Deep-Learning-Empowered Holographic Metasurface with Simultaneously Customized Phase and Amplitude.

Metasurfaces with simultaneously and independently controllable amplitude and phase have provided a higher degree of freedom in manipulating electromagnetic (EM) waves. Compared with phase- or amplitude-only modulation, the capability of simultaneously controlling the phase and amplitude of EM waves can enable holography with a higher resolution. However, this drastically increases the design complexity of holographic metasurfaces, and the design process is usually quite time-consuming. In this paper, we propose an inverse design of meta-atoms that can simultaneously and independently tailor the phase and amplitude of transmitted waves using customized deep ResNet while eliminating the coupling of parameters. To demonstrate the design method, two holographic metasurfaces were designed using the trained network without the need for parameter sweeping, which will significantly enhance design efficiency. Prototypes were fabricated and measured. Both the simulated and measured results show that high-resolution holography is obtained, which sufficiently verifies the reliability of the design method. Our work paves the way for the intelligent design of metasurfaces and can also be applied to the design of other artificial materials or surfaces.

Genetic-algorithm-empowered metasurface design: simultaneous realization of high microwave frequency-selection and low infrared surface-emissivity.

With the improvement of equipment integration, it is difficult to meet the increasing functional requirements with the function of a single spectrum. In this work, a multispectral functional metasurface (MFM) is designed to achieve multispectral compatibility between microwave and infrared using multi-optimization. For microwaves, a frequency selective surface (FSS) is designed to achieve frequency selectivity. And for infrared, a twice genetic algorithm (GA) is employed to further increase the metallic filling ratio, thus reducing the infrared emissivity while maintaining the performance of microwave FSS. In order to verify our design and method, the MFM is fabricated and measured, and all the results are consistent with the theoretical design. The performance of FSS can achieve 3dB bandwidth in 7.2-11.2GHz with low insertion losses and stability, and meanwhile the mean infrared emissivity has been reduced to 0.24 in 3-14µm. In summary, the designed multispectral compatible metasurface has wide application value in radome. What's more, the multi-optimization method for designing the multispectral metasurface can also be extended to other fields.

Phase-to-pattern inverse design paradigm for fast realization of functional metasurfaces via transfer learning.

Metasurfaces have provided unprecedented freedom for manipulating electromagnetic waves. In metasurface design, massive meta-atoms have to be optimized to produce the desired phase profiles, which is time-consuming and sometimes prohibitive. In this paper, we propose a fast accurate inverse method of designing functional metasurfaces based on transfer learning, which can generate metasurface patterns monolithically from input phase profiles for specific functions. A transfer learning network based on GoogLeNet-Inception-V3 can predict the phases of 28×8 meta-atoms with an accuracy of around 90%. This method is validated via functional metasurface design using the trained network. Metasurface patterns are generated monolithically for achieving two typical functionals, 2D focusing and abnormal reflection. Both simulation and experiment verify the high design accuracy. This method provides an inverse design paradigm for fast functional metasurface design, and can be readily used to establish a meta-atom library with full phase span.

Overcome chromatism of metasurface via Greedy Algorithm empowered by self-organizing map neural network.

Chromatism generally exists in most metasurfaces. Because of this, the deflected angle of metasurface reflectors usually varies with frequency. This inevitably hinders wide applications of metasurfaces to broadband signal scenarios. Therefore, it is of great significance to overcome chromatism of metasurfaces. With this aim, we firstly analyze necessary conditions for achromatic metasurface deflectors (AMD) and deduce the ideal dispersions of meta-atoms. Then, we establish a Self-Organizing Map (SOM) Neural Network as a prepositive model to obtain a diversified searching map, which is then applied to Greedy Algorithm to search meta-atoms with the required dispersions. Using these meta-atoms, an AMD was designed and simulated, with a thickness about 1/15 the central wavelength. A prototype was fabricated and measured. Both the simulation and measurement show that the proposed AMD can achieve an almost constant deflected angle of 22° under normal incidence within 9.5-10.5GHz. This method may find wide applications in designing functional metasurfaces for satellite communications, mobile wireless communications and others.

Deep Learning: A Rapid and Efficient Route to Automatic Metasurface Design.

Metasurfaces provide unprecedented routes to manipulations on electromagnetic waves, which can realize many exotic functionalities. Despite the rapid development of metasurfaces in recent years, the design process of metasurface is still time-consuming and computational resource-consuming. Moreover, it is quite complicated for layman users to design metasurfaces as plenty of specialized knowledge is required. In this work, a metasurface design method named REACTIVE is proposed on the basis of deep learning, as deep learning method has shown its natural advantages and superiorities in mining undefined rules automatically in many fields. REACTIVE is capable of calculating metasurface structure directly through a given design target; meanwhile, it also shows the advantage in making the design process automatic, more efficient, less time-consuming, and less computational resource-consuming. Besides, it asks for less professional knowledge, so that engineers are required only to pay attention to the design target. Herein, a triple-band absorber is designed using the REACTIVE method, where a deep learning model computes the metasurface structure automatically through inputting the desired absorption rate. The whole design process is achieved 200 times faster than the conventional one, which convincingly demonstrates the superiority of this design method. REACTIVE is an effective design tool for designers, especially for laymen users and engineers.

Transparent and broadband absorption-diffusion-integrated low-scattering metamaterial by standing-up lattice.

In this paper, a transparent absorption-diffusion-integrated metamaterial (ADMM) based on standing-up lattice structure is proposed which takes full advantage of electromagnetic absorption and destructive interference simultaneously for the suppression of broadband backward scattering within a wide angular domain, especially for the lower-frequency scattering. The proposed ADMM is constituted by two kinds of rhombic and squared ITO lattices arranged in a pseudorandom distribution and then backed with ITO film. Calculation, simulation, and experimental measurement show that the proposed ADMM can achieve low scattering with normalized reflection less than 0.1 in the frequency band of 6.1-21.0GHz. In addition, owing to the standing-up lattice structure, the averaged optical transmittance of our ADMM reaches the optimal value of around 82.1% in the visible wavelength range of 380-780nm, promising an excellent optical transparency. The proposed comprehensive scheme provides an effective way to achieve broadband scattering suppression and high compatibility with optical transparency, enabling a wide range of applications in the window glass of stealth armament, electromagnetic compatibility facility and photovoltaic solar device.

Fast optimization method of designing a wideband metasurface without using the Pancharatnam-Berry phase.

Arbitrary control of electromagnetic waves remains a significant challenge although it promises many important applications. Here, we proposed a fast optimization method of designing a wideband metasurface without using the Pancharatnam-Berry (PB) phase, of which the elements are non-absorptive and capable of predicting the wideband and smooth phase-shift. In our design method, the metasurface is composed of low-Q-factor resonant elements without using the PB phase, and is optimized by the genetic algorithm and nonlinear fitting method, having the advantages that the far field scattering patterns can be quickly synthesized by the hybrid array patterns. To validate the design method, a wideband low radar cross section metasurface is demonstrated, showing good feasibility and performance of wideband RCS reduction. This work reveals an opportunity arising from a metasurface in effective manipulation of microwave and flexible fast optimal design method.