Stimmy

The idea ist't completely out of reach. in 2010, there was a cooler called "Danamics LMX Superleggera", which uses a liquid metal as a coolant and wasn't a commercial success. The alloy isn't gallium-based though, it's a sodium potassium (NaK) alloy. NaK is way cheaper than Galinstan, but it's also highly reactive, oxidizes and may catch fire in air, and explodes when it touches water. So it can only be used in hermetrically sealed systems, and filling a water cooling loop with NaK would be outright insane. Usually, electromagnetic pumps are used for NaK: A high current is flowing through the coolant while it is passing between two magnets. According to the right hand rule which every electrical engineer should know, a current flowing through a conductor within a magnetic field creates a mechanical force, which causes the coolant to move. This pump principle would also work for the much safer Galinstan, but since it has a lower electrical conductivity, it will be less efficient and require more power. I just think a Galinstan loop offers no real benefits when compared to water. With a good block and radiator design, the lower thermal conductivity of the water shouldn't be a big issue. The water is constantly mixed when passing through the block, making sure that every portion of the water comes very close to the cooling fins at some point. It also won't look that cool, the tubes will just look like they have solid metal rods in them. Also, considering Linus' frequent water colling spills, I'm not sure if putting a few kilograms of liquid metal in his hands would be wise . So in short I'd say: It may work, but it's expensive and the performance may even be worse than with water. A classical example ot "Not worth it".

Thanks for the answers! I've looked around a bit more, and found out that some liquid metal thermal compounds have a significantly higher thermal conductivity than Galinstan. According to the manufacturer, "Coollaboratory Liquid Pro" has a thermal conductivity of 80 W/mK, which would be two times as much as pure gallium. This thermal conductivity might even surpass that of the solder alloys for soldered CPUs (not really sure about this). As stated in the MSDS, Coollaboratory Liquid Pro contains silver, which may explain the higher thermal conductivity. So, using pure gallium would be quite pointless.

Hi guys, while surfing around on the web, I found out that pure gallium has a much higher thermal conductivity than the usual Galinstan liquid metal which contains gallium, indium, and tin. Galinstan has a thermal conductivity of 16.5 W/mK, pure gallium has 40.6 W/mK. It's also quite easy to apply, since it has a melting point of 30 °C and wets the surfaces just like liquid metal. If chip and IHS/heatsink are slighty pre-warmed, it can be applied lust like liquid metal. The only obvious disadvantage is that one has to warm the heatsink up to >30°C before one can remove it again. Also, the gallium melts and solidifies regularly during everyday use. This may put thermal expansion stress on the chip, and as an unlikely worst-case scenario might even crack it. I can also imagine that the liquid metal layer is so extremely thin that a higher thermal conductivity reduces the temps only by a neglible amount. Does anyone have experience with gallium as a thermal compound, or know a reason while using it would be a bad idea?