科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Journal of Nanophotonics2025-12-03· Materials science

Broadband solar energy harvesting with a machine-learning-optimized nickel–quartz metamaterial absorber

Md. Murad Kabir Nipun, Shilpa Rana, Md Golam Morshed, Khandaker Mohammad Raisul Amin, F. A. Sabbir Ahamed, Habeeb Faruk Khan

原始摘要(英文原文)· Original abstract
This work presents a high-performance metamaterial absorber (MMA) designed for efficient solar energy harvesting across a broad spectral range from 100 to 1000 nm. The absorber features a compact unit cell measuring 100×100×12 nm3, incorporating nickel as both the resonator and ground layers, with quartz serving as the dielectric substrate. High broadband absorption is achieved through a multi-faceted mechanism: First, the structural parameters are carefully optimized to realize near-perfect impedance matching with free space, thereby suppressing reflection at the interface. Second, the metal–insulator–metal configuration facilitates the strong hybridization of localized surface plasmon resonances, which collectively broaden the absorption spectrum. Finally, the introduction of the dielectric spacer enables strong magnetic resonance due to anti-parallel surface currents, which confine the magnetic field and significantly enhance dissipation within the nickel layers. The proposed MMA demonstrates exceptional broadband absorption exceeding 90% across the 130- to 1000-nm wavelength range, with an average absorption of 96.63% considering the entire 100- to 1000-nm span. Notably, a peak absorption of 97.57% is observed at 201 nm, whereas the proposed structure achieves over 99% absorption consistently within the 700- to 900-nm near-infrared region. Electromagnetic field analyzes encompassing electric field distribution, magnetic response, surface current flow, and effective parameter retrieval offer insights into the underlying absorption mechanisms. Excellent polarization insensitivity up to 60 deg has also been demonstrated by the proposed MMA. In addition, a machine learning framework was developed to enhance predictive modeling of the absorber’s behavior, using five advanced algorithms across three test scenarios. Among these, the extra trees regressor achieved superior performance with a validation accuracy of 99.91% and 99.96% for both resonator and substrate thickness, respectively. The absorber also yields a high solar absorption efficiency of 95.87%, indicating strong potential for use in next-generation solar energy harvesting applications.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Broadband solar energy harvesting with a machine-learning-optimized nickel–quartz metamaterial absorber — 科研速览 Science Skim