Background: solar cells work by a photon creating an electron-hole pair, which are then shuttled apart by the built-in electric field before they can recombine.
Would this be possible using thermal energy? Doing some rough calculations kT at T=300K is 0.026eV, which is ridiculously small, but how about high temperature semiconductors with a low band gap?
Wow, but those seem to use IR instead of heat? I wonder if heat > electricity violates the second law of thermodynamics.
Only if you’re trying to reduce entropy. If you have a concentrated heat source and a cold place to spread the heat into, then the 2LoT is on your side.
Fossil fuel and nuclear power generation plants mostly convert heat into electricity. You just can’t ever get 100% efficiency, is all.
It is possible to convert heat directly into electricity. That’s how thermocouples work. Google “Seeback effect” if you want the gory technical details.
Peltier devices, aka thermoelectric coolers, are found in CPU coolers, portable drink coolers, and all sorts of things. If you run electricity through one, one side gets hot and the other side gets cold. Reverse the current and the hot side gets cold and the cold side gets hot. They work in reverse though, too. If you apply heat to one side and cold to the other, they will generate electricity.
A more mechanical way to convert heat into electricity is to use a Sterling engine to drive a generator.
The second law of thermodynamics is just fine and happy with all of these.
Doh, I read your OP too quickly and thought that was what you meant.
Yeah, the key is that you always need a temperature difference between two things.
Hmm I think if charge carrier creation occurs by temperature, that’s a conductor.
There are also thermoelectric generators, which are based on vacuum tube principals. This is what is running some of our deep space probes, the ones that venture so far from the sun that solar cells are not feasible.