect the presence of a species in gases. Thermal conductivity detectors (TCDs) are a common type of gas sensors, which use the difference in thermal conductivity of gases to detect the composition of gas mixtures. Miniature TCDs have been demonstrated to have excellent sensitivity, upto and lower than 1 ppm. However, these TCDs often require high power consumption (~100 mW to ~5 W) to provide low sensitivity which limits the usage of these sensors for remote applications. Furthermore, some of these TCDs operate at high temperatures, which necessitate routine re-calibration and may be destructive to the detected sample. We have developed a gas sensing technique that can provide sensitivities better than 50 ppm while consuming significantly low power and operating at near room temperature. In this work, the 3-Omega measurement technique was applied to a fabricated metal-coated glass microfiber for the purpose of low power gas sensing. The metal coated glass fiber was fabricated by depositing a thin layer of gold (~150 nm) onto a glass microfiber (diameter ~30 microns), using a custom designed deposition lathe installed in a standard sputtering system. The uniformity of the metal coating was verified using Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Spectroscopy (EDS). The sensor was wired in a four-point probe configuration for 3-Omega analysis. In a standard 3-Omega measurement, a sinusoidal electric current at frequency ω is driven through the thin metal heater layer, resulting in a Joule heating at frequency 2ω. The periodic heating creates a thermal wave that penetrates the surrounding medium. The thermal wave attenuates over a penetration depth, which depends on the thermophysical properties of the surrounding medium. This gives rise to a temperature oscillation at the source, which has a frequency 2ω but lags the heating current by a phase lag ϕ due to the finite time it takes for a temperature response. The temperature oscillation causes the resistance of the heater to oscillate at 2ω. Since the current is driven at 1ω, this resistance oscillation results in a voltage component at 3ω. The amplitude, V3ω and phase lag, ϕ of the voltage signal can be directly measured and the voltage amplitude can be directly related to the amplitude of temperature oscillation.

The sensor performance was evaluated for mixtures of CO2, Ar, He and CH4 in N2 in an isothermal chamber, where mass flow controllers precisely controlled concentrations. A custom 3-Omega conditioning circuit controls the AC heating current and detection of the 3-Omega voltage signal. The amplitude and phase lag, and the in-phase and out-of-phase components of the 3-Omega voltage signal are measured for different gas mixtures and are related directly to their concentrations. Using this gas sensing technique, we have demonstrated the uncertainty in concentration (i.e., sensitivity) to be better than 50 ppm, and as low at 10 ppm for some gases. The variation between the different 3-Omega signals for different gas mixtures can be related to their the thermophysical properties, particularly specific heat and thermal conductivity. The sensor operates at near room temperatures with temperature oscillations less than 10 K, while consuming power as low as 20 mW, which is ~10% of the power consumed by the conventional TCDs. Furthermore, the ease of fabrication of the metal coated microfiber is an advantage compared to some of the more complicated sensor designs, which adds to the robustness of the sensor. While the sensor performance has been evaluated on binary mixtures, it is possible to extend the analysis to ternary mixtures as well, without significant reduction in sensitivity.