Kitty Litter And Broken Light Bulbs Power This Homebrew Gas Chromatograph

We’re always on the lookout for unexpected budget builds here at Hackaday, and stumbling across a low-cost, DIY version of an instrument that sells for tens of thousands of dollars is always a treat. And so when we saw a tip for a homebrew gas chromatograph in the tips line this morning, we jumped on it. (Video embedded below.)

For those who haven’t had the pleasure, gas chromatography is a chemical analytical method that’s capable of breaking a volatile sample up into its component parts. Like all chromatographic methods, it uses an immobile matrix to differentially retard the flow of a mobile phase containing the sample under study, such that measurement of the transit time through the system can be made and information about the physical properties of the sample inferred.

The gas chromatograph that [Chromatogiraffery] built uses a long stainless steel tube filled with finely ground bentonite clay, commonly known as kitty litter, as the immobile phase. A volatile sample is injected along with an inert carrier gas – helium from a party balloon tank, in this case – and transported along the kitty litter column by gas pressure. The sample interacts with the column as it moves along, with larger species held back while smaller ones speed along. Detection is performed with thermal conductivity cells that use old incandescent pilot lamps that have been cracked open to expose their filaments to the stream of gas; using a Wheatstone bridge and a differential amp, thermal differences between the pure carrier gas and the eluate from the column are read and plotted by an Arduino.

The homebrew GC works surprisingly well, and we can’t wait for [Chromatogiraffery] to put out more details of his build.

Thanks to [Heye] for the tip.

29 thoughts on “Kitty Litter And Broken Light Bulbs Power This Homebrew Gas Chromatograph

  1. Chromatograph.

    I’ve the hunch that what you’re describing there is called “Chromatograph”. Chromatograms are its products. It’s a meta-chromatogram, or in OO parlance a chromatograph factory :-)

  2. Balloon tanks are actually only a small portion helium from what I have been told. The remainder is just air to save money. They are a great source of helium for leak checking but I wonder if it will mess things up for a GC.

    1. I believe the air content is to reduce the risk of asphyxiation by inhaling the helium. As I understand it High concentration of helium displaces CO2 which inhibits the breath reflex.

      1. We breath in oxygen but don’t feel its absence. We breath out CO2 and or bodies sense its excess, so as long as we exhale into large enough volume so that co2 concentration does not Spike, we feel ok, even if we are not getting oxygen.
        In large volume foreign gas or high altitude asphyxiation area we simply lose consciousness in a minute or two.

        1. I remember that was a plot point in the Larry Niven stories about the time travelers in Rainbow Mars and the connected short stories.

          The people from the distant polluted future had to wear inflated CO2 concentrating bags over their heads when traveling in the past because the lower concentration of carbon in the atmosphere would not activate the breathing reflex.

    1. If you’re using an already characterized carrier and matrix, then it’s just a matter of identifying peaks and looking up the elution time in the database/table. If no such database is available, then you need pure (or known mixtures) reagents and create your own table… Then you can compare unknown samples to your lookup table.

      1. unless you have a mass spectrometer (and even then some tinkering is required for reasons i am not going to go into) you almost always inject standard samples to characterize the system response (retention time, peak area) to known substances and concentrations and then use that data to make measurements. a lot of the time the system has very individual responses depending on the exact make and exact length of the column, its impurification content, the state of wear of your valves, minute possible leaks that might or might not develop, etc…

        unless, in a very reduced set of applications, you are not interested in identifying or measuring the concentration of something or other. like, for example, you just want to know that SOMETHING is different. say, you take continuous samples of gas from a gas duct and just want to know if something about it changed or not. you take a baseline reading and then compare to that. but even that is finnicky to draw conclusions from, with drift, and saturation, and sensibility changes over time and whatnot.

        so…. you can rarely import characterization data from samples from one system to another.

  3. This is amazing! I wonder what else you can do with thermal conductivity cells?

    If you pass the air between peltier effect coolers and cycle the temperature, will you see peaks when you get above the boiling point of the various VOCs in the air?

    What if you use a laser to heat something? Is there anything interesting about the graph of temperature vs what boils off?

    How about identifying plastics or measuring alcohol content by looking at the temperature vs thermal conductivity graph of a nichrome wire?

    This opens a lot of possibilities now that hobbyists know these methods are DIYable.

      1. That’s really cool. I wonder if there’s any hidden information there besides just pressure?

        If you sweep the temperature, will you see anything interesting in the temperature vs heat loss curve? Perhaps when passing the decomposition or boiling points of various unwelcome VoCs?

  4. We had a similar (commercial) device in school (“Realgymnasium” in German, not university). I was in voluntary extra chemistry lessons where I once had the opportunity to use the machine.
    It for sure used a hot wire detector and helium. I am not sure any more about the filling of the column, perhaps some silica gel material – which would also be available as cat litter.

  5. I worked on an oven temp controller for the column as a tech at Varian Aerograph in ’75 (19).
    It used a platinum temp sensor and it’s function was to precisely ramp the oven temp at a programmable rate, maintaining the linearly increasing setpoint to within .2 C.
    The controller was a big board full of TTL counters and other chips which was laid out by a job-shopper using pieces of tape. (She was paid $10/hr, pretty nice in those days.)
    –Probably just a few lines of code for an Arduino.
    The effect of ramping the temperature was to spread out the resolution of the peaks as the run progressed.
    They also had flame ionization detectors as well as the TC kind you built.

  6. Perhaps to heat the coil he could pass a currant through it directly rather than heat it inductively or wrapping an insulated coil around it. That way it would be heated nice and evenly and be readily controllable.

  7. Very nice. Don’t apologize for the dry-cell reference. They were standard in all the old equipment and worked great. Often as a kid I could get surplus or free equipment from Boeing Research like chart recorders, and turned out to only need a new ref battery. A real next challenge would be to make the output of your chromatograph the input to a mass spec. I think a time-of-flight mass spec is today within fairly easy reach of the DIY builder and it avoids all the magnet stuff. Good vacuum and the detector end might be major hurdles and probably means getting hold of surplus microchannel plates and stuff.

  8. You can also use powdered detergent as a sep column material. I’ve done demos where you pack a column with powdered laundry detergent and then hook it to the lab gas. When you inject an organic compound at the front end (use a T – fitting and a septum) you can see the elution happen as the flame at the far end of the column changes over time. With a little practice it’s a very cool effect.

  9. Two older construction articles on the subject, both from Scientific American’s “Amateur Scientist” column:

    How to build a gas chromatograph and use it to separate the components of mixtures – June 1966
    How to construct a gas chromatograph that can measure one part in a million – September 1967

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.