S. Shepard, L. Patel, F. Rubaiat
Louisiana Tech University, US
solar energy, photovoltaics, light trapping, nanoparticle scattering
We consider the light-trapping enhancement of a silicon photovoltaic with conductive nanoparticles of various shapes, sizes and spacing as top-reflectors; used in conjunction with a periodic and a pseudo-random back-reflectors. The nanoparticles in the top-reflector are silver (with the appropriate Drude model) and the back-reflectors are perfect electric conductors. The power spectral density then multiplied by the responsivity of crystalline-Si and by the input spectral density (blackbody solar radiation herein, normalized to one) to form Jsc. We see that for a given top-reflector the choice of a periodic back-reflector can yield a higher Jsc for a crystalline-Si PV than the choice of a pseudo-random one. This (perhaps initially surprising) result: that a periodic structure can outperform a pseudo-randomly rough surface as a back-reflector is due to resonances that can be setup between it and the top-reflector. At a single (resonant) frequency it’s easy to see how this can happen. The surprise is that we can sometimes produce enough resonances (perhaps manifesting as a sufficiently broad one) to provide sufficient enhancement across the 700nm broad band of silicon’s responsivity to surpass Lambertian (although Lambertian probably is still optimal for the infinitely broadband case).