Strategical Deep Learning for Photonic Bound States in the Continuum (Laser Photonics Rev. 16(10)/2022)

Ma, Xieyao; Ma, Yanhe; Cunha, Preston; Liu, Qiushi; Kudtarkar, Kaushik; Xu, Dianguo; Wang, Jiafei; Chen, Yixin; Wong, Zi Jing; Liu, Ming; Hipwell, M. Cynthia; Lan, Shoufeng

doi:10.1002/lpor.202270049

Cited by 3 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alternatively, DNN‐based nanophotonic structure design is a promising approach nowadays, which can offer a more efficient tool to investigate the multi‐dimensional BIC behavior. [ 52 ] By applying a seven‐layers fully connected neural network (FCNN) with the parameters of 5D structure as input and a high resolution 1000 points transmittance spectrum from 1100 to 1400 nm as output, we ran 12 705 simulations to construct the dataset, and only 80% of them were required for a well‐converged model training (the training data can even be much smaller, see Figure S8, Supporting Information) as illustrated in Figure 3d,e. More detailed information about the DNN algorithm can be found in the Experimental section.…”

Section: Resultsmentioning

confidence: 99%

Customizing 2.5D Out‐of‐Plane Architectures for Robust Plasmonic Bound‐States‐in‐the‐Continuum Metasurfaces

Wang

Sun

et al. 2023

Advanced Science

View full text Add to dashboard Cite

Bound states in the continuum (BICs) have a superior ability to confine electromagnetic waves and enhance light–matter interactions. However, the quality‐factor of quasi‐BIC is extremely sensitive to structural perturbations, thus the BIC metasurfaces usually require a very‐high precision nanofabrication technique that greatly restricts their practical applications. Here, distinctive 2.5D out‐of‐plane architectures based plasmonic symmetry protected (SP)‐BIC metasurfaces are proposed, which could deliver robust quality factors even with large structural perturbations. The high‐throughput fabrication of such SP‐BIC metasurfaces is realized by using the binary‐pore anodic aluminum oxide template technique. Moreover, the deep neural network (DNN) is adapted to conduct multiparameter fittings, where the 2.5D hetero‐out‐of‐plane architectures with robust high quality‐factors and figures of merit are rapidly predicted and fabricated. Finally, owning to its large second‐order surface sensitivity, the desired 2.5D hetero‐out‐of‐plane architecture demonstrates a detection limit of endotoxin as low as 0.01 EU mL‐1, showing a good perspective of biosensors and others.

show abstract

Section: Resultsmentioning

confidence: 99%

Customizing 2.5D Out‐of‐Plane Architectures for Robust Plasmonic Bound‐States‐in‐the‐Continuum Metasurfaces

Wang

Sun

et al. 2023

Advanced Science

View full text Add to dashboard Cite

show abstract

“…However, since a limited number of structural parameters fails to provide sufficient degrees of freedom, achieving multi‐objective designs for color routers proves challenging. Thanks to the rapid development of machine learning and deep learning, some optical designs that require great wisdom and complex mechanisms can be realized using artificial intelligence (AI), [ 26–32 ] especially generative AI. Notably, image‐based generative networks offer a high degree of design freedom and have played a significant role in optical design endeavors, such as unit cells in metasurfaces.…”

Section: Introductionmentioning

confidence: 99%

Design of Multifunctional Color Routers with Kerker Switching Using Generative Adversarial Networks

Yan,

Zhu,

Bao

et al. 2024

Laser & Photonics Reviews

View full text Add to dashboard Cite

To achieve optoelectronic devices with high resolution and efficiency, there is a pressing need for optical structural units that possess an ultrasmall footprint yet exhibit strong controllability in both the frequency and spatial domains. For dielectric nanoparticles, the overlap of electric and magnetic dipole moments can scatter light completely forward or backward, which is called Kerker theory. This effect can expand to any multipoles and any directions, re‐named as generalized Kerker effect, and realize controllable light manipulation at full space and full spectrum using well‐designed dielectric structures. However, the complex situations of multipole couplings make it difficult to achieve structural design. Here, generative artificial intelligence (AI) is utilized to facilitate multi‐objective‐oriented structural design, wherein the study leverages the concept of “combined spectra” that consider both spectra and direction ratios as labels. The proposed generative adversarial network (GAN) is named as DDGAN (double‐discriminator GAN) that discriminates both images and spectral labels. Using trained networks, the simultaneous design for scattering color and directivities, RGB color routers, as well as narrowband light routers is achieved. Notably, all generated structures possess a footprint <600 × 600 nm indicating their potential applications in optoelectronic devices with ultrahigh resolution.

show abstract

Inverse design in photonic crystals

Deng,

Liu,

Shi

2024

Nanophotonics

View full text Add to dashboard Cite

Photonic crystals are periodic dielectric structures that possess a wealth of physical characteristics. Owing to the unique way they interact with the light, they provide new degrees of freedom to precisely modulate the electromagnetic fields, and have received extensive research in both academia and industry. At the same time, fueled by the advances in computer science, inverse design strategies are gradually being used to efficiently produce on-demand devices in various domains. As a result, the interdisciplinary area combining photonic crystals and inverse design emerges and flourishes. Here, we review the recent progress for the application of inverse design in photonic crystals. We start with a brief introduction of the background, then mainly discuss the optimizations of various physical properties of photonic crystals, from eigenproperties to response-based properties, and end up with an outlook for the future directions. Throughout the paper, we emphasize some insightful works and their design algorithms, and aim to give a guidance for readers in this emerging field.

show abstract

Strategical Deep Learning for Photonic Bound States in the Continuum (Laser Photonics Rev. 16(10)/2022)

Cited by 3 publications

References 0 publications

Customizing 2.5D Out‐of‐Plane Architectures for Robust Plasmonic Bound‐States‐in‐the‐Continuum Metasurfaces

Customizing 2.5D Out‐of‐Plane Architectures for Robust Plasmonic Bound‐States‐in‐the‐Continuum Metasurfaces

Design of Multifunctional Color Routers with Kerker Switching Using Generative Adversarial Networks

Inverse design in photonic crystals

Contact Info

Product

Resources

About