Integrated Photonic On-chip Sensor for Methane Fugitive Emissions Monitoring

The oil and gas industry represent a tremendous global energy growth sector, with over 5×10⁵ active well-pads in the United States (US) alone. Natural gas (NG) has received particular interest, given its potential as an intermediate "bridge fuel" prior to full migration into renewable energy sources; however, the primary constituent of NG is methane (CH₄) which presents a radiative forcing > 20× higher than CO₂. Given the significant expansion of oil and gas drilling operations in the United States, the efficacy of NG as a clean fuel is dependent upon minimization of fugitive emissions during NG extraction, processing, and delivery, which requires source magnitude estimation and localization capability for leak detection and repair (LDAR). The principal challenge therefore involves the development of high-fidelity source estimation technologies for spatially and temporally resolved CH₄ emissions monitoring, to enable fast and site-specific LDAR competency.

In this paper, we demonstrate the first photonic integrated chip sensor (PICS) for high-resolution tunable diode-laser absorption spectroscopy (TDLAS) of the near-infrared 2ν₃ band of CH₄, which may be deployed at scale in a wide-area sensor network for fugitive emissions monitoring. Our PICS incorporates a III-V laser/detector chip bonded to a silicon photonic substrate, where light is guided through a long (20 cm) waveguide for evanescent field sensing of ambient CH₄. The III-V gain is utilized in an external cavity laser (ECL) configuration to provide narrow-linewidth, broadband tunability under ambient conditions, without need for thermal control. In addition to its compact spatial footprint, our PICS are fabricated in a high-volume wafer-scale silicon photonics process, thus enabling disruptive SWaP-C (size, weight, power, and cost) as compared to conventional free-space optical sensors. Using the fully integrated PICS, we demonstrate detection of time-varying CH₄ concentration mimicking leaks from NG infrastructure. Sensitivity analysis indicates 1.9×10^-4 Hz^-1/2 minimum fractional absorbance and bandwidth normalized CH₄ sensitivity of 92.8 ppmv•Hz^-1/2, thus demonstrating our PICS as a viable TDLAS platform for environmental trace-gas monitoring.

The authors acknowledge funding support under ARPA-E Award Number DE-AR0000540.