Light Management Optimization of Upconversion-Based Solar Cell Using Genetic Algorithms

J. Bork1, M. F. Doty1, J. M. O. Zide1
1University of Delaware, Newark, DE, USA
Published in 2020

We demonstrate a COMSOL Multiphysics®-based approach to light management for upconversion-based solar cells (UBSCs). Single-junction solar cells are of high importance to sustainable energy production, however the cells are limited in their performance by the bandgap of the solar cell. Moreover, solar cells are unable to absorb light having an energy less than the bandgap of the solar cell. This parasitic loss can be addressed by photon upconversion, a process that converts below-bandgap (BB) light into above-bandgap (AB) light through sequential absorption of two BB photons. Use of photon upconversion in solar cells makes a once-inaccessible region of the solar spectrum accessible to the solar cell, thereby increasing the efficiency limit of the solar cell.

In order to achieve this increase in efficiency, however, the solar cells must be designed specifically for upconversion. Moreover, a UBSC has additional considerations compared to that of a typical solar cell. Much like a light-emitting diode should be designed for maximum emission and a photodetector needs to be designed for maximum absorption, a UBSC needs to be designed for both the absorption of BB light and the emission of AB light for best performance. In this work, the Wave Optics Module in COMSOL Multiphysics® is used to optimize the performance of a single-junction UBSC, taking into consideration absorption and emission effects. Absorption of the BB light in the upconverter layer is modeled by an adaptive frequency sweep of a planewave port excitation. The emission of the AB light is modeled as a single-frequency dipole surrounded by perfectly matched layers. Use of LiveLink™ for MATLAB® enables these separate simulations to be combined to determine the efficiency of the UBSC structure (while accounting for effects such as Purcell enhancement and photon recycling efficiency). Further use of LiveLink™ for MATLAB® in conjunction with parametric sweeps enables the use of genetic algorithms to find an ideal combination of device parameters (layers thicknesses, grating periods, etc) for maximal performance.

The results of this model verify that the consideration of both absorption and emission effects is critical to optimizing the performance of UBSCs. Preliminary results suggest performance enhancements of ~1% are feasible. This performance limit increase is considerable for a competitive market and such a well-established technology that already operates within a few percent of the existing performance limit. Additionally, the model provides paths for future enhancement and design rules for the upconverter itself. The understanding provided by this model is ultimately important to the improvement of third-generation solar cells and for advancing the development of sustainable energy production.