Elastomers in CO2?

so…who has the skinny on leaking valves when using scCO2?

I noticed I had a leaking valve stem, tore the valve down, and failure mode seems to be disintegration of the Viton o-ring.


OEM reports this as a “wear item”.

they don’t appear to be wrong.

https://www.osti.gov/servlets/purl/1592964

wondering how folks skilled in the art are dealing with this?

just wait till they leak and swap in a freshly rebuilt one from your “spares” collection?

proactively swap them out on a schedule?

Are there elastomers that work better here that the “ultra low temperature Fluorocarbon” swagelok installs in this valve?

@JonaaronbrayAzoth?

4 Likes

Unfortunately, just about any elastomer will degrade over time when exposed to supercritical CO2 due to cyclic swelling/contraction of the material and damage caused by “rapid” depressurization effects. That said, some materials are better than others.

Whenever possible, avoid using NBR materials (such as Buna-N) if they will come in contact with SCO2 as they will tend to fail rather quickly (especially at high temperatures and if exposed to rapid depressurization). For SCO2 exposed seals, I tend to use EPDM as it lasts substantially longer than most seal materials (anywhere from 10x to 100x longer), that said, like all elastomers, they will eventually fail.

Having used Swagelok in the past, I know they offer EPDM seals as an alternative on most of their valves, you just have to ask when ordering.

If you want to delve deeper, here’s a couple of papers on the topic:
Solubility of Fluorinated Polymers in Supercritical Carbon Dioxide.pdf (5.3 MB)
The_mechanical_properties_of_elastomers_in_high-pr.pdf (557.3 KB)
Modelling the Permeability of Nitrile Rubber.pdf (7.3 MB)
Rapid Gas Decompression Performance of elastomers.pdf (1.3 MB)

12 Likes

Thank you.

I’ll get to the recommended reading tonight or tomorrow.

Don’t think swagelok is gonna sell me just the three o-rings per valve, so I’m pricing out seat & seal kits right now. And tracking down the part number for the other style of valve I’ve got on here…

3 Likes

Yep, @Kork hit the nail on the head. Unfortunately, there really is no suitable elastomer for use with SCO2, it just diffuses too easily into material and causes them to become bubble wrap. You can look into “explosive decompression resistant” materials, that rabbit hole lead me into several thousand dollars in seals which ultimately last only 50-100% longer than the 4 cent buna ones. Polyurethane works reasonably well as far as run of the mill materials.

Swagelok seems to be in denial about this. Their SK series needle valves or butech are really the only options I found that last. Ball valves are out.

4 Likes

I’m getting less and less fond of this machine…

Lulz. I updated my post, it’s the SK series.

Teflon tape can be a bit tricky. Fo joints that don’t get broken, loctite 567 with the correct activator and a long cure is the way to go for SCO2. You need a really tidy technician though to not get globs in the system (it doesn’t dissolve in the stuff at all but just mechanically nobody wants pipe dope in their dope dope)

Edit for further clarity: the activator is loctite SF 7471

2 Likes

Doesn’t look like the PTFE likes SCCO2 duty either…

#15/16/17 from exploded view below

Edit: yep, not unexpected if I’ve done my reading…

2 Likes

The PTFE should be fine, usually when that fails it’s thermally or mechanically related. I’ve had zero issues with it being compromised by SCO2

Maybe I’m missing it but that PTFE seal looks 5x5 to me. That o ring looks nibbled to fuck though

2 Likes

From article shared in first post. And why I said “not unexpected”…but I have no business interpreting those graphs :shushing_face:

2 Likes

I also have no experience looking at FTIR. The bumps look pretty similar to me, and that level of technical understanding has gotten me farther than a lot of paid “analytical chemists” so far. I will say, I’ve made a lot of things die with SCO2 and I think the only time we killed PTFE was when we tried to use it for SAE o-ring boss style seals (fun fact: that dog don’t hunt)

2 Likes

I haven’t poked at it hard, but it looks to be “flaking”, something transferred to my fingers…

As for the oring, it IS clearly split, rather than simply disintegrating. Too much wrench might be the problem…possibly while frozen. there was was evidence of hash laden solvent leakage at that seal.

This right here. I always ask the vendor what their normal “wear” schedule is based on hours of use. Then I cut that in half for my first change out to see how much wear there was. If everything looks good still - then I go back up to their limit for the next change out. And then at that change out, if everything looks good, I increase by 50% of the time until things are about 75% “worn”. Then I call it good and my updated PM continues for all systems/valves that are the same style and on similar systems.

They have been super helpful to me in this department and been able to get me things shipped on three continents without any hiccups. I’d go ahead and ask them - they might even already have a gasket kit just for this - I know they do for some valves I use of theirs on my steam system. And I also know that they have “PM” kits for their valves which include a nice set of “spares” if you have to get something quickly. If you haven’t already asked for their reliability centered maintenance plan - ask them for it. Its pretty good and is specific for each valve. <3

I’ve used Swag in all kinds of situations, high and low pressure, temperature - halogenated solvents, steam, sulfuric acid piping. Just give their tech group a poke and I’m sure they will come through.

Whenever I have a valve/seal/instrument need that is more specialized than something I’ve used before I reach out to Paul at TruFlow. Even if he doesn’t have exactly what I need he’s connected to all the other people who build these things and makes the connection in a friendly and productive manner. Freaking love this guy.

Those IR fingers are super different and are basically showing that there is a pretty vigorous series of chemical reactions happening and that those reactions continue to happen as usage progresses. Pretty cool, IMO - but also obviously a wear part. hehe. :smiley:

I’ve definitely seen this flaking too. Its like the reactions happen along micro fissures in the material and then you get chunks falling off, like tire separation. One of the things I hate to see happen, cause then I have to do a batch investigation to make sure we didn’t get any of those incidental product contact parts shedding into our material. Freaking hate those investigations!

2 Likes

NITRILE (BUNA-N) 90 DUROMETER

Only 90 and above durometer rubbers are resistant to Co2 impregnation.

3 Likes

I’ve found higher durometer elastomers to fail more severely if the depressurization is rapid after long cycles. Sure, they take longer to saturate, but if you pull the rug out they sure do bubble wrap like nothing else

1 Like

I run 36" orings of nitrile buna-n 90 that expand slightly after a few cycles but can be run for months after that without issues.

High supercritical pressures might effect the rubber different but those pressures are of minimal use in the extraction cannabinoids so not really of a concern.

I wouldn’t suggest anyone use any rubber on less than 90 hardness on any co2 system. The 90 has never depressurized a vessel on me but 70 has on many occasions.

The problem is that elastomers necessarily have void space in the material. Avoiding gaskets and o rings in sections of the plumbing that experience rapid decompression and the accompanying temp swings is your best bet. You can also play with some different fluorinated compounds and see if they hold up to the conditions of various points in the machine. Avoid over tightening, and use smaller cross section diameter o rings where you can. You can also soak your gaskets in mineral oil before installation for a day or so if it’s not an “oh fuck it’s leaking” swap. The oil impregnates the material and reduces void space where co2 can penetrate.

2 Likes