Investigation of efficient spectral splitting for concentrator modules using luminescent materials

One approach to high efficiency photovoltaic (PV) involves the manipulation of the incident radiation spectrum. The idea is to present an optical spectrum to the PV that is well matched to the bandgap of the PV material. Solar thermophotovoltaics (TPV) is one example. Another is the use of optics to spatially separate components of the solar spectrum and direct each toward individual PV cells [1]. Wavelength selective mirrors based on dielectric stacks is one solution, but workable designs based on them do not permit a large number of spectral channels. This study investigates the potential of a new concept for dividing solar radiation into spectral channels using an optical design that (1) is simple, easily manufactured, and extensible to many spectral channels, and (2) does not achieve high geometric concentration. The concept is based on the approach of stacked luminescent solar concentrators (LSCs) for dividing the solar spectrum using fluorophores that are tuned to different spectral bands [2]. However, whereas multicolor LSCs are perform two functions using the same optical component - spectral division and concentration - our concept uses luminescent materials for the sole purpose of spectral splitting. The particular design we investigate uses a cylindrical optical matrix that guides light by total internal reflection. Along the length of this light guide are sections that are doped with a luminescent material - in this case semiconductor nanocrystals (see figure). We find that the optical efficiencies can be quite high (QE >; 90%, PE >; 80%) compared to what one might intuitively expect. Furthermore, when we couple the output to a PV model based on experimental cell performance parameters we find that solar-to-electric conversion could exceed 30% using existing materials. Although this does not exceed what can be achieved by HCPV designs on multijunction epitaxially grown stacks, the concept is easily extensible to an arbitrarily large number of sp- ctral channels. This advantage could be used as a route to many more junctions than other approaches permit. In addition to modeling results, preliminary analysis of engineering challenges is also presented (heat rejection, sensitivity to fluorescence quantum yield, etc).