Optimization of Tandem Solar cells by Deep Learning

Abstract

This study explores the use of deep learning to predict the optimal optical design for the top cell in tandem solar cells, to maximize power conversion efficiency. The study also investigates the effects of metasurfaces on tandem solar cell architecture, including the use of active layers as metasurfaces. The study proposes three stacked layers of metasurfaces, each with a specific function. Computational techniques are developed to analyze the optical responses of metasurfaces, including using simulation tools and algorithms. The study uses Artificial Neural Networks (ANN) and 2D Convolutional Neural Networks (CNN) to analyze a dataset of 10,578 TiO2/CH3NH3PBr3/ZnO metasurfaces. Nine different CNN models were used with different architectures to identify the best hyperparameters that give the low mean square error. The results show that CNN shows high prediction accuracy of resonator geometry aside from the training dataset. The CNN-based predictions were generated much faster, taking an average of 0.3 ± 0.05 seconds per prediction, whereas each FDTD simulation (n=9) took approximately 25 minutes hence, the CNN is 85 times faster than conventional solvers. The Deep SHapley Additive Explanations (SHAP) framework was used to gain insights into CNN's predictions and understand the behavior of complex nanophotonic devices. The designed metasurfaces were integrated into a typical reference tandem solar cell architecture, and the study concludes that the proposed metasurfaces can significantly enhance the efficiency of tandem solar cells. The active layer comprises of near 90% absorption of solar spectrum. The average absorption of the top cell increased in the UV-vis region (650-800nm) up to 93.4%. The bandwidth of absorption in the the Silicon bottom cell also increased which shows that metasurface transmits or scatters the unabsorbed NIR light to bottom cell. The estimated Jsc of the top and bottom cell is 19.5mA/cm2 and 20mA/cm2 respectively. The Jsc in the top cell increased by 2mA/cm2 by integrating optimized metasurface with the bottom cell. The recorded Voc of the solar cell is 0.7397V. The total Jsc of 35.91 mA/cm2 . The fill factor observed is 82.2%