I am a teacher, not an engineer, trying to dramatize for nineth-graders that, even without evaporation, variations of pressure alone can change R effect. We are building a device with H2O in a chamber inside a sealed jacket to be filled with a refrigerant connected to a reservoir under controllable air pressure. We want to effect freezing of the H2O by reducing the pressure on the refrigerant, then remelt the ice by increasing the pressure on the refrigerant. High efficiency would be great, but clarity of principle is our key interest. What amateur-safe refrigerant should we use to keep the pressure within the range of a small air compressor and vacuum pump (preferably between zero psia to 100 psia, but other minimums and maximums are no problem)? We are all excited at making this work. Any other suggestions?