I’m thinking a lot about methane sensing these days. There’s a theme in the world of methane reduction, where lots of people have ideas for reducing methane emissions, and there’s some funding for trying out ideas, but there’s not a great way to actually measure if the ideas work*. Generally, there are two big categories of methane sensor – cheap and crappy (like the MQ4), or nice and really expensive. A lot of the science efforts that set out to do an experiment testing an idea on methane reduction actually end up doing two experiments – the experiment that they actually wanted to do, and a second experiment trying to cobble together and validate some kind of decent measurement system that doesn’t require them to take out a second mortgage. This situation is slowing down the science around methane reduction, and we don’t want that science to slow down.
It’s helpful to talk about some specs, so let’s say that a helpful system would measure down to ppm concentration, be well calibrated against various moisture/humidity conditions, and if I had my druthers, have an accuracy of ± 10ppm. It would be ideal if such a sensor cost between $100-$1000.
OK, so that’s good context – what about actually building it?
A lot of the low-cost methods for methane detection are based on optical measurements. This review paper gives a good overview of various, current optical methods for detection. My spidey-sense says that many of these can be made much more accessible by:
just building it yourself rather than paying a 10-50x markup
finding and using some more accessible optical components, rather than just buying the most expensive thing on Thorlabs
developing new tricks that haven’t been explored in the current designs
I like this idea, and I’m getting into it. I’m curious, first if there are any other existing open-source projects to make a better methane sensor – a quick search didn’t find anything, but maybe y’all have seen stuff that I haven’t. Second, would anyone be interested in a collab?
Interesting paper. Building good quality spectroscopes is an interesting challenge. To be useful in many capacities it will be essential to have high quality components and good quality calibration.
In the past I have looked at a number of methods such as getting DVDs and then removing the film to create a uniform diffraction grating. If you are looking in IR then a CD would be better. But you do have some curvature issues.
There are likely some great high quality components on AliExpress. As always with AliExpress the issue is the number of similar low quality options and the difficulty of distinguishing. Again the only way you know for sure is testing.
I do think that with the increase of high quality translation stages in open hardware that FTIR is something that is achievable. We had an masters project making a Michelson interfermometer that was planned to become an FTIR (I think the most recent version is in the MPhys_2019 branch). This specific design suffered from parasitic tilt if I remember correctly.
Looking through the paper, this analyser is awesome. Not probably what I would choose to build
My spidey sense says that the useful thing here is a low-cost, high-accuracy system that draws in air and then measures the methane content in a measurement chamber. This might start out as a tabletop thing, because miniaturizing is its own process, but I think this could be miniaturized and made handheld once you have a working prototype – small and low-cost tend to go hand in hand.
There are existing handheld units with prices starting around $11k that people use for measuring gas leaks from pipelines in free space – guy is walking along a pipeline, waving a laser around and seeing if there are any methane leaks. The common design for these handheld systems is a tunable laser shooting out into space, sweeping the laser frequency past a methane absorption frequency, and measuring backscatter from particles in the air. The weird thing about these systems is that they have weird units (ppm - meter), and they don’t give really reliable results in freespace, because you don’t exactly know how far the beam travels before it bounces back, or the shape/thickness of the invisible methane plume where the laser intersects it, and you need those distance numbers in order to get a measurement in ppm. Hard to do in free space.
There are also even fancier handheld units that suck in an air sample and measure it with a fancy spectroscope or contact gas sensor. These tend to be even pricier
It’s easier to get solid ppm measurements when you have a known path length for the laser. That could be a tabletop setup with a laser on one side and a detector on the other, or you could try to quantify distance in the handheld units by pointing the laser at an object on the other side of your methane plume and adding a time of flight sensor, like they have in those laser tape measures. A big tabletop setup also lets you use tricks like bouncing light back and forth several times between parallel mirrors to get a longer path length through the gas, which boosts your signal:noise ratio and improves accuracy.
Thinking about the problem a bit, I think there are four lego pieces in an optical methane measuring system: the light source, the gas handling, the optical handling, and the detection system. Different combinations of legos give you better accuracy, precision, size and price characteristics. Some lego features are additive (for example, adding a second detector with a narrowband optical filter right at a methane absorption peak lets you control for light absorption by humidity and particulates and other gases, adding to your accuracy), and some are multiplicative (a longer optical path length through gas has a linear relationship with the signal:noise ratio). There are several different combinations of legos that would give the right price/accuracy/precision. I think the initial design process would be all abort exploring these combinations.
Thanks that is a really good description - makes me think we should go with something like a 1 m light saber maybe with mirrors held at end mechanically for multiple passes and hold the laser and sensor in your hand…
Hi Alex,
Sorry, this is a very slow response. I would be interested in a collaboration.
I would wonder if a better hand-held remote unit could be built by leveraging recent advances and miniaturization of pulsed laser time-of-flight (ToF) sensors.
Lidar sensing is used for space-based Methane and CO2 sensing:
I work on applications (often non-conventional) of SPAD-based depth sensors and would be excited to investigate overlap of these sensors and remote gas detection.
I have interest in development of open source sensor of Methane, I’m boitata’s developing, this a scientific open hardware anaerobic digester of food waste.
In this time my research is focused in developer a simplified autonomus control, inclusive this sensor. My hypothesis is optical methods.
I think that reaching the 1-5ppm level might be very hard without a lot of optics, which to get stable will always be expensive and involve a lot of prototyping (I’d happily be proved wrong here).
The opening post mentioned the “crappy” MQ-4, Looking at the datasheet there is a fair bit of sensitivity. But clearly it fairly temperature and humidity sensitive (as well as sensitive to certain other gases).
One route to improving an MQ-4 sensor may be to add a humidity and a temperature sensor, and then do some calibration of a few MQ-4 units with varying temperature and humidity. Having onboard temperature and humidity correction would help, as might having a few sensors and doing some averaging. This wont get you down to the 10s or even 100 ppm level, but can probably push a long way to something less crappy but still affordable and easy to replicate, I suppose it depends on the application.
