Liquid Metal Cooling a PHONE!?!?!

It's a really bad idea to put Liquid Metal in your phone - so we're going to try it.

Buy Liquid Metal:

Intro Screen Music Credit:
Title: Laszlo - Supernova

[JerryRigEverything]

Dang man! Sweet project.
I would just recommend always disconnecting the battery first, before anything else. (And plugging it in last) probably not a big deal. But better safe than sorry. Phones are expensive.

[tinglydingle]

"That's a super risky, maybe even bad idea. So we're gonna show you how to do it!"

Linus Tech Tips in a nutshell.

[myjy]

"After we tell you to-"

*10 sec >>>

"Through our link be-"

*10 sec >>>

[Arkane.]

“You feel it getting warm and hard and harder”... Linus 2018

[SamaelSmith]

whats next? "Liquid Metal Cooling a SMARTWATCH!?!?!"

[victorwei]

"Liquid metal, it's a hell of a drug you know" - Linus 2018

[Nick]

Do I hear liquid metal consoles next...?

[Minitomate]

"You notice something getting kind of hot in your pants."
Me: **Note 7 flashbacks coming off again to my mind.**

[rodrigokiefe]

loved the ending btw "do you think the editor is going to go through this fottage and listen to it??" xD 8:36

[GENKI_INU]

I never understood the logic behind the CPU power/speed wars on mobile phones...

What's the point of having flagship phones that throttle down to the speed of regular phones within a few minutes of load/gaming due to overheating? Are we going to see fans strapped to our smartphones in the future? Why not just make something that performs efficiently and sustainably, but still fast?

The smartphone industry should have a main focus on efficiency and battery life instead, since we already have enough processing power on them to do pretty much anything you'd want on a phone anyway. Pure speed shouldn't come before efficiency in this case.

[nardisdassler]

I always feel ripped off when the sponsor isn't Tunnel Bear :(

[Major_Mason]

Don't make the squirty mistake like last time...

[justinluce]

Next video: Liquid Metal in a human???

[elster]

Now finally I can play flappy bird at a constant 100 FPS! Thank you, liquid metal!

[pyry]

Always disconnect the battery ribbon first /JerryRigEverything

[marcun]

They should test the phone with only better thermal paste. This would show how much liquid metal helped.

[preston]

0:23








That's all I'm gonna say

[ekdromoi]

I did more reading about this stuff since I'm interested in using LM for my own laptop, and it seems corrosion is not out of the norm.

Typically when LM is used under the IHS of a desktop CPU (often necessary for intel chips since they bond the IHS to silicon with toothpaste instead of solder) the results are pretty good even after a year. The IHS (which is made out of copper, then nickel plated) does get stained by the LM but nowhere near what happened to you. This is because your contact (in a laptop) is silicon die <-> LM <-> copper heatsink. In the desktop scenario, LM is often applied underneath the IHS so the contact is silicon die <-> LM <-> nickel plating <-> copper.

To be even more clear, the key differences in your application of LM is that
1. Your application was silicon-copper instead of being silicon-nickel as you would find in the 'delidding' of a desktop CPU
2. Your application was not insulated against air (read: oxygen)

1 is important because the electrode potential of gallium is -0.53V, nickel -0.24V, and copper +0.34V. Obviously, all liquid metals are made out of gallium (because Ga is liquid at room temp, and is also nontoxic unlike mercury) plus small percentages of other metals to reduce the melting point. When gallium is in contact with pure copper, the differences in electrode potential favors the gallium and copper to alloy, which will eventually consume the gallium completely[1] (Ga + Cu → CuGa2 [67%] + Cu3Ga [11%], and both products are stable until 175C)[1][2].
The liquid metal will literally into the copper until the gallium is gone, which causes the copper to turn silver-ish. The non-gallium components (indium, tin) of the liquid metal[3], which are solid at room temperature, get left behind - and that stuff is hardened deposit that you were trying to scrub off the heatsink.
Note that at higher temperatures, the reaction between gallium and the copper heatsink only gets faster.

2 is important because the gallium based LM components can and do oxidize. This is the reason why people generally recommend sealing your contact as much as possible. With a desktop chip under an IHS that's already halfway taken care of by the IHS itself, so if you've never seen a liquid metal tutorial for a laptop you might have not even known that sealing the LM was important. I'm not making this stuff up, gallium begins to oxidize at 25C and fully oxidizes at 75C[2] (Ga2O3, mass fraction 0.12) which is a totally achievable temperature range. This is not even considering the fact that the other components on the LM can oxidize, and they likely will.

Other interesting things to note:
1. The thermal conductivity of CuGa2 (the principal alloy of gallium and copper) at 20C is 98 W/(m⋅K). Copper's thermal conductivity at 20C is 400 W/(m⋅K).
2. We have no idea what exactly is in LM, the formulations are proprietary - but it's been said that the consistencies of Coolaboratory CLU & grizzly conductonaut are not the same so clearly whatever is in this stuff varies and lead to different results.
3. Simply buffing the residue off the copper heatsink and reapplying the LM might be OK. The GaCu alloy is obviously not as good of a heat conductor as pure copper but it's not clear how deep the gallium will attack. If eventually you keep reapplying LM and the gallium can't penetrate its own pitting then it's effectively self-limiting and you end up with a 'stable' LM application.

These papers are interesting if you want to learn more about the process:

TL;DR
Liquid metal pastes react with copper heatsinks (aggravated by oxidation and higher temperatures) until the gallium in the paste is totally absorbed by the copper heatsink and you end up in thermal runaway. The damage to the copper is permanent, but if you keep reapplying the LM whenever this happens you may end up with a stable LM application. LM is much more stable when used under the heatspreader of a desktop CPU, since the heatspreader is nickle-plated and way less reactive to the LM.


This is an excerpt of a comment on another video
bobsagget823

[rogtech]

Love how he always says "it's a bad idea" then proceeds to tell you how to do it.

[cojsxd]

Flash custom kernel and overclock it XD