Here is an invention I'd like to share with anyone who has the money to develop and market it. Furthermore, I don't even have the technical skill to determine if it's viable on paper so I suppose it's reasonable to say that anyone who pursues this is entitled not to share anything with me. Picture a black PVC water tank, shaped to maximise its surface area to the Sun. Let's say it can be filled and topped up by rain water from a typical house roof and can contain 4 ton of liquid water or about 4 kilolitres. When at capacity, there remains several litres of space above the liquid water, say 10 litres. It can be sealed tight to prevent gas exchange (and prevent water leaks - it is a water tank after all). Except that there is a miniature turbine that sits atop the tank. As the water heats up from solar radiation during the day, the vapour pressure above the water increases and escapes through the turbine to create energy that could be stored in a battery. During the night, the opposite occurs: vapour pressure decreases. As you can now see, the device relies on the diurnal thermal gradient. In practice, I imagine one would need to build two miniature valve turbines: one that did work due to increased vapour pressure during the day and one that did work due to the decrease in vapour pressure during the night. But some clever engineer may calculate that a bidirectional turbine is more practical and or cost effective. Let's say the diurnal temperature change in the tank is 5 degrees Kelvin. I think it's safe to assume that the change in the volume of the tank is negligible. But now comes the hard part. The system is not adiabatic and not isothermal. Work is done as the internal pressure equalises with the atmospheric pressure. So I think it's safe to assume it's isobaric. To roughly calculate the potential work that can be done (obviously energy will be lost to moving parts and electrical resistance of the circuit), am I correct in saying I would need only calculated the difference in partial pressure of water vapour between a temperature difference of 5 degrees multiplied by the volume (10 litres)? If so, would the system work better if the tank held only 10 litres of liquid water and 4 kiloliters of partial water vapour? Am I even in the ballpark of attempting to solve this? And would anyone care to comment on whether this could be a viable energy source given that most people can't put ten thousand of these things on their properties. The other problem is whether more energy is expended in the manufacture of such a tank than it can reasonably be expected to return in its lifetime. Thanks for your help in advance.