Machine-Learning-Assisted Design of Multiband Terahertz Metamaterial Absorber

Abstract

The design of metamaterials involves engineering the electromagnetic properties on a subwavelength scale, making their behavior challenging to predict with models that are timeconsuming and computationally expensive. In recent years, there has been growing interest in developing various machine learning (ML) models for designing metamaterials. In this research, we demonstrate an ML-based framework for the optimal design of metamaterial absorbers in the terahertz range. Herein, the optical properties of a metamaterial structure with an open polygonshaped meta-atom are studied using numerical simulations. In addition, machine learning models are developed, specifically, a forward and an inverse model to predict absorption spectra from the structure and vice versa using different machine learning algorithms, including neural networks, decision trees, KNN, and linear regression. The mean square error was as low as 0.003 for the forward model and 0.02 for the inverse model. This work can be expanded for designing other complex fabricated electromagnetic structures for use in metamaterials and metasurfaces.

AI-Based Metamaterial Design

“…In another research by Soni et al, the authors demonstrated an AI-driven framework for optimizing terahertz-range metamaterial absorbers, which also involved the development of ML models to predict absorption spectra based on structure and vice versa . Metamaterial absorbers (MMAs) designed for operation in the terahertz range (spanning from 0.1 to 10 THz) garnered substantial research attention due to their diverse applications including energy harvesting. , The research employed COMSOL Multiphysics’s wave optics module to simulate the proposed metamaterial perfect absorber (Figure F), which featured ten structural parameters, six of which were variables, encompassing four height parameters (h1, h2, h3, and h4), split-ring thickness (t), and split dimension (s). The simulations, operating in the frequency domain with Maxwell’s equations, established the relationship between electric and magnetic fields and their sources to gain insights into electromagnetic wave behavior, generating absorbance plots for about 6000 structural parameter configurations of the metamaterial structure.…”

“…Subfigures (A–E) are reproduced with permission from Elsevier . (F–H) Subfigures are reproduced with permission from the American Chemical Society …”