Back-reflector design in thin-film silicon solar cells by rigorous 3D light propagation modeling

Abstract

Purpose – The purpose of this paper is to find an optimized thin-film amorphous silicon solar cell design by numerically optimizing the light trapping efficiency of a pyramid-structured back-reflector using a frequency-domain finite element Maxwell solver. For this purpose short circuit current densities and absorption spectra within the investigated solar cell model are systematically analyzed. Furthermore, the authors employ a topology simulation method to accurately predict the material layer interfaces within the investigated solar cell model. The method simulates the chemical vapor deposition (CVD) process that is typically used to fabricate thin-film solar cells by combining a ballistic transport and reaction model (BTRM) with a level-set method in an iterative approach. Predicted solar cell models are far more realistic compared to solar cell models created assuming conformal material growth. The purpose of the topology simulation method is to increase the accuracy of thin-film solar cell models in order to facilitate highly accurate simulation results in solar cell design optimizations. Design/methodology/approach – The authors perform numeric optimizations using a frequency domain finite element Maxwell solver. Topology simulations are carried out using a BTRM combined with a level-set method in an iterative fashion. Findings – The simulation results reveal that the employed pyramid structured back-reflectors effectively increase the light path in the absorber mainly by exciting photonic waveguide modes. In using the optimization approach, the authors have identified solar cell models with cell periodicities around 480 nm and pyramid base widths around 450 nm to yield the highest short circuit current densities. Compared to equivalent solar cell models with flat back-reflectors, computed short circuit current densities are significantly increased. Furthermore, the paper finds that the solar cell models computed using the topology simulation approach represent a far more realistic approximation to a real solar cell stack compared to solar cell models computed by a conformal material growth assumption. Research limitations/implications – So far in the topology simulation approach the authors assume CVD as the material deposition process for all material layers. However, during the fabrication process sputtering (i.e. physical vapor deposition) will be employed for the Al:ZnO and ITO layers. In the framework of this ongoing research project the authors will extend the topology simulation approach to take the different material deposition processes into account. The differences in predicted material interfaces will presumably be only minor compared to the results shown here and certainly be insignificant relative to the differences the authors observe for solar cell models computed assuming conformal material growth. Originality/value – The authors systematically investigate and optimize the light trapping efficiency of a pyramid nano-structured back-reflector using rigorous electromagnetic field computations with a 3D finite element Maxwell solver. To the authors’ knowledge such an investigation has not been carried out yet in the solar cell research literature. The topology simulation approach (to the best of the authors’ knowledge) has previously not been applied to the modelling of solar cells. Typically a conformal layer growth assumption is used instead.