High-Q optical metasurfaces: from quasi-BIC modulation to ultra-narrowband absorption in the near-infrared spectrum

In this paper, an all-dielectric metasurface composed of silicon and silica is proposed. By breaking the symmetry of each cell structure, two quasi-bound states in the continuum (BICs) with high quality (Q) factor can be excited. The maximum Q value for quasi-BICs can reach up to 15,815, and their modulation depths and spectral contrasts can be as high as 100%. We investigate the performance using the finite element method, and found that the Fano formula based on temporal coupled mode theory fits well. Multipole decomposition results indicate that transmission peak FR1 and transmission peak FR2 are mainly controlled by electric quadrupole (EQ), while transmission peak FR3 is mainly dominated by magnetic dipole (MD). By changing the polarization angle alpha, switching between BIC and quasi-BIC can be achieved, and thus this metasurface exhibits effective bidirectional optical switching performance. In addition, dynamic modulators are a hot topic of research. A dynamic modulator was designed by adding two rows of single-layer graphene on both sides above the silicon layer. As a matter of application, an ultra-narrowband absorber with a dielectric-dielectric-metal (DDM) structure was constructed by introducing a sufficiently thick aluminum metal layer below the original structure. The absorptivity of absorption peak A1, absorption peak A2, and absorption peak A3 is observed to be 70%, 99%, and 82%, respectively, with Q-factors of 977, 2799, and 1351. At the same time the structure can be used for refractive index sensors. The maximum sensitivity is measured at 286 nm/RIU, and the maximum figure of merit (FOM) reaches 534 RIU-1. In conclusion, the metasurface provides exploration for multifunctional devices such as optical switches, modulators, ultranarrowband absorbers, and refractive index sensors. (c) 2024 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI)training, and similar technologies, are reserved.