CH4 Detection using an Interband Cascade LED Coupled to a Hollow-Core Fiber

Abstract:

Mid-infrared, interband-cascade, light-emitting devices (ICLEDs) have the potential to improve the performance of low-cost, trace-gas sensors for air quality and greenhouse gas measurements, but the applicability of ICLEDs has not yet been widely demonstrated. Optical-based detection methods allow for fundamental concentration measurements through the Beer-Lambert law with high-sensitivity, selectivity, and stability. ICLEDs are broadband, incoherent, high-optical-power devices (up to 3 mW continuous-wave at room temperature) that are similar to light-emitting diodes (LEDs), but have active cores similar to interband cascade lasers. ICLEDs have large potential for a high-power optical source at low cost in moderate quantities. We present an ICLED-based, methane sensor using a hollow-core fiber and direct absorption spectroscopy. An ICLED with a peak, emission wavelength of 3.4 µm and an output power of 0.5 mW was coupled to a plastic, hollow-core fiber (1500 µm diameter, 1 m length) for atmospheric methane detection. A two-lens system coupled light into the fiber with ≈50% coupling efficiency. A beamsplitter and reference channel were used to normalize for any changes in the output of the light source (e.g. thermal drift) for high-precision measurements. The detectors were 3-stage, photovoltaic, HgCdTe detectors. Calibrated amounts of methane in nitrogen and water in nitrogen were added to the hollow-core fiber by standard dilution of flows over a range of 0-50 ppmv CH₄ and 10-23,000 ppmv H₂O to test for sensitivities in this spectral region. A LICOR LI-7810 measured water vapor and methane downstream of the fiber for a direct intercomparison. An Allan-deviation experiment yielded precision of 0.34 ± 0.01 ppmv CH₄ at 1-s integration time. Water vapor sensitivity at 1 Hz was 6.8 ± 0.3 ppmv. Longer integration times will yield improvements in precision. Sub-ppmv methane detection is relevant for monitoring emissions near sources such as petrochemical infrastructure, agricultural feedlots, landfills, and wastewater treatment plants as a standalone sensor or part of a sensor network. Future work includes minimizing the influence of water vapor on the retrieved CH₄ signal through narrow spectral filters as well as improving long-term drift.