I had an idea recently. But I don't have the physics knowledge to know if it makes sense or not. I've become a bit of a PC enthusiast lately and one thing that is big in the PC world lately is water cooling.

Here is my theory. If I take a regular PC water cooling loop, and lower the air pressure (custom loops typically have a bit of air at the top of the reservoir). I could lower the boiling point of the water to just above room temp (lets say about 30°C). Then when the system is running, the liquid water will hit the CPU block and vaporize, since CPU's can easily be 75°C. The water vapor will then travel to the radiator and re-condense to liquid. And continue back to the reservoir, etc.

Lets just assume the PC cooling parts can handle the lower pressure for now. Also I assume the system pressure would increase once some water is turned to vapor, but I should be able to continue lowering the pressure while the system is running until it reaches a steady state, right? Since heat pipes evidently don't have this problem.

Would this system remove heat from the CPU any faster then a regular water loop? From what I understand, the phase change will remove heat much faster then normal. It's the basic principal that heat pipes use, which are already widely used in PC's. But heat pipes don't have a pump to force the circulation of a large amount of water.

Are there factors I'm not aware of. Will the pressure in the system just skyrocket once water begins to vaporize? Would it require some kind of active pressure regulator mechanism? I'm guessing not since this isn't a problem for heat pipes.

I was thinking of testing the idea out in the next fews months, though maybe using a small heater element as a stand-in for a CPU, just in case things go sideways.

Also, if you are thinking that all this might be needlessly complicated for no real gain. Well, that's the PC enthusiast world right now! regular PC water cooling already isn't necessary right now. many air coolers are cheaper, quieter, more reliable and keep things cooler as it is. it's just fun to play with.

  • $\begingroup$ You should look into fully immersive phase change cooling like this: pcper.com/news/Cases-and-Cooling/… $\endgroup$ Dec 7, 2017 at 20:56
  • $\begingroup$ @A.C.A.C. Actually, projects like that are partly what inspired my idea. I think one of the main drawbacks with them is I've heard the fluid is something like $100-$200 per litre. But yes, It's very cool and I think is a better idea then mine, mainly because the fluid cools every component, not just the CPU. $\endgroup$ Dec 7, 2017 at 21:23
  • $\begingroup$ It sounds to me like you are trying to re-invent the heat pipe. $\endgroup$ Oct 18, 2020 at 22:15
  • $\begingroup$ @David yes basically. Call it like an active or powered heat pipe. Where the fluid is pumped instead of wicked. $\endgroup$ Oct 21, 2020 at 9:48

2 Answers 2


basically, what you are describing here is a closed-circuit phase-change refrigeration system that runs below atmospheric pressure and uses water as the working fluid. this was common 100 years ago before the invention of working fluids (freons of various molecular weights) more suited to temperature ranges that are typical for systems intended to freeze foods; it is called the steam refrigeration cycle and is extensively documented on the web and elsewhere.

assembling from scratch a working miniature steam cycle refrigerator and installing its cold side inside a computer case will be a difficult engineering task. You would be better off, in my opinion, considering a peltier-effect thermojunction cooler or a once-through water system that runs on the cold water tap and exhausts the warm water through a drain hose into a flower bed in your back yard.

  • $\begingroup$ Hey, Did a small amount of reading about the steam cycle refrigeration system. It sounds like a much more complicated system then what I'm proposing. I'm basically making a heat pipe that has a pump to circulate (not compress) the water instead of a wick and capillary action to circulate like in a heat pipe. 1-act.com/resources/heat-pipe-fundamentals/… $\endgroup$ Dec 12, 2017 at 17:25
  • $\begingroup$ Also, My main interest is high efficiency. which peltier coolers are not. I'm not looking for sub-zero temperature cooling. just higher heat transfer rate to room temp. $\endgroup$ Dec 12, 2017 at 17:27
  • $\begingroup$ OK, forget the peltiers. you can get the high heat transfer rate with fast-moving cold water... BTW are you familiar with the overclock community? they have developed all sorts of CPU cooling devices to prevent their machines from croaking... $\endgroup$ Dec 12, 2017 at 19:23
  • $\begingroup$ Well I'm not sure that increasing the flow rate of the water in the loop is going to make a big difference, as typically with these systems the water temperature rises until a balance is reached. I'd then need to add more radiators to dissipate more heat. Which is not more efficient, and takes much more space $\endgroup$ Dec 12, 2017 at 19:37
  • $\begingroup$ Yeah I've seen all the solutions the overclocking community has. I just want to know if this particular modification to an already common system will increase efficiency $\endgroup$ Dec 12, 2017 at 19:40

You are right in thinking that the pressure would shoot back up when the water vaporises, as the amount of gas in the tube is related to the pressure (in a rough approximation by the ideal gas law). If you kept pumping it to maintain the pressure, you’ll be pulling the water vapour out of your system, and it wouldn’t go back to your reservoir.

Moreover, while you might be able to remove more heat per unit mass of water used, the rate of heat removal shouldn’t be visibly affected, as it depends more on the contact surface and thermal conductivity of your exchanger.

  • $\begingroup$ So to get it to maintain the correct pressure, I might need some kind of large expansion chamber? possibly larger than what's going to easily fit in a PC. $\endgroup$ Dec 7, 2017 at 21:39
  • $\begingroup$ A larger volume for the gas to fill would certainly postpone the problem, by lowering the rate at which pressure increases, and it might give enough time for the water to re-condense. One thing to be careful of is that, just as low pressure “helps” with boiling, it will also hinder condensation. $\endgroup$
    – LvdT
    Dec 7, 2017 at 22:23
  • $\begingroup$ So basically your main point is. It won't make a lot of difference because lowering the boiling point will be more like increasing the heat capacity of the water more than increasing it's conductive properties? I don't suppose you have any links I could read which explain this further? $\endgroup$ Dec 8, 2017 at 14:18
  • $\begingroup$ No specific links, sorry. $\endgroup$
    – LvdT
    Dec 8, 2017 at 15:48
  • $\begingroup$ Why is it then, that systems like the one linked in A.C.A.C's comment work so efficiently? From what I've read, the fluid it typically much less thermally conductive, lower latent heat of vaporization, and less thermal capacity then water, it's only advantage over water is a boiling point between 30-40 degrees. yet it cools efficiently without even having a heat sink. probably reducing surface area several thousand times. $\endgroup$ Dec 8, 2017 at 16:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.