Saturday, December 31, 2011

Caffeine extraction from green coffee with supercritical CO2



I finally succeeded in extracting caffeine from green coffee beans by using supercritical CO2. I built a high pressure chamber from 2" steel pipe fittings, and poured in 200mL of water. There is an aluminum screen above the water line, which held 0.75 lbs of moisturized green coffee beans in the upper part of the chamber. I added liquid CO2 to the chamber, then closed all valves and raised the temperature, making the CO2 pass into the supercritical phase. I left the system overnight at about 60*C, 3000 psi, then drained the water. It was very black due to impurities and some bean burning that occurred where my electric strip heater caused localized overheated zones in the chamber. The water was highly caffeinated, and tasted somewhat like coffee. I used a typical hydrocarbon extraction process to isolate the caffeine from the water (will show this in a later video).

16 comments:

  1. Is this a proof-of-concept or do you have plans for the caffeine?

    ReplyDelete
  2. Those are some scary sounding pressures, but really impressive. What is it about supercritical fluids that makes them such good solvents compared to regular liquids?

    ReplyDelete
  3. Very impressive, Ben.

    Though caffeine has a carbon skeleton and many hydrogen atoms, most chemists would not refer to it as a "hydrocarbon," reserving that term for organic compounds that are without nitrogen, oxygen, and other so-called "heteroatoms." In the context of extraction from water, "hydrocarbon" sounds especially weird, because what we generally think of as "hydrocarbons" won't dissolve in water in the first place, since they are basically oils without heteroatoms to polarize their bonds. A more useful term for a nitrogen-containing small biologically active molecule from a natural source (e.g. caffeine, cocaine, morphine) is "alkaloid."

    ReplyDelete
  4. Wouldn't the supercritical CO2 alone extract the caffeine? Forget the water. Then release the dry ice, let it sublime and you will have caffeine. NO?

    ReplyDelete
  5. The supercritical CO2 alone cannot dissolve much caffeine. I would need a very large chamber relative to the coffee mass in order to extract caffeine without water. Commercial decaffeination plants use hot water to wash the CO2 of dissolved caffeine, then recirculate the CO2 back through the coffee beans. It would not be practical to only use CO2, AFAIK.

    ReplyDelete
  6. timelessbeing, I only wanted to test the ability to use supercritical CO2 as a solvent in a home shop. However, I will continue with caffeine experiments.

    Ptolom, that is a very good question. I don't know, but I plan to search for answers soon.

    Sean, thanks for the info. I was referring to the solvent used to extract the caffeine (methylene chloride) as a hydrocarbon. Maybe I misspoke.

    ReplyDelete
  7. Amazing blog Ben,

    You do some sensational stuff, the nitrogen liquefier, the supercritical CO2 chamber etc.

    A safe way to check for leaks is to put a soapy water mixture into a spray bottle and spray over the pipe joints.(after pressurising the pipeline with a gas) The leaks should form bubbles.

    You indicated the pressure could have got up to 3,000 psi. I was wondering where you attached the valves into the 2 inch end plugs, did you tap the threads for the valves yourself or did you obtain the plugs pre tapped. If they shot out at that pressure they could do some damage.

    ReplyDelete
  8. Have you ever approached the idea of supercritical extraction from a theoretical point of view?
    I am currently trying to design an industrial scale caffeine extraction column for my final degree project. I have read most of the patents but they only give rough values for extraction eg. %extraction, time, co2 flow rate etc.
    Have you come across ways of calculating your own values or are you just basing it on the rough values from the patents?

    ReplyDelete
  9. Hugh, I am only basing my process on rough guesses from the information in patents, although some of the patents are quite detailed. I have no experience with industrial-scale supercritical CO2. Good luck -Ben

    ReplyDelete
  10. Hi Ben,
    I am looking into making something similar to your rig. and along with engineeringtoolbox.com are there any other sites you would suggest? Also do you think that super critical carbon dioxide will erode some coating on the steel pipes? I saw that in your Aerogel experiment the methanol ate away the alumium screen and left a black soot in the chamber.

    Again your experiments are amazing. Thanks for the videos.

    ReplyDelete
  11. Ben, search for pressure ratings for standard steel pipe. Use schedule 80 or 160 from mcmaster.com.

    The aluminum reacted with the methanol chemically, and this was accelerated by the temperature. I don't think CO2 would react with steel or any common surface coating. However, supercritical CO2 will degrade nitrile O-rings, so use Viton if you plan to keep the vessel supercritical for longer periods of time.

    ReplyDelete
  12. What is a long time? I would like to make a chamber like your acrylic rig. It is an exceptional proof of concept. With what I have seen though keeping it pressurized for a long time would not be wise.

    I plan on varying the time on my super critical extraction rig. Starting with a fast wash(less than an hour) and increasing the time. Do you have any suggestions on a electric heater. Yours looks salvaged. If so what from? I suppose I could use a heating blanket. But I was wondering what you used.

    ReplyDelete
  13. Ben, I am not really sure about plastic degradation times, but I would avoid using acrylic or buna-n for an extractor vessel, since the supercritical CO2 will dissolve some of those polymers. For my extraction vessel, I would use only steel, PTFE (Teflon), and Viton.

    The heater was bought from McMaster. Search for flexible heaters, and strip heaters.

    ReplyDelete
  14. i'm very interest about this project. and i wanna to know, can co2 make caffeine back to coffe bean, i mean inject a something (e.g. flavor,caffeine) to coffe bean?
    thanks

    ReplyDelete
  15. You do realize Sch 80 is only good to about 2500 psi at 2"?

    Sch80 sounds like a potential pipe bomb in this application.

    ReplyDelete
  16. You are awesome! Keep up the good work

    ReplyDelete