Thermal contact resistance considerations - indium, conduction, TEC, 1

[zdunolPOL]

Hello,

we're facing a problem at work regarding thermal contact resistance. We use water heat exchanger to take away heat generated by TEC (Peltier module), about 100W. So the question is basically how should we contact TEC with heatsink so Tec's hotside temperature is the lowest. We tries pure indium sheet (200 um thick) but several problems arise. Firstly Thermal contact resistance (Rc) depends on pressure, second thing is the oxidation of indium with oxygen from air, next thing is Indium's cold welding capabilities, wettability, boiling ponit.

So we did some measurements, we put TEC, indium and heatsink together, pressed it and heated up up to 80degC and left compressed ( about 100kg/36 cm^2 = 140 PSI) and left for a night (heating up was conducted in vacuum). I runned the thermal tests and the results were terrible, I deattached everything from each other and indium was covered with a lot of white residue - it was indium's oxide. So right now we will do the same thing but with no oxygen - in enclosure filled with nitrogen and see what happens.

From thermal point of view i dont think that layer of indium's oxygen would matter in heat transfer due to its low thickness (100 Angstroms max?), but I am more concerned (and more ignorant) about all the mechanical phenomena like contact with surfaces, cold welding (heatsink is made from copper).

so my question is - did anybody tried (and was succesfull) with using indium as thermal interface material? Do you have any other materials which you would recommend (the requirements are pretty stiff - applicable in vacuum, operational temperature about -50degC, absolutely no pumping out or drying out, very resilitent to thermal cycling)? What do you think about this issue with indium's oxide? Do you think we should go above indium boiling point during the assembly?

There is a company - indium dot com which uses indium and its alloys for thermal interface materials, have any of you used their product maybe?

Kind regards,
PZ

 

[StoneCold]

Why not use a heat transfer paste. Like Thermon. Bake it until it is dry then put it under vacuum.

http://www.thermon.com/us/heattransfer.aspx

Regards
Stonecold

 

[Compositepro]

Thermal pastes work by eliminating the air gap between interfaces. It is important to keep in mind that any material added in series to the heat flow path can only add resistance to heat flow. Even copper, which has excellent thermal conductivity, has a resistance to heat flow.

Thermal pastes have much lower thermal resistance than air but their thermal conductivity is much less than most metals. I am not very familiar with the behavior of indium, but if it oxidizes like aluminum foil, then the oxide layers will prevent the indium from wetting the surfaces (required to get rid of air gap) and the oxide will act as a spacer.

It seems you should simulate a soldering process using a flux to remove the oxide layer. It would also perhaps be better to start with a bead of indium that can spread and displace air and flux, rather than using foil. There are solder pastes. Perhaps someone makes one with indium or similar.

Also, are the surfaces but extremely smooth and flat, so that the gap is small?

 

[zdunolPOL]

ok, so firstly thanks for engaging in the dialogue guys (this forum seems great), now

@StoneCold
We decided to go with indium because it is good for vacuum and low temperatures. The only grease I could find that is guaranteed to work in -50degC was arcticsilver 5 (im in europe, maybe you guys have better greases, I will look for them) so this is the first issue with greases. Moreover as far as i know greases tend to pump out, degass and spill - all of theese are unacceptable in our product. The product itself will move and rotate, TIM will be under vacuum and electronics attached is pretty expensive . I have no idea how baking the grease works but i suspect it should get rid off all the violatile stuff, and there is another constraint - we cannot heat up the electronics above 40degC so, wouldn't baked thermal grease just become a powder and wouldn't this powder generate high contact resistance and we would go back to square one ?

@Compositepro
Yeah this is exactly what i suspect - I'm afraid that indium oxide give raise to high contact resistance, so next thing i will do is attach all the stuff toghether with unoxidized indium and will give an update. I will talk with the guys who know more about soldering maybe it is a good idea, but we are talking here about 62x62 mm surfaces and TEC max temperature is 150degC so I have some doubts. The thing about oxide layer is that i can easily get rid off it by putting the foild into HCl acid, it gets rid off them really nicely. Then I just literally put the foil on the copper heatsink and it sticks really well ( I think this process is called cold welding), I am more concerned about the interface between TEC and Indium (TEC is made of ceramics). I cannot recall parameters of the surfaces but we didn't polish them with any high tec equipments so I definitely don't think they are extremely flat and without waviness. Indium's surface is definitely not flat.

To sum up
I think the key is to eliminate air (gas) gaps between surfaces and also make indium wet the surfaces.
Wetting copper with indium seems to work well, indium does not stick to the ceramics though.
We can fight with air gaps by polishing and assembling the stuff under vacuum.
We can get rid off the oxides with acid and asembling things in Nitrogen/Vaccum.
We can go above indium's melting point.

I am also thinking about this one thing - let's say that we assemble the stuff in 20degC, there are some gas gaps between surfaces, then we heat things up to facilitate wetting, so shouldn't the bubbles expand and if there is oxygen in there, oxidize more surface? Also expansion would rise pressure between stacked elements.. Maybe I should assemble stuff in hot nitrogen, or in hot, high pressure nitrogen?

 

[zdunolPOL]

*hot low pressure nitrogen at the end (editing doesn't work)

 

[lilliput1]

-50C? How can it be cooler than the water (can't be below 0C) cooling it?

 

[IRstuff]

there's this:

http://www.apiezon.com/products/vacuum-greases/h-grease

http://www.aremco.com/news-item/hea...illed-vacuum-compatible-grease-now-available/

http://www.intlvac.com/products/ion...ssories/thermal-management/mung-thermal-paste

http://www.lesker.com/newweb/deposition_materials/depositionmaterials_targetbonding_silverpaste.cfm

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529

 

[zdunolPOL]

@lilliput1

TEC's hotside touches liquid cooling block, its temperature is above 0degC (I'm guessing around 60 with bad thermal contact with the heatsink)

@irstuff thanks for help!

Apiezon H Grease: operates in range -10 to +something -<won't work we need -50degC

Heat-Away™ 641-EV looks really nice, probably will order on monday!!

MUNG THERMAL PASTE also looks good but there is no info about temperature range, will ask the producent and give you guys an update (or is it illegal to do that?)

 

[IRstuff]

You needed to read further down the page:

"Apiezon H grease has been approved by NASA as a material suitable for lubricating the gold-plated threads of small variable capacitors required to operate under high vacuum from -65°C to +125°C to prevent galling."

The -10ºC statement was referring to "optimum" consistency, whatever that might be.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529

 

[zdunolPOL]

oh, thanks, I just looked briefly on those (it's midnight here in Poland), thanks once again, I don't know why but i couldn't find those kind of products on my own. Maybe it's my googling skills, definitely lack of experience too.

 

[EdStainless]

Use a vacuum grade grease loaded with sub-micron silver particle, but before you apply it lapp the surfaces extremely flat, like in optical grade flat. Then add only enough grease to exclude the air.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[zdunolPOL]

Yeah well it turns out the TEC was malfunctioning, I don't think we can use grease due to the reliability (pumping out), or maybe we should use some kind of gasket in between the surfaces? I'm looking into low temp. melting alloys like the ones indium corp. sells.

 

[IRstuff]

Grease in general is a bad thing in vacuum, but vacuum-rated grease is intended for vacuum, which means that it doesn't outgas and may even help the vacuum by gettering other contaminants. The vapor pressure of the first grease listed above is 1.7E-9 Torr at 20ºC

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529