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.

Conventional metasurface design methods usually require a lot of computational resources and time, meaning they fail to satisfy the efficient, rapid design on demand. On account of this, we branch out of the conventional metasurface design methods by attempting to relate the emerging discipline of artificial intelligence to a traditional physical area. With our method, named AMID, metasurface structures are designed inversely where they can be computed directly by simply proposing and inputting the desired design targets into AMID. AMID greatly simplifies conventional methods that call for not only sufficient professional knowledge but also trial and error through simulation softwares. According to the design results, unit cells of metasurfaces are successfully computed, which verifies the availability of AMID and improves the design efficiency in the meanwhile.

In this paper, we propose a method of designing programmable coding metasurface with a reconfigurable number of reflected beams. Each unit cell of the metasurface is loaded with one varactor to enable adjustable phase responses. The bias, under beneath the ground drilled with small holes, is connected to unit cells by thin via through the holes, to avoid adverse influences on unit cells. To simplify the bias network, a super unit composed of 5×5 unit cells is taken as the coding element of realizing 2D coding sequences. With different bias voltages, the unit cell will produce different reflection phases. By programming the bias voltages, the number and directions of reflected beams can be tailored due to different coding sequences. Under normal incident, reflected waves will be split into multiple beams whose directions can be flexibly controlled. We designed a prototype, implemented our design and carried out experiments. Both simulation and experiment results show that the number of reflected beams can be reconfigured between 1 and 5. Reflect beam number reconfigurability of such metasurfaces can be readily used as planar reflector of antennas and may find applications in wireless communications.

The modern design of advanced functional materials or surfaces has increasingly undergone miniaturization and integration, so as to implement multiple functions using the same aperture. Metasurfaces, as an emerging kind of artificial functional surface, have provided unprecedented freedom in manipulating electomagnetic (EM) waves upon two-dimensional surfaces. It is desirable that the aperture of metasurfaces can be multiplexed with a number of functions for EM waves. Based on previous research, an artificial neural network-forward design can achieve an inverse design. In this paper, we propose an inverse design method of multiplexing metasurface apertures. A deep learning network (DLN) is trained to serve as a forward model for a genetic algorithm (GA), that is, a deep-leaning-forward genetic algorithm (DLF-GA). With the DLF-GA model, the phase of meta-atoms can be predicted for orthogonally linearly polarized waves simultaneously. The DLN was trained by a data set consisting of 70 000 samples with an accuracy of 95% on the test data. To demonstrate the competency of this method, we demonstrate the design of a multiplexed metasurface that can achieve focusing and diffuse scattering for polarization. The metasurface, which consists of 24 × 24 meta-atoms, can be generated monolithically with the input phase profile. A prototype was fabricated and measured. The predicted, simulated, and measured results are well consistent and underscore the validity of this inverse design method. This method provides an efficient and accurate method for the fast design of multiplexed metasurfaces and will find applications in microwave engineering such as in satellite communications.

This study proposes a novel controllable reflection-enhancement metasurface (CREM) design method to amplify electromagnetic (EM) signals in a reflective manner. In contrast to previous passive structures, a field-effect transistor (FET) amplification circuit is integrated into the unit cell of a reflective metasurface. The FET circuit amplification shows that the metasurface can magnify the reflected wave intensity. A CREM composed of FET-connected dual-patch unit cells is designed for the specific implementation of the proposed method. The simulation and experiment results indicate that an apparent gain can be achieved in the operating band. Moreover, the reflection coefficient can be edited as required by the amplification level of the FET circuit. The tunable range of the reflection amplitude is approximately 10 dB. Different from the traditional signal amplifier, signal amplification is achieved herein in a reflective manner, which simplifies the amplification process and provides a new method of controlling EM waves. Our work provides an effective route to reflection enhancement and enables the design of devices and systems for real-world applications on demand.

In this article, we propose a design method of aperture-multiplexing metasurfaces using back-propagation neural network (BPNN), which can achieve independent wavefront modulation for orthogonally-polarized waves. To this end, we first propose a modified Jerusalem Cross (MJC) structure as the metasurface unit cell, which decouples orthogonal interactions by increasing the effective inductances of each of the two Jerusalem Cross (JC) branches. Due to the reduced orthogonal couplings, the MJC can achieve nearly independent control of orthogonally-polarized waves. Then, via BPNN, a dictionary mapping between reflection phase and structural parameters of MJC is established to facilitate metasurface design using MJCs. It is verified that the fitting degree of dataset exceeds 99.99% and that the prediction error of reflection phase is less than 0.01°. Using this design method, we demonstrated a focusing metasurface that can focus reflected waves for both $x$ - and $y$ -polarized waves simultaneously, with focus lengths of 150 and 300 mm, respectively. The simulated and measured results are well consistent, which prove the feasibility of this method. This work provides an efficient method of designing multiplexing and multifunctional metasurfaces, which may find applications in fields, such as satellite communication and base stations.