Optimized scattering power spectral density of photovoltaic light trapping patterns

We present a generic approach for the optimization of light-trapping patterns for thin-film solar cells. The optimization is based on tailoring the spatial frequencies in the light-trapping pattern to the waveguide modes supported by the thin-film solar cell stack. We calculate the dispersion relations for waveguide modes in thin-film Si solar cells and use them to define the required spatial frequency band for light trapping. We use a Monte Carlo algorithm to optimize the scattering power spectral density (PSD) of a random array of Mie scatterers on top of a-Si:H cells. The optimized particle array has a PSD that is larger in the desired spatial frequency range than the PSD of a random array and contains contributions at more spatial frequencies than the PSD of a periodic array. Three-dimensional finite-difference time-domain simulations on thin-film solar cells with different light-trapping patterns show that the optimized particle array results in more efficient light trapping than a random array of Mie scatterers. We use the same approach to design a random texture and compare this to the Asahi-U-type texture. We show that the optimized texture outperforms the Asahi-U pattern and an optimized periodic pattern. The light-trapping patterns presented avoid the ohmic absorption losses found in metallic (plasmonic) patterns. They can be tailored to specific spatial frequency ranges, do not contain materials that are incompatible with high-temperature processes, nor require patterning of the active layer. Therefore, they are applicable to nearly all types of thin-film solar cells.