I do understand the principles behind refrigeration. I've even fixed a few fridges and air conditioners in my time. But I'll readily admit that I'm no refrigeration engineer (and my mechanical knowledge kinda sucks a little, too).
My question is whether a subzero cooling system can be implemented within a PC chassis without the use of a compressor. No compressor at all. It seems to me that if the coolant pressure is sufficiently low then it will not require as much "push" back from the cold side (condenser, radiatior, etc) towards the hot side (evaporator, CPU cooler, etc) - perhaps gravity acting on a sufficiently large liquid-state coolant mass in the reservoir would be enough.
I understand that no cooling system can take more heat away than is produced and thus can at best reach equilibrium with the ambient (cannot cool below ambient) unless more energy (some electrical power, etc) is added to the system. I guess that's the real difference between active refrigeration and passive heatpipes. It's also why TEC cooling can be spectacular for spot-cooling but overall it's judged to be energy-intensive and inefficient. There are always thermal inefficiencies and losses, anyhow, I understand that.
But is no-compressor subzero phase-change cooling viable, or is the whole notion fundamentally flawed? I suppose the liquid coolant reservoir might need to be excessive large and not really workable for PC applications, but if it is workable then what sorts of scales would be involved?