Effect of Copper oxidation on thermal conductivity?

[Jimbo Mahoney]

Just been thinking about this after doing some googling and finding a paper detailing the decrease in thermal conductivity of copper when it oxidises...

Bearing in mind lapping is a popular technique, will this not increase the amount of copper oxide forming on the base of the block?

Is there any way to prevent the oxidation or reduce it? The only way I know is to coat the copper in nail varnish, but this is self defeating as far as thermal conducitivty goes...

Can Copper Oxide be beneficial in any way?

[BillA]

what paper ?

[Butcher]

You're not going to get much oxide in the TIM joint as the TIM itself will tend to exclude moisture (they're usually silicone based). Oxide on the rest of the block surface won't affect anything as it's not being used for heat conduction.

[Jimbo Mahoney]

Bill - a largely irrelevant paper here:

http://www.mri.psu.edu/conferences/s...hnson_3_10.pdf

If you read it, you'll realise it's unconnected with anything in the realms of watercooling, but let's face it, we are all here to reduce temperatures and improve the cooling, even if it's 0.5'C.

Well, I am anyway!

Butcher - Fair enough about the TIM joint, but what if the block has been left out unused for a period? In addition, the internal surface of the block, ie the part in contact with the coolant, will oxidise too.

I would like to see actual lower temperature data on the reduction in thermal conductivity with increased oxidation as it may turn out to be stupidly low.

It's just a thought.

[BillA]

for me 0.5° is HUGE, but I live here

that paper is concerned with sintering, extrapolation to 'other' things is questionable

oxidation includes the faint discoloration over time to the thick black crust that forms on copper when it is cherry red and exposed to oxygen (air)

so does the discoloration affect the thermal resistance of the TIM joint ?
based on my (relatively crude) testing I would say not,
if I cannot measure it I lump its effects with all the other uncontrolled variables associated with the procedure and component - see 'fouling' of HXs

Inchoerent is generating good values to 4 decimal places, perhaps he would be interested

[Jimbo Mahoney]

Bill - I guess 1'c for me is significant (anything less is immeasurable for me anyway).

I know the paper is unrelated almost, but it is true that copper oxide has a reduced thermal conductivity.

It's good to know it seems largely irrelevant.

Thanks.

[|kbn|]

Originally Posted by Jimbo Mahoney

Bill - I guess 1'c for me is significant (anything less is immeasurable for me anyway).

I know the paper is unrelated almost, but it is true that copper oxide has a reduced thermal conductivity.

It's good to know it seems largely irrelevant.

Thanks.

Not really an issue. It will only affect the surface, a very thin layer. It wont affect the inside structure of the copper unless it is exposed...
Between the cpu and block, the tim should stop oxide, so pollish the block first and you wont get problems from that..
The internal surface should have enough surface area to make any normal amount of oxide unnoticable even with the best test equipment...

Oxide is only an issue for heated blocks, IE Those which are soldered or braised..

However, what about annealed copper? I expect it could make quite a difference for machined o-ring sealed blocks..
Does machining a block in a miill acctually affect the hardness much?

[Susquehannock]

Machining shouldn't effect hardness as you are only removing metal. You need pressure (ie forging etc ...) to increase hardness in Copper.

The unregistered one is spot on about the deep red oxidation. I am a Machinist/Metalsmith
and have experienced this first hand just last night while designing my own heatsink
from scratch. And I do mean from scratch. Well, I didn't mine the Copper myself, but that's about it.

* smelting copper *