Monday, September 19, 2011

Supercritical CO2 caffeine extraction (negative result -- more work needed)

I tried to extract caffeine from green coffee beans using supercritical CO2, but I had no success. The beans underwent a strange transformation, becoming white and rubbery after 6 hours at 80*C in supercritical CO2. I also used water and ethanol as a cosolvent, thinking that the caffeine would end up in solution in the water/ethanol mix after the CO2 became subcritical.

Do you have any advice about how this process is supposed to work?


  1. You are doing it in batch, and should be doing it with a continuous stream of CO2 passing through the beans. The beans are absorbing the water and ethanol.

  2. Usually you do not use a co-solvent, but rather pass the CO2 through activated carbon after it travels through the beans, where presumably it sticks to the carbon. The carbon dioxide then goes back through the beans, I guess until equillibrium is reached. I don't know how they get the caffeine out of the filters.

    Caffeine is really cool because it is easy to purify by sublimation, so often whatever muck you get out by co2 or solvent extraction, once the solvent is gone, heating the stuff with a cold finger in the flask will yield beautify needle like pure white crystals growing from the cold finger.

    Ethanol is perhaps too miscible with water, and it really doesn't dissolve caffeine any better. Maybe the beans are getting water logged or something. You could weigh them before and after, which would tell you whether you gain mass, and how much.

  3. Same Anon as before here-

    How about if you used a cold and hot end of the vessel to get some convection in there? Also, keeping the beans off the bottom, and warming the bottom and cooling the top may do the trick to get some of that caffiene out :)

    Glad you liked the Aerogel idea, Ben! I can't wait to see what you do with it. Any base material of choice? Is the tank made of steel?

    What if you were to cut off the bottom, insert a thick threaded piece into the bottom (solder it in, maybe? It could have sufficient surface area) so that the interior of the cylinder is threaded... That might make a safe enough vessel.

  4. Anonymous, thanks for the comments. The tank is about 3/8" thick aluminum. The alloy is unknown, so attempting to weld it makes me nervous. My welder tops out at 200 amps, so 3/8" aluminum will be quite a challenge. I think I will try something else.

    For the aerogel, I will probably just use steel pipe fittings to make the chamber. I will also need a large CO2 tank, since the supercritical drying requires lots of CO2 liquid changes.

    I am trying to buy TMOS for methyl-based silica aerogel, but Sigma-Aldrich is still "reviewing" my order. Amateur chemists face many more challenges than amateur electrical/mechanical engineers.

  5. Same Anon as before-
    That's why I was recommending soldering a threaded endcap in :) I know they make materials which can solder aluminum well enough, and with sufficient surface area for the soldering, it should be just fine.

  6. Ben-

    TEOS would be a bit safer to work with, since the hydrolysis you will do will yield ethanol rather than methanol. Secondly, is usually a bit more reasonable for silicon compounds. Get a chemistry teacher to order you some, if you know one. Generally schools can get small quantities of stuff without too many issues, especially something relatively mild like an orthosilicate.

  7. Ben,

    I think you would have better success if you used another vessel as an expansion and collection cell. The co2 will expand and become supersaturated allowing your analyte to percipitate out.

    Also if you use ethanol in your extractor you'll never know if your co2 did the extraction or the ethanol. As for the water.. idk, but I don't think it's helping?

    Here's a calculator if you want to know exactly how much dry ice to add to achieve a certain pressure.,%20enthalpy%20and%20entropy%20of%20carbon%20dioxide.htm

    Wish you luck, much respect. My own extractor is on the way. I'll have to share it with you when I have it up and running.