Abstract:Objective Nanostructures based on metallic materials can modulate the amplitude, phase, and polarization of electromagnetic waves owing to their surface plasmon resonance (SPR) properties. The interference between bright and dark modes forms Fano resonances in metamaterials. Excitation of the dark mode can effectively suppress far -field radiation and enhance near -field radiation. However, the significant heat loss of metallic materials limits their application in optics; therefore, only a few superconfigurable materials based on surface plasma excitations can be used in practical applications. Recent studies have shown that highly refractive index all -dielectric nanostructures with low absorption properties do not undergo heat loss, thus facilitating the realization of high- performance compact devices. In this study, we designed a fully dielectric nanopillar supersurface with a high Fano resonance quality factor, Q, and modulation depth. We hope our design can provide innovative ideas for asymmetric transmission, polarization angle detection, and super -surface multifunctional multiplexing. Methods In this study, the Fano resonance theory was simulated around a fully dielectric supersurface material. Maxwell s equations describe the electromagnetic -wave propagation law in space, and the equations can be solved to determine the response of the supersurface to the incident light. However, the analytical solution of Maxwell s equations cannot be obtained in general; therefore, the simulation results are typically obtained by solving a system of equations using numerical methods. The two widely used solution methods are the finite element method (FEM) and the finite difference in the time -domain method (FDTD). We used the FDTD Solutions software to simulate the supersurface and perform high -precision simulations to replace the more expensive prototype experiments. The periodic boundary conditions were set in the x- and y -directions owing to the periodicity of the superlattice structure, and a perfect matching layer (PML) was set in the z -direction. In addition, the polarization plane wave was vertically incident in the negative direction of the axis. Simulations were performed sequentially by changing the nanopillar structure to analyze the Fano resonance generation mechanism. Results and Discussions The designed full -dielectric supersurface has a high -quality factor, Q, and modulation depth. Flexible modulation from single-Fano resonance to double-Fano resonance can be achieved by increasing the number of nanocolumn rows. The transmission spectrum of the first simulated single -row nanocolumn and the electromagnetic field distribution show that the Fano resonance (Fig. 3) was generated by a toroidal dipole but with a decreased quality factor. The coupling between the nanocolumns can be modulated by increasing the number of nanorows such that the toroidal dipole (TD) and magnetic dipole (MD) jointly dominate the dark mode, thus increasing the quality factor and enhancing the near -field coupling (Fig. 6). The final increase to the three rows of nanopillars achieves a double-Fano resonance. The first Fano resonance peak is formed by the TD and electric dipole (ED) resonance when the scattering power values are equal, and both interfere to cancel out each other to produce a radiation -free anapole mode. The second Fano resonance peak is formed by the resonant interference of the TD and MD to form the dark mode, whereas the remaining resonant modes act in the dark mode. The interference between the two modes forms the Fano resonance peak (Fig. 9). The sensitivity of methane volume fraction and the background refractive index can be measured simultaneously, and the simulation calculations show that the sensor has a high sensitivity (Fig. 12). Conclusions Based on the high Fano resonance quality factor, Q, of the Fano resonance metasurface and the local characteristics of the electric field, this study designed a fully dielectric structure and investigated the relationship between the hypersurface structure parameters and the quality factor Q and modulation depth. Effective control of the single-Fano resonance to double-Fano resonance is achieved via essential parameter optimization. By comparing the sensitivities of the rectangular column and elliptical cylindrical metasurface structures with the same parameters, the rectangular column structure is selected, and its optimal parameters are determined. The quality factor of the rectangular column metasurface structure reaches 3408, and the modulation depth of the double resonance peak is close to 99%. Through simulations, CH4 volume fraction sensitivity values can reach 1.57 nm/ % (for dip1) and 1.66 nm/ % (for dip2). The background refractive index sensitivities are 419.45 and 395.7 nm/RIU, and the figure of merit (FOM) values are 524.3 and 542.8 RIU-1, respectively. Knowledge of linear algebra proves that the sensitivity error of the sensor is slight. In addition, the sensor can measure the deflection angle according to the magnitude of the resonance peak, and the manufacturing error tolerance of the sensor did not exceed 1.3 nm.

Inverse design of nanohole all-dielectric metasurface based on deep convolutional neural network

Neural Network-Based Surrogate Model for Inverse Design of Metasurfaces

Long short-term memory neural network for directly inverse design of nanofin metasurface.

Stacking integrated learning-based inverse design of four-nanopore high-Q all-dielectric metasurface

Inverse Design of Dual‐Band Optically Transparent Metasurface Absorbers with Neural‐Adjoint Method

End-to-End Diverse Metasurface Design and Evaluation Using an Invertible Neural Network

An Efficient Artificial Neural Network Model for Inverse Design of Metasurfaces

Multifunctional Sensor Design Based on Fano Resonance Metasurface

Direct Field-to-pattern Monolithic Design of Holographic Metasurface Via Residual Encoder-Decoder Convolutional Neural Network

Symmetry Enhanced Network Architecture Search for Complex Metasurface Design

Inverse design of plasmonic metasurfaces by convolutional neural network.

Inverse Design of Coupled Subwavelength Dielectric Resonators with Targeted Eigenfrequency and Q Factor Utilizing Deep Learning.

Deep-Learning-Empowered Inverse Design for Freeform Reconfigurable Metasurfaces

Design of optical meta-structures with applications to beam engineering using deep learning

Robust Large-scale Metasurface Inverse Design Using Phaseless Data with Invertible Neural Network

On-Demand Design of Metasurfaces through Multineural Network Fusion

Deep Learning-Enhanced Inverse Modeling of Terahertz Metasurface Based on a Convolutional Neural Network Technique

Harnessing artificial neural networks for inverse design and analysis of MEMS-based Fabry–Pérot filters

Multifunctional Metasurface Inverse Design Based on Ultra‐Wideband Spectrum Prediction Neural Network

Neural-adjoint method for the inverse design of all-dielectric metasurfaces

Inverse design of metasurface optical filters using deep neural network with high degrees of freedom