Optics and Lasers in Engineering, Journal Year: 2025, Volume and Issue: 186, P. 108837 - 108837
Published: Jan. 16, 2025
Language: Английский
Citations
0Tristate Switching of Terahertz Metasurfaces Enabled by Transferable VO2
Laser & Photonics Review, Journal Year: 2025, Volume and Issue: unknown
Published: Jan. 22, 2025
Abstract Achieving dynamic switching among absorption (A), reflection (R), and transmission (T) states is not only essential for advancing the understanding of light‐metasurface interactions but also holds significant potential practical applications, such as selective electromagnetic shielding smart windows. However, at terahertz higher frequencies, implementing active elements in multilayer configurations presents challenges that are straightforward those encountered microwave range. In this work, it demonstrated tristate ART tuning can be realized a single‐layer, free‐standing metasurface by between dual dipolar mode (electric dipole magnetic dipole) single dipole). By transferring flexible vanadium dioxide (VO 2 ) thin film onto dielectric Huygens’ metasurface, modulation achieved, transitioning from near‐unity state to near‐perfect state, finally high‐reflection with up 0.65 during insulator‐to‐metal transition induced heating phase‐change material. The results may lead new approaches designing reconfigurable metasurfaces based on materials wavefront control applications.
Language: Английский
Citations
0Empowering nanophotonic applications via artificial intelligence: pathways, progress, and prospects
Nanophotonics, Journal Year: 2025, Volume and Issue: 14(4), P. 429 - 447
Published: Feb. 3, 2025
Abstract Empowering nanophotonic devices via artificial intelligence (AI) has revolutionized both scientific research methodologies and engineering practices, addressing critical challenges in the design optimization of complex systems. Traditional methods for developing are often constrained by high dimensionality spaces computational inefficiencies. This review highlights how AI-driven techniques provide transformative solutions enabling efficient exploration vast spaces, optimizing intricate parameter systems, predicting performance advanced materials with accuracy. By bridging gap between complexity practical implementation, AI accelerates discovery novel functionalities. Furthermore, we delve into emerging domains, such as diffractive neural networks quantum machine learning, emphasizing their potential to exploit photonic properties innovative strategies. The also examines AI’s applications areas, e.g., optical image recognition, showcasing its role device integration. facilitating development highly efficient, compact devices, these AI-powered paving way next-generation systems enhanced functionalities broader applications.
Language: Английский
Citations
0Physics-informed learning in artificial electromagnetic materials
Applied Physics Reviews, Journal Year: 2025, Volume and Issue: 12(1)
Published: March 1, 2025
The advent of artificial intelligence—deep neural networks (DNNs) in particular—has transformed traditional research methods across many disciplines. DNNs are data driven systems that use large quantities to learn patterns fundamental a process. In the realm electromagnetic materials (AEMs), common goal is discover connection between AEM's geometry and material properties predict resulting scattered fields. To achieve this goal, usually utilize computational simulations act as ground truth for training process, numerous successful results have been shown. Although demonstrated successes, they limited by their requirement lack interpretability. latter because black-box models, therefore, it unknown how or why work. A promising approach which may help mitigate aforementioned limitations physics guide development operation DNNs. Indeed, physics-informed learning (PHIL) has seen rapid last few years with some success addressing conventional We overview field PHIL discuss benefits incorporating knowledge into deep process introduce taxonomy enables us categorize various types approaches. also summarize principles critical understanding Appendix covers AEMs. specific works highlighted serve examples Finally, we provide an outlook detailing where currently what can expect future.
Language: Английский
Citations
0可重构超表面研究进展及应用（特邀）
Infrared and Laser Engineering, Journal Year: 2025, Volume and Issue: 54(3), P. 20240620 - 20240620
Published: Jan. 1, 2025
Citations
0Graph-based design of irregular metamaterials
International Journal of Mechanical Sciences, Journal Year: 2025, Volume and Issue: unknown, P. 110203 - 110203
Published: April 1, 2025
Language: Английский
Citations
0A dual-band programmable metasurface for terahertz beam steering
Applied Physics Letters, Journal Year: 2025, Volume and Issue: 126(19)
Published: May 12, 2025
Terahertz programmable metasurfaces hold significant promise for next-generation communications due to their capability steer electromagnetic waves. However, most existing terahertz operate at only a single frequency, leaving much of the vast spectrum underutilized. In this study, we introduce “butterfly” dual-band metasurface, integrated with liquid crystals, designed efficient beam steering. By applying bias voltages, metasurface achieves phase change nearly 270° two distinct frequencies, ∼400 and ∼700 GHz. Our experimental results demonstrate that butterfly is compatible both binary ternary coding schemes these remarkably enhancing beam-steering performance expanding spatial coverage. This advancement in technology marks step forward harnessing full potential spectrum, opening another pathway broadband communication imaging applications.
Language: Английский
Citations
0Advancements in Metasurfaces for Polarization Control: A Comprehensive Survey
Next research., Journal Year: 2025, Volume and Issue: unknown, P. 100407 - 100407
Published: May 1, 2025
Language: Английский
Citations
0Advances in artificial intelligence for artificial metamaterials
APL Materials, Journal Year: 2024, Volume and Issue: 12(12)
Published: Dec. 1, 2024
The 2024 Nobel Prizes in Physics and Chemistry were awarded for foundational discoveries inventions enabling machine learning through artificial neural networks. Artificial intelligence (AI) metamaterials are two cutting-edge technologies that have shown significant advancements applications various fields. AI, with its roots tracing back to Alan Turing’s seminal work, has undergone remarkable evolution over decades, key including the Turing Test, expert systems, deep learning, emergence of multimodal AI models. Electromagnetic wave control, critical scientific research industrial applications, been significantly broadened by metamaterials. This review explores synergistic integration metamaterials, emphasizing how accelerates design functionality materials, while novel physical networks constructed from enhance AI’s computational speed ability solve complex problems. paper provides a detailed discussion AI-based forward prediction inverse principles metamaterial design. It also examines potential big-data-driven methods addressing challenges In addition, this delves into role advancing focusing on progress electromagnetic optics, terahertz, microwaves. Emphasizing transformative impact intersection between underscores improvements efficiency, accuracy, applicability. collaborative development process opens new possibilities innovations photonics, communications, radars, sensing.
Language: Английский
Citations
2Physics-informed deep learning for 3D modeling of light diffraction from optical metasurfaces
Optics Express, Journal Year: 2024, Volume and Issue: 33(1), P. 1371 - 1371
Published: Dec. 4, 2024
We propose an alternative data-free deep learning method using a physics-informed neural network (PINN) to enable more efficient computation of light diffraction from 3D optical metasurfaces, modeling corresponding polarization effects, and wavefront manipulation. Our model learns only the governing physics represented by vector Maxwell's equations, Floquet-Bloch boundary conditions, perfectly matched layers (PML). PINN accurately simulates near-field far-field responses, impact polarization, meta-atom geometry, illumination settings on transmitted light. Once trained, PINN-based electromagnetic field (EMF) solver scattering response for multiple inputs within single inference pass several milliseconds. This approach offers significant speed-up compared traditional numerical solvers, along with improved accuracy data independence over data-driven networks.
Language: Английский
Citations
0