I spent a year participating in a study on solar energy for electricity, broadly defined. This was a while ago, but some of the basic facts are invariant.
1) For direct solar-to-electricity, one must look at lifetime costs, and include the costs of building and maintaining the collection array and 'power conditioning'. For 'solar panels' this came out to a large enough cost that (at then prices ~1978) if the solar cell itself was less than 10% efficient and was in Boston (a pretty good location from the point of view of clear days) it would not compete econmicaly with fossil fuel even if the solar cells were free.
2) At low market penetration, solar requires relatively little storage, because its availability is reasonably matched to the power demand--in particular, peak air conditioning needs drive the east coast peak power demand in the US. At high market penetration, storage is a big issue, and increases cost. If you had a bacterial system that generated fuel, preferably reversibly, handling both storage and conversion, it would be a real help. Methanogenic bacteria would be one creature to try engineering better through evolution or design. One big advantage of biomass is that is has no needed electrical storage.
3) As a physicist, you might enjoy the draft introduction that Jerry Gollub and I wrote to the APS report. It shows how to think about a solar cell as a form of Carnot engine. I am sending you a copy by snail mail, since it exists only in prehistoric paper form, and much of it was cut out of the final report (as I recall).