I think this refrigeration system is particularly ingenious and interesting as it is very reliable, has no moving parts to wear out, there is virtually nothing that can go wrong with it, completely silent, no motor, compressor, pump. Perhaps the confusion arises from the role of the hydrogen gas. If you have what amounts to a continuous loop of pipe with a reservoir of ammonia and water at the bottom and if the boiler is releasing ammonia gas which is building up pressure on one side of the loop, then why doesn't the back-pressure push the water/ammonia solution backwards out of the boiler ? Because the other side of the loop is filled with hydrogen gas to equalize the pressure. Well then, if the pressure is equalized then what am I talking about as far as the ammonia boiling and building up pressure. It doesn't seem to make any sense right ? Well, here is the ingenious part IMO. As far as atmospheric pressure is concerned, if you put a pressure gauge anywhere on the system the pressure would read the same. There does not seem to be any pressure differential, so how can it possibly work ? The answer is that the hydrogen molecules and the ammonia molecules are different in size. The hydrogen completely fills the space on one side to equalize the over-all pressure, but the ammonia molecules still have plenty of room to expand and move around between the hydrogen molecules as if nothing were there. So if the pressure builds up throughout the system to 10 atmospheres and the hydrogen is at ten and the ammonia is at ten, when the ammonia is released into the hydrogen as far as the ammonia is concerned, there is no pressure there at all, it is free to move around BETWEEN the hydrogen molecules, like pouring sand from an hour-glass into a box tightly packed with bowling balls. The box is filled with bowling balls packed so tight that no more could fit, but as far as the sand is concerned the box is practically empty. The hydrogen is an inert gas that does nothing but prevent back pressure and provide space between its molecules for the pressurized ammonia to expand and depressurize. If you follow the actual refrigerant, the ammonia, the cycle is actually the same as any other refrigerating cycle. Basically: Compress a fluid, drive off the heat then decompress the fluid so it can re-absorb heat. There are many different ways of accomplishing that, some simple, some very complicated but fundamentally all based on the same principle.