To The Point
University of British Columbia led study combines ultrabroadband (~7-15 μm, 1-2 μm) frequency comb with an 81 m pathlength cavity. A wedged diamond plate is used as the input and output coupler to the cavity in contrast to the traditional use of a partially transmissive input coupling mirror. While 1 ppm methanol in air data are reported, researchers predict a 12 parts per trillion detection limit enabled in part by a free-space electro-optical sampling system (EOS).
Journal Article Source
Nature Photonics
Cavity-enhanced field-resolved spectroscopy
This study effectively has three technological components that make it unique. It combines an ultrabroadband frequency comb, an optical cavity that implements a diamond wedge instead of the traditional partially transmissive input mirror, and an EOS method for collecting and reporting frequency resolved absorbance spectra. The use of a 0.5 mm thick, 1° wedge plate enables the use of unprotected gold-coated mirrors within the cavity.
This EOS method is implemented in contrast to the typical use of Fourier-transform infrared spectrometers that are used to measure the ultrbroadband output of frequency combs. The study reports an average finesse over the ~7-15 μm of about 50 and greater than 40 in the 1-2 μm range.
Opinion
Admittedly, even for me who’s a subject matter expert in gas spectroscopy, this paper was tough to get through. My biggest hang-up was the frequency comb aspect of the work. I’ve never had the opportunity to use one. The implications of this work seem to be the leap in sensitivity that these researchers have accomplished.
The effective cavity pathlength of the system used is 81 m. The EOS system used in the study has the ability to detect relative amplitudes on the order of 10-6 at 30 THz for an averaging time of 25 s and a scanned-time window of 4.5 ps. Given the 50 signal-to-noise ratio reported on the detection of 1 ppm of methanol, the study claims a detection limit of 12 parts per trillion. Wowza!
Research papers are notorious for claiming sensitivity limits beyond what was actually demonstrated. However, that aside, this potentially opens the door for detecting elusive gas species like H2S. So what’s the catch? You need an expensive femtosecond laser system, a high-precision optical cavity with a diamond wedge, and serious data analysis know-how. I’m not familiar with processing EOS data but it seems non-trivial at first glance.
What’s a bit unique about this work in the context of traditional gas spectroscopy is the use of a free-space EOS device as the signal collection device. Typically we characterize the laser output in the time domain and convert that into the frequency domain using optical devices like etalons. When dealing with femtosecond timescales this approach is prohibitive. It would be interesting to explore the use of free-space EOS systems in traditional tunable diode laser applications. I certainly would love to!
