Cooling of IC's: Paste, pad, pudding or nothing?

[eTobey]

Since i havent found much satisfying information about thermal conductivity expecially between pad and paste, i thought i just try it for my self. The information i found in the webs were even the opposite which made it more confusing...

Anyone that has got some good links are welcome to post them here. Some proper terms for further research are also welcome.

The resistor is this one (2 Ohm with 2.25A going through):
https://www.mouser.de/datasheet/2/414/OBSOLETE-MHP20-1671194.pdf

Measurements are of course not scientific, and should be seen only a rough reference. The only things that were reasonably stable, were the thickness, ambient temperature and the power that the resistor dissipated. The temperature was measured with a GUID PC210 IR-Cam. The pads are from an ebike motor, and the paste is "Silicone cooler paste - Heatgrease20" from Startech.com.

What i find most interesting, is, that olive oil does a equal good job as the paste, but does drop quickly with thickness.

[Siwastaja]

What i find most interesting, is, that olive oil does a equal good job as the paste, but does drop quickly with thickness.

Remember that the ages old wisdom that paste is only meant to fill microscopic gaps is still true. Pastes have pretty poor thermal conductivity - datasheets of proper products do show it. Not much better than any random skin lotion.

This is because paste won't be mechanically strong to offer any kind of reliable electrical insulation even in thicker layer, so there is no point of making it thick. Therefore, there is no incentive to make it thermally more conductive. In very thin layers, what it is is enough.

This is totally different with thermal pads and especially "gap pads". Unlike grease they respond to the market need of providing electrical insulation, and part of that is adding thickness to safeguard against tiny dirt particles etc. puncturing through the pad, compromising the electrical safety. Gap pads specifically are designed to fill in gaps e.g. when heatsinking different height parts to one heatsink.

Now with layers that are intentionaly tens if not hundreds of times thicker than any thermal paste layer, it is worth engineering effort to get into thermal conductivity range of 2 - 20 W/mK which still sucks compared to pure aluminum but isn't that bad at all for something which is electrically insulating and soft.

Paste can be 0.2-0.5 W/mK and everybody's happy with it. Except those who try to fill gaps with it against the ages old wisdom.

[SiliconWizard]

Well, oil does transfer heat and will fill up the tiny gaps. Not as effective as water for heat transfer, but using water here wouldn't be very practical, at least not in direct contact.
Unfortunately, olive oil will "dry up" rather quickly, especially when heated, and will make a big mess. Also, the device while operating may smell funky.