Measuring CO2

Australia & New Zealand Homebrewing Forum

Help Support Australia & New Zealand Homebrewing Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Lyrebird_Cycles

Well-Known Member
Joined
10/7/16
Messages
1,439
Reaction score
777


Carbodoseur


Measuring CO2 without very expensive equipment has always been difficult: The method of choice for DCO2 is an Anton Paar Carbo QC analyser, but at something like $10k it’s a little out of reach for most. The Zahm and Nagel comes nowhere near the AP for accuracy and precision but they still see fit to charge around $2k. The method we used at the big blue involved a specially made vessel, a vacuum pump, a mercury manometer, concentrated caustic soda and sulphuric acid. Not for the faint of heart.

Fortunately for us there’s an easy method that comes close to the performance of the Zahm and Nagel at less than 1/50th the price. The idea isn’t new, it’s based on the Du Jardin Salleron Carbodoseur but I’ve added a couple of tweaks that make it more practical for brewers.

The cheap ass version of this uses a 100 ml measuring cylinder with a 24 / 29 conical fitting (AKA a mixing cylinder), available for about $20 , a 24 / 29 stopper with a through fitting (known as a thermometer adapter), available for about $10 , a piece of 6mm glass tube and a thermometer (which you should already have).

A slightly more convenient version is the same cylinder with a ready made plastic stopper with tube and valve integrated, available from Alla or the original from Dujardin Salleron, but like everything they make expect to pay over the top .

To make this work for beer, we add water. For every beer I’ve tried 50 : 50 cold water and beer works a treat but super carbonated beers may require further dilution (see below). The water needs to be cold (preferably less than 5 oC) and substantially free of CO2. The easy way to achieve this is to fill hot water into a bottle and refrigerate it. The method works best if the beer is also cold.

Fill the cylinder to the 50 ml mark with the water, put the cap on, close the tube (with your thumb if using the cheap ass version) and shake it baby. Open the tube (or take your thumb off) and check whether the level has changed. If it has, there is dissolved CO2 in your water and you need to get water with none: try boiling it first.

If the level didn’t change, tip that lot of water out and refill the cylinder to exactly 50 ml with water from the same source. Now fill the cylinder to exactly 100 ml with your beer, pouring very carefully so as to minimise CO2 loss. Measure the temperature of the beer / water mix carefully and note the result. If it is more than 12 oC or less than 0 oC start again.

But the cap on, close the tube and shake. Open the tube and let the beer / water mix spurt out. Repeat until no more comes out. Since it is really important to minimise gas leaking, it is best to start with small steps and vent often.Take the cap off and add a drop of antifoam (or vegetable oil if you are a real tightass) and place the cylinder in the fridge until the foam settles (no more than 3 minutes).

Once settled, read the remaining volume of the mix. Remember to do this with the cap / tube removed.

If the volume is more than 94 ml go back to the start and omit the water addition (just use beer). If the volume is less than 40 ml go back to the start and reduce the beer / water ratio.

If the volume is between 94 ml and 40 ml note the result and apply the following algorithm:

DCO2 = sqrt (98 – Vol) x e^(1652 / (T + 273.15) – 6.75) * dilution factor (= 2 for a 50 / 50 mix).

If you don’t feel like doing the maths, I've attached a spreadsheet with a calculator and a table:

The DCO2 calculation returns a result in g/l CO2: if you prefer to work in volumes leave out the dilution factor and you’ll be near enough.

To validate this I tried a few beers including some VB cans. Unless the leopard has changed its spots VB will be ~5.3 g/l CO2. My measured values were between 5.0 and 5.1 g/l which is reasonable assuming a small gas loss on opening the can.







*The algorithm is based on the volume loss being proportional to the square of the dissolved CO2 content but tweaked to fit the empirical curve derived by the OEM (which is why it starts at 98 not 100). The exponential term is derived from simple Arrhenius kinetics and takes account of the change of CO2 solubility with temperature. The dilution factor is obvious.
 

Attachments

  • CO2_Chart.xls
    54 KB · Views: 104
Last edited:
Ah, the old simple Arrhenius kinetics... :what::what::what:

Amazing work though, turning science/maths out of reach for the average HB'er and making it simple and easy to follow. This is why I love AHB and the HB community so much. Hats off to you kind sir!
 
Who are you really, lyrebird? Do I know you from a Nobel Prize award?
 
On the odd occasion I've had to measure dissolved CO2
Just put a 1L beaker and the can/bottle on the scales, weigh to 0.1g (limit of available scale that will also weigh can/bottle beaker)
Open can bottle, pour into beaker, put in ultrasound until all sign of bubbling end (around 10-15 minutes)
Reweigh, do the obvious.
A 330mL beer at 4.5g/L of CO2 would be expected to loose ~1.5g
Probably good enough for most people.
Have to work through the above by LC, there are a couple of things I need to get my head around (how much impact the headspace over the beer matters for one) fascinating tho.
Mark
 
Clever stuff.
We send our agriculture gas samples to a lab where they analyse CO2 and N2O using gas chromatography. About $10 a sample.
Interesting, nitrous oxide is Far worse than CO2 in terms of a GHG but it gets little mention in the funny pages...
 
Who are you really, lyrebird? Do I know you from a Nobel Prize award?
art.jpg
 
A 330mL beer at 4.5g/L of CO2 would be expected to loose ~1.5g
Probably good enough for most people.

1.5 g +/- 0.1 g (stated limit of scales) is +/- 0.3 g/l CO2.

Stated accuracy of the carbodoseur technique (per manufacturer's spec) is +/- 0.05 g/l. The dilution step probably doubles that to 0.1 g/l CO2
 
Last edited:
If that is supposed to mean anything - it escapes me, is it meaningful or just noise?
1.5 g +/- 0.1 g (stated limit of scales) is +/- 0.3 g/l CO2.

Stated accuracy of the carbodoseur technique (per manufacturer's spec) is +/- 0.05 g/l. The dilution step probably doubles that to 0.1 g/l CO2
Possibly in theory, and yes I am quite familiar with the concept of error.
If you went and bought the made for purpose equipment maybe the sort of error you are talking about might be realistic.
Go buy some random measuring cylinder - most say something like 100mL approximately, never seen one that I would trust to be more accurate to a couple of mL without calibrating it. Even a class A measuring cylinder is rated to 0.6mL error
Measure two 50mL aliquots (beer/water), how, a 50mL pipette, Class A 0.05mL error, use the measuring cylinder and that goes through the roof.
Pressure is dependant on volume, the amount of head space will make a big difference, once you established how much you could do a work around, but how is a home brewer going to accurately do that (on a set of scales)
Temperature measurement is going to be critical (well maybe just important), measured how with what and to what precision?
Even the height of the blow off tube would make a difference....

It's a fascinating bit of kit, and for a lab probably gives consistent and repeatable results, on a home brew scale I strongly suspect it's results would be pretty suspect, the biggest problem being how do you standardise the results so you know they are meaningful.
Mark
 
Actually headspace seems to have very little effect, neither does the height of the blow off tube. There is no pressure term as the idea is to vent pressure every step.

Temperature dependent accuracy can be calculated from the algorithm, an ordinary $40 thermometer is more than acceptable.

The accuracy of the cylinder graduations is the major determinant of the error which is, as stated, 0.05 g/l. That's about the same as Zahm and Nagel achieves (assuming the pressure gauge on your Z & N is spot on).
 
Pressure is dependant on volume, the amount of head space will make a big difference, once you established how much you could do a work around, but how is a home brewer going to accurately do that (on a set of scales)
Temperature measurement is going to be critical (well maybe just important), measured how with what and to what precision?
Even the height of the blow off tube would make a difference....

Actually headspace seems to have very little effect, neither does the height of the blow off tube. There is no pressure term as the idea is to vent pressure every step.

What and where are we taking these measurements, are we talking about in the fermentation of the wort?
 
No doubt the factory made set ($200+ most places) does a pretty good job.
My comments mainly relate to a home made version, if anyone is seriously expecting accuracy of 0.05g/l - well it would be time to change hands.
Just the number of steps and the cumulative error would make that imposable, not to mention all the other variables.

I actually like the idea, it is simple and if one had a good enough piece of equipment no doubt reliable and accurate. I just know home brewers too well to think many people would either build or use a carbodoseur well enough for it to be worth the cost. I can see a 1+mL error measuring cylinder with a silicone up rubber bung being used to report results to three decimal places.
If you used a 250mL measuring cylinder (as head space, height of the blow off... don't matter) why couldn't you just use 100 mL samples of beer - the only thing that comes to mind is that there might be enough evolved CO2 to blow the whole sample out of the measuring cylinder.

One other point in the OP, you are attributing up to 0.3g/L from a benchmark 5.3g/L loss just to opening the package
0.3/5.3*100= >5% error. Same sort of error you don't like with the method I posted.

I'm sure either method will if used properly with decent equipment give very similar results. Anyone needing to get a pretty good idea of dissolved CO2 (if you are entering comps - a very good idea - Jesus talk about a crap shoot), go with which ever method you have the equipment for or an inclination to invest in.
Mark
 
If you used a 250mL measuring cylinder (as head space, height of the blow off... don't matter) why couldn't you just use 100 mL samples of beer - the only thing that comes to mind is that there might be enough evolved CO2 to blow the whole sample out of the measuring cylinder.

You can scale it up and the same algorithm applies, just use percentage volume. As you surmised, if you use 100% beer you lose so much volume you go off the chart.

It's normal to lose some CO2 when you open a package, lab grade package samplers use a sealed spear system to prevent this.
 
Last edited:
Back
Top