Boris,

Well:

http://www.dtic.mil/dtic/tr/fulltext/u2/a191755.pdf

My feeling is that any anodizing (which creates a porous structure in the outer skin of the metal) will be a poorer contact than a method which plates or converts.

Ah, I've seen a similar number before. The catch is that they are using a mineral oil fill between the two metal surfaces, so in effect this number is the theoretical limit for the case where the contact surfaces are perfectly smooth.

Working under vacuum with environmental controls don't allow greases or lubricants, I won't be able to achieve conductivity even close to that.

Thank you for that data point!

If you look at the "List of thermal conductivities" on Wikipedia, you will find that it is measured in watts per linear metre per kelvin. As any alO2 coating is generally less than a few microns, a quick calculation will show that it will make virtually no difference what you use.

Be careful with chromate coatings, these are highly toxic and are generally only used to improve electrical conductivity on critical aerospace applications.

Nickel plating can peel off aluminium at cryo temperatures due to differential expansion. Usually only used with AlSi type alloys where the CTE's are well matched.

The pressure will of itself make no real difference either, more important will be the integrity of the joint in creating a thermal bond (need to be mated to a high standard) to maximise heat flow.

A better joint for pipework comes from cryofit type concentric couplings which are fitted using liquid nitrogen. Commonly used in oxygen supply systems.

There is a process called 'hard anodising' which avoids the sponginess that James alludes to. This is widely used in aerospace applications.

Edit

I have just seen your response to James and that suggests you are joining two flat surfaces. I have used a fine layer of bismuth to join dissimilar metals. Apply it to one face and warm it to melting point (270C) preferably in an argon atmosphere (your welding shop will have bottles of argon).Bolt together while warm. Bismuth is a poor solder, but an excellent filler. This will make an excellent thermal joint. I used it as a high standard pressure seal.

Since you need to estimate thermal conductivity through a joint, at a very low temperature, you should consider the surface topology of joining pieces of metal, It may affect the measurements drastically.

Thanks for the feedback, everyone!

The joint is quite large, about 10" in diameter, and would have to be taken apart many times, so I'd rather not use indium foil. Though I do use it elsewhere, in places where total area is fairly small and the two parts are rarely, if ever, taken apart.

For this joint, the material on both sides is Aluminum 6061.

Ian - you may be right about electroless nickel plating. For cryo applications, I think it requires an intermediate layer. I am not familiar with details of the application process, but we certainly have fairly large parts that are electroless-nickel-plated and do not suffer degradation after thermal cycling.

I am somewhat surprised at the assertion that joint pressure does not affect conductivity, or that a thin low-conductivity coating won't affect overall conductivity. See, for example, this paper: http://matthewwturner.com/uah/IPT2008_summer/baselines/LOW%20Files/Thermal/Spacecraft%20Thermal%20Control%20Handbook/16.pdf

Thermal conductivity of electroless nickel is about 8 W / m K, while for anodized aluminum it's less than 1 W / m K:

http://www.techmetals.com/wp-content/uploads/2015/03/Engineering-Properties-of-EN-Coatings.pdf

Article Thermal Conductivity of Hard Anodized Coatings on Aluminum

(Hard anodizing is used for vacuum systems because it outgasses less, I think.)

James, could you tell me more about graphoil? Do you have a distributor for this material that you like?

Boris,

I last used Graphoil in '96, '97 when I had to seal against gaseous methane at 90K or so. We acquired some from Rutherford Appleton Labs, but it appears to be a commodity item now:

https://www.ebay.com/itm/Flexible-Graphite-Foil-Graphoil-Gasket-Sheet-Plate-250-200-1mm-/271657373246

Mind, it's not vanishingly thin - fraction of a mm IIRC. It's possible that the anodizing will introduce as much thermal resistance as the addition of a ductile gasket.

Hi Boris

I guess my answers are all about expectations.

If you have a system where the temperature differential is say 200K then the heat dissipation through 5 microns of anodising is about the same as the same as through a typical electroless nickel coating of 50 microns.

Given your parameters then roughly 96watts of heat will be transferred by either system, if the nickel plating is flawless!

I don't know about your manufacturing capabilities, but a deposited coating like electroless nickel can have enough issues to reduce the thermal transfer coefficient by roughly 25%; inclusions, bubbles, thickness variation etc. However, safe to say the difference will be tiny given the overall system transfer rate.

Pressure per-se makes no real difference to heat transfer. Hard coatings like nickel or alo2 will not compress much.

Of much more importance is the conformity of the joint. How closely the parts fit together. Tightening bolts simply squashes the parts together more closely and so improves the contact between mating surfaces. This is the substance of the matthew Turner paper. However the same bolts / pressure can easily induce distortion and significantly depracate the mating of the surfaces.

I don't know if you have ever replaced a car cylinder head but induced distortion through applying too much pressure (over-tightening the bolts - or even tightening them in the wrong sequence) can destroy a cars performance.

The point about the bismuth, or the graphoil mentioned by James is to improve the conformity of the parts. Fill in all those tiny voids that exist between any machined surfaces with a heat conducting material and you improve the system enormously.

Thanks for the additional answers, guys.

It's a good point conformity. I've never replaced a car cylinder, but tightening in a gradual sequence is certainly familiar to me from working on vacuum systems. You want the two flanges come together smoothy and evenly.

Gold probably accomplishes something similar to bismuth and indium foil, since it presents a softer surface that will conform more readily under ordinary pressure.

Something that concerns me is compatibility with clean-room environment. Do you guys know if graphoil is acceptable in such places?