Night Sky Heat Sink For Electric Generation

Low grade heat is cheap. If you paint something black and leave it in sunlight, it can easily reach 40-60C temperatures. If you drill down a kilometre, the rocks are 40C. Not enough to boil water, but enough to have a bath or heat homes. Which takes up a significant fraction of our energy use in Britain, so that's quite something as it is.

However, we can't generate appreciable electricity from it. The temperature differences between the heat source and ambient are too low to get much power, and at a very low efficiency.

But what if we could instead use far colder heat sinks? Say, by radiating away excess heat to the night sky?

If it wasn't for our atmosphere (I am NOT advocating we remove the atmosphere), the night sky would make an excellent sink, at a temperature of only 3K above absolute zero. Unfortunately for radiative cooling, we have an atmosphere, and clouds, and these absorb heat and reradiate it at night. On cloudless nights when the temperature is low enough but above freezing, it is possible to produce ice, but getting below this without help is difficult.

Enter the help. In 2014, researchers at Stanford University demonstrated it was possible to achieve radiative cooling in the daytime with an appropriately doped radiator. If it can be made to emit infrared in the window the atmosphere is transparent to, 8-14 micrometres, and reflect enough of the other light hitting it, it is possible to radiate enough energy to reduce temperatures by *42 degrees celcius at midday*. In California.

So. What if we were to use this advance to improve nighttime radiative cooling, and use it to freeze a large heatsink, a block of mud perhaps? During the day this could then be used with an appropriate heat source, whether geo or solar thermal, to produce electricity. It wouldn't be small, but we're not trying to power a vehicle here (someone will suggest we try to power a vehicle).

If, say, we can create a heat sink temperature of 240K, and a heat source of 330K (achievable by simple flat plate solar collectors), the Carnot efficiency would be 21%. Real world heat engines achieve, afaict from searching the internet, about 75% of the theoretical maximum, so we might achieve 15% efficiency of converting the sunlight into electricity. Not as high as the 20% that is typical for solar PV, but this system would incorporate thermal storage by design, enabling it to deliver power consistently and when it is needed (not *right* when its needed, not like gas turbines, but the demand is relatively predictable). Perhaps further optimisations will give us colder sinks and hotter sources. We could also use goethermal, avoiding any reliance on sunlight entirely.

This wasn't something that we could have built nine years ago. I'm somewhat excited by it. Partly because it's unusual enough to flummox the usual crowd of people who don't want Britain to build any energy generation.