Quick Summary
A TDLAS sensor incorporates a 4 cm disk to generate a 1 m pathlength and measure atmospheric oxygen levels using the 763 nm line. Researchers use a WMS-2f peak detection scheme to measure atmospheric oxygen concentration over a period of 24 hours. The VCSEL-detector package weighs 75 g and measures 90 x 40 x 20 mm (3.5 x 1.6 x 0.8 in). The sensor achieves an Allen deviation of 0.06% when using a 318 s sample integration time.
Journal Article Source
Sensors (published 5 October 2023)
Abbreviations
TDLAS: Tunable Diode Laser Absorption Spectroscopy
WMS: Wavelength Modulation Spectroscopy
VCSEL: Vertical Cavity Surface Emitting Laser
Opinion
The researcher’s WMS-2f approach seems interesting but their correlation equation on page 5 doesn’t match up with the data they show.
C = (Peak(WMS-2f) – 0.02148)/0.15
This is an interesting approach since all they’re doing is looking at the peak of the WMS-2f signal and correlating that with a known oxygen concentration. The paper doesn’t actually mention which peak they’re using for the calculation. It’s a little bit of a head scratcher but I can see how using a correlation trace could potentially be helpful if you’re making measurements in a relatively stable environment. I think there are some serious issues with this approach (i.e. varying temperature and pressure) but it might work for unique use cases.
The authors make reference to using this kind of oxygen sensor in rocket propulsion, space station functionality, or confined cabin environments. There’s just no way the 318 s (5.3 min) sample integration time they recommend is a viable option for measuring oxygen levels in these environments unless folks or rocket engines get real good at holding their breath.
The novelty I see in this paper is the size and weight of the diagnostic. Not quite the size of a common coin (USD quarter is just under 2.5 cm) but small enough I guess. Maybe they have large coins in China… I’m not sure. I’ve been meaning to visit. In this study they use a transparent film to enclose the inner surface of the disk to prevent contamination. This would hinder real world applications but I think that’s a solvable problem depending on the application. Overall it was an interesting read. It gives me some ideas around how to dramatically increase the pathlength of a sensor in a terribly small form factor. I’ll definitely try experimenting with this design.
