How the Purity of Sparged Carbon Dioxide Affects the Oxygen Concentration of Beer


My last blog post discussed the importance of carbon dioxide purity when using injected CO2 to increase the CO2 concentration of your beer. The thing we learned is that the purity of CO2 must be very high (99.99% or better) when using injection, or you will at the same time significantly increase your dissolved oxygen levels. However, injection isn’t the only method for adding CO2 to beer. Sparging, in which CO2 is bubbled through beer (usually in a tank with slight over-pressure) is another common practice for boosting CO2. So in this post we will explore how sparging a finished beer tank with carbon dioxide impacts the final oxygen concentration of your product.

First, let’s do a quick review of what happens when you inject CO2. When you inject gas (usually into a pipe,) you are forcing a given weight of gas into the liquid under pressure. All of the carbon dioxide, plus any trace oxygen and nitrogen, gets pushed into the beer and dissolves completely, allowing you to calculate the weight of gas used and extrapolate from there your various gas concentrations.

On the other hand, when you sparge gas into a liquid the dissolved concentration of gas will be bound by Henry’s Law. Henry’s Law tells us that the amount of gas that will dissolve in a liquid will be proportional – at a constant temperature – to the partial pressure of gas in equilibrium with the liquid.  This means that the gasses dissolved in your beer will never be more concentrated that the partial pressure of the gas you are using to sparge.

The consequence is that, with any given CO2 concentration outcome desired, you will have significantly lower oxygen concentrations for sparged beer then for injected beer. For example, in injected beer the oxygen pickup from injecting one volume of 99.95% CO2 (at 0oC) into the beer when the oxygen concentration is 0.01% is 143 ppb. But a theoretical sparging of that same CO2 into the beer at atmospheric pressure would follow Henry’s Law, and your oxygen pickup would be about 7 ppb.

In real brewing situations, however, most brewers use tank overpressure to help get sparged CO2 into solution, so you would probably be picking up about 2 times the above amount, or 14 ppb. The table below shows the expected oxygen pickup given varied percentages of O2 traces (in your CO2) when measured at sea level and at 0oC:

Sparged CO2 at 1 V/V

0.001% O2

0.005% O2

0.01% O2




My final thought is that CO2 purity isn’t nearly as important if you are sparging rather than injecting, since the amount of gas that will dissolve into your liquid is much lower. This also applies to the purity of the gas you use to flush air from tanks before filling.

How the Purity of Injected Carbon Dioxide Affects the Oxygen Concentration of Beer

When an industrial supplier sets a minimum purity for the CO2 they supply to your brewery, you need to be aware of the ramifications of that purity and whether there is any chance it will increase the dissolved oxygen concentration of your beer. CO2 specified at 99.5% or better may sound very pure, but when we do the math we find this is actually a problem. This post is specifically about carbon dioxide that is “injected” into beer. Another post will address CO2 that is “sparged” into beer.

If your CO2 has a 99.5% purity, then the impurity is 0.5%. For purchased CO2, the assumption is that the impurity is always air, so only 1/5 of the impurity – 0.1% — should be oxygen. If you were to add to your beer one Volume of CO2 with an impurity of 0.1% O2, it would increase your oxygen concentration by a whopping 1,420 ppb.

In 1985, Nick Huige and his Miller Brewing Company co-workers published a paper on this subject in the MBAA Tech Quarterly. Their findings on the affect of injecting impure CO2 are astounding and can been seen in the table below.


Amount of added CO2

Concentration of O2 impurity in CO2




0.5 V/V

7 ppb

35 ppb

142 ppb

1.0 V/V

14 ppb

71 ppb

284 ppb

2.0 V/V

28 ppb

142 ppb

567 ppb

Dissolved oxygen added to the beer


So if you want to inject CO2 into your beer, you need to be mindful of the actual purity. I know a brewer whose CO2 specification from his supplier was 99.5%. Most of the time the supply was much purer >99.998%, which is excellent. But when they had an unexpected increase in their dO2 levels, the cause was eventually traced back to the CO2. What was the purity of the “problem” C02? A number that still sounds good on a cursory level – 99.97%! – but was not acceptable in the context of dissolved oxygen in the product.

Most brewers specify a minimum CO2 purity of 99.990%. This equates to an oxygen impurity of about 0.002%. If one V/V of CO2 with this oxygen content were injected into beer, the resulting increase to the beer dO2 would be about 28 ppb.

My final thought is that if you are injecting CO2 into your beer, be sure your purity specification is not too low. You don’t ever want to be in a position where you’ve been getting great CO2 but then have one “bad” batch – still within the manufacturer’s specification – adding too much oxygen to your product.


Brewery Inline Sensor Placement: Validation Techniques


My last couples of posts have dealt with sensor validation. I recently did some coaching on inline oxygen sensor placement, so I’d like to continue that discussion with those examples. One placement was pre-and-post centrifuge in a regional brewery and the other was in a process pipe flowing at about 500 gallons-per-minute. While the flow rate may make a difference in how far a sensor needs to be from pumps and other interferences, the same techniques can still be used to determine sensor placement.

The key to great dissolved gas measurement is making sure to measure in places where all of the gas is in solution. The same validation tips used for double-checking instrumentation can help you decide where to place an inline analyzer. In general we can say that if you measure close to a known – or at least probable — source of air ingress and yet can’t detect the full impact, but measure farther down the process and find higher values, then you probably want to place your sensor further away from that source.

Let’s talk first about the pre-centrifuge application. Since a centrifuge creates a small amount of vacuum, I don’t recommend putting a sensor immediately before it. Where there is any type of a vacuum, there is the possibility of a false low reading, especially if the vacuum can cause any foaming or bubbles forming in the beer due to degassing. If possible, I’d place a sensor where I’m fairly sure there is no suction on the beer.

You can get a false low reading post centrifuge as well, but for different reasons. The first is that any dO2 that ingresses via the centrifuge may not be fully dissolved in the beer in a short run to a sensor. The second is that any CO2 forced out of solution in the centrifuge may lower the dissolved O2. (The CO2 will carry O2 out with it, but then they will both reabsorb farther down the line.) In both cases, you can do better by placing the sensor at a distance post centrifuge.

Validation of inline sensor placement is easy if you use a portable instrument to help, even if you are measuring over a long distance. First take readings with both the inline and the portable at the farthest reasonable sample port from your Bright Beer Tank. If you get similar readings on both, you’ll know that your inline is correct and that you can trust both instruments. Then use your portable to take another sample at a port just before your bright beer tank. If there is more dO2 than you found at your inline placement, then you may have an ingress problem that needs correcting. (You may be tempted to test right in your BBT, but don’t. Tank filling nearly always results in pickup or loss of DO2, so measuring for that variable is a whole different matter.)

The advice I gave to the customer who was evaluating the 500 gallons-per-minute process pipe was similar, and can likewise be applied to most placement situations. Here’s a summary of my basic guidelines for the best inline probe locations:

–        As far as possible from the outlet of pumps

–        At least five pipe diameters from bends

–        As far as possible from CO2 or O2 injection

–        Never in a descending pipe

My final thought is to always validate an inline sensor location if possible. Remember to look for the places in your process where all of the gas in your beer is in solution: that’s usually the best place to put a sensor.


%d bloggers like this: