RTV copper silicone as thermal compound?

[smile]

Hello, might be stupid question but given the price are these RTV silicone sealants that claim they have copper in them can be used as liquid thermal pad? Thermal adhesive silicone?

Supra Copper Gasket Maker - Permatex
Copper rtv silicone gasket maker -DoneDeal

many more brands, the MSDS does not mention copper maybe because it is no harmful? Or it's just marketing crap?

[wraper]

Liquid thermal pads do not cure like RTV silicones, they remain mushy. If you need adhesive in particular, RTV silicone can be used for that. You need neutral cure silicone to use anywhere near the electronics and most of them emit acetic acid when curing. Permatex appears to be fine. DoneDeal says "Sensor safe when cured" which is very suspicious wording, no datasheet too. Normal RTV silicone can be used as thermal interface if heat dissipation and thickness are not high. Generally you can expect about 0.2-0.3 W/mK thermal conductivity, although some are significantly worse. Don't use it between two wide pieces of metal as it will take forever to cure in the middle. Wide plastics are generally fine as they pass moisture to some extent. If you need high thermal conductivity, there are some special silicones. Do not expect any outstanding thermal conductivity unless there is exact number in the datasheet. If silicone say contains only 10% of copper, it won't improve thermal conductivity too much.

[smile]

Well the wmk is very unreliable.

ARCTIC G-1 - Thermal Compound. 4.3 wmk.. 6.3Eur for 9grams
No longer manufactured, the manufacturer must have unlearned how to make thermal glue in 2023. Not that it would be true 4.3wmk but was not bad in year 2011, at the price is sold it great too. Why Arctic no longer makes it they make no alternative is very strange.

Regarding RTG copper gasket makers - same thing how much copper nobody knows. I think better use copper powder etc. and RTV silicone or better two component silicone. But it's not available at small quantities

[wraper]

Filling Silicone with powder won't do the job. You can add only so much until it becomes unusable. If you say add 30% copper, it should be like 1mm thick layer of this concoction performing like 0.7mm of pure silicone. Which is not great at all.

[smile]

How so 30% only, if the particle size is 1/5microns then the copper content can be about 70% at least. This is based on my experience with aluminum powder and epoxy. The aluminum powder is coated with oil and will not cure in silicone sealant of any kind.

[wraper]

Even if you add 70%, the improvement will be still nowhere near real high performance thermal pads and adhesives.

[BrokenYugo]

One of the flavors of that Chinese electronics silicone you can get in small tubes from the usual places is supposed to be thermally conductive, "silicone thermal conductive adhesive" I'd bet on that sooner than "copper" RTV. AFAIK the "copper" in something like Permatex ultra copper or whatever they're calling it now is more a marketing thing to make you think of copper gaskets in an automotive context, heavy duty/high temp.

[rdl]

Yeah, I wonder, is copper metal actually better than the usual aluminum oxide/boron nitride, or is it just a marketing thing?

[jonpaul]

RTV silicone rubber may have corrosive as it cures.

The thermal conductivity is poor regardless of filling.

Avoid

j

[smile]

One of the flavors of that Chinese electronics silicone you can get in small tubes from the usual places is supposed to be thermally conductive, "silicone thermal conductive adhesive" I'd bet on that sooner than "copper" RTV. AFAIK the "copper" in something like Permatex ultra copper or whatever they're calling it now is more a marketing thing to make you think of copper gaskets in an automotive context, heavy duty/high temp.

I tested this myself, the chinese silicone vs. viktor reinz REINZOSIL. There is no difference. Seems like from china they sell you simple silicone.

[TimNJ]

I would think a thermal pad would generally be higher performing, but if for some reason, you need some sort of flexible liquid filling instead, something like this might be reasonable. Don't mind the bastard thermal conductivity unit - It equates to ~1W/m-k.

https://www.cotronics.com/catalog/09%20%204538.pdf

The viscosity is probably in the honey/molasses range which means it will still want to flow, and not stay in place like an RTV silicone might.

[thm_w]

One of the flavors of that Chinese electronics silicone you can get in small tubes from the usual places is supposed to be thermally conductive, "silicone thermal conductive adhesive" I'd bet on that sooner than "copper" RTV. AFAIK the "copper" in something like Permatex ultra copper or whatever they're calling it now is more a marketing thing to make you think of copper gaskets in an automotive context, heavy duty/high temp.

I tested this myself, the chinese silicone vs. viktor reinz REINZOSIL. There is no difference. Seems like from china they sell you simple silicone.

Have you compared to a real thermally conductive RTV though? What was the layer thickness?
Its not the same as regular white silicone as you can see the white powder/material in it. The two materials will separate given time. Not saying its actually any good though.

Found a few from 0.6 up to a max of 2.5W/mk:
0.6W/mK: https://www.aerospace-sealants.com/products/as-rtv-tc?variant=1260232528
1.8 W/mK: https://docs.rs-online.com/d8cf/0900766b80f8df6e.pdf
Normal RTV is 0.3 as noted above: https://www.sciencedirect.com/science/article/abs/pii/S001122759700146X

Kafuter K-5203k claims to be 1.6W/mK

[smile]

This is result of testing DIY thermal glue. I think it works good enough if it matches thermal pad from memory of GPU.
So no need need to go crazy buying expensive stuff. This is about 1mm thick

[smile]

NO replies wow.

[thm_w]

NO replies wow.

Your test is interesting but not conclusive at all.
You haven't even listed the temperature numbers, and you haven't compared two products at the same thickness in a typical use case (eg heat source -> TIM -> heat sink).

[smile]

I have provided thermal images of simultaneous tested 3 test samples of original msi gpu thermal pad, then silicone compound itself, then the sample with added aluminum. They are all 1mm think.

Did you read the thread? seem not.

[PlainName]

The thermal images don't say much to me. It would depend on when they were taken, the temperature of the iron, etc. For instance, if they are all at the iron temperature then they could be made of cardboard for all the image is showing.

I think you need some way of determining what it takes to keep the cold side some K below the hot side. Alternatively, timing how long it takes a large heatsink to get up to some specific temperature (obviously, the same hot/cold conditions for each test item).

[smile]

Iron temperature is 100C
All samples placed at once, time taken is 15sec.