Deep learning: a new tool for photonic nanostructure design

Hegde, Ravi S.

doi:10.1039/c9na00656g

Cited by 107 publications

(65 citation statements)

References 88 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There have been several review articles that introduce the recent progresses in the inverse design of nanophotonic devices with the aid of deep learning. [172][173][174] In addition, optimization methods have also been widely employed in inverse design. Mature and efficient algorithmic optimization methods have been developed for finding optimal solutions with affordable computational sources, especially for large-scale and complex-shaped nanophotonic devices.…”

Section: ≈532 Nm Directional Raman Scatteringmentioning

confidence: 99%

Directional Control of Light with Nanoantennas

Lai

Lam

et al. 2020

Advanced Optical Materials

View full text Add to dashboard Cite

research fields, such as medicine, biology, and materials science. [1] These different types of optical components usually utilize the reflection, refraction, and diffraction of light to reshape the wavefronts and control the propagation direction of light in real space. [2] Although great achievements have been made in controlling light in various scopes, it is still challenging to manipulate the light direction at the subwavelength scale with the conventional optical devices. [3] The major constraint of controlling light at nanoscale is the diffraction limit of light, where the smallest resolvable feature is at the wavelength scale. The limited resolution hinders the observation of many intriguing phenomena and features at the subwavelength scale, such as biomolecules, cell components, nanoparticles and nanoscale devices. [4-7] New techniques are therefore highly demanded for subwavelength light manipulation. Nanoparticles have been introduced as a bridge to connect the gap between the macroscopic and the microscopic scale owing to their extraordinary optical properties in the manipulation of light. [8,9] In order to design nanoantennas with nanoparticles, many ideas have been borrowed from conventional macroscopic schemes. One of the most famous examples is Yagi-Uda antennas. This antenna structure was first invented by Hidetsugu Yagi and Shintaro Uda in 1926. [10] The traditional Yagi-Uda antennas are usually composed of multiple parallel elements in a line, usually halfwave dipoles made of metal rods, which function as a reflector, a driving element (feed), and several directors. Each director reradiates the electromagnetic waves from the driving element at a different phase, and the reradiated waves from the multiple directors superpose and interfere to enhance the radiation in a single direction. The Yagi-Uda antennas have been widely used in analog television, shortwave communication links, radar antennas, and broadcasting stations. Inspired by this idea, Yagi-Uda nanoantennas consisting of a reflector, a feed, and several directors have also been invented. As the nanoscale counterpart, Yagi-Uda nanoantennas can redirect light emission in a desired direction with a narrow angle and enhanced signal intensity in a nanoscale region. [11,12] The detailed analysis of different Yagi-Uda nanoantennas will be introduced in Sections 3-5. The inspiring work has opened a new era for directional light control at the subwavelength scale. After the development for a decade, the techniques for directional light scattering and emitting have been enriched. Many different Light manipulation has been widely employed in lighting, display, and energy storage, becoming an inseparable part of human lives. However, the conventional optical devices suffer from the diffraction limit of electromagnetic waves. To overcome the limitation, plasmonic and dielectric nanoantennas are introduced for the control of light direction at nanoscale. The directionality of the nanoantennas stems from their electromagnetic resonance properties or ...

show abstract

Section: ≈532 Nm Directional Raman Scatteringmentioning

confidence: 99%

Directional Control of Light with Nanoantennas

Lai

Lam

et al. 2020

Advanced Optical Materials

View full text Add to dashboard Cite

show abstract

“…The inverse mapping from the RS to DS has a nonunique nature [ 34 ] as there are potentially many sets of design parameters that can meet the design criteria. In unsupervised inverse‐design strategies, an ML model is trained to learn the characteristics of the good designs gathered in a training dataset and synthesize similar superior samples.…”

Section: Resultsmentioning

confidence: 99%

“…Recently, there has been a deluge of insightful reports regarding the ML-assisted inverse design of photonic nanostructures, [34][35][36] encouraged by the enormous advancements of nanofabrication technology and empowered by the massive computational resources that have become available to nanophotonic researchers and designers. Consequently, various machine learning algorithms have been applied to several applications.…”

Section: Introductionmentioning

confidence: 99%

TCO‐Based Active Dielectric Metasurfaces Design by Conditional Generative Adversarial Networks

Jafar‐Zanjani

Salary

Huynh

et al. 2020

Advcd Theory and Sims

View full text Add to dashboard Cite

While researchers in the field of active flat optics continue to make groundbreaking progress by seeking novel materials and control systems, the complexity and sensitivity of the nanostructures that they aspire to design are unavoidably increasing. Inverse design of the popular class of transparent conducting oxide (TCO)‐based active metasurfaces is particularly challenging, largely due to the limited choice of the active materials, and sensitive physical mechanisms that give rise to their tunability. In this contribution, a new machine learning method based on a combination of the K‐means clustering algorithm and conditional Wasserstein generative adversarial networks (cWGANs) for broadband multi‐modal inverse design of TCO‐based active metasurfaces is developed. Multi‐objective evolutionary optimization is adopted to efficiently generate a diverse training dataset of high‐performance active metasurfaces. The training dataset includes samples that operate at specific wavelengths throughout the optical telecommunications (telecom) band. K‐means algorithm is then used to extract the clusters (modes) present in the training dataset, and exclusive cWGAN models are fit on each of them. The model is capable of generating designs operating at wavelengths that are not present in the training dataset. It also provides a clear picture of the feasibility and interplay between the design objectives.

show abstract

“…Although, they have been widely used in nanophotonics design and optimization problems, they suffer from remarkable computation complexity and often result in local optimum (rather than global optimum) solutions; (ii) algorithms employing artificial neural networks (ANNs) to optimize topologies of nanophotonic structures. While being more reliable in providing the global optimum designs, such data-driven algorithms require a large amount of training instances to be practical for the real-world applications [249][250][251][252][253][254][255][256][257][258][259][260][261][262][263][264][265][266][267].…”

Section: Emergence Of Deep Learning In Analysis Design and Optimizamentioning

confidence: 99%

Tunable nanophotonics enabled by chalcogenide phase-change materials

et al. 2020

View full text Add to dashboard Cite

AbstractNanophotonics has garnered intensive attention due to its unique capabilities in molding the flow of light in the subwavelength regime. Metasurfaces (MSs) and photonic integrated circuits (PICs) enable the realization of mass-producible, cost-effective, and efficient flat optical components for imaging, sensing, and communications. In order to enable nanophotonics with multipurpose functionalities, chalcogenide phase-change materials (PCMs) have been introduced as a promising platform for tunable and reconfigurable nanophotonic frameworks. Integration of non-volatile chalcogenide PCMs with unique properties such as drastic optical contrasts, fast switching speeds, and long-term stability grants substantial reconfiguration to the more conventional static nanophotonic platforms. In this review, we discuss state-of-the-art developments as well as emerging trends in tunable MSs and PICs using chalcogenide PCMs. We outline the unique material properties, structural transformation, and thermo-optic effects of well-established classes of chalcogenide PCMs. The emerging deep learning-based approaches for the optimization of reconfigurable MSs and the analysis of light-matter interactions are also discussed. The review is concluded by discussing existing challenges in the realization of adjustable nanophotonics and a perspective on the possible developments in this promising area.

show abstract

Deep learning: a new tool for photonic nanostructure design

Abstract: We review recent progress in the application of Deep Learning (DL) techniques for photonic nanostructure design and provide a perspective on current limitations and fruitful directions for further development.

Cited by 107 publications

References 88 publications

Directional Control of Light with Nanoantennas

Directional Control of Light with Nanoantennas

TCO‐Based Active Dielectric Metasurfaces Design by Conditional Generative Adversarial Networks

Tunable nanophotonics enabled by chalcogenide phase-change materials

Contact Info

Product

Resources

About