Mercury cooling

Associate
Joined
16 Nov 2011
Posts
1,057
Location
127.0.0.1
Has anyone ever thought about, know someone that has or even cooled a system themselves with mercury. Mercury is 13x more thermally conductive than water, so in theory to get a substance's temp down, mercury would be 13x relatively warmer than water. At 60'C 21'C water would be equivalent to 57'C mercury and if you cooled the mercury down to -30'C it would be equivalent to water at -1110'C.

So presuming that you could get pumps and tubing that are mercury safe, you could in theory have quite a nice cooling system.

And don't tell me not to do this, it is just a theoretical discussion.
 
Liquid metal is an interesting concept. If someone designed a safe, reliable, sealed unit that couldn't leak, then it would be a very good system - in theory.

Jim, that one uses NaK, which I'd say is probably even worse than mercury! As we know, mercury is very toxic. NaK is toxic and highly flammable/explosive!
 
Yeah, gallium is a very soft metal, but it needs to be at 30C or above to be a liquid. Galinstan however, is a liquid at room temp and freezes at -19C to +11C depending on the blend. This stuff is an alloy of gallium, indium and tin, far less dangerous than Mercury or NaK. Unfortunately some versions have antimony in them.
 
Mercury doesnt play nice with other metals, same way as gallium. All metals which are liquid at room temperature have this property, no way around it i'm afraid.

Also you want to avoid any highly electrically conductive liquid.

All the complications mercury would bring, itd be cheaper, easier, safer and more effective to just use a phase change unit.

Mercury's biggest safety concern isnt its toxicity but how impossible it is to clean up, since it goes EVERYWHERE. It doesnt seem to slow down or stick to anything so if you drop a blob on hard floor from a height you can see it split to a thousand tiny silver marble drops and scatter into every nook and cranny near it. You can have contact with mercury without any issue. I have even put fingers in mercury before. The oddest feeling ever, its surface tension and weight feel like the liquid is pressing on you but because it is a metal and does not stick to non metals like other liquids, it doesn't feel even a little wet.

it would be equivalent to water at -1110'C.

Over 800 degrees colder than absolute 0, made me laugh a little xD

At the end of the day, you are just increasing efficiency of heat transfer between a hot chip and ambient temperature. The biggest gains to be had in a situation like this, with ambitious cooling, is to cool below ambient temperature.
 
Last edited:
I worked with Cerrobend (woodss metal) for about 6 years, interesting product as it's only liquid at 70c, not sure on it's thermal conductiveness but we used it for bending aluminium sections.

Obviously no good for any sort of liquid based cooler as it's a solid below 70c, but I wonder how it would behave as part of an aircooled heatsink as you can cast it into any shape.

It's nasty stuff though, made up of cadmium and lead, although we used to use it without any form of PPE or risk assessment, I've breathed in more fumes from that stuff and had cuts gone manky from it than I care to remember.
 
Cadmium and lead, what a great combination. Definitely not ROHS compliant! It could definitely be made into an intricate shape, but if it reaches 70C, it'll melt over your motherboard. I wonder how you'd make an RMA claim on that?

Off-topic I know, but I work as a design engineer for a defence contractor. You'd be amazed at how much cad, lead and asbestos we're trying to get out of service.
 
I meant poured into a vessel as part of a cooler, like heat pipes or something.

And yes I hear you RE: asbestos, my current job has me working in a lot of schools and it's amazing how many of them are still full of the stuff, in fact the protocol seems to be that if it's in a location where it is highly likely to stay undisturbed then they would rather leave it in situ than go through the danger and cost of removing it.
 
Sorry, didn't fully understand what you meant. Fill a pipe with it and it's a sealed unit. It just needs 70C to melt it.

Asbestos, urghh. My garage roof is made of asbestos. If it's not going to be disturbed, then it's just better to leave it be. I remember some stupid contractors whose idea of asbestos control was sunglasses and hawaian shorts. And of course lobing the stuff down to the floor from the top of a cherry picker...
 
I'm always puzzled why no-one has come up with a small refridgeration unit it'd be far safer than liquid metal and more thermally efficient as you can cool it to virtually any temperature regardless of ambient. Someone has even built one themselves theres a video on youtube somewhere or other.
 
I'm always puzzled why no-one has come up with a small refridgeration unit it'd be far safer than liquid metal and more thermally efficient as you can cool it to virtually any temperature regardless of ambient. Someone has even built one themselves theres a video on youtube somewhere or other.

You would get condensation.
 
I'm always puzzled why no-one has come up with a small refridgeration unit it'd be far safer than liquid metal and more thermally efficient as you can cool it to virtually any temperature regardless of ambient. Someone has even built one themselves theres a video on youtube somewhere or other.

There a number of waterchillers and phase change units available which are practically heavy duty fridges.

Plastidip/Putty on the board and insulated tubing and blocks make condensation a minor inconvenience rather than a problem.

Many modern waterchillers can have the temperatures monitored and set around ambient to avoid condensation.

These things are incredibly efficient at cooling but they are also noisy and power hungry and will dump a ton of heat into the room if you don't have it dumping outside through a window or vent.

Heatpipes are very clever simple things, they don't rely on just the conductivity of copper but have a bit of water sealed inside at a pressure which lets it turn to vapour very easily and condense at the top of the heat-pipe, transfering heat and allowing mavity to take the cooled water back down to the bottom of the heat-pipe. Replacing water with mercury would allow the heat to transfer quicker between materials but you couldn't get it to turn to vapour or condense again at the top of the pipe. It would be the difference between having a waterloop with a slow working pump and a mercury loop with no pump. The water-loop would work far better since you are physically moving the material which contains the heat.
 
Last edited:
I'm always puzzled why no-one has come up with a small refridgeration unit it'd be far safer than liquid metal and more thermally efficient as you can cool it to virtually any temperature regardless of ambient. Someone has even built one themselves theres a video on youtube somewhere or other.

Check out Linus Tech Tips on Youtube, he recently done a video on just this.

Has anyone ever thought about, know someone that has or even cooled a system themselves with mercury. Mercury is 13x more thermally conductive than water, so in theory to get a substance's temp down, mercury would be 13x relatively warmer than water. At 60'C 21'C water would be equivalent to 57'C mercury and if you cooled the mercury down to -30'C it would be equivalent to water at -1110'C.

So presuming that you could get pumps and tubing that are mercury safe, you could in theory have quite a nice cooling system.

Nobody has ever thought about doing it because if you want an impractical way of cooling your system and ultra low temps without worrying the toxic nature of Mercury they use Dry Ice or liquid nitrogen both of which give better results then Mercury.

Besides as you say 'presuming that you could get pumps....' a pump that can shift Mercury around a full loop? Mercury is heavy like lead and more dense, I doubt you could get a consumer pump that could move that much mercury easily.
 
Back
Top Bottom