Cooling Nitrogen Water-Cooling

[Michael641]

Just as water flows through a liquid cooling system in a computer, can liquid nitrogen do the same thing without escaping the tubes (vaporizing)? If nitrogen can't do this, can liquid helium do it? The pressure may build up but can you release the pressure without releasing any of the liquid nitrogen / liquid helium?

 

[Zerk2012]

No to much pressure would build up!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
To cold for the pump to function!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

 

[Solandri]

You can use liquid nitrogen or helium for cooling. Some of the early supercomputers were cooled this way. The computer would sit on one floor, and the floor underneath was filled with refrigeration equipment pumping liquid nitrogen or helium upstairs. I'm not sure if they still do that, but it's certainly been done before.

There are two problems with this.

1) Large temperature gradients cause more thermal expansion and contraction. Stuff changes size, causing cracks and breaks which eventually leads to leaks and equipment failure.

2) Water has substantially higher specific heat. 4.184 kJ / kg*K - raising the temperature of 1 kg of water 1 degree K requires 4.184 kJ of energy. Compare that to 2.04 kJ / kg*K for liquid nitrogen. In other words, to get the same amount of cooling, you need roughly 2x as much nitrogen as you do water. Or if you pump the same amount of liquid through the pipes, the the nitrogen has to rise in temperature up 2x as much to carry out the same amount of heat as water. Liquid helium's specific heat varies from about 2-4 kJ / kg*K, but that's down around 2-4 Kelvin, which is a ridiculously expensive temperature to maintain. (You attain it by basically hooking up 4-6 refrigerators in sequence, each one chilling the "ambient" temperature that the next refrigerator operates in.)

On top of that, if you do generate enough heat to initiate a phase change into a gas, water is ridiculously better at absorbing heat that way. Its heat of vaporization is 2257 kJ/kg (energy required to turn it from liquid into gas). N2 is just 199 kJ/kg. Helium is only 21. So water is *much* safer in a closed loop because only a tiny amount of it needs to vaporize to absorb a huge amount of heat. In the same emergency overheat circumstances, a large amount of nitrogen and a massive quantity of helium would have to vaporize to absorb the same amount of heat, causing much higher pressures and a much greater chance of containment failure.

Really, water is pretty much the ideal coolant. Liquid at room temperature, non-toxic, absorbs a huge amount of heat, and is very safe (from an energy absorption standpoint) if you let it get too hot. Its properties are exceeded only by certain exotic coolants, like liquid salts.

 

[TJ Hooker]

@Solandri
The specific heat by itself isn't really what's important. What you care about is the thermal conductivity of your cooling solution and the temperature delta. Now, thermal conductivity will be related to the specific heat of the coolant, but even if conductivity is halved with nitrogen, the temperature delta will be way larger, so it will be a more effective cooling solution.

Also, the heat of vaporization is largely irrelevant, as water coolers don't rely on phase change. You'd have to let your CPU get to 100 C. Unless you're talking about a heat pipe which is in a partial vacuum, but in that case the energy of vaporization is different anyway.

Not that I'm recommending liquid nitrogen/helium cooling. As far as consumer PCs are concerned, it's too impractical to use anywhere other than one-off extreme overclocking attemps.

 

[Solandri]

Agreed that temperature delta matters (rate of heat transfer is proportional to temperature delta). But specific heat will determine how quickly that temperature delta diminishes (how quickly the liquid coolant heats up). So (temperature delta) * (coolant flowrate) determines the rate at which you can carry heat away. But specific heat determines how long you can maintain a temperature delta for a fixed flowrate, so all three matter.

Thermal conductivity is mostly relevant for solids. For liquids and gases, convection plays a much greater role in heat transfer (which is why air cooling works, despite air having just about the worst thermal conductivity there is). The exception is if you can create micro-channels (microscopic tunnels) for the coolant to run through. This vastly increases the surface area for heat transfer, while being small enough to pretty much eliminate convection, and also increases the resistance of the coolant against pumping. In that situation, the liquid starts to act more like a solid when it comes to heat transfer.

http://www.eetimes.com/document.asp?doc_id=1279032

I only mentioned heat of vaporization because once you're talking about nitrogen or helium cooling, you're getting into pretty exotic technology. I assumed OP was interested in theoretical knowledge of cooling systems in general, not planning to actually nitrogen cool his PC. So for sake of completeness I felt it was important to mention heat of vaporization (and liquid salt), since it's actually a very important part of the safety of other cooling systems (pressure relief valves in particular)..