This shows the solar spectrum: https://en.wikipedia.org/wiki/File:Solar_Spectrum.png https://en.wikipedia.org/wiki/File:Solar_Spectrum.png Visible light ends

This shows the solar spectrum: https://en.wikipedia.org/wiki/File:Solar_Spectrum.png https://en.wikipedia.org/wiki/File:Solar_Spectrum.png Visible light ends at about 700nm, so it looks to me (doing an approximate area comparison) as if there is about twice the total power in the spectrum available in the infrared as compared to in the visible. The article (sorry, it's behind a journal subscription) doesn't say outright in words what frequencies are most absorbed, but from looking at their absorption spectra I would assume that the band gap is between about 650nm and 850nm. I am a materials scientist, but I don't know enough of the acronyms they're using to say more than that. The only other major part I understand and knew to look for is that those efficiencies are measured under one sun of illumination (a lot of the really good looking numbers are done under tens or hundreds of suns). Convoluting those absorbance spectra with the solar spectrum to get an approximate maximum power input looks to me like something along the lines of 1/2-2/3rds of the total power as is in the visible spectrum. Definitely not negligible.