Modding Mercury viable water cooling alternative?

Besides the obvious safety and monetary concerns, whats to stop someone from using mercury in place of water to cool a CPU or GPU? Water (k=.58) is 3x better than alcohol (k=.17) and mercury (k=8) is 14x better than water at removing heat. (k = W/mK).

It would obviosuly be heavy and require a more powerful pump than a water loop. Also, spilling it would cause major chaos. But besides the caveats, why haven't I seen someone do this? I mean, I've even seen ln2. using mercury isn't much of a stretch.

 

The viscosity it probably one of the main concerns; since you wouldn't be able to achieve as high a flowrate (compared to H2O), which would likely negate the higher thermal conductivity.

Not to mention that it's be near impossible to get the necessary volume of mercury

 

That, and you would probably die.

It has been discussed here before, for example - http://forums.bit-tech.net/showthread.php?t=119857&highlight=mercury

 

and also wudnt u need liquid mercury...now correct me if im rong and forgive me but i gave up science in the 9th grade coz i cudnt fit it into my skwl schedule....but isnt mercury sumtimes solid and therefore needs 2 b at heactic temps to be fluid?

 

no mercury is liquid at room temperature its the only metal that is.

 

This pops up as regular as a very regular thing on All-Bran.

I'll add that mercury reacts with aluminium, copper, brasses, solders, etc, to form amalgams, so what do you make the radiator from? Steel (or iron-plated copper) would work, but I doubt you can buy them to fit an ATX case.

 

shot down!

Oh well, I was sitting in Heat Transfer class and the professor was talking about using mercury in palce of water to cool nuclear reactor rods and I had a thought.

 

Your post makes babies cry. how old are you? 13?

 

apparently he dropped english too...

 

I thought nuclear reactors did use metal cooling systems? Anyway, i don't think its done because for the gains, its not worth it

Plus its hard to get, hard to pump, hard to store, hard to put in a PC, etc etc

I believe it is (or was) used in nuclear reactors just because water couldn't stand the heat

 

Normally, the reactor rods aren't too hot. But in the 3 mile island accident, the reactor lost its coolant (for whatever reason, I dont recall). I don't remember the specifics, but my professor said the rods were so hot that the water couldnt come in contact with them. A layer of steam forms on superheated surfaces and doesn't allow the water to actually touch it. I thought he said it was called like the Linden Frost effect, but I guess that's incorrect because I can't find anything by searching that.

EDIT: here it is: http://en.wikipedia.org/wiki/Leidenfrost_effect I was close, its the Leiden Frost Effect.

 

I think something is just so hot that water is vaporized before it can touch it. Coolant is pumped through the reactor at high pressure, and seveal coolant loops lead the heat outside the reactor without bringing with it radiation, which boils water and turns a turbine with steam. If the high pressure pipes burst, the coolant floods out very quickly and the rods get into thermal runaway. The hotter they get the less water can do to cool them, and they melt right through the reactor. Tom Clancy's The Hunt for Red October (which I'm currently reading) had an excellent explanation of how it works, right before it all went wrong.

 

that's basically correct, check my edit (you replied before I could find the article!).

 

No, Three Mile Island was a fairly standard Loss of Coolant Accident - a pressure relief valve opened and then would not close. By the time they worked out what had happened, more than half the core was exposed. As steam is much more transparent to fission-inducing neutrons than water is, the reactor began to melt down.

 

If i remember correctly, nuclear reactors use liquid Lithium as a primary coolant and run this through a heat exchanger to transfer the heat to water for power production.

If you did make a mercury cooling loop, you'd want to be carefull not to get any leaks. Otherwise you could end up as mad as a hatter

 

hm........ in order to make something like this viable......... I would probably recommend bent stainless steel lines of maybe 3/8ths size(similar to whats used for car brakes) and everything else being along the same vein of materials......... in order to disassemble everything, since you will want to do that at some point, you would need a professional mini clean room setup to do it safely, you could probably make something to that effect(however getting filter systems small enough to take any mercury vapor out of the air would be exspensive as hell. To safely run this system(since sh!t happens) the preferred method would be to have a similar clean room setup to run the tower itself with everything installed inside.... As for viscosity......... well sometimes thicker is better, ie lots of rotary engine enthusiasts used a really light motor oil instead of water and antifreeze in their engines with much better results, if you had the right pump setup I'd say go for it, what I would probably do....... is to have a strong pump, but gear so it pumps abit slower than average, more time to pickup heat through a stainless block, and since it does end up picking up more heat........ I think you would probably end up seeing on average lower temps than water but again......... everything comes back to ambient tempuratures...... Since you have what is essentially a liquid, It will ultimately come to a median temperature.

Now, If you have a 2-3 radiator setup all with 120mm fans, you might see slightly lower temps than can be found in a pure water cooling system, reason being, as you point out, mercury can hold more heat so if you can dissipate the heat faster as well......... yada yada. So basically, on paper this is something that looks good, but no offense this reminds me of when i work at genie ind. 5 idiot engineers arguing over a bolt length for almost 3 hours when we were building the first unit of a new product line..... throw some common sense in there and it comes down to the same thing...... median temperature will never change that much with any liquids, you might get better results with certain types of coolants in machinery, ie I use coolant thats 1/10th oil, 9/10ths water in the machinery at work for better results in cooling capacity, but that oil sits in a resevoir for probably at least an hour before it flows back on through the system....... getting the drift here? Its kinda late so I just keep rambling on lol

anyway that is essentially the issue here, not much more to it. plus If you chill the mercury hoping to get better results(ie a waterchilled system) the mercury will just flow slower as it solidifies

 

Mercury, being a metal, has a greater thermal conductivity than water.

But in a liquid cooling system, it's not thermal conductivity that matters so much as specific heat capacity, and water is superior to essentially every other material in this regard.

 

What about Gallium?
http://en.wikipedia.org/wiki/Gallium

BTW, a Gallium/Aluminum alloy, dropped into water separates the hydrogen freakishly fast.

 

How abouts some kind of giant heatpipe system (mercury and mercury vapour can be used for this iirc) and do away with the pump

 

I thought you could lower boron rods into the reactor to stabilise it, or failing that, blow boron gas into to completely kill it

At least i think so, maybe during the 3 mile incident that wasn't standard implementation ....