It seems to me that direct conversion of sunlight to electricity is expensive and (presently) a very low efficiency process.
My question is: is there any way to convert it to microwave or RF frequency radiation?
Like an AM radio-using “heterodyning” to convert the RF to an intermediate frequency.
If we could do this, a simple rectifier /detector circuit could convert the radiation to DC power. Has any one very tried this?
It’s certainly possible in principle: That’s what fluorescence is. Ultraviolet light is absorbed, and the energy is re-emitted as multiple photons of visible light.
That said, to do what you’re suggesting, you’d need atomic or molecular transitions in the microwave energy range, and I don’t think there are any of those (at least, not useful ones).
Hyperfine energy levels are usually spaced in the radio-wave or microwave range (as in the famous 21-cm line), but perhaps that’s what you meant when you said “not useful ones.”
Right. Fluorescence depends on having multiple low-energy gaps next to each other, so you can get all of the energy of the original high-energy photon split up into low-energy photons. If you tried to use the hyperfine transition, at best, you’d take one high-energy photon in, and get out one high-energy photon with nearly all of the original energy, plus one or two very low-energy ones.
Wow, learn something new today. I thought Heterodyning was this and not a “real” thing. 
Four wave mixing occurs all the time in fiber optic systems. This can either up- or down-convert light. Four wave mixing is also used with coherent light (lasers) in crystals to create sources with wavelengths not obtainable in other ways. with the right crystal and sufficiently high powers, the conversion efficiency can be surprisingly good.