Generative Model for the Inverse Design of Metasurfaces

How scanning probe microscopy can be supported by artificial intelligence and quantum computing?

The impact of Artificial Intelligence (AI) is rapidly expanding, revolutionizing both science and society. It is applied to practically all areas of life, science, and technology, including materials science, which continuously requires novel tools for effective materials characterization. One of the widely used techniques is scanning probe microscopy (SPM). SPM has fundamentally changed materials engineering, biology, and chemistry by providing tools for atomic-precision surface mapping. Despite its many advantages, it also has some drawbacks, such as long scanning times or the possibility of damaging soft-surface materials. In this paper, we focus on the potential for supporting SPM-based measurements, with an emphasis on the application of AI-based algorithms, especially Machine Learning-based algorithms, as well as quantum computing (QC). It has been found that AI can be helpful in automating experimental processes in routine operations, algorithmically searching for optimal sample regions, and elucidating structure-property relationships. Thus, it contributes to increasing the efficiency and accuracy of optical nanoscopy scanning probes. Moreover, the combination of AI-based algorithms and QC may have enormous potential to enhance the practical application of SPM. The limitations of the AI-QC-based approach were also discussed. Finally, we outline a research path for improving AI-QC-powered SPM. RESEARCH HIGHLIGHTS: Artificial intelligence and quantum computing as support for scanning probe microscopy. The analysis indicates a research gap in the field of scanning probe microscopy. The research aims to shed light into ai-qc-powered scanning probe microscopy.

Deep learning-assisted inverse design of metasurfaces for active color image tuning

Metasurfaces, artificial planar nanostructures, offer numerous advantages for color printing applications, including ultra-high resolution, resistance to fading, wide color gamut coverage, and multifunctional capabilities. Due to the sensitivity of their resonance spectra to the external environment, metasurfaces have garnered significant interest for color tuning applications. However, most existing approaches are limited to passive color tuning, wherein only the color changes passively while the composite color image remains unaltered. Active color image tuning, on the other hand, requires precise matching of both color and intensity to the designed targets before and after the tuning process. In this study, we propose a novel approach for active metasurface color image tuning by modulating the environmental refractive index. Building upon a forward neural network that establishes the relationship between the metasurface geometric parameters and color/intensity information, we employ a multi-objective inverse adjoint neural network. This network not only overcomes the inherent 'one-to-many' problem in inverse design using neural networks but also facilitates active color image tuning under three distinct environmental conditions. Our work provides a new approach for the inverse design of metasurfaces and opens up possibilities for applications in dynamic color printing, information encryption, and other related fields.

Localized Plasmonic Structured Illumination Microscopy Using Hybrid Inverse Design

Super-resolution fluorescence imaging has offered unprecedented insights and revolutionized our understanding of biology. In particular, localized plasmonic structured illumination microscopy (LPSIM) achieves video-rate super-resolution imaging with ∼50 nm spatial resolution by leveraging subdiffraction-limited nearfield patterns generated by plasmonic nanoantenna arrays. However, the conventional trial-and-error design process for LPSIM arrays is time-consuming and computationally intensive, limiting the exploration of optimal designs. Here, we propose a hybrid inverse design framework combining deep learning and genetic algorithms to refine LPSIM arrays. A population of designs is evaluated using a trained convolutional neural network, and a multiobjective optimization method optimizes them through iteration and evolution. Simulations demonstrate that the optimized LPSIM substrate surpasses traditional substrates, exhibiting higher reconstruction accuracy, robustness against noise, and increased tolerance for fewer measurements. This framework not only proves the efficacy of inverse design for tailoring LPSIM substrates but also opens avenues for exploring new plasmonic nanostructures in imaging applications.

On-Demand Design of Metasurfaces through Multineural Network Fusion

In this paper, a multineural network fusion freestyle metasurface on-demand design method is proposed. The on-demand design method involves rapidly generating corresponding metasurface patterns based on the user-defined spectrum. The generated patterns are then input into a simulator to predict their corresponding S-parameter spectrogram, which is subsequently analyzed against the real S-parameter spectrogram to verify whether the generated metasurface patterns meet the desired requirements. The methodology is based on three neural network models: a Wasserstein Generative Adversarial Network model with a U-net architecture (U-WGAN) for inverse structural design, a Variational Autoencoder (VAE) model for compression, and an LSTM + Attention model for forward S-parameter spectrum prediction validation. The U-WGAN is utilized for on-demand reverse structural design, aiming to rapidly discover high-fidelity metasurface patterns that meet specific electromagnetic spectrum responses. The VAE, as a probabilistic generation model, serves as a bridge, mapping input data to latent space and transforming it into latent variable data, providing crucial input for a forward S-parameter spectrum prediction model. The LSTM + Attention network, acting as a forward S-parameter spectrum prediction model, can accurately and efficiently predict the S-parameter spectrum corresponding to the latent variable data and compare it with the real spectrum. In addition, the digits "0" and "1" are used in the design to represent vacuum and metallic materials, respectively, and a 10 × 10 cell array of freestyle metasurface patterns is constructed. The significance of the research method proposed in this paper lies in the following: (1) The freestyle metasurface design significantly expands the possibility of metamaterial design, enabling the creation of diverse metasurface structures that are difficult to achieve with traditional methods. (2) The on-demand design approach can generate high-fidelity metasurface patterns that meet the expected electromagnetic characteristics and responses. (3) The fusion of multiple neural networks demonstrates high flexibility, allowing for the adjustment of network structures and training methods based on specific design requirements and data characteristics, thus better accommodating different design problems and optimization objectives.

Gated Recurrent Neural Network for Predicting the Plasmonic Colloid Composition from Spectra

In current research on the synthesis of colloidal nanostructures, the size and morphology of nanoparticles still exhibit certain dispersion and variation from batch to batch. Characterization of size distribution and morphology distribution of nanoparticles often requires techniques such as scanning electron microscopy or transmission electron microscopy, which involve high vacuum environments, are time-consuming, and costly. Experienced researchers can roughly estimate the size and distribution of nanostructure from spectra for a given synthetic route, but the accuracy is often limited. This paper reports the potential of using neural networks to accurately predict the composition of colloidal nanostructures from spectra. We address several fundamental issues in neural network prediction of colloidal composition. We first demonstrate the prediction of the composition of a colloidal binary mixture of gold nanoparticles using a gated recurrent neural network (GRU). The evolution of prediction errors for scattering, absorption, and extinction spectra of nanostructures with sizes ranging from 5 to 120 nm are analyzed. Furthermore, we demonstrate that the neural network model operates robustly under white noise in experimental testing scenarios. Compared to fully connected neural networks, the gated recurrent unit exhibits better testing accuracy in spectral prediction. When confronted with experimental data that deviates from simulation outputs, minor adjustments to the training set can allow the predictions to align closely with the experimental spectra, paving the way for the characterization of complex colloidal compositions with artificial intelligence.

Tbps wide-field parallel optical wireless communications based on a metasurface beam splitter

Optical wireless communication (OWC) stands out as one of the most promising technologies in the sixth-generation (6G) mobile networks. The establishment of high-quality optical links between transmitters and receivers plays a crucial role in OWC performances. Here, by a compact beam splitter composed of a metasurface and a fiber array, we proposed a wide-angle (~120°) OWC optical link scheme that can parallelly support up to 144 communication users. Utilizing high-speed optical module sources and wavelength division multiplexing technique, we demonstrated each user can achieve a communication speed of 200 Gbps which enables the entire system to support ultra-high communication capacity exceeding 28 Tbps. Furthermore, utilizing the metasurface polarization multiplexing, we implemented a full range wide-angle OWC without blind area nor crosstalk among users. Our OWC scheme simultaneously possesses the advantages of high-speed, wide communication area and multi-user parallel communications, paving the way for revolutionary high-performance OWC in the future.

Physics-Informed Inverse Design of Programmable Metasurfaces

Emerging reconfigurable metasurfaces offer various possibilities for programmatically manipulating electromagnetic waves across spatial, spectral, and temporal domains, showcasing great potential for enhancing terahertz applications. However, they are hindered by limited tunability, particularly evident in relatively small phase tuning over 270°, due to the design constraints with time-intensive forward design methodologies. Here, a multi-bit programmable metasurface is demonstrated capable of terahertz beam steering facilitated by a developed physics-informed inverse design (PIID) approach. Through integrating a modified coupled mode theory (MCMT) into residual neural networks, the PIID algorithm not only significantly increases the design accuracy compared to conventional neural networks but also elucidates the intricate physical relations between the geometry and the modes. Without decreasing the reflection intensity, the method achieves the enhanced phase tuning as large as 300°. Additionally, the inverse-designed programmable beam steering metasurface is experimentally validated, which is adaptable across 1-bit, 2-bit, and tri-state coding schemes, yielding a deflection angle up to 68° and broadened steering coverage. The demonstration provides a promising pathway for rapidly exploring advanced metasurface devices, with potentially great impact on communication and imaging technologies.

A surrogate-assisted extended generative adversarial network for parameter optimization in free-form metasurface design

Metasurfaces have widespread applications in fifth-generation (5G) microwave communication. Among the metasurface family, free-form metasurfaces excel in achieving intricate spectral responses compared to regular-shape counterparts. However, conventional numerical methods for free-form metasurfaces are time-consuming and demand specialized expertise. Alternatively, recent studies demonstrate that deep learning has great potential to accelerate and refine metasurface designs. Here, we present XGAN, an extended generative adversarial network (GAN) with a surrogate for high-quality free-form metasurface designs. The proposed surrogate provides a physical constraint to XGAN so that XGAN can accurately generate metasurfaces monolithically from input spectral responses. In comparative experiments involving 20000 free-form metasurface designs, XGAN achieves 0.9734 average accuracy and is 500 times faster than the conventional methodology. This method facilitates the metasurface library building for specific spectral responses and can be extended to various inverse design problems, including optical metamaterials, nanophotonic devices, and drug discovery.

Generative models struggle with kirigami metamaterials

Generative machine learning models have shown notable success in identifying architectures for metamaterials-materials whose behavior is determined primarily by their internal organization-that match specific target properties. By examining kirigami metamaterials, in which dependencies between cuts yield complex design restrictions, we demonstrate that this perceived success in the employment of generative models for metamaterials might be akin to survivorship bias. We assess the performance of the four most popular generative models-the Variational Autoencoder (VAE), the Generative Adversarial Network (GAN), the Wasserstein GAN (WGAN), and the Denoising Diffusion Probabilistic Model (DDPM)-in generating kirigami structures. Prohibiting cut intersections can prevent the identification of an appropriate similarity measure for kirigami metamaterials, significantly impacting the effectiveness of VAE and WGAN, which rely on the Euclidean distance-a metric shown to be unsuitable for considered geometries. This imposes significant limitations on employing modern generative models for the creation of diverse metamaterials.

Specific wavelength peak emulation with amorphous metastructures

The conventional design process for metasurfaces is time-consuming and computationally expensive. To address this challenge, we utilize a deep convolutional generative adversarial network (DCGAN) to generate new nanohole metastructure designs that match a desired transmittance spectrum in the visible range. The trained DCGAN model demonstrates an exceptional performance in generating diverse and manufacturable metastructure designs that closely resemble the target optical properties. The proposed method provides several advantages over existing approaches. These include its capability to generate new designs without prior knowledge or assumptions regarding the relationship between metastructure geometries and optical properties, its high efficiency, and its generalizability to other types of metamaterials. The successful fabrication and experimental characterization of the predicted metastructures further validate the accuracy and effectiveness of our proposed method.

Recent advances in the metamaterial and metasurface-based biosensor in the gigahertz, terahertz, and optical frequency domains

Recently, metamaterials and metasurface have gained rapidly increasing attention from researchers due to their extraordinary optical and electrical properties. Metamaterials are described as artificially defined periodic structures exhibiting negative permittivity and permeability simultaneously. Whereas metasurfaces are the 2D analogue of metamaterials in the sense that they have a small but not insignificant depth. Because of their high optical confinement and adjustable optical resonances, these artificially engineered materials appear as a viable photonic platform for biosensing applications. This review paper discusses the recent development of metamaterial and metasurface in biosensing applications based on the gigahertz, terahertz, and optical frequency domains encompassing the whole electromagnetic spectrum. Overlapping features such as material selection, structure, and physical mechanisms were considered during the classification of our biosensing applications. Metamaterials and metasurfaces working in the GHz range provide prospects for better sensing of biological samples, THz frequencies, falling between GHz and optical frequencies, provide unique characteristics for biosensing permitting the exact characterization of molecular vibrations, with an emphasis on molecular identification, label-free analysis, and imaging of biological materials. Optical frequencies on the other hand cover the visible and near-infrared regions, allowing fine regulation of light-matter interactions enabling metamaterials and metasurfaces to offer excellent sensitivity and specificity in biosensing. The outcome of the sensor's sensitivity to an electric or magnetic field and the resonance frequency are, in theory, determined by the frequency domain and features. Finally, the challenges and possible future perspectives in biosensing application areas have been presented that use metamaterials and metasurfaces across diverse frequency domains to improve sensitivity, specificity, and selectivity in biosensing applications.

Full-Color Imaging System Based on the Joint Integration of a Metalens and Neural Network

Lenses have been a cornerstone of optical systems for centuries; however, they are inherently limited by the laws of physics, particularly in terms of size and weight. Because of their characteristic light weight, small size, and subwavelength modulation, metalenses have the potential to miniaturize and integrate imaging systems. However, metalenses still face the problem that chromatic aberration affects the clarity and accuracy of images. A high-quality image system based on the end-to-end joint optimization of a neural network and an achromatic metalens is demonstrated in this paper. In the multi-scale encoder-decoder network, both the phase characteristics of the metalens and the hyperparameters of the neural network are optimized to obtain high-resolution images. The average peak-signal-to-noise ratio (PSNR) and average structure similarity (SSIM) of the recovered images reach 28.53 and 0.83. This method enables full-color and high-performance imaging in the visible band. Our approach holds promise for a wide range of applications, including medical imaging, remote sensing, and consumer electronics.

Tandem neural network-assisted inverse design of highly efficient diffractive slanted waveguide grating

Virtual reality devices featuring diffractive grating components have emerged as hotspots in the field of near-to-eye displays. The core aim of our work is to streamline the intricacies involved in devising the highly efficient slanted waveguide grating using the deep-learning-driven inverse design technique. We propose and establish a tandem neural network (TNN) comprising a generative flow-based invertible neural network and a fully connected neural network. The proposed TNN can automatically optimize the coupling efficiencies of the proposed grating at multi-wavelengths, including red, green, and blue beams at incident angles in the range of 0°-15°. The efficiency indicators manifest in the peak transmittance, average transmittance, and illuminance uniformity, reaching approximately 100%, 92%, and 98%, respectively. Additionally, the structural parameters of the grating can be deduced inversely based on the indicators within a short duration of hundreds of milliseconds to seconds using the TNN. The implementation of the inverse-engineered grating is anticipated to serve as a paradigm for simplifying and expediting the development of diverse types of waveguide gratings.

A deep learning method for empirical spectral prediction and inverse design of all-optical nonlinear plasmonic ring resonator switches

All-optical plasmonic switches (AOPSs) utilizing surface plasmon polaritons are well-suited for integration into photonic integrated circuits (PICs) and play a crucial role in advancing all-optical signal processing. The current AOPS design methods still rely on trial-and-error or empirical approaches. In contrast, recent deep learning (DL) advances have proven highly effective as computational tools, offering an alternative means to accelerate nanophotonics simulations. This paper proposes an innovative approach utilizing DL for spectrum prediction and inverse design of AOPS. The switches employ circular nonlinear plasmonic ring resonators (NPRRs) composed of interconnected metal-insulator-metal waveguides with a ring resonator. The NPRR switching performance is shown using the nonlinear Kerr effect. The forward model presented in this study demonstrates superior computational efficiency when compared to the finite-difference time-domain method. The model analyzes various structural parameters to predict transmission spectra with a distinctive dip. Inverse modeling enables the prediction of design parameters for desired transmission spectra. This model provides a rapid estimation of design parameters, offering a clear advantage over time-intensive conventional optimization approaches. The loss of prediction for both the forward and inverse models, when compared to simulations, is exceedingly low and on the order of 10-4. The results confirm the suitability of employing DL for forward and inverse design of AOPSs in PICs.

A universal inverse design methodology for microfluidic mixers

The intelligent design of microfluidic mixers encompasses both the automation of predicting fluid performance and the structural design of mixers. This article delves into the technical trajectory of computer-aided design for micromixers, leveraging artificial intelligence algorithms. We propose an automated micromixer design methodology rooted in cost-effective artificial neural network (ANN) models paired with inverse design algorithms. Initially, we introduce two inverse design methods for micromixers: one that combines ANN with multi-objective genetic algorithms, and another that fuses ANN with particle swarm optimization algorithms. Subsequently, using two benchmark micromixers as case studies, we demonstrate the automatic derivation of micromixer structural parameters. Finally, we automatically design and optimize 50 sets of micromixer structures using the proposed algorithms. The design accuracy is further enhanced by analyzing the inverse design algorithm from a statistical standpoint.

Data-Driven Design for Metamaterials and Multiscale Systems: A Review

Metamaterials are artificial materials designed to exhibit effective material parameters that go beyond those found in nature. Composed of unit cells with rich designability that are assembled into multiscale systems, they hold great promise for realizing next-generation devices with exceptional, often exotic, functionalities. However, the vast design space and intricate structure-property relationships pose significant challenges in their design. A compelling paradigm that could bring the full potential of metamaterials to fruition is emerging: data-driven design. This review provides a holistic overview of this rapidly evolving field, emphasizing the general methodology instead of specific domains and deployment contexts. Existing research is organized into data-driven modules, encompassing data acquisition, machine learning-based unit cell design, and data-driven multiscale optimization. The approaches are further categorized within each module based on shared principles, analyze and compare strengths and applicability, explore connections between different modules, and identify open research questions and opportunities.

More Than 30 000-fold Field Enhancement of Terahertz Nanoresonators Enabled by Rapid Inverse Design

The rapid development of 6G communications using terahertz (THz) electromagnetic waves has created a demand for highly sensitive THz nanoresonators capable of detecting these waves. Among the potential candidates, THz nanogap loop arrays show promising characteristics but require significant computational resources for accurate simulation. This requirement arises because their unit cells are 10 times smaller than millimeter wavelengths, with nanogap regions that are 1 000 000 times smaller. To address this challenge, we propose a rapid inverse design method using physics-informed machine learning, employing double deep Q-learning with an analytical model of the THz nanogap loop array. In ∼39 h on a middle-level personal computer, our approach identifies the optimal structure through 200 000 iterations, achieving an experimental electric field enhancement of 32 000 at 0.2 THz, 300% stronger than prior results. Our analytical model-based approach significantly reduces the amount of computational resources required, offering a practical alternative to numerical simulation-based inverse design for THz nanodevices.

Intelligent Phase Contrast Meta-Microscope System

Phase contrast imaging techniques enable the visualization of disparities in the refractive index among various materials. However, these techniques usually come with a cost: the need for bulky, inflexible, and complicated configurations. Here, we propose and experimentally demonstrate an ultracompact meta-microscope, a novel imaging platform designed to accomplish both optical and digital phase contrast imaging. The optical phase contrast imaging system is composed of a pair of metalenses and an intermediate spiral phase metasurface located at the Fourier plane. The performance of the system in generating edge-enhanced images is validated by imaging a variety of human cells, including lung cell lines BEAS-2B, CLY1, and H1299 and other types. Additionally, we integrate the ResNet deep learning model into the meta-microscope to transform bright-field images into edge-enhanced images with high contrast accuracy. This technology promises to aid in the development of innovative miniature optical systems for biomedical and clinical applications.

Designing Metasurfaces for Efficient Solar Energy Conversion

Metasurfaces have recently emerged as a promising technological platform, offering unprecedented control over light by structuring materials at the nanoscale using two-dimensional arrays of subwavelength nanoresonators. These metasurfaces possess exceptional optical properties, enabling a wide variety of applications in imaging, sensing, telecommunication, and energy-related fields. One significant advantage of metasurfaces lies in their ability to manipulate the optical spectrum by precisely engineering the geometry and material composition of the nanoresonators' array. Consequently, they hold tremendous potential for efficient solar light harvesting and conversion. In this Review, we delve into the current state-of-the-art in solar energy conversion devices based on metasurfaces. First, we provide an overview of the fundamental processes involved in solar energy conversion, alongside an introduction to the primary classes of metasurfaces, namely, plasmonic and dielectric metasurfaces. Subsequently, we explore the numerical tools used that guide the design of metasurfaces, focusing particularly on inverse design methods that facilitate an optimized optical response. To showcase the practical applications of metasurfaces, we present selected examples across various domains such as photovoltaics, photoelectrochemistry, photocatalysis, solar-thermal and photothermal routes, and radiative cooling. These examples highlight the ways in which metasurfaces can be leveraged to harness solar energy effectively. By tailoring the optical properties of metasurfaces, significant advancements can be expected in solar energy harvesting technologies, offering new practical solutions to support an emerging sustainable society.

All-optical geometric image transformations enabled by ultrathin metasurfaces

Image processing plays a vital role in artificial visual systems, which have diverse applications in areas such as biomedical imaging and machine vision. In particular, optical analog image processing is of great interest because of its parallel processing capability and low power consumption. Here, we present ultra-compact metasurfaces performing all-optical geometric image transformations, which are essential for image processing to correct image distortions, create special image effects, and morph one image into another. We show that our metasurfaces can realize binary image transformations by modifying the spatial relationship between pixels and converting binary images from Cartesian to log-polar coordinates with unparalleled advantages for scale- and rotation-invariant image preprocessing. Furthermore, we extend our approach to grayscale image transformations and convert an image with Gaussian intensity profile into another image with flat-top intensity profile. Our technique will potentially unlock new opportunities for various applications such as target tracking and laser manufacturing.

A Design Strategy for Surface Nanostructures to Realize Sensitive Refractive-Index Optical Sensors

Refractive-index optical sensors have been extensively studied. Originally, they were surface plasmon resonance sensors using only a flat gold film. Currently, to develop practically useful label-free optical sensors, numerous proposals for refractive index sensors have been made using various nanostructures composed of metals and dielectrics. In this study, we explored a rational design strategy for sensors using surface nanostructures comprising metals or dielectrics. Optical responses, such as reflection and transmission, and resonant electromagnetic fields were computed using a numerical method of rigorous coupled-wave analysis combined with a scattering-matrix algorithm. As a result, good performance that almost reached the physical limit was achieved using a plasmonic surface lattice structure. Furthermore, to precisely trace the refractive-index change, a scheme using two physical quantities, resonant wavelength and reflection amplitude, was found to be valid for a 2D silicon metasurface.

Deep Learning-Based Metasurface Design for Smart Cooling of Spacecraft

A reconfigurable metasurface constitutes an important block of future adaptive and smart nanophotonic applications, such as adaptive cooling in spacecraft. In this paper, we introduce a new modeling approach for the fast design of tunable and reconfigurable metasurface structures using a convolutional deep learning network. The metasurface structure is modeled as a multilayer image tensor to model material properties as image maps. We avoid the dimensionality mismatch problem using the operating wavelength as an input to the network. As a case study, we model the response of a reconfigurable absorber that employs the phase transition of vanadium dioxide in the mid-infrared spectrum. The feed-forward model is used as a surrogate model and is subsequently employed within a pattern search optimization process to design a passive adaptive cooling surface leveraging the phase transition of vanadium dioxide. The results indicate that our model delivers an accurate prediction of the metasurface response using a relatively small training dataset. The proposed patterned vanadium dioxide metasurface achieved a 28% saving in coating thickness compared to the literature while maintaining reasonable emissivity contrast at 0.43. Moreover, our design approach was able to overcome the non-uniqueness problem by generating multiple patterns that satisfy the design objectives. The proposed adaptive metasurface can potentially serve as a core block for passive spacecraft cooling applications. We also believe that our design approach can be extended to cover a wider range of applications.

Real-data-driven real-time reconfigurable microwave reflective surface

Manipulating the electromagnetic (EM) scattering behavior from an arbitrary surface dynamically on arbitrary design goals is an ultimate ambition for many EM stealth and communication problems, yet it is nearly impossible to accomplish with conventional analysis and optimization techniques. Here we present a reconfigurable conformal metasurface prototype as well as a workflow that enables it to respond to multiple design targets on the reflection pattern with extremely low on-site computing power and time. The metasurface is driven by a sequential tandem neural network which is pre-trained using actual experimental data, avoiding any possible errors that may arise from calculation, simulation, or manufacturing tolerances. This platform empowers the surface to operate accurately in a complex environment including varying incident angle and operating frequency, or even with other scatterers present close to the surface. The proposed data-driven approach requires minimum amount of prior knowledge and human effort yet provides maximized versatility on the reflection control, stepping towards the end form of intelligent tunable EM surfaces.

Unifying the design space and optimizing linear and nonlinear truss metamaterials by generative modeling

The rise of machine learning has fueled the discovery of new materials and, especially, metamaterials-truss lattices being their most prominent class. While their tailorable properties have been explored extensively, the design of truss-based metamaterials has remained highly limited and often heuristic, due to the vast, discrete design space and the lack of a comprehensive parameterization. We here present a graph-based deep learning generative framework, which combines a variational autoencoder and a property predictor, to construct a reduced, continuous latent representation covering an enormous range of trusses. This unified latent space allows for the fast generation of new designs through simple operations (e.g., traversing the latent space or interpolating between structures). We further demonstrate an optimization framework for the inverse design of trusses with customized mechanical properties in both the linear and nonlinear regimes, including designs exhibiting exceptionally stiff, auxetic, pentamode-like, and tailored nonlinear behaviors. This generative model can predict manufacturable (and counter-intuitive) designs with extreme target properties beyond the training domain.

Inverse design of metasurfaces with customized transmission characteristics of frequency band based on generative adversarial networks

In recent years, deep learning (DL) has demonstrated significant potential in the inverse design of metasurfaces, and the generation of metasurfaces with customized transmission characteristics of frequency band remains a challenging and underexplored area. In this study, we propose a DL-assisted method for the inverse design of transmissive metasurfaces. The method consists of a generative adversarial network (GAN)-based graph generator, an electromagnetic response predictor, and a genetic algorithm optimizer. By integrating these components, we can obtain customized metasurfaces with desired transmission characteristics of frequency band. We demonstrate the effectiveness of the proposed method through examples of inverse-designed three-layer cascaded transmissive metasurfaces with wideband, dual-band, and stopband responses in the 8∼12 GHz frequency range. Specifically, we realize three different types of dual-band metasurfaces, namely double-wide, front-wide and rear-narrow, and front-narrow and rear-wide configurations. Additionally, we analyze the accuracy and reliability of the inverse design method by employing data from the training dataset, self-defined objectives, and bandwidth-reduced target responses scaled from the wideband type as design inputs. Quantitative evaluation is performed using metrics such as mean absolute error and average precision. The proposed method successfully achieves the desired effect as intended.

Tensor completion algorithm-aided structural color design

In recent years, structural color has developed rapidly due to its distinct advantages, such as low loss, high spatial resolution and environmental friendliness. Various inverse design methods have been extensively investigated to efficiently design optical structures. However, the optimization method for the inverse design of structural color remains a formidable challenge. Traditional optimization approaches, such as genetic algorithms require time-consuming repetitions of structural simulations. Deep learning-assisted design necessitates prior simulations and large amounts of data, making it less efficient for systems with a small number of features. This study proposes a tensor completion algorithm capable of swiftly and accurately predicting missing datasets based on partially obtained datasets to assist in structural color design. Transforming the complex physical problem of structural color design into a spatial structure relationship problem linking geometric parameters and spectral data. The method utilizes tensor multilinear data analysis to effectively capture the complex relationships associated with geometric parameters and spectral data in higher-order data. Numerical and experimental results demonstrate that the algorithm exhibits high reliability in terms of speed and accuracy for diverse structures, datasets of varying sizes, and different materials, significantly enhancing design efficiency. The proposed algorithm offers a viable solution for inverse design problems involving complex physical systems, thereby introducing a novel approach to the design of photonic devices. Additionally, numerical experiments illustrate that the structural color of cruciform resonators with diamond can overcome the high loss issues observed in traditional dielectric materials within the blue wavelength region and enhance the corrosion resistance of the structure. We achieve a wide color gamut and a high-narrow reflection spectrum nearing 1 by this structure, and the theoretical analysis further verifies that diamond holds great promise in the realm of optics.

Deep-Learning-Assisted Simultaneous Target Sensing and Super-Resolution Imaging

Metasurfaces have recently experienced revolutionary progress in sensing and super-resolution imaging fields, mainly due to their manipulation of electromagnetic waves on subwavelength scales. However, on the one hand, the addition of metasurfaces can multiply the complexity of retrieving target information from detected electromagnetic fields. On the other hand, many existing studies utilize deep learning methods to provide compelling tools for electromagnetic problems but mainly concentrate on resolving one single function, limiting their versatilities. In this work, a multifunctional deep learning network is demonstrated to reconstruct diverse target information in a metasurface-target interactive system. First, a preliminary experiment verifies that the metasurface-involved scenario can tolerate the system noises. Then, the captured electric field distributions are fed into the multifunctional network, which can not only accurately sense the quantity and relative permittivity of targets but also generate super-resolution images precisely. The deep learning network, thus, paves an alternative way to recover the targets' information in metasurface-target interactive systems, accelerating the progression of target sensing and superimaging areas. Besides, another new network that allows forward electromagnetic prediction is also proposed and demonstrated. To sum up, the deep learning methodology may hold promise for inverse reconstructions or forward predictions in many electromagnetic scenarios.

Revisiting the Design Strategies for Metasurfaces: Fundamental Physics, Optimization, and Beyond

Over the last two decades, the capabilities of metasurfaces in light modulation with subwavelength thickness have been proven, and metasurfaces are expected to miniaturize conventional optical components and add various functionalities. Herein, various metasurface design strategies are reviewed thoroughly. First, the scalar diffraction theory is revisited to provide the basic principle of light propagation. Then, widely used design methods based on the unit-cell approach are discussed. The methods include a set of simplified steps, including the phase-map retrieval and meta-atom unit-cell design. Then, recently emerging metasurfaces that may not be accurately designed using unit-cell approach are introduced. Unconventional metasurfaces are examined where the conventional design methods fail and finally potential design methods for such metasurfaces are discussed.

Convolutional neural networks for mode on-demand high finesse optical resonator design

We demonstrate the use of machine learning through convolutional neural networks to solve inverse design problems of optical resonator engineering. The neural network finds a harmonic modulation of a spherical mirror to generate a resonator mode with a given target topology ("mode on-demand"). The procedure allows us to optimize the shape of mirrors to achieve a significantly enhanced coupling strength and cooperativity between a resonator photon and a quantum emitter located at the center of the resonator. In a second example, a double-peak mode is designed which would enhance the interaction between two quantum emitters, e.g., for quantum information processing.

End-to-End Diverse Metasurface Design and Evaluation Using an Invertible Neural Network

Employing deep learning models to design high-performance metasurfaces has garnered significant attention due to its potential benefits in terms of accuracy and efficiency. A deep learning-based metasurface design framework typically comprises a forward prediction path for predicting optical responses and a backward retrieval path for generating geometrical configurations. In the forward design path, a specific geometrical configuration corresponds to a unique optical response. However, in the inverse design path, a single performance metric can correspond to multiple potential designs. This one-to-many mapping poses a significant challenge for deep learning models and can potentially impede their performance. Although representing the inverse path as a probabilistic distribution is a widely adopted method for tackling this problem, accurately capturing the posterior distribution to encompass all potential solutions remains an ongoing challenge. Furthermore, in most pioneering works, the forward and backward paths are captured using separate models. However, the knowledge acquired from the forward path does not contribute to the training of the backward model. This separation of models adds complexity to the system and can hinder the overall efficiency and effectiveness of the design framework. Here, we utilized an invertible neural network (INN) to simultaneously model both the forward and inverse process. Unlike other frameworks, INN focuses on the forward process and implicitly captures a probabilistic model for the inverse process. Given a specific optical response, the INN enables the recovery of the complete posterior over the parameter space. This capability allows for the generation of novel designs that are not present in the training data. Through the integration of the INN with the angular spectrum method, we have developed an efficient and automated end-to-end metasurface design and evaluation framework. This novel approach eliminates the need for human intervention and significantly speeds up the design process. Utilizing this advanced framework, we have effectively designed high-efficiency metalenses and dual-polarization metasurface holograms. This approach extends beyond dielectric metasurface design, serving as a general method for modeling optical inverse design problems in diverse optical fields.

Design of a High-Performance Titanium Nitride Metastructure-Based Solar Absorber Using Quantum Computing-Assisted Optimization

Metastructures of titanium nitride (TiN), a plasmonic refractory material, can potentially achieve high solar absorptance while operating at elevated temperatures, but the design has been driven by expert intuition. Here, we design a high-performance solar absorber based on TiN metastructures using quantum computing-assisted optimization. The optimization scheme includes machine learning, quantum annealing, and optical simulation in an iterative cycle. It designs an optimal structure with solar absorptance > 95% within 40 h, much faster than an exhaustive search. Analysis of electric field distributions demonstrates that combined effects of Fabry-Perot interferences and surface plasmonic resonances contribute to the broadband high absorption efficiency of the optimally designed metastructure. The designed absorber may exhibit great potential for solar energy harvesting applications, and the optimization scheme can be applied to the design of other complex functional materials.

Emulating the Deutsch-Josza algorithm with an inverse-designed terahertz gradient-index lens

An all-dielectric photonic metastructure is investigated for application as a quantum algorithm emulator (QAE) in the terahertz frequency regime; specifically, we show implementation of the Deustsh-Josza algorithm. The design for the QAE consists of a gradient-index (GRIN) lens as the Fourier transform subblock and patterned silicon as the oracle subblock. First, we detail optimization of the GRIN lens through numerical analysis. Then, we employed inverse design through a machine learning approach to further optimize the structural geometry. Through this optimization, we enhance the interaction of the incident light with the metamaterial via spectral improvements of the outgoing wave.

Adaptive penalty method with adam optimizer for enhanced convergence in optical waveguide mode solvers

We propose a cutting-edge penalty method for optical waveguide mode solvers, integrating the Adam optimizer into pseudospectral frequency-domain (PSFD) frameworks. This strategy enables adaptable boundary fluctuations at material interfaces, significantly enhancing numerical convergence and stability. The Adam optimizer, an adaptive algorithm, is deployed to determine the penalty coefficient, greatly improving convergence rates and robustness while effectively incorporating boundary conditions into the interfaces of subdomains. Our solver evaluates the numerical performance of optical waveguides by calculating effective indices of standard benchmark waveguides with high accuracy. This method diminishes numerical boundary errors and provides a marked increase in convergence speed and superior accuracy when compared to conventional methods and even metaheuristic optimization methods, all while maintaining the inherent global spectral accuracy of the PSFD.

Correlating metasurface spectra with a generation-elimination framework

Inferring optical response from other correlated optical response is highly demanded for vast applications such as biological imaging, material analysis, and optical characterization. This is distinguished from widely-studied forward and inverse designs, as it is boiled down to another different category, namely, spectra-to-spectra design. Whereas forward and inverse designs have been substantially explored across various physical scenarios, the spectra-to-spectra design remains elusive and challenging as it involves intractable many-to-many correspondences. Here, we first dabble in this uncharted area and propose a generation-elimination framework that can self-orient to the best output candidate. Such a framework has a strong built-in stochastically sampling capability that automatically generate diverse nominations and eliminate inferior nominations. As an example, we study terahertz metasurfaces to correlate the reflection spectra from low to high frequencies, where the inaccessible spectra are precisely forecasted without consulting structural information, reaching an accuracy of 98.77%. Moreover, an innovative dimensionality reduction approach is executed to visualize the distribution of the abstract correlated spectra data encoded in latent spaces. These results provide explicable perspectives for deep learning to parse complex physical processes, rather than "brute-force" black box, and facilitate versatile applications involving cross-wavelength information correlation.

Artificial intelligence enhanced sensors - enabling technologies to next-generation healthcare and biomedical platform

The fourth industrial revolution has led to the development and application of health monitoring sensors that are characterized by digitalization and intelligence. These sensors have extensive applications in medical care, personal health management, elderly care, sports, and other fields, providing people with more convenient and real-time health services. However, these sensors face limitations such as noise and drift, difficulty in extracting useful information from large amounts of data, and lack of feedback or control signals. The development of artificial intelligence has provided powerful tools and algorithms for data processing and analysis, enabling intelligent health monitoring, and achieving high-precision predictions and decisions. By integrating the Internet of Things, artificial intelligence, and health monitoring sensors, it becomes possible to realize a closed-loop system with the functions of real-time monitoring, data collection, online analysis, diagnosis, and treatment recommendations. This review focuses on the development of healthcare artificial sensors enhanced by intelligent technologies from the aspects of materials, device structure, system integration, and application scenarios. Specifically, this review first introduces the great advances in wearable sensors for monitoring respiration rate, heart rate, pulse, sweat, and tears; implantable sensors for cardiovascular care, nerve signal acquisition, and neurotransmitter monitoring; soft wearable electronics for precise therapy. Then, the recent advances in volatile organic compound detection are highlighted. Next, the current developments of human-machine interfaces, AI-enhanced multimode sensors, and AI-enhanced self-sustainable systems are reviewed. Last, a perspective on future directions for further research development is also provided. In summary, the fusion of artificial intelligence and artificial sensors will provide more intelligent, convenient, and secure services for next-generation healthcare and biomedical applications.

A Comprehensive and Versatile Multimodal Deep-Learning Approach for Predicting Diverse Properties of Advanced Materials

A multimodal deep-learning (MDL) framework is presented for predicting physical properties of a ten-dimensional acrylic polymer composite material by merging physical attributes and chemical data. The MDL model comprises four modules, including three generative deep-learning models for material structure characterization and a fourth model for property prediction. The approach handles an 18-dimensional complexity, with ten compositional inputs and eight property outputs, successfully predicting 913 680 property data points across 114 210 composition conditions. This level of complexity is unprecedented in computational materials science, particularly for materials with undefined structures. A framework is proposed to analyze the high-dimensional information space for inverse material design, demonstrating flexibility and adaptability to various materials and scales, provided sufficient data are available. This study advances future research on different materials and the development of more sophisticated models, drawing the authors closer to the ultimate goal of predicting all properties of all materials.

Metasurface-Enhanced Infrared Spectroscopy: An Abundance of Materials and Functionalities

Infrared spectroscopy provides unique information on the composition and dynamics of biochemical systems by resolving the characteristic absorption fingerprints of their constituent molecules. Based on this inherent chemical specificity and the capability for label-free, noninvasive, and real-time detection, infrared spectroscopy approaches have unlocked a plethora of breakthrough applications for fields ranging from environmental monitoring and defense to chemical analysis and medical diagnostics. Nanophotonics has played a crucial role for pushing the sensitivity limits of traditional far-field spectroscopy by using resonant nanostructures to focus the incident light into nanoscale hot-spots of the electromagnetic field, greatly enhancing light-matter interaction. Metasurfaces composed of regular arrangements of such resonators further increase the design space for tailoring this nanoscale light control both spectrally and spatially, which has established them as an invaluable toolkit for surface-enhanced spectroscopy. Starting from the fundamental concepts of metasurface-enhanced infrared spectroscopy, a broad palette of resonator geometries, materials, and arrangements for realizing highly sensitive metadevices is showcased, with a special focus on emerging systems such as phononic and 2D van der Waals materials, and integration with waveguides for lab-on-a-chip devices. Furthermore, advanced sensor functionalities of metasurface-based infrared spectroscopy, including multiresonance, tunability, dielectrophoresis, live cell sensing, and machine-learning-aided analysis are highlighted.

Deep-Learning-Enabled Intelligent Design of Thermal Metamaterials

Thermal metamaterials are mixture-based materials that are engineered to manipulate, control, and process the flow of heat, enabling numerous advanced thermal metadevices. Conventional thermal metamaterials are predominantly designed with tractable regular geometries owing to the delicate analytical solution and easy-to-implement effective structures. Nevertheless, it is challenging to achieve the design of thermal metamaterials with arbitrary geometry, letting alone intelligent (automatic, real-time, and customizable) design of thermal metamaterials. Here, an intelligent design framework of thermal metamaterials is presented via a pre-trained deep learning model, which gracefully achieves the desired functional structures of thermal metamaterials with exceptional speed and efficiency, regardless of arbitrary geometry. It possesses incomparable versatility and is of great flexibility to achieve the corresponding design of thermal metamaterials with different background materials, anisotropic geometries, and thermal functionalities. The transformation thermotics-induced, freeform, background-independent, and omnidirectional thermal cloaks, whose structural configurations are automatically designed in real-time according to shape and background, are numerically and experimentally demonstrated. This study sets up a novel paradigm for an automatic and real-time design of thermal metamaterials in a new design scenario. More generally, it may open a door to the realization of an intelligent design of metamaterials in also other physical domains.

Deep learning-based inverse design optimization of efficient multilayer thermal emitters in the near-infrared broad spectrum

This study proposes a deep learning architecture for automatic modeling and optimization of multilayer thin film structures to address the need for specific spectral emitters and achieve rapid design of geometric parameters for an ideal spectral response. Multilayer film structures are ideal thermal emitter structures for thermophotovoltaic application systems because they combine the advantages of large area preparation and controllable costs. However, achieving good spectral response performance requires stacking more layers, which makes it more difficult to achieve fine spectral inverse design using forward calculation of the dimensional parameters of each layer of the structure. Deep learning is the main method for solving complex data-driven problems in artificial intelligence and provides an efficient solution for the inverse design of structural parameters for a target waveband. In this study, an eight-layer thin film structure composed of SiO₂/Ti and SiO₂/W is rapidly reverse engineered using a deep learning method to achieve a structural design with an emissivity better than 0.8 in the near-infrared band. Additionally, an eight-layer thin film structure composed of 3 × 3 cm SiO₂/Ti is experimentally measured using magnetron sputtering, and the emissivity in the 1-4 µm band was better than 0.68. This research provides implications for the design and application of micro-nano structures, can be widely used in the fields of thermal imaging and thermal regulation, and will contribute to developing a new paradigm for optical nanophotonic structures with a fast target-oriented inverse design of structural parameters, such as required spectral emissivity, phase, and polarization.

Systematic modification of functionality in disordered elastic networks through free energy surface tailoring

A combined machine learning-physics-based approach is explored for molecular and materials engineering. Specifically, collective variables, akin to those used in enhanced sampled simulations, are constructed using a machine learning model trained on data gathered from a single system. Through the constructed collective variables, it becomes possible to identify critical molecular interactions in the considered system, the modulation of which enables a systematic tailoring of the system's free energy landscape. To explore the efficacy of the proposed approach, we use it to engineer allosteric regulation and uniaxial strain fluctuations in a complex disordered elastic network. Its successful application in these two cases provides insights regarding how functionality is governed in systems characterized by extensive connectivity and points to its potential for design of complex molecular systems.

Artificial Intelligence and Advanced Materials

Artificial intelligence (AI) is gaining strength, and materials science can both contribute to and profit from it. In a simultaneous progress race, new materials, systems, and processes can be devised and optimized thanks to machine learning (ML) techniques, and such progress can be turned into innovative computing platforms. Future materials scientists will profit from understanding how ML can boost the conception of advanced materials. This review covers aspects of computation from the fundamentals to directions taken and repercussions produced by computation to account for the origins, procedures, and applications of AI. ML and its methods are reviewed to provide basic knowledge of its implementation and its potential. The materials and systems used to implement AI with electric charges are finding serious competition from other information-carrying and processing agents. The impact these techniques have on the inception of new advanced materials is so deep that a new paradigm is developing where implicit knowledge is being mined to conceive materials and systems for functions instead of finding applications to found materials. How far this trend can be carried is hard to fathom, as exemplified by the power to discover unheard of materials or physical laws buried in data.

Exploring the limits of metasurface polarization multiplexing capability based on deep learning

Metasurfaces provide a new approach for planar optics and thus have realized multifunctional meta-devices with different multiplexing strategies, among which polarization multiplexing has received much attention due to its convenience. At present, a variety of design methods of polarization multiplexed metasurfaces have been developed based on different meta-atoms. However, as the number of polarization states increases, the response space of meta-atoms becomes more and more complex, and it is difficult for these methods to explore the limit of polarization multiplexing. Deep learning is one of the important routes to solve this problem because it can realize the effective exploration of huge data space. In this work, a design scheme for polarization multiplexed metasurfaces based on deep learning is proposed. The scheme uses a conditional variational autoencoder as an inverse network to generate structural designs and combines a forward network that can predict meta-atoms' responses to improve the accuracy of designs. The cross-shaped structure is used to establish a complicated response space containing different polarization state combinations of incident and outgoing light. The multiplexing effects of the combinations with different numbers of polarization states are tested by utilizing the proposed scheme to design nanoprinting and holographic images. The polarization multiplexing capability limit of four channels (a nanoprinting image and three holographic images) is determined. The proposed scheme lays the foundation for exploring the limits of metasurface polarization multiplexing capability.

Deep learning-based inverse design of microstructured materials for optical optimization and thermal radiation control

Microstructures with engineered properties are critical to thermal management in aerospace and space applications. Due to the overwhelming number of microstructure design variables, traditional approaches to material optimization can have time-consuming processes and limited use cases. Here, we combine a surrogate optical neural network with an inverse neural network and dynamic post-processing to form an aggregated neural network inverse design process. Our surrogate network emulates finite-difference time-domain simulations (FDTD) by developing a relationship between the microstructure's geometry, wavelength, discrete material properties, and the output optical properties. The surrogate optical solver works in tandem with an inverse neural network to predict a microstructure's design properties that will match an input optical spectrum. As opposed to conventional approaches that are constrained by material selection, our network can identify new material properties that best optimize the input spectrum and match the output to an existing material. The output is evaluated using critical design constraints, simulated in FDTD, and used to retrain the surrogate-forming a self-learning loop. The presented framework is applicable to the inverse design of various optical microstructures, and the deep learning-derived approach will allow complex and user-constrained optimization for thermal radiation control in future aerospace and space systems.

Automatic optimization of miniaturized bound states in the continuum cavity

Bound states in the continuum (BICs) provide, what we believe to be, a novel and efficient way for light trapping. However, using BICs to confine the light into a three-dimensional compact volume remains a challenging task, since the energy leakage at the lateral boundaries dominates the cavity loss when its footprint shrinks to considerably small, and hence, sophisticated boundary designs turn out to be inevitable. Conventional design methods fail in solving the lateral boundary problem because a large number of degree-of-freedoms (DOFs) are involved. Here, we propose a fully automatic optimization method to promote the performance of lateral confinement for a miniaturized BIC cavity. Briefly, we combine a random parameter adjustment process with a convolutional neural network (CNN), to automatically predict the optimal boundary design in the parameter space that contains a number of DOFs. As a result, the quality factor that is accounted for lateral leakage increases from 4.32 × 10⁴ in the baseline design to 6.32 × 10⁵ in the optimized design. This work confirms the effectiveness of using CNNs for photonic optimization and will motivate the development of compact optical cavities for on-chip lasers, OLEDs, and sensor arrays.

A knowledge-inherited learning for intelligent metasurface design and assembly

Recent breakthroughs in deep learning have ushered in an essential tool for optics and photonics, recurring in various applications of material design, system optimization, and automation control. Deep learning-enabled on-demand metasurface design has been the subject of extensive expansion, as it can alleviate the time-consuming, low-efficiency, and experience-orientated shortcomings in conventional numerical simulations and physics-based methods. However, collecting samples and training neural networks are fundamentally confined to predefined individual metamaterials and tend to fail for large problem sizes. Inspired by object-oriented C++ programming, we propose a knowledge-inherited paradigm for multi-object and shape-unbound metasurface inverse design. Each inherited neural network carries knowledge from the "parent" metasurface and then is freely assembled to construct the "offspring" metasurface; such a process is as simple as building a container-type house. We benchmark the paradigm by the free design of aperiodic and periodic metasurfaces, with accuracies that reach 86.7%. Furthermore, we present an intelligent origami metasurface to facilitate compatible and lightweight satellite communication facilities. Our work opens up a new avenue for automatic metasurface design and leverages the assemblability to broaden the adaptability of intelligent metadevices.

Physics-model-based neural networks for inverse design of binary phase planar diffractive lenses

The inverse design approach has enabled the customized design of photonic devices with engineered functionalities through adopting various optimization algorithms. However, conventional optimization algorithms for inverse design encounter difficulties in multi-constrained problems due to the substantial time consumed in the random searching process. Here, we report an efficient inverse design method, based on physics-model-based neural networks (PMNNs) and Rayleigh-Sommerfeld diffraction theory, for engineering the focusing behavior of binary phase planar diffractive lenses (BPPDLs). We adopt the proposed PMNN to design BPPDLs with designable functionalities, including realizing a single focal spot, multiple foci, and an optical needle with size approaching the diffraction limit. We show that the time for designing single device is dramatically reduced to several minutes. This study provides an efficient inverse method for designing photonic devices with customized functionalities, overcoming the challenges based on traditional data-driven deep learning.

Deep Learning Assisted Optimization of Metasurface for Multi-Band Compatible Infrared Stealth and Radiative Thermal Management

Infrared (IR) stealth plays a vital role in the modern military field. With the continuous development of detection technology, multi-band (such as near-IR laser and middle-IR) compatible IR stealth is required. Combining rigorous coupled wave analysis (RCWA) with Deep Learning (DL), we design a Ge/Ag/Ge multilayer circular-hole metasurface capable of multi-band IR stealth. It achieves low average emissivity of 0.12 and 0.17 in the two atmospheric windows (3~5 μm and 8~14 μm), while it achieves a relatively high average emissivity of 0.61 between the two atmospheric windows (5~8 μm) for the purpose of radiative thermal management. Additionally, the metasurface has a narrow-band high absorptivity of 0.88 at the near-infrared wavelength (1.54 μm) for laser guidance. For the optimized structure, we also analyze the potential physical mechanisms. The structure we optimized is geometrically simple, which may find practical applications aided with advanced nano-fabrication techniques. Also, our work is instructive for the implementation of DL in the design and optimization of multifunctional IR stealth materials.

Inverse design of photonic crystal filters with arbitrary correlation and size for accurate spectrum reconstruction

Spectroscopic technique based on nanophotonic filters can recover spectral information through compressive sensing theory. The spectral information is encoded by nanophotonic response functions and decoded by computational algorithms. They are generally ultracompact, low in cost, and offer single-shot operation with spectral resolution better than 1 nm. Thus, they could be ideally suited for emerging wearable and portable sensing and imaging applications. Previous work has revealed that successful spectral reconstruction relies on well-designed filter response functions with sufficient randomness and low mutual correlation, but no thorough discussion has been performed on the filter array design. Here, instead of blind selection of filter structures, inverse design algorithms are proposed to obtain a photonic crystal filter array with predefined correlation coefficients and array size. Such rational spectrometer design can perform accurate reconstruction for a complex spectrum and maintain the performance under noise perturbation. We also discuss the impact of correlation coefficient and array size on the spectrum reconstruction accuracy. Our filter design method can be extended to different filter structures and suggests a better encoding component for reconstructive spectrometer applications.

Inverse design of polarization conversion metasurfaces by deep neural networks

To address the problem of multiple solutions and improve the calculating speed, we construct a tandem architecture consisting of a forward modeling network and an inverse design network. Using this combined network, we inversely design the circular polarization converter and analyze the effect of different design parameters on the prediction accuracy of the polarization conversion rate. The average mean square error of the circular polarization converter is 0.00121 at an average prediction time of 1.56×10^-2 s. If only the forward modeling process is considered, it takes 6.15×10^-4 s, which is 2.1×10⁵ times faster than that using the traditional numerical full-wave simulation method. By slightly resizing the network input and output layers, the network is adaptable to the design of both the linear cross-polarization and linear-to-circular polarization converters.

Machine learning assisted inverse design of microresonators

The high demand for fabricating microresonators with desired optical properties has led to various techniques to optimize geometries, mode structures, nonlinearities, and dispersion. Depending on applications, the dispersion in such resonators counters their optical nonlinearities and influences the intracavity optical dynamics. In this paper, we demonstrate the use of a machine learning (ML) algorithm as a tool to determine the geometry of microresonators from their dispersion profiles. The training dataset with ∼460 samples is generated by finite element simulations and the model is experimentally verified using integrated silicon nitride microresonators. Two ML algorithms are compared along with suitable hyperparameter tuning, out of which Random Forest yields the best results. The average error on the simulated data is well below 15%.