Last time I worked in solar, the processes needed for the conversion typically worked above 600C (or much higher). If you can do conversion at atmospheric pressure, then you don't need pressure vessels (which are susceptible to fatigue because of cycling from 100C to 600C+ when a 5 minute cloud passes overhead, for example), but then your gas is at a lower density and now your process is less efficient.
Also, the higher the temperature, the higher the radiative heat losses become. T1^4 - T2^4 = much much higher. So you want to minimize the view factor of the heat transfer tubes/materials. Which means you want a cavity receiver...yada yada. Google solar air receiver and you'll get some interesting reading material, for example:
http://www.sciencedirect.com/science/article/pii/S1876610214...
I'm a lot more familiar with salt systems, tbh:
http://www.esolar.com/wp-content/uploads/2013/10/SolarPACES-...
Yeah, the stuff I'm familiar with is solar thermal water splitting using iron oxide at ~1000C, though I can see why the Caltech group preferred ceria (my favorite ceramic).
I think my favorite cute trick with solar thermal is solar thermal reforming of biomass. By piggybacking on the existing chemical reaction you actually end up with solar energy directly added to the chemical energy! First time I saw that I was struck by how clever it was.
I imagine using CO2 as a reactant is a lot more difficult, that's not a simple molecule to break apart so I very much appreciated that the solar thermal reforming process takes advantage of the plants to do the hard part while extracting even more solar energy.
What do you mean here? What aspect of the sun's energy cycles on the minute/second timescale? And why would variability at that time scale do anything to a solar thermal system?
I'm pretty sure the reason it is hard is because you need to go above and below phase transitions that change the oxidation state, achieve high rates of oxygen absorption and release, and maintain a high enough surface area to not lose your oxygen diffusion rate.
Or are you talking about some new issue?