Optimization of spectrally selective Si/SiO2 based filters for thermophotovoltaic devices

- A design procedure is introduced and applied for TPV filters for improved system performance.
- 1D Si/SiO2 based spectrally selective filter designed to enhance overall efficiency of TPV systems with InGaSb, Si, GaSb, and. Ga0.78In0.22As0.19Sb0.81 cells.
- Near zero transmittance is achieved over a wide wavelength range above band gap wavelength of the TPV cell improving overall efficiency.
- For an emitter temperature of 1800 K the overall system efficiency, reaches 50% for InGaSb cell by using the optimized filter, where maximum efficiency for the limiting case is 58%, and the efficiency without filter is 25%.
- For the considered emitter temperature range (1000-1800 K), Si type cell has the worst and Ga0.78In0.22As0.19Sb0.81 type cell has the best performance.

Design of a spectrally selective filter based on one-dimensional Si/SiO2 layers is considered for improved performance of thermo-photovoltaic devices. Spectrally selective filters transmit only the convertible radiation from the emitter as non-convertible radiation leads to a reduction in cell efficiency due to heating. The presented Si/SiO2 based filter concept reflects the major part of the undesired range back to the emitter to minimize energy required for the process and it is adaptable to different types of cells and emitters with different temperatures since its cut-off wavelength can be tuned. While this study mainly focuses on InGaSb based thermo-photovoltaic cell, Si, GaSb, and Ga0.78In0.22As0.19Sb0.81 based cells are also examined. Transmittance of the structure is predicted by rigorous coupled wave approach. Genetic algorithm, which is a global optimization method, is used to find the best possible filter structure by considering the overall efficiency as an objective function that is maximized. The simulations show that significant enhancement in the overall system and device efficiency is possible by using such filters with TPV devices. The methodology described in this paper allows for an improved filter design procedure for selected applications.