Building Multifunctional Metasystems <i>via</i> Algorithmic Construction

Zhu, Dayu; Liu, Zhaocheng; Raju, Lakshmi; Kim, Andrew S.; Cai, Wenshan

doi:10.1021/acsnano.0c09424

Cited by 47 publications

(16 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The stack—rather than individual metasurfaces—must be designed as one entity to perform the desired set of functions (Table S2 ). Such systems often prove computationally intensive to devise 28 – 30 , require precise lateral alignment of the constituent metasurfaces 29 , 31 , and do not support closely spaced operating wavelengths 31 , 32 .…”

Section: Resultsmentioning

confidence: 99%

Multifunctional resonant wavefront-shaping meta-optics based on multilayer and multi-perturbation nonlocal metasurfaces

et al. 2022

View full text Add to dashboard Cite

Photonic devices rarely provide both elaborate spatial control and sharp spectral control over an incoming wavefront. In optical metasurfaces, for example, the localized modes of individual meta-units govern the wavefront shape over a broad bandwidth, while nonlocal lattice modes extended over many unit cells support high quality-factor resonances. Here, we experimentally demonstrate nonlocal dielectric metasurfaces in the near-infrared that offer both spatial and spectral control of light, realizing metalenses focusing light exclusively over a narrowband resonance while leaving off-resonant frequencies unaffected. Our devices attain this functionality by supporting a quasi-bound state in the continuum encoded with a spatially varying geometric phase. We leverage this capability to experimentally realize a versatile platform for multispectral wavefront shaping where a stack of metasurfaces, each supporting multiple independently controlled quasi-bound states in the continuum, molds the optical wavefront distinctively at multiple wavelengths and yet stay transparent over the rest of the spectrum. Such a platform is scalable to the visible for applications in augmented reality and transparent displays.

show abstract

Section: Resultsmentioning

confidence: 99%

Multifunctional resonant wavefront-shaping meta-optics based on multilayer and multi-perturbation nonlocal metasurfaces

et al. 2022

View full text Add to dashboard Cite

show abstract

“…[50][51][52] Well trained on precollected data, these delicately constructed deep learning models can produce meta-atoms upon given optical requirements with high efficiency, fidelity, and diversity. By further pairing deep learning with genetic algorithms, [53,54] topology optimization, [55] or adjoint optimization, [56] the metaatoms generated by deep learning models can be refined with improved performance or assembled in fully functional metasurface devices.…”

Section: Introductionmentioning

confidence: 99%

“…In this work, we conceptually propose a statistical perspective to estimate the design capability of multifunctional metasurfaces and, consequently, demonstrate an end‐to‐end design pipeline to experimentally realize such physically limited optimal design in a given parameter space. So far, the multifunctional metasurface design, with either conventional parameter sweeps or algorithm‐based inverse methods, [ 54 ] has followed a two‐step, top‐down design strategy. First, the design targets are manually decoupled into several subfunctions, and the ideal optical response distribution on the metasurface plane is retrieved.…”

Section: Introductionmentioning

confidence: 99%

Pushing the Limits of Functionality‐Multiplexing Capability in Metasurface Design Based on Statistical Machine Learning

Xiong

et al. 2022

Advanced Materials

105

View full text Add to dashboard Cite

metasurface research has spawned new types of flat optical components including beam deflectors, [4] high-quality-factor diffractors, [5] wave plates, [6] lenses, [7,8] and holograms. [9,10] Compared with their bulky counterparts that often rely on the phase accumulation of light propagating through conventional media, metasurface-based components can efficiently manipulate light by a deep subwavelength interface, making them compact, easy for integration, and relatively low loss, especially when dielectric materials are used. [11] In addition, the large flexibility in metasurface design offers unique ability to control light for multipurposed tasks that are usually inaccessible based on natural materials, yielding prescribed optical responses to different combinations of light characteristics such as the frequency, phase, angle of incidence, and polarization. [12][13][14][15][16][17][18][19][20] The exotic optical properties of metasurfaces originate from the engineered light-structure interactions. The conventional design approach heavily relies on template-based parameter sweep via numerical simulations to find eligible meta-atom structures, which could be facilitated by physical principles such as plasmonic resonance, [21] Mie theory, [22] or Pancharatnam-Berry (PB) phase for circular polarized light. [23] In the case of multifunctional metasurfaces, the design routines often utilize an empirical decoupling of the meta-atom response using multimode resonance in dielectric pillars, [24] As 2D metamaterials, metasurfaces provide an unprecedented means to manipulate light with the ability to multiplex different functionalities in a single planar device. Currently, most pursuits of multifunctional metasurfaces resort to empirically accommodating more functionalities at the cost of increasing structural complexity, with little effort to investigate the intrinsic restrictions of given meta-atoms and thus the ultimate limits in the design. In this work, it is proposed to embed machine-learning models in both gradientbased and nongradient optimization loops for the automatic implementation of multifunctional metasurfaces. Fundamentally different from the traditional two-step approach that separates phase retrieval and meta-atom structural design, the proposed end-to-end framework facilitates full exploitation of the prescribed design space and pushes the multifunctional design capacity to its physical limit. With a single-layer structure that can be readily fabricated, metasurface focusing lenses and holograms are experimentally demonstrated in the near-infrared region. They show up to eight controllable responses subjected to different combinations of working frequencies and linear polarization states, which are unachievable by the conventional physics-guided approaches. These results manifest the superior capability of the data-driven scheme for photonic design, and will accelerate the development of complex devices and systems for optical display, communication, and computing.

show abstract

“…Nevertheless, the edge roughness of the designed 2D pattern meta‐atom, brought by the GAN feature, will bring great difficulty for the meta‐device fabrication when the target wavelength is extended from infrared light to visible light. Another scheme added the number of layers of meta‐atoms to realize the complex‐amplitude metasurfaces, [ 15 ] but the fabrication of meta‐device composed of multi‐layer structures is more difficult than its counterpart with the single‐layer structure. It can be found that the current deep learning schemes for designing complex‐amplitude metasurfaces often set higher requirements on fabrication technology, which hinders the practical development of metasurface devices.…”

Section: Introductionmentioning

confidence: 99%

Complex‐Amplitude Metasurface Design Assisted by Deep Learning

et al. 2022

View full text Add to dashboard Cite

Metasurfaces can enable powerful manipulations for electromagnetic waves, thus many exotic functionalities have been realized. However, constrained by the complexity of metasurface-modulated complex amplitudes (amplitude and phase), the design of complex-amplitude metasurfaces sets a high threshold for researchers because of the requirement of plenty of specialized knowledge. In this paper, a deep learning scheme that uses a forward surrogate network to assist complex-amplitude metasurface design is proposed. The model is simple to construct and easy to converge in training. Accordingly, two complex-amplitude multiplexing devices, which can simultaneously display a nanoprinting image at the device surface and one/two holographic images in the far field, are designed with the proposed network using the cross-shaped meta-atom. The results show that the demonstrated scheme can be used to design the complex-amplitude metasurface devices easily and effectively once training of the network is completed, and the single-layer smooth structure holds the advantage for the fabrication. The proposed method here is promising to realize designs of more complex-amplitude meta-devices, multi-polarization, and multi-wavelength multiplexing meta-devices.

show abstract

Building Multifunctional Metasystems via Algorithmic Construction

Cited by 47 publications

References 48 publications

Multifunctional resonant wavefront-shaping meta-optics based on multilayer and multi-perturbation nonlocal metasurfaces

Multifunctional resonant wavefront-shaping meta-optics based on multilayer and multi-perturbation nonlocal metasurfaces

Pushing the Limits of Functionality‐Multiplexing Capability in Metasurface Design Based on Statistical Machine Learning

Complex‐Amplitude Metasurface Design Assisted by Deep Learning

Contact Info

Product

Resources

About