There exists a concept in Thermodynamics, which describes theoretical limits in the efficiency of all possible heat-engines. This principle states, that if we have a heat-source and a heat-sink, each has an absolute temperature. The ratio between these temperatures defines the highest-possible output of free energy from a heat-engine, as well as the lowest-possible consumption of free energy by a refrigeration-device.
The principle is based on the axiomatic assumption, that there exist no perpetual-motion machines, which simply convert ambient heat into free energy. If we could connect a heat-engine to an air-conditioner, and if these limits could be exceeded, we would have such a perpetual-motion machine.
This also explains why in practice, air-conditioners, refrigerators and heat-pumps can transfer heat from a colder source to a warmer sink, with the energy in heat far-exceeding the electricity consumed. They are all examples in which the ratio of absolute temperatures is close to 1.0 . Actually, what matters is the ratio of the temperatures of the working-fluid in each case, which is actually more oblique than the ratio for air temperatures, because heat-exchangers are never perfectly efficient. And the working-fluids used tend to be similar, because the temperatures at which those systems are designed to work, are also similar.
This also implies that if we wanted to build a heat-engine that uses small temperature-differences to generate electricity, large reserves of heat would be needed as a source, and sent to the sink, before even small amounts of electricity result – which might sometimes be available – but which constrains the system, regardless of what type of heat-engine is used.
Well, in Industrial Power Generation, the temperatures which the heat-source can be run at, depend firstly on what type of fuel is burning, but also depend on the range of temperatures at which water will boil. At 1 atmosphere of pressure, water only boils at 100⁰C, which is also 373K, while the external temperature tends to be around 273-300K .
Actually, by keeping the water boiling at much higher pressures, its boiling-point can also be increased. But it is generally not boosted beyond 200⁰C , which corresponds to about 473K . And so, according to basic principles, no power station based on water and steam, can be more than 50% efficient.
(Edit 05/12/2017 : Additionally, my late father, who was a professional Engineer, used to tell me, that something prevents a steam turbine from being more than 50% efficient. But, this is not a subject I know about, even though it would additionally limit the maximum efficiency of steam-turbine-based power-stations, to approximately 25%. )
In theory, if we could operate our heat-engine at 1000K, and its heat-sink still at 300K, we could achieve efficiencies closer to 70%. Mind you, that that point our heat-source might resemble a lightbulb, more than what we are used to, but this would still obey the rules of Thermodynamics.
My only point being, that the use of water, and its associated boiling-points, is an arbitrary decision. There is no magical reason why we must use it. We could use vaporous sodium if we knew how to work with it safely.
If one breaks out, a sodium-fire is a nasty hazard, much more dangerous in its nature than wood or oil-fires already are.