Galinstan as thermal paste

Viscosity ≈2'4x that of water (source: High Heat Flux, Single-Phase Microchannel and Minichannel Cooling with Water and Liquid Metal @ Electronics COOLING).

Of course, this stuff may not do nice on aluminum, yet for copper & brass could be very welcome. Maybe mixed a bit with cheap thermal goo if viscosity is a bit lacking…

Video I just found of a guy who made some liquid metal paste with recycled indium solder plus gallium:

Enjoy! :-)

Cheers ^:)

Interesting material I wasn’t aware of. Better keep it away from my flashlights, it might not be the best thermal paste option… :TIRED:

“…dicussion by Burton indicated that galinstan does react with copper, even at low temperatures.
Liquid gallium at ~30°C also slowly reacts with copper to leave the copper surface pitted.”

“Galinstan is corrosive against metals…”

“It is unsuitable to be used with Aluminum as it becomes extremely brittle in the presence of Galinstan.”

It’s always good to look for improvements, but a proper thermal paste joint is pretty good. With a 20mm MCPCB, a thermal paste thickness of 0.075mm, and thermal paste conductivity of 5 W/m*K, the thermal resistance of the thermal paste joint is 0.05 K/W. This could certainly be improved upon, but it is already small compared to other thermal resistances in the path such as the LED junction to solder point resistance (1-3 K/W).

Gallium based thermal paste has been commercially available for quite a while.

I used Coollaboratory Ultra in my last PC build (on a nickel plated CPU cooler) with very good results.

Unfortunately, gallium absolutely cannot be used on aluminium , and apparently it’s not wise to use it on copper either (although the reaction isn’t anywhere near as dramatic as it with aluminium). So that pretty much rules it out for flashlight builds…

Interestingly though, when I went to the Coollaboratory site to get that link, I found they’re now selling a product made from silicone and copper:

Only possible downside for flashlight use is that it’s likely electrically conductive, but that’s not a big deal if you’re careful.
(I once used ‘Permatex copper anti-seize’ in a pinch on a laptop cpu cooler, and it hasn’t exploded in the 2 years since)

Seems to work well, but unless you’re chasing every last degree it may not be worth the effort over more readily available thermal pastes:

Isn’t a silver based thermal paste a better solution? Like Arctic Silver 5? It already has a higher thermal conductivity than copper and doesn’t react with typical flashlight materials.

“Silver based” ≠ silver. You could add 1% of micronized silver to a paste and claim it to be silver based, though I doubt it would have any significant impact on its performance.

Source: Thermal Paste Roundup 2015 – 47 products tested with air-cooling and liquid nitrogen (LN2)



Coollaboratory Liquid Copper seems to be the king of conventional pastes.

Cheers ^:)

160W load though. Divide all the results by 10 and the differences aren’t quite as significant.

You can use tin-zinc, tin-zinc-gallium or tin-indium solder to just solder copper to aluminium without any major hassle (outside of obtaining or mixing the solder itself and the aluminium-special flux, though you can get by without one). The latter two compositions even have lower melting point than the common tin-lead solder. And the tin-zinc-gallim should bond easily to aluminium surface.

If you’re talking about the charts in the previous post then where do you see 160W figure? I don’t think modern CPU’s simply can take that much power for any prolonged period of time.
And shouldn’t the surface area of their hood be bigger than a 20mm LED MCPCB?

Well, right now I have some special aluminium soldering flux in the mail; check out the thread Aluminum soldering fluxes: The Way It's Meant to be Soldered :D for related info. I'm yet to decide what kind of solder to use among Sn63/Pb37 and Sn99'3/Cu0'7; I'm somehow leaning towards the lead free alloy, as it is said to work better on aluminium, though I don't know why.

Cheers ^:)