Waiting for photovoltaic cells that use a phosphor coating to translate photons into energy levels that can be more readily absorbed. I'd made a note to look into this, and when I did, it looks like there were some promising results about 5 years ago with UV to visible translation, and a factor of 14 improvement in electric output. But, that's just for the UV component. Maybe some sort of prism could be used to split visible light off to one PV cell, and UV to another, while IR could go to either another, or maybe used for a thermocouple of sorts. Something this big would really only be efficient if you could use multiple reflectors to focus more light onto the PV cells. Hrmmm.