A temperature-insensitive graphene-water-based ultra-wideband terahertz metamaterials absorber designed using deep neural networks
Jing Li, Huanyang Chen, S. L. Yap, Binzhen Zhang
Abstract
To address the limitations of traditional wave-absorbing materials in electromagnetic wave absorption, this study utilizes the electromagnetic properties of water and graphene in the terahertz (THz) band, coupled with deep neural networks (DNN), to propose a temperature-insensitive, ultra-wideband (UWB) THz metamaterials absorber (MAs). Simulation results show that when the graphene Fermi level is E f = 0.9 eV, the absorber achieves an absorption rate exceeding 90 % over the 3.832 ∼ 9 THz frequency range. Analysis of the water-graphene composite structure reveals that the ultra-wideband absorption is primarily attributed to the coupling between the top-layer graphene and the dielectric water layer. A comprehensive investigation of the absorption mechanism is carried out using transmission line theory , impedance matching theory, and the analysis of field distribution and power loss. Moreover, the study demonstrates that adjusting the graphene Fermi level (0.01 ∼ 0.9 eV) enables flexible tuning of the absorber’s bandwidth and absorption performance . Additionally, the absorber remains stable across a temperature range of 0 ∼ 100 °C and exhibits wide-angle and polarization-insensitive absorption characteristics. With its simple structure, compact size, superior absorption performance, and tunability, this absorber shows great potential for applications in THz thermal imaging, radar stealth, smart switches, and electromagnetic radiation protection.