Metasurfaces and their applications

2.1 High impedance surface

Modern antenna systems require more compact geometries and lower profiles. However, this is often restricted by the presence of a conducting ground plane beneath the radiating elements. Specifically, a quarter wavelength distance is required between an antenna its ground plane because of the effect of image currents. If the antenna is too close to the ground plane, radiation from the image currents will destructively interfere with the direct radiation from the antenna, leading to an impedance mismatch.

A high impedance surface [43] is composed of periodic subwavelength metallic patches backed by a ground plane, with conductor via the center of each unit cell, as shown in Figure 1A. These planar high impedance metasurfaces are related to the corrugated surfaces [44], [45], which have been studied previously and which also provide ways to control scattering and surface wave propagation [46]. The corrugated surface is a metal slab with periodic vertical slots with subwavelength spacing. With a quarter of wavelength deep slots, the short circuit at the bottom end of the surface is transformed to be open circuit at the top end, yielding a high impedance. High impedance surface is similar in performance but with a subwavelength thickness. The behavior of high impedance surfaces can be understood as follows: each patch is coupled to its neighboring patches, providing an effective surface capacitance. They are also linked to their neighbors through a conductive path that includes the vias and the ground plane, which provides an effective surface inductance. The side view of a high impedance surface with a zoomed unit cell is shown in Figure 1B. The equivalent circuit model is thus a parallel LC circuit, where L and C are controlled by the geometry of the metallic patches and the thickness of the substrate [47]. The fringing field at the gap between neighboring patches contributes to the capacitance, while the current path from patch to the ground through vias contributes to the inductance. At the resonance frequency f0=12πLC, the parallel LC circuit is equivalent to an open circuit showing infinite impedance, so the reflection phase equals 0 at normal incidence indicated in Figure 1C, in contrast to an ordinary metallic reflector, which has a reflection phase of π. This property is equivalent to a magnetic conductor but with limited bandwidth and is often called an artificial magnetic conductor [48], because the tangential magnetic field at the surface is zero, rather than the electric field, as seen on an electric conductor. This unique property is useful for applications such as low-profile antennas [49], because the image current at the surface adds in phase with the current on the antenna, eliminating the need for a quarter-wavelength separation between antenna and the ground plane. In addition to providing an unusual reflection phase, high impedance surfaces can also be used to manipulate surface waves. These structures possess a band gap between the first band, which supports transverse magnetic waves, and the second band, which supports transverse electric waves, plotted and shown in Figure 1D. The TM band does not reach the TE band edge but stops slightly below it. For TE band, it slopes upward before crossing the light line, resulting in a band gap between TE and TM surface waves below the light line. In other words, no surface waves are supported within the band gap, and this interesting property can be used in applications such as de-coupling of nearby antennas.

Figure 1:

High impedance surfaces: properties and surface wave suppression.

(A) Top view of the high impedance surface. (B) Side view of high impedance surface array with zoomed view of resonating unit cell. (C) Reflection phase of the high impedance surface at normal incidence. (D) Dispersion curve of the high impedance surface, showing a band gap between TE and TM mode.

A variety of applications of high impedance surfaces have been investigated, such as low-profile antennas [50], absorbers [51], scattering reduction [52], field enhancement [53], antenna de-coupling [54], etc. Detailed examples of high impedance surface applications are introduced in the following paragraph.

Figure 2A [55] shows a low-profile antenna application, where the surface is composed of elements of a dog bone shape. The high impedance region is designed at the resonance frequency of the folded dipole. The surface behaves as a perfect magnetic conductor so that the image currents add in phase, resulting in a unidirectional radiation pattern shown in Figure 2B [55]. High impedance surfaces have advantages of low profile, light weight, and low cost, which make them extremely suitable to be coated on vehicles, buildings, etc. as microwave absorbers to protect sensitive antennas or electronics from undesired interferences. One example of a conductor-backed frequency-selective surface that provides a high impedance response being used as a microwave absorber is shown in Figure 2C [56]. An ultra-thin, high impedance surface-based microwave absorber using a genetic algorithm is designed, which is composed of frequency-selective surfaces backed with a ground plane with subwavelength thickness. Note that vias are not needed for normally incident waves because currents are not induced in the center via. Because high impedance surfaces do not support surface waves within the band gap region, they can be used in planar antenna decoupling applications, as demonstrated in Figure 2D [57], where two patch antennas are shown with and without the electromagnetic band gap structure patterned between them. Results shown in Figure 2E [57] illustrate the advantage of stopping surface wave propagation to lower coupling between the antennas. Other applications including a transverse electromagnetic (TEM) waveguide is depicted in Figure 2F [58], where the sidewalls of the waveguide are coated with a metasurface, which provides a magnetic conducting boundary, thus allowing the waveguide to support TEM modes over a limited frequency range.

Figure 2:

Applications of high impedance surfaces.

(A) Low profile folded dipole on high impedance. (B) Simulated and measured radiation pattern of dipole over high impedance surfaces. (C) PEC backed FSS screen as a thin absorber. (D) Mutual coupling reduction of patch antennas with high impedance surfaces. (E) Coupling between patch antennas is largely reduced due to surface wave suppression of high impedance surface. (F) An artificial TEM wave guide by using uniplanar band gap structure as sidewalls.

2.2 Active impedance surfaces

Despite the unique properties of high impedance surfaces, they have limitations including angular dependence and limited bandwidth [59]. However, some studies have investigated polarization and angle-insensitive surfaces [60], [61] by applying uniaxial magneto-dielectric medium with extreme parameters and series-resonance grids without vias in the slab, respectively. Other limitations of passive high impedance surfaces are linear response and fixed working frequency and usually hard to be addressed with only passive structures. However, it is possible to overcome these limitations by loading the surface with electronic devices including pin diodes, varactor diodes, and transistors [62]. A review of advances in nonlinear, active, and tunable metasurfaces in Ref. [63] demonstrated their exotic properties comparing with passive approaches.

Varactor diodes provide voltage-dependent capacitance, which can be applied at the gaps between unit cells to tune capacitance. A tunable impedance surface reflector is depicted in Figure 3A [64], in which the reflection phase of the surface is manipulated by tuning the bias voltages on the varactor diodes. An electronically steerable reflector is created by applying a reflection phase gradient, which can be programmed to control the angle of the reflected beam. At THz frequencies, graphene can be used for tunable and switchable impedance surface applications. Figure 3B [65] shows a reconfigurable antenna based on a graphene high impedance surface. The conductivity of the graphene is controlled by a bias voltage, thus influencing the radiation property of the antenna deployed on top of it, creating a reconfigurable antenna system. Furthermore, Figure 3C [66] shows a non-linear, self-tuning impedance surface created by attaching power-sensing circuits beneath the surface and feeding back the corresponding control voltages to the varactor diodes. The impedance of the surface changes with different levels of incident power, allowing for properties such as self-focusing of surface waves or power-dependent absorption.

Figure 3:

Active high impedance surfaces.

(A) An electronic steerable high impedance surface with varactor diodes between neighboring unit cells. (B) Beam reconfigurable THz antenna with a tunable high impedance surface grounding. (C) Power-dependent high impedance surface with pin didoes across outer ring to inner via. (D) Graphene non-foster circuit loaded high impedance surface with enhancement of bandwidth. (E) Photo of the wide band non-foster circuit (NFC) based artificial impedance surface supporting TM surface waves. (F) Comparison of impedance bandwidth of non-Foster loaded surface (green and blue curves) and passive approach (red curve).

Although, with varactor or pin diodes, the resonance frequency or impedance of the surface can be tuned, there are fundamental bandwidth limitations with such resonant metasurfaces. Impedance surfaces loaded with non-Foster circuits can exceed this limitation using negative capacitors and/or negative inductors. Figure 3D [67] shows the concept of loading non-Foster circuits on an artificial impedance surface, significantly expanding the bandwidth compared to a passive surface with the same substrate thickness. A low-profile and low-dispersion non-Foster loaded artificial impedance surface is shown in Figure 3E [68]. The active circuits enable the impedance surface to have constant impedance and low dispersion over nearly an octave of bandwidth with an electrically thin structure, indicated in Figure 3F. These bandwidth benefits come at the expense of low power handling and low operating frequency (usually below 1 GHz) of the non-Foster circuits and require careful design to avoid instability.

Metasurfaces loaded with active electronics can also be applied to other areas such as imaging [69], antennas [70], absorbers, etc. and will be introduced in subsequent sections.

3.1 Passive metasurface absorbers

Classic examples of microwave absorbers, including Dallenbach and Kleinsteuber [75] and Salisbury [76] absorbers, are designed with a thickness of one-quarter wavelength, so that the electric field of the incoming wave is maximum at the lossy layer. Metasurface-based absorbers can reduce the thickness of the absorber greatly by using resonant structures to increase the electric field at the surface. A review of thin absorbers for electromagnetic waves is shown in Ref. [77]. Metasurface absorbers are broadly studied for their low profile and simplicity of fabrication. High impedance surfaces have been investigated as microwave absorbers for radar applications [78], with a stable response to different incidence angles for TM-polarized waves [51], [79]. Near the resonance frequency, the interaction between electromagnetic waves and the surface is increased, enhancing the loss associated with resistive materials in the surface. A high impedance surface based microwave absorber is shown in Figure 4A [80]. Other metamaterial absorbers that also show high microwave absorption are presented in Figure 4B [13]. This example contains two metamaterial resonators coupled separately to the electric and magnetic fields to provide high absorption within a certain angle and frequency range. This near-unity high-absorbing rate is because, when ϵ and μ are manipulated independently, both incident electric field and magnetic field could be absorbed. Simulation results are shown in Figure 4C [13] with 100% absorption, while measurement results show a good correlation of 96%. Based on these early studies, there are numerous other works on metasurface absorbers with high performance over wide ranges of angles and polarizations.

Figure 4:

Passive metasurface absorbers from microwave to optical frequencies.

(A) Resistive sheet toped high impedance surface as a microwave absorber. (B) Unit cell of perfect metamaterial absorber. (C) Absorption simulation results of surfaces in (A). (D) Infrared perfect absorber. (E) Broadband solar absorber. (F) THz all-dielectric metasurface absorber.

Research in THz and optical frequency absorbers [81] draws a lot of interest in the community because of its broad applications in plasmonic sensors, infrared detectors, bolometers, and photovoltaic structures. Figure 4D [82] shows an array of gold material patches on MgF₂-substrate that are capable of absorbing energy independent of polarization and with a wide angle range of up to ±80°. The proposed structure is designed to be polarization independently in x and y directions at normal incidence, thus yielding a polarization-insensitive property. Moreover, as the direction of the magnetic field of the incident wave stays unchanged, the absorption rate peak is independent of the incident angle for TM polarization as high as 96% even at 80° incidence. Besides angle and polarization dependence, a broad band solar metasurface absorber is depicted in Figure 4E [83], where 90% of absorption is achieved in the visible and near-infrared range. Ohmic losses and thermal effects of metallic unit cells composing the metasurface absorbers can be eliminated with all-dielectric absorbers, as shown in Figure 4F [84], where electric and magnetic resonances overlap in the same frequency range, achieving over 97.5% absorption experimentally.

In general, metasurface-based absorbers have advantages of low profile and light weight with simple metallic structures. There are, however, limitations in terms of their absorption bandwidth, absorption rate, and air breakdown at high power [53]. Theoretical analysis of the limitations of TM surface wave attenuation by lossy slabs and impedance sheets is described in Ref. [85]. The fundamental limitations of the absorbing bandwidth of a given passive surface is analyzed, indicating that the minimum thickness of a 10-dB broad-band dielectric radar absorber is related to the largest operating wavelength [59]. A recent research on metasurface absorbers that targeted but is not bound to these limitations is focusing on active loading and tuning aimed at gaining more flexibility in response to incident waves, which will be introduced in the following subsection.

3.2 Active metasurface absorber

To overcome the intrinsic limitations of passive metasurfaces, active electronics play an important role in advanced metasurfaces by enabling switchable absorption, tunable resonant frequency, and tunable non-linear response. The following outlines current research on tunable metasurfaces for absorber applications.

Figure 5A [86] shows a metasurface that includes diodes within each unit cell that is tunable from total reflection to absorption, taking the advantage of turning the diode’s channel ON and OFF with different bias voltages. When the diodes have a bias voltage of zero, the junction appears as an open circuit and the elements of the metasurfaces are isolated. In this state, the electric and magnetic responses are balanced, causing it to work as a perfect absorber. When the diodes are biased at 0.75 V, the response of the surface changes dramatically by shorting adjacent unit cells together. In this state, the sheet is highly reflective. Because the elements of the metasurfaces are orthogonally arranged, it is possible to bias only a subset of the diodes to make the absorption and reflection properties polarization-dependent. The absorption rate and bandwidth of a metasurface can also be manipulated by tuning the substrate dielectric constant. For example, if the substrate contains liquid crystal, a bias voltage can be used to change the orientation of the molecules within the liquid crystal from random to aligned, thereby changing the dielectric constant. Figure 5B [87] shows the unit cell of a metasurface containing liquid crystals, allowing tuning of the absorption rate by up to 30% and the resonant absorption bandwidth by over 4%. A tunable metasurface absorber that can be used from tens of GHz to near infrared due to the broad band optical response of graphene is depicted in Figure 5C [88].

Figure 5:

Active and tunable metasurface absorbers.

(A) Switchable metasurface between total absorption to total reflection. (B) Liquid-crystal filled absorption tunable metasurface absorber at THz. (C) Optical tunable metasurface absorber. (D) Transistor-based high-power metasurface absorber. (E) Waveform-dependent microwave absorber. (F) Electrically tunable metasurface absorber.

For some applications, not only normal and angular incidences of spatial wave need to be absorbed, but also, surface currents on the metallic body of the system needs to be carefully absorbed, as a simple gap or any discontinuity of the surface may work as a slot antenna and allow surface waves penetrating and further radiating into the systems being enclosed, resulting in interference. An absorber that absorbs propagating surface waves is called a surface wave absorber. A transistor-based surface wave metasurface absorber is demonstrated in Figure 5D [38]. By manipulating the bias voltages to the gates of the FET transistors within each unit cell, the impedance of the drain-to-source channel is controlled, thereby changing the impedance of the metasurface. FETs have advantages compared to diodes because they do not draw current from the bias circuitry and they can provide higher on/off ratio. Figure 5E [89] shows a waveform-dependent metasurface absorber. Full-wave rectifiers are positioned at the gaps between patches, where electric fields are concentrated. Incoming microwave signals are rectified and transformed to static charges and stored in capacitors within the diode bridge circuit and then dissipated by resistors. This surface will preferentially absorb short pulses. By using inductors instead of capacitors, it is also possible to create a surface that absorbs only long pulses or continuous-wave signals. These materials represent the first waveform-dependent absorbers, which respond differently to various modulation waveforms even at the same frequency. Figure 5F [90] shows a metasurface that includes microwave power sensors at each unit cell. Feedback circuits change the surface impedance in response to the local microwave power level by tuning varactor diodes on the top surface. This nonlinear metasurface can enable self-focusing of surface waves, analogous to the self-focusing effect that occurs in some nonlinear optical materials [91]. It can also behave as a power-dependent surface wave absorber. A self-tuning metasurface working as a broadband surface wave absorber is introduced in Ref. [92], where frequency sensing circuit is designed and embedded on the surface, tuning the peak absorbing frequency to the incoming wave’s frequency by biasing corresponding DC voltages to the varactor diodes.

In addition to those discussed here, there are numerous other designs with active electronics loaded on traditional passive metasurface absorbers [93], [94], [95], [96], [97], [98], [99], [100], [101]. In conclusion, a tunable metasurface absorber is capable of dynamically modulating the surface impedance and material dispersion properties, which enable a metasurface absorber to provide different responses to incident waves by tuning the resonance frequency or absorption rate and to respond differently to the waveform or power level of the incoming wave.

4.1 Metasurface waveguides

This subsection discusses applications of impedance surfaces related to guiding surface waves, followed by recent pattern synthesis techniques to obtain the desired structures.

Impedance surfaces, in their simplest forms, are subwavelength patterned metal surfaces printed on a dielectric media. The small electrical thickness of the substrate allows one to model them as simple 2D surfaces, and their often periodic or pseudo-periodic nature ensures precise and rapid modelling of even large structures with ease by using periodic boundary conditions and eigenmode solvers. The size and shape of the periodic patterning determine the surface impedance and dispersion characteristics. If the wave-vector along the surface (k) is greater than its free space counterpart (k_o), the impedance surface will support the wave propagating along the surface and bound to it, creating a planar waveguide.

Simple waveguide structures consisting of a narrow strip of high impedance bordered by two low impedance sheets have been demonstrated [102]. Similar to guiding light along an optical fiber, the wave is confined to the waveguide by total internal reflection at the impedance discontinuity. Figure 6A shows a high impedance region bounded by two low impedance surfaces, along with the simulated wave propagating within the guide. In this case, the guide is made using an anisotropic impedance region, where the transverse impedance matches the surrounding surfaces. This creates a guide that is transparent to waves impinging from outside [30]. The patterns for these earlier structures are often simple square or rectangular metal patches, with the separation between adjacent unit cells providing the capacitance that controls the surface impedance.

Figure 6:

Guiding wavefronts through metasurfaces.

(A) Prototype and simulation results for a simple transparent waveguide. (B) Smooth and sharp bending of waves from line-wave structures. (C) Anisotropic metasurfaces to guide waves towards/away from edges. (D) H-shaped metasurfaces capable of coupling propagating modes to surface modes. (E) Huygens’ surface capable of reflecting two asymmetrical, co-polarized waves with gold nanorods. (F) Omega-type bianisotropic metasurface (O-BMS) unit cell with three distinct metasurface layers.

Experiments based on the high-low impedance principle were also performed to develop transparent surface waveguides that could bend an incoming wave. It was determined in Ref. [30] and similar works that below a certain radius of curvature, it was impossible to bend the propagating wave without suffering from significant power leakage. This occurs due to the difference in propagation speeds along the inner and outer bends to support a wave with a constant phase front. More recent research, however, has been able to show bending to arbitrary angles by controlling the impedance profile of the guide [103].

In relation to the concept of photonic topological insulators, it is possible to build unidirectional, polarization locked waveguides where light is guided along an arbitrary path at the interface between two complementary impedance surfaces, such as a capacitive surface in contact with an inductive surface. Referred to as “line waves”, these structures support a composite TE/TM mode at the interface. The mode has a singularity at the interface and is therefore highly confined. Its polarization locked propagation ensures reflection-free, quasi non-reciprocal behavior. By controlling the impedance on each side of the guide, it is possible to curve EM waves around bends smoothly. It is also possible to build a variety of conventional microwave and optical devices, such as the magic-T structure seen in Figure 6B.

Impedance sheets are capable not only of guiding surface modes but also of converting propagating waves into surface waves by eliminating the momentum mismatch discussed previously [104]. While similar coupling can be performed by using prisms and gratings, these structures are based on resonant couplings between the waves and therefore provide high conversion efficiency only for the angle of incidence with matching momenta. Elimination of mismatch has been achieved through the gradient of the reflection phase using “H-shaped” etchings on the substrate for various incidence angles (Figure 6D). Interestingly, non-homogeneous, anisotropic impedance surfaces have also been used as scattering controllers by redirecting surface waves towards or away from edges in a controlled fashion, as depicted in Figure 6C [105].

As geometry and patterning determine the characteristics of the impedance surfaces, much work has gone into analytically determining shapes and sizes of unit cells. The impedance sheet has been analyzed by some as a discontinuity in space, and a closely formed solution characterizing the electric and magnetic susceptibilities has been put forward [106]. It characterizes any given impedance surface based on its susceptibility tensor where transmission, reflection, and incident angles can be arbitrarily specified. Anisotropy can also be designed with highly elongated cells with arbitrary orientations, which provide smooth transitions between regions of varying impedance. As opposed to abrupt transitions or impedance discontinuities, which help confine the wave and can be used to build surface waveguides, certain applications require the wave to propagate over smoothly varying impedances allowing it to “bend”. A method proposed in Ref. [107] aims to solve this problem by means of a mathematical starting function defining the desired anisotropy. An equi-spaced point grid is created and shifted based on the gradient of the starting function. A Voronoi function is then applied to this separated point grid and provides the base for the anisotropic surface. Providing extra degrees of freedom through variation in cell shape, size, and orientation, the authors show better results through this technique than those based on conventional mathematical techniques such as conformal mapping, mesh generation, or point density. The concept is illustrated and discussed in depth in a subsequent section on imaging.

4.2 Huygens’ surfaces and beam shaping

One of the most fundamental tenets of electromagnetics, Huygens’ principle, states that every point on a wavefront acts as a secondary source by creating its own wavefronts. By using Schelkunoff’s equivalence principle, it is possible to develop ultra-thin surfaces that generate arbitrary field patterns for any given incident illumination, requiring it to be both electrically and magnetically tunable. Such surfaces are called Huygens’ surfaces and have special properties allowing control over reflection, beam steering, and polarization manipulation [33]. A schematic representation of such a surface can be seen in Figure 6E.

Huygens’ surfaces are being used to obtain high aperture illumination cavity backed antennas to replace Fabry-Perot leaky wave antennas. By eliminating the need to couple the partially reflecting surface modes to that of the cavity, Huygens’ cavity backed metasurface antennas provide additional degrees of freedom for cavity design. Designers are free to choose cavities with any specifications of aperture size or transmission angle and then use the equivalence principle to generate equivalent electrical surface impedance or magnetic surface admittance required to support the cavity modes [108]. Also used at optical frequencies, Huygens’s surfaces have shown capabilities of independently tuning both the magnitude and the phase of the reflection coefficient [109] of co-polarized, asymmetrical reflected waves. Using “gap-surface plasmon resonators” as unit cells, they are designed by tunable arrays of gold nanorods on silica spacers and display worst case power efficiency of 45% at 800 nm while providing a 360° arbitrary phase shift and 0–0.67 magnitude variation. Special Huygens’ surfaces that show not only electrical and magnetic polarizability but also magneto-electric coupling are commonly referred to as omega-type bianisotropic metasurfaces and can be used in beam splitting applications. Figure 6F shows such a surface consisting of three impedance sheets stacked together to achieve better impedance match from both sides, ensuring very low reflections for a wide range of incident angles using simple passive structures [110], [111].

5.1 Uniform metasurface antenna

Uniform metasurfaces have a homogeneous propagation constant over the entire surface area. Visually, the surface is composed of periodic, constant-dimension unit cells in a 2D plane. High impedance surfaces are widely adopted for these types of leaky wave metasurfaces due to their simplicity and low-profile properties. Figure 7A [113] indicates a 2D composite right-/left-handed (CRLH) structure with a coaxial probe feed at the center of the surface. When excited at the center, a circular surface wave radiates a conical beam in a leaky-wave mode when the phase velocity is greater than the speed of light. Like traditional leaky-wave antennas, when the frequency changes, the angle of the conical beam varies according to the dispersion relation β, which leads to beam scanning. Applying varactor diodes on the surface allows it to be electronically tuned, thereby controlling both the reflection phase and surface wave propagation properties. By varying the pattern of bias voltages, it is possible to achieve beam steering from leaky waves in both the forward and backward directions. By simply controlling the bias voltage, the amplitude and phase can be tuned with multiple degrees of freedom per half wavelength, as shown in Figure 7B [114]. The tunable band diagram of the same is plotted in Figure 7C [114]. When alternating rows of varactors are biased at two different voltages, the TE band is folded into the reduced Brillouin zone, with the upper region supporting backward TE modes. This tunability allows a wide-angle beam steering at a fixed frequency, which could be applied to low-cost, low-profile antenna systems. Figure 7D [113] demonstrates a one-dimensional (1D) CRLH structure that can also provide full-angle scanning by electrically controlling inductors and capacitors at a single frequency point. For THz metasurfaces performing as leaky-wave radiators, most research is carried out only in simulation. Figure 7E [115] depicts a sinusoidally modulated graphene leaky-wave antenna with electronic beam scanning. By applying bias voltages to the different grating pads beneath the graphene substrate, the graphene surface reactance can be modulated, resulting in a versatile beam-scanning capability at THz frequencies. Another 2D leaky-wave antenna operating at THz frequencies with tunable frequency and beam angle is shown in Figure 7F [116]. It is based on tuning the conductivity of the graphene, and it can be analyzed using a simple transmission line model.

Figure 7:

Passive and tunable uniform metasurface antenna.

(A) For 2D CRLH structure radiation a conical-beam, three topologies are shown: (1) a patch mushroom structure, (2) an inter-digital mushroom structure, (3) a stepped-impedance structure. (B) Two-dimensional varactor diodes loaded textured metasurface as a tunable beam steering metasurface. (C) Dispersion diagram of (B) with different biasing to the varactor diodes, fulfilling backward/forward TE modes. (D) A 1D backfire-to-end-fire fan-beam CRLH leaky-wave antenna. (E) A THz grapheme-based leaky-wave metasurface with electronic beam scanning capability at a fixed frequency. (F) A THz reconfigurable graphene based HIS leaky-wave antenna.

5.2 Holographic leaky-wave metasurfaces

Holographic antennas borrow their design concept from optical holography [117], which is used for recording and recreating a complex optical wavefront. In the microwave region, this technique has been applied for planar antenna designs. The principle of the holographic leaky-wave metasurface can be described in two steps. First, the interference pattern of the reference wave (exciting source) and the object wave (desired radiation pattern) is calculated, represented as follows:

where ψ_ref is the reference wave generated by the feed and ψ_obj is the object wave corresponding to the desired radiation pattern. The impedance of the surface is then designed according to this interference pattern. The second step is using the same ψ_ref to interfere with |ψ²|, so that the desired object wave is recreated, as |ψ_ref|²ψ_obj. The interference pattern is represented by varying the effective surface impedance, shown in Figure 8A [37]. A surface covered with an array of subwavelength metallic patches on a grounded dielectric substrate provides variable surface impedance controlled by the geometry of the patches. A closer view of the pattern is shown in Figure 8B [37]. Small gaps provide higher impedance, while large gaps correspond to lower impedance. Furthermore, by varying the shape of the cells, anisotropic impedance values are achieved, enabling polarization control. For example, a surface that can generate circular polarization from a linearly polarized excitation source is shown in Figure 8C [37]. Each patch is divided into two regions by a slit with a variable direction, and the anisotropy of the patches creates a tensor impedance surface, which enables coupling between different polarizations.

Figure 8:

Holographic leaky-wave metasurfaces.

(A) Demonstration of leaky-wave being excited with a single feed. (B) A partial close view of the isotropic holographic metasurface. (C) Anisotropic holographic metasurface for circular polarization applications. (D) Dual-beam holographic metasurface. (E) Isotropic holographic metasurface with a single feed in the center. (F) Anisotropic holographic metasurface with a coplanar feeder with circular polarization radiation for satellite communications. (G) Synthesis of holographic metasurface design in controlling magnitude, phase and polarizations. (H) One-dimensional sinusoidally modulated reactance surface. (I) Top and bottom views of a polarization insensitive holographic surface with complementary unit cells on either side.

A multifunctional surface supporting dual beam and dual polarization radiation is achieved with two orthogonal excitations, as shown in Figure 8D [118]. Two different functions are written on the same holographic interference pattern without mutual coupling, and frequency-dependent beam scanning with orthogonal polarizations is achieved. With simple isotropic patches, the impedance can be modulated by a sinusoidal spiral function and fed with single-point feed at the center, allowing a TM surface wave to be transformed to a circularly polarized leaky wave with broadside radiation, shown in Figure 8E [119]. An anisotropic holographic metasurface with a coplanar feed and circularly polarized radiation is shown in Figure 8F [120], which can be applied in space to ground satellite communications. A synthesis method is presented in Ref. [121] that gives control to the amplitude, phase, and polarization of the aperture field by adjusting the boundary conditions imposed by the metasurface. Some examples of the patterns with single or four probe feeding are given in Figure 8G [121]. A simple 1D sinusoidally modulated reactance surface is demonstrated in Figure 8H [122], which allows for nearly independent control of the leakage and phase constants along the surface with arbitrary off-broadside radiation. A method of mapping the gaps between metallic strips is demonstrated as well.

An interesting polarization-insensitive holographic high impedance surface is demonstrated in Ref. [123]. Designed with the potential to focus incoming radiation to the center of the antenna, regardless of polarization, it can serve as an important step towards metasurface-based energy scavenging. Like the line wave structure discussed previously, it is based on the overlapping dispersion curves of complementary TE/TM surfaces around the frequency of operation. Two complementary loop-wire structures are printed on either side of the substrate with matching refractive indices, enabling the wave to follow the same surface path for both LHCP and RHCP. Figure 8I shows a close-up schematic of the loop-wire unit-cell based metasurface.

In conclusion, metasurface-based leaky wave antennas can provide advantages in point-to-point communications due to their planar structure, ease of fabrication, simple feeding, as well as frequency and beam steering control.