I'm very interested in building this cool device, and was wondering if water could be used instead of air as the injected fluid. The main differences I notice are the much larger density, and thus possibly smaller angular velocity. And higher viscosity, which could mean a smaller radial gradient of temperature. I've also had no luck finding people using water instead of air, but alas no explanation why not either. If the physical principle why it works was known, one could predict if water would work or not. What do you think or know?