That might be looking at things a bit too simplistically. The way things get certified by BAM or others is mostly as a unit, not as a single component. That’s because how the parts are used is often much more important than what material they’re made of.
Oxygen fires are never the result of a single failure, but a chain of things going wrong. Start with slight contamination on a polymer surface, couple that with adiabatic heating from rapid pressurisation, and add a thin or porous piece of polymer nearby and you’ve got the beginning of an incident. If that happens to be close to a thin metal spring, it might be the next in line to catch fire, and so on it goes.
Zytel is a proprietary blend of a polyamide, meaning it belongs to the same class as nylon, as rjack pointed out. Nylons do not exhibit great properties for oxygen service. Nylon 6/6, for example, has a comparatively low auto-ignition temperature (~240°C in 103bar oxygen), relatively high heat of combustion (~32,000J/g), and a poor oxygen index (~26%).
Zytel comes in many different varieties. Most often you’ll find Zytel 42 or Zytel 101 used as seat materials. Most Zytel blends fall into the same ballpark as other nylon or polyamide blends. Zytel 42, for example, has an auto-ignition temperature of ~180°C in 103bar oxygen, produces ~34,000J/g of heat, and has an oxygen index of around 34%. By any measure, these numbers are poor.
Yet many manufacturers specifically use Zytel blends in their oxygen-compatible product lines. Sherwood uses Zytel 101 in their medical post valves (KVAB and KVAC series). Poseidon uses Zytel as a seat, though I don’t know which blend, in their oxygen-capable regulators. Aqualung uses Zytel in the gaskets of the burst disc assemblies on their combat swimmer bailout cylinders. Cavagna uses a polyamide as the seat material in their oxygen valves. And the list goes on.
So why are polyamides with such poor oxygen compatibility characteristics so often used in oxygen service? The answer lies in how and where the parts are placed within the assembly. Leaving aside their oxygen characteristics, Zytel has excellent mechanical properties and is relatively stable in static oxygen environments.
Take the Sherwood KVAB medical post valve, for example:
View attachment 899107
If you follow the trail of gas through the valve, it becomes clear that the Zytel seat surface isn’t where adiabatic heating would occur. Nor is it directly in the gas path when the cylinder is being filled, meaning particle impacts are of no significance. Particle impacts could be an issue when the cylinder is emptied, as the seat surface is then directly in the gas path. However, thanks to the Joule–Thomson effect, the gas cools down during expansion, which significantly reduces the risk. The whole valve is incredibly well-engineered, keeping the polyamide seat away from danger zones. Where adiabatic heating could occur, under the packing nut, you’ll find PTFE, Viton, and copper as sealing materials. Even the spring is cleverly hidden for the most part within the seat.
A similar logic applies to the Poseidon seat when it’s inside an Xstream. The alloy ball takes most of the “heat,” so to speak. It isn't the Zytel that would take is exposed to the worst.
Now contrast that with this Z-Valve:
View attachment 899108
The valve’s seat is directly exposed to the gas flow when the cylinder is filled. Particle impacts and adiabatic heating can both hit the seat surface head-on. No matter what super-material you use, the engineering of this valve almost rules it out from ever being oxygen-compatible.
I always find it very helpful to mentally map the gas flow and follow it step by step to identify critical areas in terms of oxygen safety.
I write all this because I found the notion that the part is certified for 300 bar of oxygen too simplistic. I can almost guarantee that the part would ignite easily if exposed to rapid adiabatic heating and a bit of contamination. The point is: the way it’s placed in a Poseidon regulator, or the way the Zytel is used in a Sherwood KVAB valve, almost completely removes the part from the genuinely dangerous areas of the assembly. The whole Poseidon, as an assembly, is resistant to 300 bar of oxygen, but that doesn't mean that all its individual components, when removed from it, will be resistant as well.
The most resistant seat materials for oxygen service are mostly halogenated, but they are seldom used in breathing scenarios, or outright banned in some countries (France) due to the risk of carbonyl fluoride or phosgene developing even under partial combustion. Polyamides and other hydrocarbons, on the other hand, develop "only" carbon monoxide as a toxic gas, but carbonyl fluoride is at least ten times more deadly, while phosgene is probably even worse, being one of the gases used in World War I.
I have no idea what your adapter looks like, so I can't give you specific recommendations on whether the part is suitable or not. I'm not saying that it isn't, or that it's not a good idea, but it might be worth to stop for a minute and think about. I would suggest that if you substituted the part, you mentally map the gas flow and ensure it isn't exposed to any dangerous zones. In the end, every single polymer today will ignite happily if the conditions are just right and it is up to the engineer to develop the assembly clever enough that it is removed from danger zones.
I apologize for the long post, but felt the distinction was worth pointing out.
Back to your original question: I know of no threadlocker that is certified for 100% oxygen at high pressure. I’d actually be very surprised if such a product existed, as the market would be tiny at best. I’d prefer a mechanical locking method. If you can fit a washer, then split washers, Belleville washers, wave washers, or Nord-Lock washers could be great alternatives. If that’s not feasible, gentle staking can work well to lock the screw in place, as happy suggested. It doesn’t take much staking to stop a screw from turning. Ever so slightly and gently bending the screw can be feasible as well.
PTFE tape wouldn’t be my first choice either, as there's still a risk the screw could come undone. If that happens, you might end up with thin strands of PTFE tape floating in the oxygen path, and those are perfect fuel for combustion.
