There are numerous discussions on HomeBrewTalk about off-flavors from CO2. Some are non-believers and insist that CO2 is CO2, while others claim to have experienced these off-flavors first-hand. I am primarily a soda brewer, and I'm a firm believer now that CO2 can have a significant impact on flavor. It is not about 99% vs 99.95% purity; it is about what the impurities are.
Co2 Can Have A Significant Flavor Impact
Our senses are exceptionally sensitive to certain compounds, having the ability to smell or taste it in the orders of parts per million or billion. A good example of this is gas odorants in flavor infused water. The additives are in the orders of several PPM, yet we're able to detect these odors as a hint of peach or rose in water, even though the essential oils are in such a small amount that they're not easily measured.
I have a Sodastream machine, and there are only two main components to it. Water and CO2. I've modified it to accept bulk tank and at first it worked great. The line goes up to the machine unregulated and as temperatures change, CO2 condenses in the line. One time I left the main valve open and came back to use the machine a few days later, and immediately noticed a nasty chemical odor from the escaping CO2. Trying a sip, the carbonated water had a nasty chemical taste to it.
Suppose you mark a piece of paper with water soluble ink. It stays nice and clean even in humid air, but as soon as it gets wet, the ink will run. Same thing happens with CO2. CO2 is a good solvent for various oil-soluble components, and even if the gas is pure, it can leach out chemicals such as plasticizers and bring them out to surface, which can then get injected into beverage as an oil mist. Unlike water, CO2 becomes super critical above 88F, so instead of CO2 exiting only in a gaseous phase, it gains the ability to diffuse whatever is dissolved in it in a "fog" like fluid.
CO2 is handled as liquid in transfer, and it will dissolve oils used in process equipment such as the machining oil residues used in fittings and such.
I tested my 2.5 lb aluminum tank which has been filled numerous times, and sure enough, it contained quite a bit of residue. It was rinsed internally with petroleum ether which was collected. The rinse was collected and solvent was allowed to evaporate. The residue was allowed to dry at 180F, which gave off an odor of oil. The residue is a mixture of viscous oily liquid mixed with metal shavings, and it was weighed out to 61 milligrams on an analytical balance. Even if this was all soluble, it would only amount to 54 PPM in 2.5 lbs of CO2.
I believe that the difference between "beverage grade" and "industrial grade" CO2 is the filling process, and measures are taken to ensure PPM level residues with toxicity or objectionable flavor/odor are not introduced.
After experiencing this off-flavor, I think the most worthwhile solutions are an oil mist filter to remove the bulk of the impurities and an absorbent filter such as charcoal to grab the last bit of contaminants. Once the system has been contaminated, I would replace the plastic lines and clean the regulator, and disassemble and solvent-wash all metal parts. I would also replace the diaphragm, as once absorbed, you can't remove trace amounts of residues from rubber parts.
A Picture Of The Residues I've Found
The View Inside Of A Co2 Tank
The inside of an aluminum tank looks like this. It sort of looks like cast iron. It doesn't have the nice and shiny look that the outside does. It wasn't the easiest thing to photograph, so please excuse the quality.
Co2 Can Have A Significant Flavor Impact
Our senses are exceptionally sensitive to certain compounds, having the ability to smell or taste it in the orders of parts per million or billion. A good example of this is gas odorants in flavor infused water. The additives are in the orders of several PPM, yet we're able to detect these odors as a hint of peach or rose in water, even though the essential oils are in such a small amount that they're not easily measured.
I have a Sodastream machine, and there are only two main components to it. Water and CO2. I've modified it to accept bulk tank and at first it worked great. The line goes up to the machine unregulated and as temperatures change, CO2 condenses in the line. One time I left the main valve open and came back to use the machine a few days later, and immediately noticed a nasty chemical odor from the escaping CO2. Trying a sip, the carbonated water had a nasty chemical taste to it.
Suppose you mark a piece of paper with water soluble ink. It stays nice and clean even in humid air, but as soon as it gets wet, the ink will run. Same thing happens with CO2. CO2 is a good solvent for various oil-soluble components, and even if the gas is pure, it can leach out chemicals such as plasticizers and bring them out to surface, which can then get injected into beverage as an oil mist. Unlike water, CO2 becomes super critical above 88F, so instead of CO2 exiting only in a gaseous phase, it gains the ability to diffuse whatever is dissolved in it in a "fog" like fluid.
CO2 is handled as liquid in transfer, and it will dissolve oils used in process equipment such as the machining oil residues used in fittings and such.
I tested my 2.5 lb aluminum tank which has been filled numerous times, and sure enough, it contained quite a bit of residue. It was rinsed internally with petroleum ether which was collected. The rinse was collected and solvent was allowed to evaporate. The residue was allowed to dry at 180F, which gave off an odor of oil. The residue is a mixture of viscous oily liquid mixed with metal shavings, and it was weighed out to 61 milligrams on an analytical balance. Even if this was all soluble, it would only amount to 54 PPM in 2.5 lbs of CO2.
I believe that the difference between "beverage grade" and "industrial grade" CO2 is the filling process, and measures are taken to ensure PPM level residues with toxicity or objectionable flavor/odor are not introduced.
After experiencing this off-flavor, I think the most worthwhile solutions are an oil mist filter to remove the bulk of the impurities and an absorbent filter such as charcoal to grab the last bit of contaminants. Once the system has been contaminated, I would replace the plastic lines and clean the regulator, and disassemble and solvent-wash all metal parts. I would also replace the diaphragm, as once absorbed, you can't remove trace amounts of residues from rubber parts.
A Picture Of The Residues I've Found
The View Inside Of A Co2 Tank
The inside of an aluminum tank looks like this. It sort of looks like cast iron. It doesn't have the nice and shiny look that the outside does. It wasn't the easiest thing to photograph, so please excuse the quality.