Friday, January 27, 2012

How to Carbonate Pretty Much Anything

Carbonated apple juice.

This has been a project I've been wanting to do for several years, but I've always lacked the confidence to try it until after watching Ben Krasnow do many things of the same ilk.  The concept is relatively simple; soda gains its bubbly taste and much of its acidity through a process called carbonation, where carbon dioxide is dissolved into a water-based solution.  Gases are actually always soluble in water (which is why fish don't drown), but if you raise the pressure of the gas pressing down on the water (or cool the water), progressively more gas will dissolve into the water.  Conversely, if you lower the pressure, the dissolved gas will come out of solution and bubble out.

Here's a fun experiment: Fill a drinking glass with tap water and leave it somewhere undisturbed for several hours.  Notice how lots of small bubble of gas form on the inside of the glass.  Why is that?

So soda is carbonated by dissolving more CO2 into it than naturally exists in the Earth's atmosphere.  This causes the soda to bubble once you release the pressure in the bottle, and the carbon dioxide actually combines with the water to form carbonic acid, which gives soda much of the tangy flavor which you'll notice is lost when soda goes flat:

The chemistry really isn't so bad; take something water-based, add excess CO2, enjoy the pleasurable bubbly and tangy sensation you traditionally associate with soda and other carbonated beverages.

Doing this at home is very easy, but potentially rather dangerous.  Getting high pressure CO2 isn't too difficult; you can go to the grocery store and buy dry ice for about $1 per pound.  Dry ice is solid CO2, so as it sublimates into a gas, it vastly expands (there is an almost 800x volume difference between CO2 as a solid and a gas), but if you instead trap it inside a pressure vessel, it becomes a high pressure gas.

Now you just need to build a pressure vessel to contain the CO2 and whatever you want the CO2 to go into.

There's a hitch; you can't just drop a chunk of dry ice into a plastic bottle with some water.  Well, you can, but this is a traditional past-time done by teenage delinquents due to the fact that the bottle inevitably explodes.  We don't want our pressure vessel to explode.  We want to hold it at something like 30PSI for awhile and then be able to pour out our freshly carbonated beverage.  This is the problem that has previously always scared me off of this project.  With Ben's help, I've managed to come up with the beginnings of a solution to this, which prevents such inconveniences as injury or death.

Meet my pressure vessel.  I built this out of 3" schedule 40 PVC pipe, a couple pipe fittings, a 100PSI pressure gauge, and a needle valve drain cock.  I horse-traded the pressure gauge from my ever-supportive father for helping to clean up his shop this winter break, but was able to buy the rest of it at an Ace Hardware for about $25.
Building the chamber out of 3" schedule 40 PVC is important.  Each diameter and schedule of pipe has a different operating pressure, which is the highest pressure at which the pipe is still safe.  larger schedule numbers indicate higher pressures, so if the 260PSI operating pressure of 3" schedule 40 was too low for this, we would move up to schedule 80 or schedule 120 NPS, at the expense of higher cost and it being much more difficult to purchase SCH80 or SCH120 pipe retail.  Although 3" SCH40 pipe has a rated 260PSI operating pressure, and a bursting pressure well above that, in my calculations I derated the end cap to half that because of the two holes drilled into it for the pressure gauge and needle valve.  The PVC parts needed for this should all be easy to find in your local hardware store's gardening or plumbing sections:
  • 16" of 3" SCH40 PVC pipe.  Many hardware stores will custom-cut this to length for you if you ask nicely.  I ordered 16" to give me a final chamber volume of about 8.5 cups, but feel free to build your chamber as big or small as you like.
  • 1x 3" slip end cap.  This is the fitting in the bottom center of the picture above, and is to be cemented onto the bottom end of the pipe.
  • 1x 3" slip to female NPT coupling.  You can get 3" pipe with pipe threads on it, but it's often much easier and cheaper to use slip fittings everywhere except where you need your access port.
  • 1x 3" male NPT plug.  This is a threaded plug which I installed the pressure gauge and needle valve into, and which I take on and off to refill the chamber between batches.
  • Teflon tape - This is often called plumber's tape, and is used to seal the threaded joints between the pressure gauge, needle valve, plug, and NPT coupling.
  • PVC cement and primer - The slip fittings can all be permanently joined, since the threaded plug can be removed to fill, empty, and clean the chamber.
This project was the first time I've ever dealt with NPT, or National Pipe Thread.  This is a special pressure-tight thread designed for pipe fittings, which has a distinctive taper to it.  As you're wandering through the hardware store, pause for a moment to appreciate the threaded pipe and notice that it gets skinnier towards the end.  This gives the thread much more compressive force when it's screwed together than normal straight threads, which causes it to be pressure tight.  Fun fact: NPT tapers off at 1° 47′ 24″, which works out to be an inch for every 16 inches of thread, and a crazy small angle.

The pressure gauge and needle valve are really important.  Since the CO2 is trying to expand to 800 times its original volume, it is very easy to add so much dry ice to the chamber such that we will go well beyond the 260PSI limit of the pipe, and create a very impressive, if unfortunate, bomb.  The pressure gauge lets us monitor how high the pressure gets inside the chamber, and the needle valve allows us to just crack it and bleed off CO2 as slowly as we want.

Fixing the valve and gauge on the plug does require access to a fairly well stocked machine shop.  Pulling out my copy of Industrial Presses Machinery's Handbook, I was able to find tables for how large of a hole to drill for both the 1/8" and 1/4" NPT taps I need for these two fittings (28th ed, p1863, table 1b, 1/8" NPT - drill letter Q, 1/4" NPT - 7/16").  Drill, tap, apply teflon tape, screw in, enjoy.

Charging the chamber
 Now for the actual contents of the chamber.  There are several possible sources of high-pressure CO2, but probably one of the simplest and cheapest is dry ice.  You will need to be a little deliberate to make sure you  choose a grocery store that carries dry ice (Many carry it, but nowhere near all of them), but once you find one that does carry it, you just need to ask the cashier for 1-2 pounds of it when you're checking out; they'll go unlock the box, weigh out a chunk, and bag it for you.  You'll want to break up the dry ice into smaller pieces, since each batch isn't going to use very much dry ice at all, and it'll go faster with small pieces.

As far as exactly what you carbonate, the sky is pretty much the limit.  The classics include water (optionally with flavored syrups to make Italian sodas), fruit juices, and reviving flat sodas or beer, but anything with a high water content works.  I sized this chamber rather deliberately such that I can easily fit apple and orange slices in it.  The carbonated fruit comes out with a very intriguing sparkling aftertaste.  I only let the fruit soak for ~20 minutes, which wasn't quite long enough (they went flat rather quickly), so solid fruit would probably be something you'd want to pressurize several hours early, get the chamber stabilized, and just let it sit until ready to serve.

When filling the chamber, you will not want to fill it much beyond half way with whatever you're carbonating.  The dry ice will sublimate very rapidly, and the less head space you have in the chamber, the faster the pressure will rise in the chamber.  By leaving several cups of head room above the liquid, it becomes much easier to throttle the needle valve to keep the pressure below the operating limit of the chamber.

Be very careful while the chamber is pressurizing.  You will find that the pressure rise is not at all linear (for a variety of reasons) and it's quite possible for the chamber to go from 50PSI to pegging my 100PSI gauge in a matter of seconds.  You simply can not walk away from the chamber while it's coming up to pressure, but once the dry ice finishes sublimating, you can seal the chamber and leave it pressurized for as long as you like.  You might even be able to install another stop cock on the bottom of the chamber and have carbonated drink on tap for as long as you like.

Fundamental problem with this setup

Generating high pressure CO2 gas from dry ice is fundamentally a problematic affair.  In the ideal world, I wouldn't have anything to do with the stuff and use a commercial gas cylinder and gas regulator to charge the chamber, but those are expensive.

The real problem is that it is very easy for the dry ice to get away from you, particularly when you're using it to charge liquids.  Water has a very high thermal mass, which is why dry ice sublimates so quickly when you drop it in the punch bowl during Halloween.  Unfortunately, the water tends to freeze on the dry ice, cause it to sublimate slower.  Once the chamber finally reaches your target pressure and you start venting it, the ice will often crack off the dry ice and you'll see the chamber start to rise in pressure quite rapidly.  During one of the batches, I cracked the needle valve when the chamber reached 60PSI, and in a matter of seconds the pressure inside the vessel pegged above 100PSI before I was able to spin the valve wide open.

A possible modification to solve this icing issue would be to separate the solid dry ice from the water entirely.  Building a second (much smaller) chamber to sublime the dry ice, and then pipe the high pressure CO2 back into the main chamber, would be much easier to control.  You would still want the dry ice to be sitting in some sort of liquid to act as a heat source, but you would then be able to use something with a much lower freezing point, such as high proof ethanol, and could even make the secondary generation chamber somehow detachable to charge multiple vessels.

In the end, I still rank this as a rather difficult and dangerous project, and you can buy carbonation tools retail which are much much safer than this, so I won't tell you not to do this, but make well sure you know what you're doing.  I found it an interesting way to spend an afternoon; anyone else in Davis want to have a sparkling fruit juice party?


  1. Nice! I noticed that the apple slices lose carbonation quickly even after being in the chamber for a day at 50 psi. It's best to eat them right after depressurizing. What else are you planning to carbonate?

  2. I'm looking forward to having access to fresh strawberries; those'll probably be pretty good. You could make some pretty dang fancy salads with carbonated cranberries or the such.

    Has anyone actually bothered trying something hard like carrots yet?

  3. I would not mess around with this PVC pressure chamber without at least one automatic gas pressure relief valve (spring-loaded, direct action relief valve). They are cheap and easy to find - just ask Google.

    Maybe it would be possible to make a pressure relief system out of some tubing, and S-bend, and a separate column of water.

    Put sushi or sashimi in this thing and tell us how it comes out.

    He He - David

  4. Thank you for this post. I believe that I understand that your model requires that the beverage (water or juice) be placed in the PVC pressure vessel you have constructed and the sublimation/carbonation occurs when the dry ice is immersed in the beverage and the cover is secured. I further understand that the liquid will initially create a thermal barrier that will inhibit the sublimation until the barrier is cracked. Additionally I understand the offering of the ethanol as a means to prevent the frozen shell/barrier on the dry ice. However, if we establish a separate sublimation chamber as you illustrated, is there still a need to accelerate the sublimation by immersion? Was the acceleration by immersion in the initial model a desirable condition or one by default of placing both the beverage and dry ice in the single vessel? I believe that the single vessel also allowed the dry ice to lower the temperature of the beverage so that it will absorb the gas more readily.

    Please advise the benefits of accelerated sublimation.

    If a separate sealable aluminum vessel were able to hold the dry ice and vent the CO2 without allowing and beverage to flow into the vessel, while the vessel is submersed in the beverage, you could still chill the beverage while carbonating it . . . right? Please provide your thoughts if you are still interested in this project.