When CO2 reacts whith the oil (or with H2O in the oil) (CO2 systems- ie. centrifugal compressor) it creates acids in the oil. At what TAN level do you think an POE oil in this case should be dissaproved? We usually see very high TAN values (often up to 3-4)

Any comments on this?

You should never use POE with CO2 (or ammonia for that matter). The oil will decompose very quickly from reaction of the ester group with the acidity from CO2 and moisture in the system. I would expect this system to have major problems with carboxylic acid salts deposited in the high side of the system up to the expansion device.

If you have completely removed the CO2 prior to testing and are getting TAN values of 3 or 4, the POE is finished. POE decomposition is autocatalytic, which means decomposition speeds up additional decomposition. TANs above 0.2 are a problem for POE stability. TANs below 0.05 promote good POE stability

PAG oils are preferred for use with CO2 because of the stability of the ether linkage to acid attack.

The oil is usually an POE especially formulated for CO2 systems (Fuchs Reniso C55E), but I now have a suspicion to what might be wrong. We do not remove the CO2 before TAN measuring (we do remove it before measuring viscosity).

I will carry out some tests on some of our earlier (high) results and see what shows up after degassing.

Would you say all TAN measurements should be done on degassed samples?
Thanks for all input.

Dissolved CO2 will definitely cause the TAN to increase. You may want to be careful when removing the CO2 to prevent any other volatile acids (formic acid, HCl etc.) from being pulled off the sample as well.

No disrespect to Fuchs formulating ability, but POE and CO2 is a recipie for eventual system problems. The only way to protect the POE from decomposition by acid attack is to manage the temperatures in the system to prevent chemical reaction. CO2 tends to have higher discharge temperatures than traditional refrigerants so this is no easy task.

CO2 is such a small molecule, you will not be able to fully protect the ester linkage from attack by using steric hindrance from acid chain branching. Acid scavenger additives (epoxides etc.) that can be used to protect against autocatalytic attack with traditional refrigerants will be consumed by CO2 itself. POEs are just not a good choice in either acidic or caustic environments.

If you are curious and have the suitable glassware, you could run a hydroxyl number (by reaction with anyhdride) on the POE to look for alcohol groups formed from POE hydrolysis. Normally, the hydroxyl number is used to measure the degree of conversion of the POE during manufacture. Expect to see hydroxyl numbers less than 5 on high quality POEs. If the hydroxyl number is going up, the POE is coming apart.